INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

CHARACTERISTICALLY, urinary concentration decreases in response to a reduction of functioning renal mass (10, 18, 21, 42). Although a variety of factors have been implicated, the pathogenesis of the polyuria and impaired urinary concentration ability in chronic renal failure (CRF), especially the cellular and molecular defects, are poorly understood. Patients with CRF exhibit a high filtered load of solute per remaining functioning nephron (23). Part of the defect in urinary concentration may be attributed to this high solute load imposed on each nephron (18, 20). The increase in filtration rate imposes an increased reabsorptive burden on the residual nephrons, if water and electrolyte balance is to be preserved. In vivo studies of rats using the remnant kidney model (i.e., 5/6 nephrectomy) have suggested that despite an increase in absolute reabsorption, delivery of water to the end of the proximal tubule of superficial nephrons and to the bend of Henle's loop of juxtamedullary nephrons was increased (5). This was a consequence of an increase in filtered load and a decrease in fractional reabsorption. Disruption of medullary architecture due to interstitial fibrosis may also contribute to the defect in urinary concentration by preventing the generation of a hypertonic medullary interstitium (16).

The collecting duct system plays a central role in the process of concentrating urine. Isolated and perfused cortical collecting ducts from kidneys of uremic rabbits with CRF exhibited impaired responsiveness to vasopressin with regard to changes in water permeability and adenylate cyclase activity (10). Also, downregulation of the V₂ receptor was demonstrated by decreased V₂ receptor mRNA levels from inner medulla in CRF rats (43). It has also been demonstrated that fractional water reabsorption is significantly reduced in the terminal collecting duct of rats with remnant kidney (5). Consistent with this, Wilson and Sonnenberg (47) found that the amount of sodium and water reabsorbed along the medullary collecting duct was significantly depressed in uremic rats with CRF induced by 5/6 nephrectomy. These studies emphasize the presence of important defects in the collecting duct in kidneys of CRF rats.

The urinary concentrating defect in CRF is apparently not due to a reduced secretion of vasopressin from pituitary gland (24). In contrast, blood vasopressin levels were increased in patients with CRF (24). Moreover, increased vasopressin levels have also been demonstrated in rats with CRF induced by 5/6 nephrectomy (4), and this is in part thought to be due to a reduced clearance of vasopressin as a consequence of diminished renal function in CRF. In patients with advanced CRF, urine remained hypotonic to plasma despite the administration of supramaximal dose of vasopressin (21, 42). This vasopressin-resistant hyposthenuria is consistent with abnormalities in the cellular mechanisms involved in vasopressin regulation of the water permeability of the collecting duct in CRF.

The renal tubule is composed of a series of distinct segments that function together to permit the regulated excretion of water and solutes. The aquaporins are a family of membrane proteins that function as water-selective channels in many water-transporting tissues (1, 26, 31, 34). Aquaporin-1 (AQP1), a 28-kDa integral protein, is expressed in the proximal tubule and descending limb of Henle's loop, where it is proposed to play an important role in fluid absorption (1, 35, 36). The water permeability of the collecting duct is regulated by vasopressin, and vasopressin-regulated water transport across the apical plasma membrane of collecting duct cells is believed to be mediated by aquaporin-2 (AQP2) (34). Water reabsorption in the collecting duct is regulated both by short-term and long-term mechanisms, both of which have been shown to depend critically on AQP2. Short-term regulation occurs as a result of exocytotic insertion of AQP2 water channels into the apical plasma membrane in response to vasopressin (30, 32, 38, 50). In addition, long-term regulation of AQP2 protein levels in collecting duct cells has been shown to play a major role in the regulation of collecting duct water reabsorption (30). Water transport across the basolateral plasma membrane of collecting duct principal cells is thought to be mediated by aquaporin-3 (AQP3) (9, 22) and aquaporin-4 (AQP4).

Dysregulation of AQP2 has been shown to be involved in multiple hereditary and acquired water balance disorders (6, 12, 13, 14, 29, 37, 44, 49). We therefore examined whether there were changes of AQP2, AQP3, and AQP1 expression in remnant kidney of CRF rats to define cellular and molecular defects that may be involved in the urinary concentrating defect and polyuria in the early phase of CRF. In particular, we wished to investigate whether the previously described functional defects in collecting ducts of kidneys from CRF rats are associated with changes in AQP2 and AQP3 expression. Specifically, the purposes of present study are as follows: 1) to examine whether there are changes in the levels of AQP2 expression in the remaining nephrons that undergo both morphological and functional adaptation after loss of renal mass; 2) to examine the subcellular distribution of AQP2 in remaining collecting duct principal cells; 3) to examine whether changes of AQP2 expression can be correlated with functional changes in renal water excretion; and 4) to examine whether changes of AQP2 expression and urine concentration ability occur in CRF rats in response to exogenous 1-desamino-[8-D-arginine]vasopressin (DDAVP) administration. 5) In addition, we examined whether there were changes in the expression and subcellular localization of AQP1 and AQP3 in CRF.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Experimental animals. Studies were performed on 46 adult Munich-Wistar rats initially weighing 265 ± 8 g (Møllegard Breeding Centre). The rats were maintained on a standard rodent diet (Altromin, Lage, Germany) with free access to water.

Induction of CRF by surgical reduction of renal mass. Experimental CRF was induced by excision of ~2/3 of the left kidney and right total nephrectomy, using the so-called excision remnant kidney model (17). The protocols used in this study are depicted in Fig. 1. After a period of acclimation, the rats were randomly divided into three groups.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Diagram of the study design. Protocol 1: chronic renal failure (CRF) induced in rats by 2/3 nephrectomy followed by contralateral nephrectomy (n = 17). Protocol 2: sham-operated rats matching protocol 1 (n = 11). Protocol 3: CRF rats treated with 1-desamino-[8-D-arginine]vasopressin (DDAVP) using implantable osmotic minipumps for 7 days (n = 6). Protocol 4: sham-operated rats treated with normal saline using implantable osmotic minipumps for 7 days (n = 6). Protocol 5: normal rats with excision of both poles of left kidney (2/3 nephrectomy) (n = 6). Protocol 6: kidneys of rats with CRF (n = 4, protocol identical to the one described in protocol 1) and of sham-operated rats (n = 4, protocol identical to protocol 2) were perfusion fixed for immunocytochemistry (protocol not indicated in diagram). In CRF groups, two-thirds of the left kidney (the upper and lower poles) was excised, and 1 wk later the right total kidney was removed. Rats were maintained in metabolic cages at the days marked with asterisk, allowing monitoring of urine excretion rates. Urine volume, osmolality, creatinine, sodium, and potassium concentrations were measured. Plasma was collected at the time of right nephrectomy, and at the time of death, for measurement of sodium and potassium concentration, creatinine, and osmolality.

