MafA Is Required for Postnatal Proliferation of Pancreatic β-Cells

The postnatal proliferation and maturation of insulin-secreting pancreatic β-cells are critical for glucose metabolism and disease development in adults. Elucidation of the molecular mechanisms underlying these events will be beneficial to direct the differentiation of stem cells into functional β-cells. Maturation of β-cells is accompanied by increased expression of MafA, an insulin gene transcription factor. Transcriptome analysis of MafA knockout islets revealed MafA is required for the expression of several molecules critical for β-cell function, including Glut2, ZnT8, Granuphilin, Vdr, Pcsk1 and Urocortin 3, as well as Prolactin receptor (Prlr) and its downstream target Cyclin D2 (Ccnd2). Inhibition of MafA expression in mouse islets or β-cell lines resulted in reduced expression of Prlr and Ccnd2, and MafA transactivated the Prlr promoter. Stimulation of β-cells by prolactin resulted in the phosphorylation and translocation of Stat5B and an increased nuclear pool of Ccnd2 via Prlr and Jak2. Consistent with these results, the loss of MafA resulted in impaired proliferation of β-cells at 4 weeks of age. These results suggest that MafA regulates the postnatal proliferation of β-cells via prolactin signaling.


Introduction
Accumulating evidence suggests that postnatal organ development and maturation are critical for future health, especially with respect to metabolic disease [1]. Pancreatic b-cells vigorously proliferate postnatally to increase insulin secretion capacity [2], which is implicated in adult b-cell mass [3]. Although the compensatory growth of b-cell mass in insulin resistance has been intensively investigated [4], the signaling pathway that regulates postnatal proliferation of b-cells is less well known [5]. Uncovering this mechanism will elucidate how b-cell mass is regulated during development and how the insulin-expressing cells that differentiate from stem cells acquire the capacity to proliferate.
Maturation of b-cells occurs concurrently with the expression of v-maf musculoaponeurotic fibrosarcoma oncogene family protein A (MafA) [9], a transcription factor that regulates the expression of insulin via the C1-A2 elements of the insulin promoter [10]. In the pancreas, MafA is expressed exclusively in mature b-cells. Forced expression of MafA with Pdx1 and Ngn3 converts pancreatic acinar cells into insulin-secreting cells [11]. MafA expression is reduced in the b-cell with compromised function [12]. In the islets of the MafA knockout (KO) mice, the ratio of the b-cell mass to the a-cell mass is normal at birth; however, this ratio is reduced during the neonatal period [13], suggesting that MafA may be involved in regulation of the postnatal b-cell mass. Thus, the role of MafA in postnatal proliferation of b-cells was investigated in this study.

Mice
This study was carried out in strict accordance with the Fundamental Guidelines for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions under the jurisdiction of Ministry of Health, Labour and Walfare.
The protocol was approved by the Animal Care and Use Committee of the National Center for Global Health and Medicine (Permission Number: 13104). Islet isolation and pancreatic dissection were performed under deep anesthesia followed by cervical dislocation, and all efforts were made to minimize suffering. The generation of MafA KO mice was described previously [13]. Male mice were analyzed in this study. Mice were genotyped by NaOH extraction methods as described previously [14]. The primers used in this analysis are listed in Table S2 in File S1.

Construction of Mouse Prolactin Reporter Luciferase Vectors
A reporter vector containing the human Prlr promoter (hPrlr) was obtained from the Promoter Reporter GoClone Collection (Promega, Madison, WI). The mouse Prlr (mPrlr) luciferase reporter vectors mPrlrP-1, mPrlrP-2 and mPrlrP-3 were generated by amplifying 2359 bp, 1304 bp and 608 bp fragments of the mPrlr promoter from high-quality mouse genomic DNA (Clontech) by PCR with the primers listed in Table S3 in File S1. An infusion cloning kit (Promega) was utilized to clone the amplified products into the pGL4.10 vector (Clontech, Palo Alto, CA), which was digested with NheI and HindIII. The reporter vectors with deletions of the putative MafA binding regions, mPrlrP-5, mPrlrP-6, mPrlrP-8, mPrlrP-9 and mPrlrP-11, were generated using the PrimeSTAR Mutagenesis Basal Kit (Takara Bio, Shiga, Japan); the reactions were performed using the primers listed in Table S3 in File S1, and the pGL4.10-mPrlrP-1 reporter vector was used as a template. MafA binding sites were predicted with TRANSFAC (BIOBASE, Beverly, MA). The sequences of the reporter vectors were confirmed by sequencing with the universal RVprimer3.

