Adipose Dipeptidyl Peptidase-4 and Obesity Correlation with Insulin Resistance and Depot-Speciﬁc Release from Adipose Tissue in Vivo and in Vitro

Pathophysiology/Complications ORIGINAL ARTICLE

Adipose Dipeptidyl Peptidase-4 and Obesity Correlation with insulin resistance and depot-speciﬁc release from adipose tissue in vivo and in vitro

1 3 HENRIKE SELL, PHD WOLFRAM TRUDO KNOEFEL, MD metabolic disorders, with adipose tissue 2 4 MATTHIAS BLUHER€ , MD ARNE DIETRICH, MD 2 5,6 being a key player in this scenario (3). The NORA KLÖTING, PHD BARBARA A. FIELDING, PHD 1 7 current view of the role of expanded adi- RAPHAELA SCHLICH, PHD PETER ARNER, MD , 5 pose tissue in obesity identifies adipo- MIRIAM WILLEMS, MD1 3 KEITH N. FRAYN, PHD 3 1 kines as a potential link between obesity FLORIAN RUPPE, MD JURGEN€ ECKEL, PHD and insulin resistance (4). This link has stimulated a further characterization of OBJECTIVEd fi the adipocyte secretome by means of di- To study expression of the recently-identi ed adipokine dipeptidyl peptidase- fi 4 (DPP4) in subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) of patients with verse proteomic pro ling approaches, various BMIs and insulin sensitivities, as well as to assess circulating DPP4 in relation to obesity leading to the discovery of such novel and insulin sensitivity. adipokines as dipeptidyl peptidase-4 (DPP4) (5). RESEARCH DESIGN AND METHODSdDPP4 expression was measured in SAT and DPP4 is a transmembrane glycoprotein VAT from 196 subjects with a wide range of BMIs and insulin sensitivities. DPP4 release was and exoprotease that cleaves N-terminal measured ex vivo in paired biopsies from SAT and VAT as well as in vivo from SAT of lean and dipeptides from various substrates (6). obese patients. Circulating DPP4 was measured in insulin-sensitive and insulin-resistant BMI- Most importantly, DPP4 also cleaves and matched obese patients. inactivates the incretins glucagon-like RESULTSdDPP4 expression was positively correlated with BMI in both SAT and VAT, with peptide 1 (GLP-1) and gastric inhibi- VAT consistently displaying higher expression than SAT. Ex vivo release of DPP4 from adipose tory polypeptide. In this context, DPP4- tissue explants was higher in VAT than in SAT in both lean and obese patients, with obese inhibitors are in clinical use as antidiabetic patients displaying higher DPP4 release than lean controls. Net release of DPP4 from adipose drugs to improve glycemic control by tissue was also demonstrated in vivo with greater release in obese subjects than in lean subjects fi stimulating pancreatic insulin secretion and in women than in men. Insulin-sensitive obese patients had signi cantly lower circulating and suppressing glucagon production DPP4 than did obesity-matched insulin-resistant patients. In this experiment, DPP4 positively (7). We recently demonstrated that adipo- correlated with the amount of VAT, adipocyte size, and adipose tissue inflammation. cytes release DPP4 in a differentiation- CONCLUSIONSdDPP4, a novel adipokine, has a higher release from VAT that is particularly dependent manner (5). Circulating DPP4 pronounced in obese and insulin-resistant patients. Our data suggest that DPP4 may be a marker concentrations are increased in obese for visceral obesity, insulin resistance, and the metabolic syndrome. subjects and correlate with fasting plasma insulin, leptin, and adipocyte size in subcutaneous adipose tissue (SAT); however, the tissue source of circulating DPP4 besity is an increasing health issue has promoted research on the endocrine is not known. This study aimed to assess Oworldwide and an economical bur- link between lipid and glucose homeosta- DPP4 expression and release in paired bi- den, and as the hallmark of the sis, demonstrating that adipose tissue is opsies of SAT and visceral adipose tissue metabolic syndrome the obese state is an endocrine organ releasing various adi- (VAT) of lean and obese patients and of frequently associated with the develop- pokines. A complex interorgan crosstalk patients with or without impaired glucose ment of chronic diseases, including type 2 scenario between adipose tissue and other tolerance, as well as DPP4 release from ad- diabetes (1,2). The association between central and peripheral organs under- ipose tissue in vivo. Because circulating the epidemics of obesity and diabetes lies the progression of obesity-related DPP4 is increased in obese patients with the metabolic syndrome (5), we hypothe- ccccccccccccccccccccccccccccccccccccccccccccccccc sized that DPP4 expression and release in From 1Paul-Langerhans-Group Integrative Physiology, German Diabetes Center, Dusseldorf,€ Germany; the VAT are more prominent than in SAT and 2Department of Medicine, University of Leipzig, Leipzig, Germany; the 3Department of General, Visceral, that VAT DPP4 could be a marker for in- and Pediatric Surgery, Heinrich-Heine-University and University Hospital Dusseldorf,€ Germany; the sulin sensitivity. 