NKX6.3 Regulates Reactive Oxygen Species Production by Suppressing NF-kB and DNMT1 Activities in Gastric Epithelial Cells

NKX6.3 plays an important role in gastric epithelial differentiation and also acts as a gastric tumor suppressor. The specific aim of this study was to determine whether NKX6.3 contributes to gastric mucosal barrier function by regulating reactive oxygen species (ROS) production. NKX6.3 reduced ROS production and regulated expression of anti-oxidant genes, including Hace1. In addition, NKX6.3 reduced DNMT1 expression and activity by down-regulating NF-kB family gene transcription. Silencing of Hace1 recovered ROS production, whereas knock-down of DNMT1 and NF-kB reduced ROS production and induced Hace1 expression by hypomethylating its promoter region. In addition, NKX6.3 inhibited CagA effects on cell growth, ROS production, and NF-kB and DNMT1 activity. In gastric mucosae and cancers, NKX6.3 and Hace1 expression was significantly reduced. The NKX6.3 expression was positively correlated with Hace1 and Nrf2 genes, but negatively correlated with DNMT1. Hypermethylation of Hace1 gene was observed only in gastric mucosae with H. pylori, atrophy and intestinal metaplasia. Thus, these results suggest that NKX6.3 inhibits ROS production by inducing the expression of Hace1 via down-regulating NF-kB and DNMT1 activity in gastric epithelial cells.

NKX6.3, Hace1, Nrf2, and DNMT1 was compared between non-neoplastic gastric mucosae and gastric cancers. Overall, we found that NKX6.3 significantly decreased ROS production and regulated the expression of ROS-related genes, including Hace1, by suppressing DNMT1 and NF-kB activity. These results suggest that NKX6.3 plays an important role in gastric epithelial protection and gastric cancer prevention by regulating the ROS levels.

NKX6.3 attenuates ROS production by regulating ROS-responsive genes.
To determine whether NKX6.3 contributes to ROS production, we performed in vitro ROS analysis in AGS Mock , MKN1 Mock , AGS NKX6. 3 and MKN1 NKX6.3 cells using DCFDA staining assay. Stable NKX6.3 expression in AGS and MKN1 cells showed reduced cellular ROS levels in a time-dependent manner, as compared to mock stable AGS and MKN1 cells ( Fig. 1A-C). To further confirm these initial observations, we measured the expression levels of ROS-related genes, including Hace1, Nox1, Noxa1 and Nrf2, and oxidative stress-responsive genes, including catalase, mnSOD, GSH, Nqo1 and Ho-1 by real-time RT-PCR and western blotting. Interestingly, NKX6.3 induced the expression of Hace1, Nrf2, catalase, and mnSOD, while decreasing the expression of Nox1 and Noxa1 proteins. In addition, it also reduced the expression of GSH, Nqo1 and Ho-1 at the mRNA level ( Fig. 1D and E). To further support our results, we recapitulated NKX6.3, Hace1 and Nrf2 gene expression patterns from large cohorts of gastric cancer patients available from the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database (accession numbers GSE27342). Interestingly, NKX6.3 expression was positively correlated with Hace1 and Nrf2 expression in gastric cancer cohort (Supplementary Figure S1A). In addition, Hace1 and Nrf2 expression showed a positive correlation (Supplementary Figure S1A). Next, we aimed to determine whether the anti-oxidant activity of NKX6.3 is dependent on Hace1 expression. In AGS and MKN1 cells, treatment with siHace1 partially recovered ROS production ( Fig. 2A) and markedly reduced Nrf2 expression (Fig. 2B). Also, Hace1 silencing in NKX6.3 stable cells decreased GSH, Nqo1 and Ho-1 mRNA expression (Fig. 2C) and showed moderate ablation of NKX6.3-induced cell growth inhibition (Fig. 2D). Thus, it is likely that NKX6.3 attenuates ROS production by Hace1-dependent or -independent regulation of ROS-mediated gene expression.