Transcriptome and Quantitative RT-PCR Analysis of Isolated Islets and Cultured Cells
The islets were isolated from MafA KO or wild-type mice at 7 weeks of age using collagenase digestion as described previously [14]. Total RNA was extracted from the isolated islets or cultured cells using the QIAshredder and RNeasy Micro Kit (Qiagen Valencia, CA) following the manufacturer's instructions. The concentration of purified RNA was measured by a NanoDrop ND 1000 Spectrophotometer (Thermo Scientific, Rockford, IL). The A260/280 of RNA from the wild-type and MafA KO islets were 1.9460.07 and 1.8660.04, respectively. RNA expression in the islets isolated from the MafA KO and wild-type mice was compared using the Mouse 430 2.0 Array (Affymetrix, Santa Clara, CA, USA, n = 2) as described previously [15]. Samples for the analysis were prepared in accordance with the manufacturer's protocol, and the results were analyzed using the DAVID 6.7 [16] and IPA programs (Ingenuity Systems, Redwood City, CA). Reverse transcription was performed using high-capacity cDNA reverse transcription kits (Applied Biosystems, Foster City, CA). Quantitative PCR amplification was performed using the Taq-Man universal PCR master mix core reagent kit (Applied Biosystems) with the probes listed in Table S4 in File S1 and was analyzed using an ABI Prism 7900 (Applied Biosystems); C t values were measured in duplicate. mRNA was quantified by normalization to b-actin expression using the 2-DDCt method. For analysis of the islets from the MafA KO mice and wild-type littermates, the expression of MafA was examined in all assays to confirm that MafA was absent in KO islets. The data are presented as the means 6 S.E.M., and statistical significance was determined using a two-tailed unpaired Student's t-test.
For the prolactin stimulation experiments, INS-1 cells were transfected with siRNA against mouse MafA, rat MafA or rat Prlr (Silencer Select siRNA s233236, s172995 or Silencer siRNA 48147, respectively) or control siRNA (Life Technologies) using Lipofectamine 2000 (Life Technologies) according to the manufacturer's protocol. The cells were plated in 6-well plates or 60 mm dishes, and the medium was changed 24 hours after transfection. The reduced expression of MafA or Prlr was confirmed by qRT-PCR 48 hours after transfection as described above. In parallel, the medium in the 60 mm dishes was changed to a defined serum-free medium consisting of RPMI-1640 medium supplemented with 10 mM glucose (Sigma), 0.1% human serum albumin, 10 mg/ml human transferrin, 0.1 nM triiodothyronine, 50 mM ethanolamine, 50 mM phosphoethanolamine (Wako, Osaka, Japan), 10 mM HEPES, 100 U/ml penicillin and 100 mg/ml streptomycin [20]. After 24 hours, mouse prolactin (R&D Systems, #1445-PL-050, Minneapolis, MN) was added at a final concentration of 1 mg/ml or at the indicated concentration, and the cells were incubated for 8 hours with or without AG490 (Millipore, Billerica, MA). The cells were then evaluated by immunoblot analysis or immunofluorescence.

Reporter Assay
HeLa cells were transfected with the indicated reporter plasmids and the pCMV-b-gal plasmid (Promega), as an internal control, using Lipofectamine 2000 (Life Technologies) according to the manufacturer's protocol. The medium was changed 24 hours after transfection. At 48 hours after transfection, the firefly luciferase activity of the cells was analyzed using the luciferase assay system (Promega), and the b-gal activity was assessed as reported previously [14]. The data are presented as the means 6 S.E.M., and statistical significance was determined using a two-tailed unpaired Student's t-test.

