A Novel Polyphenol Conjugate Sensitizes Cisplatin-Resistant Head and Neck Cancer Cells to Cisplatin via Nrf2 Inhibition

The integration of chemotherapy and targeted therapy with radiation has become more popular in many human cancers in the emerging era of personalized therapy (1). Cisplatin is a very old antineoplastic agent first described in 1844 (2), but it is still an important therapeutic tool for various types of human solid neoplasms, including testicular, bladder, ovarian, lung, colorectal, and head and neck cancer (HNC; ref. 3). As a first-line chemotherapeutic agent, cisplatin is currently used in combination with radiotherapy in the organ preservation protocol for HNC (4). Cisplatin binds with high affinity to nuclear DNA, causing a DNA damage response and inducing apoptosis in cancer cells (5). Cisplatin also provokes undesirable moderate-to-severe adverse effects involving the immune system, kidney, gastrointestinal tract, and nervous system (6). This has prompted the use of other platinum derivatives of oxaliplatin and carboplatin and the continuous development of other novel platinum-containing antineoplastic agents to broaden their clinical use (7). Moreover, the elevated incidence of chemoresistance is the main limitation of cisplatin use in clinical practice, leading to a failure in cancer management (8). There are several strategies for circumventing cisplatin toxicity and resistance in cancer patients, including its combined use with targeted agents. These combinatorial regimens have been applied as novel therapeutic strategies for many human cancers, but they have largely failed to improve the therapeutic profile in randomized controlled trials (9, 10). These observations indicate the need for other novel anticancer agents to reduce toxicity and overcome resistance in cisplatin-based chemotherapy.

HNC, the eighth most common cancer worldwide, is frequently associated with gains of mutations in oncogenes and losses of mutations in tumor suppressor genes that are generally related to resistance to cancer therapy (11–13). Oncogene addiction and inactivation of tumor suppressor genes result in cell deregulation due to increased cellular stress, which are not ordinarily observed in normal cells (14). Targeting the cancer-specific deregulation may result in the selective death of cancer cells (15). Cancer cells are characterized by high levels of oxidative stress involving reactive oxygen species (ROS), resulting in the alterations of cellular metabolic pathways and the activation of the antioxidant defense mechanism (16, 17). Although the altered cell metabolism plays a crucial role in cancer development, this may represent the most critical weak point in most cancer types and provide a therapeutic approach to killing cancer cells (18). Combinations of antioxidant inhibitors with radiotherapy or chemotherapeutic agents induce increased oxidative stress in cancer cells but not normal cells, causing selective cancer cell death (19, 20). Therefore, an increase in the levels of ROS is increasingly accepted as a valuable anticancer strategy for future anticancer drug discovery (21, 22).

DPP-23 [(E)-3-(3,5-dimethoxyphenyl)-1-(2-methoxyphenyl)prop-2-en-1-one] is a novel synthetic polyphenol conjugate recently developed to be selectively toxic to cancer cells but not normal cells (23). DPP-23 induces the production of ROS in cancer cells and subsequently targets the unfolded protein response in the endoplasmic reticulum, resulting in caspase-dependent apoptosis. The selective generation of ROS in cancer cells is proposed to be the main mechanism of action of DPP-23, but its mode of action requires further elucidation. The effects of DPP-23 in reversing multidrug resistance have also rarely been tested in HNC cells in experimental or clinical studies. Therefore, further investigation of the mechanism of action of DPP-23 and its synergy with conventional chemotherapeutic agents is needed. In our current study, we show that DPP-23 selectively kills HNC cells by targeting the oxidative stress response and increases cisplatin antitumor activity in resistant HNC cells both in vitro and in vivo.

Asan Medical Center (AMC; Seoul, Republic of Korea) cancer cell lines previously established from the primary tumors of head and neck in our institution (24) and other cancer cell lines were obtained from the Korea Cell Line Bank and ATCC. Cell lines were authenticated by STR-based DNA fingerprinting and multiplex PCR. The cancer cells were grown in Eagle's minimum essential medium (Thermo Fisher Scientific) supplemented with 10% FBS. Cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. Normal oral keratinocytes or fibroblasts were obtained from patients undergoing oral surgery and used in in vitro assays. Cisplatin-resistant cells were developed from their parental cisplatin-sensitive HN3, HN4, and HN9 cells by exposing them to continuously increasing concentrations of cisplatin (Sigma-Aldrich; ref. 25). Cisplatin resistance was evaluated by using cell viability assays performed in both the resistant and parental cells, and the IC₅₀ of this drug for each HNC cell type was calculated.

