Roles of voltage‑gated potassium channels in the maintenance of pancreatic cancer stem cells.

The targeting of membrane proteins that are activated in cancer stem cells (CSCs) represents one of the key recent strategies in cancer therapy. The present study analyzed ion channel expression profiles and functions in pancreatic CSCs (PCSCs). Cells strongly expressing aldehyde dehydrogenase 1 family member A1 (ALDH1A1) were isolated from the human pancreatic PK59 cell line using fluorescence‑activated cell sorting, and PCSCs were identified based on tumorsphere formation. Microarray analysis was performed to investigate the gene expression profiles in PCSCs. ALDH1A1 messenger RNA levels were higher in PCSCs compared with non‑PCSCs. PCSCs were resistant to 5‑fluorouracil and capable of redifferentiation. The results of the microarray analysis revealed that gene expression related to ion channels, including voltage‑gated potassium channels (Kv), was upregulated in PCSCs compared with non‑PCSCs. 4‑Aminopyridine (4‑AP), a potent Kv inhibitor, exhibited greater cytotoxicity in PCSCs compared with non‑PCSCs. In a xenograft model in nude mice, tumor volumes were significantly lower in mice inoculated with PK59 cells pre‑treated with 4‑AP compared with those in mice injected with non‑treated cells. The present results identified a role of Kv in the persistence of PCSCs and suggested that the Kv inhibitor 4‑AP may have potential as a therapeutic agent for pancreatic carcinoma.

Introduction

Pancreatic cancer is one of the most intractable human malignancies and the seventh leading cause of cancer-related death worldwide (1-3). The 5-year survival rate of pancreatic ductal adenocarcinoma, which is a major histological subtype of pancreatic tumors, remains poor (8% in the USA in 2017, and 15-20% in Japan in 2018), even after surgery, due to its strongly invasive and metastatic characteristics (1,4). Chemotherapy, including gemcitabine, S-1 and 5-fluorouracil (5-FU), and radiation treatment are the current options for patients with unresectable, recurrent or metastatic cancer (1-4). Although potential drug combinations have been tested in clinical trials, their efficacy is limited due to the drug resistance of pancreatic cancer (5-7). A detailed understanding of the molecular mechanisms regulating the tumorigenesis and progression of pancreatic cancer is needed for the development of more effective treatments.

Accumulating evidence indicates the significance of cancer stem cells (CSCs), which are resistant to current anticancer medicines and radiation, in tumor initiation, progression, recurrence, metastasis and, ultimately, patient death (8-10). Therefore, specific and effective therapies against CSCs are needed, and previous studies have reported the potential of drugs that target CSC markers or signaling pathways (11-15). The targeting of membrane transporters that are specifically upregulated in CSCs represents one of the key novel strategies in cancer treatment.

Membrane transporters and transporters/ion channels have been demonstrated to be involved in the biological processes of cancer cells, and a cellular physiological approach exhibits potential as a potential strategy in specific cancer therapies (16-18). Our previous studies reported that a number of ion channels, including transient receptor potential vanilloid 2 (TRPV2), were highly expressed in squamous cell carcinoma (ESCC) CSCs (19,20). The cytotoxic concentration of tranilast, an analog of a tryptophan metabolite that specifically inhibits TRPV2, needed to suppress cell proliferation was lower for CSCs compared with non-CSCs. Tranilast is typically used in the treatment of inflammatory diseases, such as allergic conjunctivitis, asthma, dermatitis, keloids and hypertrophic scars (21). The results of our aforementioned studies suggested that TRPV2 may serve a role in the maintenance and homeostasis of CSCs and may act as a targeted therapeutic agent for ESCC (20). However, transporter/ion channel expression profiles and their oncogenic functions in pancreatic CSCs (PCSCs) have not been examined in detail to date.

Therefore, the aims of the present study were to determine the expression levels and function of transporters/ion channels in PCSCs obtained from pancreatic carcinoma cell lines. The results demonstrated that various genes related to ion channels, including the voltage-gated potassium channel (Kv), were upregulated in PCSCs compared with non-PCSCs. The study also examined whether 4-aminopyridine (4-AP), a Kv inhibitor widely used in the treatment of multiple sclerosis (MS) and Charcot-Marie-Tooth disease, exerted specific inhibitory effects on PCSCs in vitro and in vivo.

Materials and methods

Cell lines, culture and materials

PK59, PANC1, PK1, PK45H, KP4-1, AsPC1 and SUIT-2 cells were obtained from the Cell Engineering Division in RIKEN BioResource Center and cultured as previously described (20). PK59, PANC1, PK1, PK45H, AsPC1 and SUIT-2 cells were maintained in rPMi-1640 medium (Nacalai Tesque, Inc.) supplemented with 100 u/ml penicillin, 100 µg/ml streptomycin and 10% FBS (Nacalai Tesque, Inc.). KP4-1 cells were cultured in DMEM plus HamF12 medium (Nacalai Tesque, Inc.) supplemented with 100 u/ml penicillin, 100 µg/ml streptomycin and 10% FbS. Cells were cultured in flasks or dishes in a humidified incubator at 37°C with 5% Co2 in air. 4-AP was purchased from Nacalai Tesque, Inc. 5-Fu was purchased from FUJIFILM Wako Pure Chemical Corporation.

Detection of CSCs using Aldefluor fluorescence and cell sorting

The expression of ALDH1A1 in PK59 cells was confirmed using an Aldefluor kit according to the manufacturer's instructions (Stemcell Technologies, Inc.) (20,22). PK59 cells were centrifuged (150 × g; 23°C; 5 min) and resuspended in Aldefluor buffer and 3 ml of the resulting cell suspension (1.0×106 cells/ml) was mixed with 45 µl activated Aldefluor substrate. The mixture with added ALDH1A1 inhibitor dieth-ylaminobenzaldehyde (DEAB) was used as a negative control. Following incubation at 37°C for 45 min away from light, the cells were centrifuged (150 × g; 23°C; 5 min), resuspended (1.0×106 cells/ml) in Aldefluor buffer and maintained on ice for 1 h. The cells were subsequently isolated by flow cytometry using the Cell Sorter SH800 (Sony Corporation) and categorized into two subgroups based on fluorescence and cell scattering. Cells expressing ALDH1A1 treated with or without 4-AP were analyzed using the BD Accuri C6 flow cytometer and the associated software (BD Biosciences).

