Reduction of serum FABP4 level by sitagliptin, a DPP-4 inhibitor, in patients with type 2 diabetes mellitus[S]

Fatty acid binding protein 4 (FABP4), also known as adipocyte FABP or aP2, is secreted from adipocytes in association with lipolysis as a novel adipokine, and elevated serum FABP4 level is associated with obesity, insulin resistance, and atherosclerosis. However, little is known about the modulation of serum FABP4 level by therapeutic drugs. Sitagliptin (50 mg/day), a dipeptidyl peptidase 4 (DPP-4) inhibitor that increases glucagon-like peptide 1 (GLP-1), was administered to patients with type 2 diabetes (n = 24) for 12 weeks. Treatment with sitagliptin decreased serum FABP4 concentration by 19.7% (17.8 ± 1.8 vs. 14.3 ± 1.5 ng/ml, P < 0.001) and hemoglobin A1c without significant changes in adiposity or lipid variables. In 3T3-L1 adipocytes, sitagliptin or exendin-4, a GLP-1 receptor agonist, had no effect on short-term (2 h) secretion of FABP4. However, gene expression and long-term (24 h) secretion of FABP4 were significantly reduced by sitagliptin, which was not mimicked by exendin-4. Treatment with recombinant DPP-4 increased gene expression and long-term secretion of FABP4, and the effects were cancelled by sitagliptin. Furthermore, knockdown of DPP-4 in 3T3-L1 adipocytes decreased gene expression and long-term secretion of FABP4. In conclusion, sitagliptin decreases serum FABP4 level, at least in part, via reduction in the expression and consecutive secretion of FABP4 in adipocytes by direct inhibition of DPP-4.


Study 2: effects of sitagliptin and exendin-4 on FABP4 expression and secretion in 3T3-L1 adipocytes
All biochemical reagents were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise indicated. Preadipocyte 3T3-L1 cells were obtained from Health Science Research Resources Bank (Osaka, Japan). Differentiation of 3T3-L1 cells was induced as previously described ( 10 ). After an overnight serum depletion by 0.5% BSA in DMEM (Invitrogen, Carlsbad, CA), differentiated 3T3-L1 adipocytes were stimulated with 0-50 µM sitagliptin, 0-50 nM exendin-4 (a GLP-1 receptor agonist), and 0-500 ng/ml recombinant DPP-4 (Prospec, East Brunswick, NJ) in the absence or presence of 10 µM isoproterenol in DMEM supplemented with 0.5% BSA for 2 or 24 h. The doses of reagents and incubation periods were varied according to the experimental protocol. Each experiment was done in at least triplicate. Cytotoxicity was assessed by lactate dehydrogenase (LDH) assay kit (Promega, Madison, WI).
Small interfering RNA knockdown. Small interfering RNA (siRNA) analysis was performed by using Stealth RNAi (Invitrogen) targeting mouse DPP-4 mRNA (cat. #MSS203657) and negative control (cat. #12935-300). Using a method of density-based separation followed by replating of enriched adipocytes in a monolayer ( 32 ), differentiated 3T3-L1 adipocytes were transfected with specifi c or control Stealth RNAi by lipofectamine 2000 (Invitrogen) and cultured for 48 h.
For treatment of type 2 diabetes mellitus, incretinbased therapies using glucagon-like peptide 1 (GLP-1) receptor agonists or dipeptidyl peptidase 4 (DPP-4) inhibitors have become available ( 26,27 ). GLP-1 receptor agonists directly activate the GLP-1 receptor, whereas DPP-4 inhibitors slow GLP-1 degradation, thereby increasing the concentration of endogenous GLP-1, which activates the GLP-1 receptor (28)(29)(30). The potential glycemic benefi ts of these treatments include glucosedependent stimulation of insulin production and secretion; suppression of inappropriate glucagon secretion; slowing of gastric emptying, which reduces the rate of glucose appearance in the circulation; and satiety, which may reduce food intake ( 29 ). In the present study, we investigated the impact of DPP-4 inhibitor therapy on serum FABP4 level in patients with type 2 diabetes mellitus and its mechanisms.

SUBJECTS AND METHODS
The present study consisted of a human study (study 1) and an in vitro study (study 2). Study 1 conformed to the principles outlined in the Declaration of Helsinki and was performed with the approval of the Ethical Committee of Fujita Health University. Written informed consent was received from all of the study subjects. Study 2 was approved by the Committee for Animal Research, Sapporo Medical University, and was conducted in accordance with the Guidelines of Sapporo Medical University for Animal Use in Research.

