Crane first proposed in 1960 that active glucose transport in the intestinal epithelium (which expresses SGLT1) was energized by the Na ⁺ gradient across cell membranes, the so-called Na ⁺ – glucose cotransport hypothesis (for review, see Wright and colleagues ). The Na ⁺ /K ⁺ -ATPase located on the basolateral membrane is the primary active and ATP-consuming transport step, which lowers cytosolic Na ⁺ concentrations and establishes the concentration gradient that drives Na ⁺ uptake, and, secondary, the uptake of other molecules from the luminal surface into proximal tubule cells ( Figs. 8.1 and 8.4 ). This concept was rapidly refined and extended to active transport processes of a diverse range of molecules and ions including Na ⁺ –glucose cotransport in the kidney.

Regulation of glucose transport in the proximal tubule.

(A) Insulin is a physiologic stimulator of SGLT2, which may serve to maximize renal glucose reabsorption capacity in situations of increased blood glucose levels (e.g., after a meal). At the same time, enhanced Na ⁺ –glucose uptake and insulin suppress renal gluconeogenesis. The latter, in contrast, is stimulated by fasting, which may involve increased catecholamine levels. In metabolic acidosis, the increase in gluconeogenesis from glutamine (Gln) is linked to the formation of (1) ammonium (NH ₄ ⁺ ) , a renally excreted acid equivalent, and (2) new bicarbonate, which is taken up into the circulation. The Na ⁺ -H ⁺ exchanger NHE3 contributes to apical H ⁺ /NH ₄ ⁺ secretion and Na ⁺ /bicarbonate reabsorption. SGLT2 and NHE3 are both stimulated by insulin to enhance Na ⁺ and glucose reabsorption, and their functions may be positively linked through the scaffolding protein MAP17. (B) Diabetes increases luminal glucose delivery to both SGLT2- and SGLT1-expressing segments. Glucose transporters GLUT2 and GLUT1 mediate glucose transport across the basolateral membrane, but GLUT2 may also translocate to the apical membrane in diabetes. Angiotensin II (Ang I) , serum, glucocorticoid-inducible kinase SGK1, hepatocyte nuclear factor HNF-1α, and protein kinase C PKCβ1 promote glucose reabsorption in the diabetic kidney, whereas hypoxia-induced HIF-1α, inflammation, and excessive intracellular glucose levels may be inhibitory. Basolateral glucose uptake via GLUT1/2 may be involved in glycolysis and tubule regeneration after injury as well as hyperglycemia-induced TGF-β.

SGLT1 and SGLT2 have been the most intensively studied members of the human SLC5 solute carrier (SLC) family, which now includes 12 members. Six of these are named as SGLTs, varying in their preferences for binding of glucose, galactose, mannose, fructose, myoinositol, choline, short-chain fatty acids, and other anions. All SGLTs have 15 exons, spanning from 8 to 72 kb, which code for 60- to 80-kDa proteins composed of 580 to 718 amino acids. The molecular nature of SGLTs has been largely pioneered by Wright and colleagues (see for review), who identified and cloned SGLT1 and SGLT2, and showed that defects in SGLT1 were associated with intestinal malabsorption of glucose-galactose. Wright’s group also defined the crystal structure of a sodium galactose bacterial isoform in Vibrio parahaemolyticus (vSGLT), which allowed better characterization of how Na ⁺ and sugar transport are coupled: Na ⁺ binds first to the outside of the transport protein to open the outside gate, thereby permitting outside sugar to bind and be trapped; this is followed by a conformational change and the subsequent opening of the inward gate releases the Na ⁺ and sugar into the cell cytoplasm. The transport cycle is completed by the change in conformation from an inward-facing ligand-free state to an outward-facing ligand-free state. ^, More recently, the structural basis of human SGLT2 and its inhibition by an SGLT2 inhibitor has been resolved.

The affinity of SGLT1 is similar for glucose and galactose, whereas SGLT2 does not transport galactose or fructose. More recent studies in transfected human embryonic kidney (HEK) 293T cells indicated that the apparent affinities ( K _m ) for D-glucose are similar for human SGLT1 and human SGLT2, with values of 2 mM and 5 mM, respectively. Sugar binding occurs in a Na ⁺ -dependent manner, and the K _m values for Na ⁺ transport by human SGLT1 and human SGLT2 are 70 mM and 25 mM, respectively. Thus under euglycemic conditions, the glucose concentration in the tubular fluid of the very early proximal tubule (reflecting blood glucose levels) is similar to the K _m of SGLT2, whereas the luminal Na ⁺ concentration of ∼140 mM is much higher than the K _m of SGLT2 and is not rate limiting.

SGLT1 and SGLT2 transport Na ⁺ and glucose with an Na ⁺ -glucose coupling ratio of 2:1 and 1:1, respectively. The greater Na ⁺ -glucose coupling ratio of SGLT1 enhances its glucose concentration power and hence the ability of the late proximal tubule to effectively reabsorb glucose despite falling luminal glucose concentrations (see Fig. 8.1 ). Na ⁺ -glucose transport is electrogenic, and the membrane potential and driving force are maintained by paracellular Cl ⁻ reabsorption or transcellular K ⁺ secretion via luminal KCNE1/KCNQ1 channels (see Fig. 8.1 ).

The mRNA expression of three other members of the SLC5 family that can transport glucose based on in vitro substrate studies has been detected in the kidney. SGLT3 (SLC5A4) is not a glucose transporter, but glucose can depolarize the plasma membrane in its presence in a saturable, Na ⁺ -dependent manner, and this effect is inhibited by phlorizin. As such, it has been proposed to be a glucose sensor; however, its expression and function in the kidney remain unclear. SGLT4 (SLC5A9) is expressed in the kidney and transports glucose in COS-7 cells but with a lower apparent affinity than mannose (Ki 8 vs. 0.15 mM). Thus SGLT4 may primarily be involved in mannose homeostasis. SGLT5 ( SLC5A10 ) is a Na ⁺ -dependent sugar transporter that has a relatively high affinity and capacity for mannose and fructose relative to glucose and galactose. ^, SGLT5 mRNA is highly kidney abundant and expressed in kidney cortex, ^, and studies in knockout mice have indicated that SGLT5 is the major luminal transporter for fructose reabsorption in the kidney. For an integrated discussion of kidney glucose and fructose transport and metabolism, see review by Vallon and Nakagawa.