Immunoblot Analysis
INS-1 cells were harvested in PBS and sonicated in 200 ml of buffer containing 10 mM Tris/HCl (pH 7.5), 150 mM NaCl, and 1% TX-100 supplemented with the protease inhibitor and phosphatase inhibitor cocktails (Nacalai Tesque, Kyoto, Japan). The lysates were centrifuged at 20,0006g rpm for 2 minutes. The supernatant was analyzed by SDS-PAGE followed by immunoblot analysis, as described previously [21]. For the nuclear fractionation, Subcellular Protein Fractionation Kit for Cultured Cells (Thermo Scientific) was used following the manufacturer's instructions. For immunoprecipitation, the supernatant was incubated with 20 ml of the indicated antibodies immobilized on 10 ml of protein G Sepharose 4 fast-flow beads (GE Healthcare, Fairfield, CT) for 3 hours at 4uC. After the beads were washed four times with 500 ml of buffer, containing 10 mM Tris/HCl (pH 7.5), 150 mM NaCl, and 0.33% Triton X-100 supplemented with protease inhibitor and phosphatase inhibitor cocktails, the immunoprecipitates were subjected to SDS-PAGE and immunoblot analysis with the antibodies listed in Table S5 in File S1. Images were obtained with ChemiDoc XRS Plus (Biorad, Hercules, CA) and quantified using Image Lab 3.0 (Biorad). The data are presented as the means 6 S.E.M., and statistical significance was determined using a two-tailed unpaired Student's t-test.

Immunofluorescence
For immunofluorescence of cells, INS-1 cells were fixed with 4% (w/v) paraformaldehyde for 15 min and were permeabilized with 0.3% (w/v) Triton X-100 in PBS for 15 min. After blocking with PBS containing 10% (w/v) BSA for 30 min, the cells were incubated overnight with the primary antibodies (Table S5 in File S1) in PBS containing 3% (w/v) BSA. The cells were then washed with PBS and incubated with Texas red-conjugated anti-rabbit IgG and FITC-conjugated anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA) in PBS containing 3% (w/v) BSA for 1 h, followed by washing with PBS and mounting with Mounting Medium containing 49,6-diamidino-2-phenylindole (DAPI) (Vector, Burlingame, CA).
Immunostaining analyses of mice pancreatic sections were performed on paraffin-embedded sections as described previously [14]. The primary antibodies used in this study are listed in Table  S5 in File S1. For amplification, biotinylated anti-mouse antibodies (Jackson ImmunoResearch) were used at a 1:400 dilution, followed by incubation with streptavidin-conjugated Alexa Fluor 488 (1:400) (Life Technologies). The secondary antibody was DyLight 594-conjugated anti-guinea pig IgG (Jackson ImmunoResearch). DAPI mounting medium (Vector) was used to label the nuclei.
For both staining, immunofluorescent images were obtained using an Olympus FV-1000 (Olympus) in confocal mode, and the acquired images were identically processed using Adobe Photoshop CS5.1. Quantification was performed using NIH ImageJ software. For the BrdU + /insulin + cells from the MafA KO and wild-type pancreata, a total of 1706 and 1262 insulin + cells, respectively, were counted from 3 mice of each genotype. The data are presented as the means 6 S.E.M., and statistical significance was determined using a two-tailed unpaired Student's t-test.

BrdU Incorporation Study
A 10 mg/ml solution of 59-bromo-29 deoxyuridine (BrdU, Sigma) in PBS was filter-sterilized and kept on ice. A dose of 100 mg BrdU/kg body weight was injected intraperitoneally into the mice, followed by ad libitum feeding for 24 hours. The mice were sacrificed, and BrdU incorporation in the pancreas was analyzed by immunofluorescence; a small section of the duodenum was used as a positive control for BrdU incorporation.

