Tomentodione M sensitizes multidrug resistant cancer cells by decreasing P-glycoprotein via inhibition of p38 MAPK signaling

In this study, we investigated the mechanism by which tomentodione M (TTM), a novel natural syncarpic acid-conjugated monoterpene, reversed multi-drug resistance (MDR) in cancer cells. TTM increased the cytotoxicity of chemotherapeutic drugs such as docetaxel and doxorubicin in MCF-7/MDR and K562/MDR cells in a dose- and time-dependent manner. TTM reduced colony formation and enhanced apoptosis in docetaxel-treated MCF-7/MDR and K562/MDR cells, and it enhanced intracellular accumulation of doxorubicin and rhodamine 123 in MDR cancer cells by reducing drug efflux mediated by P-gp. TTM decreased expression of both P-gp mRNA and protein by inhibiting p38 MAPK signaling. Similarly, the p38 MAPK inhibitor SB203580 reversed MDR in cancer cells by decreasing P-gp expression. Conversely, p38 MAPK-overexpressing MCF-7 and K562 cells showed higher P-gp expression than controls. These observations indicate that TTM reverses MDR in cancer cells by decreasing P-gp expression via p38 MAPK inhibition.


INTRODUCTION
Multidrug resistance (MDR) is the phenomenon by which neoplastic cells display resistance to many chemotherapeutic drugs that are chemically dissimilar with different cytotoxic targets [1]. MDR negates the efficacy of cancer chemotherapy and results in therapeutic failure, which is detrimental for patient survival. Hence, MDR is the biggest challenge for strong and effective cancer treatment.
Currently, numerous chemotherapeutics are used in the clinic for cancer treatment such as doxorubicin (DOX), Daunorubicin (DNR), Epirubicin (EPI), Docetaxel (Doc) and Cisplatin (DDP) [2]. However, intrinsic or acquired drug resistance such as MDR leads to failure of chemotherapy. In most cases, chemotherapy becomes ineffective over time because of resistance. Moreover, the acquired resistance of the neoplastic cells is not limited to the original drug, but, also affects a wide variety of structurally and mechanistically unrelated drugs. Therefore, new therapeutic strategies are needed to overcome chemotherapy resistance in MDR cancer cells.
The mechanisms underlying MDR in cancer cells are complicated and not completely understood. Some of the MDR-related mechanisms include decreased drug accumulation, alterations in cell cycle checkpoints, resistance to cellular apoptotic mechanisms and repair of damaged cellular targets [3]. The most common and well studied mechanism of MDR in cancer cells is over-expression of ATP-binding cassette (ABC) transporters, whose regulation is highly complex and results in decreased drug accumulation [4,5]. The most extensively studied MDR transporters include ABCB1 (also known as MDR1 or P-gp), ABCC1 (also known as MRP1) and ABCG2 (also known as BCRP or MXR). ABCA1 was originally found as the primary gatekeeper www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 60), pp: 101965-101983 Research Paper for eliminating cholesterol that played an important role in protecting against cardiovascular disease [6]. It has also been implicated in multi-drug resistance apart from its role in reverse cholesterol metabolism [7,8]. Moreover, many groups have reported P-gp as a target for antitumor therapy in drug-resistant tumors [9][10][11].
To date, P-gp is the best studied among the ABCtransporter-mediated MDR proteins [12][13][14]. The hallmark of chemo-resistance in cancer cells is P-gp overexpression. High P-gp levels are associated with poor survival rate in osteosarcoma [15], lymphomas [16], small-cell lung cancer [17], breast cancer [18] and leukemias [19]. Therefore, P-gp inhibitors are necessary to treat cancer patients that demonstrate MDR. Many P-gp inhibitors such as Valspoder (PSC-833), dofequidar fumarate (MS-209), thiosemicarbazone derivative (NSC73306) Zosuduidar (LY335979) and Tariquidar (XR9567) have been found to antagonize P-gp function [20][21][22][23]. However, phase III trials of these drugs have been disappointing, primarily because of low host tolerance to these experimental MDR modulators that precludes the attainment of active intracellular levels [24,25]. Moreover, these drugs demonstrate adverse effects in patients and alter the pharmacokinetics of co-administered anticancer drug [26]. Also, none of these inhibitors have achieved significant survival benefits [27]. Therefore, there is an urgent need for novel MDR antagonists that demonstrate low toxicity and high efficacy.
Traditional Chinese Medicine (TCM) has offered many alternative natural products that demonstrate great potential as MDR inhibitors with low adverse effects. These natural products decrease P-gp expression [21,28,29]. Recent studies have shown that natural terpenoids reverse MDR in cancer cells [30,31]. However, their mechanism of action is unknown. Therefore, in this study, we investigated the mechanism of action of a novel meroterpenoid, tomentodione M (TTM) ( Figure 1A) in reversing MDR in MCF-7 and K562 cancer cells.

