B-cell translocation gene 2 (BTG2) is a member of an emerging gene family that is involved in cellular functions. In this study, we demonstrate that BTG2 regulates glucose homeostasis via upregulation of Nur77 in diabetic mice. Hepatic BTG2 gene expression was elevated by fasting and forskolin. Overexpression of Btg2 increased the expression of hepatic gluconeogenic genes and blood glucose output and subsequently impaired glucose and insulin tolerance. Upregulation of the transcriptional activity of Nur77, gluconeogenic genes, and glucose production by forskolin was observed by Btg2 transduction, but not in Btg2 knockdown. BTG2-stimulated glucose production and glucose-6-phosphatase promoter activity were attenuated by dominant-negative Nur77. Coimmunoprecipitation and chromatin immunoprecipitation assays showed that BTG2 induced Nur77 occupancy on the glucose-6-phosphatase promoter via a physical interaction. Btg2 gene expression was increased in streptozotocin-treated and db/db mice. Finally, impairment of glucose homeostasis, such as the increase of blood glucose, glucose intolerance, and insulin intolerance, was elevated in diabetic mice, whereas this phenomenon was abolished in knockdown of Btg2. Together, these data suggest that BTG2 participates in the regulation of hepatic glucose homeostasis, which means that BTG2 might serve as a potential therapeutic target for combating metabolic dysfunction.
The liver plays a crucial role in glucose homeostasis by maintaining a balance between uptake and storage of glucose via glycogenesis and in the release of glucose by both glycogenolysis and gluconeogenesis (1,2). Gluconeogenesis is commonly triggered by key hormones including insulin, glucagon, and glucocorticoid, which contribute to the expression of key metabolic enzymes, such as phosphoenolpyruvate carboxykinase (Pck1), fructose biphosphatase (Fbp) 1, and glucose-6-phosphatase (G6pc) (3). A variety of transcriptional factors and cofactors regulated by the cAMP signaling pathway such as cAMP-responsive element–binding protein (CREB), FoxA 2 (also known as hepatocyte nuclear factor [HNF]-3β), HNF-4α, glucocorticoid receptor (GR), forkhead box protein O1, peroxisome proliferator-activated receptor γ coativator-1α (PGC-1α), and transducer of regulated CREB activity 2 control the expression of hepatic gluconeogenic genes (4–9). Especially CREB acts as a critical transcriptional checkpoint for the upregulation of gluconeogenesis by binding cAMP-response element. Multiple genes involved in the gluconeogenic genes including Pck1, G6pc, and Pgc-1α were shown to be increased by recruiting CREB to the promoters of these genes (4,10,11).
The nuclear receptor Nur77/NR4A1 is a member of a family of the NR4A receptor, which has highly conserved DNA- and ligand-binding domains. Nur77 binds to its cognate DNA sequence, called NGFI-B (Nur77)-binding response elements (NBREs) or Nur response elements (NurREs), and induces transcription of target genes involved in the gluconeogenesis and steroidogenesis (12,13). In addition, hepatic expression of NR4A receptors induced by cAMP in response to glucagon and fasting is increased in diabetic mice with elevated gluconeogenesis, and ectopic overexpression of the Nur77 gene induces upregulation of hepatic gluconeogenic genes, glucose production, and increased blood glucose level (12). Consistent with these results, Nur77-null mice on a high-fat diet also show decrease of hepatic glucose production and insulin resistance in the liver (14).
B-cell translocation gene 2 (BTG2) is a member of the BTG/TOB gene family of antiproliferative genes, which are highly conserved among various species (15). BTG2 is reported to be the initial gene induced by growth factors and tumor promoters in several cell types (16,17). Recently, we demonstrated that BTG2 acts as a crucial coactivator of CREB to regulate hepatic gluconeogenesis, and the increase of BTG2 by glucagon–CREB signaling resulted in the elevation of glucose production from hepatocytes (18). However, the interrelationship between BTG2 and Nur77 in terms of the regulation of hepatic gluconeogenesis has not been addressed yet.
In the current study, we identified that transcriptional coregulator BTG2 acts as a novel coactivator of hepatic gluconeogenesis and elucidated a molecular mechanism that links to Nur77 in the regulation of glucose homeostasis in the livers of fasted and diabetic mice.
Research Design and Methods
Forskolin and streptozotocin were purchased from Sigma-Aldrich (St. Louis, MO). Cell-culture media were purchased from Gibco-Brl (Grand Island, NY).
