dCK negatively regulates the NRF2/ARE axis and ROS production in pancreatic cancer

Abstract Objectives Decreased deoxycytidine kinase (dCK) expression is a reported indicator of gemcitabine efficacy in pancreatic cancer, due to the impact of this kinase on gemcitabine metabolism. The transcription factor NF‐E2 p45‐related factor 2 (NRF2, also called Nfe2l2), a master regulator of redox homoeostasis, has been reported to tightly control the expression of numerous ROS‐detoxification genes and participates in drug resistance. However, the contribution of dCK to the NRF2 signalling axis has seldom been discussed and needs investigation. Materials and methods By overexpressing dCK in pancreatic cancer cells, we assessed the impact of dCK on NRF2 transcriptional activity. Furthermore, we measured the impact of dCK expression on the intracellular redox balance and reactive oxygen species (ROS) production. By utilizing immunohistochemical staining and tissues from pancreatic cancer patients, we assessed the correlation between dCK and NRF2 expression. Through proliferation and metastasis assays, we examined the impact of dCK expression on cell proliferation and metastasis. Results dCK negatively regulates NRF2 transcriptional activity, leading to the decreased expression of ARE‐driven antioxidant genes. In addition, dCK negatively regulates intracellular redox homoeostasis and ROS production. Negative correlations between dCK and NRF2 levels in pancreatic cancer cell lines and patient samples were observed. In vitro cell line studies suggested that dCK negatively regulated proliferation and metastasis. Conclusion Decreased dCK expression promotes NRF2‐driven antioxidant transcription, which further enhances gemcitabine treatment resistance, forming a feedback loop.


