Npas4 Is a Novel Activity–Regulated Cytoprotective Factor in Pancreatic β-Cells

Chemicals were purchased from Fisher Scientific or Sigma-Aldrich. Cell culture reagents and disposables were obtained from Hyclone, LifeTech, BD-Falcon, and Corning.

All procedures were approved by either the University of British Columbia (UBC) or l’Université de Montréal animal care committees. For timed matings, noon on the day the vaginal plug was discovered was considered e0.5. All glucose and intralipid infusions were performed as described by Fontés et al. (29). Islets were isolated from mice at 8–15 weeks of age through standard collagenase digestion.

Immunostaining was performed on paraformaldehyde (PFA)-fixed, paraffin-embedded tissues as previously described (30). Primary antibodies included rabbit anti-Npas4 (1:350; Abcam), mouse antiglucagon (1:2,000; Sigma-Aldrich), and guinea pig anti-insulin (1:2,000; Millipore). Fluorescently conjugated secondary antibodies were from Jackson ImmunoResearch. Sections were imaged on either a Leica SP5 II confocal imaging system or an Olympus BX61 equipped for widefield fluorescence.

MIN6 cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, 1× glutamine, and 1× penicillin/streptomycin (pen/strep). Cells were passaged once a week and fed every 2nd day. Islets were cultured in RPMI supplemented with 10% FBS, 1× glutamine, and 1× pen/strep.

Npas4 adenovirus was constructed using the AdEasy system (31). MIN6 cells were infected for 2 h, and cells were then washed with PBS prior to further growth. Islets were picked into experimental plates, and after a 4-h recovery, virus was added and infection took place overnight; media was changed daily. As previously described, all cells were infected with a multiplicity of infection of 100:1 (32).

RNA was extracted and purified with TRIzol (LifeTech) according to the manufacturer’s protocol. cDNA synthesis was performed using Superscript II (LifeTech). For qPCR, 40 ng of cDNA was amplified using PrimeTime primers/probes (IDT) in a ViiA7 real-time PCR system (ABI); samples were run in triplicate and the expression level was quantified through ∆∆CT determination with GusB as a reference gene (33,34). The probe/primer1/primer2 used were as follows: 18s, TGCTCAATCTCGGGTGGCTGAA/GAGACTCTGGCATGCTAACTAG/GGACATCTAAGGGCATCACAG; β-actin, CCCATTCCCACCATCACACCCT/GCCTCGTCACCCACATAG/CTGTATTCCCCTCCATCGTG; Ddit3, CTTGACCCTGCGTCCCTAGCTTG/GCACCTATATGTCATCCCCAG/TGCGTGTGACCTCTGTTG; Gapdh, TGCAAATGGCAGCCCTGGTG/GTGGAGTCATACTGGAACATGTAG/AATGGTGAAGGTCGGTGTG; GusB, TCTAGCTGGAAATGTTCACTGCCCTG/CACCCCTACCACTTACATCG/ACTTTGCCACCCTCATCC; GusB (rat) TGGGCGATCAGCGTCTTGAAGTAAT/GGTCGTGATGTGGTCTGTG/TGTCTGGCGTCATATCTGGTATTG; Ins1, TGTTGGTGCACTTCCTACCCCTG/ATCAGAGACCATCAGCAAGC/GTTTGACAAAAGCCTGGGTG; Ins1 promoter, AGACCTAGCACCAGGCAAGTGTTT/GTGTTGATGTCCAATGAGTGC/CCCATTAAGGTGTCCAGGTG; Ins2, CCTCCACCCAGCTCCAGTTGT/TGATCTACAATGCCACGCTTC/GGCTTCTTCTACACACCCATG; Ins2 promoter, CAGGGAAGTGTTTGGAAACTGCAGC/GGACTAAGTAGAGGTGTTGA/CTGGACTTTGCTGTTTGACC; Ins2 (rat), AAGAATCCACGCTCCCCACACA/CCCAGGCTTTTGTCAAACAG/CACCCAGCTCCAGTTGTG; MafA, ACTTCTCGCTCTCCAGAATGTGCC/AGTCGTGCCGCTTCAAG/CGCCAACTTCTCGTATTTCTCC; NeuroD1, TCATGAGTGCCCAGCTTAATGCCA/TCTTTCGATAGCCATTCGCATC/ATAGTGAAACTGACGTGCCTC; Npas4 (human), ACGCTTCTGCTCAACACTACCGC/GTCCTAATCTACCTGGGCTTTG/CCATCTCTGCCTGAATATCTCC; Npas4, CGTTGGTTCCCCTCCACTTCCAT/GTCTCAACATTCCCCTACGAAG/GACAATATGCCATACTTGAAGACG; Npas4 (rat), CTTGAGCAGAGAGAAGCCCCGTG/ATGAGTCTTGCCTGCATCTAC/ACAAGTAGAAATCCAGGTAGTGC; Pdx1, TTCCGCTGTGTAAGCACCTCCTG/GTACGGGTCCTCTTGTTTTCC/GATGAAATCCACCAAAGCTCAC; Rgs2, AGCGGGAGAAAAGAAGCGGACA/CTGAGAATGCAAAGTGCCATG/AAGTAGCTCAAACGGGTCTTC; Rgs2 enhancer, ATTCGCTCTGGTCAGGCTTGTGT/ATTGTCTTCCCACTGTCACG/GCTGTTTCCTCTATGAGCTGG; Rgs2 intron 1, AGCTCAGAACTAACAATGCAACGTCCA/TGTAGCTGAATAACGTCCACC/TCAACTCAAACATCTCCACCC; Rplp0, TGTCTTCCCTGGGCATCACGTC/TGACATCGTCTTTAAACCCCG/TGTCTGCTCCCACAATGAAG; TBP (human), TCCAAAGGATGCAGAGAAAGCCATCA/GGGAGCATTGGTGAAATAGAATC/GGTTTTGGTCTGGAGTCTGAG.

