UT-A urea transporters are expressed in numerous extra-renal tissues, including the heart, liver, brain and testis [7]; however their role in these tissues remains unclear. Therefore, we performed whole animal necroscopy and pathological analysis of 33 different organs and tissues to determine if deletion of urea transporters resulted in a non-renal phenotype. Tissues analyzed: Kidneys, spleen, skeletal muscle, liver, testis, reproductive tract, brain, GI tract (all levels), lungs, heart, thymus, pancreas, sciatic nerve, trachea, spinal cord, esophagus, femur, teeth, nasal sinuses, gall bladder, ears, urinary bladder, bone marrow, eyes, tongue, skin, parathyroid glands, Harderian glands, pituitary gland, thyroid glands, salivary glands, adrenal glands and lymph nodes. For analysis, 3 male wildtype mice and 3 male knockout mice were compared. Animals were euthanized and for some organs (see Table 1) wet weight was measured. All other tissues were emersion fixed in formaldehyde, embedded in paraffin and histopathologic diagnosis was performed. Kidney images were captured using a Leica MZ FL III Fluorescence Stereomicroscope.

Measurements of renal blood flow (RBF) were performed in male UT-A1/3^−/− mice (10–12 weeks old). Mice were anesthetized with inactin (100 mg/kg) and ketamine (100 mg/kg). The trachea was cannulated. A cannula was placed in the jugular vein for substance injection and the infusion of isotonic saline at a rate of 0.35 ml/h. Blood pressure monitoring was performed using the left femoral artery. After a flank incision, the left renal artery was carefully dissected free to permit placement of a 0.5PSB nanoprobe connected to a T402-PB flowmeter (Transonic Systems Inc.). The probe was held in place with a micromanipulator. The flow signal was digitized and analyzed using PowerLab software (ADInstruments). RBF was determined for 10 minutes, and values represent the 10-minute average. Following baseline measurements, the response of RBF to intravenous bolus injections of NG-Nitro-L-arginine Methyl Ester Hydrochloride L-NAME (1μg/g BW) was assessed.

As an index of total kidney nitric oxide production, we determined the levels of both nitrate and nitrite in urine. A colorimetric assay (Cayman Chemical) was used according to the manufacturer’s guidelines.

Serum vasopressin levels were measured by radioimmunoassay (Alpco Diagnostics). Mice were euthanized by decapitation, the blood collected and serum processed as described above. Since the assay required 1ml of serum, and the amount of serum collected from each mouse by decapitation was approximately 250μls, four mice were used for one basal measurement (12 total animals per group, n=3).

Littermate wildtype and UT-A1/3^−/− male mice from heterozygous crosses were used for all studies. Animals were maintained in mouse metabolic cages (Hatteras Instruments) for the duration of the study, under controlled temperature and light conditions (12 hour light and dark cycles). Several experimental manipulations were performed as follows:

Mice received free access to pelleted diet containing 4, 20 or 40% protein by weight (as casein) for 7 days before metabolic cage studies (the starch and sucrose content of the diet was altered in inverse proportion of casein in order to ensure an equivalent calorie intake). Subsequently, mice received a fixed daily ration of 5g of gelled diet per 20g body weight (BW) per day, also with either 4, 20 or 40% protein. The gelled diet was made up of 1ml of deionized water, 4g of special low NaCl synthetic food (0.001% Na w/w, Research Diets), 0.2 mmole NaCl and 25mg agar. Pre-weighed drinking water was provided ad libitum during the initial period of the study. After three days of adaptation to the cages, urine was collected under mineral oil in pre-weighed collection vials for successive 24hr periods. Urine volume was measured gravimetrically assuming a density of one. After the initial collection period, each mouse received a fixed daily ration of 5.7g of gelled diet per 20g BW per day for 24hrs containing 1.7mls of deionized water. Mice did not have access to supplemental drinking water during this period. Urine was collected under mineral oil for 24hrs. Blood was collected as detailed below and kidneys were processed as detailed in Solute content of kidney.

