The prevalence of diabetes is correlated with that of obesity worldwide and has reached epidemic proportions. Type 1 diabetes (T1D) is a metabolic disease that occurs when the body is unable to produce enough of the hormone insulin, due to the destruction of insulin-producing β-cells in pancreatic islets (Apostolou et al., 2003). Streptozotocin (STZ) is widely used to induce experimental T1D in rodent models (Szkudelski, 2001) and does so by inducing pancreatic β-cell apoptosis (Hayashi et al., 2006; Le May et al., 2006). Also, STZ-induced hyperglycemia has been reported to accelerate endoplasmic reticulum (ER) stress-mediated pancreatic β-cell apoptosis in a murine model (Ahn et al., 2015). In addition, recent studies indicate ER stress plays a crucial role in pancreatic β-cell apoptosis, and have established that ER stress-mediated apoptosis is responsible for initiating β-cell apoptosis in T1D (Eizirik et al., 2008; Oyadomari et al., 2002).

ER stress response is mediated by three sensors, that is inositol-requiring protein-1α (IRE-1α), activating transcription factor 6 (ATF6), and PKR-like ER kinase (PERK), located on the ER membrane. ER stress is transduced by unfolded protein responses (UPRs) initiated by the dissociation of Grp78 from these three ER membrane sensors (Oyadomari and Mori, 2004; Sano and Reed, 2013; Szegezdi et al., 2006). First, activation of IRE1α causes the recruitments of TRAF2 and ASK1, which lead to the downstream activation of JNK. Second, activation of ATF6 causes its translocation to Golgi where its cytosolic domain is sequentially cleaved by two proteases, S1P and S2P, to produce the active form (Sano and Reed, 2013). Third, PERK activation leads to the phosphorylation of eukaryotic initiation factor 2 (eIF2α), which inhibits protein synthesis. Phosphorylation of eIF2α also enables the preferential translations of UPR-dependent genes, such as, activating transcriptional factor 4 (ATF4). The next phase of ATF4-mediated C/EBP homologous protein (CHOP) activation can also induce apoptosis via a transcriptional route (Oyadomari and Mori, 2004; Szegezdi et al., 2006). In particular, chronic and persistent ER stress induces the degenerative ER stress pathway and leads to apoptosis, and a growing number of authors have suggested CHOP is a key regulator of this degenerative pathway. In addition, ER stress is considered to promote the transcription of CHOP (Oyadomari and Mori, 2004; Szegezdi et al., 2006), an important regulator of ER stress-mediated apoptosis and metabolic gene expression (Oyadomari and Mori, 2004). Interestingly, CHOP knockout mice have been reported to be protected from renal (Lakshmanan et al., 2011), pancreatic (Song et al., 2008), heart (Nam et al., 2015), and liver (Tamaki et al., 2008) dysfunction in murine models of metabolic disease, indicating CHOP deficiency improves pancreatic β-cell survival and ameliorates diabetic complications.

Extracellular-signal-regulated kinase 5 (ERK5; also called big MAP kinase 1) is a member of the MAPK family and is activated by MEK5 upstream kinase via phosphorylation of TEY motif and C-terminal transcriptional domain. In a previous study, we reported diabetic stimuli reduced the ERK5 transcriptional activity through ERK5-SUMOylation in endothelial cells, and that ERK5-SUMOylation was increased in a mouse model of STZ-induced diabetic cardiac dysfunction induced by myocyte apoptosis (Le et al., 2012; Woo et al., 2008). Recently, it was reported that high fat diet Adipo-ERK5 deficient mice exhibited increased body weight as compared with control littermates, thus suggesting ERK5 plays a protective role of ERK5 in metabolic disease (Shishido et al., 2008; Zhu et al., 2014). However, the role played by ERK5 in the diabetic pancreas has not been addressed. In the present study, we examined the mechanisms underlying the ERK5-mediated anti-diabetic effects of pancreatic β-cells in a mouse model of type 1 diabetes.

Reagents and antibodies

Chemicals and materials were obtained from commercial sources as follows: streptozotocin (Sigma, USA) and BIX02189 (a specific inhibitor of MEK5 and ERK5 (Tatake et al., 2008) were purchased from Selleck Chemicals (USA). Antibodies were purchased from the following vendors: KDEL (Enzo Life Sciences, Lörrach, Germany), GADD, p-PERK, PERK, ATF4, and HA-probe (Santa Cruz, USA), p-elF2 α, eIF2α, p-ERK5, ERK5, caspase-3, and PARP-1 (Cell Life Sciences, USA) and α-tubulin (Sigma, USA).

