HRD1, an Important Player in Pancreatic β-Cell Failure and Therapeutic Target for Type 2 Diabetic Mice

Type 2 diabetes (T2D) is a global chronic metabolic disease characterized by continuous elevation of plasma glucose. As of 2017, there were ∼425 million patients with diabetes worldwide, >90% of whom have T2D; by 2045, the global number of individuals with diabetes is estimated to reach 629 million (1). Although insulin resistance is a vital prerequisite for T2D, recent studies have shown that unmet insulin demand due to pancreatic β-cell failure, which is mainly caused by hyperglycemia/glucotoxicity, hyperlipidemia/lipotoxicity, and inflammation, is strongly linked to T2D development and progression (2). Therefore, β-cell function maintenance is of great significance in global glucose homeostasis.

The ubiquitin-proteasome system is one of the main mechanisms of cellular protein degradation. Proteins targeted for degradation are labeled through ubiquitination and then degraded by 26S proteasomes. To bind ubiquitin to proteins, three distinct enzymes are required: E1 (ubiquitin-activating enzyme), E2 (ubiquitin-conjugating enzyme), and E3 (ubiquitin ligase); E3 ubiquitin ligases recognize certain protein substrates and catalyze the transfer of activated ubiquitin to client proteins. The targeting specificity of proteins degraded by ubiquitin proteasomes is mediated by E3 ubiquitin ligase, which contains a large family of members (3). Although the protein targets of E3 ligase are yet to be identified in most cases, some of them are associated with human diseases, such as tumors, autoimmunity, cardiovascular disease, as well as pancreatic β-cell failure in diabetes (4–9). Some E3 ligases are relatively abundant in islets and β-cells (10). In fact, several E3 ligases have been reported to be involved in the regulation of β-cell function, survival, proliferation, and apoptosis (11–13).

HMG-CoA reductase degradation protein 1 (HRD1), also called synoviolin, is a representative of the endoplasmic reticulum (ER) stress–associated E3 ubiquitin ligase; HRD1 is localized in the ER, together with Sel1, Derlin1, Herp, and other proteins, as a component of the ER-associated degradation (ERAD) complex (14,15). Misfolded proteins in the ER bind to BIP and other chaperones during ERAD and are retrovirally translocated into the cytosol, where these misfolded proteins are ubiquitinated by the cytoplasmic RING domain of HRD1 (14) and ultimately destroyed via the ubiquitin-proteasome system. Although HRD1 is required to clear misfolded proteins during ERAD, new evidence suggests that HRD1 also regulates cell functions by controlling the availability of specific proteins such as p53 and Nrf2 (16,17), which serve as non-ERAD substrates for HRD1. HRD1 is suggested to support pancreatic β-cell function by targeting misfolded proinsulin for efficient degradation, and loss of HRD1 leads to blunted glucose-stimulated insulin secretion (GSIS) (18). However, Allen et al. (19) have recently reported that HRD1 expression is upregulated in the pancreatic islets of Akita diabetic mice. Genomic data from the National Center for Biotechnology Information Gene Expression Omnibus (GEO) database also show that HRD1 mRNA level is increased in pancreatic islets of patients with T2D (GEO accession number GSE20966). However, the β-cell–specific pathological roles of HRD1 remain largely uncharacterized.

A large number of β-cell transcription factors play an essential role in the maintenance of mature β-cell function; however, the expression levels and/or activities are reduced when β-cells are chronically exposed to diabetogenic stress, such as hyperglycemia (20–23). Among them, the transcription factor v-maf musculoaponeurotic fibrosarcoma oncogene homolog A (MafA) is expressed only in pancreatic β-cells and is the key factor required for β-cell formation and function (24–26). MafA deletion in mice caused glucose intolerance and induced diabetes; by contrast, attenuation of MafA reduction in db/db mouse β-cells overcame β-cell failure and diabetes (27). Therefore, MafA expression levels must be tightly regulated for the maintenance of β-cell function. MafA activity is regulated by multiple mechanisms, including modulation of mRNA transcription, posttranslational modifications, and subsequent changes in subcellular localization (nucleus or cytoplasm) (26,28–30). Although intense ubiquitin-induced degradation accounts for MafA reduction in β-cells under diabetic conditions (31–33), the specific role of E3 ubiquitin ligases in the regulation of MafA activity remains unexplored.