Cells were lysed in standard lysis buffer, and protein was denatured by boiling. Crude lysates were processed using sonication (S-4000 with cuphorn; Misonix) for 2 min (85%) and centrifuged at 10,000g for 10 min. Proteins were separated using standard SDS-PAGE and blotted to polyvinylidene fluoride membrane (Bio-Rad). Membranes were blocked with 5% milk powder in Tris-buffered saline with Tween and probed with 1:10,000 rabbit anti-Npas4 (M.E. Greenberg, Harvard University, Cambridge, MA), 1:500 mouse anti-CHOP (Santa Cruz Biotechnology), 1:1,000 chicken anti-Rgs2 (Sigma-Aldrich), 1:2,000 mouse anti-Pdx-1 (Developmental Studies Hybridoma Bank), 1:2,000 mouse anti-NeuroD1 (Cell Signaling), 1:5,000 rabbit anti-MafA (Abcam), or 1:125,000 mouse anti-GAPDH (Sigma-Aldrich) overnight at 4°C. Membranes were then probed with horseradish peroxidase–conjugated secondary antibodies used at a dilution of 1:10,000 (Jackson ImmunoResearch) and visualized using ECL Prime (GE Biosciences). Densiometry was used to quantify protein amounts (ImageJ).

MIN6 cells were seeded at 8.0 × 10⁵ cells/well in 12-well plates (or 2.0 × 10⁵ cells/12-mm coverslip) and cultured for 36 h; media was then changed to low-glucose DMEM. The next morning, cells were stimulated in high-glucose DMEM supplemented with 40 mmol/L KCl. Alternatively, cells were washed with PBS before a 1-h preincubation in HEPES-buffered Krebs-Ringer (KRBH; 114 mmol/L NaCl, 4.7 mmol/L KCl, 1.2 mmol/L KH₂PO₄, 20 mmol/L HEPES, 2.5 mmol/L CaCl₂, 0.2% BSA, pH 7.4) supplemented with 2.8 mmol/L glucose. Cells were stimulated with 2.8 or 25 mmol/L glucose-KRBH for 2 h. For cycloheximide (CHX) and EGTA treatment, the preincubation step contained 1.5 µg/mL CHX (60 min) or 5 mmol/L EGTA (10 min). Palmitate was prepared by dissolving palmitic acid in 30 mmol/L NaOH at 70°C, complexing with 20% fatty acid–free BSA (6:1 molar ratio), and dissolution in DMEM. MIN6 cells or mouse islets were cultured overnight and treated with thapsigargin or palmitate as indicated. Groups of 40 overnight-cultured islets were stimulated as described above; 3-h to 24-h stimulations were performed in RPMI. Groups of 50 human islets were handpicked and cultured in Connaught Medical Research Laboratories (5.5 mmol/L glucose) supplemented with pen/strep, glutamine, and 10% FBS. The next morning, islets were treated as above. RNA was extracted or cells were fixed with 4% paraformaldehyde (PFA) for 15 min, permeabilized, and immunostained as described above.

