Laura Marroqui,1 Reinaldo Sousa Dos Santos,1 Tina Fløyel,2 Fabio A. Grieco,1

Izortze Santin,1,3 Anne Op de beeck,1 Lorella Marselli,4 Piero Marchetti,4

Flemming Pociot,2 and Decio L. Eizirik1

TYK2, a Candidate Gene for Type 1Diabetes, Modulates Apoptosis and theInnate Immune Response in HumanPancreatic b-CellsDiabetes 2015;64:3808–3817 | DOI: 10.2337/db15-0362

Pancreatic b-cells are destroyed by an autoimmune attackin type 1 diabetes. Linkage and genome-wide associationstudies point to >50 loci that are associated with the dis-ease in the human genome. Pathway analysis of candidategenes expressed in human islets identified a central rolefor interferon (IFN)-regulated pathways and tyrosine ki-nase 2 (TYK2). Polymorphisms in the TYK2 gene predictedto decrease function are associated with a decreased riskof developing type 1 diabetes. We presently evaluatedwhether TYK2 plays a role in human pancreatic b-cell ap-optosis and production of proinflammatory mediators.TYK2-silenced human b-cells exposed to polyinosinic-polycitidilic acid (PIC) (a mimick of double-strandedRNA produced during viral infection) showed lesstype I IFN pathway activation and lower production ofIFNa and CXCL10. These cells also had decreased ex-pression of major histocompatibility complex (MHC) classI proteins, a hallmark of early b-cell inflammation in type 1diabetes. Importantly, TYK2 inhibition prevented PIC-induced b-cell apoptosis via the mitochondrial path-way of cell death. The present findings suggest thatTYK2 regulates apoptotic and proinflammatory path-ways in pancreatic b-cells via modulation of IFNa sig-naling, subsequent increase in MHC class I protein,and modulation of chemokines such as CXCL10 thatare important for recruitment of T cells to the islets.

Type 1 diabetes is a chronic autoimmune disease char-acterized by islet inflammation (insulitis) and specificdestruction of pancreatic b-cells. Insulitis occurs in the

context of a “dialog” between invading immune cells andthe target pancreatic b-cells (1), which includes upreg-ulation of islet human leukocyte antigen (HLA) class Iexpression in b-cells (2) and production of chemokinessuch as CXCL10 by the islet cells (3–5).

Susceptibility to type 1 diabetes is strongly linked tothe genetic background. Recent linkage and genome-wideassociation studies have identified .50 genetic variantswith association to type 1 diabetes, explaining ;80% ofthe heritability (6). It has been generally assumed thatcandidate genes for type 1 diabetes modify risk for thedisease by acting at the immune system level (7). Recentdata (8,9 and present findings), however, indicate thathuman pancreatic b-cells express .80% of type 1 diabetescandidate genes. Furthermore, another study comparingsingle nucleotide polymorphism (SNP) locations againstchromatin maps for different cell types indicates a pri-mary signature of T1D SNPs in T-cell enhancers but alsoa highly significant (P , 1027) enrichment in pancreaticislet enhancers (10). These genes may contribute to type 1diabetes by regulating important pathways in the b-cells,such as antiviral responses, innate immunity, and activa-tion of apoptosis (6,11–13).

Tyrosine kinase 2 (TYK2) is a member of the Januskinase (JAK) family of tyrosine kinases. These kinasesplay a critical role in the intracellular signaling ofseveral cytokines and type I interferons (IFNs) throughphosphorylation and activation of signal transducersand activators of transcription (STATs) (14). TYK2 hasbeen associated with several autoimmune diseases, such as

1ULB Center for Diabetes Research, Medical Faculty, Université Libre de Brux-elles, Brussels, Belgium2Department of Pediatrics, Herlev University Hospital, Herlev, Denmark3Endocrinology and Diabetes Research Group, BioCruces Health Research In-stitute and Spanish Biomedical Research Centre in Diabetes and AssociatedMetabolic Disorders, Barakaldo, Spain4Department of Clinical and Experimental Medicine, Pancreatic Islet Laboratory,University of Pisa, Pisa, Italy

Corresponding author: Laura Marroqui, lmarroqu@ulb.ac.be.

Received 17 March 2015 and accepted 12 June 2015.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db15-0362/-/DC1.

© 2015 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, andthe work is not altered.

3808 Diabetes Volume 64, November 2015

ISLETSTUDIES

systemic lupus erythematosus, multiple sclerosis, rheu-matoid arthritis, and type 1 diabetes (15,16). Six SNPs(rs34536443, rs2304256, rs280523, rs280519, rs12720270,and rs12720356) of the TYK2 gene have been studied inrelation to autoimmunity and inflammation, and two ofthese, rs34536443 and rs2304256, are significantly as-sociated with multiple autoimmune and inflammatorydiseases, with the minor allele decreasing susceptibilityto these diseases (15). Importantly, the SNP rs2304256:C.A (odds ratio for A vs. C = 0.86), located in exon 8 atchromosome 19p13.2, is associated with protectionagainst type 1 diabetes. rs2304256 is a nonsynonymousSNP causing a missense mutation in TYK2, where the Aallele leads to a substitution of valine for phenylalanine atposition 362 in the JAK-homology 4 (JH4) region (15,16).This region is critical for both the interaction betweenTYK2 and IFNAR1 (17) and preserving expression ofthe IFNAR1 at the cell membranes (18). Thus, it hasbeen suggested that this SNP reduces TYK2 function,resulting in a decreased susceptibility to IFN-relatedautoimmune diseases (15).

We presently analyzed the role of TYK2 in immune-mediated human pancreatic b-cell apoptosis and localinflammation. The data obtained indicate that TYK2inhibition prevents double-stranded RNA (dsRNA) (a by-product of viral proliferation, tested here as polyinosinic-polycitidilic acid [PIC]) induced apoptosis in humanpancreatic b-cells through the reduction of the type IIFN–STAT signaling and consequent prevention of theincrease of major histocompatibility complex (MHC)class I protein expression.

