Ubiquitin D Regulates IRE1α/c-Jun N-terminal Kinase (JNK) Protein-dependent Apoptosis in Pancreatic Beta Cells*

Pro-inflammatory cytokines contribute to pancreatic beta cell apoptosis in type 1 diabetes at least in part by inducing endoplasmic reticulum (ER) stress and the consequent unfolded protein response (UPR). It remains to be determined what causes the transition from “physiological” to “apoptotic” UPR, but accumulating evidence indicates that signaling by the ER transmembrane protein IRE1α is critical for this transition. IRE1α activation is regulated by both intra-ER and cytosolic cues. We evaluated the role for the presently discovered cytokine-induced and IRE1α-interacting protein ubiquitin D (UBD) on the regulation of IRE1α and its downstream targets. UBD was identified by use of a MAPPIT (mammalian protein-protein interaction trap)-based IRE1α interactome screen followed by comparison against functional genomic analysis of human and rodent beta cells exposed to pro-inflammatory cytokines. Knockdown of UBD in human and rodent beta cells and detailed signal transduction studies indicated that UBD modulates cytokine-induced UPR/IRE1α activation and apoptosis. UBD expression is induced by the pro-inflammatory cytokines interleukin (IL)-1β and interferon (IFN)-γ in rat and human pancreatic beta cells, and it is also up-regulated in beta cells of inflamed islets from non-obese diabetic mice. UBD interacts with IRE1α in human and rodent beta cells, modulating IRE1α-dependent activation of JNK and cytokine-induced apoptosis. Our data suggest that UBD provides a negative feedback on cytokine-induced activation of the IRE1α/JNK pro-apoptotic pathway in cytokine-exposed beta cells.

protein interaction trap) coupled to functional genomic analysis of human and rodent beta cells exposed to IL-1␤ ϩ IFN-␥ to identify novel cytokine-induced IRE1␣-interacting proteins that modulate UPR activation in pancreatic beta cells (13). Based on this analysis, we selected two candidates for detailed functional studies: NMI (13) and ubiquitin D (UBD; this study). UBD (also know as FAT10) is expressed in the immune system and in some cancer cells (14 -17), and its expression is up-regulated by IFN-␥ and TNF-␣ (18,19). UBD interacts non-covalently or covalently with different proteins, respectively, modifying their activity (20,21) or triggering their degradation in the proteasome (22,23). Of particular interest in the context of T1D, the Ubd gene maps to the telomeric region of the human major histocompatibility complex (MHC), the most important susceptibility locus for T1D (24,25). Polymorphisms in the region of the Ubd gene have been associated with autoimmune diabetes in rat and human (26 -29), but this remains to be confirmed.
We presently show that UBD expression is induced by proinflammatory cytokines in rat and human pancreatic beta cells, and it is also present in beta cells of inflamed islets from NOD mice. Of particular importance, we show that UBD interacts with IRE1␣ in cytokine-treated human and rodent beta cells, providing a negative feedback for IRE1␣-induced activation of JNK and consequent apoptosis.

Culture of Human Islet Cells, FACS-purified Rat Beta Cells, INS-1E Cells, the Human Beta Cell Line EndoC-␤H1, and
HEK293T Cells-Human islets from 13 non-diabetic donors were isolated in Pisa using collagenase digestion and density gradient purification (30). The donors (seven women and six men) were 67.1 Ϯ 4.7 years old and had a body mass index of 25.0 Ϯ 1.0 (kg/m 2 ) ( Table 1). Beta cell purity, as evaluated by immunofluorescence for insulin, using a specific anti-insulin antibody ( Table 2), was 52 Ϯ 5.4%. The islets were cultured as described previously (25,31).