MIN6 cells were plated onto glass coverslips and infected with virus as above. Forty-eight hours after infection, cells were treated for 24 h with thapsigargin. Cells were fixed, and TUNEL-positive cells were identified using the in situ cell death detection kit (Roche). At least 500 cells from each coverslip were counted to assess TUNEL positivity.

MIN6 cells were transfected with 1 μg pFOX-RIPI-Luc (35,36) or pFOX-Luc (promoterless) and 100 ng Npas4-pcDNA3.1 or pcDNA3.1 and 50 ng pCMV-Renilla using Lipofectamine 2000 (LifeTech). After 48 h, cells were lysed with passive lysis buffer (PLB), and luciferase activity was quantified with the Dual Luciferase Reporter assay (Promega) on a SpectraMaxL luminometer (Molecular Devices).

After a 1-h preincubation in KRBH, cells were stimulated for 2 h and KRBH was collected and centrifuged at 2,500g for 10 min to remove cellular debris. Supernatants were analyzed with a mouse insulin ELISA (Alpco). Total insulin was determined by acid-ethanol extraction and normalized to total protein content. For cAMP assays, cells were washed once with PBS and stimulated in 0.25 mmol/L isobutylmethylxanthine with or without exendin-4 or forskolin (10 μmol/L) for 30 min in KRBH. Samples were analyzed with a cAMP screen assay kit (LifeTech), and luminescence was detected as above.

Chromatin immunoprecipitation (ChIP) assays were performed as previously described (37,38). Cells were washed briefly in PBS and then incubated in growth media with or without 40 mmol/L KCl. Cells were then fixed by the addition of 1% formaldehyde in cross-linking buffer (0.1 mol/L NaCl, 1 mmol/L EDTA, 0.5 mmol/L EGTA, 25 mmol/L HEPES-KOH, pH 8.0) for 10 min at room temperature with gentle agitation. Fixation was stopped with the addition of glycine to a final concentration of 125 mmol/L, and incubation continued for 5 min with gentle agitation. Cells were washed three times with ice-cold PBS containing protease inhibitors (Complete EDTA-Free; Roche) and then collected by scraping. Cells were pelleted by centrifugation at 2,000g for 10 min at 4°C (PBS decanted), snap frozen in liquid nitrogen, and stored at −80°C. Cells were lysed in 20 cell pellet volumes of L1 buffer (50 mmol/L HEPES-KOH, pH 7.5, 140 mmol/L NaCl, 1 mmol/L EDTA, 10% glycerol, 0.5% NP-40, 0.25% Triton X-100, protease inhibitors) for 10 min at 4°C. The nuclei were pelleted by centrifugation at 3,000g for 10 min at 4°C and then rinsed with 20 cell pellet volumes of L2 (200 mmol/L NaCl, 1 mmol/L EDTA, 0.5 mmol/L EGTA, 10 mmol/L Tris-HCl, pH 8.0, protease inhibitors) for 10 min at room temperature and then repelleted as above. Nuclei were resuspended in five cell pellet volumes of L3 (1 mmol/L EDTA, 0.5 mmol/L EGTA, 10 mmol/L Tris-HCl, pH 8.0, protease inhibitors) and sonicated in a Misonix Cup Horn sonicator at 80% power for 24 cycles, 30-s cycles on ice, which resulted in genomic fragments of 200–1 kb in size. Insoluble material was removed by centrifugation at 20,000g for 10 min at 4°C. The nuclear lysate was then adjusted to 1 mL by adding L3 buffer supplemented with 0.3 mol/L NaCl, 1% Triton X-100, and 0.1% deoxycholate. The lysate was then precleared by adding 50 μL of prerinsed sheep anti-rabbit Dynabeads (LifeTech) and incubating at 4°C for 1 h with agitation. After preclearing, 5% of the ChIP sample was set aside as input sample, and the remainder of the lysate (950 μL) was incubated with 4 μg of antibody overnight at 4°C. The next day, 40 μL of prerinsed sheep anti-rabbit Dynabeads was added to the immunoprecipitates, and immunoprecipitation was carried out over a second night at 4°C. The bound beads were then washed twice with each of the following: low-salt buffer (150 mmol/L NaCl, 0.1% SDS, 1% Triton X-100, 2 mmol/L EDTA, 20 mmol/L Tris-HCl, pH 8.1), high-salt buffer (500 mmol/L NaCl, 0.1% SDS, 1% Triton X-100, 2 mmol/L EDTA, 20 mmol/L Tris-HCl, pH 8.1), LiCl buffer (0.25 mol/L LiCl, 1% NP-40, 1% deoxycholate, 1 mmol/L EDTA, 10 mmol/L Tris-HCl, pH 8.1), and TE (1 mmol/L EDTA, 10 mmol/L Tris-HCl, pH 8.0). All washes were carried out for 5 min at 4°C. Immunoprecipitated materials were eluted from the beads by adding 200 μL of elution buffer (1% SDS, 0.1 mol/L NaHCO3) and incubating the sample at 65°C with continuous vortexing for 1 h; 150 μL of elution buffer was added to the input material and processed in parallel. Cross-linking was reversed by overnight incubation at 65°C, and resultant DNA was phenol-chloroform purified, ethanol precipitated, and resuspended in 50 μL TE. One microliter was used for standard Taqman analyses.

