DPP4 IN CARDIOMETABOLIC DISEASE: RECENT INSIGHTS FROM THE LABORATORY AND CLINICAL TRIALS OF DPP4 INHIBITION

Abstract

The discovery of incretin-based medications represents a major therapeutic advance in the pharmacologic management of Type 2 diabetes (T2DM), as these agents avoid hypoglycemia, weight gain and simplify the management of T2DM. Dipeptidyl peptidase-4 (CD26, DPP4) inhibitors (DPP4i) are the most widely used incretin-based therapy for the treatment of T2DM globally. DPP4i are modestly effective in reducing HbA1c (≈0.5%) and while these agents were synthesized with the understanding of the role that DPP4 plays in prolonging the half-life of incretins such as glucagon like peptide-1 (GLP-1) and gastric inhibitory peptide (GIP), it is now recognized that incretins are only one of many targets of DPP4. The widespread expression of DPP4 on blood vessels, myocardium and myeloid cells and the non-enzymatic function of CD26 as a signaling and binding protein, across a wide range of species, suggest a teleological role in cardiovascular regulation and inflammation. Indeed, DPP4 is up regulated in pro-inflammatory states including obesity, T2DM and atherosclerosis. Consistent with this maladaptive role, the effects of DPP4 inhibition appear to exert a protective role in cardiovascular disease at least in pre-clinical animal models. Although 2 large clinical trials suggest a neutral effect on cardiovascular end-points, current limitations of performing trials in T2DM over a limited time horizon on top of maximal medical therapy, must be acknowledged before rendering judgment on the cardiovascular efficacy of these agents. This review will critically review the science of DPP4 and the effects of DPP4i on the cardiovascular system.

INTRODUCTION

The growing burden of Type 2 Diabetes Mellitus (T2DM) worldwide represents one of the most important challenges for global health. Over 50% of the risk of mortality from T2DM is attributable to cardiovascular (CV) causes, with T2DM contributing significantly to death and disability adjusted life years (DALY)¹. While diet and lifestyle approaches are fundamental to the treatment of risk in T2DM, these treatments often fail in many patients necessitating pharmacologic approaches. Dipeptidyl peptidase-4 (DPP4, also known as CD26) is a widely expressed glycoprotein that has gained attention owing to its role in the catalytic degradation of incretins such as glucagon-like peptide-1 (GLP-1) and as a receptor for the Middle Eastern Respiratory Syndrome (MERS) virus. The development of DPP4 inhibitors (DPP4i) as a class of anti-diabetic medications was predicated on the simple notion that these drugs would raise GLP-1/GIP levels resulting in enhancement of insulinotropic effects of glucose. This rather simple construct is now replaced with a much more nuanced understanding of this protein. In this review, we will summarize the structure and function of DPP4 and its known role in physiology. We will review its importance in the pathophysiology of cardiometabolic disorders and provide the evidence to date, on the effects of DPP4 inhibition, both from the context of experimental models, mechanistic human studies and recent clinical trial evidence on CV outcomes.

OVERVIEW OF DPP4 BIOLOGY AND REGULATION

Human DPP4 is a 766 amino acid single-pass type II integral transmembrane glycoprotein that belongs to S9b dipeptidyl peptidase family which also include quiescent cell proline dipeptidase (QPP, also called DPP2), fibroblast activation protein (FAP), DPP8, and DPP9. Monomeric DPP4 has a short N-terminal cytoplasmic portion (6 residues, AA 1–6), a 22-residue transmembrane domain (AA7–29) and an extra-cellular domain, comprised of an eight-blade β-propeller (Arg54-Asn497) and a large α/β-hydrolase domain (Gln508-Pro766) responsible for binding to adenosine deaminase (ADA) and matrix proteins such as fibronectin and collagen^2,3. Residues 630, 708 and 740 in the α/β-hydrolase domain are critical for catalytic function of DPP4. While the C terminal α/β-hydrolase domain is relatively conserved, the N-terminal eight-blade β-propeller domain demonstrates sequence variability. The catalytic pocket responsible for cleavage of a number of protein substrates is about 8 Å and contained within a 15Å wide opening at the interface between its hydrolase and propeller domains⁴. Residue 294 and residues 340–343 within the cysteine-rich segment of DPP4 is essential for ADA binding⁹. DPP4 exists as a monomer, homodimer, or homotetramer on the cell surface, with the homodimer representing the predominant catalytically active form⁷. DPP4 is subject to post-translational modification via glycosylation and N-terminal sialytion both of which have been suggested to regulate catalytic activity^7,8.

Current understanding of DPP4 molecular regulation is incomplete. The promoter of DPP4 contains a consensus GAS (interferon gamma-activated sequence) sequence at −35 to −27, which has a binding motif for STAT1α. Administration of both interferons and retinoic acid results in the tyrosine phosphorylation of STAT1α, and subsequent nuclear translocation and binding to the GAS motif resulting in DPP4 transcription⁹. In addition to IL-12 and TNFα, IL-1α has also been shown to regulate the expression of DPP4¹⁰. The promoter for DPP4 also contains consensus sites for nuclear factor-κB, SP-1, EGFR, AP-1 factor and hepatocyte nuclear factor-1. IL-12 enhances the translation but not transcription of DPP4 in activated lymphocytes, while TNFα decreases cell surface expression of DPP4, suggesting a regulatory role of IL-12 and TNFα in the translation and translocation of DDP4¹¹.

DPP4 also circulates as a soluble form in the plasma. Soluble DPP4 (sDPP4) lacks the cytoplasmic and transmembrane domain with preserved catalytic activity. Whether sDPP4 is cleaved from the membrane or is secreted is unclear. For instance, studies investigating viral liver infection suggest that sDPP4 is cleaved from membrane bound DPP4¹². sDPP4 is also detected in the lumen of secretory granules in pancreatic A cells and in the exocytic secretory lysosomes of natural killer cells suggesting that it may also be processed intracellularly^13,14.

