Increased Pancreatic -Cell Proliferation Mediated by CREB Binding ...

3 downloads 51 Views 1MB Size Report
Dec 9, 2005 - Genotyping was performed by Southern blot hybridization or PCR as previously ... Imaging was performed using an Olympus BH-2 (Optical Analysis Corp.,. Nashua, NH) .... entirety to ensure sequence fidelity. Luciferase ...
MOLECULAR AND CELLULAR BIOLOGY, Oct. 2006, p. 7747–7759 0270-7306/06/$08.00⫹0 doi:10.1128/MCB.02353-05 Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Vol. 26, No. 20

Increased Pancreatic ␤-Cell Proliferation Mediated by CREB Binding Protein Gene Activation䌤 Mehboob A. Hussain,1‡* Delia L. Porras,2† Matthew H. Rowe,2† Jason R. West,2† Woo-Jin Song,1 Weston E. Schreiber,1 and Fredric E. Wondisford1‡* Departments of Pediatrics and Medicine, Metabolism Division, Johns Hopkins University, Baltimore, Maryland,1 and Department of Medicine, Section of Endocrinology, University of Chicago Hospital, Chicago, Illinois2 Received 9 December 2005/Returned for modification 3 February 2006/Accepted 28 July 2006

The cyclic AMP (cAMP) signaling pathway is central in ␤-cell gene expression and function. In the nucleus, protein kinase A (PKA) phosphorylates CREB, resulting in recruitment of the transcriptional coactivators p300 and CREB binding protein (CBP). CBP, but not p300, is phosphorylated at serine 436 in response to insulin action. CBP phosphorylation disrupts CREB-CBP interaction and thus reduces nuclear cAMP action. To elucidate the importance of the cAMP-PKA-CREB-CBP pathway in pancreatic ␤ cells specifically at the nuclear level, we have examined mutant mice lacking the insulin-dependent phosphorylation site of CBP. In these mice, the CREB-CBP interaction is enhanced in both the absence and presence of cAMP stimulation. We found that islet and ␤-cell masses were increased twofold, while pancreas weights were not different from the weights of wild-type littermates. ␤-Cell proliferation was increased both in vivo and in vitro in isolated islet cultures. Surprisingly, glucose-stimulated insulin secretion from perfused, isolated mutant islets was reduced. However, ␤-cell depolarization with KCl induced similar levels of insulin release from mutant and wild-type islets, indicating normal insulin synthesis and storage. In addition, transcripts of pgc1a, which disrupts glucose-stimulated insulin secretion, were also markedly elevated. In conclusion, sustained activation of CBP-responsive genes results in increased ␤-cell proliferation. In these ␤ cells, however, glucose-stimulated insulin secretion was diminished, resulting from concomitant CREB-CBP-mediated pgc1a gene activation. turn binds to the regulatory subunit of protein kinase A (PKA) and releases the PKA catalytic subunit. Elevation of cAMP also activates the cAMP-regulated guanine nucleotide exchange factors (also known as EPAC) (28). Both PKA and EPAC are intermediary links between GLP-1 signaling and insulin secretion. The cAMP-PKA-CREB signal transduction pathway also increases pancreatic ␤-cell survival by stimulating IRS-2 production (13, 21). PKA links cAMP formation to gene transcription by phosphorylating nuclear cAMP response element binding protein (CREB) at serine 133. Phospho-CREB, bound to the CRE promoter element, recruits the coactivators CREB binding protein (CBP) and the closely related protein p300. CBP and p300 activate gene transcription through their intrinsic histone acetyltransferase activities and through recruitment of other transcriptional coactivator molecules (12, 25, 32). CBP, but not p300, is phosphorylated by the insulin signaling pathway at a unique serine residue (serine 436) (42). This phosphorylation site is adjacent to the CREB binding domain of CBP. Phosphorylation at S436 interferes with and reduces CREB/CBP interaction, even if CREB is also phosphorylated, resulting in reduced transcription rates of target genes. The removal of the CBP phosphorylation site by a serine-to-alanine mutation of CBP (forming the mutant CBP-S436A) results in enhanced basal and cAMP-stimulated transcription rates of CREB-responsive genes that are not suppressed by cellular insulin action (42, 43). Several in vivo models of negative regulation of the cAMPPKA-CREB pathway in ␤ cells (21, 31, 36) are documented. Studies of these models suggest that disruption of the cAMPPKA-CREB pathway is detrimental to ␤-cell function, survival,

Under circumstances of increased metabolic demand, such as pregnancy (2, 4), or in insulin-resistant states (3, 30), islet and ␤-cell masses increase to meet metabolic requirements. The failure of ␤ cells to adapt to metabolic requirements results in relative insulin deficiency and diabetes mellitus. Within the pancreas, three conceptual mechanisms may lead to an increase in ␤-cell mass: (i) neogenesis from nonendocrine pancreatic tissue, such as pancreatic-duct epithelium (2, 3, 27, 28), pancreatic acinar cells (␤ cells derived from non-␤ cells) (3, 38, 39), or undifferentiated cells within the islet (1, 9, 15, 44); (ii) proliferation of ␤ cells (␤ cells derived from existing ␤ cells) (6, 11, 38); and (iii) reduced ␤-cell apoptosis in the context of normal ␤-cell turnover (5, 26). Cyclic AMP (cAMP) signaling is critical in the physiologic function of ␤ cells. This is exemplified by the effects of the incretin hormone glucagon-like peptide-1 (GLP-1), which improves glucose-stimulated insulin secretion from pancreatic ␤ cells, in part by raising intracellular cAMP levels (13, 18, 28). In pharmacologic studies, GLP-1 also stimulated PDX-1 gene expression in pancreatic-duct epithelial cells and stimulates proliferation of ␤ cells (5, 7, 22, 26, 38). Binding of GLP-1 to its ␤-cell receptor elevates intracellular cAMP levels; cAMP in

* Corresponding author. Mailing address: Metabolism Division, Department of Pediatrics and Medicine, Johns Hopkins University, 600 N. Wolfe Street, CMSC 10-113, Baltimore, MD 21287. Phone: (410) 502-5761. Fax: (410) 502-5776. E-mail for Mehboob A. Hussain: [email protected]. E-mail for Frederic E. Wondisford: fwondisford @jhmi.edu. † These authors contributed equally to this work. ‡ These authors contributed equally to this work. 䌤 Published ahead of print on 14 August 2006. 7747

7748

HUSSAIN ET AL.

and proliferation (20, 21). Furthermore, based on studies of pharmacologic stimulation of cAMP production in ␤ cells with GLP-1 or its analogue exendin-4, it is probable that activation of the cAMP-PKA-CREB pathway improves ␤-cell function and increases ␤-cell proliferation (16–18, 22), in part by increasing cAMP-responsive IRS-2 expression (16, 20). However, direct investigations of in vivo models that specifically upregulate the nuclear cAMP signal and gene regulation downstream of CREB-CBP in ␤ cells are lacking. We describe here a unique model of constitutively activated nuclear cAMP signaling in ␤ cells due to expression of a phosphorylation mutant of CBP (CBP-S436A) in mice. The S436A mutation presents in vivo and ex vivo with increased ␤-cell proliferation. Furthermore, in contrast to conclusions based on the models mentioned above, persistent nuclear activation of the cAMP signal at the CREB-CBP level results in reduced ␤-cell function, which is associated with increased production of the transcription coactivator peroxisome proliferator-activated receptor-␥ coactivator-1␣ (PGC-1␣), responsible for the inhibition of glucose-responsive insulin secretion in ␤ cells (41). MATERIALS AND METHODS Mouse model. The CBP-S436A mice have been previously described (43), and they were kept in a mixed background (C57BL/6 ⫻ 129Sv ⫻ DBA/2) for this study. All animal protocols were approved by the Institutional Animal Care and Use Committee of the University of Chicago and that of Johns Hopkins University. Genotyping was performed by Southern blot hybridization or PCR as previously described (43). Studies were performed with heterozygous CBPS436A mutants and wild-type littermates. Homozygous CBP-S436A knock-in mice have not reproduced efficiently and were not used for the present studies. Care was taken to use littermates or animals of similar ages for comparisons. Care was taken to use male and female mice in the studies. No differences between male and female mice were detectable for the parameters measured. Immunohistochemistry and islet morphometry. Pancreata were immediately removed from euthanized mice, weighed, fixed in 4% paraformaldehyde for 30 min or 10% formalin for 4 h, and thereafter kept in 70% ethanol until paraffin embedding. Five-microgram serial sections were prepared for further processing as previously described (19a, 19b). Primary antibodies used were guinea pig anti-insulin (1:2,000; Linco), guinea pig antiglucagon (1:200), guinea pig antipancreatic polypeptide (1:50), rabbit antisomatostatin (1:200), Ki-67 (monoclonal rat anti-mouse Ki-67 antigen TEC-3; DAKO Cytomation, Carpinteria, CA), and rabbit anti-cleaved caspase-3 (Cell Signaling Technology, Danvers, MA). Affinity-purified rabbit anti-phospho-serine 436-CBP (1:25) was generated by Invitrogen (Carlsbad, CA) against peptide PVCLPLKNA(pS)DKRNQQTIL (pS indicates phosphoserine). Phospho-S436-CBP localization was performed on the pancreas removed from a wild-type CD-1 mouse. The pancreas was immediately frozen and embedded into optimal-cutting-temperature medium. Eight-microgram-thick sections were fixed in 4% paraformaldehyde containing phosphatase inhibitor cocktails I and II (Sigma-Aldrich, St. Louis, MO) for 5 min. To test antibody specificity, control studies were performed by coincubating the phospho-CBP antibody with synthesized antigen (1 ␮g/␮l) before proceeding with immunohistology. Antigen localization was performed using appropriate secondary antibodies linked to either horseradish peroxidase or alkaline phosphatase (all from Jackson Immunoresearch, West Grove, PA) and subsequent detection with either 3,5diaminobenzidine (DAB) or alkaline phosphate color reaction, respectively, and signal amplification using a double-immunostaining kit (Envision double-stain system; DAKO). For antigen retrieval before Ki-67 detection, slide sections were treated with citrate buffer, pH 6.0, at 95°C for 30 min. For Ki-67 detection, a rabbit anti-rat secondary immunoglobulin and a catalyzed signal amplification (CSA) procedure (CSA kit; DAKO) were used following the manufacturer’s instruction. For Ki-67 staining, pancreata from four wild-type mice and four CBP-S436A mice approximately 4 months of age were examined. For Ki-67 staining, at least three 5-␮m sections per mouse at least 150 ␮m apart were analyzed. For phospho-S436-CBP detection, catalyzed signal amplification (CSA kit; DAKO) was also necessary for visualization. For fluorescence immunochem-