Drug cytotoxicity was measured by assessing cell viability using Trypan blue exclusion, crystal violet staining, and MTT (3-[4,5-dimethyl-2-thiazolyl]-2,5-diphenyl-2H-tetrazolium bromide) and clonogenic assays. All of the cell viability tests were performed after the treatment of cells with DPP-23 (a kind gift from Professor Y. Lim at Konkuk University, Seoul, South Korea; ref. 23), cisplatin, their combination, or an equivalent amount of DMSO (control) for 72 hours. For Trypan blue exclusion, the cells were trypsinized, stained with 0.4% Trypan blue (Thermo Fisher Scientific), and counted using a hemocytometer. For crystal violet staining, the cells were fixed in ice-cold 100% methanol and stained with 0.5% crystal violet solution (Sigma-Aldrich). For MTT assays, the cells were exposed to the tetrazolium compound MTT (Sigma-Aldrich) for 4 hours, followed by solubilization buffer for 2 hours. Absorbance was then measured at 570 nm using a SpectraMax M2 microplate reader (Molecular Devices). For clonogenic assays, the cells were stained with 0.5% crystal violet solution, and the number of colonies was counted. All assays were performed with triplicate samples and repeated three times. The interaction of cisplatin and DPP-23 was considered synergistic when growth suppression was greater than the sum of the suppression induced by either drug alone; the combination index (CI) was used to evaluate drug interactions: CI = 1, additive interaction; CI < 1, synergistic interaction; CI > 1, antagonistic interaction (ComboSyn, Inc.; ref. 26).

Cell cycle, cell death, and caspase activity assays were performed in HNC cells exposed to DPP-23, cisplatin, their combination, or an equivalent amount of DMSO (control) for 72 hours. For cell-cycle assays, the cells were fixed overnight in ice-cold ethanol and stained for 30 minutes with propidium iodide (Sigma-Aldrich) at 37°C. The cellular DNA content was measured using a FACSCalibur flow cytometer (BD Biosciences). For cell death assays, the cells were washed in ice-cold PBS and resuspended in binding buffer. The cells were stained using an Annexin V-FITC Apoptosis Detection Kit (BD Biosciences) and then analyzed with flow cytometry and CellQuest software (BD Biosciences). Caspase activity assays were performed in triplicate wells using the fluorimetric Homogeneous Caspase Assay (Roche). The substrate working solution was added, and the plate was incubated in the dark at 37°C for 2 to 8 hours. The fluorescence in each well was measured at an excitation wavelength of 485 nm and an emission wavelength of 520 nm using a SpectraMax M2 microplate reader. For the measurement of the mitochondrial membrane potential (ΔΨm), the cells were stained with 200 nmol/L tetramethylrhodamine ethyl ester (Life Technologies) for 20 minutes and then analyzed by flow cytometry. The median fluorescent intensity of each treatment group was normalized to that of the control group. The statistical significance of differences among different treatment groups was assessed using the Mann–Whitney U test with Bonferroni post hoc adjustment.

To assay the total cellular glutathione (GSH) levels, tumor and normal cells were exposed to DPP-23 and N-acetyl-L-cysteine (NAC; 3 mmol/L; Sigma-Aldrich). The cell lysates were discarded and GSH levels were assayed in the supernatant using a Glutathione Colorimetric Detection Kit (BioVision Inc.). Glutathione disulfide (GSSG) quantification was performed using a GSH/GSSG Detection Kit (Abcam) according to the manufacturer's protocol. The change in GSSG levels in the DPP-23–treated samples compared with the DMSO-treated control samples was expressed as the fold change. To measure ROS levels, the cells were exposed to 5 μmol/L DPP-23 or an equivalent amount of DMSO (control) for 3 and 6 hours, and ROS generation was detected with 2′,7′-dichlorofluorescein diacetate (DCF-DA; Enzo Life Sciences). The cells were incubated with 10 μmol/L DCF-DA for 30 minutes at 37°C and analyzed in a FACSCalibur flow cytometer. In a different experiment, the cells were pretreated with 3 mmol/L NAC for 1 hour or catalase (2,000 U/mL; Sigma-Aldrich) for 2 hours before exposure to DPP-23. The ROS levels were measured by flow cytometry using DCF-DA and were then shown as the fold change over the DMSO treatment levels.

