Dipeptidyl Peptidase-4 Inhibition Stimulates Distal Tubular

Diabetes Care 1

Julie A. Lovshin,1,2 Harindra Rajasekeran,1 Dipeptidyl Peptidase-4 Inhibition Yulyia Lytvyn,1 Leif E. Lovblom,3 Shajiha Khan,1 Robel Alemu,1 Amy Locke,1 Stimulates Distal Tubular Vesta Lai,1 Huaibing He,4 Lucinda Hittle,4 Weixun Wang,4 Daniel J. Drucker,2,3 and Natriuresis and Increases in David Z.I. Cherney1,5,6 Circulating SDF-1a1-67 in Patients With Type 2 Diabetes https://doi.org/10.2337/dc17-0061

OBJECTIVE Antihyperglycemic agents, such as empagliflozin, stimulate proximal tubular natriuresis and improve cardiovascular and renal outcomes in patients with type 2 diabetes. Because dipeptidyl peptidase 4 (DPP-4) inhibitors are used in combination with sodium–glucose cotransporter 2 (SGLT-2) inhibitors, we examined whether and how sitagliptin modulates fractional sodium excretion and renal and systemic he- PATHOPHYSIOLOGY/COMPLICATIONS modynamic function. 1Division of Nephrology, Department of Medi- RESEARCH DESIGN AND METHODS cine, University Health Network, University of We studied 32 patients with type 2 diabetes in a prospective, double-blind, random- Toronto, Toronto, Ontario, Canada 2 ized, placebo-controlled trial. Measurements of renal tubular function and renal and Division of Endocrinology and Metabolism, De- partment of Medicine, University of Toronto, systemic hemodynamics were obtained at baseline, then hourly after one dose of Toronto, Ontario, Canada sitagliptin or placebo, and repeated at 1 month. Fractional excretion of sodium and 3Lunenfeld-Tannenbaum Research Institute, lithium and renal hemodynamic function were measured during clamped euglyce- Mount Sinai Hospital, University of Toronto, mia. Systemic hemodynamics were measured using noninvasive cardiac output Toronto, Ontario, Canada 4Pharmacokinetic Pharmacodynamics and Drug monitoring, and plasma levels of intact versus cleaved stromal cell-derived factor Metabolism, Merck and Co., Inc., Rahway, NJ (SDF)-1a were quantified using immunoaffinity and tandem mass spectrometry. 5Institute of Medical Science, University of Toronto, Toronto, Ontario, Canada RESULTS 6Department of Physiology, University of Toronto, Sitagliptin did not change fractional lithium excretion but significantly increased total Toronto, Ontario, Canada fractional sodium excretion (1.32 6 0.5 to 1.80 6 0.01% vs. 2.15 6 0.6 vs. 2.02 6 Corresponding author: Julie A. Lovshin, lovshin@ lunenfeld.ca. 1.0%, P = 0.012) compared with placebo after 1 month of treatment. Moreover, a1-67 Received 10 January 2017 and accepted 3 May sitagliptin robustly increased intact plasma SDF-1 and decreased truncated 2017. a3-67 plasma SDF-1 . Renal hemodynamic function, systemic blood pressure, cardiac Clinical trial reg. no. NCT02406443, clinicaltrials output, stroke volume, and total peripheral resistance were not adversely affected .gov. by sitagliptin. This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/ CONCLUSIONS suppl/doi:10.2337/dc17-0061/-/DC1. DPP-4 inhibition promotes a distal tubular natriuresis in conjunction with increased © 2017 by the American Diabetes Association. levels of intact SDF-1a1-67. Because of the distal location of the natriuretic effect, Readers may use this article as long as the work DPP-4 inhibition does not affect tubuloglomerular feedback or impair renal hemo- is properly cited, the use is educational and not for profit, and the work is not altered. More infor- dynamic function, findings relevant to using DPP-4 inhibitors for treating type 2 mation is available at http://www.diabetesjournals diabetes. .org/content/license. Diabetes Care Publish Ahead of Print, published online May 26, 2017 2 Sitagliptin Promotes Distal Tubular Natriuresis Diabetes Care

Cardiovascular and renal complications renal tubular sodium handling and also Merck Sharpe & Dohme) by the research are leading causes of premature death renal and systemic hemodynamic func- trial pharmacist with an allocation of and morbidity in patients with type 2 di- tion after acute and sustained administra- one to one and a block size of four using abetes. Diabetic nephropathy is indepen- tion. Here we demonstrate that sitagliptin computer-generated lists (Supplementary dently associated with an increased risk enhances natriuresis independently of Fig. 1). for adverse cardiovascular outcomes and significant changes in renal or systemic is partly driven by ambient hyperglycemia hemodynamic parameters, likely through Measurement of Renal Tubular (1). Unfortunately, preventing renal sites and mechanisms distinct from those Function and Hemodynamics dysfunction or slowing progression of regulated by SGLT-2 inhibitors in the kid- Oneweekbeforethefirst study visit chronic kidney and/or cardiovascular dis- ney. Furthermore, we demonstrate that (study visit 1), study participants were in- ease in type 2 diabetes remains a major sitagliptin robustly augments plasma lev- structed to maintain a minimum sodium therapeutic shortcoming of current dia- els of bioactive stromal cell–derived intake of 150 mmol/day, a protein diet of betes clinical care. factor (SDF)-1a1-67, a chemokine with 1.5 g/kg/day, and to avoid alcohol and Dipeptidyl peptidase 4 (DPP-4) inhi- proven natriuretic actions in preclinical tobacco for at least 48 h. In the 24 h lead- bitors are antihyperglycemic agents studies. Collectively, our findings high- ing up to study visit 1, study participants that potentiate incretin