Insulin infusion suppresses while glucose infusion induces Toll-like ...

2 downloads 0 Views 773KB Size Report
Feb 12, 2013 - and high-mobility group-B1 protein expression in mononuclear cells of type 1 diabetes ... while glucose infusion induces Toll-like receptors and high-mobility .... nine. In this assay system, the release of superoxide radical, as.
Am J Physiol Endocrinol Metab 304: E810–E818, 2013. First published February 12, 2013; doi:10.1152/ajpendo.00566.2012.

Insulin infusion suppresses while glucose infusion induces Toll-like receptors and high-mobility group-B1 protein expression in mononuclear cells of type 1 diabetes patients Paresh Dandona, Husam Ghanim, Kelly Green, Chang Ling Sia, Sanaa Abuaysheh, Nitesh Kuhadiya, Manav Batra, Sandeep Dhindsa, and Ajay Chaudhuri Division of Endocrinology, Diabetes, and Metabolism, State University of New York at Buffalo and Kaleida Health, Buffalo, New York Submitted 19 November 2012; accepted in final form 17 January 2013

Dandona P, Ghanim H, Green K, Sia CL, Abuaysheh S, Kuhadiya N, Batra M, Dhindsa S, Chaudhuri A. Insulin infusion suppresses while glucose infusion induces Toll-like receptors and high-mobility group-B1 protein expression in mononuclear cells of type 1 diabetes patients. Am J Physiol Endocrinol Metab 304: E810 –E818, 2013. First published February 12, 2013; doi:10.1152/ajpendo.00566.2012.—The purpose of this study was to determine whether an insulin infusion exerts an anti-inflammatory effect and whether the infusion of small amounts of glucose results in oxidative and inflammatory stress in patients with type 1 diabetes. Ten patients with type 1 diabetes were infused with either 2 U/h of insulin with 100 ml 5% dextrose/h to or just dextrose (100 ml/h) or physiological saline (100 ml/h) for 4 h after an overnight fast on three separate days. Blood samples were collected at 0, 2, 4, and 6 h. Insulin with glucose infusion led to the maintenance of euglycemia and a significant suppression of reactive oxygen species (ROS) generation, p47phox expression, Toll-like receptor (TLR)-4, TLR-2, TLR-1, CD14, high-mobility group-B1 (HMGB1), p38 mitogen-activated protein (MAP) kinase, c-Jun NH2terminal kinase (JNK)-1, and platelet/endothelial cell adhesion molecule expression and a fall in serum concentrations of C-reactive protein, HMGB1, and rapid upon activation T cell expressed and secreted. Glucose infusion led to an increase in plasma glucose concentration from 115 (fasting) to 215 (at 4 and 6 h) mg/dl and to an increase in ROS generation, the expression of TLR-4, TLR-2, TLR-1, HMGB1, p38 MAP kinase, and JNK-1, and plasma concentrations of HMGB1. While insulin reduces indexes of oxidative and inflammatory stress in patients with type 1 diabetes, even small amounts of glucose (20 g over 4 h) induce oxidative and inflammatory stress. These effects are reflected in TLR, p38 MAP kinase, and HMGB1 expression. The induction of significant oxidative and inflammatory stress by small amounts of glucose in patients with type 1 diabetes may have important pathophysiological and therapeutic implications. inflammation WE HAVE PREVIOUSLY DEMONSTRATED the anti-inflammatory effect of insulin both in vitro and in vivo (1, 7). Thus, insulin at physiologically relevant concentrations suppresses the key proinflammatory transcription factor nuclear factor-␬B (NF-␬B), as well as the expression of the adhesion molecule intercellular adhesion molecule (ICAM)-1 and the chemokine monocyte chemotactic protein (MCP-1) in human aortic endothelial cells in vitro. Intravenously infused insulin at a rate of 2 U/h, which doubles the circulating insulin concentrations in vivo, has also been shown to suppress intranuclear NF-␬B binding in peripheral blood mononuclear cells (MNC) along with the cellular

Address for reprint requests and other correspondence: P. Dandona, State Univ. of New York at Buffalo, 3 Gates Circle, Buffalo, NY 14209 (e-mail: [email protected]). E810

expression of chemokines and their receptors and matrix metalloproteinases. In addition, intravenously delivered insulin also suppresses the expression of Egr-1, another proinflammatory transcription factor, and the plasma concentrations of ICAM-1, MCP-1, rapid upon activation T cell expressed and secreted (RANTES, CCL-5), tissue factor and plasminogen activator inhibitor-1 (18, 25). These experiments were all conducted in patients with either obesity or type 2 diabetes but not patients with type 1 diabetes. We have recently demonstrated that insulin suppresses the expression of Toll-like receptors (TLRs)-1, -2, -4, -7, and -9 in the MNC of patients with type 2 diabetes as a part of its comprehensive anti-inflammatory effect (15). These observations have important implications for various inflammatory processes, including infections, metabolic inflammation, and atherosclerosis, since TLRs may be involved in all these processes. In view of the fact that type 1 diabetes patients are already on treatment with subcutaneously injected insulin and the fact that they often suffer from infections and other inflammatory conditions (28, 31), we thought it important to specifically investigate the hypothesis that intravenously infused insulin exerts an anti-inflammatory and a TLR-suppressive effect in this population. This is relevant, since subcutaneously injected insulin does not exert an anti-inflammatory effect independently of a fall in glucose concentrations (34); it is possible that the plasma concentrations achieved after subcutaneous injections of insulin are not sufficient to exert this effect. Because intravenous insulin infusions result in an increase in insulin concentrations several times higher than those at baseline, they have to be carried out with concomitant glucose infusion to maintain euglycemia. Thus, similar amounts of glucose have to be infused as controls on another day. When carrying out these control infusions with small amounts of glucose in patients with type 1 diabetes, we noticed that blood glucose concentrations increased significantly. This was unlike what had happened in the obese and type 2 diabetes patients infused with similar small amounts of glucose; they demonstrated no changes in glucose concentrations, probably because of endogenous insulin secretion (7, 15). This observation allowed us to hypothesize additionally that even relatively small and brief increases in glucose concentrations induce an increase in the expression of TLRs and other comprehensive inflammatory changes at the cellular and molecular level in patients with type 1 diabetes who have no insulin reserve in their ␤-cells. We have previously shown that the intake of glucose and other macronutrients leads to the induction of

