Long-Term IKK2/NF-Kb Signaling in Pancreatic B-Cells Induces Immune-Mediated Diabetes

960 Diabetes Volume 63, March 2014

Heba H. Salem,1,2 Bernadette Trojanowski,1 Katja Fiedler,1 Harald J. Maier,1 Reinhold Schirmbeck,3 Martin Wagner,3 Bernhard O. Boehm,4,5 Thomas Wirth,1 and Bernd Baumann1

Long-Term IKK2/NF-kB Signaling in Pancreatic b-Cells Induces Immune-Mediated Diabetes

Type 1 diabetes is a multifactorial inflammatory virus-induced diabetes and allergic inflammatory disease in genetically susceptible individuals responses as well as increased major characterized by progressive autoimmune histocompatibility complex class I/II expression destruction of pancreatic b-cells initiated by yet by b-cells that could collectively promote unknown factors. Although animal models of type 1 diabetes development. Potential novel diabetes diabetes have substantially increased our candidate genes were also identified. Interestingly, understanding of disease pathogenesis, animals successfully recovered from diabetes heterogeneity seen in human patients cannot be upon transgene inactivation. Our data give the reflected by a single model and calls for additional first direct evidence that b-cell–specific models covering different aspects of human IKK2/NF-kB activation is a potential trigger of SIGNAL TRANSDUCTION pathophysiology. Inhibitor of kBkinase immune-mediated diabetes. Moreover, IKK2- (IKK)/nuclear factor-kB(NF-kB) signaling is CAPdx-1 mice provide a novel tool for studying a master regulator of inflammation; however, its critical checkpoints in diabetes pathogenesis role in diabetes pathogenesis is controversially and mechanisms governing b-cell degeneration/ discussed by studies using different inhibition regeneration. approaches. To investigate the potential Diabetes 2014;63:960–975 | DOI: 10.2337/db13-1037 diabetogenic effects of NF-kBinb-cells, we generated a gain-of-function model allowing conditional IKK2/NF-kB activation in b-cells. A Type 1 diabetes is an autoimmune disease in genetically transgenic mouse model that expresses predisposed individuals and presumably triggered and/or a constitutively active mutant of human IKK2 accelerated by environmental factors (1). Analyses of dependent on Pdx-1 promoter activity (IKK2-CAPdx-1) animal models of type 1 diabetes have greatly improved spontaneously develops full-blown immune- our knowledge about disease pathophysiology and ge- mediated diabetes with insulitis, hyperglycemia, netic contribution. However, there is still an unmet need and hypoinsulinemia. Disease development to understand islet cell pathology and ongoing in- involves a gene expression program mimicking flammatory processes within the islets of Langerhans.

1Institute of Physiological Chemistry, Ulm University, Ulm, Germany Corresponding author: Bernd Baumann, [email protected]. 2 Department of Biochemistry, Faculty of Pharmacy, Cairo University, Cairo, Egypt Received 2 July 2013 and accepted 27 November 2013. 3Center for Internal Medicine, Ulm University Medical Center, Ulm University, This article contains Supplementary Data online at http://diabetes Ulm, Germany .diabetesjournals.org/lookup/suppl/doi:10.2337/db13-1037/-/DC1. 4Division of Endocrinology, Diabetes and Metabolism, Ulm University Medical Centre, Ulm University, Ulm, Germany © 2014 by the American Diabetes Association. See http://creativecommons 5Lee Kong Chian School of Medicine, Nanyang Technological University, Imperial .org/licenses/by-nc-nd/3.0/ for details. College London, Singapore diabetes.diabetesjournals.org Salem and Associates 961

In type 1 diabetes, inflammation contributes to the The supernatant was collected, and the pellet was re- induction and amplification of the immune insult extracted. The pooled supernatant was used for insulin against b-cells and, at later stages, to the stabilization measurement. Insulin was determined in plasma samples and maintenance of the insulitic process, thus pro- containing protease inhibitor and pancreatic extracts moting disease development and progression (2). using the Ultra-Sensitive Mouse Insulin ELISA Kit Furthermore, b-cell response to stress and in- (Chrystal Chem Inc.) following the manufacturer’s flammation is thought to be a critical determinant in instructions. disease outcome (2). Canonical inhibitor of kB(IkB) kinase 2 (IKK2)/nuclear factor-kB(NF-kB) signaling is Protein Extraction, Western Immunoblotting, Luciferase Assay, and Electrophoretic Mobility Shift the master regulator of inflammatory responses and Assay innate immunity (3), and it is activated by viral and bacterial pathogens, various cytokines, and general Pancreata were snap-frozen in liquid nitrogen and pul- stress factors. Apart from its antiapoptotic function in verized, and proteins were extracted for Western im- most cell types, the predominant role of NF-kBacti- munoblotting and luciferase activity measurement (7). vation in b-cells appears to be of a death-promoting For immunoblotting, antibodies against IKK2 (Abcam) – nature (4,5). and extracellular signal related kinase-2 (Santa Cruz The transcription factor NF-kBexistsashomo-and Biotechnology) were used. Electrophoretic mobility shift heterodimers of five different subunits. In resting cells, assay was performed as described (7) using whole-cell NF-kB dimers are sequestered in the cytoplasm by IkB extract from isolated islets. proteins. A key step in connecting extracellular stimuli Histology and Immunostaining to NF-kB induction is the activation of the IKK complex For paraffin sections, pancreata were excised, fixed composed of the catalytic subunits IKK1 and IKK2 and overnight in 3.8% buffered formalin, dehydrated, paraf- the regulatory component NF-kB essential modulator. fin embedded, cut in 3-mm sections, and further pro- This complex phosphorylates IkB proteins, thereby cessed as previously described (8). For cryosections, initiating their ubiquitination and subsequent protea- 10-mm slices from natively frozen pancreata were fixed somal degradation, thus allowing nuclear translocation with 4% paraformaldehyde. Sections were incubated with of NF-kB(4,5). the primary antibodies: rabbit (Cell Signaling Technol- To mimic an inflammatory insult selectively in pancre- ogy) or guinea pig (Abcam) anti-insulin, goat anti-human atic b-cells, we created a mouse model that allows the IKK2 (Santa Cruz Biotechnology), rabbit anti-RelA/p65 conditional expression of a constitutively active IKK2 allele (Laboratory Vision), anti–Pdx-1 and antichromogranin A in b-cells. Interestingly, prolonged IKK2/NF-kBactivation (Abcam), anti-Ki67 (Thermo Scientific), rat anti-CD4 and in b-cells is sufficient to induce insulitis with marked hy- anti-CD8 (Abcam), anti–CD11c-phycoerythrin and anti- b perglycemia, hypoinsulinemia, and reduced -cell mass in B220 (BD Biosciences), and anti–CD25-phycoerythrin transgenic animals. Intriguingly, upon switching off trans- and anti–major histocompatibility complex (MHC) class gene expression, diabetes could be efficiently reversed. II (MHC II)–FITC (eBioscience). Secondary antibodies were coupled with Alexa Fluor (Invitrogen) for immu- RESEARCH DESIGN AND METHODS nofluorescence and with horseradish peroxidase in Mice immunohistochemistry that was developed by 3-amino- Male mice were housed under specific pathogen-free 9-ethylcarbazole (DakoCytomation). Immunofluorescent conditions at the animal facility of the University of Ulm. stainings were visualized as before (8), and other stain- Pdx-1.tTA mice (C57BL/6) (6) and (tetO)7.IKK2-CA ings were analyzed on a Leica DM IRB microscope (Leica (constitutively active mutant of human IKK2) mice Microsystems) equipped with ProgRes C14 digital camera (NMRI) (7) were described previously. Pdx-1.tTA mice are (Jenoptik) and Openlab software (Improvision). knockin animals in which the coding sequence of tetracycline-dependent transactivator (tTA) has replaced Detection of Apoptosis the endogenous Pdx-1 gene, thereby rendering this mouse In situ detection of DNA strand breaks was performed 2 line heterozygous for Pdx-1 (Pdx-1+/ ). Control mice using the TUNEL labeling method with the FragEL DNA Fragmentation Detection Kit and colorimetric-TdT en- group include wild-type and single-transgenic (tetO)7. IKK2-CA mice unless otherwise stated. Experiments were zyme (EMD Millipore) according to the manufacturer’s performed in accordance with institutional guidelines and instructions. German animal protection law. Islet Isolation Metabolic Studies Pancreata were inflated in situ with 0.5 mg/mL ice-cold Blood glucose was measured using the One-Touch Ultra collagenase XI solution (Sigma-Aldrich, St. Louis, MO) glucometer (LifeScan Inc., Mipitas, CA). Pancreatic dissolved in PBS (with Ca/Mg). Pancreatic tissue was insulin was extracted by overnight agitation with cold dissected out and digested for 19–24 min at 37°C with acid ethanol (0.18 mol/L HCl in 70% ethanol) at 4°C. subsequent washing and centrifugation. Islets in the 962 NF-kB Activation in b-Cells Induces Diabetes Diabetes Volume 63, March 2014

