Diabetes Volume 64, October 2015 3425

Daniela Strodthoff,1,2 Zuheng Ma,3 Tina Wirström,3 Rona J. Strawbridge,4 Daniel F.J. Ketelhuth,1 David Engel,1 Robert Clarke,5 Sture Falkmer,6 Anders Hamsten,4 Göran K. Hansson,1 Anneli Björklund,3 and Anna M. Lundberg 1

Toll-Like 3 Influences Glucose Homeostasis and b-Cell Insulin Secretion

Diabetes 2015;64:3425–3438 | DOI: 10.2337/db14-0838

Toll-like receptors (TLRs) have been implicated in the pathogenesis characterized by the progressive development pathogenesis of type 2 diabetes. We examined the of hyperglycemia, dyslipidemia, and impaired insulin secretion function of TLR3 in glucose metabolism and type 2 from pancreatic b-cells (1). Experimental and clinical studies diabetes–related phenotypes in animals and humans. have demonstrated that inflammation mediates its patho- TLR3 is highly expressed in the , suggesting physiology, with Toll-like receptors (TLRs) playing critical that it can influence metabolism. Using a diet-induced functions (2). fi obesity model, we show that TLR3-de cient mice had In addition to the recognition of pathogen-derived enhanced glycemic control, facilitated by elevated insu- structures, TLRs can activate innate immunity by recog- Tlr32/2 lin secretion. Despite having high insulin levels, nition of endogenous molecules, such as lipids, fatty acids mice did not experience disturbances in whole-body in- (FAs), and other mediators, that are elevated during tissue METABOLISM sulin sensitivity, suggesting that they have a robust met- stress and cell death in chronic inflammatory diseases (3). abolic system that manages increased insulin secretion. Certain TLRs have been implicated in glucose metabolism Increase in insulin secretion was associated with upre- — fi gulation of islet glucose phosphorylation as well as exo- and type 2 diabetes TLR2- and TLR4-de cient mice are cytotic VAMP-2 in Tlr32/2 islets. TLR3 deficiency protected against diet-induced obesity and insulin resis- – also modified the plasma lipid profile, decreasing VLDL tance (4 9), and polymorphisms in human TLR4 have levels due to decreased triglyceride biosynthesis. More- been linked to a lower risk of type 2 diabetes (10,11). over, a meta-analysis of two healthy human populations TLRs are expressed predominantly in immune cells, showed that a missense single nucleotide polymorphism but their expression in nonhematopoietic cells suggests in TLR3 (encoding L412F) was linked to elevated insulin that they have “nonimmune” functions. Notably, TLR3, levels, consistent with our experimental findings. In conclu- which is activated by viral double-stranded RNA (dsRNA) sion, our results increase the understanding of the function and mRNA from dying cells, is highly expressed in pan- of innate receptors in metabolic disorders and implicate creatic b-cells (12,13). Further, infectious agents, such as TLR3 as a key control system in metabolic regulation. dsRNA, that bind to TLR3 accelerate b-cell dysfunction and apoptosis, causing insulitis and autoimmunity (14). Although TLR3 has been examined in type 1 diabetes, its The prevalence of obesity and type 2 diabetes is radically role in autoimmune diabetes remains unclear (15–18). rising worldwide, causing serious socioeconomic problems. Whether TLR3 has a role in obesity and the development Type 2 diabetes is a heterogeneous disorder with a complex of type 2 diabetes is unknown. In this study, we examined

1Cardiovascular Research Unit, Center for Molecular Medicine, Department of 6Laboratory of Pathology and Clinical Cytology, Ryhov Hospital, Jönköping, Medicine, Karolinska Institutet, Stockholm, Sweden Sweden 2Metabolism Unit, Department of Medicine, and Department of Endocrinology, Corresponding author: Daniela Strodthoff, [email protected]. Metabolism and Diabetes, Karolinska Institutet at Karolinska University Hospital Received 26 May 2014 and accepted 17 April 2015. Huddinge, Karolinska Institutet/AstraZeneca Integrated Cardio Metabolic Center and Center for Innovative Medicine, NOVUM, Stockholm, Sweden This article contains Supplementary Data online at http://diabetes 3Endocrinology and Diabetes Unit, Department of Molecular Medicine and .diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0838/-/DC1. Surgery, Karolinska Institutet, Karolinska University Hospital Solna, Stock- © 2015 by the American Diabetes Association. Readers may use this article as holm, Sweden long as the work is properly cited, the use is educational and not for profit, and 4Atherosclerosis Research Unit, Center for Molecular Medicine, Department the work is not altered. of Medicine Solna, Karolinska Institutet, Karolinska University Hospital Solna, See accompanying article, p. 3345. Stockholm, Sweden 5Clinical Trial Service Unit and Epidemiological Studies Unit, University of Oxford, Oxford, U.K. 3426 Role of TLR3 in Glucometabolic Regulation Diabetes Volume 64, October 2015

