Synergistic Activations of REG Iα and REG I β Promoters by IL-6 and Glucocorticoids through JAK/STAT Pathway in Human Pancreatic β Cells

Reg (Regenerating gene) gene was originally isolated from rat regenerating islets and its encoding protein was revealed as an autocrine/paracrine growth factor for β cells. Rat Reg gene is activated in inflammatory conditions for β cell regeneration. In human, although five functional REG family genes (REG Iα, REG Iβ, REG III, HIP/PAP, and REG IV) were isolated, their expressions in β cells under inflammatory conditions remained unclear. In this study, we found that combined addition of IL-6 and dexamethasone (Dx) induced REG Iα and REG Iβ expression in human 1.1B4 β cells. Promoter assay revealed that a signal transducer and activator of transcription- (STAT-) binding site in each promoter of REG Iα (TGCCGGGAA) and REG Iβ (TGCCAGGAA) was essential for the IL-6+Dx-induced promoter activation. A Janus kinase 2 (JAK2) inhibitor significantly inhibited the IL-6+Dx-induced REG Iα and REG Iβ transcription. Electrophoretic mobility shift assay and chromatin immunoprecipitation revealed that IL-6+Dx stimulation increased STAT3 binding to the REG Iα promoter. Furthermore, small interfering RNA-mediated targeting of STAT3 blocked the IL-6+Dx-induced expression of REG Iα and REG Iβ. These results indicate that the expression of REG Iα and REG Iβ should be upregulated in human β cells under inflammatory conditions through the JAK/STAT pathway.


Introduction
Decreased functional cell mass is one of hallmarks in both type 1 and type 2 diabetes. Regeneration of pancreatic cells has been shown to occur at a basal rate in normal adult tissues and to increase under certain conditions such as pregnancy and obesity [1][2][3]. However, the mechanism of human pancreatic cell regeneration still remains to be elucidated.
Reg (Regenerating gene) gene product, Reg protein, could be responsible for the regenerative process [4][5][6]. The Reg gene was originally isolated from regenerating pancreatic islets from depancreatized rats, and it encodes a 16 kDa autocrine/paracrine growth factor for cells [7,8]. The Reg and Reg-related genes were isolated and revealed to constitute a multigene family, the Reg family, which consists of four subtypes (types I, II, III, and IV) based on the primary structures of the encoded proteins of the genes [4,5,9,10]. The type I Reg gene, Reg I, was the first isolated Reg gene from the regenerating islets [7]. In rat and mouse, Reg I protein is not expressed in pancreatic cells in physiological conditions but is induced during islet regeneration [4,7,11,12]. Therefore, the Reg I gene induction, which occurs in response to inflammatory mediators such as interleukin-(IL-) 6 and glucocorticoid analogue dexamethasone (Dx) in rat RINm5F cells [13], is considered to be one of the crucial events in cell regeneration. In human, five functional REG family genes (REG I , REG I , REG III, HIP/PAP, and REG IV) were isolated [4,7,[14][15][16][17][18][19]. However, which human REG gene is expressed in cells during regeneration is still obscure mainly because of their restrict supply of human islets. Induction of proliferation/regeneration in human cells must be beneficial for both type 1 and type 2 diabetes treatment/prevention [20]. The human REG genes would . For construction of the promoter region mutants used in this study, the indicated sequences in the "−146" REG I /luciferase constructs and "−155" REG I /luciferase constructs were substituted by primers using PCR. The cells were grown in 24-well plates and were transfected with reporter plasmids by lipofection. Briefly, 0.8 g of each reporter plasmid and 0.08 g of pCMV-SPORT--galactosidase (Life Technologies, Carlsbad, CA), as an internal control, were mixed with Lipofectamine 2000 (Life Technologies) per well  in a 24-well plate. After 24 h incubation, the cells were treated  with indicated amount of stimulants, 100 nM Dx, 20 ng/mL  IL-6, 300 U/mL IL-1 , 370 U/mL TNF-, 100 U/mL IFN-,  60 U/mL IL-1 + 185 U/mL TNF-+ 14 U/mL IFN-,  10 ng/mL IL-22, 1500 U/mL IFN-, 100 nM IL-8, and 1 U/mL PDGF, amino acids at a concentration five times greater than that in basal RPMI, 100 M palmitate, 10 M H 2 O 2 , 250 M SNAP, 2 mM STZ, and 2 mM 3AB, and combinations thereof. After 24 h treatment, the cells were harvested and processed for luciferase assay as described [13,26]. In cases of STZ and STZ + Dx, the media containing STZ or STZ + Dx were changed with fresh media after 2 h treatment, and cells were further incubated for 22 h before harvest.
DNA-protein binding reactions were performed by incubation of the nuclear extracts in a solution containing 10 mM Journal of Diabetes Research 3 HEPES (pH 7.8), 50 mM KCl, 5 mM MgCl 2 , 1 mM EDTA, 10% glycerol, 5 mM DTT, 1 mg/mL BSA, 0.7 mM PMSF, 50 ng/ L of poly(dI-dC) (Roche) for 10 min at room temperature, followed by an additional 30 min incubation with 32 P-endlabeled probe at room temperature. For supershift assays, 1 g of anti-signal transducer and activator of transcription (STAT)3 antibody (Santa Cruz Biotechnology, SC-482X, Santa Cruz, CA) [30] was added to the samples and incubated for 15 min before incubation with the labeled probe. DNAprotein complexes were separated on 4% nondenaturing acrylamide gels and detected by autoradiography.

