Urine concentration in the kidney is a complicated process involving many solutes, such as urea, Na⁺, K⁺, Cl⁻, etc. (Chou and Knepper, 1989). Of all these solutes, urea plays a key role by recycling in the renal medulla (Sands, 2002).

Figure 1 illustrates the intrarenal urea recycling (Lei et al., 2011). As urine flows along the collecting duct, water is absorbed through water channels (AQPs), resulting in a high urea concentration in the terminal IMCD. There, the urea is rapidly reabsorbed by UTs UT-A1 and UT-A3. This process is regulated by vasopressin and is important to maintain high urea concentration in inner medullary interstitium and to set up the intrarenal osmotic gradient. The reabsorbed urea enters ascending vasa recta (AVR) where the endothelium has micropores permeable to urea. Urea is transferred from AVR to DVR through UT-B in a countercurrent exchange that helps to preserve the urea concentration gradient in IM (Yang and Bankir, 2005). In addition, UT-A2 expressed in TDL also mediates urea transport from the lumen to the interstitium, where the urea eventually goes back to the IM through DVR. The entire urea recycling is carried out in concert by UT-A1, UT-A3 in terminal IMCD, UT-A2 in TDL, and UT-B in DVR, and enables the kidney to establish the hyperosmotic gradient in the IM.

The vital roles of UTs in intrarenal urea recycling and the urine concentrating mechanism have been confirmed by knockout mouse models. In UT-B null mice, the urea clearance rate is reduced by 25% while the urea concentration increases by 30% and urea concentration in urine decreases by 35%. The UT-B null mice demonstrate higher urine output than wild-type mice (Figure 2; Bankir et al., 2004; Yang and Bankir, 2005; Lei et al., 2011). Moreover, the urinary urea concentration and the osmolality are not elevated in response to acute urea load (Figure 3), indicating that urea cannot be efficiently concentrated and water cannot be preserved in the absence of the UT-B. UT-B deletion causes a urea-selective urinary concentrating defect (Yang et al., 2002; Yang and Bankir, 2005). This fact could prove very promising in developing diuretics that specifically inhibit UT-B activity. Of note, UT-B expressed on the erythrocyte membrane also significantly contributes to the urine concentrating mechanism, which will be discussed later.

Urea transporter-A2 null mice are not different from wild-type mice in urine concentrating ability under normal conditions. A reduction in urea accumulation in the IM was observed only when a low-protein diet was applied, which reduced urea production (Yang et al., 2002). UT-A2 was thought to mediate urea exit from AVR to TDL. However, comparing the phenotype between UT-B null mice and UT-A2 null mice, the previously proposed role of UT-A2 needs to be further assessed (Uchida et al., 2005). The UT-A2 and UT-B double knockout (UT-A2/B−/−) mouse is a good model for determining the role of UT-A2 by comparing UT-A2/B−/− and UT-B−/− (Lei et al., 2011). As shown in Figure 2, urine output, urine osmolality, and urea concentration in IM at both basal and water-deprived conditions in UT-A2/B−/− mice were between those in UT-B−/− mice and wild-type mice. The phenotype of UT-A2/B−/− mice after acute urea load was also between that of UT-B−/− mice and wild-type mice (Figure 3). These results show that UT-A2 and UT-B contribute to urea cycling in an opposing way, that is, UT-A2 mediates urea transport from the lumen to the interstitium. This hypothesis requires an outward-directed concentration gradient from TDL to the interstitium, which might be provided by active secretion to the pars recta. Further studies are needed to clarify this hypothesis.

Urea transporter-A1 and UT-A3 double knockout (UT-A1/3−/−) mice were generated by gene targeting strategy (Fenton et al., 2004). UT-A1/3−/− mice had significantly higher urine output and water intake than wild-type mice in high-protein and normal-protein feeding conditions but not with low-protein feeding. After an 18-h water deprivation, reduced urine output was observed in UT-A1/3−/− in low-protein fed mice but not in mice under high-protein and normal-protein conditions (Fenton and Knepper, 2007; Fenton, 2008). These findings suggest that the diuresis that occurs in UT-A1/3−/− mice is due to the high urea concentration accumulated in the lumen, which supports the hypothesis that UT-A1 and UT-A3 mediate the rapid urea transport from the lumen of the IMCD to the interstitium. Fenton has an excellent review discussing the UT-A1/A3 knockout mouse (Fenton and Knepper, 2007).

