Prevention of DNA damage in Barrett’s esophageal cells exposed to acidic bile salts

Summary

Components of gastroesophageal reflux, bile and acid, induce strong DNA damage in Barrett’s esophageal cells increasing the risk of cancer development. These tumorigenic alterations can be prevented by natural compounds that inhibit ROS and accelerate DNA damage repair.

Abstract

Esophageal adenocarcinoma (EA) is one of the fastest rising tumors in the USA. The major risk factor for EA is gastroesophageal reflux disease (GERD). During GERD, esophageal cells are exposed to refluxate which contains gastric acid frequently mixed with duodenal bile. This may lead to mucosal injury and Barrett’s metaplasia (BE) that are important factors contributing to development of EA. In this study, we investigated DNA damage in BE cells exposed to acidic bile salts and explored for potential protective strategies. Exposure of BE cells to acidic bile salts led to significant DNA damage, which in turn, was due to generation of reactive oxygen species (ROS). We found that acidic bile salts induce a rapid increase in superoxide radicals and hydrogen peroxide, which were determined using electron paramagnetic resonance spectroscopy and Amplex Red assay. Analyzing a panel of natural antioxidants, we identified apocynin to be the most effective in protecting esophageal cells from DNA damage induced by acidic bile salts. Mechanistic analyses showed that apocynin inhibited ROS generation and increases the DNA repair capacity of BE cells. We identified BRCA1 and p73 proteins as apocynin targets. Downregulation of p73 inhibited the protective effect of apocynin. Taken together, our results suggest potential application of natural compounds such as apocynin for prevention of reflux-induced DNA damage and GERD-associated tumorigenesis.

Introduction

Esophageal adenocarcinoma (EA) has overtaken other histological types of esophageal cancer in the USA with an incidence rate that is more rapid than other types of cancer (1,2). The major risk factor for EA is gastroesophageal reflux disease (GERD). Because of the disease, esophageal cells are exposed to a gastric refluxate that is frequently mixed with duodenal bile. This causes significant cellular and DNA damage and induces inflammation, which in turn, exacerbates the mucosal injury. If the damage persists, it can cause Barrett’s esophagus (BE), a condition in which the normal squamous epithelial lining is replaced by a metaplastic intestinal type of epithelium. Further damage of Barrett’s epithelium induced by reflux may cause tumorigenic alterations leading to EA (3–6).

Since Barrett’s esophagus is a precancerous lesion that may carry an increased risk of EA, preventive strategies have been proposed, including endoscopic and surgical ablative therapies. Current analyses suggest that BE patients with high-grade dysplasia may benefit from these procedures (7). However, ablative therapies do not warrant the risk and excessive cost for the vast majority of BE patients. The existing medical treatment for GERD almost exclusively uses drugs intended to suppress production of acid, such as proton pump inhibitors. There is no doubt that proton pump inhibitors provide symptomatic relief, but this treatment has a limited effect on EA (8). Proton pump inhibitor treatment does not correct the underlying reflux damage in the Barrett’s esophagus, and gastroesophageal reflux of weakly acidic material occurs in Barrett’s patients who take proton pump inhibitors (9–11). An attractive alternative for these patients is chemoprevention. The idea behind chemoprevention is to halt or reverse progression from pre-malignant lesions, such as BE and low-grade dysplasia to high-grade dysplasia or cancer (12). Natural compounds are of particular interest because of their safety, low toxicity, general acceptance as dietary supplements and lack of deleterious side effects.

Inactivation of p53 is a strong indicator of neoplastic transformation of BE (13,14). p53 protein is a regulator of important cellular processes such as DNA damage response, cell cycle control and DNA damage repair (15). Prolonged exposure to acidic bile salts has been shown to cause inactivation of p53 and transformation of Barrett’s cells (16) as well as p53 protein degradation (17). Other members of the p53 protein family, p73 and p63, have also been shown to play significant roles in regulation of DNA damage response and repair (18,19). Although much less is known about their roles in EA, recent studies have shown that levels and activities of p73 and p63 are altered by acidic bile salts (18,20).

In this study, we aim to understand the mechanisms of bile acid-induced DNA damage and determine interventions to prevent damage induced by gastroesophageal reflux.