action and are light distinct mechanisms for how DPP-4 completed a 24-h timed urine collection. widely used in the treatment of type 2 inhibitors (vs. effects previously eluci- The night before study visit 1, study par- diabetes (2) as monotherapy or, more dated for SGLT-2 inhibitors) control natri- ticipants fasted for a minimum of 12 h commonly, in combination with metfor- uresis and tubuloglomerular function, and avoided caffeine intake. Participants min, sulfonylureas, and sodium–glucose results with implications for the use of were also instructed to take one lithium cotransporter 2 (SGLT-2) inhibitors. DPP-4 these drugs as combination therapy in carbonate tablet (300 mg) at 2200 h to inhibitors lower glucose by preventing subjects with type 2 diabetes. allow for measurement of fractional lith- the N-terminal cleavage and inactivation ium excretion (FELi). FELi is a marker of of the incretin hormones glucagon-like RESEARCH DESIGN AND METHODS proximal tubular sodium reabsorption, peptide-1 (GLP-1) and glucose-dependent which allows for determination of proxi- Study Population insulinotropic peptide (3,4). The cardio- Local advertising was used to enroll 36 mal versus distal natriuresis localization vascular safety of DPP-4 inhibition with patients with type 2 diabetes (A1C 6.5– within the renal tubules. On the morning sitagliptin (5), alogliptin (6), and saxaglip- of study visit 1, study participants with- 9.0% [48–75 mmol/L], or HbA1c ,6.0– tin (7) has been demonstrated in large 9.0% [742–75 mmol/mol] if treated held all oral antihyperglycemic agents outcome studies; however, saxagliptin with oral hypoglycemic agents or insulin and prandial insulin but continued all and alogliptin may increase hospitaliza- and diagnosed with type 2 diabetes .2 other routine medications, including tion for heart failure, particularly in pa- years’ duration), receiving stable doses antihypertensives. tients with preexisting heart or kidney (within 1 month) of renin-angiotensin- Study participants arrived at the disease (8). Remarkably, little is known aldosterone system (RAAS) blockade Renal Physiology Laboratory (Toronto about the renal hemodynamic or natri- (ACE inhibitor or angiotensin receptor General Hospital, Toronto, Ontario, uretic actions of DPP-4 inhibitors in sub- blocker if ACE inhibitor intolerant), with Canada), at ;0800 h for a brief physical jects with type 2 diabetes. a systolic blood pressure (SBP) of at examination and body weight measure- fl The results of the Empagli ozin, Car- least .120 mmHg and estimated glomer- ment. After a rest period of ;15 min, diovascular Outcomes, and Mortality in ular filtration rate (GFR) .50 mL/min/ blood work (HbA1c, blood glucose, elec- Type 2 Diabetes (EMPA-REG OUTCOME) 1.73m2 (SupplementaryFig.1;seeSupple- trolytes, lipids, creatinine, urea, uric acid, trial (9) with the SGLT-2 inhibitor, empa- mentary Table 1 for exclusion criteria). aldosterone, plasma renin concentration) fl gli ozin, highlight the clinical importance The study was approved by the University and urine parameters (albumin, protein, of understanding how different glucose- Health Network Research Ethics Board electrolytes) were collected, and vital lowering agents modulate natriuresis and and conducted in accordance with the signs were measured at the bedside. A renal hemodynamic function. Although Declaration of Helsinki and International standardized liquid meal (2 MightyShakes fl sitagliptin increased fractional sodium ex- Conference on Harmonization on Good Vanilla 50/6 oz [single serving, 240 cal- cretion in a single study of subjects with Clinical Practice. All patients provided ories: fat, 13 g; carbohydrate, 23 g; pro- type 2 diabetes (10), the mechanism(s) written informed consent. tein, 8 g); Hormel Health Laboratories, and tubular location for these effects Savannah, GA) was given to all study partic- and associated changes in renal hemody- Study Design ipants to allow for endogenous incretin namic or cardiovascular function have This was a single-center (University hormone secretion and to control dietary not been comprehensively elucidated. Health Network, University of Toronto, intake during the study visits. For partici- Remarkably, cardioprotection with empa- Toronto, Ontario, Canada) prospective, pants who could not tolerate the standard- gliflozin was evident within weeks, high- double-blind, randomized, placebo- ized liquid meal (three patients), a substitute lighting the importance of understanding controlled trial, conducted over two meal with equal proportions of macronutri- the acute nonglycemic actions of current study visits separated by 1 month ents was given (ad libitum water consump- therapies used to treat type 2 diabetes. (Supplementary Fig. 2). Thirty-six patients tion was allowed during the experimental Accordingly, we have now studied were randomized to sitagliptin (100 mg period, up to a maximum of 500 mL). whether and how the DPP-4 inhibitor, per os daily) or to identical placebo per Aeuglycemicclampwasinitiated, sitagliptin, modifies proximal and distal os daily (manufactured and supplied by and once euglycemia was reached and care.diabetesjournals.org Lovshin and Associates 3