PHYSIOLOGICAL ROLE OF CATALYTIC FUNCTION OF DPP4

DPP4 exerts its peptidase function by removing N-terminal dipeptides displaying proline, alanine or serine as the penultimate (P1) amino acid residue. Inactivation of gastric inhibitory polypeptide (GIP) and GLP-1 is responsible for the anti-hyperglycemic effect of DPP4 inhibition. Mice lacking the gene encoding DPP4 are refractory to the development of obesity and hyperinsulinemia and demonstrate improved post-prandial glucose control^14,15. GLP-1 and GIP suppress glucagon release, decreases gastric empting, promotes β-cell proliferation, and suppresses β-cell death^16–18. Pharmacological inhibition of DPP4 enzymatic activity improves glucose tolerance in wild-type, but not in Dpp4^−/− mice. Interestingly DPP4 inhibition improves glucose tolerance in Glp1r^−/− mice, indicating that DPP4 contributes to blood glucose regulation by additional substrates such as GIP or through GLP-1R-independent mechanisms¹⁵. In addition to gut derived peptides GLP-1 and GIP, the other substrates include a variety of neuropeptides and chemokines. A recent study suggests in addition to chemokines other cytokines such as GM-CSF, G-CSF, IL-3, and Erythropoietin (Epo) could also be cleaved by DPP4¹⁶. The catalytic activity of DPP4 and its substrates (Figure 1) have been extensively reviewed elsewhere^7,17 and we will not go into detail here.

Figure 1. DPP4 functions and structure.

DPP4 consists of a 6-amino-acid cytoplasmic tail, a 22-amino-acid transmembrane domain and a large extracellular domain. The extracellular domain is responsible for the dipeptidase activity and binding to its ligands such as adenosine deaminase (ADA) and fibronectin. AA, amino acid; ADA, adenosine deaminase. Some concepts of this figure were adapted from Zhong J et al. Atherosclerosis. 2013;226:305–314 with modifications.

PHYSIOLOGY OF NON-CATALYTIC FUNCTION OF DPP4

In addition to its well-known peptidase activity, DPP4 also possesses non-catalytic functions through its interaction with a ligands including ADA, caveolin-1, fibronectin, and CXCR4.

Interaction with ADA and Mediation of T cell Co-stimulation

The best known non-catalytic function of DPP4 is to provide co-stimulation for T cells by interacting with ADA¹⁸. Activation of DPP4 induced tyrosine phosphorylation of molecules downstream of TCR/CD3, such as p56^1ck, p59^fyn, ZAP-70, MAP kinase, c-Cbl, and phospholipase Cγ¹⁹. Interestingly, murine and rat DPP4 do not bind ADA²⁰. By using site-directed mutagenesis, Leu-294 and Val-341 were identified as two ADA-binding sites in Human DPP4⁶. Leu-294 and Val-341 are positioned at the outer strand of the tetramerization blade IV and blade V, respectively²⁰. Therefore, ADA binding has been thought to interfere with tetramerization. Similarly, the glycosylation of Asn281 is likely to influence ADA binding^3,20. Thus tetramerization and glycosylation may serve as control mechanisms for ADA binding and may explain why murine DPP4, which lacks glycosylation sites and may not bind ADA.

Novel Roles for DPP4

In 2013, DPP4 was identified as a functional receptor for the spike protein of a novel betacoronavirus species, the Middle Eastern Respiratory Syndrome Coronavirus (MERS-CoV) in human and bat cells²¹. The engagement of the MERS-CoV spike protein S with CD26 mediates viral attachment to host cells initiating infection. The viral receptor-binding domain recognizes blades IV and V of the DPP4 β-propeller²². The residues identified in the virus-DPP4 interface are also involved in ADA binding²². As the ADA-DPP4 interaction has been shown to induce co-stimulatory signals in T cells, this may indicate a possible manipulation of the host immune system by MERS-CoV through competition for the ADA-recognition site. DPP4 receptor binding domain may thus represent a potential treatment strategy for MERS-CoV infection²³.

DPP4 on T cells may also interact with caveolin-1 on APCs, resulting in its phosphorylation and leads to activation of downstream NFκB^24,25. Activated NFκB in turn upregulates CD86^26,27. This process has been suggested to be involved in the pathogenesis of inflammatory disorders²⁸. DPP4 has been reported to bind multiple components of extracellular matrix such as collagen, fibronectin, and the HIV-1 Tat protein^2,17. Interactions with matrix components may play a role in sequestration of DPP4 and allow additional functions such as matrix remodeling, metastasis and chemotaxis.

DPP4 IN DIABETES

DPP4 Expression and Diabetes

DPP4 is an important regulator of post-prandial glucose via degradation of GLP-1 and GIP¹⁷. Both GLP-1 and GIP are rapidly inactivated by DPP4, resulting in a short half-life (less than 2 min for GLP-1, and less than 2 min in rodents or 7 min in human for GIP)^29–31. In contrast to early reports suggesting that GLP-1 levels are decreased in T2DM, recent studies suggest that, in general, T2DM patients do not exhibit reduced GLP-1 secretion in response to an OGTT or meal test. However deteriorating glycemic control may be associated with reduced GLP-1 secretion³². DPP4 enzymatic activity correlates with the degree of glucose homeostasis in T2DM³³. DPP4 expression in both visceral and subcutaneous adipose tissue positively correlates with BMI^34,35, with visceral adipose tissue (VAT) consistently displaying higher expression³⁶. DPP4 expression positively correlates with the amount of VAT, adipocyte size, inflammation and HbA1c and negatively correlates with glucose infusion rates during euglycemic-hyperinsulinemic clamp³⁶. Within adipose, adipocytes appear to secrete abundant DPP4 with levels exceeding that from macrophages³⁵. Ex vivo release of DPP4 from adipose tissue explants were higher in VAT than in subcutaneous adipose tissue with obese patients displaying higher DPP4 release than lean controls. Secreted DPP4 may in turn exert paracrine effects on insulin signaling in adipocytes and skeletal muscle cells³⁵. Amongst inflammatory cells, T cell expression exceeds that of macrophages and dendritic cells³⁷. Expression of DPP4 increases in peripheral T cells and in the liver of patients with non-alcoholic fatty liver disease with levels correlating with insulin resistance^38,39. Interestingly, several widely-used glucose-lowering medications such as metformin and thiazolidinedione have been reported to reduce circulating DPP4^40–42 and DPP4 on T cells³⁸. DPP4i also appear to exert effects on brown adipose tissue metabolism by increasing UCP-1 and PGC-1a levels, through increase in GLP-1⁴³.

