Inhibition of Small Maf Function in Pancreatic β-Cells Improves Glucose Tolerance Through the Enhancement of Insulin Gene Transcription and Insulin Secretion

Abstract

The large-Maf transcription factor v-maf musculoaponeurotic fibrosarcoma oncogene homolog A (MafA) has been found to be crucial for insulin transcription and synthesis and for pancreatic β-cell function and maturation. However, insights about the effects of small Maf factors on β-cells are limited. Our goal was to elucidate the function of small-Maf factors on β-cells using an animal model of endogenous small-Maf dysfunction. Transgenic (Tg) mice with β-cell-specific expression of dominant-negative MafK (DN-MafK) experiments, which can suppress the function of all endogenous small-Mafs, were fed a high-fat diet, and their in vivo phenotypes were evaluated. Phenotypic analysis, glucose tolerance tests, morphologic examination of β-cells, and islet experiments were performed. DN-MafK-expressed MIN6 cells were also used for in vitro analysis. The results showed that DN-MafK expression inhibited endogenous small-Maf binding to insulin promoter while increasing MafA binding. DN-MafK Tg mice under high-fat diet conditions showed improved glucose metabolism compared with control mice via incremental insulin secretion, without causing changes in insulin sensitivity or MafA expression. Moreover, up-regulation of insulin and glucokinase gene expression was observed both in vivo and in vitro under DN-MafK expression. We concluded that endogenous small-Maf factors negatively regulates β-cell function by competing for MafA binding, and thus, the inhibition of small-Maf activity can improve β-cell function.

Although various factors affect the transcription, synthesis and secretion of insulin in pancreatic islet β-cells, some pancreatic transcriptional factors, such as pancreatic and duodenal homeobox factor 1 (Pdx-1), neurogenic differentiation factor 1 (NeuroD1), and v-maf musculoaponeurotic fibrosarcoma oncogene homolog A (MafA), have been certified to be intimately involved in insulin transcription under the conditions of glucolipotoxicity (1–4). These transcriptional factors bind to conserved enhancer elements in the promoter region of the insulin genes and regulate glucose-responsive insulin gene transcription and, consequently, insulin secretion and synthesis. Pdx-1 and MafA are selectively expressed in pancreatic β-cells, whereas NeuroD1 is expressed in all pancreatic endocrine cells. All 3 factors are involved in both insulin gene expression and islet and pancreas development and maturation (5, 6).

In particular, the transcription factor MafA has been reported to be a key regulator of insulin gene transcription and β-cell maturation (7–10). Maf transcription factors belong to the basic leucine zipper family, and the Maf family is divided into 2 groups, large-Maf factors and small-Maf factors. Large-Maf factors include MafA, c-Maf, MafB, and neural retina-specific leucine zipper protein (11, 12). Large-Mafs possess a DNA-binding domain and an N-terminal transactivating domain; therefore, they play key roles in gene regulation and transcription.

On the other hand, small-Maf transcription factors, including MafF, MafG, and MafK, are expressed in a wide variety of tissues at various levels (13–15). Although small-Maf factors lack a transactivation domain, they act as transcriptional regulators by binding to a DNA sequence known as the Maf recognition element (MARE) (16). Small-Maf factors form heterodimers with the CNC family of proteins, including NF-E2-related factor 1 (Nrf1), Nrf2, Nrf3, BTB and CNC homology 1 (Bach1), and Bach2, which further interact with Fos and FosB, but not with large-Maf factors (16–18). Homodimer of small-Maf factors suppress transcriptional activity of large-Maf factors via MARE, but small-Maf heterodimers can act as either suppressors or activators depending on their dimerization partners (16). It has been reported that MafK expression inhibited insulin transcription competing with MafA; moreover, in pancreatic islets, β-cell-specific overexpression of MafK was reported to result in the impairment of glucose-stimulated insulin secretion (GSIS) only at a young age and resulted in reciprocal islet hypertrophy and compensatory increase in the DNA-binding activity of MafA in adult age (19).