CSC culture

Cells expressing high levels of ALDH1A1 were separated from PK59 cells using FACS and cultured in tumorsphere medium containing rPMi-1640 medium with 100 u/ml penicillin, 100 µg/ml streptomycin, 2% b27 supplement (gibco; Thermo Fisher Scientific, inc.), 10 ng/ml epidermal growth factor and 10 ng/ml fibroblast growth factor (both Invitrogen; Thermo Fisher Scientific, Inc.) for 7 days in ultra-low attachment 6-well plates (Corning, Inc.) (20). The presence of tumorspheres in plates was detected under an inverted light microscope (magnification, ×40). Tumorspheres were recovered by centrifugation (300 × g; 23°C; 10 min), counted, and dissociated into single cells by processing with trypsin-EDTA and gentle mechanical crushing using a glass pipette (20). The obtained single cells were re-plated to allow the reformation of spheres. The spheres were passaged every 4-7 days when their diameter reached ~100 µm (20).

Drug sensitivity test

PCSCs separated from adherent and spheroid PK59 cells were seeded in 96-well microplates at a concentration of 2,000 cells/well. The cells were treated with increasing concentrations of 5-FU or 4-AP for 72 h (5-FU, 0.16-80 µM; 4-AP, 0.08-40 mM). WST-8 assay (Cell Count Reagent SF; Nacalai Tesque, Inc.) was performed to evaluate the viability of cells treated with 5-FU or 4-AP. A total of 10 µl of this reagent was added per well and the cells were incubated for 90 min at 23°C. The number of viable cells was determined by measuring absorbance at 490 nm using microplate reader.

Reverse transcription-quantitative (RT-qPCR)

Total RNA was extracted from PK59 cells and PK59 CSCs using an RNeasy kit (Qiagen, Inc.); the concentration of the cells before RNA extraction was not measured. The concentration of RNA was measured after extraction, and the RNA concentration was adjusted to 50 ng/µl. Reverse transcription was performed using a High-Capacity cDNA Reverse Transcription kit (Applied Biosystems; Thermo Fisher Scientific, inc.) as follows: 25°C for 10 min, 37°C for 120 min and 85°C for 5 min). The 7300 real-Time PCr System (Applied biosystems; Thermo Fisher Scientific, inc.) with TaqMan gene expression Assays (Applied biosystems; Thermo Fisher Scientific, inc.) were used according to the manufacturer's instructions. The PCR thermocycling conditions were as follows: initial denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and at 60°C for 1 min. The expression levels of the following mRNAs were determined: ALDH1A1 (cat. no. HS00946916_m1), potassium voltage-gated channel (KCN) subfamily B member 1 (KCNB1; cat. no. Hg00270657_m1), KCNC1 (cat. no. Hg00428197_m1), KCND1 (cat. no. Hg01085825_m1), SOX 2 (cat. no. Hs 010530 49_ s1), CD 4 4 (cat. no. Hs00153304_m1), CD133 (cat. no. Hs01009259_m1), and CXCR4 (cat. no. Hs00607978) (all Applied Biosystems; Thermo Fisher Scientific, inc.). The expression levels were normalized to those of the housekeeping gene β-actin (cat. no. Hs01060665_g1; Applied Biosystems; Thermo Fisher Scientific, Inc.). Assays were performed in triplicate.

Small interfering RNA (siRNA) transfection

PK59 cells were seeded at a density of 1.0×105 cells/well on 6-well plates and were transfected with 20 nM Aldh1A1 or KCNB1 siRNA (Stealth RNAi™; cat. nos. HSS100366 and hSS180043; Invitrogen; Thermo Fisher Scientific, Inc.) using the Lipofectamine® RNAiMAX reagent (Invitrogen; Thermo Fisher Scientific, inc.) according to the manufacturer's instructions. The cells were incubated with the transfection mixture at 37°C for 24 h, following which the medium was replaced. The Stealth RNAi™ siRNA negative control (cat. no. 12935112; invitrogen; Thermo Fisher Scientific, inc.) was used as the negative control. The siRNA sequences were as follows: ALDH1A1 siRNA sense, 5′-CAG GAA CAG UGU GGG UGA AUU GCU A-3′ and antisense, 5′-UAG CAA UUC ACC CAC ACU GUU CCU G-3′; and KCNB1 siRNA sense, 5′-CCU AAG UUC UUA AGG CAG AAC UGU A-3′ and antisense, 5′-UAC AGU UCU GCC UUA AGA ACU UAG G-3′.

Assessment of overexpression

Unsorted PK59 cells were transfected with the control HaloTag plasmid (cat. no. G6591), ALDH1A1 HaloTag plasmid (cat. no. FHC09770) or KCNB1 HaloTag plasmid (cat. no. FXC27060) using FuGENE HD transfection reagents (cat. no. E2311) (all Promega Corporation) according to the manufacturer's instructions. Vector transfection was confirmed by fluorescence microscopy to detect the haloTag fusion protein stained by the tetramethylrhodamine-conjugated HaloTag ligand (cat. no. G8252; Promega Corporation) according to the manufacturer's protocol. A cell proliferation assay was subsequently conducted using ALDH1A1-expressing cells. Briefly, PK59 cells were seeded in 6-well plates at a density of 0.75×105 cells/well and incubated at 37°C with 5% Co2 for 24 h. Subsequently, the cells were transfected with plasmid as aforementioned, detached from the flasks with trypsin-EDTA at 48 and 72 h post-transfection and counted with a hemocytometer.

Cell counting

Unsorted PK59 cells were seeded in 6-well plates at a density of 0.75×105 cells/well and incubated at 37°C with 5% CO2 for 24 h. Subsequently, the cells were transfected with siRNA as aforementioned, detached from the flasks with trypsin-EDTA at 72 h post-transfection and counted with a hemocytometer.

Sample preparation and hybridization to microarrays

Total RNA (50 ng/µl) was extracted from Pk59 cells and Pk59 CSCs using an RNeasy kit (Qiagen, inc.), aforementioned. Microarray analyses were performed by Takara Bio, Inc. Using the Agilent SurePrint G3 Human Gene Expression 8×60 K microarray (Agilent Technologies, Inc.). RNA quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc.). Total RNA was labeled with Cyanine-3 (Cy3) using a Low Input Quick Amp Labeling kit (Agilent Technologies, Inc.). Samples were purified on RNeasy columns (Qiagen, inc.). The Cy3-labeled cRNA was fragmented and subsequently hybridized for 17 h. Following hybridization, the slides were washed and immediately scanned with an Agilent DNA Microarray Scanner (cat. no. G2565CA; Agilent Technologies, Inc.) using the one-color setting for 8×60K array slides.