Mice were anaesthetized with isoflurane and blood was collected using retro-orbital eye bleeding. Serum was separated from whole blood using centrifugation at 2,000g and StatSampler collection tubes (StatSpin Inc). Sodium, potassium, chloride, creatinine, urea, glucose, calcium, albumin, magnesium, phosphorus, total protein and uric acid levels in serum were determined using an autoanalyser. Urine samples were centrifuged at 14,000g for 5 min and sodium, potassium, chloride, creatinine and urea levels were determined using an autoanalyser. Serum and urine osmolalities were determined using a vapor pressure osmometer (Wescor).

Kidneys were removed and rapidly dissected into cortex, outer-medulla, inner medulla base (IM1) and papilla (IM2/3) as shown in Figure 9. Segments were blotted on filter paper and placed into pre-weighed microfuge tubes. The entire procedure for dissection and weighing the tissue is estimated to have been less than 3min per kidney. Samples were weighed immediately and subsequently dried over a desiccant at 60° C for 8hrs (after which weight remained constant). After re-weighing, 25μls of deionized, distilled water was added to each tube, tubes were capped, placed at 90° C for 3mins and, after brief centrifugation, stored at 4° C for 24hrs. After centrifugation for 1min at 8,000g the supernatant was analyzed as follows; sodium and potassium concentrations were determined in 5μl samples (diluted in 5mls of 100ppm lithium solution) using a flame spectrophotometer (Model 2655-10, Sherwood Scientific), osmolality was determined in a 10μl sample using a vapor pressure osmometer (Wescor) and urea concentration was measured using a commercially available assay (Biotron Diagnostics). To determine initial solute concentrations, calculations were performed as detailed in the work of Schmidt-Nielsen et al. [8].

GFR measurements were made in conscious mice using FITC-inulin clearance and a modified protocol based on the work of Qi et al. [9]. Mice received free access to pelleted diet (Research diets) containing either 4 or 40% protein by weight (as casein) for 7 days before metabolic cage studies. Mice were housed individually in metabolic cages for three days (as detailed above) before surgery and received gelled food. On the fourth day mice were anaesthetized with isoflurane and an osmotic minipump (Model 2001, Alzet), containing ~3% FITC-inulin, was implanted subcutaneously. Inulin solution was prepared exactly as detailed in [9]. Mice were returned to metabolic cages and continued to receive free access to water and gelled food. On day 6 post-surgery, urine was collected under mineral oil for 24hrs. During urine collection, the collection vessel and metabolic cage base were covered with aluminum foil to minimize exposure of the urine to light. Cages were washed twice with 5mls of 500mM HEPES (pH 7.4) to collect residual fluorescence. Serum was collected by retro-orbital eye bleeds as described above. Fluorescence was measured in 10μl samples as previously described [9] using a Victor-3 1420 Multi-label Counter (Wallac) and the GFR estimated by the 24hr urinary FITC-inulin excretion rate i.e. urinary fluorescence counts per 24hr divided by the concentration of plasma FITC-inulin.

In order to assess whether altered expression of other transporters involved in salt balance or urine concentration occurred in UT-A1/3^−/− mice, animals were housed in metabolic cages as detailed above and received a 20% protein diet throughout the study. After 3 days adaptation, half the animals studied were switched to a water restricted diet, containing 1.7 ml water per day, for 36hrs and did not have access to supplemental drinking water during this period. Mice were killed by decapitation and the kidneys processed for immunoblotting.

Measurements of total urinary nitrate and nitrite (NOx) excretion rates under basal conditions on a 20% protein diet (24 hour urine collections) were performed to determine if the blood congestion observed in UT-A1/3^−/− mice kidneys may be related to abnormal nitric oxide (NO) production by the kidney. In knockout mice, NOx excretion (127.0 ± 7.8 nmol/gBW/day) was more than double that in wildtype controls (52.5 ± 9.5 nmol/gBW/day) (Figure 2).