However, little is known about the function of endogenous small Maf factors in pancreatic β-cells in vivo, and the association between small-Maf factors and the diabetic state is also not well understood. To clarify the role of small-Maf factors in vivo, we aimed to repress endogenous small-Maf functions using dominant-negative MafK (DN-MafK), which lacks the part of the DNA-binding domain of endogenous MafK that reportedly decreases nuclear factor-erythroid 2 DNA-binding activity (20). In this report, we describe the generation of pancreatic β-cell-specific DN-MafK transgenic (Tg) mice and characterize their metabolic phenotype.

Research Design and Methods

Generation of Tg mice

Construction of the expression vector, including the 1.9-kb human insulin promoter used to generate Tg mice, has been described previously (21). The vector was provided by Dr Yamaoka (Institute for Genome Research, University of Tokushima, Tokushima, Japan). The DN-MafK mutant construct described elsewhere (20) was provided by Dr Orkin (Children's Hospital, Boston, MA). This DN-MafK construct was inserted into the multiple cloning sites in the cytomegalovirus expression vector with N-terminal 3 tandem Flag tags (Sigma-Aldrich). Flag-DN-MafK was subcloned into the cloning site flanking the exon-intron organization and a polyadenylation signal of the rabbit β-globin gene. The BssHII-excised fragment of this vector, excluding the plasmid-derived sequence, was used as the transgene. Integration of the transgene into the mouse genome was detected by PCR, between a sense primer in exon 1 of the human insulin promoter (5′-GCATCAGAAGAGGCCATCAA-3′) and an antisense primer in exon 3 of the rabbit β-globin gene (5′-ACTCACCCTGAAGTTCTCAG-3′), and by Southern blot analysis. The SalI-NotI fragment of the transgene was used as a probe and compared with indicator bands of 1, 10, and 100 copies of the transgene. Three lines of Tg mice (numbers 72, 23, and 53) were established on the C57BL/6J background.

Animal care and diet

All mice were housed at 2–4 animals per cage under controlled ambient conditions and a 12-hour light, 12-hour dark cycle, with lights on at 7 am. The animals were maintained in accordance with standard animal care procedures based on the institutional guidelines at Hokkaido University Graduate School of Medicine and were given free access to drinking water and diet. Both wild-type (Wt) and DN-MafK Tg male mice were fed standard chow (Oriental Yeast) until 5 weeks or 6 weeks of age and were subsequently switched to a high-fat diet (HFD) for 10 weeks, 14 or 15 weeks and an additional 10 weeks. The HFD contains 56.7% calories from fat and 20.1% calories from protein (High Fat Diet 32; Clea Tokyo).

Measurement of biochemical markers

Body weight was monitored weekly from 6 weeks of age, and a random blood glucose test was performed every 2 weeks using a One Touch Ultra blood glucose meter (Johnson & Johnson). Blood samples were also collected from the tail vein every 2 weeks. For glucose tolerance testing (GTT), the plasma was separated and stored at −80°C until use for insulin measurement. The concentration of insulin in the plasma was measured using an ELISA kit (Morinaga Institute of Biological Science).

Intraperitoneal GTT (ipGTT) and oral GTT (OGTT)

All mice underwent the OGTT at 16 weeks of age or the ipGTT at 20 weeks of age. After a 16-hour overnight fast, the mice were ip or orally loaded with glucose at a concentration of 1.0 mg/g body weight. We obtained blood samples at 0, 15, 30, 60, 90, and 120 minutes after glucose loading. Glucose and plasma insulin levels were measured as described above.

Insulin tolerance testing

After the mice were given free access to diet, human insulin (Humalin R; Eli Lilly) was injected ip at a concentration of 0.75 mU/g body weight at 16 weeks of age. Blood samples were collected from the tail vein every 30 minutes, and blood glucose was determined immediately as described above.

Immunohistochemical analysis

Isolated pancreatic tissues were immersion fixed in 4% formalin at 4°C overnight. Tissues were then roughly paraffin-embedded, and 5-μm sections were mounted on glass slides. Sections were immersed for 15 minutes in methanol containing 0.3% (vol/vol) hydrogen peroxide to deactivate endogenous peroxidase activity. After rinsing with PBS, the sections were immunostained with a specific antibody, including rabbit antihuman insulin (diluted 1:1000), anti-MafF/G/K (1:200), and anti-Flag (1:1000) antibodies (Santa Cruz Biotechnology, Inc). The sections were counterstained with hematoxylin.

