Altered Expression Profile of Transporters in the IMCD of Aquaporin 1 Knockout Mice

Aquaporin-1 is the major protein responsible for transport of water across the epithelia of the proximal tubule and thin descending limbs. Rapid water efflux across the thin descending limb is required for the normal function of the countercurrent multiplier mechanism. Therefore urinary concentrating capacity is severely impaired in aquaporin-1 knockout (AQP1 -/-) mice. Here, we have investigated the long-term consequences of deletion of the aquaporin-1 gene product by profiling abundance changes in transporters expressed in the inner medullas of AQP1(-/-) mice versus heterozygotes [AQP1(+/-)], which have a normal concentrating capacity. Semiquantitative immunoblotting demonstrated marked suppression of two proteins strongly expressed in the inner medullary collecting duct (IMCD): UT-A1 (a urea transporter) and AQP4 (a basolateral water channel). Furthermore, the urea permeability of the IMCD was significantly reduced in AQP1(-/-) mice. In contrast, there was increased expression of three proteins normally expressed at higher levels in the cortical collecting duct (CCD) than in IMCD: AQP3 (another basolateral water channel) and the epithelial sodium channel subunits β-ENaC and γENaC. Changes in expression of these proteins were confirmed by immunocytochemistry. Messenger RNA profiling (real-time RT-PCR) revealed changes in UT-A1, β-ENaC, γ−ENaC, and AQP3 transcript abundance that paralleled the changes in protein abundance. Thus, from the perspective of transport proteins, the IMCDs of AQP1 (-/-) mice have a significantly altered phenotype. To address whether these changes are specific to AQP1 (-/-) mice we profiled IMCD transporter expression in a second knockout model manifesting a concentrating defect, that of ClC-nK1, a chloride channel in the ascending thin limb important for urinary concentration. As in the AQP1 knockout mice, ClC-nK1 (-/-) mice showed decreased expression of UT-A1 and increased expression of β-ENaC and γ−ENaC vs. WT controls. In conclusion, the expression profile of IMCD transporters is markedly altered in AQP1 -/mice and this manifestation is related to the associated concentrating defect. Index words: AQP 1 ClC-nK1 Inner medulla INTRODUCTION The renal collecting duct (CD) exhibits considerable physiological and histological heterogeneity along its length. The solute transport process and cellular composition of the CD change as the segment descends from the renal cortex, through the hypertonic medullary interstitium, down to the papillary tip. The phenotypic heterogeneity along the CD is readily appreciated by contrasting the cortical CD (CCD) with the inner medullary CD (IMCD). The principal cells of the rat, rabbit, and mouse CCD and IMCD have distinct transport characteristics that are a reflection of the differential expression of transporter proteins in these segments. For example, the CCD is well known as a site of active sodium reabsorption; a process dependent on the epithelial sodium channel, ENaC (12). In contrast, direct studies in isolated perfused tubules from rats have detected little or no active sodium transport in the IMCD [unpublished observations], which may be related to the much lower expression of ENaC protein in this region (10; 10). Additionally, the IMCD of the rabbit, rat and mouse have significant urea permeability (Purea) that is absent in the CCD owing the restricted expression of specific urea transporters(21) (7). Finally, there is differential expression of the basolateral water channels, in the rat CD with AQP3 being more abundant in the CCD while AQP4 is the predominant basolateral water channel of the IMCD (24). In summary, although the entire CD is derived from a common embryological precursor there is significant phenotypic diversity in this segment of the renal tubule resulting from a differential expression of solute transporters. Here, we have addressed the role urinary concentrating ability may play in determining the expression pattern of IMCD transporters using aquaporin 1 (AQP1) null mice as well as another mouse line in which the major chloride channel of the thin ascending limb (ClC-nK1) has been deleted (16). AQP1 water channels and ClC-nK1 chloride channels are normally expressed in renal structures that participate in the formation of