Available online at www.sciencedirect.com

www.elsevier.com/locate/ybbrc

Biochemical and Biophysical Research Communications 366 (2008) 710–716

c-Jun amino terminal kinase 1 deficient mice are protectedfrom streptozotocin-induced islet injury

Kyoichi Fukuda a, Greg H. Tesch a,b, David J. Nikolic-Paterson a,b,*

a Department of Nephrology, Monash Medical Centre, 246 Clayton Road, Clayton, 3168 Vic., Australiab Monash University Department of Medicine, Monash Medical Centre, Clayton, 3168 Vic., Australia

Received 29 November 2007Available online 17 December 2007

Abstract

In vitro studies have implicated the c-Jun amino terminal kinase (JNK) in cytokine-induced pancreatic injury leading to a loss of insu-lin production and hyperglycemia. We examined the role of JNK1 in the multiple low dose streptozotocin (MLD-STZ) model in whichislet injury and hyperglycemia are dependent upon T cell immunity and pro-inflammatory cytokines. MLD-STZ in wild type miceinduced islet leukocyte infiltration, cytokine production, b-cell apoptosis, and hyperglycemia. In contrast, Jnk1�/� mice were substan-tially protected from a loss of insulin producing cells and hyperglycemia in the MLD-STZ model despite a marked islet T cell and mac-rophage infiltrate. Based upon several lines of evidence, this protection was attributed to a reduction in TNF-a production by infiltratingJnk1�/� macrophages leading to reduced b-cell apoptosis. In conclusion, JNK1 signaling plays an essential role in macrophage inducedb-cell apoptosis and the development of hyperglycemia in MLD-STZ induced pancreatic injury.� 2007 Elsevier Inc. All rights reserved.

Keywords: Apoptosis; Beta cell; Diabetes; Hyperglycemia; Islet; JNK; Macrophage; Streptozotocin; T-cell; TNF-a

The loss of pancreatic b-cells leads to insulin insuffi-ciency and the development of hyperglycemia. A wide vari-ety of insults can lead to b-cell loss, including killing by Tcells, cytokine-mediated b-cell cytotoxicity, and cellulardamage due to oxidative stress [1]. In particular, pro-inflammatory cytokines, such as IL-1, TNF-a, and IFN-c, are an important mechanism of leukocyte-mediatedb-cell injury and apoptosis. IL-1b induces b-cell cytotoxicityin vitro [2], an effect that is augmented by TNF-a [3,4]. In

vivo treatment with the IL-1 receptor antagonist (IL-1ra)prevented insulitis and hyperglycemia in the multiple low-dose streptozotocin (MLD-STZ) model and IL-1ra treat-ment prolonged survival of mouse islets in an allograft

0006-291X/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.bbrc.2007.12.007

* Corresponding author. Address: Department of Nephrology, MonashMedical Centre, 246 Clayton Road, Clayton, 3168 Vic., Australia. Fax:+61 3 9594 6530.

E-mail address: david.nikolic-paterson@med.monash.edu.au (D.J.Nikolic-Paterson).

model [5,6]. Also, blockade of TNF-a production pre-vented hyperglycemia in the MLD-STZ model [7].

An important mechanism of cytokine-induced b-celldamage is activation of the c-Jun amino terminal kinase(JNK). Activation of the JNK pathway is implicated inapoptotic cell death in a variety of cell types [8], and studieshave shown a key role for JNK signaling in IL-1 induceddeath of cultured b-cells [9]. The two major JNK isoforms,JNK1 and JNK2, are both expressed in pancreatic islets[10]. There is redundancy between JNK1 and JNK2 insome functions of this signaling pathway [11], while otherfunctions, such as TNF-a induced apoptosis in embryonicfibroblasts, operate via JNK1 only [12].

The present study investigated the role of JNK1 signal-ing in pancreatic b-cell damage leading to hyperglycemiaby examining JNK1 gene deficient mice in the MLD-STZmodel. This model was selected since both pro-inflamma-tory cytokines and the T cell immune response play a path-ogenic role in islet inflammation leading to b-celldestruction and hyperglycemia [13,14].