In the healthy kidney, the glucose that is being reabsorbed by proximal tubule cells is not linked to appreciable glucose metabolism in these cells. This is due to the fact that most glucose is reabsorbed in the early PCT (S1 segment), but these cells lack significant capacity for aerobic and anaerobic glycolysis. Thus glucose that is taken up across the luminal membrane or formed within proximal tubule cells (see later) exits across the basolateral membrane into the interstitium by concentration-driven facilitative glucose transporters, GLUT2 and GLUT1 (see Fig. 8.1 ) and subsequently enters the peritubular capillaries by convection through fenestrated endothelial cells. GLUT2, the low-affinity (Km; 15–20 mM) “liver” transporter, is primarily expressed in the PCT (S1/S2 segments), but GLUT2 mRNA has also been detected in the proximal straight tubule (PST) (S3 segment). GLUT2 is thought to be the dominant transporter involved in basolateral exit of glucose derived from apical glucose uptake or gluconeogenesis in the PCT. In comparison, GLUT1, the high-affinity (1–2 mM) “erythroid/brain” transporter, is expressed along the entire proximal tubule and has been implicated in transcellular glucose transport, particularly in the S3 segments. Notably, GLUT1 is also expressed in the basolateral membrane of further distal tubule segments and at higher levels than in S3 segments. This includes expression in the medullary thin and thick ascending limbs (TAL) of the rat kidney and at the highest levels in connecting segments and collecting ducts. In the latter, GLUT1 was expressed at the highest level in intercalated cells and to a lesser extent in principal cells. These findings indicated a good correlation between the level of GLUT1 expression and the glycolytic activity of the different nephron segments, indicating that, in particular, the more distal tubule segments are taking up glucose for energy supply via basolateral GLUT1. Some studies have used positron emission tomography and α-methyl-4-[F-18]-fluoro-4-deoxy-d-glucopyranoside to monitor glucose transport in mouse kidneys lacking either SGLT1, SGLT2, or GLUT2. The studies confirmed prominent contributions of SGLT2 and SGLT1 to renal glucose uptake. Moreover, renal glucose reabsorption appeared absent in mice lacking GLUT2, consistent with a more prominent role of GLUT2 versus GLUT1 with regard to basolateral glucose exit of glucose in the proximal tubule (see Fig. 8.1 ). This is in line with the renal phenotype of patients with mutations in GLUT2 and GLUT1. Loss-of-function mutations in GLUT2 is the basis of the Fanconi-Bickel syndrome, which includes a renal Fanconi syndrome, a proximal tubulopathy consisting of glycosuria, phosphaturia, aminoaciduria, proteinuria, and hyperuricemia. The observed proximal tubulopathy may be due to intracellular glucose accumulation and glucotoxicity that occurs when basolateral glucose exit is blocked. In comparison, patients with GLUT1 mutations have primarily neurologic symptoms and no renal phenotype has been documented. ^,

In addition to GLUT1 and GLUT2, some of the other 12 members of the SLC2 gene family have been found in the kidney and may contribute to glucose transport, but little is known about their quantitative contribution. For example, GLUT4 mRNA and immunoreactivity were focally localized in the TALH of the loop of Henle, coexpressed with IGF-I and increased by vasopressin treatment, indicating a potential role in local fuel control. GLUT5 is strongly expressed in the apical membrane of the S3 segment in the rat kidney but is proposed to transport primarily fructose. ^, GLUT12 can transport glucose, and an apical localization has been reported in distal tubules and collecting ducts, but its quantitative role remains to be determined.

The kidneys reabsorb the filtered glucose and use glucose as an energy source, but they also generate new glucose. Gluconeogenesis involves the formation of glucose-6-phosphate from precursors such as lactate, glutamine, alanine, and glycerol with its subsequent hydrolysis by glucose-6-phosphatase to generate free glucose that can exit the cell. The healthy human kidneys produce ∼15–55 g glucose per day by gluconeogenesis, similar to the liver in the postabsorptive state (i.e., 12–16 hours after the last meal). Renal gluconeogenesis in the postabsorptive state is stimulated by epinephrine and falling levels of insulin (see Fig. 8.4 ). Insulin-induced suppression of gluconeogenic gene expression in the proximal tubule was accompanied by phosphorylation and inactivation of forkhead box transcription factor 1 (FoxO1). In contrast to the liver, renal gluconeogenesis is probably insensitive to glucagon. Studies in humans indicated that in the postabsorptive state, renal gluconeogenesis primarily uses lactate as substrate, followed by glutamine, glycerol, and alanine.

In contrast to the uniform stimulation of gluconeogenesis along the entire proximal tubule by starvation, metabolic acidosis enhances gluconeogenesis primarily in S1 and S2 segments. ^{, ,} Furthermore, gluconeogenesis in response to metabolic acidosis primarily uses glutamine as substrate. During the process of renal gluconeogenesis from glutamine, the conversion of glutamine to glutamate and α-ketoglutarate produces ammonium (NH ₄ ⁺ ), which is excreted into the urine as an acid equivalent. The subsequent pathway from α-ketoglutarate to glucose forms new bicarbonate, which is returned as buffer to the systemic circulation (see Fig. 8.4 ). The link between proximal tubular ammonium, bicarbonate, and glucose formation explains why acidosis is a prominent stimulus for renal gluconeogenesis. ^,

Apical glucose uptake via SGLT1 or SGLT2 can have an inhibitory influence on the expression of renal gluconeogenic genes (see Fig. 8.4 ). This effect may serve to prevent glucose overload in the cells of the proximal tubule and has been proposed to involve glucose-induced and sirtuin 1–mediated deacetylation of peroxisome proliferator–activated receptor gamma coactivator 1-α, a coactivator of FoxO1.

In general, it is expected that cytosolic glucose in proximal tubule S3 segments is used for metabolism or leaves the cell via basolateral GLUT1/GLUT2. It has also been hypothesized that glucose generated from lactate in the medullary S3 segment forms part of an intrarenal Cori cycle; glucose enters the lumen by reversed transport through SGLT1 and is taken up by downstream tubular segments, where it is used as energy substrate for glycolysis (e.g., in medullary TAL), and the formed lactate is being returned to the neighboring S3 segment as a substrate for gluconeogenesis. Studies in human proximal tubule segments indicated that in contrast to S1 segments, lactate appears to be a better gluconeogenic precursor than glutamine in S2 and S3 segments. Moreover, studies in mice, rats, and humans indicated that the luminal membrane of the thick ascending limb (and the adjacent macula densa) express SGLT1. ^{, ,} Further studies are needed to determine the potential role of SGLT1 in these structures including a proposed Cori cycle.

Taken together and under normal conditions, the PCT is the site of greatest renal glucose reabsorption and generation. The ability of early proximal tubular segments for gluconeogenesis but inability to metabolize glucose prevents a futile cycle. The reabsorption of filtered glucose and renal gluconeogenesis provide an energy source to distal tubular segments, primarily in the renal medulla, and returning the glucose to the systemic circulation helps to maintain blood glucose levels, particularly in the postabsorptive state. In addition, renal gluconeogenesis is closely linked to the renal response to metabolic acidosis. See Chapter 5 for additional discussion of the role of gluconeogenesis in renal metabolism.

Therapies for type 2 diabetes mellitus (T2DM) have included drugs that target the liver, small intestine, adipose tissue, skeletal muscle, and/or pancreas. Many of these therapies, including insulin, have difficulties to establish adequate glycemic control without the potential for relevant unwanted side effects like hypoglycemia and weight gain and may not reduce cardiovascular complications. The following sections outline how glucose transport changes in the diabetic kidney, its implications for diabetic kidney function, and how targeting renal glucose transport is a antihyperglycemic and kidney-protective therapy.

