点击显示 收起
【摘要】 To investigate the underlying causes for aldose reductase deficiency-induced diabetes insipidus, we carried out studies with three genotypic groups of mice. These included wild-type mice, knockout mice, and a newly created bitransgenic line that was homozygous for both the aldose reductase null mutation and an aldose reductase knockin transgene driven by the kidney-specific cadherin promoter to direct transgene expression in the collecting tubule epithelial cells. We found that from early renal developmental stages onward, urine osmolality did not exceed 1,000 mosmol/kgH 2 O in aldose reductase-deficient mice. The functional defects were correlated with significant renal cellular and structural abnormalities that included cell shrinkage, apoptosis, disorganized tubular and vascular structures, and segmental atrophy. In contrast, the transgenic aldose reductase expression in the bitransgenic mice largely but incompletely rescued urine concentrating capacity and significantly improved renal cell survival, cellular morphology, and renal structures. Together, these results suggest that aldose reductase not only plays important roles in osmoregulation and medullary cell survival but may also be essential for the full maturation of the urine concentrating mechanism.
【关键词】 urine concentration diabetes insipidus
NEWBORN MAMMALS ARE NOT CAPABLE of producing concentrated urine. In rats, the urine concentrating mechanism develops in the first few weeks of postnatal life. Urine osmolality rises from 300 mosmol/kgH 2 O at birth to nearly 2,000 mosmol/kgH 2 O by the age of 3 wk ( 39 ). The maturation of this mechanism is believed to be developmentally regulated and may be regulated by corticosteroid hormones ( 37 ) and/or hypertonicity and the osmoresponsive transcriptional factor tonicity-responsive element binding protein (TonEBP)/osmotic response element binding protein (OREBP) ( 15, 24 ), the latter being the only known osmotic-responsive transcriptional activator. Loss of TonEBP/OREBP resulted in renal atrophy and lack of tonicity-responsive gene regulation, defective urine concentration, and the development of polyuria and polydipsia ( 28 ). Kidney-specific transgenic expression of a dominant negative form of TonEBP/OREBP resulted in similar phenotypes as that of the TonEBP/OREBP knockouts ( 26 ). These data suggest that TonEBP/OREBP, and the target genes it regulates, are essential for renal osmoregulation and contribute to the development of the urine concentrating mechanism.
Aldose reductase (AR; AKR1B1, EC1.1.1.21) is one of the osmoresponsive genes transcriptionally activated by TonEBP/OREBP ( 26, 28 ). Other genes include the Na + /Cl - -coupled betaine/ -aminobutyric acid transporter ( 31 ), the Na + / myo -inositol cotransporter ( 33 ), and the taurine transporter ( 18 ), heat shock protein 70 ( 46 ), and the vasopressin-regulated urea transporters ( 34 ). In the kidney, AR catalyzes the conversion of glucose to form sorbitol, which serves as a major organic osmolyte in the renal medulla. A large body of convincing evidence suggests that AR is of central importance to renal osmoregulation ( 2, 4, 5, 11, 38, 42 ). AR's role in the development of the urine concentrating mechanism itself, however, has not been clearly demonstrated. Our laboratory previously demonstrated that AR-deficient mice develop symptoms of polydipsia and polyuria, a condition resembling that of diabetes insipidus in humans ( 16 ). This finding was later confirmed by another research group ( 1 ). Although these previous studies firmly established the importance of AR in urine concentration, the mechanisms through which AR deficiency leads to the development of polyuria and polydipsia remain obscure. Mechanistically, it was not clear whether urine concentrating ability became defective after full development, or, alternatively, if it was never been fully established in knockout (KO) mice. Moreover, it was also unclear whether AR deficiency locally within the kidney is sufficient to cause diabetes insipidus. The present study was conducted to address these questions.
MATERIALS AND METHODS
Animals. A 3.6-kb fragment containing the mouse kidney-specific cadherin promoter (Ksp) sequence from -3458 to +129 ( 17 ) with restriction sites Sph I and Sbf I on its 5'- and 3'-ends, respectively, was PCR-amplified from mouse genomic DNA (catalog no. 6650-1, Clontech, Palo Alto, CA) and inserted into the pGEM-T vector (catalog no. A1360, Promega, Madison, WI). After confirmation of the DNA sequence immediately around the promoter region by sequencing, the Sph I- Sbf I fragment was released and inserted into plasmid pHARb at corresponding restriction sites. Plasmid pHARb is a pBluescript-based promoterless vector that contains a 1.3-kb cDNA for human AR (hAR) and a SV40 poly-A signal sequence ( 7 ). A 6.3-kb Sph I- Nde I DNA fragment encompassing the Ksp, hAR cDNA, and SV40 poly-A signal was released from the resultant expression vector pKspAR ( Fig. 1 ) and used for mouse egg pronuclear microinjection. Transgenic founders were identified by PCR amplification of transgene DNA, with the two primer sequences being 5'-GTCAAGGGGGCAGTGACAGA-3' (kspAR_f) and 5'-CTGAGTGTCTTCTGGCAGGC-3' (kspAR_r) and confirmed by Southern hybridization. Radioactive 32 P-labeled hAR-specific DNA probes were prepared by random-priming amplification of a 254-bp cDNA fragment using the Megaprime labeling System (catalog no. RPN1607, Amersham Biosciences) as instructed. The cDNA fragment containing the 254-bp sequence immediately downstream of human AR exon 10 was released from pBluescriptSK-hARexon10 by Bam HI and Eco RI. Positive B6/CBA founders were backcrossed with C57BL/6 for six consecutive generations to obtain N6 mice to improve genetic homogeneity. The transgenic allele (KspARtg) was then introduced into our previously created AR KO mice ( 16 ) that have been backcrossed with C57BL/6 for seven generations. Subsequent intercrossing generated bitransgenic (BT) mice homozygous for both KspARtg and the KO alleles, as well as their controls. Mice were bred, weaned, and maintained as described previously ( 16 ). Only male mice were used for characterizations and analyses. Animal care and experimental procedures were in accordance with the guidelines by the University of Hong Kong Committee on the Use of Live Animals in Research and Teaching.