K. Fukuda et al. / Biochemical and Biophysical Research Communications 366 (2008) 710–716 711

Materials and methods

Animals. Jnk1+/� (Mapk8) mice on the C57BL/6J background wereimported from Jackson Labs, Bar Harbor, ME and bred at MonashAnimal Services, Clayton, Australia. Jnk1�/� mice were bred fromJnk1+/� mice, with Jnk1+/� littermates used as controls in someexperiments.

MLD-STZ model. Male mice (22–26 g) received 5 intraperitonealinjections of 40 mg/kg streptozotocin (Sigma–Aldrich, St. Louis, MO)dissolved in 0.1 M sodium citrate, pH 4.5, over 5 consecutive days. Bloodglucose was measured once weekly following a 16 h fast using the glucoseoxidase method. Hyperglycemia was defined as >7 mmol/L fasting bloodglucose on two or more occasions by week 4. Animals were killed 5 days(peak of STZ-induced b-cell apoptosis), 2 weeks (peak of islet leukocyteinfiltration) or 4 weeks (assessment of hyperglycemia and b-cell loss) afterthe start of STZ injections. Details of group sizes for each parameteranalyzed are provided in the figure legends. Animal experimentation wasapproved by the Monash Medical Centre Animal Ethics Committee.

Pancreas insulin content. Groups of four normal mice or mice 4 weeksafter MLD-STZ were killed, pancreas removed and weighed, and theninsulin was extracted from pancreatic tissue using acid ethanol andquantified by ELISA (Linco Research, St. Charles, MO).

Histochemistry. Pancreatic tissue was fixed in 4% neutral-buffer for-malin for 3 h (immunostaining and PAS staining) or 24 h (aldehydefuchsin staining). Gomori’s aldehyde fuchsin staining was performed on4 lm paraffin sections as previously described [15]. The percent islet areastained was determined by image analysis using Image Pro Plus 4.0 soft-ware (Media Cybernetics, CA). PAS staining of 4 lm sections was used toanalyze the degree of insulitis using a semi-quantitative scoring system asfollows: 0, normal; 1+, minor peri-islet mononuclear cell infiltration; 2+,moderate intra-islet mononuclear cell infiltration (<50% of islet area); 3+,severe intra-islet cell infiltration (>50% of islet area) with damage to isletarchitecture.

Immunohistochemistry. Immunoperoxidase staining was performed on4 lm paraffin sections using an avidin–biotin complex system with thefollowing primary antibodies: F4/80 (macrophages) (Serotec, Oxford,UK); GK1.5 (CD4); YTS169.4 (CD8); guinea pig anti-insulin (Dako,Glostrup, Denmark), and; rabbit anti-cleaved caspase-3 (Cell SignalingTechnology, Beverly, MA). For each islet, the islet area was measured andthe number of cells in the islet plus surrounding positive cells touching theislet was counted and expressed as cells/mm2. In the case of enumeratingapoptotic cells, the number of caspase-3+ cells in islets was normalized bythe area of insulin staining in each islet on serial sections. This was per-

Table 1Sequences of primers and probes for real time RT-PCR

Molecule Sequences of primers and minor groove binder (mgb)probes

IL-1b CAA GAT AGA AGT CAA GAG CAA ATAG AAA CAG TCC AGC CCA TACCAC AAG CAG AGC ACA AG mgb

TNFa TCT ACT CCC AGG TTC TCT TCGCA GAG AGG AGG TTG ACT TTTCA CCC ACA CCG TCA G mgb

IFN-c CAG CAA CAA CAT AAG CGT CAACC TCA AAC TTG GCA ATA CTCCAA CAG CAA GGC GAA A mgb

IL-4 TGA ACG AGG TCA CAG GAG AAACC TTG GAA GCC CTA CAG ACAC AGC AAC GAA GAA CAC mgb

iNOS ACT ACT AAA TCT CTC TCC TCT CCTCT CTG CTC TCA GCT CCA ATCC CTC CCC TCT CTC C mgb

Cyclophilin GAA GGT GAA AGA AGG CAT GAAGCC CGC AAG TCA AAA GAA ACAA GAC CAG CAA GAA GA mgb

formed to confirm that apoptotic cells were insulin positive and to makethis relative to the number of b-cells remaining in the islets.