Erythrocyte membranes have high urea permeability, which is 1.1 ± 0.2 × 10⁻⁵ cm/s in human and 3.3 ± 0.4 × 10⁻⁵ cm/s in rodents (Sands, 2002). To maintain the cell osmotic stability, urea must transverse erythrocyte membranes rapidly (Macey and Yousef, 1988; Fenton et al., 2005) when erythrocytes pass through vasa recta from the cortex to IM. The high efficiency of UT-B expressed in erythrocytes prevents the cells from shrinking in the IM where urea concentration accumulation is 50- to 200-fold higher than that in peripheral blood (Sands, 2002), and subsequently swelling after the cells leave the hyperosmotic environment. However, the role of UT-B in erythrocyte survival is currently speculative. Theoretical calculations and in vivo studies show that only about 0.5% of the total cardiac output of erythrocytes pass through the vasa recta (Lieberthal et al., 1987). Moreover, erythrocyte survival is not affected by either increased or decreased urine concentrating ability (Liu et al., 2011).

Urea transporter-B in erythrocytes also plays an important role in the urine concentrating mechanism. Lieberthal et al.’s (1987) study shows that urine cannot be well concentrated in the isolated kidney perfused without erythrocytes. The concentrating ability is restored after adding erythrocytes to the perfusate. The presence of UT-B enables rapid urea intake and release into and out of erythrocytes in the IM as they circulate away from the IM. This process helps build the urea concentration gradient. In addition, the countercurrent exchange between AVR and DVR, which conserves urea in the kidney, also depends on rapid urea transport through erythrocytes.

Urea transporter-B transports a number of urea analogs, such as formamide, acetamide, thiourea, and methylurea, with various rates (Figure 4; Zhao et al., 2007). By competing with high affinity to UT-B, urea analogs decrease UT-B urea permeability and result in UT-B blockage (Goodman, 2002). Thiourea, one of the analogs of urea, was shown to have a 10-fold greater binding affinity to UT-B than urea (Zhao et al., 2007). Of note, erythrocytes have high water permeability mainly mediated by water channel AQP1 (Yang et al., 2001). UT-B is also permeable to water (Figure 5A; Yang and Verkman, 1998). The low water permeability found in AQP1/UT-B double knockout erythrocytes was also seen after HgCl₂ treatment of UT-B null erythrocytes or phloretin treatment of AQP1 null erythrocytes (Figure 5B; Yang and Verkman, 2002). The single channel water permeability of UT-B is 7.5 × 10⁻¹⁴ cm³/s, similar to that of AQP1 (Yang et al., 2001), which provides direct functional evidence for UT-B-facilitated water transport in erythrocytes and suggests that urea traverses an aqueous pore in the UT-B protein (Figure 5C; Yang and Verkman, 1998).

Three different sizes of UT-A proteins, 39, 51, and 56 kDa in rats and 51, 56, and 97 kDa in human, are found in heart when blotting the heart tissue with UT-A1 antibody (Duchesne et al., 2001). However, the type of UT-A isoforms is still undetermined. The 56-kDa protein is up-regulated in the uremia and hypertensive rat models. In human, both of 56 and 51 kDa protein expression are increased in heart failure caused by dilated cardiomyopathy.

Urea transporter-B is highly expressed in hearts. ECG recording shows that the UT-B knockout mouse has a prolonged P-R interval from early (6 week) to old ages (52 weeks), indicating that the conduction from atrium cordis to cardiac ventricle is delayed (Figures 6A,B). However, UT-B deletion results in increased type II and III atrial ventricular heart block attack only in the aged mice (>52 weeks). The significantly reduced APA and V_max demonstrate that the cardiocyte excitability and conductibility are impaired in UT-B null mice (Yu et al., 2009). The accumulated urea in cells might be involved in the progressive heart block in UT-B null mice (Meng et al., 2009). However, the causative role of urea accumulation in cardiac disease is still in debate (Scheuer and Stezoski, 1973). One in vitro study shows that urea at 3–300 mM had a protective effect to heart damage induced by ischemia and electrolysis (Wang et al., 1999). Other possible mechanisms may exist. Interestingly, TNNT2 (Troponin T Type 2) is significantly increased in UT-B null hearts (Hershberger et al., 2009). TNNT2 is an important factor in heart that acts as a Ca²⁺ sensor (Yu et al., 2009) and affects the contractility of myocardium (Sehnert et al., 2002). The increased TNNT2 might be involved in the development of heart block in UT-B null mice. Atrial natriuretic peptide (ANF) is increased in aged UT-B knockout mice (>52 weeks; Figure 6C; Meng et al., 2009). ANP is regarded as a hypertrophy protective factor (Oliver et al., 1997; Horio et al., 2000) and hypertrophy is related to the heart block (Moak et al., 2001).