The rats were anesthetized with halothane (Christian Friis, Copenhagen, Denmark), and during surgery, the rats were placed on a heated table to maintain rectal temperature at 37-38°C. The left kidney was exposed through left flank incision, gently dissected free from the adrenal gland, and approximately two-thirds of the left kidney including the upper and lower pole was excised (consistently ~50% of the kidney mass was removed; see RESULTS). One week later, these rats were again anesthetized with halothane, and the right kidney was removed through right flank incision after dissecting it free from the adrenal gland. The wound was closed with 4-0 Mersilene and metal clamps. Immediately after each surgical procedure, buprenorphium, 0.2 mg/kg sc (Temgesic; Reckitt and Colman), was injected to relieve pain for 1 day, and during this time rats were allowed to recover from anesthesia and surgery in cages with free access to water and standard rat chow.

As a control group, rats were subjected to sham operations identical to the ones used for CRF rats, except that kidneys or poles of kidney were not removed. Sham-operated rats were monitored for 2 wk in parallel with 5/6 nephrectomized rats. All rats were killed under light halothane anesthesia, and kidneys were rapidly removed and processed for membrane fractionation and immunoblotting at the same day.

We have chosen 2 wk follow-up after induction of 5/6 nephrectomy to 1) produce severe chronic renal insufficiency, 2) to allow significant hypertrophy of the remnant kidney, and 3) to minimize interstitial fibrosis. Bonilla-Felix et al. (3) used 1 wk follow-up and found tubular enlargement but only very mild azotemia and no interstitial fibrosis. In contrast, Teitelbaum and McGuinness (43) followed rats for three or more weeks and found marked azotemia, marked hypertrophy with increased tubular diameter and significant interstitial fibrosis. Hence, we have chosen 2 wk of follow-up. The functional data revealed marked chronic renal insufficiency, and the light microscopy evaluation confirmed a mild fibrosis (see RESULTS).

Clearance studies. The rats were maintained in the metabolic cages, allowing quantitative urine collections and measurements of water intake (Fig. 1). Urine volume, osmolality, creatinine, sodium, and potassium concentration were measured. Plasma was collected from tail vein at the time of right nephrectomy, and from abdominal aorta at the time of death, for measurement of sodium and potassium concentration, creatinine, and osmolality.

Experimental protocol. The following protocols were followed (Fig. 1). For protocol 1, CRF was induced in rats by excision remnant kidney model (n = 17). For protocol 2, animals were sham-operated rats (n = 11). For protocol 3, CRF rats were treated with DDAVP (Sigma) using implantable osmotic minipumps (Alzet, Palo Alto, CA) for 7 days (n = 6). For protocol 4, sham-operated rats were treated with 0.9% NaCl solution using implantable osmotic minipumps for 7 days (n = 6). For protocol 5, in normal rats, both poles of left kidney were excised (2/3 nephrectomy, n = 6). In protocol 6, for immunocytochemistry, kidneys of rats with CRF (n = 4, protocol identical to the one described in protocol 1) and of sham-operated rats (n = 4, protocol identical to protocol 2) were perfusion fixed (see below).

Implantation of osmotic minipumps. To study whether changes of AQP2 expression and urine concentrating ability take place in CRF rats in response to exogenous DDAVP administration, osmotic minipumps were implanted in CRF (n = 6) and sham-operated (n = 6) rats (protocols 3 and 4; Fig. 1). For implantation, minipumps (model 2002; Alzet) were filled with DDAVP (Sigma V1005) in a carrier solution containing 5% dextrose and 0.05% acetic acid for CRF rats or 0.9% NaCl solution for sham-operated rats. The pumps were equilibrated with normal saline for 4 h before insertion (29). Rats were anesthetized with inhalation of halothane, and the minipump was inserted into the subcutaneous tissue under the skin of back. Under these conditions, the pumps delivered 0.1 µg/h sc of DDAVP to CRF rats or 0.5 µl/h of 0.9% NaCl solution to sham-operated rats for 7 days.

Membrane fractionation for immunoblotting. The kidneys were homogenized (0.3 M sucrose, 25 mM imidazole, 1 mM EDTA, pH 7.2, containing 8.5 µM leupeptin, 1 mM phenylmethylsulfonyl fluoride) using an Ultra-Turrax T8 homogenizer (IKA Labortechnik), at maximum speed for 10 s, and the homogenate was centrifuged in an Eppendorf centrifuge at 4,000 g for 15 min at 4°C to remove whole cells, nuclei, and mitochondria. The supernatant was then centrifuged at 200,000 g for 1 h to produce a pellet containing membrane fractions enriched for both plasma membranes and intracellular vesicles (29, 30). Gel samples (Laemmli sample buffer containing 2% SDS) were made of this pellet.

Electrophoresis and immunoblotting. Samples of membrane fractions from total kidney were run on 12% polyacrylamide minigels (Bio-Rad Mini Protean II). For each gel, an identical gel was run in parallel and subjected to Coomassie staining to assure identical loading (44). The other gel was subjected to immunoblotting. After transfer by electroelution to nitrocellulose membranes, blots were blocked with 5% milk in PBS-T (80 mM Na₂HPO₄, 20 mM NaH₂PO₄, 100 mM NaCl, and 0.1% Tween 20, pH 7.5) for 1 h and incubated with affinity-purified anti-AQP2 (40 ng/µl IgG) (30, 32, 33), or with AQP1 immune serum (diluted 1:2,000; generously provided by Dr. Peter Agre) (36), or with affinity-purified anti-AQP3 (0.5 µg/ml) (9). The labeling was visualized with horseradish peroxidase (HRP)-conjugated secondary antibody (P448, diluted 1:3,000; DAKO, Glostrup, Denmark) using enhanced chemiluminescence system (ECL, Amersham International).

