To examine the mechanisms responsible for the increase in glucose and ketone production caused by empagliflozin in patients with type 2 diabetes mellitus (T2DM).
Twelve subjects with T2DM participated in two studies performed in random order. In study 1, endogenous glucose production (EGP) was measured with 8-h infusion of 6,6,D2-glucose. Three hours after the start of 6,6,D2-glucose infusion, subjects ingested 25 mg empagliflozin (n = 8) or placebo (n = 4), and norepinephrine (NE) turnover was measured before and after empagliflozin ingestion with 3H-NE infusion. Study 2 was similar to study 1 but performed under pancreatic clamp conditions.
When empagliflozin was ingested under fasting conditions, EGP increased by 31% in association with a decrease in plasma glucose (−34 mg/dL) and insulin (−52%) concentrations and increases in plasma glucagon (+19%), free fatty acid (FFA) (+29%), and β-hydroxybutyrate (+48%) concentrations. When empagliflozin was ingested under pancreatic clamp conditions, plasma insulin and glucagon concentrations remained unchanged, and the increase in plasma FFA and ketone concentrations was completely blocked, while the increase in EGP persisted. Total-body NE turnover rate was greater in subjects receiving empagliflozin (+67%) compared with placebo under both fasting and pancreatic clamp conditions. No difference in plasma NE concentration was observed in either study.
The decrease in plasma insulin and increase in plasma glucagon concentration caused by empagliflozin is responsible for the increase in plasma FFA concentration and ketone production. The increase in EGP caused by empagliflozin is independent of the change in plasma insulin or glucagon concentrations and is likely explained by the increase in NE turnover.
Introduction
Sodium–glucose cotransport 2 inhibitors (SGLT2i) lower the plasma glucose concentration by inhibiting renal glucose reabsorption and producing glucosuria (1,2). We (2–5) and others (6,7) have shown that members of this class of drugs exert multiple metabolic actions in patients with type 2 diabetes mellitus (T2DM), including an increase in basal endogenous (primarily hepatic) glucose production (EGP) (3,7), plasma free fatty acid (FFA) concentration, and ketone production (4,6). These metabolic actions of SGLT2i have important clinical implications. The stimulation of ketogenesis results in increased risk for euglycemic ketoacidosis (8,9), while the increase in EGP offsets by ∼50% the amount of glucose excreted in the urine (3), thereby blunting decrease in the plasma glucose concentration and attenuating the clinical efficacy of SGLT2i. Therefore, strategies that offset these actions of SGLT2i can amplify the clinical efficacy of the drug and decrease the risk of untoward side effects associated with their use.
The primary site of SGLT2i action is the proximal tubule of the kidney (1). Thus, the systemic metabolic actions of SGLT2i (2) are not a direct effect of sodium–glucose cotransporter 2 (SGLT2) inhibition but rather due to other mechanisms that are not fully understood. The decrease in plasma glucose concentration caused by SGLT2i is associated with an increase in plasma glucagon concentration and a decrease in plasma insulin concentration (3,5). However, we previously have shown that the change in plasma insulin and glucagon concentrations caused by SGLT2i does not play a role in the stimulation of EGP (10). Further, we demonstrated that removal of both native kidneys in renal transplant subjects attenuates the increase in EGP caused by dapagliflozin (11), indicating an important role of the renal nerves in the increase in EGP. Further, coadministration of pioglitazone (12) or glucagon-like peptide 1 receptor agonists (GLP-1 RAs) (13), which decrease the plasma FFA concentration, with dapagliflozin prevents the increase in plasma ketone concentration, while GLP-1 RA, which stimulates insulin secretion and inhibits glucagon secretion, failed to block the increase in EGP (14), suggesting that distinct mechanisms underlie the increase in EGP and ketone production by SGLT2i.
The aim of the current study was to elucidate the mechanisms responsible for the increase in EGP and plasma ketone concentration following SGLT2i administration. To achieve this goal, we measured EGP, total body norepinephrine (NE) turnover rate, and plasma insulin, glucagon, FFA, and β-hydroxybutyrate (b-HB) concentrations before and after empagliflozin administration during the fasting state and under pancreatic clamp conditions.