Fig. 1. Generation of transgenic mice and Northern analyses. A : construction of Ksp-driven aldose reductase (AR) expression vector pKspAR. The 6.3-kb Sph I- Nde I DNA fragment containing Ksp, hAR cDNA, and a poly-A signal (SV40 poly A) was released and used in pronuclear microinjection. Locations for the pair of PCR primers (kspar_f and kspar_r) as well as the hAR exon10 region are indicated. Amp r, ampicillin resistance gene cassette. B : Northern blot analysis of transgenic mice (representative of repeated results). Sizes of the RNA bands are in numbers of nucleotides. Lanes A and C : RNA samples from 2 KspARtg +/- mice. Lanes B and D : RNA samples from 2 wild-type (WT) littermates. C : kidney-specific transgene expression (representative of repeated results). Total RNA samples (20 µg/lane loading) were extracted from tissues of a bitransgenic (BT) mouse. 18S, 18S rRNA; 28S, 28S rRNA. See text for a more detailed description.
Genotyping, metabolic cage experiments, and urine and serum analyses. Presence of KspARtg was tested as described above by tail genomic DNA PCR amplification. Homozygosity for KspARtg was tested by Southern hybridization of Bam HI-digested tail DNA utilizing the above-mentioned hAR-specific probe. A 600-bp DNA fragment from cDNA encoding TonEBP/OREBP ( 23 ) was used as a control probe to calibrate for DNA loading. On the other hand, PCR primers HHA (5'-GGGCTATACGGAGAAACTGTGT-3'), HHB (5'-TGACCTTCCTCTAGAGGCTCTT-3') and neoPA (5'-ATCAGCAGCCTCTGTTCCAC-3') were used not only to test for the presence but also the homozygosity of the KO allele and the wild-type (WT) allele. Urine and serum collections and analyses were conducted with 5- and 8-wk-old male mice as described ( 16 ). For the 3-wk-old mice, abdominal massaging was performed on 21-day-old mice on the day of weaning to stimulate urination to collect small aliquots of urine. For the 2-wk-old mice, urine samples were collected by bladder aspiration immediately following cervical dislocation.
Transgene expression by Northern and in situ hybridizations. Northern hybridizations were conducted with total RNA samples extracted from the kidney and other tissues. In addition to the above-mentioned hAR-specific probe, full-length cDNAs for hAR and mouse AR (mAR) ( 43 ) were also used for the preparation of probes. In situ hybridization was performed with kidneys dissected from male BT and KO mice. Processed 10-µm kidney cryostated sections were hybridized with 35 S-UTP (Amersham Biosciences)-labeled riboprobes prepared from in vitro transcription of pBluescriptSK-hARexon10 using T3 and T7 RNA polymerases (Promega). The vector was linearized by Eco RI and Bam HI for synthesizing the antisense and sense probe, respectively. Ilford K.5 emulsion (catalog no. 1355127, Ilford, Cheshire, UK) was used for autoradiography.
Analyses of renal glucose, sorbitol, and myo-inositol. For HPLC analyses, kidney samples were dissected and quickly frozen and stored in liquid nitrogen until use. Monosaccharides and polyols were extracted from whole kidney as described ( 16, 25 ). Samples were subsequently loaded and analyzed on a DX-600-IC HPLC System (Dionex, Sunnyvale, CA) using a Carbopac (CPMA-1) anion exchange column, an ED50 electrochemical detector, and an AS50 automated sample injector.
Histological analyses of paraffin-embedded kidney sections. For periodic acid-Schiff (PAS) and immunohistochemical staining, kidneys were fixed overnight in 2% paraformaldehyde at 4°C following the removal of the capsules. The fixed tissues were paraffin embedded and sectioned at 5-µm thickness.
Transmission electron microscopic studies. Kidneys were dissected from 5-wk-old euhydrated mice. Tissue blocks of the terminal segment of inner medulla/renal papilla were cut into 1-mm 3 pieces using a razor blade and fixed in 2.5% glutaraldehyde for 4 h at 4°C. After several washes in 0.1 M cacodylate buffer, the tissue blocks were osmicated in 1% osmium tetroxide for 1 h at room temperature. The blocks were then processed for routine dehydration, incubated in propylene oxide, and embedded in epoxy resin (Poly/Bed 812, Polyscience, Warrington, PA). Ultrathin (90 nm) sections were cut and mounted on 150-mesh copper grids, stained with 2% aqueous uranyl acetate followed by Reynold's lead citrate. The sections were then examined with a Philips EM208S electron microscope operating at 80 kV. Images were recorded with DigitalMicrograph 3.3.0 software. Photoshop 7.0 was used for subsequent image handling. In addition, a few ultrathin sections were mounted on slides and stained with toluidine blue for morphological examinations under light microscopy.
Statistical analyses. All data are expressed as means ± SE. One-way ANOVA with Tukey's posttest was used in multigroup comparisons. Probability values <0.05 (*) were considered to be statistically significant, those <0.01 (**) very significant, and <0.001 (***) extremely significant.
