NCI-H1299 cells (CRL-5803, 70026320, human lung carcinoma) purchased from the American Type Culture Collection (ATCC) were cultured in RPMI-1640 medium (MilliporeSigma) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (Biosera), 100 units/ml penicillin and 100 mg/ml streptomycin (FUJIFILM Wako Pure Chemical Corporation) at 37°C in a humidified atmosphere containing 5% CO₂. A549 cells (RCB-0098, 010, human lung carcinoma, ATCC cat. no. CCL-185) purchased from RIKEN Cell Bank were cultured in the same manner as H1299 cells, except in DMEM (D5796, MilliporeSigma) instead of RPMI-1640 medium (MilliporeSigma). The authenticity of the cell lines was confirmed by ATCC or RCB. Mycoplasma negativity was confirmed using PCR for parental cells and VR cells prior to use. All cells were preserved with CELLBANKER 1 (Nippon Zenyaku Kogyo Co., Ltd.) in liquid nitrogen.

The cells were seeded into 6-well plates at 4×10⁴ cells per well and incubated for 24 h at 37°C in a humidified atmosphere containing 5% CO₂ with or without 500 nmol/l vinorelbine. ROS Assay Kit-Highly sensitive DCFH-DA (Dojindo Laboratories, Inc.) was used to detect ROS according to the manufacturer's protocol. Fluorescent images were obtained using a BZ-X810 fluorescence microscope (Keyence Corporation). The fluorescence intensity was determined using ImageJ 1.53 software (National Institute of Health).

Vinorelbine (FUJIFILM Wako Pure Chemical Corporation) was dissolved in dimethyl sulfoxide (DMSO; Nacalai Tesque, Inc.). VR cells were cultured with the indicated concentrations from 5 to 500 nmol/l of vinorelbine for >4 months. Bardoxolone (Selleck Chemicals), a Keap1 inhibitor, was dissolved in DMSO. Bardoxolone prevents the ubiquitination and degradation of Nrf2 and, therefore, functions as an Nrf2 activator. For the inhibition of Keap1, the cells were exposed to the indicated concentrations from 0.1 to 2.0 µmol/l of bardoxolone for 24 h until they were harvested for lysate preparation using the same procedure as described below in ‘Western blot analysis’.

The passage number of each cell line was <50. The H1299 cells were seeded at 1×10⁶ cells per 10-cm dish and cultured at 37°C for 96 h until just before confluency. Whole-cell lysates were then fractionated into cytoplasmic, membrane and nuclear protein extracts using the Subcellular Protein Fractionation kit for Cultured Cells (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. To confirm that the nuclear and cytoplasmic fractions predominantly consisted of proteins derived from the nuclear and cytoplasmic compartments, respectively, the extracts were probed with nucleus-specific anti-lamin B1 antibody (cat. no. 12586, Cell Signaling Technology, Inc.) and cytoplasm-specific anti-α-tubulin antibody (017-25031, FUJIFILM Wako Pure Chemical Corporation).

