Interleukin-1β-induced Rat Pancreatic Islet Nitric Oxide Synthesis Requires Both the p38 and Extracellular Signal-regulated Kinase 1/2 Mitogen-activated Protein Kinases*

Interleukin-1β (IL-1β) is cytotoxic to rat pancreatic β-cells by inhibiting glucose oxidation, causing DNA damage and inducing apoptosis. Nitric oxide (NO) is a necessary but not sufficient mediator of these effects. IL-1β induced kinase activity toward Elk-1, activation transcription factor 2, c-Jun, and heat shock protein 25 in rat islets. By Western blotting with phosphospecific antibodies and by immunocomplex kinase assay, IL-1β was shown to activate extracellular signal-regulated kinase (ERK) 1/2 and p38 mitogen-activated protein kinase (p38) in islets and rat insulinoma cells. Specific ERK1/2 and p38 inhibitors individually reduced but in combination blocked IL-1β-mediated islet NO synthesis, and reverse transcription-polymerase chain reaction of inducible NO synthase mRNA showed that ERK1/2 and p38 controlled IL-1β-induced islet inducible NO synthase expression at the transcriptional level. Hyperosmolarity caused phosphorylation of Elk-1, activation transcription factor 2, and heat shock protein 25 and activation of ERK1/2 and p38 in islets comparable to that induced by IL-1β but did not lead to NO synthesis. Inhibition of p38 but not of ERK1/2 attenuated IL-1β-mediated inhibition of glucose-stimulated insulin release. We conclude that ERK1/2 and p38 activation is necessary but not sufficient for IL-1β-mediated β-cell NO synthesis and that p38 is involved in signaling of NO-independent effects of IL-1β in β-cells.

Inhibitors-The highly specific inhibitor of p38 (compound VK-19.577, identical to SB203580 and p38i) (26,27) was from Vertex Pharmaceuticals Inc. (Cambridge, MA). Compound PD 098059 from New England Biolabs is the specific inhibitor of MAPK/ERK kinase activation (hereafter termed MEK inhibitor (MEKi)) (25,28). p38i and MEKi were both dissolved in DMSO to stock concentrations of 25 and 100 mM, respectively. Inhibitors were added 1 h before islet stimulation, and a final 0.14% (v/v) DMSO was added to all conditions as control for the DMSO used to dissolve MEKi and p38i.
Islet Culture-150 randomly picked islets/300 l of CM ϩ 0.5% human serum (final osmolarity, 300 mosM) were placed in 4-well dishes (Nunc) treated with IL-1␤, with or without inhibitors, or exposed to hyperosmolarity (addition of hyperosmolar saline to CM as described previously (17)), as indicated in the figure legends. For immunoprecipitation and Western blotting experiments, each condition was made in duplicate, and islets were pooled prior to lysis.
RIN Cell Preculture and Culture-RIN-5AH-T2B cells of low passage number (10 -17) were maintained in 80-cm 2 tissue culture flasks (Nunc) in CM supplied with 10% heat-inactivated FCS (HyClone, Logan, UT) at 37°C in a 5% CO 2 /95% air mixture. When confluent RIN cells were trypsinized, 150,000 cells seeded in 96-well dishes (Costar, Cambridge, UK) containing 200 l of CM ϩ 10% heat-inactivated FCS and precultured for 24 h before experimentation, by which time the cell number had doubled. The RIN cells were then exposed to IL-1␤ as described in the Fig. 2
Whole Cell Lysate Kinase Assay-The GST-Elk-1, GST-ATF2, and Hsp25 phosphotransferase reactions were carried out in a final volume of 25 l at 30°C for 30 min after addition of 5 l of whole cell lysate, 17 l of reaction buffer (2 g of GST-Elk-1, 2 g of GST-ATF2, 1 g of Hsp25, 25 mM Tris-HCl, pH 7.5, 0.1 mM EGTA, 0.1 mM Na 3 VO 4 , 1 M cAMP-dependent protein kinase inhibitor peptide, and 10 mM Mgacetate), and 3 l of ATP mixture (1 mM ATP and 3 Ci [␥-32 P]ATP). Reactions were terminated by addition of 25 l of SDS sample buffer (125 mM Tris-HCl, pH 6.8, 4% SDS, 0.1 M dithiothreitol, 10% glycerol, and 0.02% bromphenol blue) and boiling for 5 min. The samples were then subjected to SDS-PAGE as described by Laemmli (30), using a 4% stacking gel and a 12% separating gel. After electrophoresis, the separating gel was washed for 15 min in a mixture of 10% acetate and 40% methanol. The gels were dried, and the proteins were visualized by autoradiography and quantitated by PhosphorImager analysis (Molecular Dynamics, Sunnyvale, CA). To asses the possible contamination of the used GST-ATF2, we performed Coomassie Blue staining and kinase assay with GST-ATF2 and ATF2 (1-96 and 1-505) from Santa Cruz Biotechnology (Santa Cruz, CA), which confirmed the substrate specificity and purity of the GST-ATF2.
