点击显示 收起
【摘要】 Chronic renal injury can be mediated by angiotensin II (ANG II) and prostanoids through hemodynamic and inflammatory mechanisms and attenuated by individual suppression of these mediators. In rats with renal ablation (Nx), we investigated 1 ) the intrarenal distribution of COX-2, ANG II, and the AT 1 receptor (AT 1 R); 2 ) the renoprotective and antiinflammatory effects of an association between the AT 1 R blocker, losartan (Los), and the gastric sparing anti-inflammatory nitroflurbiprofen (NOF). Adult male Munich-Wistar rats underwent Nx or sham operation (S), remaining untreated for 30 days, after which renal structure was examined in 12 Nx rats (Nx pre ). The remaining rats were followed during an additional 90 days, distributed among 4 treatment groups: Nx V (vehicle), Nx Los (Los), Nx NOF (NOF), and Nx Los/NOF (Los/NOF). Nx pre rats exhibited marked albuminuria, hypertension, glomerulosclerosis, interstitial expansion, and macrophage infiltration, accompanied by abnormal glomerular, vascular, and interstitial COX-2 expression. ANG II appeared in interstitial cells, in contrast to S, in which ANG II was virtually confined to afferent arterioles. Intrarenal AT 1 R distribution shifted from mostly tubular in S to predominantly interstitial in Nx pre. All these changes were aggravated at 120 days and attenuated by Los and NOF monotherapies. Los/NOF treatment arrested renal structural injury and ANG II expression and reversed hypertension, albuminuria, and renal inflammation. In conclusion, abnormal expression of COX-2, ANG II, and AT 1 R may be key to development of renal injury in Nx. Concomitant COX-2 inhibition and AT 1 R blockade arrested renal injury and may represent a useful strategy in the treatment of chronic nephropathies.
【关键词】 prostaglandinendoperoxide synthase angiotensin II antiinflammatory agents kidney failure chronic inflammation
BOTH HEMODYNAMIC AND INFLAMMATORY factors are involved in the pathogenesis of chronic progressive nephropathies, whereas angtiotensin II (ANG II) and prostanoids are thought to participate in both kinds of events. Accordingly, cyclooxygenase inhibitors, and especially suppressors of the reninangiotensin-aldosterone system (RAS), were shown to retard the progression of chronic experimental nephropathies. RAS inhibitors have also been successful in the treatment of human nephropathies ( 3, 22 ), suggesting that ANG II plays a central pathogenic role in these processes.
The beneficial effect of RAS suppressors was initially attributed to amelioration of the glomerular hemodynamic dysfunction associated with progressive nephropathies. However, recent observations suggest that the nonhemodynamic effects of RAS suppressors may be equally important, given the strong proinflammatory and profibrotic effects of ANG II ( 35 ). A substantial fraction of this proinflammatory ANG II may originate in the renal parenchyma, rather than in renal vessels or in the systemic circulation ( 44 ). Increased intrarenal production of ANG II was described in various models of renal fibrosis ( 12, 30, 34 ). A preliminary report has suggested that, in the renal ablation (Nx) model, ANG II is expressed in renal interstitial cells, paralleling the severity of renal injury ( 28 ).
Both the hemodynamic and proinflammatory effects of ANG II are mediated by AT-1 receptors (AT 1 R) ( 35 ), extensively expressed in renal tissue. In the normal rat kidney, AT 1 R are predominantly expressed in tubular cells and vessels ( 15 ). Recent data obtained with the Nx model have suggested that AT 1 R expression is shifted from the glomerular to the tubulointerstitial compartment 4 wk after ablation ( 6 ). However, the renal distribution of AT 1 R in this model and its temporal evolution have not been established.
Cyclooxygenase (COX) derivatives may play an important role in the pathogenesis of progressive nephropathies, comparable to their role in arthritis. Increased renal expression of isoforms 1 and 2 of COX has been reported in nephropathies of immunological and nonimmunological origin, such as systemic lupus erythematosus ( 43 ), glomerulosclerosis in Fawn-Hooded hypertensive rats ( 50 ), Heymann nephritis ( 40 ), and renal ablation ( 9, 49 ). In a recent study by this laboratory, increased renal expression of COX-2, which correlated significantly with the extent of renal damage, was shown in Nx rats ( 9 ). In addition, chronic use of COX inhibitors greatly attenuated renal injury in the Nx model ( 9, 10, 48 ).
In view of the complexity and the large number of events leading to progressive renal fibrosis, the interruption of two or more pathogenic pathways by a combination of drugs with different mechanisms of action is likely to be more effective than the respective monotherapies. Three independent studies showed that the combination of a RAS suppressor with an immunosuppressive agent exerted a much stronger protective effect on Nx rats than either drug alone ( 11, 13, 33 ).
In the present study, we evaluated in Nx rats the renal distribution of both the AT 1 R and the COX-2 isoform, as well as the variation of their renal expression with time. In addition, we estimated the amount of ANG II present in glomerular arterioles and in the cortical interstitium. To evaluate the role of these mediators in progression, Nx rats were treated with nitroflurbiprofen (NOF), a nonsteroidal anti-inflammatory drug (NSAID) with low gastrointestinal toxicity, or losartan potassium (Los), an AT 1 R blocker. In an attempt to obtain more effective renal protection, a group of Nx rats receiving a combined NOF/Los treatment was studied as well.
METHODS
Seventy-seven adult male Munich-Wistar rats, obtained from a local breeding colony, with initial weights of 240 to 260 g, were used in this study. Under anesthesia with pentobarbital sodium (50 mg/kg ip), nephrectomy was performed by removal of the right kidney and ligation of the appropriate left renal artery branches, thus ensuring the infarction of at least two-thirds of the left kidney. Twelve sham-operated (S) rats underwent anesthesia, ventral laparotomy, and manipulation of the renal pedicles, without any removal of renal mass. After recovering from anesthesia, the animals were returned to their original cages, given free access to tap water and standard chow (0.5% Na, 22% protein), and maintained at 23 ± 1°C under a 12:12-h light-dark cycle. All experimental procedures were in strict accordance with our institutional guidelines and with international standards for manipulation and care of laboratory animals, being previously approved by the local Research Ethics Committee.