For knockdown experiments, HN3-cisR and HN9-cisR were seeded and transfected 18 hours later with 50 nmol/L siRNA targeting human TP53, Nrf2, PUMA, or a scrambled control siRNA (Integrated DNA Technologies). After 48 hours, the cells were exposed to DPP-23 for an additional period of 24 hours and then protein expression was analyzed. For stable Nrf2 overexpression, the HN3-cisR cells were stably transfected with vector control or Nrf2 plasmid (TransOMIC). The knockdown and overexpression were confirmed by Western blotting using the specific antibodies.

Cells were lysed at 4°C in RIPA buffer (Thermo Scientific). Immunoblotting was then performed using standard procedures. Briefly, a total of 50 μg protein was resolved by SDS-PAGE on 10% to 12% gels, transferred to nitrocellulose or polyvinylidene difluoride membranes, and probed with primary and secondary antibodies. The following primary antibodies were used: p53 (DO1; Santa Cruz Biotechnology) and p21^WAF1/CIP1, cleaved PARP, cleaved caspase 3, phospho-p53-Ser15, PUMA, and BAX (Cell Signaling Technology), and Mcl-1, Nrf2, NQO1, HO-1, and Keap1 (Abcam). β-Actin (Sigma-Aldrich) was used as a loading control. All antibodies were diluted to between 1:250 and 1:5,000.

All animal study procedures were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee of our institution. Six-week-old athymic BALB/c male nude mice (nu/nu) were purchased from Central Lab Animal Inc. AMC-HN9-cisR cells (5 × 10⁶) were injected subcutaneously into the flank. Tumor volumes and body weights were measured every 3 days. Tumors were measured using calipers, and volume was calculated by (length × width²)/2. Treatment began when the cell implants became palpable nodules (day 0). The mice were randomized into four treatment groups: vehicle, cisplatin, DPP-23, and cisplatin plus DPP-23 (n = 10 each). The oral food intake was measured daily in different treatment groups after treatment, and the intake of each group was compared as the fold change over that of the control group.

Mice were treated by intraperitoneal injection of 5 mg/kg cisplatin once per week or by intraperitoneal injection of 10 mg/kg DPP-23 once per day (23) or with a combination of cisplatin and DPP-23 according to the same schedules. The mice were sacrificed on day 34, and the tumors were isolated, weighed, and compared among different groups. Tumors were also analyzed by immunoblotting and an in situ terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay (R&D Systems). The number of apoptotic bodies was counted in a blind manner in 10 randomly selected high-power fields. Whole-blood samples were collected from the tail veins of DPP-23–treated or control mice and processed using an automated hematology analyzer (Beckman Coulter). For histologic evaluation, normal tissues from vital organs, namely, the lung, liver, kidney, spleen, and small/large intestines, were isolated, fixed in formalin, embedded in paraffin, sectioned, and stained by hematoxylin and eosin. The statistical significance of the differences among different treatment groups was assessed using the two-tailed Mann–Whitney U test or Student t test.