Diabetes mellitus is associated with increased blood glucose levels. This enhances the amounts of glucose filtered by the kidneys. The early phase of diabetes is often associated with an increase in GFR or glomerular hyperfiltration (see later), which further increases the tubular glucose load. At the same time, the tubular capacity to reabsorb glucose increases by ∼20%–30% to ∼500–600 g/day in patients with T2DM ^, and type 1 diabetes (T1DM). Moreover and despite increased blood glucose levels, diabetes also enhances renal gluconeogenesis. The latter can be the consequence of diabetes-associated metabolic acidosis, and the induced gluconeogenesis involves metabolism of glutamine to glucose associated with the generation of ammonia and bicarbonate (see Fig. 8.4 ). Other potential triggers for gluconeogenesis in diabetes include the activation of the sympathetic nervous system, the reduced insulin levels observed in T1DM, enhanced circulating fatty acids, or cells that are glucose “starved” due to insulin resistance (T2DM).

Increasing renal glucose reabsorption in response to a rise in filtered glucose makes sense with regard to energy substrate conservation. Moreover, the further distal segments may need/use more glucose as an energy substrate to reabsorb the load of salt and other compounds, which is increased due to glomerular hyperfiltration. Renal glucose retention and enhanced glucose formation become maladaptive in diabetes, however, when they sustain hyperglycemia (see Fig. 8.4 ). When blood glucose levels increase to the point that the filtered load exceeds the Tm or tubular transport capacity for glucose, then the surplus is excreted in the urine. In this regard, the kidney provides a safety valve that can prevent extreme hyperglycemia. The safety valve, however, only opens at rather high blood glucose levels (>15 mmol/L) (see Fig. 8.2 ) and only works as long as glomerular filtration is maintained. SGLT2 inhibitors cause this valve to open at lower blood glucose levels.

The levels of protein expression and activity of SGLT2, SGLT1, GLUT2, and potentially GLUT1 determine the capacity of renal glucose reabsorption, and their upregulation may explain the increased glucose transport maximum that can be observed in diabetes. The available preclinical and human studies reported increased, unchanged, or reduced renal glucose transporter expression and/or activity in diabetes or under high-glucose conditions. The observed different responses may reflect different diabetes models, metabolic states, levels of kidney injury, other factors that regulate the expression of these transporters, the use of nonselective antibodies, or dissociations between mRNA and protein expression.

Using knockout mice as gold standard negative antibody controls, the renal protein expression of SGLT2 was found to be increased by 40% to 80% in the early hyperglycemic stages of genetic mouse models of T2DM (db/db) and T1DM (Akita). ^,

Consistent with a potential concerted regulation of luminal and basolateral glucose transport, upregulation of GLUT2 expression has been reported in renal proximal tubules in diabetic rats. Notably, studies in streptozotocin (STZ)-induced T1DM in rats proposed targeting of GLUT2 (but not GLUT1) also to the brush border membrane of proximal tubules. ^, The latter may be linked to protein kinase C PKCβ1 activation and may implicate facilitative glucose transport, together with SGLT2 and SGLT1 in the increased glucose reabsorption across the apical membrane of proximal tubules in the diabetic kidney (see Fig. 8.4 ). On the other hand, renal clearance studies in genetic models of T1DM and T2DM in mice found that coinhibition of SGLT2 and SGLT1 prevented any net glucose reabsorption in the kidney.

The available data on changes in glucose transporters in diabetic patients are sparse and also variable. Primary cultures of human exfoliated proximal tubular epithelial cells harvested from fresh urine of patients with T2DM showed an increased glucose uptake associated with increased protein expression of SGLT2 and GLUT2. An increase in SGLT2 protein expression has also been reported in fresh kidney biopsies of patients with T2DM and advanced nephropathy. On the other hand, the mRNA expression of SGLT2 and GLUT2 was slightly lower in 19 patients with T2DM and preserved kidney function as compared with 20 nondiabetic patients matched for age and estimated GFR (eGFR), all being subjected to nephrectomy. Similar results were reported for SGLT2 and GLUT2 mRNA in another set of patients with T2DM, but the results did not reach statistical significance.

If an increase in SGLT2 expression occurs in the diabetic kidney, then it may simply reflect overall growth and hypertrophy of the diabetic proximal tubule and the associated increase in transport machinery, ^, and this may be exaggerated on the single-nephron level with advanced nephropathy when nephrons are lost and the remaining nephrons aim to compensate. Moreover, hyperinsulinemia associated with extrarenal insulin resistance in obesity and T2DM (as well as postprandial hyperinsulinemia) may phosphorylate SGLT2 at Ser624 and contribute to enhanced renal SGLT2 activity (see Fig. 8.4 ). Furthermore, upregulation of SGLT2 expression in diabetic rats has been linked to activation of Ang II AT1 receptors and the transcription factor, hepatocyte nuclear factor HNF-1α. The latter, as well as HNF-3β, have also been implicated in renal GLUT2 upregulation (see Fig. 8.4 ). Notably, pharmacologic inhibition of SGLT2 in normoglycemic mice also increased renal membrane SGLT2 protein expression, possibly reflecting negative feedback regulation by intracellular glucose levels. Along this line, if renal SGLT2 expression is reduced in the diabetic kidney, this may be due to enhanced diabetes-induced proximal tubular gluconeogenesis (see Fig. 8.4 ) or reflect more severe tubular hypoxia or damage.

The renal expression of SGLT1 protein appears to vary among genetic mouse models of diabetes: Renal SGLT1 protein expression was found to be increased in leptin-deficient ob/ob mice, a model of T2DM, and reduced in Akita mice, a model of T1DM; the latter study used knockout mice as negative antibody control. In contrast to SGLT2 (see earlier), insulin stimulation slightly decreased SGLT1-mediated Na ⁺ –glucose transport in HEK-293T cells, indicating differences in the regulation of these two transporters. In contrast to the strong increase in SGLT2 (see earlier), SGLT1 protein was not significantly changed in fresh kidney biopsies of patients with T2DM and nephropathy in comparison with nondiabetic controls. The interpretation of renal SGLT1 mRNA expression data may be complicated by the observation that mRNA and protein expression can dissociate, at least in mouse kidneys.

GLUT1 protein expression was downregulated in proximal tubules isolated from rat cortices at 2 and 4 weeks after STZ but increased in kidneys of rats at 30 weeks after STZ. A study in patients with T2DM and preserved kidney function reported that in whole renal tissue, GLUT1 mRNA expression was slightly lower as compared with nondiabetic patients.