Immunoprecipitation, Immunocomplex, and JNK Kinase Assay (Sol-id Phase Assay)-100 l of whole cell lysates from RIN cells or 300 islets, diluted 1:4 in washing buffer (lysis buffer with 0.1% Triton X-100), were immunoprecipitated by incubation overnight with anti-MAPKAP-K2 or anti-ERK1/2 antibodies and then with protein A-Sepharose beads (Amersham Pharmacia Biotech) for 3 h at 4°C (MAPKAP-K2 and ERK1/2) or incubated with GST-c-Jun (5 g) coupled to glutathione-Sepharose beads for 3 h at 4°C (JNK kinase assay). The beads were washed three times in washing buffer and twice in kinase buffer (20 mM HEPES (pH 7.5), 20 mM ␤-glycerophosphate, 10 mM MgCl 2 , 1 mM dithiothreitol, 50 M Na 3 VO 4 ). Kinase reactions were carried out for 30 min at 30°C in 30 l of kinase buffer containing 10 Ci [␥-32 P]ATP and 1 g of Hsp25 (MAPKAP-K2) or 5 g of GST-Elk-1 (ERK1/2). Reactions were terminated with 30 l of SDS sample buffer, and the samples were analyzed as in the whole cell lysate kinase assay. Western Blotting-Pooled duplicate experiments (a total of 300 islets/ condition) were lysed in 50 l of lysis buffer. Twenty l (ϳ20 g) of whole cell lysates were added to 20 l of SDS sample buffer and boiled for 5 min. SDS-PAGE (12%) was performed, and Western blotting was carried out according to standard protocols (31). Anti-total p38, antitotal ERK1/2, anti-phosphospecific p38, or anti-phosphospecific ERK1/2 antibodies were used. Enhanced chemiluminescence was used for detection. Lysate from the same experiment was separated on two gels and probed with either phosphospecific or total antibody as control for sample variation in protein content.
Nitric Oxide Synthesis and Insulin Release-Islet NO production was measured as nitrite accumulation in conditioned media determined by the Griess reaction (32). In brief, 150 l of medium were mixed with an equal volume of the Griess reagent (one part 0.1% naphtylethylene diamine dihydrochloride and one part 1% sulfanilamide in 5% H 3 PO 4 (Merck, Darmstadt, Germany)) and incubated for 10 min at room temperature. The absorbance at 550 nm was measured on an immunoreader (Nippon Inter Med., Tokyo, Japan). The detection limit was 1 M, equal to 2 pmol/islet in our conditions. Values below the detection limit were assigned the value of 2 pmol/islet. Intra-and interassay coefficients of variation calculated from three points on the standard curve were as follows: 1 M, 2.3 and 16.8%; 10 M, 1.8 and 8.1%; and 25 M, 1.9 and 11.2%. Accumulated insulin release in the conditioned media was measured by radioimmunoassay (33). The detection limit was 35 fmol/ml. Intra-and interassay coefficients of variation between three known controls were as follows: A, 6.1 and 12.2%; B, 4.3 and 11.4%; and C, 3.1 and 8.2%.
Reverse Transcription-PCR-Total RNA from snap frozen islets was extracted and cDNA was prepared with cDNA Cycle® kit (Invitrogen, Leek, The Netherlands) as described previously (34). Reverse transcription-PCR was performed using [␣-32 P]dCTP (Amersham Pharmacia Biotech) and a fixed volume (5 l) of cDNA dilution. Each analysis was performed with a set of iNOS primers in combination with a set of primers for TATA-binding protein (35) as an internal standard. The PCR products were separated on a 6% polyarylamide gel (Life Technologies, Inc.), visualized by autoradiography, and quantitated by Phos-phorImager. Expression data are given as ratios to the co-amplified internal standard (TATA-binding protein).
Statistical Analysis-Results are presented as mean Ϯ S.E. (n Ͼ 2) or as mean Ϯ range (n ϭ 2). Wilcoxon's matched-pair test was used, and p Ͻ 0.05 was chosen as the level of significance.