Catalytic Inhibition of DPP4 Function in Diabetes

There are currently four DPP4i approved by the FDA and one DPP4i approved by EU: sitagliptin (FDA approved in 2006), saxagliptin (FDA approved in 2009), linagliptin (FDA approved in 2011), alogliptin (FDA approved in 2013), and vildagliptin (EU approved in 2007). They can be broadly divided into two classes based on structure: DPP4 dipeptide structure mimetics and non-peptidomimetics. The first class includes sitagliptin (β-amino acid-based), vildagliptin, and saxagliptin (nitrile containing), whereas the non-peptidomimetics include alogliptin (modified pyrimidinedione) and linagliptin (xanthine-based). There are several others in advanced clinical trials including dutogliptin⁴⁴ and gemigliptin and at least 2 once weekly DPP4i are in advanced clinical trials. DPP4i demonstrate approximately a 0.4–0.8% lowering of HbA1C, with the degree of lowering depending on the baseline HbA1c, concomitant therapy and patient population^45,46. In general the glycemia lowering efficacy with DPP4i is lower than that with sulfonylureas, insulin and thiazolidinediones, but they are significantly better tolerated and devoid of weight gain^47–49. The most frequently reported side effects of DPP4i include nasopharyngitis, upper respiratory tract infection, urinary tract infection and headache. In contrast to other members of the class, linagliptin has a primarily non-renal route of excretion and does not need dose adjustment in renal impairment⁵⁰. Several studies corroborate improvement in β cell function indices including homeostasis model assessment beta cell function index (HOMA-b), fasting proinsulin:insulin ratio, and insulin secretion rates in response to mixed meal challenge^51–54.

Non-catalytic Interactions of DPP4 and Inflammation

The mechanism by which DPP4 may potentiate inflammation in T2DM likely involves both its catalytic and non-catalytic function. There is good evidence suggesting a role for DPP4 non-catalytic function in the regulation of T cell activation^5,18,25,27. DPP4 expression increases with functional maturation of dendritic cells (DCs) and macrophages from monocytes with a role for upregulation of CD86 expression and activation of the nuclear factor-κB pathway^24,26,37. The interaction between DPP4 and ADA may also facilitate T cell activation through regulation of pericellular adenosine levels. Jurkat cell (a T cell line) with a DPP4 mutant devoid of ADA binding activity is sensitive to adenosine-mediated inhibition of T cell proliferation⁵, indicating that DPP4 on immune cells facilitates adenosine clearance by binding to ADA (Figure 2). In assays using T-cell/DC co-cultures, adenosine dose-dependently suppresses T-cell proliferation, while pre-incubation of DCs with ADA relieved the inhibitory effect of adenosine in a dose-dependent manner. siRNA knockdown of DPP4 on DC suppresses proliferation of both CD4⁺ and CD8⁺ T cells while sitagliptin did not³⁷. Both T-cell receptor signaling and Signal transducer and activator of transcription 3 (STAT3), an important signaling molecule that regulates T-cell proliferation and activation, are regulated by DPP4-expressing DC’s³⁷. STAT3 phosphorylation increases on T-cell activation, an effect suppressed by adenosine³⁷. Lymphocyte-specific protein tyrosine kinase (Lck), a Src family tyrosine kinase that is activated with TCR activation and allows binding and phosphorylation of ZAP-70 to mediate downstream signaling events, is prevented from being activated in T cells co-cultured with DPP4 siRNA-transfected DC [as suggested by an increase in phospho-Lck (Tyr505), the inactive form of Lck]³⁷. These results suggest that DPP4 on antigen presenting and T cells may promote adipose inflammation and insulin resistance through its non-catalytic function. However, the precise extent to which non-catalytic function regulates inflammation independent of the catalytic function in T2DM will require experiments in animals expressing human DPP4 that additionally have mutations in enzymatic function as murine DPP4 lacks the ADA binding site^20,55.

Figure 2. Mechanisms by which DPP4 modulates immune response of relevance to cardiovascular disease.

DPP4 regulates immune response via multiple mechanisms that involve both adaptive and innate immune function. These include: 1) Control of pericellular adenosine levels: adenosine triphosphate (ATP) or ADP is converted into AMP by membrane-bound CD39. AMP is further catalyzed by CD73 and produce adenosine. Adenosine is degraded by ADA bound to membrane-anchored DPP4. Accumulation of adenosine suppresses T cell activation and proliferation; 2) ADA-DPP4 interaction provides co-stimulatory signaling for T cell activation. By signaling through CD45, DPP4 enhances T cell receptor signals to promote T cell activation; 3) Inactivation of Glucagon-Like Peptide-1 (GLP-1). GLP-1 binds to GLP-1R and increases cAMP resulting in Protein Kinase A (PKA) mediated suppression of ERK, JNK, and NFκB; 4) Activation of APCs by interaction with caveolin-1. DPP4 binds to caveolin-1 and activates IRAK and NFκB, leading to the activation of APCs.