Cell culture

INS-1 cells (a rat insulinoma cell line) from passage 15 to 30 were seeded at equal densities and cultured in RPMI 1640 medium supplemented with 10% FBS, 50 U/ml penicillin, 50 μg/ml streptomycin, and 50 μM β-mercaptoethanol. Cells were incubated in a humidified 5% CO₂/95% air atmosphere at 37°C and grown until 80% confluent.

Protein extraction and Western blot analysis

Cells were lysed in RIPA buffer (pH 7.4) supplemented with 50 mM Tris, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, and 0.01 mM PIC (protease inhibitor cocktail). For Western blotting, proteins were separated by SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes, which were immunoblotted with indicated primary antibodies and then with corresponding secondary antibodies. Signals were visualized using chemiluminescence detection regents (Millipore, USA), according to the manufacturer’s instructions.

Quantitative real time RT-PCR

The mRNA levels were determined by quantitative real time RT-PCR (qRT-PCR). Briefly, total RNA was isolated using TRIzol® Reagent (Invitrogen, USA), and reverse transcription reaction was conducted using TaqMan reverse transcription reagents (Applied Biosystems, USA), according to the manufacturer’s instructions. qRT-PCR was conducted with 1 μl of template cDNA and Power SYBR Green (Applied Biosystems) in an ABI PRISM 7500 unit (Applied Biosystems). Quantification was carried out using the efficiency-corrected ΔΔCq method. The primers used to amplify DNA sequences were as follows: rat CHOP, forward 5′-AGCTGAGTCTCTGCCTTTCG-3′ and reverse 5′-AATCCGCCCTCTCTTTTAGAG-3′; rat GAPDH, forward 5′-TGGCACCCAGCACAATGAA-3′ and reverse 5′-CTAAGTCATAGTCCGCCTAGAAGCA-3′ (Liu et al., 2013).

Small interfering RNA (siRNA) and Adenovirus vector

Cells were transiently transfected with control siRNA or siRNA against ERK5 or CHOP using Lipofectamine 2000 (Invitrogen, USA), according to the manufacturer’s instructions. Rat ERK5 siRNA and rat CHOP siRNA were purchased from Bioneer (Korea). The targeting sequences of the siRNAs were as follows: rat CHOP siRNA, sense, 5′-CUGGGAAACA GCGCAUGAA-3′; antisense, 5′-UUCAUGCGCUGUUUCCCAG-3′ (He et al., 2012) and rat ERK5 siRNA, 5′-GAAAGGGTGC GAGCCTATAUU-3′ and human ERK5 siRNA, 5′-GGGCCTAT ATCCAGAGCUU-3′³⁵. Non-specific control siRNA (Bioneer) was used as a negative control. The constitutively active form of MEK5α (CA-MEK5α) was cloned as previously described (Le et al., 2013).

Animal experiments

Male C57BL/6J mice (WT, 8 week old, Jackson Laboratories, USA) and male CHOP^−/−mice (C57BL/6J background, 8 week old, Jackson Laboratories) were injected IP with 150 mg/kg body weight STZ. To determine the effect of ERK5 inhibitor, 10 mg/kg body weight of BIX02189 in 100 μL saline was injected IP every 2 days for 6 days. To measure blood glucose levels, mice were fasted for 12 h with free access to water. Blood samples were obtained by tail tip puncture using a glucometer (Accu-Chek; Roche, USA). All animal experiments were handled in accordance with the protocol approved by the Institutional Animal Care and Use Committee at Yeungnam University College of Medicine, Daegu, Republic of Korea (The approved number of IACUC: YUMC-AEC2015-030).