IL-1␤ Activates p38 and ERK1/2 in Rat
Islets-To investigate whether p38 and ERK1/2 were activated by IL-1␤ in rat islets, we first assayed the kinase activities toward Elk-1, ATF2, and Hsp25 in whole cell lysates of IL-1␤ stimulated islets. Detectable kinase activity using Elk-1, ATF2, and Hsp25 as substrates was present in control islets (0Ј, baseline) (Fig.  1A). IL-1␤-enhanced phosphorylation of Elk-1 was evident within 20 min, peaked at 90 min with a 4.1-fold increase over baseline, and was sustained at 12 h. Enhanced ATF2 phosphorylation was found within 1 min, with a maximum 3.8-fold activation of phosphotransferase activity at 20 min, and was sustained at 12 h. Increased Hsp25 phosphorylation was apparent within 1 min, with a 3.1-fold peak activity at 20 min, and was sustained at 12 h.
To demonstrate that the IL-1␤-stimulated kinase activities toward Elk-1, ATF2, and Hsp25 in islets were associated with the presence and increased phosphorylation of ERK1/2 and p38, we performed Western blotting with phosphospecific an-tibodies recognizing only Tyr 204 -and Tyr 182 -phosphorylated ERK1/2 and p38, respectively. As seen in Fig. 1B, a time-dependent IL-1␤-mediated enhancement of the phosphorylation of ERK1/2 (pERK1/2) and p38 (pp38) was detected. The time courses of IL-1␤-mediated phosphorylation of ERK1 and 2 were similar: a weak phosphorylation over baseline (0Ј) at 2.5-5 min, followed by a peak phosphorylation at 20 min, with only a slight decrease until 3 h, after which a marked decline was observed. However, a weak phosphorylation over baseline level was seen even at 24 h. Enhanced p38 phosphorylation by IL-1␤ was visible at 1 min, peaking and reaching a plateau between 20 min and 12 h and not detectable after 24 h. Control experiments using antibodies to total ERK1/2 and p38 (ERK1/2 and p38) showed no effects of IL-1␤.
Detection of ERK1/2 and p38 Activity in RIN Cells-To investigate whether the ERK1/2 and p38 activities found in intact islets were due to the presence of these kinases in ␤-cells, the RIN ␤-cell line was assayed for ERK1/2 and p38 activities. The RIN ␤-cell line is comparable to primary ␤-cells in terms of IL-1␤-mediated NO production, iNOS expression, and cytotoxicity, albeit at a higher concentration than is needed in islets (36,37). Based on the IL-1␤ time course from intact islets (Fig.  1A), an IL-1␤ exposure period of 20 min was used. IL-1␤ stimulated ATF2 and Hsp25 kinase activities in a dose-dependent manner, whereas IL-1␤-induced Elk-1 kinase was not further activated by an increased concentration of IL-1␤ above 150 pg/ml ( Fig. 2A). The binding of phosphospecific antibodies (pERK1/2 and pp38) showed that both ERK1/2 and p38 were phosphorylated in RIN cells after a 20-min exposure to 1500 pg/ml IL-1␤ (Fig. 2B). p38 and ERK1/2 activation by IL-1␤ was

FIG. 2. IL-1␤ causes phosphorylation of MAPK-substrates (A) and ERK1/2 and p38 (B) in RIN 5AH-T2B cells.