Another factor that could alter renal blood flow dynamics is vasopressin. Consequently, we measured plasma vasopressin levels in UT-A1/3^−/− and wildtype mice. Under basal conditions (free access to water) on a 20% protein diet, we observed no difference in plasma vasopressin levels between the groups (Figure 3).

Since greater blood congestion was observed in the kidneys of UT-A1/3^−/− mice, we carried out measurements of total renal blood flow in anesthetized mice (Figure 4) using an ultrasonic flow probe. Total renal blood flow in the UT-A1/3^−/− mice was not significantly different from that observed in age-matched wildtype control mice. Inhibition of NO production by infusion of L-NAME resulted in a significant decrease in total renal blood flow in both knockout mice and controls. However, the percentage decrease due to L-NAME was significantly greater in UT-A1/3^−/− mice (39.2 ± 2.4% versus 27.1 ± 4.9%), suggesting a greater role for NO in the maintenance of blood flow.

We determined GFR in conscious UT-A1/3^−/− and wildtype mice on two levels of protein intake, 4% and 40%, using FITC-inulin clearance (Figure 5). Increasing the protein content of the diet more than doubled the FITC-inulin clearance in both UT-A1/3^−/− mice and wildtype controls (Figure 5 and Table 2). However, no significant differences were observed in inulin clearance between UT-A1/3^−/− and wildtype mice under either dietary condition, even when corrected for bodyweight (Figure 5 and Table 2).

UT-A1/3^−/− mice have a urinary concentrating defect that is dependent on the level of urea excretion [4]. Since the main determinant of urea excretion rate is protein intake, in the present study, the effects of three different dietary protein intakes on urinary concentrating ability were determined. Initial observations were made without restriction of water intake (Figure 6). On a low protein intake (4% protein diet), there were no significant differences in fluid consumption, urine flow or urine osmolality between wildtype and UT-A1/3^−/− mice. However, on a normal protein intake (20% protein diet), UT-A1/3^−/− mice exhibited significantly greater fluid consumption and urine flow than wildtype mice, resulting in a decreased urine osmolality. This decrease in urinary concentrating ability was even greater on a high protein intake (40% protein diet). Furthermore, after an 18hr water restriction (2ml of water per day per 20gBW), UT-A1/3^−/− mice on either a 20% or 40% protein diet (but not a 4% protein diet) were unable to reduce their urine flow and could not raise their maximal urinary osmolality above that observed under basal conditions (data not shown). During this 18hr water restriction, the body weight of UT-A1/3^−/− mice on a 20% or 40% protein diet decreased by 18.2 ± 0.4% and 24.6% ± 0.3%, respectively. In contrast, UT-A1/3^−/− mice on a 4% protein diet were able to maintain fluid balance without a marked loss of body weight (3.0 ± 0.2%).

It is well recognized the renal NO production plays an important role in the regulation of water transport in the renal collecting duct, although it appears to have both inhibitory [17] and stimulatory [18] effects on collecting duct water permeability. Furthermore, NO plays an important role in regulation of thick ascending limb NaCl transport, a process critical to the urinary concentrating mechanism. Hence, it should not seem surprising that changes in NO excretion were found in the UT-A1/3^−/− mice in association with the markedly abnormal water conservation capacity of these animals. However, these studies do not directly determine the cause or the consequences of the increased NO excretion in these mice, or of the associated medullary vascular congestion.