4Department of Surgery, University of Leipzig, Leipzig, Germany; the 5Oxford Centre for Diabetes, En- 6 docrinology & Metabolism, Churchill Hospital, Oxford, U.K.; the Faculty of Health and Medical Sciences, RESEARCH DESIGN AND University of Surrey, Guildford, U.K.; and the 7Department of Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, Stockholm, Sweden. METHODS Corresponding author: Henrike Sell, [email protected]. Received 27 February 2013 and accepted 20 July 2013. Patients DOI: 10.2337/dc13-0496 © 2013 by the American Diabetes Association. Readers may use this article as long as the work is properly For all studies, protocols were approved cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ by local ethics committees. All partici- licenses/by-nc-nd/3.0/ for details. pants gave written, informed consent. care.diabetesjournals.org DIABETES CARE 1 Diabetes Care Publish Ahead of Print, published online October 15, 2013 Adipose DPP4 and obesity

DPP4 expression in paired biopsies explorative laparotomy, or gastric reagents (Life Technologies). From each from SAT and VAT. Paired samples of sleeve resection were selected and allo- RT-PCR, a 2-mL sample was amplified in a VAT and SAT were obtained from 196 cated to two experimental groups of 26-mL PCR with the Brilliant SYBR green Caucasians (97 men and 99 women) insulin-sensitive and insulin-resistant obe- QPCR Core reagent kit from Stratagene undergoing open abdominal surgery for sity with 30 subjects each, as described pre- (La Jolla, CA) according to the manufac- various reasons, including gastric band- viously (13). On the basis of the GIR in a turer’s instructions. Samples were incu- ing, cholecystectomy, appendectomy, euglycemic hyperinsulinemic clamp, pa- bated in the ABI PRISM 7000 sequence and weight-reduction surgery (8). Pa- tients were defined as either insulin sen- detector for an initial denaturation at tients with severe conditions, including sitive (GIR .70 mmol/kg z min) or insulin 958C for 10 min, followed by 40 PCR cy- generalized inflammation or end-stage resistant (GIR 60 , mmol/kg z min). Both cles, each cycle consisting of 958C for 15 s, malignant disease, were not included. groups were matched for sex, age, and 608C for 1 min, and 72C for 1 min. Human Age ranged from 24 to 86 years, and BMI. Clinical parameters were assessed as DPP4 mRNA expression was calculated rel- BMIrangedfrom20.8to54.1kg/m2. described previously (13). ative to the mRNA expression of 18s rRNA, Sixty-seven patients had impaired glu- In vivo release of DPP4 from SAT. all determined by premixed assays on de- cose tolerance or type 2 diabetes. Adipose Twenty-seven healthy volunteers (15 mand for DPP4 and 18s rRNA (Applied tissue specimens were frozen in liquid ni- women, 12 men; BMI 21–41.5 kg/m2; Biosystems). Amplification of specifictran- trogen immediately after excision. Clinical age 32–56 years) were recruited from the scripts was confirmed by melting curve parameters were assessed as described pre- greater Oxford community by advertise- profiles (cooling the sample to 688Cand viously (9). Insulin sensitivity was assessed ment or from the Oxford BioBank (14). heating slowly to 958C with measurement with the euglycemic-hyperinsulinemic None of the subjects were taking medica- of fluorescence) at the end of each PCR. clamp method (10) as previously de- tion known to affect lipid metabolism, and The specifi city of the PCR was further ver- scribed (11). In brief, after an overnight all were normoglycemic. Arteriovenous dif- ified by subjecting the amplification prod- fast and supine resting for 30 min, intra- ferences were measured across abdominal ucts to agarose gel electrophoresis. venous catheters were inserted into ante- SAT in the fasting state. A superficial epi- cubital veins in both arms. One was used gastric vein draining abdominal SAT (15) Measurement of DPP4 in serum and for the infusion of insulin and glucose; the and an arterialized dorsal hand vein (with conditioned medium other was used for the frequent sampling. the hand kept in a warming box at 608C) DPP4 release by adipose tissue explants After a priming dose of 1.2 nmol/m2 insu- were cannulated. The cannulas were kept and plasma or serum concentration were lin, the infusion with insulin (Actrapid patent with an intravenous infusion of determined by ELISA (R&D Systems, 100 U/mL; Novo Nordisk, Bagsvaerd, Den- 0.9% saline solution. Adipose tissue blood Wiesbaden, Germany). The assay was mark) was started with a constant infusion flow was measured in abdominal SAT as performed in duplicates according to the rate of 0.28 nmol/m2 body surface per described previously (16) and calculated manufacturer’s instructions. The intra- minute and continued for at least 120 from the washout of 133Xe, assuming a par- and interassay variations were 5.4 and min. After 3 min, a variable 20% glucose tition coefficient of 10 mL/g (17). After the 8.1%, respectively. infusion rate (GIR) was