RESULTS
Generation of transgenic mice overexpressing AR kidney specifically. We reasoned that if the KO phenotypes can be rescued to a certain degree by coexpressing a kidney-specific AR transgene, then local kidney deficiency of AR would probably be sufficient to cause diabetes insipidus. Plasmid pKspAR ( Fig. 1 A ) was constructed to target AR expression to the kidney collecting ducts. Following the pronuclear microinjection of mouse eggs with transgene DNA, nine offspring mice (including both males and females) were identified to be the transgenic founders (KspARtg +/? ). Transgenic status was indicated by the presence of a 653-bp DNA fragment from PCR amplification of the tail genomic DNA using the two primers indicated in Fig. 1 A and further confirmed by presence of a 1,759-bp Bam HI-digested tail genomic DNA fragment that can hybridize with 32 P-labeled probes prepared from a DNA fragment 3' to hAR exon 10 in Southern hybridization (Yang JY et al., unpublished observations). The transgenic founders all appeared to live and breed normally. Transgenic line 1618, with the highest level of AR expression in the kidney, was selected for further breeding and characterization.
Kidney-specific AR expression and transgene in vivo functionality. A prominent band that is 2,400 nucleotides in size was detected by hAR exon 10-flanking, sequence-specific, radiolabeled probes ( Fig. 1 B ) in total RNA extracted from kidneys of heterozygous transgenic mice (KspARtg +/- ). Additionally, there was a minor band that is 1,400 nucleotides in size. This smaller band does not seem to be the endogenous mAR mRNA as no band was detected for RNA sample from the WT littermates. Rather, the transgene seemed to give rise to two transcripts, one stopping at the end of the hAR cDNA (1.4 kb) and the other at the SV40 polyadenylation signal (2.4 kb).
After backcrossing with C57BL/6 mice for six generations, the sixth-generation mice (N6) KspARtg mice were bred with a line of KO mice to generate BT mice, as well as WT and KO mice. No body weight difference was observed among these genotype groups at the age of 8 wk. Kidney-specific AR expression was first demonstrated by Northern analyses of total RNA samples extracted from tissues of BT mice ( Fig. 1 C ). As expected, the kidney-specific cadherin promoter directs AR expression only in the kidney but not in all other tissues tested, including the brain, heart, intestine, liver, lung, spleen, and testis. Within the kidney, transgenic AR expression was shown to be predominantly localized in the tubule epithelial cells in the inner medulla and the renal papilla/pelvis ( Fig. 2 ). There are, however, weak positive signals found in a small fraction of cells in the outer medulla/cortex region. Immunostaining with anti-hAR antibody also detected similar patterns of gene expression (Yang JY et al., unpublished observations). In contrast, no expression was detected in kidney glomeruli. These data on the patterns of gene expression are in accordance with other studies using the same promoter fragment in our laboratory ( 26 ) as well as others ( 17, 40 ).
Fig. 2. In situ hybridization analysis demonstrating KspARtg expression in the kidney medulla. The data are representative of repeated results of antisense probe hybridization. Hybridization with sense probes did not lead to positive signals, and the results are not included. Positive signals are represented by the white spots. The red background staining is due to eosin under bright-light illumination. A : knockout (KO) whole kidney. B : BT whole kidney. C : BT medulla-papilla. IM, inner medulla; CO, cortex; RP, renal papilla. Original magnification x 100. See text for a more detailed description.
To demonstrate the in vivo functionality of KspARtg, kidneys from BT mice and their controls were isolated. Extraction of the renal monosaccharides and polyols was performed using the whole kidney. HPLC analyses showed substantial production of renal sorbitol in BT mice, indicating that the transgene is indeed functional in catalyzing glucose-sorbitol conversion ( Fig. 3 A ). Because, in the medulla, only a portion of the epithelial cells of the collecting tubules are expected to express the transgene, it is not surprising to see that the renal sorbitol level in BT mice accounts for <10% of that for their WT littermates (0.070 nmol/mg wet kidney wt, n = 3 for BT mice vs. 0.758 nmol/mg kidney wet wt, n = 4 for WT mice, P < 0.001), whereas the level for KO mice is not completely detectable. Differences in levels of renal glucose and myo -inositol among the three genotypic groups, on the other hand, are not statistically significant ( Fig. 3 B ).
Fig. 3. Renal content of sorbitol ( A ) and glucose myo -inositol ( B ) as determined by HPLC. Euhydrated mice were used in these assays ( n = 3-4). ND, not detectable for KO mice. See text for a more detailed description. *** P < 0.001.
Renal expression of water channels AQP2 and AQP3 and urea transporters UT-A1, UT-A2, and UT-B. Our previous studies indicated that renal expression/secretion of antidiuretic hormone (AVP), the V 2 vasopressin receptor (V 2 R), and the water channel proteins AQP1 and AQP2, all appeared to be normal or upregulated in AR-deficient mice ( 16 ). In this current study, we performed immunohistochemistry analyses on paraffin-embedded kidneys with antibodies against AQP2, AQP3, UT-A1, and UT-A2. These analyses mostly indicated no significant difference in the expression of these four renal transporters between WT and KO mice (supplemental Figs. 2 and 3; the online version of this article contains all supplementary data), although in some cases, some of these channels/transporters (e.g., UT-A1) appeared to be slightly upregulated in BT mice. The trend of AQP2, UT-A1, and UT-A2 expression was further confirmed by Western blotting and/or quantitative real-time PCR analyses with whole kidney protein samples (data not included). Furthermore, Western blotting also did not show significant differences in UT-B expression among the three genotypic groups of mice at the age of 5 wk (data not included). The functional defect of urine concentration in KO mice is unlikely to be caused by altered expression of these water channels or urea transporters.