Research Design and Methods
Subjects
Twelve patients with T2DM participated in the current study. Except for diabetes, all subjects were in good general health based on medical history, physical examination, blood chemistries, complete blood count, thyroid function, urinalysis, and electrocardiogram. Patients had stable (±1.5 kg) body weight over the 3 months before the study, and no subject participated in any excessively heavy exercise program. Patients were drug naive (n = 4) or on a stable dose of metformin with (n = 4) or without (n = 4) a sulfonylurea. The distribution of background medications was similar in the empagliflozin and placebo groups (Table 1). All background medications were continued without change throughout the study. Subjects with evidence of proliferative diabetic retinopathy or serum creatinine of >1.4 mg/dL (women) or >1.5 mg/dL (men), or with estimated glomerular filtration rate (eGFR) of <60 mL/min/1.73 m2 were excluded.
Baseline patients characteristics
. | Empagliflozin . | Placebo . | P value . |
---|---|---|---|
Age (years) | 52 ± 4 | 54 ± 3 | NS |
Sex (male/female) | 4/4 | 3/1 | NS |
BMI (kg/m2) | 32.1 ± 0.9 | 31.3 ± 1.1 | NS |
FPG (mg/dL) | 121 ± 5 | 126 ± 4 | NS |
HbA1c (%) | 7.3 ± 0.6 | 7.6 ± 0.9 | NS |
Diabetes duration (years) | 5.2 ± 1.3 | 5.8 ± 1.7 | NS |
GFR (mL/min/1.73 m2) | 96 ± 6 | 98 ± 7 | NS |
Baseline medications (patients, n) | |||
Drug naive | 2 | 1 | |
Metformin | 3 | 1 | |
Metformin + sulfonylurea | 2 | 2 |
. | Empagliflozin . | Placebo . | P value . |
---|---|---|---|
Age (years) | 52 ± 4 | 54 ± 3 | NS |
Sex (male/female) | 4/4 | 3/1 | NS |
BMI (kg/m2) | 32.1 ± 0.9 | 31.3 ± 1.1 | NS |
FPG (mg/dL) | 121 ± 5 | 126 ± 4 | NS |
HbA1c (%) | 7.3 ± 0.6 | 7.6 ± 0.9 | NS |
Diabetes duration (years) | 5.2 ± 1.3 | 5.8 ± 1.7 | NS |
GFR (mL/min/1.73 m2) | 96 ± 6 | 98 ± 7 | NS |
Baseline medications (patients, n) | |||
Drug naive | 2 | 1 | |
Metformin | 3 | 1 | |
Metformin + sulfonylurea | 2 | 2 |
Data are mean ± SEM or n.
The study protocol was approved by The University of Texas Health Science Center at San Antonio Institutional Review Board (San Antonio, TX), and informed written consent was obtained from all subjects before their participation.
Experimental Design
All studies were performed at the Texas Diabetes Institute Clinical Research Center (San Antonio, TX) at 6:00 a.m. after a 10-h overnight fast. After confirming eligibility, individual subjects were consecutively randomized to receive empagliflozin or placebo in 2:1 ratio. Subject stratification was done according to the following parameters: age, BMI, diabetes duration, fasting plasma glucose (FPG), eGFR, and hemoglobin A1c (HbA1c). Each subject participated in two studies that were performed in random order with a 7- to 10-day interval between studies. Background medications (metformin and/or sulfonylurea) were withheld in the morning of the study day. In study 1, EGP was measured with a prime-continuous 6,6,D2-glucose infusion, and total body NE turnover was measured with 3H-NE infusion before and after empagliflozin administration. In study 2, EGP and NE turnover were measured under pancreatic clamp conditions.
