Renal Effects of SGLT2 Inhibitors: An Update

Josselin Nespoux; Volker Vallon


Curr Opin Nephrol Hypertens. 2020;29(2):190-198. 

In This Article

SGLT2 Inhibitors Lower Glomerular Hyperfiltration and Preserve Glomerular Filtration Rate in the Long-term

SGLT-mediated glucose transport is coupled with sodium reabsorption. Thus, tubular growth and enhanced glucose reabsorption in the proximal tubule of the diabetic kidney is associated with enhanced sodium and fluid reabsorption (Figures 2 and 3). The latter decreases fluid and NaCl delivery to the downstream juxtaglomerular apparatus (JGA). In the JGA, a group of specialized cells, the macula densa cells, release ATP into the interstitium of the JGA in proportion to the luminal NaCl concentration. The released ATP is broken down to adenosine. Adenosine is the mediator of the tubuloglomerular feedback (TGF). Adenosine lowers hydrostatic glomerular capillary pressure and GFR primarily via activation of vasoconstrictor adenosine A1 receptors on the afferent arteriole, with a smaller contribution of vasodilator adenosine A2 receptors on the efferent arteriole under some conditions.[34,35] By establishing an inverse relationship between NaCl delivery to the macula and GFR, the TGF helps to stabilize distal tubule NaCl delivery. In diabetes, the reduced delivery of NaCl to the macula densa causes an increase in GFR through the physiology of TGF (Figure 2). The increase in proximal tubule reabsorption also reduces tubular back pressure, which further increases GFR. The increase in GFR restores fluid and electrolyte load to the distal tubule and helps to restore sodium and fluid balance. This forms the basis for the tubular hypothesis of diabetic glomerular hyperfiltration and the immediate and desirable GFR-lowering effect of SGLT2 inhibition[4,5] (Figure 2).

Figure 3.

Proposed role for SGLT1 and NOS1 in macula densa in the modulation of glomerular filtration rate and potential interactions with SGLT2 inhibition. Depicted are effects of diabetes/hyperglycemia (black arrows) and the effect of SGLT2 inhibition (grey arrows). Depicted is the sensing of luminal glucose at the macula densa by SGLT1 and the link to GFR via NOS1. Shown are also the complex influences of SGLT2 inhibition on this pathway, including effects on macula densa glucose delivery and effective circulating volume (ECV). Depicted is also the shift in Na transport activity to the outer medulla upon SGLT2 blockade and the consequences on renal medullary oxygenation. Thus, SGLT2 inhibition may mimic systemic hypoxia to the oxygen-sensing structures in the deep cortex and outer medulla, and through an increase in erythropoietin release enhance hematocrit and oxygen delivery to the kidney and other organs. The asterisks indicate that increased UV also contributes to the increase in HCT, that is, the plasma concentration effect. EPO, erythropoietin; GFR, glomerular filtration rate; HCT, hematocrit; MD, macula densa; mTAL, medullary thick ascending limb; NO, nitric oxide; NOS1, nitric oxide synthase 1; PO2, oxygen tension; SGLT1, sodium-glucose cotransporter 1; SNGFR, single nephron glomerular filtration rate; Tm, tubular maximum transport capacity; T2DM, type 2 diabetes mellitus; UK, urinary excretion of potassium; UNa, urinary excretion of sodium; UV, urinary flow rate.

Diabetic hyperfiltration contributes to the development and progression of DKD as it increases the tubular transport burden. Lowering GFR at the onset of treatment forms the therapeutic basis for therapies targeting the renin–angiotensin system and SGLT2 inhibition, thereby alleviating tubular workload. The large SGLT2 inhibitor cardiovascular outcomes trials in T2DM patients demonstrated that empagliflozin, canagliflozin, and dapagliflozin reduce hyperfiltration at the onset of therapy and slow the decline in eGFR in the long-term.[36–38] Most importantly, the GFR reduction in response to SGLT2 inhibition is reversible after drug discontinuation, consistent with the functional mechanisms described above. Similar responses were observed in subgroups of patients with lower kidney function, with a rapid and reversible decrease in eGFR followed by better preservation of GFR in the long-term.[36,37] Similarly, new data in CREDENCE showed that during the first 3 weeks of treatment, eGFR reduction was greater in the canagliflozin group vs. placebo (–3.72 ± 0.25 vs. –0.55 ± 0.25 ml/min/1.73 m2). Thereafter, eGFR decline was slower in the canagliflozin group vs. placebo (–1.85 ± 0.13 vs. –4.59 ± 0.14 ml/min/1.73 m2/year).[9] Additional analyses of the CANVAS-Program trial showed that canagliflozin slowed eGFR reduction across T2DM patients with different levels of albuminuria, and canagliflozin absolute effect was greater in patients with the highest albuminuria level vs. placebo.[39]

In accordance with the discussed role of the TGF mechanism in diabetic hyperfiltration, the increase in GFR in response to streptozotocin (STZ)-diabetes was blunted in adenosine A1 receptor knockout mice, which are TGF-less in response to increasing or decreasing (in diabetes) the macula densa NaCl signal.[40] Moreover, SGLT2 inhibition with empagliflozin increased urinary adenosine excretion in T1DM patients.[41] Using direct in-vivo visualization technique, Kidokoro et al. demonstrated that glomerular afferent arteriole diameter is 50–60% greater and associated with increased SNGFR in Akita T1DM mice compared with nondiabetic mice. Empagliflozin acutely lowered glomerular hyperfiltration within 30 min and reduced afferent arteriole diameter by ~15% compared with vehicle. This was associated with greater urinary adenosine levels. Empagliflozin effects were abolished by pharmacological blockade of the A1 receptor but not by COX-2/NOS1 inhibition, consistent with the notion that SGLT2 inhibition reduces diabetic hyperfiltration predominantly via adenosine-mediated vasoconstriction of the afferent arteriole[42] (Figure 2). Preliminary studies in patients with T2DM suggested that SGLT2 inhibition may also lower GFR by reducing filtration fraction,[43] which may reflect the vasodilator effect of adenosine on the efferent arteriole (Figure 2).

By reducing glomerular hyperfiltration and inhibiting proximal tubular hyperreabsorption, SGLT2 inhibition reduces active tubular transport work and, thereby reduces the energy demand and oxygen consumption, primarily in the kidney cortex, as predicted by mathematical modeling of the rat nephron with normal and reduced nephron number,[13,44] and consistent with experimental data using the SGLT2/SGLT1 inhibitor phlorizin[45] (Figure 2). In this regard, recent studies by Prujim et al. indicated that reduced cortical oxygenation predicts progressive decline of renal function in CKD.[46] By reducing glomerular hyperfiltration and hyperglycemia, SGLT2 inhibition also reduces albuminuria, tubular growth, and tubulointerstitial inflammation.[4,5] Studies by Cassis et al.[47] proposed that podocytes may express SGLT2 in protein overload conditions.