TLR3 function in glucose metabolism and type 2 diabetes– (Crystal Chem Inc., Downers Grove, IL), adiponectin (R&D related phenotypes in animals and humans. Systems, Minneapolis, MN), leptin (PeproTech, Rocky Hill, NJ), and serum amyloid A (SAA) (Life Technologies, RESEARCH DESIGN AND METHODS Carlsbad, CA) were measured by ELISA per the manufac- Animal Studies turers’ instructions. Plasma TG and cholesterol lipopro- All mice were housed and used as specified by the Stockholm tein profiles were examined as previously reported (21). North Committee for Experimental Animal Ethics and the Lipoprotein Biosynthesis Swedish National Board for Laboratory Animals. The hyper- Mice were fed the HFD for 34 weeks and fasted 8 h before glycemic clamp technique was performed at the National the experiment. VLDL synthesis was assessed after irrevers- Mouse Metabolic Phenotyping Center, University of ibly blocking VLDL catabolism using 10% (w/v) tyloxapol Massachusetts Medical School, with approval by the local (500 mg/kg i.v.; Sigma-Aldrich).Micewerebledfromthetail Institutional Animal Care and Use Committee (Worcester, into EDTA-coated tubes at the indicated times. Plasma TGs MA). Male C57BL/6 (Charles River Laboratories, Sulzfeld, 2 2 were measured as described above, normalized to baseline Germany) and Tlr3 / mice fully backcrossed (N14) onto TGs, and expressed as fold increase. C57BL/6 (19,20) were weaned at 4 weeks and fed a high- fat diet (HFD) (34.9% fat; Altromin, Lage, Germany) or Insulin Signaling normal chow diet from age 5 weeks for 20 weeks, unless Insulin signaling in the liver was assessed by Western otherwise stated. blot analysis of basal phosphorylated protein kinase B (PKB)/AKT in TLR3-deficient and control mice. In a sep- Glucose, Insulin, and Pyruvate Tolerance Test arate experiment, 10 min before freezing the tissue in liq- Intraperitoneal glucose tolerance test (ipGTT), intraperitoneal uid nitrogen, mice were injected with insulin (2 units/kg insulin tolerance test (ipITT), and intraperitoneal pyruvate body weight; Actrapid; Novo Nordisk). Number of mice and tolerance test were performed in overnight-starved mice duration of HFD are indicated in the figure legends. (unless otherwise stated) by injecting glucose (1 g/kg body weight; B. Braun Melsungen, Melsungen, Germany), insulin Histopathology and Immunohistochemistry (0.8 units/kg body weight; Actrapid; Novo Nordisk, Bagsværd, Pancreatic tissue was dissected from the surrounding Denmark), or pyruvate (1.5 g/kg body weight; Sigma-Aldrich, tissues, and specimens from the corpus/cauda regions fi fi St. Louis, MO), respectively, and monitoring glucose concen- were xed in formalin and embedded in paraf n. Sections – m trations (Abbott Scandinavia AB,Solna,Sweden)intail-vein (4 10 m) were stained with hematoxylin and eosin, insulin, blood. HOMA-insulin resistance (IR) was calculated as fasting glucagon, and somatostatin by immunohistochemistry, as glucose (mg/dL) 3 fasting insulin (mU/L)/405. described (22). The average islet area was calculated per total pancreatic area in three sections per sample. The detection of Hyperglycemic Clamp and Insulin Clearance apoptotic cells was performed by TUNEL staining technique After the HFD for 26 weeks, a survival surgery was using an In Situ Cell Death Detection Kit (Roche, Mannheim, – performed at 4 5 days before clamp experiments to estab- Germany) per the manufacturer’s instructions. lish an indwelling catheter in the jugular vein. Mice were fasted overnight before the start of the experiment. A hy- Isolation, Incubation, and Perifusion of Pancreatic Islets perglycemic clamp was conducted in conscious mice, start- 2 2 Islets of Langerhans were isolated from Tlr3 / and con- ing with an infusion of 20% dextrose to quickly reach trol mice by collagenase digestion in Hanks’ balanced salt a target hyperglycemia (;300 mg/dL glucose level) and solution, followed by sedimentation, and then cultured for maintain hyperglycemia by adjusting glucose infusion 20–24 h as described (23). After culture, equal-sized islets rates. Plasma samples were collected before the start of infusion (baseline) and at indicated time points to measure were preincubated as previously reported (23), followed by 60-min batch incubation (3 islets per tube in triplicate, glucose, insulin, and C-peptide levels. Insulin clearance was 2.8 or 16.7 mmol/L glucose), perifusion, RNA extraction, estimated by the ratio of fasted C-peptide to insulin. At the or Western blot. Islets from the batch incubations were end of the clamps, mice were killed. treated with acid-ethanol to extract cellular insulin (23). Tissue Processing Secreted insulin and islet insulin content were analyzed 125 Mice were killed with CO2 after overnight starvation, and by RIA using I-labeled insulin and anti-porcine insulin blood was collected for plasma analysis. Organs were dissected (Endocrinology and Diabetes Unit, Karolinska University Hos- after vascular perfusion with sterile RNase-free PBS, unless pital Solna, Stockholm, Sweden (24)). Perifusion experiments otherwise indicated. Tissues were snap-frozen for RNA and were performed as described (23). Perfusate samples were protein analysis or processed for immunohistochemistry. collected every minute, and secreted insulin was analyzed Blood and Plasma Analysis by radioimmunoassay. Whole blood was analyzed by a scil Vet abc hemocounter. Polyinosinic-Polycytidylic Acid Treatment Total plasma triglycerides (TGs) and cholesterol were Polyinosinic-polycytidylic acid (polyI:C) was administered measured using a colorimetric per the manufacturer’s three times per week simultaneously with the HFD in young instructions (Randox Laboratories Ltd., Crumlin, U.K.). Insulin C57BL/6 mice (age 6 weeks). Mice received 12 intraperitoneal diabetes.diabetesjournals.org Strodthoff and Associates 3427 injections of polyI:C (100 mg/mouse, InvivoGen, Toulouse, SNP&SEQ Technology Platform, Uppsala University, France) or NaCl (Fresenius Kabi, Uppsala, Sweden). Chow- Uppsala, Sweden, or Centre National de Génotypage, fed C57BL/6 mice (age 18 weeks) were treated 17 times with Paris, France (Supplementary Table 1 reports the clinical 2 2 polyI:C (100 mg/mouse) or NaCl. Tlr3 / mice received 11 characteristics and genotyping methods). SNPper (28) was intraperitoneal injections of polyI:C or NaCl during HFD. To used to identify single nucleotide polymorphisms (SNPs) in assess insulin secretion, C57BL/6 mice (age 9 weeks) re- the TLR3, of which 24 were available in the Polca and Olivia ceived 17 intraperitoneal injections of polyI:C or NaCl dur- data sets. Because rs3775291 was the only SNP in the cod- ing HFD. ing region, we focused on this. The association between rs3775291 and metabolic phenotypes was examined. Real-Time PCR Total RNA was isolated using the RNeasy Lipid Tissue Statistical Analysis Mini Kit (Qiagen, Hilden, Germany) or RNeasy Micro Kit Results from the animal studies are expressed as mean 6 (Qiagen) and reverse-transcribed with Superscript II, random SEM. The Mann-Whitney U test was used for group com- hexamers (pdN6), and RNasin (Life Technologies). RNA parisons, and the paired t test of log-transformed values concentration was measured by spectrophotometry (Thermo was used for pairwise observations. Associations between Scientific), and RNA quality was assessed on a Bioanalyzer rs3775291 and metabolic phenotypes in the human (Agilent Technologies, Waldbronn, Germany). cDNA was cohorts were analyzed by linear regression. Skewed varia- amplified on an ABI 7900HT (Applied Biosystems, Foster City, bles were natural log-transformed before analysis. An ad- CA) by real-time PCR using primers and probes (Applied ditive genetic model was assumed, and adjustments were Biosystems) for the selected . Data were calculated as made for age and BMI. Fixed-effects inverse variance meta- -DDC 2 T,whereDDCT = DCT (sample) – DCT (calibrator = average analyses were performed using METAL software (29). A , fi CT values of all samples in each group) and DCT is the target P value of 0.05 indicated a signi cant association. CT minus the CT of Hprt. RESULTS Western Blot 2/2 Pooled islets from two to three mice per sample were used Tlr3 Mice Have Enhanced Glucose Tolerance for Western blot as described (25). and Increased Circulating Insulin Upon Glucose Stimulation For insulin-signaling analysis, a piece of liver was lysed 2/2 and homogenized in ice-cold buffer (137 mmol/L NaCl, After being fed the HFD or chow diet for 20 weeks, Tlr3 and control mice experienced increased body weight, which 2.7 mmol/L KCL, 1 mmol/L MgCl2, 1% Triton X-100, 10% glycerol, 20 mmol/L Tris [pH 7.8], 10 mmol/L NaF, 1 mmol/L did not differ between mice with either diet (Fig. 1A). To EDTA, 5 mmol/L sodium pyrophosphate, 0.5 mmol/L examine the effects of TLR3 deletion on glucose homeosta- sis, we performed ipGTT. TLR3-deficient mice had better Na3VO4,1mg/mL leupeptin, 0.2 mmol/L phenylmethyl sulfonyl fluoride, 1 mg/mL aprotinin, and 1 mmol/L responses to glucose versus control mice, irrespective of diet (Fig. 1B). Baseline glucose did not differ between microcystin). Homogenates rotated 30 min at 4°C before 2 2 control and Tlr3 / mice fed chow or the HFD (Fig. 1C). centrifugation at 12,000g for 15 min at 4°C. 2/2 Primary antibodies were used at the following dilu- Tlr3 mice fed the HFD had elevated circulating tions: mouse anti–VAMP-2, 1:10,000 (Synaptic Systems insulin under fasting conditions versus control mice GmbH, Goettingen, Germany); rabbit anti-glucokinase (Fig. 1D and E), which rose further after the glucose challenge (Fig. 1D). Similarly, incremental insulin was (Gck), 1:500 (Santa Cruz Biotechnology, Dallas, TX); 2/2 mouse anti–b-actin, 1:20,000 (Sigma-Aldrich); and rabbit increased in HFD-fed Tlr3 mice compared with con- – trol mice (Fig. 1F). Insulin responses did not differ be- anti caspase-3 and rabbit anti-cleaved caspase-3, both 2 2 tween chow-fed Tlr3 / and control mice (Fig. 1D and 1:1,000 (Cell Signaling Technology, Danvers, MA). After 2/2 incubation with horseradish peroxidase–conjugated anti- E). HOMA-IR increased in HFD-fed Tlr3 mice versus mouse or anti-rabbit, bands were visualized by chemilu- control mice (Fig. 1G). Circulating levels of adiponectin minescence (Pierce, Rockford, IL). and leptin (Supplementary Fig. 1A and B, respectively) were similar between groups. Human Studies To determine the effects of TLR3 deficiency on glucose Male participants from the Polca (n = 625) and Olivia (n = metabolism, the hyperglycemic clamp technique was 306) cohorts were included. Polca comprises healthy 50-year- performed in HFD-fed mice (Fig. 1H–J). Body weight 2 2 oldindividualswhowerefreeofcoronaryheartdiseaseand was similar in HFD-fed control and Tlr3 / mice, and recruited at random using a population registry (26), and plasma glucose rose quickly and plateaued at ;300 mg/dL Olivia includes healthy individuals aged 33–80 years, recruited during infusion in both genotypes (Fig. 1H). Glucose infu- 2 2 for the Precocious Coronary Artery Disease (PROCARDIS) sion rates were significantly higher in HFD-fed Tlr3 / 2 2 study (27). Subjects with type 2 diabetes (defined as diagno- versus control mice (Fig. 1I). Insulin secretion in Tlr3 / sis, use of antidiabetic medication, or fasting glucose $7 mice climbed, as evidenced by elevated plasma insulin mmol/L) were excluded. All participants were genotyped compared with control mice (Fig. 1J). Further, plasma 2 2 using the Illumina Infinium 1M or 610K platforms at the C-peptide levels increased slightly during clamping in Tlr3 / 3428 Role of TLR3 in Glucometabolic Regulation Diabetes Volume 64, October 2015