Small interfering RNAs (siRNAs; 100 nmol/L) (SMARTpool, ON-TARGETplus no. L-054722; Thermo Scientific) were transfected into MIN6 cells 24 h after plating using HiPerFect (Qiagen) as described in the manufacturer’s protocol. Experiments were performed 48 h posttransfection.

Prism 5 (GraphPad Software) was used for statistical analysis. Student t test and one-way ANOVA with Dunnett test were used, with P values ≤0.05 considered significant.

Npas4 was initially detected during full-transcriptome analyses of mouse pancreatic islets (39) and is highly expressed compared with other bHLH-PAS domain factors (Fig. 1A). Real-time PCR was used to determine when Npas4 is expressed during β-cell development. As depicted in Fig. 1B, Npas4 is dramatically induced just prior to birth. Immunofluorescence analyses of adult pancreatic sections demonstrated that Npas4 is expressed in both α- and β-cells but not in the exocrine tissue (Fig. 1C and D and Supplementary Fig. 1). We noted significant heterogeneity in both the intensity of Npas4 staining between islets and the amount of nuclear staining between islets, suggesting that Npas4 is dynamically regulated.

To understand the dynamics of Npas4 expression in β-cells, MIN6 cells were depolarized with 40 mmol/L potassium chloride (KCl). This glucose-mimicking stimulus caused significant induction, 5- and 50-fold at 20 min and 2 h, respectively, in Npas4 (Fig. 2A and G). Elevated glucose was unable to significantly increase Npas4 expression in MIN6 cells after a 12-h preincubation in 5.5 mmol/L glucose (Fig. 2B and G). Depolarization-induced Npas4 induction proved calcium dependent as the calcium chelator EGTA prevented induction (Fig. 2C and G). Induction at the message level was not impacted by blocking protein synthesis, suggesting that Npas4 is an IEG (Fig. 2D).

Although glucose did not regulate Npas4 in MIN6 cells, it dose-dependently increased Npas4 expression in mouse and human islets with 15- and 40-fold respective increases over basal at 25 mmol/L glucose (Fig. 2E, H, and J). Glucose-induced Npas4 expression was temporary, returning to baseline levels by 6 h (Fig. 2F). Npas4 protein was undetectable under basal conditions, but it was significantly induced by depolarization of both islets and MIN6 cells (compare Fig. 2G and H). As we noted in both cytoplasmic and nuclear staining from sections of mouse pancreas (Fig. 1C and D), we depolarized MIN6 with KCl and noted markedly increased nuclear Npas4 immunostaining (Fig. 2I).