A, lysates of RIN cells exposed to 0, 150, or 1500 pg/ml IL-1␤ for 20 min were analyzed by the whole cell lysate kinase assay using [␥-32 P]ATP and GST-Elk-1 (Elk-1), GST-ATF2 (ATF2), and Hsp25. Following SDS-PAGE, the products were visualized by autoradiography. A representative autoradiogram is shown (upper panel). Phosphorylation was quantified by Phos-phorImager analysis and presented as mean Ϯ S.E. (n ϭ 3) relative to control cells (0) treated without IL-1␤ (lower panel). B, lysates of RIN cells exposed to 0 (c) or 1500 (IL-1␤) pg/ml IL-1␤ for 20 min were analyzed by Western blotting as described in Fig. 1B. Total ERK1/2 and p38 and phosphorylated ERK1/2 (pERK1/2) and p38 (pp38) are indicated. Results shown are representative of n ϭ 2. C, ERK1/2 (upper panel) and MAPKAP-K2 (lower panel) were immunoprecipitated from lysates of RIN cells stimulated with (IL-1␤) or without (c) 1500 pg/ml IL-1␤, and their activities were measured in an immunocomplex kinase assay as described in Fig. 1C with GST-Elk-1 (Elk-1) and Hsp25 as substrates, respectively. Phosphorylation was quantified by Phosphor-Imager analysis and is indicated below the autoradiograms as ERK1/2 and MAPKAP-K2 activity relative to control cells (c) treated without IL-1␤.  ). B, lysates of IL-1␤-exposed islets were analyzed by Western blotting. Lysates from pooled duplicate experiments were separated on two gels and blotted onto two membranes. One of the membranes was probed with phosphospecific ERK1/2 (pERK1/2) antibody, washed, and then reprobed with phosphospecific p38 (pp38) antibody. The other membrane was probed with ERK1/2 antibody, washed, and then reprobed with p38 antibody. Results shown are representative of two individual experiments. C, ERK1/2 (upper panel) and MAPKAP-K2 (lower panel) were immunoprecipitated from lysates of IL-␤-exposed islets with specific antibodies, and their activities were measured in immunocomplex kinase assay with GST-Elk-1 (Elk-1) and Hsp25 as substrates, respectively. Phosphorylation reactions were initiated by the addition of [␥-32 P]ATP. Following SDS-PAGE, phosphorylated proteins were visualized by autoradiography. Each experiment was performed twice, and representative autoradiograms are shown. Phosphorylation was quantified by PhosphorImager analysis and indicated below the autoradiograms as mean Ϯ range.
Inhibition of ERK1/2 and p38 in IL-1␤-exposed Islets-To dissect the relative contributions of ERK1/2 and p38 signaling pathways in the IL-1␤-induced Elk-1, ATF2, and Hsp25 kinase activities, islets were incubated with p38i and MEKi 1 h prior to IL-1␤ exposure (Fig. 3A). The MEKi inhibited both basal and IL-1␤ stimulated Elk-1 kinase activity without affecting the ATF2 and Hsp25 kinase activities. A maximal inhibition was seen at 100 M of MEKi. The p38i inhibited both basal and IL-1␤-induced ATF2 and Hsp25 kinase activities; it was more effective in Hsp25 kinase inhibition, which was completely blocked at 10 M, whereas this concentration of p38i inhibited IL-1␤-induced ATF2 kinase activity by 71%. When the inhibitors were combined, the kinase activities toward the three substrates were completely blocked.
The specificity of the inhibitors was further evaluated in an immunocomplex kinase assay, where the activities of immunoprecipitated ERK1/2 and MAPKAP-K2 in lysates of islets that had been preincubated with MEKi and/or p38i prior to IL-1␤ exposure were determined. MEKi inhibited IL-1␤-induced ERK1/2 activity and did not affect the activity of MAPKAP-K2, whereas p38i inhibited IL-1␤-stimulated MAPKAP-K2 activity but not that of ERK1/2 (Fig. 3B, top and middle panels). Because JNK has both Elk-1 and ATF2 kinase activities (20,38), the complete inhibition of ATF2 and Elk-1 kinase activities in whole cell lysates of IL-1␤ stimulated islets preincubated with MEKi and p38i questioned the involvement of JNK in IL-1␤ signaling in islets. However, we found substantial IL-1␤-stimulated JNK activity determined by c-Jun phosphorylation in an in vitro solid-phase kinase assay in islets, and as expected, JNK activity was unaffected by either of the inhibitors (Fig. 3B,  bottom panel). This indicates that JNK is neither an IL-1␤activated Elk-1 nor ATF2 kinase in islets.

Involvement of ERK1/2 and p38 in IL-1␤-mediated ␤-Cell
Dysfunction-MEKi and p38i were then used to evaluate the impact of ERK1/2 and p38 on IL-1␤-mediated ␤-cell dysfunction. NO production was not detectable from untreated islets, and neither MEKi nor p38i added alone or in combination resulted in NO production (Table I). At a low (25 pg/ml) IL-1␤ concentration, p38i caused a 35% decrease in the IL-1␤-induced islet NO production, whereas MEKi blocked NO production at that concentration of IL-1␤. At a high (150 pg/ml) IL-1␤ concentration, p38i and MEKi caused a 25 and 33% reduction, respectively of IL-1␤-induced islet NO production. However, in the presence of the two inhibitors, a synergistic effect was found, and IL-1␤-induced NO synthesis was completely blocked.