To prepare total cell lysates, the cells were lysed in radioimmunoprecipitation assay buffer (20 mM Tris, pH 7.4, 150 mM NaCl, 1% Nonidet p-40, 1% deoxycholic acid and 0.1% sodium dodecyl sulfate) supplemented with protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 mM sodium vanadate, 1 ng/ml aprotinin, 1 ng/ml leupeptin and 1 ng/ml pepstatin A). Protein concentrations were determined using Bio-Rad Protein Assay Dye Reagent Concentrate (#5000006JA; Bio-Rad Laboratories, Inc.). The primary antibodies used were as follows: Anti-ABCB1 (1:1,000; cat. no. 13978; Cell Signaling Technology), anti-Fyn (1:1,000; cat. no. A0086, ABclonal Biotech Co., Ltd.), anti-phosphorylated Fyn (p-Fyn) (1:1,000; cat. no. AP0510, ABclonal Biotech Co., Ltd.), anti-Nrf2 (1:1,000; cat. no. sc-722, Santa Cruz Biotechnology, Inc.), anti-caspase-3 (1:1,000; cat. no. 9665, Cell Signaling Technology, Inc.), anti-cleaved caspase-3 (1:500; cat. no. 9664, Cell Signaling Technology, Inc.), anti-Bcl-2 (1:1,000; sc-7382, Santa Cruz Biotechnology, Inc.), anti-MRP1/ABCC1 (1:1,000; cat. no. 72202, Cell Signaling Technology, Inc.), anti-MRP2/ABCC2 (1:1,000; cat. no. 12559, Cell Signaling Technology, Inc.), anti-MRP3/ABCC3 (1:1,000; cat. no. 14182, Cell Signaling Technology, Inc.) and anti-β-actin (1:1,000; cat. no. A5441, MilliporeSigma). Gel electrophoresis was performed using PowerPac^™ HC High-Current Power Supply (1645052; Bio-Rad), and 10% gel was used for all antibodies except for cleaved caspase-3 and 15% gel was used for cleaved caspase-3. The polyvinylidene difluoride membranes (IPVH00010; Merck Millipore Ltd.) were incubated with 2% bovine serum albumin (A9647-100G; MilliporeSigma) for 60 min at room temperature (20–22°C) for blocking. Following incubation with the primary antibodies overnight at 4°C, the polyvinylidene difluoride membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (1:10,000; 711-035-152; donkey anti-rabbit IgG, 715-035-150; donkey anti-mouse IgG, Jackson ImmunoResearch Laboratories, Inc.) for 60 min at room temperature (20–22°C). The protein bands were visualized using an Ez West Lumi Plus detection kit (ATTO Corporation) and the LuminoGraph II imaging system (ATTO Corporation). The protein expression was quantified using ImageJ 1.53 software (National Institute of Health).

The cells were seeded at 2×10⁴ cells/well of a 2-chamber slide and were then fixed in 4% paraformaldehyde for 10 min at 37°C, permeabilized with 0.1% Triton X-100 for 15 min at room temperature (20–22°C) and blocked in 2% bovine serum albumin (Biosera) in phosphate-buffered saline for 60 min at room temperature (20–22°C). Antigen recognition was performed by incubation with primary antibodies against p-Fyn (1:400; cat. no. AP0510, ABclonal Biotech Co., Ltd.) and integrin β3 (1:200; cat. no. 336402, BioLegend, Inc.) overnight at 4°C, followed by incubation with Alexa Fluor 568- (1:600; cat. no. ab175471, Abcam) and 488- (1:600; A-21441, Invitrogen; Thermo Fisher Scientific, Inc.) conjugated secondary antibodies. Nuclei were counterstained with Invitrogen Prolong Gold Antifade Mountant with DAPI (P36931; Invitrogen; Thermo Fisher Scientific, Inc.). Fluorescent images were obtained using a TCS SP8 microscope (Leica Microsystems GmbH). Co-localization analysis was performed using ImageJ 1.53 software (National Institute of Health). Pearson's R value was determined as follows:

where Xi or Yi is the individual intensity in the pixel indexed with i, and X¯ or Y¯ is the mean intensity. R value ranges between −1 and 1 continuously, and 0 < R < 1 and −1 < R < 0 suggest correlation and anti-correlation, respectively.

The cells were seeded in 96-well microplates at 5×10³ cells/well and cultured at 37°C for 48 h. The cells were then incubated with increasing concentrations of vinorelbine from 0.1 nmol/l to 50 µmol/l for 96 h. The cells were additionally incubated with 10 µl Cell Counting Kit-8 reagent (Dojindo Laboratories, Inc.) for 2 h, and the absorbance at the 450 nm wavelength was measured using Microplate Manager 6 (Bio-Rad Laboratories, Inc.). The half-maximal inhibitory concentration (IC₅₀) was calculated using Prism 7 software (GraphPad Software, Inc.) with a three-parameter sigmoidal curve fit.

