Empagliflozin, via Switching Metabolism Toward Lipid Utilization, Moderately Increases LDL Cholesterol Levels Through Reduced LDL Catabolism

Specific sodium glucose cotransporter (SGLT) inhibitors represent an emerging and promising new class of glucose-lowering drugs in the management of type 2 diabetes. The unique mode of action of this class of novel agents can effectively decrease blood glucose levels, independently of the insulin pathway, via increasing glucose excretion in urine, i.e., glucosuria (1,2). Besides improved glycemic parameters, SGLT2 inhibitors have shown additional benefits such as body weight loss and blood pressure–lowering, with low risk of hypoglycemia (3). However, an increase in LDL cholesterol (LDL-C) plasma levels has also been observed in patients treated with SGLT2 inhibitors (1). The mechanism by which SGLT2 inhibition raises LDL-C levels remains unclear. It has been suggested that the increase in LDL-C may be partly due to hemoconcentration, as SGLT2 inhibitors induce volume contraction subsequent to increased urinary volume (4,5). However, the transient diuretic effect of SGLT2 inhibitors may not completely contribute to the observed LDL-C increase. We therefore investigated the effects of the SGLT2 inhibitor empagliflozin in the diet-induced insulin-resistant dyslipidemic golden Syrian hamster, a validated preclinical model with cholesterol metabolism similar to that of humans (6,7).

All animal protocols were approved by the local (Comité régional d’éthique de Midi-Pyrénées) and national (Ministère de l’Enseignement Supérieur et de la Recherche) ethics committees. Male golden Syrian hamsters (91–100 g, 6 weeks old; Elevage Janvier, Le Genest Saint Isle, France) were fed ad libitum over 4 weeks with a high-fat/high-cholesterol diet (0.5% cholesterol, 0.25% deoxycholate, 11.5% coconut oil, and 11.5% corn oil) with 10% fructose in the drinking water as previously described (7). After 2 weeks of diet to induce dyslipidemia, hamsters were randomized into two sets of nonradioactive (set 1) or radioactive (set 2) experiments, according to blood glucose and LDL-C levels in fed or overnight fasting conditions (fasting starting at 5:00 p.m. and blood collection at ∼8:00 a.m.), and were then treated orally for 2 weeks with vehicle or empagliflozin 30 mg/kg once daily. The dose was selected from a pilot study where glucose urine excretion was measured in this hamster model treated acutely with empagliflozin 3, 10, and 30 mg/kg. The 30 mg/kg dose was found to increase glucose urine excretion by 1,200-fold versus vehicle, while the 3 and 10 mg/kg doses showed a slighter effect (80- and 200-fold, respectively). At the end of the treatment period, a first set of hamsters was used to measure biochemical parameters using commercial kits in fed or overnight fasting conditions. Lipoprotein total cholesterol profile was assessed using fast protein liquid chromatography analysis using one pooled plasma sample (one pool per treatment group); Western blot analyses for LDL receptor protein expression and fecal cholesterol mass excretion were performed as previously described (7). A second set of hamsters underwent radioactive tracer–based in vivo experiments to measure intestinal cholesterol absorption, LDL cholesteryl esters kinetics, or macrophage-to-feces reverse cholesterol transport as previously described (6,7). Intestinal cholesterol absorption was assessed after administration of ¹⁴C-cholesterol–labeled olive oil by oral gavage and intraperitoneal injection of poloxamer-407 (a lipase inhibitor) to measure ¹⁴C-tracer plasma tracer appearance at time 3, 5, and 6 h after oral gavage (6). Kinetics of LDL cholesteryl oleate were performed by intravenously injecting ³H-cholesteryl oleate–labeled LDL in overnight fasted hamsters, previously isolated from hamsters fed the same high-fat/high-cholesterol diet (7). Hamsters were kept fasted for the first 6 h of the kinetic experiment and were then kept in individual cages with access to food and water for feces collection over 72 h. Plasma ³H-tracer decay curve was monitored over 72 h after injection to calculate ³H-cholesteryl oleate LDL fractional catabolic rate using Simulation Analysis and Modeling (SAAM II) software. Liver (collected after 72 h) and feces were used to measure ³H-tracer recovery in cholesterol and bile acid fraction after chemical extraction (6,7).

Macrophage-to-feces reverse cholesterol transport was measured over 72 h after intraperitoneally injecting ³H-cholesterol–labeled/oxidized LDL–loaded J774 macrophages (6,7). In this experiment, hamsters were not fasted and had constant access to food and water over 72 h. Plasma ³H-tracer appearance was measured every 24 h, and liver (collected after 72 h) and feces (collected over 72 h) were used to measure ³H-tracer recovery in cholesterol and bile acid fraction after chemical extraction.