NKX6.3 induces expression of Hace1-HECT E3 ligase. The Hace1-HECT E3 ligase is a tumor sup-
pressor that directly regulates ROS production, and its reduced expression due to promoter hypermethylation is frequently observed in several cancers 17,18 . Since NKX6.3 induced re-expression of Hace1 protein (Fig. 1D), we hypothesized that NKX6.3 functions as a hypomethylating agent. Expectedly, expression of Hace1 mRNA was markedly increased in NKX6.3 stable transfectants (Fig. 2E). Notably, hypermethylation of CpG islands on Hace1 promoter was observed in mock stable AGS and MKN1 cells, whereas de-methylation on Hace1 promoter was detected in NKX6.3 stable transfectants (Fig. 2F). Next, we evaluated the regulatory role of NKX6.3 in DNMT1 expression. As shown in Fig. 3A, NKX6.3 significantly down-regulated expression of DNMT1 mRNA and protein in AGS NKX6.3 and MKN1 NKX6.3 cells. Knockdown of NF-kB, a transcription factor for DNMT1 19 , with siNF-kB markedly decreased DNMT1 expression at the mRNA and protein levels (Fig. 3B). ChIP assay results indicated that p50 binding to the DNMT1 gene promoter was significantly inhibited by NKX6.3, comparable with the effect of siNF-kB (Fig. 3C). In addition, NKX6.3 expression as well as NF-kB silencing significantly decreased DNMT1 activity (Fig. 3D). Knock-down of DNMT1 with shDNMT1 increased Hace1 mRNA and protein expression (Fig. 3E) and induced de-methylation of Hace1 gene in AGS and MKN1 cells (Fig. 3F). These findings collectively suggest that NKX6.3 induces Hace1 expression by suppressing DNMT1 activity.
Hace1 regulates NF-kB expression and activity. Next, we determined the regulatory role of Hace1 in NF-kB expression. As shown in Figure 5A-D, silencing of Hace1 with siHace1 increased expression of p65, p50 and DNMT1 at the mRNA and protein levels in NKX6.3 stable transfectants. In addition, DNMT1 activity was also increased in siHace1 transfected AGS NKX6.3 and MKN1 NKX6.3 cells (Fig. 5E), suggesting that Hace1 is involved in NF-kB inactivation.
H. pylori CagA is an important factor for ROS production. Here, we examined the effects of H. pylori CagA on ROS production and expression of ROS-related genes. Expectedly, CagA significantly increased cell growth and ROS production in AGS and MKN1 cells, whereas NKX6.3 ablated the effects of CagA ( Fig. 6A-C). To further confirm that CagA induces ROS production, we examined ROS levels in AGS cells by using the strain of H. pylori with or without CagA. As shown in Figure 6D, H. pylori with CagA significantly increased ROS production, but H. pylori without CagA did not affect ROS levels in AGS cells. Additionally, CagA increased the expression of NF-kB p65, p50 and DNMT1 proteins and induced loss of Hace1 expression, whereas NKX6.3 markedly inhibited CagA effects on protein expression (Fig. 6E). Furthermore, NKX6.3 suppressed CagA-induced DNMT1    the large cohorts of gastric cancer patients available from the NCBI GEO database (accession numbers GSE27342; Supplementary Figure S1).
In 55 non-neoplastic gastric mucosae, DNMT1 and Hace1 expression was increased and reduced, respectively, in gastric mucosae with H. pylori infection, atrophy and intestinal metaplasia ( Fig. 7F and G). Non-neoplastic gastric mucosae also showed positive correlation between NKX6.3 and Hace1 expression and inverse correlation between DNMT1 and NKX6.3 expression (Fig. 7H and I). The methylation status of Hace1 promoter region was examined by MSP to confirm that reduced Hace1 mRNA expression in gastric mucosae is caused by hypermethylation of Hace1. Interestingly, Hace1 hypermethylation was observed only in the gastric mucosae with H. pylori infection, atrophy, and/or intestinal metaplasia (Fig. 7J), suggesting that NF-kB-induced DNMT1 expression caused by NKX6.3 inactivation may reduce Hace1 expression by hypermethylating the promoter region of Hace1 in the gastric mucosa.