| INTRODUC TI ON
Despite a low incidence rate, pancreatic cancer remains the fourth leading cause of cancer-related deaths and is regarded as one of the most malignant and lethal cancer types. 1,2 Significant progress has been made in the past few decades in solid cancer screening and treatment, which has greatly increased patient chances for a cure.
Despite the tremendous progress in pancreatic cancer research, the ratio of mortality to incidence has changed little, and the 5-year survival rate remains desperately low at approximately 5%-7%. 3,4 Surgical resection is considered the only curative treatment for pancreatic cancer. However, most patients have distal organ metastasis at diagnosis, and approximately only 20% of patients have the chance to undergo surgical resection. Thus, chemotherapy treatment or chemotherapy in combination with radiotherapy remains the main option for patients with advanced and metastatic pancreatic cancer. 5,6 Despite considerable toxicity, 5-fluorouracil (5-FU) and its analogs, or combinations thereof, have been widely used for the treatment of advanced pancreatic cancer but are moderately effective at improving a patient's life. 7 The anti-cancer agent gemcitabine (2′, 2′-difluorodeoxycytidine, Gemzar, Eli-Lilly, Indianapolis, IN) is a cell cycle-dependent deoxycytidine analog of the antimetabolite class. Since 1997, gemcitabine has been accepted as a reference first-line therapy drug for patients with a good performance status. 8 Since then, combinational trials with gemcitabine have been conducted and reported. These combinations included cytotoxic agents (5-FU, cisplatin, oxaliplatin and capecitabine) and biological agents (erlotinib, Cetuximab and bevacizumab). Although higher clinical benefits and relatively longer survival have been achieved, none of these combination regimens have been proven to be significantly more effective than gemcitabine alone as the first-line therapy. The overall survival rate remains unchanged. 9 Gemcitabine has modest clinical benefits and might not improve overall survival to a clinically significant degree due to the inherent chemoresistance of pancreatic cancer cells and the impaired drug delivery system. 10 Thus, a better understanding of the molecular mechanisms underlying drug resistance in pancreatic cancer is necessary for developing new effective treatments for this lethal disease.
Gemcitabine is a proto-drug and needs to be taken up and catalysed by a series of enzymes to form the active drug. Gemcitabine is strongly hydrophilic and efficient gemcitabine cell permeation requires specialized integral membrane transport proteins. The major mediators of gemcitabine trafficking are the human equilibrative nucleoside transport (hENT1) and, to a lesser degree, the human concentrative nucleoside transport 3 (hCNT3). [11][12][13] As a proto-drug, intracellular gemcitabine must be phosphorylated into its mononucleotide form by deoxycytidine kinase (dCK) for subsequent metabolism. This step is the rate-limiting step of gemcitabine metabolism. Subsequent nucleotide kinases convert gemcitabine monophosphate to its active metabolites: gemcitabine diphosphate and gemcitabine triphosphate. 14,15 Gemcitabine exerts its cytotoxicity by blocking de novo DNA synthesis through inhibiting ribonucleotide reductase, which is required for the production of the deoxyribonucleotide precursors needed for DNA synthesis.
Ribonucleotide reductase contains a larger subunit, ribonucleotide reductase subunit (RRM)1, and a smaller one, RRM2, that are inactivated by difluorodeoxycytidine-5-phosphate. 16,17 The triphosphorylated form of gemcitabine is incorporated into DNA and leads to chain termination during DNA synthesis. hENT1, dCK and RRM1 are important determinants of gemcitabine activity and gemcitabinebased chemotherapy efficacy. 18 Living cells operate optimally within certain pH and temperature ranges; furthermore, the biochemical and physiological processes within a living cell also require an optimal redox balance for the sufficient flux of metabolic processes. The ability of a living cell to adapt rapidly to redox homoeostasis perturbations is essential for survival.
Cancerous cells are continuously threatened by ROS and by toxic secondary metabolites generated from ROS-mediated cell damage, leading to oxidative stress. NRF2 acts as one of the most versatile mechanisms for adapting to cellular oxidative stress and regulates redox homoeostasis to provide proliferative and progressive advantages to cancerous cells. 19,20 NRF2 plays vital and decisive roles in pancreatic cancer oncogenesis. In a transgenic K-Ras knock-in mouse pancreatic ductal adenocarcinoma (PDAC) model with NRF2 simultaneously deleted, pancreatic intraepithelial neoplasia (PanIN), cell proliferation and the tumour burden were reduced. 21 NRF2 also sustains metabolic reprogramming in cancerous cells. For example, in non-small cell lung cancer, NRF2 has been reported to regulate serine biosynthesis, providing a growth advantage to cancerous cells. 22 Highly proliferative cancerous cells require a large quantity of nutrients to maintain high anabolism levels. NRF2 has been reported to be a decisive regulator, redirecting glucose and glutamine anabolism into anabolic pathways, especially under sustained PI3K-Akt signalling pathway activation, which increases nuclear NRF2 accumulation and NRF2/ARE signalling. 23 NRF2 overexpression in pancreatic cancer has also been reported to participate in gemcitabine resistance, and inhibiting NRF2 expression and NRF2 transcriptional targets has been reported to improve gemcitabine sensitivity in pancreatic cancer cells. 24,25 However, the impact of gemcitabine metabolic regulators on the NRF2 signalling pathway has seldom been discussed.

| Cell culture
The human pancreatic cancer cell lines PANC-1 and MIA PaCa-2 were obtained from the American Type Culture Collection (ATCC, USA). The cells were cultured according to standard protocols provided by ATCC. In brief, PANC-1 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% foetal bovine serum (FBS), 100 U/mL penicillin and 0.1 mg/mL streptomycin. For MIA PaCa-2 cells, an additional 2.5% horse serum was used in the culture. These cells were maintained in a humidified incubator at 37°C with 5% CO 2 .

| Establishment of dCK-overexpressing cell lines
To overexpress dCK in PANC-1 and MIA PaCa-2 cells, a lentivirus- Stable dCK-overexpressing cell lines were obtained by infecting PANC-1 and MIA PaCa-2 cells and subsequent puromycin selection.