0193-1849/13 Copyright © 2013 the American Physiological Society

http://www.ajpendo.org

INSULIN INFUSION VS. GLUCOSE INFUSION

oxidative stress and inflammation at the cellular and molecular level. Consistent with this concept, it has been shown recently that the incubation of MNC in high glucose concentrations in vitro leads to an increase in the expression of TLR-4 on their surface as measured by flow cytometry (10). Thus, glucose appears to induce an increase in TLR-4 expression as a part of its proinflammatory effect. Furthermore, an increase in TLR-2 and TLR-4 receptors has been demonstrated in the monocytes of patients with type 1 diabetes (13, 14). This observation has important implications in terms of the inflammatory response of patients with diabetes to infections. It is well known that patients with diabetes and uncontrolled hyperglycemia are vulnerable to infections and that their ability to recover from infections is impaired (3, 16, 28). It is, therefore, clearly important to determine whether an increase in glucose concentrations in vivo in patients with type 1 diabetes would induce an increase in TLR expression and whether insulin infusion would suppress TLR expression. These data are hitherto not available in patients with type 1 diabetes. Thus, the hypotheses tested are 1) that insulin infusion suppresses the expression of TLR and high-mobility group-B1 (HMGB1) and plasma concentrations of HMGB1 in patients with type 1 diabetes as a part of its comprehensive anti-inflammatory effect; and 2) glucose infusion and the resulting hyperglycemia induce TLR and HMGB1 expression and increase in plasma HMGB1 concentrations. SUBJECTS AND METHODS

Subjects. Ten patients (7 males, 3 females) with type 1 diabetes [average age: 48 ⫾ 4 yr; mean duration of diabetes: 25 ⫾ 8 yr; body mass index: 25.4 ⫾ 1.3 kg/m2, and hemoglobin A1c (HbA1c): 7.4 ⫾ 0.7] were included in this cross-over controlled study. None of the patients had any evidence of microangiopathic complications, including microalbuminuria. Five patients were on stable treatment with statins, and six were on treatment with angiotensin-converting enzyme-inhibitors or angiotensin receptor blockers. Six patients were taking low-dose aspirin. All patients were on treatment with four injections per day of insulin [rapidly acting insulin before each meal and basal insulin (glargine) at night]. They were all maintaining stable glycemic control and had predictable fasting glucose concentrations between 90 and 130 mg/dl. They were instructed to take their basal insulin at night but no further insulin. All patients presented at the clinical research center of the Diabetes Endocrinology Center of New York at 8:00 –9:00 A.M. after an overnight fast. The experimental protocol involved three infusions on 3 days, separated by at least 1 wk. Each patient had an indwelling intravenous cannula inserted in the antecubital vein that was kept open with heparin. Through another intravenous cannula, they were infused with either 1) 2 U/h insulin and 5% dextrose (100 ml/h) for 4 h on one day; or 2) 100 ml/h of 5% dextrose for 4 h on another day; and 3) 100 ml/h of normal physiological saline only for 4 h on the 3rd day. Blood samples were collected at baseline and at 2, 4, and 6 h following the start of the infusions. The study was approved by the Human Research Committee of the State University of New York at Buffalo, and an informed consent was obtained from all patients volunteering for the study. MNC isolation. Blood samples were collected in sodium-EDTA and carefully layered on Lympholyte medium (Cedarlane Laboratories, Hornby, ON). Samples were centrifuged, and two bands were separated out at the top of the red blood cell pellet. The MNC band was harvested and washed two times with Hanks’ balanced salt solution. This method yields ⬎95% MNC preparation. Reactive oxygen species generation measurement by chemiluminescence. Five hundred microliters of MNC (2 ⫻ 105 cells) were delivered in a

E811

Chronolog Lumi-aggregometer cuvette. Luminol was then added, followed by 1.0 ␮l of 10 mM formylmethionyl leucinyl phenylalanine. In this assay system, the release of superoxide radical, as measured by chemiluminescence, has been shown to be linearly correlated with that measured by the ferricytochrome C method (9). NF-␬b DNA-binding activity. Nuclear NF-␬B and Oct-1 DNAbinding activity was measured by electrophoretic mobility shift assay. Nuclear extracts were prepared from MNC by high-salt extraction, as previously described (2, 14). The specificity of the bands was confirmed by supershifting these bands with specific antibodies against Rel-A (p65), p50, and Oct-1 or with nonspecific antibodies (Santa Cruz Biotechnology). NF-kB binding was normalized to that from Oct-1. Quantification of c-Jun NH2-terminal kinase-1, TLR-2, TLR-4, TLR-1, and CD14 expression. The mRNA expression of c-Jun NH2-terminal kinase (JNK)-1, TLR-2, TLR-4, TLR-1, and CD14 was measured in MNC by RT-PCR. Total RNA was isolated using a commercially available RNAqueous-4PCR Kit (Ambion, Austin, TX). Real-Time RT-PCR was performed using the Stratagene Mx3000P QPCR System (La Jolla, CA), Sybergreen master mix (Qiagen), and genespecific primers (Life Technologies). All values were normalized to the expression of a group of housekeeping genes, including actin, ubiquitin C, and cyclophilin A. Western blotting. MNC total cell lysates and nuclear extracts were prepared, and electrophoresis and immunoblotting were carried out as described before (7). Monoclonal antibody against JNK-1, TLR-2, TLR-4, p38 mitogen-activated protein (MAP) kinase, p47phox and HMGB1 (Abcam, Cambridge, MA), Oct-1, and actin (Santa Cruz Biotechnology) were used, and all values were corrected for loading to actin levels when total cell lysate is used or to Oct-1 when nuclear extract was used. Plasma measurements. Glucose concentrations were measured in plasma by the YSI 2300 STAT Plus glucose analyzer (Yellow Springs). Enzyme-linked immunosorbent assay (ELISA) was used to measure plasma concentrations of insulin and C-reactive protein (CRP) (Diagnostic Systems Laboratories, Webster, TX). Plasma concentration of free fatty acids (FFAs) was measured by HR series reagents (Wako Chemicals, Richmond, VA). HMGB1 concentrations were measured by ELISA from IBL International (Hamburg, Germany) as developed by Shino-Test Japan with a limit of detection of the high sensitive range assay of 0.2 ng/ml. Statistical analysis. Statistical analysis was conducted using SigmaStat software (SPSS, Chicago, IL). All data are represented as means ⫾ SE. Sample size was calculate based on our previous data in type 2 diabetes patients. Assuming a mean change in TLR-4 expression from baseline of 25% and a standard deviation of 20%, a sample size of 10 should be sufficient to detect such change with a test power of 0.8 (␤ ⫽ 0.2). Baseline measurements were normalized to 100%, and changes from baseline were calculated as percent change from baseline. Tests will be two-sided, and significance is tested at the ␣ ⫽ 0.05 level. Statistical analysis was carried out using one-way repeated-measures analysis of variance (RMANOVA) with the Holm-Sidak post hoc test. Two-factor RMANOVA followed by Dunnett’s post hoc was used for multiple comparisons between different treatments. RESULTS