RESEARCH DESIGN AND METHODS

Culture of Human Islets and EndoC-bH1 Humanb-Cells and Cell TreatmentsHuman islets were isolated from 16 organ donors withoutdiabetes (Supplementary Table 1) with approval from thelocal ethical committee in Pisa, Italy. Isolation of humanislets was done by collagenase digestion and density-gradientpurification (19). Subsequently, isolated islets were culturedin M199 medium containing 5.5 mmol/L glucose (19).Within 1–5 days of isolation, the human islets were shippedto Brussels. After arrival in Brussels and overnight recov-ery, the human islets were dispersed and cultured in Ham’sF-10 medium containing 6.1 mmol/L glucose, 2 mmol/LGlutaMAX, 50 mmol/L 3-isobutyl-1-methylxanthine, 1%charcoal-absorbed BSA, 10% FBS, 50 mg/mL streptomycin,and 50 units/mL penicillin. The proportion of b-cells inthe preparations was determined by immunocytochem-istry for insulin (9).

The EndoC-bH1 human b-cell line (provided by Dr. R.Scharfmann, Centre de Recherche de l’Institut du Cerveau etde la Moelle Épinière, Paris, France) was cultured in platescoated with Matrigel-fibronectin (respectively 100 and 2mg/mL) in low-glucose DMEM as previously described (20).

Rat INS-1E cells were used for experiments to evaluateTYK2 stability. These cells were provided by Dr. C. Wollheim

(University of Geneva, Geneva, Switzerland) and werecultured in RPMI 1640 GlutaMAX-I medium (Invitrogen)(21).

The cells were treated with human IFNa (specific ac-tivity 1.8 3 108 units/mg; PeproTech Inc., Rocky Hill, NJ)at 2,000 units/mL or transfected with 1 mg/mL of thesynthetic dsRNA analog PIC (InvivoGen, San Diego, CA)(22). For the exposure to IFNa, medium without FBS wasused, whereas for small interfering RNA (siRNA) and PICtransfection, medium without BSA and antibiotics wasused.

Culture of B Lymphoblastoid Cell Lines and CellTreatmentB lymphoblastoid cell lines (BLCLs) from 12 HapMapCEPH founders were obtained and cultured as describedpreviously (12). Out of the 12 BLCLs, 7 had the CC geno-type and 5 the AA genotype corresponding to the TYK2SNP rs2304256. BLCLs were seeded in six-well cultureplates (1.0 3 106 cells/well), precultured for 1 day, andthen left untreated or stimulated with 1,000 units/mLhuman recombinant IFNa (PeproTech, Rocky Hill, NJ)for 30 min. BLCLs were washed in Hanks’ balanced saltsolution and lysed in M-PER Mammalian Protein Extrac-tion Reagent supplemented with Halt Protease and Phos-phatase Inhibitor Cocktail (all from Pierce, Rockford, IL).Protein concentrations were measured using the DC Pro-tein Assay (Bio-Rad, Hercules, CA) and Western blot forSTAT1 performed as described below.

RNA InterferenceDispersed human islets or the EndoC-bH1 human b-cellswere transfected with 30 nmol/L of two different siRNAstargeting TYK2 (TYK2#1, 59-CCAUCUGGUAAUAAACUCATT-39, and TYK2#2, 59-GAUGCUAUAUUUCCGCAUATT-39; Qiagen) or Allstars Negative Control siRNA(siCTRL; Qiagen) using the Lipofectamine RNAiMAXlipid reagent (Invitrogen) in a two-step transfectionprotocol. In this approach, cells were exposed for 16 h to30 nmol/L siTYK2 or siCTRL, washed, and allowed to re-cover in culture for 24 h. The cells were then exposed againfor 16 h to the same siRNAs, allowed to recover for 24 hin culture, and then used for the subsequent experi-ments. The siCTRL used does not affect b-cell gene ex-pression, function, or viability in both human islets (23)and EndoC-1bH1 cells (data not shown). In additionalexperiments, human islet cells were transfected with a pre-viously validated siRNA targeting the ubiquitin-specificpeptidase 18 (USP18) (22).

Assessment of Cell ViabilityMeasurements of living, apoptotic, and necrotic cells weredetermined after incubation with the DNA-binding dyesHoechst 33342 (HO) and propidium iodide (PI) as previouslydescribed (13). These measurements were performed bytwo different observers, one of them unaware of sam-ple identity (the level of agreement between the twoobservers was always .90%). At least 500–600 cells were

diabetes.diabetesjournals.org Marroqui and Associates 3809

counted in each experimental condition. In some experi-ments, apoptosis was confirmed by a second method,namely, Western blot for cleaved (activated) caspase 3, asdescribed below.

mRNA Extraction and Real-Time PCRPoly(A)+ mRNA was extracted using Dynabeads mRNADIRECT kit (Invitrogen) and reverse transcribed as pre-viously described (24). Quantitative real-time PCR wascarried out using SYBR Green and compared with a stan-dard curve (25). The housekeeping gene b-actin was usedto correct expression values. b-Actin mRNA expressionwas not modified by the different treatments presentlyused (data not shown). The primers used herein are listedin Supplementary Table 2.

Western Blot AnalysisCells were washed with cold PBS and lysed in Laemmlibuffer. Immunoblot analysis was performed by over-night incubation with the antibodies listed in Sup-plementary Table 3. Membranes were incubated withsecondary peroxidase-conjugated antibody (anti-IgG(H+L)-HRP; Invitrogen) for 1 h at room temperature.Immunoreactive bands were visualized using theSuperSignal West Femto chemiluminescent substrate(Thermo Scientific), detected using ChemiDoc XRS+(Bio-Rad), and quantified with the Image Laboratorysoftware (Bio-Rad).

Measurement of Chemokine Secretion by ELISASupernatants from dispersed human islets and EndoC-bH1 human b-cells were collected after treatments todetermine the levels of CXCL10, IFNa, and IFNb usingcommercially available ELISA kits (R&D Systems, Abing-don, U.K.).