EFECTS OF DPP4 INHIBITION ON CARDIOVASCULAR SYSTEM

Effects of DPP4i on Endothelial Cells, NO and Blood Pressure

DPP4 is expressed on endothelial cells and appears to play a physiologic role in the regulation of vascular tone and angiogenesis⁵⁶. Studies have suggested DPP4/incretin axis has substantial implication in cardiovascular system (Figure 3). Incubation of human umbilical vein endothelial cells (HUVECs) with DPP4i such as alogliptin and vildagliptin has been shown to result in eNOS and Akt phosphorylation paralleled by a rapid increase in NO^57,58. DPP4 inhibition relaxes pre-constricted aortic segments in a dose dependent manner, responses unaffected by the GLP-1R antagonist, exendin 9–39⁵⁷. In one study, vascular relaxation was reduced by endothelial denudation, L-NG-monomethyl-arginine citrate (L-NMMA) and by the soluble guanylate cyclase inhibitor ODQ, while a combination of L-NMMA, and blockers of calcium activated potassium channels completely obviated relaxation in response to DPP4 inhibition, suggesting that both eNOS and potassium channels mediated effects on vascular function⁵⁷. Inhibition of Src kinase decreased eNOS and Akt phosphorylation in response to DPP4-inhibition NO^57,58. DPP4-inhibition-mediated GLP-1R activation may also represent a mechanism for the blood pressure-lowering effect of DPP4i⁵⁹. However the locus of GLP-1R activation is controversial. Although endothelial cells have been shown to express GLP-1R at least in some studies, at least in one study the antihypertensive effects of GLP-1R activation, by liraglutide did not directly relax preconstricted aortic rings or increase cyclic GMP (cGMP). However, conditioned medium from liraglutide-treated hearts relaxed aortic rings in an endothelium-independent, GLP-1R-dependent manner. The decrease in BP involved an indirect natriuretic effect dependent on ANP release from the atria in a GLP-1R dependent manner⁵⁹. Several other studies in animal models support a favorable effect of DPP4 inhibition in improving endothelial function and blood pressure^57,60,61. In contrast to the positive effects in animals, the effects of DPP4i on endothelial function in humans has been inconsistent^62–64. While some of the inconsistency may relate to differences in end-points, methodology and patient characteristics, it is possible that these conflicting results may reflect differential alteration in the levels of substrates and their metabolites by DPP4i, which may vary depending on the study participants. In a double blind study in T2DM, treatment with vildagliptin for 4 weeks improved forearm blood flow in response to intra-arterially delivered aceytylcholine⁶². In contrast, 2 studies using flow-mediated dilation (FMD) of the brachial artery have shown opposite results with one study showing that both sitagliptin and alogliptin worsen FMD, while in another study sitagliptin improved FMD in association with an increase in CD34⁺ cells in another study^63,64.

Figure 3. Cardiovascular Effect of DPP4/Incretin Axis.

ADA, adenosine deaminase; ANP, atrial natriuretic peptide; CM, chylomicrons; DPP4, dipeptidyl peptidase 4; FFAs, free fatty acids; NO, nitric oxide; sDPP4, soluble DPP4

While earlier studies in small cohorts appear to suggest a favorable effect of DPP4i in reducing BP these effects have not been consistently observed^65,66. An obvious explanation is that the magnitude of effect which is typically small (1–3 mm Hg) may not be easily observed in studies utilizing clinic systolic blood pressure, particularly as a secondary end-point. Indeed a large meta-analysis of published data on DPP4i utilizing clinic BP does not support an effect of on blood pressure⁶⁷. An alternate hypothesis that has been proffered relates to differential alteration of substrates of DPP4 such as NPY, which shifts the receptor affinity from Y1 to non-Y1 receptors⁶⁸. DPP4 inhibition has been demonstrated to enhance sodium excretion through interaction with the renal NHE3 exchange protein and does so in both wild-type and Glp1r^−/− mice⁶⁹. The mechanisms may relate to the physical association of the sodium hydrogen exchanger type 3 (NHE3) with DPP4 with inhibition of DPP4 catalytic activity suppressing NHE3-mediated NaHCO₃ reabsorption in rat renal proximal tubule, resulting in increased Na⁺ excretion⁷⁰. Further the redistribution of the complex of DPP4-NHE3 is believed to represent an adaptive mechanism in chronic hypertension⁷¹. It has also been suggested that catalytic inactivation of natriuretic peptides such as brain natriuretic peptides by DPP4 may regulate blood pressure through natriuretic and vasodilatory effects. DPP4 converts BNP to produce BNP(3–32)⁷², which has reduced vasodilation and natriuretic effects⁷³. In humans however inhibition of DPP4 with sitagliptin acutely does not potentiate BNP levels or its effects on forearm blood flow⁷⁴. In a post-hoc analysis of SAVOR-TIMI, NT-proBNP levels were not increased with Saxagliptin treatment but were lower when compared to placebo⁷⁵.

DPP4 Inhibition in Angiogenesis and Stem Cell Homing

Expression of DPP4 on Sca-1⁺ lin⁻ donor hematopoietic cells negatively regulates homing and engraftment vasculogenesis⁷⁶. By inhibiting DPP4 catalytic function or deleting DPP4, the transplantation and engraftment efficacy of hematopoietic stem cells was greatly enhanced⁷⁶. SDF-1 is a substrate for DPP4 and has been implicated in the mobilization and homing of hematopoietic cells in response to G-CSF treatment in experimental ischemia/infarction^77,78. DPP4-truncated SDF-1 not only loses its chemotactic activity, but also blocks chemotactic effects of full-length SDF-1⁷⁹. Sitagliptin treatment in patients with T2DM results in a 2-fold increase of circulating EPC with concomitant increase in plasma SDF-1⁸⁰. Short term treatment with DPP4i has also been demonstrated to increase SDF-1 levels and CD34⁺ cells^80,81. The increase in CD34⁺ cells corresponded to increased homing and deposition to an ischemic hind limb preparation⁸¹ and infarcted myocardium⁸². DPP4 may also regulate HSCs and HPCs by truncating multiple CSFs (other than SDF-1) with consequent loss of their activity. DPP4 knockout or pretreatment of HPCs from human cord blood or mouse bone marrow with DPP4i enhances the proliferative action of GM-CSF, G-CSF, IL-3, and Erythropoietin (Epo)¹⁶. DPP4 deficiency or catalytic inhibition promotes hematopoiesis and bone marrow engraftment in mice after radiation or chemotherapy¹⁶. Interestingly, DPP4-truncated CSFs blunts the activity of their respective full-length CSF, both in vitro and in vivo with the truncated GM-CSF, demonstrating a higher affinity to GM-CSF receptor¹⁶. However, there were also studies showing DPP4 inhibition reduces angiogenesis through inactivation of NPY(1–36). Truncation of NPY by DPP4 leads to a shift of receptor subtype specificity with cleaved NPY(3–36) binding to non-Y1 (Y2, Y3, and Y5) receptors⁸³. Production of NPY(3–36) is required for angiogenic activity as DPP4 inhibition by neutralizing antibody suppresses NPY-mediated endothelial cell migration in an endothelial wound assay⁸⁴.