Chromatin Immunoprecipitation (ChIP) Assay.
ChIP assay was performed using a ChIP-IT Express kit (Active Motif, Carlsbad, CA) following the manufacturer's instructions as described previously [31]. Briefly, 1.1B4 cells were treated with or without IL-6 + Dx and cross-linked with 1% formaldehyde for 10 min at room temperature. After washing and treatment with glycine Stop-Fix solution, the cells were lysed and nuclei were subjected to ultrasonic disruption for preparing DNA fragments ranging in size from 200 to 1500 bp. Immunoprecipitation was carried out overnight at 4 ∘ C using an incubation mixture containing sheared chromatin, protein G magnetic beads, and 3 g of anti-STAT3 antibody (Santa Cruz Biotechnology, SC-482X). The immune complexes were precipitated, eluted, reverse cross-linked, and treated with proteinase K. The resulting DNA samples were subjected to PCR amplification (1 cycle of 94 ∘ C for 3 min, 35 cycles of 94 ∘ C for 30 s, 62 ∘ C for 30 s, 72 ∘ C for 1 min) of the REG I promoter using specific forward (5 -ACCTTGGACTTAGACAGCTTG-3 ) and reverse (5 -ACCACGTCATTTAAGCAAAAGG-3 ) primers designed to amplify the promoter region (−212 to −46 nucleotides in relation to the transcription start site). The PCR products were resolved by electrophoresis in a 2.5% agarose gel containing ethidium bromide for visualization.

Data Analysis.
Results are expressed as mean ± SE. Statistical significance was determined by Student's -test using GraphPad Prism software (GraphPad Software, La Jolla, CA).