Immunofluorescence and Flow CytometryImmunofluorescence was performed as previously de-scribed (26). In brief, cells were plated on polylysine-coated coverslips, treated with intracellular PIC or IFNaduring 24 h, and fixed with 4% paraformaldehyde. Afterpermeabilization with 0.3% Triton X-100, cells wereincubated overnight with the primary antibody rabbitanti–MHC class I (W6/32) (1:1,000) or rabbit anti–cleavedcaspase 3 (1:100), or for 1 h with mouse monoclonal anti-insulin (1:1,000). Alexa Fluor 568 goat anti-rabbit IgGor rabbit anti-mouse IgG and Alexa Fluor 488 goat anti-mouse IgG were respectively applied for 1 h (the anti-bodies used are listed in Supplementary Table 3). Afternuclear staining with Hoechst, coverslips were mountedwith fluorescent mounting medium (DAKO, Carpintera,CA) and immunofluorescence was visualized on a Zeissmicroscope equipped with a camera (Zeiss-Vision, Munich,Germany). Images were acquired at340 magnification andanalyzed using AxiVision software.

The same protocol used for immunofluorescence, butwithout permeabilization, was used for flow cytometry.Cells were detached by a mild trypsin treatment, suspendedin 2% paraformaldehyde-containing PBS, and then analyzed

by flow cytometry (FacsCalibur; BD Biosciences, San Jose,CA). Analysis was performed using CellQuest Pro softwareversion 6.0 (BD Biosciences, San Jose, CA). The cellularpopulations were selected based on size and cell granularityand analyzed by red fluorescence.

Statistical AnalysisData are presented as means 6 SEM. Comparisons wereperformed by two-tailed paired Student t test or by ANOVAfollowed by Student t test with Bonferroni correction, asindicated. A P value ,0.05 was considered as statisticallysignificant.

RESULTS

Ingenuity Pathway Analysis of Candidate Genes forType 1 Diabetes Indicates IFN Signaling as anImportant Pathway in b-CellsTo better understand how candidate genes for type 1diabetes affect human pancreatic b-cells, we analyzed51 previously described candidate genes (SupplementaryTable 4) using Ingenuity Pathway Analysis (Ingenuity Sys-tems, http://www.ingenuity.com) (Supplementary Fig. 1).The expression of these genes was compared against ourprevious RNA-seq data of five human islet preparations(9). Forty-two out of 51 genes (82%) were found expressed(i.e., reads per kilobase of transcript per million mappedreads [RPKM] .0.5) in human b-cells. Ingenuity PathwayAnalysis of the type 1 diabetes candidate genes expressedin human pancreatic b-cells identified as the three top ca-nonical pathways “interferon signaling,” “role of JAK1, JAK2and TYK2 in IFN signaling,” and “role of pattern recogni-tion receptors in recognition of bacteria and virus.”

TYK2, a key regulator of IFN signaling (14), was listedin three out of the four top canonical pathways identified,namely, “interferon signaling,” “role of JAK1, JAK2 andTYK2 in IFN signaling,” and “Tec kinase signaling.” Fur-thermore, and in line with a previous bioinformatics pre-diction (15), BLCLs obtained from patients with genotypeAA of the TYK2 SNP rs2304256 (the variant protectiveagainst type 1 diabetes) showed a trend for less markedIFNa-induced STAT1 phosphorylation, as compared withpatients with genotype CC (Fig. 1). Thus, whereas AApatients increased by 3.5-fold STAT1 phosphorylation ascompared with basal levels, CC patients showed a 5.7-foldincrease.

TYK2 Knockdown Protects Human b-Cells AgainstPIC-Induced ApoptosisPIC, but not IFNa, induced a mild increase in TYK2 ex-pression in human islet cells. On the other hand, onlyIFNa increased TYK2 expression in human EndoC-bH1cells (Supplementary Fig. 2A and B). Culture of dispersedhuman islets and EndoC-bH1 under different glucoseconcentrations (6 or 28 mmol/L) did not change TYK2expression (Supplementary Fig. 2C and D). We nextattempted to study the putative role of TYK2 in humanb-cells by using specific siRNAs to inhibit TYK2 expression(knockdown [KD]) but faced major technical problems to

3810 TYK2 Regulates Pancreatic b-Cell Apoptosis Diabetes Volume 64, November 2015

obtain an adequate (i.e., .50%) inhibition of the protein.To evaluate if this could be explained by a prolonged sta-bility of the TYK2 protein, we treated INS-1E cells withcycloheximide and followed TYK2 and a-tubulin expres-sion over a 24-h period. Both proteins had a similar andrather long half-life (around 20 h) (Supplementary Fig. 3Aand B), explaining the observed difficulties in inhibitingTYK2 by a single period of exposure to the siRNA. Toovercome this problem, we designed a two-step transfec-tion protocol (see RESEARCH DESIGN AND METHODS) with thesiRNAs targeting TYK2. By this approach, TYK2 mRNAand protein expression were inhibited by .50% (Fig. 2A,B, E, and F).

Exposure to intracellular PIC increased apoptosis ofdispersed human islet cells and EndoC-bH1 cells (Fig. 2Cand D), and this was partially prevented by TYK2 KD.Whereas basal apoptosis was not modified by TYK2 KD,PIC-induced apoptosis was reduced by 30% (Fig. 2C andD). The protective effects of TYK2 KD against PIC-induced apoptosis were confirmed by a lower expressionof cleaved caspase 3 (Fig. 2E and F). We confirmed theidentity of the apoptotic cells in dispersed human islets byimmunofluorescence using cleaved caspase 3 as an apo-ptotic marker. Thus, PIC treatment led to an increase of7.0% in apoptosis of insulin-positive cells, whereas TYKinhibition prevented by 37% this increase in cell death(Supplementary Fig. 4).