added. The GIR was subjects had rested for 45 min, blood sam- adjusted during the clamp to maintain the ples were taken simultaneously from the Statistical analysis blood glucose at 5.0 mmol/L. Bedside blood two sites. Samples were stored at 2808C Data are expressed as mean 6 SEM. The glucose measurements were performed until analysis. Arteriovenous differences Shapiro-Wilcoxon test was used to test every 5 min by means of the glucose dehy- were calculated as arterialized plasma the Gaussian distribution of biological pa- drogenase technique with Hemocue B DPP4 concentration 2 adipose venous rameters. The Student t test was used for (Hemocue, Angelholm, Sweden). plasma DPP4 concentration. Net release comparison between two groups, and In vitro release of DPP4 from paired was calculated as arteriovenous difference ANOVA followed by P for linear trend biopsies from SAT and VAT. SAT and 3 (21) 3 (1-hematocrit) 3 blood flow, posttest was used for comparison between VAT biopsies were obtained during plan- where there is a factor (21) that converts multiple groups. The Mann-Whitney test ned abdominal surgery (surgery for hernia, to net release and a factor (1 2 hematocrit) was used for variables that were not nor- gall bladder, and other noninflammatory that converts to whole blood concentra- mally distributed. Correlations were as- and nonmalignant causes) from 12 lean tions (assuming no DPP4 in erythrocytes). sessed with the Pearson correlation and 11 obese patients (BMI 22 6 2and Positive values for net release indicate a net coefficient or as indicated in the graphs. 38 6 3 kg/m2, respectively; age 62 6 6 release of DPP4 from adipose tissue in vivo; All statistical analyses were done with and 58 6 6 years, respectively). Adipose negative values indicate a net uptake. JMP statistics software (SAS Institute tissue specimens were immediately trans- Inc., Cary, NC) or Prism (GraphPad Soft- ferred to medium, and explants were gen- Measurement of DPP4 expression ware,Inc.,LaJolla,CA),consideringa erated as described earlier (12). Briefly, fat Human DPP4 mRNA expression was P value , 0.05 as statistically significant. explants were cultured in serum-free me- measured by quantitative real-time RT- dium. After 24 h, the conditioned medium PCR in a fluorescent temperature cycler RESULTS was collected and stored in aliquots at with the TaqMan assay, and fluorescence 2808C until further use. was detected on an ABI PRISM 7000 se- DPP4 expression correlates with BMI, Serum DPP4 and adipose DPP4 expres- quence detector (Applied Biosystems, is higher in VAT than SAT, and is sions in insulin-sensitive and insulin- Darmstadt, Germany). Total RNA was increased in VAT of lean patients resistant obese subjects. Sixty morbidly isolated with TRIzol (Life Technologies, with impaired glucose tolerance obese men and women (BMI 45 6 1.3 kg/m2) Grand Island, NY), and 1 mgRNAwas DPP4 expression in SAT and VAT was scheduled for elective cholecystectomy, reverse transcribed with standard measured in 196 individuals of various

2 DIABETES CARE care.diabetesjournals.org Sell and Associates

BMIs that were characterized by impaired inSATandVATofnormoglycemicpa- In vitro DPP4 release is most or normal glucose tolerance, as assessed tients was positively associated with sev- prominent in VAT from obese patients by oral glucose tolerance test. DPP4 ex- eral measures of obesity, including body Paired biopsies of SAT and VAT from lean pression in both depots correlated posi- fat, VAT and SAT areas, circulating leptin and obese subjects were used to study tively with BMI (Fig. 1A and B)and and interleukin 6, and the amount of DPP4 release in vitro. DPP4 release was correlated with each other (r = 0.28, macrophages in adipose tissue; it was significantly higher from VAT in both P , 0.0001). DPP4 expression was signif- negatively associated with circulating adi- lean and obese subjects (Fig. 2A). The icantly higher in VAT than in SAT. Fur- ponectin. With respect to the metabolic highest DPP4 release was measured from thermore, DPP4 expression significantly state of normoglycemic patients, DPP4 VAT of obese subjects. Dividing the group increased in both depots with increasing expression in adipose tissue was posi- of obese subjects in those with type 2 di- BMI from lean to obese subgroups (Fig. tively correlated with fasting plasma insu- abetes and those without, we observed a 1C). In lean individuals with impaired lin and HbA1c and negatively correlated significant increase of DPP4 release from glucose tolerance, DPP4 was significantly with GIR obtained during the euglycemic- VAT in those obese patients with type 2 increased in the VAT but not in the SAT hyperinsulinemic clamp. In addition, diabetes (BMI and age not different be- (Fig. 1D). This difference was not related DPP4 expression was associated with tween obese subgroups) (Fig. 2B). DPP4 to BMI or sex in either group. In over- total cholesterol, free fatty acids (FFAs), release from