Empagliflozin and Renal Sodium-Hydrogen Exchange in Healthy Subjects

Abstract

Context

Sodium glucose co-transporter-2 inhibitors exert clinically relevant cardiorenal protection. Among several mechanisms, inhibition of sodium-hydrogen exchanger-3 (NHE3) in proximal renal tubules has been proposed in rodents. Demonstration of this mechanism with the associated electrolyte and metabolic changes in humans is lacking.

Objective

The present proof-of-concept study was designed to explore the involvement of NHE3 in modulating the response to sodium glucose co-transporter-2 inhibitors in humans.

Methods

Twenty healthy male volunteers received 2 tablets of empagliflozin 25 mg during a standardized hydration scheme; freshly voided urines and blood samples were collected at timed intervals for 8 hours. Protein expression of relevant transporters was examined in exfoliated tubular cells.

Results

Urine pH levels increased after empagliflozin (from 5.81 ± 0.5 to 6.16 ± 0.6 at 6 hours, P = .008) as did urinary output (from median, 1.7; interquartile range [IQR, 0.6; 2.5] to 2.5 [IQR, 1.7; 3.5] mL/min⁻¹, P = .008) and glucose (from median, 0.03 [IQR, 0.02; 0.04] to 34.8 [IQR, 31.6; 40.2] %, P < .0001), and sodium fractional excretion rates (from median, 0.48 [IQR, 0.34; 0.65] to 0.71 [IQR, 0.55; 0.85] %, P = .0001), whereas plasma glucose and insulin concentrations decreased and plasma and urinary ketones increased. Nonsignificant changes in NHE3, phosphorylated NHE3, and membrane-associated protein 17 protein expression were detected in urinary exfoliated tubular cells. In a time-control study in 6 participants, neither urine pH nor plasma and urinary parameters changed.

Conclusions

In healthy young volunteers, empagliflozin acutely increases urinary pH while inducing a substrate shift toward lipid utilization and ketogenesis, without significant changes in renal NHE3 protein expression.

Sodium-glucose cotransporter-2 inhibitors (SGLT2i) exert cardioprotective effects. Among the multiple mechanisms proposed to explain these effects, an interaction with the sodium-hydrogen exchanger (NHE) can be postulated (1, 2). More than 9 isoforms of NHE exist. NHE1 is the predominant isoform in the human heart (3) and has been implicated in the pathophysiology of heart failure (4). SGLT2i could directly decrease NHE1 activity and restore physiological calcium handling (5, 6). This hypothesis appears to be of great interest because functional SGLT2 is expressed almost exclusively in the kidney (7). NHE3, on the other hand, is expressed only in gastrointestinal and renal tubular cells (8). In the kidney, NHE3 colocalizes with SGLT2 in the S1 segment of the proximal tubule, where it affects reabsorption of sodium and bicarbonate (9). NHE3 and SGLT2 show functional interaction. Under normal conditions, 25% of filtered sodium is reabsorbed in the proximal tubule coupled to bicarbonate; this percentage can increase to 80% when glucose transport is saturated (10). High glucose levels in the proximal tubule have been shown to inhibit NHE3-dependent bicarbonate reabsorption (11). Furthermore, SGLT2i can directly inhibit NHE3 activity by binding to the enzyme and promoting phosphorylation of 2 specific serine sites, S552 and S605, resulting in natriuresis that is not related to the magnitude of glucose excretion (9). Evidence from preclinical studies have shown that, in murine models of diabetes, chronic use of a SGLT2i did not affect total renal expression of NHE3 but significantly increased phosphorylation at S605 (12). This effect is absent in wild-type mice, supporting the existence of a functional interaction between SGLT2 and NHE3 (9, 12).