Discussion
Gastric cancers develop as a consequence of chronic inflammation from persistent mucosal or epithelial cell colonization by microorganisms, including H. pylori 20 . Chronic inflammatory cells such as macrophages/monocytes, lymphocytes, and plasma cells are present in the gastric mucosa of chronic gastritis and lead to the generation of several ROS and RNS 21 . Persistent ROS can damage cellular DNA, RNA, and proteins by chemical reactions, which subsequently cause proto-oncogene activation, oncogene/tumor suppressor gene mutations, and chromosomal aberrations 22,23 . In stomach, H. pylori and ROS collaborate in the gastric epithelium to activate the NF-κB and AP-1 transcription factors that up-regulate the expression of chemokines, including IL-8 24,25 , and in turn, NF-kB-dependent genes play a major role in regulating the cellular ROS levels 26 . Recently, we and others reported   that NKX6.3 acts as a master regulator in gastric differentiation and proliferation [12][13][14] . Since the gastric mucosal barrier may protect the host from potentially harmful agents to maintain cell survival, we focused on the role of NKX6.3 in the protection of the gastric mucosal epithelia from harmful ROS.
It is well known that Hace1 functions as an important component of the cellular ROS detoxification and anti-oxidative stress responses by facilitating optimal activation of Nrf2 27 . Here, we showed that NKX6.3 reduced intracellular ROS and modulated the expression of ROS-related genes (Fig. 1). Of these, NKX6.3 induced expression of Hace1 and Nrf2 proteins in gastric cancer cells (Fig. 1D) and showed positive correlation with Hace1 and Nrf2 in the gastric cancer cohort (Supplementary Figure S1). In addition, Hace1 silencing with siHace1 recovered ROS production and reduced Nrf2, GSH, Nqo and Ho-1 expression ( Fig. 2A-C). Furthermore, knock-down of Hace1 ablated NKX6.3-induced cell growth inhibitory activity (Fig. 2D). Taken together, these results suggest that NKX6.3 plays an important role in suppression of ROS production by regulating expression of ROS-related genes, especially Hace1.
NKX6.3 increased Hace1 expression and was positively correlated with Hace1, suggestive of direct regulation of Hace1 expression. Since Hace1 expression is significantly reduced by promoter hypermethylation in most cancer patients 17 , we investigated whether DNMT1 regulates Hace1 expression in gastric epithelial cells. NF-kB reportedly binds to one possible NF-kB binding element in the DNMT1 promoter and DNMT1 mediates NF-kB-dependent p16 gene promoter hypermethylation 19 . Here, we showed that NKX6.3 induced Hace1 promoter demethylation and increased its expression (Fig. 2E and F) by down-regulating DNMT1 expression via inhibiting p50 binding in the promoter region of DNMT1 gene (Fig. 3C). In addition, the effect of NKX6.3 on DNMT1 activity was comparable with that of siNF-kB treatment (Fig. 3D). Furthermore, DNMT1 silencing with siDNMT1 led to demethylation of Hace1 and increased its expression at the mRNA and protein levels ( Fig. 3E and F). These results suggest that NKX6.3 induces Hace1 expression by inhibiting DNMT1 expression and activity.
It has been reported that H. pylori induces the production of ROS and DNA damage in gastric epithelial cells and frequently causes chromosomal aberrations 33,34 . Previously, we showed that H. pylori CagA increased the expression of NF-kB proteins and ROS levels in gastric cancer cells 35 . In the current study, CagA significantly increased cell growth and ROS production in AGS Mock and MKN1 Mock cells, whereas NKX6.3 ablated these CagA effects by down-regulating NF-kB p65, p50 and DNMT1 and up-regulating Hace1 expression (Fig. 6A-F). These findings suggest that NKX6.3 may counteract CagA-induced NF-kB activity in gastric epithelial cells.
Damage to cellular components results in increased mutations and altered functions of important proteins in premalignant tissues, thereby contributing to the multi-stage carcinogenesis 36 . Chronic inflammation of gastric mucosa triggers a pathway of chronic gastritis, atrophy, intestinal metaplasia, dysplasia, which finally progress to gastric cancer 37 . H. pylori, especially CagA strains, is considered as the most important risk factor of atrophy and intestinal metaplasia 38,39 . In gastric cancer tissue, NKX6.3 and Hace1 expression showed a positive correlation, while DNMT1 expression was inversely correlated with NKX6.3 and Hace1 (Fig. 7A-E). The NCBI GEO database also showed reduced NKX6.3, Hace1 and Nrf2 expression and a positive correlation between these genes (Supplementary Figure S1). In non-neoplastic gastric mucosae with H. pylori infection, atrophy and intestinal metaplasia, DNMT1 mRNA was increased, whereas Hace1 mRNA was reduced. (Fig. 7F and G). NKX6.3 expression was positively and inversely correlated with Hace1 and DNMT1, respectively (Figure H and I). Interestingly, Hace1 hypermethylation was observed only in the gastric mucosae with H. pyori infection, atrophy, and/or intestinal metaplasia (Fig. 7J). Thus, NKX6.3 inactivation in gastric mucosa may increase activity of NF-kB and DNMT1 and reduce Hace1 expression, subsequently progressing to atrophy, intestinal metaplasisa and cancer.