SAT only tended to be higher weight and obese subjects, there was no and triglycerides. In both normoglycemic in obese patients irrespective of their di- difference in DPP4 expression in either patients and patients with impaired glu- abetes state (data not shown). depot according to glucose tolerance cose tolerance, DPP4 expression in VAT (Fig. 1E). DPP4 mRNA in both depots positively correlated with VAT area In vivo release of DPP4 is higher in correlated with various clinical parame- and FFAs, whereas a negative correla- obese patients and in females ters and measures of adipose tissue in- tion was found for GIR and circulating Arteriovenous differences of DPP4 were flammation (Table 1). DPP4 expression adiponectin. measured across abdominal SAT in lean

Figure 1dDPP4 expressions in paired SAT (white bars) and VAT (black bars) of lean, overweight (OW), and obese patients in relation to obesity and glucose tolerance. DPP4 expression in adipose tissue was measured in 197 individuals. A and B: Linear regression analysis of DPP4 expression in SAT and VAT with respect to BMI. Statistical evaluation is indicated in each graph. C–E: Comparison of SAT and VAT DPP4 expressions in lean, overweight, and obese subjects (C); in lean subjects with (IGT) and without (NGT) impaired glucose tolerance (D); and in overweight or obese subjects with and without impaired glucose tolerance (E). *P , 0.05 compared with SAT; xP , 0.05 compared with VAT of subjects without impaired glucose tolerance. care.diabetesjournals.org DIABETES CARE 3 Adipose DPP4 and obesity

Table 1dPearson’s correlation coefficients of SAT and VAT DPP4 expressions with effects within adipose tissue and systemic clinical parameters and adipose tissue measures effects via the blood circulation. For a bet- ter understanding of the regulation of Normal glucose tolerance Impaired glucose tolerance DPP4 in humans with different degrees (n = 129) (n =67) of obesity and insulin sensitivity, in this study we measured DPP4 mRNA expres- Variable SAT DPP4 VAT DPP4 SAT DPP4 VAT DPP4 sion in adipose tissue and correlated it VAT area 0.38*** 0.73*** 0.11 0.54*** with clinical parameters and adipose tis- DPP4 SAT area 0.27** 0.54*** 0.31* 0.00 sue measures. expression is system- Waist-to-hip ratio 0.21* 0.50*** 0.01 0.15 atically lower in SAT irrespective of the Macrophages in VAT (%) 0.50** 0.73*** 0.20 0.19 body fat level, suggesting that there is a fi DPP4 Macrophages in SAT (%) 0.49** 0.61*** 0.19 0.41 depot-speci c control of expres- Body fat (%) 0.34*** 0.62*** 0.20 20.03 sion. The fact that circulating DPP4 and DPP4 Fasting plasma glucose (mmol) 20.04 20.03 20.19 20.29* expression in adipose tissue both Fasting plasma insulin (pmol) 0.27** 0.50*** 0.10 20.18 correlate with adipocyte size and adipose fl Oral glucose tolerance test (2 h) 0.07 0.05 0.04 20.01 tissue in ammation also suggests that fl GIR (clamp) 20.26** 20.57*** 0.12 20.27* proin ammatory adipokines released 2 2 from enlarged adipocytes could regulate HbA1c 0.22* 0.37*** 0.13 0.02 Cholesterol 0.22* 0.43*** 20.06 0.17 DPP4 release. fi DPP4 HDL 20.09 20.07 0.03 20.02 The ndings with expression in LDL 20.08 0.07 0.30* 0.07 adipose tissue in relation to BMI have fi Triglycerides 0.28** 0.50*** 0.20 0.09 been divergent, with a rstreporton DPP4 FFAs 0.17* 0.35*** 20.09 0.33** this subject demonstrating higher Leptin 0.36*** 0.48*** 0.24 0.14 expression in adipose tissue from obese Adiponectin 20.20* 20.40*** 20.14 20.38** patients than in that from lean controls Interleukin 6 0.24** 0.53*** 0.14 0.20 (18) and data from a second study describing higher DPP4 expression in lean *P , 0.05. **P , 0.01. ***P , 0.001. subjects than in obese ones (19). Together with our previous publication describing and obese subjects as described before obese patients. DPP4 expression was sig- DPP4asanoveladipokine(5),wenow (15). In 16 out of all 27 subjects studied nificantly elevated in both SAT and VAT of show in different groups of patients that there was net release of DPP4 from adi- insulin-resistant patients relative to insulin- both DPP4 mRNA expression and DPP4 pose tissue (Fig. 2C). This was related to sensitive subjects (Fig. 3A). DPP4 circu- protein levels are increased in both SAT the arterial DPP4 concentration, with the lating concentrations were significantly and VAT from obese subjects. By using net release of DPP4 was negatively corre- elevated in insulin-resistant versus insulin- different groups of patients, such as con- lated with the arterial level such that the sensitive subjects (Fig. 3B). There was secutive patients with a continuous spec- greatest net release of DPP4 was associ- no difference in circulating DPP4 in fe- trum of BMIs, as well as carefully ated with the lowest arterial concentra- males compared with males (data not characterized insulin-sensitive and insu- tion (Fig. 2C). In patients with lower shown). Serum DPP4 correlated with the lin-resistant morbidly obese patients, we