The role of renal NHE3 in modulating sodium exchange is more controversial. In NHE3 knock-out mice, empagliflozin increased urinary pH and excretion of sodium, chloride, and bicarbonate compared with wild-type mice, despite similar glycosuria (9). In healthy subjects, empagliflozin has been shown to increase the fractional excretion of sodium and endogenous lithium 90 and 180 minutes after an oral administration (13). However, in the DAPASALT trial (14), no significant difference emerged in 24-hour sodium excretion in patients with type 2 diabetes treated with dapagliflozin for 14 days under controlled sodium intake. In this study, during the first days of treatment, patients showed an increase in aldosterone excretion, possibly reflecting hyperactivity of the renin-angiotensin-aldosterone system, leading to sodium reabsorption through distal epithelial channels (14).

This appears to conflict with previous observation that angiotensinogen-to-creatinine ratio, a marker of intrarenal renin-angiotensin-aldosterone system activation, declines during treatment with SGLT2i (15, 16). Moreover, using single and multiple doses of empagliflozin 25 mg in individuals with type 2 diabetes, Heise et al (17) have reported a transient increase in 24-hour natriuresis without any significant change in plasma renin activity and serum aldosterone level.

Few studies have analyzed the in vivo effect of SGLT2i on electrolyte and acid-base balance. In a post hoc analysis of the RED trial (Renoprotective Effects of Dapagliflozin in Type 2 Diabetes), dapagliflozin was associated with slight increases in the plasma concentrations of magnesium, chloride, phosphate, and sulphate without affecting either absolute or fractional electrolyte excretions or urinary pH (18).

Thus, main aim of our study was to explore the acute effect of empagliflozin on urinary pH and the associated metabolic changes in healthy subjects. Exfoliated tubular cells were collected from the urine in the attempt to assess whether changes in pH might be associated with changes in NHE3 protein expression and the membrane-associated protein 17 (MAP17), an essential auxiliary subunit of SGLT2 (19).

Methods

Subjects

Twenty healthy male subjects, aged 18 to 35 years, were recruited on a volunteer basis (Supplementary Table S1) (20). All of them were normotensive, had normal glucose tolerance, preserved kidney function, and no history of medications or substance abuse.

Design

The day before the study, participants were asked to avoid eating vegetables and fruit and to practice physical exercise, and to stop eating after 8:00 Pm. The metabolic study started at 8:00 Am with participants in the overnight fasted state (10-12 hours). After voiding their bladder, subjects received a standardized hydration scheme: at start, water (300 mL) was ingested and urine was collected for the following 2 hours (time 0); then empagliflozin 25 mg was given with 300 mL of water; after 4 hours, a new urine collection was performed (time 240 minutes), followed by a second administration of empagliflozin 25 mg with an additional 300 mL of water; after another 2 hours (time 360 minutes) urine was again collected (Fig. 1A). Urine volume and pH level were measured immediately after voiding at each collection time. No food was allowed during the whole study period. During the protocol, blood samples were drawn at timed intervals (time −120, time 120, and time 360) for the measurement of serum creatinine and plasma glucose and electrolytes.

Figure 1.

Flowchart of the study protocol (A), urine output (B), and pH (C) at baseline and after empagliflozin administration; fractional glucose (D) and sodium (E) excretion; ß-hydroxybutyrate (F) and acetoacetate (G) excretion at baseline and after empagliflozin administration. Plots are boxplots.

Six of the 20 volunteers, randomly selected in a ratio of 1:3, repeated the study in the fasting state with the same hydration scheme but without assuming empagliflozin.

The study protocol was approved by the Tuscany Area Vasta Nord Ovest Ethics Committee. All participants provided written informed consent before enrollment.

Blood and Urine Analysis and Renal Cell Isolation

Blood samples were centrifuged (3000 rpm at 4 °C for 10 minutes) and serum and plasma aliquots were stored at −80 °C. At each collection time, immediately after voiding urine volume was measured, and urine pH was determined 3 times over a 15-second time interval (Supplementary Table S2) (20). Two urine aliquots of 5 mL each were frozen and stored at −80 °C for later analysis, whereas the residual volume was centrifuged at 4500 rpm and 12 °C for 10 minutes. Given the large volume, the urine samples were spun sequentially in the same tube, discarding the supernatant after each centrifugation step and storing the cellular pellet.

