点击显示 收起
【摘要】 In the rat, urinary concentrating ability develops progressively during the third postnatal (P) week and nearly reaches adult level at weaning ( P21 ) governed by a rise in circulating glucocorticoid. Elevated extracellular osmolality can lead to growth arrest of epithelial cells. We tested the hypothesis that supranormal exposure of rat pups to glucocorticoid before the endogenous surge enhances urinary concentrating ability but inhibits renomedullary cell proliferation. Proliferating-cell nuclear antigen (PCNA)-positive cells shifted from the nephrogenic zone in the first postnatal week to Tamm-Horsfall-positive thick ascending limb (TAL) cells at the corticomedullary junction at P10 - 14. Renal PCNA protein abundance was stable in the suckling period and decreased 10-fold after weaning. Renal PCNA protein abundance decreased in response to dexamethasone (DEXA; 100 µg·kg -1 ·day -1, P8-12 ). Prolonged administration of DEXA ( P1-P11 ) reduced selectively the area and thickness of the outer medulla and the number of PCNA-positive cells. DEXA ( P8 - 12 ) increased urinary and papillary osmolality in normohydrated and water-deprived pups and led to osmotic equilibrium between interstitium and urine, whereas apoptotic and GADD153-positive cells increased in the inner medulla. TAL-associated NaCl transporters Na-K-2Cl cotransporter, Na-K-ATPase- 1, Na/H exchanger type 3, and ROMK increased significantly at weaning and in response to DEXA. We conclude that a low level of circulating glucocorticoid is permissive for proliferation of Henle's loop and the outer medulla before weaning. A reduced papillary tonicity is a crucial factor for the reduced capacity to concentrate urine during postnatal kidney development. We speculate that supranormal exposure to glucocorticoid in the suckling period can alter kidney medullary structure and function permanently.
【关键词】 mitosis urea osmolality apoptosis
IN THE RAT, NEPHROGENESIS continues for 7-8 days after birth and during this time urinary concentrating ability is virtually absent ( 12, 15, 31 ). A rapid and progressive increase in concentrating capacity develops during the third postnatal week, in particular just before weaning at 21 days ( 27, 36 ). Plasma glucocorticoid is important for this transition: plasma concentration increases 10- to 15-fold at weaning compared with postnatal day 10 ( 13, 27, 36 ) and adrenalectomy before weaning prevents the increase in concentrating capacity while supplementation with glucocorticoid accelerates it ( 27 ). By contrast, early exposure of the fetus to glucocorticoid leads to pathological changes in kidneys with a reduced number of glomeruli, cysts, and development of hypertension in adult animals, referred to as "programming" of adult hypertension ( 4, 24, 25, 29, 35 ). Preterm children treated postnatally with glucocorticoid display elevated blood pressure at an early age ( 14 ). Thus subnormal glucocorticoid exposure is mandatory for normal intra- and early extrauterine stages of rodent kidney development, whereas a rise in glucocorticoid to the normal adult level is essential for maturation of renal medullary function in later developmental stages after nephron formation. Little information is available about the effect of glucocorticoid on kidney growth and differentiation just after nephrogenesis has finished. After the surge at parturition, circulating glucocorticoids are at a nadir in the second postnatal week in rodents coincident with marked expansion of the renal medulla ( 5, 16, 13, 20 ). Thus only short-loop nephrons exist at birth ( 5, 16 ). Before maturation of urinary concentrating ability, loops of Henle undergo considerable elongation to achieve the adult conformation in the second and third postnatal weeks. This occurs through mitotic activity and apoptosis in descending and ascending limbs ( 5, 22, 33 ). It is unknown whether proliferation of cells in Henle's loop at this stage depends on low circulating glucocorticoid and whether proliferating cells are undifferentiated. Glucocorticoid increases concentrating ability at weaning ( 27 ), and elevated extracellular osmolality is genotoxic and may hamper cell proliferation ( 6, 18 ). This is one mechanism by which supranormal glucocorticoid potentially disturbs cell proliferation. The relevance of these in vitro observations for proliferation of Henle's loop in the renal medulla in vivo has not previously been examined.
In the present study, we determined changes in thick ascending limb (TAL)-associated cell proliferation and NaCl transport proteins in a postnatal period that covered the interval with maximal changes of endogenous glucocorticoid ( 13, 20 ). Moreover, we supplemented rats with glucocorticoid in a postnatal developmental "window" after nephrogenesis where endogenous plasma glucocorticoid is low and proliferation is intense [postnatal (P) days 8 - 12 ]. We explored the effect of "precocious" glucocorticoid on NaCl transporters, cell proliferation, apoptosis, and urinary concentrating ability. We tested the hypothesis that glucocorticoid enhances urinary concentrating ability but exerts negative effects on mitotic activity, induces a DNA damage response, and accelerates apoptotic cell death in medullary cells.
MATERIALS AND METHODS
Animal Experiments
All in vivo experiments and animal handling were approved by the Animal Experimentation Inspectorate under the Danish Ministry of Justice and were in accordance with the published guidelines from the National Institutes of Health. Female Sprague-Dawley rats used for breeding had free access to standard pathogen-free rat chow [Altromin-1310, Lage, Germany, Na + (2 g/kg), Cl - (5 g/kg)] and tap water. The dams and pups were housed with a 12:12-h light-dark cycle at 20°C until weaning at 3 wk ( P21 ) of age.
Developmental series. We examined rat pups at developmental stages that covered the period of maximal changes in endogenous glucocorticoid: the birth peak ( P0 ), nadir at days 8-12, and the 10- to 15-time increase at weaning on day 21 ( 13, 20 ). Pups were decapitated, and trunk blood was collected in EDTA-coated tubes and plasma was isolated by centrifugation and stored at -20°C. The kidneys were snap-frozen in liquid nitrogen and stored at -80°C. For immunohistochemistry, rats were anesthetized with intraperitoneal injections of pentobarbital sodium (mebumal, 50 mg/kg) and perfusion fixed with 4% phosphate-buffered paraformaldehyde for 4 min through the left cardiac ventricle at P2-P20. Kidneys were fixed for 4 h, then trimmed in coronal blocks with the papilla, washed in PBS, dehydrated, and then paraffin-embedded.
