Cardamonin Enhances Cisplatin Chemosensitivity of Nasopharyngeal Cancer Cells via Attenuating c-Myc-Mediated β-Catenin/ABCG2 Signaling

Purpose: Resistance to chemotherapeutic drugs in nasopharyngeal carcinoma(NPC) remains a major obstacle of clinical therapy. To address the issue, screening for natural low-toxicity products as chemosensitizers has become a promising strategy for cancer therapy. In this study, we investigated chemosensitizing effects of cardamonin (CM), a plant-derived chalcone, on cisplatin (DPP)-resistant NPC cells, and explored the molecular mechanism for its antitumor activity. Methods: The chemotherapeutic ecacy of cardamonin, cisplatin and their combination in cisplatin-resistant NPC cells were analyzed using MTT assay, apoptosis assay, and cell cycle analysis. Real-time PCR, western blotting, and cell transfection analysis were performed to assess the synergistic inhibitory action of cardamonin supplemented with cisplatin on Wnt/β-catenin/ABCG2 signaling. The effect of cardamonin on ABCG2 drug eux function was analyzed by doxorubicin accumulation assay. A CNE2/DPP nude mouse model was used to determine the combinatorial effects of cardamonin on tumor growth in vivo. Results: Cardamonin increased cisplatin-induced cytotoxicity, accompanied by induction of apoptosis and cell cycle arrest in DPP-resistant NPC cells. Moreover, cardamonin could synergized with cisplatin to downregulate β-catenin, c-Myc, and ABCG2. Specically, cardamonin inhibited Wnt/β-catenin/ABCG2 signaling through c-Myc-mediated transcription inactivation, thereby suppressing the expression of ABCG2 in cisplatin-resistant NPC cells. These ndings were conrmed in vivo, wherein cardamonin treatment with cisplatin resulted in reduced tumor growth in a CNE2/DPP xenograft animal model. Conclusions: Taken together, our data rstly demonstrated that cardamonin increased chemosensitivity of nasopharyngeal cancer cells to cisplatin through inactivation of Wnt/β-catenin/ABCG2 signaling, more specically by inhibition of β-catenin/ABCG2 signaling through c-Myc-mediated transcriptional inactivation, thereby downregulation of ABCG2 and reversal of cisplatin resistance. Thus, in addition to its chemotherapeutic potential, cardamonin may serve as a useful chemosensitizer to conventional chemotherapeutic drugs in the treatment of nasopharyngeal carcinoma.


Introduction
Nasopharyngeal carcinoma (NPC) or nasopharynx cancer, is the most common head and neck tumor that is prevalent in southern China and Southeastern Asia [1,2]. Although NPC is generally a good response to radiation therapy, cisplatin-based combinational chemoradiotherapy is required for locally advanced cancers [3,4,5] and the 5-year overall survival rate still remains low [6]. One serious challenge with cisplatin-based chemothrapy is the acquisition of chemoresistance due to increased expressions of ATPbinding cassette (ABC) transporters that is a key mechanism that contributes to multidrug resistance [7,8,9]. Consequently, there is an urgent need for developing new therapeutic agents for circumventing cisplatin resistance in the treatment of NPC.
ABCG2 has been reported as one of at least three human ABC transporters that facilitate e ux of conventional chemotherapeutic drugs including DPP [18]. Additionally, ABCG2 was shown to mediate DPP resistance, which was reversed by inhibiting wnt/β-catenin signaling in ovarian cancer [16]. Therefore, a combination of chemotherapy targeting abnormal Wnt/β-catenin/ABCG2 signaling has become a promising cancer therapy to improve tumor chemosensitivity and eliminate cancerous cells.
Recently, the use of selective, potent, and relatively non-toxic natural compounds to impede tumors and enhance chemosensitivity has gained immense importance [19]. Moreover, some natural plant products exhibit excellent in inhibiting Wnt/β-catenin signaling and improving chemosensitivity in some cancers [20]. Thus, the natural low-toxicity compounds might be promising agents to enhance the therapeutic e cacy of conventional chemotherapeutic drugs for drug-resistant tumors. Cardamonin, a natural chalcone extracted from cardamom spice, has been reported extensively to have antiin ammatory and anti-tumor activities [21]. However, whether CM can chemosensitize resistant NPC cells to DPP remains unexplored. In this report, we rstly demonstrated that CM could circumvent chemoresistance to DDP in resistant NPC cells by blocking β-catenin/ABCG2 signaling pathway.
Furthermore, we identi ed major ABC transporter ABCG2 to be downregulated following co-treatment of CM and DPP via c-Myc-mediated transcription inactivation in DPP resistant NPC cells. Finally, a xenograft mouse model was established to validate our in vitro results and further revealed that the combination of cardamom and DPP signi cantly reduced tumor growth more effectively than when used alone in vivo.
Collectively, these data indicate that CM inhibits NPC growth and reverses the DPP resistance of NPC through inactivation of c-Myc-mediated β-catenin/ABCG2 signaling.