Measurements

All measurements were performed at the Metabolism Laboratory of the University of Pisa; samples from each subject were assayed together to reduce intrasubject variability. Plasma and urine glucose, β-hydroxybutyrate (β-HB), acetoacetate (AcAc), sodium, potassium, and chloride concentrations were measured on a Synchron systemCX4 (Beckman Coulter, Fullerton, CA, USA). Serum and urine creatinine were measured by standard clinical chemistry methods. Fasting plasma insulin was measured by Chemiluminescence Access (Beckman Coulter, Brea, CA, USA).

Protein Expression

Total proteins were extracted from urine cellular pellet with NP40 Lysis Buffer (FNN0021, Invitrogen-ThermoFisher MA-USA) added with Protease inhibitor cocktail (P2714, Sigma-Merck, Darmstadt-Germany). For western blot analysis, the following primary antibodies were used: NHE-3 RRID: AB_2928052; p-NHE-3 RRID: AB_2928053; MAP-17 RRID: AB_2643736; Kidney Injury Molecule-1 (KIM-1) RRID: AB_2746461. In brief, the same volume of protein extract was diluted in SDS-PAGE buffer, heated at 100 °C for 5 minutes, and separated on Any kD Mini-Protean TGX gels (Bio-Rad Laboratories, Italy). Samples were then transferred to a polyvinylidene difluoride membrane (BioRad), treated with a blocking solution (3% milk in TTBS) and incubated overnight with the primary antibodies. After a final step with specific secondary horseradish peroxidase-conjugated antibodies, bands were identified by an enzymatic chemiluminescence reaction (Clarity Western ECL, BioRad). Protein signals were acquired on ChemiDoc Instrument (BioRad), and band intensity was evaluated using ImageJ v1.49 software; each sample value was normalized by the intensity of KIM-1, a specific marker of human proximal tubular cells (hPTCs). Data were expressed as units of OD.

RNA Extraction and Quantitative RT-PCR

In a subgroup of 5 subjects, RNA was extracted from the cellular pellet, containing putative hPTCs. The procedure was carried out on the QIAcube (Qiagen, Hilden, Germany), a robotic workstation for automated nucleic acids purification, loaded with RNeasy Micro kit (cat. 74004, Qiagen). Because of the small number of cells collected, and consequently to the low quantity of isolated RNA, the spectrophotometric quantification of nucleic acids by NanoDrop 2000c (Thermo Fisher Scientific) was not reliable; therefore, the same volume of total RNA (10 µL) was used for the RT, determined with High-capacity cDNA Reverse Transcription kit (Applied Biosystems, Thermo Fisher Scientific). RT-PCR was performed in triplicate on an Eco real time instrument (Illumina Inc., San Diego, CA, USA) according to the standard procedure for TaqMan Gene Expression Assays (Thermo-Fisher); transcripts were quantified using the following assays: KIM-1/HAVCR1: Hs00930379_g1; APN: Hs00174265_m1; SGLT2: Hs00894642_m1; NHE3: Hs00903842_m1; MAP17/PDZK1/P1: Hs00173779_m1; GAPDH: Hs02758991_g1. Their values were expressed by means of ΔΔC_t calculation, where Ct is the threshold cycle; the amount of the target gene, normalized respect to the mean of 2 hPTCs specific markers (KIM-1 and APN), is given as 2^−ΔΔCt.

Calculations

Urinary solute excretion rate was calculated as the product of urine solute concentration and urine volume; renal solute clearance rate was calculated as the ratio of urine solute excretion to plasma solute concentration. Solute filtered rate was calculated as the product of creatinine clearance and plasma solute concentration. Fractional solute excretion was obtained as the ratio of urinary solute excretion to solute-filtered load.