MOL. CELL. BIOL. istry, appropriate secondary antibodies with fluorescent tags as indicated in the figures (1:500 dilution; Jackson Immunoresearch) were used. Fluorescence nuclear counterstaining was performed with DAPI (4⬘,6⬘-diamidino-2-phenylindole) by using Vector mounting medium containing DAPI. Imaging was performed using an Olympus BH-2 (Optical Analysis Corp., Nashua, NH) upright microscope or a Diaphot (Nikon Instruments Inc., Melleville, NY) inverted microscope, both equipped with charge-coupled-device digital cameras (CCD cameras) and software. Islet area and size were determined using either the MetaMorph (Universal Imaging Corp., Downingtown, PA) or Image-Pro Plus (Medica Cybernetics, Silver Spring, MD) imaging software, following the manufacturer’s instructions. The imaging systems and software were calibrated to calculate area in ␮m2 or selected regions of interest. The images shown in the figures (magnification, ⫻200) were captured as. TIFF files and processed for display with Adobe Illustrator 10.0.03 (Adobe Systems, San Jose, CA). For histological islet size determination, pancreata from three wild-type mice and four CBP-S436A mice (4-month-old littermates) were used. Three 5-␮m sections, 150 ␮m apart from each other, from each mouse were viewed in their entirety. The numbers of islets found per pancreas were not different between wild-type (56 islets) and mutant (60 islets) mice. Totals of 118 wild-type islets and 120 heterozygote CBP-S436A mutant islets per mouse were analyzed for histological size determination and morphometry on two sections 300 ␮m apart. Cross-sectional islet area and ␤-cell were measured with a 10 ⫻ objective and calibration on an Olympus microscope outfitted with a CCD camera and operated with ImagePro software. Islet isolation. Islets were isolated as previously described (19). Briefly, the pancreas was injected with 1 ml of collagenase (Roche Diagnostics Corp., Indianapolis, Indiana) and incubated at 37°C in a total of 5 to 7 ml of digestion solution. Digested and vigorously shaken islets were subsequently washed twice in RPMI 1640 with 10% fetal calf serum (FCS) and 1% penicillin-streptomycin (Gibco-BRL) added and were hand picked under a dissecting microscope and/or purified on a Ficoll gradient. Islets were then cultured in RPMI 1640 (Life Technologies, Inc.), 5.5 mM glucose, and 10% FCS in a humidified incubator (95% air, 5% CO2) at 37°C. Sizes of isolated islets taken in 96-well culture plates were determined within 24 h of islet isolation on an inverted microscope. At least 300 each of wild-type and CBP-S436A isolated islets were sized. Pancreata from four age-matched mice in each group were used. Sizes of the islets were determined by measuring the cross section of each islet viewed on an inverted microscope equipped with a CCD camera and operated with MetaMorph software. In vivo BrdU labeling. One hundred microliters of 10 mg/ml bromodeoxyuridine (BrdU), in phosphate-buffered saline, was administered intraperitoneally 4 h before harvesting, fixing, embedding, and serial sectioning of the pancreas. Detection of BrdU incorporation was performed using a monoclonal anti-BrdU antibody (Zymed, South San Francisco, CA) using the manufacturer’s instructions. Care was taken to include intestinal or spleen sections on the slides as internal positive controls. Four wild-type mice and four CBP-S436A mice 4 months of age were used for BrdU studies. Formalin-fixed, paraffin-embedded 5-␮m sections 150 ␮m apart were examined. In vitro proliferation. For in vitro proliferation assays, islets were cultured for 4 days. This period of in vitro culture time is required for islets to adjust to a new standardized environment without confounding extrapancreatic metabolic effects (33). Assays were performed on five independent isolations of wild-type and mutant islets from mice approximately the same age. Care was taken to pick an equal number (n ⫽ 25 or 50) of wild-type and mutant islets of approximately equal sizes, and these islets were cultured for 4 days in RPMI 1640 with 5.5 mM glucose, 10% FCS, and 1% penicillin-streptomycin prior to the assay. Proliferation assays were performed using a colorimetric assay based on tetrazolium salt conversion to formazan (XTT; Roche Biosciences, Indianapolis, IN) (37). The formazan concentration was determined by direct measurement on a plate reader at the 455-nm wavelength, from which the background wavelength at 650 nm was subtracted according to the manufacturer’s instructions, after 4 h of incubation with the supplied reagents. Islet perifusion. Islet perifusion was performed according to the methods described in reference 13, with minor modifications. In brief, a fast protein liquid chromatography apparatus (Pharmacia-LKB, NJ) with two pumps was used to mix a 2.8 mM buffer with a 20 mM glucose Krebs Ringer buffer (pH 7.4, containing 2.2 mM Ca2⫹, and 10 mM HEPES salt buffer) to the desired glucose concentrations as follows: 2.8 mM for 30 min followed by 11.1 mM, then 2.8 mM, 16.7 mM, and 2.8 mM for 7 or 10 min each. The buffer was heated to 37°C and islets were perifused (1 ml/min) in a plastic perifusion chamber (Swinnex 25; Millipore) that was kept in a 37°C water bath. The perifusate was collected in 1 ml fractions in which insulin concentrations were determined. At the end of the