Immunostaining and TUNEL assay

For immunohistochemical examinations, we used 4% paraformaldehyde fixed, paraffin-embedded pancreas tissue sections. Sections were stained with a rabbit anti-insulin polyclonal antibody (Santa Cruz, USA) or a mouse anti-Chop monoclonal antibody (Santa Cruz, USA) as the primary antibody; Cy3-labeled anti-rabbit IgG (Amersham Pharmacia Biotech, USA) or Alexa Fluor 488–labeled anti–mouse IgG (Molecular Probes Inc., USA) were used as the secondary antibody. Apoptosis was measured by TUNEL (terminal deoxyribonucleotide transferase (TdT)-mediated dUTP nick end labeling) staining, which targets in situ DNA fragmentation, using the In Situ Cell Death Detection Kit (Roche, USA) as described previously (Nam et al., 2015).

Statistical analysis

Results in the bar graph are expressed as means ± SD. Statistical evaluations were performed using and ANOVA and Student t-tests wherever indicated. All results are presented as the means of at least three independent experiments, and probability values (p values) of < 0.05 were considered significant.

ERK5 inhibition exacerbated STZ-induced hyperglycemia and pancreatic apoptosis

Recent reports suggest that ERK5 activation protects against diabetic stress-induced endothelial inflammatory responses and diabetic cardiac dysfunction after myocardial infarction (Le et al., 2012; Shishido et al., 2008). However, the mechanism by which ERK5 affects hyperglycemia and pancreatic β-cell apoptosis has not been addressed. To assess the role of ERK5 in diabetes, we used BIX02189, a specific ERK5 inhibitor, in a STZ-induced mouse model of type 1 diabetes. Fasted blood glucose levels were found to progressively increase in STZ treated mice (Fig. 1A). In STZ treated mice co-treated with BIX02189, a significant increase in blood glucose level was observed at only 2 days post-treatment, indicating ERK5 inhibition exacerbates STZ-induced hyperglycemia. During the 6-day experimental period, the body weights of STZ treated mice were lower than those of treatment naïve controls, and body weight losses by STZ plus BIX02189 treated mice were significantly greater than those of STZ treated mice, suggesting ERK5 inhibition exacerbated hyperglycemia and body weight loss in our STZ-induced model (Fig. 1B).

We next investigated whether ERK5 inhibition increases β-cell apoptosis induced by STZ. Immunoblotting data showed that cleaved PARP-1 and caspase-3 levels were slightly higher in STZ treated mice than in treatment naïve controls and that these levels were higher in STZ plus BIX02189 treated co-treated mice than in STZ treated mice (Fig. 1C). Consistent with our immunoblotting observations, numbers of TUNEL-positive cells in pancreatic islets were significant higher in STZ plus BIX02189 co-treated mice than STZ treated mice (Fig. 1D). These results suggest that ERK5 inhibition augments STZ-induced β-cell apoptosis, hyperglycemia, and body weight loss.

ER stress was involved in the ERK5 inhibition-mediated exacerbation of STZ-induced pancreatic β-cell apoptosis

It has been established that ER stress contributes to the pathogenesis of diabetes and β-cell apoptosis (Oyadomari et al., 2002; Papa, 2012), and thus we evaluated UPRs in pancreas tissues from mice treated with STZ or STZ plus BIX02189. As shown in Fig. 2A, protein expression levels of GRP78/94, CHOP, ATF4, phospho-PERK, and phosphoeIF2 α were significantly higher in STZ treated mice than in treatment naïve controls, and significantly higher in STZ plus BIX02189 co-treated mice than in STZ treated mice.

Immunohistochemical analysis of pancreatic tissues showed that CHOP was expressed in insulin-producing cells (Fig. 2B). CHOP expression was moderate in islets from STZ treated mice, but was significantly higher than this in STZ plus BIX02189 co-treated mice. These results suggested that ERK5 inhibition-induced β-cell apoptosis involved the participation of CHOP.

ERK5 regulates STZ-induced apoptosis in a UPRdependent manner

To confirm the role of ERK5 in the regulation of UPRs and apoptosis in pancreatic β-cells, we examined the effect of ERK5 inhibitor and the siRNA knockdown of ERK5 in INS-1 cells. It has been reported that STZ causes apoptosis at low doses but necrosis at high doses in INS-1 cells (Saini et al., 1996) and thus, we selected a STZ concentration that induced INS-1 cell apoptosis (data not shown). Western blot analysis revealed that STZ inductions of the cleaved forms of PARP-1 and caspase-3 were significantly increased in the presence of BIX02189 (Fig. 3A). In addition, STZ-induced apoptosis was also enhanced by transfecting cells with ERK5 siRNA (Fig. 3B). Furthermore, STZ plus BIX02189 co-treatment increased the expressions of proteins associated with UPRs, that is, of GRP 94/78, CHOP, ATF4, phospho-PERK, and phospho-eIF2α as compared with STZ alone (Fig. 3A). Transfection with ERK5 siRNA had the same effects as BIX02189 on these UPR associated proteins, suggesting that ERK5 regulates their expressions. These observations suggested that ERK5 exerts protective effects by ameliorating the inductions of UPR-associated proteins and apoptosis in pancreatic β-cells.