Figure 1—Gain-of-function mouse model for conditional activation of IKK2 in pancreatic b-cells. A: Transgenic approach for doxycycline (Dox)-regulated expression of IKK2-CA in pancreatic b-cells. tTA protein is expressed under the control of Pdx-1 promoter and can bind to a bidirectional promoter [(tetO)7; 7xtetO] driving the transcription of luciferase and IKK2-CA transgene. This binding is inhibited by Dox, thereby shutting off transgene expression. The Pdx-1.tTA mice are knockin animals in which the coding sequence of tTA has replaced the endogenous Pdx-1 gene, thereby rendering this mouse line heterozygous for Pdx-1 (Pdx-1+/2). Animals were bred under Dox (0.1 g/L Dox in drinking water) to avoid developmental defects, which then had been withdrawn after weaning (3 to 4 weeks) to activate transgene expression. B: Luciferase activity was measured in pancreas (Pa), spleen (Sp), liver (Li), kidney (Ki), stomach (St), intestine (Int), lung (Lu), thymus (Thy), thalamus/hypothalamus (Tha), and brain (Br) of 26- to 32-week-old control (white bars) and IKK2-CAPdx-1 mice (black bars) (n =2). diabetes.diabetesjournals.org Salem and Associates 963 sediment were purified with Histopaque 10771 (Sigma- transgenic mice expressing the tTA under the control of +/2 Aldrich), washed, and frozen in liquid N2. Pdx-1 promoter (Pdx-1 ) (6) with mice carrying the A RNA Extraction, Quantitative PCR, and Microarray luciferase-(tetO)7-IKK2-CA minigene (7) (Fig. 1 ). The transgenic expression system was not activated until RNA was extracted with RNeasy kit (Quiagen) and cDNA weaning in order to avoid any effects of IKK2-CA activity synthesis was done using the Transcriptor High Fidelity on pancreas development. Measurement of luciferase cDNA Synthesis Kit (Roche). Quantitative real-time PCR activity, the coexpressed reporter gene, revealed strong was performed with the Roche LightCycler 480 (Roche) Pdx-1 transgene activity in the pancreas of IKK2-CA mice using gene-specific primers and hydrolysis probes after doxycycline withdrawal and only minor activity in designed by the Roche Universal Probe Library system the intestine (Fig. 1B). Pancreatic IKK2-CA expression (Roche). Microarray analysis was performed with the was confirmed by immunoblot and could be switched off Mouse Gene 1.0 ST array (Affymetrix) and evaluated by doxycycline (Fig. 1C), thus allowing conditional regu- with the “Genesifter” sofware (Geospiza). The expression lation of IKK2 activity. Immunofluorescence staining data are available at Gene Expression Omnibus super demonstrated a mosaic expression of IKK2-CA exclu- series (accession number GSE47504). Pdx-1 sively in the islets of IKK2-CA mice (Fig. 1D), cor- Flow Cytometry relating with obvious reduction in insulin Pancreatic cells were isolated and purified as previously immunoreactivity (Fig. 1D). described (9) and subsequently stained with monoclonal Electrophoretic mobility shift assays showed strong Pdx-1 antibodies using standard procedures. The following basal NF-kB activity in IKK2-CA mice relative to +/2 antibodies were purchased from eBioscience: anti–MHC their Pdx-1 littermates, demonstrating transgene II (M5/114.15.2), anti–MHC class I (MHC I; AF6– functionality (Fig. 1E). NF-kB activity was also elevated +/2 88.5.5.3), anti-CD11b (M1/70), anti-CD3e (17A2), anti- in Pdx-1 mice compared with controls. Furthermore, CD25 (PC61.5), and anti-CD69 (H1.2F3); from BD IKK2 activation led to nuclear translocation of the NF-kB Pdx-1 Biosciences: anti-CD11b (M1/70), anti-CD8a (53–6.7), subunit RelA in b-cells of IKK2-CA mice, while al- anti-CD4 (RM4–5), CD44 (IM7), anti-NK1.1 (PK136), most no nuclear RelA was detected in islets of either +/2 and anti-CD11c (HL3). Anti-CD45 (30-F11), anti-CD19 Pdx-1 or control mice (Fig. 1F). (6D5), anti-F4/80 (BM8), and anti-Ly6G (1A8) were from IKK2/NF-kB Activation in b-Cells Induces Full-Blown BioLegend, while Ly6C (1G7.G10) was from Miltenyi Diabetes Biotec. Fixable viability dye (eBioscience) was used to We then assessed the physiological and histological exclude dead cells. FACS was performed on an FACS- consequences of IKK2/NF-kB signaling in pancreatic Canto II (BD Biosciences), and data were analyzed with b-cells. Interestingly, at the age of 11 weeks, some ani- FACSDiva 6.2 (BD Biosciences) and FlowJo softwares mals started to develop hyperglycemia, and at ;24 weeks (Tree Star). Pdx-1 of age, all IKK2-CA mice showed substantial eleva- 2 Statistical Analysis tion in fed (538.5 6 16 vs. 154.9 6 7 mg/dL in Pdx-1+/ Values are given as mean 6 SEM. Statistical analysis was mice) and fasting (423.3 6 34 vs. 96.9 6 4 mg/dL in +/2 performed with the Prism software (GraphPad) using Pdx-1 mice) blood glucose levels, with several animals two-tailed Student t test. Data with P values of #0.05 .600 mg/dL (Fig. 2A and B). Hyperglycemia was ac- fi companied by a 57% reduction of plasma insulin levels were considered statistically signi cant. 2 compared with Pdx-1+/ mice (Fig. 2C), indicating that RESULTS these mice were overtly diabetic. Animals displayed Generation of Conditional, Gain-of-Function Mouse clinical signs of diabetes, including polyuria, polydipsia, Model for Canonical NF-kB Signaling in Pancreatic weight loss, and general sickness, and in severe cases, b-Cells mortality was recorded. Furthermore, immunohisto- To directly explore the biological consequences of ca- logical analyses indicated an extensive loss of insulin- nonical NF-kB activation in pancreatic b-cells, we gen- positive b-cells in IKK2-CAPdx-1 mice, with the remaining erated IKK2-CAPdx-1. This was achieved by crossing b-cells appearing degranulated with only faint residual

