Background
Three studies reported an endemic area of blackfoot disease (BFD) in the southwestern region of Taiwan [
1‐
3], and one report showed high levels of arsenic and fluorescent substances in the underground water in this area [
4]. Arsenic in the drinking water correlates with high incidence of bladder, lung, renal, and skin cancers in the areas of endemic BFD [
5‐
9]. Furthermore, the correlation between arsenic exposure and occurrence of bladder cancer has been demonstrated by epidemiological studies [
8,
9]. This speculation was supported by setting up the tap-water supply system in these arsenic endemic regions and bladder cancer mortality rate reduced afterword [
10].
Inorganic arsenic has been defined as a human carcinogen [
11]. Natural and anthropogenic sources may release arsenic into the atmosphere. Exposure to arsenic may induce either cell proliferation or apoptosis depending on the dose of arsenic [
12]. In addition, the duration of arsenic exposure is another parameter involved in the carcinogenesis.
Mitotic serine/threonine kinase Aurora family consists of Aurora-A, B and C three members with 403 amino acids and the molecular weight of 46 kDa [
13]. Aurora-A participates in cell mitosis and meiosis, and proliferation through regulation of cell cycle [
14]. In proliferative cells, Both of the mRNA and protein levels of Aurora-A are dynamically expressed at the G1, S, and G2 phase. Briefly, its expression decreased between mitotic exit and entrance of the G1 phase of the next cell cycle [
15]. Aurora-A dysfunction may lead to genomic damage including centrosome amplification and chromosomal instability. It further correlates with the tumorigenesis of diverse cancer cells. Vader and Lens reported that Aurora-A can be regulated by transcriptional upregulation or gene amplification [
14]. Abnormal Aurora-A dysfunction causes centrosome amplification and transformation of two murine fibroblasts [
16,
17], implying that it may play similar role in human cancers. Elevated expression of Aurora-A is associated with diverse clinical pathogenic parameters, including tumor grade, invasion, metastasis, and low overall survival of bladder cancer patients [
18].
Comparison of bladder cancer tumor cells from BFD endemic area and other regions reveals higher Aurora-A expression in the former [
19]. Arsenic increases the activity of c-
myc and E2F-1 [
20] and selective activation of NF-kB and E2F by low concentration of arsenite in U937 human monocytic leukemia cells [
21]. Aurora-A acts as a direct target of E2F3 during G2/M cell cycle progression [
22]. Increased E2F1 protein level accompanied with Aurora-A overexpression was detected in breast cancer specimens. Further analysis reveals that Aurora-A increased E2F1 protein stability by suppressing its degradation [
23]. Currently, how arsenic-related Aurora-A dysfunctions through gene amplification or epigenic modification remain unknown.
This study aimed to reveal the molecular mechanism of arsenic-induced tumor development. We established an immortalized human uroepithelial cell line model system, and set up a mouse-arsenic exposure model to validate our cell line investigation.
Methods
Cell line and culture
The immortalized bladder urothelial E7 cells (ATCC, #CRL-2017) contain HPV E7 oncogene, which binds with phosporylated tumor suppressor RB protein (provided by Nan-Haw Chow; National Cheng Kong University Hospital) [
24]. This cell line was maintained in F12 medium (GIBCO, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum at 37 °C in a 5% CO
2 incubator.
Arsenic treatment
The immortalized E7 cells were treated with different amount of sodium arsenite (NaAsO2; Fluka, St. Louis, MO, USA) for various times and the protein was collected using lysis buffer (50 mM Tris-HCl, pH 7.4, 1% Nonidet P-40, 150 mM NaCl, 0.5% Sodium deoxycholate). RNA was extracted by TRizol™ (Invitrogen, Carlsbad, CA, USA), and genomic DNA was extracted by the commercial kit, YGB100 (RBC Bioscience, Taipei, Taiwan).
Immunofluorescent assay (IFA)
E7 cells (1 × 105 or 5 × 104/well) were plated in 6-well plates. After incubation with arsenic for one week, cells were fixed with 3.7% formaldehyde for 30 min followed by washing with 1X PBS for 30 min and 0.1% Triton X-100 treatment for 30 min. Cells were washed again with 1X PBS, immersed with blocking buffer (Thermo, Rockford, IL, USA) for 30 min, and then stained with mouse anti-BrdU antibody (#RPN20AB, Amershan Biosciences, Buckigamshire, England), mouse anti-Aurora-A antibody (NCL-L-AK2, Novocastra, Bannockburn, IL, USA) and mouse anti-α-tubulin antibody (Sigma Chemical Co., St, Louis, MO, USA) at 4°Covernight. The next day, cells were washed and stained with Fluirescein (FITC)-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA, USA) for 1 h. To stain the nuclear DNA, cells were incubated with Propidium Iodide (PI, 5 μg/ml; Sigma), or Hochest 33,258 (50 ng/ml, Sigma).
