VIP exerts a spectrum of effects as a potent anti-inflammatory factor. In addition, VIP increases expression of MUC2, a major intestinal secretory mucin. We therefore investigated the effects of VIP on the promoter activity of the 5′-flanking region of the MUC2 gene. VIP activated MUC2 transcription in human colonic epithelial cells via cAMP signaling to ERK and p38. cAMP/Epac/Rap1/B-Raf signaling was not involved in MUC2 reporter activation. Furthermore, activation of MUC2 transcription was independent of many of the reported downstream effectors of G protein-coupled receptors, such as PKC, Ras, Raf, Src, calcium, and phosphoinositide 3-kinase. VIP induced cAMP response element-binding protein (CREB)/ATF1 phosphorylation, and this was prevented by treatment with inhibitors of either MEK or p38 and by PKA and MSK1 inhibitor H89. CREB/ATF1 and c-Jun were shown to bind to an oligonucleotide encompassing a distal, conserved CREB/AP1 site in the 5′-flanking region of the MUC2 gene, and this cis element was shown to mediate promoter reporter activation by VIP. This study has identified a new, functional cis element within the MUC2 promoter and also a new pathway regulating MUC2 expression, thus providing further insight into the molecular mechanism of VIP action in the colon. These findings are relevant to the normal biology of the colonic mucosa as well as to the development of VIP as a therapeutic agent for treatment of inflammatory bowel disease.
- G protein-coupled receptor
- inflammatory bowel disease
- protein kinase A
crohn's disease and ulcerative colitis are two forms of chronic human inflammatory bowel disease (IBD) that primarily affect the gastrointestinal tract. These are complex diseases caused by multiple environmental and genetic factors involving the immune system, microbial factors, and the intestinal epithelial barrier (7, 25). The intestinal epithelium is covered by a continuous layer of mucus that provides a physical barrier for the underlying cells. The physical and chemical properties of mucus are attributed largely to its constituent mucins, which exhibit high viscoelasticity and hydrodynamic volume, physical properties arising from their size, glycosylation, charge, and disulfide cross-linking. In IBD, quantitative and qualitative changes in mucins have been reported to occur. The number of goblet cells that synthesize mucin is reduced in active disease, so the mucus layer is thinner (9, 19, 60, 65). It has also been shown that MUC2 synthesis, secretion, and sulphation are all reduced in active ulcerative colitis, resulting in increased exposure of mucosal cells to toxic agents and pathogens and impaired wound-healing capability (67). Muc2 knockout mice may be more susceptible to dextran sulfate sodium-induced colitis; IL-10 knockout mice express less Muc2 and are also more susceptible (19).
Mucins are very large glycoproteins featuring O-glycosylated, tandemly repeated serine- and threonine-rich regions (14, 46). Genomic and cDNA sequencing have identified at least 15 different mucin genes, which encode either secretory or membrane-associated proteins. MUC2 encodes a large (>5,000-amino acid residue) gel-forming mucin and is the most abundant secretory mucin in the intestine, playing a key role as a barrier against bacterial toxins and small molecules. As for all mucins, the expression of MUC2 is cell and tissue specific and, within the intestinal epithelium, it is expressed by goblet cells. MUC2 levels can vary during cell differentiation and inflammation and are altered during carcinogenesis (5, 26). MUC2 transcription can be enhanced by growth factors, cytokines, and bacterial components, indicating that diverse ligands, receptors, and signaling pathways are involved in its regulation (reviewed in Ref. 68).
Colonic goblet cells are located in the vicinity of enteric nerves, which suggest the possibility of regulation by neurotransmitters such as VIP, a 28-amino acid peptide abundantly expressed by both colonic nerve cells and enteroendocrine cells (29, 61). This neuropeptide hormone has been reported to regulate the immune, central nervous, and gastrointestinal systems (58). In particular, it has been reported to antagonize inflammatory responses by inhibiting the production of proinflammatory cytokines such as TNF-α, IFN-γ, IL-6, and IL-12, to reduce the activity of inducible nitric oxide synthase, and to enhance the production of the anti-inflammatory cytokines IL-10 and IL-1 receptor antagonist (IL-1Ra) (12, 16). In cultured colonic epithelial cells, which are known to express the VIP receptor VPAC1, VIP increased mucin production at both the mRNA and protein level (24, 41, 50, 56). Interestingly, it has been reported that administration of VIP ameliorated trinitrobenzene sulfate (TNBS)-induced colitis in mice (1). Hence, the addition of VIP to the current treatment regimen could potentially enhance the effectiveness of therapy and prevention of recurrence of IBD.