It has been known for several years that diets rich in protein increase whole kidney GFR [19, 20]. Studies by Seney and Wright, determined that this increase in GFR results from changes in the tubuloglomerular feedback (TGF) system [21, 22]. Furthermore, they determined that it is not an actual change in the sensing mechanism of the TGF response that is affected by high protein diet, but rather it is an effect on the signal causing the TGF response. They concluded that this diminished TGF response is due, at least in part, to a reduced early distal NaCl concentration, without a change in early distal tubule osmolality [22]. Bankir et al. [23, 24] have proposed that the reduction in early distal NaCl concentration is due to increased concentrations of urea consequent to the high protein intake. The urea concentrations in the late thick ascending limb and early distal tubule are dependent both on the urea concentration of the glomerular filtrate and the extent of urea recycling, a result of passive urea secretion into the loop of Henle from urea reabsorption in the IMCD [25–27]. Since urea recycling is likely to be virtually eliminated in the UT-A1/3−/− mice, it would be predicted that the increase in GFR in response to high protein feeding would be markedly attenuated. However, in our studies, a large protein-dependent increase in GFR was observed in both UT-A1/3^−/− and wildtype mice and we observed no significant difference in inulin clearance between the groups under either dietary condition examined. Therefore, these data suggest that urea reabsorption from the IMCD and the process of urea recycling is not an important determinant of the protein-induced increases in GFR observed.

The excretion of urea, in classical thinking, is thought to depend on two factors: the filtered load of urea and the amount of urea reabsorption that occurs along the nephron. Although secretion of urea into the loop of Henle occurs as part of the urea recycling process discussed in the previous paragraph, this secretion is thought to occur passively. However, several pieces of evidence support the notion that active urea secretion may occur at some point along the renal tubule (reviewed extensively in [28]). Evidence for active urea secretion in rodents was initially provided by Bodil Schmidt-Nielsen [29] and microperfusion studies by Kawamura and Kokko indicated that a low rate of active urea secretion could occur in the rabbit proximal tubule [30], although net urea secretion was not detectable in another study of the rabbit proximal straight tubule [31]. More recently, Kato and Sands have shown that in rats with a low dietary protein intake, urea can be actively secreted in the terminal IMCD [32]. In our studies, the FE_urea in UT-A1/3^−/− mice on a high protein diet was virtually 100% (Table 2). Conclusive evidence for active urea secretion from clearance studies would require a net fractional urea excretion significantly greater than 100%. Although this benchmark was not achieved, given that at least 30–40% of the filtered load of urea is normally reabsorbed in the proximal tubule [26, 33], the finding of a net fractional excretion of 102.0 ± 8.8 percent (Table 2) strongly suggests the presence of active urea secretion in the mouse renal tubule. To determine where the postulated active urea secretion occurs in the mouse kidney nephron will require micropuncture studies.

In this study, we carried out an analysis of solute concentrations in cortex, outer medulla and two levels of the inner medulla to assess the separate effects of changes in dietary protein intake and/or deletion of the collecting duct urea transporters on corticomedullary solute gradients. Several striking observations were apparent. First, in wildtype mice, a change in dietary protein intake from 4% to 40% resulted in increased tissue osmolality that was due solely to greater urea accumulation in the inner medulla (Figure 10), the site of expression of the UT-A1 and UT-A3 transporters [37, 38]. Sodium concentrations at all levels of the corticomedullary axis were unaffected by the change in dietary protein intake. Second, in agreement with studies in rats by Schmidt-Nielson et al [39], the concentration of urea in the papillary tip of wildtype mice fed a low protein diet was equivalent to the concentration of urea in the urine, whereas in mice fed a high protein diet, urea concentration in the urine is greater than in the papillary tip, consistent with a failure of urea to completely equilibrate between lumen and interstitium. Third, in contrast to wildtype mice, in knockout mice there was a substantially attenuated corticomedullary osmolality gradient and there was no urea gradient on either diet. However, the corticomedullary sodium gradients were virtually equivalent in wildtype and knockout mice on either level of dietary protein intake. Thus, neither marked medullary urea depletion due to dietary protein restriction, or marked medullary urea depletion due to deletion of collecting duct urea transporters affected the ability of the kidney to form a corticomedullary sodium gradient. Therefore, in contrast to predictions of the passive concentrating hypothesis proposed by Stephenson [5] and by Kokko and Rector [Kokko, 1972 #474, we conclude that NaCl accumulation in the IM is not dependent on either IMCD urea transport or the accumulation of urea in the IMCD interstitium.