arterial DPP4 (,288 ng/mL, n =14), amount of VAT, as did adipocyte size and can furthermore demonstrate that DPP4 obese subjects were characterized by a adipose tissue inflammation expressed as expression, especially in VAT, is nega- significantly higher net release of DPP4 the percentage of macrophages in VAT tively associated with insulin sensitivity than were lean patients (Fig. 2D), and (Fig. 3C–E). Furthermore, DPP4 corre- in both lean and obese subjects. there was a correlation between DPP4 related with insulin resistance because it To extend our understanding of how lease and BMI (Spearman rank correlation was negatively related to GIR and posi- DPP4 is not only expressed in adipose 0.72, P = 0.01, data not shown). Women tively associated with fasting insulin and tissue but also released from the tissue, we (n = 16) showed net release (P = 0.014 for HbA1c (Fig. 3F–H). also studied DPP4 release from adipose difference from no net release, as assessed tissue explants ex vivo. SAT biopsies from by Wilcoxon signed rank test) and CONCLUSIONSdWe recently iden- obese patients were characterized by showed significantly more release than tified DPP4 as a new adipokine that may higher DPP4 release than seen in those men (Fig. 2E). Among the women, be a missing link between increased from lean controls. This set of data cor- DPP4 release was significantly related to adipose tissue mass in obesity and obe- roborates our earlier study showing that BMI (r = 0.59, P = 0.021, data not shown). sity-associated metabolic diseases (5). Al- enlarged subcutaneous adipocytes from though much attention has focused on obese patients release higher amounts of Circulating DPP4 is lower in BMI- theroleofDPP4inthedegradationof DPP4 than do adipocytes from lean con- matched healthy obese patients and GLP-1, our earlier data suggest that trols (5). Additionally, we have now correlates with insulin resistance and DPP4 also exerts direct effects, as it is shown that adipose tissue explants from adipose tissue inflammation able to induce insulin resistance in adipo- VAT release more DPP4 than do SAT ex- DPP4 expression and circulating levels were cytes and skeletal muscle cells in concen- plants, pointing to a possible higher rela- also measured in healthy, insulin-sensitive trations that can be found in the tive contribution by VAT to circulating obese patients and compared with BMI-, circulation of overweight and obese sub- DPP4 levels. DPP4 release from VAT age-, and sex-matched insulin-resistant jects (5). DPP4 thus may also have local from obese patients is again higher than

4 DIABETES CARE care.diabetesjournals.org Sell and Associates

Figure 2dIn vitro DPP4 release from paired SAT (white bars) and VAT (black bars) samples from lean and obese patients and in vivo DPP4 release from abdominal SAT samples from lean and obese men and women. A and B: DPP4 release was measured in paired adipose tissue specimens from 12 lean and 11 obese (5 with type 2 diabetes) patients by ELISA. *P , 0.05 compared with respective SAT sample or between groups as indicated. C–E: DPP4 was measured in arterialized blood and adipose tissue venous blood in lean and obese patients, and the net release of DPP4 was calculated (n =27).C: Correlation of arterial DPP4 with DPP4 net release from adipose tissue, taking blood flow into account. Linear regression was performed with Spearman correlation. D: Net release of DPP4 from adipose tissue in lean and obese subjects with low arterial DPP4 (, 288 ng/mL, n =14). *P , 0.05 by Mann-Whitney test. E: Net release of DPP4 from adipose tissue in men and women. *P , 0.05 by Mann-Whitney test. ND, nondiabetic. that in VAT from lean controls. Although in association with fibrosis in expanding morbidly obese patients compared with DPP4 release is significantly increased adipose tissue in obesity may contribute insulin-resistant BMI-matched obese sub- from VAT, SAT constitutes the major ad- to increased DPP4 release and its depot- jects, showing that circulating DPP4 lev- ipose tissue in humans and therefore may specific differences (21,22). els and DPP4 expression in VAT are be the primary source of adipose-derived DPP4 expression and release are higher in insulin-resistant patients de- circulating DPP4, whereas VAT may con- higher in obese patients with the meta- spite similar adiposity. Some studies tribute to increased circulating levels in bolic syndrome and type 2 diabetes. have already attempted to explain why morbidly obese patients with insulin re- Animal data support the notion that DPP4 is overexpressed in adipose tissue sistance and more pronounced visceral DPP4 expression in adipose tissue increa- of patients with the metabolic syndrome. obesity. It is not yet known whether ses with developing type 2 diabetes, A first study analyzing single nucleotide higher release of DPP4 from VAT in obe- because rats with streptozotocin-induced polymorphisms revealed that DPP4 poly- sity simply reflects higher