CREB-CBP INTERACTION AND ␤-CELL PROLIFERATION

VOL. 26, 2006

perifusion assay, islets were subjected to 30 mM KCl (1 ml/min for 1 min) and perifusate was collected for an additional 10 min at 1 min intervals. Perfusate insulin was measured by enzyme-linked immunosorbent assay (ultrasensitive rat insulin enzyme-linked immunosorbent assay kit; Crystal Chem, Downers Grove, IL), following the manufacturer’s instructions. After perifusion, islets were collected from the perifusion chamber for phenol-chloroform DNA extraction and determination. Insulin secretion was normalized to total islet DNA content. IRS-2 and PGC-1␣ real-time PCR detection. Islets were isolated and cultured for 48 h as described above. For forskolin stimulation tests, islets were incubated with or without (controls) forskolin 10 ␮M for the entire incubation period. Islet RNA was harvested using a standard protocol. cDNA was prepared with 0.7 ␮g of total RNA using iQ SYBR green supermix (Bio-Rad Laboratories, Inc., Hercules, CA). Real-time PCR for IRS-2 and PGC-1␣ was performed using 18S rRNA as a control on a MyiQ single-color real-time PCR detection system (Bio-Rad). Primers used were as follows: 18S forward (fw), 5⬘-GGCGGCTTG GTGACTCTAGAT-3⬘; 18S reverse (rv), 5⬘-CTTCCTTGGATGTGGTAGCC G-3⬘; IRS-2 fw, GCGGCCTCATGTTCTTCACT-3⬘; IRS-2 rv, AACTGAAGT CCAGGTTCATATAGTCA-3⬘; PCG-1␣ fw, ATCCGAGCGGAGCTGAAC AAG; and PGC-1␣ rv, GCGGTGTCTGTAGTGGCTTGA-3⬘. Cycle conditions for both IRS-2 and PGC-1␣ were as follows: 95°C for 3 min followed by 50 cycles of 95°C for 10 s, 60°C for 30 s, and 72°C for 30 s, followed by 95°C for 1 min and 55°C for 1 min, and then 80 cycles at 55°C for 10 s with an increase of 0.5°C in the set-point temperature after cycle 2. Cell culture, transfection, and luciferase expression studies. To test the specificity of the mutation of CBP-S436A with CREB interaction, we compared the effects of wild-type and S436A mutant CBP on transcriptional activity of CREB and HIF1␣ using a GAL4 binding domain (GAL4-BD) protein fusion system. 293T cells were grown at 37°C in 5% CO2 in Dulbecco’s modified Eagle’s medium (Gibco-BRL) supplemented with 10% fetal calf serum and 1% penicillin-streptomycin (Gibco-BRL). Cells were transfected with the respective plasmids four hours after they were split into 24-well culture plates (Biocrest, Cedar Creek, TX) with Fugene (Roche Biosciences, Indianapolis, IN) according to the manufacturer’s instructions. The plasmids used were the reporter plasmid carrying four concatemeric GAL4 DNA sequences upstream of a minimal promoter and the firefly luciferase cDNA pFR-Luc (Biocrest) (50 ng/well), CMV-Gal4BD-CREB (Biocrest) (5 ng/well), CMV-GAL4-BD-HIF1␣ (5 ng/well), pCMVDNA3.1-CBP (73) (150 ng/well), and pcDNA3.1-CBP-S436A (43) (150 ng/well). CMV-Gal4-BD-HIF1␣ was generated by standard cloning procedures, by insertion into the CMV-Gal4-BD pFA plasmid (Biocrest). A full-length HIF1␣ cDNA generated by reverse transcription-PCR (RT-PCR) from mouse liver mRNA (FastTrack; Invitrogen). The primers used were 5⬘-ATCACCATGGAG GGGCGCCGGCGGCGAG-3⬘ and 5⬘-ATGGTACCTCAGTTAACTTGATCC AAAGCTCTGAG-3⬘. The cloned HIF1␣ cDNA insert was sequenced in its entirety to ensure sequence fidelity. Luciferase activity in cells was measured as previously described (43). Measurements were made in duplicate. At least three series of transfection studies were conducted for each GAL4-transcription factor fusion construct. Results were normalized to luciferase activity without the presence of transcriptional coactivators or GAL4 constructs. The total amount of transfected DNA per well was adjusted by adding a cytomegalovirus promoterdriven plasmid expressing dsRed fluorescent protein (Clontech). Statistical analysis. Results are shown as averages and standard errors of the means (SEM). Where appropriate, Student’s t test was used to analyze differences between wild-type and heterozygous CBP-S436A mutant islets. A P value of ⬍0.05 was considered significant. Areas under the curves from islet perifusion studies were calculated with Prism 4 software (Graph Pad Software Inc., San Diego, CA).

RESULTS Islet morphometry. Pancreas weights at the time of excision for 4-month-old hemizygous CBP-S436A (mutant) and wildtype littermates were not different (121 ⫾ 13 mg versus 124 ⫾ 15 mg) (Fig. 1a). The total numbers of islets, determined from histological analysis of pancreas sections, were also not different between wild-type (56 islets) and mutant (60 islets) mice. In mutant mice, however, both the islet cross-sectional area (22,112 ⫾ 140 ␮m2, versus 12,323 ⫾ 73 ␮m2 [for wild type]; P ⬍ 0.05) (Fig. 1b) and the number of ␤ cells per islet cross section (140 ⫾ 11, versus 72.7 ⫾ 8.5 [for wild type]; P ⬍ 0.05) (Fig. 1c) were approximately twice those found from wild-type

7749

FIG. 1. Comparison of pancreas weights (a), islet areas (b), numbers of ␤-cell nuclei/islet (c), and calculated ␤-cell areas (d) of wildtype and CBP-S436A mutant littermates. Pancreas weights are not different. An increased number of ␤ cells/islet in the face of unchanged ␤-cell area (␤-cell size) reflects ␤-cell hyperplasia. The numbers of ␤ cells are increased in mutants. NS, nonsignificant.

littermates. In contrast, calculated ␤-cell areas of mutant (158 ⫾ 41 ␮m2) and wild-type (188 ⫾ 26 ␮m2) mice were similar (no significant difference) (Fig. 1d) and within the range of expected normal mouse ␤-cell sizes (40). This result suggests that the enlarged size of the islets was due to ␤-cell hyperplasia and not hypertrophy. By use of histological pancreatic sections, islet size histograms (Fig. 2a) and islet size distribution (Fig. 2b) for wild-type and mutant mice were determined. Mutant animals have a higher percentage of larger islets with a biphasic distribution. A third population of even larger islets, ranging in size from 50,000 to 75,000 ␮m2, was also found in mutant mice. Mutant animals also had altered islet architecture, with an increased number of non-␤ cells found in the center of the islet (Fig. 2a). Thirty percent of ⬎50,000-␮m2 islets in the CBP-S436A mutant mice have such altered architecture, while no such changes were found in their wild-type counterparts. Next, intact islets from pancreata obtained from wild-type and mutant mice were isolated and cultured, and over 300 islets of each genotype were examined. The cross-sectional areas of islets were determined using an inverted micro-

7750

VOL. 26, 2006

CREB-CBP INTERACTION AND ␤-CELL PROLIFERATION

7751

FIG. 3. Phospho-serine 436-CBP in islets. Immunohistochemistry on frozen murine pancreas sections treated with phosphatase inhibitors and fixed with 4% paraformaldehyde. Sections were incubated overnight with antibody directed against phospho-serine 436-CBP (1:25) without preincubation (a) or with a preincubation with antigen (1 ␮g/␮l) (b). Detection was performed with a secondary antibody linked to horseradish peroxidase and catalyzed signal amplification. Antigen localization was performed with a DAB reaction. Bright-field images (magnification, ⫻200) captured on a color CCD camera with appropriate software are shown. Antigen detection in tissue reveals phospho-serine-CBP in nuclei of islet cells.

scope as described in Materials and Methods. Figure 2c shows representative images of isolated and cultured islets. The size distribution of islets, expressed as percentage of all islets, is shown for wild-type and mutant mice in Fig. 2d. As in the histological analysis of islets, mutant islets have a higher percentage of larger islets with a biphasic distribution, with, again, a small increase in islets with a size range of 50,000 to 100,000 ␮m2. CBP phosphorylation in pancreatic islet cells. To assess whether CBP is phosphorylated in islet cells, we performed an immunohistochemical analysis of frozen pancreatic tissue from wild-type mice with an antibody directed against phosphoS436-CBP. Frozen tissue was sectioned and immediately fixed and treated with phosphatase inhibitors. We detected positive staining for phospho-S436-CBP in nuclei of pancreatic islet cells (Fig. 3a). The specificity of this immunostaining was ensured by coincubating the antigen, which was used to generate the antibody; use of this antigen resulted in no detectable islet staining (Fig. 3b). Proliferation markers. A representative stain for the proliferation marker Ki-67 is shown in Fig. 4. Panels a to c show sections from wild-type mice, and panels e to g show sections from mutant littermates. Figure 4h was obtained by coimmunostaining for Ki-67 and insulin, and it demonstrates that Ki-

67-positive cells also stain positively for insulin (Fig. 4h). In coimmunostaining, we found that all Ki-67-positive islet cells were also positive for insulin. We were unable to detect any insulin-negative, Ki-67-positive cells in islets. Ki-67-positive nuclei were found at rates of 0.71 ⫾ 0.19 nuclei per wild-type islet and 1.71 ⫾ 0.19 nuclei per islet in the mutant mice (means ⫾ SEM, P ⬍ 0.05, Student’s t test) (Fig. 4d). Representative stains for BrdU incorporation is shown in Fig. 5. Panels a and b show sections from wild-type mice, and panels d and e show sections from mutant littermates. BrdU incorporation was observed in 0.28 ⫾ 0.11 nuclei per islet in wild-type mice, as opposed to 0.61 ⫾ 0.16 in mutant mice (means ⫾ SEM, P ⬍ 0.05) (Fig. 5e). Apoptosis marker. On histological analysis, we did not detect any morphological signs of increased cell death (e.g., small, dense nuclei, fragmented nuclei, necrosis, or infiltration with leukocytes) in islets of wild-type or CBP-S436A mutant littermates. Immunohistochemistry for cleaved caspase-3 revealed no indication of cellular apoptotic activity in islets of wild-type or heterozygous CBP-S436A mutant littermates (Fig. 6c to f), whereas cleaved caspase-3 was clearly detectable on histological sections of islets from mice, which had been subjected to streptozotocin treatment 48 h prior to pancreas tissue harvest (Fig. 6a and b).