Microarray data processing

Scanned images of the microarray slides were examined using Feature Extraction Software 10.10 (Agilent Technologies, Inc.) with default parameters to obtain signal intensities with background subtraction and spatial detrending. Gene expression profiles and functions were assessed with Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems, Inc.) as previously described (17,19,20). The obtained datasets have been submitted to the public curated database Gene Expression Omnibus (GSE172185) (https://www.ncbi.nlm.nih.gov/geo).

Formation of tumorspheres

PCSCs were suspended in medium with or without 5 mM of 4-AP, seeded at 100 or 200 cells/well in 96-well plates and incubated at 37°C with 5% CO2 in air for one week. To exclude any potential disturbances to the formation of tumorspheres, no changes were made to the culture medium. The formed spheres were counted using a phase-contrast microscope under ×40 magnification as previously described (20,23).

Measurement of the intracellular concentration of chloride [(Cl−)]i

[Cl]i was assessed using the MQAE reagent (dojindo laboratories, inc.), a chloride-sensitive fluorescent probe (24). PK59 cells were seeded in 24-well plates at a density of 4×104 cells/well and incubated at 37°C with 5% Co2 for 48 h. Subsequently, MQAE reagent dissolved in complete RPMI-1640 medium was applied, and the plates were incubated at 37°C in a Co2 incubator for 12 h. The plates were washed five times with PBS, and the fluorescence intensity of MQAE was measured by fluorescence microscopy (BZ-X810; Keyence Corporation); three fields of view were analyzed per sample at ×100 magnification. Quantification was performed using a BZ-X810 analyzer and accompanying software (v.1.1.1.8; Keyence Corporation).

Animal experimental protocol

All animal protocols were approved by the by the Institutional Review Board of the Kyoto Prefectural university of Medicine (Kyoto, Japan; approval no. M2019-267), and all experiments were strictly conducted in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals. A total of 27 female BALB/c nude mice (age, 4 weeks) were obtained from Shimizu Laboratory Supplies Co., Ltd. and maintained under pathogen-free barrier conditions. The mice were provided with ad libitum access to sterile food and water and housed with a 12-h light-dark cycle at 24°C and 40-70% humidity. Suspensions of 5×105 Pk59 cells in 50 µl rPMi-1640 and 50 µl Matrigel matrix (Corning, Inc.) incubated with 5 mM 4-AP for 48 h in vitro were subcutaneously injected into the right side of the lower flanks of 4-week-old female nude mice, and suspensions containing the same number of PK59 cells incubated without 4-AP were injected into the left side (n=3). To investigate the influence that the expression of Aldh1A1 has on tumor growth, suspensions of 5×105 PK59 cells transfected with control siRNA, ALDH1A1 siRNA, control plasmid or ALDH1A1 plasmid in vitro were injected subcutaneously into the lower flanks of 4-week-old female nude mice (n=3 mice/group). For the assessment of drug combination treatment, suspensions of 5×105 PK59 cells incubated with 4-AP, 5-FU or 5-FU combined with 4-AP for 48 h in vitro were injected subcutaneously into the lower flanks of 4-week-old female nude mice (n=3 mice). All 27 mice were sacrificed using pentobarbital (120 mg/kg) 28 days or 35 days after the injection, and the volumes of resected tumors were measured. The volumes of tumors were calculated according to the following formula (25): Tumor volume (mm3)=1/2 × length × width2.

Humane endpoints were reached when the xenograft tumor reached >10% of the animal body weight, the tumor diameter was >20 mm, tumors metastasized or grew such that it led to rapid body weight loss (>20%), or signs of immobility, a huddled posture, the inability to eat, ruffled fur, self-mutilation, ulceration, infection or necrosis were observed. All 27 animals reached the study endpoints (on day 28 or 35) and were euthanized by cervical dislocation under anesthesia by intraperitoneal injection of pentobarbital (70 mg/kg). Death was verified by the cessation of a heartbeat and dilated pupils.

Immunohistochemistry

Xenograft samples were embedded in paraffin following 12-h formalin fixation at 4°C. Immunohistochemistry for the ALDH1A1 protein was performed on 4-µm-thick paraffin sections of tumor tissues using the avidin-biotin-peroxidase method. After dewaxing paraffin sections with xylene and hydration in a graded ethanol series (99.5, 90, 70 and 50%), endogenous peroxidases were blocked by an incubation in 0.3% H2O2 for 30 min at 23°C. An Avidin/biotin blocking kit (vector laboratories, Inc.) was used to block endogenous biotin, biotin receptors and avidin-binding sites. The sections were subsequently treated with a protein blocker and incubated at 4°C overnight with an anti-ALDH1A1 antibody (1:300; cat. no. 611194; BD Pharmingen; BD Biosciences), followed by visualization of the avidin-biotin-peroxidase complex using the Vectastain ABC Elite kit (Vector Laboratories, Inc.) with diaminobenzidine tetrahydrochloride. The sections were counterstained with hematoxylin for 4 min at 23°C and subjected to dehydration in a graded ethanol series (50, 70, 90 and 99.5%), cleared in xylene and mounted in Entellan new (Sigma-Aldrich, Inc.). Quantification of staining intensity was performed using bz-x800 All-in-one fluorescence microscope and analyzer software (magnification, ×400; n=3 mice/group).

Correlation analysis in databases

Correlation analysis was performed using cBioportal (www.cbioportal.org). Using the human genome assembly hg19/grCh37, the gene expression at the mRNA level of each case in databases was investigated. After selection of the database, TCGA, PanCancer Atlas, the gene name 'ALDH1A1' was entered and referred to the co-expression. This co-expression function enables the investigation the relationships of gene expression between ALDH1A1 and other genes. Spearman correlation between the expression levels of ALDH1A1 and Kv in primary tumor samples of human pancreatic cancer was performed using cBioPortal.