Materials and methods

Cell cultures

Human telomerase-immortalized esophageal cell lines—CP-A and CP-B (both from American Type Culture Collection, Manassas, VA) and BAR-T (hTERT immortalized human Barrett’s metaplasia (BE) cells, a kind gift from Dr. Souza (21)) were cultured in keratinocyte SFM (KSFM) medium supplemented with 40 µg/ml bovine pituitary extract, 1.0ng/ml epidermal growth factor (Life Technologies, Carlsbad, CA) and 5% fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA). Human SV40 T antigen immortalized non-tumorous esophageal epithelial cells HET-1A (ATCC) were cultured in Dulbecco's modified Eagle's medium (Life Technologies) supplemented with 10% fetal bovine serum. Cell lines were authenticated and characterized by the suppliers (21). ATCC uses morphology, karyotyping and PCR-based approaches to confirm the identity of cell lines.

siRNA, antibodies and chemicals

p73 was inhibited either by lentiviral transduction with shRNA or transfection with siRNA, as described previously (22). Both approaches targeted the same p73 sequence (5′-GCAATAATCTCTCGCAGTA-3′) found in all p73 isoforms (22).

Cells were transfected with Lipofectamine 2000 (Invitrogen) following the manufacturers’ protocols.

The following antibodies were used: p53 and p21 were from Calbiochem (Billerica, MA); p73 from Bethyl (Montgomery, TX); p63(A4A) and p73(H-79) from Santa Cruz Biotechnologies (Dallas, TX); Rad51 from Abcam (Cambridge, MA); BRCA1 and the double strand breaks Repair Antibody Sampler Kit from Cell Signaling Technologies (Danvers, MA). The following chemicals were used: resveratrol (3,4′,5-Trihydroxy-trans-stilbene, 5-((1E)-2- (4-Hydroxyphenyl)ethenyl)-1,3-benzenediol), curcumin (1,7-bis(4-Hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione, Diferuloylmethane), apigenin (4′,5,7-Trihydroxyflavone,5,7-Dihydroxy-2-(4-hydroxyphenyl)-4-benzopyrone), epigallocatechin gallate (2-(3,4,5-Trihydroxyphenyl)-3,4-dihydro-1(2H)-benzopyran- 3,5,7-triol 3-gallate) and apocynin (4-hydroxy-3-methoxyacetophenone), all from Sigma-Aldrich (St. Louis, MO).

Acidic bile salts (BA/A) treatment

Cells were treated with a bile salt cocktail consisting of a 20-μM equimolar mixture of glycocholic, taurocholic, glycodeoxycholic, glycochenodeoxycholic and deoxycholic sodium salts (all reagents were from Sigma-Aldrich); total bile salt concentration was 100 μM. For cell treatment, the bile salt cocktail was diluted in Dulbecco's modified Eagle's medium, pH 4.0 (BA/A); pH was adjusted with HCl. Due to different sensitivity to bile salts, CP-A and Het-1A were treated with BA/A for 30min, CP-B for 15min and BAR-T for 5min.

DNA damage repair

We developed a novel assay for analyses of DNA damage repair in esophageal cells exposed to acidic bile salts based on approaches previously described by Iliakis et al. (23). Genomic DNA was collected from untreated and BA/A-treated cells using chloroform/isoamyl alcohol and then purified using the High Pure PCR Product purification kit from Roche (Branchburg, NJ). Nuclear proteins were extracted from control (Cntr) or apocynin-treated (Apo) cells. Briefly, control and apocynin-treated cells were collected and resuspended in the cytoplasmic buffer [HEPES 10mM pH 7.9, MgCl₂ 1.5mM, KCl 10mM, TritonX-100 0.01%, DTT 0.5mM, 1% protease inhibitors cocktail (Sigma-Aldrich)] at 4°C. The cells were lysed on ice for 20min with gentle vortexing. The lysates were centrifuged at 400g for 10min at 4°C and supernatants was collected as cytoplasmic fractions. The pellets were then washed once with cytoplasmic buffer and centrifuged for 5min. The pellets were then resuspended in the nuclei buffer (HEPES 10mM pH 7.9, MgCl₂ 1.5mM, KCl 400mM, EDTA 0.2mM, glycerol 25%, DTT 0.5mM, 1% protease inhibitors cocktail) and kept on ice for 20min. During incubation, samples were frozen and thawed four times to assist disruption of nuclear membrane and collection of nuclear proteins. The samples were then centrifuged at 21 000g for 40min and supernatant was collected as nuclei extract. Protein concentration was determined with the Bradford assay. For repair reaction, 5 µg of nuclear protein extract was incubated with 1 µg of DNA purified from BA/A-treated cells in the presence of reaction buffer (4mM HEPES-KOH, pH 7.5 at 37°C, KCl 2mM, MgCl₂ 0.3mM, dNTPs 10 μM, ATP 1.5mM and 1mM β-mercaptoethanol) for 24h. The reaction mixture was then separated on agarose gel to visualize high molecular weight and fragmented DNA. DNA damage repair was measured as a ratio of fragmented DNA in test and control (BA/A-treated only) samples using the ImageJ image analysis software (NIH).