FIG. 2. Islet histology and morphometry (a and b). Three 5-␮m sections 150 ␮m apart from one another were viewed in their entirety in each mouse. The numbers of islets found per pancreas of wild-type (56) and mutant (60) mice were not different. A total of 118 wild-type and 120 heterozygote CBP-S436A mutant islets were analyzed. Cross-section islet area and ␤-cell numbers were measured with a 10⫻ objective and calibration on an Olympus microscope outfitted with a CCD camera and operated with ImagePro software. (a) Representative images of islets of wild-type (left) and heterozygote CBP-S436A mutants (right). Islets are immunostained with an antibody cocktail against glucagon, somatostatin, and pancreatic polypeptide and a green fluorescent secondary antibody. Nuclei are counterstained with DAPI. Mutant islets have disturbed islet architecture with increased non-␤ cells distributed in the center of the islets. (b) Islet size distributions, as the percentages of all islets viewed, are shown for wild-type and heterozygote CBP-436A mutants. Mutant islets have a higher proportion of larger islets with a biphasic distribution with a small third rise at islet sizes of 50,000 to 75,000 ␮m2. (c and d) Isolated islet size. Islets were isolated and cultured overnight in RPMI 1640 medium–5.5 mmol/liter glucose, supplemented with 5% FCS and 100 U/ml penicillin-streptomycin at 37°C in a humidified chamber with 95% O2–5% CO2. The next day islet size cross-sectional area was viewed on an inverted microscope with 20⫻ objective attached to a CCD camera and MetaMorph image analysis software. Totals of 304 wild-type and 316 CBP-S436A islets were analyzed. (c) Representative images of cultured isolated islets from wild-type (left) and CBP-S436A (right) mice. (d) Islet size distribution in percentage of all islets viewed is shown for wild-type mice and heterozygote CBP-S436A mutants. Mutant islets have a higher proportion of larger islets. Mutant islets have a higher proportion of larger islets with a biphasic distribution with a third rise at islet sizes of 50,000 to 100,000 ␮m2.

7752

HUSSAIN ET AL.

MOL. CELL. BIOL.

FIG. 4. Proliferation marker Ki-67 in four-month-old wild-type (a to c) and heterozygous CBP-S436A (e to g) littermates. Ki-67 was detected on 4% paraformaldehyde-fixed, paraffin-embedded tissue sections with a monoclonal rat anti-mouse Ki-67 antibody. Detection was performed with a rabbit anti-rat antibody with an amplification step with goat anti-rabbit antiserum. The latter antibody was tagged with horseradish peroxidase. Antigen localization was performed with a DAB reaction. (h) Double immunostaining for insulin was performed with a guinea pig anti-insulin antiserum; detection and localization were performed with an alkaline phosphate-tagged goat anti-guinea pig secondary antibody and a fast red alkaline phosphate stain. Bright-field images (magnification, ⫻200) captured on a color CCD camera with appropriate software are shown. Ki-67-positive nuclei are indicated by arrows. Double immunostaining with insulin shows that Ki-67 localizes in the nuclei of insulin-containing ␤ cells in the islets of CBP-S436A mice. (d) Islets in heterozygous CBP-S436A mutant mice have an average of 1.71 ⫾ 0.19 Ki-67-positive (pos.) nuclei per islet, compared to 0.71 ⫾ 0.19 per wild-type islet.

In vitro proliferation assay. Figure 7a to e show results of five separate in vitro proliferation assays (XTT assay) performed on cultured, isolated islets. The proliferation rate is normalized to the proliferation rate of wild-type islets that were studied in parallel. The proliferation measurements were conducted after 4 days of ex vivo culture of isolated islets, in order to account for possible in vivo metabolic effects on proliferation (33). The proliferation rate of mutant islets (1.47 ⫾ 0.39) was nearly twofold higher than that found for wild-type

islets (0.81 ⫾ 0.21; P ⬍ 0.05) (Fig. 7f). Since islets of similar sizes and numbers from wild-type and mutant mice were compared, this result suggests that mutant islets contain an increased number of proliferating cells. NADH is required for the formation of formazan in the in vitro XTT proliferation assay (53), and total cellular NADH content has been shown to be a reliable marker for cellular proliferation. PGC-1␣ elevation in ␤ cells results in reduced glycolysis (41) and, consequently, reduced ATP synthesis,

VOL. 26, 2006

CREB-CBP INTERACTION AND ␤-CELL PROLIFERATION

7753

FIG. 5. BrdU uptake in four-month-old wild-type (a and b) and heterozygous CBP-S436A (d and e) mice. BrdU was injected intraperitoneally four hours before the extraction of tissue. The tissue sections used were 4% paraformaldehyde-fixed, paraffin-embedded tissue sections with an anti-mouse BrdU antibody. Detection was performed with a secondary antibody linked to horseradish peroxidase. Antigen localization was performed with a DAB reaction. Bright-field images (magnification, ⫻200) captured on a color CCD camera with appropriate software are shown. BrdU-positive nuclei are indicated by arrows. (c) Islets in heterozygous CBP-S436A mice have an average of 0.61 ⫾ 0.16 BrdU-positive (pos.) nuclei per islet, compared to 0.28 ⫾ 0.11 BrdU-positive nuclei per wild-type islet (mean ⫾ SEM, P ⬍ 0.05).

which in turn results in a reduced NADH/NAD ratio (41) (see also Discussion). Despite elevated PGC-1␣ levels in mutant pancreatic ␤ cells (see below) and, therefore, presumably a reduction in the NADH/NAD ratio, we find increased formazan formation in the XTT assay of cultured mutant islets. Therefore, we conclude that the ex vivo proliferation assay may actually be demonstrating less proliferation in mutant islets than would actually be present if NADH levels were unaffected. Glucose-stimulated insulin secretion. Using ex vivo perifusion of isolated islets, we extended the previous findings with mutant mice by use of static incubation of islets (43). Figure 8 are results obtained from three different perifusion assays of isolated islets. Results are normalized to the number of islets (panels a to c) or to total DNA content (panels d to f), as an indirect marker of total cell number, to adjust for any differences in cell content of perifused islets. Baseline and glucosestimulated insulin secretion was blunted in mutant islets. In contrast, wild-type and mutant islets released similar amounts of insulin after use of a 30 mM KCl stimulus, indicating that

the cellular machinery for insulin production, storage, and secretion is likely intact in mutant islets. Areas under the curves of insulin secretion from perifused islets from three independently performed studies are summarized in Table 1. The results suggest that both first (immediate response after glucose elevation) and second phases of glucose-stimulated insulin secretion are impaired in mutant islets. Potassium chloride-stimulated insulin secretions were similar in wild-type and mutant mice. This finding suggests that insulin production and storage and the secretion apparatus are not significantly impaired in the mutant islets. Specificity of CREB-phospho-S436-CBP interaction. Using GAL4-BD fusion proteins expressing GAL4-CREB and GAL4-HIF1␣, we tested the specificity of the phosphorylation state of CBP at S436 for its interaction with CREB. S436 of CBP resides within the CH1 domain, which is immediately adjacent to the KIX domain of CBP. We therefore tested whether phosphorylation of CBP at S436 could also alter the activity of transcription factors that interact with the CH1 domain of CBP. Alterations in the transcription factors related

7754

HUSSAIN ET AL.

MOL. CELL. BIOL.