Effect on glucose, insulin, and FFA concentrations. Fasting blood glucose did not change significantly in the insulin and saline infusion visits while it increased significantly from 117 ⫾ 21 to 219 ⫾ 32 mg/dl (P ⬍ 0.001; Fig. 1A) during the dextrose infusion visit. Plasma insulin concentration increased significantly (from 1.4 ⫾ 0.3 to 22.1 ⫾ 3.2 ␮U/ml, P ⬍ 0.001; Fig. 1B) during the insulin infusion, whereas it did not change significantly during the dextrose and saline infusions. There was a significant decrease in plasma FFA by 56 ⫾ 6% (from

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00566.2012 • www.ajpendo.org

E812

INSULIN INFUSION VS. GLUCOSE INFUSION

A

B

C

D

E

F

Fig. 1. Change in plasma concentrations of glucose (A), insulin (B), and free tatty acid (FFA, C), C-reactive protein (CRP, D), RNATES (E), and high-mobility group-B1 (HMGB1, F) following 2 U/h insulin/dextrose infusion (Insulin), dextrose alone (Dextrose), or saline alone (Saline) in obese patients with type 1 diabetes mellitus (T1DM) for 4 h. Data are presented as means ⫾ SE. *P ⬍ 0.05 following insulin and **P ⬍ 0.05 following dextrose by 1-way repeated-measures analysis of variance (RMANOVA). ^P ⬍ 0.05 following saline by 1-way RMANOVA (compared with baseline). P ⬍ 0.05 by 2-way RMANOVA comparing insulin with dextrose (#) and saline (%) groups. $P ⬍ 0.05 by 2-way RMANOVA comparing dextrose and saline groups.

0.34 ⫾ 0.04 to 0.14 ⫾ 0.02 mM, P ⫽ 0.009; Fig. 1C) during insulin infusion followed by an increase (by 84 ⫾ 26% to 0.54 ⫾ 0.05 mM) 2 h after cessation of the infusion. FFA increased significantly by 87 ⫾ 28% (from 0.29 ⫾ 0.04 to 0.49 ⫾ 0.05 mM, P ⫽ 0.018) during dextrose infusion and by 45 ⫾ 14% (from 0.37 ⫾ 0.05 to 0.51 ⫾ 0.06 mM, P ⫽ 0.027) during saline infusion. The increase in FFA in the glucose arm was significantly higher than that in the saline arm (P ⫽ 0.033, by 2-way RMANOVA). Effect on circulating proinflammatory mediators. Circulating levels of CRP fell significantly by 27 ⫾ 9% (from 0.71 ⫾ 0.16 to 0.45 ⫾ 0.1 mg/l) at 6 h (P ⫽ 0.008; Fig. 1D), and RANTES fell by 25 ⫾ 10% (from 36.0 ⫾ 4.3 to 26.1 ⫾ 4.7 ng/ml at 2 h, P ⫽ 0.012; Fig. 1E) following insulin infusion. There was no significant change in these indexes following dextrose or saline infusions. HMGB1 concentrations fell significantly following insulin infusion by 22 ⫾ 5% (from 1.28 ⫾ 0.21 to 1.08 ⫾ 0.22 ng/ml at 4 h, P ⫽ 0.016; Fig. 1F) while it increased significantly following glucose infusion by 19 ⫾ 8% (from 1.29 ⫾ 0.19 to 1.45 ⫾ 0.21 ng/ml at 2 and 4 h, P ⫽ 0.024; Fig. 1F). Saline infusion did not cause any significant change in HMGB1 concentrations. Effect on reactive oxygen species generation. There was a significant reduction by 23 ⫾ 7% below the baseline (P ⫽ 0.026) in reactive oxygen species (ROS) generation by poly-

