Kidney-Protective Effects of SGLT2 Inhibitors

Abstract

The sodium-glucose cotransporter 2 (SGLT2) inhibitors have become an integral part of clinical practice guidelines to slow the progression of CKD in patients with and without diabetes mellitus. Although initially developed as antihyperglycemic drugs, their effect on the kidney is multifactorial resulting from profuse glycosuria and natriuresis consequent to their primary site of action. Hemodynamic and metabolic changes ensue that mediate kidney-protective effects, including (1) decreased workload of proximal tubular cells and prevention of aberrant increases in glycolysis, contributing to a decreased risk of AKI; (2) lowering of intraglomerular pressure by activating tubular glomerular feedback and reductions in BP and tissue sodium content; (3) initiation of nutrient-sensing pathways reminiscent of starvation activating ketogenesis, increased autophagy, and restoration of carbon flow through the mitochondria without production of reactive oxygen species; (4) body weight loss without a reduction in basal metabolic rate due to increases in nonshivering thermogenesis; and (5) favorable changes in quantity and characteristics of perirenal fat leading to decreased release of adipokines, which adversely affect the glomerular capillary and signal increased sympathetic outflow. Additionally, these drugs stimulate phosphate and magnesium reabsorption and increase uric acid excretion. Familiarity with kidney-specific mechanisms of action, potential changes in kidney function, and/or alterations in electrolytes and volume status, which are induced by these widely prescribed drugs, will facilitate usage in the patients for whom they are indicated.

Introduction

The development of drugs to inhibit the sodium-glucose cotransporter 2 (SGLT2) was designed to provide an additional means to control the plasma glucose in patients with type 2 diabetes mellitus. In the course of development, regulatory agencies required long-term cardiovascular outcome trial evidence to exclude an increased risk of major adverse cardiovascular events. Results of such trials demonstrated risk reductions for myocardial infarction, heart failure, kidney disease progression, and cardiovascular mortality. Strikingly, these benefits were demonstrated in patients receiving renin-angiotensin-aldosterone blockade at baseline. Sodium-glucose cotransporter 2 inhibitors (SGLT2is) have become a critical part of clinical practice guidelines for the management of patients with cardiovascular disease with and without diabetes. The clinical trials justifying the use of these drugs in patient care have been the subject of recent reviews, but the mechanisms by which this class of compounds provides kidney protection and how they affect various aspects of kidney homeostasis have not been fully reviewed (1).

Kidney Handling of Glucose

Under normal circumstances, approximately 180 g of glucose is filtered daily across the glomerular capillaries and is nearly completely reabsorbed by the proximal tubule, leaving the urine free of glucose. The SGLT2 is a high-capacity, low-affinity transporter responsible for the bulk of glucose reabsorption in the early part of the tubule. In the later portions of the tubule, the remaining amount of glucose is reabsorbed by the low-capacity, high-affinity SGLT1.

Glucose reabsorption is a saturable process exhibiting a tubular maximum of approximately 180 mg/dl. Plasma glucose concentrations above this level exceed the transport systems, and excess glucose begins to appear in the urine. Glycosuria in the absence of hyperglycemia can occur when the filtered load of glucose is increased due to a very high GFR, as in pregnancy or with compensatory hyperfiltration of a solitary kidney or when the tubular maximum is reduced as with proximal tubulopathies.

In patients with diabetes mellitus, the threshold for glucose excretion is increased from approximately 180 to 250 mg/dl due to glucose-mediated hypertrophy of the proximal tubule and increased expression of SGLT2 (2). These changes lead to a doubling in the percentage of filtered Na⁺ reabsorbed and a 50% increase in daily kidney ATP turnover (3,4). Increased Na⁺ reabsorption leads to volume expansion and accounts for a low renin form of hypertension in patients with type 2 diabetes mellitus.

SGLT2is reduce the tubular maximum and increase urinary excretion of glucose, causing a lowering of plasma glucose concentration. The increase in urinary glucose is <50% of the filtered load, even though SGLT2 is responsible for the reabsorption of approximately 80%–90% of the filtered glucose load. The difference between the amount filtered and excreted is due to increases in SGLT1-mediated transport in the more distal portions of the proximal tubule (5). Studies in SGLT2i knockout mice demonstrate that SGLT1 has the capacity to reabsorb approximately 30%–40% of filtered glucose, a value similar to estimates in humans (5). Simultaneous inhibition of SGLT2/SGLT1 produces a greater glycosuric effect in the kidney as compared with SGLT2i alone and can potentially limit postprandial hyperglycemia.