Activation of Human REG Family Gene Promoters in
Cells. We first tested various stimuli on human REG family gene transcription in human 1.1B4 cells. As Reg gene expression was observed in the phase of transient cell proliferation: for example, in pancreatic islets of diabetic BB/Wor//Tky rats during the remission phase of diabetes [32], in islets of nonobese diabetic mice during active diabetogenesis [33] and pancreatic ductal cells, which are thought to be progenitor cells of cells, during differentiation and proliferation in a mouse model of autoimmune diabetes [34,35], and inflammation and cell death in and/or around islets was involved in these cases, we used inflammatory cytokines such as IL-1 , TNF-, and IFN- [36], as well as pancreatic cell toxic agents, such as palmitate [37,38], H 2 O 2 [39,40], SNAP [41,42], and STZ [43,44]. Other factors, which were reported to upregulate Reg gene expression in cells, such as IL-6 [13,31], glucocorticoid [13,31], IL-22 [6,45], IFN- [35], PDGF [46], and amino acids [46], were tested. We also used IL-8, which was reported to stimulate REG I promoter activity in gastric endocrine carcinoma cells [47]. As shown in Figure 1(a), IL-6 and IL-22 modestly increased the REG I promoter activity (columns 3 and 13). When IL-6 and the glucocorticoid analogue Dx were added together, the promoter activity was remarkably increased (columns 3 and 4). In contrast, the combination of IL-22 and Dx did not further increase the promoter activity (columns 13 and 14). The combination of IL-22 and IL-6 + Dx did not result in the synergistic response (columns 14 and 15). Treatments with other stimulants were ineffective or resulted in only small changes in promoter activity of REG I gene compared to IL-6 + Dx. These results showed that IL-6 + Dx was the most effective inducer for human REG I gene transcription in human cells. IL-6 + Dx also enhanced the promoter activities of REG I and HIP/PAP significantly; however, the fold increases were much smaller compared to that of REG I (Figures 1(a), 1(b), and 1(d)). Although the transcription of REG III and REG IV was induced to some extent by IL-6 + Dx, the increments were quite small (Figures 1(c) and 1(e)). The other treatments did not induce REG I , REG I , REG III, HIP/PAP, nor REG IV transcription as much as IL-6 + Dx did.
To determine whether the IL-6 + Dx-induced expression of REG family gene was restricted only in 1.1B4 cells, we investigated the effects of IL-6 + Dx on the promoter activities of human REG genes in rat RINm5F cells. Similarly in 1.1B4 cells, the transcription of REG I was synergistically increased by the combined addition of IL-6 and Dx (Figure 2(a)). The REG I transcription was also significantly induced by IL-6 + Dx in RINm5F cells (Figure 2(b)); however, the increase was not so large as that in 1.1B4 cells. The transcriptional activity of HIP/PAP was also increased by IL-6 and IL-6 + Dx about 1.5-fold (Figure 2(d)). On the other hand, the transcriptional activities of REG III and REG IV were never induced by Dx, IL-6, nor IL-6 + Dx (Figures 2(c) and 2(e)). These results indicated that activation of human type I REG gene transcription by IL-6 + Dx is general feature in pancreatic cells.  human REG family genes in 1.1B4 cells treated with IL-6 and Dx by real-time RT-PCR. The mRNA levels of REG I and REG I were significantly increased by the addition of IL-6 + Dx (Figures 3(a) and 3(b)). On the other hand, the mRNA levels of REG III, HIP/PAP, and REG IV were not significantly increased by the addition of neither IL-6, Dx, nor IL-6 + Dx (Figures 3(c), 3(d) and 3(e)). These results strongly suggest that type I REG family genes (REG I and REG I ) were induced in human cells under inflammatory conditions by increased levels of circulating IL-6 and glucocorticoids.  (Figure 4(a)). In contrast, further deletion to position −130 caused the loss of inducibility. These data suggest that the region between −146 and −131 is essential for the IL-6 + Dx-induced REG I transcription. The sequence analysis revealed the presence of elements homologous to the STAT-binding site between position −142 and −134 (TGCCGGGAA) (Figure 4(b)). Site-directed mutagenesis was conducted within the luciferase construct "−146, " and the influence of the mutation ("−146 M" in Figure 4(b)) on the amplitude of luciferase induction by IL-6 + Dx was monitored. The replacement of the STAT-binding sequence to the sequence CAGTAGGAA, which disrupted the binding of STAT ("−146 M"), significantly blocked the IL-6 + Dxinduced transcriptional activation of REG I in 1.1B4 cells (Figure 4(c)).

