Interleukin-12 (IL-12)/STAT4 Axis Is an Important Element for β-Cell Dysfunction Induced by Inflammatory Cytokines

Pathology driving β-cell loss in diabetes is poorly defined. Chronic subclinical inflammation is associated with β-cell dysfunction. Acute in vitro exposure of islets and β-cells to an inflammatory cytokine cocktail (IL-1β/TNF-α/IFN-γ) results in loss of cell function and viability. The contribution of each cytokine alone or in combination has been evaluated in homogeneous mouse β-cell lines and primary mouse islets. Cytokine cooperation is required for β-cell apoptosis with the most potent combinations including IL-1β. Single cytokine exposure did not induce β-cell apoptosis. Expression of endogenous interleukin-12 in β-cells correlated with inflammatory cytokine combinations that induced β-cell apoptosis. Uncoupling of the IL-12 axis by a block of IL-12 production, inhibition of IL-12 receptor/ligand interaction or disruption of IL-12 receptor signaling conferred protection to β-cells from apoptosis induced by inflammatory cytokine stimulation. Signaling through STAT4 is indicated since disruption of IL-12 concomitantly reduced inflammatory cytokine stimulation of endogenous IFN-γ expression. Primary mouse islets isolated from mice deficient in STAT4 show resistance to inflammatory-cytokine-induced cell death when compared to islets isolated from wild type mice. Collectively, the data identify IL-12 as an important mediator of inflammation induced β-cell apoptosis. Modulation of IL-12/STAT4 signaling may be a valuable therapeutic strategy to preserve islet/β-cell viability in established diabetes.


Introduction
Worldwide diabetes incidence is predicted to exceed 592 million by 2035 [1].Diabetes is a complex metabolic disease being influenced by numerous factors.A core feature is the failure of insulin producing β-cells for both type 1 (T1DM) and type 2 (T2DM) diabetes [2,3].Causes of β-cell failure are poorly understood, but chronic sub-clinical inflammation is a contributing factor.Inflammation is a feature of both T1DM and T2DM [4][5][6][7][8][9][10][11][12].Acute exposure of islets to inflammatory cytokines ex vivo promotes islet stress and dysfunction, including loss of glucose-stimulated insulin secretion, increased apoptosis and elevated expression of various marker genes, including monocyte chemoattractant protein-1 (MCP-1) [13,14].Elevated MCP-1 in islets occurs during early insulitis in experimental diabetes mouse models and is used clinically to assess transplantable human islets [15].Induction of islet dysfunction by inflammatory cytokines, especially the triple cytokine combination of IL-1β/TNF-α/IFN-γ, is extensively reported [16].The cellular responses in islets and β-cells to inflammatory cytokine exposure are less well characterized.
Several cellular effects have been associated with exposure of β-cells to inflammatory cytokines [17,18].A candidate mediator of β-cell dysfunction is interleukin-12 (IL-12).Local production of IL-12 has been reported and may establish an islet:immune interface for targeted βcell destruction [19].IL-12, a heterodimeric ligand composed of subunits, p35 (IL-12 p35) and p40 (IL-12 p40), coordinates a Th1 immune response by inducing expression of IFN-γ.Principally considered an immune factor, IL-12 has also been identified in non-immune cells, including islets [19].Being a key mediator in disease pathologies, several approaches to uncouple IL-12 action have been identified.STA-5326 (Apilimod1) is a small molecular weight compound that inhibits c-Rel translocation from the cytoplasm to the nucleus and disrupts transcription of both IL-12 p35 and IL-12 p40 [20][21][22][23].Lisofylline (LSF) is a methylxanthine metabolite of Pentoxifylline that inhibits IL-12 signaling activity.LSF limits commitment to T-helper 1 cell development and IFN-γ production [24].LSF stopped onset of Type 1 diabetes in NOD mice [25].Antibodies that bind, sequester and neutralize IL-12 p40, eg Usterkinumab1 and Bria-knumab1 have proven clinical efficacy in the autoimmune condition psoriasis [26][27][28][29].Antibody-mediated neutralization of IL-12 p40 in islets conferred protection to β-cell dysfunction mediated by inflammatory cytokines [19].Ligation of the IL-12 ligand to its heterodimeric receptor primarily activates (phosphorylates) signal transducer and activator of transcription 4 (STAT4).Genetic deletion studies show STAT4 is an important factor in elevating susceptibility to several autoimmune diseases.In terms of diabetes, NOD mice deficient in STAT4 do not develop spontaneous diabetes unlike wild-type NOD mice [30,31].
Exposure of islet β-cells to pro-inflammatory cytokines results in β-cell dysfunction [14,19].The current report has identified a pivotal role for IL-12 and IL-12 mediated STAT4 signaling in the development of β-cell apoptosis.These data identify potential therapeutic targets for preservation of β-cell function and/or β-cell survival in established diabetes.

