Anks4b, a Novel Target of HNF4α Protein, Interacts with GRP78 Protein and Regulates Endoplasmic Reticulum Stress-induced Apoptosis in Pancreatic β-Cells*

Background: Target genes of HNF4α in β-cells are largely unknown. Results: Expression of Anks4b is decreased in the βHNF4α KO islets. HNF4α activates Anks4b promoter activity. Anks4b binds to GRP78 and regulates sensitivity to ER stress. Conclusion: HNF4α novel target gene, Anks4b, regulates the susceptibility of β-cells to ER stress. Significance: Anks4b is a novel molecule involved in ER stress. Mutations of the HNF4A gene cause a form of maturity-onset diabetes of the young (MODY1) that is characterized by impairment of pancreatic β-cell function. HNF4α is a transcription factor belonging to the nuclear receptor superfamily (NR2A1), but its target genes in pancreatic β-cells are largely unknown. Here, we report that ankyrin repeat and sterile α motif domain containing 4b (Anks4b) is a target of HNF4α in pancreatic β-cells. Expression of Anks4b was decreased in both βHNF4α KO islets and HNF4α knockdown MIN6 β-cells, and HNF4α activated Anks4b promoter activity. Anks4b bound to glucose-regulated protein 78 (GRP78), a major endoplasmic reticulum (ER) chaperone protein, and overexpression of Anks4b enhanced the ER stress response and ER stress-associated apoptosis of MIN6 cells. Conversely, suppression of Anks4b reduced β-cell susceptibility to ER stress-induced apoptosis. These results indicate that Anks4b is a HNF4α target gene that regulates ER stress in β-cells by interacting with GRP78, thus suggesting that HNF4α is involved in maintenance of the ER.

Mutations of the HNF4A gene cause a form of maturityonset diabetes of the young (MODY1) that is characterized by impairment of pancreatic ␤-cell function. HNF4␣ is a transcription factor belonging to the nuclear receptor superfamily (NR2A1), but its target genes in pancreatic ␤-cells are largely unknown. Here, we report that ankyrin repeat and sterile ␣ motif domain containing 4b (Anks4b) is a target of HNF4␣ in pancreatic ␤-cells. Expression of Anks4b was decreased in both ␤HNF4␣ KO islets and HNF4␣ knockdown MIN6 ␤-cells, and HNF4␣ activated Anks4b promoter activity. Anks4b bound to glucose-regulated protein 78 (GRP78), a major endoplasmic reticulum (ER) chaperone protein, and overexpression of Anks4b enhanced the ER stress response and ER stress-associated apoptosis of MIN6 cells. Conversely, suppression of Anks4b reduced ␤-cell susceptibility to ER stress-induced apoptosis. These results indicate that Anks4b is a HNF4␣ target gene that regulates ER stress in ␤-cells by interacting with GRP78, thus suggesting that HNF4␣ is involved in maintenance of the ER.
Hepatocyte nuclear factor (HNF) 4 4␣, a transcription factor belonging to the nuclear receptor superfamily (NR2A1), is expressed in the liver, pancreas, kidney, and intestine (1,2). HNF4␣ has multiple functional domains, including the N-terminal A/B domain associated with the transactivation domain (AF-1), a DNA binding C domain, a functionally complex E domain that forms a ligand binding domain, a dimerization interface and transactivation domain (AF-2), and an F domain with a negative regulatory function (3,4). HNF4␣ predominantly binds to a 6-bp repeat (AGGTCA) with a 1-bp spacer (mainly A) called direct repeat (DR1).
Maturity-onset diabetes of the young (MODY) is a genetically heterogeneous monogenic disorder that accounts for 2-5% of type 2 diabetes (5). We discovered that mutations of the human HNF4A gene cause a particular form of MODY known as MODY1 (6). The primary pathogenesis of MODY1 involves dysfunction of pancreatic ␤-cells (5). In addition, it has been shown that targeted disruption of HNF4␣ in pancreatic ␤-cells leads to defective insulin secretion in mice (7,8). These findings have demonstrated that HNF4␣ has an important role in ␤-cells.