Comet assay

Alkaline comet assay was performed to determine DNA damage induced by BA/A (24). In brief, the cells were collected after appropriate treatment, mixed with LM Agarose (Trevigen, Gaithersburg, MD) and allowed to solidify on flare comet slides (Trevigen) at 4°C. The slides were then immersed in lysis buffer (10mM Tris, pH 10, 2.5 M NaCl, 100mM EDTA, 1% Triton X-100, 10% DMSO) for 2h at 4°C. Electrophoresis was then performed on slides in alkaline conditions (300mM NaOH, 1mM EDTA, pH > 13). The slides were then washed and fixed with 100% cold ethanol, and visualized with Olympus BX41 fluorescent microscope (Olympus, Pittsburg, PA). Tail DNA content was measure in a minimum of 50 cells per treatment using the Open Comet software.

Analysis of DNA damage using p-H2AX antibody

Briefly, CP-A cells treated with acidic bile salts (100 uM for 30min) were collected 12h after BA/A treatment using RIPA buffer. The lysates were separated on SDS-PAGE gel, incubated with primary p-H2AX antibody (Millipore, Billerica, MA; 1:100) overnight at 4°C and analyzed by Western blotting.

Reactive oxygen species (ROS) assay

Cellular ROS levels were determined using 2′, 7′-dichlorofluorescin diacetate (DCFDA) (Sigma-Aldrich) dye according to manufacturers’ protocol. Briefly, 1×10⁶ cells were collected after appropriate treatment and stained with DCFDA dye (20 µM) for 30min at 37°C. ROS levels were analyzed using flow cytometry. Dead cells were detected using propidium iodide staining.

Measurements of intracellular superoxide using TMH spin probe and electron paramagnetic resonance (EPR)

Production of cellular superoxide was quantitatively measured by TMH spin probe and EPR (25). Spin probe stock solutions was prepared in ice cold 0.9% NaCl in the presence of the chelating agent [diethylene triamine penta-acetic acid, 0.2 mM]. Production of O₂· in cells was measured in Krebs HEPES buffer containing 5.8g/l NaCl, 0.35g/l KCl, 0.37g/l CaCl₂, 0.29g/l MgSO₄, 2.1g/l NaHCO₃, 0.14g/l K₂HPO₄, 5.21g/l Na-HEPES, 2g/l d-glucose, pH 7.3, in the presence of 0.1mM diethylene triamine pentaacetic acid. Cells were incubated for 30min at 37°C with 0.5mM TMH. The cells were then collected in 0.6ml of fresh Krebs HEPES buffer, snap frozen in liquid nitrogen and analyzed for EPR spectra. The signal was quantified using calibration with standard concentrations of the 3-carboxy-proxyl nitroxide. The portion of the signal due to O₂· was determined by pre-incubation of duplicate samples with PEG-SOD (100 U/ml) for 3h which inhibited more than 75% of nitroxide formation (26).

Detection of cellular hydrogen peroxide production by Amplex Red assay

Cellular H₂O₂ was measured by Amplex Red assay (Thermo Fisher Scientific) which is based on oxidation of the non-fluorescent molecule N-acetyl-3, 7-dihydroxyphenoxazine (Amplex Red) to resorufin (excitation at 530nm and emission at 590nm) (27). Amplex Red provides specific and sensitive H₂O₂ detection of extracellular H₂O₂. Since H₂O₂ is diffusible, Amplex Red assay provides an index of total cellular H₂O₂ production (28).