Cell Treatment and NO Measurement-Cells were exposed to the following cytokine concentrations, based on previous dose-response experiments performed by our group (31,(35)(36)(37): recombinant human IL-1␤ (R&D Systems, Abingdon, UK) 10 units/ml for INS-1E cells or 50 units/ml for human islet cells, primary rat beta cells, and the EndoC-␤H1 cells; recombinant rat IFN-␥ (R&D Systems) 100 units/ml for INS-1E cells or 500 units/ml for primary rat beta cells; and human IFN-␥ (PeproTech, London, UK) 1000 units/ml for human islet cells or the EndoC-␤H1 cells. Lower cytokine concentration and shorter time points were used in the rodent experiments because rat beta cells are more sensitive to cytokine-induced damage than human islets (38,39). Culture medium was collected for nitrite determination (nitrite is a stable product of NO oxidation) by the Griess method.  ArrayMAPPIT and Binary MAPPIT-To identify IRE1␣-interacting proteins, an ArrayMAPPIT screen was performed as described (40), using as MAPPIT bait the cytoplasmic portion of IRE1␣ (amino acids 571-977) (13). The same bait was also used in the binary MAPPIT analyses, and it was generated based on the original pSEL MAPPIT bait construct (41). MAPPIT preys were cloned in the previously described pMG1 vector (42). Briefly, HEK293T cells were transfected with bait, prey, and STAT3 reporter plasmids (Table 3), and luciferase activity was measured 48 h after transfection using the luciferase assay system kit (Promega, Leiden, The Netherlands) on a TopCount illuminometer. Cells were stimulated with erythropoietin (5 ng/ml) 24 h after transfection. To provide further support for the specificity of the IRE1␣/UBD interaction, we performed additional MAPPIT two-hybrid analyses, using irrelevant protein baits in addition to the empty bait. From the Cytokine Receptor Laboratory bait collection, we randomly selected a number of baits that were cloned in the same vector backbone (pSEL ϩ 2L) as the IRE1␣ bait. In addition, because PERK, like IRE1␣, is an ER stress sensor protein with a very similar topology to that of IRE1␣, we generated a MAPPIT bait consisting of the PERK cytosolic portion, similar to the IRE1␣ bait that was used to show the UBD interaction.
RNA Interference-The siRNAs (30 nM) used in this study are described in Table 4. They were transfected into beta cells using previously defined optimal conditions (31,43). After 16 h of transfection, cells were cultured for a 24-or 48-h recovery period before exposure to cytokines.
Assessment of Cell Viability-The percentage of viable, apoptotic, and necrotic cells was determined after staining the cells with the DNA-binding dyes propidium iodide (5 g/ml; Sigma, Bornem, Belgium) and Hoechst dye 33342 (5 g/ml; Sigma). A minimum of 600 cells was counted for each experimental condition by two independent observers, with one of them unaware of sample identity. The agreement between findings obtained by the two observers was always Ͼ90%. This fluorescence assay for single cells is quantitative and has been validated by systematic comparisons against electron microscopy observations, ladder formation, and caspase 3/9 activation (31, 44 -46).
Western Blot, mRNA Extraction, and Real Time PCR-For Western blot, cells were washed with cold PBS and lysed using Laemmli Sample Buffer. Total protein was extracted and resolved by 8 -10% SDS-PAGE, transferred to a nitrocellulose membrane, and immunoblotted with the specific antibodies for the protein of interest (Table 2) as described (31). The densitometric values were corrected by the housekeeping proteins ␣-tubulin or ␤-actin.
Poly(A) ϩ mRNA was isolated from cultured cells using the Dynabeads mRNA DIRECT TM kit (Invitrogen) and reversetranscribed as described previously (47,48). The real time PCR amplification reactions were done using iQ SYBR Green Supermix on a LightCycler instrument (Roche Diagnostics, Vilvoorde, Belgium) and on a Rotor-Gene Q (Qiagen, Venlo, Netherlands), and the concentration of the gene of interest was calculated as copies per l using the standard curve method (47,49). Gene expression values in human and rat cells were corrected by the housekeeping genes ␤-actin and Gapdh, respectively. Expression of these genes was not affected by cytokine treatment (31,50). The primers used in this study are provided in the Table 5. Immunoprecipitation-Proteins from the EndoC-␤H1 cells were collected with cold lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1% Triton X-100, 2 mM MgCl 2 , 25 mM NaF, 2 mM Na 3 VO 4 , 2 mM Na 4 P 2 O 7 , 1 mM sodium Gly-phosphate, and Complete Protease inhibitor mixture, Roche Diagnostics) and then pre-cleared for 1 h at 4°C with Dynabeads Protein G (Novex by Life Technologies, Inc., Oslo, Norway). The same amounts of protein were then incubated overnight at 4°C, either with the anti-IRE1␣ antibody ( Table 2) or normal rabbit IgG (Santa Cruz Biotechnology, sc-2027) as negative controls.