In the liver, HNF4␣ plays a critical role in nutrient transport and metabolism by regulating numerous target genes, including phosphoenolpyruvate carboxykinase (PCK1), glucose-6phosphatase (G6PC), apolipoprotein AII (APOA2), and microsomal triglyceride transfer protein (MTTP) (9,10). In contrast, we have little information about the target genes of HNF4␣ in pancreatic ␤-cells. Previous in vitro studies have suggested that HNF4␣ regulates the expression of pancreatic ␤-cell genes involved in glucose metabolism, such as insulin (INS), solute carrier family 2 (SLC2A2), and HNF1A (11). However, the expression of these genes was unchanged in the islets of ␤-cellspecific HNF4␣ knock-out (␤HNF4␣ KO) mice (7,8), indicating that such genes are not targets of HNF4␣ in vivo, at least in ␤-cells.
In the present study, we investigated the mRNA expression profile of ␤HNF4␣ KO mice and found that ankyrin repeat and sterile ␣ motif domain containing 4b/harmonin-interacting, ankyrin repeat-containing protein (Anks4b/Harp) is a target of HNF4␣ in ␤-cells. We also demonstrated that Anks4b interacts with glucose-regulated protein 78 (GRP78), a major chaperone protein that protects cells from endoplasmic reticulum (ER) stress in vitro and in vivo. Gain-and loss-of-function studies of Anks4b revealed that it regulates sensitivity to thapsigargin (TG)-induced ER stress and apoptosis in MIN6 ␤-cell line. Our results suggest that HNF4␣ plays an important role in the regulation of ER stress and apoptosis in pancreatic ␤-cells.

Microarray Expression Profiling and HNF4␣ Motif Scan-
Mice were maintained on a 12-h light/12-h dark cycle and allowed free access to food and water. All animal experiments were conducted according to the guidelines of the Institutional Animal Committee of Kumamoto University. Pancreatic islets were isolated from 45-week-old female ␤HNF4␣ KO mice (n ϭ 5) and control flox/flox mice (n ϭ 5) by collagenase digestion (12). Total RNA was prepared from the isolated islets with an RNeasy micro kit (Qiagen) according to the manufacturer's instructions, and its quality was confirmed by using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). DNA microarray analysis was performed by the Kurabo GeneChip custom analysis service with GeneChip mouse genome 430 2.0 array (Affymetrix Inc., Santa Clara, CA). For identification of potential HNF4␣ binding sites, 5 kb of the promoter sequence upstream of the transcriptional start site was retrieved from the University of California Santa Cruz Genome Browser, and the sequence was analyzed by using the Transcription Element Search System (TESS) and the HNF4 Motif Finder generated by Sladek and colleagues (38).
Quantitative RT-PCR-Total RNA was extracted using an RNeasy micro kit (catalog number 74004, Qiagen, Valencia, CA) or Sepasol-RNA I super reagent (Nacalai Tesque, Kyoto, Japan). Then 1 g of total RNA was used to synthesize firststrand cDNA with a PrimeScript RT reagent kit and gDNA Eraser (RR047A, TaKaRa Bio Inc., Shiga, Japan) according to the manufacturer's instructions. Quantitative real-time PCR was performed using SYBR Premix Ex Taq II (RR820A, TaKaRa) in an ABI 7300 thermal cycler (Applied Biosystems, Foster City, CA). The specific primers employed are shown in supplemental Table 1. Relative expression of each gene was normalized to that of TATA-binding protein.