For fluoroimmunostaining, tissue sections were incubated overnight at 4°C with rabbit antihuman insulin (1:1000), anti-Maf F/G/K (1:200) (Santa Cruz Biotechnology, Inc), antimouse insulin monoclonal antibody (1:1000), anti-Flag (1:1000) (Sigma-Aldrich), and antiproliferative cell nuclear antigen (PCNA) monoclonal antibody (Nichirei). After rinsing with PBS, Alexa Fluor 488 goat antimouse antibody and Alexa Fluor 594 donkey antigoat antibody (Invitrogen) were added, and the mixture was incubated for 30 minutes. To estimate β-cell mass, the area of insulin-positive cells was measured with BZ-II analyzer (Keyence, Osaka, Japan) according to the manufacturer's instructions, and β-cell mass was calculated by the next formula: β-cell mass (mg) = the pancreas weight (mg) × percent pancreatic islet area × percent β-cell count. PCNA-positive β-cells were counted separately from insulin-positive islet cells.

Islet isolation

Islets were isolated using collagenase XI (Sigma-Aldrich) according to the manufacturer's instructions, as described elsewhere (22, 23).

Glucose-stimulated insulin secretion

Insulin secretion was measured after culturing islets from Wt and DN-MafK Tg mice for 4 hours in RPMI 1640 medium containing 11mM glucose supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (Sigma-Aldrich). Size-matching 5 islets were preincubated at 37°C for 30 minutes in Krebs-Ringer bicarbonate HEPES buffer containing 2.8mM glucose, followed by incubation with 2.8mM, 5.6mM, or 11.2mM glucose solution for 90 minutes. The isolated islets were extracted in acid-ethanol, and their insulin content was measured. Insulin was immunoassayed as described above.

Construction of adenovirus (Ad)-DN-MafK

An Ad vector containing DN-MafK and green fluorescent protein (GFP) genes was constructed with the help of O.D. 260 Inc. Briefly, DN-MafK cDNA along with rabbit β-globin polyA was cloned into a pE1.2 shuttle plasmid, and a GFP fragment along with rabbit β-globin polyA was inserted into a pE3.1 shuttle plasmid. These plasmids were then further modified as described previously (24). Ad that possessed the cytomegalovirus-GFP expression cassette in the E1 region of the virus genome was used as a control virus (O.D. 260 Inc.). The Ad titer was determined using the OD 260-sodium dodecylsulfate (SDS) method as described previously (24).

Cell culture and transduction

Cells from the MIN6 cell line (passage 43–50) were grown in DMEM containing 15% FBS or in glucose-free DMEM (Invitrogen) containing 10% dialyzed FBS (Invitrogen) and 1% penicillin-streptomycin with the indicated concentration of glucose (Sigma Chemical Co). The cells were then transduced with Ad-DN-MafK or Ad-GFP at multiplicity of infection of roughly 20. They were incubated for 2 hours, followed by washing and further culturing for 48–60 hours. Efficacy of infection was confirmed by fluorescence microscopy, and confirmation of flag-DN-MafK expression was performed by Western blotting using anti-Flag antibody. The collected cells were used for protein and RNA extraction.

Luciferase assay

The insulin promoter lesion (∼238–0 bp) containing plasmid and the reporter plasmid were generated. The Flag-DN-MafK cDNA was subcloned into the pcDNA 3.1 vector and these plasmids were transfected into Ad-GFP or Ad-DN-MafK infected MIN6 cells using Lipofectamine 2000 (Invitrogen). pcDNA plasmid was used to adjust the dose of DNA. Dual-Luciferase reporter assays were performed 48 hours after transfection according to manufacturer's protocol (Promega), then absorbance was measured using Glomax Luminometer (Promega). The firefly luciferase data normalized by Renilla was used for analysis.