Isolation of pancreatic islets. After killing, the common bile duct wasligated at the distal end with nylon suture and then cannulated with a 30Gneedle and syringe to infuse 3 ml of 1 mg/ml ice-cold collagenase P (RocheBiochemicals, Mannheim, Germany) in Hank’s Balanced Salt Solution(HBSS). The inflated pancreas was excised and incubated in 2 ml of col-lagenase P solution at 37 �C for 12 min. After a brief vigorous shake todisperse the digested pancreas, islets were washed three times in HBSSwith centrifugation for 30 s at 200g. The islet pellet was either resuspendedin RNAlater (Ambion) for RNA extraction, or in culture media forapoptosis studies, and then the material put in petri dishes under a dis-secting microscope with reflected lighting and a dark base. Individual isletswere recovered using sterile forceps and then used for RNA extraction orcell culture studies.

Real time RT-PCR. Total RNA was extracted from isolated pancreaticislets, or from the pancreatic draining lymph node, using the RNAeasyMicro kit (Qiagen, Doncaster, Vict., Australia) and reverse transcribedusing the Superscript First-Strand Synthesis kit (Invitrogen) with oligo-dTprimers. Real-time PCR was performed with the primers listed in Table 1using the Rotor-Gene 3000 system (Corbett Research, Sydney, Australia)using the QuantiTect Probe PCR kit (Qiagen) with thermal cycling condi-tions of 37 �C for 10 min to activate uracil-DNA glycosylase, 95 �C for15 min, followed by 45 cycles of 95 �C for 15 s and 60 �C for 60 s. Standardcurves were established for target PCR products including cyclophilin, usingserial dilutions of each purified PCR product. The relative abundance ofeach mRNA was calculated using the DDCt method and normalized againstthe cyclophilin mRNA level. All samples were analyzed twice and theaverage result taken.

TNF-a production by peritoneal macrophages. Peritoneal cells wereextracted from normal WT and Jnk1�/� mice, plated into 24-well cultureplates, and incubated at 37 �C for 2 h after which non-adherent cells wereremoved. Adherent macrophages were cultured overnight in DMEM/10%FCS and then stimulated with a cytokine mix (10 ng/ml rat IL-1b, 10 ng/mlhuman TNF-a, 10 ng/ml rat IFN-c) or 10 ng/ml human TNF-a for 3 h,media changed, and then incubated for 24 hand the supernatant collectedand stored at�80 �C until samples were assayed using an ELISA specific formouse TNF-a (BD Biosciences Pharmingen). The cell pellet was extracted toquantify the DNA content (Quant-iT DNA assay kit, Invitrogen). Macro-phage production of TNF-a was expressed relative to DNA content.

Apoptosis in cultured islets. Islets were isolated from the pancreas ofnormal mice, as described above, were cultured for 72 h in DMEM with10% FCS, and then incubated for 24 h with either 0.5 mM STZ or with acytokine mix (10 ng/ml IL-1b, 10 ng/ml TNF-a, 10 ng/ml IFN-c).Apoptosis was quantified by cell death detection ELISA kit (Roche) andnormalized against cellular DNA content.

Statistical analysis. Data are presented as means ± SEM. Data wereanalyzed by one way ANOVA with Newman–Keuls Multiple Comparisonpost-test unless otherwise stated.