In the current study, we demonstrated that E3 ubiquitin ligase HRD1 is considerably induced to trigger impaired β-cell insulin secretion in diabetic status. We observed that HRD1 expression is dramatically increased in β-cells of islets from humans with T2D and mice. Using mass spectrometric analysis, we identified a large group of potential interactomes for HRD1 in β-cells, including the β-cell–specific transcription factor MafA. Moreover, we found that HRD1 serves as a novel E3 ubiquitin ligase that is responsible for MafA ubiquitination and degradation, resulting in the reduced nuclear function of MafA, leading to β-cell dysfunction.

Male C57BL/6 and db/db mice were purchased from the Model Animal Research Center of Nanjing University. To establish a T2D model, the mice were housed in cages, reared under a 12-h light/12-h dark cycle with free access to water, and fed with either a normal diet consisting of standard laboratory chow or with a high-fat diet (HFD) (D12492; Research Diets) containing 60% kcal from fat for at least 3 months. All animal experiments were carried out in strict accordance with the guidelines and rules formulated by the Animal Care Committee of Nanjing Medical University.

Mice were injected with adeno-associated virus serotype 8 vector (AAV8)–mouse Insulin1 promoter (MIP)–specific shRNA for HRD1 (shHRD1)–GFP or AAV8-MIP-HRD1-GFP or with control virus (Vigene Biosciences, Inc., Jinan City, China). The viruses (10¹² genome copy particles/mL) were infused at a rate of 6 μL/min by pancreatic intraductal viral infusion as described previously (34).

Fasting blood glucose was measured in mice 2 and 4 weeks after viral infusion. Intraperitoneal glucose tolerance tests (IPGTTs) were performed by intraperitoneal injection of 1 g/kg body weight of glucose (Sigma-Aldrich) after overnight fasting. Blood samples were collected at all time points for plasma insulin measurements with ELISA test kits (Ezassay Biotech Co., Ltd., Shenzhen City, China). For insulin tolerance tests (ITTs), the mice were intraperitoneally injected with 1 unit/kg body weight of insulin after a 4-h fasting.

The use of human primary islets was approved by the Research Ethics Committee of Nanjing Medical University (Nanjing, China). Islet isolation and culture, islet/β-cell line GSIS assay, and islet perfusion analysis were performed as described by Huang et al. (35).

The mouse pancreatic β-cell line MIN6 was established as described previously (36) and cultured in DMEM (Invitrogen) containing 15% FBS (Gibco), 10 mmol/L HEPES, 1 mmol/L sodium pyruvate, 100 units/mL penicillin, 100 μg/mL streptavidin, and 50 μmol/L β-mercaptoethanol. Chronic high-glucose treatment was performed as previously described (35).

Human islets were dispersed into single cells by adding 0.08% trypsin solution containing 10 μg/mL DNase (Sigma-Aldrich) at 37°C for 8 min followed by pipetting 20 times for 30 s and trypsin inactivation by addition of 10% FBS. Human islet single cells were subsequently plated in 35-mm dishes. The Sine+DC lock-in function of an EPC 10 amplifier and the Patchmaster software (HEKA Electronik, Lambrecht/Pfalz, Germany) were used in the standard whole-cell technique. To measure the voltage-dependent Ca²⁺ currents, we used an intercellular solution consisting of 118 mmol/L NaCl, 20 mmol/L tetraethylammonium chloride, 5.6 mmol/L KCl, 1.2 mmol/L MgCl₂, 2.6 mmol/L CaCl₂, 5 mmol/L d-glucose, and 5 mmol/L HEPES (adjusted to pH 7.4 with NaOH). The pipette solution used contained 125 mmol/L Cs-glutamate, 10 mmol/L CsCl, 10 mmol/L NaCl, 1 mmol/L MgCl₂, 0.05 mmol/L EGTA, 3 mmol/L Mg-ATP, 0.1 mmol/L cAMP, and 5 mmol/L HEPES (adjusted to pH 7.1 with CsOH). The whole-cell Ca²⁺ current was normalized to the initial cell size and expressed as picoamperes per picofarad. In all experiments, human β-cells were positively identified by immunostaining for insulin.