Glucocorticoid series. Litters were reduced to the sex-matched size of 10 pups at the day of birth ( n = 3 litters). Intervention started at the nadir of endogenous glucocorticoids, P8 ( 13 ). At P8, rats were divided into four groups receiving either high-dose dexamethasone [DEXA; 100 µg·kg -1 ·day -1 (Decadron, 4 mg/ml, Merck)]; low-dose DEXA (10 µg·kg -1 ·day -1 ); mifepristone, a glucocorticoid-receptor antagonist (5 mg·kg -1 ·day -1, Sigma); or vehicle (0.9% saline or sesame oil) once daily by subcutaneous injections. DEXA was diluted in isotonic saline, and all rats were injected with 5 µl/g. After 5 days of treatment, rats were decapitated, and blood, urine, and kidneys were collected as described above. Urine osmolality was determined by freeze-point depression (Osmomat 030-D, Gonotec, Bie and Berntsen) of 50-µl samples. Na and K concentration was determined by flame photometry (model IL 943, Instrumentation Laboratory, Lexington, MA). In another series ( n = 1 litter with 11 pups), rats were treated continuously from P1 to P11 with DEXA (100 µg·kg -1 ·day -1 ) or vehicle (isotonic saline), and kidneys were perfusion fixed for morphological analysis as described below.
Fluid deprivation series. Three litters of 10 sex-matched pups were given either high-dose DEXA (100 µg·kg -1 ·day -1 ) or vehicle (0.9% saline) in 5 µl/g once daily from P8 to P12. On P11, the rats were removed from the dams for 12-14 h. Rat pups at P12 do not urinate spontaneously ( 29 ). Pups voided by the gentle striking of the pubic region accompanied by application of suprapubic pressure after 6 h. Urine was collected again after 12 h when the pups were decapitated. Four pups from each group were left for additional 3 h (15 h in total), and then the pups were anesthetized with an intraperitoneal injection of pentobarbital sodium (mebumal, 50 mg/kg), and the urine was collected directly from the bladder before perfusion fixation.
mRNA Assays
Whole kidneys were used for isolation of total RNA with midi- or mini-RNeasy columns (Qiagen, Albertslund, Denmark) according to the manufacturer's instructions. RNA was quantified by spectrophotometry and stored at -80°C. RNA was analyzed by ribonuclease protection assay as described previously ( 1 ) or by quantitative PCR. For both assays, PCR amplification of cDNAs was performed with primer sets shown in Table 1 (Invitrogen). PCR products were cloned in vector pSP73 as described previously ( 1 ). cDNA was obtained by reverse transcription of 1 µg total RNA (iScript cDNA synthesis kit, Bio-Rad). Cloned cDNAs were sequenced (Applied Biosystems). Plasmids carrying cDNA inserts were linearized with Hin dIII and used as a template for synthesis of radiolabeled antisense RNA probes ( 1 ) or as standards for real-time quantitative PCR after serial dilution. For quantitative PCR, 50 ng cDNA in duplicate were used as a template and mixed with the respective primers ( Table 1 ) and iQ-SYBR Green Supermix (Bio-Rad) in a final volume of 25 µl. The mixture was denatured for 3 min at 95°C, and 44 cycles were run on an iQ-Thermocycler (Bio-Rad) as follows: denaturation, 30 s at 95°C, and annealing and extension, 45 s at 60°C. The standard curve was constructed by plotting threshold cycle (C t values) against serial dilutions of the linearized plasmid (10 -9 -10 -15 g). Specificity was established postrun for each plate setup by melting curve analysis. Random samples from each plate were loaded on agarose gels to confirm the amplification product. Plasmids for generating probes for Na/H exchanger type 3 (NHE3), renin, and -actin have previously been validated ( 1, 23 ).
Table 1. Primers used for PCR and subsequent cloning in plasmids
Western Blotting
Proteins were isolated from kidney tissue by homogenization of tissue for 30 s in buffer [100 mM Tris·HCl, pH 7.2, 10 mM EDTA, 1 mM DTT, and 0.1% Triton X-100, containing a complete protease inhibitor tablet (Roche)/10 ml homogenization buffer]. After 10-min incubation on ice, the samples were centrifuged at 10,000 g for 10 min at 4°C. For determination of Na-K-2Cl cotransporter (NKCC2), freshly dissected tissue was homogenized in a sucrose-imidazole buffer [0.3 M sucrose, 25 mM imidazole, 1 mM EDTA, pH 7.2, with protease inhibitors (0.02 M leupeptin, 0.4 M pefabloc) and phosphatase inhibitors (0.2 M ortho-vanadate, 0.2 M NaF, and 0.082 µg/µl okadaic acid)]. Cell debris was removed by centrifugation of the samples for 15 min at 4,000 g at 4°C. Protein concentration of the supernatants was determined by spectrometry according to the Bradford method (Bio-Rad protein assay reagent) using BSA as a standard. Aliquots (20-40 µg) were denatured for 5-10 min at 95°C, separated on a 4-12% PAGE gel, and blotted on to a polyvinylidene difluoride membrane. An additional gel was run in parallel and stained with simple blue (Bio-Rad) to ensure uniform loading. The membranes were blocked over night with 5% dry milk in TBST (20 mM Tris·HCl, 137 mM NaCl, 0.05% Tween 20, pH 7.6) and then incubated with the specific primary antibody appropriately diluted for rabbit anti-NKCC2 (1:2,000, Chemicon), rabbit anti-ROMK (1:2,000, Alomone), rabbit anti-Na-K-ATPase- 1 (1:1,000, Upstate), rabbit-anti-PCNA (1:2,000, Santa Cruz Biotechnology), and rabbit anti-GADD153 (1:500, Santa Cruz Biotechnology). After additional washes in TBST, the membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody (DAKO) diluted 1:2,000 in 5% milk-TBST for 1 h and visualized with an ECL chemiluminescence detection kit (Amersham). The relative densities of immunoreactive protein bands were evaluated by QuantityOne software (Bio-Rad).