MafA Is Involved in the Expression of Prolactin Receptor in b-Cells
To examine the role of MafA in b-cell, transcriptome analysis of MafA KO islets at 7 weeks of age was performed. Several downregulated genes were detected, including Slc30a8 (ZnT8), Urocortin 3, Prlr and Ccnd2 (Table S1 in File S1). Among the 9 molecules that were downregulated in both sets (Table 1), prolactin receptor (Prlr) was our primary focus because Prlr KO neonates have reduced b-cell mass [8]. The expression of Ccnd2, one of the downstream targets of prolactin signaling [22], was also markedly reduced in MafA KO islets (  Fig. 1C). Reduced protein expression of Prlr and Ccnd2 was also detected in MafA KO islets ( Fig. 1D and E). Quantification and statistical analyses showed that the expression of Prlr and Ccnd2 in KO islets was 24.8615.9% and 50.3615.9% of those in wild-type islets (p = 0.04 and 0.05), respectively (Fig. 1F). These results suggest that MafA is directly or indirectly involved in the expression of Prlr and Ccnd2 in b-cells.
In silico analysis detected six conserved MafA recognition elements (MAREs) between the transcription start site (TSS) and 23000 bp in the mouse Prlr promoter (Fig. 2B), and these sites were also preserved in the human and rat genes (data not shown). A luciferase reporter assay using the human and mouse Prlr (hPrlr and mPrlr, respectively) promoters clearly showed that overexpression of MafA transactivated both the hPrlr (p,0.01) and mPrlr (p,0.05) promoters in HeLa cells ( Fig. 2A and B). MafA activated the 22232 to +127 mPrlr promoter fragment to a greater extent than the 21177 to +127 fragment (p = 0.05) and the 2481 to +127 fragment (p = 0.06) (Fig. 2B). Moreover, the relative luciferase activity of various mPrlr promoter deletion constructs concomitant with forced expression of MafA revealed that the regions from 2217 to 2207 (IV; p,0.01) and from 2 2026 to 21409 (I to III; p,0.01) were involved in the MafAmediated transcriptional activation of the mPrlr promoter (Fig. 2C). These results suggest that MafA transcriptionally regulates the expression of Prlr.   (Fig. 3B). In INS-1 cells, stimulation with 1 mg/ml prolactin resulted in the tyrosine phosphorylation of Stat5B (Fig. 3C-F), and this effect on Stat5B phosphorylation was dose-dependent (Fig. 3G). Inhibition of Prlr expression with siRNA resulted in reduced phosphorylation of Stat5B but did not affect its total protein level (Fig. 3C-F). However, Inhibition of MafA expression with siRNA, resulting in 15.2% reduction in Prlr expression (Fig. 3A), did not significantly alter the tyrosine phosphorylation of Stat5B in b-cells (Fig. 3C and D). Therefore, an siRNA targeting Prlr was used to examine the effect of reduced Prlr expression.
In INS-1 cells, the majority of Stat5B was localized in the cytoplasm (Fig. 3H) and was translocated into the nucleus after prolactin stimulation (Fig. 3I). However, this prolactin-induced redistribution of Stat5B was impaired when Prlr expression was inhibited by siRNA ( Fig. 3J and K). Inhibition of the prolactininduced translocation of Stat5B was also observed after the addition of AG490, a Jak2 inhibitor (Fig. 3L-N). These results confirm the previous findings that prolactin-induced phosphorylation and nuclear translocation of Stat5B are mediated by Prlr and Jak2 [20]. Using these systems, the effect of prolactin signaling on the expression of Ccnd2 was examined. With upregulation in the phosphorylation of Stat5 in INS-1 cells, prolactin stimulation increased the amount of nuclear Ccnd2 (Fig. 3O). Quantitative analysis demonstrated that nuclear Ccnd2 protein was increased to 139.2610.6% after prolactin stimulation (p = 0.01). Ccnd2 protein relative to control Lamin A/C was 189.3644.4% of the control level (p = 0.09; Fig. 3P). We also examined the mRNA expression of Ccnd2 in INS-1 cells with siRNA targeting Prlr compared to those with control siRNA. mRNA expression of Prlr and Ccnd2 in INS-1 cells with siRNA targeting Prlr was 15.360.5% (p,0.01) and 84.962.9% (p = 0.01) respectively, relative to the controls (Fig. 3Q). These results verify the previous data [7,24,25], suggesting that prolactin signaling regulates the expression of Ccnd2 in b-cells.