MDR cancer cells overexpress P-gp
We first evaluated the cytotoxicity of several anticancer agents such as DOX, DNR, EPI, DDP and Doc in MDR and parental MCF-7 and K562 cell lines. As shown in Table 1 (Table 1).
Since overexpression of ABC transporters is the primary determinant of the MDR phenotype, we examined the expression of three main ABC transporters, MDR1, MRP1 and BCRP by qRT-PCR. As shown in Figure 1B, MDR1 mRNA was highly expressed in MCF-7/MDR and K562/MDR cells than the corresponding parental cells. However, MRP1 or BCRP mRNA levels were similar in both MDR and parental MCF-7 and K562 cells ( Figure 1B).
The chemotherapy agents including DOX, DNR, EPI and Doc are exported out of the cells by P-gp [32][33][34]. Therefore, we examined the levels of P-gp protein in MDR and parental MCF-7 and K562 cells. Our data showed high P-gp protein levels in MCF-7/MDR and K562/MDR cells than the corresponding parental cells ( Figure 1C) and was consistent with the previous study [35,36]. Therefore, we selected MCF-7/MDR and K562/ MDR cells to investigate if TTM reversed P-gp-mediated MDR.

TTM is not cytotoxic to MDR cancer cells and non-tumor cells
We determined cytotoxicity of TTM on MDR and parental MCF and K562 cells by MTT and Cell Counting Kit-8 (CCK-8) assays. We observed that TTM concentrations below 100 μM did not significantly inhibit proliferation of MDR and parental MCF and K562 cells ( Figure 1D and 1E). Moreover, treatment with 0-100 μM TTM did not affect cell survival in MDR and parental MCF-7 and K562 cells between 0-48 hrr ( Figure 1D and 1E). We also showed that 0-100 μM TTM showed no cytotoxicity on non-cancer human mammary epithelial MCF-10A cells at 48 hr ( Figure 1F). Therefore, we selected 10 μM and 30 μM TTM for further experiments since they showed minimal growth inhibition of <5% and <10%, respectively.

TTM reverses P-gp-mediated MDR in vitro
We investigated if TTM treatment enhanced chemosensitivity of MDR cells to anticancer drugs such as DOX, DNR, EPI, Doc and DDP. We used Verapamil, a P-gp inhibitor and calcium channel blocker as a positive control [37][38][39]. TTM decreased the drug resistance in MDR cells in a dose-dependent manner. As shown in Table 1

TTM increases apoptosis in Doc-treated MDR cancer cells
We evaluated the effects of TTM on Doc-induced apoptosis. As shown in Figure 2A-2B, MDR cells treated with (+) or without (-) 30 μM TTM or 10 μM Ver combined with 1 μM Doc for 48 hr. The results indicated that treatment of the MDR cells with 1 μM Doc, especially with 30 μM TTM or 10 μM Ver alone, did not cause notable apoptosis, while combination treatment dramatically increased Doc-induced apoptosis (30.4% ± 3.68% and 37.6% ± 1.97% for Ver and TTM combined with Doc respectively, versus 12.65% ± 3.74% for Doc alone in MCF-7/MDR cells; 27.65% ± 2.33% and 30.15% ± 2.62% for Ver and TTM combined with Doc respectively, versus 7.8% ± 1.84% for Doc alone in K562/MDR cells). In the parental MCF-7 and K562 cells, Doc treatment alone induced significant cell death ( Supplementary Figures 1 and 2). Moreover, TTM increased Doc-induced apoptosis in K562 parental cells, but had no effect on MCF-7 cells.
Caspases are important mediators of cellular apoptosis that cleave various proteins that are necessary for cell survival and function [40]. We assessed the status of cleaved caspases 3 and 9 as well as PARP in MDR cells treated with TTM and Doc in combination. As shown in Figure 2C, MCF-7/MDR and K562/MDR cells treated with a combination of Doc and TTM showed increased cleaved caspases-3, -9 and PARP levels. This suggested that TTM treatment enhances Doc-induced apoptosis in MDR cells by activating the apoptotic machinery.