The cells were seeded in 96-well microplates at 5×10³ cells/well and cultured at 37°C for 48 h. The cells were additionally incubated with 10 µl Cell Counting Kit-8 reagent ((Dojindo Laboratories, Inc.) for 1 h, and the absorbance at 450 nm wavelength was measured using Microplate Manager 6 (Bio-Rad Laboratories, Inc.). The ratio of cell proliferation was determined by comparing the measured absorbance.

Ambion's Silencer^™ Select Pre-Designed siRNAs (ID nos. s9491, 9492 and 9493) targeting Nrf2 were purchased from Thermo Fisher Scientific, Inc. The sequences of the siRNAs used are presented in Table SI. The cells were seeded at ~60% confluency in 6-cm dishes with antibiotic-free medium containing 12.5 µl Lipofectamine 2000^® (cat. no. 11668019; Thermo Fisher Scientific, Inc.) and 1,000 µl Opti-MEM (cat. no. 31985062; Thermo Fisher Scientific, Inc.), which gave a final siRNA concentration of 20 nmol/l, transfected for 48 h, and then harvested for lysate preparation.

In total, 104 surgical specimens were obtained from patients who underwent lung cancer resection followed by adjuvant cisplatin-vinorelbine treatment for pathological stage II to IIIA (UICC, 7th edition) (22) between December, 2006 and June, 2018 at the Department of Thoracic Surgery, Kyoto University Hospital (Kyoto, Japan). The follow-up period ranged from 3 to 159 months (median, 40 months). The characteristics of the 104 patients are presented in Table SII. Disease-free survival and overall survival were available for all patients. The study was conducted in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans and was approved by the Ethics Committee of the Graduate School of Medicine, Kyoto University (approval no. G0028-7, R1706, R1486). Informed consent was obtained from all patients.

Immunohistochemistry was performed to evaluate the expression of Nrf2, ABCB1 and p-Fyn in formalin-fixed, paraffin-embedded tissue sections from 3 to 4 µm in thickness of the 104 surgical samples mentioned above. The slides were stained using the VECTASTAIN Elite ABC HRP kit (Vector Laboratories, Inc.) according to the manufacturer's protocol. Diaminobenzidine (049-22831, FUJIFILM Wako Pure Chemical Corporation) was used as the chromogen.

The log-rank test was used to compare two Kaplan-Meier survival curves, and statistical analyses were performed using Prism 7 (GraphPad) and JMP12 software (SAS Institute Inc.). Fisher's exact test was used to compare the patient clinicopathological characteristics. An unpaired t-test with Welch's correction was used to compare the fluorescence strength. A value of P<0.05 was considered to indicate a statistically significant difference.

VR cell lines were established by exposing H1299 cells to increasing concentrations of vinorelbine from 5 to 500 nmol/l. These cells were classified into three groups according to the resistance level as follows: VR5, resistant cells cultured at 5 nmol/l; VR50, resistant cells cultured at 50 nmol/l; and VR500, resistant cells cultured at 500 nmol/l (Fig. 1A). According to the datasheet of Navelbine^® Injection (Kyowa Kirin Co., Ltd.), 5 nmol/l corresponds to the plasma concentration of vinorelbine at 24–48 h following the intravenous injection of 20 mg/m² vinorelbine in humans. The vinorelbine IC₅₀ values for the H1299 parental, VR5, VR50 and VR500 cells were 11.3, 43.9, 99.9 and 137.3 nmol/l, respectively (Fig. 1B). In Fig. S1, the ratio of H1299 VR500 over H1299 parental cells in number were plotted at 2, 3, 4 and 5 days after seeding in 96-well microplates at 5×10³ cells/well using cell proliferation assay. The ratio of cells declined as the time passed to day 5, and this result suggested that the H1299 VR500 cells were less proliferative than the H1299 parental cells. To evaluate the apoptosis of the H1299 cells treated with vinorelbine, lysates obtained from H1299 parental cells treated with or without 500 nmol/l vinorelbine and the VR500 cells were immunoblotted with anti-caspase-3, anti-cleaved caspase-3 and anti-Bcl-2 antibodies. Western blot analysis revealed that the expression of cleaved caspase-3 was augmented in the H1299 parental cells treated with 500 nmol/l vinorelbine, and conversely, Bcl-2 expression was increased in the VR500 cells (Fig. S2). The A549 VR500 cells were established in the same manner as the H1299 VR500 cells.