Results

MLD-STZ induced islet inflammation, loss of b-cells, and

hyperglycemia

MLD-STZ in WT mice caused marked insulitis, with thepresence of peri-islet and intra-islet mononuclear cell infil-tration (Fig. 1A–C), which peaked at 2 weeks after MLD-STZ. This caused a substantial loss of insulin granules inislet b-cells leading to a dramatic reduction in the insulincontent of the pancreas (Fig. 1D–F), with hyperglycemiaevident in 80% of mice by week 4 (Fig. 1G and H).

Jnk1�/� mice developed peri-islet and intra-islet mono-nuclear cell infiltration following MLD-STZ, equivalent tothat seen in WT mice (Fig. 1A–C). In contrast, Jnk1�/�

Fig. 2. Leukocyte infiltration in MLD-STZ. (A) Immunostaining of F4/80+ macrophages, CD4+ T cells, and CD8+ T cells in pancreatic islets in normaland MLD-STZ treated mice. (B) Quantification of islet infiltration by CD4+ T cells (open bars), CD8+ T cells (hatched bars), and F4/80+ macrophages(closed bars). Data are based upon groups of 4–6 mice. No significant differences were evident between WT and Jnk1�/� MLD-STZ mice.

Fig. 1. Jnk1�/� mice are protected from MLD-STZ induced hyperglycemia. (A) PAS staining shows normal islet architecture in a WT mouse, and; (B)peri-islet and intra-islet infiltration by mononuclear cell and a loss of islet architecture at 2 weeks after MLD-STZ in a WT mouse (grade 3+ lesion). (C)The degree of insulitis was assessed in groups of 10–16 mice at 2 weeks after MLD-STZ using a semi-quantitative scoring system. No significant differencewas evident between WT and Jnk1�/� mice. (D) Gomori’s aldehyde fuchsin staining for b-cell granulation in normal mice or in WT and Jnk1�/� mice 4weeks after MLD-STZ. (E) Quantification of the area of islet aldehyde fuchsin staining in mice with (closed bars) and without (open bars) MLD-STZbased upon groups of 8 to 12 mice. *P < 0.05 versus MLD-STZ WT. (F) Insulin content of pancreas (lg insulin/100 mg pancreas tissue) based upongroups of 4 mice. *P < 0.05 versus MLD-STZ WT. (G) Fasting blood glucose following MLD-STZ. Data are based upon groups of 20–26 mice. *P < 0.05versus WT. (H) Cumulative incidence of hyperglycemia by week 4 in the MLD-STZ model. *P < 0.05 versus WT mice by v2 test.

712 K. Fukuda et al. / Biochemical and Biophysical Research Communications 366 (2008) 710–716

K. Fukuda et al. / Biochemical and Biophysical Research Communications 366 (2008) 710–716 713

mice showed significant protection from MLD-STZinduced loss of pancreatic b-cells and islet insulin contentcompared to WT mice (Fig. 1D–F). This translated intoa marked protection of Jnk1�/� mice from hyperglycemia,with only 20% of animals becoming diabetic by week 4(Fig. 1G and H). This protection from hyperglycemiawas maintained up to week 8 (data not shown). No genedosage effect was apparent since Jnk1+/� mice exhibitedthe same susceptibility as WT mice to MLD-STZ inducedhyperglycemia (data not shown).

Islet leukocyte infiltration and cytokine mRNA expression

MLD-STZ induced an infiltrate of F4/80+ macro-phages, CD4+ T cells, and CD8+ T cells in the isletsof WT mice (Fig. 2). Consistent with the results ofthe PAS staining, MLD-STZ Jnk1�/� mice also showeda macrophage and T cell infiltrate, the composition ofwhich was not different to that seen in WT mice(Fig. 2).

An increase in the mRNA levels for TNF-a, IL-1b,IFN-c, IL-4, and iNOS was evident in pancreatic islets iso-lated after 2 weeks in MLD-STZ WT mice compared tonormal WT mice (Fig. 3A–E). A similar increase in mRNAlevels for all of these cytokines was evident in islets isolatedfrom MLD-STZ Jnk1�/� mice, except for a significantreduction in islet TNF-a mRNA levels (Fig. 3A). Giventhat macrophages are a major source of TNF-a, we exam-ined the role of JNK1 in TNF-a production in vitro.Jnk1�/� mouse peritoneal macrophages showed a signifi-cant reduction in TNF-a secretion in response to stimula-tion by a cytokine mixture or stimulation by humanTNF-a compared to WT macrophages (Fig. 3G).