TTM decreases clonogenicity of Doc or DOXtreated MDR cancer cells
MDR cancer cells with high P-gp expression are highly malignant and resistant to traditional chemotherapeutic agents [41,42]. Therefore, we tested the effects of TTM on clonogenicity of parental and MDR MCF-7 cells treated with 1 μM Doc or DOX with or without 30 μM TTM. MCF-7/MDR cells formed greater number of colonies than the parental MCF-7 cells ( Figure  2D  Since P-gp plays an important role in DOX efflux, we examined the effects of TTM by assessing the accumulation and efflux of Rh123, which is a specific P-gp substrate [43]. We found that treatment with 10 μM Ver and 10, 30 or 50 μM TTM increased Rh123 accumulation by 1.

TTM decreases P-gp levels in MDR cells
Next, we analyzed if TTM decreased P-gp expression in MDR cells. The P-gp protein levels decreased in a dose-dependent manner in MCF-7/ MDR and K562/MDR cells when treated with 0-30 μM TTM for 48 hr ( Figure 4A-4B). Moreover, P-gp levels decreased in a time-dependent manner when treated 30 μM TTM for 48 hr ( Figure 4C). We transfected MCF-7 cells with P-gp overexpression plasmid to determine if TTM effects on MDR were related to P-gp protein levels. We observed that 30 μM TTM decreased P-gp levels in MDR1-overexpressing MCF-7 cells ( Figure 4D). We also www.impactjournals.com/oncotarget As shown, MDR cells were resistant to Doc, and both TTM and Ver enhanced Doc-induced apoptosis. ** denotes P < 0.01 and *** denotes P < 0.001 compared to Doc alone treatment. (C) Quantitative western blot analysis of procaspase-9, cleaved caspase-9, procaspase-3, cleaved caspase-3, PARP and cleaved PARP in MCF-7/MDR and K562/MDR cells incubated with 1 μM Doc with or without 30 μM TTM for 48 hr. Note: GAPDH was used as internal control. ## denotes P < 0.01 and ### denotes P < 0.001 compared to control. (D) Histogram plots showing total number of colonies in MCF-7 and MCF-7/MDR cells treated with 1 μM Doc with or without 30 μM TTM for 48 hr. Note: Cells were plated in 6-well plates containing RPMI 1640 plus 10% FBS at a density of 1000 cells per well. Colonies were counted after 15 days. The data are presented as the mean ± SEM from three independent experiments. ### denotes P < 0.001 compared to control. *** denotes P < 0.001 compared to Doc alone treatment. www.impactjournals.com/oncotarget

Figure 3: Effect of TTM on the accumulation and efflux of DOX and Rh123 in MDR cancer cells. (A) Histogram plots
showing flow cytometry analysis of intracellular DOX accumulation in MCF-7/MDR and K562/MDR cells pre-incubated for 4 hr with or without 10 μM TTM or 10 μM Ver (positive control) followed by incubation with 10 μM DOX as determined by flow cytometry. Also shown are histogram plots for DOX retention assay, wherein MDR cells were pre-incubated for 3 hr with 10 μM DOX followed by incubation for 4 hr in medium with or without 10 μM TTM or 10 μM Ver. Subsequently, intracellular DOX levels were determined by flow cytometry. (B) Histogram plots showing flow cytometry analysis of Rh123 dye accumulation in MDR and parental MCF-7 and K562 cells pretreated with or without 10, 30, 50 μM TTM or 10 μM Ver for 1.5 hr followed by incubation with 5 μM Rh123 in the dark for 1.5 hr. (C) Histogram plots showing flow cytometry analysis of Rh123 efflux in MDR and parental MCF-7 and K562 cells that were first incubated with 5 μM Rh123 in the dark for 1.5 hr followed by incubation with or without 10, 30, 50 μM TTM or 10 μM Ver for 1.5 hr. Intracellular fluorescence of Rh123 was detected by flow cytometry. Note: The data are represented as the mean ± SEM from three independent experiments. * denotes P < 0.05, ** denotes P < 0.01 and *** denotes P < 0.001 compared to DOX or Rh123 alone treatments. www.impactjournals.com/oncotarget showed by laser scanning confocal microscopy that P-gp levels were higher in MCF-7/MDR cells than MCF-7 cells; TTM reduced P-gp levels in MCF-7/MDR cells in a concentration dependent manner ( Figure 4E). These data demonstrated that TTM decreased MDR by lowering P-gp expression.