Developmental arrest of the urine concentrating mechanism in KO mice and phenotypic rescue of polyuria and polydipsia in BT mice. To determine whether urine concentration became defective after full establishment of the urine concentrating mechanism or whether it had never been fully developed in KO mice, we compared the urine osmolality for three groups of mice at different ages ( Fig. 4 A ). Our data showed that the urine concentrating ability in WT mice matured to a similar level ( 2,400 mosmol/kgH 2 O urine osmolality) at the same age (3 wk) as occurs in rats ( 39 ). Signs of underdevelopment of the urine concentrating mechanism in KO mice, however, appeared as early as 2 wk postnatally and became statistically significant 1 wk later. At the age of 3 wk, urine osmolality in KO mice only reached at 800 mosmol/kgH 2 O. Urine osmolality did not increase further from 3 wk to 13 mo. Therefore, this result suggests that the urine concentrating mechanism in KO mice had never been fully developed. For BT mice, urine osmolarity was gradually increased starting at the age of 3 wk and became significantly improved from the age of 5 wk on, indicating that the time-dependent transgenic AR expression and intracellular accumulation of sorbitol in the collecting tubule epithelial cells (CTECs) was able to stimulate the development of the urine concentrating mechanism. Water intake and urine output for WT and KO mice at the ages of 5 and 8 wk ( Fig. 4, B and C ) were similar to what were reported previously ( 16 ), whereas both water intake and urine output were incompletely normalized in BT mice as a result of transgene expression. These data, therefore, indicated that AR/sorbitol is essential for maximal urine concentration.
Fig. 4. Urine osmolality, water intake, and urine output in WT, KO, and BT mice. A : urine osmolality at the ages of 2, 3, 5, and 8 wk and 8-13 mo ( n = 4-6). B : water consumption at the ages of 5 and 8 wk ( n = 7-8). C : urine output at the ages of 5 and 8 wk ( n = 7). See text for a more detailed description. * P < 0.05. ** P < 0.01. *** P < 0.001.
Analyses of serum collected from 8-wk-old mice ( Table 1 ) did not show significant differences among the three genotypic groups of mice for all the serum solutes tested except urea. These results ruled out the involvement of hypokalemia ( 29 ). In accordance with the trends of water intake and urine excretion, urine osmolarity and concentrations of urinary solutes change dramatically among genotypic groups ( Table 2 ). When 8-wk-old mice were provided with water ad libitum, urine from BT mice was more than twice as concentrated as that of KO mice, suggesting significant phenotypic rescue as a result of KspARtg expression. Indeed, the concentrations of most urinary analytes were normalized to the levels of WT mice, ranging from 57.9 to 89.4% ( Table 2 ). Although we did not observe hypercalcemia in KO mice as reported by Aida et al. ( 1 ), there did appear to be a small but not significant trend toward an increase in 24-h urinary Ca 2+ excretion in KO mice, the effects of which probably cannot be completely excluded ( 13 ). No significant difference in urinary excretion was observed for most of other solutes, suggesting again that AR deficiency primarily affects the urine concentrating mechanism. Interestingly, the concentrations of major renal solutes, including Na +, Cl -, K +, and urea, were reduced in KO mice, all to 70%, and were normalized to a similar degree of 60-65% as a result of the presence and expression of KspARtg in BT mice. In other words, there is a parallel decrease in the urine concentration of all major renal solutes in KO mice and a parallel increase in BT mice, suggesting again it is the urine concentrating mechanism that was mostly affected. When data on urine osmolarity, water consumption, and urine excretion and the analyses of serum and urine samples are taken together, it is clear that AR deficiency alone caused defective urine concentration in KO mice that was largely but incompletely rescued by transgenic AR expression in the CTECs in BT mice.
Table 1. Serum analyses
Table 2. Urine analysis
Renal cell shrinkage and structural abnormalities in KO and improvement by transgenic AR expression in BT mice. To look for the cellular and renal structural basis of the functional defects in KO mice and phenotypic rescue in BT mice, renal histological and morphological examinations were performed. Kidneys dissected from the three genotypes of mice at all ages did not show any apparent differences in wet weight or kidney shape. However, cell shrinkage and renal structural abnormalities were observed in kidneys of both KO mice and BT mice as early as 2 wk postnatally ( Fig. 5 ). At the age of 2 wk, significant cell shrinkage was evident throughout the entire medulla (Yang JY et al., unpublished observations), especially in the renal papillary segments ( Fig. 5, B and E ). CTECs in KO mice lacked the characteristic cuboidal shape of WT mice, with small cytoplasmic volume and greatly reduced cellular sizes, indicating possible CTEC atrophy. In the kidneys of 2 ( Fig. 5, B and E )- and 5-wk ( Fig. 7, B and E )-old KO mice, the tubular/vascular structures in the inner medulla also appeared to have been greatly disorganized, especially in the terminal segment of the inner medulla toward the papillary tips. For BT mice, cell shrinkage and tubular and vascular structures generally appeared to be less abnormal ( Figs. 5, C and F, and 7, C and F ), probably as a consequence of time-dependent rescue via transgene expression. In contrast to the inner medulla, little or no difference was observed for the cortex among three genotypes of mice (Yang JY et al., unpublished observations).