Experimental groups. Thirty days after Nx, the tail-cuff pressure (TCP) was measured by an indirect method ( 11 ). The animals were then placed in metabolic cages for determination of daily urinary albumin excretion rate (U alb V). Animals that at this time had failed to develop hypertension (defined as TCP 140 mmHg) or albuminuria (U alb V 50 mg/day) were excluded from the study. The kidneys of 12 Nx rats were then perfusion-fixed with Duboscq-Brazil solution (0.45% picric acid in a mixture of ethanol, formaldehyde, and acetic acid) after a brief saline washout and prepared for light microscopic and immunohistochemical analysis as described below. This group, designated Nx pre, was used to evaluate the extent of renal injury at 30 days after Nx and served to assess the therapeutic efficacy of treatments started thereafter. The remaining Nx rats were then followed for an additional 3 mo (up to 4 mo after nephrectomy) after having been distributed among four experimental groups: Nx V ( n = 12), Nx rats receiving inert vehicle; Nx NOF ( n = 14), Nx rats receiving NOF (Nicox, Sophia Antipolis), 7.5 mg/kg dissolved in a mixture of DMSO 5% in olive oil and given by gavage twice daily; Nx Los ( n = 14), Nx rats receiving Los (Merck Sharp & Dohme, Rahway, NJ) dissolved in the drinking water at 20 mg/dl, corresponding to a daily ingestion of 50 mg/kg; and Nx Los/NOF ( n = 13), rats given simultaneous NOF and Los treatments. The concentration of Los in the drinking water was adjusted to compensate for variations in daily water intake to ensure that both Nx Los and Nx Los/NOF groups received a constant dosage of 50 mg·kg -1 ·day -1. A group of S rats ( n = 12) was also followed. To mimic the situation usually found in clinical practice, all drug treatments were started only 30 days after renal ablation, when the process of progressive renal injury associated with this model is already in motion. The distribution of rats into the four Nx groups was performed in such a way as to ensure that both mean TCP and albumin excretion rates were similar among groups. The mean initial blood pressure values (in mmHg) and albuminuria values (in mg/day) are shown in Table 1. None of these differences was statistically significant.
Table 1. Renal and systemic functional parameters before treatment (30 days after renal mass reduction) and at the end of the study (120 days after renal mass reduction)
Experimental protocol. To evaluate the efficiency of NOF as a COX blocker, the daily urinary excretion of thromboxane B 2 (TxB 2 ), a derivative of thromboxane A 2 (TxA 2 ), was measured by radioimmunoassay (NEN, Boston, MA) 60 days after renal ablation (30 days of treatment). Because, after renal mass reduction, most of the urinary prostanoids is believed to originate from renal synthesis ( 26 ), we also estimated the thromboxane production per nephron by factoring urinary TxB 2 by the total number of nephrons, assumed to be 6 x 10 4 in S and 10 4 in Nx.
All groups were followed until 120 days after surgery (90 days of treatment), with monthly determination of TCP and U alb V ( 11 ). At the end of the study, rats were anesthetized with pentobarbital sodium (50 mg/kg ip), and a blood sample was collected from the abdominal aorta to assess serum creatinine (S creat ). The renal tissue was then perfusion-fixed with Duboscq-Brazil solution after saline washout and prepared for morphological analysis as described below.
Histological analysis. Sections 4-µm thick were stained by periodic acid-Schiff reaction or Masson trichrome. All morphometric evaluations were performed in a blinded fashion by a single observer. The extent of glomerular sclerosis (GS) was evaluated by attributing a score to each glomerulus according to the apparent extent of the tuft area affected by the sclerotic injury, as described previously ( 11 ). At least 100 glomeruli were evaluated for each rat. A GS index (GSI) was calculated for each rat as the weighted average of all individual glomerular scores thus obtained ( 11 ). To evaluate the extent of renal interstitial expansion, the fraction of renal cortex occupied by interstitial tissue staining positively for extracellular matrix components (%INT) was quantitatively evaluated in Masson-stained sections by a point-counting technique in 25 consecutive microscopic fields, at a final magnification of x 100 under a 176-point grid ( 11 ).
Immunohistochemical analysis. All immunohistochemical studies were performed in 4-µm-thick, paraffin-embedded renal sections. Sections were mounted on glass slides coated with 2% gelatin, deparaffinized, and rehydrated using standard techniques. Sections were then exposed to microwave irradiation in citrate buffer to enhance antigen retrieval and preincubated with 5% normal rabbit (for ED-1) or horse (for COX-2, AT 1, and ANG II) serum in Tris-buffered saline to prevent nonspecific protein binding. Incubation with the primary antibody was always carried out overnight at 4°C in a humidified chamber. Negative control experiments for all antigens were performed by omitting incubation with the primary antibody.
For macrophage detection, a monoclonal mouse anti-rat ED-1 antibody (Serotec, Oxford, UK) was used. After being washed, sections were incubated with rabbit anti-mouse Ig (Dako, Glostrup, Denmark), then with an alkaline phosphatase anti-alkaline phosphatase (Dako) complex. Finally, sections were developed with a fast-red dye solution, counterstained with Mayer's hemalaum, and covered with Kaiser's glycerin-gelatin (Merck, Darmstadt, Germany).
COX-2-, ANG II-, and AT 1 R-positive cells were detected by an indirect streptavidin-biotin alkaline phosphatase technique. The primary (monoclonal mouse anti-rat) antibody for COX-2 was purchased from Transduction Laboratories (Lexington, KY). For ANG II detection, a monoclonal rabbit anti-human ANG II (Peninsula Lab, San Carlos, CA) was used, whereas AT 1 R was detected with a monoclonal rabbit anti-rat AT 1 R antibody (RDI, Flanders, NJ). Sections were preincubated with avidin and biotin solutions to block nonspecific binding of these compounds (Blocking Kit, Vector Labs, Burlingame, CA). After being washed, the sections were incubated at room temperature with rat-adsorbed biotinylated anti-mouse or anti-rabbit IgG (Vector Labs) for 45 min, then with a streptavidin-biotin-alkaline phosphatase complex (Dako) for an additional 30 min. Sections were finally incubated with a freshly prepared substrate, consisting of naphthol AS-MX-phosphate and developed as described above.
The extent of renal infiltration by macrophages, ANG II-positive cells, and COX-2-positive cells was evaluated in a blinded manner at x 250 magnification and expressed as cells per square millimeter. For each section, 25 microscopic fields, each corresponding to an area of 0.06 mm 2, were examined. Because interstitial AT 1 R in Nx rats was often so densely expressed as to preclude the individualization of positively stained cells, AT 1 R expression had to be estimated by a point-counting technique similar to the one employed to determine %INT. This technique allowed us to assess the distribution of AT 1 R among several compartments of the renal cortex (glomeruli, vessels, tubules, and interstitium). The glomerular expression of COX-2 was evaluated by counting positively stained cells in a total of 100 glomeruli/rat, and the results were expressed as cells/100 glomeruli. The fraction of stained macula densa regions was also estimated. The expression of COX-2 in arteries/arterioles was evaluated by counting the number of stained cells in a total of 50 cortical vessels and given as cells/50 vascular profiles.