High luciferase activity was detected in pancreas of IKK2-CAPdx-1 mice and only minor activity in their intestine. C: Strong IKK2 expression was detected by Western blot in pancreatic extracts of IKK2-CAPdx-1 mice after doxycycline (Dox) withdrawal and disappeared after Dox readministration; extracellular signal–related kinase-2 (ERK2) was used as loading control. D: Immunofluorescence staining of paraffin- embedded pancreatic sections showing mosaic expression of transgenic human IKK2 in islets of 12-week-old IKK2-CAPdx-1 mice. The staining shows that IKK2-positive cells have reduced or even no insulin immunoreactivity. Sections were costained with DAPI (blue) for nuclei. Scale bar, 50 mm. E: An electrophoretic mobility shift assay of whole-cell islet extracts with NF-kB and SP1-specific probes showing strong activation of NF-kB in 12-week-old IKK2-CAPdx-1 mice. SP1 serves as a quality control. F: Immunohistological staining of paraffin- embedded pancreatic sections showing IKK2-CA–induced nuclear localization of RelA in b-cells of 12-week-old IKK2-CAPdx-1 mice. Scale bar, 50 mm. Controls denoted in this figure were single transgenic for IKK2-CA transgene. RLU, relative light unit. 964 NF-kB Activation in b-Cells Induces Diabetes Diabetes Volume 63, March 2014

Figure 2—Expression of IKK2-CA in b-cells results in diabetes development. Animals were kept under doxycycline during breeding, doxycycline was removed after weaning, and mice were then analyzed at the indicated ages (A)orat24–28 weeks (B–F). A: Fed blood glucose levels of 8, 11–13, 14–17, and 24- to 28-week-old control (n = 9, 13, 7, and 28), Pdx-1+/2 (n = 6, 12, 3, and 17), and IKK2-CAPdx-1 mice (n = 5, 12, 7, and 28). The upper detection limit of the glucometer is 600 mg/dL. B: Overnight-fasted blood glucose levels of control (n = 21), Pdx-1+/2 (n = 16), and IKK2-CAPdx-1 mice (n = 22). C: Fed plasma insulin level of control (n = 12), Pdx-1+/2 (n = 6), and IKK2-CAPdx-1 mice (n = 12). D: Representative immunoﬂuorescent images of parafﬁn sections of IKK2-CAPdx-1 and Pdx-1+/2 pancreata stained for insulin (red) and IKK2 (green) and costained with DAPI for nuclei in blue. Scale bar, 50 mm. E: Pancreatic insulin content of control, Pdx-1+/2, and IKK2-CAPdx-1 mice (n =6–9/group). F: Quantitative RT-PCR of Ins2 mRNA of control, Pdx-1+/2, and IKK2-CAPdx-1 mice (fold upregulation vs. control; n =8–12/group). Hprt was used as a reference gene. Results were analyzed by Student t test and- presented as the mean 6 SEM. **P < 0.01; ***P < 0.001.

insulin immunoreactivity (Fig. 2D) and reduced Pdx-1 This implicates that IKK2/NF-kBactivationinb-cells expression (Supplementary Fig. 1A). Costaining with leads to suppression of insulin expression and thus loss IKK2 revealed that most of the IKK2-CA–expressing cells of their functionality in addition (or prior) to their were no longer positive for insulin. Interestingly, few demise. In accordance, pancreatic insulin content and insulin-negative cells still expressed nuclear Pdx-1, while insulin transcription had dropped to ;3 and 26% of 2 more such cells retained the expression of the endocrine the values of the Pdx-1+/ littermates, respectively marker chromogranin A (Supplementary Fig. 1A and B). (Fig. 2E and F). diabetes.diabetesjournals.org Salem and Associates 965

Figure 3—IKK2/NF-kB activation in b-cells induces inflammation, insulitis, and increased islet antigen presentation. A: Hematoxylin-eosin staining of paraffin sections from pancreas of Pdx-1+/2 and IKK2-CAPdx-1 mice showing peri-insulitis (I) and invasive insulitis (II)in IKK2-CAPdx-1 mice. B: Flow cytometric analysis of pancreatic cells from control (white bars) and IKK2-CAPdx-1 (black bars). The data are represented as relative to the number of CD45+ cells in the control samples, which is set to one (n = 5/group from two independent experiments). Immune cells were analyzed for the expression of surface markers CD3+ (T cells), CD19+ (B cells), and CD11b+/CD11c+ (dendritic cells [DCs]), CD11b+/F4/80+ (macrophages [Mf]), and NK1.1+ (natural killer cells [NK]). MHC II expression on pancreatic cells of the myeloid (C) and lymphoid lineage (D), analyzed by flow cytometry and gated as indicated. Elevated MHC II levels show that infiltrated 966 NF-kB Activation in b-Cells Induces Diabetes Diabetes Volume 63, March 2014