As a further measure of the immune response, the Th1/Th2 cytokine ratio was examined in the pancreatic draininglymph node (PLN). There was an increase in the ratio ofIFN-c/IL-4 mRNA in WT mice after MLD-STZ treatment(Fig. 3F). However, there was no significant differencebetween the IFN-c/IL-4 mRNA ratio in the PLN ofMLD-STZ WT mice compared to MLD-STZ Jnk1�/�mice (Fig. 3F).

Fig. 3. Cytokine mRNA expression in MLD-STZ. Pancreatic islets wereisolated from normal mice (n = 4–5) or 2 weeks after MLD-STZadministration (n = 7–11) in WT and Jnk1�/� mice. Islet mRNA levelswere analyzed for: (A) TNF-a, (B) IL-1b, (C) IFN-c, (D) IL-4, and (E)iNOS. *P < 0.05 versus MLD-STZ WT mice. (F) The pancreatic draininglymph node (PLN) was isolated from normal mice (n = 3) or 2 weeks afterMLD-STZ administration in WT and Jnk1�/�mice (n = 5–10), and IFN-c and IL-4 mRNA levels were analyzed. The graph shows the ratio ofIFN-c/IL-4 mRNA levels. (G) Peritoneal macrophages were preparedfrom groups of 4–6 normal WT and Jnk1�/� mice and stimulated with acytokine mix (rIL-1b, hTNF-a, rIFN-c) or with hTNF-a alone for 3 h,washed, and then the culture supernatant collected after 24 h and assayedfor mouse TNF-a secretion by ELISA. *P < 0.05 compared to WT.

c

Islet cell apoptosis

Islet cell apoptosis was identified by immunostaining forcleaved caspase-3 (Fig. 4A). Significant apoptosis of isletcells was identified in MLD-STZ WT mice at the twotime-points examined: day 5 when islet cell apoptosis isdue to the acute effects of STZ, and day 14 when apoptosis

714 K. Fukuda et al. / Biochemical and Biophysical Research Communications 366 (2008) 710–716

is leukocyte-dependent (Fig. 4B). Analysis of serial tissuesections indicated that most cleaved caspase-3+ apoptoticcells were b-cells based upon insulin staining. WhileJnk1�/�mice showed no difference in islet cell apoptosis atday 5 of MLD-STZ, there was a significant reduction inislet cell apoptosis at day 14 of MLD-STZ in Jnk1�/�mice(Fig. 4B).

To determine whether islet cells from Jnk1�/� micehave an intrinsic resistance or susceptibility to apoptosis,we performed in vitro studies using pancreatic islets isolatedfrom normal mice. As shown in Fig. 4(C), islets fromJnk1�/� mice showed equal sensitivity to STZ and cyto-kine induced apoptosis compared to islets from WT mice.

Fig. 4. Islet cell apoptosis in MLD-STZ. (A) Immunostaining for cleaved caspapoptotic cells (nuclear staining) are evident in WT mice at 14 days after MLDcleaved caspase-3 stained cells in normal mice (n = 3), on day 5 (n = 4–5), or(closed bars) mice. *P < 0.05 versus WT on day 14. (C) Pancreatic islets werecultured, and then stimulated with 0.5 mM STZ or with a cytokine mix (IL-1bdetection ELISA kit and normalized against cellular DNA content and backgbetween WT and Jnk1�/� mice.

Discussion

This study has identified an essential role for JNK1 sig-naling in the MLD-STZ model of hyperglycemia. There arethree mechanisms which could explain the protection ofJnk1�/� mice from MLD-STZ induced hyperglycemia:(1) an intrinsic resistance of Jnk1�/� b-cells to STZ and/or cytokine induced apoptosis; (2) suppression of leuko-cyte-mediated islet damage through a reduction in TNF-a production, or; (3) a shift in the Th1/Th2 balance inthe immune response.