Statistical analysis. One-way ANOVA with paired comparisons according to the Newman-Keuls formulation was used in this study ( 9 ). The Spearman correlation coefficient was used to evaluate the existence of significant linear correlation between parameters obtained in individual rats. Because GSI and albumin excretion rates behaved as continuous variables with nonnormal distribution, an approximately Gaussian distribution was obtained in all groups by performing log transformation of these data before statistical analysis. For similar reasons, parameters expressed as proportions underwent arcsine transformation before analysis ( 9 ). P values < 0.05 were considered significant.
RESULTS
Renal and systemic parameters obtained at 30 (before treatment) and 120 days after Nx are given in Table 1. Nx groups exhibited limited growth compared with S. In all Nx groups except Nx Los/NOF, body weights were not statistically different from those observed before treatment. Average food intake was similar among groups. The left kidney weight was similar among Nx groups 120 days after Nx. Figure 1 shows, in a graphic manner, TCP as a function of time. TCP rose markedly in the Nx V group, reaching 216 ± 6 mmHg 120 days after surgery (vs. 112 ± 3 in S, P < 0.05, and 172 ± 3 mmHg in Nx pre, P < 0.05). After 30 days of treatment (60 days after Nx), TCP was unchanged by NOF monotherapy but fell by 40 mmHg in both Nx Los and Nx Los/NOF. Los monotherapy lost part of this initial antihypertensive effect with time, TCP returning to pretreatment levels 120 days after renal ablation. Because NOF monotherapy had no antihypertensive effect, TCP reached similar values in Nx Los and Nx NOF rats 120 days after Nx. Although these values were significantly lower than in Nx V, TCP did not differ statistically from the respective Nx pre values, remaining markedly elevated compared with S. In the Nx Los/NOF group, the initial antihypertensive effect was maintained throughout the study ( Fig. 1 ). As a consequence, final TCP values in this group were only moderately elevated compared with S (144 ± 7 mmHg, P < 0.05 vs. S and Nx V ) and significantly lower than those verified before treatment ( P < 0.05 vs. Nx pre ).
Fig. 1. Time course of tail-cuff pressure (TCP) as a function of time in sham-operated rats (S), rats 30 days after nephrectomy (Nx), Nx rats before additional treatement (Nx pre ), and Nx rats treated with vehicle (Nx V ), losartan (Nx Los ), nitroflurbiprofen Nx NOF, and combination therapy (Nx Los/NOF ) groups.
As expected, U alb V ( Table 1 ) was markedly increased at 30 days of surgery (Nx pre group), reaching 104.3 ± 7.1 mg/day ( P < 0.05 vs. S). Albuminuria was aggravated at 120 days after Nx (178.5 ± 43.1 mg/day, P < 0.05 vs. S and P 0.05 vs. Nx pre ). Monotherapy with LOS numerically decreased U alb V relative to untreated Nx (112.9 ± 14.7 mg/day, P 0.05 vs. Nx V ) and prevented further increases in albuminuria, which remained at levels similar to those observed in the Nx pre group. Monotherapy with NOF significantly reduced albuminuria relative to pretreatment values (72.6 ± 11.2 mg/day, P < 0.05 vs. Nx V, Nx Los, and Nx pre ). Combined Los+NOF treatment exerted a much more efficient antiproteinuric action than any of the monotherapies, reversing albuminuria to 26.2 ± 3.5 mg/day ( P < 0.05 vs. Nx V, Nx NOF, Nx Los, and Nx pre ).
As expected, S creat levels were markedly increased by renal mass reduction. None of the monotherapies promoted a significant decline in S creat relative to either Nx V or Nx pre. In the Los+NOF group, S creat was significantly reduced compared with either Nx V or Nx pre (1.0 ± 0.1 vs. 1.2 ± 0.1 in Nx pre, P < 0.05 and 1.4 ± 0.2 mg/day in Nx V, P < 0.05).
As shown in Table 2, urinary excretion of TxB 2, measured at 30 days of treatment, was slightly increased in Nx V compared with S (20.6 ± 1.7 vs. 17.8 ± 1.9 ng/day in S, P 0.05). Calculated TxB 2 excretion per nephron was increased seven-fold in Nx rats 60 days after surgery (2.1 ± 0.2 vs. 0.3 ± 0.1 pg·nephron -1 ·day -1 in S, P < 0.05). As expected, total urinary TxB 2 excretion was markedly depressed compared with S and Nx V in Nx NOF and Nx Los/NOF rats (7.0 ± 1.0 and 8.6 ± 0.8 ng/day, respectively, P < 0.05 vs. S and Nx V ). TxB 2 excretion in the Nx NOF and Nx Los/NOF groups appeared even more depressed when factored by the estimated number of nephrons, reaching values close to those verified in S (0.7 ± 0.1 and 0.9 ± 0.1 ng·nephron -1 ·day -1, respectively, P < 0.05 vs. Nx V and P 0.05 vs. S). Monotherapy with Los resulted in a slight but statistically significant reduction in the calculated TxB 2 excretion per nephron relative to Nx (1.7 ± 0.2 ng·nephron -1 ·day -1, P < 0.05 vs. Nx V ).
Table 2. Urinary thromboxane B 2 excretion 60 days after renal ablation (30 days of treatment)
Glomerular segmental sclerotic lesions were evident 30 days after surgery (Nx pre group), the GSI reaching values almost 20-fold higher than in S ( Table 3 ). Ninety days later (120 days after surgery), considerable progression of glomerular injury had occurred. In Nx V rats, the GSI attained values almost 10-fold as high as in Nx pre and 200-fold higher than in S. Treatment with any of the monotherapies was associated with a less pronounced increment of the GSI, although the respective differences relative to the Nx V group were not statistically significant. Combined Los/NOF treatment arrested the progression of glomerular injury, keeping the GSI at levels close to those verified in the Nx pre group (43 ± 11 vs. 19 ± 4 in Nx pre, P 0.05). There was a significant correlation between GSI and TCP ( r = 0.74, P < 0.01). Interstitial expansion was also a prominent component of renal injury after Nx ( Table 3 ), %INT reaching values more than threefold higher than in Nx pre 120 days after Nx. Unlike NOF and Los monotherapies, combined Los/NOF treatment significantly attenuated the progression of interstitial expansion (4.8 ± 0.5 vs. 7.8 ± 0.7 in Nx V, P < 0.05).
Table 3. Parameters of renal injury and inflammation 120 days after renal ablation (90 days of treatment)
Immunohistochemistry. Thirty days after Nx (Nx pre group), macrophage infiltration in the renal tissue ( Table 3 ), assessed by the density of ED-1-positive cells, more than quadrupled compared with S values (131 ± 12 vs. 29 ± 3 cells/mm 2 in S, P < 0.05). This was aggravated 120 days after surgery, macrophage density reaching values of 178 ± 23 cells/mm 2 in the Nx V group. None of the monotherapies had any significant influence on macrophage infiltration. However, the combined Los/NOF treatment reduced the renal macrophage density to levels significantly lower than in the Nx pre group (81 ± 6 cells/mm 2 vs. Nx pre, P < 0.05). A significant linear correlation ( r = 0.67, P < 0.001) was observed between the macrophage density and the GSI.