Statistical Analysis

Data are shown as mean ± SD (or median [interquartile range] for not-normally distributed variables). To determine data distribution, we used the Shapiro-Wilk test. Paired comparison analysis was conducted using ANOVA for repeated measures or Wilcoxon rank test depending on data distribution. A P value ≤.05 was considered significant. A sample size of 20 subjects was calculated to provide at least 80% power to detect a mean difference in urinary pH induced by SGLT2 inhibition of 0.3, with an SD of difference of 0.45, at a 2-sided α level of 0.05 (effect size, 0.667).

Results

Circulating Parameters

Plasma glucose and insulin concentrations decreased throughout the 6 hours of the study, whereas plasma β-HB and AcAc rose, the former more quickly and to a greater extent than the latter (Table 1). Serum electrolytes showed only minor changes.

Table 1.

Effect of empagliflozin on biochemical parameters^a

	Time −120	Time 120	Time 360	P1	P2	P3
Plasma glucose, mmol/L	5.36 [5.22; 5.94]	5.00 [4.89; 5.24]	4.56 [4.35; 4.93]	.0009	<.0001	<.0001
Plasma insulin, µU/mL	7.7 [4.8; 10.0]	4.2 [2.5; 6.1]	2.3 [1.4; 3.2]	<.0001	<.0001	<.0001
Plasma ß-HB, µmol/L	101 [72; 186]	190 [60; 330]	630 [283; 1167]	<.0001	<.0001	<.0001
Plasma AcAc, µmol/L	17 [12; 22]	21 [13; 30]	37 [14; 73]	NS	.0005	.0002
Plasma β-HB/AcAc	7.6 [4.2; 12.3]	9.4 [4.5; 10.4]	13.3 [9.5; 26.1]	NS	<.0001	<.0001
Serum Na+, mEq/L	137 ± 3	135 ± 3	135 ± 3	.0416	NS	NS
Serum K+, mEq/L	4.0 ± 0.3	4.2 ± 0.4	3.9 ± 0.3	.0301	NS	.0009
Serum Cl−, mEq/L	104 ± 3	101 ± 3	101 ± 3	.0051	.0004	NS

Abbreviations: AcAc, acetoacetate; ß-HB, β-hydroxybutyrate; β-HB/AcAc, β-hydroxybutyrate/acetoacetate ratio; NS, not significant.

^aEntries are mean ± SD or median [25% CI; 75% CI]; P¹ = t 0 vs t 240; P² = t 0 vs t 360; P³ = t 240 vs t 360 by Wilcoxon signed-rank test.

Urine Parameters

Urinary output rose ∼40%, from 1.7 [0.6; 2.5] to 2.5 [1.7; 3.5] mL/min⁻¹ (P = .008) (Fig. 1B), whereas creatinine clearance initially tended to decrease (by ∼25%), then returned to baseline.

Urine pH increased significantly after both the first and the second dose of empagliflozin compared with baseline (from 5.81 ± 0.5 to 6.11 ± 0.6 at t 240 and 6.16 ± 0.6 at t 360; P = .0087 and P = .0079, respectively) (Fig. 1C). In contrast, in the control study (n = 6), neither urine pH nor urinary output changed significantly (Supplementary Fig. S1) (20). In these 6 subjects, the mean change in urine pH at 360 minutes was 6.4% after empagliflozin and 0.2% after the water load alone (P = .28; Supplementary Fig. S2) (20).

Urinary glucose excretion increased markedly in absolute amount, rate, and fractional excretion (from 0.03 [0.02; 0.04] to 34.8 [31.6; 40.2] %, P < .0001) (Fig. 1D). Sodium excretion rate changed minimally over the 6 hours of empagliflozin stimulation; as a fraction of the filtered load, however, sodium excretion rose ∼50%, from 0.48 [0.34; 0.65] to 0.71 [0.55; 0.85] % (P = .0001) (Fig. 1E), as did also potassium and chloride excretion (Table 2). Empagliflozin increased both β-HB and AcAc excretion (Fig. 1F and 1G) and their concentration ratio in the urine (Table 2).