DPP4 expres- diabetes display significantly increased morphisms are probably not modulating sion and therefore higher presence at the DPP4 activity in epididymal fat (23). Fur- DPP4 expression and the association of plasma membrane or enzymes involved thermore, recent animal studies with DPP4 expression with cardiovascular in DPP4 shedding from the membrane DPP4 inhibitors support the notion that risk (26). Conversely, DPP4 expression are also regulated in obesity. The mecha- DPP4 may play a functional role within seems to be mediated by epigenetic ef- nisms of DPP4 release from the cell mem- adipose tissue, because DPP4 inhibition fects. Methylation levels of the DPP4 pro- brane are unknown and involve enzymes has been seen to prevent adipose tissue moter were found to be negatively not yet identified. It could be speculated inflammation and development of glu- associated with DPP4 mRNA expression that metalloproteinases and ADAMs (a cose intolerance in high fat diet–induced in VAT in obese women with and without disintegrin and metalloproteinase do- obesity (24,25). How DPP4 is regulated in the metabolic syndrome and also to be mains) could be shedding partners for this context is not known. We previously associated with HDL cholesterol (27). DPP4. In fact, the first so-called sheddase, demonstrated higher DPP4 release and Furthermore, three DPP4 polymor- ADAM17, is involved in tumor necrosis higher circulating DPP4 levels in obese phisms were found to be significantly factor-a processing at the cell surface, patients with the metabolic syndrome rel- associated with methylation levels. Inter- thereby leading to tumor necrosis factor- ative to obese controls (5). In this study, estingly, we observed here a significant a release into the circulation (20). Dereg- we further refined these findings in a well- increased DPP4 expression in VAT of ulation of metalloproteinases and ADAMs characterized cohort of insulin-sensitive lean patients with impaired glucose care.diabetesjournals.org DIABETES CARE 5 Adipose DPP4 and obesity

Figure 3dDPP4 expression in adipose tissue and circulating DPP4 in insulin-sensitive (IS) and insulin-resistant (IR) morbidly obese subjects. A: DPP4 expression in adipose tissue was measured in insulin-sensitive (n = 30) and insulin resistant (n = 30) obese subjects. *P , 0.05 compared with respective SAT (white bars) or between groups as indicated. Black bars, VAT. B: Circulating DPP4 concentrations in insulin-sensitive and insulin-resistant groups. *P , 0.05 compared with insulin-sensitive group. C–H: Linear regression analysis of DPP4 serum concentration and measures of adipose tissue morphology and inflammation or insulin sensitivity. Statistical evaluation is indicated in each graph. tolerance, suggesting that mechanisms soluble DPP4 are relatively low. When significantly increased not only in epididy- explaining higher DPP4 expression in concentrations are high, this fat depot can mal fat but also in the liver (23). In fact, obese subjects with the metabolic syn- extract DPP4 from the circulation. The other tissues in addition to adipose tissue drome could also be translated to lean pa- data are similar to those seen for steroid have been discussed as sources of circu- tients who have not yet fully developed hormones: for both estradiol and testos- lating DPP4 including the liver, vascular insulin resistance. It should be noted terone, uptake by abdominal SAT is seen cells, and immune cells (29,30). Data that this phenomenon was restricted to when arterial concentrations are high, but from humans also indicate that the liver VAT, which again illustrates the major dif- release when arterial concentrations are maybeaprimarysourceofDPP4inad- ferences between the SAT and VAT depots low (28). The data show a strong effect of dition to adipose tissue, because as pa- in conferring an increased metabolic risk. obesity on net release. Interestingly, albeit tients with nonalcoholic fatty-liver Data on in vivo DPP4 release measur- unexplained, DPP4 net release is higher in disease (NAFLD) have higher circulating ing arteriovenous differences suggest that women, although circulating DPP4 and DPP4 activity than do controls (31). It abdominal SAT is not the only source of adipose DPP4 expression are not different should be noted, however, that those circulating DPP4, although there is net release between males and females. In animals, NAFLD patients had significantly higher in most of the subjects. In fact, net release DPP4 expression and activity in rats BMI, and DPP4 was not adjusted for BMI is seen when the arterial concentrations of with streptozotocin-induced diabetes are in that study. In addition, the authors