morphonuclear neutrophils (PMN) at 6 h following insulin infusion while it increased by 59 ⫾ 18% above the baseline (P ⫽ 0.032) at 6 h following dextrose infusion. ROS generation by PMN did not change during saline infusion (Fig. 2A). The protein expression of p47phox, a key subunit of NADPH oxidase, in MNC was significantly suppressed by 31 ⫾ 11% (P ⬍ 0.025; Fig. 2, B and C) below the baseline during insulin infusion while there was no significant change in p47phox following dextrose or saline infusions. Effect on intranuclear NF-␬B binding. NF-␬B binding was diminished (by 12 ⫾ 7%) but not significantly after the infusion of insulin. It did not increase significantly after glucose or saline infusion (data not shown). Effect on Toll-like receptors in MNC. The mRNA expression of TLR-1, TLR-2, TLR-4, and CD14 in MNC fell by 14 ⫾ 5, 29 ⫾ 8, 18 ⫾ 7, and 35 ⫾ 9% below baseline, respectively (P ⫽ 0.001– 0.034 for all; Fig. 3, A–D), following insulin infusion. TLR-1, TLR-2, and TLR-4 mRNA expression increased by 51 ⫾ 18, 37 ⫾ 10, and 48 ⫾ 21%, respectively, above the baseline following dextrose infusion (P ⫽ 0.011– 0.030 for all; Fig. 3, A–D). CD14 expression did not alter after dextrose infusion. There was no change in these mediators following saline infusion. There was a significant increase in the MNC TLR-4 protein levels by 34 ⫾ 12% above baseline (P ⫽ 0.018; Fig. 3, E and F) following dextrose infusion while

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00566.2012 • www.ajpendo.org

E813

INSULIN INFUSION VS. GLUCOSE INFUSION

A

C Fig. 2. Percent change in reactive oxygen species (ROS) generation by polymorphonuclear neutrophils (PMN, A) and in p47phox protein levels in mononuclear cells (MNC, C) as measured by Western blotting following 2 U/h insulin/dextrose infusion, dextrose alone, or saline alone in patients with T1DM for 4 h. Densitometric measurements are corrected for loading with that of actin. B: representative Western blots chosen to demonstrate the average change in p47phox and actin in each study group. Panels are separated by spaces to indicate that samples presented are from more than one gel. Data are presented as means ⫾ SE. *P ⬍ 0.05 following insulin and **P ⬍ 0.05 following dextrose by 1-way RMANOVA (compared with baseline). #P ⬍ 0.05 by 2-way RMANOVA comparing insulin with dextrose group. $P ⬍ 0.05 by 2-way RMANOVA comparing dextrose and saline groups.

B

it did not change following either insulin or saline infusion. TLR-2 protein levels did not change during any infusion (data not shown). Effect on other proinflammatory mediators in MNC. There was a significant fall in p38 MAP kinase and HMGB1 protein in MNC by 18 ⫾ 8% (P ⫽ 0.024) and 36 ⫾ 7% (P ⫽ 0.011) below baseline, respectively (Fig. 4, A and C-D), following insulin infusion while the protein levels of p38 MAP kinase, HMGB1, and JNK-1 increased significantly by 27 ⫾ 11% (P ⫽ 0.017), 39 ⫾ 18% (P ⫽ 0.01), and 26 ⫾ 11% (P ⫽ 0.019) above baseline, respectively (Fig. 4, A-D), following dextrose infusion. HMGB1 protein was also suppressed in the nuclear extract from MNC by 26 ⫾ 9% (P ⫽ 0.021) at 4 h following insulin infusion while glucose infusion increased nuclear levels of HMGB1 by 29 ⫾ 9% (P ⫽ 0.025) at 6 h (Fig. 4, E and F). There was no significant change in these mediators following saline infusion. The expression of the adhesion molecule platelet/endothelial cell adhesion molecule mRNA fell by 18 ⫾ 7% (P ⫽ 0.032) at 6 h following insulin infusion without any significant change following dextrose or saline only infusions (Fig. 4G). DISCUSSION

Our data show clearly for the first time that insulin exerts a comprehensive anti-inflammatory and a suppressive effect on the expression of TLR-4, TLR-2, TLR-1, and CD14 in patients with type 1 diabetes. In addition, there was a reduction in plasma CRP and RANTES concentrations. It is of interest that, while a trend toward the reduction in TLR expression started at 4 h, it became significant only at 6 and 2 h after the cessation of insulin infusion. In this respect, the response to insulin in type 1 diabetes patients was significantly slower than that in type 2 diabetes patients since, in the latter, TLR expression was diminished within 2 h of the initiation of insulin infusion (15). At 6 h, TLR expression in type 2 diabetic patients had reverted to the baseline. Saline-infused controls did not show a change in any of the indexes.

This trend of a prolonged action 2 h after the cessation of the infusion was also observed with the suppression of HMGB1 expression. The suppression of HMGB1 was rapid and was significant at 2 h. HMGB1 is an intranuclear protein that binds to histones and during inflammation helps “open up” the nucleosome such that the gene promoters are available for binding to proinflammatory transcription factors like NF-␬B and RNA polymerase II to initiate transcription (5, 22, 37). In addition, in severe inflammatory conditions associated with apoptosis and cell death, HMGB1 is released in tissue spaces and plasma (35). In this setting, it acts as a proinflammatory cytokine through its binding to the receptor for advance glycation end products (RAGE) (20). This novel effect of insulin in type 1 diabetes patients shows that, quite apart from suppressing proinflammatory transcription factors, as previously demonstrated, insulin has distinct additional effects on the level of TLR and HMGB1 both of which have been shown to be increased in patients with type 1 diabetes (12, 13). Its suppressive action on HMGB1 would potentially inhibit the action of proinflammatory transcription factors and the transcription of proinflammatory genes. In addition, the suppression of plasma concentrations of HMGB1 would result in the inhibition of its proinflammatory cytokine-like effects through the RAGE, TLR-2, and TLR-4. Equally important, glucose and hyperglycemia exert potent proinflammatory effects through the increase in TLR and HMGB1 expression as well as the plasma concentrations of the latter, as discussed below. We also investigated the effect of insulin and glucose on the expression of p38 MAP kinase and JNK-1. Whereas p38 MAP kinase was significantly suppressed by insulin during this short period of infusion, JNK-1 did not change. Both kinases can phosphorylate inhibitory factor ␬B␣ and thus lead to its ubiquitination and proteasomal degradation. This releases cytosolic NF-␬B so that it translocates into the nucleus and initiates proinflammatory transcription (19, 36). These kinases also mediate serine phosphorylation of insulin receptor substrate-1 and thus prevent its binding to phosphatidylinositol 3-kinase