Quantitation of total kidney AQP2, AQP3, and AQP1 expression. ECL films with bands within the linear range were scanned (30) using a Vista scanner and Corel Photopaint Software to control the scanner. For AQP2 and AQP1, both the 29-kDa and the 35- to 50-kDa bands (corresponding to nonglycosylated and the glycosylated species; Ref. 40) were scanned, as described (29, 44). For AQP3, the 27-kDa and the 33- to 40-kDa bands (corresponding to nonglycosylated and the glycosylated species) were scanned as described previously (9, 44). The labeling density was quantitated (29) of blots where samples from CRF kidneys were run on each gel with control kidneys from sham-operated animals. The labeling density was corrected by densitometry of Coomassie-stained gels. The total kidney expression of AQP in the experimental animals was calculated as a fraction of the mean sham control value for that gel after correction for the fraction that was loaded of the total kidney mass. The loading fraction of total kidney mass was calculated by using the loading volumes of the gel samples to get identical densities of the Coomassie-stained lanes and by correction for body weight and kidney weight as follows. Loading fraction of total kidney mass = [(loading volume of gel sample to get identical Coomassie staining)/(total volume of entire gel sample)] × (the volume fraction that the entire gel sample constituted out of total kidney homogenate used for preparation of gel sample) × [(total kidney wt per 100 g body wt corresponding to the left remnant kidney in CRF and the left plus right kidneys in sham-operated rats)/(number of kidneys being 1 in CRF and 2 in sham-operated rats)].¹

Values were presented in the text as means ± SE. Comparisons between groups were made by unpaired t-test. P < 0.05 was considered significant.

Preparation of tissue for immunocytochemistry. The kidneys from four CRF rats and four sham-operated rats were fixed by retrograde perfusion via the aorta with 0.2% glutaraldehyde plus 2% paraformaldehyde in 0.1 M sodium cacodylate buffer, pH 7.4, as previously described (7, 32, 33). Kidneys were postfixed for 1 h, and tissue blocks prepared from kidney inner medulla were infiltrated for 30 min with 2.3 M sucrose containing 2% paraformaldehyde, mounted on holders, and rapidly frozen in liquid nitrogen (36). For light microscopy, the frozen tissue blocks were cryosectioned (0.8-1 µm, Reichert Ultracut S cryoultramicrotome), then sections were incubated with affinity-purified anti-AQP2 antibodies (LL127AP or LL358AP), or anti-AQP3 antibody (LL178AP), or AQP1 immune serum, and the labeling was visualized with HRP-conjugated secondary antibody (P448 1:100, DAKO), followed by incubation with diaminobenzidine. For fluorescence microscopy, the label was visualized using goat-anti-rabbit IgG (Z0421, 1:50, DAKO) and FITC-conjugated rabbit-anti-goat antibody (F250, 1:50, DAKO).

Freeze substitution of kidney tissue. The frozen samples were freeze substituted in a Reichert Auto Freeze-substitution Unit (Reichert, Vienna, Austria) as described before (35). Briefly, the samples were sequentially equilibrated over 3 days in methanol containing 0.5% uranyl acetate at temperatures gradually increasing from 80°C to 70°C, then rinsed in pure methanol for 24 h while increasing the temperature from 70°C to 45°C. At 45°C, the samples were infiltrated with Lowicryl HM20 and methanol 1:1, 2:1 and, finally, pure Lowicryl HM20 before ultraviolet polymerization for 2 days at 45°C and 2 days at 0°C. For electron microscopy, immunolabeling was performed on ultrathin Lowicryl HM20 sections (60-80 nm), which were incubated overnight at 4°C with affinity-purified anti-AQP2 diluted in PBS with 0.1% BSA or 0.1% skimmed milk. The labeling was visualized with goat-anti-rabbit IgG conjugated to 10-nm colloidal gold particles (GAR.EM10; BioCell Research Laboratories, Cardiff, UK) diluted 1:50 in PBS with 0.1% BSA. The sections were stained with uranyl acetate and lead citrate before examination in Philips CM100 or Philips 208 electron microscopes.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Five-sixths nephrectomy is associated with renal hypertrophy and severe CRF. As shown in Table 1, the levels of plasma creatinine in CRF rats (135.7 ± 15.1 µmol/l, n = 17) were significantly higher (P < 0.05) than in sham-operated rats (33.9 ± 1.1 µmol/l, n = 11). Consistent with this, the creatinine clearance was significantly reduced in rats with CRF (0.54 ± 0.09 in CRF vs. 2.21 ± 0.25 ml/min in sham-operated rats; P < 0.05). Thus the CRF rats had significant chronic renal insufficiency.

Another known feature following 5/6 nephrectomy is renal hypertrophy of the remnant kidney (18). The weight of kidneys where both poles have been removed (i.e., corresponding to the initial 2/3 left nephrectomy) is only 0.18 ± 0.01 g/100 g body wt (protocol 5). Two weeks after induction of 5/6 nephrectomy, remnant kidneys in CRF rats were significantly hypertrophied and weighed 0.41 ± 0.02 g/100 g body wt compared with 0.31 ± 0.02 g/100 g body wt of left kidney of sham-operated rats (P < 0.05). Despite the hypertrophy, the total renal mass of the remnant kidney is only 64.6 ± 2.8% of total renal mass in sham-operated animals (P < 0.05). Light microscopy examination revealed extensive hypertrophy of a all nephron and collecting duct segments. Very mild interstitial fibrosis with few fibroblasts was only observed in the inner medulla but not in cortex and outer medulla of remnant kidneys at the light and electron microscopic level (not shown).

Urine production is increased with urinary concentration defect in rats with CRF. During the period of acclimation, daily urine production averaged 40 ± 2 µl · min¹ · kg¹ in rats with CRF (n = 17). This rose to 111 ± 3 µl · min¹ · kg¹ after induction of 5/6 nephrectomy (P < 0.05). There was a parallel increase in water intake, from 94 ± 4 to 155 ± 5 µl · min¹ · kg¹ per day (P < 0.05). In contrast, urine production and urine osmolalities were unchanged in sham-operated rats (n = 11). Thus the urine output of a single remnant kidney was approximately four times that of one kidney of sham-operated control rats. These data are summarized in Table 2 and Fig. 2A.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Time course of the changes in urine output (A) and urine osmolality (B) in CRF rats (n = 17, open bars) and sham-operated rats (n = 11, solid bars). Urine output significantly increased and urine osmolality significantly decreased after induction of 5/6 nephrectomy in CRF rats, whereas there were no changes in urine output and urine osmolality after sham operation in control rats. * P < 0.05.