Figure 2 A shows a typical staining pattern for ANG II in an afferent arteriole obtained from an S rat (the efferent arteriole, not stained, is shown as well). In Fig. 2 B, ANG II-positive cells unrelated to vascular tissue are shown in the renal cortical interstitium of an Nx rat 120 days after ablation. Most of these ANG II-positive cells appeared in association with inflamed areas. Figure 3 shows the intensity of both modalities of ANG II expression in graphic form. Afferent arteriolar ANG II expression was deeply depressed in untreated Nx (0.04 ± 0.02 positively stained arterioles/mm 2 in Nx pre and 0.03 ± 0.02 in Nx V vs. 1.55 ± 0.16 in S, P < 0.05) and was not influenced by any of the treatments. By contrast, the density of ANG II-positive cells in the interstitium was already markedly increased at 30 days after ablation (Nx pre group) compared with S. These values were tripled 120 days after Nx (Nx V group). The increasing interstitial ANG II expression was completely arrested and kept at pretreatment levels by the combined Los/NOF treatment. None of the monotherapies had any significant effect on interstitial ANG II expression.
Fig. 2. Renal ANG II expression evidenced by immunohistochemistry (positive cells appear heavily stained). A : an afferent arteriole (AA), positively stained for ANG II, is shown next to a nonstained efferent arteriole (EA). B : ANG II-positive cells in an inflamed area.
Fig. 3. Graphical representation of the expression of ANG II in afferent arterioles ( ) and ANG II-positive cells in the renal interstitium ( ). a P < 0.0 5 vs. S. b P < 0.0 5 vs. Nx pre. c P < 0.0 5 vs. Nx V. d P < 0.0 5 vs. Nx Los. e P < 0.0 5 vs. Nx NOF.
AT 1 R expression patterns are shown in Figs. 4 and 5. In the S group, AT 1 R expression was almost entirely confined to the tubular compartment and only sparsely expressed at glomeruli, vessels, and interstitium. Renal mass reduction increased only numerically the total renal expression of AT 1 R but promoted a drastic change in its intrarenal distribution. In the Nx pre group (30 days after Nx), total renal AT 1 R expression was clearly shifted to the interstitial area. This pattern became even more marked 120 days after ablation (Nx V group), when AT 1 R expression at the tubules declined to very low levels ( Fig. 5 ). None of the treatments promoted any significant change in the intensity or distribution of AT 1 R expression in Nx rats.
Fig. 4. AT 1 receptor (AT 1 R) expression (positive structures appear heavily stained). A : in S rats, AT 1 R was typically expressed in tubules (T), vessels (V) and, to a much lesser extent, glomeruli (G). B : in Nx pre rats (30 days after ablation), AT 1 R expression was massively shifted to the interstitium. C : in Nx V rats (120 days after ablation), interstitial AT 1 R expression was even more intense, whereas tubular and vascular AT 1 R expression was much weaker than in the Nx pre group.
Fig. 5. Graphical representation of renal AT 1 R distribution among tubular (dark gray bar sections), interstitial (light gray bar sections), glomerular (open bar sections), and vascular (filled bar sections) compartments. Error bars (SE) refer to total height of column (total renal AT 1 R). * P < 0.0 5 vs. S (tubular compartment). # P < 0.05 vs. S (interstitial compartment).
As described previously ( 9, 49 ), COX-2 was constitutively expressed in cells of the macula densa region of S rats ( Fig. 6 A ). Only rare COX-2-positive cells were found in glomeruli, vessels, or interstitium of intact kidneys. In accordance with previous observations of this laboratory ( 9 ), renal mass reduction numerically increased the frequency of macula densa staining positively for COX-2 ( Table 4 ). Monotherapy with NOF or Los further augmented the expression of COX-2 at the macula densa, although only with the latter was this change statistically significant ( Table 4 ). Combined Los/NOF treatment nearly doubled COX-2 expression at the macula densa ( Table 4 ).
Fig. 6. Renal cyclooxygenase 2 (COX-2) expression (positive cells appear heavily stained). In S, COX-2 was typically expressed at the macula densa ( A ), whereas in Nx V (120 days after renal ablation), COX-2-positive cells appeared in glomeruli ( B ), injured vessels ( C ), and at inflamed interstitial areas ( D ).
Table 4. Renal COX-2 expression and its distribution
As previously reported by this laboratory ( 9 ), there was a dramatic elevation in the density of COX-2-positive cells at glomeruli ( Fig. 6 B ), vessels ( Fig. 6 C ), and interstitium ( Fig. 6 D ) 30 days after renal mass reduction, which was exacerbated after 120 days ( Table 4 ). Monotherapies had no effect on the frequency of glomerular, vascular, or interstitial COX-2-positive cells, whereas combined Los/NOF treatment reduced vascular and glomerular COX-2 expression to values indistinguishable from those seen in S ( Table 4 ). There was a strong positive correlation ( r = 0.83, P < 0.001) between the glomerular density of COX-2-positive cells and the GSI.
DISCUSSION
As expected, renal ablation promoted growth retardation, systemic arterial hypertension, impaired renal function, and severe albuminuria. These functional changes were accompanied by severe glomerulosclerosis, as well as expansion and intense macrophage infiltration of the interstitial area. Mounting evidence indicates that these renal structural abnormalities, which are characteristic of the Nx and other models of progressive nephropathies, are a consequence of the concerted action of mechanical stress, caused by glomerular hypertension and hypertrophy ( 2, 9 ), and inflammatory phenomena, comprising cell infiltration and/or proliferation and extracellular matrix accumulation ( 8, 9 ). Moreover, a causal relationship appears to exist between these phenomena, because the distension of the glomerular walls due to intracapillary hypertension may trigger the local release of cytokines, growth factors, and, particularly, ANG II and prostanoids ( 1, 23 ).