Fig. 5. Periodic acid-Schiff (PAS) staining of 2-wk-old mouse medulla, illustrating renal cell shrinkage, structural abnormalities in KO mice (original magnification x 100; D - F were cropped from the corresponding black-bordered boxes in A-C and further magnified 2-fold digitally). CT, collecting tubules; LH, loops of Henle (mainly identified by their characteristic squamous epithelial cells); UW, urothelial walls. Orange arrows indicate epithelial cells/tubular segments that appear to be abnormal or damaged in the collecting tubules. Yellow arrows indicate cells/limbs of loops of Henle that appear to be abnormal or damaged. In the medulla of WT mice, purple PAS staining was primarily found in the basolateral sides of the collecting tubule epithelial cells (CTECs), bordering the loop of Henle tubule cells or blood vessels ( A and D ). Strong staining, however, was observed for KO mouse medulla in similar extracellular spaces and many other seemingly empty spaces ( B and E ). Significant cellular shrinkage was observed for medullary cells, in particular the CTECs in the KO mice ( B and E ). Also noticeable are cells with small dark-staining spots (probably representing disintegrated nuclei in apoptotic cells) and segmental atrophy in both the collecting tubules and the loops of Henle ( E ). Most of the cellular and structural abnormalities were to some degree improved in BT mice ( C and F ).
Fig. 6. Medullary cellular, nuclear, and mitochondrial shrinkage, apoptosis, and renal structural abnormalities as revealed by transmission electron microscopy. The data were typical of multiple sections involving multiple mice for each genotype. Original magnifications were indicated. Scale bars in sections with x 36,000 magnification = 0.2 µm; those in sections with x 8,900 magnification = 1 µm; those in sections with x 4,400 magnification = 2 µm; and those in sections with x 1,400 magnification = 5 µm. CT, collecting tubule; EC, collecting tubule epithelial cells; NU, nucleus; MT, mitochondria; VA, vacuoles. Significant cellular and nuclear shrinkage ( B, E, and H ) and atrophic renal tubular/vascular layers that are greatly disorganized with segmental atrophy are shown for KO mice in B and E (up and rightward arrows). CTEC apoptosis in KO mice was characterized by the presence of a large number of vacuoles and greatly shrunken cell and nuclei ( E and H ), chromatin condensation and dispersion of nuclear material toward the nuclear membrane (up arrow in H ), and possible presence of apoptotic bodies (down arrows in H ). Mitochondrial shrinkage was shown for both KO and BT mice in K and L (up arrows). Furthermore, more microvillous projections or vesicles were found in the luminal sides of the CTECs from KO and BT mice compared with those from WT mice (up and leftward arrows in H and I ). All the morphological and structural abnormalities found in KO mice were shown to be incompletely improved in BT mice ( C, F, I, and L ). Results are representative of at least 3 mice/genotype.
Fig. 7. Urothelial cell apoptosis, urothelial wall segmental atrophy, and disorganized vascular compartments in KO mice kidneys (unperfused kidneys). A - C : transmission electronic microscopic images. Original magnification, x 1,400. D - F : light microscopic images of toluidine blue-stained ultrathin (90 nm) sections, Original magnification, x 400. CT, collecting tubule; UW, urothelial wall; VR, vasa recta. Vascular compartments were mainly identified by the presence of red blood cells. Rightward arrows indicate apoptotic/atrophic urothelial cells ( B ), whereas leftward arrows indicate red blood cells in the VR. Notice the disorganized vascular structures in sections from KO mice ( B and E ) and slight improvement in BT mice ( C and F ). Results are representative of at least 3 mice/genotype.
In addition to significant tubular/vascular structural abnormalities, transmission electron microscopic studies also indicated significant cellular, nuclear, and mitochondrial shrinkage in KO mice ( Figs. 6 and 7 ). Moreover, renal cell apoptosis was found in a significant number of CTECs ( Fig. 6, E and H ) and urothelial cells ( Fig. 7 B ) in the terminal segment of the inner medulla toward the papillary tip, but only sporadically in WT mice. These CTECs showed typical characteristics of apoptotic cells, which include cell and nuclear shrinkage and deformation, chromatin condensation, dispersion of DNA toward the nuclear membrane, presence of a great number of vacuoles around the nucleus, and formation of apoptotic bodies, etc. ( Figs. 6, E and H, and 7 B ). More surprisingly, the mitochondria in both KO and BT mice were markedly shrunken ( Fig. 6, K and L ), suggesting potential mitochondrial dysfunction, which might be involved in the induction of apoptosis. Also noteworthy is the presence of a large number of microvillous projections or vesicles on the luminal surface of CTEC principal cells in KO ( Fig. 6 H ) and BT mice ( Fig. 6 I ) but relatively few in WT mice. Both the nature of these ultramicroscopic structures and the significance of their differences among the three genotypes of mice are not clear. In contrast to CTECs, relatively few apoptotic cells were observed for other types of renal cells. Probably as a consequence of the apoptotic loss of CTECs and urothelial cells, a certain degree of structural injury/disorganization such as a thin tubule cell layer, tubular segmental atrophy, and irregular disruption or obstruction of the tubules and/or vasculature was observed in KO mice ( Figs. 5, B and E, 6, B and E, and 7 B ).
DISCUSSION
Kidney local deficiency of AR is sufficient to cause polyuria and polydipsia. In both human and rodents, AR was found to be ubiquitously expressed in almost all tissues including brain, lens, retina, adrenal gland, testis, lung, kidney, and heart ( 6, 14 ). A general deficiency in AR, henceforth, will potentially affect one or more of the mechanisms and/or factors that are involved in urine concentration and kidney function ( 22, 30 ). These include, but are not limited to, blood volume and pressure-regulatory mechanisms, levels of glucocorticoids and mineralocorticoids ( 8, 12, 20, 36 ), renin/angiotensin ( 3, 10, 21 ), and TonEBP/OREBP ( 26 ). Some of these urine concentration-relevant factors are renal, and some are nonrenal. Furthermaore, some are arginine vasopressin (AVP)-dependent and some are independent. Because the effects of AR deficiency on most of these factors remain undetermined, their involvement in the development of polyuria and polydipsia in KO mice has not been definitively excluded.