This marked increase in urine output was associated with significant impairment of urinary concentrating ability in CRF rats: urine osmolality was 493.8 ± 15.6 in CRF vs. 1,325.9 ± 56.1 mosmol/kgH₂O in sham-operated rats (P < 0.05, Fig. 2B). Consistent with changes in urine output and urine osmolalities, the urine-to-plasma osmolality ratio was 1.6 ± 0.1 in CRF vs. 4.6 ± 0.4 in sham-operated control, and the solute-free water reabsorption (T^cH₂O) was 65 ± 12 in CRF vs. 141 ± 13 µl · min¹ · kg¹ in sham-operated rats, respectively (P < 0.05, Table 1). Even when T^cH₂O was corrected for the fractional osmolal clearance (C_osm/C_Cr, where C_osm is osmolal clearance and C_Cr is creatinine clearance), the value of T^cH₂O in rats with CRF was much lower (124 ± 16 vs. 258 ± 13 µl · min¹ · kg¹, P < 0.05). Thus this suggests that factors other than the increased solute load per nephron are important determinants of the impaired concentrating ability in CRF rats.

Total kidney AQP2 expression is decreased in rats with CRF. The affinity-purified anti-AQP2 antibody exclusively recognizes the 29-kDa and the 35- to 50-kDa band (Fig. 3), corresponding to nonglycosylated and glycosylated AQP2. As shown in Fig. 3, densitometric analysis of all samples from CRF and sham-operated rats revealed a marked decrease in total kidney AQP2 expression in CRF rats to 42.5 ± 11.6% (n = 13) of levels in sham-operated rats (100 ± 11%, n = 11, P < 0.05).² In Fig. 4, A and B, the AQP2 expression was compared with the changes in urine osmolality and solute-free water reabsorption determined 2 wk after induction of 5/6 nephrectomy. The significant correlations between the AQP2 expression and these parameters are consistent with the functional association among these parameters. Also, there was a significant inverse correlation between the AQP2 expression and the severity of azotemia measured by plasma creatinine levels (Fig. 4C).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Immunoblot of membrane fractions of total kidneys from CRF and sham-operated rats. A: immunoblot was reacted with affinity-purified anti-aquaporin-2 (anti-AQP2) and revealed 29-kDa and 35- to 50-kDa AQP2 bands, representing nonglycosylated and glycosylated forms of AQP2 (densities were 65.9 ± 10.2% in CRF rats and 100 ± 5% in sham-operated rats, respectively; P < 0.05). B: densitometric analysis of all samples from CRF and sham-operated rats (corrected according to loading and loading fraction of total kidney mass) reveals a marked decrease in AQP2 expression from 100 ± 10.7% in sham-operated controls (n = 11) to 42.5 ± 11.6% in CRF rats (n = 13). * P < 0.05.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Correlations of AQP2 expression in CRF and sham-operated rats with urine osmolality (A), solute-free water reabsorption (B), and plasma creatinine concentrations (C). Linear regression analysis revealed significant correlations between the AQP2 expression and urine osmolality or solute-free water reabsorption, consistent with a functional association among these parameters. Also, there was a significant inverse correlation between the AQP2 expression and the severity of renal failure as determined by increases in plasma creatinine levels.

Total kidney AQP3 and AQP1 expressions are decreased in rats with CRF. The expressions of two other water channels, AQP3 and AQP1, were also examined. This was done to test whether there was a selective downregulation of AQP2 in CRF or a more general decrease in water channel expression associated with the urinary concentration defect in polyuric CRF rats. As shown in Fig. 5, densitometric analysis of all samples from CRF and sham-operated rats revealed a marked decrease in total kidney AQP3 expression to 25.1 ± 10.5% in CRF rats (n = 13, P < 0.05) of levels in sham-operated rats (100 ± 10.1%, n = 11). Also, there was significant decrease in total kidney AQP1 expression in CRF rats; the expression was 22 ± 6% in rats with CRF (n = 13, P < 0.05; Fig. 6) compared with 100 ± 7.5% in sham-operated controls (n = 11).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   A: immunoblot of total kidney membrane fraction from CRF rats and sham-operated rats. Blot was reacted with affinity-purified anti-AQP3 and revealed 27-kDa and 33- to 40-kDa AQP3 bands, representing nonglycosylated and glycosylated forms of AQP3 (densities were 52.3 ± 17.1% in CRF rats and 100 ± 9.6% in sham-operated rats, respectively, P < 0.05). B: densitometric analysis of all samples from CRF and sham-operated rat kidneys (corrected according to loading and loading fraction of total kidney mass) revealed a marked decrease in AQP3 expression from 100 ± 10.1% in sham-operated controls (n = 11) to 25.1 ± 10.5% in CRF rats (n = 13). * P < 0.05.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   A: immunoblot of total kidney membrane fraction from CRF rats and sham-operated rats. Immunoblot was reacted with anti-AQP1 serum and revealed 29-kDa and 35- to 50-kDa AQP1 bands, representing nonglycosylated and glycosylated forms of AQP1 (densities were 46.2 ± 7.9% in CRF rats and 100 ± 6.9% in sham-operated rats, respectively, P < 0.05). B: densitometric analysis of all samples from CRF and sham-operated rats (corrected according to loading and loading fraction of total kidney mass) revealed a marked decrease in AQP1 expression from 100 ± 7.5% in sham-operated controls (n = 11) to 22.0 ± 6.1% in CRF rats (n = 13). * P < 0.05.

Total kidney AQP2 expression is reduced in CRF compared with immediate 2/3 nephrectomy (polectomy). To evaluate whether AQP2 expression in the remaining collecting ducts changes during the hypertrophic phase in the remnant kidneys, we compared the total kidney AQP2 levels in remnant kidneys after 2 wk of CRF (protocol 1) with the levels in kidneys immediately taken out after 2/3 nephrectomy in normal rats (protocol 5).

