C57BL/6J wild-type (WT) mice and Akita mice were obtained from Jackson laboratory (Bar Harbor, ME). The genotyping of Akita mice was confirmed using tetra-primer ARMS-PCR approach (46). Mice were housed on a 12 h light (6:00 a.m. to 6:00 p.m.)−12 h dark (6:00 p.m. to 6:00 a.m.) cycle at an ambient temperature of 22°C and fed normal chow diet and water ad libitum. All procedures involving animals were performed in accordance with the protocol approved by the Institutional Animal Care and Use Committee of the University of Oklahoma Health Science Center. All experiments were performed with age-matched female mice.

Akita mice at 5-6 weeks of age were randomly grouped for the injection i.p. with either vehicle (n = 9 mice), STF (10 mg/kg body weight; 2 mg/ml in 10% DMSO in saline buffer; n = 9 mice) or 4μ8C (10 mg/kg of body weight) once daily. These doses were chosen based on previously reported efficacy shown on mice via IP injection (47, 48). Compounds were dosed approximately 3−4 h before the initiation of the dark cycle (2−3 p.m.). Blood glucose levels were measured using the OneTouch Ultra2 glucometer after fasting for 6 h. Body weights were measured weekly. At the end of treatment, mice were fasted for 4 h and euthanized, and pancreata were removed and weighted. A tail end portion of the pancreata was saved for insulin and proinsulin content measure while the remaining pancreata were formalin fixed and paraffin-embedded.

Intraperitoneal glucose tolerance test (ipGTT) and intraperitoneal insulin tolerance test (ipITT) were performed after 16-h and 4-hour fasting, respectively. Blood glucose levels were measured at 0, 15, 30, 60, and 120 minutes after intraperitoneal administration of glucose (1.5 g/kg body weight) for ipGTT or insulin (0.75 IU/kg body weight) for ipITT.

Islets were isolated using the standard collagenase digestion method. Briefly, the common bile duct was cannulated and distended with Collagenase P (0.5 mg/ml, Sigma-Aldrich, USA) in 1x Hank’s balanced salt solution. Pancreata were removed and incubated in water bath at 37C for 25 m. Islets were separated using Histopaque-1077 (Sigma-Aldrich, USA) and cultured overnight at 37°C in RPMI1640 media containing 10% FBS.

Islets isolated from 6-week Akita or WT B/6J mice were treated with STF 20 μM or DMSO vehicle (0.1%) for 3 h. Proteins were extracted with RIPA buffer (10 mM Tris pH 7.4, 150 mM NaCl, 0.1% SDS, 1% NP40, 2 mM EDTA) plus protease inhibitor/phosphatase inhibitor cocktail (Sigma-Aldrich) and centrifuged at 4°C for 10 min at 10,000 g. Total protein concentration in the cell lysate was determined by BCA. Samples of ~ 20 μg protein prepared in Laemmli sample buffer without (non-reducing) or with (reducing) 5% b-mercaptoethanol were resolved on 4-12% Bis-Tris NuPAGE gels (Invitrogen) at 100 V for 60 min. The nonreducing gels were incubated in 25 mM dithiothreitol (DTT) solution for 10 min at room temperature before being transferred to PVDF membranes for 25 min at 4°C at 40 V. Membranes were probed with anti-proinsulin antibody (CCI-17, NOVUS) and HRP-conjugated secondary antibodies (1:3000; Santa Cruz Biotechnology, CA, USA).

Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. 2 μg of total RNA was reverse transcribed using Superscript kit (Invitrogen). Real-time PCR was performed with a CFX96 Real-Time PCR detection system (Bio-Rad, CA) using SYBR Select Master Mix (Applied Biosystems, CA). Relative mRNA levels were normalized against Cyclophilin A. Primers used are shown in Table S1 . Regular RT-PCR was performed and the products was resolved by agarose gel electrophoresis. The unspliced and spliced XBP1 mRNA levels were quantified using ImageJ software (National Institutes of Health, Bethesda, MD).

20 primary islets isolated from Akita mice treated with STF or vehicle were seeded in 96-well plates overnight and then incubated in fresh KRBH buffer (115 mM NaCl, 5 mM KCl, 24 mM NaHCO₃, 2.5 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, 2% w/v BSA, pH 7.4) containing 2.5 mM glucose for 1 h. Islets were incubated for an additional hour in KRBH buffer containing 2.5 or 16.7 mM glucose. Secreted insulin and proinsulin levels were measured with insulin ELISA kits (ALPCO, Salem, NH) and Rat/Mouse Proinsulin kit (Mercodia), respectively and normalized to total protein of cell lysates.