Why should diabetes reduce renal SGLT1 expression? Although this would permit the renal glucose “valve” to open earlier (and make SGLT2 inhibitors more efficacious, see later), this may not be the kidneys’ intention. A reduced renal SGLT1 protein expression was also observed in response to genetic or pharmacologic SGLT2 inhibition in nondiabetic mice. ^, These conditions and diabetes share an enhanced glucose load to the late proximal tubule. In vitro studies in proximal tubule cells indicated that high glucose can reduce SGLT expression and Na ⁺ –glucose cotransport activity via enhanced oxidative stress. Studies in a model of pig epithelial tubular cells (LLC-PK1) showed that hypoxia can diminish SGLT1 (and SGLT2) protein expression by activation of hypoxia-inducible factor-1α (HIF-1α). Thus an increased glucose load to the outer medullary S3 segment enhances Na ⁺ –glucose reabsorption and thereby hypoxia, which may downregulate SGLT1 to limit oxygen-consuming transport work and glucotoxicity in this segment, which has a high sensitivity to acute injury (see Fig. 8.4 ).

In comparison, an increase in SGLT1 expression in the diabetic kidney would further increase the renal glucose reabsorption capacity but may put the S3 segment at risk of hypoxia and enhanced glucotoxicity. Studies in Akita diabetic mice indicated that the serum and glucocorticoid-inducible kinase SGK1 may stimulate SGLT1 activity and glucose reabsorption in PSTs. SGK1 could also promote proximal tubular glucose reabsorption by enhancing the activity of luminal K ⁺ channels (KCNE1/KCNQ1), which maintain the electrical driving force during electrogenic Na ⁺ –glucose cotransport (see Fig. 8.4 ). SGK1 was upregulated in proximal tubules in patients with diabetic nephropathy.

Transport functions in proximal tubules require high turnover of ATP, which, under normal conditions, is derived primarily through mitochondrial oxidative phosphorylation. ^, This may change in pathophysiologic situations with impaired mitochondrial function, when glycolysis may be enhanced and contribute to maintaining ATP. For example, a shift to glycolysis has been proposed in proximal tubules regenerating from acute kidney injury (AKI), as well as proximal tubules undergoing atrophy. This metabolic switch to glycolysis occurred early during proximal tubule regeneration and was reversed during successful tubular recovery, but it persisted and became progressively more severe in tubule cells that failed to redifferentiate. Tubular upregulation of HIF-1α in mice enhanced renal GLUT1 mRNA expression; this was associated with less oxygen consumption and increased glycolysis. Thus hypoxia may increase basolateral GLUT1-mediated facilitative uptake of glucose, which is then used for glycolysis and recovery. Hypoxia-induced GLUT1 likely applies to distal tubule segments but may also be relevant for medullary S3 segments. In this regard, studies in the proximal tubular cell line LLC-PK1, cultured and polarized on porous tissue culture inserts, showed that basolateral exposure to 25 mmol/L D-glucose enhanced glucose uptake via GLUT1 and the subsequent intracellular metabolism of glucose enhanced TGF-β 1 synthesis and secretion; this was not observed in response to apical glucose exposure. These in vitro studies suggest that it may be the hyperglycemia-induced persistent uptake of glucose via basolateral GLUT1 (or GLUT2), rather than the filtered glucose, that affects the tubular synthesis of TGF-β 1 and thereby the development of tubulointerstitial fibrosis and tubular growth (see Fig. 8.4 ).

By reducing hyperglycemia, SGLT2 inhibitors have the potential to reduce glucotoxicity in the kidney and extrarenal organs. ^, In accordance, studies in diabetic rodent models with severe hyperglycemia have shown that SGLT2 inhibition can reduce growth, lipid accumulation, inflammation, and injury of the diabetic kidney and improve autophagy secondary to strong blood glucose–lowering effect ^{, , , ,} ( Fig. 8.5 ). The observed smaller effect of SGLT2 inhibitors on blood glucose in the clinical trials contributes to, but cannot fully explain, the profound and rapid impact on cardiac and kidney outcomes observed in nondiabetic patients and appears insufficient to fully explain the rapid beneficial effect on heart failure detectable within a few months.

Kidney-protective mechanisms of SGLT2 inhibition.

SGLT2 inhibition reduces the reabsorption of glucose and Na ⁺ in the early proximal tubule. This increases the delivery of NaCl and K ([Na-Cl-K] _MD ) and fluid (V) to the macula densa, which lowers glomerular filtration rate (GFR) and glomerular capillary pressure (P _GC ) through tubuloglomerular feedback (TGF) and by increasing hydrostatic pressure in the Bowman space (P _Bow ). Lowering P _GC & GFR and hyperglycemia protects glomerular and tubular function by reducing filtration of albumin, decreased tubular growth, inflammation, and reduced tubular transport work, which lowers cortical oxygen demand. Tubular transport work and toxicity are also reduced by inhibition of Na-H-exchanger NHE3. Shifting glucose and Na ⁺ reabsorption to SGLT1 and medullary thick ascending limb (mTAL) lowers oxygen level in the outer medulla, which may activate hypoxia-inducible factor (HIF) and enhance erythropoietin release. The resulting increase in hematocrit (Hct)(12) improves O ₂ delivery to kidney medulla and cortex. More delivery of NaCl and fluid downstream of early proximal tubule may facilitate responsiveness to atrial natriuretic peptide (ANP) and enhance the diuretic effect of SGLT2 inhibition, leading to reduced blood pressure. SGLT2 inhibitors are uricosuric, which has the potential to further protect the kidney and heart.

This figure was modified with permission from Vallon V. Glucose transporters in the kidney in health and disease. Pflugers Arch . 2020;472[9]:1345–1370.

In patients with T2DM, the glucosuric effect of SGLT2 inhibition is associated with a 2- to 3-kg lower body weight. Although the diuretic effect and fluid loss may contribute to the initial weight loss, the majority of the steady-state weight loss with SGLT2 inhibitor treatment is due to fat loss, including visceral and subcutaneous fat (see Fig. 8.5 ), as a consequence of a shift in substrate utilization from carbohydrates to lipids. ^{, ,} The released free fatty acids are used by the liver to form ketone bodies and thus increase ketogenesis. SGLT2 inhibitors may improve cardiac and kidney outcomes in part by increasing plasma levels of ketone bodies like β-hydroxybutyrate, which are used as additional energy substrates to improve the performance of cardiac myocytes (or kidney cells) in diabetes mellitus ^, (see Fig. 8.5 ). Importantly, SGLT2 inhibitors can increase the risk of diabetic ketoacidosis, particularly when the drugs are used off label in patients with T1DM.

In contrast to insulin therapy, SGLT2 inhibitors do not increase the incidence of hypoglycemia. ^, This is because they become ineffective at lowering blood glucose any further once the filtered glucose load falls to ∼80 g/day, which can be handled by renal SGLT1 (see Fig. 8.1 and 8.5 ). In addition, SGLT2 inhibitors leave the metabolic counterregulation intact, which can increase plasma glucagon concentrations and subsequently endogenous hepatic but also kidney glucose production (gluconeogenesis) in patients with T2DM. ^{, ,} This is potentially relevant for cardiovascular outcomes because episodes of hypoglycemia can reduce the cardioprotective effects of antihyperglycemic therapy.