| Cell viability assay
Cell Counting Kit-8 (Dojindo, Japan) was used to measure cell viability. Briefly, 200 μL of medium containing cells (3000/well) was added to 96-well plates. After culturing for the indicated times, CCK-8 solution was added into each well and incubated at 37°C. After 2 hour, the optical density at 450 nm of each well was measured using a microplate reader.

| Cell apoptosis analysis
Flow cytometric techniques were used to measure cell apopto-

sis. The percentage of apoptotic cells was analysed by fluorescein
isothiocyanate-conjugated Annexin V and propidium iodide (Invitrogen, Carlsbad, CA, USA) staining, followed by flow cytometric analysis.

| Quantitative real-time PCR
Total RNA was extracted using TRIzol reagent (Invitrogen, USA). A TaKaRa PrimeScript RT reagent kit was used for reverse transcription to obtain cDNA (TaKaRa, Japan). The expression status of candidate genes and β-actin was determined by quantitative real-time PCR using an ABI 7900HT real-Time PCR system (Applied Biosystems, Foster City, CA, USA). The primer sequences are listed in Table 1.
The protein concentration of the whole cell lysate was measured using a Thermo Pierce BCA Protein Assay kit. Equal amounts of total protein were separated with SDS-PAGE and then transferred to PVDF membranes. Antibodies against dCK and NRF2 were purchased from Abcam. The Keap1 antibody was obtained from Proteintech.

| Transwell invasion assay
Invasion assays were conducted using a 24-transwell chamber with a Matrigel-coated membrane (BD, Franklin Lakes). The lower chamber was filled with 800 μL of media containing 10% FBS. Subsequently, approximately 1 × 10 5 cells were seeded in 200 μL of medium without serum in the top chamber for the invasion assays. For 24 hour, the cells were incubated at 37°C with 5% CO 2 and allowed to invade the lower chamber. After removing the non-migrating or non-invading cells, the remaining cells were washed, fixed and stained with crystal violet. We counted the number of migrating and invading cells in six fields randomly selected at 100× magnification. Experiments were performed at least in triplicate.

| Promoter activity assessment by dualluciferase assay
PANC-1 and MIA PaCa-2 cells were seeded in 96-well culture plates and transfected with the indicated vectors using Lipofectamine™ TA B L E 1 Primers sequences used in the text 2000 (Invitrogen). The antioxidant NRF2 activity response was assessed using pGMARE-lu firefly luciferase constructs (Genomeditech, China). The pRL-TK plasmid (Promega) was used as the internal control. Firefly and Renilla luciferase activities were measured using a dual-luciferase system (Promega) according to the manufacturer's protocol.
An immunohistochemical score >6 was defined as high expression, whereas a score ≤6 was considered a low expression level.

| ROS measurement and intracellular GSH activity assay
The

Intracellular GSH activity was determined by a GSH/GSSG Ratio
Detection Assay kit from Abcam to assess the oxidative status of the pancreatic cancer cells.

| Statistical analysis
All data are presented as the means ± SD; experiments were repeated at least three times. Two-tailed unpaired Student's t-tests and one-way analysis of variance were used to evaluate the data. SPSS version 16.0 (IBM) was used for the data analysis.

| dCK regulates Keap1/NRF2/ARE activation in pancreatic cancer
Decreased dCK expression has been reported to participate in gemcitabine resistance in pancreatic cancer, which is correlated with NRF2/ARE activation. However, the impact of dCK on NRF2/ARE activation has seldom been discussed. First, we overexpressed dCK in PANC-1 and MIA PaCa-2 cells, and the overexpression efficacy was validated by western blotting ( Figure 1A).
Then, we assessed the impact of dCK expression on intracellular ROS production. Through using a reactive oxygen species assay kit, we demonstrated that dCK overexpression decreased intracellular ROS levels in PANC-1 and MIA PaCa-2 cells ( Figure 1B).