Immunohistochemistry and Morphometry
Coronal sections through all regions including the tip of the papilla (4 µm) were prepared for immunostaining as previously described ( 32 ). Some antigens required retrieval by boiling for 20 min in either 10 mM citrate buffer, pH 6, or TEG buffer (100 mM Tris·HCl, pH 9, 10 mM EGTA). The sections were blocked for 1 h in 5% milk TBST, then incubated overnight at 4°C with primary antibodies diluted in 5% milk-TBST: anti-ROMK (1:500), anti-NKCC2 (1:1,000), anti-Na-K-ATPase- 1 (1:100), PCNA (1:100), and sheep anti-Tamm-Horsfall glycoprotein (THP; 1:200, Upstate). The following day, the sections were washed and incubated for 1 h with HRP-conjugated goat anti-rabbit antibody or HRP-conjugated rabbit anti-sheep antibody (DAKO) diluted 1:1,000 in 5% milk-TBST. The immunolabeling was visualized by incubation with 0.01% diaminobenzidine and 0.02% H 2 O 2 for 1-20 min. A negative control with omission of primary antibody was always run in parallel. After antigen retrieval by heating for 5 min in a pressure cooker, GADD153 labeling (1:1,000) was performed with the CSA-labeling amplification system (DAKO) according to instructions. A series of negative controls was performed: a primary antibody was incubated overnight at 4°C with 20 µg/ml peptide used for immunization and applied to the tissue sections and amplified exactly as described above; primary antibody was omitted and reacted as above; and, finally, an anti-rabbit secondary antibody (DAKO) instead of a primary antibody was applied and reacted. For double-labeling experiments for PCNA and THP, the basic procedure was as described above. The secondary Alexa Fluor 488-conjugated donkey anti-sheep (1:300, Molecular Probes) and Alexa Fluor 568-conjugated goat anti-rabbit antibodies (1:100, Molecular Probes) were incubated in the dark for 1 h and inspected with an Olympus BX51 microscope. Photos were captured with a Olympus DP50 camera and processed using Photoshop 6.0. For morphometry, kidney sections labeled for THP and NKCC2 were photographed and compound pictures of whole sections were produced using Corel Draw. Kidney cortex, outer medulla, and papilla area and thickness were determined by defining a field extending from a line between juxtamedullary glomeruli (cortex-outer medullary border) to a line that connects points where THP- and NKCC2-positive TALs disappear (outer medullary-inner medullary junction). PCNA-positive nuclei were counted within this area with the aid of Image Tool version 3.0 (National Institutes of Health).
Osmolality, Na +, and Urea Concentration in Papillary Tissue
Papillas were snap-frozen in liquid nitrogen. Osmolytes were extracted using a protocol from Kusano et al. ( 19 ). Papillas were weighed and vacuum dried until the weight had stabilized (2.5 h, difference was taken as water content), boiled in 150 µl ultrapure water in a water bath for 1 h, and centrifuged for 10 min at 13,000 g. The osmolality of the supernatants was measured by freeze-point depression (Osmomat 030-D, Gonotec, Bie & Berntsen), and sodium and potassium were measured by flame photometry (IL 943, flame photometer, Instrumentation Laboratory). Urea content was determined kinetically as the amount of NADH used over time measured by spectrometry (340 nm) after hydrolysis of urea by urease and formation of L -glutamate by glutamate dehydrogenase (ABX Pentra Urea CP, ABX Diagnostics).
Detection of Apoptosis
Apoptotic cells were identified in kidney sections using an immunohistochemical approach that recognizes DNA double-strand breaks [transferase-mediated dUTP nick-end labeling (TUNEL) technology]. The procedure was performed in accordance with the manufacturer's instructions (in situ cell death detection kit, Roche).
Statistics
All values are presented as means ± SE. We found no variations with gender in any of the analyzed data sets, and data were pooled. For multiple comparisons, one-way ANOVA was performed. When there was a significant difference between the groups in sets of data, a post hoc test by unpaired Student's t -test with Bonferroni's correction was used. Two groups were compared with an unpaired Student's t -test. P < 0.05 was considered statistically significant.
RESULTS
Localization and Quantification of Proliferating Cells in the Developing Kidney
At P2, immunoreactivity for PCNA was associated with cell nuclei in the nephrogenic zone (not shown), and at P7 weak labeling for PCNA was observed in the nephrogenic zone but also appeared in tubular cells, always restricted to nuclei, at the border between the cortex and outer medulla ( Fig. 1 A ). A shift in localization appeared at P10, when PCNA immunoreactivity increased in intensity and accumulated in medullary rays in the juxtamedullary cortex and tubules in the outer stripe of outer medulla ( Fig. 1 A ). At P21, PCNA-positive cells were observed predominantly in the kidney cortex and exhibited an even distribution ( Fig. 1 A ). PCNA immunoreactivity was scarce in the inner medulla and papilla at all stages (not shown). Western blotting for PCNA using whole kidney protein showed that PCNA abundance in adult animals ( P56 ) was 7-8% of the level at birth ( P0, Fig. 1 B ). Of note, PCNA decreased most markedly after weaning between P21 and P56 ( n = 2-3 at each stage, Fig. 1 B ). Serial sections of kidney at P10 showed that the majority of PCNA-positive cells in medullary rays displayed apical labeling for NKCC2 ( Fig. 1 C ). At P10, double labeling for PCNA (red) and the TAL marker THP (green) showed that the majority of PCNA-positive cells were coexpressed with THP ( Fig. 2 C ). Not all proliferating cells in medullary rays colocalized with THP, indicating either undifferentiated TAL cells or cells in non-TAL segments ( 5 ).
Fig. 1. A : distribution of proliferating-cell nuclear antigen (PCNA) in kidney through postnatal (P) developmental stages. PCNA was associated with cell nuclei at all analyzed stages of development. At P7, the nephrogenic zone displayed signals, whereas signals in more mature tubules were scarce. At P10, PCNA signal shifted to medullary rays in cortex and tubules at the corticomedullary junction in outer medulla. At weaning ( P21 ), PCNA-positive cells were fewer and scattered across cortex and outer medulla. Scale bars = 200 µm. B : Western blot analysis for PCNA using whole kidney protein revealed a band at expected size at 35 kDa ( n = 2 at each stage). Columns represent average densitometric units at each stage. C : serial sections of kidney at P10 labeled for PCNA ( a ) and Na-K-2Cl cotransporter (NKCC2; a' ). NKCC2 and PCNA were not mutually exclusive. Arrowheads point at 3 typical cells positive for both PCNA and NKCC2. Scale bars = 50 µm.