Study 1
Subjects reported to the Clinical Research Center at 6:00 a.m., after an overnight fast. A catheter was placed into an antecubital vein for infusion of all test substances, and a second catheter was inserted retrogradely into a vein on the dorsum of the hand, which was placed in a heated box (70°C) for sampling of arterialized blood. A prime (4 mg/kg)-continuous (0.4 mg/min) infusion of 6,6,D2-glucose (Cambridge Isotope Laboratories, Andover, MA) was started at 6:00 a.m. and continued until study end at 2:00 p.m. At 8:00 a.m., a prime (3.8 µCi)-continuous (0.38 µCi/min) infusion of 3H-NE (levo-7-3H-NE; Perkin-Elmer, Waltham, MA) was started and continued for 60 min. Arterialized blood samples were collected before the start and between the 40- to 60-min time period after the start of 3H-NE infusion. Previous studies (15) have demonstrated that the plasma 3H-NE specific activity reaches a steady state within 30 min without significant change (<5%) in plasma 3H-dihydroxyphenylglycol (the primary intraneuronal metabolite of NE) radioactivity.
At 8:30 a.m. (2.5 h after the start of 6,6,D2-glucose infusion), arterialized blood samples were drawn for the measurement of baseline plasma 6,6,D2-glucose enrichment and plasma glucose, insulin, glucagon, FFA, and β-HB concentrations. At time zero (9:00 a.m.), subjects ingested empagliflozin 25 mg (n = 8) or placebo (n = 4). After 9:00 a.m. (time zero), arterialized blood samples were obtained every 10 to 30 min for 300 min for determination of plasma glucose, insulin, glucagon, FFA, and β-HB concentrations and 6,6,D2-glucose enrichment. At 240 min, a second prime-continuous infusion of 3H-NE was started and continued for 60 min (i.e., from 240 to 300 min). Arterialized blood samples were collected before the start (at 240 min) and between 280 and 300 min for the measurement of plasma NE concentration and 3H-NE specific activity.
At 6:00 a.m., subjects voided, and the urine was discarded. Urine was collected from 6:00 a.m. to 9:00 a.m. (baseline period) and from 9:00 a.m. to 2:00 p.m. (drug treatment period). Urinary volume and glucose concentration were measured to determine the urinary glucose excretion rate. At 2:00 p.m., subjects received a meal and returned home.
Study 2
EGP and NE turnover were measured as described in study 1, with the following exceptions. Somatostatin infusion (750 µg/h) was started 5 min before the start of the 6,6,D2-glucose infusion, and the basal plasma insulin and glucagon concentrations were maintained with infusion of insulin (0.1 mU/kg ⋅ min) and glucagon (0.3 ng/kg ⋅ min). Somatostatin, insulin, and glucagon infusions were continued until the end of study (minute 300).
Analytical Methods
Plasma glucose was measured using the glucose oxidase method (Analox Reagent Instruments, International Point of Care, Toronto, Ontario, Canada). Plasma insulin (IBL America, Minneapolis, MN) and glucagon (MilliporeSigma, Burlington, MA) concentrations were measured by radioimmunoassay. Plasma NE concentration was determined by ELISA (ALPCO, Salem, NH) as previously described (11). Plasma 3H-NE radioactivity was determined on deproteinized barium/zinc plasma samples.
Plasma isotope enrichment of 6,6,D2-glucose was measured as previously described (11) on an AB Sciex API 4000 triple quadrupole mass spectrometer (AB Sciex, Concord, Ontario, Canada), equipped with ESI Turbo-V source and coupled with an Agilent 1290 Infinity HPLC system (Agilent, Santa Clara, CA).
Data Analysis
Under steady-state postabsorptive conditions, the basal rate of EGP equals the 6,6,D2-glucose infusion rate divided by steady-state plasma enrichment of 6,6,D2-glucose. After drug administration, nonsteady conditions for 6,6,D2-glucose prevail, and the total body Ra is calculated using the Steele equation. The change in EGP during the last hour of the study (i.e., 240–300 min) from baseline was considered the drug effect on EGP and was compared between empagliflozin and placebo with ANOVA.