Sodium-Glucose Cotransporter 2 Inhibitor and Glucose Handling

In SGLT2i clinical trials, reductions in hemoglobin A_1c range from 0.6% to 1.0% versus placebo regardless of background therapy (6,7). The ability to lower blood glucose and glycated hemoglobin (HgA_1c) levels is proportional to the filtered load of glucose, which, in turn, is determined by the ambient glucose concentration and the GFR. Greater reductions are seen in patients with poor glycemic control because high plasma glucose concentrations along with hyperfiltration increase the filtered load of glucose to the site of action of the drug. As plasma glucose concentrations fall, the glycosuric effect decreases, contributing to a low risk of hypoglycemia with these drugs. In addition, the tubular maximum is reduced to a range of 76–90 mg/dl, which is above the plasma glucose concentration at which hypoglycemic symptoms typically occur (8,9). A decrease in GFR at the same plasma glucose concentration also lowers the filtered load of glucose and diminishes the glycosuric effect of these drugs. This attenuation in glucose lowering is evident at a GFR of <60 ml/min and is nearly complete at a GFR <30 ml/min (10).

Mechanisms of Kidney Protection

Improved Glycemic Control

SGLT2is diminish the complications of diabetes mellitus by lowering the plasma glucose concentration, which leads to improvement in both insulin sensitivity and secretion (11,12) (Table 1). These drugs do not act on skeletal muscle; therefore, favorable effects on insulin action presumably reflect reversal of glucotoxicity. Improvement in insulin sensitivity is associated with a shift from glucose to fatty acid oxidation in skeletal muscle and a reduction in intramyocellular fat content (13). Although reducing glucotoxicity contributes to improvement in microvascular complications of diabetes mellitus, the reduction in HgA_1c is modest in patients with CKD and therefore unlikely to be the sole explanation for cardiorenal protection. For example, the reduction in the risk of kidney failure is found irrespective of the baseline HgA_1c to include patients with a baseline value of <7% (14). Additionally, SGLT2is reduce heart and kidney failure in patients with and without diabetes mellitus.

Table 1.

Kidney-protective effects of sodium-glucose cotransporter 2 inhibitors

Protective Effect	Mechanism	Comment
Hemodynamic benefits	Increased tubuloglomerular feedback; increased delivery of NaCl to macula densa causes adenosine-mediated vasoconstriction of the afferent arteriole and lowers intraglomerular pressure; efferent vasodilation may predominate in type 2 diabetes mellitus in the setting of renin-angiotensin blockade	A 3- to 6-ml/min reduction in GFR is common in first 2–3 wks of therapy caused by reduction in intraglomerular pressure; increased pressure in Bowman’s space may also contribute to the early eGFR decline
Improved glycemic control	Reduction in TM leads to glycosuria, the glucose-lowering effect diminishes as plasma glucose level declines and/or GFR falls due to diminished filtered load	In clinical trials, HgA1c decreases by 0.6%–1.0% versus placebo; the glycosuric effect is no longer evident as eGFR approaches 30–40 ml/min
Decreased glucose flux across cell	Decreased proximal tubular cell glucose entry limits abnormally high rates of glycolysis, potentially limiting kidney fibrosis	Increased glycolysis linked to the activation of HIF1α and the suppression of Sirt3 and increased epithelial-mesenchymal transition, activation of NLRP3 inflammasome
Natriuresis	Inhibition of SGLT2 is accompanied by reduced NHE3 activity, plasma volume is reduced	Contributes to the 4/2–mm Hg reduction in BP, tissue-bound Na+ is also reduced, despite reduced BP and plasma volume heart rate is not increased consistent with decreased sympathetic outflow
Perirenal fat	Decreased paracrine release of adipokines and proinflammatory cytokines	Decreased leptin release decreases central afferent input and lowers sympathetic outflow, albuminuria and glomerular injury are decreased
Weight loss	Urinary glucose loss corresponding to 200–400 kcal/d, increased energy expenditure associated with beiging of adipocytes	Visceral and subcutaneous fat mass is reduced likely to include perirenal fat, increases in fibroblast growth factor 21 contributes to reduction in fat mass
Improved metabolic flexibility	Loss of glucose in urine leads to a fasting-like state with a decreased insulin-glucagon ratio and increased ketogenesis	Decreased respiratory exchange ratio reflects increased fat oxidation, mTORC1 is suppressed and autophagy is restored in tubular cells and podocytes
Decreased tubular workload	Decrease Na+ entry into proximal tubular cell reduces ATP and O2 consumption	SGLT2is are associated with less risk of AKI
Increase hemoglobin concentration (3%)	Increased O2 consumption to reabsorb Na+ in downstream segments result in ↑ erythropoietin production	Hypoxia stimulates HIF2α, causing ↑ autophagy and ↓ inflammation
Increase plasma magnesium (Mg2+)	↑ glucagon and PTH stimulate Mg2+ reabsorption in thick limb, upregulation of TRPM6/7 in DCT	SGLT2is may be useful in Mg2+ wasting disorders, decreased risk of arrhythmias and risk of diabetes mellitus
Decrease plasma uric acid	Increased tubular glucose competes with uric acid for reabsorption via GLUT9	Reduction in risk of gout flares, decreased risk of CKD due to hyperuricemia
Decrease risk of hyperkalemia from RAASi	Increased flow and Na+ delivery augment K+ secretion in the distal nephron	SGLT2is alone have minimal effects on plasma K+ concentration