Urea transporter-B is expressed in brain, mainly in astrocytes and ependymal cells. Brain contains a high concentration of urea, the same as seen in liver (Buniatian and Davtian, 1966). Intriguingly, UT-B is not detected in endothelial cells of blood vessels which form the blood-brain barrier (BBB). Indeed, the permeability of urea across BBB is relatively slow and reversible (Kleeman et al., 1962). Urea mediated by UT-B enters into the astrocytes, passes through the ependymal cells, and finally gets into the cerebrospinal fluid (CSF) to be excreted (Couriaud et al., 1996). In support of this, arginase II, the key enzyme for urea synthesis in brain, was detected to have a larger distribution than UT-B (Braissant et al., 1999).

In 5/6 nephrectomized uremic rat models, both UT-B mRNA and protein in brain was significantly reduced (Hu et al., 2000). In contrast, AQP4 and AQP9 were remarkably increased (Trinh-Trang-Tan et al., 2005). This might be the reason for the rapid development of brain edema in dialysis. The reduced UT-B in astrocytes impairs urea effusion and results in urea accumulation in brain. Subsequently, the elevated osmolarity drives water to flow into the brain through the water channels AQP4 and AQP9.

Urea transporter-B is expressed in brain (Figure 7A) and controls urea concentration in brain tissue (Figure 7B; Guo et al., 2007). UT-B may also play an important role in the regulation of nitric oxide (NO) production in brain. Urea has an inhibitive effect on l-arginine transport into the epithelial cells (Xiao et al., 2001) and phloretin can block this inhibition (Wagner et al., 2002). UT-B likely mediates urea transport and regulates intracellular l-arginine concentration. UT-B null mice present both depression-like behavior (Li et al., 2012) with an extended immobility time in forced swim test (Figure 7C) and less interest in sucrose in a sucrose preference test (Figure 7D). Decreased rCBF (cerebral blood flow) supports the hypothesis of “vascular depression” (Figure 7E). Histologically, swelling of unmyelinated fibers was observed in hippocampus. This change may reflect the degeneration of membrane structures and impaired axoplasmic transport which contribute to the depressed-like behavior (Figure 7F). Notably, the UT-B null mouse has decreased NO production in brain and increased urea concentration especially in hippocampus in both acute urea load tests and normal condition (Figure 7B). Acute urea load decreases nNOS expression (Li et al., 2012).

Urea transporter-A5 is reported in testis. However, Northern blot analysis revealed different UT-A mRNA transcripts at 4.0, 3.3, 2.8, and 1.7 kb in testis (Tsukaguchi et al., 1997), which are all longer than UT-A5. This suggests that other types of UT-A isoforms may also exist. UT-A proteins were found in the nuclei of Sertoli cells throughout development of the seminiferous tubules (SMT) as well as in the residual bodies at stage VIII of SMT (Fenton et al., 2002a). Considering that the Sertoli cell nuclei do not produce urea, the function of UT-A in the nuclei is unclear and may not be related to urea transport. The residual bodies are involved in the formation of the spermatozoa head while spermatids transform into spermatozoa. This process requires a reduction of cell volume. UT-A in residual bodies probably participates in reducing cell volume by rapidly excreting urea, while water channel AQP7 and AQP8 excrete fluid. In support of this, UT-A1 and UT-A3 double knockout mice have enlarged testis (Fenton et al., 2005). Interestingly, UT-A5 protein increases at day 20 after birth, the same time when SMT begins to produce fluid to drive sperm to reproductive tracts.

Urea transporter-B mRNA also expresses in Sertoli cells and increases on stage II–III of SMT development. In vitro study shows phloretin-inhibited urea diffusion across Sertoli cells and SMT cell membranes following the concentration gradient of urea (Guo et al., 2007). In UT-B knockout mice, urea was significantly accumulated in the testis (Figure 8A). After 5 min of injection, [¹⁴C] urea accumulation in testis was significantly less in UT-B null mice compared to the wild-type mice (Figure 8B). This suggests that urea transport across SMT epithelia and exit from Sertoli cells is decreased in the absence of UT-B (Guo et al., 2007). UT-B null mouse has a larger testis and earlier breeding time (Figures 8C,D).