stabilized for 2 h, a third venous cath- based noninvasive cardiac output moni- methods by the Department of Clinical eter was placed to facilitate bolus infu- toring methods (NICOM; Cheetah Medi- Biochemistry at the University Health sion of p-aminohippurate sodium 20% cal, Newton Center, MA). In brief, four Network, Toronto, Ontario, Canada. (PAH; Bachem, Bubendorf, Switzerland) electrodes were placed externally on Urine lithium was measured by induc- and inulin (Inutest; Fresenius Kabi Austria the chest (at the clavicle) and lower ab- tively coupled plasma spectroscopy by GmbH, Graz, Austria) to assess effective domen (below costal margin) bilaterally. n-Common Laboratories, North York, On- renal plasma flow (ERPF) and GFR, respec- Calibration and baseline measurements tario, Canada. tively, according to previously described were completed before study drug ad- Urinary inflammatory cytokines, che- methods (11). After a 90-min equilibra- ministration, and then measurements mokines,andgrowthfactorsandco- tion period, urine was collected by spon- were done hourly (first 10 min) for 3 h gnate receptors were measured using taneous voiding, and blood was sampled after study drug administration. Patients a Discovery Assay (Human Cytokine Array/ every 45 min. were positioned supine during measure- Chemokine Array 42-Plex; Eve Technologies After baseline renal function tests ments and were instructed to remain still Corp., Calgary, Alberta, Canada), using were completed at ;1300 h, study par- and quiet. previously described methods, corrected ticipants were given an identical fixed- for urinary creatinine (15). Plasma ANP, calorie liquid meal (as described above) Study End Points plasma and urinary NO, plasma and uri- to stimulate endogenous incretin hormone The primary study end point was the nary cGMP, urine 8-isoprostane, and secretion. Participants were then given change in the fraction of excreted sodium urine 8-hydroxy-2-deoxyguanosine were their first dose of study drug (sitagliptin (FENa) at 3 h after a single dose and after measured by ELISA, as described previ- or placebo). Participants were asked short-term administration of sitagliptin ously (16,17). to empty their bladder (they were up- (1 month). Secondary end points included right and asked to void until a subjective the change in gold standard measured re- Plasma SDF-1a and Sitagliptin feeling of empty was achieved), and nal hemodynamic function, GFRINULIN and Quantification urine was collected under spontaneous ERPFPAH (12), under clamped euglycemia, Levels of intact and truncated SDF-1a in voiding for up to 3 h during a timed urine and derived measures of intrarenal he- plasma were measured using a hybrid collection. PAH and inulin were continu- modynamics (renal blood flow, filtration method coupling immunoaffinity en- ously infused at maintenance levels fraction [FF], renal vascular resistance, richment and liquid chromatography– through this 3-h period, and blood sam- glomerular hydrostatic pressure, renal af- tandem mass spectrometry (18). Briefly, 1-67 3-67 ples were drawn hourly for measurement ferent resistance, renal efferent res- plasma SDF-1a and SDF-1a were istance) using previously described first captured with anti–SDF-1a antibody of GFRINULIN and ERPFPAH in response to the study drug. methods (13,14). FELi, systemic hemody- (MAB350; Bio-Techne, R&D Systems, Inc., After study visit 1, subjects were in- namic function (cardiac output, stroke Minneapolis, MN), immobilized on mag- structed to continue all routine medica- volume, total peripheral resistance, netic beads, and then quantified using a tions in addition to the daily study drug. mean arterial pressure, SBP, diastolic multiple reaction monitoring method An interim telephone call was made to blood pressure, heart rate, and pulse on a triple quadrupole mass spectrome- study participants at ;14 days to screen pressure). FENa and FELi excretion were ter. Plasma sitagliptin concentrations for adverse events (study visit 2). Approx- calculated according to FE(electrolyte) = were quantified as described previously imately 28–30 days later (1 month), all 100 3 ([(electrolyte urine) 3 (creatinine (19). study participants returned to the Renal plasma)]/[(electrolyte plasma) 3 (creati- Physiology Laboratory for a final study nine urine)]). Statistical Analyses Statistical analyses were performed by an visit, study visit 3. All prestudy procedures Acute and chronic changes in plasma 1-67 independent statistician using SAS 9.2 (e.g., dietary sodium and protein recom- SDF-1a ,urinaryinflammatory cyto- (SAS Institute Inc.). Data are reported as mendations) and experimental proce- kines and chemokines, plasma renin mean 6 SD, median (interquartile range), dures on study visit 3 were identical to concentration, aldosterone, plasma and or as frequency (%). The treatment effect study visit 1, with the exception that on urinary nitric oxide (NO), plasma atrial on all study end points was assessed using the morning of study visit 3, participants natriuretic peptide (ANP), and urinary repeated-measures linear mixed-effects withheld study drug administration until reactive oxygen species (8-isoprostane, models. The sample size calculation was given by study personnel after the second 8-hydroxy-2-deoxyguanosine, cyclic gua- based on the estimated effect of study standardized liquid meal (;1300 h) be- nosine monophosphate [cGMP]) were medication on FE Preclinical studies in- fore the timed urine collection during secondary exploratory end points. Ter- Na. dicated the most conservative estimate study visit 3. tiary confirmatory end points consisted for a DPP-4 inhibitor on FE is 14% of changes in HbA1c,lipids(totalcholes- Na