A meta-analysis of patients with T2DM treated with SGLT2 inhibitors found a consistent decrease in systolic blood pressure of 3 to 6 mm Hg, The magnitude of this blood pressure effect is expected to have cardiovascular protective consequences, particularly in high-risk patients. The blood pressure–lowering effect of SGLT2 inhibition relates to the reduction in body weight and a modest glucose-based osmotic diuresis (100–470 mL/day) and a small natriuretic effect. ^, The lower blood pressure and an associated modest reduction in plasma volume may rapidly reduce cardiac preload and afterload and thereby contribute to the rapid beneficial effects in heart failure patients (see Fig. 8.5 ). SGLT2 is functionally linked to the Na-H-exchanger NHE3, ^{, ,} such that SGLT2 inhibition induces natriuresis and lowers blood pressure in part by inhibiting NHE3 in the proximal tubule (see Figs. 8.4 and 8.5 ).

Beneficial renal and cardiovascular effects of SGLT2 inhibition may also be due to a plasma uric acid–lowering effect. The uricosuric effect of SGLT2 inhibitors is positively related to the increase in tubular and urinary glucose delivery, as observed in healthy subjects and patients with T2DM ^, (see Fig. 8.5 ) and may involve an interaction with the proximal tubule luminal urate transporter URAT1. ^,

With regard to cardiac protection, more evidence is accumulating that supports the concept that SGLT2 inhibitors may also work through off-target effects on cardiac transporters like NHE-1, Nav1.5, or SGLT1 (see Fig. 8.5 ), but much needs to be learned.

Glomerular hyperfiltration, which is observed in a subset of patients at the onset of T1DM and T2DM, can increase the risk of developing diabetic nephropathy later. Glomerular hyperfiltration increases transport work and oxygen consumption in the diabetic kidney, and lowering GFR has opposite effects.

According to the “tubular hypothesis,” glomerular hyperfiltration in diabetes is explained by primary tubular hyperreabsorption (for review, see Vallon and Thomson ). Moderate levels of hyperglycemia increase proximal tubular reabsorption by providing more substrate for Na ⁺ –glucose cotransport via SGLT2 and SGLT1 and by causing the tubule to grow, which enhances the transport machinery and capacity. The increased reabsorption reduces the NaCl and fluid delivery to the downstream macula densa, which senses this reduction and causes GFR to increase through the normal physiologic action of tubuloglomerular feedback (TGF) ( Fig. 8.6 ). The primary role of the TGF is to stabilize the NaCl and fluid delivery downstream of the macula densa and thereby facilitate the fine regulation of NaCl and fluid balance in the distal nephron by neurohumoral control. A secondary consequence of this TGF physiology is that the mechanism contributes to the autoregulation of GFR and renal blood flow. Moreover, it makes GFR responsive to primary changes in tubular transport upstream of the macula densa, like in the diabetic kidney. A primary increase in proximal reabsorption also reduces the distal tubular flow rate, which increases GFR by lowering tubular back pressure (i.e., the hydrostatic pressure in Bowman space) and thereby increasing the effective glomerular filtration pressure (see Fig. 8.6 ). Mathematic modeling indicates that TGF and the changes in tubular back pressure contribute equally to the increase in GFR in diabetes.

The tubular hypothesis of diabetic glomerular hyperfiltration: effect of SGLT2 inhibition.

(A, B) In vivo micropuncture studies in rats with superficial glomeruli were performed in nondiabetic and streptozotocin-diabetic rats. Small amounts of blue dye were injected into the Bowman space to determine nephron configuration, including the first proximal tubular loop and the early distal tubule close to the macula densa. Tubular fluid was collected close to the macula densa to determine 1. the tubuloglomerular feedback signal ([Na-Cl-K] _MD ) and 2. single-nephron glomerular filtration rate (SNGFR) by inulin clearance. The Bowman space was punctured to determine the hydrostatic pressure (P _Bow ) . Measurements were performed under control conditions and following application of the SGLT2/SGLT1 inhibitor phlorizin into the early proximal tubule (i.e., without changing systemic blood glucose levels). (C) Basal measurements (bsl) revealed that glomerular hyperfiltration in diabetes was associated with reductions in [Na-Cl-K] _MD and P _Bow . Adding phlorizin (P) had a small effect in nondiabetic rats but normalized [Na-Cl-K] _MD , P _Bow , and SNGFR in diabetes. (D) Diabetes induces a primary hyperreabsorption in the proximal tubule due to tubular growth and enhanced Na ⁺ –glucose cotransport, which, through tubuloglomerular feedback ([Na-Cl-K] _MD ) and reducing tubular back pressure (P _Bow ), causes glomerular hyperfiltration. SGLT2 contributes to tubular hyperreabsorption, and as a consequence, SGLT2 inhibition mitigates these changes and lowers glomerular hyperfiltration.

This figure was modified with permission from Vallon V, Thomson SC. Targeting renal glucose reabsorption to treat hyperglycaemia: the pleiotropic effects of SGLT2 inhibition. Diabetologia 2017;60:215–225.

Conversely, SGLT2 inhibition attenuates proximal tubule hyperreabsorption in the diabetic kidney and thereby lowers diabetic glomerular hyperfiltration (see Figs. 8.5 and 8.6 ). This has been shown in micropuncture studies in rats using direct application of phlorizin into the Bowman space and by acute or chronic systemic application of selective SGLT2 inhibitors. In accordance, pharmacologic or genetic inhibition of SGLT2 suppressed hyperfiltration on the whole-kidney level in diabetic mice. ^, Consistent with the proposed local mechanism, the suppression of diabetic hyperfiltration in response to SGLT2 inhibition was associated with an increase in the NaCl concentration at the macula densa ^, and in the hydrostatic pressure in the Bowman space, and it was independent of effects on blood glucose ^{, ,} (see Fig. 8.5 ).

When macula densa cells sense an increase in luminal NaCl, the resulting TGF-induced ATP release promotes local formation of adenosine, which primarily constricts the afferent arteriole via adenosine A ₁ receptors but can also dilate the efferent arteriole via adenosine A ₂ receptors (see Fig. 8.5 ). Both effects are expected to lower glomerular capillary pressure (P _GC ). Micropuncture of glomerular capillaries in rats with T1DM established that the SGLT2 inhibitor ipragliflozin indeed reduced GFR and P _GC . However, the changes in P _GC and GFR were not strictly correlated, consistent with the asymmetry of afferent and efferent arteriolar TGF responses and their consequences on GFR and P _GC. Thus SGLT2 inhibition can induce a robust reduction in P _GC even when GFR decreases only slightly and vice versa. This may have implications in advanced CKD (GFR<30 mL/[min × 1.73m ² ]), where the initial GFR drop in response to SGLT2 inhibition can be smaller, but the kidney-protective effect is preserved, possibly due to a robust effect on the efferent arteriole and predominant reduction in P _GC.