Immunohistochemistry

After institutional review board approval, 11 BE patients and 4 normal esophageal samples resected at Vanderbilt University Medical Center were histologically verified, and representative regions were selected for immunohistochemical analyses. The use of all human pathology specimens for research was approved by the Institutional Review Board (IRB) of Vanderbilt University Medical Center. Since only de-identified tissues were included in this retrospective study, the IRB has waived requirements for informed consent. Immunohistochemical staining was done using the following antibodies: p73(H-79) and p63(4A4) from Santa Cruz Biotechnology, and p53(DO-1) from Calbiochem. Specificity of staining was verified by omitting a primary antibody step in the protocol. Immunohistochemical results were evaluated for staining intensity. The intensity of staining was graded as 0 (negative), 1 (weak), 2 (moderate) and 3 (strong).

Statistics analyses

Statistical analyses were performed using the Student’s t-test. Results are expressed as averages ± SE, if not specifically indicated. Results were considered significant at values of P < 0.05.

Results

Acidic bile salts induce DNA damage and ROS production in esophageal cells

Since DNA damage in BE is an important contributing factor to tumorigenesis, we sought to understand the regulation of DNA damage by acidic bile salts (BA/A) in Barrett’s metaplastic cells. For our analyses, human BE cells, CP-A, were exposed to acidic growth medium (pH 4.0) supplemented with 100 uM bile salts cocktail for 30min. The composition, total bile salts concentration and pH were selected based on previous measurements that allowed us to mimic a typical episode of reflux in GERD patients (29–31). DNA damage in BA/A treated cells was analyzed using alkaline comet assay and phosphorylation of H2AX. We found significant increase in comet tail DNA content and phosphorylation of H2AX histone (p-H2AX) in BA/A treated cells compared to untreated controls (Figure 1A).

Figure 1.

BA/A induces DNA damage and accumulation of reactive oxygen species. (A) CP-A cells treated with BA/A (100 µM for 30min, pH 4.0) were analyzed for DNA damage using an alkaline comet assay. Graph represents levels of tail DNA content in untreated and BA/A treated cells analyzed 18h after BA/A treatment (*P < 0.05, n = 3) (top left panel). Right panels show representative images of comet formation in untreated and BA/A treated CP-A cells. BA/A induces phosphorylation of H2AX (bottom left panel) in CP-A cells (12h post-BA/A treatment). (B) CP-A cells treated with BA/A (100 µM for 30min, pH 4.0) were labeled with the DCFDA dye and analyzed for cellular ROS levels at the indicated time using flow cytometry. Graph represents alteration in ROS levels (*P < 0.05, n = 3). Bottom panel shows ROS levels in untreated and BA/A treated cells 1h after treatment with BA/A. (C) Similar to (B) but hydrogen peroxide (left panel) and superoxide (right panel) levels were analyzed using the Amplex Red assay and EPR spectroscopy respectively, at the indicated times (*P < 0.05, n = 3).

To understand the mechanism of DNA damage induced by acidic bile salts, we analyzed ROS in Barrett’s esophageal cells as previous reports have shown an induction of ROS in reflux conditions (32,33). CP-A cells were treated with BA/A and collected at the indicated time (Figure 1B). The cells were then stained with the DCFDA fluorogenic dye and analyzed for ROS using flow cytometry. We found that BA/A treatment significantly elevated ROS levels in CP-A cells. The ROS levels remained elevated for 3h, following which the levels returned to those observed in untreated cells. Since ROS constitutes a variety of reactive species, we then sought to determine how BA/A treatment affect the induction of the specific ROS. Amplex Red assay and EPR spectroscopy were used to determine the levels of hydrogen peroxide and superoxide radicals, respectively. Our analyses revealed a rapid increase in the levels of hydrogen peroxide and superoxide radical in cells treated with BA/A compared to untreated cells (Figure 1C). The levels of hydrogen peroxide (left panel) and superoxide radicals (right panel) returned to near baseline between 3 and 6h after BA/A treatment which is similar to total ROS levels profile determined with DCFDA dye (Figure 1B). Thus, our data show that exposure to acidic bile salts results in a rapid upregulation of ROS, such as hydrogen peroxide and superoxide radicals in BE cells.