In conclusion, NKX6.3 induced Hace1 expression by suppressing DNMT1 expression and activity via down-regulating the NF-kB signaling pathway. In addition, NKX6.3 ablated CagA effects on cell proliferation, ROS production, and activities of DNMT1 and NF-kB signaling pathway. Overall, we conclude that NKX6.3 protects gastric mucosal epithelia by regulating harmful ROS production. Modulation of NKX6.3 anti-oxidant activity could contribute to the prevention of precancerous changes in gastric mucosa and gastric cancer.

Materials and Methods
Cell culture and generation of NKX6.3 stable cells. AGS  Gastric mucosa and cancer tissue specimens. A total of 65 patients with sporadic gastric cancer who underwent a gastrectomy at the Chonnam National University Hwasun Hospital were enrolled. In addition, a total of 55 non-neoplastic frozen gastric mucosa remote (>5 cm) from gastric cancer were included in this study. Adjacent gastric mucosal tissues to each frozen specimen were also fixed in formalin and stained with hematoxylin-eosin. Informed consent was provided according to the Declaration of Helsinki. Written informed consent was obtained from all subjects. The study was approved by the Institutional Review Board of The Catholic University of Korea, College of Medicine (MC15SISI0015). There was no evidence of familial cancer in any of the patients.
Histological assessment of 55 non-neoplastic gastric mucosae was performed independently by two pathologists. According to the updated Sydney system 40,41 , gastritis was determined by polymorphonuclear leukocyte infiltration, mononuclear cell infiltration, glandular atrophy and intestinal metaplasia, as previously described 14 .
Infection with a CagA-positive strain of H. pylori was determined by the presence of CagA protein in 55 gastric mucosa tissue samples using western blot analysis, as described previously 35 .

Measurement of cell viability and proliferation.
Cell viability was determined in AGS and MKN1 gastric cancer cells after treatment with siHace1 or H. pylori CagA transfection. To assess cell viability, a MTT [3-(4,5 dimethylthiazol-2-yl)−2,5-diphenyltetrazoliumbromide] assay was performed at 24, 48, 72, and 96 hrs following transfection with siHace1 and CagA in mock, AGS Mock and MKN1 Mock cells, and NKX6.3 stable transfectants, AGS NKX6.3 and MKN1 NKX6.3 cells. Absorbance was measured at 540 nm using a spectrophotometer and viability was expressed relative to the mock control.
For cell proliferation assay, a BrdU incorporation assay was performed at 24, 48, 72, and 96 hrs following transfection with CagA in mock and NKX6.3 stable cells using the BrdU cell proliferation assay kit (Millipore, Billerica, MA, USA), according to the manufacturer's instruction. Absorbance was measured using a spectrophotometer at 450 nm and proliferation was expressed relative to the mock control.

Expression of ROS-related genes in gastric cancer cell lines and tissues. Expression of NKX6.3,
Hace1, and DNMT1 was examined in 65 frozen gastric cancers, corresponding non-cancerous gastric mucosal tissues and 2 gastric cancer cell lines by real-time RT-PCR and Western blot analysis. After quantification of mRNA extracted from cancer tissues and non-cancerous gastric mucosae, cDNA was synthesized using the reverse transcription kit from Roche Molecular System (Roche, Mannheim, Germany), according to the manufacturer's protocol. For QPCR, 50 ng cDNA was amplified using Fullvelocity SYBR Green QPCR Master Mix (Stratagene, La Jolla, CA, USA) and 20 pmol/μl of each primer (forward and reverse) using Stratagene Mx 3000 P QPCR system, according techniques previously published 39 . NKX6.3, Hace1, and DNMT1 mRNAs were quantified by SYBR Green Q-PCR and normalized to mRNA of the β-actin. Sequences of the primers are described in Supplementary  Table S1. Data are reported as relative quantities according to an internal calibrator using the 2 −△△CT method 42 . All samples were tested in duplicate, and the mean values were used for quantification. In addition, the relation between expression levels of ROS-related genes and gastritis parameters, such as atrophy, intestinal metaplasia and H. pylori infection was also examined in 55 non-neoplastic gastric mucosae.