Noninvasive Cardiac Output terol, LDL, HDL, triglycerides), and body (20). No variances were provided for the Monitoring weight. preclinical studies, so we used an “in- Cardiovascular parameters were assessed flated” SD of 14 (equal to the effect to detect potential systemic physiologi- Biochemical Analyses size) in our sample size calculation. With cal effects that might arise secondary Routine blood (HbA1c, electrolytes, liver this variance and effect size, a sample size to DPP4 inhibitor–stimulated natriure- enzymes, and urine biochemistry) were of 16 per group provided .80% power to sis. Systemic hemodynamic function was performed at baseline and after 1 month detect significant between-group differ- measured using electrical bioreactance- of treatment using conventional assay ences (two-tailed a = 0.05) in response 4 Sitagliptin Promotes Distal Tubular Natriuresis Diabetes Care

to DPP-4 inhibition. Unblinding occurred Table 1—Demographics and clinical characteristics of study population after completion of the study and locking Placebo (n = 16) Sitagliptin (n =16) of the database, and analyses were done on a per-protocol basis. One participant Day 1 1 month Day 1 1 month was excluded from analysis of NICOM Demographics measurements due to an inability to obtain Age (years) 59.3 6 8.8 d 60.4 6 7.6 d d d reliable NICOM measurements because Males 9 (56.3) 11 (68.8) Ethnicity of large body habitus (allocated to placebo), Caucasian 8 (50) d 12 (75) d and one participant’s 3-h NICOM measure- South Asian 4 (25) d 2 (12.5) d ments were discarded due to the inability Black 1 (6.3) d 1(6.3) d to capture data (allocated to placebo). An Other 3 (18.7) d 2(6.2) d a level of 0.05 was used to determine Diabetes duration (years) 8.50 (5.5, 14.0) d 6.0 (3.0, 10.0) d statistical significance for all comparisons. Metformin use 14 (87.5) 14 (87.5) 12 (75) 12 (75) Sulfonylurea use 6 (37.5) 6 (37.5) 3 (18.8) 3 (18.8) RESULTS Insulin use 4 (25) 4 (25) 4 (25) 4 (25) Hypertension duration (years) 11.4 6 10.1 d 6.8 6 8.9 d Clinical Characteristics RAAS inhibition 16 (100) 16 (100) 16 (100) 16 (100) In total, 48 patients were screened be- Diuretics 6 (37.5) 6 (37.5) 5 (31.3) 5 (31.3) tween July 2015 and March 2016 at the CCB 5 (31.3) 5 (31.3) 3 (18.8) 3 (18.8) Renal Physiology Laboratory, Toronto Statin 11 (68.8) 11 (68.8) 9 (56.3) 9 (56.3) General Hospital, Toronto, Ontario, Can- CVD 1 (6.3) 1 (6.3) 4 (25) 4 (25) ada (Supplementary Fig. 2), and 36 pa- Clinical characteristics 6 6 6 6 tients were ultimately randomized to Hematocrit 0.38 0.04 0.36 0.04 0.39 0.02 0.37 0.03 HbA (%) 7.31 6 0.84 7.18 6 0.97 7.18 6 0.79 6.82 6 0.82 sitagliptin (n = 18) or placebo (n = 18). 1c HbA1c (mmol/mol) (56 6 9.2) 55 6 10.6 55 6 8.6 51 6 9.0 After randomization but before study visit Creatinine (mmol/L) 68.6 6 9.0 66.9 6 7.5 70.6 6 7.8 67.8 6 8.9 1, one patient (allocated to sitagliptin) eGFR (CKD-EPI) withdrew from the study. Three patients (mL/min/1.73m2)94.26 11.4 95.6 6 12.3 94.1 6 7.2 96.4 6 6.7 were withdrawn during the study be- BMI (kg/m2) 30.22 6 6.96 30.16 6 7.02 31.69 6 5.49 31.70 6 5.63 cause of adverse events (dyspnea, allo- SBP (mmHg) 124 6 14 123 6 14 123 6 13 121 6 9 6 6 6 6 cated to placebo; vasovagal event, DBP (mmHg) 71 8737718737 HR (bpm) 74 6 14 68 6 12 74 6 14 68 6 12 allocated to placebo; lower gastrointesti- MAP (mmHg) 86 6 9896 8876 9886 7 nal bleeding, allocated to sitagliptin). Data are expressed as n (%), mean 6 SD, or median (interquartile range). CCB, calcium channel Baseline characteristics were similar be- blocker; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; CVD, cardiovascular disease; tween both groups (Table 1). Intravenous DBP, diastolic blood pressure; eGFR, estimated GFR; MAP, mean arterial pressure; HR, heart rate. infusion volumes of dextrose (5%), insulin, and maintenance fluids did not differ significantly between treatment groups compared with placebo after 1 month of or urinary cGMP, plasma or urinary NO, or (data not shown). administration at 3 h after study drug plasma renin or aldosterone concentrations (Supplementary Table 3). compared with placebo (Supplementary Renal Tubular Effects Table 4). Urinary 8-isoprostane concen- FE was blunted with sitagliptin com- a Na Plasma SDF-1 trations did not change in response to pared with placebo after single-dose ad- Acute administration of sitagliptin in- DPP-4 inhibition with sitagliptin; how- ministration (Fig. 1A)onexperimental creased circulating levels of intact SDF- ever, small but statistically significant day 1 but not after 1 month of adminis- 1a1-67 compared with placebo after a increases in 8-hydroxy-2-deoxyguanosine tration. In contrast, an increase in FE single dose and after 1 month (P , 0.0001) Na concentrations were observed at 1 month was observed with sitagliptin compared (Fig. 2A). In contrast, truncated SDF-13-67 (Supplementary Table 4). Most concen- with placebo at 3 h after 1 month of ad- decreased acutely after single-dose si- trations of urinary cytokines, chemokines, ministration relative to after a single dose tagliptin administration compared with growth factors, and cognate receptors on experimental day 1 (Fig. 1B). No con- placebo (P , 0.0001) (Fig. 2B). Among were unchanged after 1 month of sitagliptin sistent change in FE occurred during the subjects randomized to sitagliptin, Li compared with placebo (Supplementary study (Supplementary Fig. 3). plasma sitagliptin