Pdx-1 is an important transcription factor for b-cell IKK2-CAPdx-1 Mice Show Signs of Apoptosis and 2 function, and its heterozygosity, present in the Pdx-1+/ Endoplasmic Reticulum Stress knockin mouse, was shown to be associated with some As apoptosis is thought to be the major cause of b-cell 2 b-cell defects (6). Indeed, Pdx-1+/ mice showed slight death in diabetes (16), we performed TUNEL assays that elevation in fed blood glucose level (Fig. 2A) and re- showed the presence of apoptotic cells only in IKK2- duction in insulin transcription and pancreatic content CAPdx-1 islets (Fig. 4A and B). This indicates that apo- compared with control littermates (Fig. 2E and F); how- ptosis may account at least in part for the reduced b-cell ever, in contrast to IKK2-CAPdx-1 mice, they were not mass in IKK2-CAPdx-1 mice. b-Cells are susceptible to diabetic. endoplasmic reticulum (ER) stress, and ER stress- mediated apoptosis in b-cells has been implicated in the IKK2-CA–Induced Diabetes Is Associated With pathogenesis of diabetes (17). Moreover, islets from Inflammation and Leukocytic Infiltration prediabetic NOD mice (18) and patients with type 1 di- Invasion of pancreatic islets by leukocytes is the hallmark abetes (19) exhibit signs of ER stress. Consistent with 2 of immune-mediated diabetes. Unlike Pdx-1+/ mice, that, the ER stress-related factors Ddit3/Chop and Atf3 islets of IKK2-CAPdx-1 showed marked peri-insulitis were upregulated in IKK2-CAPdx-1 mice (Fig. 4C). In ad- (perivascular, periductal, and peri-islet infiltrates) and dition, other stress-associated apoptosis-inducing factors insulitis (Fig. 3A), reminiscent of the insulitic process in like Nos2 (iNOS) and Myc were also elevated in IKK2- humans. Of note, similar to human histopathology, not CAPdx-1 pancreata (Fig. 4D and E). all islets showed mononuclear infiltration (Supplemen- tary Fig. 2). IKK2-CA–Induced Diabetes Is Reversible Flow cytometry revealed that hematopoietic (CD45+) To examine the possibility of reverse remodeling the cells in IKK2-CAPdx-1 pancreata increased fourfold (Fig. diabetic phenotype, doxycycline was readministered for 3B) and that infiltrating leukocytes were activated CD4+ 30 days to diabetic IKK2-CAPdx-1 mice (Fig. 5A). Diabetes and CD8+ T, B, and dendritic cells as well as macrophages in IKK2-CAPdx-1 mice was confirmed by elevated fed and and few natural killer cells (Fig. 3B–D and Supplementary fasted blood glucose values and reduced plasma insulin Fig. 3A). This was further confirmed on the histological levels (Fig. 5B and C). We observed a clear reduction in level (Fig. 3E and F and Supplementary Figs. 3B and C fed blood glucose values already within the first 10 days, and 4). The majority of infiltrating T cells was of the which virtually completely normalized in all animals by CD4+ type (Fig. 3B and F). Furthermore, we detected 30 days (Fig. 5B). Consistently, fasting blood glucose and hyperexpression of MHC I and II by b-cells of IKK2- fed plasma insulin levels were also restored (Fig. 5C). CAPdx-1 mice (Fig. 3G, Supplementary Fig. 5A–C), as Doxycycline-dependent transgene inactivation was con- previously reported in diabetic patients (10). This sug- firmed by immunofluorescence staining (Fig. 5D) and gests increased antigen presentation capacity of b-cells in Western blot (Fig. 1C). Diabetic IKK2-CAPdx-1 mice also these mice, which might be critical for phenotype de- regained normal structured islets showing virtual ab- velopment. In accordance, there was transcriptional sence of infiltrating cells upon doxycycline treatment upregulation of the MHC I and II components, H2-Q4 (Fig. 5D and E). Insulin immunostaining demonstrated and H2-Aa (Fig. 3H), respectively. the reappearance of b-cell–rich islets and prominent The insulitic process in IKK2-CAPdx-1 mice was asso- b-cell regranulation, which strongly stained for insulin ciated with elevated expression of inflammatory cyto- (Fig. 5D). Importantly, the reversion of the diabetes kines and chemokines like tumor necrosis factor (Tnf), status was accompanied by increased levels of Ki67 im- Ccl5, Ccl2, and Cxcl10 and the adhesion molecule in- munoreactivity in islets of IKK2-CAPdx-1 mice (Fig. 5F tracellular adhesion molecule-1 (Fig. 3I). This in- and Supplementary Fig. 6), indicating that proliferation flammatory profile is reminiscent of those found in islets is involved in b-cell regeneration in this model. Fur- from patients, in animals with type 1 diabetes (11,12), thermore, transcription of insulin, MHC I/II molecules, and in normal human islets subjected to cytokines or and various inflammatory markers was effectively nor- enteroviruses (13–15). malized (Fig. 5G–I).

cells in IKK2-CAPdx-1 were evidently activated. Data are representative of two experiments with five mice per group (C) or one experiment with two mice per group (D). Pdx-1+/2 mice were included in the control group in B–D, as they show the same pattern. E: Immunohis- tochemical staining of paraffin sections from Pdx-1+/2 and IKK2-CAPdx-1 mice (n = 3/group) for the B-cell marker B220. F: Immunofluo- rescence staining of cryosections from pancreas of Pdx-1+/2 and IKK2-CAPdx-1 mice (n = 3/group) for CD4, CD8, or CD11c (DC marker) together with insulin and DAPI. G: Immunofluorescence staining of cryosections from pancreas of Pdx-1+/2 and IKK2-CAPdx-1 mice for insulin (red) and MHC II (green) (n = 3/group) showing the expression of MHC II by some b-cells (arrows, inset). Quantitative RT-PCR for mRNA transcripts involved in antigen presentation (H) and inflammation (I) from control (white bars), Pdx-1+/2 (gray bars), and IKK2-CAPdx-1 (black bars) mice (fold upregulation vs. control; n =8–12). Hprt was used as a reference gene. Results were analyzed by Student t test and presented as the mean 6 SEM. Animals were 24–28 weeks old. Scale bars, 50 mm. *P < 0.05; **P < 0.01; ***P < 0.001. Icam1, intracellular adhesion molecule-1. diabetes.diabetesjournals.org Salem and Associates 967

Figure 4—Islets of IKK2-CAPdx-1 mice show signs of apoptosis and ER stress. A: TUNEL assay on parafﬁn sections of pancreas from 24- to 26-week-old control, Pdx-1+/2, and IKK2-CAPdx-1 mice (n = 3/group). Sections were counterstained with methyl green. Arrowheads indicate the presence of apoptotic nuclei (TUNEL-positive nuclei) in the islets of IKK2-CAPdx-1 mice. Dashed line marks the islet area. Scale bar, 50 mm. B: Quantiﬁcation of TUNEL staining in A. A total of >400 islets from three animals per group were analyzed. C–E: Quantitative RT-PCR for mRNA transcripts encoding stress and apoptosis-inducing genes in pancreata of 24- to 26-week-old control, Pdx-1+/2, and IKK2-CAPdx-1 (fold upregulation vs. control; n =6–10/group). Hprt was used as a reference gene. *P < 0.05; **P < 0.01; ***P < 0.001. Results were analyzed by Student t test and presented as the mean 6 SEM. ND, not detectable.

IKK2-CAPdx-1 Mice Exhibit Gene Expression Profile (threshold, 1.5-fold; P , 0.05 in t test with Benjamini- Mimicking Antiviral and Allergic Inflammatory Hochberg correction) (Supplementary Table 1, selected Responses genes, and Table 1). Several of those genes were verified To gain insight into molecular events induced by IKK2/ by quantitative PCR (qPCR) as depicted in Fig. 6B. The NF-kB activation and possibly involved in diabetes de- increased expression of various cytokines, chemokines, velopment, we performed gene expression profiling of adhesion molecules, and antigen presentation molecules pancreatic islets isolated from 11-week-old IKK2-CAPdx-1 clearly demonstrated an activation of innate and adap- animals. This age represents a critical checkpoint in tive immune responses in islets of IKK2-CAPdx-1 mice phenotype development at which some animals already (Table 1, Supplementary Table 1, and Fig. 6B). However, showed hyperglycemia, while others were still normo- aside from these factors typically detected in other type 1 glycemic (Fig. 6A). In this analysis, including samples diabetes models, other less-characterized genes were from both types, 288 transcripts were found to be highly upregulated (Table 1, Supplementary Table 1, and upregulated, 46 of which are known NF-kB targets Fig. 6B). The T-cell–directed chemokine Ccl17 (also (Supplementary Table 2) and 28 to be downregulated known as thymus and activation-regulated chemokine) 968 NF-kB Activation in b-Cells Induces Diabetes Diabetes Volume 63, March 2014