Study 1: effects of sitagliptin on FABP4 level in patients with type 2 diabetes mellitus
A total of 24 type 2 diabetic patients treated with only diet and exercise therapy [n = 14; male/female (M/F): 7/7] or glimepiride (0.5-2 mg/day), a sulfonylurea, for more than 8 weeks (n = 10, M/F: 8/2) were enrolled from outpatient clinics affi liated with Fujita Health University. Entry criteria were no treatment with antidiabetic agents except for sulfonylurea before the enrollment and no evidence of complications such as hepatic, cerebrovascular, cardiovascular, or renal disease. Before and after treatment with sitagliptin (50 mg/day), a DPP-4 inhibitor, for 12 weeks, blood samples were collected, and plasma and serum samples were analyzed immediately or stored at Ϫ 80°C until biochemical analyses.
Serum concentration of FABP4 was measured using a commercially available enzyme-linked immunosorbent assay kit for FABP4 (Biovendor R and D, Modrice, Czech Republic). The accuracy, precision, and reproducibility of the kit have been described previously ( 8 ). The intra-and interassay coeffi cient variances in the kits were <5%. Plasma glucose was determined by the glucose oxidase method. Fasting plasma insulin was measured by a radioimmunoassay method. Hemoglobin A1c (HbA1c) was determined by a latex coagulation method and was expressed

Statistical analysis
Numeric variables are expressed as means ± SEM. The distribution of each parameter was tested for its normality using the Shapiro-Wilk W test, and nonnormally distributed parameters were logarithmically transformed for regression analyses. The correlation between two variables was evaluated using Pearson's correlation coeffi cient. Comparison between two groups was done with Wilcoxon signed-rank test for paired samples and Mann-Whitney's U test for unpaired samples. One-way ANOVA and the Tukey-Kramer post hoc test were used for detecting signifi cant differences in data between more than two groups. A P value of <0.05 was considered statistically signifi cant. All data were analyzed by using JMP 9 for Macintosh (SAS Institute, Cary, NC). was present in both the CL and CM of 3T3-L1 adipocytes and that GAPDH, a nonsecretory protein, was not present in the CM ( Fig. 2C-F ), indicating that FABP4 is secreted from adipocytes but not leaked by injury of cell membranes. No cytotoxic effect, measured as LDH leakage, was observed in any of the experiments performed (data not shown). Treatment with sitagliptin or exendin-4 did not affect short-term (2 h) secretion of FABP4 from adipocytes in the absence or presence of isoproterenol ( Fig.  2C, D ). However, treatment with sitagliptin for 24 h significantly decreased FABP4 secretion from adipocytes, though the effect of sitagliptin was not mimicked by exendin-4 ( Fig. 2E, F ). DPP-4 was present in both the CL and CM of 3T3-L1 adipocytes incubated for 24 h ( Fig. 3A ), indicating that DPP-4 is released from adipocytes as previously reported ( 34 ). Treatment of differentiated 3T3-L1 adipocytes with recombinant DPP-4 for 24 h increased gene expression of FABP4 in a dose-dependent manner ( Fig. 3B ), and the effect of DPP-4 was inhibited by sitagliptin ( Fig. 3C ). Similar results were obtained for gene expression of PPAR ␥ 2 and C/EBP ␣ (supplementary Fig. 2A-D). Western blot analysis showed that treatment with recombinant DPP-4 significantly increased long-term (24 h) secretion of FABP4 from adipocytes, which was inhibited by sitagliptin ( Fig. 3D ). However, sitagliptin and/or DPP-4 did not affect shortterm (2 h) secretion of FABP4 ( Fig. 3E ) or adiponectin (supplementary Fig. 3).

DISCUSSION
The present study showed for the fi rst time that sitagliptin, a DPP-4 inhibitor, decreased serum FABP4 concentration in patients with type 2 diabetes mellitus. Treatment of 3T3-L1 adipocytes with sitagliptin for 24 h signifi cantly decreased gene expression and consecutive secretion of FABP4. We also confi rmed that DPP-4 is released from 3T3-L1 adipocytes as previously reported ( 34 ). Treatment with recombinant DPP-4 for 24 h increased gene expression and consecutive secretion of FABP4 in 3T3-L1 adipocytes, the effect of which was cancelled by sitagliptin. Furthermore, knockdown of DPP-4 in 3T3-L1 adipocytes decreased gene expression and consecutive secretion of FABP4. It has recently been reported that DPP-4 is one of the adipocyte-derived bioactive molecules known as adipokines, though the receptor for soluble DPP-4 remains obscure ( 35 ). Exogenous DPP-4 decreased insulin signaling in adipocytes, skeletal muscle cells, and smooth P = 0.096), HbA1c ( r = Ϫ 0.37, P = 0.083), or other variables.