concentrations corre- Table 5). A small but statistically significant lated with chronic percentage change in increase in interferon-g was observed Renal and Systemic Hemodynamic FE and with SDF-1a1-67 concentrations, Na (Supplementary Table 5). Function Effects although these associations did not reach Despite changes in natriuresis, acute and statistical significance (Supplementary 1-month administration of sitagliptin did CONCLUSIONS Fig. 4A and B). not influence renal hemodynamic function, Although type 2 diabetes is predomi- blood pressure, or NICOM parameters Neurohormones, Markers of Oxidative nantly recognized as a metabolic disease (Supplementary Tables 2 and 3). A small Stress, and Urinary Cytokines and characterized by hyperglycemia and but statistically significant increase in Chemokines insulin resistance, the introduction of heart rate was observed with sitagliptin Acute and 1-month sitagliptin administra- new medications, such as the SGLT-2 in- (+5.7 6 3.3 vs. 2.2 6 4.5 bpm, P = 0.019) tion did not change plasma ANP, plasma hibitors, GLP-1 receptor (GLP-1R) agonists, care.diabetesjournals.org Lovshin and Associates 5

Figure 1—Acute and chronic changes in FENa in response to sitagliptin compared with placebo in patients with type 2 diabetes. A:FENa on day 1 at 3 h (compared with baseline) after the first dose of sitagliptin or placebo. B:FENa at 1 month, 3 h after sitagliptin or placebo (compared with day 1). C: Percentage change in FENa at 3 h (compared with baseline) on day 1 and after 1 month of sitagliptin or placebo. D:PercentagechangeinFENa at 1 month (compared with day 1) at baseline and at 3 h after sitagliptin or placebo. The horizontal bars in A and B indicate percentage changes in group means. and DPP-4 inhibitors, has refocused at- from human studies linking DPP-4 inhibi- transient, reverses over time, and natri- tention on nonglycemic aspects of tion to natriuresis and changes in renal uresis ultimately predominates after type 2 diabetes pathophysiology. Nota- physiologyisverylimited(10).Here 1 month of treatment of sitagliptin. Fifth, bly, the action of SGLT-2 inhibitors has we describe several important features despite changes in FENa, no changes in reemphasized the importance of exagger- of sitagliptin action in the diabetic intrathoracic fluid volume, cardiac out- ated proximal tubular sodium reabsorp- human kidney, with immediate clinical put, or systemic hemodynamic function tion in type 2 diabetes. Furthermore, relevance. were observed. enhanced sodium absorption may lead First, sitagliptin increased urine FENa by Agents that induce a proximal tubular to maladapted physiological responses 40%, findings detectable by 1 month of natriuresis, such as SGLT-2 inhibitors and within the diabetic kidney that contribute therapy. In contrast, FELi,amarkerof carbonic anhydrase inhibitors, activate to diabetic nephropathy by promoting proximal tubular sodium reabsorption, tubuloglomerular feedback, leading to hyperfiltration and intraglomerular was unchanged, suggesting that sitaglip- afferent arteriolar vasoconstriction and hypertension. Apart from systemic blood tin induces natriuresis by blocking distal reductions in GFR and ERPF (24). Con- pressure control, blockade of the RAAS, tubular sodium reabsorptive mechanisms versely, agents that act distal to the mac- and more recently, SGLT-2 inhibition, rather than proximal sodium uptake, ula densa are not coupled to these few therapeutic strategies have attenu- which is modulated by SGLT-2 inhibitors. intrarenal autoregulatory mechanisms ated diabetes-related complications in The increased FENa described here is con- and should not affect tubuloglomerular the kidney. sistent with a previous report that urinary feedback or renal hemodynamic function. The demonstration that SGLT-2 inhibi- sodium excretion increases by 2 weeks of Consistent with distal rather than proxi- tion with empagliflozin in patients with sitagliptin therapy in subjects with type 2 mal natriuretic pathways, we showed type 2 diabetes and existing cardiovascu- diabetes (10). Second, we reveal, based that renal hemodynamic function re- lar disease attenuates nephropathy (9,12, on FELi studies, that these renal natri- mained unchanged, both acutely and 21) has fostered interest in understanding uretic effect(s) occur distal to the macula chronically, in response to sitagliptin. whether and how other antihyperglyce- densa. Third, we show that natriuresis These findings have clinical relevance, mic agents modify renal tubular and was acutely suppressed after 3 h after implying that the focus of sitagliptin ac- natriuresis-related pathways in the di- the first dose of sitagliptin, suggesting tion is the distal tubule, thereby leaving abetic kidney (22). Although preclinical that initial sodium retention may occur tubuloglomerular feedback, afferent resis- studies demonstrate that DPP-4 inhibi- with DPP-4 inhibitors in some individuals. tance, and intraglomerular pressure unal- tion promotes natriuresis (23), evidence Fourth, the initial acute suppression is tered (24). In contrast, GLP-1R agonists 6 Sitagliptin Promotes Distal Tubular Natriuresis Diabetes Care