Reduced Proliferation of b-Cells in MafA KO Mice at 4 Weeks of Age
Postnatal proliferation of b-cells is critical for the b-cell mass in adults [3]. Our above results prompted us to examine if the reduced expression of Prlr and Ccnd2 in the b-cells of MafA KO mice affected postnatal proliferation, as Ccnd2 KO neonates exhibit reduced proliferation of b-cells and reduced b-cell mass [26]. The b-cell to a-cell ratio in the islets of the MafA KO mice is also reduced after birth [13]. A BrdU incorporation assay was performed in MafA KO mice and their wild-type littermates at 4 weeks of age. BrdU-positive staining was observed in 0.3560.0%

Discussion
This study described transcriptome analysis of the islets isolated from MafA KO mice. The results revealed the downstream candidates of MafA, and Prlr was a focus of this study. In the embryonic pancreas, Prlr is expressed primarily in acinar cells and ductal epithelium during early gestation. Later in gestation and in the postnatal period, Prlr is expressed predominantly in pancreatic islets [27], when MafA is expressed in b-cells [21]. The results from this study collectively suggest that MafA is critical for the expression of Prlr and that Prlr/Jak2/Stat5B signaling can induce the expression of Ccnd2 in b-cells. Consistent with these results, loss of MafA expression resulted in the impaired proliferation of postnatal b-cells. Thus, prolactin signaling may play an important role in the proliferation of neonatal b-cells under the control of MafA, in addition to its role in b-cell proliferation during gestation. Because the use of transformed b-cell lines may hamper the analysis of the promoter activity or the expression of Ccnd2 in detail, more analysis is needed to clarify the role of prolactin signaling on cell cycle in b-cells and to exclude the possibility that MafA directly activates the Ccnd2 promoter [28]. In addition to Prlr and the previously reported potential target genes of MafA/ MafB, such as ZnT8 [29], Granuphilin [30] and Glut2 [13,31], transcriptome analysis of the MafA KO islets in this study showed the downregulation of Vdr, Pcsk1 and Urocortin 3, which are supposedly critical for b-cell function; however, the direct MafA binding sites in the promoters of these genes remain unknown.
In rodents, prolactin and placental lactogen bind only to Prlr [8]. During gestation, the expression of Prlr in pancreas and the serum prolactin level increase [5,32], although the action of prolactin is antagonized by progesterone [33]. b-Cell-specific expression of placental lactogen-I results in accelerated b-cell proliferation, increased b-cell mass and number and increased insulin production, leading to hypoglycemia and elevated plasma insulin [34]. In contrast, Prlr KO neonates have reduced b-cell mass [8]. During pregnancy, Prlr +/2 mice have reduced b-cell replication, but there is no increase in b-cell apoptosis, resulting in reduced b-cell mass. In Prlr +/2 mice, impaired glucose clearance, decreased glucose-stimulated insulin release, higher post-prandial blood glucose, lower insulin levels and attenuated increases in islet density, b-cell number and mass are also observed throughout pregnancy, but not in the absence of pregnancy [35]. These results suggest the important role of Prlr in b-cell proliferation during pregnancy.
At birth, maternal serum prolactin continues to rise to facilitate mammary gland function, while secretion of placental lactogen from the placenta peaks during mid-gestation [5,34]. Thus, prolactin, which may originate from the mother's feeding, may have a dominant effect on the neonatal proliferation of b-cells, not placental lactogen. An earlier study showed that prolactin and placental lactogen increase neonatal islet proliferation and insulin secretion [36]. Our results provide molecular evidence that Prlr is important for the postnatal proliferation of b-cells. Because b-cell replication in neonates plays a major role in b-cell mass in adult humans [3], the regulation of b-cell mass by prolactin signaling in postnatal pancreas may be implicated in an individual's susceptibility to diabetes.