The initial GFR-lowering effect has been confirmed in humans. The SGLT2 inhibitor empagliflozin decreased GFR by 19% in T1DM patients with baseline hyperfiltration independently of lowering blood glucose levels. The SGLT2 inhibitor canagliflozin initially lowered eGFR in patients with T2DM and basal eGFR of ≥55 mL/min/1.73 m ² . Following this initial dip, eGFR remained stable over the following weeks and months in the canagliflozin-treated group such that eGFR was better preserved after 2 years of follow-up and associated with reduced urinary albumin-to-creatinine ratios than in the control group, which had been treated with glimepiride to achieve similar blood glucose control. Similar biphasic eGFR profiles (i.e., initial dip followed by subsequent better preservation) have been shown for other SGLT2 inhibitors.

Surviving nephrons in advanced stages of CKD are assumed to hyperfilter as a way of compensation for the reduced nephron number and thus maintain a high-glucose load on the level of the single nephron. This should preserve the acute GFR-lowering effect of SGLT2 inhibition, even if the effect on overall glucose homeostasis is attenuated. Importantly, lowering single-nephron glomerular hyperfiltration in CKD and thereby the oxygen-consuming transport work may help to preserve the integrity of the remaining nephrons and overall kidney function over the long term (see Fig. 8.5 ). The additive effect of inhibitors of the renin-angiotensin system and of SGLT2 inhibitors is consistent with the concept that angiotensin II blockade primarily dilates the efferent arteriole, whereas SGLT2 inhibition primarily constricts the afferent arteriole.

It has been discovered that the macula densa senses an increased glucose delivery via luminal SGLT1, which then activates nitric oxide synthase 1 (NOS1), and the resulting increase in NO formation blunts the afferent arteriolar vasoconstrictor effect of TGF and thereby contributes to diabetic hyperfiltration. ^, Thus the increase in macula densa glucose delivery in response to an SGLT2 inhibitor can activate the macula densa-SGLT1-NOS1 pathway and, thereby, potentially limit the initial fall in GFR. On the other hand and considering the close proximity of the macula densa to both afferent and efferent arterioles, this pathway may also dilate the efferent arteriole, potentially in settings with endothelial dysfunction and low efferent NO tone.

In patients with T2DM and albuminuria, dapagliflozin increased urinary metabolites linked to mitochondrial metabolism, potentially indicating that dapagliflozin improves mitochondrial function in the diabetic kidney. Consistent with this, scRNA sequencing of proximal tubules in db/db mice indicated that while RAS blockade is more antiinflammatory/antifibrotic, SGLT2 inhibition affected genes related to mitochondrial function. ScRNA sequencing from research kidney biopsies from young persons with T2DM indicated that SGLT2 inhibition mitigated diabetes-induced metabolic perturbations via suppression of mTORC1 signaling in kidney tubules. Studies in nondiabetic mice provided evidence that SGLT2 inhibition causes distinct effects on kidney metabolism, reflecting responses to partial NHE3 inhibition, and as well as urinary loss of glucose and NaCl; this included upregulation in renal gluconeogenesis and using tubular secretion of the tricarboxylic acid (TCA) cycle intermediate, α-ketoglutarate, to communicate to the distal nephron the need for compensatory NaCl reabsorption.

Mathematic modeling predicted that inhibition of SGLT2 in the diabetic kidney reduces oxygen consumption in the PCT and renal cortex, in part by lowering GFR ^, (see Fig. 8.5 ). The predicted increase in cortical O ₂ pressure and availability has been observed in a diabetic rat model using phlorizin, a dual SGLT1/SGLT2 inhibitor. Preserving renal cortical oxygenation may be important to preserve kidney function in patients with CKD.

SGLT2 inhibition causes more equal distribution of renal transport work as it shifts glucose uptake downstream to the S3 segments (see earlier) and enhances transcellular Na ⁺ reabsorption in distal segments, including the S3 segment and medullary TAL. This may further reduce the already physiologically low O ₂ availability in the renal outer medulla. The latter has been proposed for SGLT2 inhibition using mathematic modeling ^, and was shown in vivo in rats in response to acute dual SGLT2/SGLT1 inhibition by phlorizin in nondiabetic and diabetic rats. The effect on medullary transport and oxygenation would be attenuated by the reduction in blood glucose and GFR in response to SGLT2 inhibition. ^, Moreover, the proposed SGLT2 inhibitor–induced reduction in oxygen pressure in the deep cortex and outer medulla may stimulate hypoxia-inducible factors HIF-1 and HIF-2 (see Fig. 8.5 ). Gene knockout of SGLT2 increased the renal mRNA expression of hemoxygenase, ^, a tissue-protective gene that is induced by HIF-1α. On the other hand, activation of HIF-2 may explain an enhanced erythropoietin release from renal interstitial cells in response to SGLT2 inhibition. Together with the diuretic effect, the latter may contribute to the observed modest increase in hematocrit and hemoglobin in response to SGLT2 inhibition. This may improve the oxygenation of the kidney outer medulla and cortex but also facilitate oxygen delivery to the heart and other organs. For a detailed discussion of the clinical use and outcomes of SGLT2 inhibitors in CKD and diabetic kidney disease, see Chapter 41 , Chapter 55 . In other words and in addition to its volume effect, SGLT2 inhibition may simulate systemic hypoxia to the oxygen sensor in the deep cortex and outer medulla of the kidney, and the induced response then helps the failing heart and also the kidney. In accordance with an overall nephroprotective effect, SGLT2 inhibitor use also reduced the risk of AKI in an analysis of >3000 patients with T2DM by ∼50%. Nevertheless, caution is warranted, as excessive volume depletion and the transport shift to the outer medulla may increase the AKI risk in individual sensitive patients.

The amount of filtered glucose determines the glucosuric and blood glucose–lowering effect of SGLT2 inhibition. As a consequence, the antihyperglycemic effects of SGLT2 inhibitors are attenuated in patients with reduced GFR. In contrast, the blood pressure–lowering and heart failure–protective effects are preserved in patients with CKD and reduced GFR (eGFR ≥30 mL/min 1.73 m ² ) ^, Modeling studies of CKD and nephron loss predicted that the increase in single-nephron GFR in remaining nephrons and the reduction of glucose reabsorption by SGLT2 inhibition increase paracellular Na ⁺ secretion in the proximal tubule Thus the model predicts that the chronic natriuretic and diuretic effects of SGLT2 inhibition persist in CKD. The modeling approach also predicts that the SGLT2 inhibition–induced changes in the oxygen signal at the renal sensor are preserved in CKD

Both OAT1 and OAT3 were among the original seven drug transporters that regulatory agencies identified as important for analysis of transport of new drug entities. This regulatory attention has perpetuated the notion that these transporters primarily transport drugs. Although it is true that the OATs transport many well-known drugs (e.g., antibiotics, antivirals, NSAIDs, diuretics) and toxins (e.g., organic mercurials and aristolochic acid), it is now clear that they, as well as the other five SLC and ABC transporters highlighted by regulatory agencies, transport many endogenous metabolites, signaling molecules, vitamins, gut microbiome, and dietary products.