Cells were lysed in lysis buffer (50 mmol/L Tris-Cl, pH 8.0, 20% glycerol, 140 mmol/L NaCl, 0.5% Nonidet P-40, 5 mmol/L MgCl₂, 0.2 mmol/L EDTA, 1 mmol/L dithiothreitol, and 1 mmol/L protease inhibitor cocktail). After centrifugation, 10% supernatants were used as inputs, and the rest were immunoprecipitated with agarose A/G beads (Roche), IgG (negative controls), or antibodies for the target proteins at 4°C overnight. The pellets were washed three times with wash buffer, resuspended in two times SDS, and then subjected to Western blot analysis, which was performed as previously described (36) using antibodies against HRD1 (Cell Signaling Technology), GAPDH (Bioworld Technology, Inc., Nanjing, China), Flag-Tag (Santa Cruz Biotechnology), Myc-Tag (Santa Cruz Biotechnology), and HA-Tag (Cell Signaling Technology).

The mass spectrometry was performed on the precipitated protein of MIN6 cells transfected with adenovirus-HRD1 (Ad-HRD1) or empty vectors. The labeled peptides were analyzed on the LTQ-Orbitrap instrument (Thermo Fisher Scientific) connecting to a nanoACQUITY UPLC System via a nanospray source. The liquid chromatography–tandem mass spectrometry (MS/MS) was operated in positive ion mode. The analytical condition was set at a linear gradient from 0 to 60% of buffer B (CH₃CN) for 150 min and flow rate of 200 nL/min. For analysis of proteins from MIN6 cells, one full MS/MS scan was followed by five MS/MS scans on those five highest peaks, respectively. The MS/MS spectra acquired from precursor ions were submitted to MaxQuant (version 1.2.2.5) using the following search parameters: the database searched was UniProt proteome (version 20120418), the enzyme was trypsin (KR/P); the dynamic modifications were set for oxidized Met (+16), carbamidomethylation of cysteine was set as static modification, MS/MS tolerance was set at 10 ppm, the minimum peptide length was 6, and the false detection rates for peptides and proteins were all set at <0.01 (37).

Pancreases obtained from mice were fixed in 10% formalin and then embedded in paraffin for sectioning. The paraffinized sections were deparaffinized and then rinsed in double-distilled H₂O. Antigen retrieval was performed by heating the slides at 100°C for 8 min in an acidic retrieval solution. The samples were blocked in 3% (w/v) BSA for 15 min at room temperature before incubating at 4°C overnight with primary antibodies against HRD1, insulin (Santa Cruz Biotechnology), or MafA (Abcam). After being washed, the specimens were incubated in fluorochrome-conjugated secondary antibody for 1 h at room temperature in the dark. Nuclei were stained with DAPI and then secured with a coverslip. Images were obtained using a laser scanning microscope (Olympus).

Chromatin was cross-linked by 1% formaldehyde and terminated by 0.125 mol/L glycine. Cells were collected after two washings with iced PBS in lysis buffer (1% SDS, 10 mmol/L EDTA, 50 mmol/L Tris-HCl, pH 8.1, and 1 mmol/L phenylmethylsulfonyl fluoride) and then sonicated on ice. After centrifugation, 10% of the supernatants were used as inputs, and the rest was diluted fivefold with chromatin immunoprecipitation (ChIP) dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mmol/L EDTA, 16.7 mmol/L Tris-HCl, pH 8.1, 167 mmol/L NaCl, and 1 mmol/L phenylmethylsulfonyl fluoride). The diluted fraction was subjected to immunoprecipitation overnight after 1 h of preclearing at 4°C with 30 μL protein A agarose beads/salmon sperm DNA beads (Millipore). Complexes were recovered through a 2- to 4-h incubation at 4°C with 60 μL protein A agarose beads/salmon sperm DNA beads. Precipitates were serially washed with low-salt immune complex wash buffer (0.1% SDS, 1% Triton X-100, 2 mmol/L EDTA, 20 mmol/L Tris-HCl, pH 8.1, and 150 mmol/L NaCl), high-salt immune complex wash buffer (0.1% SDS, 1% Triton X-100, 2 mmol/L EDTA, 20 mmol/L Tris-HCl, pH 8.1, and 0.5 mol/L NaCl), and LiCl immune complex wash buffer (LiCl 0.25 mol/L, 1% Nonidet P-40, 0.1 mmol/L EDTA, 24.1 mmol/L C₂₄H₃₉O₄Na, and 10 mmol/L Tris-HCl, pH 8.1) and twice with TE buffer (1 mmol/L EDTA and 10 mmol/L Tris-HCl, pH 8.1). Precipitated chromatin complexes were eluted with ChIP elution buffer (2.9% SDS and 28.6 mmol/L NaHCO₃). DNA was extracted, purified, and then subjected to PCR using designed primers (38) (shown in Supplementary Table 3). All of the buffers mentioned above contained protease inhibitor cocktail (Roche).