To determine how much of AQP1, AQP2, and AQP3 is removed during 2/3 nephrectomy (polectomy), membrane fractions of polectomized and sham kidneys were subjected to immunoblotting and densitometry for AQP1, AQP2, and AQP3. The results revealed that the density of AQP2 was unchanged between 2/3 nephrectomy (polectomy) and sham control kidneys (93 ± 8% vs. 100 ± 11%, respectively; not significant, n = 6 in each group). Similarly, there were no differences in densities of AQP1 and AQP3 (not shown). By correction for kidney weight³ and the fraction loaded out of total kidney (see METHODS), 51 ± 8% of AQP2 was left after 2/3 nephrectomy (polectomy). These results demonstrated that proportional amounts of AQP2 (and of AQP1 and AQP3) were removed in response to 2/3 nephrectomy (polectomy).

The kidneys that underwent 2/3 nephrectomy with both poles removed (the procedure is identical to the initial 2/3 nephrectomy of the CRF rats) therefore have a similar number of collecting ducts remaining, but the kidneys in CRF have been subjected to hypertrophy during the development of CRF. Densitometric analysis of samples from CRF rats (n = 7) revealed a marked decrease in total kidney AQP2 levels to 45 ± 17.5% (P < 0.05) compared with 100 ± 25.6% in the two kidneys with 2/3 nephrectomy (n = 6; Fig. 7, A and B). This suggests that there is a significant downregulation of AQP2 in the remaining collecting ducts in CRF. Densitometric analysis also showed a significant decrease (P < 0.05) in AQP3 expression (36.2 ± 8.8% in CRF vs. 100 ± 20.2% in 2/3 nephrectomized rats) and AQP1 (50.6 ± 13.7% in CRF vs. 100 ± 21.3% in 2/3 nephrectomized rats; Fig. 7, C and D).

View larger version (25K): [in this window] [in a new window]   Fig. 7.   A and C: immunoblots of total kidney membrane fraction from CRF rat kidneys and kidneys from rats with 2/3 nephrectomy, showing expression of AQP2 (A) and AQP1 (C). B and D: densitometric analysis of samples from CRF rats and rats with simple excision of both poles (corrected according to loading and loading fraction of total kidney mass) revealed significant decrease in AQP2 (B) and AQP1 (D) expression.

Long-term DDAVP treatment does not change the increased urine output or the reduced AQP2 expression associated with CRF. Previously, we have shown that in lithium-induced nephrogenic diabetes insipidus (NDI), DDAVP treatment for 7 days resulted in a marked reduction in urine output and a change in the subcellular localization of AQP2 but no changes in AQP2 expression (29). In vitro studies have indicated a resistance to vasopressin action in the collecting duct of CRF rat kidneys (10, 43). We therefore investigated the effects of long-term DDAVP treatment of CRF rats on urine production rates and AQP2 expression levels.

As shown in Fig. 8, the significantly increased urine output in CRF rats persisted despite 7 days treatment with high doses of DDAVP (0.1 µg/h sc) compared with sham-operated rats treated with 0.9% NaCl solution (0.5 µl/h sc) (96 ± 4 vs. 40 ± 2 µl · min¹ · kg¹, n = 6, P < 0.05; Fig. 8A). Consistent with this, there was no significant difference in urine output between CRF rats with (n = 6) or without (n = 17) treatment with DDAVP (96 ± 4 vs. 116 ± 4 µl · min¹ · kg¹, P > 0.05; Fig. 8B). In parallel, there was a maintained significant difference in urine osmolality between CRF rats treated with DDAVP and sham-operated rats treated with 0.9% NaCl solution (574.8 ± 19.7 vs. 1,534.0 ± 62.4 mosmol/kgH₂O, P < 0.05; Table 3) and no difference between CRF rats with or without DDAVP treatment (574.8 ± 19.7 vs. 532.1 ± 17.6 mosmol/kgH₂O, P > 0.05; Table 3). Thus there was a DDAVP-resistant polyuria in rats with CRF.

View larger version (30K): [in this window] [in a new window]   Fig. 8.   A: time course of changes in urine output in CRF rats (n = 6, open bars) treated with DDAVP (0.1 µg/h sc for 7 days) and sham-operated rats (n = 6, solid bars) treated with normal saline (0.5 µl/h sc for 7 days). Polyuria persisted despite high dose of DDAVP treatment in CRF rats. * P < 0.05 vs. sham-operated rats. B: time course of changes in urine output in CRF rats treated with (n = 6, open bars) or without (n = 17, solid bars) DDAVP. There was no significant reduction in urine output in CRF rats treated with DDAVP compared with nontreated CRF rats.

Next, we examined whether 7 days of DDAVP treatment induced any changes in AQP2 expression. Densitometric analysis revealed a persistent downregulation in total kidney AQP2 expression in DDAVP-treated CRF rats of 38.2 ± 2.7% (n = 6, P < 0.05; Fig. 9A) compared with 100.0 ± 21.6% (n = 6) in sham-operated controls. Consistent with this, there was no difference in AQP2 expression between CRF rats with or without DDAVP treatment. Also, there was persistent decrease in total kidney AQP3 expression with DDAVP-treated CRF rats compared with sham-operated rats [14.8 ± 4.0% in DDAVP-treated CRF rats (n = 6, P < 0.05) compared with 100.0 ± 23.3% in sham-operated controls (n = 6); Fig. 9B]. Thus long-term DDAVP treatment, known to increase AQP2 in normal rats and vasopressin-deficient Brattleboro rats, did not produce changes in AQP2 and AQP3 expression in CRF rats.

View larger version (28K): [in this window] [in a new window]   Fig. 9.   Immunoblots of total kidney membrane fraction from CRF rats treated with DDAVP, 0.1 µg/h sc, for 7 days (n = 6) and sham-operated rats treated with normal saline, 0.5 µl/h sc, for 7 days (n = 6). A: immunoblot was reacted with affinity-purified anti-AQP2. B: densitometric analysis (corrected according to loading and loading fraction of total kidney mass) revealed a persistent decrease in AQP2 expression from 100 ± 21.6% in sham-operated controls to 38.2 ± 2.7%. C: immunoblot was reacted with affinity-purified anti-AQP3. D: densitometry revealed a persistent decrease in AQP3 expression from 100 ± 23.3% in sham-operated controls to 14.8 ± 4.0%. * P < 0.05.