Total body NE turnover rate was calculated as the 3H-NE infusion rate (decay per minute [dpm]/min) divided by the steady-state plasma 3H-NE specific activity (dpm/pg). NE clearance rate was calculated as 3H-NE infusion rate (dpm/min) divided by the steady-state plasma 3H-NE (dpm/mL) radioactivity.
All values are presented as mean ± SEM. The comparison of empagliflozin effect (before and after drug administration) versus placebo was performed with two-way ANOVA using the SPSS software package. P value < 0.05 was considered statistically significant.
Data and Resource Availability
The data sets generated and analyzed during the current study are available from the corresponding author upon reasonable request.
Results
Subjects in the empagliflozin and placebo groups were well matched in age, sex, BMI, duration of diabetes, FPG, HbA1c, and eGFR (Table 1).
In studies 1 and 2, urinary glucose excretion during the measurement of EGP increased to 2.82 ± 0.62 g/h and 2.0 ± 0.31 g/h, respectively. Urine volume was greater in subjects receiving empagliflozin compared with subjects receiving placebo in both study 1 and study 2 (966 ± 63 vs. 360 ± 42 mL, respectively, in study 1, and 1,113 ± 82 mL vs. 667 ± 86 mL, respectively, in study 2, both P < 0.05).
Study 1
During the last hour (i.e., 240–300 min) following empagliflozin ingestion, the FPG decreased by 36 ± 4 mg/dL compared with a 13 ± 3 mg/dL decrease in subjects receiving placebo (Fig. 1A). The placebo-subtracted decrease in plasma glucose concentration following empagliflozin was 23 mg/dL (P = 0.03). Following empagliflozin, EGP progressively increased, and, during the last hour (240–300 min), was 0.73 ± 0.12 mg/kg ⋅ min above baseline (Fig. 1B) compared with a decrease of −0.78 ± 0.15 mg/kg ⋅ min in subjects receiving placebo. The placebo-subtracted difference in the change in EGP from baseline to end of study following empagliflozin was 1.55 ± 0.22 mg/kg ⋅ min (P = 0.005). Consistent with previous studies (3), empagliflozin caused a small but significant increase in plasma glucagon concentration (50 ± 5 ng/mL vs. 42 ± 4 ng/mL, P = 0.003) compared with a small decrease (−3 ng/mL) in subjects receiving placebo (Fig. 1D). The difference in the change in plasma glucagon between the two groups was highly statistically significant (P < 0.0001, two-way ANOVA). Plasma insulin concentration decreased significantly following empagliflozin (12 to 3 μU/mL) and did not change following placebo (11 vs. 9 μU/mL) (P < 0.05 by ANOVA) (Fig. 1C).
Change from baseline in plasma glucose concentration (A) after empagliflozin and placebo (time 0) administration. The basal rate of EGP (B). Plasma insulin (C), glucagon (D), FFA (E), and β-hydroxybutyrate (F) concentrations during EGP measurement. bHGP, basal hepatic glucose production. *P < 0.05.
Change from baseline in plasma glucose concentration (A) after empagliflozin and placebo (time 0) administration. The basal rate of EGP (B). Plasma insulin (C), glucagon (D), FFA (E), and β-hydroxybutyrate (F) concentrations during EGP measurement. bHGP, basal hepatic glucose production. *P < 0.05.
Empagliflozin caused a significant increase in plasma FFA concentration (Fig. 1E), from 0.65 ± 0.06 to 0.84 ± 0.04 mmol/L (P < 0.05) during the last hour (240–300 min) of the study, compared with a small decrease (0.48 ± 0.10 to 0.42 ± 0.08 mmol/L) in subjects receiving placebo. The change in plasma FFA concentration from baseline to last hour in subjects receiving empagliflozin versus placebo (+0.22 ± 0.09 vs. −0.04 ± 0.09) was statistically significant (P < 0.05). Empagliflozin caused a 43% increase in plasma β-HB concentration (from 0.21 ± 0.02 to 0.31 ± 0.03 mmol/L, P < 0.05) compared with a small decrease (from 0.21 ± 0.02 to 0.17 ± 0.02 mmol/L, P = NS) in subjects receiving placebo. The difference in the change in plasma β-HB concentration from baseline to end of study between the two groups (+0.11 ± 0.03 vs. −0.04 ± 0.06) was statistically significant (P < 0.05).