Proteomic Identification of Anks4b-interacting Proteins-Silver-stained gels were subjected to in-gel digestion followed by extraction of peptides and proteomic analysis by LC-MS/MS. Gel digestion and peptide extraction were performed as reported previously (17). The peptide samples thus obtained were analyzed in an ESI-Q-TOF tandem mass spectrometer (6510; Agilent) with an HPLC chip-MS system, consisting of a nano pump (G2226; Agilent) with a four-channel microvacuum degasser (G1379B; Agilent), a microfluidic chip cube (G4240; Agilent), a capillary pump (G1376A; Agilent) with degasser (G1379B; Agilent), and an autosampler with thermostat (G1377A; Agilent). All modules were controlled by Mass-Hunter software (version B.02.00; Agilent). A microfluidic reverse-phase-HPLC chip (Zorbax 300SB-C18; 5-m particle size, 75-mm inner diameter, and 43 mm in length) was used for separation of peptides. The nano pump was employed to generate gradient nano flow at 600 nl/min, with the mobile phase being 0.1% formic acid in MS-grade water (solvent A) and 0.1% formic acid in acetonitrile (solvent B). The gradient was 5-75% solvent B over 9 min. A capillary pump was used to load samples with a mobile phase of 0.1% formic acid at 4 l/min. The Agilent ESI-Q-TOF was operated in the positive ionization mode (ESIϩ), with an ionization voltage of 1,850 V and a fragmentor voltage of 175 V at 300°C. Fragmentation of protonated molecular ions was conducted in the auto-MS/MS mode, starting with a collision energy voltage of 3 V that was increased by 3.7 V per 100 Da. The selected m/z ranges were 300 -2,400 Da in the MS mode and 59 -3,000 Da in the MS/MS mode. The data output consisted of one full mass spectrum (with three fragmentation patterns per spectrum) every 250 ms. The three highest peaks of each MS spectrum were selected for fragmentation. Mass lists were created in the form of Mascot generic files and were used as the input for Mascot MS/MS ion searches of the National Center for Biotechnology Information nonredundant (NCBI nr) database using the Matrix Science Web server Mascot version 2.2. Default search parameters were as follows: enzyme, trypsin; maximum missed cleavage, 1; variable modifications, carbamidomethyl (Cys); peptide tolerance, Ϯ 1.2 Da; MS/MS tolerance, Ϯ 0.6 Da; peptide charge, 2ϩ and 3ϩ; instrument, ESI-Q-TOF. For positive identification, the result of (Ϫ10 ϫ log (p)) could not exceed the significance threshold (p Ͻ 0.05).
Immunoprecipitation-Mouse Anks4b cDNA was amplified by PCR using a pair of primers (5Ј-CGGATCCCCATGTC-TACCCGCTATCACCAA-3Ј and 5Ј-CGGATCCTTAGAG-GCTGGTGTCAACCAACT-3Ј) and was subcloned in-frame into the pcDNA3-HA and pcDNA3-FLAG expression vectors. The GRP78 expression vector (pCMV-BiP-Myc-KDEL-wt) was a gift from Dr. Ron Prywes (Addgene plasmid 27164). After transfection into COS-7 cells, the cells were lysed in immunoprecipitation buffer (20 mM Tris-HCl (pH 7.4), 175 mM NaCl, 2.5 mM MgCl 2 , 0.05% Nonidet P-40, 1 mM PMSF, and protease inhibitor mixture (Nacalai Tesque)) and incubated on ice for 30 min. Then 700 g of cell lysate and FLAG tag antibody beads (Wako) were mixed and stirred at 4°C for 18 h. After washing with immunoprecipitation buffer, proteins were eluted by using DYKDDDDK peptide (Wako). A sample of the eluate and 2% of the cell lysate (from before processing) were subjected to Western blotting analysis.
Flow Cytometric Analysis-An annexin V-FITC apoptosis detection kit (BioVision Research Products, Mountain View, CA) was used for the apoptosis assay according to the manufacturer's instructions. MIN6 cells were cultured in DMEM for 30 h with or without 1 M thapsigargin (Nacalai Tesque). After incubation in trypsin/EDTA for 10 min at 37°C, cells were centrifuged at 6,000 rpm for 10 min. The pellet was resuspended in 1ϫresuspension buffer, and the cells were stained with annexin V-FITC antibody. After incubation for 5 min at room temperature in the dark, stained cells were analyzed using a FACSCalibur (BD Biosciences) and FlowJo software (Tomy Digital Biology, Tokyo, Japan).
Statistical Analysis-Statistical analyses were performed using Statview J-5.0 software (SAS Institute, Cary, NC). The significance of differences was assessed with the unpaired t test, and p Ͻ 0.05 was considered to indicate statistical significance.