TM, tubular maximum; HgA_1c, glycated hemoglobin; HIF1α, hypoxia-inducible factor 1α; Sirt3, sirtuin 3; NLRP3, NOD-, LRR-, and pyrin domain–containing protein 3; SGLT2, sodium-glucose cotransporter 2; NHE3, Na⁺-H⁺ antiporter 3; mTORC1, mammalian target of rapamycin complex 1; SGLT2i, sodium-glucose cotransporter 2 inhibitor; HIF2α, hypoxia-inducible factor 2α; PTH, parathyroid hormone; TRPM6/7, transient receptor potential melastatin 6/7; DCT, distal convoluted tubule; GLUT9, glucose transporter 9; RAASi, renin-angiotensin-aldosterone system inhibitor.

Reducing cellular flux of glucose may contribute to kidney protection. In addition to being transported across the basolateral membrane into the blood, increased glucose uptake by proximal tubular cells fuels increased rates of glycolysis. Increased glycolysis has been linked to the accumulation of hypoxia-inducible factor 1α (HIF1α) and suppression of sirtuin 3, changes linked to the epithelial-mesenchymal transition and kidney fibrosis (15,16). In mice with streptozotocin-induced diabetes, SGLT2is decrease kidney fibrosis in association with suppression of the proximal tubule glycolytic flux and restoration of sirtuin 3. This alteration in tubular cell metabolism also leads to downregulation of the NOD-, LRR-, and pyrin domain–containing protein 3 (NLRP3) inflammasome, contributing to an additional antifibrotic effect (17).

Hemodynamics

Glomerular hyperfiltration characterized by a 25%–50% increase in GFR is an early hemodynamic abnormality in patients with diabetes mellitus, and elevations in intraglomerular pressure and kidney blood flow due to dilation of the afferent arteriole contribute to these findings. Afferent vascular tone is regulated by changes in delivery of NaCl to the macula densa through the tubuloglomerular feedback mechanism. Glomerular hypertrophy and increased kidney size typically coexist with hyperfiltration. These changes can be reversed over several weeks as plasma glucose is normalized with intensive insulin therapy, suggesting that these early phenomena are influenced by the metabolic state (18). In this regard, increased glucose concentration at the macula densa also participates in regulating the tone of the afferent arteriole.

Inhibition of Na⁺-coupled glucose reabsorption in the proximal tubule promotes increased NaCl delivery to macula densa cells activating the tubuloglomerular feedback response, leading to increased tone of the afferent arteriole (Figure 1). This effect decreases intraglomerular pressure and reduces the degree of hyperfiltration. A study of patients with poorly controlled type 1 diabetes mellitus found that 8 weeks of therapy with empagliflozin decreased renal plasma flow and attenuated hyperfiltration in association with increased urinary adenosine, suggesting that activation of tubuloglomerular feedback was responsible for the hemodynamic changes (22–24).