To investigate whether ablation of NKX6.3 is associated with ROS production, the expression of NF-kB, Hace1, DNMT1, GSH, Nqo1, and Ho-1 mRNA were analyzed using real-time RT-PCR in AGS Mock , MKN1 Mock , AGS NKX6.3 , and MKN1 NKX6.3 cells. The effects of Hace1 silencing with siHace1 on the expression of ROS-related genes were also examined. To further confirm that NKX6.3 regulates NF-kB activity, we analyzed mRNA transcript expression of IL-6, IL-8, and TNF-α, which are downstream targets of the NF-kB 43 . Each mRNA was quantified by SYBR Green QPCR and normalized to mRNA of the housekeeping gene, β-actin. The primer sequences are described in Supplementary Table S1.
For Western blot analysis, the samples were ground to very fine powder in liquid nitrogen using a pestle and mortar, suspended in an ice-cold Nonidet P-40 lysis buffer supplemented with a 1x protease inhibitor mix (Roche). Cell lysates were separated on 10% polyacrylamide gel and blotted onto a Hybond-PVDF transfer membrane (Amersham), which had been subsequently probed with specific antibodies, and then incubated with anti-mouse IgG conjugated with horseradish peroxidase. The list of antibodies is described in Supplementary  Table S2. The protein bands were detected using enhanced chemiluminescence Western blotting detection reagents (Amersham).
Methylation status of Hace1. Methylation analysis was carried out in 55 frozen non-neoplastic gastric mucosae and 2 gastric cancer cell lines after transfection with NKX6.3. Methylation status of the promoter region of the Hace1 gene was determined using sodium bisulfite treatment of the DNA followed methylation specific PCR (MSP), as described in the literature with minor modifications 17 . 5 μl of the bisulfite-modified DNA was subjected to MSP using two sets of primers for methylated and unmethylated Hace1. The primer sequences are described in Supplementary Table S1. PCR was performed in a total volume of 30 μl, containing 5 μl of the template DNA, 0.5 μM of each primer, 0.2 μM of each dNTP, 1.5 mM MgCI 2 , 0.4 unit of Ampli Taq gold polymerase (Perkin-Elmer) and 3 μl of 10X buffer. The reaction solution was initially denatured for 1 min at 95 °C. Amplification was carried out for 40 cycles of 30 s at 95 °C, 30 s at 58 °C and 30 s at 72 °C, followed by a final 5 min extension at 72 °C. Each PCR product was loaded directly onto 2% agarose gels, stained with ethidium bromide and visualized under UV illumination.
Generation of CagA gene deleted H. pylori strains. The isogenic mutant H. pylori 26695 (∆cagA::aphA), in which most of the cagA gene was replaced by a aphA (kanamycin resistant gene from pIP1433) cassette, was made using PCR products generated with primers "kanF" (5′-GATAAACCCAGCGAACCAT) a n d "a p h A R" ( 5 ′ -G TA C TA A A A C A AT T C AT C C A G TA A ) ( 1 4 0 2 b p ; a p h A k a n a m ycin resistance cassette); "cagA F1" (5′ -ATCGT TGATAAGAACGATAGGG)and "cagA R2" (5′-ATGGTTCGCTGGGTTTATCATTGATTGCTTCTTTGACATCGGTACCAAGCGACCCAAATAG) (552 bp, upstream of deleted cagA segment); "cagA F5" ( 5′T TA CT GG AT GA AT TG TT TT AG TA CA TC AA AT  AG CA AG TG GT T T GG GA ATGACCTACT TAACAAAATCATG-) and "cagA R6" (5′ -AT TGCT ATTAATGCGTGTGTGG) (425 bp; downstream of deleted cagA segment). Natural transformation was carried out by adding 7 μl of purified PCR product containing this ΔcagA::aphA allele to a lawn of cells (wild type H. pylori 26695) growing exponentially on nonselective medium, and re-streaking the population on selective medium (containing 15 μg/ml of kanamycin) after 6-8 hrs or overnight incubation to obtain transformant colonies. PCR tests and sequencing of representative kan r transformants demonstrated expected replacement of cagA by aphA in each case.