Fig. 2. Effect of dexamethasone (DEXA; 100 µg·kg -1 ·day -1, P8-P12 ) on PCNA. A : Western blots for PCNA in whole kidney ( left, n = 5-6) and in deep cortex-outer stripe fraction ( right, n = 6). OD, optical density. * P < 0.05, DEXA compared with control. B : effect of DEXA (100 µg·kg -1 ·day -1, P8-P12 ) on intrarenal distribution of immunoreactive PCNA protein. Histochemical PCNA signal is seen without counterstain. PCNA immunoreactivity was less widely distributed and less intense in response to DEXA ( right ). Scale bars = 200 µm. C : double labeling of control kidney ( P12, left ) and DEXA-treated kidney ( P12, right ) for PCNA (red) and the thick ascending limb marker Tamm-Horsfall glycoprotein (THP; green). The vast majority of PCNA-positive cells were associated with the thick ascending limb in both conditions. Scale bars = 50 µm.
Effect of DEXA on Cell Proliferation, Apoptosis, and DNA Damage Proteins in Kidney During Postnatal Development
Body weight gain in response to DEXA treatment was 2.3 ± 0.6 g ( P = 5.7 x 10 -5 compared with control) and 5.7 ± 0.7 g ( P = 0.021 compared with control) in the two treatment groups (control pups 8.6 ± 0.8 g). Mifepristone (5 mg/kg) did not change body weight gain between P8 and P12 (9.4 ± 0.9 g, P = 0.53 compared with control, n = 7-8 pups/condition in 3 litters). Total kidney mass was not significantly changed by DEXA treatment from P8 to P12 (287 ± 7.7 and 305 ± 4.6 mg in DEXA groups, respectively, and 304 ± 12.8 mg in mifepristone vs. 298 ± 5.3 mg in controls, n as above). Thus the kidney weight-to-body weight ratios were significantly increased in response to DEXA (12.5 ± 0.3 mg/g, P = 0.011 and 13.7 ± 0.5 mg/g, P = 0.001) compared with controls (11.0 ± 0.3 mg/g) and mifepristone (10.9 ± 0.3 mg/g, P = 0.87 vs. control).
PCNA abundance decreased after DEXA when measured in whole kidney homogenate and in the deep cortex-outer stripe of outer medulla fraction ( Fig. 2, A and B, n = 5-6). Both the area of distribution and the intensity of PCNA labeling appeared decreased compared with control kidneys ( Fig. 2 B ). PCNA-immunopositive nuclei were very rarely observed in the inner medulla in all conditions ( Fig. 2 B ). After DEXA treatment, the majority of PCNA-positive cells in the outer medulla still colocalized with THP ( Fig. 2 C, right, n = 2).
In kidney sections from control pups, labeling for apoptosis by TUNEL reaction yielded very few positive cell nuclei and apoptotic bodies scattered primarily in the inner medulla near the border to the outer medulla ( Fig. 3 A ). After DEXA, the number of TUNEL-positive nuclei increased within the inner medulla ( Fig. 3 A ). TUNEL-labeled cells were observed in bending tubules ( Fig. 3 A ), most likely thin limbs of Henle's loop because they did not colocalize with THP-positive cells (not shown). At P12, the urea-inducible growth-arrest and DNA damage protein GADD153 was associated with the nuclei of collecting ducts, interstitial cells, and loop of Henle cells ( Fig. 3 B ). GADD153 immunoreactivity was strongest in the inner medulla ( Fig. 3 B ). After DEXA treatment, GADD153 immunolabeling was more intense particularly in the inner medulla ( Fig. 3 B, right ). Preabsorption of GADD153 antibody with peptide used for immunization prevented immunolabeling (not shown). Western blotting for GADD153 confirmed that GADD153 abundance was enhanced after DEXA ( Fig. 3 B, n = 5). As previously published ( 38 ), GADD153 immunoblotting produces a slower-migrating nonspecific band ( Fig. 3 B ) that was not changed by experimental conditions.
Fig. 3. A : labeling for transferase-mediated dUTP nick-end labeling (TUNEL) shows an increase in number of apoptotic cells and cell bodies after DEXA ( right, P8-P12 ) compared with control ( left ). Apoptotic cells and bodies were primarily located in inner medullary loop of Henle cells at the border with the outer medulla. Counting of TUNEL-positive cells showed a higher number after DEXA treatment (bar graph, n = 2). Scale bars = 50 µm. B : in control kidney at P12, GADD153 immunoreactivity was found in the nuclei of collecting duct cells, interstitial cells, and loop of Henle cells with strongest intensity in inner medulla ( n = 2). DEXA treatment yielded more intense labeling of cell nuclei in inner medulla only ( right ). Bar graph shows the effect of DEXA on GADD153 protein abundance in whole kidney homogenate ( n = 5). * P < 0.05, DEXA compared with control.
Effect of Prolonged DEXA on Postnatal Kidney Development
Administration of DEXA through P1 - 11 resulted in a significantly shorter corticomedullary axis (control 4.6 ± 0.2 mm, n = 3 vs. DEXA 3.8 ± 0.4 mm, n = 4, P < 0.04). Labeling for THP was used to delineate the borders between the outer and inner medulla in morphometric analysis of coronal sections ( Fig. 4 A ). DEXA did not change the area (5.9 ± 0.5 vs. 5.6 ± 0.6 mm 2 ) or thickness (0.69 ± 0.05 vs. 0.63 ± 0.01 mm) of the cortex. DEXA also had no effect on the length (2.7 ± 0.2 vs. 2.9 ± 0.1 mm, from tip to base) or area of the inner medulla (2.8 ± 0.2 vs. 3.2 ± 0.07 mm 2 ). In contrast, DEXA treatment significantly reduced the area and thickness of the renal outer medulla ( Fig. 4, A and B ). DEXA reduced significantly the number of PCNA-positive cell nuclei in the outer medulla ( Fig. 4 B ). To exclude the possibility that a change in THP labeling by DEXA caused the effect on outer medullary growth, all kidney sections were labeled for TAL-specific NKCC2. NKCC labeling yielded that same pattern of intrarenal distribution (not shown). The use of these markers is not likely to result in underestimation of outer medullary dimensions.
Fig. 4. Effect of DEXA on kidney morphology. A : compound micrographs show coronal sections of kidneys from rats treated with vehicle (control, left ) and DEXA ( right ) from P1 through P11 and labeled for THP. The top dashed line indicates the border between the cortex and outer medulla as determined by the localization of juxtamedullary glomeruli. The bottom dashed line indicates the border between the outer and inner medulla determined as the sites of disappearance of THP labeling. Bars = 500 µm. B : quantitative evaluation of kidney dimensions. The thickness of outer medulla (OM) was determined as the distance between the 2 dashed lines that mark the borders of the outer medulla in A. Outer medullary area was determined as the area of the field indicated in A, and the no. of PCNA-positive nuclei within this area was determined. * P < 0.05 [ n = 3 (control), n = 4 (DEXA)].