Figure 1—Lack of TLR3 improves response to glucose and increases circulating insulin after glucose stimulation. A: Body weight was measured in control and Tlr32/2 mice fed chow or HFD. B: ipGTT was performed by injecting glucose (1 g/kg i.p.), and glucose concen- tration was analyzed in blood from the tail vein. C: Baseline glucose. Levels of circulating insulin (D) and baseline insulin (E) were measured by ELISA. F: Incremental insulin was calculated by subtracting basal insulin. G: HOMA-IR. Hyperglycemic clamps were performed on mice fed an HFD. H: Plasma glucose. I: Glucose infusion rate. J: Plasma insulin. K: C-peptide. L: Estimated insulin clearance calculated as fasted C-peptide-to-insulin ratio. Number of animals included: Body weight and ipGTT, n =21–26 each genotype; insulin release during ipGTT chow-fed mice, n = 6 + 6; HFD-fed mice, n =18–19; HOMA-IR, n =6–8 each genotype; hyperglycemic clamps and insulin clearance, n =8– 12 each genotype. Mice were fed a diet for 20 weeks except for hyperglycemic clamps (26 weeks). Areas under the curve (AUC) for glucose, insulin kinetic, and hyperglycemic clamps were calculated. Data are means 6 SEM. **P < 0.01; ***P < 0.001; ****P # 0.0001. n.s., not significant. diabetes.diabetesjournals.org Strodthoff and Associates 3429 mice (Fig. 1K), reflecting greater insulin production. The for 20 weeks. To rule out the influence of starvation, ipITT C-peptide-to-insulin ratio did not differ between the geno- was repeated in freely HFD-fed mice, yielding the same types, indicating that insulin clearance was similar be- result (Fig. 2C). Aged mice that were fed the HFD for tween the groups (Fig. 1L). 12 months still responded better to the glucose challenge (Fig. 2D) and had increased insulin levels upon glucose Tlr32/2 and Control Mice Have Similar Responses to stimulation (Fig. 2E), without becoming insulin resistant Insulin During ipITT and Display No Impairment of (Fig. 2F). Despite the rise in the HOMA-IR index, the in- 2 2 Hepatic Insulin Sensitivity sulin responses did not differ between Tlr3 / and control 2 2 By ipITT, insulin responses did not differ between Tlr3 / mice by ipITT, indicating that the increased insulin in 2 2 and control mice fed chow (Fig. 2A)ortheHFD(Fig.2B) Tlr3 / mice did not affect whole-body insulin resistance,

Figure 2—Control and Tlr32/2 mice respond similarly to insulin and show no impairment of hepatic insulin sensitivity. ipITT was performed on overnight-starved mice (except C, which were not starved) by injecting insulin (0.8 units/kg body weight i.p.), and glucose concentration was analyzed in blood from the tail vein. Mice were fed chow (A)andHFD(B and C)for20weeks.D: ipGTT was performed on aged mice, fed the HFD for 12 months, by injecting glucose (1 g/kg i.p.), and glucose concentration was analyzed in blood from the tail vein. E: Levels of circulating insulin were measured by ELISA. F: ipITT on aged mice (13 months old). G: Basal insulin signaling as revealed by Western blot analysis of phosphorylated (p)AKT in liver. Age at HFD start was 4–5 weeks. HFD duration was 20 weeks. H: Insulin stimulated (2 units/kg body weight) AKT phosphorylation. AgeatHFDstartwas14–16 weeks. HFD duration was 18 weeks. G and H: Graphs are shown as AKT phosphorylation normalized to total AKT. Gluconeogenesis was assessed by analyzing PEPCK mRNA levels normalized to Hprt (I) and by intraperitoneal pyruvate tolerance test in liver of mice (J) fed an HFD for 20 and 15 weeks, respectively. Diet and number of animals included in each experiment: chow diet, n =15–20 (A); HFD, n = 18 (B); freely fed, n = 11 each genotype (C); aged HFD-fed, n =4–5eachgenotype(D–F); HFD, n =6+6(G); HFD, n =3+4eachgenotype(H); HFD, n =18pergenotype(I); HFD, n =9pergenotype(J). The graphs for ipITT are shown as the percentage reduced from baseline glucose. Areas under the curve (AUC) for glucose kinetic and circulating insulin were calculated. Data are means 6 SEM. *P < 0.05; **P < 0.01; ***P < 0.001. 3430 Role of TLR3 in Glucometabolic Regulation Diabetes Volume 64, October 2015 irrespective of age or metabolic conditions. Liver-specific in- In examining the factors influencing lipoprotein levels, we sulin sensitivity was assessed by analyzing insulin signal- found that mRNA expression of 3-hydroxy-3-methyl-glutaryl-CoA ing (Fig. 2G and H) and gluconeogenesis (Fig. 2I and J). reductase (HMGCR), the rate-limiting enzyme in choles- Basal phosphorylated AKT was reduced in TLR3-deficient terol biosynthesis, was significantly reduced in the liver of 2 2 mice compared with control mice fed the HFD (Fig. 2G). Tlr3 / mice (Fig. 4F). Further, sortilin-1, which has been Insulin-stimulated AKT activation was similar between the implicated in the regulation of VLDL catabolism, rose in 2 2 genotypes (Fig. 2H). Reduced phosphoenolpyruvate the liver of Tlr3 / mice (Supplementary Fig. 2A). By carboxykinase (PEPCK) mRNA in TLR3-deficient mice inhibiting lipoprotein lipase–dependent TG hydrolysis 2 2 (Fig. 2I) was accompanied by a trend toward reduced with tyloxapol, we show that Tlr3 / mice had significantly glucose output from the liver compared with control lower TG synthesis (Fig. 4G). Interestingly, circulating TG mice (Fig. 2J), demonstrating the ability of insulin to was raised significantly upon TLR3 stimulation with the suppress glucose production in TLR3-deficient mice. TLR3 ligand polyI:C (Supplementary Fig. 2B).