Fasting plasma NE concentration was 241 ± 47 pg/mL and 219 ± 72 pg/mL in subjects receiving empagliflozin and placebo, respectively (Table 2), and remained unchanged during the last hour (240–300 min) of study (299 ± 76 and 206 ± 25 pg/mL, respectively). Total body NE turnover was 440 ± 92 ng/min at baseline and increased to 735 ± 90 ng/min in subjects receiving empagliflozin (P = 0.02), while it decreased slightly in subjects receiving placebo (346 ± 55 to 298 ± 44 ng/min, P = NS). The difference in the change from baseline to end of study between the two groups in NE turnover rate was statistically significant (P = 0.04). NE clearance rate was not affected by empagliflozin or placebo (Table 2).
Kinetic parameters of total body NE turnover
. | Empagliflozin . | Placebo . | Two-way ANOVA . |
---|---|---|---|
NE concentration (pg/mL) | |||
Study 1 | |||
Baseline | 241 ± 47 | 219 ± 72 | |
240–300 min | 299 ± 76 | 206 ± 25 | |
Change from baseline | +58 ± 62 | −9 ± 58 | NS |
Study 2 | |||
Baseline | 287 ± 91 | 229 ± 61 | |
240–300 min | 247 ± 81 | 111 ± 65 | |
Change from baseline | −40 ± 85 | −118 ± 66 | NS |
Counts (dpm/mL) | |||
Study 1 | |||
Baseline | 74 ± 2 | 109 ± 3 | |
240–300 | 63 ± 2 | 112 ± 4 | |
Change from Baseline | −11 ± 3 | +3 ± 2 | NS |
Study 2 | |||
Baseline | 108 ± 4 | 82 ± 3 | |
240–300 min | 88 ± 3 | 70 ± 2 | |
Change from baseline | −20 ± 3 | −12 ± 3 | NS |
NE specific act (dpm/ng) | |||
Study 1 | |||
Baseline | 310 ± 81 | 490 ± 91 | |
240–300 min | 210 ± 42 | 541 ± 81 | |
Change from baseline | −100 ± 28 | +50 ± 22 | 0.03 |
Study 2 | |||
Baseline | 376 ± 78 | 352 ± 62 | |
240–300 min | 352 ± 66 | 631 ± 87 | |
Change from baseline | −24 ± 18 | +279 ± 98 | 0.02 |
NE turnover (ng/min) | |||
Study 1 | |||
Baseline | 440 ± 82 | 346 ± 59 | |
240–300 min | 735 ± 90 | 280 ± 44 | |
Change from baseline | +295 ± 72 | −66 ± 85 | 0.04 |
Study 2 | |||
Baseline | 437 ± 96 | 481 ± 112 | |
240–300 min | 456 ± 108 | 298 ± 58 | |
Change from baseline | +18 ± 21 | −183 ± 22 | 0.05 |
NE clearance (L/min) | |||
Study 1 | |||
Baseline | 1.81 ± 0.21 | 2.25 ± 0.22 | |
240–300 min | 2.14 ± 0.24 | 2.48 ± 0.26 | |
Change from baseline | +0.33 ± 0.12 | +0.23 ± 0.11 | NS |
Study 2 | |||
Baseline | 1.84 ± 0.24 | 2.73 ± 0.34 | |
240–300 min | 2.09 ± 0.31 | 2.92 ± 0.36 | |
Change from baseline | +0.25 ± 0.14 | 0.19 ± 0.10 | NS |
. | Empagliflozin . | Placebo . | Two-way ANOVA . |
---|---|---|---|
NE concentration (pg/mL) | |||
Study 1 | |||
Baseline | 241 ± 47 | 219 ± 72 | |
240–300 min | 299 ± 76 | 206 ± 25 | |
Change from baseline | +58 ± 62 | −9 ± 58 | NS |
Study 2 | |||
Baseline | 287 ± 91 | 229 ± 61 | |