Indeed, the primary function of these multispecific transporters may not be the handling of drugs and toxins but rather the modulation of local and systemic metabolism and signaling. ^, Much of this change in our understanding of “what drug transporters really do” is the result of “omics” analyses of knockout mice, human genome-wide association studies (GWAS), and the identification of heritable mutations that cause or modulate well-known metabolic diseases. According to this new systems biology view (explained in more detail later), the multispecificity of these transporters for endogenous substrates enables a range of drugs and toxins to coopt these transporters expressed in the gut, liver, kidney, and many other tissues. But the pharmaceutical and commercial relevance of these transporters can create the misconception that these widely expressed and evolutionarily conserved genes primarily exist to handle synthetic drugs. Although this is apparently their key role from the perspective of clinical pharmacology and pharmacokinetics, it is becoming clear that the pharmaceutical lens is extremely limited even from the clinical point of view, since it is now evident that OAT1, OAT3, and other “drug” transporters are central to endogenous physiology and are important for understanding the pathophysiology of uremia, hyperuricemia, carnitine deficiency, and a number of genetic conditions. Importantly, short-term blockade in humans of OAT1 and OAT3 with the drug probenecid results in impressive metabolic alterations.

Oat1 knockout mouse tissue has defective uptake of the classic organic anion transporter probe, PAH, whereas the Oat3 knockout mouse tissue has defective uptake of estrone sulfate. ^, Consistent with in vitro data, the Oat1 and/or Oat3 knockout mice have altered in vivo or ex vivo (e.g., embryonic kidney organ cultures) handling of diuretics (e.g., loop and thiazides), antibiotics (e.g., penicillin and ciprofloxacin), a wide range of antiviral agents, and methotrexate.

Knockout mice responses to loop and thiazide diuretics provide a useful illustration ( Fig. 8.9 ). These albumin-bound drugs in the peritubular capillaries must be transported by basolateral uptake transporters OAT1 and OAT3 in the proximal tubule, transit the cell, and exit through apical transporters (including members of the ABCC or MRP families)—all before flowing down the luminal (urinary) space to be excreted or, for the loop and thiazide diuretics, to inhibit salt reabsorption in later nephron segments. Three to five times more diuretic is required to achieve the same degree of natriuresis after deletion of either Oat1 or Oat3 in mice. ^,

(A, B) Marked attenuation of thiazide and loop diuretic effects in Oat1 and Oat3 knockout mice.

In Oat1 or Oat3 knockout, which transport diuretics ( Table 8.1 ), luminal sodium elimination is markedly attenuated. ED ₅₀ , Half-maximal effective dose; IV, intravenous; UN aV , urinary sodium excretion.

Modified from Eraly SA, Vallon V, Vaughn DA, et al. Decreased renal organic anion secretion and plasma accumulation of endogenous organic anions in OAT1 knockout mice. J Biol Chem. 2006;281:5072–5083; and Vallon V, Rieg T, Ahn SY, et al. Overlapping in vitro and in vivo specificities of the organic anion transporters OAT1 and OAT3 for loop and thiazide diuretics. Am J Physiol Renal Physiol. 2008;294:F867–F873.

OATs are implicated in renal organic mercurial toxicity because mercury binds to glutathione and other thiol-containing compounds, many of which are “effectively” seen as organic anions by the transporter. When the Oat1 knockout mouse was treated with high-dose mercury, the kidneys were surprisingly well protected from injury (histologically and by renal indices), consistent with the inability of the organic mercurial to be taken up by the proximal tubule due to the absence of Oat1.

Perhaps most revealing, and somewhat unexpected, have been the results from performing metabolomics analyses on the plasma of the Oat knockout animals. ^{, ,} These have revealed a somewhat surprising range of endogenous OAT substrates. Generally speaking, both OAT1 and OAT3 appear to play a key role in regulating the flow of organic anions through the so-called gut–liver–kidney axis. This includes the renal handling of many compounds derived from the gut microbiome—either products of the gut microbiome or due to the action of the gut microbiome on dietary components such as phytochemicals. For example, in the Oat3 knockout, among the greatest changes are in flavonoids that have been acted upon by phase 2 liver enzymes (e.g., glucuronidation). This highlights the important connection between OATs (especially OAT3) with liver metabolism (via so-called drug-metabolizing enzymes) of endogenous compounds, as well as drugs and toxins. Other groups of metabolites elevated in the Oat3 knockout include primary and secondary bile acids. In Oat1 and Oat3 knockouts, fatty acids, TCA cycle intermediates, and vitamins are also elevated. A double knockout of Oat1 and Oat3 in rats has been reported. Many of the analyses performed, including testing the diuretic response and the measurement of uremic toxins, were found to be consistent with results described earlier in the single knockout mice. Interestingly, the double knockout rats had uremic toxin elevation and renal dysfunction at 4 weeks that apparently resolved by 7 weeks. This compensation appears due to increased expression of another proximal tubule organic anion transporter of uremic toxins, slco4c1.

With respect to drugs, OAT1 and OAT3 strongly favor anionic drug substrates, but both transporters (especially OAT3) can bind a limited number of cationic/zwitterionic drugs, such as cimetidine. Importantly, physiochemical properties of metabolites interacting with the OATs appear different from the drug data. OAT3 metabolite substrates are larger, less polar, more chemically complex, and contain more ring structures than OAT1, though there is considerable overlap in substrates. ^,

In addition, metabolic reconstructions based on changes in gene expression in the knockouts have been performed. ^{, ,} These indicate that OATs regulate many biochemical pathways, including purine metabolism, TCA cycle, and metabolism of fatty acids, eicosanoids, amino acids, and vitamins (e.g., see top pathways affected by Oat1 loss in knockout mice in Table 8.2 ). Together, the reconstructions and the metabolomics data support the view that OATs are not simply “drug” transporters but have major impact on many aspects of systemic and proximal tubule physiology.

Together, the metabolic reconstructions based on knockout omics data and the analysis of endogenous substrates (e.g., metabolites and signaling molecules) from the perspective of chemical properties call into question the oft-discussed “redundancy” of OAT1 and OAT3 in the proximal tubule. From a practical pharmacokinetic perspective, this view may still be a useful first approximation for many drugs that can interact with both OAT1 and OAT3. But considering the endogenous metabolite preferences alone (without considering drugs), a more appropriate view might be that the two transporters have distinct roles in many metabolic processes, although they work together to handle certain substrates like uric acid. On the basis of current data, renal OAT1 appears somewhat more linked to local and systemic aerobic metabolism, whereas renal OAT3 appears more linked to flow of metabolites that originate in the gut or liver (e.g., primary and secondary bile acids). Focused analyses have shown the critical role OAT1 plays in regulation of metabolism involving tryptophan, lipids, and the gut microbiome. There is also some evidence to suggest that OAT3 could modify phenotypes, for instance, in diabetic disease, blood pressure, and in the setting of treatment with SGLT2 inhibitors.