Reagents
Cardamonin, cisplatin, and other agents were purchased from Sigma-Aldrich Company. Primers were obtained from Thermo Fisher Scienti c. Antibodies against ABCG2, β-catenin, α-tubulin, c-Myc, SOX2, and cyclin D1 were from Cell Signaling Technology. The secondary antibodies were purchased from Santa Cruz.

Cell culture
The human nasopharyngeal carcinoma cell line CNE2 and its cisplatin-resistant cell line (CNE2/DDP) selected by continuous exposure of parental CNE2 cells to increasing concentrations of DPP over a Western blot analysis Cells were collected after treatment for 24h under different conditions and total cell lysates using RIPA lysis buffer were obtained from the treated cells. The protein concentrations were determined using Bio-Rad protein assays (Sigma-Aldrich, Shanghai, China) and all experimental samples were normalized by αtubulin as a reference protein. Proteins from cell lysates were separated on 12% SDS-PAGE, following by electrotransferring to PVDF membranes. The membranes were blocked with 5% BSA and incubated with speci c primary antibodies at 4°C overnight. Then it was incubated with secondary antibodies at room temperature for 1h. Finally, proteins of interest were visualized with the ECL detection system (Millipore Co.) following the manufacturer's protocol.

Cell transfection
The pcDNA3-β-catenin, pcDNA3-c-Myc, and empty plasmids (pcDNA3) were constructed by GenePharma (Shanghai, China). The small interfering RNAs (siRNAs) targeting β-catenin and c-Myc or their scramble siRNAs were purchased from Ribobio Company (Guangzhou, China). For transient transfection, overexpressing plasmids or siRNAs of indicated genes were individually transfected into CNE2/DPP cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. The control groups were generated by infecting CNE2/DPP cells with empty plasmid or scramble siRNAs.

Drug accumulation assay
The effect of CM on ABCG2 activity was assessed by measuring intracellular doxorubicin(DOX) accumulation that is a conventional substrate chemotherapeutic agent of ABCG2 [24]. CNE2 and CNE2/DPP(5×10 5 per well) cells were seeded into six-well plates respectively and treated by 4uM doxorubicin with or without various concentrations of CM for 1h. Then, the cells were collected and washed twice with ice-cold PBS and nally resuspended in 100μL of ice-cold PBS for DOX-associated MFI measured by ow cytometry. Fumitremorgin C (FTC, a speci c inhibitor of ABCG2 that can block the pumping out of doxorubicin in overexpressing ABCG2 cells [24]) was used as the positive control.

In vivo xenograft assay
To evaluate the in vivo effects of CM on CNE2/DPP cells, a CNE2/DPP xenograft model was established.
Six-week-old BALB/c nude mice were purchased from the Medical Laboratory of the University of Zhengzhou (Zhengzhou, China). CNE2/DPP cells (5×10 6 ) were injected subcutaneously into the left ank of the mice. Three days after injection, the mice were randomized into four groups: Control (PBS, 0.1ml/10kg), DPP (5mg/kg), CM (25mg/kg), DPP (5 mg/kg) and CM (25 mg/kg). Five mice were set up for each group and every group was administered twice a week by intraperitoneal injection. Tumor volume was measured at 3-day intervals with a caliper and calculated as (length×width2)/2 for four weeks. After 4 -week drug treatment individually or in combination, nude mice were humanely euthanized and xenograft tumors were harvested and measured.