12 (CXCL12) (4), is widely expressed and also localized to glomerular podocytes and distal tubular cells in mouse and human kidneys (26). We focused on assess- ment of SDF-1 because preclinical studies revealed that GLP-1R signaling does not mediate the natriuretic actions of DPP-4 inhibitors (20,23). SDF-1 expression is upregulated in the diabetic rat nephron (23), and SDF-1/CXCR4–dependent signaling suppresses renal oxidative stress and fibrosis through cAMP-mediated pathways. Here we applied immunocapture coupled with mass spectrometry to demonstrate, for the first time in humans with diabetes, that DPP-4 inhibition robustly increases intact plasma SDF-1a1-67 and markedly decreases truncated SDF-1a3-67. In preclinical studies, blockade of SDF-1a1-67/CXCR4 signaling with the antagonist AMD 3100 reversed the natriuretic effects(s) of DPP-4 inhibition, mechanistically linking DPP-4 inhibition to enhanced urine sodium excretion via increased SDF-1a1-67/CXCR4 signaling (23). Our findings associating increased levels of bioactive SDF-1a1-67 with enhanced natriuresis in humans with diabetes treated with sitagliptin suggest conservation of biology across species and highlight SDF-1a1-67 as a new mechanistic target for control of natriuresis in human subjects. The extent of renal dysfunction and progression of nephropathy in subjects with diabetes has been partly linked to changes in oxidative stress, inflammation, Figure 2—Immunoquantification of plasma SDF-1a in response to 1 month of sitagliptin adminis- or enhanced growth factor activity (27). tration compared with placebo in patients with type 2 diabetes. Individual patient data are indicated Although DPP-4 inhibitors have been re- for immunoquantification of intact plasma SDF-1a1-67 (A) and truncated plasma SDF-1a3-67 (B) ported to exert anti-inflammatory effects concentrations over the course of this 1-month clinical study in type 2 diabetes patients adminis- in preclinical and clinical studies (4,28), tered placebo or sitagliptin. Inset (upper right corner) depicts metabolism of intact SDF-1a1-67 to 3-67 sitagliptin therapy was not associated truncated SDF-1a by the proteolytic actions of DPP-4 and suppression of DPP-4–mediated SDF- 1a metabolism by DPP-4 inhibitors. with changes in urinary inflammatory cytokines, chemokines, growth factors, or markers of oxidative stress (isoprostane), attenuate diabetes-associated sodium modify blood pressure or systemic hemo- aside from small changes in 8-hydroxy-2- avidity through distinct, more proximal dynamic parameters assessed by nonin- gaunosine and interferon-g at 1 month. mechanisms (16) and may therefore vasive cardiac monitoring. Hence, these Hence, although we cannot exclude the modify tubuloglomerular feedback. Con- physiological findings predict that DPP-4 possibility that DPP-4 inhibition may pro- sistent with this possibility, liraglutide re- inhibitors may exert clinically useful ef- duce chronic anti-inflammatory actions in duced the composite renal microvascular fects on renal hemodynamics in combina- longer studies or reduce inflammation in outcome by 22% in patients with type 2 tion with SGLT-2 inhibitors without the different tissue compartments from diabetes at heightened cardiovascular burden of excess volume contraction, hy- unique patient populations, the current risk, an effect predominantly driven by potheses that should be carefully tested data suggest that clinically detectable reductions in macroalbuminuria (25). in future mechanistic human studies. changes in inflammatory status or oxida- Although SGLT-2 inhibitors produce Amajorfinding of clinical relevance tive stress are not a prominent feature of more robust natriuresis, blood pressure is the demonstration that SDF-1a1-67 short-term therapy with DPP-4 inhibition reduction, and potentially volume deple- is a physiological substrate of DPP-4 in patients with type 2 diabetes. tion, our current findings demonstrate in humans with diabetes. SDF-1a1-67, The lack of effect of DPP-4 inhibitors on that sitagliptin administration did not also known as C-X-C motif chemokine renalfunctioninlargecardiovascular care.diabetesjournals.org Lovshin and Associates 7

outcome studies is consistent with our increasing use of DPP-4 inhibitors and DPP-4 inhibition to the distal convoluted short-term study results, because modifi- SGLT-2 inhibitors in combination therapy tubule. Hence, we predict that DPP-4 in- cation of distal tubular natriuresis via for type 2 diabetes. Both drugs modify hibitors should not interfere with or di- DPP-4 inhibition should not affect renal renal tubular sodium handling, with po- minish the effects of other drugs that function or tubuloglomerular feedback tential implications for renal and systemic block proximal renal tubular sodium re- pathways. Furthermore, the clinical rele- hemodynamic function. Our current find- absorption. Further mechanistic studies vance of our findings is highlighted by the ings localize renal natriuretic effects of examining the simultaneous renal tubular

Figure 3—Putative natriuretic mechanisms for DPP-4 inhibitors (DPP-4i). DPP-4i promote a distal tubular natriuresis that does not affect renal or systemic hemodynamic function in patients with type 2 diabetes without nephropathy. Although these DPP-4i–mediated natriuretic mechanisms remain in- completely understood, they occur in conjunction with a signiﬁcant increase in circulating intact SDF-1a1-67 and suppression of truncated SDF-1a3-67. Physiologically, SDF-1a1-67 is a substrate of the DPP-4 enzyme. Administration of DPP-4i in patients with type 2 diabetes prevents the N-terminal cleavage and inactivation of intact plasma SDF-1a1-67 to truncated plasma SDF-1a3-67 (inset) by inhibiting the proteolytic activity of DPP-4. Potential distal tubular ion transport channels that may link DPP-4 inhibition to stimulating distal natriuresis include the Na+/Cl– thiazide-sensitive channel (3), and the epithelial sodium channel (ENac) (4). In contrast, SGLT-2 inhibitors (SGLT-2i) (1) and GLP-1RA (2) inhibit more proximal sodium reabsorption in the proximal convoluted tubules (PCTs) and inﬂuence tubuloglomerular feedback by increasing sodium delivery at the macula densa. SGLT-2i modify renal hemodynamic function. CCD, cortical collecting duct; DCT, distal convoluted tubule . 8 Sitagliptin Promotes Distal Tubular Natriuresis Diabetes Care