It has been established that ER stress-mediated apoptosis plays an important role in the destruction of INS-1 cells (Oyadomari et al., 2002; Yusta et al., 2006). The protective effect of ERK5 activation was investigated by transducing INS-1 cells with adenovirus encoding a constitutive active form of MEK5α (CA-MEK5α). ERK5 activation resulted in a significant decrease in the expressions of apoptotic marker proteins and UPR-associated proteins induced by STZ (Fig. 3C). In addition, the overexpression of CA-MEK5α prevented mRNA induction of GRP78 and CHOP by STZ (Fig. 3D). These results suggest that ERK5 protected INS-1 cells from ER stress.

CHOP deficiency ameliorated the ERK5 inhibition-mediated exacerbation of STZ-induced hyperglycemia and pancreatic cell death

Previous studies have suggested that CHOP may be involved in STZ-induced apoptosis mediated by ER stress pathway (Araki et al., 2013; Zhang et al., 2011). To assess the potential mediatory role of CHOP in pancreatic beta cell death, INS-1 cells were transfected with CHOP siRNA or control siRNA. As shown in Fig. 4A, CHOP knockdown significantly reduced the induction of apoptosis by STZ plus BIX02189, which suggested that CHOP induction is responsible for the ERK5 inhibition-mediated exacerbation of STZ-induced β-cell apoptosis.

To clarify the physiological relevance of the CHOP-mediated degenerative pathway in ERK5 inhibition-mediated exacerbation of STZ-induced hyperglycemia and body weight loss, we used CHOP deficient mice. It was found the number of days required to reach the diabetic state and average glucose levels after STZ administration differed significantly in CHOP deficient and control C57BL/6J mice. As was expected, fasting blood glucose concentrations showed that ERK5 inhibition-induced hyperglycemia was significantly suppressed in CHOP deficient mice (Fig. 4B). In addition, C57BL/6J control mice had significantly lower body weights than CHOP deficient mice (Fig. 4C). Moreover, immunoblotting data showed that STZ plus BIX02189 cotreatment induced a dramatic increase in apoptotic markers and TUNEL-positive cell numbers, and that these changes were attenuated in CHOP deficient mice (Figs. 4D and 4E), suggesting the protective effects of ERK5 on STZ-induced pancreatic β-cell damage are mediated by interrupting a CHOP-dependent degenerative pathway. Interestingly, total amounts of ATF4 were significantly upregulated by the combination of STZ and BIX02189 in littermate control mice, but these inductions were not inhibited in CHOP KO. Since CHOP is a downstream molecule in PERK-eIF2α-ATF4-CHOP signaling pathway, CHOP deficiency might not affect the upstream signaling pathway. These results suggest that CHOP deficiency reduced the pancreatic apoptosis induced by the combination of STZ and BIX02189 without affecting upstream signaling pathway

In the present study, we investigated the involvement of ERK5 in STZ-induced apoptosis (Figs. 1C and 1D) and ER stress (Fig. 2) in pancreatic islets. Our findings indicate that ERK5 protects against STZ-induced β-cell apoptosis and hyperglycemia by interrupting the ER stress-mediated apoptotic pathway (Fig. 3), and suggest ERK5 activation has therapeutic potential for the management of ER stress-dependent diabetes and its complications.

STZ has been widely used to produce animal models of diabetes (Szkudelski, 2001) and published observations suggest that ER stress-associated apoptosis underlies STZinduced hyperglycemia (Saini et al., 1996) and that ER stress induces pancreatic β-cell apoptosis and peripheral insulin resistance (Araki et al., 2003). Recently, ERK5-SUMOylation was detected in a murine model of diabetic cardiac dysfunction induced by STZ following myocardial infarction (Shishido et al., 2008; Woo et al., 2008). In addition, the authors showed that CA-MEK5α overexpression could activate ERK5 by inhibiting its SUMOylation (Shishido et al., 2008). Furthermore, in another study, blood glucose concentrations were found to be elevated in adipocyte specific ERK5 deficient mice and insulin sensitivity to be attenuated as compared with wild-type mice (Zhu et al., 2014).