As described previously, intrarenal ANG II distribution was profoundly changed after renal mass reduction ( 28 ). Thirty days after ablation there was a marked decrease in ANG II expression in afferent arterioles, whereas large amounts of ANG II-positive cells appeared in the inflamed interstitial areas. This intrarenal shift of ANG II-positive cells was intensified 120 days after Nx, which may have contributed substantially to aggravate renal injury. These findings support and extend previous observations made in Nx and chronic nitric oxide (NO) inhibition models, which showed the presence of ANG II-positive cells in inflamed areas of the renal interstitium ( 28, 34 ). The pattern of intrarenal ANG II distribution suggests completely distinct roles for ANG II located in each of these "compartments." It is possible that "vascular" ANG II, the expression of which is reduced after renal ablation, is primarily linked to regulation of sodium balance. According to this view, vascular ANG II would be depressed by the extracellular fluid volume expansion known to occur in the Nx model ( 17 ). By contrast, "interstitial" ANG II seems to be insensitive to volume expansion and, given its association with interstitial expansion and cellular infiltration, it appears to be linked to renal inflammation. The origin of this interstitial ANG II was not addressed in the present study. ANG II may have been produced locally, because tubular epithelial cells, macrophages, and myofibroblasts all possess the appropriate biochemical machinery ( 12, 29, 39 ). Alternatively, ANG II may have originated outside the renal interstitium and undergone internalization by local cells after binding to AT 1 R ( 44, 57 ). Local ANG II may have had a proinflammatory and profibrotic effect at the renal interstitium. ANG II stimulates inflammatory cells such as lymphocytes ( 25 ) and activates nuclear factor- B in monocytes ( 36 ). In addition, ANG II stimulates the release of MCP-1 by vascular smooth muscle cells ( 36 ). In renal tissue, ANG II stimulates the proliferation of mesangial cells, glomerular endothelial cells ( 53 ), and myofibroblasts ( 55 ), as well as the secretion of chemokines and growth factors such as RANTES ( 52 ), PDGF ( 8 ), and MCP-1 ( 36 ). In addition, ANG II strongly stimulates collagen synthesis and renal fibrosis by activating TGF- ( 54 ) and the MAPK/ERK pathway ( 41 ). Blockade of interstitial ANG II is a possible explanation for the well-known beneficial effect of RAS suppressors in progressive nephropathies, also observed in the Nx Los and Nx Los/NOF groups.
In agreement with previous observations ( 15 ), renal expression of AT 1 R in S rats appeared mostly in tubular cells, and to a lesser extent, at the interstitial area, whereas weaker expression was seen in vessels and glomeruli. This pattern was completely disrupted 30 days after Nx, when dense AT 1 R expression could be demonstrated in interstitial cells, far exceeding in intensity the expression of AT 1 R in tubules. The exact meaning of this finding and the cell types involved are uncertain. Several inflammatory cells known to infiltrate the renal interstitium in the Nx model have the potential to express AT 1 R, such as lymphocytes ( 25 ) and macrophages ( 29 ). In addition, AT 1 R may be expressed by myofibroblasts originating from tubular cell transdifferentiation ( 27 ). This hypothesis is particularly attractive because it helps to explain the progressive shift in AT 1 R expression, from tubules to the interstitial area, observed in Nx rats, and also because tubular cells already express AT 1 R under normal conditions. The simultaneous presence at the interstitial area of large amounts of ANG II and of the AT 1 R may accelerate the progression of the nephropathy by a positive-feedback mechanism. Consistent with this view is the aggravation of the renal structural injury at 120 days of Nx, which was paralleled by the intensity of the inflammatory infiltration and of the interstitial expression of ANG II.
The potential role of prostanoids in the pathogenesis of progressive nephropathies has long been acknowledged. The stimulation of podocytes by complement fractions can increase the local synthesis of prostanoids ( 40 ). Similarly, nonimmune mechanisms such as mesangial stretching can augment the expression of COX and enhance the production of its derivatives ( 1 ). Accordingly, studies of the Nx model showed increased urinary excretion of prostanoids per nephron ( 26 ).
Increased production of prostanoids can enhance inflammation and, therefore, accelerate renal injury. Prostanoids derived from COX-2 are thought to modulate proliferation and activation of T lymphocytes ( 20 ). Dendritic cells, described in the remnant kidney ( 33 ), constitutively express COX-2 and utilize prostanoids as an autocrine stimulus for cytokine secretion and for their own proliferation ( 51 ). In addition to its well-known vasoconstrictor effect ( 24 ), TxA 2 stimulates the expression of adhesion molecules and of MCP-1 in endothelial cells ( 21 ), promotes the proliferation of mesangial cells ( 4 ), and enhances platelet aggregation and extracellular matrix production ( 5, 24 ).
Previous studies have shown that the renal cortical expression of COX-2 increases after renal ablation, whereas the expression of COX-1 remains unchanged ( 18, 49 ). We showed recently that a large fraction of the excess COX-2 expressed in remnant kidneys localizes in glomeruli, vessels, and the interstitium, especially in areas of injury and inflammation ( 9 ). The present study confirms these observations, reinforcing the concept that COX-2 can exert a dual role in this model: at the MD, COX-2 appears to exert a physiological effect, possibly related to sodium homeostasis. At "anomalous" locations such as glomeruli and vessels, COX-2 and its products would mediate inflammation and structural injury. The consistent presence of COX-2 in damaged areas, and the strong correlation between the intensity of its expression and parameters of renal injury, strengthens the notion that COX-2-derived prostanoids play an important pathogenic role in this model. The mechanisms by which COX-2 may have been induced in these areas are obscure. COX-2 may have been activated by ANG II anomalously produced in the interstitium ( 56 ), by stretching of mesangial cells resulting from glomerular hypertension ( 1 ) and/or by the action of other mediators such as TNF- and interleukin-1 ( 7, 31 ). Once synthesized, prostanoids can further activate COX-2, thereby contributing to amplify and perpetuate the inflammatory process ( 42 ). The protective effect of chronic treatment with either COX-2-specific inhibitors ( 9, 48 ) or NOF ( 10 ) lends further support to the notion that prostanoids play a fundamental role in the pathogenesis of progressive renal injury in the Nx model.
Consistent with previous observations, Los monotherapy lowered blood pressure by 40 mmHg 1 mo after treatment was started, although TCP returned to pretreatment levels at the end of the study. In addition, Los limited U alb V, GSI, interstitial damage, macrophage infiltration, and ANG II-positive cell infiltration ( 11 ). However, protection conferred by losartan monotherapy was only partial, because progression of renal inflammation and of renal structural injury was not arrested in the Nx Los group. There are at least three possible reasons for the limited efficacy of Los treatment in this study. First, rats failing to attain blood pressures higher than 140 mmHg or albuminuria in excess of 50 mg/dl 30 days after nephrectomy were excluded from the study to ensure that the attending nephropathy had a progressive nature. Second, unlike in most previous studies of this model, treatments were started only 30 days after nephrectomy, when renal injury was already established. Third, rats were followed up to 4 mo after renal mass reduction, whereas in most other studies of this model observations were ended at 2 mo or less. In the face of the unusual severity of renal injury, the relative resistance to Los treatment was not unexpected. At any instance, these findings agree with previous experimental observations, obtained in this laboratory and elsewhere ( 11, 19 ), as well as in large clinical trials ( 3 ), all of which indicate that, once set in motion, progressive nephropathies can be attenuated, but not detained, by isolated treatment with AT 1 R blockers or ACE inhibitors. As a whole, these observations suggest that events antecedent to the initiation of treatment may be of crucial importance in the pathogenesis of renal injury associated with this model. Additionally, the relative inefficiency of Los monotherapy in the present study may reflect the presence of ANG II-independent inflammatory events, as well as the recrudescence of arterial hypertension after the first few weeks of treatment. Finally, it is conceivable that the "conventional" dose of Los employed in the present study, although high enough to exert full vascular effect, was insufficient to effectively block the enormous amount of AT 1 R already present at the renal tissue when treatment was started.