RESULTS
Anks4b Is a Novel Target of HNF4␣-To identify target genes of HNF4␣ in pancreatic ␤-cells, DNA microarray analysis was performed using islets from ␤HNF4␣ KO mice and control mice. Body weight and blood glucose levels were similar for these two strains of mice (body weight was 32.7 Ϯ 1.7 g (n ϭ 5) versus 34.1 Ϯ 1.9 g (n ϭ 5) and random blood glucose was 128 Ϯ 27 mg/dl (n ϭ 5) versus 114 Ϯ 24 mg/dl (n ϭ 5) for ␤HNF4␣ KO versus control mice). Microarray analysis identified 56 up-regulated genes (signal log ratio Ն2) and 100 down-regulated genes (signal log ratio ՅϪ1.5) in ␤HNF4␣ KO islets (supplemental Table 2). Expression of the majority of the genes known to be involved in glucose metabolism was unchanged. To validate these results, expression of mRNA for genes randomly chosen from both the down-regulated and the up-regulated groups was assessed by quantitative real-time PCR in an independent group of 12-week-old male mice. As a result, differential expression of most genes was confirmed ( Fig. 1A and supplemental Fig. 1). Gupta et al. (19) reported that ST5, a regulator of ERK activation, is a direct target of HNF4␣ in ␤-cells. Expression of ST5 mRNA was reduced by 24.6% in ␤HNF4␣ KO islets (supplemental Fig. 2).
Next, we performed a computational scan of the HNF4␣ binding motif in the down-regulated genes. This identified 22 high affinity HNF4␣ binding sequences in the mouse promoter. In 3 out of 22 genes, the HNF4 motif was also conserved in the corresponding human genome. These three genes encoded Anks4b, guanylate cyclase 2c (Gucy2c), and peroxisome proliferator-activated receptor ␥ coactivator-1␣ (Ppargc1a). Quantitative real-time PCR analysis confirmed a significant decrease of Anks4b expression in the islets of 12-week-old ␤HNF4␣ KO mice (17.3% of the control level, p Ͻ 0.01) (Fig. 1B). In contrast, the reduction of Gucy2c mRNA expression was marginal (21.7% of the control level, p ϭ 0.06), and Ppargc1a mRNA levels were unchanged. The difference of sex and age of mice or different detection systems might have contributed to the different results. To elucidate the direct effect of HNF4␣ on the expression of these three genes, we established MIN6 ␤-cells that stably expressed HNF4␣-specific shRNA (HNF4␣ KD-MIN6) by retroviral infection. Suppression of endogenous HNF4␣ was confirmed at both the mRNA and the protein levels (Fig. 1C). Decreased expression of Anks4b, Gucy2c, and Ppargc1a was found in HNF4␣ KD-MIN6 cells (Fig. 1D). Because Anks4b gene expression was most markedly decreased in both ␤HNF4␣ KO islets and HNF4␣ KD-MIN6 cells (35.2% of the control level, p Ͻ 0.001), we focused on Anks4b for further investigation.
Screening of the promoter region of the mouse Anks4b gene by using a genomic databank revealed an HNF4␣ binding site (nucleotides Ϫ108 to Ϫ120 relative to the translation start codon when A is designated as ϩ1). We cloned a 190-bp promoter region upstream of a luciferase reporter gene and co-expressed it with the HNF4␣ expression vector in HEK293 cells. Induction of HNF4␣7 (an isoform expressed in pancreatic ␤-cells (4)) increased Anks4b promoter activity in a concentration-dependent manner ( Fig. 2A), whereas overexpression of the HNF4␣ mutant lacking AF-2 had no effect (Fig. 2B). When the putative HNF4␣ binding site in the Anks4b promoter was subjected to mutation (H4m), transcriptional activation by HNF4␣7 was significantly reduced by 64.0% (p Ͻ 0.001) (Fig.  2B). Disruption of the HNF4␣ binding site was also associated with a 48.5% reduction of promoter activity in MIN6 cells (p Ͻ 0.001) (Fig. 2C). To assess the binding of HNF4␣ to the Anks4b promoter, a chromatin immunoprecipitation (ChIP) assay was performed using MIN6 cells. This assay revealed binding of HNF4␣ to the Anks4b promoter of MIN6 cells (Fig. 2D). Specific binding of HNF4␣ to the putative binding site was also demonstrated by the electrophoretic mobility shift assay (EMSA) (supplemental Fig. 3). Thus, both in vivo and in vitro data indicated that Anks4b is a direct target of HNF4␣ in ␤-cells.