Systemic Inflammation Was Increased in Tlr32/2 Mice Increased Insulin Secretion in Pancreatic Islets From fi but Not Locally in the Pancreas TLR3-De cient Mice 2/2 Systemic inflammation was assessed by analyzing periph- After glucose challenge, islets of Langerhans from Tlr3 fi eral blood cells and levels of SAA protein. Whereas total mice secreted signi cantly more insulin than control mice blood lymphocyte, granulocyte, and monocyte numbers (Fig. 5A), but islet insulin content was unchanged (Fig. did not differ between the genotypes (Fig. 3A), TLR3- 5B). Pancreatic islet structure was similar between TLR3- fi deficient mice fed the HFD displayed increased levels of de cient and control mice, wherein insulin-producing cells serum SAA (Fig. 3B), indicating augmented systemic inflam- were central and glucagon- and somatostatin-expressing mation despite the lack of TLR3. cells lay in the periphery (Fig. 5C). Further, islet morphom- Histopathological analysis of the pancreas showed no etry showed the islet area was similar between groups difference in the basic structure of the Langerhans cells (Fig. 5D). 2 2 between control and Tlr3 / mice (Fig. 5C). Thus, the Because TLR3 stimulation with synthetic dsRNA might size, number, and shape of islets were in the normal increase cell death, we performed the TUNEL assay (Fig. range. No bleeding, necrosis, or infiltration of granulo- 5E) and analyzed cleaved caspase-3 protein (Fig. 5F) but cytes could be detected. could not detect increased islet apoptosis. Glucose is the principal signal for insulin release from Tlr32/2 Mice Have Reduced Circulating VLDL b-cells. To determine the mechanism of increased insulin 2 2 Dyslipidemia is a common comorbidity in patients with secretion in Tlr3 / mice, we measured Gck protein expres- obesity and type 2 diabetes, characterized by increased FA sion, the rate-limiting component of glucose metabolism in influx into the liver and higher plasma TG. Despite no b-cells that phosphorylates glucose to glucose-6-phosphate 2 2 2 2 difference in weight between control and Tlr3 / mice, (Fig. 6A). Islets in Tlr3 / mice expressed more Gck com- the latter had decreased circulating TG (Fig. 4A). This was pared with control mice fed the HFD, indicating improved attributed to reduced TG concentrations in the VLDL glucose metabolism in the islets, which could affect greater fraction (Fig. 4B). Similarly, total cholesterol was lower insulin secretion. Furthermore, we measured mRNA levels 2 2 overall and in the VLDL fraction in Tlr3 / mice (Fig. 4C of GLUTs in pancreatic islets and peripheral tissues (Fig. 6B 2 2 and D). Moreover, Tlr3 / mice had decreased circulating and Supplementary Fig. 3A–C). GLUT2 did not differ sig- 2 2 free FAs (FFAs), the products of VLDL-TG hydrolysis, ver- nificantly between Tlr3 / and control mice in the islets or sus control mice (Fig. 4E). No differences in LDL, HDL, TG, liver, and GLUT4 expression was similar in adipose tissue and cholesterol were observed (Fig. 4B and D). and muscle.

Figure 3—Systemic inflammation was increased in Tlr32/2 mice, but not locally, in the pancreas. Blood analysis of total lymphocytes, granulocytes, and monocytes per white blood cells (WBC) (A) and SAA in mice fed an HFD for 20 weeks (B). Total blood count, n =8–18 per genotype; for SAA, n =15–17 per genotype. *P < 0.05. diabetes.diabetesjournals.org Strodthoff and Associates 3431

Figure 4—Tlr32/2 mice have reduced TG in VLDL due to decreased TG biosynthesis. Control and Tlr32/2 mice fed an HFD were analyzed for plasma levels of TG (A) and cholesterol (C). Size analysis of lipoprotein profile was performed on plasma, and the concentrations of TG (B) and cholesterol (D) were measured in each fraction and plotted against the retention fraction number. Two to three plasma samples were pooled per genotype, and a total of five pools per group were analyzed. Mean profiles are shown. Areas under the curve were calculated for each fraction of VLDL, LDL, and HDL. FFA (n =31–32 each genotype) were measured in plasma (E), and the relative mRNA level of HMGCR (n = 18 each genotype) in liver was analyzed and normalized to Hprt (F). G: Newly synthesized VLDL-TG was assessed in HFD-fed (34-week) mice by intravenous injection of tyloxapol. Graph is shown as the fold-increase ratio of newly synthesized VLDL-TG normalized to baseline TG (n = 6 each genotype). Data are means 6 SEM. *P < 0.05; **P < 0.01; ***P < 0.001. OD, optical density.

TLR3 Deficiency Increases Insulin Secretion on control mice (Fig. 7A, D, E, and H), indicating that the rise in Glucose and K+ Stimulation, Which is Reversed by insulin was not restricted to stimulation by glucose. TLR3 Stimulation The fusion of insulin-containing vesicles with the b-cell Next, we examined glucose-stimulated insulin secretion membrane and subsequent release of insulin constitute 2/2 from islets in vitro by perifusion of islets from Tlr3 and the last crucial step of glucose-stimulated insulin secre- control mice. Stimulation with high glucose (16.7 mmol/L) tion. By Western blot, the exocytotic protein VAMP-2 was 2 2 induced biphasic insulin secretion from chow-fed, under upregulated in the pancreatic islets of Tlr3 / mice (Fig. physiological conditions, (Fig. 7A)andHFD-fed(Fig.7E) 7I), which we speculate facilitates enhanced insulin secre- 2 2 control and TLR3-deficient mice. The enhancement in tion in Tlr3 / islets. 2 2 Tlr3 / mice involved the first and second phases of insulin To examine TLR3 function in our model, we treated secretion in response to high glucose in chow-fed (Fig. 7A–C) C57BL/6 chow-fed (Fig. 7J and K) and HFD-fed mice and HFD-fed mice (Fig. 7E–G). In addition, depolarization (Fig. 7L–N) with the TLR3 ligand polyI:C and analyzed 2 2 with K+ increased insulin secretion in Tlr3 / islets versus its effects on glucose control and VAMP-2 expression. 3432 Role of TLR3 in Glucometabolic Regulation Diabetes Volume 64, October 2015

Figure 5—Tlr32/2 mice secrete more insulin from pancreatic islets. Insulin secretion (A) and insulin content (B) were measured in islets from control and Tlr32/2 mice fed an HFD. C: Representative immunohistochemical staining for insulin, glucagon, and somatostatin in the islets of Langerhans of control and Tlr32/2 mice fed an HFD. Scale bar = 100 mm. D: Average islet area was calculated per total pancreatic area on three separate pancreatic sections per sample. E: Detection of apoptosis by TUNEL staining was performed on pancreatic sections from control and Tlr32/2 mice fed an HFD. The circle indicates islet area, and duodenum villi were used as the control for positive staining (n =6–8 each genotype). Scale bar = 50 mm. F: Cleaved caspase-3 was assessed by Western blot and normalized to caspase-3 in mice fed an HFD for 18 weeks (n =5–6 each genotype). Data are means 6 SEM. ctr., control. *P < 0.05; ***P < 0.001.