FIG. 6. Cleaved caspase-3 immunohistochemistry of pancreas tissue from a streptozotocin-treated mouse (a) and from four-month-old mutant CBP-S436A (b) and wild-type (c) littermates. Sections were incubated after antigen retrieval with rabbit anti-cleaved caspase-3 antibody (1:200) at 4°C overnight. Detection was performed with a secondary antibody linked to horseradish peroxidase. Antigen localization was performed with a DAB reaction. Bright-field images (magnification, ⫻200) captured on a color CCD camera with appropriate software are shown. (a) Islet cells of a mouse treated with streptozotocin 48 h prior to pancreas harvest show activated apoptosis as indicated by cleaved caspase-3 immunodetection. Islets from four-month-old mutant CBP-S436A (b) and wild-type (c) littermates show no evidence for ongoing apoptosis.

to HIF1␣ have recently been reported to be important in the development of diabetes (14). Using the GAL4 eukaryotic protein-protein interaction assay (Pathdetect; Biocrest), we tested the effects of cotransfected wild-type CBP and mutant CBP-S436A on the transcriptional activities of fusion proteins expressing the GAL4-BD fused to CREB or HIF1␣ in the presence of the catalytic subunit of PKA. Wild-type CBP significantly increased the transcriptional activity of GAL4-BD– CREB. This effect was, as expected (43), dramatically enhanced by the cotransfection of mutant CBP-S436A. In

contrast, wild-type CBP only slightly increased GAL4-BD– HIF1␣ transcriptional activity (8), and mutant CBP-S436A had no additional effect. The results of these studies are summarized in Fig. 9 and Table 2. Thus, the S436A mutation of CBP does not have any effect on the interaction of HIF1␣ with CBP. Considering also our previous results for the lack of effects of mutant CBP-S436A on the transcriptional activity of the transcription factor FOXO1 (43), we conclude that the S436A mutation is specific in enhancing CREB-CBP interaction and transcriptional activation. IRS-2 and PGC-1␣ expression levels. By use of real-time quantitative RT-PCR, IRS-2 and PGC-1␣ mRNA levels were measured, and they are shown in Fig. 10. IRS-2 (Fig. 10a and b) and PGC-1␣ (Fig. 10c and d) mRNA levels were significantly increased by forskolin in wild-type islets. Basal IRS-2 and PGC-1␣ mRNA levels in CBP-S436A mutant islets, however, were both elevated compared to wild-type islets and were further stimulated by forskolin treatment.

DISCUSSION

FIG. 7. In vitro proliferation (XTT) assay of cultured, isolated islets. Isolated wild-type and mutant islets of approximately the same size were cultured for four days before the assay. Results of five separate studies performed on different days are shown separately in panels a to e. The numbers of islets cultured are indicated in the graphs. Panel f shows the means ⫾ SEM of results from panels a to e.

Thc cAMP signaling pathway is central to ␤-cell function, regeneration, and proliferation and prevention of apoptosis (7, 22, 28). Previous in vivo mouse studies on the cAMP signaling in pancreatic ␤ cells have used distinct genetic disruptions of this pathway. These include studies using GLP-1 (36) and/or GIP (31) receptor knockout mouse models and mouse models which overexpress a dominant negative regulator of CREB (21) or which overexpress a transcriptional CREB repressor, inducible cAMP early repressor (20). Activation of the cAMP pathway in ␤ cells, however, has been studied using pharmacologic treatment with GLP-1 and analogues in rodent models (38). In contrast to these previous models, the present work describes a novel in vivo model of a point mutation in the cbp gene resulting in activation of cAMP-responsive transcription in pancreatic ␤ cells. CBP, but not p300, is phosphorylated by the insulin signaling pathway at a unique serine residue (serine 436) (42). In the present study, we demonstrated that CBP is phosphorylated at serine 436 in the nuclei of normal mouse pancreatic islet cells (Fig. 3) and that the mutation of this

CREB-CBP INTERACTION AND ␤-CELL PROLIFERATION

VOL. 26, 2006

7755

FIG. 8. Islet perifusion. Isolated islets from wild-type and mutant animals were perifused (1 ml/min) with Krebs Ringer buffer at different concentrations of glucose as indicated. After glucose perifusion, islets were stimulated with 30 mM KCl. Insulin concentration was measured in the eluate at one-minute intervals. Results of studies performed at three independent occasions are shown (y axes) and are normalized to the number of islets (a to c) and total DNA of the perifused islets (d to f). See also Table 1.

TABLE 1. Areas under the curves of perifused isletsa Curve from indicated expt normalized to:

Area under the curve for indicated islets Glucose stimulated

30 mM KCl stimulated

wt

CBP-S436A

wt

CBP-S436A

No. of isletsb 1 2 3 Mean ⫾ SEM

159.4 190.1 175.7 175.1 ⫾ 8.9

41.3 55.0 48.3 48.2 ⫾ 3.6

52.1 74.1 51.3 59.2 ⫾ 7.5

68.7 66.7 76.4 70.6 ⫾ 3.0

Total islet DNAc 1 2 3 Mean ⫾ SEM

33.9 36.6 46.2 38.9 ⫾ 3.7

5.5 9.0 6.8 7.1 ⫾ 1.0

11.1 14.2 13.5 12.9 ⫾ 0.9

9.2 10.9 10.8 10.3 ⫾ 0.6

a Areas are given for curves depicted in Fig. 8; they are shown separately for the three separate studies for the glucose-stimulated and KCl-stimulated insulin secretion. wt, wild type. b Normalized to the number of islets 关(ng/ml) ⫻ 100 islets ⫻ min兴. c Normalized to total islet DNA 关(ng/ml) ⫻ total islet DNA ⫻ min兴.

serine 436 to alanine results in enhanced CREB-responsive transcriptional activity and changes in pancreatic islet size and ␤-cell proliferation and function. In heterozygous CBP-S436A mutant mice, compared to wild-type littermates, the average islet size, an indicator of islet mass (2), is increased by approximately twofold in the face of unchanged whole-pancreas weight (Fig. 1 and 2a and b). Similarly increased islet size is also found by measuring islets that were isolated from the pancreata of mutant mice and wild-type control littermates (Fig. 2c and d). The size distribution analysis of islets indicates that in mutant mice, the distribution of islets is shifted toward populations of larger islets (Fig. 2b and d). This difference in size distribution, taken together with the finding that the numbers of islets detected per pancreas section of mutant and wild-type animals do not differ, indicates that the increased islet and ␤-cell masses are not due to formation of new islets but to increased ␤-cell hyperplasia within existing islets. In addition, mutant animals also had altered islet architecture, with an increased number of non-␤ cells in the center of the islet (Fig. 2a). Thirty percent of the ⬎50,000-␮m2 islets

7756

HUSSAIN ET AL.

MOL. CELL. BIOL.

FIG. 9. Transfection studies. 293T cells cultured with standard methods were transfected in 24-well plates. The plasmids used were the reporter plasmid carrying the GAL4 DNA binding sites upstream of the firefly luciferase cDNA (pFR-Luc; Biocrest) (50 ng/well), CMVGAL4-BD-CREB (Biocrest) (5 ng/well), CMV-GAL4-BD-HIF1␣ (5 ng/well), pCMV-DNA3.1-CBP (150 ng/well), and pcDNA3.1-CBPS436A (150 ng/well). CMV-GAL4-BD-CREB stimulates transcriptional activity, which is further amplified by overexpression of wild-type (wt) CBP. This effect is dramatically augmented by transfection of mutant CBP-S436A. CMV-GAL4-BD-HIF1␣ slightly increases transcriptional activity, which is not further significantly altered by wildtype or mutant CBP (means ⫾ SEM; significance is calculated with Student’s t test). NS, nonsignificant. See also Table 2. RSV, Rous sarcoma virus.

in the CBP-S436A mutant mice have such altered architecture. We have not found any such description of altered islet architecture, in particular in models of increased islet size as a result of insulin resistance. This finding indicates that the present model reveals, at least in part, an islet phenotype which is not found in models of increased islet mass as a result of peripheral insulin resistance.

TABLE 2. Relative luciferase activity in 293T cells grown and transfected in 24-well culture plates with plasmids in a eukaryotic GAL4-transcription factor coactivator assay a Plasmid description

Relative luciferase activity

Plasmids from the GAL4-BD–CREB study pFR-Luc RSV-PKA............................................................................... 1 pFR-Luc RSV-PKA GAL4-BD–CREB.............................................. 3.3 ⫾ 0.5 pFR-Luc RSV-PKA GAL4-BD–CREB CMV-CBP ......................... 8.9 ⫾ 2.1 pFR-Luc RSV-PKA GAL4-BD–CREB CMV-CBP-S436A.............35.2 ⫾ 11.6 Plasmids from the GAL4-BD–HIF1␣ study pFR-Luc RSV-PKA............................................................................... 1 pFR-Luc RSV-PKA GAL4-BD–HIF1a.............................................. 1.2 ⫾ 0.1 pFR-Luc RSV-PKA GAL4-BD–HIF1a CMV-CBP.......................... 1.7 ⫾ 0.2 pFR-Luc RSV-PKA GAL4-BD–HIF1a CMV-CBP-S436A ............. 1.5 ⫾ 0.2 a Luciferase activity was normalized to levels detected in cells not transfected with either GAL4-BD construct or transcriptional coactivator (wild-type CBP or CBP-S436A). Plasmids used were the reporter plasmid carrying four concatemeric GAL4 DNA sequences upstream of a minimal promoter and the firefly luciferase cDNA (pFR-Luc) (50 ng/well), CMV-GAL4-BD-CREB (5 ng/well), CMV-GAL4-BD-HIF1␣ (5 ng/well), pCMV-DNA3.1-CBP (150 ng/well), and pcDNA3.1-CBP-S436A. Wild-type CBP significantly increased the transcriptional activity of GAL4-BD–CREB. See also Fig. 9. See the text for details.