Statistical analysis

All statistical analyses were conducted using the statistical software JMP (version 12; SAS Institute, Inc.). Statistical analysis was performed using the Mann-whitney u test for two-group comparisons. One-way ANOVA was used to compare the differences among multiple groups, followed by Tukey's multiple comparisons post-hoc test. In the drug sensitivity assay, IC50 values were calculated based on a non-linear regression. Data are presented in the graphs as the mean ± SEM. P<0.05 was considered to indicate a statistically significant difference.

Results

Acquisition of CSCs

The Aldefluor assay was performed using PK59 cells to isolate cells that expressed high levels of ALDH1A1 by FACS (Fig. 1A). The proportion of CSCs in the Pk59 cell line was 3.9%. To confirm the characteristics of CSCs, tumorsphere formation assay was performed, and the formation of tumorspheres was observed under a microscope (Fig. 1B). Total RNA was extracted from tumorspheres, and RT-qPCR was used to compare ALDH1A1 mRNA levels in adherent PK59 cells (non-CSCs) and spheres (CSCs). Among PK59 cells, CSCs exhibited higher ALDH1A1 mRNA levels compared with those in non-CSCs (Fig. 1C). The Aldefluor assay was also performed using other pancreatic cell lines, including PANC1, PK1, PK45H, KP4-1, AsPC1 and SUIT-2 (Fig. S1). Cells that strongly expressed ALDH1A1 were only isolated from three cell lines, namely PANK1, PK4-1 and SUIT-2 (Fig. S1B). Among them, only cells isolated from SUIT-2 proliferated. However, SUIT-2 cells expressing high levels of ALDH1A1 did not form fine spheres in the sphere formation assay (Fig. S1C). Thus, only CSCs from the PK59 cell line were examined in the present study.

Figure 1

Acquisition of CSCs in PK59 cells. (A) Aldefluor assay in PK59 cells. Cells strongly expressing ALDH1A1 were isolated using FACS. A specific inhibitor of ALDH, DEAB, was used to control for background fluorescence. (B) Sorted PK59 cells strongly expressing ALDH1A1 formed spheres in the sphere formation assay. Magnification, ×40. Scale bar, 500 µm. (C) Aldh1A1 mRNA expression levels were increased in CSCs compared with those in non-CSC. Data are presented as the mean ± SEM. n=3. *P<0.05 vs. non-CSCs. (D) The obtained CSCs exhibited the ability to redifferentiate. CSCs cultured on normal plates acquired adhesive properties and a typical morphology. Magnification, ×40. Scale bar, 500 µm. (E) Anticancer drug sensitivity of CSCs. 5-FU was more cytotoxic at a lower concentration in non-CSCs compared with in CSCs in PK59 cells. n=4. CSC, cancer stem cell; ALDH1A1, aldehyde dehydrogenase 1 family member A1; ALDH, aldehyde dehydrogenase; 5-FU, 5-fluorouracil; DEAB, diethylaminobenzaldehyde.

To clarify whether established CSCs were capable of redifferentiation, CSCs were seeded on normal plates in culture medium. Three days later, these cells were observed to possess adhesive properties and typical morphologies (Fig. 1D). The CSCs and non-CSCs of PK59 cells were treated with 5-FU to determine their resistance to anticancer drugs. The IC50 values were ~3.47 and 6.07 µM in non-CSCs and CSCs, respectively (Fig. 1E). These results suggested that the cytotoxicity of 5-FU was greater at lower concentrations in non-CSCs compared with that in CSCs, which reflected the properties of CSCs. The properties of CSCs were further assessed using comparative analyses between sorted PK59 cells and unsorted PK59 cells of the expression profiles of other previously reported PCSC markers, such as CD44, CD133 and SOX2 (4,26,27). In addition, ALDH1A1 overexpression and knockdown experiments were performed to investigate its effects on cell proliferation. Aldh1A1 mRNA levels were significantly increased by the overexpression of ALDH1A1 in PK59 cells compared with those in the control group (Fig. 2A), and CD44 and CD133 mRNA levels were also significantly elevated (Fig. 2b). The number of ALDH1A1 plasmid-transfected PK59 cells at 72 h post-transfection was significantly higher compared with that of the control cells (Fig. 2C). ALDH1A1 and SOX2 mRNA levels were markedly decreased by the transfection of ALDH1A1 siRNA into PK59 cells (Fig. 3A and B). Tumorigenicity following transfection was slightly weaker in nude mice inoculated with ALDH1A1 siRNA-transfected PK59 cells compared with those injected with the control cells (Fig. 3C). By contrast, tumorigenicity was slightly stronger in nude mice inoculated with ALDH1A1 plasmid-transfected PK59 cells compared with that in mice injected with control cells (Fig. 3D). These results suggested a potential relationship between ALDH1A1 and tumor formation, although not statistically significant in the three mice, which supported the rationale for PK59 cells strongly expressing ALDH1A1 to be defined as PCSCs.

Figure 2

Effect of ALDH1A1 overexpression in PK59 cells. (A) The ALDH1A1 plasmid effectively increased ALDH1A1 mRNA levels in PK59 cells. n=3. (B) CD44 and CD133 mRNA levels were significantly increased by the Aldh1A1 plasmid compared with those in cells transfected with the control plasmid. (C) The overexpression of ALDH1A1 promoted the proliferation of PK59 cells. Cell numbers were counted 72 h after ALDH1A1 plasmid transfection. n=4. Data are presented as the mean ± SEM. *P<0.05 vs. control plasmid. ALDH1A1, aldehyde dehydrogenase 1 family member A1; ACTB, β-actin.

Figure 3

Effect of ALDH1A1 knockdown in PK59 cells. ALDH1A1 siRNA effectively reduced (A) ALDH1A1 and (B) SOX2 mRNA levels in PK59 cells. n=3. *P<0.05 vs. control siRNA. (C) A xenograft model in nude mice using Pk59 cells transfected with ALDH1A1 siRNA. The volumes of resected tumors were slightly lower in the ALDH1A1 siRNA group compared with those in the control group. n=3. (D) A xenograft model in nude mice using PK59 cells transfected with the ALDH1A1 overexpression plasmid. The volumes of resected tumors were slightly higher in the ALDH1A1 plasmid group compared with those in the control plasmid group. n=3. Data are presented as the mean ± SEM. ALDH1A1, aldehyde dehydrogenase 1 family member A1; siRNA, small interfering RNA; ACTB, β-actin.