TTM inhibits p38 MAPK signaling
P-gp expression is regulated by MAPK signaling pathway and therefore inhibition of MAPK signaling is an effective therapeutic approach to overcome MDR [44]. Therefore, we analyzed the phosphorylation levels of JNK, ERK and p38 MAPKs in TTM-treated MDR cells. First, we observed increased phospho-p38 and phospho-JNK levels in MCF-7/MDR and K562/MDR cells than the parental MCF-7 and K562 cells ( Figure 5A). Moreover, TTM treatment decreased phospho-p38 and phospho-JNK levels, whereas phospho-ERK levels remained unchanged in MDR cells ( Figure 5A).
Next, we tested the effects of 10 μM SB203580 in MDR cells. SB203580 reduced phospho-p38 MAPK levels in both MCF-7/MDR and K562/MDR cells ( Figure  5C). We observed 1.38-and 1.67-fold reduced P-gp expression in SB203580 treated MCF-7/MDR and K562/ MDR cells than the corresponding untreated cells ( Figure  5C). This demonstrated that the p38 MAPK pathway regulated P-gp expression.
We then investigated if SB203580 treatment enhanced apoptosis in Doc treated MCF-7/MDR and K562/MDR cells by FACS analysis of AnnexinV-FITC/ PI staining. SB203580 treatment alone did not induce apoptosis, but, it reduced phospho-p38 levels in MDR cells ( Figure 5D and Figure 5C). We observed slight karyopyknosis by laser confocal microscopy in MCF-7 and MCF-7/MDR cells treated with 10 μM SB203580 alone (data not shown). However, when MDR cells were pretreated with 10 μM SB203580 for 1 h, they demonstrated increased apoptosis when treated with Doc than Doc alone (40.77% ± 2.87% for SB203580 combined with Doc, versus 14.4% ± 1.25% for Doc alone in MCF-7/ MDR cells; 32.6% ± 1.75% for SB203580 combined with Doc, versus 8.33% ± 0.88% for Doc alone in K562/MDR cells) ( Figure 5D). TTM did not further enhance apoptosis induced by SB203580 in Doc-treated MDR cells ( Figure  5D). These results indicated that inhibition of p38 MAPK enhanced chemosensitivity of MDR cells by activating apoptosis.

TTM inhibits p38 MAPK-induced P-gp in MCF-7 and K562 cells
Finally, we verified if TTM decreased P-gp expression in MDR cells by inhibiting p38 MAPK signaling pathway. First, we transfected MCF-7 and K562 parental cells with p38 overexpression vector (pMT3 p38) for 24 hr. Then, we treated the cells with or without TTM for an additional 48 hr and examined the levels of P-gp, p38 and phospho-p38. Western blot analysis demonstrated that total and phosphorylated p38 MAPK as well as P-gp levels were higher in p38 MAPK-overexpressing MCF-7 and K562 cells that were not treated with TTM ( Figure  6A). However, TTM-treated p38 MAPK-overexpressing MCF-7 and K562 cells showed decreased levels of phospho-p38 and P-gp ( Figure 6A). Moreover, qRT-PCR data demonstrated that TTM inhibited MDR1 mRNA levels in p38 MAPK-overexpressing MCF-7 and K562 cells (Supplementary Figure 7). We also observed that p38 MAPK-overexpressing MCF-7 and K562 cells treated with TTM showed increased DOX and Rh123 accumulation ( Figure 6B). These results confirmed that TTM decreased P-gp levels by inhibiting p38 MAPK signaling.