Flow cytometry analysis
Cell cycle distribution was determined by flow cytometry. Cells (1 × 105/well) were plated in 6-well plates. After incubation with different doses of arsenic for one week, cells were collected and fixed with 70% ethanol at −20 °C overnight. The cell cycle distribution was analyzed after PI (40 μg/ml) staining for 1 h.
Western blotting
Cells were lysed in lysis buffer, and 50 μg of lyset was loaded onto a SDS-PAGE followed by to a PVDF membrane (Millipore, Billerica, MA, USA) transferring. Aurora-A, and β-actin levels were determined by anti-Aurora-A (Cell signaling, Boston, MA, USA) and anti-β-actin (Sigma) antibodies.
cDNA preparation and RT-PCR
Total RNA (1 μg) was used to prepare cDNA according to the manufacturer’s instructions (Improm-IITM Reverse Transcriptase; Promega, Madison, WI, USA). The cDNA (1 μg) was used for PCR according to the manufacturer’s instructions (YEAtaq DNA polymerase; Yeastern Biotech, Taipei, Taiwan). Primers were used as follows,
Aurora-A(F): GAAATTGGTCGCCCTC;
Aurora-A(R): TGATGAATTTGCTGTGATCC;
18 s rRNA(F): AAACGGCTACCACATCCAAG;
18 s rRNA(R): CCTCCAATGGATCCTCGTTA.
The plasmids, including pGL2-AAP (provided by Dr. Liang-Yi Hung), pRLTK (at the molar ratio of 10:1) and pCMV-E2F1 (provided by Dr. Ju-Ming Wang), were co-transfected into 1 × 105 E7 cells seeded in a twelve-well plate using Lipofactamin 2000™ (Invitrogen). After arsenic treatment, cell lysate was collected. The lysate in the eppendorf tubes was centrifuged at 13000 rpm at 4 °C for 1 min. LAR II was added to measure the firefly luciferase activity. Renilla luciferase activity was determined and used to normalize firefly luciferase activity.
Chromatin immunoprecipitation (ChIP) assay
ChIP assay was conducted following the manufacture’s protocol (Active Motif). Briefly, after arsenic treatment, cells (5 × 10
7) were fixed with 1% formaldehyde at RT for 5 min, followed by nuclei extraction. ChIP assay was conducted overnight at 4 °C using 7 μg of DNA and 3 μg of control IgG or E2F1 (C-20, Santa Cruz) after enzyme digestion of chromatin at 37 °C for 30 min. We used protein G magnetic beads (Active Motif) to capture the ChIP product, which was digested with proteinase K after reversion of cross-links for 15 min at 95 °C. The isolated DNA was used as the template for PCR amplification. The primers used for Aurora-A promoter were: site 1 (−268 to −80), sense 5′-TGGTCCGGTTCTCTTGGTAT-3′ and antisense 5′-AAGCTTGACGCATTGGAGAT-3′; site 2 (−104 to +106), sense 5′-CGACGCGTTGGCTCCACCACTTCCGG-3′ and antisense 5′-CCAGGAGCTCAGCCGTTAGAATTCAAAGG-3′ (He, [
22]).
E2F1 silencing using shRNA
The lentivirus of hairpin (sh) RNA against E2F1 (target sequence: 5′-cgtggactcttcggagaactt-3′) was from National RNAi Core Facility of Academia Sinica (Taipei, Taiwan). We infected the bladder cancer E7 cell with this sh RNA for 1 day. We then used 1 μg/ml puromycin for one week to screen stable clones.
Mice feeding
Four weeks-old female BALB/c mice (Laboratory Animal Center, National Cheng Kung University, College of Medicine, Tainan, Taiwan) were used. All animal experiment protocols were approved by the Laboratory Animal Committee at National Cheng Kung University. Mice were feed with different doses of arsenic (0 and 1 mg/L) in their drinking water. The mice were treated with arsenic containing water for two months. The mice were i.p. injected with BrdU (50 mg/kg) for 24 h before sacrificed. Their bladder, uterus, lung, and kidney were collected for further study.