Because VIP generally exerts inhibitory, antagonistic effects against proinflammatory cytokines and LPS, MUC2 upregulation by VIP was expected to occur along pathways distinct from those previously described for proinflammatory cytokines, phorbol ester, and LPS (2, 33, 32, 68). A detailed examination of the intracellular signaling mechanisms involved in VIP-induced MUC2 transcription reveals that cAMP signaling, PKA, p38, and ERK activate cAMP response element-binding protein (CREB)/ATF1, which exerts its effects at a distal, conserved cAMP-responsive element (CRE).
MATERIALS AND METHODS
TriReagent, VIP (1–28), forskolin, 8-bromoadenosine 3′,5′-cyclic monophosphate (8-Br-cAMP), PMA, Ro 318220, PD153035, SB 203580, Clostridium difficile toxin B, adenosine, 5-hydroxytrypamine, and bradykinin were obtained from Sigma. 8-(4-chlorophenylthio)-2′-O-methyladenosine-3′,5′-cyclic monophosphate (8-CPT-2Me-cAMP) was obtained from BIOLOG. Dual-Luciferase Reporter Assay System was from Promega. H89, PP2, SU6656, PD98059, SP600125, wortmannin, and bisindolylmaleimide I were purchased from Calbiochem. AG1478 [4-(3-chloroanilino)-6,7-dimethoxyquinazoline HCl], D609 (tricyclodecan-9-yl xanthogenate·K), SB202190, BAPTA-AM, Y27632, and PGE1, PGE2, and PGF2 were purchased from Biomol. Antibodies against CREB (sc186X) and c-Jun (sc45X) were purchased from Santa Cruz Biotechnology. Antibodies to ERK1/2(p44/42 MAPK) (#9101), SAPK/JNK (#9251), p38 (#9211), phospho-ERK1/2 (#9101), phospho-JNK (#9251), phospho-CREB/ATF1 (Ser133) (#9191), phospho-p38 (#9211), and UO126 were purchased from Cell Signaling. MUC2 antibody Ccp58 was from Biomeda. Secondary antibody-horseradish peroxidase conjugates were purchased from Zymed Laboratories. Oligonucleotides were synthesized by Operon/Qiagen.
HM3 cells were maintained at 37°C in 5% CO2 atmosphere in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, with added pyruvate, penicillin, and streptomycin. HM3M2 cells are an HM3 cell line stably transfected with MUC2 promoter region −2864/+19 inserted into the pGL2 luciferase vector (Promega) (33).
RNA isolation, RT-PCR, and Western blot analysis.
These experiments were performed as described previously (32).
Transient transfection and luciferase reporter assays.