To confirm these in vitro findings, we used a 72-h in vivo perfusion system. As previously described (29), the jugular vein and carotid artery of 6-month-old Wistar rats were cannulated under general anesthesia and, after a 5-day recovery, infused with saline, 70% glucose, or a 20% lipid emulsion for 72 h. RNA was extracted from freshly isolated islets, cDNA was prepared, and qPCR was performed. Npas4 expression was significantly elevated in the islets of rats perfused with glucose (Fig. 3A), and a trend toward significance was also observed in intralipid-infused animals. Concomitant with the glucose-stimulated induction of Npas4, we noted a significant reduction in Ins2 expression (Fig. 3B). To understand if Npas4 is important in diabetes disease progression, we examined Npas4 expression in an obese transgenic mouse model of type 2 diabetes expressing human IAPP under the rat insulin promoter (RIP) in addition to the dominant agouti viable yellow mutation (A^vy) (40). Compared with control animals (A^vy/A), there was markedly higher expression of Npas4 in the transgenic mice at 14 weeks of age (Fig. 3C). By 24 weeks of age, however, there was no appreciable difference observed. In support of these data, Npas4 is upregulated in islets of diabetic mice profiled by the Attie laboratory (http://diabetes.wisc.edu/kelleretal2008/search.php) (41).

To address whether the in vivo reduction of insulin 2 expression (Fig. 3B) could be attributed to an increased expression of Npas4, an adenovirus was created to drive Npas4 expression (Ad-Npas4). When compared with KCl-stimulated MIN6 cells infected with a control adenovirus (Ad-Cerulean, variant of GFP), Ad-Npas4 caused an ∼40-fold increase in Npas4 protein levels (Supplementary Fig. 2A). Ad-Npas4–transduced MIN6 cells showed significantly reduced insulin 1 (38% reduction) and insulin 2 (31% reduction) message levels (Fig. 4A and B) when normalized to β-glucuronidase expression, an appropriate reference gene for these experiments (Supplementary Fig. 2B–F). Insulin content was also reduced by 57% (Fig. 4C). Similar results were observed in mouse islets (Fig. 4D–F). To determine if endogenous Npas4 tempers insulin expression, we used siRNAs to knock down Npas4 in MIN6 cells; with a 40% reduction in Npas4 protein/mRNA, we observed increases in both Ins1 and Ins2 expression (Fig. 4G and H).

To determine if the negative regulation of insulin by Npas4 was due to decreased transcription, we assessed the ability of Npas4 to regulate expression of RIP1-driven luciferase transcription (35,36). Although there was high luciferase activity in transfected MIN6 cells, cotransfection with Npas4 expression plasmid attenuated luciferase activity to background levels without impacting control, Renilla luciferase activity (Fig. 4I). As the decreased promoter activity could be due to either direct or indirect effects, we measured the expression of the three main activators of insulin expression in β-cells: Pdx-1, NeuroD1, and MafA. Although Npas4 overexpression did not alter the mRNA levels of Pdx-1 or NeuroD1 (Fig. 4J and K), both were significantly reduced at the protein level (Fig. 4M–P). Interestingly, there were significant increases in both MafA mRNA and protein levels (Fig. 4L, Q, and R). To test whether Npas4 also had direct effects on insulin promoter activity, we also measured Npas4 enrichment at the insulin promoters with ChIP (38). After a 2-h depolarization with KCl, Npas4 was enriched 11-fold at the insulin 1 (Fig. 4S and T) promoter and 15-fold at the insulin 2 promoter (Fig. 4S and T) compared with nondepolarized cells.