After incubation for 1 h with Dynabeads Protein G, the complexes were washed five times with cold lysis buffer and then resuspended in 5ϫ Laemmli Sample Buffer. Immunoprecipitates and total proteins were run in 8 -12% SDS-PAGE, transferred to a nitrocellulose membrane, and immunoblotted with specific antibodies for the protein of interest (Table 2).
Histology-Pancreas from NOD-SCID (control) mice and diabetes-prone NOD mice were collected at 8, 9, and 13 weeks of age for NOD-SCID and 3, 8, and 12 weeks for NOD mice and processed and stained as described previously (13). After overnight incubation at 4°C with the primary antibody against UBD, slides were double-stained for insulin. Alexa Fluor fluo- rescent secondary antibodies (Molecular Probes-Invitrogen) were applied for 1 h at room temperature (Table 2). After nuclear staining with Hoechst, pancreatic sections were mounted with fluorescence mounting medium (DAKO, Glostrup, Denmark). Immunofluorescence was visualized on a Zeiss microscope (Axio ImagerA1, Zeiss-Vision, Munich, Germany) equipped with a camera (AxioCAM Zeiss) (6).
Ethic Statements-Human islet collection and handling were approved by the local Ethical Committee in Pisa, Italy. Male Wistar rats were housed and used according to the guidelines of the Belgian Regulations for Animal Care. All experiments were approved by the local Ethical Committee. NOD mice have been housed at the KULeuven animal facility since 1989. NOD-SCID mice were also from a stock colony at the KULeuven. All experiments in mice were approved and performed in accordance with the Ethics Committee of the KULeuven, Leuven, Belgium.
Statistical Analysis-Data are expressed as means Ϯ S.E. A significant difference between experimental conditions was assessed by one-way ANOVA followed by a paired Student's t test with Bonferroni correction. p values Ͻ 0.05 were considered statistically significant. The figures are shown as a box plot indicating lower quartile, median, and higher quartile, with whiskers representing the range of the remaining data points, when the number of experiments is Ն4 for each conditions. Alternatively, data are represented as points indicating individual experiments plus the average and the S.E. or bar graph with indicated S.E., when the number of experiments is Ͻ4.

Results
UBD Interacts with IRE1␣-Candidate proteins that interact with IRE1␣ and are modified by pro-inflammatory cytokine treatment in pancreatic beta cells were identified using Array-MAPPIT (13). UBD was chosen for detailed signal transduction studies following additional selection based on the review of the literature.
A binary MAPPIT analysis confirmed the interaction between UBD and IRE1␣ (Fig. 1A). HEK293T cells were cotransfected with an erythropoietin (EPO) receptor-based MAPPIT bait protein containing the cytoplasmic portion of IRE1␣ (pSEL-IRE1␣), a FLAG-tagged UBD prey (pMG1-UBD), and a STAT3-responsive luciferase reporter (pXP2d2-rPAP1). After stimulation with EPO, the UBD prey induced a 10-fold increase in luciferase signal as compared with cells co-transfected with the empty prey vector. The binary MAPPIT analysis was also used to assess the ability of UBD to bind to the kinasedefective K599A IRE1␣ mutant (51) and to the different domains of IRE1␣ (Fig. 1A, kinase domain (Kin) and endoribonuclease domain (Ken)). After stimulation with EPO, the luciferase signals observed in both kinase-defective K599A IRE1␣ mutant and IRE1␣ Kin were comparable with the signal observed with the wild type IRE1␣ when combined with the UBD prey (Fig. 1A). However, the signal obtained with the IRE1␣ Ken was significantly reduced compared with the IRE1␣ wild type, suggesting that the endoribonuclease domain of IRE1␣ is less important for the interaction between IRE1␣ and UBD. These results indicate that the UBD-IRE1␣ interaction is independent of the phosphorylated state of IRE1␣.