STZ-induced apoptosis in pancreas islets was observed as early as 3 days after the administration of STZ plus BIX02189 in control mice (Fig. 1D). Morphological observations in the mice administered STZ plus BIX02189 revealed marked involution of pancreatic islets and marked increases in the number of CHOP protein expression from 2 days after administration (Fig. 2B). We also found that STZ plus BIX02189-induced ER stress and apoptosis were prevented in CHOP deficient mice but not in littermate controls (Fig. 4). Collectively, these results show that the islets of STZ-induced diabetic mice are under ER stress and that the protection afforded by ERK5 in our diabetic model is in part derived from the inhibition of CHOP-mediated apoptosis.

Although STZ-induced pre-prandial glucose levels were dramatically blocked in CHOP KO compared to littermate controls, decent amount of apoptosis was still found in STZ-treated islet in CHOP KO. The sum of our findings support the idea that CHOP is a fundamental factor causing β-cell dysfunction and apoptosis in response to the chronic ER stress that coincides with β-cell compensation for insulin resistance. Our findings suggest that in the absence of a death signal, UPR signaling can improve protein secretory capacity and preserve the functional integrity of the ER in β-cells. It remains to be examined whether β-cell failure and loss of glucose homeostasis in vivo result from an inadequate UPR adaptive response or activation of an apoptotic response due to chronic, unresolved ER stress.

To determine whether ERK5 is directly involved in STZ-induced ER stress, we transduced INS-1 cells with Ad-CAMEK5 α to activate ERK5 (Fig. 3C). A previous report suggested that CHOP is induced by PERK/eIF2α/ATF4 and activates the expressions of proapoptotic genes (Marciniak et al., 2004). We found ERK5 activation inhibited STZ-induced PERK pathway regulators and apoptosis in INS-1 cells, and that the STZ-induced mRNA expressions of GRP78 and CHOP were significantly diminished by ERK5 activation. These results suggest ERK5-dependent suppression of CHOP expression is regulated in a transcriptional level, but not post-translational modification. However, it is not clear whether ERK5 activation regulates mRNA expression of CHOP through promoter regulation or not.

ERK5 activation also inhibited the STZ-induced protein expressions of GRP78 and GRP94 and of GRP78 at the mRNA level. These results suggest that ERK5 activation inhibits not only PERK pathway but also GRP78-dependent pathway. It was previously shown ERK5 regulates Akt activation, (Roberts et al., 2010) and that the PI3K/Akt pathway protects against ER stress-induced apoptosis via GRP78 regulation (Dai et al., 2010). Together these results suggest MEK5-ERK5 signaling modulates ER stress-mediated apoptosis via the PI3-kinase-Akt pathway.

The in vivo data from a murine model in the present study showed the role of ERK5-dependent suppression of CHOP expression in STZ-induced pancreatic β-cell dysfunction. Since CHOP deficiency rescued pancreatic β-cell dysfunction and hyperglycemia in Akita mouse model of diabetes which has genetic mutation of pro-insulin (Oyadomari et al., 2002). This mutation induced ER stress and CHOP induction. We thus believe that ERK5-dependent suppression of CHOP might act as a therapeutic target for ER stress-dependent diabetes. It remains to be determined in our further study whether ERK5 has similar mechanistic pathway in autoimmune response during the progression on type 1 diabetes.

In conclusion, the results of the present study demonstrate that ERK5 protects against STZ-induced hyperglycemia by regulating pancreatic β-cell apoptosis. In addition, they show that ERK5 inhibition exacerbated ER stress and β-cell apoptosis in STZ-induced murine model of diabetes and that a genetic deficiency in CHOP reduces pancreatic β-cell apoptosis in the co-treatment of STZ and an ERK5 inhibitor. Our observations support the proposition that ERK5 activation-mediated inhibitions of ER stress and apoptosis should be considered a potent therapeutic target for the treatment of diabetes and its complications.