Natural compounds inhibit BA/A induced ROS production and DNA damage

To further investigate the mechanism of BA/A induced DNA damage and develop a strategy to prevent it, we tested a panel of natural compounds that are known to have antioxidant properties. We selected natural compounds because of their safety, availability and low toxicity, making them ideal for drug development. CP-A cells were treated with 10 uM of either compound for 3h followed by washing to remove drug and short exposure to BA/A, as described above. The cells were then collected 1h post BA/A treatment, stained with DCFDA dye and analyzed for cellular ROS levels using flow cytometry. We found that pre-treatment with either apigenin, apocynin or EGCG significantly reduced the ROS levels (Figure 2A). Next, we investigated the effect of natural compounds on BA/A induced DNA damage using comet assay. Cells were collected 18h after BA/A treatment and analyzed for DNA damage. Compared to BA/A treated CP-A cells, we found significantly less damage in cell pre-treated with either resveratrol, apigenin or apocynin (Figure 2B). We confirmed our observations using BAR-T (another human Barrett’s esophageal cell line) and found that apigenin and apocynin significantly reduce DNA damage induced by BA/A in these cells (Figure 2C, top left panel). Our results suggest that inhibition of ROS protects esophageal cells from DNA damage induced by acidic bile salts. However, the protective effect varies between cell lines.

Figure 2.

Apocynin prevents DNA damage induced by acidic bile acids. (A) CP-A cells were treated with the indicated natural compounds (10 µM) for 3h followed by BA/A treatment (100 µM, pH 4.0) for 30min. The cells were labeled with the DCFDA dye and analyzed for cellular ROS levels using flow cytometry. Graph represents changes in DCFDA intensity relative to BA/A treated control (*P < 0.05, n = 3). Levels of ROS in BA/A treated cells was arbitrarily set at 60. (B) Similar to (A) but DNA damage was analyzed 18h after BA/A treatment using alkaline comet assay (*P < 0.05, n = 3). Representative comet images are shown. (C) DNA damage was analyzed in BAR-T (top left; 100 µM, pH 4.0 for 5min), CP-B (top right; 100 µM, pH 4.0 for 15min) and Het-1A (bottom left; 100 µM, pH 4.0 for 30min) cells (*P < 0.05, n = 3) using comet assay. Table 1 summarizes data on reducing DNA damage by antioxidants in tested cells lines. ‘X’ denotes antioxidants effective in tested cell line.

Further, we analyzed the protective effect of natural compounds on other esophageal cell lines (CP-B and Het-1A). Het-1A cells are immortalized normal squamous epithelial esophageal cells and CP-B cells were derived from Barrett’s patient with high grade dysplasia. Thus, tested cell lines represent different stages of progression to BE-associated EA. Interestingly, all natural compounds protected CP-B cells from BA/A induced DNA damage (Figure 2C, top right), while only apocynin significantly reduced DNA damage induced by BA/A in Het-1A cells (Figure 2C, bottom left). Taken together, our results show that although a number of antioxidants confer protection, apocynin is the most effective in reducing DNA damage induced by BA/A in esophageal cells (see summary table in Figure 2).

Apocynin reduces BA/A induced damage by promoting DNA repair

To investigate whether the protective effect of apocynin was solely due to inhibition of ROS production, CP-A cells treated with BA/A were allowed to recover for 12h and then treated with 1 uM apocynin for 3h. The 12-h time point represents the time by which the ROS levels in BA/A treated cells returned to near baseline, thus negating the protective effect of apocynin associated with inhibition of ROS. Using comet assay, we investigated the dynamics of DNA damage and found that cells treated with apocynin recover faster from DNA damage than control BA/A treated cells (18 and 24h, respectively) suggesting that apocynin enhances/accelerates the process of DNA damage repair (Figure 3A).

Figure 3.

Apocynin induces DNA damage repair. (A) CP-A cells treated with BA/A (100 µM, pH 4.0 for 15min) were allowed to recover for 12h followed by apocynin treatment (1 µM, 3h). Control cells were left untreated. DNA damage was measure using comet assay at the indicated times. Graph represents the percentage of tail DNA content (*P < 0.05, n = 3). (B) A schematic view of the in vitro repair assay protocol. This assay was used to analyze the effect of apocynin (Apo) on the repair capacity of esophageal cells. (C) Graph represents the percentage of fragmented DNA in test samples relative to DNA fragmentation in BA/A treated (BA/A) cells (*P < 0.05, n = 2) (top panel). A representative gel image is shown (bottom panel). DNA isolated from CP-A cells treated with camptothecin (CPT, 5 µM) was used as a positive control. Nuclear extract from control (Cntr) and Apo treated cells were separated on agarose gel as negative controls and to compare the DNA content in cellular extracts. DNA isolated from untreated (Un) CP-A cells showed a weak fragmentation pattern due to its mechanistic shearing during purification.