FIG. 10. Real-time RT-PCR of IRS-2 (a and b) and PGC-1␣ (c and d) transcripts in isolated islets. Islets were isolated and cultured for 48 h in RPMI 1640 medium with 5.5 mmol/liter glucose, supplemented with 5% FCS and 100 U/ml penicillin-streptomycin at 37°C in a humidified chamber with 95% O2–5% CO2. Forskolin, where indicated, was added to the medium at a concentration of 10 ␮M for the entire incubation period. Islet RNA was harvested and cDNA was made with 0.7 mg of total RNA using iQ SYBR green supermix (Bio-Rad). Real-time PCR was performed using 18S rRNA as a control by use of the MyiQ single-color real-time PCR detection system (Bio-Rad). A linear plot of the numbers of IRS-2 and PGC-1␣ transcripts relative to wild-type littermate islets in wild-type islets stimulated with forskolin, heterozygous CBP-S436A islets, and heterozygous CBP-S436A islets stimulated with forskolin is shown. IRS-2 and PGC-1␣ transcription rates are elevated by forskolin in wild-type islets. Baseline IRS-2 and PGC-1␣ transcription rates are significantly elevated in mutant CBPS436A islets and further elevated under forskolin stimulation. Results are shown as scatter plots of individual results (a and c) and means ⫾ SEM (b and d).

Next, we verified that CBP is phosphorylated in pancreatic islet cells. An immunopurified antibody raised against phospho-S436-CBP was used to detect phospho-CBP on frozen pancreatic sections of wild-type mouse samples (Fig. 3a). This staining was neutralized by coincubating the antiserum with the synthesized antigen to mimic phospho-S436 CBP (Fig. 3b). This finding indicates that CBP is indeed phosphorylated in pancreatic islet cells in normal mice. In CBP-S436A mice, larger islet size is accompanied by increased ␤-cell proliferation, as assessed by in vivo BrdU uptake (Fig. 5), more abundant nuclear Ki-67 staining in mutant cells than in wild-type ␤ cells (Fig. 4), and increased ex vivo proliferation rates of isolated and cultured mutant islets as measured by the XTT assay (Fig. 7). In addition, apoptosis was not detected in either wild-type or CBP-S436A pancreatic islet cells (Fig. 6). Overall, these results indicate that in CBP-S436A heterozygous mice, the increased ␤-cell mass results from increased ␤-cell proliferation with no indication of increased ␤-cell turnover (i.e., lack of increased apoptosis). As a comparison to the present model, the heterozygote, combined insulin receptor and IRS-1 (IRhet/IRS-1het) knockout mice exhibit an approximately fourfold increase in islet size at 4 months of age (24). A second model of substantially enlarged islet size due to hepatic insulin resistance, the liverspecific insulin receptor knockout mouse (30), has a 10-fold

VOL. 26, 2006

CREB-CBP INTERACTION AND ␤-CELL PROLIFERATION

increase in islet mass. While these models show increased islet and ␤-cell masses, they do not distinguish between increases in ␤-cell number (hyperplasia) and ␤-cell size (hypertrophy) (24, 30). ␤-Cell mass can be increased solely due to increased ␤-cell size rather than number, as has recently been demonstrated in a ␤-cell-specific Akt/PKB␣ transgenic mouse (40). In CBPS436A mice, ␤-cell cross-sectional areas, indicators of ␤-cell sizes (40), are not different in mutant and wild-type islets and are within the normal range for those previously reported in the literature (40). This suggests that increased ␤-cell mass in our model is primarily due to ␤-cell hyperplasia. In vivo-stimulated ␤-cell proliferation is most likely multifactorial in the present model, which also exhibits increased hepatic insulin resistance (43). Kulkarni et al. (24) have reported averages of 0.5 proliferating cell nuclear antigen (PCNA)-positive cells/islet in wild-type animals, 0.3 cells/islet in IRhet knockout mice, and 0.3 cells/islet in IRS-1het knockout mice. IRhet/IRS-1het knockout mice have 28 PCNA-positive cells/islet. ␤-Cell mass is increased 2- to 3-fold in IRS-1het knockout mice, 2- to 3-fold in IRhet knockout mice, and 2- to 30-fold in compound IRhet/IRS-1het knockout mice (24). In comparison, CBP-S436A mice have moderate insulin resistance (43) and increased ␤-cell mass, traits which are similar to those found in IRhet or IRS-1het knockout mice. However, compared to IRhet and IRS-1het knockout mice, CBP mutants exhibit a larger proportion of proliferating (i.e., Ki-67-positive) cells per islet. In addition, isolated and cultured (in 5.5 mM glucose) mutant islets continue to proliferate at a higher (by approximately 1.5 times) rate than wild-type controls even after 4 days of culture in vitro, which is considered to be sufficient time for islets to no longer be influenced by in vivo extrapancreatic metabolic changes found in the organism of their origin (33). Continued elevated ex vivo ␤-cell proliferation rates (XTT assay) (Fig. 7) indicate that in CBP-S436A mutants, ␤-cell proliferation is ␤-cell autonomous and only partially due to the moderate insulin resistance in vivo. Increases in ␤-cell mass can be stimulated by IRS-2 overexpression (16), and ␤-cell mitosis in insulin-resistant states is, at least in part, mediated by IRS-2 (16). Transgenic overexpression of IRS-2 in mice results in increased ␤-cell mass and an increase in the absolute number of islets per histological section, while little change is found in islet size (16). In contrast, in the present study, we do find increased islet sizes but unchanged islet numbers per pancreas. The reason for this discrepancy remains unclear but nevertheless indicates that while cAMP-CREB-CBP signaling increases IRS-2 expression (see below), the effects of cAMP-CREB-CBP are distinct from direct IRS-2 overexpression. Overexpression of an artificial dominant negative CREB analogue (A-CREB) in pancreatic ␤ cells reduces IRS-2 production and increases ␤-cell susceptibility to apoptosis (21). In the present model, with enhanced CREB-CBP-responsive transcription, the IRS-2 transcript is increased. The increases in islet mass and ␤-cell number in the present model are therefore likely due not only to increased proliferation but possibly also to reduced ␤-cell apoptosis (Fig. 6) secondary to elevated IRS-2 levels. Taken together with the results of the proliferation studies in the present model, the lack of apoptosis marker cleaved caspase 3 (Fig. 6) indicates that the increased

7757

islet mass is not found in context of elevated cell turnover (proliferation combined with apoptosis) but as a consequence of ␤-cell proliferation. Perfused CBP-S436A mutant islets show reduced glucosestimulated insulin secretion, while KCl-stimulated release of stored insulin is comparable to wild-type islets (Fig. 8; Table 1). This indicates that insulin production in mutant ␤ cells is intact, which is consistent with the CREB-responsiveness of insulin gene transcription (7, 22, 28), whereas the glucose stimulus-coupled secretion is defective. The latter finding matches those previously reported by Yoon et al. (41), results showing overexpression of PGC-1␣ in pancreatic islets, which also showed reduced glucose-stimulated insulin secretion. Elevated PGC-1␣ in ␤ cells has been shown to reduce ␤-cell glucose metabolism and subsequent ATP production, which in turn results in a reduced NADH/NAD ratio (41). These changes, associated with elevated PGC-1␣, as are found in the present model (Fig. 10; see below), result in reduced glucose-stimulated insulin secretion (41). The serine 436 is immediately upstream and adjacent to the CBP KIX domain, which is considered to interact with CREB. Because this serine residue does not lie within the KIX domain, we tested the specificity of the mutation of CBP-S436A on the interaction with CREB and HIF1␣. HIF1␣ was chosen because of its known interaction with the CH1 domains of CBP/p300 (8) and the recent demonstration of altered HIF/ ARNT expression levels in islets of diabetic mice (14). By use of a eukaryotic protein-protein interaction assay with GAL4-BD fusion proteins (see Materials and Methods), we found that the CBP-S436A mutant increases transcription activity of a GAL4-BD-CREB more than wild-type CBP, whereas the activities of GAL4-BD-HIF1␣ are similar in the presence of CBP and CBP-S436A. Together with our previous finding that mutant CBP did not alter transcriptional activity of FOXO1 in the same assay (43), our results indicate that CBP phosphorylation at S436 specifically interferes with CREBCBP and not HIF1␣-CBP or FOXO1-CBP interaction. Defective insulin secretion in models of diabetes and increasing islet mass, including the partial pancreatectomy model, have recently been attributed to increased PGC-1␣ levels in ␤ cells (41). PGC-1␣, also a CREB-responsive protein (34), is elevated in ␤ cells of diabetic animals with defective insulin secretion (41). Furthermore, overexpression of PGC-1␣ suppresses glucose-stimulated insulin release from islets in culture by an average of 60% compared to normal islets (41). Transplantation of PGC-1␣-overexpressing islets into streptozotocin-treated diabetic mice does not improve glucose metabolism as control islets do (41). Similar to the case in the present model, PGC-1␣ overexpression has been shown to not reduce total islet insulin content (41). PGC-1␣-overexpressing islets exhibit reduced glucose-induced elevations in intracellular ATP levels, which negatively impacts glucose-induced membrane depolarization, and subsequent insulin release by ␤ cells (41). Taken together, these data indicate that pgc1a is a CREB-responsive gene that is upregulated in proliferating CBP-S436A islets. Elevated PGC-1␣ and its downstream consequences are likely the basis for the functional defect found in the present model. Although the link between cAMP and elevated PGC-1␣ levels in ␤ cells has not been directly demonstrated, the present