Transient transfection with plasmids and siRNA was performed with Lipofectamine 2000 (Invitrogen) as previously described (36). Adenovirus-MafA (Ad-MafA) (mouse), Ad-HRD1 (human/mouse species), and control virus were purchased from GeneChem (Nanjing, China). Specific siRNAs for HRD1 (si-HRD1) and MafA (si-MafA) were purchased from RiboBio (Guangzhou, China). The siRNA sequences are shown in Supplementary Table 1. MafA transcriptional activity was assessed using the reporter construct MARE-Luc (39) and the Dual-Glo Luciferase Assay System (Promega) on a TD-20/20 Luminometer (Turner BioSystems).

Quantitative RT-PCR (qRT-PCR) was performed with SYBR Green PCR Master Mix (Vazyme Biotech Co., Ltd, Nanjing, China) and analyzed using a Roche Real-Time PCR System as previously described (36). The primer sequences are shown in Supplementary Table 2.

HRD1-DsRed/-Flag/-HA and truncated forms, MafA-Myc/-EGFP, were generated using a QuikChange Site-Directed Mutagenesis kit (Stratagene) according to the manufacturer’s instructions. The primers for plasmids are shown in Supplementary Table 4.

HEK293A cells were transfected with wild-type (WT) or C291S mutant HRD1-Flag, MafA-Myc, and Ubiquitin-HA for 48 h. The cells were incubated with the proteasome inhibitor MG132 (Sigma-Aldrich) for 4 h before being lysed. Equal amounts of total cell lysates were incubated with MafA antibody overnight at 4°C. Immunocomplexes were collected overnight at 4°C using protein A-sepharose beads (Roche). The immunoprecipitates were washed with lysis buffer and subjected to Western blot analysis with anti-HA antibody.

All of the experiments above were performed independently at least three times. The results are presented as means ± SEM. Statistical difference between groups was determined by Student t test, and comparisons among groups were performed using ANOVA. A P value <0.05 indicated statistical significance.

The data sets and adenoviruses and plasmids generated and/or analyzed during the current study are available from the corresponding authors upon reasonable request.

A GEO data set (GSE20966) shows that HRD1 mRNA level is increased in pancreatic islets of patients with T2D. In this study, HRD1 was significantly upregulated in db/db and HFD-fed diabetic mouse islets (Fig. 1A and B); similar results were obtained in the islets isolated from humans with T2D (Fig. 1C). Consistent with this finding, isolated human and mouse islets with prolonged exposure to high glucose in vitro showed an obviously upregulated HRD1 expression compared with the normal glucose group (Fig. 1D and E). Moreover, ER-resident HRD1 was markedly increased in the pancreatic β-cells of db/db and HFD-fed mice, accompanied by a dramatic reduction in insulin (Fig. 1F), while no significantly elevated HRD1 could be observed in other islet cells, including α cells and pancreatic polypeptide cells (Supplementary Fig. 1A and B). We also observed an increased fluorescence intensity of HRD1 in chronic high glucose–treated primary β-cells in cell-immunostained dispersed human islets (Fig. 1G). Although long-term high-glucose incubation increased HRD1 expression, cotreatment with the antioxidant N-acetylcysteine prevented this effect in mouse islets and MIN6 cells (Supplementary Fig. 1C). In addition, H₂O₂ alone obviously upregulated HRD1 expression in β-cells (Supplementary Fig. 1D), indicating that oxidative stress induced by diabetic hyperglycemia (40) leads to abnormally elevated HRD1 expression in β-cells. Taken together, these results demonstrate that HRD1 is elevated under diabetic conditions and is likely to be associated with β-cell failure and diabetes.