Pharmacologic inhibition of DPP4 catalytic function stimulates angiogenesis, with enhances endothelial cell migration, aortic sprouting and angiogenesis in in vivo assays^58,85. Src kinase mediated eNOS-Akt activation in response to DPP4 inhibition appears to play an important role in the angiogenic response to DPP4 inhibition⁸⁶. The effects of DPP4 inhibition in enhancing angiogenesis in-vitro are paralleled by an improvement in hind limb blood flow in a hind-limb ischemia model⁵⁸. DPP4 inhibition has also been shown to improve healing of chronic foot ulcer in T2DM patients, in parallel with increased HIF-1α and VEGF in the periphery of the ulcer and ulcer capillary density⁸⁷. Similarly the angiogenic capacity of circulating pro-angiogenic cells in T2DM patients has been shown to be increased in-vivo in response to DPP4 inhibitors⁸⁸.

Effects on Lipid Metabolism

Administration of DPP4i to hyperlipidemic mice reduces post-prandial lipoprotein levels⁸⁹. The effects of DPP4i on lipoprotein metabolism been reviewed extensively^90,91. A single dose of sitagliptin following a high-fat liquid formula administered under pancreatic clamp conditions, decreased triglyceride (TG)-rich lipoproteins apoB-48 concentration by reducing production rates in healthy subjects⁹². The magnitude of lipid lowering seen with different DPP4i appears to be comparable and of a similar magnitude to that seen with the GLP-1R agonist exenatide⁹¹. In a 4-week single center, randomized double blind study in T2DM, vildagliptin (n=31, 50 mg bid) reduced post-prandial plasma TG, chylomicron apoB-48 following mixed meal challenge, without changes in apoB-100. Vildagliptin also increased intact GLP-1, suppressed inappropriate glucagon secretion⁹³. In a randomized cross over study, alogliptin (25 mg/d, for 7 days) suppressed post-prandial elevation of TG, ApoB-48 and remnant lipoproteins⁹⁴. In a double-blind cross-over study, sitagliptin over 6 weeks (100 mg/day) significantly decreased post-prandial area plasma apoB-48, TG and VLDL-C⁹⁵. In a randomized cross over study over 7 days of vildagliptin vs. placebo, with microdialysis catheter based analysis of metabolites, vildagliptin treatment was associated with a reduction in post-prandial lactate and pyruvate production in the skeletal muscle with evidence of enhanced lipolysis. These effects may reflect improvement in hepatic glucose production by the liver due to obviation of insulin resistance by DPP4i and was supported by calorimetry suggesting increased energy expenditure and lipid oxidation rates⁹⁶.

Effects on Atherosclerosis

A number of DPP4i have been shown to reduce atherosclerosis in experimental models (Table I)^97–101. These studies have all consistently noted a reduction in foam cell formation and inflammatory content. In at least 2 of these studies a reduction in migratory capacity of monocytes has been implicated^97,98. Pre-treatment with DPP4i for 2 weeks in ApoE^−/− mice, markedly reduced the migration of labeled bone marrow derived monocytes to atherosclerotic plaque in response to exogenously administered TNFα and sDPP4⁹⁸. In-vitro Boyden chamber assays using alogliptin and sitagliptin performed in the presence of nanomolar concentration of TNFα confirmed reduced monocyte migration in response to a CCL-2 gradient, which was not inhibited by GLP 9–39⁹⁸. The effects of DPP4i include reduction in inflammatory gene expression and a reduction in monocytes and T cells^98,100,102. Given the effects of DPP4 in co-stimulation and regulation of adenosine, it may suggest a role for DPP4, independent of its effects on substrates such as GLP-1⁹⁷. GLP-1 has been shown to decrease monocyte migration in response to CCL-2 and regulated on activation, normal T cell expressed and secreted (RANTES) in a concentration-dependent manner⁹⁷. GLP-1 may also exert anti-inflammatory effects via cAMP dependent mechanisms which may reduce NFkB and Erk activation⁶⁰. Infusion of exendin-4 for 4 weeks reduces neointimal formation¹⁰³, foam cells¹⁰⁴, and atherosclerotic lesion size^104,105. In support of a role for GLP-1 in DPP4 mediated effects, administration of antagonists to GLP-1 and GIP [Exendin(9–39) and Pro(3)GIP] attenuated anti-atherosclerotic effects of vildagliptin on atherosclerosis⁹⁹ (Table I). In humans, one study noted a reduction in carotid intima media thickness progression with sitagliptin compared to placebo over 12 months in newly diagnosed patients with type 2 diabetes with coronary artery disease¹⁰⁶.

Table 1.