response(s) to coadministered SGLT-2 in- Chun, technician, and Paul Yip, clinical biochemist 6. White WB, Cannon CP, Heller SR, et al.; hibitors and DPP4 inhibitors will further (Department of Clinical Biochemistry, Toronto EXAMINE Investigators. Alogliptin after acute clarify the mechanistic rationale for com- General Hospital). coronary syndrome in patients with type 2 diabe- Funding. The study was an investigator-initiated tes. N Engl J Med 2013;369:1327–1335 bination therapy in the treatment of study, entirely planned and conducted under the 7. Scirica BM, Bhatt DL, Braunwald E, et al.; type 2 diabetes. scientific supervision of J.A.L., D.J.D., and D.Z.I.C. SAVOR-TIMI 53 Steering Committee and Investi- The current study is not without limi- Funding for the study and supply of sitagliptin and gators. Saxagliptin and cardiovascular outcomes tations. The study was adequately pow- placebo was provided by Merck Sharp & Dohme, in patients with type 2 diabetes mellitus. N Engl J as well as Diabetes Action Canada, a Strategic Pa- Med 2013;369:1317–1326 ered to determine the effect of sitagliptin tient Oriented Research initiative supported by 8. U.S. Food and Drug Administration . FDA Drug on FENa, but the small sample size limited the Canadian Institutes of Health Research. The and Safety Communication: FDA adds warnings our ability to perform in-depth analyses funders had no role in the study design, the anal- about heart failure risk to labels of type 2 diabetes around clinical factors such as duration of ysis or interpretation of the data or drafting the medicines containing saxagliptin and alogliptin, diabetes, degree of albuminuria, or the manuscript. The funders had no role in the de- 2016. Available from https://www.fda.gov/ cision to submit this manuscript for publication. Drugs/DrugSafety/ucm486096.htm. Accessed 12 effect of coadministration of other drugs. Duality of Interest. J.A.L. has received consulting May 2017 We attempted to reduce variability by fees or speaking honorarium, or both, from Novo 9. Zinman B, Lachin JM, Inzucchi SE. Empagliflo- using a placebo-controlled study design Nordisk, Eli Lilly & Co., Merck Sharp & Dohme, and zin, cardiovascular outcomes, and mortality in that included a prestudy protocol specify- AstraZeneca. H.H., L.H., and W.W. are employees type 2 diabetes. N Engl J Med 2016;374:1092– ing limits around dietary factors that can of Merck Co., Inc. D.J.D. has received consulting 1094 fl fees or honoraria from Arisaph Pharmaceuticals 10. Tonneijck L, Smits MM, Muskiet MH, et al. in uence GFR, including dietary protein Inc., Intarcia Therapeutics, Merck Research Labo- Renal effects of DPP-4 inhibitor sitagliptin or content, dietary sodium consumption, ratories, MedImmune, Novo Nordisk Inc., and Re- GLP-1 receptor agonist liraglutide in overweight background RAAS inhibition, and ambient ceptos, Inc. D.Z.I.C. has received consulting fees or patients with type 2 diabetes: a 12-week, random- glucose levels. Furthermore, there are speaking honorarium, or both, from Janssen, ized, double-blind, placebo-controlled trial. Dia- – currently no available assays to measure Boehringer Ingelheim-Eli, Lilly, AstraZeneca, betes Care 2016;39:2042 2050 Merck, and Sanofi, and has received operating 11. Schnurr E, Lahme W, Kuppers¨ H. Measure- tubuloglomerular feedback in humans funds from Janssen, Boehringer Ingelheim-Eli ment of renal clearance of inulin and PAH in the aside from measuring changes in renal Lilly, AstraZeneca, and Merck. No other potential steady state without urine collection. Clin Nephrol function in response to changes in tubular conflicts of interest relevant to this article were 1980;13:26–29 sodium delivery; therefore, we were un- reported. 12. Cherney DZI, Perkins BA, Soleymanlou N, Author Contributions. J.A.L., H.R., Y.L., and V.L. et al. Renal hemodynamic effect of sodium- able to directly quantify the effect of sita- conducted the study and performed the measure- glucose cotransporter 2 inhibition in patients gliptin on tubuloglomerular feedback. ments. J.A.L., D.J.D., and D.Z.I.C. designed and with type 1 diabetes mellitus. Circulation 2014; In conclusion, DPP-4 inhibition pro- planned the entire study, interpreted the data, 129:587–597 motes a distal tubular natriuresis after discussed the intellectual content, and wrote the 13. Gomez DM. Evaluation of renal resistances, short-term therapy that does not modify manuscript. J.A.L. and D.Z.I.C. served as the study with special reference to changes in essential hy- investigators and were present at and supervised all pertension. J Clin Invest 1951;30:1143–1155 renal or systemic hemodynamic function clinical visits during the study. H.R., Y.L., S.K., R.A., 14. Bjornstad P, Skrtiˇ cM,LytvynY,MaahsDM,´ (Fig. 3). Sitagliptin administration in- A.L.,and V.L.criticallyreviewedthemanuscript.L.E.L. Johnson RJ, Cherney DZ. The Gomez equations creased plasma levels of SDF-1a1-67,a was an independent statistician who performed and renal hemodynamic function in kidney dis- chemokine that promotes natriuresis in the statistical analysis and contributed to the graph- ease research. Am J Physiol Renal Physiol 2016; – preclinical studies. We conclude that ical representation of the data. S.K., R.A., and A.L. 311:F967 F975. 