Plasmids were prepared using the Plasmid EndoFree MAXI Prep Kit from Qiagen. pFLAG-CMV2-RapGAP (RapGAP) and p-CGN-Rap1A 63E (Rap63E) were generous gifts from Dr. L. A. Quillam (Indiana University) (37). pRK5-HA-Rap1A V12 (RapV12) and pRK5-HA-Rap1A-S17N (Rap17N) were gifts from Dr. J. Gunzburg (Institut Curie, INSERMU-528) (21). pCDNA3-B-Raf-CAAX (BRafCAAX) and pCDNA3-Nter-B-Raf (DNBRaf) were generously provided by Dr. R. Busca (INSERM U385) (8). Epac·cAMP was a kind gift Dr J. L. Bos (University Medical Center) (13). Dominant negative CREB (KCREB) was kind gift from Dr. R. H. Goodman (Oregon Health Science University) (71). Expression vector for dominant negative Ras (RasN17) was obtained from Dr. G. Cooper (Boston University); expression vector for dominant negative Raf (DNRaf) was obtained from J. D. Li (University of Southern California, Los Angeles, CA). MUC2 promoter reporter assays employed pGL2 vector (Promega) containing various regions of the MUC2 gene 5′-flanking sequence, as described previously (22, 32). Cells were typically transfected in 24-well plates with 5 μl of Superfect (Qiagen), plus 4 μg of test plasmid and 0.2 μg of pRLO (a promoterless Renilla luciferase vector; Promega) as an internal control. After 1 day, the cells were serum starved overnight with 1% fetal bovine serum and then treated with VIP. Luciferase activities were measured using Dual Luciferase Reporter Assay System (Promega), with the Monolight 2010 Luminometer (Analytical Luminescent Laboratory). Firefly luciferase activity measurements were normalized with respect to pRLO Renilla luciferase activity to correct for variations in transfection efficiency. Inhibitor experiments using stably transfected HM3M2 cells included a 1-h inhibitor pretreatment before VIP activation. Data is presented graphically as the average of triplicates with statistical significance determined by two-tailed, paired Student's t-test.
EMSAs were performed as described previously (2). Double-stranded oligonucleotide probes used in this study contained MUC2 promoter sequence from −2583 to −2554 [5′- TCAGGTGCCATGTGACGTCAGTGCTGCCT-3′; wild-type (WT), putative CRE site is underlined]. Mutant −2583 to −2554 CREB oligonucleotide (mut) had sequence: 5′-TCAGGTGCCATGTGGTACCAGTGCTGCCT -3′ (altered bases in boldface). Probes were labeled by annealing single-stranded reverse-complimentary strands with overhanging 5′-ends and then filling in with Klenow fragment and [α32P]dCTP.
HM3M2 cells were seeded onto 1-cm glass coverslips in a 24-well plate. Cells were serum starved and treated with 10−7 M VIP for indicated times, then fixed for 10 min with 4:1 methanol/acetic acid. Wells containing cells on coverslips were rinsed with methanol, treated with 3% peroxide, and rinsed with PBS. Fixed cell monolayers were blocked and stained using Zymed Histostain Plus Broad Spectrum kit (Zymed, South San Francisco, CA) according to the manufacturer's instructions, using Ccp58 (1:200) antibody against MUC2, with final detection using Streptavidin-Alexa488 (1:1,000; Molecular Probes). Nuclei were counterstained red using propidium iodide, and then coverslips were mounted onto slides using Prolong Antifade mounting medium (Molecular Probes). Fluorescently labeled cells were viewed and photographed using a charge-coupled device video camera attached to a Zeiss Universal Research microscope illuminated using a mercury arc lamp, with image acquisition using Photoshop 3.0 and Scion Image software.
VIP stimulates MUC2 protein, gene expression, and transcriptional activity.
Immunohistochemical staining for newly synthesized, nonglycosylated MUC2 protein showed an increase in the number of positively stained HM3M2 cells (Fig. 1A). This result was confirmed by RT-PCR analysis of RNA isolated from VIP-treated and untreated cells, which showed increased levels of MUC2 message within 4 h (Fig. 1B). Consistent with these results, VIP upregulated transcriptional activity of the MUC2 luciferase reporter in HM3M2 cells in a dose-dependent manner up to 10−7 M (Fig. 1C). MUC2 promoter reporter activity showed a perceptible increase within 2 h, peaking at 4–6 h (Fig. 1D).
VIP induced phosphorylation of ERK, p38, and JNK in HM3M2 cells.
To investigate the involvement of downstream signaling pathways such as the MAP kinase pathways, we undertook Western blotting studies using antibodies against MAP kinase proteins. As shown in Fig. 2, ERK1/2 showed basal phosphorylation, which was increased slightly by VIP. p38 was also phosphorylated in response to VIP treatment. SAPK/JNK, which typically occurs as multiple bands due to alternate splicing of long (p54) and short (p46) forms, exhibited slight basal phosphorylation but also showed increased phosphorylation after VIP treatment. Whereas p38 and JNK both showed rapid, distinct activation by VIP, ERK1/2 phosphorylation was delayed and less pronounced.