Studies that evaluate CV effect of DPP4i

Intervention	Dose	Duration	Subject	Disease	Major findings	Reference
Angiogenesis
Diprotin A or Val-Pyr	5mM Diprotin A or Val-Pyr	15 min in vitro treatment	Mouse	Bone marrow transplantation	DPP4 inhibition or deletion improved hematopoietic stem cell homing and engraftment.	Christopherson KW et al. 200476
Diprotinin-A or P32/98	10 µM diprotinin-A or P32/98	3–4 days	Mouse, HMVEC	Angiogenesis	Pharmacological inhibition of CD26/DPP4 enhanced endothelial growth both in vitro and in vivo.	Takasawa W 201085
Sitagliptin	10 or 20 mg/kg/d (oral gavage)	7 weeks	Mouse	Hind limb ischemia	Sitagliptin treatment augmented ischemia-induced increases in SDF-1 and improved blood flow in ischemic limb. In addition, sitagliptin promoted EPC mobilization and homing to ischemic tissue.	Huang CY et al. 201281
Linagliptin	10 nM or 0.5 µM	4 h in vitro treatment	HUVEC	endothelial cell damage	Linagliptin inhibition inhibits AGE-induced ROS production in HUVEC.	Ishibashi Y et al. 2013139
Vildagliptin	1 nM, 5 nM or 10 nM	6 h	Mouse, HUVEC	Hind limb ischemia	Vildagliptin stimulated ischemia-induced revascularization through an eNOS signaling.	Ishii M et al. 201458
Myocardial infarction
Sitagliptin	250 mg/kg/d (in vivo) or 5 µmol/L (in vitro infusion)	8 weeks feeding or 20 min in vitro infusion	Mouse	MI	Sitagliptin improved functional recovery after I/R injury via increasing cardiac pAKT, pGSK3beta, and atrial natriuretic peptide	Sauvé M et al. 2010107
Diprotin A	10 mM	6 h	Mouse, HUVEC	Thrombosis	Diprotin A enhanced the amount of Tissue Factor encountered and induced the adherence of platelets under flow conditions.	Krijnen PA et al. 2012140
Sitagliptin	300 mg/kg/d	2 weeks	Fischer F344 rat	MI	Sitagliptin improved passive left ventricular compliance, increased endothelial cell density, reduced myocyte hypertrophy, and decreased the abundance of collagen 1.	Connelly KA et al. 2013108
Kidney disease
Sitagliptin	1µmol/L	In vitro treatment	Rat	Renovascular response	Sitagliptin enhanced angiotensin II-induced increase of perfusion pressure in isolated kidneys from both lean and obese ZSF1 rats	Tofovic et al. 2010125
Vildagliptin	1 or 10 mg/kg (i.v.)	One dose	Rat	Renal I/R injury	Vildagliptin reduced tubular necrosis, Bax/Bcl-2 and CXCL10 mRNA expression, and serum creatinine level after renal I/R.	Glorie LL et al. 2012129
Vildagliptin	4 or 8 mg/kg/d	24 weeks	Rat	Kidney injury	Vildagliptin decreased proteinuria, albuminuria, and urinary albumin/creatinine ratio, improved creatinine clearance, and inhibited interstitial expansion, glomerulosclerosis, and the thickening of the glomerular basement membrane in diabetic rats.	Liu et al. 2012127
Sitagliptin	200 mg/kg/day	8 weeks	Rat	Kidney injury	Sitagliptin suppressed nephrectomy-induced activation of PI3K–Akt and FoxO3a. Sitagliptin treatment also reduced apoptosis by decreasing cleaved caspase-3 and −9 and Bax levels.	Joo et al. 2013128
MK0626	33 mg/kg chow (~ 10 mg/kg/d)	16 weeks	Mouse	Obesity-induced renal injury	MK0626 prevented high-fructose/high-fat diet-induced glomerular and tubular injury independent of blood pressure/insulin sensitivity.	Nistala R et al. 2014126
Linagliptin	83 mg/kg rat chow	8 weeks	Zucker obese rat	Obesity-related glomerulopathy	Linagliptin enhanced filtration barrier remodeling, improved proteinuria, increased active GLP-1 and SDF-1α, and improved oxidant markers.	Nistala R et al. 2014130
Linagliptin	5 mg/kg/d in drinking water	4 weeks	Mouse	kidney fibrosis	Linagliptin ameliorated kidney fibrosis in diabetic mice without altering the blood glucose levels associated with the inhibition of EndMT and the restoration of microRNA 29s.	Kanasaki et al. 2014131
Blood pressure
Alogliptin	Various doses (in vitro)	In vitro treatment	Mouse aorta, HUVEC	Vascular function	Alogliptin relaxed pre-constricted aortic segments in a dose dependent manner. Alogliptin induced eNOS and Akt activation in HUVEC cells is independent of GLP-1.	Shah Z et al. 201157
Sitagliptin	40 mg/kg twice daily	8 days	SHR rat	Hypertension	Sitagliptin decreased blood pressure in young SHR rats but not adult SHRs.	Pacheco BP et al. 2011141
Sitagliptin	10 mg/kg/d	2 weeks	SHR rat	Hypertension and hypertensive kidney disease	Sitagliptin treatment improved endothelium-dependent relaxation in renal arteries, restored renal blood flow, and reduced systolic blood pressure in SHR rats.	Liu L et al. 2012142
Saxagliptin	10 mg/kg/d	8 weeks	SHR rat	Hypertension and endothelial dysfunction	Saxagliptin treatment reduced blood pressure in SHRs, an effect that was accompanied an increase in aortic and glomerular NO release with reductions in peroxynitrite levels.	Mason RP et al. 2012143
Linagliptin	83 mg/kg in diet (~ 4 mg/kg/d)	8 weeks	ZO rat	Hypertension	Linagliptin blunted elevated blood pressure progression in ZO rats without reducing left ventricular hypertrophy, fibrosis, or oxidative stress	Aroor AR et al. 201361
Atherosclerosis
Alogliptin	15 mg/kg/d gavage	24 weeks	Mouse	Diabetic atherosclerosis	Alogliptin reduced atherosclerotic lesions and TLR4-mediated upregulation of IL-6 and IL-1β in diabetic ApoE−/− mice.	Ta NN et al. 2011144
Alogliptin	40 mg/kg/d	12 weeks	Mouse	Atherosclerosis	Alogliptin reduced atherosclerotic plaque and vascular inflammation in atherosclerosis-prone Ldlr−/− and ApoE−/− mice.	Shah Z et al. 201198
PKF275-055 (vildagliptin analogue)	100 µm/[kg d] in drinking water	4 weeks	Mouse	Atherosclerosis	PKF275-055 reduced foam cell formation, atherosclerotic plaque, and macrophage accumulation in the aortic wall.	Terasaki et al. 2012 101
Sitagliptin	1.1% in diet	12 weeks	Mouse	Atherosclerosis	Sitagliptin reduced plaque macrophage infiltration and plaque MMP-9 levels, increased plaque collagen content but did not change overall lesion size	Vittone F et al. 2012 97
Anagliptin	0.3% in diet	16 weeks	Mouse	Atherosclerosis	DPP4 inhibition reduced accumulation of monocytes and macrophages in the vascular wall and SMC content in plaque areas.	Ervinna N et al. 2013 102
vildagliptin	0.003% w/v in drinking water (equal to 3 mg/kg/d)	4 weeks	Mouse	Diabetic atherosclerosis	Anagliptin confers a substantial anti-atherosclerotic effect in both nondiabetic and diabetic mice, which was abolished by incretin blockers Exendin(9–39) or (Pro(3))GIP.	Terasaki et al. 2013 99
Sitagliptin	0.3% in diet	16 weeks	Mouse	Atherosclerosis	Sitagliptin treatment reduced atherosclerotic plaque size, collagen fiber, MCP-1, and IL-6 in plaques, serum levels of soluble vascular cell adhesion molecule-1 and P-selectin, and increased activation of AMPK and Akt.in aortas.	Zeng et al. 2014 100
Heart failure
Vildagliptin	30 mg/kg/d	4 weeks	Rat	Heart failure	Vildagliptin reversed diabetic diastolic left ventricular dysfunction and pressure- overload-induced left ventricular dysfunction.	Shigeta et al. 2012 112
Sitagliptin	30 mg/kg/d	3 weeks	Pig	Overpacing-induced heart failure	Sitagliptin increased stroke volume and preserved glomerular filtration rate in pigs with pacing-induced heart failure	Gomez et al. 2012 113
Saxagliptin	10 mg/kg/d	Up to 7 weeks	Mouse	Dilated cardiomyopathy	Saxagliptin treatment improved glucose tolerance but not survival in a transgenic murine model of dilated cardiomyopathy.	Vyas et al. 2012 114
Vildagliptin	10 mg/kg/d	4 weeks	Mouse	Heart failure	Vildagliptin ameliorated TAC-induced left ventricular enlargement and dysfunction, and improved survival rate on day 28.	Takahashi et al. 2013 115
Vildagliptin	30 mg/kg/d	7 days	Rat	Cardiac hypertrophy	Vildagliptin attenuated the β-adrenergic stimulation-induced cardiac hypertrophy as well as cardiomyocyte hypertrophy and perivascular fibrosis.	Miyoshi et al. 2014 116
MK0626	33 mg/kg in diet (~ 10 mg/kg/d)	16 weeks	Mouse	Diastolic dysfunction	MK0626 improved western diet-induced insulin resistance and diastolic relaxation, accompanied by reduced myocardial oxidant stress and fibrosis.	Bostick et al. 2014 117
Linagliptin Sitagliptin	1.5 mg/kg/d (linagliptin) 20 mg/kg/d (sitagliptin)	7 weeks	Rat	Heart failure with preserved ejection fraction	DPP4 inhibition prevented the development of cardiac diastolic dysfunction induced by subtotal nephrectomy, without change in renal function or structure improvement	Connelly et al. 2014 118