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00566.2012 • www.ajpendo.org

E814

INSULIN INFUSION VS. GLUCOSE INFUSION

A

B

C

E

D

F

Fig. 3. Percent change in Toll-like receptor (TLR)-2 (A), TLR-4 (B), TLR-1 (C), and CD14 (D) mRNA expression and representative Western blot (E) with percent change (F) of TLR-4 protein in MNC following 2 U/h insulin/dextrose infusion, dextrose alone, or saline alone in obese T1DM patients for 4 h. Representative Western blots are chosen to demonstrate average change in TLR-4 and actin in each study group. Panels are separated by spaces to indicate that samples presented are from more than one gel. Densitometric measurements are corrected for loading with that of actin. Data are presented as means ⫾ SE. *P ⬍ 0.05 following insulin and **P ⬍ 0.05 following dextrose by 1-way RMANOVA (compared with baseline). P ⬍ 0.05 by 2-way RMANOVA comparing insulin with dextrose (#) and saline (%) groups. $P ⬍ 0.05 by 2-way RMANOVA comparing dextrose and saline groups.

during the process of insulin signal transduction. They can thus potentially mediate insulin resistance (21, 39). The suppression of p38 MAP kinase by insulin is also relevant because, over the past 20 years, it has been believed that insulin induces p38 MAP kinase (23, 26). These data were produced in vascular smooth muscle cells in vitro using very high insulin concentrations. There have been no data showing that insulin affects p38 MAP kinase in vivo. Insulin induced a suppression of ROS generation by PMNs and a reduction in the expression of p47phox in MNC. p47phox is an essential component of the enzyme NADPH oxidase and serves as a surrogate for the expression of this enzyme. As with the suppression of TLRs, these changes were seen at 6 h, rather late compared with our previous data in the obese and type 2 diabetes patients. Similarly, glucose infusion also induced an increase in ROS generation as expected, but, again, it was observed only at 6 h. The reason for this delayed response to insulin and glucose in type 1 diabetes patients is not clear and will require further investigation.

TLRs recognize specific pathogen-associated molecular patterns and through specific signal transduction mechanisms activate NF-␬B, the major proinflammatory transcription factor (17, 43). TLR-4 is the specific receptor for endotoxin [lipopolysaccharide (LPS)], whereas TLR-2 specifically recognizes lipopeptides from Gram-positive bacteria. There is evidence now that FFAs also exert their proinflammatory effect, which we first described through the mediation of fetuin A which helps palmitate to bind to TLRs (33). It is intriguing that there was no significant reduction in NF-␬B binding. The absence of significant suppression may have been because of the relatively small number of patients included in this study. It is, however, possible that the antiinflammatory effects observed were probably independent of any alteration in NF-␬B binding. The transcription of TLRs is regulated by PU.1, a distinct transcription factor that has been shown to be suppressed by insulin by us previously (15). HMGB1 expression is upregulated by the transcription factor p73␣ and downregulated by p53, which in turn is also modu-

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00566.2012 • www.ajpendo.org

INSULIN INFUSION VS. GLUCOSE INFUSION

E815

A

B

C

D

E

F

G

Fig. 4. Representative Western blot (A) from total cell lysate and nuclear extracts (NE) (E) and percent change in c-Jun NH2-terminal kinase (JNK-1, B), p38 (C), and HMGB1 (D) protein, HMGB1 protein in NE (F), and in platelet/endothelial cell adhesion molecule (PECAM) mRNA (G) in MNC following 2 U/h insulin/dextrose infusion, dextrose alone, or saline alone in obese T1DM patients for 4 h. Representative Western blots are chosen to demonstrate average change in JNK-1, p38, HMGB-1, Oct-1, and actin in each study group. Panels are separated by spaces to indicate that samples presented are from more than one gel. Densitometric measurements are corrected for loading with that of actin (total cell lysate) or Oct-1 (NE). Data are presented as means ⫾ SE. *P ⬍ 0.05 following insulin and **P ⬍ 0.05 following dextrose by 1-way RMANOVA (compared with baseline). P ⬍ 0.05 by 2-way RMANOVA comparing insulin with dextrose (#) and saline (%) groups. $P ⬍ 0.05 by 2-way RMANOVA comparing dextrose and saline groups.

lated by HMGB1 (42). It would, therefore, be of interest to examine the effect of insulin on these two regulators of HMGB1 expression. In contrast to the actions of insulin, glucose infusion induced marked proinflammatory changes. It is of interest that, in these patients with type 1 diabetes maintained on glargine insulin overnight but with no additional rapidly acting insulin in the morning, the infusion of 5% dextrose at a rate of 100 ml/h for 4 h (⫽20 g glucose ⫽ 5 g/h) led to an increase of plasma glucose concentrations from 115 to 215 mg/dl over 6 h. This change was associated with an increase in TLR-2 and TLR-4, JNK-1, and HMGB1 expression. Although an increase in TLR

expression has been shown in monocytic cell lines in vitro at very high glucose concentrations using flow cytometry (10) and a relationship between glycemia and TLR expression has been shown in patients with type 1 diabetes (13), this is the first demonstration of a hyperglycemia-induced increase in TLR expression in vivo. These data also point out that any glucose or macronutrient load in the type 1 diabetic patient without a significant ␤-cell reserve and without an insulin bolus of rapidly acting insulin could suffer significant hyperglycemia, inflammation, and the induction of mediators that would cause not only inflammation but also potential interference with insulin signal transduction. The increase in ROS generation

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00566.2012 • www.ajpendo.org