Figure 1.

Mechanism of glucose transport in the proximal tubule and effects of sodium-glucose cotransporter 2 (SGLT2) inhibition on the tubuloglomerular feedback mechanism and transport of uric acid and phosphate. (A) Glucose reabsorption via SGLT2 and SGLT1 is an example of secondary active transport where ATP hydrolysis fuels the activity of the basolateral Na⁺-K⁺-ATPase in order to maintain a low intracellular Na⁺ concentration and provide a favorable inward gradient for Na⁺-coupled glucose transport across the apical membrane. The SGLT2 couples one Na⁺ ion for each glucose molecule reabsorbed as opposed to two Na⁺ ions for SGLT1. Glucose passively exits the cell into the bloodstream via glucose transporter 2 (GLUT2) in the S1/S2 segment and via GLUT1 in the S3 segment of the proximal convoluted tubule. The amount of glucose entering the blood exceeds what is reabsorbed due to gluconeogenesis in the proximal tubule. Sodium-glucose cotransporter 2 inhibitors (SGLT2is) increase glucose and Na⁺ concentration in the tubular lumen. Increased glucose concentration competes for urate reabsorption via GLUT9b, leading to uricosuria and a reduction in plasma urate concentration. The increased availability of Na⁺ drives activity of Na⁺-dependent phosphate transport protein 2a (NaPi-2a) and accounts for a small increase in the plasma phosphate concentration. (B) ATP released from macula densa cells in response to NaCl delivery is hydrolyzed by ectonucleotidases to form adenosine, which is responsible for the tubuloglomerular feedback response (19). In untreated patients with diabetes mellitus, tubular hypertrophy and upregulation of SGLT2 decrease NaCl delivery to the Na⁺-K⁺-2Cl⁻ (NK2CL) cotransporter located on the apical surface of macula densa cells. This decrease diminishes the liberation of ATP and subsequent formation of adenosine, resulting in vasodilation of the afferent arteriole. The development of hyperfiltration serves to restore distal NaCl delivery toward normal, but does so at the expense of increased intraglomerular pressure, which contributes to the development and progression of diabetic kidney disease. Increased glucose concentration sensed by SGLT1 located on the luminal side of macula densa cells upregulates neuronal nitric oxide synthase, which, in turn, inhibits the tubuloglomerular feedback response, providing a parallel pathway promoting hyperfiltration (20,21). The increase in NaCl delivery following SGLT2i leads to increased production and binding of adenosine to the adenosine receptor (A₁), causing preferential vasoconstriction of the afferent arteriole in type 1 diabetes mellitus and vasodilation of the efferent arteriole in patients with type 2 diabetes mellitus. Because glucose delivery is also increased following SGLT2i and would be predicted to worsen hyperfiltration, the effect of NaCl delivery presumably is dominant. (C) Depiction of changes in vascular tone of the afferent and efferent arteriole and intraglomerular pressure in patients with untreated diabetes mellitus and following SGLT2i. MAP17, 17-kD membrane-associated protein; NHE3, Na⁺-H⁺ antiporter 3; nNOS, neuronal nitric oxide synthase; NO, nitric oxide; P_GC, glomerular capillary pressure.

In a setting where the preglomerular arteriolar diameter is already reduced and/or has limited vasodilatory capacity due to hyalinosis or fibrosis, administration of SLGT2is can lead to efferent vasodilation. In a study of patients with type 2 diabetes mellitus treated with dapagliflozin, the measured GFR decreased in association with a reduced filtration fraction, suggesting a vasodilatory effect on the efferent arteriole (25). The baseline GFR and renal plasma flow were lower and renal vascular resistance was much higher in these participants as compared with the hyperfiltering patients with type 1 diabetes treated with empagliflozin (22). In addition, the majority of these participants were receiving renin-angiotensin system blockers. In the presence of these drugs, increased production of adenosine can induce postglomerular vasodilation via adenosine receptor activation (25,26). In patients with type 1 diabetes mellitus without evidence of hyperfiltration, adding empagliflozin to ramipril modestly reduces GFR without decreasing kidney vascular resistance beyond that caused by the converting enzyme inhibitor alone, suggesting that augmentation of tubuloglomerular feedback is most evident under conditions of hyperfiltration (27). Interestingly, urinary 8-isoprostane levels significantly decreased in these patients, raising the possibility that the combination of the two drug classes may improve oxidative stress.