NOF is a nonsteroidal anti-inflammatory compound with low gastrointestinal toxicity, presumably due to its NO-releasing properties ( 10, 47 ). It appears unlikely that NO released by NOF has a direct therapeutic effect because the NOF molecule is rapidly degraded in the intestinal lumen, releasing flurbiprofen ( 10 ). However, we cannot exclude the possibility that nitroso proteins, which have a much longer half-life than NO itself, propagate a possible protective effect of NO released into the intestinal lumen ( 37 ). Flurbiprofen is a potent inhibitor of both COX isoforms ( 45 ). In the present study, this property was demonstrated by the marked reduction in urinary TxB 2 excretion obtained in the Nx NOF and Nx LOS/NOF groups. Monotherapy with NOF reduced proteinuria and attenuated the progression of glomerular injury and the intensity of local macrophage infiltration as effectively as monotherapy with Los, even though NOF exerted no effect on blood pressure. However, as in the case of Los treatment, progression of renal injury was not arrested. Moreover, no protective effect was observed regarding interstitial expansion or glomerular and vascular COX-2 expression. In addition, the effects of NOF were much more modest than those obtained when treatment was begun immediately after surgery ( 10 ), once again underlining the pathogenic importance of early events in this model.
Treatment of Nx rats with the Los/NOF combination promoted a significant regression of hypertension, albuminuria, and inflammatory signs such as macrophage infiltration and tissue COX-2 expression, whereas the parameters of structural tissue injury remained stable, or were strongly attenuated, compared with pretreatment levels. The protection achieved with combined treatment was much greater than that obtained with either drug alone. On the basis of the present study, one cannot exclude the hypothesis that the success of combined treatment was due to a particularly effective hemodynamic action, although previous observations from this laboratory ( 10 ) indicated that NOF had no significant effect on glomerular hemodynamics. Because treatment with NOF alone had no effect on blood pressure, it seems unlikely that the hemodynamic effect of LOS was directly intensified by its association with NOF. Therefore, the efficacy of combined treatment was likely due to the simultaneous blockade of the hemodynamic and proinflammatory actions of ANG II and COX derivatives and by abrogation of the complex interplay between hypertension and inflammation ( 34 ). The present findings support previous observations of the Nx model, which similarly indicated the superiority of the combination of a RAS suppressor with an anti-inflammatory agent ( 11, 13, 33 ). It is noteworthy that combined treatment afforded partial regression of the nephropathy associated with Nx even though it was started 1 mo after surgery, when renal injury was already established. This observation suggests that both continued stimulation of AT 1 receptors and production of prostanoids continue to play an important pathogenic role even during the late phases of the process, necessitating vigorous and persistent treatment to prevent further renal deterioration.
Despite the promising response of Nx rats to combined treatment, it must be noted that NSAIDs are potentially nephrotoxic drugs, especially in the setting of chronic renal insufficiency. NOF promoted no glomerular filtration rate decline in these and in previous studies ( 10 ). However, there have been numerous anecdotal reports of acute renal failure, hyperkalemia, hypertension, and edema attributed to NSAIDs ( 14, 16 ), although few studies have directly assessed the objective risk associated with the clinical use of these compounds ( 38, 46 ). It is possible that simultaneous administration of a RAS suppressor, by limiting or suppressing the vascular effect of ANG II, reduces the risk of excessive renal vasoconstriction in the presence of a COX inhibitor ( 32 ). Nevertheless, any future systematic use of COX inhibitors in chronic nephropathies, alone or in combination, will necessitate the development of careful clinical studies to assess the safety of these compounds in this particular group of patients.
In summary, combined treatment with Los/NOF partially reversed the nephropathy and renal inflammation associated with the Nx model, showing much more effective protection than with either drug alone. Clinical studies are needed to establish whether this scheme may eventually become a new weapon in the limited arsenal currently available to attenuate or prevent human progressive nephropathies.
ACKNOWLEDGMENTS
We thank Gláucia Rutigliano Antunes, Luciana Faria de Carvalho, and Claudia Ramos Sena for expert technical assistance.
Preliminary results of this study were presented at the American Society of Nephrology/International Society of Nephrology World Congress of Nephrology, San Francisco, CA, October 10-17, 2001, and published in abstract form ( J Am Soc Nephrol 12: 814A, 2001).
GRANTS
This work was supported by Grants 95/4710-2 and 98/09569-4 from the State of São Paulo Foundation for Research Support. During these studies, R. Zatz was the recipient of a Research Award (326.429/81) from the Brazilian Council of Scientific and Technologic Development.
【参考文献】
Akai Y, Homma T, Burns KD, Yasuda T, Badr KF, and Harris RC. Mechanical stretch/relaxation of cultured rat mesangial cells induces protooncogenes and cyclooxygenase. Am J Physiol Cell Physiol 267: C482-C490, 1994.
Brenner BM. Nephron adaptation to renal injury or ablation. Am J Physiol Renal Fluid Electrolyte Physiol 249: F324-F337, 1985.
Brenner BM, Cooper ME, de Zeeuw D, Keane WF, Mitch WE, Parving HH, Remuzzi G, Snapinn SM, Zhang Z, and Shahinfar S. Effects of losartan on renal and cardiovascular outcomes in patients with type 2 diabetes and nephropathy. N Engl J Med 345: 861-869, 2001.
Bresnahan BA, Wu S, Fenoy FJ, Roman RJ, and Lianos EA. Mesangial cell immune injury. Hemodynamic role of leukocyte- and platelet-derived eicosanoids. J Clin Invest 90: 2304-2312, 1992.
Bruggeman LA, Horigan EA, Horikoshi S, Ray PE, and Klotman PE. Thromboxane stimulates synthesis of extracellular matrix proteins in vitro. Am J Physiol Renal Fluid Electrolyte Physiol 261: F488-F494, 1991.
Cao Z, Bonnet F, Candido R, Nesteroff SP, Burns WC, Kawachi H, Shimizu F, Carey RM, de Gasparo M, and Cooper ME. Angiotensin type 2 receptor antagonism confers renal protection in a rat model of progressive renal injury. J Am Soc Nephrol 13: 1773-1787, 2002.