2 2 PolyI:C reversed the glucose response in C57BL/6 mice function in Tlr3 / mice was due to enhanced insulin under physiological conditions (Fig. 7J) and on the HFD sensitivity and plasma lipids by reducing b-cell lipotoxic- (Fig. 7L). This impairment was accompanied by downregu- ity. Upon stimulation of islets in vitro with palmitate, 2 2 lation of VAMP-2 in pancreatic islets (Fig. 7K and M)and glucose-induced insulin secretion increased in Tlr3 / a trend toward reduced insulin secretion in polyI:C-treated versus control mice (data not shown). Previous treatment C57BL/6 mice fed the HFD (Fig. 7N). Treatment of TLR3- with palmitate inhibited insulin secretion and decreased deficient mice with polyI:C had no effect on glucose response islet insulin content (Supplementary Fig. 5A–C) in the (Supplementary Fig. 4A), baseline glucose (Supplementary control and knockout mice. Insulin secretion was inhibited 2 2 Fig. 4B), or body weight (Supplementary Fig. 4C). by 60% in Tlr3 / animals versus 26% (as a percent of no We next studied the function of lipotoxicity in TLR3- previous exposure to palmitate) in the control mice (Sup- deficient mice, determining whether the improved b-cell plementary Fig. 5A and B). Islet insulin content in the diabetes.diabetesjournals.org Strodthoff and Associates 3433

glucose. Insulin content in b-cells and islet architecture 2 2 and area were similar between Tlr3 / and control mice. 2 2 Enhanced insulin secretion in Tlr3 / mice was regis- tered during the first and second phases of insulin secretion in response to high glucose by perifusion, a phenomenon that was glucose dependent and induced by depolarization with high K+. VAMP-2 is a member of the v-SNARE complex of the exocytic apparatus and mediates insulin secretion from b-cells (34). Gck (proximal event) and VAMP-2 (distal event) were upregulated in the pancreatic islets of 2 2 Tlr3 / mice. Notably, VAMP-2 is downregulated in islets Figure 6—Lack of TLR3 increases Gck expression. A: Gck protein from patients with diabetes with attenuated glucose expression was analyzed in islets from control or Tlr32/2 mice fed responses (35), contrary to our murine model in which an HFD with Western blot and normalized to b-actin (n = 6 each increased VAMP-2 protein expression parallels the ame- genotype). B: Relative Glut2 mRNA levels in islets was analyzed in mice fed an HFD and normalized to Hprt (n =7–9 each genotype). liorated response to glucose. Further, stimulation of TLR3 Data are means 6 SEM. *P < 0.05. signaling by polyI:C impairs this response and decreases VAMP-2 expression, accompanied by decreased insulin secretion in mice fed the HFD and under physiological, chow-fed conditions. However, polyI:C treatment did not knockout fell to a similar extent as in the control on culture affect glucose response in TLR3-deficient mice, suggesting in palmitate (66% vs 62%), likely due to inhibition of pro- that TLR3 modulates the processes that effect increased insulin biosynthesis. insulin secretion in islets. Our perifusion data demonstrate that insulin release is A TLR3 Polymorphism Is Associated With Metabolic 2 2 enhanced in the Tlr3 / mice in response not only to Risk Factors in Humans elevated glucose but also to elevated potassium levels. To determine whether TLR3 is linked to metabolic This kind of potassium increase functions as an artificial phenotypes in humans, we investigated the effect of secretagogue that bypasses the regulation of secretion genetic variants in TLR3 on insulin-related traits. We through K+-ATP channels, which is the normal mecha- searched for all SNPs in TLR3 that could be identified by nism for control of glucose signaling. Because TLR3 de- SNPper (28) and had been previously genotyped in the ficiency enhanced insulin secretion in response also to population studies (Fig. 8). We found 24 SNPs (Fig. 8); potassium, TLR3 modulation of insulin secretion may however, only 1, rs3775291, was found within the coding operate on distal steps in the pathway that are not di- region. rs3775291 is predicted to damage TLR3 function rectly dependent on glucose. A glucose-specific effect is (30) and has been studied primarily with regard to TLR3 not excluded, however, because glucose is known to function in viral infection (31–33). We measured the effects influence insulin secretion also at steps distal to that of of rs3775291 on insulin-related traits. K+-ATP channels (36). A meta-analysis of rs3775291 in two healthy Swedish Dysfunctional lipid metabolism causes insulin resis- cohorts was performed, and its minor allele (T) was tance, reduced insulin secretion (lipotoxicity), and in rare associated with elevated fasting insulin levels (b = 0.048, cases, enhanced insulin secretion accompanied by hypo- SE = 0.023, P = 0.0340) but not blood glucose levels glycemia (37), which primed us to analyze the effect of (b = 20.005, SE = 0.005, P = 0.3365), consistent with 2/2 2 2 Tlr3 deficiency on islet lipid metabolism. We tested our findings in Tlr3 / mice. the effects of palmitate on insulin secretion to determine 2 2 whether b-cells in Tlr3 / mice are resistant to lipotox- DISCUSSION icity, thereby mitigating the damage of an HFD. Our Although TLRs have been implicated in type 2 diabetes, results did not support this hypothesis—the inhibitory 2 2 there is little evidence of their involvement, other than effects of palmitate were more extensive in Tlr3 / TLR2 and TLR4. In this study, we examined the function mice versus control islets. 2 2 of TLR3 in the progression of diet-induced obesity in Tlr3 / mice fed the HFD had elevated basal fasting a mouse model and its effects on insulin and type 2 insulin levels and HOMA-IR values, which could indicate diabetes–related phenotypes in humans. greater insulin resistance. However, the increased insulin Our data show that the absence of functional TLR3 response to intraperitoneally administered glucose was ac- protects against metabolic disturbances due to fat intake. companied by better glucose tolerance versus controls—a TLR3-deficient mice have enhanced glycemic control on typical primary effect on insulin secretion, not a secondary glucose challenge, which might be attributed to increased effect in response to insulin resistance. The primary effect circulating insulin, caused by amplified insulin secretion of TLR3 deficiency on insulin secretion was confirmed by 2 2 from b-cells in Tlr3 / mice, accelerating the response to the hyperglycemic clamp technique, which demonstrated 3434 Role of TLR3 in Glucometabolic Regulation Diabetes Volume 64, October 2015