The results from this study clearly suggest that kidney local deficiency is sufficient to cause diabetes insipidus because CTEC-specific AR expression was able to rescue 70% of urine concentrating ability in BT mice at the age of 8 wk. Because the nature of the kidney-specific cadherin promoter can be expected to differ from that of the endogenous AR promoter in terms of development, tonicity regulation, and cellular specificity, the pattern and level of transgenic AR expression to a certain degree will be different from those of an endogenous gene temporally or spatially. These differences probably can explain a large part of the incomplete phenotypic rescue. Despite the incomplete rescue, our data from BT mice indicated that lack of AR expression in the CTECs probably is sufficient to induce polyuria and polydipsia.
AR is essential for renal cell survival in vivo. AR is the primary osmolyte-producing gene in cells in the renal medulla. Inhibition of AR by the chemical inhibitors tolrestate and sorbinil had previously been shown to cause a reduction in the viability of inner medullary epithelial cells (the same set of cells found to be apoptotic in our current study) in culture, due to a failure of high extracellular NaCl to induce sorbitol accumulation ( 47 ). Furthermore, renal medullary apoptosis in cyclosporine A-treated rats was also shown to be due, at least in part, to the downregulation of TonEBP/OREBP-regulated osmoprotective genes including AR ( 27, 41 ). Thus our in vivo demonstration that loss of AR results in significant renal cell shrinkage and apoptosis is consistent with the notion that cell survival is dependent on the intracellular accumulation of protective organic solutes, especially sorbitol ( 2, 9, 32, 45 ). The improvements in cellular apoptosis and structural lesions and the rescue of phenotypes in the BT mice, on the other hand, reconfirmed the essential role of AR/sorbitol in renal cytoprotection in vivo.
AR is required for the full maturation of the urine concentrating mechanism. The phenotypic loss of the ability to concentrate urine in KO mice probably has two components, one related to the defective maturation of the urine concentrating mechanism and the other related to the improper functioning of the mechanism due to the lack of AR/sorbitol. At the present time, the contribution from each component has not been clearly distinguished and more work will be needed. The functional contribution, however, does not appear to be very significant. In support of this, sorbinil treatment in adult animals whose urine concentrating mechanism had been fully established led only to slight changes in urine flow and osmolality ( 44 ). Most rat and mouse studies conducted in many laboratories, including our repeated use of AR inhibitors such as zopolrestate and sorbinil in a variety of experiments, did not lead to a similar degree of urine concentrating defect as that of the AR-deficient mice. Furthermore, humans given aldose reductase inhibitors have not been shown to develop severe urine concentration problems in clinical trials. The putative defective maturation, either alone or in combination with a functional defect, thus may account for a significant portion of the phenotype in KO mice. Our present findings strongly suggest that AR/sorbitol is essential for the full postnatal maturation of urine concentrating ability, because the AR KO mice never develop the ability to concentrate their urine, whereas BT mice do develop concentrated urine, albeit more slowly than WT mice.
The development of neonatal kidneys and maturation of the urinary concentrating system involve, among other factors, increases in medullary osmolality, vascularization, and cytodifferenciation of the inner medulla and branching/lengthening of the loop of Henle and the inner medullary collecting ducts, etc. Our demonstration that 2-wk-old KO mice started to show signs of defective urine concentrating capacity and morphological abnormalities indicated that AR is required for the full developmental maturation of the urine concentrating mechanism. This also suggested that AR expression has to precede the initiation of the development of the urine concentrating mechanism, a finding that is consistent with previous AR developmental expression studies ( 39 ). As suggested by others ( 19, 24 ), local hypertonicity and cell shrinkage could be important triggering factors for renal cell apoptosis during kidney development. Indeed, apoptosis was suggested to be part of the mechanism for lengthening of the thin ascending limb and the separation of the inner and outer medulla ( 19 ). Extensive cell shrinkage for the majority of medullary cells in kidneys of the KO mice from 2 wk onward, however, was associated with significant CTEC apoptosis, raising the possibility that tissue injury due to this kind of nonprogrammed cell death might hinder proper tubular development. The presence of a thin tubular layer, more abnormally long, straight papillary collecting ducts, and irregularities of the vascular and tubular structures in these mice thus might represent disturbances to the normal branching/lengthening morphogenesis that leads to the maturation of the ascending thin limbs of loop of Henle and/or inner medulla collecting tubules in normal mice. Because the maturation of these renal structures is part of the maturation of the urine concentrating mechanism, these abnormal structural features of the inner medulla in KO mice could be related to the impairment in water conservation and the arrest of the maturation of the urine concentrating mechanism. On the other hand, medullary cells, especially CTECs, are the cells expressing aquaporins and urea transporters and other genes that are directly involved in the mechanisms of water reabsorption. Shrinkage of these cells might compromise their ability to reabsorb water due to reduced production of functional solute transporters and/or their luminal membrane targeting. Together, these defects appear to have prevented the full establishment of the urine concentrating mechanism in KO mice.
Inhibition of the polyol pathway and AR is one of the proposed therapeutic and preventive strategies for diabetic complications ( 35 ). One practical clinical implication that can be derived from this study is that care has to be taken to avoid potential damage to the kidneys by treatments that have to be initiated early and sustained for a long duration.