E816

INSULIN INFUSION VS. GLUCOSE INFUSION

induced by these small amounts of glucose is consistent with the previous observations of Monnier et al. (30) demonstrating an increase in oxidative stress induced by even small variations in plasma glucose concentrations in type 2 diabetics as reflected in urinary isoprostane levels and those of Mohanty et al. (29) in NADPH oxidase-mediated ROS generation by leucocytes in normal subjects following a glucose challenge. As expected, insulin infusion resulted in the suppression of plasma FFA concentration. This reflects the anti-lipolytic effect of insulin. The cessation of insulin infusion led to a rapid rebound of plasma FFA concentrations to preinfusion levels. In this context, it is of interest that glucose infusion also resulted in a progressive increase in FFA concentration that was maximal at 4 h but was still higher than the baseline at 6 h, 2 h after the cessation of glucose infusion. The mechanism underlying the glucose-induced increase in FFA concentrations is not clear. An increase in FFAs would also contribute to interference with insulin signal transduction and insulin resistance. The increase in FFA concentrations was not only the result of the lack of rapidly acting insulin, since the infusion with saline alone was associated with an increase in FFA of a much smaller magnitude. Thus, the increase following glucose infusion was the result of the combination of the lack of insulin and the infusion of glucose. Because plasma FFA concentrations reflect adipose tissue lipolysis, the effect of the infusion of glucose on adipose tissue lipolysis in type 1 diabetes patients needs to be investigated. Furthermore, an increase in FFA concentrations has previously been shown by us to rapidly induce oxidative stress and inflammation (41). The increase in FFA concentration and the increases in the expression of JNK-1 and TLRs following small amounts of glucose also have implications for insulin resistance, since all three potentially interfere with insulin signal transduction and may contribute to insulin resistance. This is relevant since over 40% of type 1 diabetes patients in the United States have the metabolic syndrome. The suppression of TLR-4, TLR-2, TLR-1, and CD14 by insulin is also of significance. While TLR-4 is the receptor for LPS, CD14 facilitates the binding of LPS to TLR-4. Thus, insulin infusion could be a potentially powerful treatment for endotoxemia-based inflammation, a feature of Gram-negative infections. Indeed, we have recently demonstrated a potent inhibitory effect of insulin on oxidative, nitrosative, and inflammatory stresses in normal subjects injected with LPS (8). Similarly, the suppression of TLR-2 and TLR-1 is relevant to inflammation induced by Gram-positive infections, since the heterodimer formed by TLR-2 and TLR-1 gets activated by lipopeptides and peptidoglycans from such bacteria. Glucose infusion and hyperglycemia induced an increase in TLR expression. This would potentially enhance the proinflammatory effects of both Gram-positive and Gram-negative bacteria and their products. It should be stated, however, that the changes in all TLRs were at the mRNA level except for TLR-4 protein, which increased after glucose infusion. It is possible that a longer duration of infusion would have led to more uniform changes in the expression of TLR proteins. The rapid and diametrically opposite effects of glucose and insulin on TLR, CD14, and HMGB1 expression have implications for the treatment of inflammation related to infections. Hyperglycemia is known to interfere with the eradication of infection-related inflammation, whereas insulin is known to

exert anti-inflammatory effects in patients in intensive care (24). The clinical significance of our observations may also be relevant to complications of type 1 diabetes, as seen in microangiopathy and macroangiopathy both of which have oxidative and inflammatory stress as a part of their pathogenesis (11, 40). Hyperglycemia is associated with an increase in microvascular complications of diabetes and the restoration of normoglycemia to a marked reduction in these complications. While hyperglycemia is associated with macrovascular complications, its reversal does not reduce cardiovascular event rates as impressively as it reduces the microvascular complications. Indeed, as shown by the DCCT-EDIC and UKPDS studies, the beneficial effects of improved glycemic control on cardiovascular events may take a very long time (41a, 32). Another relevant issue in this context is the fact that, while HbA1c was a major determinant of the occurrence of microvascular complications in the DCCT study, patients on intensive treatment with multiple injections or continuous subcutaneous insulin infusion had significantly less complications for any given level of HbA1c. This suggests that small diurnal fluctuations of glucose may contribute to microangiopathy. We have previously shown that intravenous insulin has anti-inflammatory and cardioprotective effects in the setting of myocardial infarction (4). Insulin infusion also reduces apoptosis and inflammatory damage in peri-infarction zones of the myocardium, as seen in myocardial biopsies taken during coronary artery bypass surgery following acute myocardial infarction (27). Perioperative insulin infusion decreased cardiovascular events in patients undergoing peripheral vascular surgery in one study (38). While such data are not available in patients specifically with type 1 diabetes, our results on the antiinflammatory effects of intravenous insulin provide a rationale for conducting such trials to improve outcomes in patients with type 1 diabetes presenting with acute cardiovascular events. The induction of hyperglycemia and potent inflammatory effects observed following the infusion of a small amount of glucose over a period of 4 h implies that patients with type 1 diabetes are in a very fragile state and are vulnerable to even minimal fluctuations in glycemic load in the absence of a corresponding dose of insulin. This has implications in terms of meal intake and missed or delayed prandial insulin doses. This issue needs to be investigated further. In conclusion, insulin exerts anti-inflammatory effects in type 1 diabetes patients, including the suppression of TLR-1, TLR-2, and TLR-4 and the expression of HMGB1 and p38 MAP kinase. These actions may mediate its protective effects in the pathogenesis of inflammation by infective agents and potentially offer a powerful mode of controlling inflammatory conditions in such patients. In addition, even mild and shortlived hyperglycemia in these patients induces an increase in TLR-1, TLR-2, and TLR-4, JNK-1, HMGB1, and p38 MAP kinase expression and other inflammatory mediators in addition to inducing comprehensive inflammation at the cellular and molecular level. These effects of small amounts of glucose in type 1 diabetes patients need to be translated into clinically relevant investigations in the future. DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the authors.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00566.2012 • www.ajpendo.org