SGLT2is also affect GFR by altering the pressure gradient from the glomerular capillary to the Bowman’s space through a mechanism independent of tubuloglomerular feedback. Following inhibition of the cotransporter, increased fluid and electrolyte load to distal nephron segments known to have high flow resistance causes increased pressure in the Bowman’s space, thereby contributing to a reduction in GFR (28,29). Increased Na⁺-glucose cotransport prior to treatment reduces the fluid load to distal segments, causing pressure in the Bowman’s space to be reduced, which results in an increased pressure gradient across the glomerular capillary and hyperfiltration. Hemodynamic changes account for the approximately 3- to 6-ml/min reduction in eGFR commonly observed after initiating SGLT2i therapy (30). Considerations of how to approach this situation are given in Table 2.

Table 2.

Practical considerations in prescribing sodium-glucose cotransporter 2 inhibitors

An acute and transient decline in eGFR is common in the first several weeks of therapy a

A decline of <30% does not warrant discontinuation

A decline of >30% should prompt the following

Assess volume status and consider a decreased dose of diuretics

Discontinue prescribed or over-the-counter nonsteroidal anti-inflammatory drugs

A reversible tubular toxicity due to osmotic injury (osmotic nephrosis) can rarely occur (31)

Hold SGLT2i in the setting of acute illness causing depletion of extracellular fluid volume (decreased intake, vomiting, and/or diarrhea)

Symptomatic drop in BP

Consider a decrease in dose of diuretics

Avoid down titration of renin-angiotensin-aldosterone blockers

Hypoglycemia

More likely to occur with eGFR >60 ml/min

Consider a 10%–20% decrease in insulin dose or decrease in the dose of sulfonylurea in collaboration with the endocrinologist

Risk attenuates as eGFR declines and is nonexistent at eGFR <30 ml/min

Given the long-term benefits, every effort should be made to maintain patients on SGLT2i therapy

SGLT2i, sodium-glucose cotransporter 2 inhibitor.

^aThe approach is similar to changes in eGFR following initiation of renin-angiotensin blockers (32).

Natriuretic Effect and Changes in Plasma Volume and BP

SGLT2is induce a natriuresis, resulting in an approximately 7% reduction in plasma volume, which contributes to the BP-lowering effect of SGLT2is that averages 4/2 mm Hg (33,34). In addition to decreasing Na⁺ entry into the cell coupled to glucose, SGLT2is also decrease Na⁺ entry by inhibiting Na⁺-H⁺ antiporter 3 (NHE3) activity. Studies in rodent models show that SGLT2 and NHE3 colocalize in the proximal tubule brush border membrane (35). Inhibition of SGLT2 decreases the activity of NHE3, and the natriuretic effect of SGLT2 is diminished in models where NHE3 is knocked down (36). The functional linkage between NHE3 and SGLT2 may be related to scaffolding and/or accessory proteins, such as the 17-kD membrane-associated protein, that keep the respective transporters in close proximity (37) (Figure 1). Inhibition of NHE3 reduces HCO₃⁻ reabsorption in the proximal tubule, but metabolic acidosis is not a feature of SGLT2i therapy due to a several-fold increase in ammoniagenesis (38). Unlike the diminishing ability to induce glycosuria with advancing CKD, the ability to reduce BP is still evident in patients with CKD stage 3 or 4 (10). This persistent BP-lowering effect with advancing CKD cannot be solely explained by Na⁺ excretion, as the natriuretic effect of SGLT2i is less evident as the eGFR falls below 45 ml/min. Favorable effects on arterial stiffness, weight loss, and decreased sympathetic nerve activity may contribute to the antihypertensive effect of the drugs (39).

Total body Na⁺ content is greater than the amount of Na⁺ in the extracellular fluid. Excess amounts of Na⁺ stored in skin and other tissues are associated with cardiovascular disease, raising the possibility that reducing skin Na⁺ content might improve cardiovascular outcomes (40). Dapagliflozin and empagliflozin both reduce skin Na⁺ accumulation, indicating that the SGLT2is not only reduce plasma volume but also reduce total body Na⁺ content (41,42).