Immunohistochemical staining
The paraffin embedded specimens were processed by deparaffinization and rehydration. The slides were incubated in distilled water for 3 min, and immersed in Tris-EDTA retrieval buffer (10 mM Tris Base, 1 mM EDTA, 0.05% Tween 20, pH 9.0). The slides were heated in a microwave for 5 min. Ice bath was used to cool down the slides to RT. The slides were rinsed in 1X PBS for three times 5 min each time, and blocked in 3% H2O2 (in methanol) for 10 min at RT. The slides were rinsed with 1X PBS for three times, and then immersed in blocking buffer (Thermo) for 30 min. After incubating with the primary antibody [rabbit anti-PCNA (#FL-261, Santa Cruz, Santa Cruz, CA, U.S.A.), mouse anti-BrdU (#RPN20AB) and mouse anti-Aurora-A antibody (NCL-L-AK2)] at 4 °C overnight, the slides were rinsed with PBS for three times followed by incubation with biotin linked secondary antibody (DakoCytomation, LSAB2 System-HRP, Glostrup, Denmark) for 10 min at RT. Again, the slides were rinsed with PBS three times, followed by adding Streptavidin reagent (DakoCytomation, LSAB2 System-HRP, USA) to cover the specimen for 10 min at RT. The slides were rinsed with 1X PBS three times followed by AEC solution treatment for 15 min at RT. The slides were rinsed with distilled water and counterstained with hematoxylin. The slides were then mounted with IHC Mounting Medium (DakoCytomation).
Statistical analysis
Student’s t test was used to analyze the significance of aurora-A promoter activation in the presence or absence of arsenic.
Discussion
The International Agency for Research on Cancer recognizes arsenic and arsenic compounds as group 1 carcinogens, and low concentration of arsenic could promote tumorigenesis [
19]. Increased expression of Aurora-A has been strongly associated with the aggressiveness of the tumor, but the factor(s) which induces the overexpression of Aurora-A is still unknown [
19]. Based on these evidence, we speculate that Aurora-A might be involved in the regulation of arsenic-induced tumorigenesis. To confirm our hypothesis, we set up the cell line and animal models to clarify the mechanism of arsenic induced Aurora-A overexpression. Our data showed that low concentration of arsenic induces Aurora-A promoter activity, mRNA and protein expression, indicating that arsenic induces Aurora-A overexpression at the transcription level. Aurora-A gene amplification was not detected under long-term, low dose of arsenic treatment. Furthermore, our mouse model reveals that long time exposure of low dose of arsenic could induce Aurora-A expression and uterus epithelial cell proliferation. However, no tumor formation was detected at the time period investigated.
Arsenic increases proliferation of the epithelial cells of various tissues including bladder and skin [
26‐
28]. Cell proliferation depends on an ordered and tightly regulated process known as cell cycle, involving multiple checkpoints which assess extracellular growth signals, cell size, and DNA integrity [
28]. In our data, increased S phase population, multiple centrosomes and increased cell proliferation suggest that normal cell cycle progression was interrupted by arsenic at low concentration. Consistent with other’s report. Arsenic delays cell cycle reentry into the G1 phase [
29]. Aberrant cell cycle progression is widely detected in many cancer cells. Therefore, the factors responsible for G1/S transition and S phase arrest are possibly involved in arsenic induced tumorigenesis [
28]. Among them, E2F family proteins and cyclins (cell cycle related proteins) are the potential candidates.
It is known that phosphorylated RB (p-RB) tumor suppressor can bind with the promoter of E2F family proteins and suppress their activity in normal cells [
30]. Here, the immortalized E7 cells containing HPV E7 oncogene, which can bind with p-RB [
24], may lead to the activation of E2F1 and increased Aurora-A expression. Nevertheless, the effect of E7 and p-RB interaction on low concentration arsenic upregulation of mRNA and protein expression of Aurora-A in the immortalized uroepithelial E7 cells can be excluded because similar events were detected in another immortalized skin cell line, HaCaT [
25].
Similarly, arsenic treatment of the mice for 2 months induced Aurora-A expression in uterus and bladder. Arsenic induces gene amplification and shows carcinogenic effects in humans have been reported [
31]. However, Aurora-A gene amplification was not detected in both bladder and skin cells (unpublished data). The possibility that longer time is needed to induce Aurora-A gene amplification cannot be excluded. It is known that inorganic arsenic disturbs the epigenetic regulation of various gene expression [
32]. In our study, low-dose arsenic treatment for 4 weeks showed no effect on DNA methylation status of Aurora-A promoter region. A total of ten samples were examined, including immortalized bladder and skin cell lines, arsenic-treated immortalized bladder cells, and clinical skin cancer specimens. The results from Methylation-specific PCR (MSP) and sequencing analysis suggest that all the cytosines in 28 CpG dinucleotides within the promoter examined are free of methylation, indicating that the expression of Aurora-A was not affected by the methylation status of the promoter. The distal region of the promoter needs to be analyzed for possibly arsenic-induced Aurora-A expression, because the analyzed region of Aurora-A is only about 1.5 kb upstream of the promoter,. The other possibility is acetylation. Our preliminary data showed that the acetylation status of lysine-9 and lysine-14 on histone H3 increased after arsenic treatment (data not shown). Whether histone acetylation is involved in arsenic induced tumorigenesis needs further confirmation.