The cisplatin-resistant HNC cells (HN3-cisR, HN4-cisR, and HN9-cisR: IC₅₀ = 29.7, 25.5, and 38.9 μmol/L, respectively) showed a 10- to 18-fold increase in the cisplatin IC₅₀ compared with the relative parental cells (HN3, HN4, and HN9; IC₅₀ = 3.0, 2.6, and 2.2 μmol/L, respectively; P < 0.001; Fig. 1A). The molecular weight of DPP-23 used in this study is 298.34 g/mol (Fig. 1B). The growth, viability, and colony-forming abilities of both cisplatin-sensitive and -resistant HNC cells were markedly inhibited by treatment with DPP-23 2.5 to 10 μmol/L for 72 hours (Figs. 1C and 2A). The DPP-23 IC₅₀ was 2.0, 2.3, and 2.4 μmol/L in cisplatin-sensitive HNC cells and 2.9, 3.2, and 3.5 μmol/L in cisplatin-resistant HNC ells (P < 0.05). The representative photographs of cisplatin-resistant HNC cells treated with cisplatin or DPP-23 was shown in Supplementary Fig. S1. The cell cycles were changed in the HNC cells, with a decreased G₂–M phase and increased sub-G₁ apoptotic population (Fig. 2B). DPP-23 also induced significant cell death in cisplatin-resistant HNC cells after DPP-23 treatment (Fig. 2C). The cell growth inhibition, cell-cycle change, and cell death were blocked by coincubation of the HNC cells with the antioxidant NAC or with the pan-caspase inhibitor zVAD-fmk (Fig. 2D).

DPP-23 selectively generates ROS in cancer cells (23). GSH and GSSG levels were measured in HN3-cisR and HN9-cisR cells and in normal oral keratinocytes (HOK). DPP-23 depleted cellular GSH levels and increased total GSSG levels in HNC cells (Fig. 3A and B). However, the GSH and GSSG levels were not significantly changed in HOK-1 cells treated with DPP-23. The drug treatment also induced intracellular accumulation of ROS in HNC cells but not in normal HOK cells (Fig. 3C). The changes in GSH, GSSG, and ROS levels were blocked by pretreatment of the cancer cells with NAC. When DPP-23 was administered to a panel of 13 different HNC and 5 normal cell lines, the drug was more sensitive in HNC cells than the normal cells (Supplementary Fig. S2).

Nrf2 and its antioxidant response element (ARE) proteins, NQO1 and HO-1, and an antiapoptotic protein, myeloid cell leukemia 1 (Mcl-1), were overexpressed in all three cisplatin-resistant HNC cell lines compared with the parental cells (Fig. 4A). DPP-23 significantly decreased the levels of Nrf2 and HO-1 proteins but increased the levels of p53 and the proapoptotic protein cleaved PARP in both cisplatin-sensitive and -resistant HNC cells (Fig. 4B). Interestingly, DPP-23 induced phospho-p53 expression and its related proteins (cleaved PARP and p21^WAF1) in the mutant p53-bearing HN3-cisR cell line (heterozygous R282W missense mutation) as well as the wild-type p53-bearing HN9-cisR cell line (Fig. 4C). DPP-23 also activated p53 and its related proteins in dose- and time-dependent manners (Fig. 4C and D).

Nrf2 overexpression provides an advantage for cancer cell growth (27). Transfection of Nrf2 plasmid induced Nrf2 accumulation in HN3-cisR cells. This effect was diminished by DPP-23 treatment, along with a decrease in cancer cell survival and an increase in the level of cleaved PARP and HO-1 proteins (Fig. 5A and B). However, Nrf2 overexpression itself was not affected by DPP-23 treatment compared with vector control (P > 0.1). TP53 and Nrf2 genes were silenced using p53 or Nrf2 siRNA transfection of HN9-cisR cells. Compared with control cells, the expression of p53-downstream proteins (cleaved PARP and p21^WAF1) decreased in the cancer cells with p53 knockdown but not in those with Nrf2 knockdown when the cells were treated with different concentrations of DPP-23 (Fig. 5C and Supplementary Fig. S3). PUMA knockdown blocked the apoptotic effect of DPP-23 treatment in the cancer cell population (Fig. 5D). In addition, p53 knockdown partly blocked the apoptotic effect of DPP-23 treatment and Nrf2 knockdown augmented the apoptotic effect (Supplementary Fig. S4). Taken together, these results showed that DPP-23 activates p53-induced cell death pathways by interfering with Nrf2 antioxidant defense mechanisms.