Table 2.

Effect of empagliflozin on renal glucose and solute handling^a

	t 0	t 240	t 360	P1	P2	P3
Urine output, mL/min−1	1.7 [0.6; 2.5]	2.1 [1.9; 2.7]	2.5 [1.7; 3.5]	.0081	.0080	NS
Creatinine clearance, mL/min−1	121 [85; 168]	103 [88; 131]	122 [99; 154]	(.0637)	NS	.0362
Glucose excretion, mg	4.3 [3.0; 7.0]	6077 [5603; 7714]	4321 [3425; 5216]	<.0001	<.0001	<.0001
Glucose excretion rate, mg/min−1	0.03 [0.03; 0.05]	26.5 [24.5; 32.1]	36.0 [28.5; 41.9]	<.0001	<.0001	.0006
Fractional glucose excretion, %	0.03 [0.02; 0.04]	30.0 [27.0; 33.0]	34.8 [31.6; 40.2]	<.0001	<.0001	<.0001
Na+ excretion rate, mEq/min−1	0.07 [0.04; 0.10]	0.10 [0.08; 0.14]	0.10 [0.09; 0.13]	NS	NS	NS
Na+ fractional excretion, %	0.48 [0.34; 0.65]	0.76 [0.67; 0.91]	0.71 [0.55; 0.85]	<.0001	.0001	.0192
K+ excretion rate, mEq/min−1	0.05 [0.03; 0.08]	0.06 [0.04; 0.09]	0.05 [0.04; 0.08]	NS	NS	NS
K+ fractional excretion, %	9.5 [8.2; 14.0]	14.2 [11.7; 17.5]	11.9 [9.1; 15.6]	.0004	NS	.0021
Cl− excretion rate, mEq/min−1	0.09 [0.07; 0.15]	0.14 [0.11; 0.19]	0.14 [0.10; 0.18]	NS	NS	NS
Cl− fractional excretion, %	0.82 [0.63; 1.13]	1.40 [1.28; 1.54]	1.17 [0.93; 1.47]	<.0001	.0046	.0001
β-HB excretion rate, μmol/min−1	0.01 [0.00; 0.018]	0.04 [0.02;0.05]	0.07 [0.04; 0.14]	.0010	<.0001	.0037
β-HB fractional excretion, %	0.05 [0.0; 0.17]	0.24 [0.07; 0.37]	0.11 [0.08; 0.16]	.0062	NS	.0172
AcAc excretion rate, μmol/min−1	0.02 [0.01; 0.03]	0.03 [0.02; 0.03]	0.05 [0.02; 0.09]	.0204	<.0001	.0073
AcAc fractional excretion, %	0.73 [0.48; 1.48]	1.35 [0.70; 2.41]	1.00 [0.55; 2.11]	.0204	NS	NS
Urine ß-HB/AcAc	0.5 [0.00; 1.4]	1.3 [0.9; 2.2]	1.6 [1.2; 1.9]	.0046	.0047	NS

Abbreviations: AcAc, acetoacetate; ß-HB, β-hydroxybutyrate; β-HB/AcAc, β-hydroxybutyrate/acetoacetate ratio; NS, not significant.

Entries are median [25% CI; 75% CI]; P¹ = t 0 vs t 240; P² = t 0 vs t 360; P³ = t 240 vs t 360 by Wilcoxon signed-rank test.

Protein and mRNA Expression

Western blot analysis of NHE3 and MAP17 showed wide interindividual variability and no significant changes in the subgroup of 13 subjects in whom sufficient material could be recovered (Table 3). Representative blots of increase and decrease are shown in Fig. 2. In a subgroup of 5 subjects, we measured hPTC mRNA expression of SGLT2, NHE3, and MAP17, and no significant differences were found for any of the 3 molecules (Table 3).