Small molecule inhibitors were employed to identify which of these three pathways was responsible for conveying signals from the VIP receptor. MEK inhibitor UO126 was able to inhibit MUC2 reporter induction in a dose-dependent manner (Fig. 3A), but p38 inhibitor SB202190 was much more effective (Fig. 3B). A similar result was obtained with p38 inhibitor SB203580 (not shown). MEK and p38 inhibitors also reduced basal promoter activity, consistent with the Western blot results, which showed basal as well as VIP-induced phosphorylation of ERK and p38 (Fig. 2). JNK inhibitor SP600125 failed to inhibit MUC2 promoter activity and, in fact, enhanced it (Fig. 3C). MUC2 reporter activation by SP600125 could be prevented by p38 inhibitor SB203580 and by MEK inhibitor PD98059 but not by protein kinase C inhibitor bisindolylmaleimide I or protein kinase A inhibitor H89 (Fig. 3D). Thus VIP's effects on MUC2 gene transcription occur primarily through p38 signaling, with some contribution from the MEK/ERK signaling pathway and with negative regulation by JNK.
cAMP, but not Epac/Rap1/BRaf, mediates MUC2 activation.
Small molecules were further employed in reporter assays to identify some of the signaling events upstream of p38 and ERK. Pretreatment of cells with PKA inhibitor H89 significantly inhibited MUC2 reporter activation by VIP (Fig. 4A). However, although H89 is frequently used to implicate PKA in signal transduction, it is also a known and effective inhibitor of mitogen- and stress-activated protein kinase (MSK1) and Rho-associated coiled-coil forming protein kinase ROCKII (10).
Treatment of HM3M2 cells with the adenylate cyclase activator forskolin enhanced MUC2 reporter activity (Fig. 4B). Several physiological mediators of cAMP signaling were also tested for their abilities to activate the MUC2 reporter. Prostaglandins PGE1, PGE2, and PGF2, as well as adenosine were found to be effective activators (Fig. 4C) as reported by others (41, 42, 56). Treatment of HM3M2 cells with a stable cAMP analog 8-Br-cAMP effectively enhanced MUC2 promoter activity in a dose-dependent manner, indicating possible involvement of adenylate cylase and cAMP (Fig. 4D). In the same experiment, 8-CPT-2Me-cAMP, a cAMP analog that selectively stimulates Epac, was unable to stimulate reporter activity, indicating that the Epac/Rap1 pathway did not mediate the effects of VIP (Fig. 4D) (20).
To confirm this result, the cAMP/Epac/Rap1/B-Raf pathway was assessed further. B-Raf is a Raf1 homolog, Rap1 is its upstream small G protein GTPase, and Epac is the Rap1 guanine exchange factor. Cotransfection of MUC2 promoter reporter with an expression vector encoding constitutively active Epac (Epac·cAMP) increased reporter activity, and reporter activity was further enhanced by treatment with VIP (Fig. 5A) (13). These results suggest that Epac may be absent or inactivated in HM3M2 cells, hence the failure of 8-CPT-2Me-cAMP to activate MUC2 reporter activity.
Cotransfection of MUC2 reporter with vectors encoding proteins expected to either inhibit Rap1 signaling (RapGAP and Rap1A-17N) or to activate Rap1 signaling (RapV12 and Rap1A-63E) had no significant effect on responsiveness to VIP, which tended to be reduced in all assays (Fig. 5B) (21, 37). We attribute these results to inhibition of Ras by active Rap1 forms and/or due to inhibition by indirect or nonspecific pathways.
Raf1 was previously shown to mediate MUC2 promoter activation via MEK/ERK signaling (32). Unlike Raf1, B-Raf may be activated by either Ras or Rap1 (8, 70). Cotransfection of the MUC2 reporter with DN-B-Raf failed to inhibit both basal and VIP-induced promoter activity, suggesting that B-Raf does not mediate the effects of VIP (Fig. 5C). On the other hand, overexpression of a constitutively active form of B-Raf, B-RafCAAX (8), dramatically enhanced MUC2 promoter activity, likely via MEK/ERK. These results indicate that B-Raf is not activated in this cell line under basal conditions or upon VIP stimulation. Cotransfection with expression vectors encoding the dominant negative forms of Ras (RasN17) and Raf1 (DNRaf) also failed to inhibit (Fig. 5D), and a similar result was obtained for the more potent activator forskolin (not shown).