6 DIABETES CARE care.diabetesjournals.org Sell and Associates reported that patients without NAFLD (Collaborative Project ADAPT, Contract No. 10. DeFronzo RA, Tobin JD, Andres R. Glu- but with type 2 diabetes had similar cir- HEALTH-F2-2008-201100; Integrated Pro- cose clamp technique: a method for culating DPP4 activity to that in controls, ject HEPADIP, Contract LSHM-CT-2005- quantifying insulin secretion and re- – which would not agree with our data. Pa- 018734), and European Union COST Action sistance. Am J Physiol 1979;237:E214 tients with type 2 diabetes were about 30 BM0602. This work was further supported E223 by the Kompetenznetz Adipositas (Com- 11. Bluher€ M, Unger R, Rassoul F, Richter V, years older on average than controls, petence Network for Obesity) funded by Paschke R. Relation between glycaemic however, and we know from our previous the Federal Ministry of Education and Re- control, hyperinsulinaemia and plasma study that circulating DPP4 levels decline search (German Obesity Biomaterial Bank; concentrations of soluble adhesion mole- with age (5). Additional studies are FKZ 01GI1128) and a grant from Deutsche cules in patients with impaired glucose needed to clarify whether DPP4 concen- Forschungsgemeinschaft the Clinical Research tolerance or Type II diabetes. Diabetologia trations in the circulation are predictive of group Atherobesity KFO 152 (project BL 833/ 2002;45:210–216 NAFLD independently of age and BMI. 1-1), both to M.B. The study was also sup- 12. Lamers D, Schlich R, Greulich S, Sasson S, Several studies indicate that the in- ported by grants from Swedish Research Sell H, Eckel J. Oleic acid and adipokines Council, EASD/Lilly, and the Novo Nordisk synergize in inducing proliferation and cretin system is altered in the obese state. fl The incretin effect is reduced in obese Foundation (PA). in ammatory signalling in human vascu- No potential conflicts of interest relevant to patients (32), and GLP-1 release and gas- lar smooth muscle cells. J Cell Mol Med this article were reported. 2011;15:1177–1188 tric inhibitory polypeptide metabolism H.S. wrote the manuscript, researched data, 13. Klöting N, Fasshauer M, Dietrich A, et al. are altered after a mixed meal test in obese and contributed to discussion. M.B. researched Insulin-sensitive obesity. Am J Physiol subjects (33). In this context, obesity and date and contributed to discussion. N.K., R.S., Endocrinol Metab 2010;299:E506–E515 circulating DPP4 levels may also affect ef- M.W., and B.A.F. researched data. F.R., W.T.K., 14. Tan GD, Neville MJ, Liverani E, et al. The ficacy of treatment with DPP4 inhibitors. and A.D. contributed to study design. P.A., K.N.F., in vivo effects of the Pro12Ala PPAR- Higher circulating DPP4 levels are associ- and J.E. contributed to discussion. All authors gamma2 polymorphism on adipose tissue ated with worse response to sitagliptin in reviewed and edited the manuscript. H.S. is the NEFA metabolism: the first use of the patients with type 2 diabetes inadequately guarantor of this work and, as such, had full Oxford Biobank. Diabetologia 2006;49: – controlled by metformin or sulfonylurea access to all the data in the study and takes re- 158 168 sponsibility for the integrity of the data and the (34). In addition, higher BMI may be pre- 15. Frayn KN, Coppack SW, Humphreys SM, accuracy of the data analysis. Whyte PL. Metabolic characteristics of dictive of less HbA1c lowering by sitaglip- human adipose tissue in vivo. Clin Sci tin treatment (35). It is not clear, (Lond) 1989;76:509–516 however, whether increased DPP4 release References 16. Samra JS, Frayn KN, Giddings JA, Clark from expanded adipose tissue is contrib- 1. James WP. WHO recognition of the global ML, Macdonald IA. Modification and uting to disturbances in the incretin sys- obesity epidemic. Int J Obes (Lond) 2008; validation of a commercially available tem and how in detail treatment with 32(Suppl. 7):S120–S126 portable detector for measurement of ad- DPP4 inhibitors is attenuated. 2. Wellen KE, Hotamisligil GS. In- ipose tissue blood flow. Clin Physiol In conclusion, we have demonstrated flammation, stress, and diabetes. J Clin 1995;15:241–248 – that DPP4 is an adipokine with signifi- Invest 2005;115:1111 1119 17. Jansson PA, Lönnroth P. Comparison of cantly higher expression in and release 3. Breitling R. Robust signaling networks of two methods to assess the tissue/blood the adipose secretome. Trends Endocrinol partition coefficient for xenon in sub- from VAT, with both parameters further Metab 2009;20:1–7 cutaneous adipose tissue in man. Clin increased in obese subjects. Circulating 4. Sell H, Dietze-Schroeder D, Eckel J. The Physiol 1995;15:47–55 DPP4 is higher in insulin-resistant than adipocyte-myocyte axis in insulin re- 18. Bouchard L, Tchernof A, Deshaies Y, et al. insulin-sensitive obese subjects, indicat- sistance. Trends Endocrinol Metab 2006; ZFP36: a promising candidate gene for ing that it may be a marker for insulin 17:416–422 obesity-related metabolic complications resistance independently of obesity. Fi- 5. Lamers D, Famulla S, Wronkowitz N, identified by converging genomics. Obes nally, DPP4 overexpression in VAT may et al. Dipeptidyl peptidase 4 is a novel Surg 