Based on the Npas4-induced reduction in insulin content, we hypothesized that there would be defects in insulin secretion. However, when Ad-Npas4–transduced MIN6 cells were stimulated with 16 mmol/L glucose, there was no significant difference in secretion compared with control-infected cells (Fig. 5A and B and Supplementary Fig. 3). In order to understand if Npas4 could modulate incretin-potentiated secretion, we stimulated cells with the GLP-1 agonist exendin-4. In control conditions, exendin-4 stimulated maximal insulin secretion at 5 nmol/L; however, in Ad-Npas4-transduced MIN6 cells, exendin-4 had no potentiating effects (Fig. 5A). Bypassing the GLP-1 receptor and stimulating cAMP production directly with a combination of forskolin and isobutylmethylxanthine (IBMX), which activate adenylyl cyclase and inhibit phosphodiesterase, respectively, was able to partially stimulate secretion in Ad-Npas4–treated cells (Fig. 5A). In islets, a significant reduction in exendin-4–potentiated secretion was also observed (Fig. 5B). In order to determine if this effect was due to alterations in GLP-1 receptor signaling, we assessed exendin-4–stimulated cAMP production. Ad-Npas4 transduction of MIN6 cells significantly reduced both exendin-4–stimulated (Fig. 5C) and forskolin-stimulated (not shown) cAMP production and shifted the EC₅₀ value of exendin-4 from 0.76 to 2.65 nmol/L. These data demonstrate that Npas4 regulates the GLP-1 receptor directly but likely plays other roles in modulating cAMP production.

Based on the ability of Npas4 to inhibit incretin-stimulated cAMP production, we hypothesized that it might induce factors that impair normal receptor–G-protein coupling and examined the GTPase-activating protein Rgs2, which has been previously demonstrated to be a negative regulator of incretin-mediated cAMP production (42). To test if Npas4 regulated Rgs2, we measured expression of Rgs2 after cell depolarization. Under these conditions, Rgs2 was induced in a time-dependent manner (Fig. 6A) with delayed kinetics compared with those for Npas4 (compare Fig. 2A and Fig. 6A). Regulation of Rgs2 message and protein levels also appeared to be quite similar to Npas4 (compare Fig. 2 and Fig. 6B and C).

In order to test if Rgs2 was a target of Npas4, we transduced cells with Ad-Npas4 and observed 16- and 11-fold increases in Rgs2 mRNA in MIN6 cells and mouse islets, respectively (Fig. 6D and E). We also observed that Ad-Npas4 increased Rgs2 protein expression in MIN6 cells (Fig. 6F). Hypothesizing that Npas4 was responsible for the increased Rgs2 expression under depolarizing conditions, we knocked down Npas4 in MIN6 cells and stimulated with KCl and noted a significantly reduced induction of Rgs2 (Fig. 6G). To assess whether this was a direct effect, we performed ChIP assays after a 2-h depolarization and noted 3- and 22-fold enrichments of Npas4 in intron 1 and in an enhancer (43) of the Rgs2 gene, respectively (Fig. 6H and I). Future studies will address how Rgs2 acts to impact GLP-1 receptor function in β-cells.

As insulin gene expression comprises ∼48% of polyadenylated transcription in the mouse β-cell (unpublished RNA-Seq data), and attenuation of transcription (44) and translation (45) are part of the UPR, we reasoned that the reduction in insulin caused by Npas4 would be advantageous during periods of ER stress.

First, to determine if ER stressors induced Npas4, we treated MIN6 cells with the SERCA pump inhibitor thapsigargin. A 24-h exposure significantly increased Npas4 expression (Fig. 7A). Elevated Npas4 message was observed throughout the 24-h exposure to 1 µmol/L thapsigargin (Fig. 7B); however, increased protein expression was only detectable at 24 h (Fig. 7C). To test a relevant physiological ER stressor, we exposed MIN6 cells to palmitate (46,47). We observed a dose-dependent increase in Npas4 expression that peaked at 1 h (15-fold over control with 500 µmol/L palmitate) but was maintained for at least 24 h (Fig. 7D). Similar results were noted at the protein level, with palmitate exposure leading to robust induction at 2 h (Fig. 7E). Npas4 mRNA levels were also elevated in mouse islets in response to thapsigargin (Fig. 7F) and palmitate (Fig. 7G).