HNF4␣ and HNF1␣ Synergistically Activate Transcription of
Anks4b-HNF1␣ is a homeodomain-containing transcription factor that is also expressed in the liver, kidney, intestine, and pancreas (20). Mutation of the HNF1A gene causes another type of MODY known as MODY3 (21). In addition to the binding site for HNF4␣, we also found an HNF1␣ binding consensus sequence in the Anks4b promoter (Fig. 3A). Therefore, we examined the role of HNF1␣ in Anks4b gene transcription. First, binding of HNF1␣ to the Anks4b gene was examined by EMSA with MIN6 nuclear extracts and a probe corresponding to the HNF1␣ binding site (Fig. 3B). The probe shifted after the addition of nuclear extracts (lane 2), and its binding was blocked by the addition of a 30-fold excess of unlabeled oligonucleotide (lane 3). Specificity of HNF1␣ binding was assessed by supershifting the DNA-HNF1␣ complex using HNF1␣ antibody (lane 5), indicating that HNF1␣ also binds directly to the Anks4b promoter. To examine the influence of HNF1␣ on Anks4b gene expression, we next performed a reporter gene assay. WT-HNF1␣ caused a dose-dependent increase of Anks4b promoter activity (Fig. 3C). Interestingly, Anks4b mRNA expression was decreased in HNF1␣ KO islets according to the results of DNA microarray analysis (22). Taken together, these results suggested that Anks4b is a target of  Table 1 using flox/flox control (white bar) and ␤HNF4␣ KO islets (black bar, male, 12 week, n ϭ 4). Expression of each gene was normalized for that of TATA-binding protein (TBP). B, expression of Anks4b, Gucy2c, and Ppargc1a in ␤HNF4␣ KO islets. Decreased expression of Anks4b was confirmed by quantitative RT-PCR. C, HNF4␣ mRNA (left) and HNF4␣ protein (right) in control (Ctrl, white bar) and HNF4␣ knockdown MIN6 cells (KD, black bar) were evaluated by quantitative PCR (n ϭ 4) and Western blotting, respectively. ␤-Actin was used as the loading control. D, expression of Anks4b, Gucy2c, and Ppargc1a was significantly decreased in HNF4␣ KD-MIN6 cells. The mean Ϯ S.D. for each group is shown (*, p Ͻ 0.05, **, p Ͻ 0.01, ***, p Ͻ 0.001). HNF1␣ as well as HNF4␣. Because it has been reported that HNF4␣ and HNF1␣ cooperatively activate target genes that have binding sites for both HNFs in the promoter region (23,24), we examined the influence on Anks4b gene expression of interaction between HNF4␣ and HNF1␣. When an Anks4b reporter construct was cotransfected into HEK293 cells with 10 ng of HNF1␣ or HNF4␣ expression plasmid, the reporter gene was activated by 2.2-and 1.7-fold, respectively (Fig. 3D). In contrast, there was a dramatic increase of promoter activity (7.9-fold) when both constructs were cotransfected simultaneously (Fig. 3D). Mutation of either HNF1␣ or HNF4␣ markedly suppressed this response (Fig. 3D). Synergistic activation of Anks4b promoter activity was significantly suppressed by disruption of either the HNF4␣ binding site (H4m) or the HNF1␣ binding site (H1m), and activation was completely abolished by both H4m and H1m (Fig. 3E). Taken together, these results indicate that Anks4b promoter activity is synergistically regulated by both HNF4␣ and HNF1␣.
Anks4b Interacts with GRP78 Both in Vitro and in Vivo-Anks4b is a scaffold protein with three ankyrin repeats and a sterile ␣ motif domain that was identified as harmonin-interacting protein (25), although its function is completely unknown. To elucidate the role of Anks4b in ␤-cells, we searched for molecules that interacted with full-length Anks4b (FL) and with its deletion mutants (N-, M-, and C-Anks4b) (Fig.  4A) by performing a GST pulldown assay of mouse liver lysates ( Fig. 4B and supplemental Fig. 4). We found a protein of ϳ75 kDa that specifically precipitated with GST-M-Anks4b (B6), and it was identified as GRP78/binding immunoglobulin protein (BiP) by mass spectrometry (Fig. 4B). GRP78 is an ER-localized chaperone protein that is induced by the unfolded protein response in response to ER stress (26,27). Binding of GRP78 to GST-FL-Anks4b and GST-M-Anks4b, but not to GST, GST-N-Anks4b, or C-Anks4b, was confirmed by Western blotting using a specific antibody for GRP78 (Fig. 4C), suggesting that GRP78 bound to the middle region of Anks4b. GST pulldown experiments using MIN6 cell lysates also demonstrated binding of GRP78 to Anks4b (Fig. 4D).