Male C57BL6 mice (8 weeks of age; Samtako, Osan, Republic of Korea) were used in the experiments. For studies in the fasting and feeding conditions, mice were fed and fasted for 12 h. Diabetes was induced in wild-type (WT) and db/db mice by intraperitoneal injections of streptozotocin (80 mg/kg body weight). For overexpression of BTG2, WT mice were infected with adenoviral vector expressing Btg2 (1× 109 plaque-forming units) by tail vein injection. To disruption BTG2, db/db mice were intravenously injected with lentivirus short hairpin Btg2 (shBtg2; a single dose of 1× 109 transducing units/mL). Total RNA was prepared from the livers of streptozotocin-treated diabetic mice and adenovirus (Ad)- or lentivirus-infected mice for quantitative reverse transcription-PCR (qPCR) analysis. Before the isolation of the liver, the blood glucose levels of all mice were measured with Glucostix Accu-Check (Roche Diagnostics, Mannheim, Germany). All animal experiments were performed in accordance with the rules and regulations of the Institutional Animal Use and Care Committee at Keimyung University School of Medicine.
Plasma glucose was measured using a glucometer (Roche Diagnostics) in blood collected from the tail of mice as previously described (19). For the glucose tolerance test, a single dose of 2.5 μg/kg glucose was injected intraperitoneally after a 14-h fast. For the insulin tolerance test, mice fasted for 4 h were administered intraperitoneally with insulin (0.5 units/kg), and glucose concentrations were measured at 0, 30, 60, 90, and 120 min. Blood glucose was analyzed under the same conditions.
Construction of Plasmids and DNA
The reporter plasmids NBRE-Luc, NurRE-Luc, Pck1-Luc, and G6pc-Luc were described previously (13,18). Expression vectors of HA-Nur77, dominant-negative (DN)-Nur77, Hnf4-α, Pgc-1α, GR, and Flag-Btg2 were described previously (13,18,20–23). The small interfering RNAs for Btg2 were prepared as described previously (18). The point mutant form of human G6PC-Luc or mouse G6pc-Luc was constructed using a site-directed mutagenesis kit (Stratagene, La Jolla, CA) and the following primers: human G6PC (hG6PC)-Luc forward, 5′-CTGTAACTTTAATTTCATTAAATG-3′ and reverse, 5′-CATTTAATGAAATTAAAGTTACAG-3′ and mouse G6pc (mG6pc)-Luc forward, 5′-TAACTT-AAAATTTCACTTCCG-3′ and reverse, 5′-CGGAAGTGAAATTTTAAGTTA-3′. All plasmids were confirmed by sequencing analysis.
Cell Culture and Transient Transfection Assays
AML-12 immortalized mouse hepatocytes were cultured in DMEM/F-12 medium (Gibco-Brl) supplemented with 10% FBS, insulin-transferrin-selenium (Gibco-Brl), dexamethasone (40 ng/mL; Sigma-Aldrich), and antibiotics in a humidified atmosphere containing 5% CO2 at 37°C (24). Transient transfections were carried out as previously described (18,24).
Isolation and Culture of Primary Mouse Hepatocytes
Mouse primary hepatocytes were isolated from the livers of 8-week-old male mice (Samtako). The protocol for isolation of hepatocytes was described previously (8).
Preparation of Recombinant Ad
Ads encoding full-length Btg2 and green fluorescence protein (GFP) have been described previously (18,24). Recombinant lentiviral delivery of Btg2-targeted short hairpin RNA (shRNA) was purchased from Dharmacon (Lafayette, CO). Ad-expressing DN-Nur77 was gifted from Dr. In-Kyu Lee at Kyungpook National University School of Medicine.
Measurement of mRNA
Glucose Production Assay
Glucose production from mouse primary hepatocytes was measured according to the manufacturer’s protocol using a colorimetric glucose oxidase assay (Sigma-Aldrich). Briefly, after designated times, the cells were washed three times with PBS and then incubated for 3 h at 37°C in an atmosphere containing 5% CO2 in glucose production buffer (glucose-free DMEM [pH 7.4] containing 20 mmol/L sodium lactate, 1 mmol/L sodium pyruvate, and 15 mmol/L HEPES without phenol red). The glucose assays were carried out in triplicate, and the intra-assay coefficient of variation was 5% as described previously (27).