As Npas4 is positively regulated by ER stressors in β-cells, we questioned whether increased Npas4 expression offered protection from ER stress. As previously demonstrated (46–48), both thapsigargin and palmitate induced the expression of the proapoptotic transcription factor Ddit3 (i.e., CHOP/GADD153) (Fig. 8A and B). However, Ad-Npas4 transduction reduced Ddit3 significantly in MIN6 cells (Fig. 8A and B). Ad-Npas4 also resulted in a 53% reduction in Ddit3 protein levels in thapsigargin-treated MIN6 cells (Fig. 8C and D). Knockdown of Npas4 did not change basal Ddit3 expression; although after palmitate exposure, Ddit3 was significantly increased upon Npas4 knockdown (Fig. 8E). Furthermore, Ad-Npas4 reduced thapsigargin- and palmitate-driven expression of ATF4 and Xbp-1 (Supplementary Fig. 4) and increased expression of the cytoprotective factors Wfs-1 and Hspa5 (Fig. 8F and G). In isolated mouse islets, Npas4 transduction reduced Ddit3 expression induced by both thapsigargin and palmitate (Fig. 8H and I). As Ddit3 is a proapoptotic transcription factor and a terminal step in the ER stress pathway, we next determined if there were any alterations in β-cell apoptosis. Ad-Npas4 significantly reduced the number of TUNEL⁺ apoptotic MIN6 cells to 68.4 and 55% of control levels with 1 and 10 µmol/L thapsigargin, respectively (Fig. 8J).

Although previously considered to be brain specific, we demonstrate that Npas4 is expressed in β-cells and induced after glucose-mediated depolarization in a calcium-dependent manner. Further, we have demonstrated that Npas4 is a novel factor capable of directly and indirectly inhibiting insulin promoter activity and reducing insulin content without significantly affecting glucose-stimulated secretion. Npas4 appears to reduce β-cell stress and prevent β-cell death by blunting incretin-stimulated cAMP production and potentiation of insulin secretion and by counteracting ER stress.

Under normal physiological conditions, elevation in blood glucose levels would stimulate β-cell insulin production and secretion and subsequent rapid (<2 h) normalization of blood glucose to basal levels. In these circumstances, we expect that Npas4 may be induced to efficiently couple insulin supply and demand and reduce the likelihood that insulin biosynthesis places undue stress on the cells (49). As demonstrated here, a reduction in incretin-potentiated secretion and insulin production could help mediate these effects. Interestingly, we did not observe a significant detriment in glucose-stimulated insulin secretion when we drove Npas4 expression in β-cells, supporting the concept that induction of Npas4 likely improves the “fuel efficiency” of β-cells.

Improvement of β-cell function through reducing insulin demand has long been an attractive approach for treating diabetes. Indeed, work published in 1976 in the Lancet demonstrated that a 7-day administration of the K_ATP channel opener diazoxide to people with type 2 diabetes improved the posttreatment insulin secretory response (50). More recently, it has been demonstrated that in people newly diagnosed with type 2 diabetes, a 2-year treatment course of insulin, but not glibenclamide, improved glycemic outcomes (51). In addition, costimulation of murine β-cell lines with diazoxide and palmitate attenuated the following: the palmitate-induced reduction in insulin secretion, the palmitate-induced ER stress, and subsequent apoptosis (52). In order to understand how the immediate early response might help maintain β-cell function, future studies should be targeted at understanding whether loss of Npas4, or other IEGs, increases β-cell susceptibility to stress in vivo.

Prolonged exposure to glucose resulted in a desensitization of the Npas4 response and a return to baseline levels, a characteristic common to activity-regulated IEGs (17,20,26,53). Under normal physiological conditions, the β-cell would not endure high glucose levels for >6 h, and it is likely that Npas4 target genes would be maximally activated under such conditions. It is possible that the gluco(lipo)toxicity/dysfunction that has previously been reported to cause β-cell dysfunction during the development of type 2 diabetes is actually a homeostatic mechanism that acts to protect β-cells. Furthermore, this response may be initiated quite early on with expression of IEGs such as Npas4.