Despite the decrease in plasma volume and BP-lowering effect, there is an absence of a compensatory increase in heart rate, suggesting that SGLT2is exert a blunting effect on sympathetic nerve activity. Studies in human kidney cells and in rodent models of obesity and hypertension support the existence of crosstalk between sympathetic nerve activity and SGLT2i therapy (43–45). Elevations in sympathetic tone increase expression of SGLT2, whereas administration of SGLT2i reduces kidney sympathetic nervous system innervation and activity. A reduction in sympathetic tone may be one way that SGLT2is provide cardiovascular benefits beyond glycemic control (46).

Weight Loss and Changes in Perirenal Fat

SGLT2i therapy is typically associated with 2–3 kg of weight reduction in the first 6 months of therapy resulting from daily glucose losses of 50–100 g corresponding to 200–400 kcal/d (47). The magnitude of weight loss is greater in those with the highest HgA_1c and attenuated in those with good glycemic control and in the setting of CKD, where the filtered load of glucose is reduced (48). The reduction in weight in both rodents and humans in association with reductions in visceral and subcutaneous fat mass is measured by dual-energy x-ray absorptiometry (49). In contrast to weight loss due to caloric restriction, SGLT2i-induced weight loss is accompanied by increased energy expenditure due to activation and recruitment of brown and beige adipocytes (50,51). Caloric deficits due to glucose loss in the urine when combined with enhanced lipid oxidation and increased energy expenditure account for reductions in fat mass (52,53). Studies in obese but nondiabetic rodents demonstrate that SGLT2is also increase fibroblast growth factor 21 (FGF21). Specifically, canagliflozin triggers a fasting-like state characterized by lipid oxidation and ketogenesis but dependent on stimulation of FGF21 to bring about the activation of lipolysis and a reduction in fat mass (54). Changes in fat metabolism in specific fat depots, such as perirenal fat, may offer additional kidney-protective effects (Figure 2).

Figure 2.

SGLT2 inhibition has favorable effects on perirenal fat. The quantity of fat located between Gerota fascia and the kidney capsule is positively correlated with the development of hypertension, urinary albumin excretion, and CKD and negatively correlated with GFR (55–57). Given the close proximity to the kidney parenchyma, release of adipokines and inflammatory factors can adversely affect kidney hemodynamics and function. In mice fed a high-fat diet and treated with the SGLT2i ipragliflozin, changes of diabetic nephropathy are ameliorated in association with healthy expansion of perirenal adipose tissue as compared with controls (58). Perirenal fat exhibited suppressed inflammation and fibrosis and improved insulin sensitivity. In addition, leptin production in the depot was decreased, a protein implicated in the development of glomerulosclerosis (59). Decreased leptin interaction with afferent nerves within perirenal fat may explain the decrease in kidney sympathetic nerve activity associated with SGLT2is (60,61). PPAR-δ, peroxisome proliferator–activated receptor-δ.

Metabolic Flexibility

Glucose in the fed (absorptive) state leads to increased glycolysis, glucose uptake and storage, and suppression of fatty acid oxidation. The fasting (postabsorptive) state is characterized by a drop in insulin levels and increased glucagon, causing activation of lipolysis and promoting fatty acid oxidation. Metabolic flexibility describes the ease and rapidity by which metabolism transitions between these two states (reviewed in ref. 62). A metabolically inflexible state occurs under conditions of chronic overnutrition caused by substrate overload and impaired fuel switching at the level of mitochondria. Unremitting delivery of glucose simultaneously along with fatty acids creates an inflexible state in the mitochondria in which electrons are force fed into the respiratory chain, eventuating in mitochondrial and cellular injury.

The forced excretion of glucose into the urine with SGLT2i leads to metabolic and adaptive responses reminiscent of caloric restriction (63,64). Daily loss of glucose into the urine leads to increased glucagon levels, a reduction in insulin secretion, increased fatty acid oxidation, and decreased intrahepatic lipid content. In both participants with and without diabetes, SGLT2is shift energy metabolism from glucose to fat oxidation and increase endogenous glucose production and ketogenesis (13,65). Empagliflozin given to obese insulin-resistant mice restores insulin-stimulated glucose utilization and metabolic flexibility in the heart in association with reduced ventricular hypertrophy and protection from ischemic stress (66). In a recent study of 26 patients with type 2 diabetes, the respiratory exchange ratio decreased to a greater extent from day to night following dapagliflozin compared with placebo, suggesting improved metabolic flexibility (67). These data suggest that loss of calories in the urine (negative energy balance) extending into the nighttime restore the fed to fasting cycle that is blunted under conditions of chronic overnutrition.