Statistical analysis
All analysis was conducted using GraphPad Prism 7.0 (GraphPad Software Inc. San Diego, CA). Data are expressed as the mean±SD of triplicate samples. Differences were analyzed by the Student t test and One-way analysis of variance. P-value < 0.05 was accepted as indicative of signi cant differences.

Cardamonin increases the chemosensitivity to DPP in DPP-resistant NPC cells
To determine the effects of CM on DPP-resistant NPC cells individually or in combination, the cytotoxicity of both drugs was analyzed using sensitive NPC cells (CNE2) and its DPP-resistant NPC cells (CNE2/DPP). The IC50 values of DPP in CNE2 and CNE2/DPP cells were found to be 4.687 and 32.896uM respectively and Resistance Index (RI) rose to more than seven-folds, which re ected therapeutic e cacy of drug-resistant cells for DPP (Table1). Furthermore, there was markedly reduced cisplatin cytotoxicity in CNE2/DPP cells in comparison to their sensitive counterparts (Figure1A). Importantly, we found treatment with CM signi cantly decreased the IC50s of DPP against CNE2/DPP cells in a dose-dependent manner ( Figure 1C). In contrast, no signi cant effects were observed in CNE2 cells ( Figure 1B). These foundings indicate that CM potentiates chemosensitivity of DPP to CNE2/DPP cells but exerts little in uence on CNE2 cells, supporting the notion that CM can attenuate chemoresistance in DPP-resistant NPC cells.

Cardamonin induces apoptosis and cell cycle arrest in DPP-resistant NPC cells
To clarify whether CM by synergizing with DPP could contribute to corresponding enhancement in apoptosis, apoptosis assays were performed following CM and DPP treatment either alone or combined. We observed that CM or DPP treatment alone signi cantly increased apoptosis while combined CM and DPP treatment further enhanced apoptosis in sensitive CNE2 cells (Figure 2A), indicating the increased apoptosis in the combination therapy may be primarily driven by DPP. On the contrary, in resistant CNE2/DPP cells, DPP treatment did not induce apoptosis, but the cotreatment of DPP and CM resulted in signi cantly increased apoptosis ( Figure 2B). These results supported that CM induced more apoptosis in resistant CNE2/DPP cells upon DPP treatment.
DPP has been reported as a G0/G1 and G2/M cell cycle arrest inducer [25]. Then, we investigated how the combined treatment with CM and DPP regulated cell cycle in both sensitive CNE2 cells and its resistant CNE22/DPP cells. Our results revealed that DPP induced G0/G1 arrest while CM induced G2/M arrest in the sensitive CNE2 cells. However, the combined CM and DPP treatment caused G0/ G1 arrest ( Figure  2C), suggesting that DPP regulates primarily the cell cycle dynamics in the sensitive CNE2 cells. On the contrary, in resistant CNE2/DPP cells, CM treatment caused a G2/M growth arrest, but DPP treatment had little effect on cell cycle activities. Notably, the combined CM and DPP treatment resulted in enhanced G2/M phase arrest ( Figure 2D), indicating that the cell cycle dynamics is primarily driven by CM in resistant CNE2/DPP cells. These results appeared to be in line with apoptosis assays above.
Cardamonin decreases the expressions of βcatenin, c-Myc, and ABCG2 in CNE2/DPP Cells To understand the molecular mechanisms of the therapeutic effects of the combined treatment of CM and DPP shown above, rstly, the expression levels of multidrug resistance-associated genes and Wnt/βcatenin-associated genes were screened and analyzed by real-time PCR analysis in CNE2 and CNE2/DPP cells. We found that the resistant CNE2/DPP cells exhibited much higher expressions of β-catenin, c-Myc, and ABCG2 than the sensitive CNE2 counterparts ( Figure 3A and Figure 3B). Then, the expressions of multidrug resistance-speci c genes (ABCB1, MRP1, ABCG2, and CD44) and Wnt/β-catenin-speci c genes (β-catenin, Nanog, Sox2, OCT4, c-Myc, KLF4, and cyclin D1) were measured using real-time PCR analysis in CNE2/DPP cells treated with or without CM for 24h. The data showed that the transcript expressions of ABCG2, β-catenin, and c-Myc were markedly downregulated by cotreatment of CM and DPP, compared with either agent alone ( Figure 3C and Figure 3D). In addition, the protein expression levels of β-catenin, c-Myc, and ABCG2 analyzed by Western Blotting in CNE2/DPP cells were also decreased in comparison to CM or DPP treatment alone ( Figure 3E) after 24h of cotreatment with CM and DPP. Overall, these results suggest that CM inhibits chemoresistance through Wnt/β-catenin/ABCG2 signaling pathway in combinational treatment.
Cardamonin attenuates c-Myc-mediated βcatenin/ABCG2 signaling To further assess the effects of CM on the downstream target genes of Wnt/β-catenin signaling in resistant CNE2/DPP cells, c-Myc, and ABCG2 expression were analyzed by real-time PCR following βcatenin overexpression or knockdown. We found that β-catenin overexpression upregulated c-Myc and ABCG2 gene expression, whereas β-catenin knockdown led to the opposite effect ( Figure 4A and 4B).
Also, the combining CM with DPP led to a signi cant decline in the expression levels of the indicated genes in β-catenin knockdown CNE2/DPP cells ( Figure 4B). Next, to determine whether c-Myc is critical for CM action, we set out to assess the transcript and protein expression of ABCG2 analyzed by real-time PCR and Western blotting after c-Myc overexpression or knockdown. We found that the expression levels of ABCG2 were dramatically upregulated in c-Myc overexpressing cells ( Figure 4C and 4D), whereas the opposite was observed after c-Myc knockdown in DPP-resistant cells ( Figure 4E and 4F). Importantly, we observed the inhibitory effect of CM on the ABCG2 signal was weakened in c-Myc silencing cells ( Figure 4E and 4F), whereas CM could at least partly counteract exogenous overexpression of c-Myc supplemented with DPP in increasing ABCG2 expression ( Figure 4C and 4D), implying that c-Myc plays a crucial role in mediating the chemosensitizing effects of CM on DPP-resistant NPC cells via the downregulation of ABCG2. Taken together, these data suggest that CM enhances chemosensitivity of DPP-resistant NPC cells to DPP by attenuating c-Myc-mediated transcription activation, resulting in reduced expression of ABCG2.