The H1299 parental cells and VR500 cells were uniformly seeded into 6-well plates at 4×10⁴ cells/well. The parental cells were incubated for 24 h with or without 500 nmol/l vinorelbine, and the VR500 cells were incubated for 24 h with 500 nmol/l vinorelbine as above. Intracellular ROS formation was detected in both parental cells and VR500 cells treated with 500 nmol/l vinorelbine (Fig. 1C). To quantify the strength of ROS formation, the area of fluorescence per cell and the mean intensity of fluorescence in the parental cells and VR500 cells were calculated. The area of fluorescence per cell in the parental and VR500 cells was 449.5±480.8 and 1,474±1,731 µm² (P<0.0001). The mean fluorescence intensity in the parental cells and VR500 cells was 51.47±10.84 and 71.83±4.674 µm² (P<0.0001; Fig. 1D).

The total expression of ABCB1, Nrf2 and p-Fyn was augmented in the crude lysate of the VR cells (Fig. 2A), indicating that Nrf2 contributes to vinorelbine resistance, in addition to ABCB1 and p-Fyn. In the MRP family, MRP3/ABCC3 expression was augmented in a manner similar to ABCB1 (Fig. S3). In the A549 VR500 cells, ABCB1 expression was augmented, whereas Nrf2 and p-Fyn expression was not upregulated (Fig. S4). Given that the diverse activities of SFKs, including Fyn, are dependent on their subcellular localization (23,24), lysates obtained from the subcellular extracts of H1299 parental, VR5, VR50 and VR500 cells were immunoblotted with anti-ABCB1, anti-Fyn, anti-p-Fyn and anti-Nrf2 antibodies to investigate the subcellular localization and expression of ABCB1, p-Fyn and Nrf2 in the VR cell lines. ABCB1 expression was observed in the membrane extracts and compared with that in the nuclear and cytoplasmic extracts (Fig. 2B). The VR cells exhibited the nuclear and cytoplasmic localization of p-Fyn.

Immunofluorescence was used to confirm the precise subcellular localization of p-Fyn downstream of integrin β3. The VR5 cells were incubated with primary antibodies against p-Fyn and integrin β3. The fluorescence images revealed that p-Fyn was located in the nucleus, as well as the cytoplasm adjacent to integrin β3 in the membrane (Fig. 2C). Co-localization analysis demonstrated a positive linear association between integrin β3 and p-Fyn (Fig. 2D). These findings suggest that p-Fyn functions in both the cytoplasm downstream of integrin β3 and the nucleus.

To elucidate whether the decreased activity of Nrf2 downregulates ABCB1 expression in VR cells, knockdown experiments were performed using siRNA. The VR500 cells in which Nrf2 was overexpressed, as described above in the paragraph entitled ‘ABCB1 overexpression, and p-Fyn and Nrf2 nuclear accumulation in VR cells’ were transfected with siRNAs targeting Nrf2 and incubated for 48 h. Scrambled RNA was used as a negative control. In the VR500 cells, siRNAs with different sequences targeting Nrf2 similarly decreased Nrf2 expression and downregulated ABCB1 expression (Fig. 3A).

To examine whether activated Nrf2 upregulates ABCB1 expression, Keap1 was inhibited using bardoxolone. The VR5 cells in which the Nrf2 level was lower than that in the VR500 cells were treated with the indicated concentrations of bardoxolone for 24 h. Bardoxolone treatment increased the expression of both Nrf2 and ABCB1 in a concentration-dependent manner up to 0.8 µmol/l (Figs. 3B and S5).