Jnk1�/� mice were not protected from STZ-inducedislet cell apoptosis on the basis that islet cell apoptosis

ase-3 shows a lack of apoptotic cells in islets of normal WT mice, whereas-STZ, which is reduced in MLD-STZ Jnk1�/� mice. (B) Quantification ofon day 14 (n = 8–10) after MLD-STZ in WT (open bars) and Jnk1�/�isolated from normal WT (open bars) and Jnk1�/� (closed bars) mice,

, TNF-a, IFN-c) for 24 h and then apoptosis determined by the cell deathround in untreated islets subtracted. No significant difference was evident

K. Fukuda et al. / Biochemical and Biophysical Research Communications 366 (2008) 710–716 715

was unaltered in Jnk1�/� compared to WT mice on day 5of MLD-STZ—a time-point at which apoptosis is STZ-dependent and prior to leukocyte infiltration [16]. In addi-tion, islets isolated from normal WT and Jnk1�/� miceshowed an equal sensitivity to both STZ and cytokineinduced apoptosis in vitro. The latter finding contrasts withprevious studies which found that blockade of JNK signal-ing inhibits IL-1 induced death of cultured b-cells, or b-celllines [9,17]. Our results suggest redundancy between JNK1and JNK2 in cytokine induced b-cell apoptosis, or that theJNK signaling pathway is not involved in cytokine inducedb-cell death in vivo.

Jnk1�/� mice were protected from hyperglycemiadespite a marked islet T cell and macrophage infiltrate.However, the inflamed islets showed a significant reductionin TNF-a mRNA levels. This is the most likely mechanismresponsible for the protection of Jnk1�/� mice fromMLD-STZ induced hyperglycemia, although functionaldeletion of TNF-a from macrophages would be neededto formally establish this. Macrophages are a major sourceof TNF-a and in vitro studies confirmed that Jnk1�/�macrophages have impaired TNF-a secretion in responseto cytokine stimulation. Cytokines operate in a combinato-rial fashion to induce b-cell cytotoxicity, with TNF-aknown to significantly augment IL-1 induced b-cell cyto-toxicity [3,4]. Therefore, despite the increase in IL-1b,IFN-c, and iNOS levels seen in islets of Jnk1�/� mice inMLD-STZ, it may the case that TNF-a is a critical playerfor the induction of b-cell death. This is supported by stud-ies in which blockade of TNF-a is sufficient to protectagainst MLD-STZ induced hyperglycemia [7]. Further-more, macrophage TNF-a production is highly sensitiveto blockade of JNK signaling [18–20], whereas macrophageproduction of other inflammatory mediators is less depen-dent upon JNK signaling [20,21]. Thus, it was not unex-pected that TNF-a mRNA levels were suppressed inJnk1�/� islets in the MLD-STZ model, while the lack ofeffect upon IL-1b and iNOS mRNA levels may simplyreflect a less important role for JNK signaling in the tran-scription of these genes.

C57BL/6J mice show a Th1 response in the MLD-STZmodel [14]. Blockade of IL-18, a Th1 cytokine, or adminis-tration of IL-10, a Th2 cytokine, has been shown to preventhyperglycemia in MLD-STZ [22,23]. We also found a Th1biased response in MLD-STZ in C57BL/6J mice which wasnot altered in Jnk1�/� mice. This was a somewhat unex-pected finding given that Jnk1�/� naı̈ve T cells preferen-tially differentiate into Th2-type T cells [24], and thatJnk1�/� mice are protected from experimental encephalo-myelitis through an increased Th2 response [25]. However,the apparent lack of effect of JNK1 deletion on the Th1/Th2balance of the T cell response may simply reflect the verydifferent environment in which the immune response occursin the MLD-STZ model; a response to STZ-induced b-cellapoptosis and necrosis compared to immunization with aspecific antigen in the presence of adjuvant or pan T cell

activation in vitro with anti-CD3 and anti-CD28 antibodies[24,25].