Apical localization of AQP2 in rats with CRF. Immunofluorescence microscopy showed strong AQP2 labeling was associated with the apical plasma membrane domains (arrows in Fig. 10, A and B) and with intracellular vesicles in the cytoplasm of sham-operated rats. In kidneys from rats with CRF, there was a marked reduction in the AQP2 staining of collecting duct principal cells, and the remaining labeling was mainly associated with apical plasma membrane domains of collecting duct principal cells (arrows, Fig. 10, C and D). The immunocytochemistry confirmed that there was an overall decrease in AQP2 expression in kidneys from CRF rats.

View larger version (98K): [in this window] [in a new window]   Fig. 10.   Immunofluorescent localization of AQP2 in 0.8-µm cryosections of kidney inner medulla. A and B: immunofluorescent localization of AQP2 in sham-operated rat kidney. Labeling was only seen in the principal cells of collecting ducts (arrows), with no labeling of intercalated cells (arrowhead). Labeling was associated with the apical plasma membrane domains and to intracellular vesicles in apical and basal parts of the cells. C and D: labeling of AQP2 was much weaker in CRF rat kidneys. AQP2 labeling was mainly associated with apical plasma membrane domains (arrows). There is a distinct increase in tubule diameter in CRF rats (e.g., compare A and C, which are of identical magnification). Exposure time was identical for A-D. Magnifications: ×470 for A and C, and ×940 for B and D.

Immunoelectron microscopy was performed on ultrathin Lowicryl HM20 sections from cryosubstituted kidney inner medulla tissue from CRF and sham-operated rats (Figs. 11 and 12). The immunogold labeling pattern confirmed the light microscopy observations that the AQP2 labeling was sparse in CRF rats. The labeling was predominantly associated with the apical plasma membrane of collecting duct principal cells (arrows, Fig. 12, A and B).

View larger version (100K): [in this window] [in a new window]   Fig. 11.   Immunoelectron microscopy localization of AQP2 in ultrathin Lowicryl HM20 sections of a kidney inner medullary collecting duct cell from sham-operated rats. Abundant AQP2 labeling was associated with the apical plasma membrane (arrows) and intracellular vesicles (arrowheads). Magnifications: ×52,000 for A, and ×70,000 for B.

View larger version (124K): [in this window] [in a new window]   Fig. 12.   Immunoelectron microscopy localization of AQP2 in ultrathin Lowicryl HM20 sections of kidney inner medullary collecting duct cells from CRF rats. Labeling was extremely sparse and was predominantly associated with apical plasma membrane. Magnifications: ×52,000 for A, and ×70,000 for B.

Immunocytochemistry confirms the downregulation of AQP1 and AQP3 in rats with CRF. Immunofluorescence microscopy showed very abundant labeling of AQP1 associated with apical and basolateral plasma membranes of proximal tubule and descending thin limbs of sham-operated rat kidneys (Fig. 13, A and C). In contrast, AQP1 labeling was substantially weaker in CRF rat kidneys, with proximal tubules displaying significantly reduced labeling of the brush border as well as of basolateral plasma membrane infoldings (Fig. 13B). Also, descending thin limbs exhibited significantly lower labeling in kidneys from CRF animals compared sham-operated controls (Fig. 13D).

View larger version (94K): [in this window] [in a new window]   Fig. 13.   Immunofluorescent localization of AQP1 in 0.8-µm cryosections. A: in sham-operated rat kidney, abundant labeling of AQP1 was associated with apical (arrow) and basolateral (arrowhead) plasma membranes of proximal tubular cells. B: in CRF rat kidneys, AQP1 labeling was weaker in proximal tubules, with significantly reduced labeling of the brush border (arrow) as well as of basolateral plasma membrane infoldings (arrowhead). C: in sham-operated rat kidney, abundant labeling of AQP1 was associated with apical and basolateral membranes of descending thin limb. D: in CRF rat kidney, AQP1 labeling was weaker in enlarged and hypertrophied descending thin limb cell. Magnification, ×970.

Abundant AQP3 labeling was associated with the basolateral plasma membrane of collecting duct principal cells in sham-operated rat kidneys (Fig. 14A). As seen with AQP2, AQP3 labeling was also reduced seen in kidneys of CRF rats (Fig. 14B, please note different exposure times). The microscopy also confirmed that the cells of proximal tubules, descending thin limbs, and collecting ducts undergo a marked hypertrophy with increased tubular diameter during CRF (Figs. 10, 13, and 14).

View larger version (72K): [in this window] [in a new window]   Fig. 14.   Immunofluorescent localization of AQP3 in 0.8-µm cryosections of inner medulla. A: in sham-operated rat kidney, abundant labeling of AQP3 was associated with basolateral membrane of collecting duct principal cells. B: in CRF rat kidneys, AQP3 labeling was associated both with basal and lateral plasma membranes of kidney inner medullary collecting duct principal cells (exposed at 2 times the exposure time used for A). Note the increased tubular diameter in CRF rats (e.g., compare A). Magnification, ×940.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The results presented here demonstrated a marked reduction in AQP2 and AQP3 expression levels in the collecting duct principal cells and of AQP1 expression in proximal tubule and descending limb in hypertrophied remnant kidneys of rats with CRF induced by 5/6 nephrectomy. The reduction in the expression of these water channel proteins was associated with the development of polyuria and urinary concentration defects in CRF. Furthermore, the reductions in aquaporin levels and of the polyuria in rats with CRF were not reversed despite 7 days treatment of high doses of DDAVP.