By mimicking a fasting-like state, SGLT2is have the potential to trigger molecular pathways involved in the maintenance of optimal cellular function (68,69) (Figure 3). Increases in plasma ketone concentration following SGLT2is may explain beneficial effects on cardiac function because they provide a readily available fuel source for oxidation by the myocardium. Increased levels of ketone bodies following SGLT2is exert a kidney-protective effect by suppressing mammalian target of rapamycin complex 1 (mTORC1) signaling in the proximal tubule (72,73). States of overnutrition lead to hyperactivation of mTORC1, which is strongly associated with structural injury in models of diabetic kidney disease. Ketone bodies exert favorable effects on cyst formation in rodent polycystic kidney disease accompanied by an inhibitory effect on mTORC1 (74).

Figure 3.

SGLT2is improve metabolic flexibility. Upregulation of SGLT2 and increases of glucose flux bring about a cellular response mimicking nutrient excess where mTORC1 signaling is activated and autophagic flux is depressed. These changes are brought about by the decreased activity of AMPK and SIRT1, both of which are important inducers of autophagy. Inhibition of SGLT2 initiates signaling pathways reminiscent of starvation to include activation of AMPK and SIRT1. The subsequent suppression of mTORC1 allows autophagy to be restored and ultimately suppresses proinflammatory and profibrotic cytokines, attenuating the development of CKD. In addition to limiting the carbon load to the mitochondria by offloading calories in the urine, SGLT2is relieve the reducing pressure resulting from carbon overload across the mitochondria by increasing energy expenditure. These drugs promote activation and beiging of brown adipose tissue and white adipose tissue, respectively. This increase in nonshivering thermogenesis is mediated by increased sympathetic nerve activity directed to adipose depots, increases in fibroblast growth factor 21 levels, and increased brain-derived neurotrophic factor (70,71). Reversal of nutrient excess combined with increases in energy expenditure restores normal levels of carbon flow through the mitochondria, preventing abnormal increases in membrane potential that can lead to the generation of toxic reactive oxygen species. AMPK, AMP-activated protein kinase; CoQ, coenzyme Q; FADH, reduced form of flavin adenine dinucleotide; H₂O₂, hydrogen peroxide; mTORC1, mammalian target of rapamycin complex 1; NAHD, reduced NAD; SIRT1, sirtuin 1 (silent mating type information regulation 2 homolog).

Tubular Workload and Oxygen Requirements

The maintenance of an Na⁺ gradient from tubular lumen to intracellular space to drive glucose reabsorption is an energy-consuming process leading to O₂ consumption. Tubular hypertrophy and increased expression of SGLT2 in patients with diabetes mellitus further increase the ATP-dependent tubular workload and oxygen requirements predisposing this segment to hypoxia. SGLT2is decrease Na⁺ entry into the cell and diminish the amount of ATP required to fuel the activity of the Na⁺-K⁺-ATPase. The estimated daily kidney ATP turnover decreases by approximately 50% in patients with diabetes mellitus treated with SGLT2is (28). Decreased tubular workload and O₂ consumption may explain why SGLT2is decrease proximal biomarkers of injury and the risk of AKI (75,76).

Although cortical oxygenation may improve with SGLT2is, hypoxia is potentially made worse in the outer medulla due to increased delivery and augmented Na⁺ reabsorption in downstream segments. On the other hand, development of hypoxia in these segments may lead to stimulation of HIF2α, providing a stimulus for erythropoietin production. Increases in hematocrit of approximately 3% are a consistent finding in studies of SGLT2is (77,78). Upregulation of HIF2α has also been linked to decreased inflammation, reductions in fibrosis, and stimulation of autophagy (79).