Subsequently, we evaluated the interaction between Anks4b and GRP78 in cultured cells. COS-7 cells were transfected with the Myc-GRP78 expression plasmid alone or with Myc-GRP78 plus FLAG-tagged wild-type Anks4b expression plasmids, and cell lysates were immunoprecipitated with FLAG resin. As shown in Fig. 4E, FLAG-Anks4b was able to coimmunoprecipitate GRP78 as well as harmonin, a protein that was previously found to interact with Anks4b (25). These results indicated that Anks4b binds to GRP78 in cells.
Anks4b Colocalizes with GRP78 in the Endoplasmic Reticulum-We next investigated the intracellular localization of Anks4b. HA-tagged Anks4b and Myc-tagged GRP78 constructs were cotransfected into HeLa cells, and an immunofluorescence study was performed. HA staining (Anks4b, red) revealed a reticular pattern in the cytoplasm, but no signals were detected in the nucleus (Fig. 5A). Double staining for Anks4b and GRP78 (Myc, green) as a marker for the ER revealed that both signals were frequently colocalized (Fig. 5, B and C). In contrast, Anks4b staining did not overlap with MitoTracker, a specific marker for the mitochondria (supplemental Fig. 5). A similar staining pattern was also detected in COS-7 cells and MIN6 cells (Fig. 5, D-I). These findings were further evidence that Anks4b interacts with GRP78. Notably, Anks4b staining was detected at the periphery of the ER lumen (Fig. 5, C, F, and I, inset), suggesting that it was localized adjacent to the ER membrane.
Anks4b Regulates Apoptosis in Response to ER Stress-GRP78 is a major chaperone protein that protects cells from ER stress, and overexpression of GRP78 reduces ER stress-mediated apoptosis by attenuating the expression of C/EBP homologous protein (CHOP) (28,29). Accordingly, detection of an interaction between Anks4b and GRP78 prompted us to investigate the role of Anks4b in both ER stress and apoptosis. TG causes ER stress by preventing calcium uptake from the cytoplasm into the ER (30), and treatment of MIN6 cells with 1 M TG for 20 h increased the expression of the ER stress-related genes (ATF4, spliced XBP1, and CHOP) (data not shown). First, we examined the effect of Anks4b overexpression on MIN6 cells (supplemental Fig. 6). Anks4b overexpression did not affect CHOP gene expression in the absence of TG, but TG-induced CHOP expression was significantly increased (1.4-fold, p Ͻ 0.05) (Fig.  6A). TG-induced ATF4 expression was also significantly augmented in Anks4b-overexpressing MIN6 cells (1.3-fold, p Ͻ 0.05) (Fig. 6B). Furthermore, the number of annexin V-positive apoptotic cells was increased by overexpression of Anks4b (1.3fold, p Ͻ 0.001) (Fig. 6C). Augmentation of apoptosis was also observed in MIN6 cells overexpressing HNF4␣7 (supplemental Fig. 7). Activation of caspase-3 mediates the induction of apoptosis downstream of CHOP (31), and activated (cleaved) caspase-3 protein expression was increased when Anks4boverexpessing MIN6 cells were treated with TG (Fig. 6D).
Next, we examined the effect of knockdown of Anks4b in MIN6 cells (supplemental Fig. 8). Suppression of endogenous Anks4b mRNA by shRNA in MIN6 (reduced to 40.5% of the control level) did not affect CHOP gene expression in the absence of TG, but TG-induced CHOP expression was significantly reduced by 32.1% (p Ͻ 0.05) (Fig. 6E). In addition, flow cytometric analysis using annexin V revealed that TG-induced apoptosis was also decreased by suppression of Anks4b (Fig.  6F). Collectively, these findings indicate that Anks4b promotes the induction of ER stress and apoptosis by TG in MIN6 cells.