MEKi and p38i used either alone or in combination did not affect the glucose-stimulated insulin release from untreated islets (Table II). A low (25 pg/ml) IL-1␤ concentration did not cause a significant decrease in glucose-stimulated insulin release, and no effect of the inhibitors was found. The 62% inhibition of glucose-stimulated insulin release mediated by a high (150 pg/ml) IL-1␤ concentration was attenuated by p38i by 33%, whereas MEKi was ineffective and no synergism between the two compounds was detected.
ERK1/2 and p38 Regulate IL-1␤-induced Islet NO Synthesis at the Transcriptional Level-To investigate whether ERK1/2 and p38 controlled IL-1␤-induced islet NO production at the transcriptional level, reverse transcription-PCR of iNOS mRNA in IL-1␤-exposed islets preincubated with/without the inhibitors was performed. iNOS mRNA was not detectable in  untreated islets (Fig. 4). IL-1␤-induced iNOS was expressed to a 3.3-fold greater magnitude at 6 h compared to 24 h of IL-1␤ exposure. IL-1␤-induced iNOS was inhibited by both of the inhibitors; the MEKi was slightly more effective at both 6 and 24 h of IL-1␤ exposure. Combined, the inhibitors reduced IL-1␤-induced iNOS by ϳ95% at both 6 and 24 h. Noncytokine Stimulation of ERK1/2 and p38 Activity and Islet NO Production-To investigate whether noncytokine activation of ERK1/2 and p38 was able to induce islet NO production, islets were exposed to hyperosmolarity. As shown in Fig. 5A, a weak phosphorylation of Elk-1, ATF2, and Hsp25 was found in control islets (0Ј, baseline). When exposed to hyperosmolarity (525 mosM) a marked time-dependent phosphorylation of the three substrates was found. Hyperosmolarity induced rapid (within 0.5 min) phosphorylation with a 2.2and 2.3-fold activation of ATF2 and Hsp25, respectively. A 3.1-fold activation of Elk-1 was evident within 2.5 min. The time points for peak values with 3.8-, 4.6-, and 3.6-fold activation of Elk-1, ATF2, and Hsp25, respectively, were very similar to those induced by IL-1␤ (see Fig. 1A). Hsp25 phosphorylation declined to baseline at 12 h, whereas the phosphorylation of ATF2 and Elk-1 was sustained until at least 12 h. Elk-1 phosphorylation was more sustained, and Hsp25 phosphorylation showed a more rapid descent when compared with IL-1␤-induced phosphorylation.
Phosphorylation of ERK1/2 and p38 by hyperosmolarity (525 mosM) was found by Western blotting (Fig. 5B), with detectable ERK1/2 (pERK1/2) and p38 (pp38) phosphorylations of 2.5 min to 12 h and of 0.5 min to 12 h, respectively, in agreement with the kinase activities found in the whole cell lysate kinase assay (see Fig. 5A). However, islets exposed to increasing osmolarity (285-525 mosM) for 24 h did not produce detectable NO (nitrite Ͻ2 pmol/islet/24 h; n ϭ 6). DISCUSSION Our data demonstrate that p38 and ERK1/2 are both activated by IL-1␤ in rat islets of Langerhans and in the rat insulinoma cell line RIN-5AH, implying that the ERK1/2 and p38 activity found in the intact islet most likely originates from the ␤-cells, in accordance with the recently reported p38 activation by IL-1␤ in the INS-1 ␤-cell line (39).
MEK-and p38-specific inhibitors individually reduced IL-1␤ stimulated islet NO synthesis (Table I) and iNOS mRNA (Fig.  4), indicating that both ERK1/2 and p38 are involved in IL-1␤mediated expression of iNOS. iNOS induction by ERK1/2 in ␤-cells may be explained by Elk-1-induced c-fos expression mediated through the serum response element (40). c-Fos com-bines with c-Jun to form the transcription factor activating protein-1 (41), and two activating protein-1 binding sites have been identified in the murine iNOS gene promoter (42). Activated p38 may induce ␤-cell iNOS via several transcription factors: 1) through activating protein-1, because p38-mediated expression of c-jun and c-fos has been observed (43); 2) through NF-B, because p38 seems to be required for NF-B-mediated transcriptional activation, but it affects neither NF-B DNA binding nor phosphorylation of its subunits (44,45) and the murine iNOS gene promoter region contains two NF-B binding sites (42), and because NF-B is involved in iNOS expression in ␤-cells (36,46); and 3) through p38-mediated cAMPresponsive element-binding protein activation (47), because cAMP-responsive element-binding protein has been reported to be involved in iNOS induction via the CAAT box in the murine iNOS promoter (48).