Pancreatic tissues or islets were incubated and homogenized in 1.5% HCl in 70% EtOH overnight at -20C, and the solution neutralized with equal volume of 1M Tris pH 7.5. Insulin and proinsulin were measured by ELISA kits as outlined above and normalized to weights of pancreas for pancreatic tissues or to protein levels for islets.

Pancreata were fixed in formalin and paraffin-embedded. 6-8 slide sections on average from each mouse for all mice were sectioned with the separation at 150 μm increments. Images covering the entire tissue sample were captured in each section. The entire pancreas tissue, glucagon⁺, and insulin⁺ areas in each image were measured using ImageJ software. Relative β-cell area = sum of all islet β-cell areas/sum of the total pancreatic area, and normalized against the β-cell area of WT B/6J mice (set as 1). All images were taken with an Olympus FV1000 confocal microscope and quantified with Image-J histogram software.

Antibodies used for staining: GP anti-insulin antibody (A0564, 1:500; Dako), mouse anti-glucagon antibody (G2654, 1:500; Sigma), mouse anti-caspase 3 (cat# 9446, 1:500, CST), rabbit anti-Ki67 (Ab15580, 1:250, Abcam), mouse anti-proinsulin (GS-9A8, 1:100, DSHB), Anti-4-Hydroxynonenal [HNEJ-2] (ab48506, Abcam), DAPI (0.5 μg/mL), and Alexa Fluor 488-, 555-, and 647-conjugated secondary antibodies (Jackson ImmunoResearch).

TUNEL staining was performed together with antibodies as above in pancreatic sections with In Situ Cell Death Detection Kit-Fluorescein (Roche) according to the manufacturer’s instructions.

Isolated mouse islets were fixed with 0.1 M sodium phosphate buffer (pH 7.2) containing 2% glutaraldehyde and 2% paraformaldehyde for 1 h, then exposed to 2% osmium tetroxide, stained with 2% uranyl acetate, dehydrated with ethanol, and embedded in Epon (TAAB). Ultra-thin sections were stained with uranyl acetate and lead citrate, and images were recorded with a Hitachi H-7600 transmission electron microscope (Hitachi).

Data were analyzed using the unpaired two-tailed Student’s t-test or one-way ANOVA for multiple comparisons. All values are reported as mean ± SEM and p<0.05 was considered statistically significant.

The C96Y missense mutation in the Ins2 gene in Akita mice causes mutant proinsulin protein misfolding that is responsible for ER stress (35). Previously, the ER stress response markers PERK and ATF6 have been reported to be up-regulated in in vitro β-cell lines carrying the Ins2^Akita/+ mutation or in islets freshly isolated from Akita mice (44, 49, 50). Consistent with this, we observed that the mRNA levels of the PERK pathway genes ATF4 and CHOP and the ATF6 target gene Bip were up-regulated in islets freshly isolated from Akita mice (herein terms Akita islets) over the age of 3 weeks ( Figures S1A–C ). However, how IRE-1α responds to this mutation in β-cells in vivo is unclear. Earlier studies using in vitro β-cell lines yielded controversial results, with one report showing an activation of the IRE-1α-XBP1 pathway in Ins2^Akita/+ β-cell lines (51) and another report showing a down-regulation of IRE-1α activity in stable β-cell lines expressing Ins2^Akita/+ mutation (52). To investigate the in vivo IRE-1α activity in Akita islets, we first examined the splicing of Xbp1 mRNA, a direct target of IRE-1α RNase. We observed a gradual increase in Xbp1 mRNA splicing in Akita islets from age of 2 weeks onwards compared to compared to the age-matched WT mice ( Figures 1A, A’ ), as assessed by the electrophoretic separation of RT-PCR products. Similarly, quantitative RT-PCR also showed that spliced Xbp1 (Xbp1-s) mRNA levels significantly increased in the Akita islets over a period of 12 weeks while total Xbp1 (Xbp1-t) mRNA levels increased only slightly during the same period ( Figures 1B, C ). Second, we investigated the transcription of several XBP1 target genes EDEM1 and P58, and observed a marked upregulation in their mRNA levels in Akita islets over WT islets ( Figures 1D, E ), as assessed by qRT-PCR. Third, as IRE-1α hyperactivation is associated with activation of IRE1-dependent decay of mRNA (RIDD) in which IRE1 cleaves mRNAs, we analyzed the mRNA levels of Blos1 and Col6a1, two typical RIDD targets, by qRT-PCR. We observed that Blos1 and Col6a1 mRNA levels decreased progressively in the Akita islets from 3-week old onwards ( Figures 1F, G ). Together, our results demonstrate that IRE1α activity was already elevated at around 2 weeks of age, prior to the development of hyperglycemia in Akita mice, and continued to elevate until the Akita mice developed overt diabetes.