Feng L, Xia Y, Garcia GE, Hwang D, and Wilson CB. Involvement of reactive oxygen intermediates in cyclooxygenase-2 expression induced by interleukin-1, tumor necrosis factor-, and lipopolysaccharide. J Clin Invest 95: 1669-1675, 1995.
Floege J, Burns MW, Alpers CE, Yoshimura A, Pritzl P, Gordon K, Seifert RA, Bowen-Pope DF, Couser WG, and Johnson RJ. Glomerular cell proliferation and PDGF expression precede glomerulosclerosis in the remnant kidney model. Kidney Int 41: 297-309, 1992.
Fujihara CK, Antunes GR, Mattar AL, Andreoli N, Malheiros DM, Noronha IL, and Zatz R. Cyclooxygenase-2 (COX-2) inhibition limits abnormal COX-2 expression and progressive injury in the remnant kidney. Kidney Int 64: 2172-2181, 2003.
Fujihara CK, Malheiros DM, Donato JL, Poli A, De Nucci G, and Zatz R. Nitroflurbiprofen, a new nonsteroidal anti-inflammatory, ameliorates structural injury in the remnant kidney. Am J Physiol Renal Physiol 274: F573-F579, 1998.
Fujihara CK, Noronha IL, Malheiros DM, Antunes GR, de Oliveira IB, and Zatz R. Combined mycophenolate mofetil and losartan therapy arrests established injury in the remnant kidney. J Am Soc Nephrol 11: 283-290, 2000.
Gilbert RE, Wu LL, Kelly DJ, Cox A, Wilkinson-Berka JL, Johnston CI, and Cooper ME. Pathological expression of renin and angiotensin II in the renal tubule after subtotal nephrectomy: implications for the pathogenesis of tubulointerstitial fibrosis. Am J Pathol 155: 429-440, 1999.
Hamar P, Peti-Peterdi J, Razga Z, Kovacs G, Heemann U, and Rosivall L. Coinhibition of immune and renin-angiotensin systems reduces the pace of glomerulosclerosis in the rat remnant kidney. J Am Soc Nephrol 10, Suppl 11: S234-S238, 1999.
Haragsim L, Dalal R, Bagga H, and Bastani B. Ketorolac-induced acute renal failure and hyperkalemia: report of three cases. Am J Kidney Dis 24: 578-580, 1994.
Harrison-Bernard LM, Navar LG, Ho MM, Vinson GP, and el-Dahr SS. Immunohistochemical localization of ANG II AT 1 receptor in adult rat kidney using a monoclonal antibody. Am J Physiol Renal Physiol 273: F170-F177, 1997.
Hay E, Derazon H, Bukish N, Katz L, Kruglyakov I, and Armoni M. Fatal hyperkalemia related to combined therapy with a COX-2 inhibitor, ACE inhibitor and potassium rich diet. J Emerg Med 22: 349-352, 2002.
Hayslett JP. Functional adaptation to reduction in renal mass. Physiol Rev 59: 137-164, 1979.
Horiba N, Kumano E, Watanabe T, Shinkura H, Sugimoto T, and Inoue M. Subtotal nephrectomy stimulates cyclooxygenase 2 expression and prostacyclin synthesis in the rat remnant kidney. Nephron 91: 134-141, 2002.
Ikoma M, Kawamura T, Kakinuma Y, Fogo A, and Ichikawa I. Cause of variable therapeutic efficiency of angiotensin converting enzyme inhibitor on glomerular lesions. Kidney Int 40: 195-202, 1991.
Iniguez MA, Punzon C, and Fresno M. Induction of cyclooxygenase-2 on activated T lymphocytes: regulation of T cell activation by cyclooxygenase-2 inhibitors. J Immunol 163: 111-119, 1999.
Ishizuka T, Kawakami M, Hidaka T, Matsuki Y, Takamizawa M, Suzuki K, Kurita A, and Nakamura H. Stimulation with thromboxane A 2 (TXA 2 ) receptor agonist enhances ICAM-1, VCAM-1 or ELAM-1 expression by human vascular endothelial cells. Clin Exp Immunol 112: 464-470, 1998.
Jafar TH, Schmid CH, Landa M, Giatras I, Toto R, Remuzzi G, Maschio G, Brenner BM, Kamper A, Zucchelli P, Becker G, Himmelmann A, Bannister K, Landais P, Shahinfar S, de Jong PE, de Zeeuw D, Lau J, and Levey AS. Angiotensin-converting enzyme inhibitors and progression of nondiabetic renal disease. A meta-analysis of patient-level data. Ann Intern Med 135: 73-87, 2001.
Lee LK, Meyer TW, Pollock AS, and Lovett DH. Endothelial cell injury initiates glomerular sclerosis in the rat remnant kidney. J Clin Invest 96: 953-964, 1995.
Lianos EA, Andres GA, and Dunn MJ. Glomerular prostaglandin and thromboxane synthesis in rat nephrotoxic serum nephritis. Effects on renal hemodynamics. J Clin Invest 72: 1439-1448, 1983.
Nataraj C, Oliverio MI, Mannon RB, Mannon PJ, Audoly LP, Amuchastegui CS, Ruiz P, Smithies O, and Coffman TM. Angiotensin II regulates cellular immune responses through a calcineurin-dependent pathway. J Clin Invest 104: 1693-1701, 1999.
Nath KA, Chmielewski DH, and Hostetter TH. Regulatory role of prostanoids in glomerular microcirculation of remnant nephrons. Am J Physiol Renal Fluid Electrolyte Physiol 252: F829-F837, 1987.
Ng YY, Huang TP, Yang WC, Chen ZP, Yang AH, Mu W, Nikolic-Paterson DJ, Atkins RC, and Lan HY. Tubular epithelial-myofibroblast transdifferentiation in progressive tubulointerstitial fibrosis in nephrectomized rats. Kidney Int 54: 864-876, 1998.
Noronha IL, Fujihara CK, and Zatz R. The inflammatory component in progressive renal disease-are interventions possible (Abstract)? Nephrol Dial Transplant 17: 363, 2002.
Okamura A, Rakugi H, Ohishi M, Yanagitani Y, Takiuchi S, Moriguchi K, Fennessy PA, Higaki J, and Ogihara T. Upregulation of renin-angiotensin system during differentiation of monocytes to macrophages. J Hypertens 17: 537-545, 1999.
Pelayo JC, Quan AH, and Shanley PF. Angiotensin II control of the renal microcirculation in rats with reduced renal mass. Am J Physiol Renal Fluid Electrolyte Physiol 258: F414-F422, 1990.