To determine whether HRD1 is involved in the development of diabetes, we used β-cell–specific HRD1 knockdown mice obtained via pancreatic intraductal AAV8-MIP-shHRD1-GFP infusion, and then we confirmed the effective delivery of AAV vectors via immunofluorescence (Supplementary Fig. 2A). To evaluate the effect of β-cell–specific HRD1 knockdown on diabetes, we injected AAV8-MIP-shHRD1 into the diabetic mice (Fig. 2A). Fasting blood glucose in HFD-induced diabetic mice treated with AAV8-shHRD1 began to drop significantly 2 weeks after operation (10.13 ± 0.53 mmol/L in AAV-shHRD1 group vs. 7.925 ± 0.83 mmol/L in AAV–specific shRNA for negative control [shNC]) and reached up to 8.05 ± 0.49 mmol/L in the AAV-shHRD1 group after 2 more weeks, while AAV-shHRD1 had no effect on fasting glycemia in chow-fed mice (Fig. 2B). Meanwhile, IPGTT results revealed a marked improvement in the glucose response and serum insulin levels of AAV8-shHRD1–infused HFD mice at 4 weeks after viral infusion (Fig. 2C and D). Improvement in glucose response was also observed in AAV8-shHRD1–injected db/db mice (Fig. 2E–G); no significant changes were observed in the body weight and intraperitoneal ITT results (Supplementary Fig. 2B–E). Thus, β-cell–specific HRD1 knockdown exerts a robust effect on T2D treatment, indicating that HRD1 plays an important role in β-cell dysfunction and diabetes.

To further determine whether HRD1 upregulation can initiate the pathogenesis of hyperglycemia and diabetes, we obtained mice with β-cell–specific HRD1 overexpression via pancreatic intraductal AAV8-MIP-HRD1 infusion. HRD1 protein levels were significantly elevated in islets after AAV infusion (Fig. 3A). The mice with β-cell–specific HRD1 overexpression showed remarkably elevated fasting and refed blood glucose levels (Fig. 3B); meanwhile, IPGTT results revealed severe impairments in glucose metabolism (Fig. 3C), and no obvious changes in insulin sensitivity were observed (Fig. 3D). In addition, plasma insulin levels in mice with β-cell–specific HRD1 overexpression were disordered after glucose stimulation compared with those in mice infused with the control virus (Fig. 3E). These observations not only demonstrate that HRD1 upregulation in β-cells can trigger hyperglycemia but also indicate that HRD1 exerts a harmful effect on insulin secretion.

We investigated whether excessive HRD1 was involved in β-cell dysfunction. High glucose– or KCl-stimulated insulin secretion was dramatically decreased in MIN6 cells after HRD1 overexpression for 48 h (Fig. 4A and B and Supplementary Fig. 3A and B). Analysis of secretion dynamics in control and Ad-HRD1–infected human islets via perfusion showed that HRD1-overexpressing β-cells secreted less insulin both at the first and second phases following stimulation with high glucose and after direct membrane depolarization with KCl (Fig. 4C–F and Supplementary Fig. 3C). We applied the whole-cell patch-clamp technique to further assess the effect of HRD1 overexpression on human β-cell function, considering that a critical component of β-cell function is the electrical activity mediated via ion channels that lead to Ca²⁺ influx and exocytosis of insulin-containing granules (41). We studied the effect of HRD1 overexpression on insulin granule exocytosis, which can be measured as the increase in cell capacitance in response to a train of depolarization (42). Exocytosis in the Ad-HRD1 group was dramatically decreased for all pulse durations compared with that in the Ad-GFP group (Fig. 4G and H), despite the fact that the voltage-dependent Ca²⁺ channel currents in the HRD1-overexpressing β-cells and in the control did not significantly differ (Supplementary Fig. 3D–F). Furthermore, we observed that a slight increase in HRD1 expression did not impair insulin secretion (Supplementary Fig. 3G and H). These data further indicate that excessive HRD1, rather than physiologically elevated HRD1, severely impaired insulin secretion in β-cells.