Immunoblot analysis

Frozen tissues or collected cells were lysed in erythrocyte lysis buffer (50mM HEPES [pH 7.0], 250mM NaCl, 0.1% Nonidet P-40, 5mM EDTA, and 0.5mM dithiothreitol supplemented with 1mM phenylmethylsulfonyl fluoride, 1 μg/mL leupeptin, 1 μg/mL aprotinin, 50mM sodium fluoride, and 0.2mM sodium orthovanadate) containing benzamidine and β-glycerophosphate. Lysates were sonicated twice on ice and cleared by centrifugation. The protein content of the whole-cell extract was measured by NanoDrop (LMS). Equal amounts (20 μg) of proteins were separated on 10% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. The primary antibodies used were anti-MafF/G/K (1:2000), antiactin (1:2000), and anti-Flag for detecting Flag-DN-MafK (1:2000) antibodies. Antiactin antibody was used as a loading control. The secondary antibodies were antirabbit IgG (MafF/G/K), antigoat IgG (actin), or antimouse IgG (Flag). Analysis was performed using Amersham ECL Advance Western blotting detection kit (GE Healthcare), and images were obtained using the CCD-camera system LAS-4000 UV mini (Fujifilm).

Chromatin immunoprecipitation (ChIP) assay

ChIP analysis was performed using a ChIP assay kit (EMD Millipore). Ad-infected MIN6 cells were preincubated in a 10-cm dish for 48 hours. The cells were formaldehyde cross-linked for 10 minutes, after which they were washed and collected with PBS-containing protease inhibitors (50 μg/mL phenylmethylsulfonyl fluoride, 10 μg/mL aprotinin, and 10 μg/mL leupeptin). The cells were suspended in SDS lysis buffer and sonicated 5 times to obtain 200- to 1000-bp fragments. Immunoprecipitation was performed using Sperm DNA/Protein A agarose slurry. Two micrograms of the next antibodies were used for immunoprecipitation: antirabbit MafA (Bethyl), antirabbit Maf F/G/K (Santa Cruz Biotechnology, Inc), and normal rabbit IgG (Santa Cruz Biotechnology, Inc). Washing and chromatin elution were performed according to the manufacturer's instructions. Primers for the insulin promoter were TAATTACCCTAGGACTAAGTAGAGGTGTTG (forward) and AGGTGGGGTAGGTCAGCAGATGGCCAGA (reverse). 30 cycles were performed for PCR analysis. Quantitation of band density is performed using an imaging densitometer and normalized to the band density of control MIN6 cells.

RNA isolation and real-time PCR

Total RNA was isolated from the isolated islets and Ad-infected MIN6 cells using the RNeasy Mini kit (QIAGEN) according to the manufacturer's recommendation and was used as the starting material for cDNA preparation. A real-time PCR study was performed in duplicate on a 7500 Fast Real Time PCR system using SYBR Green PCR Master Mix (Applied Biosystems). The results were quantified using the comparative cycle threshold method, and the expression was normalized to glyceraldehyde 3-phosphate dehydrogenase.

Statistical analysis

Results are expressed as mean ± SE. Differences between the 2 groups were assessed using Student's t tests. Individual comparisons between more than 2 groups were analyzed by ANOVA. P < .05 was considered statistically significant. Data were analyzed using Ekuseru-Toukei 2012 (Social Survey Research Information).

Results

Expression of small-Maf factors in the pancreatic islets of mice fed an HFD

The role of endogenous small Maf factors in regulating pancreatic β-cell function is unknown. Therefore, we first confirmed the expression pattern of small Maf factors in islets. Pancreatic islets were isolated from 2 groups of C57BL/6J mice at 12 weeks of age, after feeding the animals either a normal diet (ND) or an HFD from 5 weeks of age. Whole-cell extracts were prepared and analyzed by Western blotting. Small Maf expression levels were significantly higher in the islets of the HFD-fed mice than in those of the ND-fed mice (Figure 1A). Pancreatic sections immunostained with insulin and small-Maf-specific antibody showed the expected increase in the islet size in the HFD-fed mice than in the ND-fed mice (Figure 1B). Furthermore, small Maf proteins were expressed and relatively highly observed in the nuclei of the β-cells in the islets (Figure 1C). These data show that the expression of small-Maf factors in pancreatic β-cells is enhanced in HFD-fed mice.