As Nrf2 regulates the expression and functions of ABCB1 (21), the present study analyzed the drug resistance profile of VR cells to examine whether Nrf2 contributes to vinorelbine resistance. Nrf2 knockdown lowered the vinorelbine IC₅₀ value in VR50 cells from 119.8 to 36.6 nmol/l, and restored vinorelbine sensitivity (P<0.001; Fig. 3C). By contrast, bardoxolone exposure for 96 h at 0.8 µmol/l increased the vinorelbine IC₅₀ value in the VR5 cells from 43.9 to 111.8 nmol/l and contributed to vinorelbine resistance (P<0.001; Fig. 3D).

To elucidate whether the expression of Nrf2, ABCB1 and p-Fyn is associated with the prognosis of patients with completely resected lung cancer, 104 tissue samples obtained from patients who underwent adjuvant cisplatin-vinorelbine treatment were stained immunohistochemically. The survival curves and clinicopathological data of the patients are presented in Fig. 4 and Table SII, respectively. The clinical samples were stained for Nrf2, ABCB1 and p-Fyn; representative staining patterns are illustrated in Fig. 5A-F. Nrf2, ABCB1 and p-Fyn were predominantly expressed in the nucleus, cytoplasm and nucleus, respectively. This finding is compatible with the subcellular localization of each protein shown in Fig. 2B. Of the 104 patients, 47 (45%) were positive for Nrf2, 53 (50%) for ABCB1, and 48 (46%) for p-Fyn (Tables I and SII). Nrf2 and ABCB1 were significantly associated with disease-free survival (P=0.029 and 0.035, respectively), and p-Fyn was associated with overall survival (P=0.040) (Fig. 4). The expression of Nrf2 was significantly associated with that of p-Fyn (P=0.005), but not with that of ABCB1 (P=0.244). The frequency of positive Nrf2 expression was significantly associated with squamous cell carcinoma (P=0.022), but not with adenocarcinoma (P=0.097) or other cancers (P=0.726) (Table I).

In the present study, Nrf2 expression was augmented in the nuclei, as well as the whole cell of VR cells, and intracellular ROS formation was detected in both parental and VR cells exposed to vinorelbine. Notably, VR cells, which had been cultured with vinorelbine for months and acquired resistance to vinorelbine, exhibited higher levels of ROS than the parental cells treated with vinorelbine for 24 h. The formation of excessive ROS is essential for the induction of apoptosis by commonly used chemotherapeutic agents such as cisplatin, bleomycin, paclitaxel, adriamycin, etoposide and vinca alkaloids (15). Vinorelbine, a semi-synthetic vinca alkaloid, arrests cell growth at the prometaphase by inhibiting microtubule polymerization and induces the accumulation of mitochondrial ROS followed by prolonged JNK activation, DNA damage, a decrease in Mcl-1 expression, mitochondrial dysfunction and caspase-mediated apoptosis (14). Under conditions of oxidative stress, increased levels of intracellular ROS promote the dissociation of Nrf2 and Keap1. Free Nrf2 translocates to the nucleus, where it binds to ARE and transactivates downstream cytoprotective genes to induce cell defense processes and enhance cell resistance (15,25). Oxidants, xenobiotics and electrophiles, including chemotherapeutic agents hamper the Keap1-mediated proteasomal degradation of Nrf2 and induce the transcription of target genes (26). The results of the present study suggested that persistent exposure to vinorelbine induced intracellular ROS in VR cells, which may subsequently stimulate Nrf2.