Figure 5—Recovery from diabetes, alleviation of the inflammatory status, and regeneration of pancreatic islets upon transgene inactivation. A: Experimental design. Mice were kept under doxycycline (Dox; 0.1 g/L Dox in drinking water) until weaning and discontinued until 24–28 weeks to activate transgene expression. The diabetic phenotype was confirmed by fed and fasted blood glucose as well as plasma insulin measurements (before Dox). Dox (1 g/L) was then readministered in drinking water for 30 days (except for F) to inactivate transgene expression. Fed blood glucose was monitored during that time after which animals were analyzed (after Dox). B: Fed blood glucose monitoring before and during Dox readministration of control (white circles), Pdx-1+/2 (white squares), and IKK2-CAPdx-1 (white triangles) mice (n =7–9/group from two independent experiments). Significance is relative to the Pdx-1+/2 group at the specified time diabetes.diabetesjournals.org Salem and Associates 969 was the most prominent one. Ccl17 together with Ccl22 IKK2/NF-kB signaling, specifically in b-cells. IKK2- (also known as macrophage-derived chemokine), another CAPdx-1 animals develop severe hyperglycemia and elevated chemokine, are known ligands for the Ccr4 hypoinsulinemia mirroring substantial b-cell loss that is receptor, which has been linked to the development associated with marked islet inflammation. Similar to the of autoimmune diabetes (20). Also, a cluster of human immunopathology, infiltrates are not seen in all inflammation-related serine proteases called serine pro- pancreatic islets and include different kinds of immune tease inhibitor clade A member 1 (Serpina1) and different cells (21). This model is the first animal model to activate members of the cationic protein products of eosinophils, IKK2/NF-kBinb-cells and to prove directly its capability the eosinophil-associated ribonucleases (Ears), were to trigger diabetes development on its own. IKK2-CAPdx-1 upregulated. Kyoto Encyclopedia of Genes and Genomes islets show a gene expression signature reflecting the analysis clearly revealed a signature reflecting antiviral activation of innate immunity and type I IFN response, responses and reactions to infections. This was repre- which could generate a proinflammatory microenviron- sented by the upregulation of different interferon (IFN)- ment sufficient to recruit different immune cells. These regulated genes, especially involved in the type 1 IFN infiltrating cells are able to contribute to and/or amplify response including signal transducer and activator of the inflammatory insult, thereby promoting b-cell de- transcription 1 (Stat1), IFN regulatory factor (Irf) 7, Irf8, struction as observed in diabetic subjects and NOD mice and Irf5 as well as IFN-inducible genes like Ifih1 and Ifit1. (1,2). Furthermore, increased expression of MHC I and II Furthermore, genes involved in oxidative stress (Nox1) in transgenic b-cells indicates that IKK2/NF-kB signaling and tissue remodeling (Mmp12) were also elevated (Table is capable of increasing islet antigen presentation to in- 1 and Supplementary Table 1). Taken together, the ob- filtrating T cells, which could participate in the autoim- served gene expression profile in the IKK2-CAPdx-1 model mune insult. In addition, MHC II expression by b-cells mimics to a great extent that detected in virus-induced can promote their death, as seen by MHC II ligation in immune-mediated diabetes and, importantly, points to antigen-presenting cells (22). This may allow for an in- other potential novel candidate genes for diabetes. tensive dialogue between b-cells and immune cells that finally promotes b-cell destruction in a T-cell–dependent Ccl17 Expression Precedes Phenotype Development manner. The main T-cell subtype in our model is the In an attempt to identify candidate genes involved in CD4+ type, in contrast to the concept that CD8+ T cells phenotype initiation, we followed up the temporal ex- predominate in humans (21), However, there is a con- pression pattern of distinct genes. Interestingly, Ccl17 siderable heterogeneity within diabetic patients (e.g., was massively upregulated in islets of 8-week-old IKK2- even no autoreactive T cells were found in a subset of Pdx-1 CA , when animals were asymptomatic, while there recent-onset patients) (23). In addition, CD4+ T cells can was only mild upregulation of Ear2, Serpina1a, Irf7, and mediate b-cell death in transgenic NOD mice (24), and Madcam1 and no elevation of MHC I/II molecules diabetes development was shown to require the presence (Fig. 7A). These data support the hypothesis that of both CD4+ and CD8+ T cells (25). Ccl17 might be a critical initial effector of IKK2/NF-kB– The IKK2-CAPdx-1 phenotype is in line with the mediated diabetes development. In addition, qPCR reported role of NF-kB as a mediator of cytokine-induced Pdx-1 analysis of highly diseased IKK2-CA mice (;24 b-cell destruction (4,5) and its activation in islets of weeks old) revealed sustained upregulation of Ccl17, prediabetic NOD mice (18). Consistent with this notion, Ear2, and Serpina1a that was completely normalized in resistance to streptozotocin (STZ)-induced diabetes was the reverse-remodeling experiment (Fig. 7B). achieved by b-cell–specific inhibition of NF-kB (26,27), and various natural products were found to inhibit STZ- DISCUSSION induced diabetes and protect against b-cell damage through In this study, we describe a novel mouse model of NF-kB inhibition (4). However, b-cell–specific repression of immune-mediated diabetes, the IKK2-CAPdx-1 model that NF-kB in normal mice elicits hyperglycemia and defective is based on genetic activation of proinflammatory glucose-stimulated insulin secretion and accelerates

point. At 20 and 30 days after Dox administration, Pdx-1+/2 mice had significantly higher blood glucose values relative to the control littermates. C: Overnight-fasting blood glucose (top) and fed plasma insulin levels (bottom) before and after doxycycline readministration. Immunofluorescence (D) and hematoxylin-eosin stainings (E) of paraffin sections from Pdx-1+/2 and IKK2-CAPdx-1 mice after Dox readministration showing regenerated b-cells with normal insulin content and near absence of infiltrating cells. F: Immunofluorescence staining of insulin (red), Ki67 (green), and DAPI (blue) of paraffin sections from Pdx-1+/2 and IKK2-CAPdx-1 mice 10 days after Dox readministration. The staining shows the presence of increased Ki67-positive b-cells (arrows) in the islets of IKK2-CAPdx-1 mice as compared with Pdx-1+/2 ones. Scale bars in D–F,50mm. Quantitative RT-PCR for insulin mRNA (G), MHC I (H2-Q4) and MHC II (H2-Aa) molecules (H), and inflammatory genes (I) in pancreata of control (white bars), Pdx-1+/2 (gray bars), and IKK2-CAPdx-1 (black bars) mice after Dox readministration (fold upregulation vs. control; n =6–10/group from 2 independent experiments). Hprt was used as a reference gene. *P < 0.05; **P < 0.01; ***P < 0.001. Results were analyzed by Student t test and presented as the mean 6 SEM. Icam1, intracellular adhesion molecule-1. 970 NF-kB Activation in b-Cells Induces Diabetes Diabetes Volume 63, March 2014

Figure 6—Validation of selected microarray data by real-time PCR. A: Fed and fasted blood glucose values from 11- to 13-week-old control (white circles), Pdx-1+/2 (white squares), and IKK2-CAPdx-1 (white and black triangles) mice (n =12–13/group). Black triangles represent samples used in the microarray analysis. B: Quantitative RT-PCR analysis of isolated islets from 11- to 13-week-old control, Pdx-1+/2,and IKK2-CAPdx-1. Samples used for the qPCR assays include those used in the array analysis and additional samples. Shown is the fold upregulation vs. control animals (n =5–8/group). Actb was used as a reference gene. Results were analyzed by Student t test and presented as the mean 6 SEM. *P < 0.05; **P < 0.01; ***P < 0.001. Il12b,interleukin-12b;Madcam1, mucosal addressin cell adhesion molecule-1. diabetes.diabetesjournals.org Salem and Associates 971