Alterations in ROS levels can affect the intracellular redox
state, which can be evaluated by the GSH/GSSG ratio. In dCKoverexpressing PANC-1 and MIA PaCa-2 cells, the GSH/GSSG ratio was increased, indicating that dCK might cause a reduced intracellular environment ( Figure 1C). NRF2/ARE activation is regarded as a critical regulator of ROS production and redox status in cancer cells. Then, we examined the changes in Keap1 and NRF2 protein levels. As shown, the introduction of dCK increased the Keap1 protein level, while the NRF2 protein levels simultaneously decreased ( Figure 1D). NRF2 drives the transcription of a series of genes that participate in ROS detoxification, and the promoter of these genes contains AU-rich element (ARE) sequences. In dCKoverexpressing PANC-1 and MIA PaCa-2 cells, we observed a decrease in ARE-driven genes, such as GCLC, GLCM, ME1, NQO1, HMOX and TXNRD ( Figure 1E). Finally, we examined the impact of dCK on ARE-driven luciferase activity. As shown, dCK decreased ARE luciferase activity in PANC-1 and MIA PaCa-2 cells ( Figure 1F).

| dCK suppressed pancreatic cancer cell proliferation
On the basis of our observations of the negative correlation be- for dCK in pancreatic cancer proliferation ( Figure 2F). Finally, we analysed the potential pathways that participate in drug resistance and anti-apoptosis. Our results demonstrated that dCK overexpression inhibited ERK1/2 activation. In addition, the protein levels of Mcl1, a well-characterized anti-apoptotic factor that participates in chemotherapy and radiotherapy resistance, were also decreased ( Figure 2G).

| Decreased dCK expression and activation of the NRF2/ARE axis are observed in gemcitabine-resistant cells
As observed above, dCK regulates the Keap1/NRF2/ARE axis in pancreatic cancer cells. We propose that dCK expression might be negatively correlated with the NRF2/ARE axis in gemcitabine-resistant cells. In gemcitabine-resistant PANC-1 and MIA PaCa-2 cells, we observed a decrease in dCK mRNA and protein levels ( Figure 3A,B). Then,

we measured the intracellular ROS levels and observed an increased
ROS level in the gemcitabine-resistant PANC-1 and MIA PaCa-2 cells ( Figure 3C). Next, we measured the intracellular GSH/GSSG ratio to assess the intracellular redox status, and the results indicated that the GSH/GSSG ratio was significant lower in the gemcitabine-resistant cells than in the parent cells, indicating that gemcitabine resistance might correlate with redox balance ( Figure 3D). Furthermore, we assessed Keap1 and NRF2 expression in gemcitabine-resistant PANC-1 and MIA PaCa-2 cells and observed a decrease in Keap1 protein levels and an increase in NRF2 levels ( Figure 3E). Finally, we assessed the expression of NRF2-targeted, ARE-driven ROS-detoxification genes and observed a significant increase in ARE-driven gene levels in the gemcitabine-resistant cells ( Figure 3F).

| NAC treatment increases dCK expression and improves cell sensitivity to gemcitabine
In cells, NAC is frequently used as a sulfhydryl source and acetylated precursor for reduced GSH. NAC also interacts directly with ROS and scavenges oxygen free radicals. Thus, we treated PANC-1 and MIA PaCa-2 cells with NAC to inhibit intracellular ROS activity and to examine the subsequent impact on dCK expression. We first