Study 2
Treatment of differentiated 3T3-L1 adipocytes with sitagliptin, but not exendin-4, for 24 h decreased gene expression of FABP4 in a dose-dependent manner ( Fig. 2A , B ). Sitagliptin signifi cantly decreased gene expression of other adipogenic markers, including PPAR ␥ 2 and C/ EBP ␣ , in a dose-dependent manner (supplementary Fig.  1A, B). For assessing secretion of FABP4, 3T3-L1 adipocytes were treated with sitagliptin or exendin-4 for 2 or 24 h in the absence and presence of 10 µM isoproterenol, a pan-␤ -adrenergic agonist, known as an inducer of FABP4 secretion ( 9,10 ). Western blot analysis showed that FABP4 the present fi ndings indicate that pharmacological inhibition of the activity of soluble DPP-4 reduces serum FABP4 level by decreasing the expression of FABP4 in adipocytes ( Fig. 4 ). In addition, exogenous DPP-4 has been reported to promote adipogenesis through increased expression of PPAR ␥ , which is upstream of FABP4 regulation ( 36 ). Taken together, Western blot analysis of DPP-4 and GAPDH using the CL and nonfi ltered CM of 3T3-L1 adipocytes incubated for 24 h. B, C: Gene expression of FABP4 determined by quantitative real-time PCR in differentiated 3T3-L1 adipocytes treated with 0-500 ng/ml DPP-4 (B) or with 500 ng/ml DPP-4 and 10 µM sitagliptin (C) for 24 h (n = 3 in each group). * P < 0.05 versus DPP-4 (0 ng/ml). † P < 0.05 versus DPP-4 (500 ng/ml). AU, arbitrary units. D, E: Western blot analysis of FABP4 and GAPDH using the CL and fi ltered CM of 3T3-L1 adipocytes treated with 500 ng/ml DPP-4 and 10 µM sitagliptin for 24 h (D) or 2 h (E) (n = 3 in each group). FABP4 secretion was relatively expressed as densitometry of FABP4 in the CM divided by those of FABP4 in the CL and GAPDH in the CL. * P < 0.05 versus control. † P < 0.05 versus DPP-4. n.s., not signifi cant. F, G: Gene expression of DPP-4 (F) and FABP4 (G) determined by quantitative real-time PCR in DPP-4-knockdown 3T3-L1 adipocytes (n = 3 in each group). * P < 0.05 versus control. H: Western blot analysis of FABP4 and GAPDH using the CL and fi ltered CM of DPP-4-knockdown 3T3-L1 incubated for 24 h (n = 3 in each group). * P < 0.05 versus control.
Inhibition of DPP-4 leads to an increase in incretins such as GLP-1 and gastric inhibitory polypeptide (GIP) and thereby an increase in insulin secretion by pancreatic ␤ cells ( 29,30 ). GLP-1 has been reported to increase cyclic AMP levels downstream of GLP-1 receptor signaling and to subsequently increase insulin secretion by pancreatic ␤ cells ( 29,30,37 ). Notably, FABP4 is secreted from adipocytes in association with lipolysis via the ␤ -adrenergic receptor-adenyl cyclase-cyclic AMP pathway ( 10 ). However, expression of the GLP-1 receptor in tissues has been obscure because of a lack of reliable antibodies ( 38 ), and a recent study demonstrated that the GLP-1 receptor is expressed in the stromal vascular fraction but not in the adipocyte fraction of adipose tissue ( 39 ). Furthermore, a signifi cant difference has been found between pancreatic GLP-1 receptor-expressing cells and 3T3-L1 adipocytes in signal transduction activated by exendin-4, a GLP-1 receptor agonist ( 40 ). In the present study, exendin-4 did not modulate gene expression and secretion of FABP4 in 3T3-L1 adipocytes ( Fig. 2B, D, F ), though doses of exendin-4 were well above the reported dissociation constant ( K d ) value of the GLP-1 receptor ( 41,42 ). Hence, the present regulatory factors of lipolysis, sympathetic tone and infl ammatory cytokines, are potentially modulated by pharmacological inhibition of DPP-4. Hence, suppression of sympathetic nerve activation and/or infl ammatory cytokines may have been an additional mechanism by which sitagliptin reduced serum FABP4 levels as its pleiotropic effects ( Fig. 4 ).
Serum FABP4 level has been reported to predict longterm cardiovascular events (22)(23)(24). Together with accumulating evidence indicating a signifi cant role of FABP4 in insulin resistance and atherosclerosis ( 4-6 ), the decrease in FABP4 level caused by DPP-4 inhibitors might have a preventive effect of cardiovascular events, though this issue has not yet been critically proved by randomized controlled trial studies using DPP-4 inhibitors.