Table 1—List of selected genes differentially regulated in diabetes development in NOD mice (4,5), which may islets of 11-week-old IKK2-CAPdx-1 mice as compared with implicate that the extent/level, context, and timing of Pdx-1+/2 littermates NF-kB activation dictate the overall outcome of diabetes Class Name Fold pathogenesis. Chemokines and their receptors Ccl17 24.59 Our gene expression data suggest that activation of Ccl22 4.68 IKK2/NF-kBinb-cells is sufficient to initiate different Ccl5 3.10 cellular mechanisms formerly shown to affect b-cell Ccl19 2.60 function/survival and induce diabetes in humans and Cxcl10 2.63 – Cxcl16 2.22 animals (16,17,28 32). These diverse factors (TNF-a, Ccr2 3.43 Atf3, C/EBP homologous protein, Nos2, c-myc, Stat1, Ccr7 3.08 and Nox-1) are known to induce inflammation, oxidative Ccr5 2.25 stress, ER stress, and nitric oxide production (Fig. 7C) Interleukins and their receptors Il1a 2.61 that together may finally funnel in b-cell dysfunction Il12b 2.25 and/or apoptosis in our model. However, we cannot ex- Il1b 1.90 clude other unknown effects of IKK2/NF-kB signaling Il6 1.72 Il1r2 2.13 that might interfere with b-cell function and promote Il2ra 2.14 disease development (Fig. 7C). Il2rg 2.77 The expression profile in our model is further pointing Il12rb2 1.52 to the involvement of innate immunity and type I IFN Adhesion molecules Itgax 4.64 response (Fig. 7C), characteristic for viral infection, and Itgae 3.64 is similarly detected in pancreata and islets from patients Itga2 2.38 Madcam1 4.19 at clinical onset and long-standing diabetes (12). Fur- Vcam1 2.51 thermore, the gene-expression program induced by en- Icam1 2.30 teroviral infection of human islets (14,15) is also Pdx-1 Antigen presentation and processing Ctss 2.11 prominently mimicked by the IKK2-CA model, sug- MHC I H2-M2 4.74 gesting that IKK/NF-kB is a critical downstream effector H2-Q8 3.17 in this context. Viral infections have been proposed as MHC II H2-Ab1 2.39 a triggering factor of type 1 diabetes in humans and H2-Aa 2.14 animals (33), and importantly, they activate Toll-like and – Antiviral response (IFN responsive) Irf7 2.60 nucleotide-binding oligomerization domain like recep- Irf8 2.34 tors, which are well-known inducers of NF-kB. Indeed, Irf5 1.84 NF-kB signaling is activated in b-cells by enterovirus Ifit1 2.28 infection (15) and double-stranded RNA treatment (34) fi I h1 1.74 and mediates at least partially the deleterious effects of Innate immunity Clec7a 3.10 these insults. Recently, Irf7, the master regulator of type Tlr3 1.53 I IFN–dependent immune responses and upregulated in STAT-mediated signal transduction Stat1 2.51 our model, has been implicated in type 1 diabetes path- Jak2 1.53 ogenesis (35). Similarly, Ifih, another candidate gene for TNF signaling Tnf 1.55 type 1 diabetes that is expressed in human islets and Tnip3 1.95 fl Traf1 1.79 can be induced by in ammatory cytokines (13) and by Tnfssf13b 1.67 double-stranded RNA in rat b-cells (36), is elevated in Tnip1 1.65 IKK2-CAPdx-1 mice. Therefore, it is well conceivable that Miscellaneous Ear2 10.06 NF-kB activation in b-cells on its own could create an Ear10 7.63 innate immune response similar to an antiviral response, Ear1 6.20 which finally culminates in diabetes development. How- Serpina1e 9.63 ever, we cannot exclude a possible role of Pdx-1 haplo- Serpina1a 9.22 fi Serpina1b 6.52 insuf ciency on promoting b-cell susceptibility to Mmp12 4.92 IKK2/NF-kB–induced inflammation and phenotype Nox1 4.06 development. We also identified other novel diabetes-associated Icam1, intracellular adhesion molecule-1; Il, interleukin; Il1r, candidate genes. Ccl17, the most upregulated one, is interleukin-1 receptor; Il2r, interleukin-2 receptor; Jak2, Janus a chemokine involved in immune and allergic in- kinase 2; Madcam1, mucosal addressin cell adhesion fl – molecule-1; Tlr, Toll-like receptor; Traf1, TNF receptor– ammatory responses (20,37 42). The early marked associated factor 1; Vcam1, vascular cell adhesion molecule-1. upregulation of Ccl17 at normoglycemia supports the idea of being a foremost effector of IKK2/NF-kB signaling that contributes to diabetes development. Consistently, viral infection of human B cells induces NF-kB–dependent 972 NF-kB Activation in b-Cells Induces Diabetes Diabetes Volume 63, March 2014