Studies using multiple rodent models of type 2 diabetes suggest that β-cell failure is driven by ER stress–induced dysfunction and death (54). In this manuscript, we demonstrate that Npas4 is induced through exposure to calcium-mobilizing β-cell ER stressors such as thapsigargin and palmitate. We reasoned that because Npas4 can reduce insulin biosynthesis, it might be able to protect against ER stress without significant detrimental physiological effects.

Ddit3 is a transcription factor that is induced by ER stress through prolonged activation of the protein kinase RNA-like endoplasmic reticulum kinase (PERK) arm of the UPR (49). Ddit3 induces apoptosis by driving expression of proapoptotic genes (e.g., Bim) and reducing the expression of antiapoptotic genes (55). The induction of Ddit3 has been demonstrated to be a key step in the development of type 2 diabetes as its genetic deletion protects from diabetes in both the db/db mouse and the high-fat diet/streptozotocin-treated mouse (56). Furthermore, a recent study suggested that the GLP-1 agonist liraglutide reduced β-cell ER stress (57); it is currently unclear whether this is a direct or secondary effect of GLP-1 signaling as an immediate effect was not observed. Additionally, it is unclear whether downregulation of β-cell GLP-1 receptor signaling after long-term stimulation might help mediate this effect.

Overexpression of Npas4 reduced the induction of Ddit3 and Xbp1 and the splicing of Xbp1 and reduced β-cell death in response to ER stress. Although we did not directly measure Atf6 expression, it is likely that this arm of the UPR is also affected by Npas4 as we observed significant induction of Xbp, the chaperone Hspa5, and Wfs-1, which may not be primarily induced by the other arms (49). MafA was induced by Npas4 and likely contributes to cytoprotection, as MafA directly inhibits Ddit3 promoter activity in β-cells (58). Interestingly, another novel cytoprotective factor (ARC) was recently described that also acts by potently suppressing Ddit3 expression in β-cells (59). It will be interesting to determine if both ARC and MafA are direct Npas4 targets.

Current treatment strategies for type 2 diabetes are targeted toward driving β-cell activity in the face of already elevated cellular stress. Although this strategy has produced drugs that increase insulin secretion and β-cell function over the short-term, recent research has suggested that this approach may exacerbate β-cell failure over the long-term. For example, Talchai et al. (11) demonstrated that β-cells in models of diabetes lose their identity after chronic stress and dedifferentiate. In conclusion, we propose that a complimentary therapeutic strategy to those currently used for type 2 diabetes would be to reduce ER stress and β-cell death while maintaining β-cell function. We posit that Npas4 may offer a therapeutic means to accomplish this goal.

This work was supported by grants to F.C.L. from the Juvenile Diabetes Research Foundation (2-2011-91) and the Canadian Institutes of Health Research (CIHR) (MOP 102628) and to V.P. from the U.S. National Institutes of Health (R01-DK-58096). V.P. holds the Canada Research Chair in Diabetes and Pancreatic Beta-Cell Function. The Michael Smith Foundation for Health Research, the Canadian Diabetes Association, and the Child and Family Research Institute (Vancouver, British Columbia) provided salary support for F.C.L. P.V.S. and N.A.J.K. were recipients of the CIHR-BC Transplantation Trainee Fellowship, and R.A.U. was a recipient of the Child and Family Research Institute (Vancouver, British Columbia) undergraduate studentship. C.Y.W.-R. was supported by a CIHR-Vanier Canada Graduate Scholarship.

B.Z. received postdoctoral fellowship support from Merck Frosst and Eli Lilly. No other potential conflicts of interest relevant to this article were reported.

P.V.S. performed experiments and wrote the manuscript. N.A.J.K., B.Z., C.Y.W.-R., C.N., and R.A.U. performed experiments. A.M.J.S. and V.P. provided reagents. F.C.L. designed, directed, and performed experiments and wrote the manuscript. All authors subsequently edited the manuscript. F.C.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors thank Dr. Bruce Verchere, University of British Columbia, Vancouver, British Columbia, Canada; Dr. Michael S. German, University of California, San Francisco, San Francisco, California; Dr. Stuart B. Smith, University of California, San Francisco, San Francisco, California; and Dr. Michael E. Greenberg, Harvard University, Cambridge, Massachusetts, for their generous contributions of reagents that made this work possible.