Ethics Statement and Mouse Islets
All protocols and procedures were performed in accordance with the "Principles of laboratory animal care" (NIH publication no.85-23), AAALAC, and approved by Institutional Animal Care and Use Committee (IACUC protocol #11-013) at Eastern Virginia Medical School.Islets were isolated from C57BL/6J (Jackson Laboratory, Bar Harbor, ME) mice between the ages of 8 to 24 weeks by common bile duct cannulation and collagenase digestion [25].STAT4ko mice (on C57BL/6J background; gift from Dr. Mark Kaplan, Indiana University) between the ages of 8 weeks to 28 weeks were used for islet isolation.Individual islets were hand picked and placed in 1 mL RPMI media (Life Technologies, Grand Island, NY) supplemented with 10% fetal bovine serum/antibiotics.
Fluorescence microscopy: Treated cells were washed in cold PBS and then incubated in cold PBS containing 0.1 μM YO-PRO-1 (Life Technologies) and 1 μg/mL Propidium Iodide (PI) at 4°C for 30 minutes.For cell lines, five random fields per well were analyzed and for islets, all islets were analyzed.The densitometric fluorescence value (green and red channels) were analyzed using ImageJ 1.42q (http://rsb.info.nih.gov/ij) to determine the fluorescence for each treatment.An apoptotic index was determined by relative expression of normalized signal with PIC being defined as unity.Images were captured with Axiophot (Zeiss, Jena, Germany) and Axiovision (Zeiss) image analysis.
Caspase-3 detection: Islets or cell line were treated for 4 hours without or with inhibitor.Cleavage of pro-caspase-3 was measured using a caspase-3 assay kit (BD Pharmigen, Franklin Lakes, NJ) according to the manufacturer's instructions.Fluorescence was measured (Spectra-Max, Molecular Devices, Sunnyvale, CA) using excitation wavelength 380 nm and emission wavelength 440 nm.

Real-time PCR
Isolated islets from wild-type mice and STAT4ko mice or βTC-3 cells were treated with varying combinations of PICs for either 4 or 24 hours without or with inhibitor.After treatment, total RNA was isolated (RNeasy Mini Kit; Qiagen, Valencia, CA) and transcribed using murine leukemia virus reverse transcriptase (Life Technologies) and random hexamers (Life Technologies) using a 20μL reaction volume.RT-PCR was performed in a CFX96 Thermal Cycler (Bio-Rad, Hercules, CA).Total reaction volume was 25μL, which consisted of 3μL cDNA (five-fold dilution) and Jump Start Taq Polymerase (Sigma-Aldrich, St Louis, MO).The primers used with SYBR Green 1 (Molecular Probes, Carlsbad, CA) probes are shown in Table 1.Taqman primers were used for IL-12 p40, IL-12 p35, and IFN-γ (Life Technologies).All RT-PCR reactions were performed in triplicate.The housekeeping gene GAPDH was used to normalize the data.The 2 -ΔΔCT method was used to analyze the data.

Data Analysis
Experiments were performed at a minimum in triplicate.For statistical analysis, student t test or one-way ANOVA with Tukey post hoc testing (Prism 4.0, Graph-Pad Software, La Jolla, CA) were used to determine statistical significance (95% CI and p < 0.05).Data set for analyses is provided (S1 Appendix).
Triple cytokine (PIC)-treatment is a more efficacious inducer of cell death in islets compared to treatment either with a single cytokine or dual combination of these pro-inflammatory cytokines.