The results presented above in conjunction with previous observations that the overexpression of IRE-1α led to cell death in transfected cells (10, 53, 54) suggest that inhibiting IRE-1α may protect β-cells from Akita mutation-induced dysfunction and death and ameliorate the diabetic condition in Akita mice. We therefore investigated the effect of pharmacological inhibition of IRE-1α activity on β-cell and diabetic conditions in Akita mice. We treated Akita mice with a specific IRE-1α RNase inhibitor STF-083010 (STF, 10 mg/Kg of BW via IP injection, a dose previously shown to significantly inhibits IRE-1α RNase activity in vivo) (47, 55) for 6 weeks and detected a gradual and significant dampening of blood glucose levels over the treatment period ( Figure 2A ). In contrast, the vehicle-treated Akita mice continued to develop increasing hyperglycemia throughout the treatment period to reach up to 400mg/dL ( Figure 2A ). Furthermore, STF treatment significantly improved glucose tolerance and decreased AUC (area under the curve) in Akita mice compared to vehicle group (p<0.05; Figures 2B, C ). On the other hand, the STF- and vehicle-treated Akita mice displayed comparable body weight ( Figure 2D ) and insulin sensitivity ( Figures 2E, F ), suggesting that STF lowers blood glucose levels not by altering insulin sensitivity. Finally, serum insulin levels in the STF-treated Akita mice were markedly increased compared to that of vehicle group; in particular, at 30 min after glucose injection (vehicle 0.2± 0.1 ng/ml vs. STF 1.5± 0.3 ng/ml; p<0.001) ( Figure 2G ). Together, these results indicate that STF treatment significantly alleviates the diabetic conditions in Akita mice.

To investigate whether the STF amelioration of diabetic conditions in Akita mice is due to the inhibition of IRE1α activity in islets, we first examined the status of IRE1α-mediated Xbp1 mRNA splicing in islets from STF-treated Akita mice. As shown in Figure 3A , the level of XBP1-s mRNA was significantly reduced in islets from STF group relative to vehicle group, as assessed by RT-PCR followed by electrophoretic separation. This result was corroborated by qRT-PCR using Xbp1 splicing-specific primers ( Figure 3B ), whereas XBP1-t mRNA levels were only moderately affected ( Figure 3C ). We further detected that the mRNA levels of XBP1 target genes EDEM1, P58, and Bip were highly suppressed in islets from STF-treated mice ( Figures 3D–F ). Moreover, the transcript levels of generic RIDD targets of IRE1-α—Col61a and Blos1— were suppressed in the Akita islets; however, their levels were significantly reversed in islets from STF-treated mice ( Figures 3G, H ). In addition, under ER stress, insulin 1 and insulin 2 mRNAs are known to be cleaved directly by IRE1-α RNase activity as β-cell-specific RIDD targets (10, 56). Both insulin 1 and insulin 2 mRNAs were expectedly down-regulated in Akita islets ( Figures 3I, J ). However, STF treatment significantly reversed their expression in the Akita islets ( Figures 3I, J ). Together, our results reveal that STF treatment suppresses Akita mutation-induced IRE-1α activation in Akita islets. Notably, STF at the dose of 10 mg/kg BW reversed hyperactivated IRE1α activity to normal level but did not completely abolish IRE-1α activity ( Figures 3A, B ). Interestingly, although STF is known to inhibit IRE-1α activity only, our results revealed that STF also suppressed the Akita mutation-induced increase in ATF4 and CHOP mRNA levels, key components of the PERK pathway ( Figures S2A, B ). The effect of STF on PERK pathway could be due to cross-talks among the branches of UPR under in vivo conditions as previously reported (57, 58).