Urea transporter-B is identified in urinary tract epithelial cells both in the cell membrane and some cytoplasmic vesicles (Lucien et al., 2005). Diets with different contents of protein or sodium do not affect the expression of UT-B in bladder and ureter. However, hydropenia causes a significant increase of UT-B expression in ureter, but not bladder. Considering that bladder and ureter have higher urea and creatinine concentration than those in plasma, the reabsorption of urea and creatinine from luminal urine to urinary tissue may exist (Spector et al., 2007). The high concentration of urea may diffuse to the epithelial cells where the urea is then transported across the basolateral membrane to the muscular layer and enter into the capillaries by UT-B UT. In terms of this, UT-B may be involved in the regulation of epithelial cell volume and osmolality (Lucien et al., 2005). Both UT-A mRNA and protein are found in rat bladders (Doran et al., 2006), suggesting that UT-A might participate in the transportation of urea from urine to the apical membranes of epithelial cells.

Interestingly, new studies reported that UT-B might be associated with urinary bladder cancer (UBC; Garcia-Closas et al., 2011; Rafnar et al., 2011). A novel UBC susceptibility locus is found on chromosome 18q12.3 where SLC14A1 localizes (Garcia-Closas et al., 2011; Rafnar et al., 2011). Urine volume and frequency of urination, which is partly regulated by UT-B, may affect the risk of UBC occurrence.

Liver is the major organ for urea synthesis. There are two different sizes of UT-A protein identified in liver and isolated hepatocytes from rats, mice, and chimpanzees (Klein et al., 1999). In uremic rats, the 49-kDa UT-A protein mainly expressed in the cell membrane was significantly increased, while the 36-kDa cytosolic UT-A protein did not show any change (Klein et al., 1999). UT-B is also reported to be expressed in liver (Sands, 2000). How the UT mediates urea transport in liver is not clear.

Urea transporter-B is expressed at the apical membrane of superficial epithelial cells and crypts in colon and small intestine. However, the expression of UT-A in intestine is still controversial (Inoue et al., 2005). Colon is believed to be permeable to urea (Moran and Jackson, 1990). Intraluminal urea infiltrated from blood is partly decomposed by intestinal flora into ammonia and carbon dioxide. Ammonia is then reabsorbed back into blood to participate in further metabolic reactions, which accomplishes nitrogen reutilization (Spector et al., 2004). Intestinal UT-B protein abundance is increased in rats when fed a low-protein diet or a 20% urea diet, but there is no change in UT-B abundance in animals fed a high-protein diet or in rats with uremia (Inoue et al., 2005). Intestinal UT-B mediates the urea movement across intestinal epithelial cells according to transepithelial urea flux assays in vitro (Spector et al., 2004).

Urea plays an important role in the urinary concentration mechanism. Development of a UT inhibitor may have clinical significance in treating water overloaded diseases. Phloretin is a commonly used UT-B inhibitor experimentally, but its broad inhibitory activity, including inhibition transport of glucose (Lefevre and Marshall, 1959) and glycerol (vom Dahl and Haussinger, 1997), impedes its usage as a UT-B inhibitor. Urea analogs, such as methylurea, thiourea, and dimethylurea, competitively inhibit UT-B urea transport, usually at high concentrations (Zhao et al., 2007). The Verkman group has published some potentially promising reagents (Levin et al., 2007). By high-throughput screening, they identified some UT-B inhibitors and successfully established an erythrocyte osmotic lysis assay to evaluate these compounds. Human RBCs were preloaded with urea or urea analogs (acetamide), then the cells were changed to an isosmotic solution without urea. Under normal conditions, urea exits the cells through UT-B. However, in the presence of UT-B inhibitors, urea accumulates in the cells. Water enters the cells through AQP1 resulting in swelling and cell lysis. This could be measured by near-infrared light scattering. Of note, UT-B not only transports urea, but also water in RBC (Figure 5), but in vitro tests show that UT-B only participates in 8% of the total transport of water, while 90% of water is transported by AQP1. The decrease in water permeability resulting from the inhibition of UT-B did not significantly affect this erythrocyte osmotic model. After optimizing chemically modified compounds, Verkman et al. reported two inhibitors blocking UT-B but not UT-A or AQP activity. Considering the crucial role of UT-B in urine concentration, this discovery may lead to the development of a new type of diuretic capable of inhibiting urea absorption without changing other solutes.