Expression of AQP2 is decreased in rats with CRF. Concentration of the urine requires 1) establishment and maintenance of a hypertonic medullary interstitium and 2) vasopressin-regulated water transport across the collecting duct epithelium for osmotic equilibration. Thus a defect in any of these mechanisms would be predicted to be associated with urinary concentrating defects. In the present study, we found a marked reduction in the expression of two collecting duct water channels, AQP2 and AQP3, indicating a defect in collecting duct water reabsorption. This is consistent with previous functional studies. In CRF, micropuncture and microcatheterization studies have indicated that the impaired urinary concentrating ability may, at least partly, be caused by impairment of vasopressin-stimulated water reabsorption in the collecting duct (5, 47). Consistent with this, patients with advanced CRF have urine that remains hypotonic to plasma despite the administration of supramaximal doses of vasopressin (21, 42). This vasopressin-resistant hyposthenuria specifically implies abnormalities in collecting duct water reabsorption in CRF patients. Fine et al. (10) observed that isolated and perfused cortical collecting ducts dissected from remnant kidneys of severely uremic rabbits (1-3 mo after surgical reduction of kidney mass) exhibited a defect in the response to vasopressin. This was demonstrated both as a decreased water flux (J_v) and decreased adenylate cyclase activity (10). Importantly, they also showed that the cAMP analog 8-bromo-cAMP failed to induce a normal hydrosmotic response in cortical collecting duct from remnant kidneys. As an extension of these observations, Teitelbaum and McGuinness (43) observed that RT-PCR of total RNA from the inner medulla of CRF rat kidneys revealed virtual absence of V₂ receptor mRNA. Thus these studies provide firm evidence for significant defects in the collecting duct. Our observations of reduced AQP2 and AQP3 expression may including demonstration of a marked reduction of AQP2 in the apical plasma membrane (i.e., the active site), at least partly, provide explanations at the molecular level for these functional defects.

Although some studies have also been presented that show little or no defects in the collecting ducts in the remnant kidney model (3), it should be noted that both the duration and severity of uremia were much less pronounced in these studies. Bonilla-Felix et al. (3) demonstrated that isolated and perfused cortical collecting duct dissected from remnant kidneys 1 wk after reduction in renal mass exhibited the same vasopressin-stimulated water permeability as tubules from normal or sham-operated rats. In their study, CRF rats only had very mild azotemia, corresponding to less than a twofold increase in serum creatinine compared with that of sham-operated rats. Moreover, these rats with mild CRF did not develop interstitial fibrosis and had normal serum electrolytes including serum calcium (3). In contrast, Fine et al. (10) reported a marked azotemia and interstitial fibrosis in the rats with severe renal failure, and the experimental conditions were comparable to those of Teitelbaum and McGuinness (43). In the present study, CRF rats developed mild interstitial fibrosis, extensive azotemia (plasma creatinine: 135.7 ± 15.1 vs. 33.9 ± 1.1 µmol/l of sham, P < 0.05), and marked reduction in creatinine clearance (0.54 ± 0.09 of CRF vs. 2.21 ± 0.25 ml/min of sham controls, P < 0.05). In the present study, the remnant kidneys also displayed marked hypertrophy and dilation of proximal and distal tubules (Figs. 10, 13, and 14). This is consistent with previous studies revealing a 50% increase in the mean internal diameter of cortical collecting ducts (10). Thus the functional data as well as the marked hypertrophy presented in this study supports the view that 2 wk after induction of 5/6 nephrectomy, rats have developed relatively severe CRF, the condition where significant functional defects has been demonstrated in the collecting ducts (10, 43). The observed reduction in AQP2 and AQP3 expression is therefore likely to play a significant role in this.

Previous morphometric studies have shown that there are some differences in the degree of hypertrophy of different tubule segments. Hayslett et al. (19) demonstrated that after unilateral nephrectomy, there was a compensatory increase of 15% in diameter and 35% in length of the proximal tubule, whereas the distal tubule increased only by 10% in diameter and 17% in length. Thus differences in hypertrophy of distinct tubule segments obviously contribute to changes in the fractional expression of channels and transporters. Other mechanisms (likely the most important ones for certain channels and transporters) may induce alterations in expression levels due to altered synthesis and/or turnover. The net result can be determined by expressing the total kidney levels as a fraction of total levels in kidneys of sham or 2/3 nephrectomized animals. This observed change in expression can be supported by expression at the cellular level as revealed by immunocytochemistry. Finally, the subcellular distribution can be determined by immunogold electron microscopy to disclose the levels in, e.g., the apical plasma membrane (being the active site). These combined data may then be related to functional data. In the present study, it has been shown that there was 1) a marked reduction in total kidney AQP2 expression, 2) a reduction in cellular levels of AQP2 labeling, and 3) a reduction in AQP2 immunogold labeling of the apical plasma membrane. Thus these data are consistent with the previous functional data showing a reduced water permeability of isolated perfused collecting ducts as well as a lower hydrosmotic response to vasopressin and cAMP (10).

Apical localization of AQP2 in rats with CRF. The collecting duct represents the final site for the control of water excretion into the urine. Water permeability of the collecting duct is tightly regulated, under the control of the antidiuretic hormone vasopressin, which causes a dramatic increase in collecting duct water permeability, allowing reabsorption of water from the tubular fluid down an osmotic gradient. Vasopressin binds to V₂ receptors present in the basolateral membrane of collecting duct principal cells. Acting through the GTP-binding protein G_s, the interaction of vasopressin with the V₂ receptor activates adenylyl cyclase, which accelerates the production of cAMP from ATP (27). Subsequently, cAMP binds to the regulatory subunit of protein kinase A, resulting in dissociation of the regulatory subunit from the catalytic subunit. This activates the catalytic subunit, which phosphorylates various proteins (27) including AQP2 (15). AQP2 is then translocated from intracellular vesicles to the plasma membrane (32, 38), thereby increasing the water permeability of the apical plasma membrane. When vasopressin is removed, water permeability returns to basal levels, reflecting endocytic retrieval of AQP2 water channels (32), which may subsequently be available for re-use (25).

As shown in Figs. 10 and 12, AQP2 was predominantly found in the apical plasma membrane domains (albeit at very low levels) of collecting duct principal cells in rats with CRF. This indicates that the signal transduction pathways involved in regulating AQP2 targeting are at least partly intact. The combination of reduced AQP2 expression and maintained targeting is similar to what is observed in the acquired NDI syndrome associated with postobstructive polyuria (12, 13) and DDAVP-treated rats with lithium-induced NDI (29).