Figure 7—Insulin release in perifusion of islets from Tlr32/2 and control mice. Islets of control and Tlr32/2 mice fed a chow diet (A–D)oran HFD (E–H) for 20 weeks were perifused with 3.3 mmol/L glucose (G), followed by 16.7 mmol/L glucose and 20 mmol/L K+. Samples of the perifused islets were collected every minute, and levels of secreted insulin were analyzed by radioimmunoassay. Areas under the curve (AUC) were calculated for the first phase (B and F) and the second phase (C and G) of insulin secretion after glucose stimulation (indicated by dashed line in A and E) and after K+ stimulation (D and H). Mean values from five mice per group are shown. I: Protein expression levels of VAMP-2 were analyzed in islets from control or Tlr32/2 mice fed an HFD with Western blot and normalized to b-actin. Samples were pooled, and each band represents islets pooled from two mice, with a total of four pools for each genotype. A representative Western blot is shown. J and K: C57BL/6 mice (age 18 weeks) fed a chow diet were treated with polyI:C (100 mg/mouse) or NaCl, respectively. J: ipGTT was performed and the AUC was calculated (n =5–6 each genotype). K: Protein expression of VAMP-2 was assessed by Western blot and normalized to b-actin. Samples were pooled, and each band represents islets pooled from two to three mice. L and M: PolyI:C and NaCl treatment were started in young mice (age 6 weeks) simultaneously with an HFD. L: ipGTT was performed and AUC were calculated (n =5–6 diabetes.diabetesjournals.org Strodthoff and Associates 3435

Figure 8—Genetic structure of human TLR3. A schematic diagram of the human TLR3 gene and the relative positions of the SNPs analyzed in this study. Linkage disequilibrium (calculated from Polca and Olivia) is given in r2 (gray and numbers).

enhanced insulin secretion at a fixed level of glycemia, was similar between the genotypes, indicating that a conclusion that was supported by the islet experiments. these effects are unlikely to be caused by impaired he- Notably, the in vitro islet perifusion experiments were per- patic insulin sensitivity. Furthermore, we observed 2 2 formed after identical culture times for control and Tlr3 / a trend toward reduced glucose output from the liver islets, minimizing residual secondary effects of the in of TLR3-deficient mice. This is most likely facilitated vivo environment. Collectively, the ipITT results indicate by the increased insulin, leading to suppression of 2 2 that the Tlr3 / mice do not become whole-body insulin PEPCK mRNA expression in the liver of TLR3-deficient resistant. The only indication was the HOMA-IR value, a crude mice. measure of resistance, the validity of which has been debated A further dissection of tissue-specific insulin resistance (38). To rule out tissue-specific insulin resistance, we assessed (e.g., in muscle and adipose tissue) might be advisable in hepatic insulin sensitivity by analyzing insulin signaling future studies to rule out a local effect of insulin in other and gluconeogenesis. Insulin-stimulated AKT activation peripheral tissues. Nevertheless, the systemic effects of

each genotype). M: Protein expression of VAMP-2 was assessed by Western blot and normalized to b-actin. Samples were pooled, and each band represents islets pooled from two to three mice. N: Insulin secretion is displayed as the percentage of insulin content in polyI:C- treated HFD-fed C57BL/6 mice. PolyI:C (100 mg/mouse) and NaCl treatment were started in 9-week-old mice simultaneously with an HFD. Mice received a total of 17 injections (n =4–5 each genotype). Data are means 6 SEM. *P < 0.05; **P < 0.01. 3436 Role of TLR3 in Glucometabolic Regulation Diabetes Volume 64, October 2015 increased insulin under several conditions (physiological Our finding that TLR3 mediates glucose tolerance is chow-fed, HFD-fed, freely HFD-fed, and aged HFD-fed supported by Wu et al. (47), who demonstrated that TLR3 mice), and the results obtained from the hyperglycemic deficiency in mice improves glucose control. In our study, clamp in addition to our data on hepatic insulin sensitivity this improvement was mediated by elevated insulin secre- 2 2 and gluconeogenesis make it unlikely that the phenotype of tion from Tlr3 / islets, at least partly due to increased 2 2 Tlr3 / mice is caused by insulin resistance. Instead, the data Gck and VAMP-2 expression in pancreatic b-cells. 2 2 suggest that Tlr3 / mice have a robust metabolic system Last, we examined whether a well-characterized poly- with adaptive measures that enables them to cope with the morphism in human TLR3 (rs3775291) was associated increased insulin secretion over a prolonged time. with glucometabolic alterations, supporting our findings Lower circulating TG, cholesterol, and FFA concen- in mice. In a meta-analysis of two healthy male popula- 2 2 trations indicated that lipid metabolism in Tlr3 / mice tions, rs3775291 was associated with increased fasting was affected. The decline in TG content occurred in the insulin levels but not fasting glucose. rs3775291 is VLDL fraction and was due to reduced VLDL secretion, as a nonsynonymous SNP in the coding region of TLR3 that is revealed by inhibiting lipoprotein lipase-dependent lipo- predicted to impair protein function (30). How rs3775291 protein catabolism. In line with this, polyI:C decreased TG affects TLR3 function is unknown, but a recent report indi- 2 2 levels. Tlr3 / mice had higher mRNA levels for sortilin-1, catesthatitreducesTLR3ligandbindingandimpairs which regulates VLDL secretion; this may account for at TLR3-mediated cell activation (48). Our results show that least some of the effects on VLDL (39,40). its relationship with fasting insulin levels is consistent with 2 2 Consistent with this model, increased plasma insulin the predicted impairment in human TLR3 and the Tlr3 / concentrations inhibit hepatic VLDL-TG production (39), mouse model findings. Unfortunately, the human studies do and insulin-resistant obese patients experience greater not allow assessment of insulin kinetics; thus, we are not basal VLDL secretion (41). The lower circulating TG levels able to determine whether the increased insulin levels were 2 2 in Tlr3 / mice were accompanied by decreased FFAs, due to increased production or reduced clearance. which are associated with reduced insulin sensitivity Although TLRs have been shown to be involved in and insulin-stimulated glucose uptake (42). In addition, diabetes (17,49), the role of innate immunity in glucose prolonged elevation of FFAs impairs b-cell secretion. homeostasis and insulin secretion is largely unknown. In- TLR3 stimulation with a viral mimic accelerates the terestingly, Krus et al. (50) demonstrated that the com- development of type 1 diabetes in rats (43,44) and indu- plement regulatory protein, CD59, modulates exocytosis ces b-cell apoptosis through TLR3 (14) and Fas-associated machinery by directly interacting with syntaxin-1 and protein with death domain recruitment, leading to acti- VAMP-2, thereby regulating insulin secretion. Interest- vation of caspase-8. A dose-dependent effect governs the ingly, TLR3-deficient mice displayed increased Gck and outcome of TLR3 stimulation—higher doses of exogenous VAMP-2 protein expression. Whether TLR3 directly inter- TLR3 ligand induce insulitis and diabetes, and lower doses acts with exocytosis or whether it modulates prevent autoimmune diabetes (15,17,45). Thus, TLR3 glucose sensing via Gck remains to be determined. TLR3 likely mediates viral recognition and protects against virally may also act distally in other metabolic tissues and initi- induced type 1 diabetes. The endogenous TLR3 ligand that ate processes that lead to changes in the islets. This no- mediates the effects in our study, in the absence of viral tion is supported by the observation that TLR3 increases infection or exogenous stimulation, is unknown. TLR3 is hepatic production of VLDL and is also in line with pub- activated by mRNA from damaged cells, but we could not lished data indicating a role for TLR3 in liver regeneration detect islet cell death. Various studies have suggested that and hepatocyte proliferation (47,51). inflammation is a driving force in the development of meta- In summary, TLR3 deficiency affects glucose homeo- bolic disturbances. The surprising finding that systemic in- stasis, insulin secretion, and lipid metabolism, likely by flammation was increased in the TLR3-deficient mice is in modulating proximal (improved glycolysis via Gck) and line with findings from a study on where in- distal events (increased VAMP-2 expression). Further, the 2 2 filtration of macrophages in atherosclerotic lesions of Tlr3 / rs3775291 polymorphism in TLR3 influences glucose mice was observed (20). However, we could not detect any homeostasis in humans. signs of local inflammation in the islets. Notably, pancreatic tissue contains and secretes copious RNases; thus, a control system that involves TLR3, an innate Acknowledgments. The authors thank Richard Flavell, Yale University, New RNA receptor, to detect and respond to abnormalities in this Haven, CT, and Dr. Claudia Monaco, Imperial College, London, for generously pro- system is logical. Patients with diabetes have increased viding the mice, and Anneli Olsson, Ingrid Törnberg, Linda Haglund, Karolinska Institutet, Stockholm, Sweden, and Linda Johansson, Ryhov Hospital, Jönköping, concentrations of circulating nucleic acids, likely due to Sweden, for technical assistance. The authors thank Jason K. Kim and Dr. Dae inhibition of plasma RNase activity (46). Circulating Young Jung, University of Massachusetts, Worcester, MA, for performing the hyper- nucleases might prevent ligand-induced stimulation of glycemic clamp. The authors thank Valerie Romer, University of Massachusetts, inflammatory responses; thus, we speculate that this Worcester, MA, for coordinating hyperglycemic clamp experiments. The authors thank impairment in patients with diabetes generates endoge- Anna Krook and Dr. Mari Björnholm, Karolinska Institutet, Stockholm, Sweden, and nous ligands of TLR3. Dr. Myriam Aouadi, Karolinska Institutet, Huddinge, Sweden, for helpful discussions. diabetes.diabetesjournals.org Strodthoff and Associates 3437