Inhibitor studies provided evidence for MSK1 involvement in MUC2 up-regulation.
The inhibitory effects of H89 indicated that the effects of VIP on MUC2 transcription could be mediated by PKA, MSK1, or ROCKII (10). ROCKII-specific inhibitor Y 27632 (IC50 = 800 nM in vitro for ROCKII vs. 8.3 μM for MSK) failed to inhibit MUC2 promoter reporter activity, whereas MSK1 inhibitor Ro 318220 (IC50 = 8 nM in vitro) inhibited MUC2 reporter response to VIP in a dose-dependent manner (Fig. 6A) (10, 11). Ro 318220 is also a potent inhibitor of PKC, but PKC involvement in VIP signaling was discounted using inhibitor bisindolylmaleimide I, which completely failed to inhibit promoter-reporter activity at a concentration (1 μM) that gave 94 ± 9% inhibition of PMA-induced activity (Fig. 6A).
Other chemical inhibitors were employed to help determine whether cAMP-independent pathways were contributing to the effects of VIP on MUC2 upregulation. Intracellular calcium chelator BAPTA-AM (10 μM) and phosphatidylinositol 3-kinase inhibitor wortmannin (250 nM) did not inhibit (not shown). EGFR signaling is known to activate MUC2 transcription (53, 64), and adenosine GPCR-mediated upregulation of MUC2 promoter activity was recently shown to involve cross-talk with the EGFR (42). In that report, 10 μM AG1478 gave 60% inhibition of reporter activity (42). Figure 6B shows that more modest but significant inhibition of MUC2 reporter activity was achieved in our system using EGFR kinase inhibitors AG1478 and PD153035. Src inhibitors SU6656 and PP2 failed to inhibit significantly (Fig. 6C), and treatment of cells with up to 50 ng/ml Rac/Cdc42 inhibitor Clostridium difficile toxin B also failed to inhibit VIP-induced MUC2 upregulation (not shown). D609, widely used as a phosphatidylcholine-dependent phospholipase C inhibitor (25 μg/ml) inhibited both basal and VIP-mediated reporter activity very effectively (53 ± 6 and 63 ± 2% inhibition, respectively), as reported by others studying nucleotide-induced mucin secretion from airway epithelia (27, 62).
A distal CRE mediates MUC2 promoter activation by VIP.
Because CRE often play a major role in cAMP-induced transcription of genes (40), we analyzed the MUC2 promoter sequence using MatInspector [Release professional 7.0; www.genomatix.de (59)], which identified only a single site located at −2571 to −2563. Transient transfection assays with truncated MUC2 promoter-reporter constructs, followed by treatment with VIP, showed greater activation of the −2864/+19 construct compared with −1628/+19, which lacked the distal CRE (Fig. 7A). Sequence alignment of mouse and human MUC2 5′-flanking sequences (MacVector 6.5.3) revealed that very little homology exists outside of the proximal 200 bp of sequence, except for three small stretches centered at −640, −1030, and −2570 (Fig. 7B). The putative CRE was located within the latter conserved region. EMSA experiments, using an oligonucleotide probe which was designed from the MUC2 sequence encompassing this site, showed binding to proteins in the nuclear extracts from VIP-treated and untreated HM3M2 cells (Fig. 7C, lanes 2–5). The intensities of the shifted bands were not increased by VIP treatment, indicating that DNA binding by nuclear proteins was constitutive. The lower band was supershifted when nuclear protein was preincubated with antibody against CREB/ATF1 (lane 6). The CRE consensus sequence is very similar to that of the TPA-responsive element (TRE) bound by AP-1, and one of the AP-1 subunits (c-Jun) is also known to bind to CRE (18). Thus it was not unexpected to find that the upper band was supershifted when nuclear protein was preincubated with antibody against c-Jun (lane 7). In addition, the shifted bands were displaced by the addition of unlabeled competitor oligonucleotide (lane 8), but addition of unlabeled competitor oligonucleotide containing mutations within the CRE/TRE consensus sequence displaced neither the bound CREB nor c-Jun (Fig. 7C; lane 9).
