2032 Diabetes Volume 65, July 2016 François Briand,1 Eric Mayoux,2 Emmanuel Brousseau,1 Noémie Burr,1 Isabelle Urbain,1 Clément Costard,1 Michael Mark,2 and Thierry Sulpice1 Empagliflozin, via Switching Metabolism Toward Lipid Utilization, Moderately Increases LDL Cholesterol Levels Through Reduced LDL Catabolism Diabetes 2016;65:2032–2038 | DOI: 10.2337/db16-0049 In clinical trials, a small increase in LDL cholesterol has Specific sodium glucose cotransporter (SGLT) inhibitors been reported with sodium–glucose cotransporter 2 represent an emerging and promising new class of glucose- (SGLT2) inhibitors. The mechanisms by which the SGLT2 in- lowering drugs in the management of type 2 diabetes. hibitor empagliflozin increases LDL cholesterol levels were Theuniquemodeofactionofthisclassofnovelagents investigated in hamsters with diet-induced dyslipidemia. can effectively decrease blood glucose levels, independently Compared with vehicle, empagliflozin 30 mg/kg/day for of the insulin pathway, via increasing glucose excretion in 2 weeks significantly reduced fasting blood glucose by urine, i.e., glucosuria (1,2). Besides improved glycemic pa- 18%, with significant increase in fasting plasma LDL rameters, SGLT2 inhibitors have shown additional benefits cholesterol, free fatty acids, and total ketone bodies by such as body weight loss and blood pressure–lowering, with 25, 49, and 116%, respectively. In fasting conditions, low risk of hypoglycemia (3). However, an increase in LDL glycogen hepatic levels were further reduced by 84% with cholesterol (LDL-C) plasma levels has also been observed in empagliflozin, while 3-hydroxy-3-methylglutaryl-CoA patients treated with SGLT2 inhibitors (1). The mechanism reductase activity and total cholesterol hepatic levels by which SGLT2 inhibition raises LDL-C levels remains were 31 and 10% higher, respectively (both P < 0.05 vs. unclear. It has been suggested that the increase in LDL-C vehicle). A significant 20% reduction in hepatic LDL PHARMACOLOGY AND THERAPEUTICS receptor protein expression was also observed with may be partly due to hemoconcentration, as SGLT2 inhib- empagliflozin. Importantly, none of these parameters were itors induce volume contraction subsequent to increased changed by empagliflozin in fed conditions. Empagliflozin urinary volume (4,5). However, the transient diuretic effect significantly reduced the catabolism of 3H-cholesteryl of SGLT2 inhibitors may not completely contribute to oleate–labeled LDL injected intravenously by 20%, in- theobservedLDL-Cincrease.Wethereforeinvestigated dicating that empagliflozin raises LDL levels through the effects of the SGLT2 inhibitor empagliflozin in the reduced catabolism. Unexpectedly, empagliflozin also re- diet-induced insulin-resistant dyslipidemic golden Syrian duced intestinal cholesterol absorption in vivo, which led hamster, a validated preclinical model with cholesterol to a significant increase in LDL- and macrophage-derived metabolism similar to that of humans (6,7). cholesterol fecal excretion (both P < 0.05 vs. vehicle). These data suggest that empagliflozin, by switching RESEARCH DESIGN AND METHODS energy metabolism from carbohydrate to lipid utilization, moderately increases ketone production and LDL choles- All animal protocols were approved by the local (Comité terol levels. Interestingly, empagliflozin also reduces intes- régional d’éthique de Midi-Pyrénées) and national (Ministère tinal cholesterol absorption, which in turn promotes de l’Enseignement Supérieur et de la Recherche) ethics LDL- and macrophage-derived cholesterol fecal excretion. committees.MalegoldenSyrianhamsters(91–100 g, 1Physiogenex SAS, Prologue Biotech, Labège, France F.B. and E.M. contributed equally to this study. 2 Cardiometabolic Diseases Research, Boehringer Ingelheim, Biberach an der © 2016 by the American Diabetes Association. Readers may use this article as Riss, Germany long as the work is properly cited, the use is educational and not for profit, and Corresponding author: François Briand, [email protected]. the work is not altered. Received 11 January 2016 and accepted 31 March 2016. diabetes.diabetesjournals.org Briand and Associates 2033 6 weeks old; Elevage Janvier, Le Genest Saint Isle, France) kinetics, or macrophage-to-feces reverse cholesterol transport were fed ad libitum over 4 weeks with a high-fat/high- as previously described (6,7). Intestinal cholesterol absorp- cholesterol diet (0.5% cholesterol, 0.25% deoxycholate, tion was assessed after administration of 14C-cholesterol– 11.5% coconut