Figure 1.

Enhanced small-Maf expression in pancreatic β-cells in C57BL/6J mice fed a HFD. A, MafF/G/K and actin expression in 3 independent experimental islets isolated from 12-week-old mice fed a ND or an HFD are detected by Western blotting. Quantitation of band density was performed using an imaging densitometer. Values are expressed as mean ± SE. *, P < .05. B, Representative images of MafF/G/K and insulin staining in islets from ND- and HFD-fed mice. MafF/G/K proteins are detected in the nuclei of pancreatic β-cells from each mouse using an anti-MafF/G/K antibody. C, Immunohistochemistry of β-cells per high power field (green, insulin; red, MafF/G/K; blue, DAPI). MafF/G/K expression is confirmed mostly in the nuclei of β-cells.

Specific inhibition of small Maf factors in pancreatic β-cells

Despite the increased small Maf expression in β-cells, serum insulin levels remained higher in HFD-fed mice than in ND-fed mice, as did blood glucose levels (Figure 1A and Supplemental Figure 1), consistent with compensatory response to the insulin resistance. This finding also indicates that relatively impaired β-cell function during the compensatory phase may be associated with enhanced small Maf expression; therefore, the inhibition of small Maf function may overcome β-cell dysfunction. To test this hypothesis, we used the Flag-DN-MafK transgene as a negative regulator of endogenous small-Maf functions and prepared the Ad-DN-MafK infected MIN6 cells and DN-MafK Tg mice. In regards to Ad-infected MIN6, efficacy of infection was equivalent to control MIN6 cells (Figure 2A) and abundant DN-MafK protein expression was confirmed (Figure 2B). Because DN-MafK lacking a basic region in the DNA-binding domain (Figure 3A) did not bind to MARE on insulin2 promoter region (Supplemental Figure 2), DN-MafK can theoretically inhibit the function of all small Maf proteins, including MafF, MafG, and MafK. Indeed, our ChIP assay results suggested the repression of insulin promoter binding of endogenous small-Maf in the DN-MafK-expressed MIN6 cells compared with that in the control cells, whereas MafA binding to MARE was significantly increased (Figure 3B). Furthermore, luciferase assay using insulin promoter resulted in significant increment of insulin transcriptional activity in DN-MafK expression (Figure 3C).

Figure 2.

Studies of Ad-infected MIN6 cells. A, Confirmation of multiplicity of Ad infection. We selected multiplicity of infection of roughly 20, and infection was confirmed in almost all MIN6 cells. B, Confirmation of DN-MafK expression of Ad-DN-MafK-infected MIN6 cells. Bands (a) and (b) indicate Flag-DN-MafK and endogenous MafF/G/K, respectively. MafF/G/K antibody, Flag antibody, and actin antibody are presented.

Figure 3.

Characteristics of DN-MafK construct. A, Schematic image of DN-MafK lacking basic regions in the DNA-binding domain. B, DNA-binding activity to the MARE on the insulin promoter 2 using a ChIP assay using MIN6 cells (n = 6). Quantitation of relative band density compared with IgG bands is performed using an imaging densitometer. White bars, control Ad (Ad-GFP)-infected MIN6 cells; black bars, Ad-DN-MafK-infected MIN6 cells. Values are expressed as mean ± SE. *, P < .05 and **, P < .01. C, Relative insulin promoter transcriptional activity using a luciferase assay. White bars, Ad-GFP-infected MIN6 cells; black bars, Ad-DN-MafK-infected MIN6 cells. Values are expressed as mean ± SE. ***, P < .001.

All 3 lines of DN-MafK Tg mice (numbers 72, 23, and 53) showed normal size and growth (data not shown). We checked the copy numbers of integrated transgene for each line using Southern blotting (Figure 4A). All lines showed between 1 and 10 copies of transgene integration, and all had similar phonotypes. The line 53, which showed the most copies, was used for the later experiment. Next, we checked DN-MafK expression in various tissues. Western blot analysis using extracts from various tissues showed that DN-MafK was expressed only in the pancreas (Figure 4B). Moreover, immunohistochemistry data using anti-Flag tag antibodies demonstrated that DN-MafK was exclusively expressed in islet cells (Figure 4C).