Table 3.

Effect of empagliflozin on protein and mRNA expression^a

	t 0	t 240	t 360	P1	P2	P3
Protein (n = 13)
NHE3/KIM1	0.57 [0.12; 4.0]	0.74 [0.21; 1.87]	0.38 [0.16; 2.41]	NS	NS	NS
pNHE3/NHE3	0.53 [0.12; 4.25]	0.50 [0.10; 4.7]	1.76 [0.55; 3.07]	NS	NS	NS
MAP17/KIM1	4.36 [1.41; 15.6]	6.0 [0.40; 12.0]	9.0 [0.14; 14.51]	NS	NS	NS
mRNA (n = 5)
SGLT2/KIM1	2.3 ± 2.5	1.7 ± 1.4	1.1 ± 0.7	NS	NS	NS
NHE3/KIM1	2.2 ± 1.3	4.0 ± 2.5	4.8 ± 1.2	NS	NS	NS
MAP17/KIM1	1.7 ± 1.1	2.5 ± 1.2	3.3 ± 3.0	NS	NS	NS

Abbreviations: KIM1, Kidney Injury Molecule-1; MAP17, Membrane-Associated Protein 17; NHE3, sodium-hydrogen exchanger 3.

Data are reported as target/reference; P¹ = t 0 vs t 240; P² = t 240 vs t 360; P³ = t 240 vs t 360.

Figure 2.

The graph on the left shows the extreme variability of normalized sodium-hydrogen exchanger 3 (NHE3) expression after empagliflozin administration in 13 study participants. Western blots of NHE3, KIM1, and MAP17 performed in exfoliated tubular cells of 7 subjects representative of the whole cohort are also shown.

Discussion

The main results of this study can be summarized as follows: (1) empagliflozin acutely increased urinary pH in healthy young subjects; (2) this effect was coupled with a rapid and sustained increase in urinary excretion of glucose and ketone bodies, and increased fractional excretion of sodium, potassium, and chloride; and (3) these changes were accompanied by variable responses of NHE3 expression in exfoliated tubular cells.

The premise of this study was that, in the kidney, inhibition of the SGLT2 cotransporter would jointly inhibit NHE3, thereby reducing hydrogen excretion in exchange for sodium. In experimental models, SGLT2i suppress renal NHE3 expression. Moreover, Onishi et al demonstrated that, in NHE3 knockout mice treated with empagliflozin, urinary pH increased and correlated with higher excretion of sodium, chloride, and bicarbonate compared with wild-type mice, despite a similar level of glycosuria (9). The current results are proof of this concept in humans.

The kidney response to strong inhibition of SGLT2 with empagliflozin was rapid and multiple, as along with pH there occurred almost simultaneous increases in urine output, glucose, sodium, and ketone excretion. These responses were interrelated (Supplementary Fig. S3) (20) but they can be tentatively lined up in a causal order. As glycosuria ensued, urine output increased by osmotic diuresis. Urinary sodium, chloride, and potassium excretion also increased, at least as fractions of the respective filtered load; in fact, creatinine clearance—and, presumably, glomerular filtration rate—initially dropped. Thus, there was a component of natriuresis to the observed diuresis. To what extent SGLT2 and NHE3 inhibition contributed to the overall restriction of sodium reabsorption is impossible to assess from the current data. With the emptying out of the glucose pool via urinary loss, plasma glucose concentrations fell, causing a ∼70% decline in plasma insulin concentrations. The latter effect was followed by a de-repression of lipolysis and a quick switch in whole-body substrate utilization from carbohydrate to fat (21), which led to increased ketogenesis (in the liver and, possibly, in the kidney). Increased ketone excretion thus became evident during the last 2 hours of the study. It is interesting that the ß-HB/AcAc ratio increased significantly in the plasma and, consequently, in the urine. Insofar as this ratio reflects the mitochondrial redox balance (21), the substrate switch was likely accompanied by an excess of reducing equivalents, typical of enhanced fat oxidation (22, 23). Furthermore, to the extent that insulin upregulates SGLT (24) as well as NHE3 expression (25, 26), the reduction in circulating insulin may have strengthened the inhibition of these cotransporters by empagliflozin.