PIC-Induced Activation of the JAK-STAT Pathway IsAbolished by TYK2 KD in Human Islet CellsTYK2 is a member of the JAK family of kinases thattransduce type I IFN signals by phosphorylating STATproteins, such as STAT1 and STAT2 (14). PIC treatmentinduced a two- and fourfold increase in phospho-STAT2and phospho-STAT1, respectively, and a nonsignificanttrend toward higher expression of total STAT1 andSTAT2 in dispersed human islet cells and EndoC-bH1 cells(Fig. 3 and Supplementary Fig. 5). The PIC-stimulatedSTAT1 and STAT2 phosphorylation was abrogated byTYK2 KD (Fig. 3A–D and Supplementary Fig. 5). Nosignificant changes were observed in total STAT1 andSTAT2 expression after TYK2 KD in dispersed human

Figure 1—TYK2 SNP rs2304256 genotype shows lower IFNa-induced STAT1 activation. BLCLs with the AA (n = 5) or the CC(n = 7) genotype were left untreated or stimulated with 1,000 units/mLhuman recombinant IFNa (PeproTech, Rocky Hill, NJ) for 30 min.Expression of phospho-STAT1 (P-STAT1), total STAT1, anda-tubulin (used as loading control) was measured by Western blot.A: The figure shows two representative BLCLs from each genotype.B: The densitometry results for P-STAT1 are represented as a boxplot indicating lower quartile, median, and higher quartile, with whiskersrepresenting the range of the remaining data points. *P < 0.05and **P < 0.01 vs. untreated (i.e., not treated with IFNa); ANOVA. Figure 2—TYK2 inhibition prevents PIC-induced apoptosis in hu-

man b-cells. Dispersed human islets (A, C, and E) or EndoC-bΗ1cells (B, D, and F ) were transfected with siCTRL or with two in-dependent siRNAs targeting TYK2 (TYK2#1 and #2) in a two-roundtransfection protocol and left to recover during 24 h. After this re-covery period, cells were left untreated or treated with intracellularPIC (1 mg/mL) for 24 h. TYK2 mRNA expression was assayed byRT-PCR and normalized by the housekeeping gene b-actin (A andB). Apoptosis (C–F ) was evaluated using HO and PI staining (C andD) and expression of cleaved caspase 3 (Casp 3) (E and F ). Proteinexpression of cleaved caspase 3, TYK2 (for KD confirmation), anda-tubulin (used as loading control) wasmeasured in dispersed humanislets (E) or EndoC-bΗ1 cells (F) by Western blot. Panels E and F arerepresentative of three independent experiments. For A–D, results aremeans 6 SEM of four to six independent experiments. *P < 0.05,**P < 0.01, and ***P < 0.001 vs. untreated (i.e., not treated with PIC)and transfected with the same siRNA; #P < 0.05, ##P < 0.01, and###P < 0.001, as indicated by bars; ANOVA.

diabetes.diabetesjournals.org Marroqui and Associates 3811

islets (Fig. 3E and F). Similar results were observed forSTAT1 phosphorylation in EndoC-bH1 cells treatedwith IFNa as a direct stimulus (Fig. 4A and B). Ofnote, IFNa treatment did not induce b-cell apoptosis(Fig. 4C).

TYK2 KD Inhibits PIC-Induced Type I IFNs and CXCL10mRNA Expression and Release by Human Islet CellsWe have previously shown that intracellular PIC upregu-lates type I IFNs and their downstream genes (IFN-stimulated genes [ISGs]) in rodent b-cells via NF-kB,

STATs, and IFN regulatory factor 3 (IRF3) activation(27,28). To evaluate the role of TYK2 in PIC-stimulatedupregulation of type I IFNs and chemokines in humanb-cells, we measured mRNA expression and release intothe medium of IFNa, IFNb, and CXCL10 by dispersedhuman islet cells (Fig. 5A–F) and EndoC-bH1 cells (Fig.5G–L). Both mRNA expression and release of IFNa, IFNb,and CXCL10 were increased in human b-cells after PICtreatment, and this was partially prevented by TYK2KD in the case of IFNa and CXCL10, but not IFNb(Fig. 5).

Figure 3—Inhibition of TYK2 decreases PIC-induced activation of the type I IFN pathway. Dispersed human islets (A) or EndoC-bΗ1 cells(B) were transfected with siCTRL or with an siRNA targeting human TYK2 as in Fig. 2. Cells were then left untreated or treated withintracellular PIC (1 mg/mL) for 24 h. Expression of phospho-STAT1 (P-STAT1), total STAT1, P-STAT2, total STAT2, TYK2 (for KD confir-mation), and a-tubulin (used as loading control) was measured by Western blot. The figures show representative Western blots of threeexperiments in dispersed human islets (A) or EndoC-bΗ1 cells (B). The densitometry results for P-STAT1 (C ), total STAT1 (E), P-STAT2 (D),and total STAT2 (F ) in dispersed human islets are means6 SEM of three independent experiments. *P< 0.05 vs. untreated (i.e., not treatedwith PIC) and transfected with the same siRNA; #P < 0.05, as indicated by bars; ANOVA.

3812 TYK2 Regulates Pancreatic b-Cell Apoptosis Diabetes Volume 64, November 2015

TYK2 KD Prevents PIC- and IFNa-Induced MHC ClassI Protein ExpressionThe mechanisms by which islet cells overexpress MHCclass I in early type 1 diabetes remains to be clarified,but it has been hypothesized that this is secondary topersistent viral infection and consequent type I IFNproduction (2,29). Since we observed that TYK2 regu-lates IFNa production and signaling in human b-cells,we next examined whether this type 1 diabetes candi-date gene plays a role for the upregulation of MHCclass I molecules in b-cells. In dispersed human isletcells, PIC and IFNa promoted a twofold increase inHLA A, B, and C (HLA-ABC) mRNA expression at 24and 48 h, respectively (Fig. 6). Interestingly, TYK2 KD(Fig. 6A and B) prevented both PIC- and IFNa-inducedHLA-ABC expression in these cells (Fig. 6C and D). PICand IFNa also stimulated an increase in HLA-ABC mRNAexpression in EndoC-bH1 cells, and this was prevented byTYK2 KD (Fig. 6E and F). These observations were con-firmed at the protein level by two different approaches,namely, flow cytometry (Fig. 6G) and immunofluorescence(Fig. 6H). Thus, 30 and 70% of the cells were positive forMHC class I protein upon treatment with PIC and IFNa,respectively, and this was decreased by ;50% in the caseof TYK2 KD (Fig. 6G). Similar results were observed byimmunofluorescence, where TYK2 KD clearly reduced thePIC- and IFNa-induced MHC class I protein expression inhuman EndoC-bH1 cells (Fig. 6H).