Being an E3 ubiquitin ligase, HRD1 usually interacts with a specific substrate and then directs its target protein to be ubiquitinated and ultimately degraded via the proteasome pathway. To determine the molecular basis of the impaired insulin secretion in β-cells with high HRD1 expression, we performed a mass spectrometric analysis to investigate the binding proteins of HRD1 in Ad-HRD1– or Ad-GFP–infected MIN6 cells. Data showed that 105 proteins were candidate substrates of HRD1 when HRD1 was highly expressed in β-cells (Supplementary Table 6). Importantly, gene ontology analysis with Database for Annotation, Visualization and Integrated Discovery (DAVID) Bioinformatics Resources 6.7 (Supplementary Fig. 4A) showed that three of the excessive HRD1-interacting proteins were involved in carbohydrate stimulus response, which may be associated with the GSIS cellular process. Surprisingly, among these proteins is MafA, which is an essential β-cell–specific transcriptional factor (Fig. 5A). Immunoprecipitation analyses confirmed that endogenously and exogenously, HRD1 and MafA coexisted in the precipitated complexes (Fig. 5B–D), and their interaction was strengthened when HRD1 was highly expressed (Fig. 5B). MafA is a positive indicator of β-cell functionality and health, and its expression in β-cells can be regulated through a transcriptional and/or posttranslational modification (26). Using human islets, we determined the inverse correlation between HRD1 and MafA. HRD1 overexpression significantly reduced MafA expression at the translational rather than transcriptional level (Fig. 5E and F); by contrast, immunoblot results revealed the accumulation of MafA in HRD1 knockdown β-cells, although its mRNA levels did not significantly change (Fig. 5G and H).

To measure the role of MafA in HRD1-induced impairment of β-cell function, we performed several additional experiments. We found that glucose- or KCl-stimulated insulin secretion was significantly inhibited in isolated mouse islets (Fig. 6A and B) and in MIN6 cells (Supplementary Fig. 5A and B) infected with Ad-HRD1 alone, but this inhibition could be restored by MafA replenishment. Furthermore, HRD1 knockdown effectively alleviated MafA reduction resulting from exposure to chronic high glucose in both mice islets (Fig. 6C) and MIN6 cells (Supplementary Fig. 5C); similarly, glucose- or KCl-stimulated insulin secretion was restored in HRD1-knockout mouse islets and MIN6 cells with long-term exposure to high glucose (Fig. 6D and Supplementary Fig. 5D). MafA expression in islets also rebounded after AAV-shHRD1 treatment for 4 weeks both in HFD-fed mice (Fig. 6E and F) and db/db mice (Fig. 6G and H). Taken together, high HRD1 expression in β-cells contributes to β-cell dysfunction via MafA downregulation.

Considering the interplay between ER-resident HRD1 and nuclear-localized MafA, we subsequently investigated their localization in HRD1-overexpressing β-cells. Surprisingly, MafA was predominantly colocalized with HRD1 in the perinuclear region rather than in the nucleus when HRD1 was overexpressed, especially in MIN6 cells (Fig. 7A). We obtained similar results in HEK293A cells in which MafA and HRD1 were co-overexpressed (Fig. 7B). By using the ER-Tracker Blue-White DPX, we confirmed that the perinuclear regions were the ER, indicating that HRD1 entraps MafA around the ER (Supplementary Fig. 6A). In addition, the interaction of HRD1–N-HA or HRD1–N-RING-HA was totally lost when the COOH-terminal domain was removed, further suggesting that HRD1 interacts with cytosolic MafA via its cytosolic COOH-terminal domain (Fig. 7C). Moreover, by using luciferase reporter assay, we found that HRD1 overexpression significantly reduced the nuclear transcriptional regulatory activity of MafA (Fig. 7D). This result was confirmed by the quantitative PCR (Fig. 7E and Supplementary Fig. 6B) and ChIP quantitative PCR (Fig. 7F and G) results for Insulin, the major gene transcriptionally regulated by MafA. In contrast, we observed that HRD1 knockdown partially rescued decreased nuclear MafA in the glucolipotoxic conditions (Supplementary Fig. 6C), which indicates that reduction of HRD1 is sufficient to retain nuclear MafA levels. These findings reveal that HRD1 regulates not only the quantity but also the localization of MafA in β-cells.