concentrated urine and are integral components of the counter-current multiplier mechanism. The genetic ablation of either of these transporters results in a severe urinary concentrating defect (13),(16). We hypothesize that the special characteristics of the IMCD (reviewed above) are to some extent dependent on the high tissue osmolality in the inner medulla. To test this, we have assessed the expression of several solute and water transport proteins in IMCD in AQP1 and ClC-nK1 knockout mice. METHODS Animals. AQP 1(-/-) breeder mice (13) in a CD1 background were kindly supplied by Dr. Alan Verkman (University of California, San Francisco). A colony of CLC-nK1 (-/-) mice (16) was also established at NIH utilizing breeding pairs provided by Dr. Shinichi Uchida (Tokyo, Japan). All mice utilized in these studies were maintained on an ad libitum diet of normal rodent chow and water. For each experiment, mice of the appropriate genotype were selected from two separate litters of approximately the same age. Semiquantitative immunoblotting. Immunoblotting procedures for comparing two sets of samples of kidney homogenates with regard to relative abundances of specific proteins were described in detail previously (11) (24). Preliminary gels were run for the entire set of samples in a given experiment on 12% polyacrylamide/SDS gels, which were stained with Coomassie blue dye to assess equality of loading as described (11) (24). We refer to this procedure as ‘semiquantitative immunoblotting’ because the relative abundance, but not the absolute abundance, of the target proteins is determined. Immunocytochemistry. Mouse kidneys were perfusion-fixed by cannulating the heart with a 26 gauge needle under isoflourane anesthesia and perfusing the mouse with 2% parafomaldehyde (PFA) in PBS (26). After perfusion-fixation, kidneys were removed and postfixed overnight in 2% PFA in PBS. Kidneys were embedded in paraffin and 5 μM sections were obtained. Sections were labeled following the immunoperoxidase method described by Hager et al. (10). Antibodies and terminology for apical Na transporters. Affinity-purified primary antibodies used for immunoblotting and immunocytochemistry were directed to the major transporter proteins in the IMCD: the ", $, and ( subunits of ENaC (14), aquaporin-1 (25), aquaporin-2 (5), aquaporin-3 (6), aquaporin-4 (24), and the urea transporter UT A-1 (18). The antibodies were prepared in our laboratory using carrier-coupled synthetic peptides as immunogens and were affinity purified. Specificity of each antibody has been demonstrated by a combination of immunoblotting showing appropriate peptide-ablatable bands, and immunocytochemistry showing localization in appropriate membrane domains. Real-time RT-PCR. Quantitative, real-time RT-PCR was used to measure relative mRNA abundances in the renal inner medulla of AQP 1 (+/-) and (-/-) mice. Inner medullas were homogenized in a guanidine thiocyanate (Sigma) solution and RNA was isolated from the homogenate by CsTFA (ICN, Aurora, OH) centrifugation. Isolated RNA was DNase (DNAfree,Ambion, TX, USA) treated and equal amounts of RNA (determined spectrophotometrically) were reverse transcribed using oligo dT and Superscript II (Invitrogen, Carlsbad, CA) according the manufacturer’s specifications. Real-time PCR was performed on an ABI Prism 7900HT system using primers designed to amplify specific mouse cDNAs and the Quantitect Syber green PCR kit (Qiagen, Valencia, CA). Specificity of the reaction was determined by melting curve analysis. Relative quantitation of gene expression was determined using the comparative CT method, (2) as outlined at: http://docs.appliedbiosystems.com/pebiodocs/04303859.pdf . Ttests were performed on transformed CT values (i.e. 2 ). The primer pairs used for the realtime PCR are listed in Table 1. Urea Permeability Measurements. The urea permeability (Purea) was measured from inner medullary collecting duct using the isolated tubule microperfusion technique. IMCD segments were microdissected from the region of the medulla 30-70% of the distance from the inner-outer medullary junction to the papillary tip of the mouse kidney. The dissection solution contained (in mM): 125 NaCl, 25 NaHCO3, 2 K2HPO4, 2 CaCl2, 1.2 MgSO4, 5.5 glucose and 5 creatinine. The tubules were transferred to a perfusion chamber mounted on an inverted