To test this hypothesis, we developed a novel in vitro repair assay that allows us to investigate the effect of apocynin on repair of damaged DNA in BA/A-treated cells in vitro. Its schematic representation is shown in Figure 3B. Genomic DNA from BA/A treated CP-A cells were purified as described in the Materials and Methods section and incubated with nuclear extracts (NE) from apocynin treated (Apo) or untreated control (Cntr) cells. Although both extracts were able to repair BA/A-treated DNA, nuclear extract from apocynin-treated cells was significantly more effective in reducing the amount of fragmented DNA than the corresponding extract from control untreated cells (Figure 3C). These results show that apocynin is able to enhance DNA damage repair of esophageal cells exposed to acidic bile salts independently of ROS inhibition.

To explore the mechanism of DNA damage repair in BA/A treated cells, we performed co-immunofluorescence analyses and assessed co-localization of DNA repair proteins in CP-A cells challenged with BA/A. We found that 53BP1 and p-H2AX proteins are co-localized in BA/A damaged cells, while no significant co-localization was observed between Rad51 and p-H2AX proteins suggesting that primarily NHEJ is activated by esophageal cells in response to acidic bile salts treatment (Supplementary Figure 1A, available at Carcinogenesis Online).

Apocynin affects expression of BRCA1 protein

To further characterize the role of apocynin in DNA repair, its effect on DNA repair proteins was analyzed using the double strand breaks Repair Antibody kit. CP-A and BAR-T cells were treated with apocynin (1 µM, 3h) in the presence or absence of BA/A treatment. We found that among tested DNA damage repair proteins, apocynin increases the levels of BRCA1 in both CP-A and BAR-T cells (Figure 4A). The levels of other DNA repair enzymes such as Rad50, Rad51 and KU80 were not increased by BA/A or apocynin treatment.

Figure 4.

Apocynin increases BRCA1 protein levels. CP-A and BAR-T cells were exposed to acidic bile salts (100 µM, pH 4.0) for 30min, allowed to recover for 12h, and then treated with 1 µM apocynin for 3h. Cellular lysates were collected after treatment with apocynin and analyzed for the indicated proteins using Western blotting.

Acidic bile salts and apocynin induce p73 protein

p53 and other members of the p53 family, p63 and p73, are involved in the regulation of DNA damage response. In addition, we and others have previously shown that expression of these proteins are altered in esophageal carcinoma (22). Hence, we sought to investigate the modulation of p53, p73 and p63 proteins by apocynin. We found strong induction of p73 protein in apocynin-treated CP-A and BAR-T cells (Figure 5A top panel and Supplementary Figure 1C, available at Carcinogenesis Online). In contrast to p73, no statistically significant changes of p53 protein levels were observed in CP-A cells, and only a weak increase in BAR-T cells (n = 3). Levels of p63 were low in both tested cell lines. Furthermore, our data show that treatment with BA/A further enhance the induction of p73 in esophageal cells (Figure 5A, compare lanes 2 with 4 and Supplementary Figure 1B, available at Carcinogenesis Online), suggesting that p73 may mediate the protective role of apocynin.

Figure 5.

Apocynin and acidic bile salts induce p73 protein. (A) CP-A cells were exposed to acidic bile salts (100µM, pH 4.0) for 30min, allowed to recover for 12h and then treated with 1 µM apocynin (Apo) for 3h. Lysates from untreated control and BA/A treated cells were prepared and analyzed for p53, p63 and p73 using Western blotting. p63 protein was not detected in CP-A cells. As a positive control for p63 protein detection, TE1 esophageal cells, which express p63 protein (22), were used. Lysates collected from CP-A cells treated with 5 µM camptothecin for 24h were used as a positive control for p53 activity. (B) Representative immunostainings for p53, p73 and p63 in esophageal squamous and BE epithelia. Upper panel: p73 protein is upregulated in the esophagi of GERD patients. Black arrows show nuclear p73 protein in the basal squamous epithelial layer and BE metaplastic cells. Middle panel: No significant changes in p53 protein levels were found in the esophagi of human patients with GERD. EA specimens with the mutant p53 gene were used as positive controls. Mutations in the p53 gene increase protein levels of p53. Bottom panel: p63 protein is strongly expressed in the basal layer of squamous esophageal epithelia (black arrows). No p63 staining was found in all cases of BE.