Cardamonin increases accumulation of Doxorubicin
To determine whether CM could affect the activities of drug e ux transporter ABCG2, the intracellular concentration of doxorubicin (DOX) that is a conventional substrate chemotherapeutic agent of ABCG2 [26] in CM-treated CNE2 and CNE2/DPP cells were detected using doxorubicin accumulation assay. We found the intracellular accumulation of doxorubicin in the resistant CNE2/DPP cells was remarkably lower than that in the sensitive CNE2 counterparts ( Figure 5). Importantly, treating with CM signi cantly increased the intracellular accumulation of doxorubicin in a concentration-dependent manner in CNE2/DPP, but no signi cant effects were observed in the sensitive CNE2 cells. In fact, the increased uorescence in the CM-treated CNE2/DPP cells was similar to that in the FTC-treated counterparts, indicating that CM can elevate the concentration of doxorubicin inside the CNE2/DPP cells by inhibition the drug e ux function of ABCG2.

Cardamonin suppresses NPC xenograft growth
To verify the in vitro ndings mentioned above, the antitumor effects of CM were evaluated using a mouse xenograft model of DPP-resistant NPC. CM could not only reduce tumor volumes ( Figure 6A) but also lower tumor weights ( Figure 6B) as compared to the untreated controls. However, the combined CM and DPP treatment markedly suppressed tumor growth in comparison with CM or DPP treatment alone as shown in Figure 6C. These ndings indicate that the combination of CM and DPP can inhibit tumor growth more effectively than when used alone in vivo. Also, these results were consistent with our in vitro ndings.