To confirm the participation of TYK2 and IFN-drivengene networks as regulators of MHC class I proteinexpression in pancreatic b-cells, we silenced USP18, anIFN-stimulated gene 15–specific protease, in dispersedhuman islets. We have previously shown that USP18 in-hibition induces inflammation and apoptosis by increas-ing IFN-induced STAT1/2 signaling in b-cells (22). Wehave now observed that USP18-silenced cells presentedincreased STAT1 phosphorylation and higher levels ofHLA-ABC mRNA expression upon IFNa treatment (Sup-plementary Fig. 6). IRF1, a transcription factor involvedin type I IFN–stimulated gene expression, contributes toIFN-stimulated expression of the immunoproteasome,which promotes antigen processing for presentation byMHC class I molecules (30). As observed in Supplemen-tary Fig. 7, PIC-induced IRF1 expression was 50 and 30%lower after TYK2 inhibition (.50% TYK inhibition, as inFig. 2) in EndoC-bH1 cells and dispersed human islets,respectively.

DISCUSSION

Pathway analysis of type 1 diabetes candidate genesexpressed in human pancreatic islets identified threetop canonical pathways, namely, “interferon signaling,”“role of JAK1, JAK2 and TYK2 in IFN signaling,” and“role of pattern recognition receptors in recognition ofbacteria and virus.” This suggests that type 1 diabetes can-didate genes play a role in regulating b-cell “self-defense”

Figure 4—Inhibition of TYK2 decreases IFNa-induced activation of the type I IFN pathway. EndoC-bΗ1 cells (A–C) were transfected withsiCTRL or with siRNA targeting human TYK2 as in Fig. 2. Cells were then left untreated or treated with IFNa (2,000 units/mL) for 24 h.A: TYK2 mRNA expression was assayed by RT-PCR and normalized by the housekeeping gene b-actin. B: Expression of phospho-STAT1(P-STAT1), total STAT1, TYK2 (for KD confirmation), and a-tubulin (used as loading control) was measured by Western blot. The figureshows a representative Western blot of three experiments in EndoC-bH1 cells. C: Apoptosis after TYK2 KD was evaluated using HOand PI staining. Results are means 6 SEM of three independent experiments. *P < 0.05 vs. untreated (i.e., not treated with IFNa) andtransfected with the same siRNA; ##P < 0.01, as indicated by bars; ANOVA.

diabetes.diabetesjournals.org Marroqui and Associates 3813

or “cell autonomous immune responses” against infection.These self-defense mechanisms are present in many nonim-mune cell types and are upregulated upon virus infection(31,32). In vertebrates, cellular self-defense synergizeswith innate and adaptive immunity to combat infections(32). The cell susceptibility/resistance of differentiatedand poorly proliferating cells, such as pancreatic b-cellsand neurons, to viral infection is an important determi-nant of clinical outcome. For instance, higher basal ex-pression and faster upregulation of IFN-induced genesimprove the survival of neurons infected by West Nilevirus (33), suggesting that candidate genes that affect

these pathways may have a major impact on the outcomeof viral infections (or exposure to other “danger signals”)in b-cells. This may determine the amplitude of the localinflammation, the degree of b-cell loss, and an eventualprogression to full autoimmunity and type 1 diabetes(2,6,29,34).

The candidate gene TYK2 may play a key role in theseself-defense pathways. TYK2 was identified in the path-way analysis described above, and it plays a critical role inthe intracellular signaling of type I IFNs via phosphoryla-tion and activation of STAT proteins (14). Of note, type IIFNs play an important role in several autoimmune

Figure 5—TYK2 KD decreases PIC-induced IFNa and CXCL10 mRNA expression and release to the medium after PIC treatment.Dispersed human islets (A–F) or EndoC-bΗ1 cells (G–L) were transfected with siCTRL or with an siRNA targeting human TYK2 as inFig. 2. The cells were then left untreated or treated with intracellular PIC (1 mg/mL) for 24 h. Expression of IFNa (A and G), IFNb (B and H),and CXCL10 (C and I) mRNAs were analyzed by RT-PCR and normalized by the housekeeping gene b-actin. Secretion of IFNa (D and J),IFNb (E and K), and CXCL10 (F and L) was measured in the supernatants by ELISA. Results are means 6 SEM of three to six independentexperiments. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. untreated (i.e., not treated with PIC) and transfected with the same siRNA;#P < 0.05, ##P < 0.01, and ###P < 0.001, as indicated by bars; ANOVA.

3814 TYK2 Regulates Pancreatic b-Cell Apoptosis Diabetes Volume 64, November 2015

diseases, including T1D (35), by, among other effects,activating dendritic cells to present self-antigens to po-tentially autoreactive T cells (36). Furthermore, very re-cent data indicate that severely reduced TYK2 expressionin pancreatic b-cells, due to a natural mutation, is respon-sible for susceptibility to virus-induced diabetes in SJLand SWR mice (37).

It has been previously shown that mice on a high-fatdiet present decreased TYK2 expression in brown adiposetissue and skeletal muscle, but not in white adipose tissueand liver, suggesting that modulation of TYK2 expressionby diet is tissue specific (38). We did not observe, however,

changes in TYK2 expression after exposure of dis-persed human islets and EndoC-bH1 human b-cellsto high glucose. These findings are in agreement withprevious RNA-seq data where TYK2 was found simi-larly expressed in human pancreatic islets from indi-viduals with different degrees of glucose tolerance(donors with normoglycemia, impaired glucose toler-ance, and type 2 diabetes) (39) or in human isletsexposed or not to palmitate (40).

One of the polymorphisms in the TYK2 gene,rs2304256, predicted to decrease its function (a predic-tion presently confirmed experimentally), is associated

Figure 6—Inhibition of TYK2 prevents PIC- and IFNa-induced MHCclass I expression. Dispersed human islets (A–D) or EndoC-bH1cells (E–H) were transfected with siCTRL or with an siRNA targetinghuman TYK2 as in Fig. 2. After this, dispersed human islets were leftuntreated or treated with intracellular PIC (1 mg/mL) for 24 h (A andC ) or IFNa (2,000 units/mL) for 24 or 48 h (B and D). EndoC-bH1cells were left untreated or treated with intracellular PIC (1 mg/mL) orIFNa (2,000 units/mL) for 24 h (E–H). mRNA expression of TYK2 (A,B, and E) and HLA-ABC (C, D, and F ) was analyzed by RT-PCR andnormalized by the housekeeping gene b-actin. G: MHC class I pro-tein levels were measured by FACS in EndoC-bH1 cells. H: ICC ofMHC class I (red) and Hoechst (blue) were performed to confirmMHC class I expression in EndoC-bH1 cells. Results are means 6SEM of four to six independent experiments. ***P < 0.001 vs. un-treated (i.e., not treated with PIC or IFNa) and transfected with thesame siRNA; #P < 0.05, ##P < 0.01, and ###P < 0.001, as in-dicated by bars; ANOVA. H: Images are representative of fourindependent experiments.