As diabetes progression in Akita mice is associated with gradual β-cell loss and IRE-1α is activated before the onset of diabetes in Akita mice, we investigated whether the STF improvement of diabetic conditions is associated with the protection of islet β-cells in the Akita mice. In pancreatic sections, Akita islets not only exhibited reduced β-cell mass but also significantly decreased insulin staining intensity in existing β-cells ( Figures 4A, B, D, E ), indicative of β-cell loss and dysfunction. In contrast, the STF-treated Akita mice possessed approximately twice the β-cell mass and significantly higher insulin staining intensity compared to that in vehicle-treated mice ( Figures 4B–E ). On the other hand, the α cell numbers, marked by glucagon immunostaining, remained comparable between STF and vehicle groups ( Figures 4A–C ). Consistent with these results, total pancreatic insulin content as quantified by ELISA was markedly higher in STF-treated Akita mice than their vehicle-treated counterparts ( Figure 4F ). To determine whether the increase in β-cell mass by STF could be attributed to an inhibition of islet cell apoptosis, we assessed apoptosis using TUNEL staining, a marker for apoptosis. An increase in TUNEL ⁺ insulin⁺ cells was observed in the vehicle-treated Akita mice relative to WT mice ( Figures 4G–J ). However, TUNEL staining was considerably reduced in STF-treated Akita to a level comparable to that of WT ( Figures 4G–J ). Treatment with STF also significantly reduced the number of CASP⁺ (a critical protein in the execution of apoptosis) in insulin⁺ cells of Akita islets compared to that in vehicle-treated islets ( Figures S3A–D ). In contrast, STF treatment appeared not to affect β-cell proliferation as the frequency of Ki67⁺ insulin⁺ cells remained comparable between the STF- and vehicle-treated mice ( Figure S4 ). Lastly, the expression of apoptotic effector genes BAX and Bak1, pro-apoptotic inducer gene p53, and negative cell-cycle regulator p21 were significantly suppressed in the islets of STF-treated Akita mice ( Figures S5A–D ). In sum, these results indicate that inhibition of increased IRE1-α activity by STF suppresses β-cell apoptosis in Akita mice which in turn leads to a preservation of β cell mass.

We interrogated whether STF also improves β-cell function. Indeed, our observations that STF heightened insulin production in β-cells ( Figures 4C–F ) and increased serum insulin levels ( Figure 2G ) suggest an improvement in Akita β-cell function. To further interrogate this, we assessed the glucose-stimulated insulin secretion in islets and found that insulin secretion was significantly higher under both basal (2.5 mM glucose concentration) and stimulated (16.7 mM glucose concentration) conditions in islets isolated from STF-treated Akita mice compared to vehicle-treated group ( Figure 5A ). Next, as high glucose and the ensuing ER stress in β-cells down-regulate the expression of β-cell-specific transcription factors (23, 59–62), which are essential for the maintenance of normal β-cell function (63–65), and there is evidence indicating β-cell dedifferentiation in Akita mice (66), we investigated the effect of STF treatment on their expression levels in Akita β-cells. As expected, the mRNA levels of several β-cell-specific transcription factors (Pdx1, MafA, NeurD1, and Nkx6.1) were significantly down-regulated in Akita islets compared to WT islets ( Figures 5B–E ). Notably, their reduced expression was markedly reversed in islets of Akita mice treated with STF ( Figures 5B–E ).

Akita mutant proinsulin tends not only to misfold but also forms heterogeneous complex with WT proinsulin, thus entrapping WT proinsulin in the ER (34, 35, 67) and limiting bioactive insulin production and secretion, thus leading to ER stress and β-cell dysfunction and death (44, 68). The disturbed ER environment in turn further exacerbates proinsulin misfolding (69). We therefore investigated whether the increased insulin production ( Figures 2G , 4C–F , 5A ) seen with STF treatment is associated with reduced proinsulin levels. As expected, proinsulin content was dramatically and significantly increased in the Akita mouse pancreas relative to WT pancreas, but was reversed to normal level in the STF-treated Akita pancreas ( Figure 5F ). Of note, this effect of STF on proinsulin is opposite to that seen for insulin ( Figure 4F ). While the STF suppression of proinsulin level in Akita islets can be interpreted as an outcome of STF-improved ER environment which permits more proinsulin conversion to mature insulin; it is also possible that this effect is mediated through increased proinsulin release from β cells or through modulation of proinsulin misfolding. To address whether STF increases proinsulin secretion, we analyzed proinsulin levels in Akita islets from Akita mice treated with STF or vehicle. We found that proinsulin content (cell lysate) and secretion (supernatant) were both attenuated in STF group, under either basal (2.5 mM) or high (16.7 mM) glucose concentration ( Figures 5G, H ). To address whether STF affects proinsulin misfolding, we examined proinsulin folding status in Akita islets under nonreducing condition. Consistent with previous reports (42), a significant port of the proinsulin in nonreduced lysates of Akita islets was detected as high molecular weight oligomers relative to WT islets ( Figure 5I ). STF treatment of Akita islets showed no apparent effect on proinsulin oligomers ( Figure 5I ).