In conclusion, this study has identified an essential rolefor JNK1 signaling in pancreatic damage leading to hyper-glycemia in the MLD-STZ model. In particular, the resultssuggest that JNK1 signaling is necessary for macrophage-derived TNF-a which causes b-cell apoptosis leading toloss of insulin production and hyperglycemia.

Acknowledgment

This study was funded by the NHMRC of Australia.

References

[1] M. Cnop, N. Welsh, J.C. Jonas, A. Jorns, S. Lenzen, D.L. Eizirik,Mechanisms of pancreatic beta-cell death in type 1 and type 2diabetes: many differences, few similarities, Diabetes 54 (Suppl. 2)(2005) S97–S107.

[2] K. Bendtzen, T. Mandrup-Poulsen, J. Nerup, J.H. Nielsen, C.A.Dinarello, M. Svenson, Cytotoxicity of human pI 7 interleukin-1 forpancreatic islets of Langerhans, Science 232 (1986) 1545–1547.

[3] T. Mandrup-Poulsen, K. Bendtzen, C.A. Dinarello, J. Nerup,Human tumor necrosis factor potentiates human interleukin 1-mediated rat pancreatic beta-cell cytotoxicity, J. Immunol. 139(1987) 4077–4082.

[4] D.L. Eizirik, S. Sandler, N. Welsh, M. Cetkovic-Cvrlje, A. Nieman,D.A. Geller, D.G. Pipeleers, K. Bendtzen, C. Hellerstrom, Cytokinessuppress human islet function irrespective of their effects on nitricoxide generation, J. Clin. Invest. 93 (1994) 1968–1974.

[5] J.O. Sandberg, A. Andersson, D.L. Eizirik, S. Sandler, Interleukin-1receptor antagonist prevents low dose streptozotocin induced diabetesin mice, Biochem. Biophys. Res. Commun. 202 (1994) 543–548.

[6] J.O. Sandberg, D.L. Eizirik, S. Sandler, D.E. Tracey, A. Andersson,Treatment with an interleukin-1 receptor antagonist protein prolongsmouse islet allograft survival, Diabetes 42 (1993) 1845–1851.

[7] M. Holstad, S. Sandler, A transcriptional inhibitor of TNF-alphaprevents diabetes induced by multiple low-dose streptozotocin injec-tions in mice, J. Autoimmun. 16 (2001) 441–447.

[8] C.R. Weston, R.J. Davis, The JNK signal transduction pathway,Curr. Opin. Cell Biol. 19 (2007) 142–149.

[9] A. Ammendrup, A. Maillard, K. Nielsen, N. Aabenhus Andersen, P.Serup, O. Dragsbaek Madsen, T. Mandrup-Poulsen, C. Bonny, Thec-Jun amino-terminal kinase pathway is preferentially activated byinterleukin-1 and controls apoptosis in differentiating pancreatic beta-cells, Diabetes 49 (2000) 1468–1476.

[10] S. Paraskevas, R. Aikin, D. Maysinger, J.R. Lakey, T.J. Cavanagh, B.Hering, R. Wang, L. Rosenberg, Activation and expression of ERK,JNK, and p38 MAP-kinases in isolated islets of Langerhans:implications for cultured islet survival, FEBS Lett. 455 (1999) 203–208.

[11] C.R. Weston, R.J. Davis, The JNK signal transduction pathway,Curr. Opin. Genet. Dev. 12 (2002) 14–21.

[12] J. Liu, Y. Minemoto, A. Lin, c-Jun N-terminal protein kinase 1(JNK1), but not JNK2, is essential for tumor necrosis factor alpha-induced c-Jun kinase activation and apoptosis, Mol. Cell Biol. 24(2004) 10844–10856.