oil, and 11.5% corn oil) with 10% fructose labeled olive oil by oral gavage and intraperitoneal injection in the drinking water as previously described (7). After of poloxamer-407 (a lipase inhibitor) to measure 14C-tracer 2 weeks of diet to induce dyslipidemia, hamsters were plasma tracer appearance at time 3, 5, and 6 h after oral ga- randomized into two sets of nonradioactive (set 1) or vage (6). Kinetics of LDL cholesteryl oleate were performed radioactive (set 2) experiments, according to blood glu- by intravenously injecting 3H-cholesteryl oleate–labeled cose and LDL-C levels in fed or overnight fasting condi- LDL in overnight fasted hamsters, previously isolated tions (fasting starting at 5:00 P.M. and blood collection at from hamsters fed the same high-fat/high-cholesterol ;8:00 A.M.), and were then treated orally for 2 weeks diet (7). Hamsters were kept fasted for the first 6 h of with vehicle or empagliflozin 30 mg/kg once daily. The the kinetic experiment and were then kept in individual dose was selected from a pilot study where glucose urine ex- cages with access to food and water for feces collection cretion was measured in this hamster model treated acutely over 72 h. Plasma 3H-tracer decay curve was monitored with empagliflozin 3, 10, and 30 mg/kg. The 30 mg/kg over 72 h after injection to calculate 3H-cholesteryl ole- dose was found to increase glucose urine excretion by ate LDL fractional catabolic rate using Simulation Anal- 1,200-fold versus vehicle, while the 3 and 10 mg/kg doses ysis and Modeling (SAAM II) software. Liver (collected showed a slighter effect (80- and 200-fold, respectively). At after 72 h) and feces were used to measure 3H-tracer the end of the treatment period, a first set of hamsters recovery in cholesterol and bile acid fraction after chem- was used to measure biochemical parameters using com- ical extraction (6,7). mercial kits in fed or overnight fasting conditions. Lipo- Macrophage-to-feces reverse cholesterol transport was protein total cholesterol profile was assessed using fast measured over 72 h after intraperitoneally injecting protein liquid chromatography analysis using one pooled 3H-cholesterol–labeled/oxidized LDL–loaded J774 mac- plasma sample (one pool per treatment group); Western rophages (6,7). In this experiment, hamsters were not blot analyses for LDL receptor protein expression and fecal fasted and had constant access to food and water over cholesterol mass excretion were performed as previously 72 h. Plasma 3H-tracer appearance was measured every described (7). A second set of hamsters underwent radio- 24 h, and liver (collected after 72 h) and feces (collected active tracer–based in vivo experiments to measure in- over 72 h) were used to measure 3H-tracer recovery in testinal cholesterol absorption, LDL cholesteryl esters cholesterol and bile acid fraction after chemical extraction. Table 1—Body weight and biochemical parameters in fed or overnight fast conditions Fed conditions Overnight fasting conditions Empagliflozin Empagliflozin Parameters Vehicle 30 mg/kg Vehicle 30 mg/kg Body weight (g) 110 6 2 114 6 2 110 6 2 111 6 1 Hematocrit (%) 49.8 6 0.7 47.9 6 0.6* 48.3 6 0.5 49.4 6 0.6 Plasma total protein (g/L) 81.2 6 1.8 81.9 6 1.8 79.6 6 2.5 76.0 6 1.0 Blood glucose (mg/dL) 86.0 6 5.5 88.6 6 2.6 73.4 6 4.0 59.9 6 2.5* Plasma total cholesterol (g/L) 4.0 6 0.2 4.0 6 0.2 3.0 6 0.1 2.9 6 0.2 Plasma LDL-C (g/L) 1.8 6 0.1 1.6 6 0.1 1.2 6 0.1 1.5 6 0.1* Plasma ketone bodies (mmol/L) 773 6 76 909 6 124 3,094 6 171 6,685 6 510‡ Plasma free fatty acids (mmol/L) 0.62 6 0.06 0.70 6 0.05 0.45 6 0.03 0.67 6 0.05† Plasma free glycerol (g/L) 0.023 6 76 0.033 6 0.004* 0.009 6 0.001 0.011 6 0.001 Liver weight (g) 5.61 6 0.13 6.04 6 0.13* 4.90 6 0.13 4.75 6 0.06 Hepatic triglycerides (mg/g liver) 15.1 6 0.9 16.9 6 0.1 16.6 6 1.3 15.3 6 0.7 Hepatic cholesterol (mg/g liver) 38.9 6 0.8 40.2 6 1.7 43.1 6 1.9 47.7 6 1.1* Hepatic fatty acids (mmol/g liver) 362 6 9 352 6 12 386 6 11 418 6 8* Hepatic ketone bodies (mmol/g liver) 12.4 6 0.5 12.1 6 0.5 14.7 6 0.6 16.8 6 0.8 Hepatic pyruvate (mmol/g liver) 6.2 6 0.5 6.4 6 0.3 6.7 6 0.4 8.0 6 0.3* Hepatic HMG-CoAred activity (mU/mg protein) 0.302 6 0.034 0.357 6 0.040 0.255 6 0.019 0.334 6 0.028* Hepatic glycogen (mg/g liver) 39.1 6 3.9 37.3 6 2.2 4.31 6 0.64 0.7 6 0.4‡ Data are mean 6 SEM. n =9–10 hamsters/group. HMG-CoAred, HMG-CoA reductase. *P , 0.05 vs. vehicle. †P , 0.01 vs. vehicle. ‡P , 0.001 vs. vehicle. 2034 SGLT2 Inhibition and LDL-C Diabetes Volume 65, July 2016 Data are expressed as mean 6 SEM.