Effect of pro-inflammatory cytokine treatment on MCP-1 gene expression
MCP-1 is a chemokine that is released from islets under conditions of stress and is a recognized marker of β-cell stress.To determine if cytokine-induced changes in MCP-1 corresponded with induction of cell death in mouse islets, expression of MCP-1 gene was measured.The greatest increase in MCP-1 expression occurred in mouse islets treated with triple cytokines (PIC; Fig 2A).Relative to MCP-1 following PIC-treatment (defined as 100%) the expression of MCP-1 by single cytokines at 4-fold concentration was 41.5% ± 17.5 IL-1β, 29% ± 9 TNF-α, and 10.7% ± 2.3 IFN-γ.Expression of MCP-1 was significantly less than PICs in islets treated with single cytokine at 1-fold, 2-fold, and 4-fold doses (p < 0.001).

Pro-inflammatory cytokine treatment induces IL-12 gene expression
A candidate pathway in the development of β-cell dysfunction in diabetes is expression and production of IL-12 and IL-12 mediated signaling.To determine the influence of pro-inflammatory cytokine treatment on the IL-12 pathway, gene expression for the IL-12 ligand (IL-12
STAT4 is a second messenger activated by IL-12 receptor ligation that transcriptionally activates the IFN-γ gene.To determine if STAT4 signaling is important for cytokine-induced islet apoptosis, the effect of cytokine combinations was compared between wild-type islets and islets from mice genetically deficient in STAT4 (STAT4ko islets).Cell death in mouse islets was determined by fluorescence (representative images in Fig 9A).Islet death occurred in WT islets treated overnight either with PICs (1.44 ± 0.11 RFU), IL-1β/TNF-α (1.37 ± 0.09 RFU), and IL-1β/IFN-γ (1.37 ± 0.08 RFU) relative to control (1.00 ± 0.16 RFU) (Fig 9C).Fluorescent intensity was measured for each treatment in the STAT4ko islets and plotted (Fig 9B).Following treatment with PICs, cell death in STAT4ko islets was significantly less than in PIC-treated WT islets (p < 0.001).Cell death that was observed in the STAT4ko islets treated with the dual cytokine combinations was not significantly different to control.Comparing islets from WT

Discussion
This study has explored the relationship of cytokine soluble mediators elevated in inflammation to β-cell dysfunction.Specifically, identification of pathways key to inflammation-  in [4]; [33]).Inflammation is an increasingly recognized feature of diabetes pathogenesis [34][35][36].Clinical focus has recently concentrated on the contribution IL-1β to β-cell dysfunction.A number of approaches have aimed to neutralize IL-1β in a bid to assess the therapeutic effectiveness in human diabetes [37][38][39].
The relative efficacy for inflammatory cytokines, in combination, to induce β-cell apoptosis have been assessed using the model systems of this study.The cytokine cocktail of IL-1β/TNFα/IFN-γ effectively induced β-cell dysfunction/death.Each single cytokine alone at equivalent concentration, or up to four-fold that used in the triple cocktail, did not induce β-cell apoptosis.These data suggest that second messengers specific for each cytokine combine to induce β-cell death rather than a mere stimulation of a critical density in common signaling pathways.Dual cytokine combinations showed an apparent hierarchy for induction of β-cell apoptosis such that pairs including IL-1β (IL-1β/TNF-α or IL-1β/IFN-γ) were more potent than the combination of TNF-α/IFN-γ.Paired cytokines were less effective than triple cytokines.These results support an important role for IL-1β for induction of β-cell dysfunction and death while additionally highlighting the necessitated contribution of other inflammatory factors.As a therapeutic in diabetes, reduction of IL-1β, either by soluble receptor antagonism or antibody sequestration and neutralization was not effective in human clinical trials [40].
The cytokine interleukin-12 (IL-12) has been shown to be upregulated in β-cells following exposure to triple cytokines [19].Exogenous IL-12 stimulated β-cell dysfunction and death Expression of IL-12 has been linked with autoimmune diabetes.In NOD mice, IL-12 plays a significant role in the transition from non-destructive to destructive insulitis [41].In human type 1 diabetes (TIDM), the genetic susceptibility locus, IDDM18, is located near a regulatory allele of the IL-12 p40 gene (IL-12B) [42], and single nucleotide polymorphisms in this region associate with an earlier age of T1DM onset and accelerated deterioration in glycemic control [43,44].Systemic daily administration of IL-12 to NOD mice increased T1DM incidence whereas addition of an IL-12 antagonist decreased T1DM incidence [41,45,46].During the development of diabetes in NOD mice and BB rats the expression of endogenous IL-12 p40 and IFN-γ increase prior to diabetes onset [47,48].To address if local expression of IL-12 in βcells could initiate diabetes, Holtz et al, generated transgenic mice expressing IL-12, or monomer IL-12 chains in the β-cell [49].RIP-IL-12, but not monomer-p35 or-p40, mice developed pancreatic islet inflammation.This was associated with an elevation in IFN-γ.Transgenic deletion of IL-12 or IFN-γ did not fully protect against diabetes development [50,51].Thus, additional pathways may contribute to β-cell dysfunction or mechanisms that are independent of IL-12 and IFN-γ may develop as a homeostatic response in transgenic mice lacking IL-12 or IFN-γ [51][52][53].Studies with the IL-12 natural antagonist (p40) 2 identified a key role for IL-12 in the development of T1DM [45,46,54].