Data are expressed as mean ± SEM. Unpaired Student t test or one-way ANOVA plus Dunnett posttest was used for statistical analysis. A P < 0.05 was considered significant.

Empagliflozin treatment significantly triggered more biochemical parameter changes in the overnight fasting condition than in the fed condition (Table 1).

Plasma LDL-C levels were found to be higher by 25% in hamsters treated with empagliflozin (P < 0.05 vs. vehicle) only in the fasting condition. Concomitantly, fasting blood glucose was reduced by 18% (P < 0.05 vs. vehicle), while plasma total ketone bodies and free fatty acids were raised by 116% (P < 0.001 vs. vehicle) and 49% (P < 0.01 vs. vehicle), respectively. Hepatic total cholesterol and fatty acid levels in overnight fasting conditions were 10 and 8% higher in hamsters treated with empagliflozin (both P < 0.05 vs. vehicle). In addition, hepatic total ketone body levels were 14% higher with empagliflozin, although not significantly. Hepatic pyruvate levels and 3-hydroxy-3-methylglutaryl (HMG)-CoA reductase activity were 19 and 31% higher, respectively, in overnight fasted hamsters treated with empagliflozin (both P < 0.05 vs. vehicle). Compared with vehicle, hepatic glycogen levels were dramatically blunted by 84% with empagliflozin (P < 0.001 vs. vehicle). In sharp contrast with the fasting condition, empagliflozin showed limited effects on biochemical parameters measured in the fed condition with the exception of minor differences on hematocrit, liver weight, and plasma free glycerol compared with vehicle.

For further confirmation of the raise in plasma LDL-C levels, total cholesterol lipoprotein profile in overnight fasted hamsters was measured by fast protein liquid chromatography (Fig. 1A). As expected, empagliflozin led to higher total cholesterol levels in fractions corresponding to LDL. Since higher plasma LDL-C may be linked to lower LDL receptor expression, Western blot analysis was also performed using liver samples collected from overnight fasted hamsters. Compared with vehicle, hepatic protein expression of the LDL receptor was found to be reduced by 20% (Fig. 1B) in overnight fasted hamsters treated with empagliflozin (P < 0.05 vs. vehicle).

As higher LDL-C levels could also be related to increased intestinal cholesterol absorption, this mechanism was also measured in vivo after oral administration of ¹⁴C-cholesterol–labeled olive oil. Strikingly, hamsters treated with empagliflozin showed a ¹⁴C-tracer plasma appearance reduced by up to 40% over 6 h after ¹⁴C-tracer administration, indicating lower intestinal cholesterol absorption (Fig. 1C). In agreement with the lower intestinal cholesterol absorption, fecal cholesterol mass excretion was 49% higher in hamsters treated with empagliflozin (Fig. 1D) compared with vehicle (P < 0.01).

We next investigated LDL-C metabolism in vivo by injecting ³H-cholesteryl oleate–labeled LDL intravenously in hamsters. Empagliflozin treatment resulted in slowed ³H-tracer decay curve over 72 h, leading to a 20% reduction in LDL cholesteryl ester catabolism (Fig. 2A), compared with vehicle (P < 0.05). At 72 h after ³H-cholesteryl oleate–labeled LDL, hepatic ³H-tracer recoveries in the whole liver and the hepatic cholesterol fraction were respectively reduced by 11% (P < 0.01 vs. vehicle) and 19% (P < 0.001 vs. vehicle) with empagliflozin treatment (Fig. 2B). As a result of reduced cholesterol absorption in the intestine, LDL-derived ³H-cholesterol fecal excretion was 26% higher (P < 0.05 vs. vehicle) in hamsters treated with empagliflozin (Fig. 2C).

For investigation of macrophage-to-feces reverse cholesterol transport in vivo, hamsters were injected intraperitoneally with ³H-cholesterol–labeled/oxidized LDL–loaded macrophages. Compared with vehicle, empagliflozin did not change plasma ³H-tracer appearance over 72 h (Fig. 2D). Hepatic ³H-tracer recoveries in the whole liver and the hepatic cholesterol fraction tended to be reduced with empagliflozin, although this was not significant (Fig. 2E). However, ³H-cholesterol fecal excretion (Fig. 2F) was increased by 29% in hamsters treated with empagliflozin (P < 0.05). These data indicate that reduced intestinal cholesterol absorption with empagliflozin treatment promotes fecal excretion of cholesterol deriving from the macrophage.