Figure 4.

Generation of Tg mice with β-cell-specific expression of DN-MafK established on a C57BL/6J background. A, Copy numbers of the integrated transgene in lines 72, 23, and 53 of the Tg mice as determined by Southern blotting. B, DN-MafK expression in various tissues as analyzed by Western blotting (1, brain; 2, heart; 3, lung; 4, liver; 5, spleen; 6, pancreas; 7, kidney; 8, intestine; and 9, fat) Flag-DN-MafK is detected only in the pancreas. C, Representative images of Flag-DN-MafK staining in islets in Wt and Tg mice Flag-DN-MafK is costained with insulin-positive cells in the Tg mice. D and E, ipGTTs were conducted at 20 weeks of age (n = 6–10). The Tg mice show a significant improvement in blood glucose (D) and augmentation of early phase insulin secretion (E) compared with the Wt mice only under the HFD condition. Black circles, Wt mice on a ND; white circles, Wt mice on a HFD; black squares, DN-MafK Tg mice on a ND; white squares, DN-MafK Tg mice on a HFD. F, DN-MafK and MafA expression in isolated islets from the ND-fed or HFD-fed Wt and Tg mice as detected by Western blotting. Actin is used as a loading control. MafA is similarly elevated under the HFD condition in both the Wt and Tg mice.

After feeding both Wt and Tg male mice either the ND or the HFD from 6 weeks of age, ipGTT was performed at 20 weeks of age. Among the ND-fed mice, there were no significant differences in glucose tolerance between the Wt and DN-MafK Tg groups (Figure 4D). However, glucose tolerance in the HFD-fed Tg mice was significantly improved compared with that in the HFD-fed Wt mice (Figure 4, D and 4E, white circles and white squares). We isolated islets from these mice and performed Western blot analyses. The results demonstrated that MafA protein expression in the islets was slightly increased in the Wt and Tg HFD-fed mice compared with that in both groups of ND-fed mice. However, similar MafA expression levels were observed in both groups of HFD-fed mice (Figure 4F).

Phenotypic analysis of HFD-fed DN-MafK Tg mice

Because glucose tolerance was significantly improved in the HFD-fed Tg mice compared with that in the Wt mice, further phenotypic analyses were performed to clarify the factors affecting this improvement. There were no significant differences in body weight, food intake, and insulin sensitivity between the Wt and DN-MafK Tg mice (Figure 5, A and C, and Supplemental Figure 3). Nonetheless, at 16 weeks of age, a significant improvement in random blood glucose levels (Figure 5B) and area under the curve (AUC) for OGTT results was observed in the DN-MafK Tg mice compared with that in the Wt mice (Figure 5, D and E). Moreover, both fasting and postglucose-loaded serum insulin levels were significantly increased in the HFD-fed DN-MafK Tg mice (Figure 5F). Because impairment in in vivo GSIS was ameliorated in the HFD-fed DN-MafK Tg mice while their insulin sensitivity remained unchanged, we postulate that the dysfunction in glucose-responsive insulin-secretion machinery in β-cells may be rectified in these Tg mice.

Figure 5.

Metabolism of Wt and DN-MafK Tg mice under the HFD condition. A and B, Body weight (A) and ad libitum-fed blood glucose levels (B) are measured in the Wt mice (white circles) and DN-MafK Tg mice (black circles) at 6–16 weeks of age (n = 18–21). Body weight is not different between the Wt and Tg mice, whereas blood glucose levels are lower in the Tg mice. C and D, The ip insulin tolerance test (C) and OGTT (D) are performed in the Wt mice (white circles) and DN-MafK Tg mice (black circles) at 16–17 weeks of age (n = 9–10 and n = 14–17, respectively). Although insulin sensitivity is not different, glucose tolerance is significantly improved in the DN-MafK Tg mice. E, Area under the glucose curve (AUC) during the OGTT in the HFD-fed mice (n = 14–17). The AUC was also significantly lower in the Tg mice. F, Serum insulin concentrations are measured during OGTT (n = 10 for each group). The Tg mice show high levels of serum insulin both before and after glucose loading. Values are expressed as mean ± SE. ***, P < .001.