DISCUSSION
In this study, we investigated the mechanism of action by which TTM, a novel meroterpenoid that is isolated from the leaves of Rhodomyrtus tomentosa reverses P-gp-mediated MDR in cancer cells. This is the first report of a syncarpic acid-conjugated terpenoid as a MDR modulator.
We demonstrated that TTM decreased the high expression of P-gp in MCF-7/MDR and K562/MDR cells, thereby reversing multidrug resistance within acceptable levels of cytotoxicity [27]. We demonstrated that TTM below 100 μM did not inhibit cell proliferation in tumor and non-tumor cells ( Figure 1D-1F), but, enhanced cytotoxicity of different chemotherapeutic drugs, DOX, DNR, EPI and Doc in the P-gp overexpressing cancer cell lines.
The ABC transporter proteins that mediate MDR are involved in cellular detoxification [20]. They pump out www.impactjournals.com/oncotarget cytotoxic drugs from cancer cells, thereby decreasing their intracellular levels and thus enhance tumor cell survival. Therefore, P-gp antagonists re-sensitize MDR cells to anticancer drugs [45]. We demonstrated that TTM increased docetaxel cytotoxicity in MDR cancer cells by enhancing apoptosis as shown by high cleaved-caspase-9/3 and cleaved-PARP levels (Figure 2A-2C). TTM treatment alone did not affect cell viability.
Increased clonogenicity is characteristic of malignant and drug-resistant cancer cells, which are responsible for cancer recurrence [46]. We demonstrated that MDR cells were more clonogenic and resistant to anticancer drugs, Doc or DOX. However, TTM in combination with Doc or DOX significantly reduced clonogenicity of MDR cancer cells ( Figure 2D and Supplementary Figure 3). We further demonstrated that TTM enhanced intracellular accumulation of DOX and Rh123 in MDR cells by reducing their efflux in a concentration-dependent manner (Figure 3, Supplementary  Figure 4 and Supplementary Figure 5). Natural products inhibit MDR by either preventing efflux activity of P-gp or downregulating P-gp/MDR1 expression [47]. We observed that TTM significantly reduced P-gp expression in MCF-7/MDR and K562/MDR