Immunoblotting and Coimmunoprecipitation Assay
Proteins were isolated from mouse primary hepatocytes and liver tissues and analyzed according to the methods described previously (19). The membranes were probed with BTG2 and Nur77 (Santa Cruz Biotechnology, Santa Cruz, CA) and then developed using an ECL Western blot detection kit (Amersham Biosciences, Piscataway, NJ). For immunoprecipitation, total protein extracts from AML-12 cells and mouse liver were probed for BTG2, Nur77, hemagglutinin (HA) (Santa Cruz Biotechnology), and FLAG antibody (Sigma-Aldrich) and then blotted with these antibodies (21). Signals were developed using an ECL Western blot detection kit (Amersham Biosciences).
Chromatin Immunoprecipitation Assay
The chromatin immunoprecipitation (ChIP) assay was performed as described previously (18,19). Briefly, 36 h after infection with Ad-Btg2 in mouse primary hepatocytes, cells were treated with forskolin (10 μmol/L). The cells were subsequently harvested, and the ChIP assay was performed with anti-Nur77. The final DNA extractions were quantified by PCR with two pairs of primers for the proximal (−1,070/−860 and −529/−310) and distal (−2,000/−1,790 and −1,679/−1,460) regions of the hG6PC and mG6pc promoters. The specific primers used for PCR are as follows: hG6PC promoter, proximal, forward, 5′-AGTGAGCCGTGATTATGCCA-3′ and reverse, 5′-TGGGAGATCAACCCCCTTCC-3′; distal, forward, 5′-TGGAATGCTGTC-AACTTTT-3′ and reverse, 5′-CTCAACCTAGCAAAGCATT-G-3′; mG6pc promoter, proximal, forward, 5′-CAGAGGTGGACCCCAGATCC-3′ and reverse, 5′-AGAGTCTGTTT-CAACCATGA-3′; and distal, forward, 5′-GGTGGCTCACAACCATCTGT-3′ and reverse, 5′-TG-AGCTGGTGTCATGTGACT-3′.
Data calculation and statistical analyses were performed using GraphPad Prism 3-5.0 software (GraphPad). The statistical significance of differences between groups were determined using the Student t test, and multiple comparisons were analyzed using one-way ANOVA under treatment and experiment as factors. All data are presented as means ± SEM. All P values <0.05 were considered significant.
BTG2 Gene Expression Is Induced by Fasting and Glucagon Signaling in the Liver
BTG2 acts as a coactivator of CREB to positively regulate hepatic gluconeogenesis in vitro (18). To investigate whether BTG2 is regulated by fasting and glucagon stimulation in vivo, we performed qPCR analysis from mouse liver. As shown in Fig. 1A, Btg2 gene expression was induced by fasting along with Creb, Nur77, Pck1, and G6pc. Also, protein levels of BTG2 and Nur77 were induced in the fasted state (Fig. 1B). Btg2 mRNA was induced by forskolin, a cAMP agonist, with expression reaching a maximal stimulation at 1 h. Similar results were observed for Creb, Nur77, Pck1, and G6pc mRNA levels (Fig. 1C), consistent with previous studies (12,18). Interestingly, the increase of BTG2 and Nur77 proteins by 1–6-h treatment of forskolin reached the peak level, and it subsequently declined after 12 h (Fig. 1D). These results suggest that BTG2 plays a pivotal role in the upregulation of gluconeogenic genes in vivo.
Overexpression of BTG2 Gene Induces Hepatic Gluconeogenesis
Our findings predict that BTG2 would be involved in the regulation of gluconeogenic genes for hepatic glucose production in liver. To test this hypothesis, we examined the effect of ectopic expression of BTG2 on the transcription of hepatic gluconeogenic genes using an Ad vector to drive expression of Btg2 (Ad-Btg2). Indeed, the expressions of Nur77, Pck1, and G6pc were significantly increased by administration of Ad-Btg2 in the liver (Fig. 2A). However, expression of other transcription factors related to gluconeogenesis like GR and Hnf-4α was not affected (Fig. 2A). Moreover, consistent with the expression of gluconeogenic genes, blood glucose level was significantly increased by Ad-Btg2 (Fig. 2B). To evaluate whether ectopic expression of BTG2 also affects glucose homeostasis and insulin sensitivity, glucose and insulin tolerance tests were performed in the Ad-Btg2–infected mice. As predicted, there was significant impairment in glucose tolerance in mice infected with Ad-Btg2 (Fig. 2C). The insulin tolerance test also showed a marked decrease in insulin sensitivity (Fig. 2D). Taken together, these results indicate that BTG2 plays a key role in hepatic glucose production and directly influences whole-body glucose metabolism.