microscope, cannulated by concentric pipets and perfused in vitro. The perfusate and the pertitubular bath solutions were identical to the dissection solution except that 5 mM creatinine was replaced by 5 mM urea in the bath solution. Therefore, the tubules were perfused with solutions of equal osmolality, but with a 5 mM bath-to-lumen urea gradient. The urea permeability was determined by measuring the urea flux resulting from the transepithelial urea gradient as described previously (4). The urea concentrations in the perfusate, bath, and collected fluid was measured fluorometrically using a continuous-flow ultramicrofluorometer (4). The Sigma BUN reagent (Kit number 64-20) was used in the continuous-flow system. Statistical analysis. Quantification of the band densities from immunoblots was carried out by densitometry using a laser densitometer (Molecular Dynamics, San Jose, CA) and ImageQuaNT software (Molecular Dynamics). Values from knockout animals were compared to controls using an unpaired t test when standard deviations were the same, or by Welch t test when standard deviations were significantly different (INSTAT; Graphpad Software, SanDiego, CA). To facilitate comparisons, we normalized the densitometry values such that the mean for the control group is defined as 100. P<0.05 was considered statistically significant. RESULTS Expression of ENaC subunits, aquaporins and urea transporter in IMCD of AQP1 (-/-) versus control AQP1(+/-) mice. Figure 1 shows immunoblots demonstrating the expression profile of the major inner medullary transport proteins in AQP1 (-/-) mice compared to control (+/-) mice. The data in Figure 1 are representative of two separate experiments. There was a marked increase in the band densities corresponding to the β and γ subunits of the amiloride-sensitive sodium channel (ENaC) in the inner medulla of AQP 1 (-/-) mice (P < 0.05 and P < 0.05 respectively); however, there was no significant change in the abundance of the α subunit. The 40kDa glycosylated form of the basolateral water channel AQP 3 (8) was significantly increased (P < 0.05) in AQP 1 (-/-) inner medullas. In contrast, the band densities corresponding to the 28 and 51kDa forms (19) of AQP 4 were both significantly decreased in AQP 1 (-/-) inner medullas (P < 0.05 and P < 0.05). AQP 2 levels were not significantly different. Finally, the abundance of the urea transporter UT-A1 was markedly suppressed in AQP1(-/-) inner medullas relative to controls (P < 0.05). In summary, there were pronounced changes in the expression levels of several of the major solute transporter proteins of inner medullary collecting ducts of AQP 1 (-/-) mice relative to those of AQP (+/-) mice and these changes altered the overall transporter abundance profile to more nearly resemble what is normally seen in the cortical collecting duct as described in the Introduction. Immunocytochemistry. We next sought to confirm changes seen in the immunoblotting studies using immunocytochemistry. Figure 2 presents immunoperoxidase labeling of UTA-1, AQP 3,and AQP 4 in the renal inner medulla of AQP 1 (-/-) and (+/-) mice. Panels A and B demonstrate reduced immunoreactive UT A-1 in AQP 1 (-/-) inner medullas (panel B) relative to the heterozygote controls (panel A). Additionally, there is a moderate increase in the basolateral AQP 3 labeling in (-/-) inner medullas (panel D) relative to controls (panel C). Basolateral labeling of AQP 4 is markedly reduced in the inner medulla of (-/-) mice (panel F) relative to control inner medullas (panel E) from heterozygotes. The results of the immunoperoxidase labeling demonstrate changes in UT A-1, AQP 3 and AQP 4, which are qualitatively similar to the changes identified in initial Western blotting experiments. Realtime RT-PCR measurements of mRNA levels. Figure 3 demonstrates the transcript abundance of individual solute transporters in the inner medulla of AQP1 (-/-) mice relative to the transcript abundance found in control, AQP1 (+/-), mice. The cumulative results of two separate studies are presented on a log linear scale to emphasize the broad range of changes in transcript abundance. Additionally, a reference line representing a ratio of 1, or no change in transcript abundance, is provided to facilitate the comparisons between groups. The only change in transcript abundance observed among the three