Figure 6: Effect of TTM on p38 MAPK-mediated increase in P-gp expression in parental MCF-7 and K562 cells. (A)
Representative western blot analysis of P-gp and total and phosphorylated p38 MAPK levels in control and p38 MAPK overexpressing MCF-7 and K562 cells treated with or without 30 μM TTM for 48 hr. MCF-7 and K562 cells were transfected with (+) or without (-) 10 μg/l pMT3-p38MAPK plasmid for 24 hr to increase p38 MAPK expression. GAPDH was used as a loading control. Note: ** denotes P < 0.01 and *** denotes P < 0.001 compared to the pMT3 p38 group. (B) Flow cytometry analysis of DOX and Rh123 accumulation in control and p38 MAPK overexpressing MCF-7 and K562 cells treated with or without 30 μM TTM for 48 hr. After 48 hr, the cells were incubated with 10 μM DOX for 4 hr or 5 μM Rh123 for 2 hr. Intracellular fluorescence was analyzed by flow cytometry. The data are presented as the mean ± SEM from three independent experiments. Note: * denotes P < 0.05, ** denotes P < 0.01 and *** denotes P < 0.001 compared to the control group. cells. Since P-gp has a long half-life [23], we examined the time-course of TTM downregulation of P-gp in the two MDR cancer cells ( Figure 4C). We also found that TTM blocked MDR1 mRNA levels in the two MDR cancer cells (data not shown). This is the first report demonstrating the ability of TTM, a natural syncarpic acid-conjugated monoterpene in reducing the levels of MDR transporter, P-gp. We observed that decreased P-gp levels were observed after 8 hr of TTM treatment in MDR cells, although increased Rh123 accumulation was observed at 4 hr post-TTM treatment. This suggests that TTM inhibits P-gp function.
Anticancer drugs activate or inhibit signaling pathways that are associated with drug resistance of tumor cells [48]. ABC transporter expression has been associated with MAPK signaling [49,50]. The MAPK signaling pathway is an important player in MDR and a promising target for systemic therapy [5,51]. Moreover, the p38 signaling pathway is associated with cellular apoptosis in cancer cells [39,52]. We demonstrated that MDR cells showed high phospho-p38 levels that correlated with P-gp overexpression in MCF-7/MDR and K562/MDR cells than the corresponding parental cells ( Figure 5A). We further demonstrated that TTM decreased phopho-p38 MAPK and P-gp levels in a concentration-dependent manner. Moreover, the p38 MAPK inhibitor, SB203580 decreased Doc resistance of the MDR cancer cells. TTM did not alter the cytotoxicity of Doc upon SB203580 pre-treatment of MDR cancer cells ( Figure 5B and Supplementary Figure  6). Moreover, SB203580 pre-treatment decreased P-gp levels in MDR cells. However, SP600125, a JNK inhibitor had no effect in combination with Doc or DOX ( Figure  5B). Hence, we showed that p38 MAPK promotes P-gpmediated MDR. SB203580 also enhanced Doc-induced apoptosis in MDR cancer cells ( Figure 5D). These results demonstrate that the p38 MAPK pathway enhances P-gp levels in MCF-7/MDR and K562/MDR cells. P-gp stability is correlated with the activator protein-1 (AP-1) transcription factor [52], which is targeted by MAPK signaling pathways [53]. Human MDR1 promoter contains an AP-1-binding site [54], and increased AP-1 binding has been observed in several multidrug-resistant cell lines [55], whereas, reduced AP-1 binding has been associated with increased drug sensitivity in other cancer cell lines [56,57]. SB203580 has been shown to block c-Fos and c-Jun expression in response to UV irradiation and anisomycin [58]. Although p38 MAPK does not directly phosphorylate or activate c-Jun, several lines of evidence support the idea that p38 MAPK contributes to c-Jun induction via AP-1, which binds in the c-Jun promoter [59,60]. In addition to AP-1, c-Jun gene expression is regulated by MEF2 family of transcription factors [61]. MEF2 is essential for LPS induction of c-Jun in macrophages, which also requires functional p38 MAPK and MEF2C transcriptional factor. Furthermore, MEF2C is directly phosphorylated and activated by p38 MAPK [62]. Thus, p38 MAPK can potentially modulate c-Jun transcriptional activity by regulating AP1 and MEF2 transcription factors that have binding sites in c-Jun. We postulate that cJun may be intrinsically involved in p38 MAPK mediated P-gp expression.
It is also worth noting that SB203580 showed more cytotoxicity than TTM due to induction of karyopyknosis induction (data not shown) and lower IC 50 values in our models ( Figure 1D-1F, and Supplementary Figure 6). We further demonstrated that p38 overexpression parental MCF-7 and K562 cells increased P-gp levels by enhancing phospho-p38 MAPK levels. Thus, higher p38 MAPK levels decreased efficacy of TTM-mediated P-gp downregulation ( Figure 6A and Supplementary Figure 7). Moreover, higher p38 MAPK levels decreased intracellular DOX and Rh123 levels in presence of TTM ( Figure 6B).
In conclusion, TTM, a novel syncarpic acidconjugated monoterpene, increases cytotoxicity of chemotherapeutic drugs in MDR cancer cells by decreasing P-gp expression via p38 MAPK inhibition ( Figure 7). Thus, our study demonstrates that TTM is a potential therapeutic drug for cancer patients demonstrating multi-drug resistance.

Rhodomyrtus tomentosa
TTM was isolated from the petroleum ether soluble fraction of 95% ethanol extract of the leaves of Rhodomyrtus tomentosa through various chromatographic techniques including column chromatography over silica gel, ODS, Sephadex LH-20 and preparative HPLC. The structure was elucidated as a novel syncarpic acidconjugated monoterpenoid by a comprehensive analysis of the spectral data (IR, UV, HRESIMS, ECD, 1Dand 2D-NMR). The detailed isolation and structural elucidation are shown in the latest article [63].

Cell lines and cell culture
Parental sensitive human breast cancer cell line MCF-7 and multidrug resistant cell line MCF-7/MDR obtained by the Institute of Hematology and Blood Diseases Hospital (Tianjin, China) were used as an experimental model. Human leukemia cell line K562 and its Doxorubicin-selected P-gp-overexpressing K562/ MDR cells were purchased from the KeyGEN BioTECH (Nanjing, China). Both, sensitive and resistant cells were grown in RPMI-1640 medium supplemented with 10% fetal bovine serum at 37 °C with 5% CO 2 incubator. DOX (1.0 μM) was added to the culture medium to maintain the MDR characteristics of the MCF-7/MDR cells and K562/ MDR cells [64].