240–300 min | 247 ± 81 | 111 ± 65 | |
Change from baseline | −40 ± 85 | −118 ± 66 | NS |
Counts (dpm/mL) | |||
Study 1 | |||
Baseline | 74 ± 2 | 109 ± 3 | |
240–300 | 63 ± 2 | 112 ± 4 | |
Change from Baseline | −11 ± 3 | +3 ± 2 | NS |
Study 2 | |||
Baseline | 108 ± 4 | 82 ± 3 | |
240–300 min | 88 ± 3 | 70 ± 2 | |
Change from baseline | −20 ± 3 | −12 ± 3 | NS |
NE specific act (dpm/ng) | |||
Study 1 | |||
Baseline | 310 ± 81 | 490 ± 91 | |
240–300 min | 210 ± 42 | 541 ± 81 | |
Change from baseline | −100 ± 28 | +50 ± 22 | 0.03 |
Study 2 | |||
Baseline | 376 ± 78 | 352 ± 62 | |
240–300 min | 352 ± 66 | 631 ± 87 | |
Change from baseline | −24 ± 18 | +279 ± 98 | 0.02 |
NE turnover (ng/min) | |||
Study 1 | |||
Baseline | 440 ± 82 | 346 ± 59 | |
240–300 min | 735 ± 90 | 280 ± 44 | |
Change from baseline | +295 ± 72 | −66 ± 85 | 0.04 |
Study 2 | |||
Baseline | 437 ± 96 | 481 ± 112 | |
240–300 min | 456 ± 108 | 298 ± 58 | |
Change from baseline | +18 ± 21 | −183 ± 22 | 0.05 |
NE clearance (L/min) | |||
Study 1 | |||
Baseline | 1.81 ± 0.21 | 2.25 ± 0.22 | |
240–300 min | 2.14 ± 0.24 | 2.48 ± 0.26 | |
Change from baseline | +0.33 ± 0.12 | +0.23 ± 0.11 | NS |
Study 2 | |||
Baseline | 1.84 ± 0.24 | 2.73 ± 0.34 | |
240–300 min | 2.09 ± 0.31 | 2.92 ± 0.36 | |
Change from baseline | +0.25 ± 0.14 | 0.19 ± 0.10 | NS |
Data are mean ± SEM.
Effect of Empagliflozin Under Pancreatic Clamp Conditions
When empagliflozin was ingested while maintaining plasma insulin (Fig. 2C) and glucagon (Fig. 2D) at basal levels, the plasma glucose concentration decreased by 37 mg/dL compared with a 19 ± 3 mg/dL decline with placebo (Fig. 2A). Thus, the placebo-subtracted decrease in plasma glucose concentration produced by empagliflozin under pancreatic clamp conditions was −18 mg/dL (P = 0.008) and was comparable to that caused by empagliflozin in study 1 (−23 mg/dL; P > 0.60 in study 2 versus study 1). Despite the constancy of plasma insulin and glucagon concentrations, empagliflozin produced a large increase (+1.14 ± 0.32 mg/kg · min, <0.001) in EGP (Fig. 2B). In subjects receiving placebo, EGP decreased slightly (−0.35 mg/kg ⋅ min, P = NS). The placebo-subtracted change in EGP with empagliflozin (1.49 mg/kg · min) was comparable to that in study 1 (1.55 ± 0.22 mg/kg ⋅ min). Neither plasma FFA (Fig. 1E) nor plasma β-HB (Fig. 2F) concentrations were affected by empagliflozin.
Pancreatic clamp conditions. Change from baseline in plasma glucose concentration (A) after empagliflozin and placebo (time 0) administration. The basal rate of EGP (B). Plasma insulin (C), glucagon (D), FFA (E), and β-hydroxybutyrate (F) concentrations during EGP measurement. **P < 0.01.