The interaction between UBD and IRE1␣ was further validated by endogenous co-immunoprecipitation experiments in INS-1E cells (Fig. 1B) and in human EndoC-␤H1 cells (Fig. 1, C  and D). Using an anti-IRE1␣ antibody, we detected UBD in the immunoprecipitate (IP), as shown by Western blot analysis (Fig. 1, B and C). In the EndoC-␤H1 lysate, UBD bound weakly also to the rabbit IgG (Fig. 1C), but the amount of UBD binding to IRE1␣ was 2-fold higher compared with that bound by the negative control rabbit IgG (Fig. 1D). To confirm the specificity of the IRE1␣/UBD interaction, we have performed additional MAPPIT two-hybrid analyses using irrelevant protein baits in addition to the empty bait (Fig. 2). Using the 10-fold cutoff value that we routinely apply for selecting specific MAPPIT interactions, only PPP3CA would be considered as a specific UBD interactor in addition to IRE1␣. To assess whether the strong signal obtained with IRE1␣ does not result from better bait expression, we measured the interaction of the different baits with UBD in parallel to their interaction with the REM2 protein. The REM2 interaction signal is a measure of the expression level of each individual bait, as it interacts with the cytokine backbone of the two-hybrid bait. The REM2 signal was not stronger for IRE1␣ as compared with the other baits, indicating that IRE1␣ more potently interacts with UBD than any of the other baits tested.
Inflammatory Signals Increase UBD Expression in Pancreatic Islet Cells-We confirmed by real time PCR (RT-PCR) our previous microarray findings (52,53) indicating that pro-inflammatory cytokines induce UBD mRNA expression in rat insulinproducing cells. There was a peak of UBD expression after 16 or 24 h of IL-1␤ ϩ IFN-␥ exposure in INS-1E cells (Fig. 3A) and FACS-purified rat beta cells (Fig. 3C), respectively, with subsequent decrease. This induction was mostly dependent on IFN-␥, with only a minor contribution by IL-1␤ (Fig. 3B). These findings were reproduced in both human islets and in the human beta cell line EndoC-␤H1 exposed for 24 -48 h to IL-1␤ ϩ IFN-␥ (Fig. 3, D and E), confirming our previous RNA sequencing data (25). The observed cytokine-induced increase in UBD mRNA expression was confirmed at the protein level in EndoC-␤H1 cells (Fig. 3, F and G).
To test whether increased UBD expression occurs during beta cell inflammation in vivo, we evaluated UBD expression in islets from diabetes-prone NOD mice. Immunofluorescence staining of pancreatic sections from NOD mice (Fig. 4A) showed UBD expression in insulin-positive cells of immuneinfiltrated islets at 8 and 12 weeks (Fig. 4A, panels h-l), although UBD was not detected in islets from age-matched, genetically similar but not diabetes-prone NOD-SCID mice (Fig. 4B, panels b, c, e, f, h, and i). There was also no detectable UBD expres-sion in islets from BALB/c mice examined at 4, 9 and 13 weeks of age (data not shown). The observed UBD expression in NOD mouse islets seems a direct consequence of islet inflammation (insulitis), because it was not detected in NOD mouse islets at 3 weeks of age (Fig. 4A, panels b and c), an age when islets are not yet invaded by immune cells, and in insulitis-free islets of NOD mice at 8 weeks of age (Fig. 4A, panels e and f).
UBD Inhibition Does Not Affect IRE1␣ Endonuclease Activity in Rat and Human Pancreatic Beta Cells-To understand the function of cytokine-induced UBD expression on IRE1␣ activity, we knocked down (KD) UBD with three different specific  siRNAs. The effects of two of them, siU2 and siU3, are shown in Fig. 5A; both siRNAs induced a Ͼ70% inhibition of UBD expression following cytokine exposure. UBD KD in INS-1E cells (Fig. 5, A-C) did not modify cytokine-induced IRE1␣ endonuclease activity as evaluated by Xbp1 splicing (Xbp1s) (Fig. 5B) and Ins-2 mRNA degradation (Fig. 5C). These data were confirmed in human beta cells, where UBD KD (60 -80% inhibition; Fig. 5D) did not change cytokine-induced Xbp1 splicing (Fig. 5E).