MTT assay
Cell proliferation was assessed by MTT colorimetric assay. Briefly, MCF-7, MCF-7/MDR and MCF-10A cells were seeded in 96-well plates at a density of 5 × 10 3 cells per well and incubated over-night to permit cell adhesion to the cell dishes. Then, the cells were treated with various concentrations of DOX, DNR, EPI, Doc and DDP in the presence or absence of TTM or Ver for 48 hr. 20 μL MTT solution was then added to each well followed incubation for an additional 4 hr. Finally, the purple formazan crystals formed were dissolved in 150 μL of DMSO by gently shaking for 10 mins. The absorbance was evaluated at a test wavelength of 570 nm, and a reference wavelength of 630 nm by ELISA reader (Spectra Max Plus384; Molecular Devices, Sunnyvale, CA). The concentration required to inhibit cell growth by 50% (IC 50 ) was calculated by using GraphPad Prism 5.0. Cell viability was calculated using the following formula: Cell viability = (At/As) × 100% At and As represent the absorbance of the test substances and solvent control, respectively [65].
The resistance index (RI) was calculated using the following formula: [28,66] resistance index (RI) = IC 50 (MDR cells) / IC 50 (their parental sensitive cells). The reversal fold change, in terms of potency of reversal, was calculated using the following formula: reversal fold change (RF) = IC 50 (MDR cells) / IC 50 (MDR cells combined with TTM or Ver treatment).
To measure the inhibitory effect of TTM on p38-MAPK signaling, MCF-7 or MCF-7/MDR cells were treated with serial dilutions of Doc with or without 10 μM SB203580 or 30 μM TTM and/or 10 μM SP600125 for 48 hr, followed by the MTT assay.
All experiments were conducted in triplicate and repeated more than three times. Ver was used as a positive control.

CCK-8 assay
K562 and K562/MDR cells were seeded in 96-well plates with a density of 5 × 10 3 cells/well in RPMI-1640 containing 10% FBS. Then the cells were separately exposed to a series of concentrations of compounds complexes for 48 hr and were added to a volume of 100 μL. After 48 hr incubation, 10 μL of the CCK-8 was dropped into each well to culture for 4 hr. The cell viabilities were monitored through measurement of the absorbance at 450 nm on an ELISA reader. The stimulation index (SI) was calculated based on the following formula [67]: SI = (OD tc -OD bl )/ (OD ntc -OD bl ). (OD tc is the mean OD of the treated cultures, OD ntc is the mean OD of the untreated cultures, and OD bl is the mean OD of the blank wells).
To measure the inhibitory effect of TTM on p38-MAPK signaling, K562 or K562/MDR cells were treated with serial dilutions of Doc with or without 10 μM SB203580 or 30 μM TTM and/or 10 μM SP600125 for 48 hr, followed by the CCK-8 assay.
All experiments were conducted in triplicate and repeated more than three times. Ver was used as a positive control.

Intracellular accumulation and efflux of DOX assay
The DOX accumulation analyses were performed by fluorescence-activated cell sorter (FACS) analysis and fluorescence microscopy analysis as described previously [68]. In the accumulation assay, both MDR cells and their parent cells were pretreated with or without 10 μM TTM or 10 μM Ver for 4 hr and then incubated with 10 μM DOX for 3 hr. Ver (10 μM) was used as a positive control. In the efflux study, sensitive cells and resistance cells were first cultured with medium containing 10 μM DOX at 37°C for 3 hr, washed 3 times with PBS, then incubated in the absence or presence of 10 μM TTM at 37°C for another 4 hr (or with 10 μM Ver as a positive control). After incubation, all cells were washed twice with icecold PBS and subjected to BD Accuri C6 flow cytometry (Becton & Dickinson Company, Franklin Lakes, NJ) to detect the fluorescence produced by DOX for efflux analysis.

Rh123 flow cytometry assay
The function of P-gp was evaluated by examining the intracellular accumulation and efflux of Rh123 as previously described [68,69]. In the accumulation analysis, MDR cells and their parent cells were treated with or without 10, 30 or 50 μM TTM or 10 μM Ver for 1.5 hr and then incubated with 5 μM Rh123 at 37 °C for 1.5 hr. In the efflux study, MDR cells and their parent cells were first cultured with medium containing 5 μM Rh123 in the dark for 1.5 hr, washed 2 times with PBS, then incubated in the absence or presence of 10, 30 or 50 μM TTM at 37 °C for additional 1.5 hr (or with 10 μM Ver as a positive control). Subsequently, the culture medium was removed and cells were washed two times with PBS. Finally, the cells were suspended into 250 μL PBS followed by FACS to examine green fluorescence produced.