IL-6 + Dx Response Element in REG Gene
In rat RINm5F cells, the cis-element of rat Reg I gene for induction in response to IL-6 + Dx was the −81∼−70  region, which corresponds to the −76∼−65 region of human REG I gene [13]. In RINm5F cells, poly(ADP-ribose) polymerase (PARP) was found to bind to the cis-element of Reg I gene and was involved in the active transcriptional DNA/protein complex formed by the stimulation of IL-6 + Dx. The DNA/protein complex formation was inhibited by the autopoly(ADP-ribosyl)ation of PARP in the complex. Thus, PARP inhibitors, such as 3AB, enhanced the DNA/protein complex formation for Reg I gene transcription [13]. To clarify whether PARP was involved in the IL-6 + Dx-induced human REG I transcription, we treated 1.1B4 cells with IL-6 + Dx + 3AB and measured the promoter activities of the constructs "−146" and "−146 M. " As shown in Figure 4(c), 3AB further enhanced the IL-6 + Dx-induced promoter activity of construct "−146. " On the other hand, the promoter activity of mutant construct "−146 M" was not induced as much as that of "−146" by IL-6 + Dx + 3AB, suggesting that STAT rather than PARP played a major role in the induction of human REG I transcription by IL-6 + Dx. STAT proteins are activated by JAKs, and IL-6 induces activation of JAKs [48,49]. Pretreatment of 1.1B4 cells with a JAK2 inhibitor, AG490 (50 M), for 4 h significantly inhibited the inducibility of REG I promoter by IL-6 + Dx (Figure 4(d)). These results suggest that the JAK/STAT signaling pathway is involved in the IL-6 + Dx-induced REG I gene transcription in cells.  Figure 5: IL-6 + Dx-increased STAT3 binding to the REG I promoter. (a) Detection of DNA-protein complexes using EMSA. Nuclear extracts (2.5 g protein) prepared from 1.1B4 cells, treated without or with Dx (100 nM), or IL-6 (20 ng/mL), or IL-6 + Dx, were incubated with a 32 P-labeled probe 1. Nuclear extracts from untreated cells were applied onto lanes 1, 5, and 9; those from Dx-treated cells were applied onto lanes 2, 6, and 10; those from IL-6-treated cells were applied onto lanes 3, 7, and 11; and those from IL-6 + Dx-treated cells were applied onto lanes 4, 8, and 12. Lanes 1-4 contain no competitor; lanes 5-8 contain 100 × unlabeled probe 1; lanes 9-12 contain 100 × unlabeled probe M1. An arrow marks IL-6 + Dx-inducible complex migration. (b) Supershift assay of STAT3 containing complex. EMSAs were performed in the presence or absence of STAT3 antibodies as indicated. An arrow marks IL-6 + Dx-inducible complex migration. An arrowhead indicates supershift complex by anti-STAT3 antibody. (c) ChIP assay showing IL-6 + Dx increases in STAT3 binding to the REG I promoter. 1.1B4 cells were treated without or with IL-6 (20 ng/mL) + Dx (100 nM). ChIPs were performed with anti-STAT3 antibody, followed by PCR with primers specific for the REG I promoter. A representative result from three experiments is shown.
As the nucleotide sequence of promoter region of REG I and REG I is very similar, we searched possible STATbinding sequence in the REG I promoter and found that there is a STAT-binding site between position −151 and −143, which corresponds to the STAT-binding site of REG I (the −142∼−134 region) as shown in Figure 4(b). The deletion and mutational analyses revealed that the STAT-binding site was essential to the IL-6 + Dx-induced REG I transcription (Figure 4(e)). Moreover, AG490 almost completely blocked the induction of REG I promoter by IL-6 + Dx (Figure 4(f)). Thus, it is suggested that IL-6 + Dx-induced REG I transcription also depends on the JAK/STAT signaling pathway.

STAT3 Binding to the REG I Promoter.
To further investigate whether STAT protein interacts with the REG I promoter in vitro and in vivo, we performed EMSA and ChIP assays, respectively. Treatment of Dx did not change the nuclear protein binding to the probe 1 containing the STAT-binding sequence (TGCCGGGAA) (Figure 5(a), lane 2). In contrast, treatment of IL-6 resulted in an increase in nuclear protein binding to the probe (lane 3), and the intensity of shift band was further increased by the combined addition of IL-6 and Dx (lane 4). The nuclear protein binding induced by IL-6 + Dx was competed with an excess of unlabeled probe (lane 8). On the other hand, excess amount of unlabeled mutated probe (probe M1), which disrupted the STAT-binding site, could not competitively block the binding to the labeled probe (lane 12). These results showed the binding specificity of the IL-6 + Dx-induced complex. An antibody for STAT3 produced a supershift band ( Figure 5(b)), suggesting that STAT3 was involved in this nuclear protein-DNA complex formed by the IL-6 + Dx stimulation. Furthermore, immunoprecipitations of cross-linked DNA-protein complexes derived from IL-6 + Dx-treated 1.1B4 cells with a STAT3 antibody contained the STAT-binding element of the human REG I gene ( Figure 5(c)). These results showed that STAT3 binding to the REG I promoter was induced by IL-6 + Dx in cells.

STAT3 Is Necessary for the IL-6 + Dx-Induced Type I REG Gene
Expression. Involvement of STAT3 in the IL-6 + Dx-induced expressions of REG I and REG I was further demonstrated by siRNA. 1.1B4 cells were transfected with scrambled siRNA control or STAT3-specific siRNA and were stimulated with IL-6 + Dx. We confirmed that STAT3-siRNA suppressed both basal and IL-6 + Dx-induced STAT3 mRNA (Figure 6(a)). In the cells treated with the STAT3-siRNA, the IL-6 + Dx-induced upregulation of REG I and REG I was abolished (Figures 6(b) and 6(c)). These results clearly showed the essential requirement of STAT3 in the IL-6 + Dxinduced type I REG gene expression in cells.