CREB is known to constitutively bind to DNA and to become phosphorylated upon activation and nuclear import of PKA (40). Because CREB activation was not evident by EMSA, we used Western blot analysis to show that CREB/ATF1 phosphorylation was observed 15 min after addition of VIP (Fig. 7D). This phosphorylation of CREB/ATF1 could be inhibited by either the MEK inhibitor U0126 or by p38 inhibitor SB202190 (Fig. 7E), indicating that both of these pathways are required for CREB phosphorylation and consistent with the results of the reporter assays using MEK and p38 inhibitors (Fig. 3D). To clarify the contribution by PKA, Western blots were carried out to determine whether the inhibitory effects of H89 were attributable to its effects on MSK1 alone or whether it affected phosphorylation of upstream MAP kinases. As shown in Fig. 7, E and F, H89 reduced the phosphoryation of ERK1/2 and, to a much lesser extent, the phosphorylation of p38 and CREB. Phosphorylation of JNK was unaffected, indicating that activation of this kinase was PKA independent.
Transient cotransfection with dominant negative CREB (KCREB) reduced MUC2 promoter-reporter activity (Fig. 7G). The functionality of the −2571 to −2563 CRE/TRE site was further tested by deleting this region from the −2864/+19 MUC2 promoter reporter construct. As shown in Fig. 7H, the responses of two different CRE-deleted mutant reporters to both VIP and forskolin was reduced by 40% relative to the parental vector.
EMSA had shown that c-Jun as well as CREB/ATF1 bound to the MUC2 CRE oligonucleotide and that these two transcription factors bound in a mutually exclusive fashion, suggesting that the activating effects of JNK inhibitor SP600125 might be mediated by this cis element. Therefore, the CRE-mutated reporter constructs Δ−2582/−2566 and Δ−2595/−2566 were tested with SP600125, with and without VIP (Fig. 7I). These mutant reporters were only slightly less responsive than the wild-type reporter, indicating that the CRE was not the SP600125-responsive element. An evaluation of a series of promoter-reporter deletion constructs showed that all were activated by SP600125, supporting this conclusion and indicating that a JNK-repressive element lies within the conserved proximal promoter region (results not shown).
MUC2 secretion is one of the major functions of mature colonocytes. VIP has been reported to enhance secretion and expression of MUC2 in colonic epithelial cell lines (50). We have shown that VIP upregulates MUC2 protein, message, and promoter-reporter activity in a time- and dose-dependent manner in the HM3 colon epithelial cell line. Further characterization of signaling events has allowed us to identify the signaling pathways and transcription factors mediating the effects of VIP on the MUC2 promoter. These have been summarized, along with previously identified cis elements and transcription factors, in Fig. 8 (22, 33, 43, 53, 52, 74).
VIP has been reported to mediate its effects through VPAC1 and VPAC2, members of the GPCR-B family (31). VPAC receptors are coupled to the Gαs isoform, which stimulates adenylate cyclase to increase intracellular cAMP levels (30, 66). The cAMP thus produced activates PKA, which has been reported to directly phosphorylate several transcription factors, including CREB. Other major cAMP-activated pathways include PKA/Rap1/B-Raf (70) and Epac/Rap1/B-Raf (13), both capable of activating MAP kinases. A cAMP/Ras pathway (38, 54), a cAMP/PKA/Src pathway (28), and a cAMP/PKA/Rac/cdc42 pathway (3, 6, 49 57) have also been described. Rap1 was reported to be a downstream target of VIP signaling in HT29 cells (23), and VIP has also been reported to stimulate nuclear translocation of PKC isoforms (51), which could also activate many signaling pathways. Finally, it has also been reported that, in gastric muscle, VIP stimulates calcium ion influx and consequent activation of nitric oxide synthase and cyclic guanosine 3′,5′-cyclic monophosphate kinase, which could also impinge many signaling pathways (47). It has also been reported that VPAC receptors can couple to Gαq and Gαi to increase intracellular calcium concentrations (31). The GβGγ dimers can also regulate adenylate cyclase, phospholipase Cβ, and ion channels; and their activation of Src and EGFR have also been described (35, 38, 55). Phospholipase Cβ produces inositol 1,4,5-trisphosphate and calcium release as well as diacylglycerol, either of which can activate PKC, a well-known activator of MUC2 transcription (32). Thus multiple pathways could mediate the effects of VIP on MUC2 synthesis.