Gene expression in PK59 CSCs

Gene expression data from PK59 CSCs and non-CSCs were obtained using microarray and bioinformatics analyses. The results of the microarray analysis revealed that the expression levels of 4,870 genes in PK59 CSCs exhibited fold-changes >3.0 compared with those in non-CSCs. Among these, the expression levels of 2,739 genes were upregulated, whereas the levels of 2,131 genes were downregulated in PK59 CSCs. A list of 50 genes with the greatest increases and decreases in expression levels in PK59 CSCs is presented in Table SI. These results revealed the upregulated expression levels of CSC markers, such as SOX2, ALDH1A1 and CXCR4, in PK59 CSCs compared with those in non-CSCs (Table SII). The expression levels of ion channel-related genes in PK59 CSCs were further analyzed using IPA software; the expression levels of 57 genes associated with ion channels were upregulated in PK59 CSCs compared with those in non-CSCs (Table I).

Table I

Ion channel-related genes with high expression levels in cancer stem cells isolated from PK59 cells.

Gene symbol	UniGene ID	Gene name	Fold-change
GJC1	Hs.712052	Gap junction protein γ1	629.762
CACNA2D1	Hs.282151	Calcium voltage-gated channel auxiliary subunit α2δ1	303.458
TMEM63C	Hs.22452	Transmembrane protein 63C	234.969
KCTD4	Hs.23406	Potassium channel tetramerization domain-containing 4	217.796
KCNB1	Hs.84244	Potassium voltage-gated channel subfamily B member 1	191.214
HCN1	Hs.353176	Hyperpolarization-activated cyclic nucleotide-gated potassium channel 1	179.023
ANXA6	Hs.412117	Annexin A6	135.343
KCNG3	Hs.352633	Potassium voltage-gated channel modifier subfamily G member 3	104.575
SCN9A	Hs.439145	Sodium voltage-gated channel α subunit 9	83.685
MCOLN2	Hs.591446	Mucolipin 2	71.243
KCNF1	Hs.23735	Potassium voltage-gated channel modifier subfamily F member 1	67.786
KCNG1	Hs.118695	Potassium voltage-gated channel modifier subfamily G member 1	60.852
TRPM8	Hs.366053	Transient receptor potential cation channel subfamily M member 8	60.393
TRPC1	Hs.250687	Transient receptor potential cation channel subfamily C member 1	55.342
SLC9A1	Hs.469116	Solute carrier family 9 member A1	39.349
GRIN2A	Hs.411472	glutamate ionotropic receptor nMdA type subunit 2A	37.432
TMEM150C	Hs.507676	Transmembrane protein 150C	37.276
MCOLN3	Hs.535239	Mucolipin 3	32.26
GABRD	Hs.113882	γ-aminobutyric acid type A receptor δ subunit	31.534
KCND1	Hs.55276	Potassium voltage-gated channel subfamily D member 1	28.252
GRIK2	Hs.98262	Glutamate ionotropic receptor kainate type subunit 2	24.759
ZACN	Hs.714919	Zinc-activated ion channel	22.824
SCN8A	Hs.710638	Sodium voltage-gated channel α subunit 8	21.842
LRRC8C	Hs.412836	Leucine-rich repeat-containing 8 VRAC subunit C	15.768
KCNC1	Hs.552896	Potassium voltage-gated channel subfamily C member 1	15.316
CACNA1G	Hs.591169	Calcium voltage-gated channel subunit α1 G	14.324
KCNH1	Hs.553187	Potassium voltage-gated channel subfamily H member 1	10.901
CACNA1H	Hs.459642	Calcium voltage-gated channel subunit α1 h	10.788
GRIN3B	Hs.660378	glutamate ionotropic receptor nMdA type subunit 3B	10.067
CACNA1I	Hs.125116	Calcium voltage-gated channel subunit α1 i	9.816
TTYH2	Hs.27935	Tweety family member 2	9.742
KCNMA1	Hs.144795	Potassium calcium-activated channel subfamily M α 1	9.038
KCNK13	Hs.510191	Potassium two pore domain channel subfamily K member 13	8.942
CACNG4	Hs.514423	Calcium voltage-gated channel auxiliary subunit γ4	8.593
KCNE2	Hs.551521	Potassium voltage-gated channel subfamily E regulatory subunit 2	8.265
KCNMB3	Hs.591285	Potassium calcium-activated channel subfamily M regulatory β subunit 3	8.138
HCN4	Hs.86941	Hyperpolarization-activated cyclic nucleotide-gated potassium channel 4	7.770
CLCN5	Hs.166486	Chloride voltage-gated channel 5	7.703
CACNA1A	Hs.501632	Calcium voltage-gated channel subunit α1 A	7.116
ABCC9	Hs.732701	ATP-binding cassette subfamily C member 9	7.108
GPM6A	Hs.75819	Glycoprotein M6A	6.524
KCTD13	Hs.534590	Potassium channel tetramerization domain-containing 13	6.269
ASIC1	Hs.274361	Acid-sensing ion channel subunit 1	5.888
CACNA2D2	Hs.476273	Calcium voltage-gated channel auxiliary subunit α2δ2	5.886
KCNQ2	Hs.161851	Potassium voltage-gated channel subfamily Q member 2	5.540
KCNT1	Hs.104950	Potassium sodium-activated channel subfamily T member 1	5.086
PKDREJ	Hs.241383	Polycystin family receptor for egg jelly	5.076
KCNK12	Hs.591586	Potassium two pore domain channel subfamily K member 12	4.632
ITPR1	Hs.567295	Inositol 1,4,5-trisphosphate receptor type 1	4.594
CACNB4	Hs.120725	Calcium voltage-gated channel auxiliary subunit β4	4.450
CNGA4	Hs.434618	Cyclic nucleotide-gated channel α 4	4.342
CLCA4	Hs.567422	Chloride channel accessory 4	4.080
KCNQ1	Hs.95162	Potassium voltage-gated channel subfamily Q member 1	3.529
TRPV2	Hs.279746	Transient receptor potential cation channel subfamily V member 2	3.415
SCN3B	Hs.4865	Sodium voltage-gated channel β subunit 3	3.412
GRINA	Hs.594634	Glutamate ionotropic receptor NMDA type subunit-associated protein 1	3.196
CLCN3	Hs.481186	Chloride voltage-gated channel 3	3.168

The top-ranking genes related to potassium voltage-gated channels, based on fold change expression in the microarray results, were selected for further analysis. RT-qPCR was conducted to validate the results obtained in the microarray analysis. In PK59 cells, KCNB1, KCNC1 and KCND1 mRNA expression levels were significantly higher in CSCs compared with those in non-CSCs (Fig. 4A). These results demonstrated that ion channels, which are involved in maintaining CSCs, exhibited high expression levels. Among these ion channels, KCNB1 was selected, and it was found that blocking this channel with 4-AP inhibited the proliferation of PK59 cells, which was suggested that inhibition of KCNB1 could be a potential therapeutic target for pancreatic cancer. Even after five passages, the high mRNA expression levels of Kv and CSC markers were maintained in PK59 CSCs (Fig. 5A and B). These results demonstrated the constant high expression of Kv in CSCs, which reinforced the possibility that Kv may be a therapeutic target for pancreatic cancer.