Up to 10 μmol/L cisplatin did not induce significant cytotoxicity of cisplatin-resistant HNC cells compared with their parental cells. However, DPP-23 alone induced a marked decrease in survival in cisplatin-resistant HNC cells. Moreover, the combination of DPP-23 and cisplatin reversed the cisplatin resistance in cisplatin-resistant HNC cells (the CIs for HN3-cisR = 0.61, 0.64, and 0.40; the CIs for HN9-cisR = 0.52, 0.54, and 0.58), which was blocked by pretreatment with NAC (Fig. 6A). DPP-23 also induced the expression of p53 and proapoptotic protein and inhibited Nrf2 expression (Fig. 6B). In combination, DPP-23 increased the cytotoxicity of cisplatin in HN9-cisR cells by exerting a synergistic effect on caspase activation (the CIs = 0.56, 0.29; Fig. 6C) and mitochondrial membrane potential (ΔΨm) reduction (the CIs = 0.46, 0.30; Fig. 6D). This effect was abrogated by pretreatment of the cells with the antioxidant NAC.

Preclinical studies were performed to determine the in vivo effects of DPP-23. BALB/c athymic nude mice were injected with HN9-cisR cells, followed by the intraperitoneal administration of cisplatin, DPP-23, their combination, or vehicle. Notably, the combination of cisplatin and DPP-23 synergistically suppressed in vivo tumor growth (Fig. 7A and B). Changes in daily oral food intake and body weight were not significantly different between the control and DPP-23–treated groups (P > 0.5; Supplementary Fig. S5). However, food intake and body weight significantly decreased in the cisplatin-treated group, which was blocked by the DPP-23/cisplatin combination (P < 0.01). In situ apoptosis assays were quantified and compared among the different treatment groups. TUNEL-positive apoptotic bodies were more frequently found in tumors treated with DPP-23 alone or cisplatin plus DPP-23 than in those treated with vehicle (Fig. 7C). Cleaved PARP and p53 protein expression increased in the tumor tissues of mice treated with DPP-23 alone or cisplatin plus DPP-23, whereas Nrf2 expression decreased in these groups compared with the control or cisplatin alone (Fig. 7D).

Blood analysis of experimental mice showed that the mean values for the red blood cell count, hematocrits, hemoglobin, white blood cells, lymphocyte, monocytes, and platelet count were not significantly different between the control and DPP-23–treated groups (P > 0.1). Histologic examination of vital organs did not reveal any significant differences between these groups (Supplementary Fig. S6).

Our current study shows that DPP-23 selectively induces ROS accumulation and cell death in cancer cells, including cisplatin-resistant HNC cells, by interfering with Nrf2 and activating cell death pathways associated with p53, and proapoptotic proteins (Fig. 8). DPP-23 significantly reversed the cisplatin resistance of HNC cells in vitro and in vivo. Cancer cells have a high demand for uncontrolled energy production to enable aberrant proliferation, resulting in the generation of higher ROS levels than normal cells (22). Cancer cells, particularly cancer stem cells, tend to survive in conditions of high ROS by increasing the expression of antioxidant systems (17, 28). High antioxidant defense mechanisms that regulate ROS might be an invaluable target that can be exploited to specifically kill cancer cells while sparing normal cells (18, 22). Much evidence has shown that the pharmacologic depletion of ROS scavengers leads to a marked decrease in cancer cell clonogenicity and results in chemo- or radiosensitization (17, 19, 20). Therefore, targeting of the ROS-scavenging systems in cancer cells in addition to other associated vulnerabilities may prove to be a promising future cancer treatment strategy. Our data also provide the first experimental evidence supporting the use of DPP-23 as a potential anticancer agent in HNC, particularly cancers with chemoresistant properties.

Our study indicates that DPP-23 exerts a selective cytotoxic effect in HNC cells by suppressing the Nrf2-mediated cellular defense response. Nrf2 controls cellular antioxidant systems responsible for GSH production (18). An oncogenic function of Nrf2 has previously been reported in esophageal, skin, and lung cancer (29). Growing evidence indicates that Nrf2 is frequently overexpressed in several types of cancers and is linked to increased resistance of cancer cells to chemotherapeutic agents and poor survival outcomes in cancer patients (30–32). Genetic or pharmacologic inhibition of Nrf2 results in a marked depletion of GSH and increases chemotherapeutic cytotoxicity (30, 31). These prior findings suggest that the Nrf2 pathway might also be a potential target of strategies that aim to control cancer cell resistance, despite its protective role in normal cells. Our current findings indicate that the activation of Nrf2–ARE pathways is also actively involved in cisplatin resistance and that pharmacologic inhibition of Nrf2 by DPP-23 leads to a cytotoxic effect in cisplatin-resistant HNC cells. Collectively, our results and those of previous studies suggest that the Nrf2 pathway is a potential target to control chemoresistance in cancer cells.