Whole-cell extracts obtained from VR cells exhibited an increased expression level of ABCB1 and Nrf2. The level of MRP3/ABCC3, which belongs to ABC sub-family C, was also augmented in VR cells, indicating that MRPs, as well as ABCB1 were associated with vinorelbine resistance. In a previous study, the authors compared the gene and protein expression of H1299 parental cells with that of VR cells using DNA microarray, and this microarray-based comparison did not reveal any specific change in MRP expression, but revealed a 10-fold or more change in ABCB1 expression (9). Therefore, the present study focused on a mechanism involving ABCB1 in vinorelbine resistance. Nrf2, a key transcription factor, protects cells from oxidative damage and harmful xenobiotics (27). Under unstimulated conditions, Nrf2 is retained in the cytoplasm by the anchor protein, Keap1, and is constantly ubiquitinated and degraded in the proteasome (20). Upon exposure to oxidative or xenobiotic stress, the Keap1-mediated proteasomal degradation of Nrf2 is inhibited, and Nrf2 translocates to the nucleus, where it forms a heterodimer with the small Maf protein and binds to ARE (28). The Nrf2 effector genes bearing ARE include those that encode the majority of antioxidant and phase II detoxifying enzymes (29). In addition to these enzyme-coding genes, Nrf2 transactivates a wide variety of other genes, including several ATP-dependent drug efflux pumps (30,31). This suggests that activated Nrf2 in cancer cells provides advantages in terms of survival and drug resistance (32). Nrf2 expression enhances the resistance of myelogenous leukemia and colorectal, breast, pancreatic, and gallbladder cancers to chemotherapeutic agents, such as imatinib, oxaliplatin, tamoxifen, gemcitabine and 5-fluorouracil, respectively (33–37). The results of the present study indicated that ABCB1 and Nrf2 are both involved in vinorelbine resistance, and that activated Nrf2 upregulates ABCB1 expression. These findings suggest that ABCB1 is dependent on Nrf2 and its downstream activity in vinorelbine resistance.

The results of the present study revealed that the suppression of Nrf2 potentiated the cytotoxicity of vinorelbine, and the upregulation of Nrf2 conferred resistance to vinorelbine. These findings suggested that Nrf2 plays a crucial role in regulating the susceptibility to vinorelbine, and that the upregulation of Nrf2 in VR cells confers vinorelbine resistance. Thus, lowering the expression of Nrf2 may present a novel therapeutic approach for VR lung cancer.

p-Fyn accumulated in the nuclei of VR cells and was located in the cytoplasm adjacent to integrin β3 in the membrane. In the focal adhesion pathway, Fyn functions downstream of several important cell surface receptors, including integrin β3, and upstream of several cellular signals important for cancer progression. Although the integrin cytoplasmic domains are short and do not have any known catalytic activity, the engagement of integrins by extracellular matrix ligands triggers outside-in signals that collaborate with growth factor-initiated signals to determine cell fate and function (38,39). Deregulated integrin signaling empowers cancer cells with the ability to proliferate without restraint, invade through tissue boundaries, and survive in foreign microenvironments (40). SFKs bind to multiple integrin β cytoplasmic domains to transmit integrin-dependent signals pivotal for cell movement and proliferation; in particular, Fyn selectively binds to the integrin β3 domain (38). Similar to other Src family members, Fyn regulates cell shape and migration (12) and is located primarily in the cytoplasm, although it is also observed in other cellular compartments, including the nucleus (41). Fyn activity is regulated in different subcellular compartments, and different equilibrium states between Fyn and the corresponding kinase are maintained in the cytoplasm and nucleus (23). Resting-state Fyn is localized near the perinuclear region in endosomes, whereas activated Fyn is trafficked to the plasma membrane via the actin cytoskeleton (42,43). The nuclear expression of p-Fyn is highly associated with the poor prognosis of patients with resected lung adenocarcinoma (44). Here, immunofluorescence images revealed the nuclear accumulation of p-Fyn in VR cells. These results are consistent with those of previous studies by the authors (9,44). The nuclear or cytoplasmic localization of Fyn in breast cancer cells has been shown to be associated with early recurrence in patients treated with endocrine therapy (13,41). However, the role of nuclear p-Fyn in vinorelbine resistance remains unclear.

In the present study, subcellular fractionation and immunofluorescence demonstrated that p-Fyn accumulated in the nucleus, as well as the cytoplasm. The regulation of Nrf2, particularly its abundance in the nucleus, is important for controlling the expression of cell-protective genes in response to oxidative stress. The continuous accumulation of Nrf2 in the nucleus can cause disease conditions (45). For example, in Keap1-deficient mice, the persistent accumulation of Nrf2 in the nucleus causes hyperkeratosis in the esophagus and forestomach, leading to neonatal death (45). As a persistent increase in the expression of cell-protective genes threatens cell survival (46), aggregated Nrf2 should be subsequently exported out of the nucleus and degraded.