Pancreatic clamp conditions. Change from baseline in plasma glucose concentration (A) after empagliflozin and placebo (time 0) administration. The basal rate of EGP (B). Plasma insulin (C), glucagon (D), FFA (E), and β-hydroxybutyrate (F) concentrations during EGP measurement. **P < 0.01.
During pancreatic clamp conditions, the plasma NE concentration decreased slightly but not significantly in subjects receiving placebo and empagliflozin, with no significant difference between the two groups (Table 2). NE turnover decreased in subjects receiving placebo (−183 ± 22 ng/min), while it increased slightly in subjects receiving empagliflozin (+18 ± 21 ng/min). The difference between the two groups in the change from baseline to the end of study in NE turnover was statistically significant (P < 0.05) (Table 1).
Conclusions
SGLT2 inhibitors exert multiple metabolic actions in patients with T2DM, including increases in EGP, plasma FFA concentration, and ketone production (2). The major finding of the current study is that the mechanisms responsible for the increase in EGP are distinct from those responsible for the increase in plasma FFA and ketone concentrations. Under pancreatic clamp conditions, in which plasma insulin and glucagon concentrations remained unchanged from basal levels, the increases in plasma FFA and ketone concentrations caused by empagliflozin were completely blocked, while the increase in EGP persisted unabated.
The increases in plasma FFA concentration and ketone production have important clinical significance, since they can lead to the development of euglycemic ketoacidosis, a rare (0.2–0.6 cases per 1,000 person treatment years) but serious complication in patients receiving SGLT2 therapy (8,9). Since preventing the change in plasma insulin and glucagon concentrations during the pancreatic clamp prevented the increase in plasma FFA and ketone concentrations, it is likely that the decrease in plasma insulin, an inhibitor of lipolysis, and/or the increase in plasma glucagon are responsible for the increase in plasma FFA concentration and subsequent increase in ketone production. Thus, agents which increase the plasma insulin concentration and/or suppress the plasma glucagon concentration, for example, GLP-1 RAs and DPP4 inhibitors (13), are likely to prevent the increase in plasma ketone concentration and, when used in combination with SGLT2i, decrease the risk of ketoacidosis. Alternatively, agents which suppress lipolysis and cause a decrease in plasma FFA concentration, for example, thiazolidinediones, have been shown to prevent the increase in plasma ketone concentration and development of ketoacidosis (12).
Unlike the increase in plasma FFA and ketones concentrations, the increase in EGP caused by empagliflozin persisted under pancreatic clamp conditions, indicating that mechanisms other than the change in plasma insulin and glucagon are responsible for the stimulation of EGP. Empagliflozin administration was associated with an increase in total body NE turnover, suggesting increased sympathetic nervous system (SNS) activation by SGLT2i. Hepatic activation of the SNS has been shown to stimulate EGP (16). This observation provides a plausible explanation that links SGLT2i to the increase in EGP and is consistent with results from experimental animals in which the increase in EGP caused by dapagliflozin was associated with increased plasma NE concentration (17). The mechanism by which empagliflozin increases SNS activity remains to be determined. Previous studies (18) have reported that a decrease in the plasma glucose concentration, even within the normal range (e.g., from 200 to 100 mg/dL), caused sympathetic activation in patients with T2DM (18). Thus, the SGLT2i-induced decrease in plasma glucose concentration caused by glucosuria could play a role in the increased SNS activity with resultant stimulation of EGP.