Glucagon Signal–Dependent Hepatic Gluconeogenic Gene Expression Is Mediated by BTG2
To determine whether Btg2 was directly regulated by the fasting hormone glucagon, as shown in Fig. 3A, Btg2 gene expression was overexpressed well by Ad-Btg2, and glucagon signaling–induced Btg2 gene expression was successfully decreased by shBtg2. In addition, forskolin-stimulated gluconeogenic gene expression was significantly reduced by knockdown of endogenous Btg2 in primary hepatocytes (Fig. 3B–D). We also tested whether BTG2 regulates glucose production in primary hepatocytes. Interestingly, like forskolin treatment, Ad-Btg2 significantly increased hepatic glucose production compared with untreated control and Ad-GFP–infected cells, whereas this effect was attenuated by silencing of Btg2 (Fig. 3E).
Because Nur77 is one transcription factor involved in gluconeogenesis, we hypothesized that BTG2 might function as a coactivator along with Nur77. To test this prediction, luciferase assays were performed with the Nur77 reporter plasmids NBRE-Luc and NurRE-Luc. Btg2 significantly increased reporter activity of both NBRE- and NurRE-mediated transcription in a dose-dependent manner (Fig. 3F). Moreover, forskolin did not activate these reporters when Btg2 was suppressed by small interfering Btg2 (siBtg2) (Fig. 3F). Ectopic expression of Btg2 significantly increased promoter activities of the Pck1 and G6pc (Fig. 3G). Interestingly, forskolin-stimulated promoter activities of G6pc and Pck1 were significantly decreased when siBtg2 was transfected, and overexpression of Btg2 rescued the promoter activities (Fig. 3G). These results indicate that BTG2 regulates glucagon-mediated gluconeogenic gene expression. This phenomenon raises an interesting possibility that BTG2 may serve as a major player in regulating Nur77-mediated transcription of gluconeogenic genes in liver.
BTG2 Physically Interacts With Nur77 and Mediates Nur77 Occupancy on the G6pc Promoter
Based on our results of the role of BTG2 regulating NBRE- and NurRE-mediated transcription, we observed a physical interaction between BTG2 and Nur77 using a coimmunoprecipitation assay. Flag-Btg2 was coprecipitated with HA-Nur77 (Fig. 4A). In the coimmunoprecipitation assay, endogenous Nur77 was coprecipitated with endogenous BTG2 (Fig. 4B). Collectively, these results demonstrate that BTG2 physically interacts with Nur77 both in vitro and in vivo.
Next, we investigated whether Btg2 modulates the promoter activity of mouse G6pc gene as well as human G6PC gene through Nur77. Like forskolin, overexpression of Nur77 significantly increased the promoter activity of G6pc in a dose-dependent manner and luciferase activity by Btg2 was synergistically increased by cotransfection of the Nur77. Introduction of mutation at the NBRE resulted in a decrease of promoter activity (Fig. 4C and D). To further confirm whether interaction of BTG2 with Nur77 affects the direct DNA binding of Nur77 protein on the promoters of these target genes, a ChIP assay was performed. The endogenous binding activity of Nur77 to its proximal region by forskolin treatment was significantly increased by Ad-Btg2 compared with that of control (Fig. 4E and F). However, the nonspecific distal region of the hG6PC and mG6pc promoter was unable to recruit Nur77 under all conditions (Fig. 4E and F). These results indicate that BTG2 could contribute to the Nur77-dependent regulation of G6pc gene expression via a physical interaction.
BTG2-Mediated Induction of Hepatic Gluconeogenesis Is Mediated by Nur77
To determine whether the BTG2-induced regulation of hepatic gluconeogenesis is mediated by Nur77, we evaluated the effect of Nur77 on BTG2-mediated hepatic glucose production and gluconeogenic gene expression in primary hepatocytes. Increased glucose production and gene expression of Creb as well as G6pc by Ad-Btg2 were attenuated by DN-Nur77, but not Pck1 gene expression (Fig. 5A and B). Next, we confirmed whether Nur77 also regulates transcriptional activity of gluconeogenic genes. G6pc and Pck1 promoter activities were synergistically increased by Btg2, Nur77, and Creb. It is notable that G6pc promoter activity was decreased by DN-Nur77 (Fig. 5C). Although Pck1 gene promoter was significantly upregulated by Btg2, its promoter activities were not affected by DN-Nur77 (Fig. 5C). Interestingly, while overexpression of Hnf-4α, GR, and Pgc-1α genes significantly increases the promoter activities of Pck1 and G6pc, which is consistent with previous reports (4,5,7,9). Remarkably, this stimulatory effect of Btg2 was not synergistically increased by cotransfection of the Hnf-4α, GR, and Pgc-1α (Fig. 5D). Thus, it is speculated that BTG2 acts as a coactivator upregulating the gluconeogenesis network with Nur77 and CREB (18), whereas it does not act as a crucial coactivator of Hnf-4α, GR, and Pgc-1α. Taken together, these results indicate that BTG2 functions together with Nur77 in the upregulation of gluconeogenesis in the primary hepatocytes.