| dCK is negatively correlated with NRF2 expression in pancreatic cancer patients
As discussed above, we observed a negative correlation between dCK and NRF2 expression in vitro in pancreatic cancer cell lines.
Next, we examined dCK and NRF2 expression in pancreatic cancer F I G U R E 5 dCK expression is negatively correlated with NRF2 levels in pancreatic cancer patients. (A) Patients with higher dCK levels exhibited lower levels of NRF2, while NRF2 expression was higher in patients with lower dCK levels. (B) dCK was negatively and significantly correlated with NRF2 expression in pancreatic cancer patients patients. As shown, patients with lower dCK levels exhibited higher levels of NRF2, indicating a negative correlation between these two proteins ( Figure 5A). Next, we increased the number of patient cases, and performed IHC staining to measure the correlation between dCK and NRF2 expression. In addition, the statistical analysis indicated that dCK expression is negatively and significantly correlated with NRF2 expression in pancreatic cancer patients ( Figure 5B).
In conclusion, our present study identifies the negative impact of the gemcitabine metabolic regulator dCK on the Keap1/NRF2/ARE axis and reveals that decreased dCK expression regulates ROS production and the intracellular redox status, which might contribute to gemcitabine resistance and regulate pancreatic cancer cell proliferation ( Figure 6). Based on the above reports and discussions, the intracellular dCK level is a promising target in pancreatic cancer. However, the regulatory mechanisms of dCK have seldom been discussed in pancreatic cancer. In idiopathic pulmonary fibrosis (IPF), dCK has been reported to be a downstream target of hypoxia and contributed to alveolar epithelial cell proliferation. 31 In chronic obstructive pulmonary disease (COPD), dCK has been also reported to be induced by hypoxia, and increased dCK levels contribute to apoptosis in chronic lung disease. 32

| D ISCUSS I ON
Hypoxia inducible factor 1α (HIF1α) is a master regulator of the hypoxic response and acts as a transcription factor that governs the expression of many hypoxia-induced genes. 33 36 In our study, we also observed that scavenging ROS production by NAC increased dCK expression. Moreover, using PROMO 3.0 to identify potential transcription factors, Blackburn MR reported that potential HIF1α binding sites exist in the dCK promoter region.
Furthermore, potential p53 and NF-κB binding sites also exist in the dCK promoter. 32 The transcriptional activities of p53 and NF-κB are also under ROS regulation, indicating that the intracellular ROS levels and redox balance might govern dCK transcript expression. 37,38 The dCK levels were also regulated at post-transcriptional levels. One recent study demonstrated that ROS detoxification and microRNA (miR)-155 suppressed post-transcriptional dCK levels, leading to chemoresistance in pancreatic cancer cells. 39 Posttranslational modifications also affect dCK enzymatic activity, regulate drug metabolism and contribute to drug resistance in cancer.
For example, ataxia telangiectasia mutated (ATM) phosphorylates and activates dCK at serine 74 in response to ionizing radiation (IR). dCK activation shifts dCK substrate specificity towards deoxycytidine, increases the intracellular dCTP pools and DNA repair activity, and contributes to chemotherapy and radiotherapy resistance. 27,40 dCK phosphorylation at serine 74 is reversed by protein phosphatase 2A, which negatively regulates dCK activity. 41 Moreover, using a mass spectrometry technique, dCK was found to exist in a complex that contains cyclin-dependent kinase 1 (Cdk1). After IR, Cdk1 interacts with dCK, and the activity of Ckd1 is inhibited by dCK both in vitro and in vivo, making dCK an important G2/M checkpoint regulator in response to DNA damage. 42 Increased basal ROS levels can induce apoptosis, and one possible mechanism for this effect is that increased basal ROS levels activate ATM. 43

| CON CLUS IONS
In conclusion, our present study uncovered novel roles for the gemcitabine metabolic enzyme dCK in ROS detoxification and NRF2/ARE transcription. In addition, our studies also demonstrated that scavenging intracellular ROS increased dCK mRNA and protein levels. Together with previous dCK reports, we have increased the understanding of the role of this enzyme in pancreatic cancer and have shed light on novel strategies for improving chemotherapy resistance in pancreatic cancer.

ACK N OWLED G EM ENTS
This work was supported by the National Natural Science Foundation Program (16YF1401800).

CO N FLI C T S O F I NTE R E S T
No potential conflicts of interest were disclosed.