The abundance of Nrf2 inside the nucleus is tightly controlled by positive and negative regulators that affect nuclear import, ARE binding, nuclear export, and degradation of Nrf2 under normal and stressful conditions (47). The nuclear export of Nrf2 is activated after Fyn accumulates in the nucleus (48). Fyn is responsive to oxidative stress (47). It is exported out of the nucleus soon after exposure to oxidative stress, which allows Nrf2 to bind to ARE and activate NAD(P)H:quinone oxidoreductase 1. The nuclear export of Fyn is an integral part of the ARE/Nrf2-mediated activation of cytoprotective genes (47). Chemical stress induces activated GSK-3β, which phosphorylates Fyn and accumulates p-Fyn in the nucleus, resulting in the ubiquitination and degradation of nuclear Nrf2 (48). The results of the present study suggested that the nuclear accumulation of p-Fyn following Nrf2 activation in VR cells triggers the nuclear export and degradation of Nrf2.

The immunohistochemical findings demonstrated that Nrf2 and ABCB1 were significantly associated with disease-free survival. Previous studies have demonstrated an association between ABCB1 expression and chemotherapy resistance in colorectal cancer (49), breast cancer (50) and chronic myelogenous leukemia (51). Nrf2 expression significantly promotes tumor size, histological grade, distant metastasis and lymph node metastasis, and reduces sensitivity to chemotherapy or radiotherapy (8,17,52). A positive Nrf2 expression in the nucleus is associated with a poor prognosis of patients with esophageal squamous cell carcinoma following chemoradiotherapy (53). The nuclear Nrf2 expression in malignant lung cancer cells is related to resistance to chemotherapy in squamous cell carcinoma (54). Consistent with these reports, the results of the present study revealed that Nrf2 expression was significantly associated with squamous cell carcinoma, implying that squamous cell carcinoma is a predictor of vinorelbine resistance. It was also found that Nrf2 and ABCB1 were closely associated with a poor susceptibility to vinorelbine and subsequent tumor relapse, and p-Fyn was not associated with disease-free survival, but with overall survival. These findings reveal that vinorelbine resistance is not a direct cause of the nuclear accumulation of p-Fyn, but rather the overall prognosis, which emerges from Nrf2 activation in the nucleus.

The results of the present study support the following model depicting the role of Nrf2 and ABCB1 in vinorelbine resistance (Fig. 5G). In the absence of stress, Nrf2 is bound to Keap1 and degraded through a proteasome-dependent pathway, whereas the presence of ROS induced by vinorelbine hinders the Keap1-mediated proteasomal degradation of Nrf2. Nrf2 released from Keap1 is translocated into the nucleus and activates the transcription of a broad spectrum of defensive genes, including ABCB1. The increase in the expression of chemoprotective genes neutralizes chemical stress and confers vinorelbine resistance to the cells, which consequently promotes tumor recurrence. As a persistent increase in defensive gene expression threatens cell survival, Nrf2 is exported out of the nucleus and degraded only after p-Fyn accumulates in the nucleus. The nuclear accumulation of p-Fyn itself is a predictor of a poor overall survival of patients with lung cancer.

The present study has the following limitations: The sample size was small, as only datasets of patients with resected stage II or IIIA lung cancer and adjuvant cisplatin-vinorelbine treatment at a single institute were analyzed. In addition, the present study was retrospective in nature. Further studies are thus required to develop improved strategies for VR non-small cell lung cancer.

In conclusion, the Nrf2-pFyn-ABCB1 axis plays a pivotal role in vinorelbine resistance in non-small cell lung cancer. Nrf2 upregulates ABCB1 and induces vinorelbine resistance, and Nrf2 activation causes the nuclear accumulation of p-Fyn, which triggers Nrf2 export out of the nucleus and subsequent degradation. The present study revealed the possible pathway underlying vinorelbine resistance in non-small cell lung cancer. It is hoped that these findings will help researchers to develop strategies with which to avoid this resistance or develop novel drug targets.