. IL-1␤ ؉ IFN-␥ induce UBD expression in rodent and human beta cells. The expression of UBD was assessed by RT-PCR (A-E) and by Western blot (F and G) in INS-1E cells (A and B), FACS-purified rat beta cells (C), human islet cells (D), and the human EndoC-␤H1 cells (E-G), and normalized by
UBD Modulates IRE1␣-dependent JNK Activation-We next evaluated whether UBD KD interferes with IRE1␣-dependent JNK phosphorylation in beta cells. A decrease of more than 70%  (panels a, d, g, and j, green), UBD (panels b, e, h, and k, red), and Hoechst for nuclear staining (panels c, f, i, and l, blue). Infiltrated lymphocytes, a sign of insulitis, are indicated by white arrowheads (panels i and l). Insulitis is present at 8 and 12 weeks of age but not at 3 weeks. The immunofluorescence analysis shows UBD expression in insulin-positive cells in NOD mice at 8 and 12 weeks (panels i and l, yellow) but not at 3 weeks (panels b and c) or at 8 weeks when insulitis is not present (panels e and f). The images shown are representative of 16 Ϯ 3 islet sections from 3 to 4 different mice per age (A). Pancreatic sections from 8-, 9-, and 13-week-old NOD-SCID mice were stained with antibodies specific for insulin (panels a, d, and g, green), UBD (panels b, e, and h, red), and with Hoechst for nuclear staining (panels c, f, and i, blue). Immunofluorescence analysis indicates absence of UBD expression in insulin-positive cells in NOD-SCID mice (panels c, f, and i, merge). The images shown are representative of 12 Ϯ 3 islet sections from 1 to 3 mice per age (B). Bars, 50 m.
in UBD protein expression was obtained with three different siRNAs (Fig. 6, A and B). The KD of UBD significantly increased cytokine-induced JNK phosphorylation in INS-1E cells after 2 h (Fig. 6, C and D) and 8 h of exposure (Fig. 6, E and F). Double KD of UBD and IRE1␣ reversed the stimulatory effect of UBD silencing on JNK phosphorylation (Fig. 7, A-D), confirming that the UBD effects on JNK are IRE1␣-dependent. This was confirmed by quantification of the area under the curve for phospho-JNK expression (Fig. 7D) and was reproduced using a second siRNA targeting UBD (data not shown).

UBD-dependent JNK Modulation Regulates Cytokine-induced Apoptosis in Pancreatic Beta
Cells-JNK has an important role in cytokine-induced beta cell apoptosis (12, 54 -58). In line with this, the KD of UBD with two different siRNAs (Fig. 8, A, C, and E) significantly increased apoptosis induced by cytokines in INS-1E cells (Fig. 8B) and rat beta cells (Fig. 8, D and F), as evaluated by nuclear dyes. KD of UBD increased basal apoptosis in INS-1E cells (Fig. 8B), but this was not confirmed in primary rat beta cells (Fig. 8, D and F) or human beta cells (see below). UBD KD increased expression of cleaved caspases 3 and 9, suggesting activation of the intrinsic pathway of apoptosis (Fig. 8, G-I). Importantly, KD of UBD by two independent siRNAs (Fig. 9, A and C) also augmented cytokine-induced apoptosis in human islet cells (Fig. 9B) and in the EndoC-␤H1 cells (Fig. 9D).
In additional experiments we observed that KD of IRE1␣ (Figs. 10, A and B, and 11A) or JNK1 (Figs. 10, C and D, and 11B) reversed the pro-apoptotic effect of UBD KD in both control and cytokine-treated INS-1E cells (Fig. 10, E and F) and FACSpurified rat beta cells (Fig. 11, E and F), confirming that cell death induced by UBD silencing in the context of cytokine exposure is mediated by up-regulation of the IRE1␣/JNK1 pathway.
UBD Inhibition Does Not Affect NO Production or Expression of the Chemokine Ccl-2-To investigate whether UBD deficiency inhibits cytokine-induced NF-B activation and reduces the induction of NF-B-regulated genes as has been described in other cell types (61), we measured two NF-B-dependent phenomena, namely induction of NO formation (measured as nitrite accumulation in the medium) and expression of the chemokine Ccl-2 (3). Neither nitrite accumulation nor Ccl-2 expression was modulated by UBD KD in cytokine-treated INS-1E cells (Fig. 12), arguing against a role for NF-B inhibition in the above described effects of UBD inhibition.