The laser confocal to examine the accumulation of Rh123 on MDR cells
To further visualize the effect of TTM on the intracellular retention of Rh123, 5 × 10 3 cells per well were seeded in the 96-well plate for overnight. The cells were treated with TTM (10, 30 μM) for 48 hr, and were then incubated with 10 μM Rh123 alone or 10 μM Rh123 combination with TTM in the fresh RPMI 1640 medium for 1.5 hr in darkness at 37 °C. After that cells were washed 3 times with cold PBS and images were acquired by the laser confocal.

Apoptosis assay
Apoptosis was evaluated by FACS consistent with the manufacturer's instructions. Sensitive cells and MDR cells were treated with (+) or without (-) 30 μM TTM combined with 1.0 μM Doc for 48 hr. Cells were then washed twice with cold PBS, suspended in 100 μL of binding buffer at a concentration of 1 × 10 6 cells/mL. After that, cells were incubated with fluorescein isothicocyannte (FITC)-conjugated annexin V reagent and PI in binging buffer for 30 mins at room temperature as described by the manufacture and finally determined by FACS.

Colony formation assays
MCF-7 and MCF-7/MDR cells were plated on 6-well plates at a density of eight hundred cells per well. Following incubation overnight, the cells were treated with TTM (10 μM) with or without Doc. The medium was replaced after 48 hr of incubation, and the colonies were further observed 15 days later. The plates were then stained with crystal violet solution (Sigma-Aldrich, St. Louis, MO), and the photographs of the colonies were taken manually [70].

Plasmid transfection assay
Plasmid preparation was performed and as previously described with some modifications [71,72]. MCF-7 cells and K562 cells were washed twice with Opti-MEM (Life Technologies, Grand Island, NY). Cells were then transfected with plasmid using Opti-MEM containing 10 μg plasmids in 100 μL of Opti-MEM and Super Electroporator NEPA21 system (NEPAGENE, Japan). After that cells were plated on 6-well plates at a density of 1 × 10 6 cells cells per well. Twenty-four hours after transfection, cells were treated with 30 μM TTM for 48 hr prior to harvesting.

Overexpression of P-gp and p38 in MCF-7 and K562 cells
Transfection was performed using the Super Electroporator NEPA21 system. In brief, 1 × 10 6 cells were transfected using Opti-MEM containing 10 μg plasmid. MCF-7 cells and K562 cells were transfected with the MDR1 plasmid, p38 plasmid or their mock plasmids for 24 hr, and treated with or without 30 μM TTM for another 48 hr, followed by Western blot analysis.

The laser confocal to detect the levels of P-gp on the cell membrane
Immunofluorescence was done as previously described before [74]. MCF-7 cells and MCF-7/MDR were cultured on a 96 wells glass culture plate. Cells were treated with TTM for 48 hr. The cells were fixed with warm 4% paraformaldehyde for 15 mins at room temperature. Then abandon the fixed liquid, washed 2 times with PBS for each 5mins and blocked with 5% BSA for 1 hr. Incubation with primary antibodies against P-gp was performed overnight at 4 °C. After incubation, washed 2 times with PBS for each 5 mins. The cells were then incubated with Alexa Flour 488-conjugated secondary antibody for 2 hr at room temperature in the dark. After incubation, washed 2 times with PBS for each 5 mins. The nuclei were stained with DAPI (Beyotime, Haimen, China) for 10 mins before imaging. Images were taken by ImageXpress ® Micro Confocal (Molecular Devices, USA).

Statistical analysis
For all experiments, quantitative results were reported as "mean ± SEM" values from at least three experiments performed in a parallel manner. Statistical analysis was performed with the GraphPad Prism 5 software. One-way ANOVA was used to compare each parameter. Next, unpaired, two tailed student's t test was www.impactjournals.com/oncotarget performed to identify which group differences accounted for significant overall ANOVA results. * denotes P < 0.05, ** or ## denotes P < 0.01 and *** or ### denotes P < 0.001 was considered as significant.

Author contributions
XZ designed, performed and analyzed experiments, and drafted the article. YX contributed to the conception, data analysis, design of the study and revised the manuscript. YZ and JL isolated and purified the compound TTM, and collected relevant data. CH, HZ and CZ analyzed and interpreted of data, and revised the manuscript. LY and LK revised the manuscript and finally approved of the version to be published. All authors read and approved the final manuscript.