The current study indicates that empagliflozin raises LDL-C levels only in fasting conditions through reduction in LDL-C catabolism and alters cholesterol metabolism at both the hepatic and intestinal levels in hamsters.

Overnight fasted hamsters treated with empagliflozin showed higher LDL-C levels concomitant with higher free fatty acids and total ketone body plasma levels. The higher level of total ketone bodies and fatty acids is in agreement with previous reports indicating that chronic treatment with SGLT2 inhibitors induces ketogenesis and a metabolism switch toward lipid oxidation to counterbalance the carbohydrate restriction in the fasting state (8–10). The excretion of glucose via urine and related calorie loss with SGLT2 inhibition therefore replicate starvation shift from carbohydrate to lipid utilization for energy in the fasting state (11). Chronic SGLT2 inhibition also seems to mimic the LDL-raising effects of ketogenic diet, in which LDL-C levels correlate with blood ketone body levels (12). In the current study, evidence for a metabolic shift toward fat utilization was also observed at the liver level (e.g., hepatic glycogen and pyruvate levels) in fasted hamsters treated with empagliflozin. The increased hepatic fatty acids levels may fuel the pool of acetyl-CoA, an important metabolic branch point, as a source for both ketone body production and hepatic cholesterol synthesis (13), with the latter associated with higher HMG-CoA reductase activity and hepatic total cholesterol levels. As hepatic levels of cholesterol regulate LDL receptor expression (14,15), empagliflozin treatment lowered LDL receptor expression and plasma LDL-C catabolism, which in turn increased LDL-C plasma levels. Although a raise in LDL-C levels is seen as an increase in cardiovascular event risk (16), it is probably not so prominent with empagliflozin. Indeed, the EMPA-REG OUTCOME study (BI 10773 [Empagliflozin] Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients) recently delivered a spectacular 38% reduction in cardiovascular mortality and 35% reduction in hospitalization with heart failure, with no change in event rate of nonfatal myocardial infarction and nonfatal stroke (17). Moreover, our study revealed that even after chronic treatment with empagliflozin, the increase in LDL-C was only observed in the overnight fasted condition. In the clinical setting, LDL-C levels are routinely assessed from plasma collected in the fasted state. Therefore, clinical investigations evaluating the effects of empagliflozin on LDL-C levels in fed conditions would be of interest. In addition, our in vivo experiments also highlighted potential antiatherogenic mechanisms induced by empagliflozin, such as LDL- and macrophage-derived fecal cholesterol excretion. Macrophage-to-feces reverse cholesterol transport is known to be inversely correlated with atherosclerosis (18), and an enhanced excretion of LDL-derived cholesterol in the feces theoretically prevents its accumulation in the arterial wall. Whether these mechanisms, besides body weight loss and blood pressure lowering, contribute to the reduced cardiovascular risk in patients treated with empagliflozin (17) remains to be further investigated.

Another point of investigation is the reduced intestinal cholesterol absorption observed in hamsters treated with empagliflozin. Since a balance exists between hepatic cholesterol synthesis and intestinal cholesterol absorption (19), the lower intestinal cholesterol absorption may therefore result from the stimulation of hepatic cholesterol synthesis by empagliflozin. However, the molecular mechanism by which empagliflozin alters intestinal cholesterol metabolism remains to be elucidated.

In conclusion, the current study suggests that empagliflozin raises LDL-C levels only in the fasting condition by reducing LDL receptor expression and LDL-C catabolism. As illustrated in Fig. 3, the proposed mechanism leading to the LDL-C increase originates from the metabolic switch toward lipid utilization, which triggers in parallel a moderate activation of ketogenesis pathway and hepatic cholesterol synthesis within the liver. Future studies to test whether SGLT2 inhibitors have a similar rhythmic effect in plasma from patients fasted versus fed would be required.

Acknowledgments. The authors thank Dominique Lopes for animal care, Marjolaine Quinsat and Hélène Lakehal for technical assistance, and Aurélie Couderc for quality control (all of Physiogenex).

Duality of Interest. This work has been funded by Boehringer Ingelheim. E.M. and M.M. are employees of Boehringer Ingelheim. F.B., E.B., N.B., I.U., C.C., and T.S. are employees of Physiogenex.

Author Contributions. F.B., E.M., M.M., and T.S. designed research. F.B., E.B., N.B., I.U., and C.C. conducted research. F.B. and E.M. analyzed data and wrote the manuscript. T.S. had the primary responsibility for the final content. All authors read and approved the final manuscript. T.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 51st Annual Meeting of the European Association for the Study of Diabetes, Stockholm, Sweden, 14–18 September 2015.