Figure 7—Proposed model for the role of the candidate gene TYK2in b-cells. Upon a viral infection, an increase in intracellular dsRNAmolecules (or other “danger signals”) is sensed by pattern recogni-tion receptors (PRRs), such as MDA5 and RIG-I (step 1). This leadsto the activation of the early antiviral response through the produc-tion and release of type I IFNs (IFNa/b) as well as an increase in thechemokine CXCL10 (step 2). While CXCL10 attracts monocytes,T lymphocytes, and natural killer cells, IFNs can exert both autocrineand paracrine effects. After binding to the IFNa receptor (IFNaR),the tyrosine kinases JAK1 and TYK2 are activated, leading to therecruitment and phosphorylation of STAT1 and STAT2 (step 3). Anegative feedback promoted by USP18 controls STAT1/2 activa-tion. STAT heterodimers translocate to the nucleus, where they bindto the IFN-stimulated response element (ISRE) and stimulate theexpression of ISGs (step 4). Among the genes induced by thissignaling pathway, there are several antiviral proteins (e.g., ISG15,Mx1, and PKR) as well as MHC class I molecules. In addition to thehyperexpression of MHC class I proteins, type I IFNs stimulateexpression of the immunoproteasome, a version of the proteosomespecialized in the production of immunogenic peptides for presen-tation by MHC class I molecules, in an IRF1-dependent manner(step 5). MHC I hyperexpression, along with higher antigen process-ing by the immunoproteasome, increases the efficiency of presen-tation of putatively modified b-cell antigens to the immune cells.Taken together, this suggests that a viral-induced increase in IFNresponse may induce an excessive inflammatory response in ge-netically predisposed individuals, leading to autoimmunity and pro-gressive destruction of pancreatic b-cells. In this context, thecandidate gene TYK2 plays a crucial role through the direct phos-phorylation and activation of STATs in response to type I IFNs.Supporting references are provided in DISCUSSION.

diabetes.diabetesjournals.org Marroqui and Associates 3815

with protection against type 1 diabetes (15,16). This is inline with our present observations, indicating that a 50%inhibition in TYK2 activity by specific siRNAs decreasesdsRNA-induced apoptosis and proinflammatory pathwaysin human pancreatic b-cells. This protection is mediatedvia inhibition of IFNa signaling, as indicated by decreasedphosphorylation of STAT1 and STAT2. Of particular rel-evance, inhibition of TYK2 prevents dsRNA- or IFNa-induced upregulation of MHC class I protein expressionand expression of the chemokine CXCL10 in humanb-cells.

MHC class I upregulation is one of the most consistentfindings in islets from patients with type 1 diabetes(41,42). This may be one of the mechanisms by whichb-cells become “visible” to the immune system in earlytype 1 diabetes (2). This MHC class I hyperexpression isobserved both in inflamed and apparently noninflamedinsulin-containing islets, suggesting that this upregu-lation precedes insulitis during diabetes development(42). The present data provide the first evidence thatthis phenomenon is regulated by a type 1 diabetes can-didate gene, namely, TYK2. In line with these findings onthe role of TYK2 in MHC class I regulation, granulocyte-macrophage colony-stimulating factor–induced downre-gulation of TYK2 and JAK1 tyrosine phosphorylation, aswell as TYK2 protein expression, contributes to decreasedSTAT1 phosphorylation and subsequently diminishedMHC class I antigen levels in hematopoietic cell lines(43). Furthermore, IFNb-stimulated MHC class I expres-sion is abrogated when JAK1, TYK2, or the IFNa/b re-ceptor is suppressed by siRNAs in human Ewing sarcomacell line (44).

The chemokine CXCL10 has been previously shown tobe upregulated in human and rodent islets exposed invitro to the proinflammatory cytokines IL-1b, TNFa,IL-17, and IFNg (3,4). Importantly, CXCL10 expressionis also upregulated in islets from patients with type 1diabetes (5,45) and in NOD mice (3,46), and its neutral-ization prevents diabetes in NOD mice (47).

As discussed above, TYK2 may play an important rolein the pathogenesis of diverse autoimmune diseases, andnovel therapeutic strategies based on specific TYK2inhibitors, such as Cmpd1, are being evaluated (48,49).The present data suggest that TYK2 inhibition could betested as a novel therapeutic approach to prevent type 1diabetes development.

In conclusion, we provide evidence that the candidategene TYK2 plays a key role in the activation of cell-autonomous immune responses through the activation ofSTATs and consequent triggering of the IFN response,that may lead to hyperexpression of MHC class I proteinsin human pancreatic b-cells (Fig. 7). Polymorphisms thatdecrease function of TYK2 (present data) and of MDA5(encoded by IFIH1 and functioning as a detector of viralinfection in b-cells and other cell types [11,50]) decreasethe risk of type 1 diabetes (51,52). This suggests thata genetically determined excessive inflammatory response

to viral infections may contribute to autoimmunity andeventual diabetes in susceptible individuals.