Reactive oxygen species (ROS) analysis. ROS levels were determined in mock or NKX6.3 stable AGS and MKN1 cells using 2′-7′-dichlorodihydrofluorescein diacetate (DCFDA). The AGS and MKN1 cells were incubated with 10 uM DCFDA at 37 °C for 20 min and rinsed twice with cold PBS, then trypsinized and subjected to FACScan flow cytometer. To determine the effects of H. pylori CagA on ROS production, the ROS levels were examined in H. pylori with/without CagA-infected AGS cells at 12 hrs as well as in mock and NKX6.3 stable AGS and MKN1 cells at 72 hrs after transfection with H. pylori CagA. DCF fluorescence was measured by FACS analysis and intensity was plotted against the number of cells. Additionally, cells were incubated with 10 uM DCFDA at 37 °C for 20 min and quickly washed with cold PBS, and photos of representative fluorescent fields were taken under an inverted microscope.
Chromatin immunoprecipitation (ChIP) analysis. For assessing the NKX6.3 binding activity in the promoter region of NF-kB p65 and p50, ChIP assays were performed using the Thermo Scientific Pierce Agarose ChIP kit (Thermo Scientific Pierce, Rockford, IL, USA), as previously described 13 . Briefly, AGS Mock , MKN1 Mock , AGS NKX6.3 and MKN1 NKX6.3 cells were cultured in a 10-cm dish for 4 days. The cells were fixed with 1% formaldehyde in PBS for 10 min, washed twice with ice-cold PBS and re-suspended in lysis buffer. Nuclei were recovered by centrifugation and MNase digestion was carried out at 37 °C for 15 min. Nuclei were lysed and the extracts were immunoprecipitated with 4 µg of antibody against NKX6.3 at 4 °C overnight. Normal rabbit IgG was used as the negative control. Protein-bound DNA was recovered using affinity chromatography purification columns according to the manufacturer's protocol (Thermo Scientific), and 5 µl of lysed nuclei were also purified under the same procedure and used as input. DNA amplification was performed by PCR using specific primers for the promoters of NF-kB p65, p50 and DNMT1 genes, as described in Supplementary Table S1. Amplification products were separated on a 2% agarose gel.
Measurement of DNMT1 activity. The DNMT1 activity was analyzed using the DNMT1 activity assay kit (Abcam) according to manufacturer's instructions. Briefly, AGS and MKN1 cells were collected and suspended in PBS. After centrifugation, the pellet was lysed in lysis buffer (10 mM Tris-Hcl pH 7.5, 10 mM NaCl, 2 mM MgCl 2 ) containing protease inhibitor mixture (Complete; Roche Molecular Biochemicals). Then, 6 ul of 20% NP-40 was added and the mixture was incubated for 10 min at 4 °C and centrifuged for 5 min at 3000 rpm. The supernatant was collected and the pellet containing the nuclei was resuspended in 50 μl of extraction buffer (20 mM Hepes pH 7.9, 420 mM NaCl, 1.5 mM Mgcl 2 , 0.2 mM EDTA and 10% glycerol) followed by incubation for 30 min at 4 °C and collection of the nuclear extract by centrifugation. All reactions were carried out in triplicate.
Statistical analysis. Student's t-test was used to analyze the effects of NKX6.3 on cellular ROS levels, cell viability and mRNA expression changes. Two-way ANOVA test was used to analyze the expression of DNMT1 and HACE1 in normal gastric mucosae and gastric cancer tissues. Linear progression test was used to analyze correlations between DNMT1, HACE1 and NKX6.3 expression levels in normal gastric mucosae and gastric cancer tissues. Chi-square test was used to analyze correlations between H. pylori CagA, atrophy, intestinal metaplasia and methylation status of the HACE1 promoter. Data are expressed as means ± S.D. from at least three independent experiments. A P-value less than 0.05 was considered to be the limit of statistical significance. All experiments were performed in triplicate to verify the reproducibility of the findings.