water channel isoforms examined in this study (AQP 2, AQP 3, and AQP 4) was a profound 20-fold increase in the abundance of AQP 3 (P < 0.05) transcript in the inner medulla of AQP 1 (-/-) mice. At the other extreme, there was a striking 90% reduction of UT A-1 and UT A-3 transcript abundance in the inner medulla of AQP1 (-/-) mice relative to controls (P < 0.05). Finally, there was a highly significant 3-fold increase in the abundance of each of the transcripts encoding the α,β, and γ ENaC subunits (P < 0.05 for each transcript). Urea permeability. Based on the marked differences in inner medullary urea transporter abundance between AQP1 (+/-) and (-/-) animals, we hypothesized there would be differences in urea permeability (Purea) of the IMCDs. A comparison of basal and vasopressin stimulated Purea of isolated, perfused inner medullary collecting ducts from AQP1 (+/-) and (-/-) mice are presented in Figure 4. The basal Purea in control, AQP1 (+/-), IMCDs (30.9 ± 2.8 x 10 cm/s) was significantly different from the basal P urea of (-/-) IMCDs (17.4 ± 8.1 x 10 cm/s) (N = 3 , P < 0.05). Additionally, after exposure to 100pM peritubular vasopressin for 30 min, the Purea of AQP1 (-/-) IMCDs was significantly reduced (N =5, P < 0.05) compared to control (17.0 ± 3.7 vs. 88.1 ± 14.1 respectively). In summary, both the basal Purea and the vasopressin-dependent Purea were significantly lower in the AQP1 (-/-) mice compared to (+/-) controls. Expression of ENaC subunits, aquaporins and urea transporter in IMCD of ClCnK1 (-/-) versus wild-type (+/+) mice. Finally, we examined the medullary transporter profile in the ClC-nK1 knockout mice, which also have impaired urinary concentrating ability resulting from defective passive chloride transport out of the thin ascending limb of Henle. As was the case in the AQP 1 (-/) model, the band densities corresponding the the β and γ ENaC subunits were significantly increased relative to wild-type controls (P < 0.05 and P < 0.05 respectively) whereas α ENaC abundance did not significantly change. Also, the levels of UT-A1 expression in ClC-nK 1 (-/-) mice were significantly reduced (P < 0.05) compared to controls. In contrast to the findings in AQP1 (-/-) mice, AQP 2 levels in the inner medulla of ClCnK1 (-/-) were significantly increased compared to wild type controls (P < 0.05) while levels of AQP 3 in the inner medulla were significantly reduced in CLC-nK1 (-/-) mice (P < 0.05). In conclusion, the expression profile of transporters in the CLC model share some similarities with that found in the AQP 1 (-/-) model, especially with regard to solute transporter expression. Discussion The physiological properties of the CD change as the tubule descends from the renal cortex to the papillary tip. The factors influencing the expression of the proteins ultimately responsible for manifesting the different transport properties of the CCD and IMCD may include the nature of physical environment the tubule resides in. Here we have examined the transporter expression profile of the IMCD in two mouse models with urinary concentrating defects. AQP1 (-/-) and ClC-nK1 (-/-) mice have urinary concentrating defects of roughly the same magnitude (13) (16) with spontaneous urine osmolalities of approximately 600-700 mOsm/kg and increasing minimally, if at all, upon water deprivation. Beyond these phenotypic similarities, impaired counter-current multiplication underlies the concentrating defect in both models albeit through different mechanisms. In the case of AQP 1 (-/-) mice, the severely diminished water permeability of the thin descending limb prevents extraction of water from this tubule segment. For CLC-nK1 (-/-) mice, the absence of the apical and basolateral chloride conductance of the thin ascending limb prevents NaCl efflux. In both instances, the inner medullary concentrating process is impaired thereby resulting in a lower inner medullary interstitial osmolalilty. In both models there were significant increases in β and γ ENaC protein expression in the inner medulla. Earlier work has shown that β and γ ENaC subunits facilitate the surface expression of functional ENaC complexes (3) suggesting sodium transport may be elevated in the IMCD of these mice. We propose that such an increase in sodium transport in the IMCD may help in maintaining extracellular fluid balance in the