Islet morphology in the HFD-fed Wt mice and DN-MafK Tg mice

Some previous studies have reported changes in the morphology of pancreatic islets in conjunction with pancreas-specific knockout or overexpression of Maf factors (10, 25). In consideration of these findings, we next performed immunostaining of pancreatic sections with antibody against insulin and PCNA to investigate the islet morphology in the HFD-fed mice. Insulin-positive pancreatic cell mass was calculated as described earlier. There were no obvious changes in the morphology of islets (Supplemental Figure 4A) or the amount of pancreatic β-cells (Supplemental Figure 4B). Double fluorescence staining with anti-PCNA and antiinsulin antibody indicated that the proliferation of β-cells was also the same in the HFD-fed Wt and DN-MafK Tg mice (Supplemental Figure 4, C and D).

Insulin secretion and gene expression of insulin and glucokinase

To evaluate changes in islet function and gene profiling in DN-MafK Tg animals, we performed GSIS and real-time RT-PCR on RNA from pancreatic islets isolated from the HFD-fed Wt and DN-MafK Tg mice. The GSIS results showed enhanced insulin secretion from the DN-MafK Tg than the Wt islets at all glucose concentrations; the insulin content of islets was also higher in the DN-MafK Tg mice (Figure 6, A–C). Moreover, DN-MafK Tg mice showed a significant increase in the expression of Insulin1, Insulin2, and glucokinase genes (Figure 7A). On the other hand, the expression levels of Mafa and Glut2 were similar in both groups. DN-MafK expressing MIN6 cells also showed significantly higher levels of insulin1 and insulin2 gene expression as well as increase in glucokinase gene expression compared with the control MIN6 cells (Figure 7B). These results suggest that the inhibition of small-Maf function causes an increase in insulin1 and insulin2 gene expressions independent of Mafa expression, possibly in part via alterations in glucose metabolism resulting from increased glucokinase expression in the islets of the HFD-fed Tg mice.

Figure 6.

GSIS assay of mouse isolated islets. A and B, GSIS using size-matching isolated islets. After preincubation with Krebs-Ringer bicarbonate HEPES (KRBH) buffer containing 2.8mM glucose for 30 minutes, the islets are incubated in the presence of 2.8mM, 5.6mM, and 11.2mM glucose for 90 minutes. Supernatant insulin concentration is measured (A). B, Insulin concentration adjusted for insulin content of each well (n = 4 for each group). White bars, Wt mice; black bars, Tg mice. C, The insulin content in pancreatic islets is determined after acid-ethanol extraction. The islets from the Tg mice contained high insulin levels (n = 20). White bars, Wt; black bars, Tg.

Figure 7.

Gene expressions of isolated islets and Ad-infected MIN6. A, Comparison of gene expression between the Wt islets (white bars) and DN-MafK Tg islets (black bars) (n = 4 for each group). insulin1, insulin2, and glucokinase were significantly elevated in the HFD-fed Tg mice compared with those in the Wt mice (n = 4 for each group). White bars, Wt; black bars, Tg. Values are expressed as mean ± SE. **, P < .01. B, Comparison of gene expression between control Ad (Ad-GFP; white bars) and Ad-DN-MafK (black bars)-infected MIN6 cells (n = 6 for each group). DN-MafK expression significantly elevated insulin1 and insulin2 and tended to increase glucokinase (n = 6 for each group). Values are expressed as mean ± SE. *, P < .05.