The involvement of ERK1/2 and p38 in IL-1 regulation of iNOS seems to be cell-specific because 1) IL-1 activation of ERK1/2 is necessary for iNOS expression in rat cardiac microvascular endothelial cells (23), 2) p38, but not ERK1/2, is involved in IL-1-mediated iNOS expression in mouse astrocytes (49), and 3) IL-1 stimulation of p38 down-regulates iNOS in rat messangial cells (50). Our data support this concept by showing that both ERK1/2 and p38 are required for IL-1-mediated iNOS expression in rat pancreatic ␤-cells.
ERK1/2 and p38 activation was found to be necessary for IL-1␤-mediated islet NO synthesis (Table I) and iNOS induction (Fig. 4). However, the observation that hyperosmolarity caused a maximal Elk-1, ATF2, and Hsp25 phosphorylation and ERK1/2 and p38 activation (Fig. 5) similar to that caused by IL-1␤ without inducing islet NO synthesis suggests that ERK1/2 and p38 activation is not sufficient to cause IL-1␤mediated islet NO synthesis. Thus, other signaling events activated by IL-1␤, but not provided by hyperosmolarity, are necessary. These signals could be involved in IB degradation and NF-B translocation to the nucleus because a recently identified IL-1␤ receptor associated kinase shares similarity with a protein kinase essential for activation of a NF-B ho-  4. p38i and MEKi inhibit IL-1␤-induced iNOS mRNA expression in islets. Islets were preincubated for 1 h with p38i and/or MEKi. Following stimulation with 150 pg/ml IL-1␤ for the indicated times, media were removed for nitrite determination, and islets were snap frozen. Total RNA was extracted, and cDNA was prepared. Reverse transcription-PCR was performed with iNOS primers in combination with TATA-binding protein primers as an internal standard. Samples were separated on polyacrylamide gels and visualized by autoradiography. Data are presented as mean of iNOS/TATA-binding protein ratios determined by PhosphorImager analysis. One of two representative experiments is shown. mologue in Drosophila (51). Another signal could be provided by interferon response factor-1 expression because interferon response factor-1 has been implicated in IL-1␤-induced iNOS expression in rat ␤-cells (52).
p38 but not ERK1/2 blockade attenuated inhibition of glucose-stimulated insulin release by 150 pg/ml of IL-1␤ (Table II), but because the MEKi and p38i were equally potent in decreasing ␤-cell NO synthesis induced by 150 pg/ml of IL-1␤, and because MEKi was even more efficient in decreasing iNOS mRNA than p38i, p38 must be involved in signaling of NOindependent effects of IL-1␤ in ␤-cells. The inhibitory effect of p38 activation on insulin release could be due to a direct action on the ␤-cell stimulus-secretion coupling, because p38 is slightly activated (1.5-fold) by 15 mM glucose (our media contained 11 mM glucose) in the INS-1 ␤-cell line (39). However, the observation that p38 inhibition did not affect glucose-stimulated insulin release from islets not exposed to IL-1␤ (Table  II), suggests that p38 is not involved in signaling pathways of glucose-stimulated insulin release. Rather, the findings that diverse classes of DNA-damaging agents activate p38 (53), that a sustained activation of p38 has been associated with apoptosis (54,55), and that IL-1␤ induces apoptosis in ␤-cells (37,56) and the present finding of a sustained (Ͼ12 h) IL-1␤-mediated p38 activation in islets make it more likely that p38 induces an apoptotic signal in ␤-cells. We are currently investigating the possible involvement of p38 on IL-1␤-mediated ␤-cell cytotoxicity.
Even though combination of MEKi and p38i inhibited IL-1␤-induced islet iNOS expression by 95% and completely blocked islet NO synthesis, the IL-1␤-mediated inhibition of insulin release was only attenuated by 33%, indicating that islet NO production is not sufficient in causing ␤-cell death and that other mediators of NO-independent signaling pathways, in addition to p38, are involved. A possible mediator could be JNK, which is activated by IL-1␤ in both islets (Fig. 3) and RIN-cells (12) and is involved in apoptosis (54,57).
In conclusion, this study suggests that ERK1/2 and p38 are involved in IL-1␤ signaling in ␤-cells, that they are necessary but not sufficient in causing IL-1␤-induced ␤-cell NO synthesis by controlling iNOS gene transcription, and, furthermore, that p38 is involved in signaling of NO-independent effects of IL-1␤ in ␤-cells. These observations make p38 inhibition a powerful approach to further elucidate the involvement of p38 in the pathogenesis of insulin-dependent diabetes mellitus.