Effect of DEXA from P8 to P12 on Urinary Concentrating Ability
Osmolality of spot urine at P12 was 347 ± 32 mosmol/kgH 2 O ( n = 6, Fig. 5 A ). In response to both doses of DEXA, spot urine osmolality increased ( Fig. 5 A shows the effect of 100 µg/kg, n = 6, control, n = 8). Mifepristone (5 mg·kg -1 ·day -1 ) had no effect on urine osmolality at P12 compared with control (380 ± 9 mosmol/kgH 2 O, n = 7 vs. 347 ± 32 mosmol/kgH 2 O, n = 7, not shown). Water deprivation (12 h) caused an increase in urine osmolality compared with hydrated rats ( Fig. 5 A ). Additional dehydration for 3 h on an empty bladder did not increase urine osmolality compared with 12 h ( Fig. 5 A ). Water deprivation of DEXA-treated rats increased urine osmolality significantly after 12 and 15 h compared with water-deprived, vehicle-injected pups and DEXA-treated, hydrated rats ( Fig. 5 A ). The papillary water content was determined to be 86 ± 0.5% ( n = 10) in hydrated pups and 76 ± 1% ( n = 8) after fluid deprivation. Papillary interstitial osmolality in control pups at P12 was 537 ± 27 mosmol/kgH 2 O ( n = 4, Fig. 5 B ). DEXA increased papillary osmolality significantly ( Fig. 5 B ). Water deprivation did not augment papillary osmolality significantly in control pups or DEXA-treated pups ( Fig. 5 B ). There was not equilibrium between spot urine osmolality and papillary osmolality in control pups ( Fig. 5, A and B ). After dehydration, equilibrium was reached (urine: 568 ± 17 mosmol/kgH 2 O, n = 15; papilla: 565 ± 23 mosmol/kgH 2 O, n = 5). DEXA led to equilibrium between interstitium and urine in hydrated rats (papilla 728 ± 48 mosmol/kgH 2 O, n = 4 vs. 614 ± 84 mosmol/kgH 2 O, n = 8, P = 0.32, Fig. 5, A and B ) and to equilibrium after dehydration (papilla 754 ± 70 mosmol/kgH 2 O, n = 5 vs. 809 ± 40 mosmol/kgH 2 O, n = 14, P = 0.49). Urea concentration was significantly higher in papillary interstitial fluid after DEXA in hydrated and water-deprived rats ( Table 2 ). There was no statistically significant difference in sodium and potassium concentrations ( Table 2 ). Renal mRNA levels for the organic osmolyte transporters sodium-dependent myo -inositol transporter (SMIT) and betaine- -amino- N -butyric acid transporter (BGT-1) were significantly elevated in response to DEXA (100 µg·kg -1 ·day -1, Fig. 5 C ).
Fig. 5. Effect of DEXA on urinary concentrating ability. A : effect of DEXA (100 µg·kg -1 ·day -1 at P8-P12 ) on osmolality of spot urine and urine collected after water deprivation. Histogram shows average value of 6-8 determinations. * P < 0.05 compared with vehicle treatment. # P < 0.05 compared with hydrated rats. B : effect of DEXA (100 µg·kg -1 ·day -1 at P8-P12 ) on papillary interstitial osmolality in control pups ( n = 4) and water-deprived pups (12 h, n = 5). *Significant difference ( P < 0.05) between controls and DEXA. C : effect of DEXA (100 µg·kg -1 ·day -1 at P8-P12 ) on osmolyte transporter mRNAs. Messenger RNA level was determined for betaine -amino- N -butyric acid transporter type 1 (BGT1) and Na + - myo -inositol transporter (SMIT) by quantitative real-time PCR using total RNA from whole kidneys ( n = 7-8). TATA box binding protein (TBP) was amplified as a reference gene. Values are normalized for TBP. SQ, starting quantity of cDNA. * P < 0.002, DEXA compared with control.
Table 2. Effect of dexamethasone (100 µg·kg -1 ·day -1 at postnatal days 8-12) and water deprivation on papillary tissue concentration of osmolytes
Developmental Regulation and Effect of DEXA on TAL-Associated Sodium and Chloride Transport Molecules
Ribonuclease protection assays were linear, at least in the range 10-40 µg of total RNA, and the hybrids and probes displayed the expected molecular sizes (data not shown). There were significant developmental increases in mRNA levels of the TAL-associated transporters ROMK, NHE3, and Na-K-ATPase 1 -subunit between P7 (when circulating glucocorticoids are at nadir) and P21 ( Fig. 6 A ). After weaning, transport molecule mRNAs remained stable or decreased slightly at P28 and P56 (data not shown). NKCC2 was developmentally regulated and increased significantly between P0 and P21 ( Fig. 6 A ), followed by stabilization at P28 and P56 (data not shown). ClCK2 mRNA expression was not developmentally regulated. Many "housekeeping" genes are developmentally regulated, and we used renin as a "biological" RNA control that is known to be downregulated during development. Renin mRNA changed inversely with development ( Fig. 6 A ) ( 10, 31 ). The differential regulation of renin and NaCl transporters suggests that the changes are not caused by variation in RNA quality ( Fig. 6 A ). Moreover, Western blot analysis confirmed that NKCC2 protein increased significantly between P7 and P21 and migrated as a band of 150 kDa ( Fig. 6 B, n = 3-4). Na-K-ATPase 1 -subunit protein migrated as a band of 110 kDa and increased between P7 and P21 ( Fig. 6 B, n = 3-4). ROMK appeared as two distinct bands, one 40 kDa, which corresponds to the theoretic size, and a product of unknown significance at 90 kDa (not shown), which has been found by several other investigators ( 9, 36 ). The 40-kDa protein significantly increased between P7 and P21 ( Fig. 6 B, n = 3-4). For all blots, a gel was run in parallel and proteins were stained with simple blue (Bio-Rad) to ensure uniform loading and migration (not shown). Immunohistochemical staining showed that NKCC2-immunoreactive protein was at the limit of detection in cross sections of the renal outer medulla at P2 ( Fig. 6 C ). At P10, NKCC2 labeling was more intense and more widely distributed along the apical aspect of tubules within medullary rays in the cortex and was especially marked in the outer medulla ( Fig. 6 C ). At P21, strong and uniform NKCC2 immunoreactivity was localized in the apical membrane of the medullary and cortical TAL cells ( Fig. 6 C ).