The subcellular localization of Stat5 has been used to characterize the activation of the Jak2/Stat5 pathway by prolactin [20]. In b-cells, after exposure to prolactin, the redistribution of Stat5B from the cytoplasm to the nucleus was much higher than the redistribution of Stat5A, indicating that Stat5B plays a major role [23]. Stat5B binds to the GAS motif to induce the expression of its target genes [6], and our data suggested that Ccnd2 can be a Stat5B target gene. Consistent with our data, dominant-negative Stat5 reduces mRNA and protein level of Ccnd2 and inhibits Sphase entry [24,25]. Constitutively active form of Stat5B binds to the GAS motif in the Ccnd2 promoter, transactivates the Ccnd2 promoter and induces the proliferation of b-cells [24]. Prolactin increases the mRNA expression of Ccnd2 in rat islets [7]. In vivo, Ccnd2 is essential for the postnatal expansion of b-cell mass and the compensatory increase in b-cell mass in response to insulinresistant states [20,37]. However, during gestation or in neonates, the regulation of Ccnd2 expression by the Prlr/Jak2/Stat5B pathway and its role in the proliferation of b-cells have not been confirmed in vivo [5]. Indeed, in pregnant Prlr +/2 mice, b-cell mass and Jak2 phosphorylation are decreased, but the expression of Ccnd2 is not changed [38], suggesting that more studies are necessary to investigate the precise role of prolactin signaling in the regulation of Ccnd2 expression and b-cell proliferation in vivo. Another target of prolactin signaling that may be implicated in bcell proliferation is Tph1, an enzyme to synthesize serotonin [39]. However, recent study showed no difference in b-cell proliferation between Tph1 KO mice and wild-types during pregnancy even in the absence of serotonin [40], suggesting the importance of Ccnd2.
In addition to its effect on proliferation, accumulating evidence suggests that prolactin signaling is also critical for b-cell function. Prolactin increases the expression of molecules such as insulin, Glut2, Gck, Tph1, FoxM1 and Prlr in rat islets or INS-1 cells [6,7]. Prolactin and placental lactogen stimulate insulin release and increase insulin content in cultured adult mouse islets and adult or newborn rat islets [41]. INS-1 cells constitutively expressing placental lactogen-II have increased Preproinsulin and Glut2 mRNA [42]. These results raise the possibility that prolactin signaling is involved in the functional maturation of bcells from the immature insulin-expressing cells found in neonates. Prolactin increases the binding of Stat5 to the GAS motif of the Gck promoter and the Gck synthesis in b-cells even in the absence of glucose [6], suggesting its action can be independent of glucose. It would be interesting to examine if the b-cell-specific expression of Prlr or Ccnd2 improves both the proliferation and function of bcells in MafA KO mice. Although Bcl2 and BclXL are also downstream targets of prolactin signaling, there is no increase in the apoptosis rate of b-cells in pregnant Prlr +/2 mice [35] and in MafA KO mice [43], suggesting that prolactin signaling does not play a major role in apoptosis of b-cells.
Clinical studies have demonstrated that men and women with hyperprolactinemia have postprandial hyperinsulinemia and an exaggerated insulin secretory response to glucose and arginine [44,45]. Thus, further studies are needed to elucidate the effects of prolactin and placental lactogen on the proliferation and functional maturation of human b-cells. Moreover, activation of prolactin signaling or inhibition of progesterone signaling [33,46] in insulin-expressing cells differentiated from human stem cells or endocrine precursor cells may enhance the proliferation and functional maturation of these cells.

Supporting Information
File S1 Tables S1-S5. Table S1. Genes that were downregulated in the islets of MafA KO mice. Table S2. Genotyping primers used in this study. Table S3. Primers used to clone the indicated promoters or to mutagenize the mPrlr promoter. Table  S4. TaqMan probes used in this study. Table S5. Antibodies used in this study. (DOCX)