Increased Plasma Magnesium and Phosphate Levels

Plasma magnesium levels increase by approximately 0.08–0.2 mEq/L in clinical trials of SGLT2is and may contribute to cardiovascular benefits by decreasing the risk of cardiac arrhythmias (80). The drop in insulin and increase in plasma glucagon levels can modulate plasma magnesium concentration. Infusion of glucagon into intact or thyroparathyroidectomized rodents results in a significant decrease in absolute and fractional magnesium excretion (81). Perfusion studies localize this effect to the thick ascending limb and the distal convoluted tubule. Increased insulin sensitivity brought about by SGLT2is may also contribute to increases in kidney magnesium reabsorption (52). In this regard, hypomagnesemia and hypermagnesuria develop in association with downregulation of the transient receptor potential melastatin 6 (TRPM6) channel in a rodent model of type 2 diabetes mellitus (82). Restoration of insulin sensitivity following SGLT2i therapy increases expression of the TRPM6 channel, thereby reducing urinary magnesium excretion. In a rodent model of insulin resistance, dapagliflozin decreases urinary magnesium excretion in association with upregulation of TRPM6/7 and claudin 16 (83).

SGLT2is alter kidney phosphate handling in a way that may contribute to enhanced kidney magnesium retention. Small increases in serum phosphate levels due to increased activity of the Na⁺-dependent phosphate transport protein 2a are consistent findings in clinical trials of SGLT2is (84,85) (Figure 1). Increased serum phosphate secondarily increases levels of parathyroid hormone and FGF23, the latter of which inhibits formation of 1,25-dihydroxyvitamin D (85–87). These hormonal changes have been linked to an increase in fracture rate noted in some clinical trials; however, no difference in fracture rate was found in a population-based cohort study comparing the use of SGLT2is and dipeptidyl peptidase-4 inhibitors (88,89). Increased parathyroid hormone levels increase serum magnesium concentration by stimulating reabsorption in the thick ascending limb (90). This effect may be mediated by the decreased abundance of claudin 14 on the basis of studies in mice where disruption of the parathyroid hormone 1 receptor increases claudin 14 expression (91). Increased levels of claudin 14 normally inhibits paracellular magnesium reabsorption by directly blocking claudin 16/19 (92). Decreased levels would be predicted to increase claudin 16/19 permeability, accounting for increased magnesium reabsorption (93).

Decreased Plasma Uric Acid Levels

The glycosuric effect of SGLT2is is linked to increased urinary uric acid excretion, although the precise mechanism is not clear. Increased levels of tubular glucose compete for uric acid reabsorption at the level of glucose transporter 9 (Figure 1). Increased activity of the urate transporter 1 (URAT1) may play a contributory role because the uricosuric effect of canagliflozin persists in rodents with tubular knockdown of glucose transporter 9 (94,95); additionally, the increase in uric acid excretion following SGLT2is in humans is attenuated following pharmacologic blockade of URAT1 (96). The uric acid–lowering effect is accompanied by a significant reduction in the incidence of gout flares and may play a role in the long-term cardiovascular benefits of these drugs (97,98).

Plasma Potassium Changes

Despite the physiologic effects of SGLT2is to increase flow and Na⁺ delivery to the distal nephron, hypokalemia has not been a common complication of these drugs. Hyperkalemia is also uncommon, even though persistent NaCl delivery to the macula densa should suppress renin release and, ultimately, lower circulating mineralocorticoid levels (99). In cardiovascular outcome trials, SGLT2is are utilized in patients receiving renin-angiotensin-aldosterone blockers. In a recent meta-analysis, SGLT2is were found to reduce both the incidence and the severity of hyperkalemia, suggesting that SGLT2is may facilitate the use of maximally tolerated doses of these drugs (100).

Summary

SGLT2is are now embedded in clinical practice guidelines to slow the progression of CKD in patients with and without diabetes mellitus. Inhibition of Na⁺-coupled glucose transport not only lowers plasma glucose concentration but initiates other hemodynamic metabolic pathways that mediate kidney-protective effects. It is important for physicians to be aware of the mechanisms by which they act and the potential changes in kidney function and/or alterations in electrolytes and volume status that may occur.

Disclosures

D.J. Clegg reports consultancy agreements with AstraZeneca. B.F. Palmer reports an advisory role and participation in the speakers bureau for AstraZeneca.

Funding

None.

Author Contributions

D.J. Clegg and B.F. Palmer conceptualized the study, wrote the original draft, and reviewed and edited the manuscript.