2007;17:372–382 be a marker of adipose tissue inflamma- adipokine potentially linking obesity to 19. Kos K, Baker AR, Jernas M, et al. DPP-IV the metabolic syndrome. Diabetes 2011; inhibition enhances the antilipolytic action, which is known to be associated – with insulin resistance and the metabolic 60:1917 1925 tion of NPY in human adipose tissue. Di- 6. Yazbeck R, Howarth GS, Abbott CA. Di- – syndrome. Further work is needed to abetes Obes Metab 2009;11:285 292 peptidyl peptidase inhibitors, an emerg- 20. Black RA, Rauch CT, Kozlosky CJ, et al. A elucidate the functional role of DPP4 fl fi ing drug class for in ammatory disease? metalloproteinase disintegrin that releases within adipose tissue and to de ne Trends Pharmacol Sci 2009;30:600–607 tumour-necrosis factor-alpha from cells. whether higher DPP4 expression and 7. Deacon CF, Carr RD, Holst JJ. DPP-4 in- Nature 1997;385:729–733 serum concentration may contribute to hibitor therapy: new directions in the 21. Chavey C, Mari B, Monthouel MN, et al. higher efficacy of DPP4 inhibitors in treatment of type 2 diabetes. Front Biosci Matrix metalloproteinases are differen- patients with a high proportion of VAT. 2008;13:1780–1794 tially expressed in adipose tissue during 8. Bluher€ M, Bashan N, Shai I, et al. Activated obesity and modulate adipocyte differen- Ask1-MKK4-p38MAPK/JNK stress sig- tiation. J Biol Chem 2003;278:11888– AcknowledgmentsdThis work was sup- naling pathway in human omental fat 11896 ported by the Ministerium fur€ Wissenschaft tissue may link macrophage infiltration to 22. Divoux A, Tordjman J, Lacasa D, et al. und Forschung des Landes Nordrhein-Westfalen whole-body Insulin sensitivity. J Clin Fibrosis in human adipose tissue: com- (Ministry of Science and Research of the State of Endocrinol Metab 2009;94:2507–2515 position, distribution, and link with lipid North Rhine-Westphalia), the Bundesministerium 9. Berndt J, Klöting N, Kralisch S, et al. metabolism and fat mass loss. Diabetes fur€ Gesundheit (Federal Ministry of Health), Plasma visfatin concentrations and fat 2010;59:2817–2825 the German Research Council (SE 1922/2-1), depot–specific mRNA expression in hu- 23.KirinoY,SatoY,KamimotoT,KawazoeK, the Commission of the European Communities mans. Diabetes 2005;54:2911–2916 Minakuchi K, Nakahori Y. Interrelationship care.diabetesjournals.org DIABETES CARE 7 Adipose DPP4 and obesity

of dipeptidyl peptidase IV (DPP4) with 27. Turcot V, Bouchard L, Faucher G, et al. 32. Muscelli E, Mari A, Casolaro A, et al. the development of diabetes, dyslipidae- DPP4 gene DNA methylation in the Separate impact of obesity and glucose mia and nephropathy: a streptozotocin- omentum is associated with its gene ex- tolerance on the incretin effect in normal induced model using wild-type and pression and plasma lipid profile in severe subjects and type 2 diabetic patients. Di- DPP4-deficient rats. J Endocrinol 2009; obesity. Obesity (Silver Spring) 2011;19: abetes 2008;57:1340–1348 200:53–61 388–395 33. Carr RD, Larsen MO, Jelic K, et al. Secre- 24. Dobrian AD, Ma Q, Lindsay JW, et al. 28. Boulton KL, Hudson DU, Coppack SW, tion and dipeptidyl peptidase-4-mediated Dipeptidyl peptidase IV inhibitor sita- Frayn KN. Steroid hormone inter- metabolism of incretin hormones after a gliptin reduces local inflammation in ad- conversions in human adipose tissue in mixedmealorglucoseingestioninobese ipose tissue and in pancreatic islets of vivo. Metabolism 1992;41:556–559 compared to lean, nondiabetic men. J Clin obese mice. Am J Physiol Endocrinol 29. Cordero OJ, Salgado FJ, Nogueira M. On Endocrinol Metab 2010;95:872–878 Metab 2011;300:E410–E421 the origin of serum CD26 and its altered 34.AsoY,InukaiT,KasaiK.Serumlevelof 25. Shirakawa J, Fujii H, Ohnuma K, et al. concentration in cancer patients. Cancer soluble CD26/DPP-4 predicts the response Diet-induced adipose tissue inflammation Immunol Immunother 2009;58:1723– to sitagliptin, a dipeptidyl peptidase-4 and liver steatosis are prevented by DPP-4 1747 (DPP-4) inhibitor, in patients with type 2 inhibition in diabetic mice. Diabetes 30. Ussher JR, Drucker DJ. Cardiovascular diabetes inadequately controlled by met- 2011;60:1246–1257 biology of the incretin system. Endocr Rev formin and/or sulfonylurea (Abstract). Di- 26. Bouchard L, Faucher G, Tchernof A, 2012;33:187–215 abetes 2012;61(Suppl. 1):A546–A546 et al. Comprehensive genetic analysis of 31. Firneisz G, Varga T, Lengyel G, et al. Serum 35. Bando Y, Kanehara H, , Aoki H, Hisada K, the dipeptidyl peptidase-4 gene and dipeptidyl peptidase-4 activity in insulin Toya A, Tanaka D. Obesity may attenuate cardiovascular disease risk factors in resistant patients with non-alcoholic fatty the HbA1c-lowering effect of sitagliptin in obese individuals. Acta Diabetol 2009; liver disease: a novel liver disease bio- Japanese type 2 diabetic patients. J Di- 46:13–21 marker. PLoS ONE 2010;5:e12226 abetes Investig 2012;3:170–174

8 DIABETES CARE care.diabetesjournals.org