The present study has several limitations. First, the number of subjects enrolled in study 1 was small, and the possibility of type 1 error in statistical tests cannot be excluded. Second, most of the study subjects had been treated with several drugs at baseline, including angiotensin II receptor blockers ( 33,49 ) and a statin ( 50 ), which have been reported to affect circulating FABP4 concentration. Therefore, such drugs might have modulated the change in FABP4 level. Third, the present study lacked a placebo control group and a GLP-1 receptor agonist group in study 1. Hence, a prospective and placebo-controlled study with larger numbers of subjects is necessary for confi rming the impact of DPP-4 inhibitor treatment on circulating FABP4 level and for determining the clinical benefi t of a DPP-4 results suggest that reduction in serum FABP4 level is independent of a canonical GLP-1 receptor signaling in adipocytes ( Fig. 4 ).
In addition to its effect on pancreatic ␤ cells, activation of the GIP receptor has been reported to increase lipogenesis, enhance secretion of adipokines, and promote expansion of fat depots in adipocytes ( 30,43 ). The extrapancreatic effects of GIP possibly increase circulating FABP4 level because there is a close correlation between serum FABP4 level and adiposity mass ( 8,16 ). Hence, it is unlikely that augmentation of GIP by DPP-4 inhibition, if any, contributes to the reduction in FABP4 concentration by sitagliptin.
Change in the FABP4 level caused by sitagliptin was not signifi cantly correlated with change in the level of glucose, HbA1c, or other variables in the present study. However, 1,5-AG level was signifi cantly increased by treatment with sitagliptin, indicating an improvement of postprandial hyperglycemia and fl uctuation of glucose level in both the monotherapy and combination therapy groups. It has been suggested that augmented fl uctuation of glucose level causes sympathetic nerve activation ( 44 ), and sympathetic tone has been reported to be reduced by sitagliptin and vildagliptin, another DPP-4 inhibitor, in an obese rat model ( 45 ). DPP-4 inhibitors have also been shown to decrease several infl ammatory cytokines, including tumor necrosis factor ␣ and interleukin 6 ( 46,47 ), which are known to increase lipolysis in adipocytes ( 48 ). Secretion of FABP4 is associated with lipolysis ( 3,9,10 ), and two of the Fig. 4. Possible mechanisms of reduced FABP4 concentration caused by sitagliptin. FABP4 is secreted from adipocytes in association with lipolysis and may act as an adipokine for the development of insulin resistance and atherosclerosis, leading to cardiovascular events ( 3,9 ). Sitagliptin, a DPP-4 inhibitor, slows GLP-1 degradation, thereby increasing endogenous GLP-1 concentration. The GLP-1 receptor has been reported to be expressed in the stromal vascular fraction of adipose tissue but not in adipocytes ( 39 ). Furthermore, exendin-4, a GLP-1 receptor agonist, did not affect expression and secretion of FABP4, suggesting that reduction in serum FABP4 level by sitagliptin is independent of a canonical GLP-1 receptor signaling in adipocytes. DPP-4 is released from cell surface in adipocytes ( 34 ), and soluble DPP-4 (sDPP-4) has been suggested to act as an adipokine ( 35 ), though the receptor for sDPP-4 remains obscure. Direct inhibition of sDPP-4 as an adipokine by sitagliptin may contribute to a decrease in the expression of FABP4 in adipocytes, consequently leading to a reduction in serum FABP4 level. Another possible explanation for the reduction in serum FABP4 levels by sitagliptin is a decrease in lipolysis via suppression of sympathetic nerve activation and/or infl ammatory cytokines as pleiotropic effects of DPP-4 inhibitors.
inhibitor. Fourth, results obtained from in vitro experiments using mouse 3T3-L1 adipocytes, a well-used cell line of adipocytes, might not be directly extrapolated to human adipocytes. Finally, we cannot exclude the possibility that the effect of sitagliptin on serum FABP4 level was partly mediated by changes in expression and secretion of FABP4 in cells other than adipocytes. There has been accumulating evidence indicating FABP4 is expressed in several types of cells, in addition to adipocytes and macrophages, under both physiological and pathological conditions ( 3,(51)(52)(53)(54)(55), though the predominant contributors of circulating FABP4 level are adipocytes rather than macrophages or other types of cells ( 3,9 ).
In conclusion, treatment with sitagliptin decreases serum FABP4 concentration in patients with type 2 diabetes mellitus, at least in part, via reduction of FABP4 expression and its secretion from adipocytes by direct inhibition of DPP-4. Reduction of serum FABP4 level as a pleiotropic effect of DPP-4 inhibitors might be benefi cial for patients with metabolic and cardiovascular diseases. A further understanding of the mechanisms of FABP4 expression and secretion in adipocytes and their pharmacological modulation may enable development of new therapeutic strategies for cardiovascular and metabolic diseases.
The authors thank Ms. Shiho Ishikawa for technical help.