Perkins DJ and Kniss DA. Tumor necrosis factor- promotes sustained cyclooxygenase-2 expression: attenuation by dexamethasone and NSAIDs. Prostaglandins 54: 727-743, 1997.
Qi Z, Hao CM, Langenbach RI, Breyer RM, Redha R, Morrow JD, and Breyer MD. Opposite effects of cyclooxygenase-1 and -2 activity on the pressor response to angiotensin II (Abstract). J Clin Invest 110: 61, 2002.
Remuzzi G, Zoja C, Gagliardini E, Corna D, Abbate M, and Benigni A. Combining an antiproteinuric approach with mycophenolate mofetil fully suppresses progressive nephropathy of experimental animals. J Am Soc Nephrol 10: 1542-1549, 1999.
Rodriguez-Iturbe B, Quiroz Y, Nava M, Bonet L, Chavez M, Herrera-Acosta J, Johnson RJ, and Pons HA. Reduction of renal immune cell infiltration results in blood pressure control in genetically hypertensive rats. Am J Physiol Renal Physiol 282: F191-F201, 2002.
Ruiz-Ortega M, Lorenzo O, Suzuki Y, Ruperez M, and Egido J. Proinflammatory actions of angiotensins. Curr Opin Nephrol Hypertens 10: 321-329, 2001.
Ruiz-Ortega M, Bustos C, Hernandez-Presa MA, Lorenzo O, Plaza JJ, and Egido J. Angiotensin II participates in mononuclear cell recruitment in experimental immune complex nephritis through nuclear factor- B activation and monocyte chemoattractant protein-1 synthesis. J Immunol 161: 430-439, 1998.
Stamler JS, Jaraki O, Osborne J, Simon DI, Keaney J, Vita J, Singel D, Valeri CR, and Loscalzo J. Nitric oxide circulates in mammalian plasma primarily as an S -nitroso adduct of serum albumin. PNAS 89: 7674-7677, 1992.
Stork JE, Rahman MA, and Dunn MJ. Eicosanoids in experimental and human renal disease. Am J Med 80: 34-45, 1986.
Sun Y, Zhang J, Zhang JQ, and Weber KT. Renin expression at sites of repair in the infarcted rat heart. J Mol Cell Cardiol 33: 995-1003, 2001.
Takano T and Cybulsky AV. Complement C5b-9-mediated arachidonic acid metabolism in glomerular epithelial cells: role of cyclooxygenase-1 and -2. Am J Pathol 156: 2091-2101, 2000.
Tharaux PL, Chatziantoniou C, Fakhouri F, and Dussaule JC. Angiotensin II activates collagen I gene through a mechanism involving the MAP/ER kinase pathway. Hypertension 36: 330-336, 2000.
Tjandrawinata RR and Hughes-Fulford M. Up-regulation of cyclooxygenase-2 by product-prostaglandin E 2. Adv Exp Med Biol 407: 163-170, 1997.
Tomasoni S, Noris M, Zappella S, Gotti E, Casiraghi F, Bonazzola S, Benigni A, and Remuzzi G. Upregulation of renal and systemic cyclooxygenase-2 in patients with active lupus nephritis. J Am Soc Nephrol 9: 1202-1212, 1998.
Van Kats JP, Schalekamp MA, Verdouw PD, Duncker DJ, and Danser AH. Intrarenal angiotensin II: interstitial and cellular levels and site of production. Kidney Int 60: 2311-2317, 2001.
Vane JR and Botting RM. A better understanding of anti-inflammatory drugs based on isoforms of cyclooxygenase (COX-1 and COX-2). Adv Prostaglandin Thromboxane Leukot Res 23: 41-48, 1995.
Wagner EH. Nonsteroidal anti-inflammatory drugs and renal disease-still unsettled. Ann Intern Med 115: 227-228, 1991.
Wallace JL, Reuter B, Cicala C, McKnight W, Grisham MB, and Cirino G. Novel nonsteroidal anti-inflammatory drug derivatives with markedly reduced ulcerogenic properties in the rat. Gastroenterology 107: 173-179, 1994.
Wang JL, Cheng HF, Shappell S, and Harris RC. A selective cyclooxygenase-2 inhibitor decreases proteinuria and retards progressive renal injury in rats. Kidney Int 57: 2334-2342, 2000.
Wang JL, Cheng HF, Zhang MZ, McKanna J, and Harris RC. Selective increase of cyclooxygenase-2 expression in a model of renal ablation. Am J Physiol Renal Physiol 275: F613-F622, 1998.
Weichert W, Paliege A, Provoost AP, and Bachmann S. Upregulation of juxtaglomerular NOS1 and COX-2 precedes glomerulosclerosis in fawn-hooded hypertensive rats. Am J Physiol Renal Physiol 280: F706-F714, 2001.
Whittaker DS, Bahjat KS, Moldawer LL, and Clare-Salzler MJ. Autoregulation of human monocyte-derived dendritic cell maturation and IL-12 production by cyclooxygenase-2-mediated prostanoid production. J Immunol 165: 4298-4304, 2000.
Wolf G, Ziyadeh FN, Thaiss F, Tomaszewski J, Caron RJ, Wenzel U, Zahner G, Helmchen U, and Stahl RAK. Angiotensin II stimulates expression of the chemokine RANTES in rat glomerular endothelial cells. Role of the angiotensin type 2 receptor. J Clin Invest 100: 1047-1058, 1997.
Wolf G, Ziyadeh FN, Zahner G, and Stahl RAK. Angiotensin II Is mitogenic for cultured rat glomerular endothelial cells. Hypertension 27: 897-905, 1996.
Wu LL, Cox A, Roe CJ, Dziadek M, Cooper ME, and Gilbert RE. Transforming growth factor 1 and renal injury following subtotal nephrectomy in the rat: role of the renin-angiotensin system. Kidney Int 51: 1553-1567, 1997.
Wu LL, Yang N, Roe CJ, Cooper ME, Gilbert RE, Atkins RC, and Lan HY. Macrophage and myofibroblast proliferation in remnant kidney: role of angiotensin II. Kidney Int Suppl 63: S221-S225, 1997.
Young W, Mahboubi K, Haider A, Li I, and Ferreri NR. Cyclooxygenase-2 is required for tumor necrosis factor- - and angiotensin II-mediated proliferation of vascular smooth muscle cells. Circ Res 86: 906, 2000.
Zou LX, Imig JD, Von Thun AM, Hymel A, Ono H, and Navar LG. Receptor-mediated intrarenal angiotensin II augmentation in angiotensin II-infused rats. Hypertension 28: 669-677, 1996.
作者单位：1 Renal Division, Department of Clinical Medicine, Faculty of Medicine, University of São Paulo, 01246-903 São Paulo; and 2 Department of Pharmacology, Faculty of Medical Sciences, State University of Campinas, 13081-970 Campinas, Brazil