In animal models of hyperglycemia and oxidative stress, MafA tended to be ubiquitinated after single and/or several sites phosphorylated, leading to MafA loss in β-cells and ultimately in impaired insulin secretion and in apoptosis (33). Considering that the E3 ubiquitin ligase HRD1 interacts with MafA and negatively regulated its expression, we next examined whether HRD1 indeed ubiquitinates MafA. Cysteine 291 within the RING domain of HRD1 is indispensable for its ubiquitin ligase activity (17); thus, we measured the effect of either the HRD1-WT or C291S-mutant HRD1 (HRD1-C291S) on ubiquitination of MafA, and we found that HRD1-WT overexpression but not HRD1-C291S mediated the obvious increase in MafA ubiquitination (Fig. 8A). Moreover, HRD1 knockdown remarkably reduced MafA ubiquitination in MIN6 cells treated with chronic high glucose (Fig. 8B). Furthermore, HRD1 deletion prevented the rapid decay of MafA protein after exposure to chronic high glucose (Fig. 8C). HRD1 deletion also restored the decrease in total MafA protein in MIN6 cells exposed to chronic high glucose, which were not different from the presence of MG132 (43) (Fig. 8D), indicating that si-HRD1 and MG132 inhibit different steps of the same MafA degradation pathway under chronic hyperglycemia. Collectively, these results clearly demonstrate that ER-resident HRD1 is an E3 ubiquitin ligase that is responsible for MafA ubiquitylation and degradation.

In this study, we explored the role of HRD1, an ER-resident E3 ubiquitin ligase, in β-cell insulin secretion disorders under diabetic conditions. Excessive HRD1 expression in β-cells leads to reduced GSIS function, whereas HRD1 inhibition effectively alleviates glucose control and response in diabetic status. HRD1 acts as a novel E3 ubiquitin ligase for MafA, which is a β-cell–specific transcription factor that maintains the homeostasis of mature β-cells. When highly expressed, HRD1 targets MafA around ER and contributes to the ubiquitination and degradation of MafA, weakening its nuclear biological function, leading to compromised insulin secretion.

In islets of patients with T2D, dramatic ER stress and oxidative stress induced by hyperglycemia and obesity play causal roles in β-cell dysfunction (40,44,45). Both ER stress and oxidative stress lead to excessive HRD1 expression (17,46,47), but this finding has not yet been investigated in β-cells. Our research demonstrated that HRD1 expression was significantly elevated in human and mouse diabetic islets, similar to that observed in β-cells exposed to chronic high glucose. In vivo and in vitro analyses revealed that ectopic overexpression of HRD1 impairs GSIS function, whereas HRD1 knockdown remarkably improves insulin secretion in β-cells under diabetogenic stress. Moreover, studies on islet perfusion and membrane capacitance measurements showed that HRD1 inhibits insulin secretion in both the first and second phases despite the fact that the Ca²⁺ influx did not significantly change, indicating that HRD1 exerts a more significant inhibitory effect on insulin synthesis and granule formation than on glucose response (48). All of these observations reveal that HRD1 upregulation is both necessary and sufficient to trigger insulin secretion disorders that further develop into diabetes.

The molecular mechanisms underlying the pathological roles of HRD1 as an E3 ubiquitin ligase are closely related to those of its substrates. We investigated for the first time the interactive proteins of HRD1 in β-cells through mass spectrometry, and we found that the essential β-cell transcription factor MafA is a bona fide substrate of HRD1. As a target for HRD1, MafA is evidently a non-ERAD substrate, just like tumor suppressor gene p53, which also serves as non-ERAD substrate for HRD1 (16). Our results reveal that MafA and HRD1 expression levels were inversely correlated because HRD1 could induce MafA ubiquitination due to its ubiquitin ligase activity, which was responsible for the aberrant MafA degradation under diabetogenic stress. Additionally, gene therapy via delivery of AAV-MIP-shHRD1 to the pancreas effectively restored MafA expression in diabetic mouse islets followed by improved fasting blood glucose and glucose tolerance; moreover, MafA replenishment significantly alleviated the impaired GSIS caused by excessive HRD1.

Posttranslational control by ubiquitination is vital for the stability and function of several β-cell transcription factors. For instance, Pdx-1, a major mediator of insulin development and a key regulator of β-cell phenotype, is targeted by Pcif1-Cul3 ubiquitin ligase complexes for ubiquitination and proteasomal degradation, impairing β-cell function and survival (11). MafA has been shown to be crucial for glucose regulation of insulin gene transcription, and its protein stability is regulated by transcriptional mechanisms, especially under stress conditions; the E3 ligases responsible for its degradation remained largely unknown. However, our findings demonstrate that E3 ligase HRD1 is responsible for MafA ubiquitination and degradation, indicating the functional importance of HRD1-mediated MafA activity in the diabetic setting.