INSULIN INFUSION VS. GLUCOSE INFUSION AUTHOR CONTRIBUTIONS Author contributions: P.D., H.G., and A.C. conception and design of research; P.D., H.G., S.A., and M.B. interpreted results of experiments; P.D., H.G., N.K., S.D., and A.C. drafted manuscript; P.D., H.G., S.D., and A.C. edited and revised manuscript; P.D. approved final version of manuscript; H.G., K.G., C.L.S., S.A., and M.B. performed experiments; H.G., K.G., C.L.S., S.A., N.K., M.B., and S.D. analyzed data; H.G. prepared figures. REFERENCES 1. Aljada A, Ghanim H, Saadeh R, Dandona P. Insulin inhibits NFkappaB and MCP-1 expression in human aortic endothelial cells. J Clin Endocrinol Metab 86: 450 –453, 2001. 2. Andrews NC, Faller DV. A rapid micropreparation technique for extraction of DNA-binding proteins from limiting numbers of mammalian cells (Abstract). Nucleic Acids Res 19: 2499, 1991. 3. Butler SO, Btaiche IF, Alaniz C. Relationship between hyperglycemia and infection in critically ill patients. Pharmacotherapy 25: 963–976, 2005. 4. Chaudhuri A, Janicke D, Wilson MF, Tripathy D, Garg R, Bandyopadhyay A, Calieri J, Hoffmeyer D, Syed T, Ghanim H, Aljada A, Dandona P. Anti-inflammatory and profibrinolytic effect of insulin in acute ST-segment-elevation myocardial infarction. Circulation 109: 849 – 854, 2004. 5. Churchill ME, Changela A, Dow LK, Krieg AJ. Interactions of high mobility group box proteins with DNA and chromatin. Methods Enzymol 304: 99 –133, 1999. 7. Dandona P, Aljada A, Mohanty P, Ghanim H, Hamouda W, Assian E, Ahmad S. Insulin inhibits intranuclear nuclear factor kappaB and stimulates IkappaB in mononuclear cells in obese subjects: evidence for an anti-inflammatory effect? J Clin Endocrinol Metab 86: 3257–3265, 2001. 8. Dandona P, Ghanim H, Bandyopadhyay A, Korzeniewski K, Ling Sia C, Dhindsa S, Chaudhuri A. Insulin suppresses endotoxin-induced oxidative, nitrosative, and inflammatory stress in humans. Diabetes Care 33: 2416 –2423, 2010. 9. Dandona P, Thusu K, Cook S, Snyder B, Makowski J, Armstrong D, Nicotera T. Oxidative damage to DNA in diabetes mellitus. Lancet 347: 444 –445, 1996. 10. Dasu MR, Devaraj S, Zhao L, Hwang DH, Jialal I. High glucose induces toll-like receptor expression in human monocytes: mechanism of activation. Diabetes 57: 3090 –3098, 2008. 11. Devaraj S, Cheung AT, Jialal I, Griffen SC, Nguyen D, Glaser N, Aoki T. Evidence of increased inflammation and microcirculatory abnormalities in patients with type 1 diabetes and their role in microvascular complications. Diabetes 56: 2790 –2796, 2007. 12. Devaraj S, Dasu MR, Park SH, Jialal I. Increased levels of ligands of Toll-like receptors 2 and 4 in type 1 diabetes. Diabetologia 52: 1665– 1668, 2009. 13. Devaraj S, Dasu MR, Rockwood J, Winter W, Griffen SC, Jialal I. Increased toll-like receptor (TLR) 2 and TLR4 expression in monocytes from patients with type 1 diabetes: further evidence of a proinflammatory state. J Clin Endocrinol Metab 93: 578 –583, 2008. 14. Ghanim H, Aljada A, Hofmeyer D, Syed T, Mohanty P, Dandona P. Circulating mononuclear cells in the obese are in a proinflammatory state. Circulation 110: 1564 –1571, 2004. 15. Ghanim H, Mohanty P, Deopurkar R, Sia CL, Korzeniewski K, Abuaysheh S, Chaudhuri A, Dandona P. Acute modulation of toll-like receptors by insulin. Diabetes Care 31: 1827–1831, 2008. 16. Gupta S, Koirala J, Khardori R, Khardori N. Infections in diabetes mellitus and hyperglycemia. Infect Dis Clin North Am 21: 617–638, 2007. 17. Hallman M, Ramet M, Ezekowitz RA. Toll-like receptors as sensors of pathogens. Pediatr Res 50: 315–321, 2001. 18. Heinola T, Kouri VP, Clarijs P, Ciferska H, Sukura A, Salo J, Konttinen YT. High mobility group box-1 (HMGB-1) in osteoarthritic cartilage. Clin Exp Rheumatol 28: 511–518, 2010. 19. Hirano M, Osada S, Aoki T, Hirai S, Hosaka M, Inoue J, Ohno S. MEK kinase is involved in tumor necrosis factor alpha-induced NFkappaB activation and degradation of IkappaB-alpha. J Biol Chem 271: 13234 –13238, 1996. 20. Hori O, Brett J, Slattery T, Cao R, Zhang J, Chen JX, Nagashima M, Lundh ER, Vijay S, Nitecki D. The receptor for advanced glycation end products (RAGE) is a cellular binding site for amphoterin Mediation of neurite outgrowth and co-expression of rage and amphoterin in the developing nervous system. J Biol Chem 270: 25752–25761, 1995.