Acknowledgments. The authors are grateful to M. Pangerl, A.M.Musuaya, N. Pachera, and I. Millard for excellent technical support,Drs. M. Cnop and M. Igoillo-Esteve for providing data on human samples, andDrs. J. Juan, O. Villate, and J.-V. Turatsinze for experimental support and RNA-seq data analysis.Funding. This work was supported by grants from Fonds National de laRecherche Scientifique (FNRS), Belgium, Communauté Française de Belgique-Actions de Recherche Concertées (ARC), and the European Union (projectsNaimit and BetaBat, in the Framework Programme 7 of the European Community).L.M. is supported by an FNRS postdoctoral fellowship. R.S.D.S. is the recipient ofa postdoctoral fellowship from Conselho Nacional de Desenvolvimento Cientificoe tecnológico (CNPq), Brazil. T.F. and F.P. were supported by a grant from theEuropean Foundation for the Study of Diabetes (EFSD/JDRF/NN). I.S. was the re-cipient of a postdoctoral fellowship from the Education Department of the BasqueCountry.Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. L.Marr. and R.S.D.S. contributed to the originalidea and the design of the experiments; researched data; contributed to discus-sion; and wrote, revised, and edited the manuscript. T.F., F.A.G., A.O.d.b.,L.Mars., P.M., and F.P. researched data and revised and edited the manuscript.I.S. researched data, contributed to discussion, and revised and edited themanuscript. D.L.E. contributed to the original idea and the design and interpre-tation of the experiments; contributed to discussion; and wrote, revised, andedited the manuscript. L.Marr. is the guarantor of this work and, as such, hadfull access to all the data in the study and takes responsibility for the integrity ofthe data and the accuracy of the data analysis.

References1. Eizirik DL, Colli ML, Ortis F. The role of inflammation in insulitis and b-cellloss in type 1 diabetes. Nat Rev Endocrinol 2009;5:219–2262. Richardson SJ, Morgan NG, Foulis AK. Pancreatic pathology in type 1diabetes mellitus. Endocr Pathol 2014;25:80–923. Cardozo AK, Proost P, Gysemans C, Chen MC, Mathieu C, Eizirik DL. IL-1band IFN-g induce the expression of diverse chemokines and IL-15 in human andrat pancreatic islet cells, and in islets from pre-diabetic NOD mice. Diabetologia2003;46:255–2664. Grieco FA, Moore F, Vigneron F, et al. IL-17A increases the expression ofproinflammatory chemokines in human pancreatic islets. Diabetologia 2014;57:502–5115. Roep BO, Kleijwegt FS, van Halteren AG, et al. Islet inflammation andCXCL10 in recent-onset type 1 diabetes. Clin Exp Immunol 2010;159:338–3436. Santin I, Eizirik DL. Candidate genes for type 1 diabetes modulate pancreatic isletinflammation and b-cell apoptosis. Diabetes Obes Metab 2013;15(Suppl. 3):71–817. Concannon P, Rich SS, Nepom GT. Genetics of type 1A diabetes. N Engl JMed 2009;360:1646–16548. Bergholdt R, Brorsson C, Palleja A, et al. Identification of novel type 1diabetes candidate genes by integrating genome-wide association data, protein-protein interactions, and human pancreatic islet gene expression. Diabetes 2012;61:954–9629. Eizirik DL, Sammeth M, Bouckenooghe T, et al. The human pancreatic islettranscriptome: expression of candidate genes for type 1 diabetes and the impactof pro-inflammatory cytokines. PLoS Genet 2012;8:e100255210. Farh KK, Marson A, Zhu J, et al. Genetic and epigenetic fine mapping ofcausal autoimmune disease variants. Nature 2015;518:337–34311. Colli ML, Moore F, Gurzov EN, Ortis F, Eizirik DL. MDA5 and PTPN2, twocandidate genes for type 1 diabetes, modify pancreatic b-cell responses to theviral by-product double-stranded RNA. Hum Mol Genet 2010;19:135–146

3816 TYK2 Regulates Pancreatic b-Cell Apoptosis Diabetes Volume 64, November 2015

12. Fløyel T, Brorsson C, Nielsen LB, et al. CTSH regulates b-cell function anddisease progression in newly diagnosed type 1 diabetes patients. Proc Natl AcadSci U S A 2014;111:10305–1031013. Marroquí L, Santin I, Dos Santos RS, Marselli L, Marchetti P, Eizirik DL.BACH2, a candidate risk gene for type 1 diabetes, regulates apoptosis in pan-creatic b-cells via JNK1 modulation and crosstalk with the candidate genePTPN2. Diabetes 2014;63:2516–252714. Babon JJ, Lucet IS, Murphy JM, Nicola NA, Varghese LN. The molecularregulation of Janus kinase (JAK) activation. Biochem J 2014;462:1–1315. Tao JH, Zou YF, Feng XL, et al. Meta-analysis of TYK2 gene polymorphismsassociation with susceptibility to autoimmune and inflammatory diseases. MolBiol Rep 2011;38:4663–467216. Wallace C, Smyth DJ, Maisuria-Armer M, Walker NM, Todd JA, Clayton DG.The imprinted DLK1-MEG3 gene region on chromosome 14q32.2 alters sus-ceptibility to type 1 diabetes. Nat Genet 2010;42:68–7117. Richter MF, Duménil G, Uzé G, Fellous M, Pellegrini S. Specific contributionof Tyk2 JH regions to the binding and the expression of the interferon a/b re-ceptor component IFNAR1. J Biol Chem 1998;273:24723–2472918. Ragimbeau J, Dondi E, Alcover A, Eid P, Uzé G, Pellegrini S. The tyrosinekinase Tyk2 controls IFNAR1 cell surface expression. EMBO J 2003;22:537–54719. Marchetti P, Bugliani M, Lupi R, et al. The endoplasmic reticulum in pan-creatic b cells of type 2 diabetes patients. Diabetologia 2007;50:2486–249420. Ravassard P, Hazhouz Y, Pechberty S, et al. A genetically engineered humanpancreatic b cell line exhibiting glucose-inducible insulin secretion. J Clin Invest2011;121:3589–359721. Ortis F, Cardozo AK, Crispim D, Störling J, Mandrup-Poulsen T, Eizirik DL.Cytokine-induced proapoptotic gene expression in insulin-producing cells is re-lated to rapid, sustained, and nonoscillatory nuclear factor-kappaB activation. MolEndocrinol 2006;20:1867–187922. Santin I, Moore F, Grieco FA, Marchetti P, Brancolini C, Eizirik DL. USP18 isa key regulator of the interferon-driven gene network modulating pancreaticb cell inflammation and apoptosis. Cell Death Dis 2012;3:e41923. Moore F, Cunha DA, Mulder H, Eizirik DL. Use of RNA interference to in-vestigate cytokine signal transduction in pancreatic b cells. Methods Mol Biol2012;820:179–19424. Liu D, Darville M, Eizirik DL. Double-stranded ribonucleic acid (RNA) inducesb-cell Fas messenger RNA expression and increases cytokine-induced b-cellapoptosis. Endocrinology 2001;142:2593–259925. Overbergh L, Valckx D, Waer M, Mathieu C. Quantification of murine cy-tokine mRNAs using real time quantitative reverse transcriptase PCR. Cytokine1999;11:305–31226. Gurzov EN, Germano CM, Cunha DA, et al. p53 up-regulated modulator ofapoptosis (PUMA) activation contributes to pancreatic b-cell apoptosis induced byproinflammatory cytokines and endoplasmic reticulum stress. J Biol Chem 2010;285:19910–1992027. Dogusan Z, García M, Flamez D, et al. Double-stranded RNA inducespancreatic b-cell apoptosis by activation of the toll-like receptor 3 and interferonregulatory factor 3 pathways. Diabetes 2008;57:1236–124528. Rasschaert J, Liu D, Kutlu B, et al. Global profiling of double stranded RNA- andIFN-g-induced genes in rat pancreatic b cells. Diabetologia 2003;46:1641–165729. Morgan NG, Richardson SJ. Enteroviruses as causative agents in type 1diabetes: loose ends or lost cause? Trends Endocrinol Metab 2014;25:611–61930. Freudenburg W, Gautam M, Chakraborty P, et al. Immunoproteasome ac-tivation during early antiviral response in mouse pancreatic b-cells: new insightsinto auto-antigen generation in type I diabetes? J Clin Cell Immunol 2013;4:14131. Randow F, MacMicking JD, James LC. Cellular self-defense: how cell-autonomous immunity protects against pathogens. Science 2013;340:701–706