Funding. This study was supported by the Swedish Research Council, project response to microbes, their products, and . J Immunol 2002;168: grants 4203 and 8691, and the Center of Excellence for Research on Inflamma- 554–561 tion and Cardiovascular disease Linnaeus support (8703), the Swedish Heart- 14. Dogusan Z, García M, Flamez D, et al. Double-stranded RNA induces Lung Foundation, Stockholm County Council, the Nanna Svartz Foundation, the pancreatic beta-cell apoptosis by activation of the toll-like receptor 3 and in- Fredrik and Ingrid Thuring Foundation, the Magnus Bergvall Foundation, terferon regulatory factor 3 pathways. Diabetes 2008;57:1236–1245 Karolinska Institutet Cardiovascular Program Career Development Grant, 15. Serreze DV, Hamaguchi K, Leiter EH. Immunostimulation circumvents di- the European Commission (AtheroRemo-201668, Athero-Flux-602222, and abetes in NOD/Lt mice. J Autoimmun 1989;2:759–776 LSHM-CT-2007-037273), the Knut and Alice Wallenberg Foundation, the Stra- 16. Moriyama H, Wen L, Abiru N, et al. Induction and acceleration of insulitis/ tegic Cardiovascular and Diabetes Programs of Karolinska Institutet and Stock- diabetes in mice with a viral mimic (polyinosinic-polycytidylic acid) and an insulin holm County Council, the Foundation for Strategic Research, the Stockholm self-peptide. Proc Natl Acad Sci U S A 2002;99:5539–5544 County Council and the Research Foundation of Ryhov County Hospital. 17. Wen L, Peng J, Li Z, Wong FS. The effect of innate immunity on autoim- Duality of Interest. No potential conflicts of interest relevant to this article mune diabetes and the expression of Toll-like receptors on pancreatic islets. were reported. J Immunol 2004;172:3173–3180 Author Contributions. D.S. and A.M.L. wrote the manuscript, performed 18. Wong FS, Hu C, Zhang L, et al. The role of Toll-like receptors 3 and 9 in the experiments, and analyzed data. Z.M., T.W., D.F.J.K., D.E., and S.F. performed development of autoimmune diabetes in NOD mice. Ann N Y Acad Sci 2008; experiments and analyzed data. R.J.S. and A.B. analyzed data. R.C., A.H., G.K.H., 1150:146–148 and A.B. contributed to discussion and critically edited the manuscript. All authors 19. Alexopoulou L, Holt AC, Medzhitov R, Flavell RA. Recognition of double- reviewed and approved the manuscript. D.S. is the guarantor of this work and, as stranded RNA and activation of NF-kappaB by Toll-like receptor 3. Nature 2001; such, had full access to all the data in the study and takes responsibility for the 413:732–738 integrity of the data and the accuracy of the data analysis. 20. Cole JE, Navin TJ, Cross AJ, et al. Unexpected protective role for Toll-like Previous Presentation. Parts of this work were presented as a poster at receptor 3 in the arterial wall. Proc Natl Acad Sci U S A 2011;108:2372–2377 the Keystone Symposia on Molecular and Cellular Biology Conference (D6: The 21. Zhang L, Ovchinnikova O, Jönsson A, et al. The tryptophan metabolite 3- Crossroads of Lipid Metabolism and Diabetes), Copenhagen, Denmark, 19–24 hydroxyanthranilic acid lowers plasma lipids and decreases atherosclerosis in April 2015. hypercholesterolaemic mice. Eur Heart J 2012;33:2025–2034 22. Ma Z, Lavebratt C, Almgren M, et al. Evidence for presence and functional References effects of Kv1.1 channels in b-cells: general survey and results from mceph/ 1. Kumar V, Abbas AK, Fausto N, Aster JC. Robbins and Cotran Pathologic mceph mice. PLoS One 2011;6:e18213 Basis of Disease. Philadelphia, Saunders Elsevier, 2009 23. Björklund A, Grill V. B-cell insensitivity in vitro: reversal by diazoxide entails more 2. Gregor MF, Hotamisligil GS. Inflammatory mechanisms in obesity. Annu Rev than one event in stimulus-secretion coupling. Endocrinology 1993;132:1319–1328 Immunol 2011;29:415–445 24. Herbert V, Lau KS, Gottlieb CW, Bleicher SJ. Coated charcoal immunoassay 3. Lundberg AM, Hansson GK. Innate immune signals in atherosclerosis. Clin of insulin. J Clin Endocrinol Metab 1965;25:1375–1384 Immunol 2010;134:5–24 25. Ma Z, Portwood N, Foss A, Grill V, Björklund A. Evidence that insulin se- 4. Shi H, Kokoeva MV, Inouye K, Tzameli I, Yin H, Flier JS. TLR4 links innate cretion influences SNAP-25 through proteasomal activation. Biochem Biophys immunity and fatty acid-induced insulin resistance. J Clin Invest 2006;116:3015– Res Commun 2005;329:1118–1126 3025 26. van ’t Hooft FM, Ruotolo G, Boquist S, de Faire U, Eggertsen G, Hamsten A. 5. Caricilli AM, Nascimento PH, Pauli JR, et al. Inhibition of toll-like receptor 2 Human evidence that the apolipoprotein a-II gene is implicated in visceral fat expression improves insulin sensitivity and signaling in muscle and white adipose accumulation and metabolism of triglyceride-rich lipoproteins. Circulation 2001; tissue of mice fed a high-fat diet. J Endocrinol 2008;199:399–406 104:1223–1228 6. Davis JE, Gabler NK, Walker-Daniels J, Spurlock ME. Tlr-4 deficiency se- 27. Farrall M, Green FR, Peden JF, et al. Genome-wide mapping of susceptibility lectively protects against obesity induced by diets high in saturated fat. Obesity to coronary artery disease identifies a novel replicated locus on 17. (Silver Spring) 2008;16:1248–1255 PLoS Genet 2006;2:e72 7. Ehses JA, Meier DT, Wueest S, et al. Toll-like receptor 2-deficient mice are 28. Riva A, Kohane IS. SNPper: retrieval and analysis of human SNPs. Bio- protected from insulin resistance and beta cell dysfunction induced by a high-fat informatics 2002;18:1681–1685 diet. Diabetologia 2010;53:1795–1806 29. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of 8. Himes RW, Smith CW. Tlr2 is critical for diet-induced metabolic syndrome in genomewide association scans. Bioinformatics 2010;26:2190–2191 a murine model. FASEB J 2010;24:731–739 30. Adzhubei IA, Schmidt S, Peshkin L, et al. A method and server for predicting 9. Kuo LH, Tsai PJ, Jiang MJ, et al. Toll-like receptor 2 deficiency improves damaging missense mutations. Nat Methods 2010;7:248–249 insulin sensitivity and hepatic insulin signalling in the mouse. Diabetologia 2011; 31. Kindberg E, Vene S, Mickiene A, Lundkvist Å, Lindquist L, Svensson L. 54:168–179 A functional Toll-like receptor 3 gene (TLR3) may be a risk factor for tick-borne 10. Kolek MJ, Carlquist JF, Muhlestein JB, et al. Toll-like receptor 4 gene encephalitis (TBEV) infection. J Infect Dis 2011;203:523–528 Asp299Gly polymorphism is associated with reductions in vascular inflammation, 32. Sironi M, Biasin M, Cagliani R, et al. A common polymorphism in TLR3 angiographic coronary artery disease, and clinical diabetes. Am Heart J 2004; confers natural resistance to HIV-1 infection. J Immunol 2012;188:818–823 148:1034–1040 33. Svensson A, Tunbäck P, Nordström I, Padyukov L, Eriksson K. Poly- 11. Steinhardt AP, Aranguren F, Tellechea ML, et al. A functional non- morphisms in Toll-like receptor 3 confer natural resistance to human herpes synonymous toll-like receptor 4 gene polymorphism is associated with metabolic simplex virus type 2 infection. J Gen Virol 2012;93:1717–1724 syndrome, surrogates of insulin resistance, and syndromes of lipid accumulation. 34. Regazzi R, Wollheim CB, Lang J, et al. VAMP-2 and cellubrevin are ex- Metabolism 2010;59:711–717 pressed in pancreatic beta-cells and are essential for Ca(2+)-but not for GTP 12. Albert ML, Pearce SF, Francisco LM, et al. Immature dendritic cells gamma S-induced insulin secretion. EMBO J 1995;14:2723–2730 phagocytose apoptotic cells via alphavbeta5 and CD36, and cross-present an- 35. Ostenson CG, Gaisano H, Sheu L, Tibell A, Bartfai T. Impaired gene and tigens to cytotoxic T lymphocytes. J Exp Med 1998;188:1359–1368 protein expression of exocytotic soluble N-ethylmaleimide attachment protein 13. Zarember KA, Godowski PJ. Tissue expression of human Toll-like re- receptor complex proteins in pancreatic islets of type 2 diabetic patients. ceptors and differential regulation of Toll-like receptor mRNAs in leukocytes in Diabetes 2006;55:435–440 3438 Role of TLR3 in Glucometabolic Regulation Diabetes Volume 64, October 2015