We previously have shown that renal transplantation in patients in whom both native kidneys were removed attenuates the increase in EGP produced by SGLT2i (11). This suggests that the renal nerves play a central role in mediating the increase in EGP caused by SGLT2i. Thus, increased renal SNS activity, triggered by glucosuria, is another possible mechanism via which SGLT2i could increase sympathetic activity. Lastly, the diuretic action of SGLT2i (19) results in reduction in the extracellular volume, and volume contraction is a strong SNS activator. Indeed, urine secretion in subjects receiving empagliflozin was greater than that in subjects receiving placebo by approximately 600 and 450 mL in study 1 and study 2, respectively, and previous studies have demonstrated that acute diuretic administration in experimental animals resulted in increased renal nerve activity (20,21). A two-hit hypothesis for the increase in ketone production and EGP rate was recently suggested by Shulman and colleagues (17), where both volume contraction and decrease in plasma insulin concentration are required to activate lipolysis and increase ketone production and EGP. The results of the current study demonstrate that preventing the change in plasma insulin and glucagon concentrations (pancreatic clamp study) is sufficient to prevent the increase in plasma FFA concentration and ketone formation. However, it failed to block the increase in EGP. It will be of great interest to examine whether prevention of SNS activation and/or volume contraction can block the increase in EGP under pancreatic clamp conditions.
The increase in total body NE turnover rate was not associated with an increase in plasma NE concentration. This can be explained by the fact that >90% of NE released is cleared by the nerve terminals within the tissue. It also is possible that the SNS activation is regional, for example, hepatic, not global. This can explain the lack of tachycardia associated with SGLT2i administration despite the decrease in blood pressure (2). Alternatively, SGLT2i also could activate the parasympathetic nervous system and prevent the increase in heart rate despite sympathetic activation. The organ(s) in which sympathetic activation by SGLT2i occurs remains to be determined. Measurement of regional NE turnover across different organs, for example, kidney and liver, will provide additional information about the mechanism of sympathetic activation by SGLT2i.
The current study has some limitations. Previous studies in experimental animals have reported a significant impact of somatostatin on sympathetic activity (22–24). This could explain the differences between study 1 and study 2 in subjects receiving placebo. Nonetheless, the difference in NE turnover between subjects receiving empagliflozin versus placebo was comparable in both study 1 and study 2.
Although the current study is a mechanistic study, it has a relatively small sample size. However, because of the robust impact that empagliflozin had on EGP (31% increase) compared with our previous study (3), an interim analysis demonstrated that 12 patients (2:1 randomization) provided sufficient power to detect a significant effect of empagliflozin on EGP.
In summary, the results of the current study demonstrate that change in plasma insulin and glucagon concentrations caused by SGLT2 inhibition with empagliflozin and subsequent increase in plasma FFA concentration are responsible for the increase in plasma ketone concentration, while increased sympathetic activity can explain, at least in part, the stimulation of EGP.
S.A. and A.K. contributed equally to this study.
Article Information
Acknowledgments. The authors would like to thank James King, RN, and Hanh Tran, LVN for their excellent care of patients through the study and Lorrie Albarado and Deena Murphy for their expert secretarial assistance in preparation of the manuscript.
Funding. This study was funded by National Institutes of Health grant R01 5R01DK097554-10 to M.A.-G. Empagliflozin and placebo were provided by Boehringer Ingelheim.
Duality of Interest. R.A.D. receives grant support from Astra Zeneca and Merck, is a member of the advisory boards of Astra Zeneca, Intarcia, Boehringer Ingelheim, and Novo Nordisk, and is a member of the speakers’ bureaus of Astra Zeneca. S.D.P. has received research funding from AstraZeneca, Boehringer Ingelheim, Novartis Pharmaceuticals Co., and Merck Sharpe & Dohme, and is a consultant for or has received honoraria from AstraZeneca, Boehringer Ingelheim, Eli Lilly and Company, GlaxoSmithKline, Janssen Pharmaceuticals, Laboratoires Servier, Merck Sharp & Dohme, Novartis Pharmaceuticals Co., Novo Nordisk, Sanofi, and Takeda. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. M.A.-G. designed the study, drafted the study protocol, analyzed the data, and wrote the manuscript. S.A., A.K., J.A., M.A.-F., G.D., and F.A.-M. contributed to data generation and collection. S.D.P. and R.A.D. reviewed and revised the manuscript. All authors read and approved the final version of the manuscript. M.A.-G. 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.