[13] Y.T. Kim, C. Steinberg, Immunologic studies on the induction ofdiabetes in experimental animals. Cellular basis for the induction ofdiabetes by streptozotocin, Diabetes 33 (1984) 771–777.

[14] A. Muller, P. Schott-Ohly, C. Dohle, H. Gleichmann, Differentialregulation of Th1-type and Th2-type cytokine profiles in pancreaticislets of C57BL/6 and BALB/c mice by multiple low doses ofstreptozotocin, Immunobiology 205 (2002) 35–50.

716 K. Fukuda et al. / Biochemical and Biophysical Research Communications 366 (2008) 710–716

[15] J.D. Peterson, B. Pike, M. McDuffie, K. Haskins, Islet-specific T cellclones transfer diabetes to nonobese diabetic (NOD) F1 mice, J.Immunol. 153 (1994) 2800–2806.

[16] B.A. O’Brien, B.V. Harmon, D.P. Cameron, D.J. Allan, Beta-cellapoptosis is responsible for the development of IDDM in the multiplelow-dose streptozotocin model, J. Pathol. 178 (1996) 176–181.

[17] C. Bonny, A. Oberson, S. Negri, C. Sauser, D.F. Schorderet, Cell-permeable peptide inhibitors of JNK: novel blockers of beta-celldeath, Diabetes 50 (2001) 77–82.

[18] K.L. Rhoades, S.H. Golub, J.S. Economou, The regulation of thehuman tumor necrosis factor alpha promoter region in macrophage,T cell, and B cell lines, J. Biol. Chem. 267 (1992) 22102–22107.

[19] E.Y. Tsai, J.V. Falvo, A.V. Tsytsykova, A.K. Barczak, A.M.Reimold, L.H. Glimcher, M.J. Fenton, D.C. Gordon, I.F. Dunn,A.E. Goldfeld, A lipopolysaccharide-specific enhancer complexinvolving Ets, Elk-1, Sp1, and CREB binding protein and p300 isrecruited to the tumor necrosis factor alpha promoter in vivo, Mol.Cell biol. 20 (2000) 6084–6094.

[20] Y. Ikezumi, L. Hurst, R.C. Atkins, D.J. Nikolic-Paterson, Macro-phage-mediated renal injury is dependent on signaling via the JNKpathway, J. Am. Soc. Nephrol. 15 (2004) 1775–1784.

[21] A. Lahti, U. Jalonen, H. Kankaanranta, E. Moilanen, c-Jun NH2-terminal kinase inhibitor anthra(1,9-cd)pyrazol-6(2H)-one reducesinducible nitric-oxide synthase expression by destabilizing mRNA inactivated macrophages, Mol. Pharmacol. 64 (2003) 308–315.

[22] F. Nicoletti, R. Di Marco, G. Papaccio, I. Conget, R. Gomis, R.Bernardini, J.E. Sims, Y. Shoenfeld, K. Bendtzen, Essential patho-genic role of endogenous IL-18 in murine diabetes induced bymultiple low doses of streptozotocin. Prevention of hyperglycemiaand insulitis by a recombinant IL-18-binding protein: Fc construct,Eur. J. Immunol. 33 (2003) 2278–2286.

[23] Z.L. Zhang, S.X. Shen, B. Lin, L.Y. Yu, L.H. Zhu, W.P. Wang, F.H.Luo, L.H. Guo, Intramuscular injection of interleukin-10 plasmidDNA prevented autoimmune diabetes in mice, Acta Pharmacol. Sin.24 (2003) 751–756.

[24] C. Dong, D.D. Yang, M. Wysk, A.J. Whitmarsh, R.J. Davis, R.A.Flavell, Defective T cell differentiation in the absence of Jnk1, Science282 (1998) 2092–2095.

[25] E.H. Tran, Y.T. Azuma, M. Chen, C. Weston, R.J. Davis, R.A.Flavell, Inactivation of JNK1 enhances innate IL-10 production anddampens autoimmune inflammation in the brain, Proc. Natl. Acad.Sci. USA 103 (2006) 13451–13456.