Fig. 6. A : ribonuclease protection assays for NaCl transporter mRNAs in kidney at various stages of postnatal development. Autoradiographs show the result of assays in which total RNA from postnatal rat kidneys hybridized with radiolabeled specific antisense probes. The histograms show average counts per minute (cpm) for the excised protected probes ( n = 3-5 at each developmental stage). a : Na-K-ATPase 1 -subunit. b : NKCC2. c : ROMK. d : Na/H exchanger type 3. e : ClCK. f : renin. * P < 0.05. B : developmental change in loop of Henle NaCl transport proteins between birth ( P0 ) and weaning ( P21 ) as determined by densitometric evaluation of Western blots. Immunoreactive proteins displayed the expected size of 150, 110, and 40 kDa, respectively ( n = 3-4 at each stage). Gels with similar amounts of protein loaded as used for Western blots always ran in parallel to ensure uniform loading by simple blue staining. * P < 0.05. C : cellular distribution of NKCC2 immunoreactive protein in kidney investigated at the indicated postnatal days. At P21 (weaning), NKCC2 protein was uniformly associated with apical membrane, whereas labeling was less pronounced at earlier stages. Scale bars = 50 µm.
Renal mRNA levels of NKCC2, ROMK, Na-K-ATPase 1 -subunit, and NHE3 were significantly increased by DEXA (100 µg·kg -1 ·day -1 ) compared with vehicle-injected controls ( Fig. 7 A, n = 7-8). At low-dose DEXA (10 µg·kg -1 ·day -1 ), there were no significant changes (not shown). There was no effect of mifepristone treatment on the analyzed renal transporters at this stage of development (data not shown). -Actin mRNA abundance was significantly lowered in response to both doses of DEXA as reported previously ( 32 ) ( Fig. 7 A ). The inverse change in NaCl transporter mRNAs and -actin mRNA indicates specific responses to DEXA treatment and not a general difference in RNA quality or loading. Experiments in whole kidney homogenate confirmed that Na-K-ATPase 1 -subunit and NKCC2 protein levels increased significantly by DEXA (100 µg·kg -1 ·day -1 ) compared with control ( Fig. 7 B, n = 5-6/condition). ROMK protein abundance was not affected significantly by DEXA treatment ( Fig. 7 B, n = 5-6). Immunohistochemical labeling showed that NKCC2-immunoreactive protein distribution changed after DEXA. NKCC2 immunoreactivity was more intense and uniformly associated with apical membranes of TAL cells in cortical medullary rays and the outer medulla ( Fig. 7 C ).
Fig. 7. A : autoradiographs display the result of protection assays for NaCl transporter mRNAs, as indicated, in response to DEXA (100 µg·kg -1 ·day -1 at P8-P12 ). Twenty micrograms total kidney RNA were used for the assays ( n = 7-8). Histograms display the quantitative evaluation of the protection assays in cpm. a : Na-K-ATPase 1 -subunit. b : ROMK. c : NKCC-2. d : NHE3. e : CLCK2. f : -actin. * P < 0.05, DEXA compared with control. B : Western blots show the effect of DEXA treatment (100 µg·kg -1 ·day -1 ) on whole kidney abundance of loop of Henle transport proteins. Twenty- to forty-microgram protein aliquots were used for the assays ( n = 5-6). * P < 0.05. C : effect of DEXA (100 µg·kg -1 ·day -1, P8-P12, n = 2) on renal distribution of immunoreactive NKCC2. In response to DEXA, NKCC2-immunoreactive protein was more widely distributed compared with control and displayed a uniform association with apical membranes. Scale bars = 200 µm.
DISCUSSION
The present study shows that brief exposure of rat kidney to a synthetic glucocorticoid receptor agonist (DEXA) in the suckling period ( P8-P12 ) accelerates loop of Henle differentiation but inhibits cell proliferation in the TAL. Continuous exposure of the kidney to DEXA from birth to P11 selectively impairs development of renal outer medulla. DEXA converts the TAL epithelium to an adultlike phenotype and increases papillary tonicity, which, together with increased water permeability in the collecting duct, leads to an enhanced urinary concentrating capacity. It is well established that intact adrenal function is crucial for development of urinary concentrating capacity at weaning ( 27 ) and glucocorticoid supplementation increases concentrating ability only in the period before weaning and not at weaning or in adult rats ( 27 ). An array of TAL-associated NaCl transporters are upregulated and display stronger membrane localization with postnatal development and by DEXA in preweaned developmental stages. This effect could be one pathway by which glucocorticoids stimulate urinary concentrating ability as indicated in previous studies ( 2, 3, 9, 11, 26, 30, 37 ).
An antagonist of the glucocorticoid receptor had no effect on growth rate, urinary concentrating ability, and transporter abundance in the second postnatal week. Thus at the nadir of corticosterone ( P8-P12 ) ( 13, 27, 25 ), the plasma level appears to be below the threshold for exerting physiological effects on kidney transport proteins and urinary concentrating ability. In this glucocorticoid-deficient stage of development referred to as the "stress-hyporesponsive period," the hypothalamic-pituitary-adrenal axis fails to elicit increases in ACTH and glucocorticoid when challenged, the adrenals are nonresponsive to exogenous ACTH, and the normal circadian rhythm is not established ( 28, 29 ). Coincident with the glucocorticoid-deficient stage, proliferating cells wane in the nephrogenic zone and accumulate in loop of Henle epithelium as indicated by the markers PCNA and THP. Of note, PCNA-positive cells in medullary rays also express NKCC2, which suggests that dividing cells are differentiated. DEXA shortened the corticopapillary axis, and low glucocorticoid exposure appears especially important for development of the outer medulla. The papilla exhibited no difference in size or area after continuous exposure to DEXA. It is not resolved by the present experiments whether these morphological changes in the outer medulla are caused primarily by decreased proliferation or increased apoptotic conversion of TALs to thin ascending limbs ( 4, 16 ), or both. Moreover, it is not possible to discriminate between an indirect effect of DEXA through, e.g., elevated interstitial osmolality or impaired angiogenesis, and a direct effect. Interstitial osmolality did not increase by the short dehydration protocol applied in the present experiments. The role of osmolality in medullary growth in the absence of changes in glucocorticoid is therefore not clear. Proliferating cells were very scarce in the immature papilla at all stages, which indicates that regions with elevated interstitial osmolality and cell proliferation are mutually exclusive in the developing kidney. DEXA is a synthetic, selective glucocorticoid-receptor agonist, and the doses used were lower than in similar studies on postnatal rat kidney function with the equipotent betamethasone ( 27, 36 ) and within the range used to treat women at risk of preterm delivery. It is not possible from the present data to extrapolate the level of endogenous glucocorticoid at which similar changes in kidney function and structure occur.