7758

HUSSAIN ET AL.

data suggest that upregulated nuclear cAMP signaling increases PGC-1␣ transcription in pancreatic ␤ cells. pgc1a is a CREB-responsive gene with CREB recognition elements in its 5⬘ regulatory region (34). In accordance with these observations, we found that forskolin increases PGC-1␣ transcripts in wild-type islets (Fig. 10c and d). Mutant islets have elevated baseline PGC-1␣ transcript levels that are further elevated by forskolin. Stimulation by forskolin of IRS-2 and PGC-1␣ expression above the baseline in CBP-S436A mice is most likely due to stimulated and enhanced phosphorylation of CREB at serine 133 (12), which increases CREB-CBP interaction (32). Proliferation of ␤ cells is not implicitly linked to a reduced insulin secretion rate. Mice with transgene overexpression of cyclin-dependent kinase 4 (29) or disruption of peroxisome proliferator-activated receptor-␥ specifically in ␤ cells exhibit enlarged islets associated with increased proliferation of ␤ cells (35). These mice do not show any reduced glucose tolerance or deficient insulin secretion in physiologic assays. Furthermore, severely insulin-resistant models, such as the liver-specific insulin receptor knockout mouse, show increased islet size and ␤-cell proliferation with apparent metabolically adequate insulin secretion (30). In the present mouse model, a specific point mutation in the cbp gene results in a ␤-cell phenotype of proliferation in the face of reduced glucose-stimulated insulin secretion. Together with a recent report that that ␤-cell-specific inducible cAMP early repressor overexpression (20) results in reduced postnatal ␤-cell numbers, the present model also indicates a central role of CREB-CBP in ␤-cell proliferation and regeneration. In addition to the report of trophic effects of insulin on pancreatic ␤ cells (10), the present work suggests that insulin action on ␤ cells is, by regulating CREB-CBP transcriptional activity, intimately involved in the genetic switch between ␤-cell proliferation and maintenance of adequate glucose-responsive insulin secretion. A lack of insulin would result in decreased phosphorylation at CBP S436 and thus increased proliferative activity. Conversely, in normal mice, in the presence of metabolically sufficient insulin levels, the CREB-CBP pathway-induced genetic switch towards proliferation is turned off by direct insulin-dependent phosphorylation of CBP at S436. In this context, it is important to note that the disruption in the present model is specific for insulin effects on nuclear action of CREB-CBP, while other signaling pathways downstream of the insulin receptor remain unaffected. In particular, it should be noted that the ␤-cell-specific homozygous insulin receptor knockout mouse (23) shows diminished postnatal ␤-cell mass. The specific sites of molecular/genetic disruption likely explain the discrepancies between the present model of the lack of insulin action on ␤-cell CBP and the ␤-cell-specific knockout of the insulin receptor or insulin receptor substrate proteins, which show reduced islet size as well as reduced insulin secretion (23, 35). The transient elevation of CREB-CBP-responsive transcription may be beneficial for ␤-cell function, proliferation, and survival, but sustained CREB-CBP-responsive transcription results in impaired ␤-cell function, and it can be hypothesized that turning off the CREB-CBP transcription, for example, by insulin-stimulated CBP phosphorylation at S436, is equally important for maintaining normal ␤-cell function.

MOL. CELL. BIOL. ACKNOWLEDGMENTS We thank Fei Fei (Cyndi) Liu for technical help, Michael Collector for streptozotocin-treated mouse pancreas tissue, and S. Radovick and A. Sidhaye for critical review of the manuscript. This work was supported in part by National Institutes of Health grant RO1 DK 64646 and Juvenile Diabetes Research Foundation grants 1-2003-162 and 5-2006-80 to M.A.H. and National Institutes of Health grant RO1 DK 63349 to F.E.W. REFERENCES 1. Abraham, E. J., C. A. Leech, J. C. Lin, H. Zulewski, and J. F. Habener. 2002. Insulinotropic hormone glucagon-like peptide-1 differentiation of human pancreatic islet-derived progenitor cells into insulin-producing cells. Endocrinology 143:3152–3161. 2. Bonner-Weir, S. 2001. Beta-cell turnover: its assessment and implications. Diabetes 50(Suppl. 1):S20–S24. 3. Bonner-Weir, S. 2000. Islet growth and development in the adult. J. Mol. Endocrinol. 24:297–302. 4. Bromley, M., D. Rew, A. Becciolini, M. Balzi, C. Chadwick, D. Hewitt, Y. Q. Li, and C. S. Potten. 1996. A comparison of proliferation markers (BrdUrd, Ki-67, PCNA) determined at each cell position in the crypts of normal human colonic mucosa. Eur. J. Histochem. 40:89–100. 5. Brubaker, P. L., and D. J. Drucker. 2004. Glucagon-like peptides regulate cell proliferation and apoptosis in the pancreas, gut, and central nervous system. Endocrinology 145:2653–2659. 6. Dor, Y., J. Brown, O. I. Martinez, and D. A. Melton. 2004. Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature 429:41–46. 7. Drucker, D. J. 2003. Glucagon-like peptide-1 and the islet beta-cell: augmentation of cell proliferation and inhibition of apoptosis. Endocrinology 144:5145–5148. 8. Ema, M., K. Hirota, J. Mimura, H. Abe, J. Yodoi, K. Sogawa, L. Poellinger, and Y. Fujii-Kuriyama. 1999. Molecular mechanisms of transcription activation by HLF and HIF1alpha in response to hypoxia: their stabilization and redox signal-induced interaction with CBP/p300. EMBO J. 18:1905–1914. 9. Fernandes, A., L. C. King, Y. Guz, R. Stein, C. V. Wright, and G. Teitelman. 1997. Differentiation of new insulin-producing cells is induced by injury in adult pancreatic islets. Endocrinology 138:1750–1762. 10. Flier, S. N., R. N. Kulkarni, and C. R. Kahn. 2001. Evidence for a circulating islet cell growth factor in insulin-resistant states. Proc. Natl. Acad. Sci. USA 98:7475–7480. 11. Georgia, S., and A. Bhushan. 2004. Beta cell replication is the primary mechanism for maintaining postnatal beta cell mass. J. Clin. Investig. 114: 963–968. 12. Gonzalez, G. A., and M. R. Montminy. 1989. Cyclic AMP stimulates somatostatin gene transcription by phosphorylation of CREB at serine 133. Cell 59:675–680. 13. Goren, H. J., R. N. Kulkarni, and C. R. Kahn. 2004. Glucose homeostasis and tissue transcript content of insulin signaling intermediates in four inbred strains of mice: C57BL/6, C57BLKS/6, DBA/2, and 129X1. Endocrinology 145:3307–3323. 14. Gunton, J. E., R. N. Kulkarni, S. Yim, T. Okada, W. J. Hawthorne, Y. H. Tseng, R. S. Roberson, C. Ricordi, P. J. O’Connell, F. J. Gonzalez, and C. R. Kahn. 2005. Loss of ARNT/HIF1beta mediates altered gene expression and pancreatic-islet dysfunction in human type 2 diabetes. Cell 122:337–349. 15. Guz, Y., I. Nasir, and G. Teitelman. 2001. Regeneration of pancreatic beta cells from intra-islet precursor cells in an experimental model of diabetes. Endocrinology 142:4956–4968. 16. Hennige, A. M., D. J. Burks, U. Ozcan, R. N. Kulkarni, J. Ye, S. Park, M. Schubert, T. L. Fisher, M. A. Dow, R. Leshan, M. Zakaria, M. Mossa-Basha, and M. F. White. 2003. Upregulation of insulin receptor substrate-2 in pancreatic beta cells prevents diabetes. J. Clin. Investig. 112:1521–1532. 17. Holz, G. G. 2004. Epac: a new cAMP-binding protein in support of glucagonlike peptide-1 receptor-mediated signal transduction in the pancreatic betacell. Diabetes 53:5–13. 18. Holz, G. G., IV, W. M. Kuhtreiber, and J. F. Habener. 1993. Pancreatic beta-cells are rendered glucose-competent by the insulinotropic hormone glucagon-like peptide-1(7-37). Nature 361:362–365. 19. Hussain, M. A., P. B. Daniel, and J. F. Habener. 2000. Glucagon stimulates expression of the inducible cAMP early repressor and suppresses insulin gene expression in pancreatic beta-cells. Diabetes 49:1681–1690. 19a.Hussain, M. A., C. P. Miller, and J. F. Habener. 2002. Brn-4 transcription factor expression targeted to the early developing mouse pancreas induces ectopic glucagon gene expression in insulin-producing beta cells. J. Biol. Chem. 277:16028–16032. 19b.Ianus, A., G. G. Holz, N. D. Theise, and M. A. Hussain. 2003. In vivo derivation of glucose-competent pancreatic endocrine cells from bone marrow without evidence of cell fusion. J. Clin. Investig. 111:843–850. 20. Inada, A., Y. Hamamoto, Y. Tsuura, J. Miyazaki, S. Toyokuni, Y. Ihara, K. Nagai, Y. Yamada, S. Bonner-Weir, and Y. Seino. 2004. Overexpression of

VOL. 26, 2006

21.