Expression of Gluconeogenic Genes in Diabetic Models Are Increased by BTG2
Excessive hepatic gluconeogenesis contributes to the fasting hyperglycemia in diabetes (28). To access the functional contribution of BTG2 to the increase of hepatic glucose production in diabetes, streptozotocin-treated and db/db mice were used. qPCR analysis showed that streptozotocin injection strongly induced expression of the Btg2 as well as the Creb, Nur77, Pck1, and G6pc genes in liver (Fig. 6A). Moreover, mRNA levels of Btg2, Creb, Nur77, Pck1, and G6pc are increased in the liver of db/db mice when compared with those of WT mice (Fig. 6B). These data indicate that BTG2 contributes to elevated gluconeogenesis in diabetes. When db/db mice were infected with shBtg2, Btg2 was significantly reduced in the liver. mRNA levels of Btg2, Creb, Nur77, Pck1, and G6pc were significantly elevated in liver of db/db mice, and these mRNAs were attenuated by shBtg2 (Fig. 6B). These findings suggest that BTG2 plays a pivotal role in transcriptional regulation of hepatic gluconeogenic genes in diabetic mouse models.
Hepatic Gluconeogenesis in db/db Mice Is Altered by BTG2
Finally, we investigated the crucial role of BTG2 on glucose homeostasis in fasted db/db mice. Administration of shBtg2 to db/db mice resulted in the amelioration of glucose tolerance (Fig. 7A). An insulin tolerance test revealed that db/db mice were significantly improved by shBtg2 (Fig. 7B). In addition, the increase of blood glucose levels in db/db mice was markedly reduced by silencing of Btg2 (Fig. 7C). These results indicate that systemic glucose homeostasis in both physiology and diabetes is mediated by BTG2.
Our recent study revealed another physiological function of BTG2, namely hepatic gluconeogenesis in primary hepatocytes (18). BTG2 gene expression was increased during fasting, resulting in elevated transactivity of hepatic gluconeogenic genes. Moreover, the function of BTG2 has not been addressed in metabolic diseases like diabetes. In this study, Nur77 was identified as a new target of BTG2 involved in hepatic gluconeogenesis. The current study demonstrates that BTG2 increases hepatic glucose production through the upregulation of G6pc gene expression by physical interaction with Nur77 during both fasted and diabetic conditions. Overexpression of Btg2 induced impaired glucose and insulin tolerances during glucose tolerance testing and insulin tolerance testing in mice. Administration of shBtg2 significantly improved glucose disposal in db/db mice.
Time-sequential expression of the Btg2 gene increased soon after forskolin treatment, with mRNA expression of Creb and Nur77 genes induced concurrently. Finally, both Pck1 and G6pc genes were expressed in primary hepatocytes (Fig. 1C). Based on these data, we speculate that BTG2 may play a key role in modulation of gluconeogenesis at the upstream level of glucagon signaling.
A recent study suggested that BTG2 increases Pck1 gene transcription through the induction of CREB in primary hepatocytes (18). CREB is a pivotal regulator of hepatic glucose production during fasting in the liver through binding to the cis-element of CREB on the target gene promoter. Gluconeogenesis is activated by regulation of Pck1 through the interaction of BTG2 and CREB as well as glucagon signaling with cAMP and dexamethasone (4,12,18). However, even though G6pc gene expression was presently increased by overexpression of Btg2 in primary hepatocytes, the regulatory mechanism by BTG2 has not been clear to explain an increase of G6pc gene expression both in vitro and in vivo. The current study expands upon these findings to show that transcription of G6pc gene is activated by BTG2 through regulation of Nur77 gene expression, which is a novel copartner of BTG2 in the liver during fasted and diabetic conditions.