Adenylate cyclase activator forskolin and cAMP analog 8-Br-cAMP were able to stimulate MUC2 reporter activity, indicating that the conventional adenylate cyclase/cAMP/PKA pathway likely transmitted signals from the VIP receptor to the MUC2 gene. The alternative Epac/Rap1/B-Raf pathway does not seem to be used, and studies using small molecule inhibitors of phosphatidylinositol 3-kinase, Ca2+, ROCKII, Rac/Cdc42, and Src did not implicate any of these cAMP-independent pathways. Western blot analysis indicated that VIP activated ERK, JNK, and p38 kinases. However, small molecule inhibitors of these three pathways indicated that only p38 and, to a lesser extent, MEK/ERK mediated the effects of VIP on MUC2 reporter activity. Although some cross-talk involving EGFR was suggested by inhibitor studies, further experiments are required to identify the signaling events immediately upstream of p38, JNK, and MEK. As reported previously, inhibition of JNK signaling served to activate rather than to inhibit MUC2 reporter activity, indicating that this pathway serves a repressive role.
CREB/ATF1 involvement in VIP-induced MUC2 upregulation was indicated by the following observations: 1) a distal CRE cis element located at bases −2571 to −2563 within the MUC2 promoter is conserved between human and mouse sequences; 2) MUC2 promoter reporters, which include this CRE site, show higher activity and responsiveness to VIP induction than reporters not containing this site; 3) EMSA studies using an oligonucleotide encompassing the putative CRE at bases −2571 to −2563 showed binding by nuclear proteins extracted from HM3M2 cells. One set of bands could be supershifted by antibody against CREB, and the other by antibody against c-Jun; 4) Western blot analysis showed that VIP treatment increased CREB phosphorylation; and 5) site-directed deletion of the CRE reduced transcriptional activation of the MUC2 −2864/+19 promoter reporter.
In addition, a recent and comprehensive analysis of the genome for CREB target genes identified clustered, conserved CRE consensus sequences within the promoters of MUC1 and MUC2 (75). This study also noted that, in general, CREB binding to DNA was constitutive and not inducible by phosphorylation, that it was prevented by DNA methylation, and that CREB phosphorylation was necessary but not sufficient for promoter activation.
In our EMSA studies, the intensity of the shifted band was not affected by VIP treatment, indicating constitutive binding by CREB. This is consistent with the classic model of CREB activation, which includes phosphorylation of the DNA-bound CREB at Ser133, rendering it competent to recruit its coactivator, CREB-binding protein (reviewed in Refs. 34, 40). Many kinases have been found to mediate CREB phosphorylation, including PKA, CaMKIV, RSK, MSK, and MAPKAP K2 (34), and recent studies have identified additional CREB phosphorylation sites and have implicated other factors that affect its activity (17). However, the requirement for both p38 and ERK activity for MUC2 promoter reporter activation points to the involvement of a CREB Ser133 kinase, which requires activation by both ERK and p38 signaling pathways. A likely candidate is MSK1, which has been shown to be phosphorylated by both ERK and p38, at Ser376 and Thr581, respectively (72, 73). A similar requirement for both MEK and p38 signaling has been found for β2-andrenergic receptor stimulation involving calcium cycling (36), for G protein agonists endothelin, for α1-andrenergic agonists in cardiac myocytes (39), for IL-1β and TNF-α induction of CREB in human airway epithelium (63), and for β3-adrenoceptors in 3T3-L1 adipocytes (45).