DEXA increased the abundance and apical membrane association of NKCC2 and Na-K-ATPase. Previous data show that glucocorticoids enhance loop of Henle mitochondrial respiratory capacity and Na-K-ATPase activity ( 7, 8, 26 ). Theoretically, each of these effects strengthens countercurrent amplification capacity. Indeed, DEXA raised medullary tonicity, urinary concentrating ability, and expression of the organic osmolyte transporters SMIT and BGT-1. Although "precocious" exposure to elevated glucocorticoid increased urinary concentrating ability (to 800 mosmol/kgH 2 O), it did not recapitulate the increase in medullary tonicity achieved spontaneously at weaning ( 2,000 mosmol/kgH 2 O) ( 27 ). The difference in absolute level of osmolality most likely reflects the incomplete elongation of Henle's loops at the early stage of P12. DEXA increased papillary tonicity by accumulation primarily of urea to a level normally encountered first at P20 ( 27 ). This finding is in accord with upregulation of urea-inducible transcription factor GADD153 specifically in inner medullary cell nuclei. The enhanced accumulation of urea could be caused by an increase in urea transporter UTA ( 17 ), by an increase of urea availability, and by an increased urea concentration in collecting duct fluid.
The short dehydration protocol applied in our experiments (12-15 h) led to equilibration between urinary and interstitial osmolality. The pups must have the capacity to activate aquaporins similar to later developmental stages. Water permeability of the collecting ducts is therefore not the limiting factor that causes a reduced urinary concentrating capacity at early stages of postnatal kidney development. Rather, the limited interstitial medullary gradient is responsible for a lower concentrating ability. Medullary tonicity and papillary sodium and urea content exhibit a developmental increase similar to urinary concentrating ability ( 27, 34 ). DEXA led to equilibration between urinary and papillary tonicity in normohydrated rats, suggesting an effect of DEXA on collecting duct water permeability that is in accord with previous observations ( 36 ).
The number of apoptotic cells in the medullary, THP-negative part of Henle's loop, was increased by DEXA. In general, apoptotic bodies were rare but have previously been found at this site and ascribed a role in the normal transformation of the TAL into thin ascending limb at the outer medullary-papillary border ( 16 ). Exposure to DEXA and/or elevated interstitial osmolality accelerates this remodeling process further during postnatal development. The present findings could be relevant in the context of "programmed" hypertension. Thus in rats with restricted protein intake during pregnancy, NKCC2 is permanently elevated in kidneys in the postnatal period ( 21 ). The role of glucocorticoid in this response is unknown. The present studies do not show whether termination of glucocorticoid exposure results in catch-up growth, but future studies should clarify whether adult rats exposed to glucocorticoid through late developmental stages exhibit changes in kidney morphology, function, and blood pressure regulation. In summary, we have observed that a low circulating concentration of glucocorticoid is permissive for proliferation of Henle's loop and development of the outer medulla in the postnatal stages that follow nephrogenesis and precede weaning. This occurs at the "cost" of reduced papillary tonicity, which appears as an important factor for the reduced capacity to concentrate urine during postnatal stages of kidney development.
GRANTS
The present work received grant support from the Danish Research Council for Health and Disease (22040305), the NOVO Nordisk Foundation, The Danish Kidney Association, the A. J. Andersen og Hustrus Fond, Jens C. Christoffersens Mindefond, and Else Poulsens Mindefond.
ACKNOWLEDGMENTS
The technical assistance of Mette Fredenslund, Inge Andersen, Lis Teusch, and Mette Svendsen is gratefully acknowledged.
【参考文献】
Andreasen D, Jensen BL, Hansen PB, Kwon TH, Nielsen S, and Skøtt O. The 1G -subunit of voltage-dependent Ca 2+ channel is localized in rat distal nephron and collecting duct. Am J Physiol Renal Physiol 279: F997-F1005, 2000.
Attmane-Elakeb A, Sibella V, Vernimmen C, Belenfant X, Hebert SC, and Bichara M. Regulation by glucocorticoids of expression and activity of rBSC1, the Na + -K + (NH 4 + )-2Cl - cotransporter of medullary thick ascending limb. J Biol Chem 275: 33548-33553, 2000.
Baum MD, Biemesderfer D, Gentry D, and Aronson PS. Ontogeny of rabbit renal cortical NHE3 and NHE1: effect of glucocorticoids. Am J Physiol Renal Fluid Electrolyte Physiol 268: F815-F820, 1995.
Celsi G, Kistner A, Aizman R, Eklof AC, Ceccatelli S, de Santiago A, and Jacobson SH. Prenatal dexamethasone causes oligonephronia, sodium retention, and higher blood pressure in the offspring. Pediatr Res 44: 317-322, 1998.
Cha JH, Kim YH, Jung JY, Han KH, and Madsen KM, and Kim J. Cell proliferation in the loop of Henle in the developing rat kidney. J Am Soc Nephrol 12: 1410-1421, 2001.
Chakravarty D, Cai Q, Ferraris JD, Michea L, Burg MB, and Kultz D. Three GADD45 isoforms contribute to hypertonic stress phenotype of murine renal inner medullary cells. Am J Physiol Renal Physiol 283: F1020-F1029, 2002.
Djouadi F, Wijkhuisen A, and Bastin J. Coordinate development of oxidative enzymes and Na-K-ATPase in thick ascending limb: role of corticosteroids. Am J Physiol Renal Fluid Electrolyte Physiol 263: F237-F242, 1992.
Djouadi F, Bastin J, Kelly DP, and Merlet-Benichou C. Transcriptional regulation by glucocorticoids of mitochondrial oxidative enzyme genes in the developing rat kidney. Biochem J 315: 555-562, 1996.
Gallazzini M, Attmane-Elakeb A, Mount DB, Hebert SC, and Bichara M. Regulation by glucocorticoids and osmolality of expression of ROMK (Kir 1.1), the apical K channel of thick ascending limb. Am J Physiol Renal Physiol 284: F977-F986, 2003.