Discussion

From this study, we were able to draw 2 major conclusions. First, the inhibition of endogenous small-Maf function using DN-MafK may alter the binding activity of other transcriptional factors, including MafA. Small-Maf factors are known to heterodimerize with the CNC transcriptional family. However, when they form homodimers, they may function as competitive inhibitory factors for MARE binding and would compete with MafA for binding to these sites (20). A heterodimer of endogenous small Maf and DN small Maf with mutations in the DNA-binding domain will suppress the DNA-binding ability of the endogenous small-Maf partner. Importantly, as the small-Maf factors do not form heterodimers with large-Maf factors, DN-MafK will not directly affect the binding and function of large-Maf factors. In a previous study, β-cell-specific overexpression of MafK was found to result in compensatory enhancement of MafA binding (19). Our results suggest that a similar underlying mechanism may exist in our study. Moreover, MafA expression levels in islets were similar between HFD-fed Wt and DN-MafK Tg mice (Figure 2G). This may represent a compensatory mechanism for β-cells to adapt to a higher insulin demand. Furthermore, the comparable expression of MafA in both groups of HFD-fed mice suggest that the amelioration of glucose tolerance in HFD-fed Tg mice likely results from the inhibition of small-Maf function, not from the enhancement of MafA expression in the β-cells. One possibility is that due to the competition between these transcriptional factors for MARE binding, DN-MafK Tg islets may have relatively higher proportion of MafA bind to the MARE.

Second, inhibition of small-Maf function resulted in significantly increased insulin secretion via the enhanced expression of insulin1 and insulin2 genes from pancreatic islets. This finding may be partially explained by the elevation of MafA binding to MARE on insulin promoters, as described above. Such an indirect increase in the binding of MafA to insulin MARE elements will lead to the induction of insulin gene transcription. Moreover, there is a possibility that DN-MafK directly enhances insulin gene expression by inhibiting the repressive effects of endogenous small-Maf factors. It is important to note that this increase in insulin gene expression occur independent of any increase in Mafa mRNA expression and MafA protein level. Another possibility is that incremental glucokinase expression in the islet of DN-MafK Tg mice may in part affect the insulin gene expression. Glucokinase is the rate-limiting enzyme of the glycolytic pathway (26), and it acts as a glucose sensor for GSIS in the pancreas (27, 28). Moreover, it was reported that glucokinase activation actually increases pancreatic β-cell proliferation (29), and in HFD-fed mice, haploinsufficiency of β-cell-specific glucokinase resulted in impaired β-cell mass and function (30). Some previous studies showed that the overexpression of MafA or PDX-1 in pancreatic islets and β-cell lines similarly resulted in the up-regulation of glucokinase mRNA expression (31, 32), and NeuroD1 was also proposed to regulate pancreatic glucokinase activity (33). In addition, MafA is known as a positive regulator of Pdx-1 and NeuroD1 (10, 31). In our study, the up-regulation of MafA binding may have partially resulted in glucokinase expression. Because of the inhibition of small-Maf factors, similar to the effect of glucokinase activation, fasting plasma insulin levels and GSIS under basal glucose conditions were elevated both in vivo and in vitro. Despite ameliorated gene expression of glucokinase, we could not confirm the obvious elevation of glucokinase protein level. To make clear these points, further studies including glucokinase activity may be needed. HFD-fed Tg mice did not show increased β-cell proliferation (Supplemental Figure 3D). One possibility is that HFD itself already increased β-cell mass and proliferation to sufficient levels where additional effects were not required to improve β-cell function. These findings suggest that small-Maf factors regulate not only insulin transcription via MARE binding but also possibly intracellular glucose metabolism and insulin release from β-cells via glucokinase expression.

Previous reports on small-Maf factors have already established that they play crucial roles in stress signaling, such as in the case of oxidative stress (16). In terms of the response to stress, small-Maf factors may suppress excessive insulin expression to avoid the accumulation of intracellular endoplasmic reticulum stress in β-cells presumably via Nrf2, which is one of the counterparts of small Maf factors. Although overexpression of both MafA and DN-MafK results in enhanced insulin gene expression, insulin synthesis, and insulin secretion, but DN-MafK can uniquely accomplish these objectives without enhancing MafA expression.

In conclusion, small-Maf factors play important roles as inhibitors of insulin transcription and secretion and, possibly, regulators of intracellular glucose metabolism. Further investigation of the function of endogenous small Maf factors in pancreatic β-cells can lead to a better understanding of the pathogenesis of diabetes.