Arsenite exerts dual effects: triggering apoptosis at relatively high concentration, whereas inducing partial differentiation at low concentration in leukemia cells. Protein/DNA array analysis showed that E2F was activated after 6 h exposure to 1 and 10 μM arsenite [
21]. Arsenic-increased activity of c-
myc and E2F-1 has been reported [
20]. However, there is no paper discussing the relationship between E2F1 and Aurora-A. Aurora-A acts as a direct target of E2F3 during G2/M cell cycle progression [
22]. Not only Aurora-A, but Aurora-B mRNA level is also regulated by cell cycle-dependent element (CDE) and cell cycle-gene homology region (CHR). A subset of E2F family proteins (E2F1 and E2F4) binds to the CDE [
33]. In this study and our previous report, we demonstrated that E2F1 increases Aurora-A promoter activity after 24 h arsenic treatment both in bladder and skin cells [
25]. In the presence of arsenic, Aurora-A expression in E2F1 silencing cells is similar to that in the cells without E2F1 silencing (Fig.
4e, lane 2 vs. lane 4). It is possible that Aurora-A induction was compromised by E2F3 transcription factor when E2F1 was silenced. It is also possible that silencing E2F1 partly affects arsenic alleviated Aurora-A expression as shown by reduced fold change (Fig.
4e, middle line and quantification diagram: 2× vs. 1.3 ×).
Furthermore, E2F1 could bind Aurora-A promoter to regulate its function. NFκB is a transcription factor which is activated by low concentration of arsenic (1 μM) [
21]. Arsenic exposure triggers PI3K/Akt/IKK/NFκB signal cascade which in turn plays essential roles in inducing cyclin D1 expression [
34]. Therefore, the relationship between Aurora-A and NFκB deserves further exploration [
35].
In our animal model, we investigated bladder and uterus after arsenic treatment because Yoshida et al. reported that chronic arsenic exposure may induce skin, lung, bladder, kidney, liver, and uterus malignancies [
36]. Our result showed that after arsenic treatment for 1 month, increase of cell proliferation and Aurora-A expression was detected in glands surrounding cells in uterus but not in bladder and skin (unpublished data). The possible reasons are as follows: first, the time of arsenic exposure is not long enough to induce Aurora-A expression and to increase cell proliferation. Longer exposure time may be necessary for induction of Aurora-A expression in many tissues such as bladder and skin. Second, the time for BrdU uptake is not enough in the bladder tissue, because BrdU was i.p. injected at 4 h before sacrifice. It may need longer time for BrdU uptake [
37]. Our speculation was confirmed after extending the arsenic treatment to 2 months and BrdU treatment to 24 h, both the uterus and bladder tissues showed increased cell proliferation and Aurora-A expression. Feeding the mice with arsenic in the drinking water mimics arsenic-induced tumorigenesis in nature. However, arsenic that acts as a co-carcinogen, does not induce tumor formation per se [
38]. Therefore, to reveal arsenic induced tumor formation in mice, some initiators must be introduced first or co-incubated with arsenic. Two mice models are proposed. Besides using UV or some known carcinogenic agents [
38‐
40] combined with arsenic, we injected low grade tumor cell line, MR4 (a MEF stable cell line, established in our laboratory), into the mice s.c. and combined with arsenic treatment for 1 month. The tumor formation, cell proliferation, and the protein expressions of Aurora-A between untreated or arsenic treated mice were investigated. Another model is transgenic mice, which carry a specific keratinocyte promoter to drive Ras oncogene expression specifically at epithelial cells.
Acknowledgements
We thank Dr. Liang-Yi Hung to provide pGL2-AAP plasmid and Dr. Ju-Ming Wang (Institute of Bioinformatics and Biosignal Transduction, College of Bioscience and Biotechnology, National Cheng Kung University) to provide pCMV-E2F1 plasmid. We thank Dr. Sheng-Hsiang Lin for providing the statistical consulting services (Biostatistics Consulting Center, National Cheng Kung University Hospital). We also thank Peter Wild for reviewing the manuscript.