GRANTS
This work was supported in part by Hong Kong University internal grants (Projects 10203108 and 10203959) and Hong Kong Research Grant Council Grants HKU7495/03M (J. Y. Yang) and HKU7259/98M (S. S. M. Chung) and by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-41707, DK-63657, and DK-61521 (J. M. Sands) and DK-62081 (J. D. Klein).
ACKNOWLEDGMENTS
The authors thank Marcella Ma, Amy Lam, and Amy Lo for technical assistance, Eric Leung for performing some metabolic cage experiments, and Amy Wong of HKU EM Unit for help with the TEM work.
【参考文献】
Aida K, Ikegishi Y, Chen J, Tawata M, Ito S, Maeda S, and Onaya T. Disruption of aldose reductase gene (Akr1b1) causes defect in urinary concentrating ability and divalent cation homeostasis. Biochem Biophys Res Commun 277: 281-286, 2000.
Bagnasco SM, Murphy HR, Bedford JJ, and Burg MB. Osmoregulation by slow changes in aldose reductase and rapid changes in sorbitol flux. Am J Physiol Cell Physiol 254: C788-C792, 1988.
Bonjour JP and Malvin RL. Stimulation of ADH release by the renin-angiotensin system. Am J Physiol 218: 1555-1559, 1970.
Burg MB. Role of aldose reductase and sorbitol in maintaining the medullary intracellular milieu. Kidney Int 33: 635-641, 1988.
Burg MB. Coordinate regulation of organic osmolytes in renal cells. Kidney Int 49: 1684-1685, 1996.
Cao D, Fan ST, and Chung SS. Identification and characterization of a novel human aldose reductase-like gene. J Biol Chem 273: 11429-11435, 1998.
Chung S and LaMendola J. Cloning and sequence determination of human placental aldose reductase gene. J Biol Chem 264: 14775-14777, 1989.
Cooke CR and Steenburg RW. Effects of aldosterone and cortisol on the renal concentrating mechanism. J Lab Clin Med 82: 784-792, 1973.
Dmitrieva NI and Burg MB. Hypertonic stress response. Mutat Res 569: 65-74, 2005.
Esther CR Jr, Howard TE, Marino EM, Goddard JM, Capecchi MR and Bernstein KE. Mice lacking angiotensin-converting enzyme have low blood pressure, renal pathology, and reduced male fertility. Lab Invest 74: 953-965, 1996.
Garcia-Perez A and Burg MB. Renal medullary organic osmolytes. Physiol Rev 71: 1081-1115, 1991.
Gertner RA, Klein JD, Bailey JL, Kim DU, Luo XH, Bagnasco SM, and Sands JM. Aldosterone decreases UT-A1 urea transporter expression via the mineralocorticoid receptor. J Am Soc Nephrol 15: 558-565, 2004.
Gill JR Jr and Bartter FC. On the impairment of renal concentrating ability in prolonged hypercalcemia and hypercalciuria in man. J Clin Invest 40: 716-722, 1961.
Gui T, Tanimoto T, Kokai Y, and Nishimura C. Presence of a closely related subgroup in the aldo-ketoreductase family of the mouse. Eur J Biochem 227: 448-453, 1995.
Han KH, Woo SK, Kim WY, Park SH, Cha JH, Kim J, and Kwon HM. Maturation of TonEBP expression in developing rat kidney. Am J Physiol Renal Physiol 287: F878-F885, 2004.
Ho HT, Chung SK, Law JW, Ko BC, Tam SC, Brooks HL, Knepper MA, and Chung SS. Aldose reductase-deficient mice develop nephrogenic diabetes insipidus. Mol Cell Biol 20: 5840-5846, 2000.
Igarashi P, Shashikant CS, Thomson RB, Whyte DA, Liu-Chen S, Ruddle FH, and Aronson PS. Ksp-cadherin gene promoter. II. Kidney-specific activity in transgenic mice. Am J Physiol Renal Physiol 277: F599-F610, 1999.
Ito T, Fujio Y, Hirata M, Takatani T, Matsuda T, Muraoka S, Takahashi K, and Azuma J. Expression of taurine transporter is regulated through the TonE (tonicity-responsive element)/TonEBP (TonE-binding protein) pathway and contributes to cytoprotection in HepG2 cells. Biochem J 382: 177-182, 2004.
Jung JY, Kim YH, Cha JH, Han KH, Kim MK, Madsen KM, and Kim J. Expression of aldose reductase in developing rat kidney. Am J Physiol Renal Physiol 283: F481-F491, 2002.
Kinsella J, Cujdik T, and Sacktor B. Na + -H + exchange activity in renal brush border membrane vesicles in response to metabolic acidosis: The role of glucocorticoids. Proc Natl Acad Sci USA 81: 630-634, 1984.
Klein JD, Le Quach D, Cole JM, Disher K, Mongiu AK, Wang X, Bernstein KE, and Sands JM. Impaired urine concentration and absence of tissue ACE: involvement of medullary transport proteins. Am J Physiol Renal Physiol 283: F517-F524, 2002.
Knepper MA. Molecular physiology of urinary concentrating mechanism: regulation of aquaporin water channels by vasopressin. Am J Physiol Renal Physiol 272: F3-F12, 1997.
Ko BC, Turck CW, Lee KW, Yang Y, and Chung SS. Purification, identification, and characterization of an osmotic response element binding protein. Biochem Biophys Res Commun 270: 52-61, 2000.
Kultz D. Hypertonicity and TonEBP promote development of the renal concentrating system. Am J Physiol Renal Physiol 287: F876-F877, 2004.