MafA is mainly localized in the nucleus of β-cells under physiological conditions. Interestingly, when HRD1 was highly expressed, HRD1 and MafA are colocalized in the ER region accompanied by the impaired nuclear activity of MafA, indicating that HRD1 controls not only the quantity but also the localization of MafA. Given that MafA could shuttle between the nucleus and the cytoplasm in β-cells under certain circumstances (30,32,38), the increased nuclear MafA translocation and/or reduced entry of cytoplasmic MafA could be responsible for cytoplasmic accumulation of MafA due to excessive HRD1. HRD1 is an ER-resident protein (17), and its localization in β-cells did not change, with its elevated expression levels owing to altered extracellular environment, suggesting that HRD1-induced cytoplasmic MafA accumulation initially occurs in the ER. Therefore, we speculate that HRD1 ubiquitinates newly translated MafA generated from the ribosomes embedded in the ER, leading to the degradation of MafA via the ubiquitin-proteasome pathway before it could enter the nucleus. As illustrated in Figs. 5, 6, and 8, HRD1 knockdown increases MafA protein levels and stability, indicating that the MafA trapping in the ER and MafA degradation in the cytoplasm also occur under physiological conditions, and this phenomenon is much more obvious when HRD1 is highly expressed (e.g., sustained high-glucose exposure). Studies have demonstrated that MafA is inactivated during its translocation from the nucleus to the cytoplasm under oxidative stress conditions induced by hyperglycemia (32,38). Our results supplement the finding above, in which HRD1 sequestrates and metabolizes MafA in the cytoplasm; this phenomenon results in the absence of MafA in nuclear extracts of glucotoxic β-cells, thereby disrupting insulin gene expression and ultimately resulting in β-cell dysfunction. However, we still lack direct evidence to support this view, and thus, further investigation that involves protein-tracing technology is warranted.

Although excessive HRD1 expression causes β-cell dysfunction, the absence of HRD1 caused by impaired GSIS results in hyperglycemia; moreover, HRD1 mediates the degradation of misfolded proteins to maintain pancreatic β-cell function under normal physiological conditions (18). In fact, the slightly elevated HRD1 expression did not induce insulin secretion disorders, although it promoted GSIS to a small extent (Supplementary Fig. 3G). All of these observations suggest that the substrates of HRD1 in β-cells in physiological and pathological conditions may differ. Thus, we reanalyzed the mass spectrometric data and found that 59 proteins interacting with highly expressed HRD1 did not exist in the interactome of the HRD1 expressed under physiological conditions. Also, 75 proteins only interact with HRD1 under physiological conditions (Supplementary Tables 5 and 6). Therefore, we tend to believe that HRD1 expression has been stabilized to maintain homeostasis of pancreatic β-cells, the dysregulation (loss or excess) of which could induce β-cell failure. Importantly, besides the MafA inhibition discussed above, there might still be other potential molecular mechanisms underlying β-cell dysfunction induced by excessive HRD1 that have not been fully studied in this work; thus, further work on this subject should be continued in the future.

In conclusion, our observations demonstrate that excessive HRD1-mediated MafA ubiquitination and degradation play a crucial role in the pathogenesis of diabetes. This identification suggests that the therapeutic targeting of HRD1 could increase MafA protein level to improve β-cell function for the treatment of diabetes.

Funding. This work was supported by the Major International (Regional) Joint Research Program of the National Natural Science Foundation of China (81420108007), the Key Program of the National Natural Science Foundation of China (81830024), and grants from the National Natural Science Foundation of China (81670705 and 81702882). This work was also supported by the Natural Science Foundation of Jiangsu Province (BK20171056 and BK20170106) and the Open Fund of the State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, China (KF-GN-201903).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. T.W., F.C., and X.H. designed this research. S.Z., J.X., Y.S., and J.W. conducted the research. Y.Z., T.S., and W.T. contributed new reagents and analytic tools. T.W., F.C., and X.H. analyzed the data. T.W. and F.C. prepared the manuscript. F.C. and X.H. are the guarantors of this work and, as such, had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.