Figure 4

Verification of gene expression and drug sensitivity of non-CSCs or CSCs. (A) verification of gene expression by RT-qPCR. The expression levels of three selected Kv-related genes, KCNB1, KCNC1 and KCND1, in CSCs were compared with those in the non-CSCs of PK59 cells using RT-qPCR. n=3. *P<0.05 vs. non-CSCs. (B) drug sensitivity of CSCs to the Kv inhibitor 4-AP. 4-AP was more cytotoxic at a lower concentration in CSCs compared with non-CSCs isolated from PK59 cells. n=4. (C) Representative images of cultured non-CSCs or CSCs treated with various concentrations of 4-AP. Magnification, ×40. Scale bar, 250 µm. Data are presented as the mean ± SEM. Kv, voltage-gated potassium channel; 4-AP, 4-aminopyridine; RT-qPCR, reverse transcription-quantitative PCR; KCN, potassium voltage-gated channel; KCNB1, KCN subfamily B member 1; KCNC1, KCN subfamily C member 1; KCND1, KCN subfamily D member 1; CSCs, cancer stem cells; ACTB, β-actin.

Figure 5

Expression levels of Kv and CSC markers in PK59 CSCs after five passages. (A) high mRNA expression levels of KCNB1, KCNC1 and KCND1 were maintained in PK59 CSCs. (B) High mRNA expression levels of ALDH1A1, CD44, CD133, SOX2 and CXCR4 were maintained in PK59 CSCs. mRNA expression levels were analyzed using reverse transcription-quantitative PCR. n=3. Data are presented as the mean ± SEM. *P<0.05 vs. non-CSCs. Kv, voltage-gated potassium channel; KCN, potassium voltage-gated channel; KCNB1, KCN subfamily B member 1; KCNC1, KCN subfamily C member 1; KCND1, KCN subfamily D member 1; CSC, cancer stem cell; ACTB, β-actin.

Effects of 4-AP on CSCs

The present study subsequently analyzed the overexpression of Kv in PK59 CSCs. To clarify the effects of the inhibition of Kv, PK59 non-CSCs and CSCs cells were both treated with 4-AP. Among PK59 cells, IC50 values were approximately 5.65 and 0.64 mM in non-CSCs and CSCs, respectively (Fig. 4B and C). Therefore, the cytotoxicity of 4-AP appeared to be greater in CSCs compared with that in non-CSCs. A sphere formation assay was further performed using PK59 CSCs treated with or without 4-AP. The number of spheres formed by the Pk59 CSCs was significantly lower in cells treated with 4-AP (Fig. 6). When non-CSCs (PK59 cell line) were treated with 4-AP, the population of cells strongly expressing Aldh1A1 significantly reduced compared with that in the control group (Fig. 7). Therefore, 4-AP specifically inhibited the activity of CSCs strongly expressing Kv. Overexpression experiments on PK59 cells using a KCNB1 plasmid were conducted in order to investigate its effects on cell proliferation and the function of Kv in PCSCs. KCNB1 mRNA levels in PK59 cells were markedly increased by KCNB1 plasmid transfection compared with those in cell transfected with the control plasmid (Fig. 8A). The number of KCNB1 plasmid-transfected PK59 cells was significantly higher compared with that of the control cells at 72 h post-transfection (Fig. 8B). The effects of 4-AP or 5-FU on cell proliferation were then assessed, and the results demonstrated that the number of 4-AP- or 5-FU-treated cells was significantly lower compared with that of untreated control cells (Fig. S2). Co-treatment with 4-AP and 5-FU revealed that 4-AP enhanced the inhibitory effects of 5-FU (Fig. S2).

Figure 6

Effects of 4-AP on sphere formation by CSCs. (A) Treatment with 5 mM 4-AP decreased the number of spheres in the CSCs isolated from PK59 cells compared with those in untreated CSCs. Data are presented as the mean ± SEM. n=6. *P<0.05 vs. control. (B) Representative images of cultured CSCs treated with or without 4-AP. Magnification, ×40. Scale bar, 500 µm. The concentration of 4-AP was based on the IC50. CSCs, cancer stem cells; 4-AP, 4-aminopyridine.

Figure 7

Effects of 4-AP on ALDH1A1 expression by non-CSCs. Treatment with 5 mM 4-AP for 48 h decreased the percentage of cells expressing high levels of ALDH1A1, as demonstrated by flow cytometry with Aldefluor staining. Data are presented as the mean ± SEM. n=3. *P<0.05 vs. control. 4-AP, 4-aminopyridine; ALDH1A1, aldehyde dehydrogenase 1 family member A1; CSC, cancer stem cell; DEAB, diethylaminobenzaldehyde.

Figure 8

Effect of KCNB1 overexpression in PK59 cells. (A) The KCNB1 plasmid effectively increased KCNB1 mRNA levels in PK59 cells. n=3. (B) The overexpression of KCNB1 promoted the proliferation of PK59 cells. The number of cells was counted 72 h post-KCNB1 plasmid transfection. n=4. Data are presented as the mean ± SEM. *P<0.05 vs. control plasmid. KCNB1, potassium voltage-gated channel subfamily b member 1; ACTB, β-actin.