To further investigate these proteins in vivo, we assessed their expression in esophageal epithelium of control subjects and GERD patients using immunohistochemistry. All GERD patients analyzed (n = 11) had elevated protein levels of nuclear p73 compared with healthy individuals (n = 4). Eight out of 11 GERD patients showed strong to moderate p73 staining (3–2), while 3 out 11 patients showed a weaker p73 immunoreactivity between 1 and 2 (Figure 5B, top panels). Notably, strong increases in the p73 protein were observed in both squamous and metaplastic Barrett’s epithelia (Figure 5B). In contrast to p73, expression of the p53 protein was slightly elevated or undetectable in squamous and metaplastic Barrett’s epithelia (Figure 5B, middle panels). As a positive control for p53 staining, esophageal tumors with elevated levels of the p53 protein due to p53 gene mutations were used (Figure 5B; middle panel). The p63 protein was expressed in the basal layer of esophageal squamous epithelium in all patients (Figure 5B, bottom panels). No p63 staining was found in all cases of BE (Figure 5B), confirming previous studies on low expression of p63 protein in BE (34–36). Thus, our data show that induction of p73 occurs both in vitro and in vivo. However, it is noteworthy that additional factors such as inflammation may contribute to upregulation of p73 protein in Barrett’s patients.

Inhibition of p73 protein diminishes the protective effect of apocynin

To elucidate the role of p73 in apocynin-treated cells, we downregulated p73 in CP-A cells using shRNA (shp73). Then, p73 deficient and control cells transfected with scrambled shRNA (scr shRNA) were treated with BA/A as described above and analyzed for DNA damage using comet assay. We found that inhibition of p73 significantly enhances DNA damage induced by acidic bile salts (Figure 6A). Then, we investigated the effect of apocynin in p73-deficient cells. We found that apocynin reduces DNA damage in scr shRNA control cells treated with BA/A. However, its protective effect was significantly diminished in p73-deficient cells showing that p73 mediates activity of apocynin (Figure 6B; DNA damage in both scr shRNA and shp73 treated with BA/A was arbitrarily set at 100%). To investigate the underlying mechanisms, we downregulated p73 using specific siRNA and analyzed the levels of DNA repair proteins. We found that inhibition of p73 reduced the levels of total and phosphorylated form of BRCA1 (Figure 6C). Notably, mRNA levels of BRCA1 were not altered by downregulation of p73 suggesting that p73 regulate BRCA1 by non-transcriptional mechanisms (data not shown). Thus, our results show that the protective effect of apocynin is dependent on cellular p73 levels and inhibition of p73 sensitizes cells to BA/A induced DNA damage by affecting BRCA1 protein.

Figure 6.

Downregulation of p73 inhibits the protective effect of apocynin. (A) CP-A cells stably transfected with either p73 shRNA or scrambled shRNA (scr shRNA) were treated with BA/A (100 µM, pH 4, 30min) for 18h and analyzed for DNA damage using comet assay. The level of DNA damage in control cells (scr shRNA) treated with BA/A was arbitrarily set at 100%. (*P < 0.05, n = 3). Representative comet images are shown. Bottom panel shows p73 levels in analyzed cells. (B) CP-A cells stably transfected with p73 shRNA or scrambled shRNA were exposed to BA/A (100 µM, pH 4.0) for 30min and allowed to recover for 12h. Then the treated cells were either treated with 1 µM apocynin for 3h or mock treated. DNA damage was measured by comet assay (*P < 0.05, n = 3). DNA damage in both scr shRNA and shp73 treated with BA/A was arbitrarily set at 100%. (C) CP-A cells were transfected with siRNA against p73 or scrambled control (scr siRNA) for 48h and analyzed for BRCA1 and p73 proteins using Western blotting.

Discussion

GERD and Barrett’s esophagus are the most prominent risk factors for the development of EA. To better understand the tumorigenic alterations induced by GERD, we investigated cellular damage induced by gastroesophageal reflux in metaplastic Barrett’s cells. Our studies showed that exposure of BE cells to acidic bile salts at physiologically relevant concentrations and pH leads to significant DNA damage. These data are consistent with the previous reports on damage induced by acidic reflux (18,37). Our studies also strongly emphasize the role ROS in DNA damage as acidic bile salts were found to be potent inducers of ROS. Analyzing the specific ROS species, our studies reveal a rapid increase in hydrogen peroxide and superoxide radicals that are known for their ability to cause DNA damage (38,39).