Discussion
Reg family genes are involved in cell proliferation and regeneration in several tissues. In rat, type I Reg family gene (Reg I) is induced during pancreatic cell regeneration [4,7]. The completion of the human genome sequencing project revealed the complete set of REG genes (REG I , REG I , REG III, HIP/PAP, and REG IV) [4,5,50,51]. However, in human, it is not clear which REG gene is involved in the cell regeneration. In this study, we found that among five functional human REG genes, REG I and REG I gene expressions were significantly enhanced by the combined addition of IL-6 and Dx in human cells. The rat Reg I gene expression is activated by IL-6 + Dx [13], and we and others have shown that human REG I protein promotes the proliferation of RINm5F cells and mouse islets [52,53]. Therefore, in human, as well as in rat [13,31], the type I REG proteins could be induced under inflammatory conditions, in which the levels of circulating IL-6 and glucocorticoids are increasing and function as a growth factor for cell proliferation.
We demonstrated that IL-6 + Dx induced the type I REG gene (REG I and REG I ) expression through the JAK/STAT3 pathway. It has been reported that in gastric cancer cells, IL-6 enhanced REG I transcription through the STAT3 activation [54,55]. In addition, the enhancement of REG I promoter activity by IL-6 was reported also in colon cancer cells [56] and in salivary ductal cells [57]. Thus, involvement of STAT3 in REG I gene expression is not restricted in pancreatic cells but rather common to diverse cell types. However, in cells, Dx is required in addition of IL-6 for the induction, since IL-6 alone only slightly induced REG I expression. We recently reported that HGF gene transcription was also induced by IL-6 + Dx through STAT3 activation in pancreatic cells [31]. It is possible that glucocorticoids are necessary for STAT3 activation in addition to IL-6 in pancreatic cells. The role of glucocorticoids in the REG I induction is currently under investigation.
The IL-6 + Dx-induced increases, in the promoter activity and the mRNA level of REG I , were much larger than those of REG I . The reason of the difference of inducibility between REG I and REG I is not clear. The expression patterns of two genes are different in pancreas [58]. In addition, we recently reported that, in 1.1B4 cells treated by intermittent hypoxia, REG I gene expression but not REG I gene expression was increased [23]. Thus, although the JAK/STAT pathway seems to be essential to the induction of both type I REG genes, these two genes should be regulated independently by other transcription factors under physiological and pathological conditions. Previous [23] and current studies suggest that REG I gene rather than REG I gene should be much more important for cells in the circumstances in which cells are damaged.
IL-22 was reported to induce REG I transcription via STAT3 activation in human colon cancer cells [59] and to induce Reg I and Reg II in NOD mouse islets [6,45,60]. In addition, Shioya et al. reported IL-22 receptor expression in human pancreatic cells [61]. However, in the present study, either IL-22 alone or IL-22 + Dx only modestly increased the REG I and REG I promoter activities in 1.1B4 cells. Therefore, STAT3 may not be fully activated by IL-22 nor IL-22 + Dx in human pancreatic cells.
In rat RINm5F cells, the cis-element of rat Reg I gene for induction in response to IL-6 + Dx was the −81∼ −70 region [13]. This region in rat Reg I promoter corresponds to the −76∼−65 region of the human REG I gene, which is downstream of the STAT-binding site (−142∼ −134). In reporter assays, the promoter activity of the construct "−130" was not enhanced by the stimulation of IL-6 + Dx. Moreover, the combined addition of 3AB with IL-6 + Dx did not enhance the promoter activity of mutant construct "−146 M" as much as that of "−146. " These results suggest that although PARP was involved in the IL-6 + Dx-induced activation of REG I promoter, STAT binding to the −142∼−134 region played a major role for the induction of REG I transcription in human cells.
In this study, we found that among five human REG family genes, the expressions of REG I and REG I were significantly enhanced by IL-6 + Dx. These results suggest that, under inflammatory conditions, human type I REG proteins could be increased and function as growth factors for cells to facilitate proliferation. Furthermore, we showed that the JAK/STAT pathway was involved in the IL-6 + Dxinduced human type I REG gene transcription in cells. Therefore, polymorphisms and/or mutations in REG I and REG I promoters as well as other polymorphisms/mutations of factors involved in the JAK/STAT pathway may be involved in diabetes incidence/pathology.