presence of polyuria by providing an alternative means of fluid reabsorption. Previous studies have demonstrated the rapid osmotic equilibration of urine in the micropuncture-accessible distal tubule (9) and the ability of the cortical collecting duct to continue to reabsorb isosmotic fluid in a process dependent on active sodium transport (22). The extension of this mechanism of fluid reabsorption from the cortical collecting duct into the inner medulla could ameliorate the urinary salt and water loss in these mice. There was a significant decrease in the expression of the UT-A1 protein and mRNA in the AQP1 (-/-) mice. The isolated perfused tubule measurements confirm that the low urea transporter expression levels are associated with low urea permeability. UT-A1 protein expression was also reduced in the IMCD of the CLC-nK1 (-/-) mice suggesting that the mechanism of suppression is somehow related to the concentrating defect present in both. Hypertonicity stimulates the UT-A1 promoter (17) implying the decreased expression of UT-A1 reported here may stem from a reduced tonicity of the inner medulla associated with the concentrating defects. Thus tonicity of the inner medulla may directly contribute to the altered expression of UT-A1 seen in these models. AQP3 levels were significantly increased and AQP 4 protein levels significantly decreased in the inner medulla of the AQP 1 (-/-) model thus reversing their normal pattern of expression. It is not readily apparent what physiological consequences this would impart as the apical rather than the basolateral membrane of the IMCD is rate limiting for water transport (8). However, the changes in AQP3 (and AQP2) expression in the inner medulla of the CLC-nK1 (-/) do not parallel the changes seen in the AQP1 model and suggests a concentrating defect per se is not the sole factor determining the final expression pattern of the transporters in the inner medulla of these models. Indeed, the dissociation of changes in protein levels and transcript abundance for AQP 4 and αENaC in the AQP 1 (-/-) model also supports the idea there are multiple regulatory mechanisms shaping the final expression profile of the transport proteins. The dissociation of the AQP4 and αENaC transcript and protein levels has not, to our knowledge, been previously described. In general, reports of increases in αENaC protein abundance are associated with elevations in αENaC transcript (1) (15) (14). However, in the A6 model of the renal collecting duct, hypotonicty is known to 1) increase transepithelial sodium transport without an increase the ENaC protein levels (20) and 2) increase the abundance of αENaC transcript (19). These effects of hypotonicity on ENaC regulation may be particularly relevant in the context of mouse models with chronic concentrating defects. In conclusion, the expression pattern of many of the transport proteins known to play prominent roles in the physiology of the IMCD is altered in AQP 1 (-/-) mice. Several, but not all of the changes in transporter expression seen in the IMCD of AQP1 (-/-) mice are also present in the IMCD of the CLC (-/-) mice; a model with a similar concentrating defect. Given the similarities between the expression profiles in AQP 1 and CLC-nK1 mouse models, as well as the underlying similarities in the urinary concentrating defect it is tempting to propose that many of the changes in transporter expression occur in response to a chronically diminished tonicity of the inner medulla. However, the changes in IMCD transporter expression in the knockout models may be present at birth and reflect differences in ontogeny rather than adaptation to, or resulting from, the concentrating defect. Furthermore, there are clearly many other possible mechanisms that may contribute to the altered expression of transporters in the IMCD of mice with concentrating defects, including urinary flow, GFR, and hormone levels, which have not been addressed in this study. For example, GFR is significantly reduced in the AQP1 (-/-) models (23). Further study will be required to establish the mechanism(s) responsible for the changes in IMCD transporter expression Acknowledgments The authors would like to thank Dr. James Wade (University of Maryland) for advice regarding mouse tissue fixation and Dr. Robert A Fenton (NIH) for the UT-A1 and UT-A3 real time PCR primers.