Regarding the molecular mechanisms by which Kv maintains stemness and cell proliferation, the present study focused on changes in the intracellular ion environment. Our previous studies reported the roles of intracellular Cl− in cancer cell proliferation (28-31); thus, we hypothesized that the inhibition of K+ channels by siRNA or 4-AP may affect the movement of Cl−, which is the counter ion of K+ (24). This hypothesis was tested by measuring the fluorescence intensity of MQAE, a Cl−-sensitive fluorescent probe, to evaluate the intracellular ion concentration (Fig. 9). The results revealed that the fluorescence intensity of MQAE was increased by the knockdown of KCNB1 or the treatment with 4-AP in PK59 cells compared with that in the corresponding control groups (Fig. 9B and C). These results suggested that the change in intracellular Cl− induced by Kv may serve a key role in the molecular mechanisms underlying the regulation of stemness and cell proliferation.

Figure 9

Analysis of intracellular Cl- levels using MQAE reagent, a Cl−-sensitive fluorescence probe. (A) KCNB1 siRNA markedly decreased KCNB1 mRNA levels in PK59 cells. Data are presented as the mean ± SEM. n=3. *P<0.05 vs. control siRNA. (B) The depletion of KCNB1 using siRNA increased the fluorescence intensity of MQAE in PK59 cells. (C) The treatment with 4-AP inhibited the increased fluorescence intensity of MQAE in PK59 cells. KCNB1, potassium voltage-gated channel subfamily B member 1; siRNA, small interfering RNA; ACTB, β-actin; 4-AP, 4-aminopyridine.

The present study also examined the effects of the inhibition of Kv on tumor growth in vivo. PK59 cells incubated with or without 4-AP for 48 h were subcutaneously injected into nude mice, and the growth of tumor nodules was assessed (Fig. 10). Tumor volumes were significantly lower in the sites injected with PK59 cells treated with 4-AP compared with those formed by untreated cells (Fig. 10). The levels of CSC markers in tumor tissues were further analyzed. The number of ALDH1A1-stained cells in tumors formed by 4-AP-treated PK59 cells was markedly lower compared with those formed by the untreated cells (Fig. 11A). Furthermore, ALDH1A1 mRNA levels were significantly lower in tumors treated with 4-AP compared with those formed by the untreated cells (Fig. 11B). Tumor weights were also significantly lower in mice injected with PK59 cells treated with 4-AP combined with 5-FU compared with those in mice inoculated with cells treated with 5-FU alone (Fig. 12). These results suggested that the inhibition of Kv with 4-AP strongly suppressed the development of pancreatic tumors in vivo.

Figure 10

PK59 cells incubated with 4-AP exhibit reduced tumor growth in a mouse xenograft model. Suspensions of 5×105 Pk59 cells incubated with 5 mM 4-AP for 48 h were injected subcutaneously into the right side of the lower flanks of 4-week-old female nude mice, and suspensions with the same amount of PK59 cells incubated without 4-AP were injected into the left side. (A) Images of tumors formed by PK59 cells incubated with or without 4-AP in mice. Tumors from Pk59 cells treated with 4-AP were smaller compared with the control tumors. (B) The volumes of resected tumors were significantly lower in the 4-AP group compared with those in the control group. Data are presented as the mean ± SEM. n=3. *P<0.05 vs. control. 4-AP, 4-aminopyridine.

Figure 11

CSC marker expression in tumor tissue. (A) The number of ALDH1A1-stained cells in the tumors of mice injected with PK59 cells treated with 4-AP was markedly lower compared with those in the control group. Magnification, ×100. Scale bar, 200 µm. (B) ALDH1A1 mRNA levels were significantly lower in tumors treated with 4-AP compared with those in the control group. Data are presented as the mean ± SEM. n=3. *P<0.05 vs. control. CSC, cancer stem cell; ALDH1A1, aldehyde dehydrogenase 1 family member A1; 4-AP, 4-aminopyridine; ACTB, β-actin.

Figure 12

Tumor growth from PK59 cells treated with 4-AP, 5-FU or 5-FU combined with 4-AP in a nude mouse xenograft model. Suspensions of 5×105 PK59 cells incubated with 2 mM 4-AP, 4 µM 5-FU or 5-Fu combined with 4-AP for 48 h were injected subcutaneously into the lower flanks of 4-week-old female nude mice. (A) Images of tumors formed by PK59 cells incubated 4-AP, 5-FU or 5-FU combined with 4-AP in mice. (B) The volumes of tumors in mice. n=3. (C) resected tumor weights. Tumor weights were significantly lower in mice injected with PK59 cells treated with 4-AP combined with 5-FU compared with those treated with 5-FU alone. Data are presented as the mean ± SEM. n=3. *P<0.05; 4-AP, 4-aminopyridine; 5-FU, 5-fluorouracil.

Correlation analysis between ALDH1A1 and KCNB1 in databases

Analysis using the cBioPortal database revealed that the expression levels of ALDH1A1 and KCNB1 are positively correlated in primary tumor samples of human pancreatic cancer (Fig. 13).

Figure 13

Correlation between ALDHA1A and KCNB1 mRNA expression levels in human pancreatic cancer in the cBioPortal database. ALDH1A1, aldehyde dehydrogenase 1 family member A1; KCNB1, potassium voltage-gated channel subfamily B member 1.

Discussion

Previous studies have demonstrated that several CSC-specific markers are present in pancreatic cancer, including ALDH1, CD133, CD24, CD44, CXCR4, EPCAM, ABCG2, c-Met and nestin (4,26,27). In addition, the high expression level of ALDH1 is associated with tumorigenic cells in pancreatic ductal adenocarcinoma (32-35). Aldefluor has been successfully applied to detect enhanced ALDH1 activity and isolate CSCs from pancreatic cancer cells (33,35). In the present study, the Aldefluor assay was used to isolate and obtain PCSCs following tumorsphere formation. The obtained results demonstrated that PCSCs strongly expressed ALDH1A1, exhibited the capacity to redifferentiate and were resistant to chemotherapy. Gene expression data confirmed the upregulated expression of 57 genes in ion channels in PK59 CSCs, indicating the potential efficacy of selective ion channel inhibitors as a targeted treatment against CSCs. Among these genes, the role and function of K+ channels in CSCs were subjected to further analysis. The Aldefluor assay was performed using various cell lines, including PANC1, PK1, PK45H, PK59, KP4-1, AsPC1 and SUIT-2. Cells that strongly expressed ALDH1A1 with FACS were isolated from four cell lines, including PK59, PANK1, PK4-1 and SUIT-2. However, tumorspheres were only obtained from cells isolated from PK59, and thus, CSCs from one cell line were examined in the present study. Our previous study demonstrated that several K+ channels were upregulated in CSCs from esophageal squamous cell carcinoma (20), suggesting that similar transporters may be expressed in other PCSCs.