1-67 helped with data collection and data management. 15. Cherney DZ, Reich HN, Scholey JW, et al. The SDF-1a is a strong candidate effector H.H., L.H., and W.W. performed the SDF-1 and effect of aliskiren on urinary cytokine/chemokine that mediates the natriuretic effect(s) of sitagliptin assays and contributedtothecritical responses to clamped hyperglycaemia in type 1 sitagliptin, highlighting potentiation of review of the manuscript. D.Z.I.C. is the guarantor diabetes. Diabetologia 2013;56:2308–2317 SDF-1 activity as a new focus for thera- of this work, as such, had full access to all the data 16. Lovshin JA, Barnie A, DeAlmeida A, Logan A, peutic strategies targeting natriuresis. in the study and takes responsibility for the integrity Zinman B, Drucker DJ. Liraglutide promotes natri- of the data and the accuracy of the data analysis. uresis but does not increase circulating levels of Our current studies predict that combina- atrial natriuretic peptide in hypertensive subjects tion therapy with SGLT-2 inhibitors and References with type 2 diabetes. Diabetes Care 2015;38:132– DPP-4 inhibitors should not abrogate 1. Cornel JH, Bakris GL, Stevens SR, et al.; TECOS 139 the beneficial renal effects observed Study Group. Effect of sitagliptin on kidney func- 17. Cherney DZ, Reich HN, Jiang S, et al. Hyper- fi fi tion and respective cardiovascular outcomes in ltration and effect of nitric oxide inhibition on with SGLT-2 inhibitors alone, ndings type 2 diabetes: outcomes from TECOS. Diabetes renal and endothelial function in humans with with implications for the treatment of Care 2016;39:2304–2310 uncomplicated type 1 diabetes mellitus. Am J type 2 diabetes. 2. Turner LW, Nartey D, Stafford RS, Singh S, Physiol Regul Integr Comp Physiol 2012;303: Alexander GC. Ambulatory treatment of type 2 R710–R718 diabetes in the U.S., 1997-2012. Diabetes Care 18. Wang W, Choi BK, Li W, et al. Quantification 2014;37:985–992 of intact and truncated stromal cell-derived fac- Acknowledgments. The authors extend their 3. Lovshin JA, Drucker DJ. Incretin-based thera- tor-1a in circulation by immunoaffinity enrich- gratitude to all study participants who took part pies for type 2 diabetes mellitus. Nat Rev Endo- ment and tandem mass spectrometry. J Am Soc in this research study for their time and com- crinol 2009;5:262–269 Mass Spectrom 2014;25:614–625 mitment to this study protocol. The authors 4. Mulvihill EE, Drucker DJ. Pharmacology, phys- 19. Zeng W, Xu Y, Constanzer M, Woolf EJ. De- thank the nursing and administrative staff that iology, and mechanisms of action of dipeptidyl termination of sitagliptin in human plasma using helped perform this study, with special thanks to peptidase-4 inhibitors. Endocr Rev 2014;35:992– protein precipitation and tandem mass spectrom- Josephine Tse, RN, and Leslie Cham, RN (Renal 1019 etry. J Chromatogr B Analyt Technol Biomed Life Physiology Laboratory, Toronto General Hospi- 5. Green JB, Bethel MA, Armstrong PW, et al.; Sci 2010;878:1817–1823 tal), Donna Doherty (administrative assistant to TECOS Study Group. Effect of sitagliptin on cardio- 20. Rieg T, Gerasimova M, Murray F, et al. Natri- D.J.D.), and Marion Butt (administrative assis- vascular outcomes in type 2 diabetes. N Engl J uretic effect by exendin-4, but not the DPP-4 in- tant to D.Z.I.C.). The authors also thank Jenny Med 2015;373:232–242 hibitor alogliptin, is mediated via the GLP-1 care.diabetesjournals.org Lovshin and Associates 9

receptor and preserved in obese type 2 diabetic peptidase-4 inhibition and has protective roles in outcomes in type 2 diabetes. N Engl J Med mice. Am J Physiol Renal Physiol 2012;303:F963– progressive diabetic nephropathy. Kidney Int 2016;375:311–322 F971 2016;90:783–796 26. Chen LH, Advani SL, Thai K, et al. SDF-1/CXCR4 21. Anders HJ, Davis JM, Thurau K. Nephron pro- 24. Heerspink HJ, Perkins BA, Fitchett DH, Husain signaling preserves microvascular integrity and re- tection in diabetic kidney disease. N Engl J Med M, Cherney DZ. Sodium glucose cotransporter nal function in chronic kidney disease. PLoS One 2016;375:2096–2098 2 inhibitors in the treatment of diabetes mellitus: 2014;9:e92227 22. Perkins BA, Udell JA, Cherney DZ. No need cardiovascular and kidney effects, potential 27. Navarro-Gonzalez´ JF, Mora-Fernandez´ C. The to sugarcoat the message: is cardiovascular mechanisms, and clinical applications. Circulation role of inﬂammatory cytokines in diabetic ne- risk reduction from SGLT2 inhibition related 2016;134:752–772 phropathy. J Am Soc Nephrol 2008;19:433–442 to natriuresis? Am J Kidney Dis 2016;68:349–352 25. Marso SP, Daniels GH, Brown-Frandsen K, 28. Makdissi A, Ghanim H, Vora M, et al. Sitaglip- 23. Takashima S, Fujita H, Fujishima H, et al. Stro- et al.; LEADER Steering Committee; LEADER Trial tin exerts an antinﬂammatory action. J Clin Endo- mal cell-derived factor-1 is upregulated by dipeptidyl Investigators. Liraglutide and cardiovascular crinol Metab 2012;97:3333–3341