Our current findings indicate that DPP-23 activates p53 and its target pathways. There is positive and negative coregulation between ROS-induced activation of p53 and Nrf2 expression (33). At low levels of oxidative stress, p21, a p53 target gene, stabilizes Nrf2 by binding to KEAP1 and interfering its ability to promote Nrf2 degradation, resulting in upregulating the Nrf2-mediated antioxidant response (34). On the contrary, along with increasing oxidative stress, activated p53 counteracts Nrf2 expression and Nrf2-induced transcription of ARE and promotes the induction of apoptosis (35, 36). In addition, substantial generation of ROS by inhibition of the Nrf2–ARE antioxidant pathway results in increased ROS-dependent p53 and PUMA activation (37). Cellular generation of ROS triggers p53 activation, as well as mediating apoptosis as a p53 downstream factor (38). Interestingly, DPP-23 increased the levels of p53-Ser-15 and PUMA in both wild-type and mutant p53-harboring cancer cells, leading to both apoptotic and nonapoptotic cancer cell death (39). Therefore, DPP-23 is associated with restoration of p53 tumor suppression in mutant p53-bearing cancer cell. This might indicate widespread applications for DPP-23, regardless of p53 mutational status.

Our present study revealed that DPP-23 acts synergistically with cisplatin and thereby overcomes cisplatin resistance in HNC cells. Because cisplatin is currently a first-line agent for HNC chemotherapy, the use of DPP-23 in combination with cisplatin may be effective in a clinical setting to circumvent drug resistance and potentially reduce toxicity. Our current data show that DPP-23 is a promising candidate that could restore the cytotoxicity of cisplatin in drug-resistant HNC cells in vitro and in vivo. DPP-23 induced a robust increase in cisplatin-mediated apoptosis by activating proapoptotic pathways in HNC cells with acquired cisplatin resistance. DPP-23 sensitized chemoresistant HNC cells to cisplatin, leading to increased cytotoxicity and a more effective therapy for aggressive HNC. DPP-23 showed no significant side effects and, moreover, blocked the appetite and weight loss seen in mice treated with cisplatin alone. Collectively, these findings may be of paramount clinical significance due to the ability of DPP-23 to induce cell death in cisplatin-resistant cells as well as minimize the potential adverse effects of cisplatin by reducing its dose.

In conclusion, DPP-23 selectively induces ROS accumulation and cell death in cancer cells, including cisplatin-resistant HNC cells, by interfering with Nrf2 antioxidant defense mechanisms and activating p53 and proapoptotic proteins. DPP-23 significantly reverses the cisplatin resistance of HNC cells in vitro and in vivo and may be a promising anticancer candidate that could overcome cisplatin resistance in HNCs. Further preclinical and clinical investigation of DPP-23 should be performed in patients with various types of treatment-resistant cancers to explore this promising cancer therapy.

No potential conflicts of interest were disclosed.

Conception and design: E.H. Kim, H.J. Jang, J.-L. Roh

Development of methodology: E.H. Kim, H.J. Jang, J.-L. Roh

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.H. Kim, H.J. Jang, J.-L. Roh

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.H. Kim, H.J. Jang, J.-L. Roh

Writing, review, and/or revision of the manuscript: E.H. Kim, H.J. Jang, J.-L. Roh

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E.H. Kim, J.-L. Roh

Study supervision: J.-L. Roh

We thank Professor Yoongho Lim for his kind gift of DPP-23.

This study was supported by a grant (no. 2015R1A2A1A15054540) from the Basic Science Research Program through the National Research Foundation of Korea (NRF), Ministry of Science, ICT and Future Planning and a grant (no. HI15C2920) from the Korean Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), Ministry of Health & Welfare, Seoul, Republic of Korea (to J.L. Roh).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.