Ro 318220 is a broad-spectrum inhibitor of PKC-α and MSK1 (10) and is one of the few chemical inhibitors that shows some selectivity for MSK1 vs. PKA (10, 11, 39). This chemical effectively inhibited reporter activation in response to VIP, whereas PKC inhibitor bisindolylmaleimide I was ineffective at concentrations sufficient to completely inhibit phorbol ester-mediated activation, further implicating MSK1 in VIP signaling. MSK-1 has also been reported to phosphorylate H3 histone and also to directly phosphorylate NF-κB (11, 69). Because both promoter methylation and NF-κB have been shown to regulate MUC2 transcription (32, 68), MSK may be found to play additional roles in MUC2 expression.
AP-1 is a homo- or heterodimeric transcription factor composed of members of the Jun and Fos families of DNA binding proteins (4). JNK is known to play an important role in AP-1 activity through phosphorylation of c-Jun (44). We have shown here and previously that the JNK inhibitor SP600125 and dominant negative c-Jun both activate MUC2 promoter reporter activity, indicating that JNK negatively regulates MUC2 through c-Jun (2). The mechanism of how c-Jun functions as a negative regulator of MUC2 transcription has not been determined. However, because c-Jun binds not only to TPA-responsive elements but also to CRE (48), it is possible that c-Jun simply competes for binding of CREB to the CRE. Our EMSA results show that separate, independent bands are shifted by antibodies to CREB and c-Jun, indicating that in vitro binding by CREB and c-Jun is mutually exclusive. An opposing effect by c-Jun and CREB at a CRE site has been reported for VIP-induced TNF-α expression (15). The ability of p38 and MEK inhibitors to completely prevent activation by SP600125 indicates that these three pathways converge at a common point in signal transductions pathways regulating MUC2 transcription. However, activation of CRE-deleted MUC2 promoter-reporter constructs by SP600125 (Fig. 7I) indicated that the distal CRE is not the SP600125-responsive element; and activation of truncated MUC2 promoter-reporter constructs indicated that the activating effects of SP600125 were mediated by the proximal promoter region. Therefore, further experiments are required to confirm and clarify the exact mode of MUC2 transcriptional repression by c-Jun.
This is the first detailed analysis of VIP signaling in colonic epithelial cells, indicating that VIP stimulates phosphorylation of JNK, ERK, and p38. An adenylate cyclase/cAMP/PKA pathway mediates ERK and p38 phosphorylation, and both MEK and p38 are required not only for the subsequent phosphorylation of CREB/ATF1 but also for MUC2 promoter-reporter activation by VIP. New insights into MUC2 gene regulation are provided by the identification a conserved, functional CRE site within its promoter, which is activated by mediators of cAMP signaling, including VIP and prostaglandins. This elucidation of VIP action in a colonic epithelial cell line may assist in the development of strategies that increase MUC2 expression via cAMP signaling pathways to augment mucous layer formation as protection against inflammatory bowel disease. Conversely, these results also provide further evidence that inhibitors of prostaglandin synthesis can reduce mucosal protection.
This work was supported by the Department of Veterans Affairs Medical Research Service, Theodora Betz Foundation Grant, and by Grant CA24321 from the National Cancer Institute.
We are deeply grateful to the following investigators for generously providing expression vectors: Dr. L. A. Quillam (Indiana University) for pFLAG-CMV2-RapGAP and p-CGN-Rap1A 63E; Dr. J. Gunzburg (Institut Curie, INSERMU-528) for pRK5-HA-Rap1A V12 and pRK5-HA-Rap1A-S17N; Dr. R. Busca (INSERM U385) for pcDNA3-B-Raf-CAAX and pcDNA3-Nter-B-Raf; Dr. J. L. Bos (University Medical Center) for Epac·cAMP; Dr. R. H. Goodman (Oregon Health Science University) for dominant negative CREB (KCREB); Dr. G. Cooper (Boston University) for dominant negative RasN17; and Dr. J. D. Li (University of Southern California) for dominant negative DNRaf.
↵* R. Hokari, H. Lee, and S. C. Crawley contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society