Discussion
IRE1␣ and its downstream targets XBP1s and JNK are potential regulatory molecules in the cross-talk between cytokineinduced ER stress and beta cell pro-inflammatory responses and apoptosis (3, 12, 13, 59). The activity of IRE1␣ is regulated by both signals generated from inside the ER and the cytosol (8,11,60). Against this background, we aim to identify novel cytokine-induced IRE1␣-interacting proteins that modulate UPR activation in pancreatic beta cells and thus contribute to inflammation and beta cell death in early T1D. For this purpose, we previously combined a high-throughput mammalian two-hybrid technology, MAPPIT with functional genomic analysis of human and rodent beta cells exposed to pro-inflammatory cytokines (13). This allowed the identification of two proteins of particular interest, namely NMI (13) and UBD (this study).
We presently confirmed by quantitative RT-PCR and Western blot that IL-1␤ ϩ IFN-␥ induce UBD expression in rat INS-1E cells, FACS-purified rat beta cells, human islet cells, and in the human beta cell line EndoC-␤H1. Induction of UBD is mostly an effect of IFN-␥, with a minor contribution by IL-1␤. Expression of the protein in both clonal rat and human beta cells indicates that cytokine-induced UBD expression takes place at least in part at the beta cell level. Immunofluorescence staining of pancreatic sections from NOD mice showed UBD expression in insulin-positive cells at 8 and 12 weeks, although the protein was not detected in islets from age-matched NOD-SCID mice or pre-insulitis NOD mice, confirming up-regulation of UBD in vivo as a consequence of local inflammation (insulitis).
It has been described in other cell types that UBD deficiency abrogates cytokine-induced NF-B activation and reduces the induction of NF-B-regulated genes (61). It is, however, unlikely that these effects explain the present observations, because KD of UBD in beta cells does not affect the expression of downstream targets of NF-B, such as inducible nitric oxide synthase and the chemokine (C-C motif) ligand 2 (CCL2) CCL2 (Fig. 12). Moreover, down-regulation of NF-B has a protective effect in cytokine-induced beta cell apoptosis (62), whereas we presently observed that KD of UBD increases apoptosis in beta cells.
UBD effects seem to be independent of the phosphorylated state of IRE1␣ because UBD interacts with the K599 kinase-defective mutant IRE1␣ in the binary MAPPIT. UBD KD in INS-1E cells and human EndoC-␤H1 cells does not modify cytokine-induced IRE1␣ endonuclease activity as evaluated by Xbp1 splicing (Xbp1s) and Ins-2 mRNA degradation. This last result is in line with the limited capacity of UBD to interact with the endoribonuclease domain (Ken) of IRE1␣ in the binary MAPPIT (Fig. 1A). The main effect induced by UBD KD was to increase JNK activity, with consequent augmentation of apo-  JUNE 3, 2016 • VOLUME 291 • NUMBER 23 JOURNAL OF BIOLOGICAL CHEMISTRY 12049 ptosis. Thus, UBD KD increased cytokine-induced JNK phosphorylation after 2 and 8 h of exposure. Double KD of UBD and IRE1␣ reversed the up-regulation of JNK phosphorylation and the increase in apoptosis, confirming the regulatory function of UBD on IRE1␣-dependent JNK activation in beta cells. Of interest, our previous findings indicate that another cytosolic IRE1␣ regulator, namely NMI, also inhibits JNK induction. The discovery that two cytosolic regulators of IRE1␣ provide specific negative feedback against IRE1␣-induced JNK activation in beta cells (present data and see Ref. 13) confirms the key biological role of JNK for ER stress-induced beta cell apoptosis.