K+ channels serve a key role in multiple cellular functions, such as the regulation of cell volume, differentiation, proliferation, migration and apoptosis (36-38). K+ channels may be classified according to several criteria, including the stimulus to which they respond and their biophysical and structural properties, into the following four main families: Kv, calcium-activated K+ channels, inward-rectifier k+ channels and two-pore-domain K+ channels (36,38). Kv are selectively permeable to K+ ions and comprise a large family of heterogeneous groups of ion channels forming 12 subfamilies (Kv1-Kv12) (36-38). Kv are widely distributed in a number of cancer cell types, and their oncogenic potential has been documented (36,38). Kv are also potential molecular targets for anticancer therapies, and their blockers and antibodies have been investigated and used in previous studies (38). In pancreatic cancer cells, epigenetic mechanisms, such as DNA methylation, have been implicated in the altered expression of Kv1.3 (39). The targeting of Kv1.3 selectively reduces tumor progression in mouse models of pancreatic ductal adenocarcinoma (40-42). Clofazimine promotes neoplastic B-cell death by inhibiting Kv1.3 in chronic lymphocytic leukemia (43). Kv1.1 blockers, such as KAaH1 and KAaH2, inhibit cell migration and adhesion in colon adenocarcinoma, breast cancer and glioblastoma (44). The tricyclic antidepressant imipramine, an antidepressant Kv10.1 antagonist, increases survival rates in patients with moderate Kv10.1 expression in brain cancers (45). The expression of Kv11.1 has been detected in pancreatic ductal adenocarcinoma (46). Furthermore, Kv11.1 has been demonstrated to participate in the P13k/Akt-dependent pathway, and its blockade inhibits tumor growth, angiogenesis, and metastasis (47).

Previous studies have analyzed the roles of Kv in various types of stem cells. For example, Wang et al (48) have reported that rat mesenchymal stem cells (MSCs) heterogeneously express distinct types of the K+ channel, and that Kv channel activity modulates the cell cycle progression, affecting the proliferation of MSCs. Zhang et al (49) have reported that Kv10.1 regulates cell proliferation and differentiation in human bone marrow-derived MSCs. Morokuma et al (50) have demonstrated that the modulation of Kv7.1 confers a hyperproliferative invasive phenotype on embryonic stem cells. Bai et al (51) have described the various types of Kv expressed in human adipose tissue-derived stem cells. However, further studies are needed to clarify the expression, specific roles and functions of kv in CSCs. To the best of our knowledge, the present study is the first to investigate the expression of Kv in PCSCs and the inhibitory effects of 4-AP, a potent Kv inhibitor, on their proliferation.

4-AP has been used as a medical agent to regulate the symptoms of MS (52,53) and has demonstrated effectiveness as a symptomatic treatment for decreased walking capacity in patients with MS (52). Phase iii trials reported an ~25% increase in walking speed in 40% of patients and improved muscle strength in the lower extremities (54,55), and 4-AP was approved as a compound by the U.S. Food & Drug Administration in 2010. Furthermore, 4-AP exhibits antitumor activities against various types of cancer cells, such as neuroblastoma (56), Schwann cells cultured from tumors that arise in neurofibromatosis type 1 (57), melanoma (58), prostate cancer (59), malignant astrocytoma (60), hepatoblastoma (61), acute myeloid leukemia and glioma (62,63) cells. Ru et al (64) have reported that 4-AP induces glioma cell apoptosis by reducing the expression of microRNA-10b-5p. These anticancer activities of 4-AP are mostly attributed to its inhibition of Kv. However, the effects of 4-AP on CSCs remain unknown. The present study elucidated the mechanisms by which 4-AP may suppress the proliferation of CSCs through its effects as an inhibitor of Kv.

Luo et al (65) have recently demonstrated that 4-AP inhibits cell proliferation, induces apoptosis and enhances the sensitivity of a cisplatin (CDDP)-resistant lung cancer cell line to CDDP by upregulating phosphatase and tensin homolog, suggesting the potential of 4-AP as a therapeutic agent for patients with resistance to anticancer agents. Although advances have been achieved in chemotherapy for patients with pancreatic cancer through the application of various key agents, such as 5-FU, S-1, CDDP and gemcitabine, the prognosis of this cancer remains poor as recurrence is common in patients with advanced disease (1-7). In the present study, 4-AP inhibited tumorsphere formation in PCSCs and tumor growth in vivo compared with those in the corresponding control groups. Furthermore, 4-AP decreased the population of cells strongly expressing ALDH1A1 among PK59 cells. These results suggested the potential of 4-AP as a candidate drug in combination with anticancer agents for treatment-resistant pancreatic cancer.

Our previous studies demonstrated that [Cl−]i controlled by Cl− channels may be an important messenger (28-31), and that a change in [Cl−]i induced cell cycle arrest at the G0/G1 phase via mitogen-activated protein kinases in cancer cells (28,29). Since Cl− is the counter ion of K+, the inhibition of K+ channels may affect its movement (24). In the present study, the depletion of kCnb1 or treatment with 4-AP altered the fluorescence intensity of MQAE. These results suggested that Kv may regulate the stemness and cell proliferation by controlling [Cl−]i in pancreatic cancer.

Pancreatic cancer is a desmoplastic tumor with fibroblasts, and a number of factors render the tumor environment a hostile milieu for antitumor immune cells, such as hypoxia, hypoglycemia and lactic acidosis (3-7). Therefore, a xenograft model does not simulate a real tumor. On the other hand, analysis of the cBioPortal database in the present study revealed that the expression levels of ALDH1A1 and KCNB1 positively correlated in primary tumor samples of human pancreatic cancer, suggesting that similar results to our in vitro analyses are obtainable under in vivo conditions.

In conclusion, the results of the present study demonstrated that several ion channels, including Kv, were strongly expressed in PCSCs. The cytotoxicity of 4-AP against Kv was greater at a lower level in CSCs compared with that in non-CSCs. Although further studies are needed on the role and function of Kv in CSCs, its inhibitor 4-AP has potential as a novel therapeutic target against pancreatic cancer.