Fig 1 )
. A significant increase in cell death in PIC-treated islets was detected using the fluorescent viability dyes YO-PRO-1 (green) and propidium iodide (red) when compared to non-cytokine treated control (Ctl) islets.Triple cytokine treatment (PIC) induced a significant increase in fluorescence when compared to untreated islets (p < 0.05).Representative images are shown in Fig 1A.The data is quantified in Fig 1C.Relative fluorescence units (RFU) for triple cytokine treated islets was 17.9 ± 3.3 RFU as compared to 2.9 ± 1.2 RFU for control islets.To determine if individual cytokine exposure was sufficient for the induction of islet death, islets were treated with single cytokines.Representative images are shown in Fig 1B.The data is quantified in Fig 1D.Islets were incubated with each cytokine at the equivalent dose used in the PIC cocktail (one-fold; 1X) and at four-fold this dose (4X) (Fig 1Band 1D).Incubation either with IL-1β, TNF-α, or IFN-γ alone (1X) did not induce elevated cell death relative to control islets.Increasing the dose of each cytokine 4-fold did not significantly induce cell death.Islets that were treated with single cytokines at either a 1X or 4X dose showed significantly less cell death than islets treated with PICs (p < 0.001) and no significant difference to control (Fig1D).

Fig 9 .
Fig 9. Islets from STAT4 deficient mice are resistant to PIC-induced apoptosis.(A) Fluorescence microscopy was used to image apoptotic islets labeled with YO-PRO-1 (green) and PI (red).Islets from WT mice (top row) or islets from STAT4ko mice (bottom row) were treated with PICs, IL-1β/TNF-α, IL-1β/IFNγ, or TNF-α/IFN-γ.(B) Graph shows quantified apoptosis from cytokine treated STAT4ko islets.(C) Graph shows quantified apoptotic index from cytokine treated WT islets vs STAT4ko islets.* p < 0.05, ** p < 0.01, *** p < 0.001 relative to ctl, +++ p < 0.001, ### p < 0.001 relative to WT PIC treatment.Three groups of each condition were treated and analyzed.doi:10.1371/journal.pone.0142735.g009 [19].IL-12 is a heterodimeric ligand formed by two protein chains IL-12 p40 and IL-12 p35.A heterodimer of IL-12 p40 and a distinct protein chain IL-23 p19, forms functional interleukin-23 (IL-23).The effect of dual cytokine stimulation was assessed on expression of IL-12 chains in β-cells.Dual cytokine pairings that included IL-1β induced expression of IL-12 to levels comparable to triple cytokines.Induction of IL-12 with TNF-α/IFN-γ was significantly lower.Overall the pattern of IL-12 expression by inflammatory cytokine combinations conformed with induction of cell death.These data show IL-12 expression is concomitant with cell death and suggested the importance of IL-12 as a mediator of cell death induced by the three cytokines.The studies in this report further explore the role of IL-12 and IL-12 mediated signaling in beta cell apoptosis.Activation of the IL-12 axis was interrupted in three separate approaches; inhibition of IL-12 ligand expression, inhibition of IL-12 receptor ligation and inhibition of IL-12 signaling.Disruption of IL-12 was effective at protecting β-cells from apoptosis associated with inflammatory cytokine stimulation.