UBD Inhibits IRE1␣/JNK in Beta Cells
Induction of ER stress in human beta cells has basic mechanistic differences as compared with rat beta cells (12). Thus, cytokine-induced ER stress in rat beta cells involves inducible NOS expression, nitric oxide (NO) production, and consequent inhibition of the ER Ca 2ϩ -transporting ATPase SERCA2b (1). However, inhibition of NO formation does not prevent cytokine-induced ER stress activation and apoptosis in human islets (12). Moreover, the expression of SERCA-2b is not modified by cytokines in human islets and in the human EndoC-␤H1 cells. Cytokines induce a biphasic JNK activation in human EndoC-␤H1 cells, with peaks at 0.5 and 8 h and a parallel activation of IRE1␣, which is more marked at 8 -16 h. KD of IRE1␣ decreases JNK phosphorylation in EndoC-␤H1 cells after both 0.5 and 8 h of IL-1␤ ϩ IFN-␥ treatment (12) and in INS-1E after both 2 and 8 h (this study) indicating that the IRE1␣ pathway contributes to both the early and late JNK activation in cytokine-exposed beta cells. Suppression of JNK1 expression by siRNAs partially  protects both rat INS-1E and primary beta cells (this study) and human EndoC-␤H1 cells against cytokine-induced apoptosis, confirming the key role for JNK in cytokine toxicity in human beta cells (12).
The concomitant activation of both JNK and IRE1␣ after cytokine exposure in human beta cells, the decrease on JNK phosphorylation after IRE1␣ KD, and the modulation of the IRE1␣-induced JNK pathway by two cytokine-induced IRE1␣-binding proteins (this study and see Ref. 13), all point to the conclusion that JNK is the major IRE1␣-regulated pathway in the cross-talk between ER-stressed/apoptotic beta cells and the immune system.
The Ubd gene lies in the telomeric region of MHC class I, a major susceptibility locus for T1D. Polymorphisms in the Ubd gene have been associated with autoimmune diabetes in rat and human (26 -29), but the role of these polymorphisms in the 48 h later, cells were left untreated or incubated with IL ϩ IFN for 16 h. The KD of JNK1 was evaluated by Western blot (C and D) and apoptosis by propidium iodide/Hoechst staining (F). One representative Western blot of five independent experiments (C) and the densitometry results for JNK1 expression (D) are shown. ␣-Tubulin expression was used as loading control (A-D). Results are represented as a box plot indicating lower quartile, median, and higher quartile, with whiskers representing the range of the remaining data points in B, D, E, and F. *, p Ͻ 0.05, and ***, p Ͻ 0.001 versus siC untreated; #, p Ͻ 0.05; ##, p Ͻ 0.01, and ##, p Ͻ 0.001 as indicated by bars. One-way ANOVA followed by Student's paired t test with Bonferroni correction.
expression of UBD remains to be clarified. This study indicates that UBD provides a negative feedback to a key pro-apoptotic branch of the UPR in the beta cells, namely the activation of JNK. During a protracted autoimmunity assault, this protective feedback will be eventually overrun, but it is conceivable that both UBD (this study) and NMI (13) play an important protec- FIGURE 11. KD of IRE1␣ or JNK1 protects rat beta cells from cytokine-induced apoptosis in the context of UBD-deficiency. FACS-purified rat beta cells were transfected with siControl (siC), siIRE1, siU2, or co-transfected with siIRE1 ϩ siU2 (A, C, and E). 48 h later, cells were left untreated or incubated with IL-1␤ ϩ IFN-␥ (ILϩIFN) for 48 h. The KD of IRE1␣ (A) and UBD (C) was assessed by RT-PCR and normalized by the housekeeping gene Gapdh. Apoptosis was evaluated by propidium iodide/Hoechst staining (E). FACS-purified rat beta cells were transfected with siC, siJNK1, siU2, or co-transfected with siJNK1 ϩ siU2 (B, D, and F). 48 h after the transfection, the cells were left untreated or incubated with IL ϩ IFN for 48 h. The KD of JNK1 (B) and UBD (D) was assessed by RT-PCR and normalized by the housekeeping gene Gapdh. Apoptosis was evaluated by propidium iodide/Hoechst staining (F). Results were normalized against the highest value in each independent experiment, considered as 1 (A-D). *, p Ͻ 0.05; **, p Ͻ 0.01, and ***, p Ͻ 0.001 versus siC untreated; #, p Ͻ 0.05; ##, p Ͻ 0.01, and ###, p Ͻ 0.001 versus siC treated; @, p Ͻ 0.05; @@, p Ͻ 0.01, and @@@, p Ͻ 0.001 as indicated by bars. One-way ANOVA followed by Student's paired t test with Bonferroni correction. Data shown are mean Ϯ S.E. of six independent experiments. tive role during putative transient periods of mild inflammation in the islets, preventing excessive beta cell lost.