CV, Cardiovascular; CVD, cardiovascular disease; I/R, ischemia-reperfusion; MI, myocardial infarction; RLP-C, remnant lipoprotein cholesterol; T2DM, type 2 diabetes; TAC, transverse aortic constriction.

Effect on Myocardial Function and Ischemia-Reperfusion

Genetic disruption of the Dpp4 in mice is not associated with baseline abnormalities in cardiac structure or function. TABLE 1 reviews the effects of DPP4i on models of ischemia re-perfusion^107–109. In response to experimental infarction, Dpp4^−/− mice exhibit improved survival with a trend towards reduction in infarct size¹⁰⁷. Hearts from Dpp4^−/− mice contained higher levels of phosphorylated AKT (pAKT), pGSK3 and ANP, well know pro-survival pathways. In a post-infarction model, sitagliptin treatment led to an improvement in passive left ventricular compliance, increased endothelial cell density, reduced myocyte hypertrophy, and reduced collagen 1¹⁰⁸. Treatment with sitagliptin resulted in improved post-infarction survival without change in infarct size, an effect associated with upregulation of pro-survival pathways HO-1 and Akt, pathways activated by GLP-1¹⁰⁷. Whether the results in response to DPP4 inhibition are due to an increase of DPP4 target proteins such as GLP-1 cannot be definitively inferred from studies to date. Recent studies seem to contest the role of direct effects of GLP-1 on myocardial cells and invoke a role for GLP-1 dependent atrial natriuretic peptide synthesis from atrial cells, since myocytes do not appear to express GLP-1 receptors⁵⁹.

DPP4 and Heart Failure

DPP4 levels correlate with heart failure with plasma DPP4 activity being significantly higher in patients with more advanced heart failure^110,111. Pharmacologic DPP4 inhibition in models of diabetic cardiomyopathy, pressure overload model of heart failure have been consistently shown to improve ventricular remodeling and survival as well as ameliorating severity of heart failure^112–118. Similarly, both pharmacological and genetic DPP4 inhibition reversed diabetic diastolic left ventricular dysfunction and pressure-overload-induced left ventricular dysfunction¹¹². Diabetes is well known to impair angiogenic ability through SDF-1α dependent pathways in cardiac tissue¹¹⁹. DPP4 inhibition (both pharmacological and genetic) reversed SDF-1α-dependent microvasculopathy and diabetes-associated diastolic left ventricular dysfunction¹¹². Improved left ventricular function assessed by ejection fraction, mitral annular systolic velocity, and peak systolic velocity has also been observed in humans with DPP4 inhibition (Sitagliptin, single dose or 4 weeks of treatment)^120,121. While the effects of DPP4 inhibition on heart failure have been directly attributed to increase in GLP-1, the extent to which these effects are driven by GLP-1 is unclear. This is especially true given recent evidence contesting the presence of GLP-1 receptors in the myocytes and dependence of atrial synthesis of ANP for GLP-1 effects⁵⁹. Moreover the effects of GLP-1 itself are controversial. GLP-1 agonism has also been shown to increase left ventricular ejection fraction^122–124 (all in short-term), however, reports on long-term CV effect of DPP4i and GLP-1 analogs are limited.

Effects on Renal Injury and CKD Progression

DPP4 inhibition has been shown to exert renoprotective effects characterized by reduction in oxidative stress, podocyte injury, mesenchymal expansion, fibrosis and proteinuria in experimental models (TABLE 1)^125–131. These effects appear to occur without changes in blood pressure or insulin sensitivity¹²⁶. Treatment with sitagliptin and linagliptin has also been shown to reduce proteinuria^132,133. In a 6-month open label trial in 35 patients, albumin creatinine ratio was reduced by at least two-fold in patients with a range of albumin excretion from normoalbunimuria to macroalbuminuria [urinary albumin-to-creatinine ratio (ACR)>300]. A pooled analysis of four completed studies (n=217) of subjects with T2DM and prevalent albuminuria (defined as a [ACR] of 30–3,000 mg/g creatinine) receiving linagliptin on top of stable doses of RAAS inhibitors, demonstrated a modest 28% placebo adjusted reduction in ACR at 6 months independent of blood pressure.