E817

21. Hotamisligil GS. Inflammation and metabolic disorders. Nature 444: 860 –867, 2006. 22. Hou CH, Fong YC, Tang CH. HMGB-1 induces IL-6 production in human synovial fibroblasts through c-Src, Akt and NF-kappaB pathways. J Cell Physiol 226: 2006 –2015, 2011. 23. Igarashi M, Yamaguchi H, Hirata A, Daimon M, Tominaga M, Kato T. Insulin activates p38 mitogen-activated protein (MAP) kinase via a MAP kinase kinase (MKK) 3/MKK 6 pathway in vascular smooth muscle cells. Eur J Clin Invest 30: 668 –677, 2000. 24. Langouche L, Vanhorebeek I, Vlasselaers D, Vander Perre S, Wouters PJ, Skogstrand K, Hansen TK, Van den Berghe G. Intensive insulin therapy protects the endothelium of critically ill patients. J Clin Invest 115: 2277–2286, 2005. 25. Li P, Cook JA, Gilkeson GS, Luttrell LM, Wang L, Borg KT, Halushka PV, Fan H. Increased expression of beta-arrestin 1 and 2 in murine models of rheumatoid arthritis: isoform specific regulation of inflammation. Mol Immunol 49: 64 –74, 2011. 26. Madonna R, Pandolfi A, Massaro M, Consoli A, De Caterina R. Insulin enhances vascular cell adhesion molecule-1 expression in human cultured endothelial cells through a pro-atherogenic pathway mediated by p38 mitogen-activated protein-kinase. Diabetologia 47: 532–536, 2004. 27. Marfella R, Di Filippo C, Portoghese M, Ferraraccio F, Rizzo MR, Siniscalchi M, Musacchio E, D’Amico M, Rossi F, Paolisso G. Tight glycemic control reduces heart inflammation and remodeling during acute myocardial infarction in hyperglycemic patients. J Am Coll Cardiol 53: 1425–1436, 2009. 28. Michalia M, Kompoti M, Koutsikou A, Paridou A, Giannopoulou P, Trikka-Graphakos E, Clouva-Molyvdas P. Diabetes mellitus is an independent risk factor for ICU-acquired bloodstream infections. Intensive Care Med 35: 448 –454, 2009. 29. Mohanty P, Hamouda W, Garg R, Aljada A, Ghanim H, Dandona P. Glucose challenge stimulates reactive oxygen species (ROS) generation by leucocytes. J Clin Endocrinol Metab 85: 2970 –2973, 2000. 30. Monnier L, Mas E, Ginet C, Michel F, Villon L, Cristol JP, Colette C. Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. J Am Med Assoc 295: 1681–1687, 2006. 31. Muller LM, Gorter KJ, Hak E, Goudzwaard WL, Schellevis FG, Hoepelman AI, Rutten GE. Increased risk of common infections in patients with type 1 and type 2 diabetes mellitus. Clin Infect Dis 41: 281–288, 2005. 32. Nathan DM, Cleary PA, Backlund JY, Genuth SM, Lachin JM, Orchard TJ, Raskin P, Zinman B. Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes. N Engl J Med 353: 2643–2653, 2005. 33. Pal D, Dasgupta S, Kundu R, Maitra S, Das G, Mukhopadhyay S, Ray S, Majumdar SS, Bhattacharya S. Fetuin-A acts as an endogenous ligand of TLR4 to promote lipid-induced insulin resistance. Nat Med In press. 34. Pradhan AD, Everett BM, Cook NR, Rifai N, Ridker PM. Effects of initiating insulin and metformin on glycemic control and inflammatory biomarkers among patients with type 2 diabetes: the LANCET randomized trial. J Am Med Assoc 302: 1186 –1194, 2009. 35. Qin S, Wang H, Yuan R, Li H, Ochani M, Ochani K, Rosas-Ballina M, Czura CJ, Huston JM, Miller E, Lin X, Sherry B, Kumar A, Larosa G, Newman W, Tracey KJ, Yang H. Role of HMGB1 in apoptosismediated sepsis lethality. J Exp Med 203: 1637–1642, 2006. 36. Schulze-Osthoff K, Ferrari D, Riehemann K, Wesselborg S. Regulation of NF-kappa B activation by MAP kinase cascades. Immunobiology 198: 35–49, 1997. 37. Stros M. HMGB proteins: interactions with DNA and chromatin. Biochim Biophys Acta 1799: 101–113, 2010. 38. Subramaniam B, Panzica PJ, Novack V, Mahmood F, Matyal R, Mitchell JD, Sundar E, Bose R, Pomposelli F, Kersten JR, Talmor DS. Continuous perioperative insulin infusion decreases major cardiovascular events in patients undergoing vascular surgery: a prospective, randomized trial. Anesthesiology 110: 970 –977, 2009. 39. Tanti JF, Jager J. Cellular mechanisms of insulin resistance: role of stress-regulated serine kinases and insulin receptor substrates (IRS) serine phosphorylation. Curr Opin Pharmacol 9: 753–762, 2009. 40. Targher G, Bertolini L, Zoppini G, Zenari L, Falezza G. Increased plasma markers of inflammation and endothelial dysfunction and their association with microvascular complications in Type 1 diabetic patients without clinically manifest macroangiopathy. Diabet Med 22: 999 –1004, 2005.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00566.2012 • www.ajpendo.org

E818

INSULIN INFUSION VS. GLUCOSE INFUSION

41. Tripathy D, Mohanty P, Dhindsa S, Syed T, Ghanim H, Aljada A, Dandona P. Elevation of free fatty acids induces inflammation and impairs vascular reactivity in healthy subjects. Diabetes 52: 2882–2887, 2003. 41a.UK Prospective Diabetes Study (UKPDS) Group. Intensive bloodglucose control with sulphonylureas or insulin compared with conventional treatment, and risk of complications in patients with type 2 diabetes (UKPDS 33) Lancet 352: 837–853, 1998.

42. Uramoto H, Izumi H, Nagatani G, Ohmori H, Nagasue N, Ise T, Yoshida T, Yasumoto K, Kohno K. Physical interaction of tumour suppressor p53/p73 with CCAAT-binding transcription factor 2 (CTF2) and differential regulation of human high-mobility group 1 (HMG1) gene expression. Biochem J 371: 301–310, 2003. 43. Zhang G, Ghosh S. Molecular mechanisms of NF-kappaB activation induced by bacterial lipopolysaccharide through Toll-like receptors. J Endotoxin Res 6: 453–457, 2000.

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00566.2012 • www.ajpendo.org