Finally, as mature insulin is formed from proinsulin processed in the Golgi complex and stored in secretory granules for release, we examined the effect of STF on insulin secretory granules and ultrastructure of islet β-cells using transmission electron microscopy. Our results showed that whereas there was marked reduction in the number of dark electron dense-core granules (mature insulin granules) and increase in the number of light or “gray” electron dense-core granules (immature insulin granules) in Akita β-cells compared to WT β-cells ( Figures 5J–L ), β-cells in the STF-treated Akita mice exhibited a marked increase in dense-core insulin granules and ( Figure 5L ), similar to those seen in the WT islets ( Figure 5J ). Together, these results demonstrate that STF normalization of IRE-1α activity facilitates insulin granule formation.

ER stress has been shown to cause and potentiate inflammation and oxidative stress that cooperatively contribute to ER stress-mediated cell death (70). We therefore investigated whether STF treatment affects these processes in the Akita islets. We found that mRNA levels of the ER stress-associated pro-inflammatory cytokine genes IL-1β, IL6, and TNF were increased in the Akita islets compared to WT islets ( Figures 6A–C ). We also detected increased transcript levels of MCP1 and CD68 ( Figures 6D, E ), markers that are highly expressed in tissue monocytes and macrophages, respectively. Strikingly, STF corrected the mRNA levels of these genes to normal levels in the Akita islets ( Figures 6A–E ).

Next, we assessed the effect of STF on oxidative stress in Akita islets. We observed an obvious nuclear accumulation of the lipid peroxidation product 4-hydroxynonenal (4-HNE), a marker of oxidative stress, in the Akita islets compared to WT islets ( Figures 6F, G ). In addition, the mRNA levels of several antioxidant genes, including those encoding the mitochondrial uncoupling protein 2 (UCP2), glutathione peroxidase 1 (Gpx1), superoxide dismutase 1 (Sod1), catalase (CAT), and heme oxygenase 1 (Hmox1), were up-regulated in the Akita islets relative to WT islets ( Figures 6I–M ), reflecting a compensatory mechanism of anti-oxidation through the antioxidant gene up-regulation (71, 72). Notably, the nuclear accumulation of 4-HNE and up-regulation of antioxidant genes were abolished in the islets of Akita mice treated with STF ( Figures 6G, H ).

To ascertain that STF improves diabetic conditions of Akita mice via inhibition of IRE1α, we utilized another structurally distinct IRE1α RNase inhibitor 4μ8C (73) for the efficacy studies. Treatment with 4μ8C improved fasting blood glucose levels in Akita mice while vehicle-treated Akita mice showed a progressive rise in blood glucose level ( Figure 7A ). 4μ8C treatment also significantly improved glucose tolerance in Akita mice ( Figures 7B, C ), with no apparent difference in body weight ( Figure S6 ) or insulin tolerance ( Figure 7D ), compared to vehicle-treated Akita mice. Moreover, there was a marked increase in serum insulin levels and pancreatic insulin content in Akita group treated with 4μ8C ( Figures 7E , S7 ). 4μ8C treatment also significantly preserved the β-cell area and restored insulin staining intensity in β-cells ( Figures 7F–J ). Furthermore, 4μ8C treatment significantly alleviated the Akita mutation-induced increase in XBP1 splicing ( Figures S8A, B ). 4μ8C also attenuated the heightened mRNA levels of the XBP1-s target genes Grp94, Bip and P58 in Akita islets ( Figures S8C–E ) and significantly reversed the repressed levels of RIDD target mRNAs Blos1, Col6a1, and INS1 in Akita islets ( Figures S8F–H ). Lastly, 4μ8C attenuated the increased levels of the PERK pathway genes ATF4 and CHOP in Akita islets ( Figures S8I, J ). Together, we conclude that both STF and 4μ8C are able to correct the diabetic conditions of Akita mice likely through the inhibition of Ire1α RNase activity.