22. 23.

24.

25.

26. 27. 28. 29.

30.

31.

32.

CREB-CBP INTERACTION AND ␤-CELL PROLIFERATION

inducible cyclic AMP early repressor inhibits transactivation of genes and cell proliferation in pancreatic beta cells. Mol. Cell. Biol. 24:2831–2841. Jhala, U. S., G. Canettieri, R. A. Screaton, R. N. Kulkarni, S. Krajewski, J. Reed, J. Walker, X. Lin, M. White, and M. Montminy. 2003. cAMP promotes pancreatic beta-cell survival via CREB-mediated induction of IRS2. Genes Dev. 17:1575–1580. Kieffer, T. J., and J. F. Habener. 1999. The glucagon-like peptides. Endocr. Rev. 20:876–913. Kulkarni, R. N., J. C. Bruning, J. N. Winnay, C. Postic, M. A. Magnuson, and C. R. Kahn. 1999. Tissue-specific knockout of the insulin receptor in pancreatic beta cells creates an insulin secretory defect similar to that in type 2 diabetes. Cell 96:329–339. Kulkarni, R. N., U. S. Jhala, J. N. Winnay, S. Krajewski, M. Montminy, and C. R. Kahn. 2004. PDX-1 haploinsufficiency limits the compensatory islet hyperplasia that occurs in response to insulin resistance. J. Clin. Investig. 114:828–836. Kwok, R. P., J. R. Lundblad, J. C. Chrivia, J. P. Richards, H. P. Bachinger, R. G. Brennan, S. G. Roberts, M. R. Green, and R. H. Goodman. 1994. Nuclear protein CBP is a coactivator for the transcription factor CREB. Nature 370:223–226. Li, Y., T. Hansotia, B. Yusta, F. Ris, P. A. Halban, and D. J. Drucker. 2003. Glucagon-like peptide-1 receptor signaling modulates beta cell apoptosis. J. Biol. Chem. 278:471–478. Lipsett, M., and D. T. Finegood. 2002. Beta-cell neogenesis during prolonged hyperglycemia in rats. Diabetes 51:1834–1841. List, J. F., and J. F. Habener. 2004. Glucagon-like peptide 1 agonists and the development and growth of pancreatic beta-cells. Am. J. Physiol. Endocrinol. Metab. 286:E875–E881. Marzo, N., C. Mora, M. E. Fabregat, J. Martin, E. F. Usac, C. Franco, M. Barbacid, and R. Gomis. 2004. Pancreatic islets from cyclin-dependent kinase 4/R24C (Cdk4) knockin mice have significantly increased beta cell mass and are physiologically functional, indicating that Cdk4 is a potential target for pancreatic beta cell mass regeneration in Type 1 diabetes. Diabetologia 47:686–694. (In German.) Michael, M. D., R. N. Kulkarni, C. Postic, S. F. Previs, G. I. Shulman, M. A. Magnuson, and C. R. Kahn. 2000. Loss of insulin signaling in hepatocytes leads to severe insulin resistance and progressive hepatic dysfunction. Mol. Cell 6:87–97. Miyawaki, K., Y. Yamada, H. Yano, H. Niwa, N. Ban, Y. Ihara, A. Kubota, S. Fujimoto, M. Kajikawa, A. Kuroe, K. Tsuda, H. Hashimoto, T. Yamashita, T. Jomori, F. Tashiro, J. Miyazaki, and Y. Seino. 1999. Glucose intolerance caused by a defect in the entero-insular axis: a study in gastric inhibitory polypeptide receptor knockout mice. Proc. Natl. Acad. Sci. USA 96:14843– 14847. Montminy, M. 1997. Transcriptional regulation by cyclic AMP. Annu. Rev. Biochem. 66:807–822.

7759

33. Piston, D., and G. Patterson. 2005. Linear relationship between glucokinase expression, metabolic redox state and insulin secretion, vol. 16. Karger, Basel, Switzerland. 34. Puigserver, P., and B. M. Spiegelman. 2003. Peroxisome proliferator-activated receptor-gamma coactivator 1 alpha (PGC-1 alpha): transcriptional coactivator and metabolic regulator. Endocr. Rev. 24:78–90. 35. Rosen, E. D., R. N. Kulkarni, P. Sarraf, U. Ozcan, T. Okada, C. H. Hsu, D. Eisenman, M. A. Magnuson, F. J. Gonzalez, C. R. Kahn, and B. M. Spiegelman. 2003. Targeted elimination of peroxisome proliferator-activated receptor gamma in beta cells leads to abnormalities in islet mass without compromising glucose homeostasis. Mol. Cell. Biol. 23:7222–7229. 36. Scrocchi, L. A., T. J. Brown, N. MacLusky, P. L. Brubaker, A. B. Auerbach, A. L. Joyner, and D. J. Drucker. 1996. Glucose intolerance but normal satiety in mice with a null mutation in the glucagon-like peptide 1 receptor gene. Nat. Med. 2:1254–1258. 37. Scudiero, D. A., R. H. Shoemaker, K. D. Paull, A. Monks, S. Tierney, T. H. Nofziger, M. J. Currens, D. Seniff, and M. R. Boyd. 1988. Evaluation of a soluble tetrazolium/formazan assay for cell growth and drug sensitivity in culture using human and other tumor cell lines. Cancer Res. 48:4827–4833. 38. Stoffers, D. A., T. J. Kieffer, M. A. Hussain, D. J. Drucker, S. Bonner-Weir, J. F. Habener, and J. M. Egan. 2000. Insulinotropic glucagon-like peptide 1 agonists stimulate expression of homeodomain protein IDX-1 and increase islet size in mouse pancreas. Diabetes 49:741–748. 39. Teitelman, G. 1996. Induction of beta-cell neogenesis by islet injury. Diabetes Metab. Rev. 12:91–102. 40. Tuttle, R. L., N. S. Gill, W. Pugh, J. P. Lee, B. Koeberlein, E. E. Furth, K. S. Polonsky, A. Naji, and M. J. Birnbaum. 2001. Regulation of pancreatic beta-cell growth and survival by the serine/threonine protein kinase Akt1/ PKBalpha. Nat. Med. 7:1133–1137. 41. Yoon, J. C., P. Puigserver, G. Chen, J. Donovan, Z. Wu, J. Rhee, G. Adelmant, J. Stafford, C. R. Kahn, D. K. Granner, C. B. Newgard, and B. M. Spiegelman. 2001. Control of hepatic gluconeogenesis through the transcriptional coactivator PGC-1. Nature 413:131–138. 42. Zanger, K., L. E. Cohen, K. Hashimoto, S. Radovick, and F. E. Wondisford. 1999. A novel mechanism for cyclic adenosine 3⬘,5⬘-monophosphate regulation of gene expression by CREB-binding protein. Mol. Endocrinol. 13:268– 275. 43. Zhou, X. Y., N. Shibusawa, K. Naik, D. Porras, K. Temple, H. Ou, K. Kaihara, M. W. Roe, M. J. Brady, and F. E. Wondisford. 2004. Insulin regulation of hepatic gluconeogenesis through phosphorylation of CREBbinding protein. Nat. Med. 10:633–637. 44. Zulewski, H., E. J. Abraham, M. J. Gerlach, P. B. Daniel, W. Moritz, B. Muller, M. Vallejo, M. K. Thomas, and J. F. Habener. 2001. Multipotential nestin-positive stem cells isolated from adult pancreatic islets differentiate ex vivo into pancreatic endocrine, exocrine, and hepatic phenotypes. Diabetes 50:521–533.