Figure 7—Gene-expression kinetics during diabetes development in IKK2-CAPdx-1 mice. A: Quantitative RT-PCR from islets isolated from 8-week-old control, Pdx-1+/2, and IKK2-CAPdx-1 mice. Shown is the fold upregulation vs. control animals (n =4–6/group). Actb was used as a reference gene. B: Quantitative RT-PCR from pancreata of 24- to 26-week-old animals before (diseased) and after doxycycline readministration (reversed). Shown is the fold upregulation vs. control animals (in diseased state, n =6–10/group; in reversed state, n =4– 5/group). Hprt was used as a reference gene. Results were analyzed by Student t test and presented as the mean 6 SEM. C: Proposed model for IKK2/NF-kB–induced diabetes. Activation of IKK2/NF-kBinb-cells is sufficient to induce gene expression programs involved in innate immunity, type I IFN response, and antigen presentation. This may subsequently result in oxidative stress, ER stress, and nitric oxide (NO) production that finally funnels in b-cell dysfunction and death that is probably mediated by apoptosis. In addition, there was a massive increase in the chemokine Ccl17, formerly shown to be involved in autoimmunity and allergic inflammatory processes. Moreover, a direct effect of IKK2/NF-kBactivationonb-cell function via unknown mechanisms may also contribute to diabetes development in this model. *P < 0.05; **P < 0.01. CHOP, C/EBP homologous protein; Madcam1, mucosal addressin cell adhesion molecule-1. diabetes.diabetesjournals.org Salem and Associates 973 expression of Ccl17 and Ccl22 (also elevated in our model) suggests that proliferation is also involved in the resto- (37), and IKK/NF-kB inhibition prevents cytokine-induced ration of b-cell mass in this model as found in other Ccl17 production in keratinocytes (38). Ccl17 is mainly models; however, other mechanisms cannot be excluded. produced by dendritic cells (39) and plays an important This finding implies that the IKK/NF-kB system might be role in T cell development, trafficking, and activation. a potential target for clinical intervention. Furthermore, Furthermore, it was highlighted as a novel biomarker for the IKK2-CAPdx-1 mouse model could be valuable for allergic inflammatory diseases like asthma (40) and atopic assessing mechanisms and identifying genes involved in dermatitis (41), in which it is strongly coexpressed with b-cell regeneration as well. Since diabetes passes through Ccl22 (42). relapsing-remitting onsets, our model is further poten- Both Ccl17 and Ccl22 preferentially attract CD4+ tially useful for evaluating clinical therapeutics at dif- T cells via the Ccr4 receptor. Cells expressing Ccl17 ferent stages of the human disease presentation through were detected within infiltrated islets from prediabetic switching the system off and on. NOD mice, and Ccr4-positive T cells were shown to be In summary, the IKK2-CAPdx-1 model represents critically involved in autoimmune diabetes de- a novel conditional mouse model of immune-mediated velopment (20). Additionally, neutralizing Ccl22 anti- diabetes with very distinct checkpoints of b-cell de- bodies inhibited insulitis and diabetes, whereas struction and regeneration. This model phenocopies transgenic Ccl22 expression accelerated disease de- major aspects of the human disease and may represent velopment (20). However, autoimmune diabetes was a valuable tool for improving preclinical drug assessment prevented by Ccl22-mediated regulatory T-cell re- for diabetes treatment. cruitment in another study (43). Furthermore, no significant difference in plasma Ccl17 was detected in type1diabetespatients(44).Yet,thisstudywas Acknowledgments. The authors thank Ute Leschik, Melanie Gerstenlauer, and Bianca Ries (University of Ulm) for excellent technical assistance and conducted in a very small cohort of diabetic Japanese Karlheinz Holzmann (Genomics-Core Facility, University Hospital Ulm) for subjects, which certainly does not exclude a different performing the microarray analysis. pattern in other patient subsets. Funding. This work was supported by grants GRK-1041-P3 (Deutsche Ears, also prominently induced in our model, re- Forschungsgemeinschaft) and BIU-C6 (Boehringer Ingelheim Ulm University present a subgroup of the RNase A family secreted by BioCenter) to B.B. rodent eosinophils (45) and also expressed by macro- Duality of Interest. No potential conflicts of interest relevant to this phages. Ears participate in host defense by their anti- article were reported. bacterial and antiviral activity together with chemotaxis to dendritic cells (45). The Serpina1 family of genes Author Contributions. H.H.S. and B.B. designed and performed the experiments, analyzed data, and wrote the manuscript. B.T. and K.F. per- encodes for inhibitors of serine proteases. In humans, it formed research and analyzed data. H.J.M. contributed to discussion and is represented by a single gene, called a1-antitrypsin reviewed and edited the manuscript. R.S. and M.W. contributed to discussion, AAT ( ) for which expression was enhanced by proin- performed the experiments, and reviewed and edited the manuscript. B.O.B. flammatory cytokines in islet cells (46). AAT is an acute- and T.W. designed research and reviewed and edited the manuscript. H.H.S. phase protein with anti-inflammatory, tissue-protective, and B.B. are the guarantors of this work and, as such, had full access to all the and antiapoptotic properties that is able to prolong data in the study and take responsibility for the integrity of the data and the islet allograft survival and to inhibit cytokine and STZ- accuracy of the data analysis. induced b-cell apoptosis (47,48) as well as diabetes development in NOD mice (49). Although decreased References functional activity of serum AAT was shown in diabetes, 1. Bluestone JA, Herold K, Eisenbarth G. Genetics, pathogenesis and clinical serum levels are variable, probably reflecting changes interventions in type 1 diabetes. Nature 2010;464:1293–1300 fl in the in ammatory status of the disease (50). To 2. Eizirik DL, Colli ML, Ortis F. The role of inflammation in insulitis and beta- fi what extent these newly identi ed genes contribute cell loss in type 1 diabetes. Nat Rev Endocrinol 2009;5:219–226 to human disease pathology, however, needs further 3. Baker RG, Hayden MS, Ghosh S. NF-kB, inflammation, and metabolic investigation. disease. Cell Metab 2011;13:11–22 One striking aspect of the IKK2-CAPdx-1 model is the fl enormous recovery potential from the diabetes status 4. Baumann B, Salem HH, Boehm BO. Anti-in ammatory therapy in type 1 – including the resolution of insulitis and the efficient re- diabetes. Curr Diab Rep 2012;12:499 509 generation of insulin-positive b-cells after transgene in- 5. Zhao Y, Krishnamurthy B, Mollah ZU, Kay TW, Thomas HE. NF-kB in type 1 activation. A possible rationale is that a portion of b-cells diabetes. Inflamm Allergy Drug Targets 2011;10:208–217 is not destroyed but, rather, has lost its functional and 6. Holland AM, Hale MA, Kagami H, Hammer RE, MacDonald RJ. Experi- differentiated state and therefore can regain its func- mental control of pancreatic development and maintenance. Proc Natl tionality upon IKK2-CA inactivation with the consequent Acad Sci USA 2002;99:12236–12241 resolution of proinflammatory mediators and disap- 7. Baumann B, Wagner M, Aleksic T, et al. Constitutive IKK2 activation in pearance of immune cells. Furthermore, the increase of acinar cells is sufficient to induce pancreatitis in vivo. J Clin Invest 2007; Ki67-positive b-cells during the regeneration phase 117:1502–1513 974 NF-kB Activation in b-Cells Induces Diabetes Diabetes Volume 63, March 2014