32. Yan N, Chen ZJ. Intrinsic antiviral immunity. Nat Immunol 2012;13:214–22233. Cho H, Proll SC, Szretter KJ, Katze MG, Gale M Jr, Diamond MS. Differentialinnate immune response programs in neuronal subtypes determine susceptibilityto infection in the brain by positive-stranded RNA viruses. Nat Med 2013;19:458–46434. Dotta F, Censini S, van Halteren AG, et al. Coxsackie B4 virus infection ofb cells and natural killer cell insulitis in recent-onset type 1 diabetic patients.Proc Natl Acad Sci U S A 2007;104:5115–512035. González-Navajas JM, Lee J, David M, Raz E. Immunomodulatory functionsof type I interferons. Nat Rev Immunol 2012;12:125–13536. Guerder S, Joncker N, Mahiddine K, Serre L. Dendritic cells in tolerance andautoimmune diabetes. Curr Opin Immunol 2013;25:670–67537. Izumi K, Mine K, Inoue Y, et al. Reduced Tyk2 gene expression in b-cellsdue to natural mutation determines susceptibility to virus-induced diabetes. NatCommun 2015;6:674838. Derecka M, Gornicka A, Koralov SB, et al. Tyk2 and Stat3 regulate brownadipose tissue differentiation and obesity. Cell Metab 2012;16:814–82439. Fadista J, Vikman P, Laakso EO, et al. Global genomic and transcriptomicanalysis of human pancreatic islets reveals novel genes influencing glucosemetabolism. Proc Natl Acad Sci U S A 2014;111:13924–1392940. Cnop M, Abdulkarim B, Bottu G, et al. RNA sequencing identifies dysre-gulation of the human pancreatic islet transcriptome by the saturated fatty acidpalmitate. Diabetes 2014;63:1978–199341. Hanafusa T, Miyazaki A, Miyagawa J, et al. Examination of islets in thepancreas biopsy specimens from newly diagnosed type 1 (insulin-dependent)diabetic patients. Diabetologia 1990;33:105–11142. Foulis AK, Farquharson MA, Hardman R. Aberrant expression of class IImajor histocompatibility complex molecules by B cells and hyperexpression ofclass I major histocompatibility complex molecules by insulin containing islets intype 1 (insulin-dependent) diabetes mellitus. Diabetologia 1987;30:333–34343. Kasper S, Kindler T, Sonnenschein S, et al. Cross-inhibition of interferon-induced signals by GM-CSF through a block in Stat1 activation. J InterferonCytokine Res 2007;27:947–95944. Shin-Ya M, Hirai H, Satoh E, et al. Intracellular interferon triggers Jak/Statsignaling cascade and induces p53-dependent antiviral protection. Biochem Bio-phys Res Commun 2005;329:1139–114645. Sarkar SA, Lee CE, Victorino F, et al. Expression and regulation of che-mokines in murine and human type 1 diabetes. Diabetes 2012;61:436–44646. Welzen-Coppens JM, van Helden-Meeuwsen CG, Leenen PJ, Drexhage HA,Versnel MA. The kinetics of plasmacytoid dendritic cell accumulation in thepancreas of the NOD mouse during the early phases of insulitis. PLoS One 2013;8:e5507147. Morimoto J, Yoneyama H, Shimada A, et al. CXC chemokine ligand 10neutralization suppresses the occurrence of diabetes in nonobese diabetic micethrough enhanced b cell proliferation without affecting insulitis. J Immunol 2004;173:7017–702448. Liang Y, Zhu Y, Xia Y, et al. Therapeutic potential of tyrosine kinase 2 inautoimmunity. Expert Opin Ther Targets 2014;18:571–58049. Menet CJ. Toward selective TYK2 inhibitors as therapeutic agents for thetreatment of inflammatory diseases. Pharm Pat Anal 2014;3:449–46650. Reikine S, Nguyen JB, Modis Y. Pattern recognition and signaling mecha-nisms of RIG-I and MDA5. Front Immunol 2014;5:34251. Downes K, Pekalski M, Angus KL, et al. Reduced expression of IFIH1 isprotective for type 1 diabetes. PLoS One. 9 September 2010 [Epub ahead ofprint]. DOI: 10.1371/journal.pone.001264652. Lincez PJ, Shanina I, Horwitz MS. Reduced expression of the MDA5 geneIFIH1 prevents autoimmune diabetes. Diabetes 2015;64:2184–2193

diabetes.diabetesjournals.org Marroqui and Associates 3817