Discussion
Chemotherapeutic drug resistance is one of the major unmet clinical challenges in NPC treatment. The natural plant products have become promising strategies in overcoming chemoresistance [20]. In this study, we for the rst time demonstrate a novel molecular mechanism for CM-induced chemosensitization to DPP in DPP-resistant NPC cells by blocking the Wnt/β-catenin/ABCG2 signaling pathway. This study reveals that the combination of CM and DPP exerted synergistic inhibitory effects on DPP-resistant NPC cell survival. Mechanismly, CM could reverse drug resistance via inhibition of βcatenin/ABCG2 signaling, which was closely correlated with suppressing c-Myc-mediated transcription in DPP-resistant NPC cells. In addition, a xenograft animal model was established to con rm the in vitro ndings, indicating CM as adjunctive therapy to DPP can inhibit tumor growth of NPC and circumvent drug resistance in vitro and in vivo.
It is crucial to understand the occurrence of chemoresistance, and eventually how to prevent it for combatting cancer effectively [8,27,28], and it is well known that the ABC transporter family are implicated in multidrug resistance and high-expressed levels of ABCG2 have a greater capacity to expel therapeutic drugs [25,29,30]. Furthermore, accumulating ndings reveal that aberrant Wnt/β-catenin signaling and increased ABCG2 expression are closely correlated to multidrug resistance in many cancers [14,15,31], including NPC [18,17]. Also, ABCG2 was reported as a downstream gene of β-catenin since its seven functional transcription factor binding sites have been identi ed in the ABCG2 gene promoter [16]. Therefore, targeting the aberrant Wnt/ βcatenin has become a novel approach to improve chemosensitivity in cancer treatment. Accordingly, in the present study, after screening for the expression levels of multidrug resistance-associated genes and Wnt/β-catenin-associated genes using real-time PCR analysis in CNE2 and CNE2/DPP cells, we found that the resistant CNE2/DPP cells exhibited much higher expressions of β-catenin, c-Myc, and ABCG2 than the sensitive CNE2 cells ( Figure 3A and 3B). Importantly, the three gene expressions in CNE2/DPP cells were markedly downregulated by cotreatment of CM and DPP, compared with either agent alone ( Figure 3C and 3D). In addition, we found that protein expressions of β-catenin, c-Myc, and ABCG2 decreased following CM and DPP co-treatment of resistant CNE2/DPP cells using western blotting analysis ( Figure 3E). These results suggest that CM overcomes chemoresistance through inhibition of Wnt/β-catenin/ABCG2 signaling.
C-Myc as an important regulator of stem cells may be associated with tumorigenesis by activating its downstream target genes [32,33]; Furthermore, recent studies have shown that c-Myc may dysregulate the transcription of ABC transporter genes such as binding to the promoter of ABCG2 and increase its expression in some cancers, thereby resulting in the multidrug resistance pro le [34,35,36]. In addition, it has been reported that c-Myc is the ultimate downstream effector of the Wnt/β-catenin pathway [37,38].
These results suggested that c-Myc may serve as a link connecting Wnt/β-Catenin pathway and multidrug resistance. Herein, our data demonstrated that c-Myc expression was upregulated following βcatenin overexpression, whereas β-catenin knockdown decreased the expression levels of c-Myc in CNE2/DPP cells ( Figure 4A and 4B). To determine whether c-Myc is critical for CM action, c-Myc overexpression or knockdown was conducted by transfecting overexpressing plasmids or siRNAs targeting c-Myc to CNE2/DPP cells. As shown in Figure 4E and 4F, the inhibitory effect of CM on ABCG2 expression was weakened in c-Myc silencing cells, whereas exogenous overexpression of c-Myc could at least partly counteract CM in inhibiting ABCG2 expression ( Figure 4C and 4D        Cardamonin decreases the intracellular accumulation of doxorubicin in DPP-resistant NPC cells. Cells were pretreated with 0, 10, 20, and 40uM of cardamonin for 1h, and then exposed to 4uM doxorubicin for another 1h. DOX-associated MFI was measured by ow cytometry. Data are expressed as means ± SD of th ree independent experiments (*P< 0.05; **p<0.01; ***p<0.001 vs. control). Each point represents the mean SEM of 5 mice per group. Data represent the mean ± SEM of three independent experiments. (n = 5 mice/group; *P<0.05, **P<0.01, ***p<0.001vs. control).