DISCUSSION
HNF4␣ plays an important role in pancreatic ␤-cells, and mutation of this gene causes MODY1 (6). However, there has been little information available about the target genes of HNF4␣ in ␤-cells. We and others have previously reported that most of the genes involved in glucose metabolism, including Slc2a2, Gck, Kcnj11, Abcc8, and Ins, are not differentially expressed in ␤HNF4␣ KO islets (7,8,19). The present large scale expression profiling analysis also demonstrated that expression of genes known to be involved in insulin secretion was largely unchanged in HNF4␣ deficient islets. Like HNF4␣, mutation of the HNF1␣ gene also causes a form of MODY (MODY3), which is characterized by ␤-cell dysfunction (21). Expression of many genes involved in insulin secretion, including Slc2a2, Pklr, and Tmem27, is decreased in HNF1␣ KO islets (22,32,33). Thus, the gene expression pattern of HNF4␣ KO islets differs markedly from that of HNF1␣ KO islets.
In the present study, we found that Anks4b gene expression was markedly reduced in both ␤HNF4␣ KO islets and HNF4␣ KD-MIN6 cells. Reporter gene assays and ChIP analysis demonstrated that HNF4␣ bound to a conserved HNF4 binding motif and activated transcription, thus indicating that Anks4b is a direct target of HNF4␣ in ␤-cells. In addition to the pancreatic islets, Anks4b is also expressed in the liver, kidney, small intestine, and colon (25). This distribution of expression is very similar to that of HNF4␣, suggesting that HNF4␣ plays a role in Anks4b gene transcription in these tissues. Furthermore, we found that Anks4b gene expression was also regulated by HNF1␣. Cotransfection of HNF4␣ and HNF1␣ dramatically stimulated promoter activity when compared with the sum of the effects of each transcription factor acting separately (Fig.  3D). Recently, Boj et al. (34) reported that HNF4␣ and HNF1␣ regulate common target genes through interdependent regulatory mechanisms. Although the mechanism of the functional interaction between HNF4␣ and HNF1␣ is still unclear, our results indicate that Anks4b gene expression is another example of such interdependent regulation.
Anks4b was originally identified as harmonin (the gene responsible for Usher deafness syndrome type 1C)-interacting protein, but its function is unknown. In this study, we showed that Anks4b binds to GRP78, a major ER chaperone protein.
We also found that Anks4b knockdown significantly inhibited TG-induced CHOP expression and apoptosis in MIN6 cells, whereas Anks4b overexpression enhanced TG-induced CHOP expression and apoptosis, strongly suggesting a direct role of Anks4b in increasing the susceptibility of ␤-cells to ER stress and apoptosis. Investigation of Anks4b knock-out mice will improve our understanding of the role of this molecule in ER stress. Anks4b does not possess the canonical ER localization signal (35), so the molecular mechanism by which Anks4b binds to GRP78 and regulates ER stress warrants further investigation.
HNF4␣ plays an important role in a number of metabolic pathways, including those for gluconeogenesis, ureagenesis, fatty acid metabolism, and drug metabolism (36 -38). Our finding that Anks4b is a target of HNF4␣ uncovers a new role for this transcription factor in regulating ␤-cell susceptibility to ER stress. ER stress is associated with ␤-cell apoptosis in common type 2 diabetes (39). Because reduced expression of Anks4b was associated with a decrease, rather than an increase, of ER stress and apoptosis, the significance of Anks4b in relation to the occurrence of MODY is unclear. However, recent genetic studies have shown that HNF4␣ has dual opposing roles in the ␤-cell during different periods of life. Although HNF4␣ deficiency results in diabetes in young adults (6), the same genetic defect occasionally causes hyperinsulinemic hypoglycemia at birth (40,41). Further studies will need to address whether reduced Anks4b expression is responsible for the hypersecretion of ␤-cells early in life.
In conclusion, we identified Anks4b as a novel molecule that controls the susceptibility to ER stress-induced apoptosis. The ER is critical for the normal functioning of pancreatic ␤-cells, and ER stress-associated apoptosis is often a contributory factor to ␤-cell death in type 2 diabetes (39). Therefore, Anks4b may be a potential target for the treatment of diabetes associated with ER stress.