To our knowledge, this is the third time that urine pH has been reported in humans treated with an SGLT2i. van Bommel et al (18) analyzed the acid-base balance in patients with type 2 diabetes chronically treated with dapagliflozin, and they found no significant change in urinary pH compared with treatment with sulfonylurea. More recently, a reduced urinary pH, responsible for changes in urine supersaturation, has been reported in healthy volunteers treated with empagliflozin for 4 weeks (27). Conflicting findings may be due to the different study settings (acute vs chronic) because the effects of empagliflozin on urine pH might be attenuated by chronic treatment. Of note, in previous studies, urine pH was determined in stored samples. pH is known to gradually increase after urine collection, and it can be altered by the storage process (28). We therefore made a point of testing urine immediately after voiding and in triplicate, to ensure a determination as accurate as possible.

It must be remarked that the observed increase in urine output with empagliflozin might have diluted the excreted hydrogen, thereby explaining the increase in urine pH. In fact, urine output did not increase significantly in the control subjects, who drank the same amount of water (Supplementary Fig. S1). We cannot rule out that empagliflozin-induced diuresis may be an additional regulatory step of urinary pH. However, the increased urine pH in our data occurred in the face of increased ketoacid excretion, which per se would lower urine pH.

A novelty of this report is to test, for the first time in humans, whether SGLT2 inhibition changes renal NHE3 expression. A study carried out in a murine model reported that chronic use of empagliflozin did not alter total renal NHE3 expression, while enhancing SGLT2 expression in both diabetic and control mice (9). Our Western blot analysis of exfoliated tubular cells did show a modulation of either NHE3 or MAP17 protein expression; however, results show a large variability. This could be due to the short duration of the study and to the scarcity of cellular material in the urine of healthy young volunteers.

A limitation of our study is that it was carried out in healthy young men only. Results may have been partly different in women or aged individuals or in the fed state. Certainly, it would be worth repeating the protocol in subjects with type 2 diabetes, who already have a metabolic setup shifted toward fat oxidation (23), overactive SGLT2 (29), and presumably overexpressed NHE3. In addition, chronic empagliflozin treatment might bring forth time-related differences. This proof-of-concept study was unblinded, not randomized, and uncontrolled. Furthermore, a sample size of 13 provided 80% power to detect large differences in NHE3 protein expression after SGLT2 inhibition (effect size, 0.847); this study was underpowered to detect smaller changes. With these caveats, our study does provide support for the notion that NHE3 is implicated in solute exchange as well as in the renal mechanism of action of SGLT2 inhibitors.

Abbreviations

AcAc

acetoacetate

β-HB

β-hydroxybutyrate

hPTC

human proximal tubular cell

KIM-1

Kidney Injury Molecule-1

MAP17

membrane-associated protein 17

NHE

sodium-hydrogen exchanger

SGLT2

sodium-glucose cotransporter-2

Funding

This study has been supported by an institutional grant from MIUR (A.S.).

Disclosures

E.F. has served on the Advisory Board of Boehringer Ingelheim/Lilly&Co., Lexicon, Oramed, and Servier; has received research grants from Boehringer Ingelheim/Lilly&Co and Janssen; and speaker fees from Boehringer Ingelheim/Lilly&Co. and MSD. A.S. has served on the Advisory Board of Novo, Sankyo, and Sanofi and has received speaker fees from Astra Zeneca, Lilly, Novo, and Sanofi. The remaining authors have nothing to disclose.

Data Availability

Some datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.