36. Henquin JC. Triggering and amplifying pathways of regulation of insulin 44. Sobel DO, Newsome J, Ewel CH, et al. Poly I:C induces development of secretion by glucose. Diabetes 2000;49:1751–1760 diabetes mellitus in BB rat. Diabetes 1992;41:515–520 37. Martins E, Cardoso ML, Rodrigues E, et al. Short-chain 3-hydroxyacyl-CoA 45. Sobel DO, Goyal D, Ahvazi B, et al. Low dose poly I:C prevents diabetes in dehydrogenase deficiency: the clinical relevance of an early diagnosis and report the diabetes prone BB rat. J Autoimmun 1998;11:343–352 of four new cases. J Inherit Metab Dis 2011;34:835–842 46. Kocic G, Bjelakovic G, Saranac Lj, et al. Altered degradation of circulating 38. Tripathy D, Almgren P, Tuomi T, Groop L. Contribution of insulin-stimulated nucleic acids and oligonucleotides in diabetic patients. Diabetes Res Clin Pract glucose uptake and basal hepatic insulin sensitivity to surrogate measures of 2008;79:204–213 insulin sensitivity. Diabetes Care 2004;27:2204–2210 47. Wu LH, Huang CC, Adhikarakunnathu S, et al. Loss of toll-like receptor 3 39. Chirieac DV, Chirieac LR, Corsetti JP, Cianci J, Sparks CE, Sparks JD. function improves glucose tolerance and reduces liver steatosis in obese mice. Glucose-stimulated insulin secretion suppresses hepatic triglyceride-rich lipoprotein Metabolism 2012;61:1633–1645 and apoB production. Am J Physiol Endocrinol Metab 2000;279:E1003–E1011 48. Zhou P, Fan L, Yu KD, Zhao MW, Li XX. Toll-like receptor 3 C1234T may 40. Chamberlain JM, O’Dell C, Sparks CE, Sparks JD. Insulin suppression of protect against geographic atrophy through decreased dsRNA binding capacity. apolipoprotein B in McArdle RH7777 cells involves increased sortilin 1 interaction FASEB J 2011;25:3489–3495 and lysosomal targeting. Biochem Biophys Res Commun 2013;430:66–71 49. Mohammad MK, Morran M, Slotterbeck B, et al. Dysregulated Toll-like 41. Sørensen LP, Søndergaard E, Nellemann B, Christiansen JS, Gormsen LC, receptor expression and signaling in bone marrow-derived macrophages at the Nielsen S. Increased VLDL-triglyceride secretion precedes impaired control of onset of diabetes in the non-obese diabetic mouse. Int Immunol 2006;18:1101– endogenous glucose production in obese, normoglycemic men. Diabetes 2011; 1113 60:2257–2264 50. Krus U, King BC, Nagaraj V, et al. The complement inhibitor CD59 reg- 42. Boden G, Chen X, Ruiz J, White JV, Rossetti L. Mechanisms of fatty acid- ulates insulin secretion by modulating exocytotic events. Cell Metab 2014;19: induced inhibition of glucose uptake. J Clin Invest 1994;93:2438–2446 883–890 43. Ewel CH, Sobel DO, Zeligs BJ, Bellanti JA. Poly I:C accelerates development 51. Zorde-Khvalevsky E, Abramovitch R, Barash H, et al. Toll-like receptor 3 of diabetes mellitus in diabetes-prone BB rat. Diabetes 1992;41:1016–1021 signaling attenuates liver regeneration. Hepatology 2009;50:198–206