8. Lattke M, Magnutzki A, Walther P, Wirth T, Baumann B. Nuclear factor kB 26. Eldor R, Yeffet A, Baum K, et al. Conditional and specific NF-kappaB activation impairs ependymal ciliogenesis and links neuroinflammation to blockade protects pancreatic beta cells from diabetogenic agents. Proc hydrocephalus formation. J Neurosci 2012;32:11511–11523 Natl Acad Sci USA 2006;103:5072–5077 9. Rajasalu T, Brosi H, Schuster C, et al. Deficiency in B7-H1 (PD-L1)/PD-1 27. Ueberberg S, Deutschbein T, Klein HH, et al. Protection from diabetes coinhibition triggers pancreatic beta-cell destruction by insulin-specific, development by single-chain antibody-mediated delivery of a NF-kB in- murine CD8 T-cells. Diabetes 2010;59:1966–1973 hibitor specifically to b-cells in vivo. Am J Physiol Endocrinol Metab 2011; – 10. Bottazzo GF, Dean BM, McNally JM, MacKay EH, Swift PG, Gamble DR. In 301:E83 E90 situ characterization of autoimmune phenomena and expression of HLA 28. Laybutt DR, Weir GC, Kaneto H, et al. Overexpression of c-Myc in beta-cells molecules in the pancreas in diabetic insulitis. N Engl J Med 1985;313: of transgenic mice causes proliferation and apoptosis, downregulation of 353–360 insulin gene expression, and diabetes. Diabetes 2002;51:1793–1804 11. Sarkar SA, Lee CE, Victorino F, et al. Expression and regulation of 29. Allagnat F, Fukaya M, Nogueira TC, et al. C/EBP homologous protein chemokines in murine and human type 1 diabetes. Diabetes 2012;61: contributes to cytokine-induced pro-inflammatory responses and apoptosis 436–446 in b-cells. Cell Death Differ 2012;19:1836–1846 12. Planas R, Carrillo J, Sanchez A, et al. Gene expression profiles for the 30. Moore F, Naamane N, Colli ML, et al. STAT1 is a master regulator of human pancreas and purified islets in type 1 diabetes: new findings at pancreatic beta-cell apoptosis and islet inflammation. J Biol Chem 2011; clinical onset and in long-standing diabetes. Clin Exp Immunol 2010;159: 286:929–941 – 23 44 31. Newsholme P, Morgan D, Rebelato E, et al. Insights into the critical role of 13. Eizirik DL, Sammeth M, Bouckenooghe T, et al. The human pancreatic islet NADPH oxidase(s) in the normal and dysregulated pancreatic beta cell. transcriptome: expression of candidate genes for type 1 diabetes and the Diabetologia 2009;52:2489–2498 fl impact of pro-in ammatory cytokines. PLoS Genet 2012;8:e1002552 32. Rabinovitch A, Sumoski W, Rajotte RV, Warnock GL. Cytotoxic effects of 14. Ylipaasto P, Kutlu B, Rasilainen S, et al. Global profiling of coxsackievirus- cytokines on human pancreatic islet cells in monolayer culture. J Clin and cytokine-induced gene expression in human pancreatic islets. Dia- Endocrinol Metab 1990;71:152–156 – betologia 2005;48:1510 1522 33. Hober D, Sauter P. Pathogenesis of type 1 diabetes mellitus: interplay 15. Ylipaasto P, Smura T, Gopalacharyulu P, et al. Enterovirus-induced gene between enterovirus and host. Nat Rev Endocrinol 2010;6:279–289 fi expression pro le is critical for human pancreatic islet destruction. Dia- 34. Liu D, Cardozo AK, Darville MI, Eizirik DL. Double-stranded RNA cooperates – betologia 2012;55:3273 3283 with interferon-gamma and IL-1 beta to induce both chemokine expression 16. Cnop M, Welsh N, Jonas JC, Jörns A, Lenzen S, Eizirik DL. Mechanisms of and nuclear factor-kappa B-dependent apoptosis in pancreatic beta-cells: pancreatic beta-cell death in type 1 and type 2 diabetes: many differences, potential mechanisms for viral-induced insulitis and beta-cell death in type few similarities. Diabetes 2005;54(Suppl. 2):S97–S107 1 diabetes mellitus. Endocrinology 2002;143:1225–1234 17. Eizirik DL, Cardozo AK, Cnop M. The role for endoplasmic reticulum stress 35. Heinig M, Petretto E, Wallace C, et al.; Cardiogenics Consortium. A trans- in diabetes mellitus. Endocr Rev 2008;29:42–61 acting locus regulates an anti-viral expression network and type 1 diabetes – 18. Tersey SA, Nishiki Y, Templin AT, et al. Islet b-cell endoplasmic reticulum risk. Nature 2010;467:460 464 stress precedes the onset of type 1 diabetes in the nonobese diabetic 36. Colli ML, Moore F, Gurzov EN, Ortis F, Eizirik DL. MDA5 and PTPN2, two mouse model. Diabetes 2012;61:818–827 candidate genes for type 1 diabetes, modify pancreatic beta-cell responses 19. Marhfour I, Lopez XM, Lefkaditis D, et al. Expression of endoplasmic re- to the viral by-product double-stranded RNA. Hum Mol Genet 2010;19: – ticulum stress markers in the islets of patients with type 1 diabetes. 135 146 Diabetologia 2012;55:2417–2420 37. Nakayama T, Hieshima K, Nagakubo D, et al. Selective induction of Th2- 20. Kim SH, Cleary MM, Fox HS, Chantry D, Sarvetnick N. CCR4-bearing attracting chemokines CCL17 and CCL22 in human B cells by latent – T cells participate in autoimmune diabetes. J Clin Invest 2002;110: membrane protein 1 of Epstein-Barr virus. J Virol 2004;78:1665 1674 1675–1686 38. Jung MR, Lee TH, Bang MH, et al. Suppression of thymus- and activation- 21. Richardson SJ, Willcox A, Bone AJ, Morgan NG, Foulis AK. Immunopa- regulated chemokine (TARC/CCL17) production by 3-O-b-D-glucopyano- thology of the human pancreas in type-I diabetes. Semin Immunopathol sylspinasterol via blocking NF-kB and STAT1 signaling pathways in TNF-a 2011;33:9–21 and IFN-g-induced HaCaT keratinocytes. Biochem Biophys Res Commun 2012;427:236–241 22. Bertho N, Drénou B, Laupeze B, et al. HLA-DR-mediated apoptosis susceptibility discriminates differentiation stages of dendritic/monocytic APC. 39. Alferink J, Lieberam I, Reindl W, et al. Compartmentalized production of J Immunol 2000;164:2379–2385 CCL17 in vivo: strong inducibility in peripheral dendritic cells contrasts selective absence from the spleen. J Exp Med 2003;197:585–599 23. Dotta F, Censini S, van Halteren AG, et al. Coxsackie B4 virus infection of beta cells and natural killer cell insulitis in recent-onset type 1 diabetic 40. Hartl D, Lee CG, Da Silva CA, Chupp GL, Elias JA. Novel biomarkers in patients. Proc Natl Acad Sci USA 2007;104:5115–5120 asthma: chemokines and chitinase-like proteins. Curr Opin Allergy Clin Immunol 2009;9:60–66 24. Amrani A, Verdaguer J, Thiessen S, Bou S, Santamaria P. IL-1alpha, IL-1beta, and IFN-gamma mark beta cells for Fas-dependent destruc- 41. Saeki H, Tamaki K. Thymus and activation regulated chemokine (TARC)/ – tion by diabetogenic CD4(+) T lymphocytes. J Clin Invest 2000;105: CCL17 and skin diseases. J Dermatol Sci 2006;43:75 84 459–468 42. Romagnani S. Cytokines and chemoattractants in allergic inflammation. – 25. Phillips JM, Parish NM, Raine T, et al. Type 1 diabetes development Mol Immunol 2002;38:881 885 requires both CD4+ and CD8+ T cells and can be reversed by non- 43. Montane J, Bischoff L, Soukhatcheva G, et al. Prevention of murine au- depleting antibodies targeting both T cell populations. Rev Diabet Stud toimmune diabetes by CCL22-mediated Treg recruitment to the pancreatic 2009;6:97–103 islets. J Clin Invest 2011;121:3024–3028 diabetes.diabetesjournals.org Salem and Associates 975

44. Aso Y, Matsuura H, Momobayashi A, et al. Profound reduction in 47. Koulmanda M, Bhasin M, Fan Z, et al. Alpha 1-antitrypsin reduces in- T-helper (Th) 1 lymphocytes in peripheral blood from patients with flammation and enhances mouse pancreatic islet transplant survival. Proc concurrent type 1 diabetes and Graves’ disease. Endocr J 2006;53: Natl Acad Sci USA 2012;109:15443–15448 377–385 48. Zhang B, Lu Y, Campbell-Thompson M, et al. Alpha1-antitrypsin protects 45. Gupta SK, Haigh BJ, Griffin FJ, Wheeler TT. The mammalian secreted beta-cells from apoptosis. Diabetes 2007;56:1316–1323 RNases: mechanisms of action in host defence. Innate Immun 2013;19: 49. Koulmanda M, Bhasin M, Hoffman L, et al. Curative and beta cell regen- 86–97 erative effects of alpha1-antitrypsin treatment in autoimmune diabetic NOD 46. Bosco D, Meda P, Morel P, et al. Expression and secretion of alpha1- mice. Proc Natl Acad Sci USA 2008;105:16242–16247 proteinase inhibitor are regulated by proinflammatory cytokines in human 50. Lisowska-Myjak B. AAT as a diagnostic tool. Clin Chim Acta 2005; pancreatic islet cells. Diabetologia 2005;48:1523–1533 352:1–13