Based on these findings, we screened a set of natural compounds, which are known to inhibit ROS. Experimental inhibition of ROS by these chemical compounds significantly reduced the levels of DNA damage in esophageal cells exposed to acidic bile salts. Interestingly, the preventive effect was varied with different antioxidants and cell lines, and did not correlate with their ability to scavenge ROS. Among tested antioxidants, apocynin, which is considered to be a weak ROS scavenger (40), most consistently reduced DNA damage in all analyzed esophageal cell lines.

Apocynin is a naturally occurring methoxy-substituted catechol that was originally isolated from the Canadian hemp (Apocynum cannabinum). This compound presents a promising potential treatment for some inflammatory, cardiovascular and neurodegenerative diseases and is known for its ability to inhibit ROS production (41–43). However, our studies revealed that apocynin is able to reduce BA/A-induced DNA damage even in post-BA/A treatment conditions when ROS levels were significantly decreased, suggesting that its protective effect cannot be explained solely by its ability to suppress ROS. To further investigate this phenomenon, we developed a novel assay that measures the repair of genomic DNA that was damaged by acidic bile salts. Using this approach, we found that treatment with apocynin not only suppresses ROS but also enhances the DNA repair capacity of esophageal cells. Apocynin significantly accelerated the repair of damaged DNA. Among DNA repair proteins, we were able to identify BRCA1 as a target of apocynin. Apocynin increases protein levels of BRCA1 (44). However, we cannot exclude that additional DDR proteins may also be involved.

Given the important role of the p53 protein family in DNA damage response and repair, we assessed the expression of the entire p53 family in condition of cellular damage induced by acidic bile salts. We found that the p73 protein is strongly up-regulated by acidic bile salts in all tested cell lines. Upregulation of p73 was also observed in the esophagi of GERD patients both in squamous and Barrett’s epithelia. In striking contrast to p73, expression of p53 was slightly increased or undetectable in GERD patients. No p63 staining was found in all cases of BE. BE cell lines also showed a low expression of p63 protein. Thus, among the p53 protein family, p73 is induced by reflux in esophageal tissues and BE cell lines after treatment with BA/A.

Our data show that p73 helps to maintain the DNA integrity as its experimental downregulation enhances DNA damage in cells exposed to acidic bile salts. p73 also mediates, at least in part, the protective effect of apocynin. When p73 protein was downregulated, apocynin became less effective in suppression of DNA damage induced by acidic bile salts. Downregulation of p73 also decreased protein levels of BRCA1 and pBRCA1. Interestingly, BRCA1 mRNA levels were not significantly changed, suggesting that p73 affects BRCA1 at the post-transcriptional level. Further studies are needed to investigate this regulatory mechanism.

In summary, a prolonged exposure to gastric acid and duodenal bile salts leads to excessive DNA damage, which causes accumulation of tumorigenic alterations and progression to EA. Mechanistically, acidic bile salts induce ROS such as hydrogen peroxide and superoxide in esophageal cells, which in turn, rapidly induce DNA damage. We found that DNA damage can be significantly reduced by suppressing ROS with natural compounds. Apocynin was found to be the most effective compound as it has a dual effect: reduction of ROS and facilitation of DNA damage repair. This process is mediated by p73 protein, which protects cells against DNA damage induced by acidic bile salts. Collectively, our studies support the concept that natural compounds can be used to prevent DNA damage and tumorigenic transformation induced by GERD.

Supplementary material

Supplementary Figure 1 can be found at http://carcin.oxfordjournals.org/

Funding

National Cancer Institute RO1 CA206564, the Department of Veteran Affairs BX002115, National Cancer Institute RO1 138833, Vanderbilt-Ingram Cancer Center P30 CA68485 and the Vanderbilt Digestive Disease Research Center DK058404.

Glossary

Abbreviations

BE

Barrett’s metaplasia

EA

esophageal adenocarcinoma

EPR

electron paramagnetic resonance

GERD

gastroesophageal reflux disease

ROS

reactive oxygen species