Gomez RA, Lynch KR, Sturgill BC, Elwood JP, Chevalier RL, Carey RM, and Peach MJ. Distribution of renin mRNA and its protein in the developing kidney. Am J Physiol Renal Fluid Electrolyte Physiol 257: F850-F858, 1989.
Gupta N, Tarif SR, Seikaly M, and Baum M. Role of glucocorticoids in the maturation of the rat renal Na + /H + antiporter (NHE3). Kidney Int 60: 173-181, 2001.
Heller H. Effects of dehydration on adult and newborn rats. J Physiol 108: 303-314, 1949.
Henning SJ. Plasma concentrations of total and free corticosterone during development in the rat. Am J Physiol Endocrinol Metab Gastrointest Physiol 235: E451-E456, 1978.
Kari MA, Hallman M, Eronen M, Teramo K, Virtanen M, Koivisto M, and Ikonen RS. Prenatal dexamethasone treatment in conjunction with rescue therapy of human surfactant: a randomized placebo-controlled multicenter study. Pediatrics 93:730-736, 1994.
Kazimierczak J. Histochemical observations of the developing glomerulus and juxtaglomerular apparatus. Acta Pathol Microbiol Scand 78: 401-413, 1970.
Kim J, Lee GS, Tisher C, and Madsen KM. Role of apoptosis in development of the ascending thin limb of the loop of Henle in rat kidney. Am J Physiol Renal Fluid Electrolyte Physiol 271: F831-F845, 1996.
Kim YH, Kim DK, Han KH, Jung JY, Sands JM, Knepper MA, and Madsen KM. Expression of urea transporters in developing rat kidney. Am J Physiol Renal Physiol 282: F530-F540, 2002.
Kültz D, Madhany S, and Burg MB. Hyperosmolality causes growth arrest of murine kidney cells. Induction of GADD45 and GADD153 by osmosensing via stress-activated protein kinase 2. J Biol Chem 273: 13645-13651, 1998.
Kusano E, Braun-Werness JL, Vick DJ, Keller MJ, and Dousa TP. Chlorpropamide action on renal concentrating mechanism in rats with hypothalamic diabetes insipidus. J Clin Invest 72: 1298-1313, 1983.
Madsen K, Stubbe J, Yang T, Skøtt O, Bachmann S, and Jensen BL. Low endogenous glucocorticoid allows induction of kidney cortical cyclooxygenase-2 during postnatal rat development. Am J Physiol Renal Physiol 286: F26-F37, 2004.
Manning J, Beutler K, Knepper MA, and Vehaskari VM. Upregulation of renal BSC1 and TSC in prenatally programmed hypertension. Am J Physiol Renal Physiol 283: F202-F206, 2002.
Marquez MG, Cabrera I, Serrano DJ, and Sterin-Speziale N. Cell proliferation and morphometric changes in the rat kidney during postnatal development. Anat Embryol 205: 431-440, 2002.
Norregaard R, Uhrenholt TR, Bistrup C, Skøtt O, and Jensen BL. Stimulation of 11 -hydroxysteroid dehydrogenase type 2 in rat colon but not in kidney by low dietary NaCl intake. Am J Physiol Renal Physiol 285: F348-F358, 2003.
Ortiz LA, Quan A, Weinberg A, and Baum M. Effect of prenatal dexamethasone on rat renal development. Kidney Int 59: 1663-1669, 2001.
Ortiz LA, Quan A, Zarzar F, Weinberg A, and Baum M. Prenatal dexamethasone programs hypertension and renal injury in the rat. Hypertension 41: 328-334, 2003.
Rane S and Aperia A. Ontogeny of Na-K-ATPase activity in thick ascending limb and of concentrating capacity. Am J Physiol Renal Fluid Electrolyte Physiol 249: F723-F728, 1985.
Rane S, Aperia A, Eneroth P, and Lundin S. Development of urinary concentrating capacity in weaning rats. Pediatr Res 19: 472-475, 1985.
Rosenfeld P, Suchecki D, and Levine S. Multifactorial regulation of the hypothalamic-pituitary-adrenal axis during development. Neurosci Biobehav Rev 16: 553-568, 1992.
Slotkin TA, Seidler FJ, Kavlock RJ, and Bartolome JV. Fetal dexamethasone exposure impairs cellular development in neonatal rat heart and kidney: effects on DNA and protein in whole tissues. Teratology 43: 301-306, 1991.
Schmidt U and Horster M. Na-K-activated ATPase: activity maturation in rabbit nephron segments dissected in vitro. Am J Physiol Renal Fluid Electrolyte Physiol 233: F55-F60, 1977.
Speller AM and Moffat DB. Tubulo - vascular relationships in the developing kidney. J Anat 123: 487-500, 1977.
Stubbe J, Jensen BL, Bachmann S, Morsing P, and Skøtt O. Cyclooxygenase-2 contributes to elevated renin in the early postnatal period in rats. Am J Physiol Regul Integr Comp Physiol 284: R1179-R1189, 2003.
Trimble ME. Renal response to solute loading in infant rats: relation to anatomical development. Am J Physiol 219: 1089-1097, 1970.
Valtin H. Sequestration of urea and nonurea solutes in renal tissues of rats with hereditary hypothalamic diabetes insipidus: effect of vasopressin and dehydration on the countercurrent mechanism. J Clin Invest 45: 337-345, 1966.
Wintour EM, Moritz KM, Johnson K, Ricardo S, Samuel CS, and Dodic M. Reduced nephron number in adult sheep, hypertensive as a result of prenatal glucocorticoid treatment. J Physiol 549: 929-935, 2003.
Yasui M, Marples D, Belusa R, Eklöf AC, Celso G, Nielsen S, and Aperia A. Development of urinary concentrating capacity: role of aquaporin-2. Am J Physiol Renal Fluid Electrolyte Physiol 271: F461-F468, 1996.
Zolotnitskaya A and Satlin LM. Developmental expression of ROMK in rat kidney. Am J Physiol Renal Physiol 276: F825-F836, 1999.
Zhang Z, Yang XY, and Cohen DM. Urea-associated oxidative stress and Gadd153/CHOP induction. Am J Physiol Renal Physiol 276: F786-F793, 1999.
作者单位：Department of Physiology and Pharmacology, University of Southern Denmark, Odense, Denmark