Vasopressin-resistant polyuria and downregulation of AQP2 in CRF. The finding of a vasopressin-resistant downregulation of AQP2 (and AQP3) expression in CRF rats raises the following possibilities. 1) Vasopressin resistance could, in part, be due to potential downregulation of the V₂ receptor protein. This would be consistent with the reduced V₂ receptor mRNA levels observed by Teitelbaum and McGuinness (43). 2) Inefficiency of vasopressin to increase adenylate cyclase activity and intracellular cAMP levels (consistent with previous reports; Refs. 10 and 43) may also be due to potential defects in the signaling cascade distal to the V₂ receptor (e.g., adenylate cyclase and protein kinase A). These two possibilities may be relevant since a cAMP response element has been reported to be present in the AQP2 promoter (46). 3) Some other factors may override the cAMP-mediated control process. This may involve a "vasopressin-independent" signaling pathway, which has previously been suggested to exist (8, 29). Each of the above three possibilities could hypothetically be involved in the previously described "vasopressin escape"-induced downregulation of AQP2 (8). Thus the downregulation of AQP2 observed in rats with CRF (where blood vasopressin levels are known to be elevated and where AQP2 expression is unchanged in response to long-term DDAVP treatment) would be consistent with a vasopressin escape-like downregulation of AQP2 in CRF.

Expression of AQP3 is decreased in rats with CRF. AQP3 is localized in the basolateral plasma membrane collecting duct principal cells (9, 22). In contrast to the lack of evidence for short-term regulation of AQP3, recent studies demonstrated long-term regulation of AQP3 expression with a marked increase in AQP3 expression in the rat inner medulla in response to water restriction or vasopressin infusion (9, 44). Thus this finding raised the possibility that changes in collecting duct water permeability in CRF may be associated with changes in AQP3 expression in addition to the observed changes in AQP2 expression. As shown in Figs. 5 and 13, there was a 75% decrease in total kidney AQP3 levels in rats with CRF compared with sham-operated rats. Also, there was a persistent decrease in AQP3 expression despite long-term treatment with DDAVP. Thus it may be speculated that the reduction in AQP3 expression seen in rats with CRF may participate in the urinary concentrating defect.

Expression of AQP1 is decreased in rats with CRF. AQP1 is present in large amounts in both the apical and basolateral plasma membranes of proximal tubule and descending thin limb cells, where it comprises nearly 4% of the total membrane protein (36). In the present study, we demonstrated that total kidney AQP1 expression was significantly reduced in rats with CRF. In previous functional studies, fluid reabsorption has been measured in the proximal tubule in the remnant kidney model of CRF (2, 11, 41, 45, 48). Fluid reabsorption determined under free-flow micropuncture conditions (2, 41, 48) and in isolated perfused proximal tubule preparations (11, 45) demonstrated, with two exceptions (41, 48), that the absolute rate of fluid reabsorption in the proximal tubule is increased compared with controls in remnant kidney. Importantly, all studies uniformly showed that the fraction of the filtered fluid reabsorbed in the proximal tubule was significantly reduced (2, 11, 41, 45, 48). The contribution of the reduced levels of AQP1 for the reduced fractional water reabsorption is currently unknown. The potential role of a relatively reduced AQP1 expression for the reduced fractional water reabsorption remains to be established. However, recent evidence from AQP1 gene knockout mice (28) have demonstrated that there is an ~80% and 90% reduction in the osmotic water permeability in the proximal tubule and descending thin limb, respectively (J. Schnermann, C. L. Chou, M. Knepper, and A. S. Verkman, personal communication). This points to an important role of AQP1 in proximal tubular water reabsorption (as well as in the descending thin limb) and raises the possibility that downregulation of AQP1 levels may play a role in the previously observed reduction in fractional water reabsorption in the proximal nephron in CRF.

The markedly reduced expression of AQP1 may relate to differences in the proximal tubule ultrastructure established during hypertrophy. Salehmoghaddam et al. (39) demonstrated an asymmetric and disproportionate growth of the hypertrophied proximal tubule in remnant kidney. The basolateral membrane area increased extensively, whereas there was little change in total luminal surface area compared with control animals. AQP1 is abundant in both the apical and basolateral membranes; however, immunoperoxidase and immunoelectron microscopy have shown that the majority of AQP1 is associated with the apical plasma membrane (35). Thus a much lower increase in total area of the apical plasma membrane (where AQP1 is abundant) during hypertrophy of the proximal tubule in contrast to an increase in the basolateral membrane (as well as other tubular segments) may be one of the possible causes of the decreased levels of AQP1 in remnant kidneys of CRF rats. However, since AQP1 levels are very low in CRF, other not yet defined mechanisms must be involved.

Reduced levels of AQP1, AQP2, and AQP3 in the remnant kidney appear not to be a nonspecific consequence of reduced renal mass. In a parallel study (unpublished observations), we have examined whether the expression levels of other transporters were also reduced in remnant kidneys using identical protocols. Total kidney levels and densities of NHE-3, Na-K-ATPase, and Tamm-Horsfall protein was significantly decreased in CRF compared with sham controls, thus showing a similar patterns as seen for AQPs. In contrast, total kidney levels of the bumetanide-sensitive Na-K-2Cl cotransporter BSC-1 and the thiazide-sensitive NaCl cotransporter TSC were not decreased, and the densities were markedly increased, which was also confirmed by immunocytochemistry. This indicates that specific mechanisms are likely to underlie the changes in the expression of channels and transporters in remnant kidneys during development of severe CRF. The underlying mechanisms require further detailed studies.

Summary. In summary, there was marked reduction in the expression levels of AQP2, AQP3, and AQP1, which was associated with the urinary concentration defect in polyuric CRF rats. The reduction was resistant to long-term treatment with DDAVP, consistent with severe NDI. Since vasopressin levels are known to be elevated in rats with CRF induced by 5/6 nephrectomy (4), this suggests that there may be a vasopressin escape-induced downregulation of AQP2 in CRF rats. The observed downregulation of the collecting duct water channels AQP2, AQP3, and the proximal nephron water channel AQP1 may provide a molecular explanation for the urinary concentration defect associated with CRF. Further studies are warranted to test whether there are alterations in the expression of NaCl transporters in CRF. Since these are responsible for establishment and maintenance of the hypertonic medullary interstitium providing the driving force for water reabsorption in the collecting duct, this represents an important area for the future studies.