Nur77 is induced by multiple extracellular signals in a tissue-specific manner. Nur77 is also involved in 3T3-L1 adipogenesis response to cAMP signaling without peroxisome proliferator–activated receptor γ agonist (29). In the liver, Nur77 modulates hepatic gene expression (10). Nur77 is also a mediator of cAMP action in the regulation of hepatic glucose metabolism (12). The expression of Nur77 in hepatocytes is increased by cAMP in vitro and by glucagon during fasting in vivo (12). In this study, Nur77 mRNA level was induced by Btg2 overexpression and abolished by its knockdown using shRNA (Fig. 3B). Moreover, Nur77 gene expression was accelerated by Creb in fasted liver, consistent with a previous report (30). These results suggest that Nur77 gene expression is stimulated by the BTG2–CREB axis during fasting and diabetes. Therefore, Nur77 is not only increased in mRNA expression, but also induces protein interaction with BTG2 to elevate hepatic glucose production.
Pei et al. (12) demonstrated that various genes involved in the gluconeogenesis including G6pc, Fbp1, and Fbp2 were upregulated by Nur77, but not Pck1. Moreover, promoter activity of these genes was increased by direct binding of Nur77 (12). In the current study, the Nur77 binding region was confirmed on the G6pc promoter, and deletion of the Nur77 response element site abolished BTG2-induced G6pc transcription. Interestingly, although a Nur77 response cis-element has been not identified on the Pck1 gene promoter, endogenous Pck1 gene expression as well as promoter activity was induced by overexpression of Btg2 (Figs. 2Aand 3G). Thus, it is speculated that Nur77 directly upregulates the expression of multiple genes involved in glucose metabolism, whereas Pck1 may be regulated by Nur77 in an indirect manner both in vitro and in vivo. In addition, a previous study addressed not only BTG2-induced physical interaction with CREB in fasting, but also discovered the binding region of the CREB protein on the Pck1 promoter (18). The data indicate that Pck1 gene expression might be induced by BTG2 through CREB, not Nur77, in liver during fasting and in diabetes. Administration of shBtg2 into db/db mice ameliorated blood glucose levels by inhibiting Pck1 and G6pc gene expression via decrease of Nur77 and CREB. In addition, gene silencing of Btg2 by lenti-shRNA significantly improved both glucose and insulin tolerance in db/db mice. These findings in vivo are consistent with our previous study that reported the function of BTG2 in vitro (18). Although the physiological role of BTG2 has been verified in diabetic mouse model, the role of BTG2 still needs to validate among human samples to develop as a new therapeutic agent in diabetes.
In conclusion, the current study demonstrates that Nur77 is a novel target of BTG2 and BTG2 regulates hepatic glucose production through Nur77-mediated G6pc gene expression. We characterized the role of BTG2 as a novel transcriptional coactivator of hepatic gluconeogenesis in vivo and elucidated the molecular mechanism of regulation of glucose homeostasis mediated by a link of BTG2 and Nur77 as a key regulator of gluconeogenesis. Moreover, disruption of BTG2 ameliorates glucose disposal and insulin resistance in diabetic animal model. Thus, as we depict in the novel schematic model shown in Fig. 7D, regulation of BTG2 may offer an effective means to control hepatic glucose homeostasis in diabetes and may guide the development of novel therapeutic target for metabolic diseases like diabetes.
Acknowledgments. The authors thank Dr. Yong-Ho Ahn (Yonsei University College of Medicine) and Dr. Timothy F. Osborne (Sanford-Burnham Medical Research Institute) for helpful discussions and thank Dr. Hye-Young Seo (Keimyung University School of Medicine) for excellent technical suggestions.
Funding. This study was supported by a grant from the Korea Health Technology R&D Project, Ministry of Health and Welfare, Republic of Korea (A120864) and by a Keimyung University Medical School Research Promoting Grant launched in 2013 (to S.-S.I.).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. Y.D.K. contributed to the conception, design, and performance of experiments; analysis and interpretation of data; and writing the draft. S.-G.K. and S.-L.H. contributed to the design and analysis and interpretation of experimental results. H.-S.C. and J.-H.B. contributed to the analysis and interpretation of data and critical review of the manuscript. D.-K.S. and S.-S.I. contributed to the conception and design of the experiments, wrote the manuscript, and performed the critical review of the manuscript. D.-K.S. and S.-S.I. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.