Lal S, Szwergold BS, Taylor AH, Randall WC, Kappler F, and Brown TR. Production of fructose and fructose-3-phosphate in maturing rat lenses. Invest Ophthalmol Vis Sci 36: 969-973, 1995.
Lam AK, Ko BC, Tam S, Morris R, Yang JY, Chung SK, and Chung SSM. Osmotic response element binding protein (OREBP) is an essential regulator of urine concentrating mechanism. J Biol Chem 279: 48048-48054, 2004.
Lim SW, Li C, Sun BK, Han KH, Kim WY, Oh YW, Lee JU, Kador PF, Knepper MA, Sands JM, Kim J, and Yang CW. Long-term treatment with cyclosporine decreases aquaporins and urea transporters in the rat kidney. Am J Physiol Renal Physiol 287: F139-F151, 2004.
Lopez-Rodriguez C, Antos CL, Shelton JM, Richardson JA, Lin F, Novobrantseva TI, Bronson RT, Igarashi P, Rao A, and Olson EN. Loss of NFAT5 results in renal atrophy and lack of tonicity-responsive gene expression. Proc Natl Acad Sci USA 101: 2392-2397, 2004.
Manitius A, Levitin H, Beck D, and Epstein FH. On the mechanism of impairment of renal concentrating ability in potassium deficiency. J Clin Invest 39: 684-692, 1960.
Meneton P, Ichikawa I, Inagami T, and Schnermann J. Renal physiology of the mouse. Am J Physiol Renal Physiol 278: F339-F351, 2000.
Miyakawa H, Woo SK, Chen CP, Dahl SC, Handler JS, and Kwon HM. Cis - and trans -acting factors regulating transcription of the BGT1 gene in response to hypertonicity. Am J Physiol Renal Physiol 274: F753-F761, 1998.
Montrose-Rafizadeh C and Guggino WB. Cell volume regulation in the nephron. Annu Rev Physiol 52: 761-772, 1990.
Na KY, Woo SK, Lee SD, and Kwon HM. Silencing of TonEBP/NFAT5 transcriptional activator by RNA interference. J Am Soc Nephrol 14: 283-288, 2003.
Nakayama Y, Peng T, Sands JM, and Bagnasco SM. The TonE/TonEBP pathway mediates tonicity-responsive regulation of UT-A urea transporter expression. J Biol Chem 275: 38275-38280, 2000.
Oates PJ and Mylari BL. Aldose reductase inhibitors: therapeutic implications for diabetic complications. Expert Opin Investig Drugs 8: 2095-2119, 1999.
Peng T, Sands JM, and Bagnasco SM. Glucocorticoids inhibit transcription and expression of the UT-A urea transporter gene. Am J Physiol Renal Physiol 282: F853-F858, 2002.
Rane S, Aperia A, Eneroth P, and Lundin S. Development of urinary concentrating capacity in weaning rats. Pediatr Res 19: 472-475, 1985.
Sands JM, Terada Y, Bernard LM, and Knepper MA. Aldose reductase activities in microdissected rat renal tubule segments. Am J Physiol Renal Fluid Electrolyte Physiol 256: F563-F569, 1989.
Schwartz GJ, Zavilowitz BJ, Radice AD, Garcia-Perez A, and Sands JM. Maturation of aldose reductase expression in the neonatal rat inner medulla. J Clin Invest 90: 1275-1283, 1992.
Shao X, Johnson JE, Richardson JA, Hiesberger T, and Igarashi P. A minimal Ksp-cadherin promoter linked to a green fluorescent protein reporter gene exhibits tissue-specific expression in the developing kidney and genitourinary tract. J Am Soc Nephrol 13: 1824-1836, 2002.
Sheikh-Hamad D, Nadkarni V, Choi YJ, Truong LD, Wideman C, Hodjati R, and Gabbay KH. Cyclosporine A inhibits the adaptive responses to hypertonicity: a potential mechanism of nephrotoxicity. J Am Soc Nephrol 12: 2732-2741, 2001.
Smardo FL Jr, Burg MB, and Garcia-Perez A. Kidney aldose reductase gene transcription is osmotically regulated. Am J Physiol Cell Physiol 262: C776-C782, 1992.
Song Z, Fu DT, Chan YS, Leung S, Chung SS, and Chung SK. Transgenic mice overexpressing aldose reductase in Schwann cells show more severe nerve conduction velocity deficit and oxidative stress under hyperglycemic stress. Mol Cell Neurosci 23: 638-647, 2003.
Springate JE, Feld LG, Van Liew JB, Fildes RD, and Acara MA. Diuretic and natriuretic effects of sorbinil, an aldose reductase inhibitor. Pharmacol Res 23: 279-283, 1991.
Strange K. Cellular volume homeostasis. Adv Physiol Educ 28: 155-159, 2004.
Woo SK, Lee SD, Na KY, Park WK, and Kwon HM. TonEBP/NFAT5 stimulates transcription of HSP70 in response to hypertonicity. Mol Cell Biol 22: 5753-5760, 2002.
Yancey PH, Burg MB, and Bagnasco SM. Effects of NaCl, glucose, and aldose reductase inhibitors on cloning efficiency of renal medullary cells. Am J Physiol Cell Physiol 258: C156-C163, 1990.
作者单位：1 Institute of Molecular Biology and Department of Physiology, University of Hong Kong and 3 Division of Clinical Biochemistry, Queen Mary Hospital, Pokfulam, Hong Kong; 2 School of Life Sciences, Xiamen University, Xiamen, China; and 4 Renal Division, Emory University, Atlanta, Georgia