Unlike the OAT subclade (of the SLC22 OAT major clade), which is large, the OCT subclade (one of three subclades of the SLC22 OCT major clade) consists of three highly homologous (both protein sequence and function) transporters: OCT1, OCT2, and OCT3. OCTs are generally held to be electrogenic uniporters. Organic cation transporter 2 (OCT2) is the main renal uptake transporter on the basolateral membrane (blood side) of the proximal tubule cell that is involved in the elimination of organic cationic drugs such as metformin and cis -platinum. ^, On the luminal (apical, urine side) of the cell, it appears that MATE (SLC47) transporter family members efflux organic cations into the urinary space. The primary liver OCT, OCT1—one of the drug transporters that, along with OAT1 and OAT3, has been highlighted by regulatory agencies as important for new drug testing to identify transport mechanisms—has been shown to be a thiamine transporter.

Thus as with the OATs, OCTs may function primarily in regulating metabolite flow into and out of tissues, and as with OATs, some of the best in vivo functional information regarding endogenous function comes from analysis of metabolites altered in the knockout mice. On the whole, the known drug and metabolite substrates for the OCTs appear less diverse than for the OATs, and it appears that based on physiochemical property analysis of drug substrates, OCT1 and OCT2 have largely overlapping specificities, at least for drugs.

The OCTs appear more polymorphic than the OATs; SNPs in OCT2 have received considerable clinical attention because, consistent with in vitro studies, they can affect levels of the antidiabetic agent metformin and the chemotherapeutic agent cis -platinum.

Nonsynonymous coding region single-nucleotide polymorphisms (SNPs) in the OATs are uncommon compared with noncoding region SNPs. OAT SNPs have also been reported that affect diuretic responsiveness, mercury toxicity, antibiotic levels, and hyperuricemia. A noncoding SNP in the OAT1 gene appears to be associated with progression of renal disease.

The importance of OATs in renal handling of gut microbiome–derived metabolites merits further discussion. Many of the metabolites accumulating in the Oat knockouts, such as indoxyl sulfate, kynurenine, p-cresol sulfate, and hippurate, are among the sets of gut microbiome–derived small molecules (organic anions); they are also frequently implicated as “uremic toxins” ^, ( Table 8.3 ).

The list of small molecules implicated in uremic toxicity is long and much debated. ^, These uremic toxins are thought to play a role in many tissue and organ toxicities and dysfunctions that occur in the uremic syndrome associated with severe CKD. It is unlikely that any single uremic toxin on the list is the key to all the manifestations of the uremic syndrome, although there is growing evidence that certain uremic toxins play a role in particular tissue toxicities. These include TMAO (trimethylamine-N-oxide), which has been implicated in cardiovascular toxicity, ^, and indoxyl sulfate, which has been implicated in multiple aspects of the uremic syndrome.

Some of these uremic toxins (e.g., indoxyl sulfate) may play a role in the actual progression of renal disease, presumably in part via OAT-mediated uptake into proximal tubule cells. ^, The pathophysiology of uremia is beyond the scope of this chapter, and readers are referred instead to Chapter 51 . But it is important to emphasize the growing appreciation of the role of OAT1 and OAT3 in regulating levels of many molecules considered uremic toxins. Thus it will be interesting to determine whether factors that affect the expression or function of OAT1 and OAT3 alter the time at which the uremic syndrome develops in CKD or the progression of renal disease.

OCT2 also appears to be a transporter of certain cationic uremic toxins, notably TMAO, a molecule accumulating in CKD that is associated with cardiovascular disease. The apical transporters of organic anions, the MATEs, appear to be involved in the efflux of TMAO. Of note, the deletion in mice of OAT3, which can transport a few organic cations despite being an organic anion transporter, results in elevated levels of TMAO, although it is not clear whether TMAO is actually transported by OAT3.

Uric acid is a product of purine metabolism and a weak organic acid with a pKa of 5.5-5.75, existing primarily as urate at physiological pH. The formation of uric acid is catalyzed by the enzyme xanthine oxidase/xanthine dehydrogenase from xanthine and hypoxanthine. In many organisms, uric acid is oxidized by the enzyme uric acid oxidase (uricase) into allantoin; however, mammalian primates demonstrate a gradient of reduced uricase function culminating in complete loss of function of the uricase gene (UOX) in the great apes. ^, Humans have accumulated three separate mutations in UOX supporting the argument of strong selective pressure for the loss of uricase function. ^, The hypothesized adaptive advantage gained with the loss of uricase function remains controversial, though roles of urate as an antioxidant and regulator of beneficial metabolic pathways have significant support. ^, The disadvantages for human health, however, are clear: Increased serum urate levels result in elevated risk of monosodium urate crystal precipitation and gout. ^, In addition to gout, hyperuricemia (clinically defined elevated serum urate levels) and urate are independently associated with the primary drivers of complex human diseases including renal diseases, hypertension, cardiovascular disease, and metabolic syndrome. ^{, ,} The root of all urate-related pathology is a derangement of the careful balance between urate production, primarily in the liver, and excretion, maintained by epithelial transport systems of the liver, intestines, and kidney.

The kidney is responsible for 70% of total urate excretion ^, ; however, the homeostatic balance between excretion of a nitrogenous waste product and prevention of urate crystal deposition has resulted in a complicated distribution of urate absorption and secretion transporters in the proximal tubule. ^, Urate is freely filtered by the glomerulus, yet the final fractional excretion of urate is between 5 and 15%, ^{, ,} suggesting active urate transport along the tubule. Early transport studies using radiolabeled uric acid and renal epithelial brush border membrane vesicle studies suggested components of active transport and active secretion along the three segments of the proximal tubule. Our current understanding of urate handling suggests that up to 95% of the filtered urate is rapidly reabsorbed in the S1 segment of the proximal tubule as part of the initial bulk reabsorption of many organic anions. ^{, ,} Next, there is active secretion of urate back into the tubule lumen, as much as 50% of the original filtered load, occurring in the S2-S3 segments of the proximal tubule. Proximal urate handling is completed with a second phase of reabsorption in the latter S2 and S3 segments. ^, Though much of this scheme was originally proposed based on human patient-based observational trials, recent identification of the urate transporter genes, mouse knockout studies, and mapping of the transporter gene expression and protein localization has confirmed this complicated urate handling physiology of the proximal tubule in humans.

A number of members of the SLC22 family (containing OATs and OCTs) transport creatinine. These include OCT2, OAT2, OAT3, and possibly OAT1. ^{, ,} All of these are considered key transporters of organic cation drugs, organic anion drugs, and/or zwitterionic drugs. Their relative importance in renal creatinine handling is debated, but it is likely that all play some role in the renal handling of creatinine, with OCT2 being particularly important. The issue is clinically relevant because a number of drugs (e.g., trimethoprim) are thought to “artificially” create the impression of renal dysfunction when creatinine is the primary measure used because of drug–metabolite interactions at the level of the transporter.