Sitagliptin protects renal function in T2DM patients as evidenced by reduced urinary albumin-to-creatinine ratio Renovascular response to angiotensin II may also be enhanced by DPP4 inhibition. Sitagliptin (1µmol/L) significantly enhanced angiotensin II-induced increase of perfusion pressure in isolated kidneys from both lean (18.2 ± 5.9 vs. 3.4 ± 1.3 mmHg) and obese (17.8 ± 8.2 vs. 5.5 ± 1.3 mmHg) ZSF1 rats¹²⁵. Improvement in filtration barrier remodeling¹³⁰ and suppression of fibrosis¹³¹ may also serve as alternative mechanisms of DPP4i in protecting renal function in chronic kidney disease.

CARDIOVASCULAR CLINICAL TRIALS WITH DPP4i

Table II lists completed and ongoing randomized controlled clinical trials with DPP4i. SAVOR-TIMI53 was designed as a superiority trial and failed to meet the pre-specified superiority outcome of saxagliptin versus placebo in a high-risk patient population with established vascular disease and/or risk factors. In the EXAMINE trial, designed as a safety trial in a very high risk post-acute coronary syndrome population, the pre-specified end-point of non-inferiority was met and alogliptin was non-inferior to placebo with regards to CV outcomes. In both EXAMINE and SAVOR-TIMI53 trials, the primary outcome composed of CV death, myocardial infarction (MI) and ischemic stroke were non-inferior in DPP4i group compared to placebo group^134,135. In both trials the median duration of follow-up was <4 years, the majority of whom were on evidence based therapies. These factors are widely felt to be to have been responsible for lack of benefit with these agents. In SAVOR-TIMI 53, baseline NT-proBNP was measured in 12,301 patients. More patients treated with saxagliptin were hospitalized for heart failure compared to placebo (HR 1.27; 95%CI 1.07–1.51; p=0.007). Differences in heart failure were seen in less than 6 months with rates at 12-months of 1.9% vs.1.3% (HR 1.46, 95%CI 1.15–1.88, p=0.002, with no difference thereafter). Subjects at greatest risk for heart failure had prior heart failure, eGFR ≤60 ml/min and/or elevated baseline levels of NT-proBNP (risk in highest NT-proBNP quartile of 2.1%)⁷⁵. Whether the association with heart failure applies to other DPP4i will need to be addressed in ongoing CV outcome trials of DPP4i¹³⁶. In consistence with this finding, a slight increase of heart failure hospitalization in patients with alogliptin compared to those with placebo (3.9% vs. 3.3%, HR = 1.19, 95% CI, 0.90–1.58) in EXAMINE trial, although no statistical significance was observed^137,138.

Table 2.

DPP4i CV outcome trials

Drug (class)	Landmark name	Study population	Primary outcome	Dosing	Estimated enrollment	Duration (years)	End date	Identifier
Alogliptin	EXAMINE	T2DM with ACS (within >15 to < 90 days)	MACE	6.25 mg, 12.5 mg, or 25 mg qd‡	5,400	4.75	2013	NCT00968708
Saxagliptin	SAVOR TIMI 53	T2DM with multiple CRF‖	Composite	2.5 mg or 5 mg qd	16,500	4	2013	NCT01107886
Sitagliptin	TECOS	T2DM with preexisting CVD	Composite	50 mg or 100 mg qd‡	14,000	5	2014	NCT00790205
Linagliptin	CAROLINA	T2DM with preexisting CVD	Composite	5 mg qd	6,000	7.5	2018	NCT01243424
Linagliptin	CARMELINA	T2DM with impaired renal function or CVD	Composite	5 mg once daily	8,300	4.5	2018	NCT01897532
Omarigliptin (MK3102)	MK-3102-015 AM1	T2DM with inadequate glycemic control on diet/exercise therapy and oral antihyperglycemic agent monotherapy	Any adverse event	25 mg once weekly	585	1.5	2014	NCT01697592
Omarigliptin (MK3102)	MK-3102-018	T2DM with preexisting CVD	Composite	25 mg once weekly	4,000	5	2017	NCT01703208

CVD, cardiovascular disease; T2DM, type 2 diabetes; MACE, Major Adverse Cardiac Events (defined as a composite of CV death, nonfatal MI and nonfatal stroke; these events were adjudicated by an independent cardiovascular endpoints committee); Composite, a composite defined as CV-related death, nonfatal MI, nonfatal stroke, or unstable angina requiring hospitalization

CONCLUSION

The development of DPP4i has evolved from the understanding of its role in regulation of incretin hormones. However the widespread expression of this molecule in a variety of cell types clearly underscores a previously unrecognized role for this molecule in physiological processes. Despite the plethora of enzymatic substrates that serve as targets, the inhibition of enzymatic function in animals and humans appears to be well-tolerated and safe with modest reduction in HbA1c. The pre-ponderance of evidence from inhibiting DPP4 enzymatic function in animal models of disease seems to suggest salutary effects including in heart failure and in totality seem to suggest the possibility of beneficial CV effects in humans. At least 2 large randomized clinical trials focused on CV outcomes in high risk patients with type 2 diabetes on best medical therapy appear to show no differences compared to placebo with regards to the composite end-point of CV death, MI and stroke over a duration less than 4 years of drug treatment. The non-catalytic function of DPP4, include its role as a co-stimulatory molecule and binding to a number of matrix proteins including ADA in humans, points to larger role for DPP4 outside of its catalytic function. Future mechanistic studies will need to focus on the relative contribution of the catalytic versus non-catalytic function in cardiometabolic disease while clinical trials will need to address the heart failure safety signals and demonstrate a beneficial effect of this class in CV complications associated with T2DM.

Non-standard Abbreviations and Acronyms

ADA

adenosine deaminase

CV

cardiovascular

DC

dendritic cell

DPP4

dipeptidyl peptidase-4

DPP4i

DPP4 inhibitors

Epo

erythropoietin

FMD

flow-mediated dilation

GIP

gastric inhibitory peptide

GLP-1

glucagon like peptide-1

HOMA-b

homeostasis model assessment beta cell function index

HUVECs

human umbilical vein endothelial cells

MERS

middle eastern respiratory syndrome

sDPP4

soluble DPP4

T2DM

type 2 diabetes

VAT

visceral adipose tissue