Abstract
ABSTRACT. The functional role of the NO synthase (NOS) isoforms in the normal or diseased kidney is uncertain. This study examined the renal expression of the endothelial (eNOS), neuronal (nNOS), and inducible (iNOS) isoforms by both immunohistochemistry and Western blot analyses in sham-operated rats (S) and in rats subjected to 5/6 nephrectomy (Nx). Primary antibodies from two different sources were used to detect iNOS. Additional S and Nx rats were chronically treated with aminoguanidine (AG), a selective iNOS inhibitor. All three isoforms were clearly expressed in S kidney. Their renal abundance, evaluated by Western blot analysis, fell in Nx rats. With the use of anti-iNOS antibodies from two distinct sources, the immunohistochemical analysis showed the presence of what appeared to be two distinct iNOS fractions: a “tubular” fraction, present in S and with decreased intensity in Nx; and an “interstitial” fraction, observed only in inflamed areas of Nx rats. AG treatment greatly attenuated renal injury in Nx rats by a direct antiinflammatory effect, likely related to iNOS inhibition, rather than to amelioration of renal hemodynamics or to reduced protein glycation. These observations suggest that: (1) the functional role of the renal iNOS isoform may vary dramatically under different physiologic conditions; (2) caution should be taken in the interpretation of immunohistochemical iNOS data, because antibodies from different sources may detect different iNOS fractions; and (3) AG treatment may become useful in the treatment of human progressive nephropathies, even those not associated with diabetes or aging.
The sequence of events leading to renal injury in progressive nephropathies is incompletely understood. In the 1980s, special attention was paid to mechanical factors such as glomerular hypertension (1) and glomerular hypertrophy (2). More recently, intermediate events, taking place between the initial insult and the final process of renal scarring, have received more attention. Among these, special emphasis has been given to chronic inflammatory manifestations such as macrophage activation, fibroblast proliferation, and excessive production of extracellular matrix (3–5). Renal inflammation is prominent in rats with 5/6 renal ablation (Nx), a widely used experimental model of progressive nephropathy. Treatment of Nx rats with a nonsteroidal antiinflammatory (6) or with the immunosuppressor mycophenolate mofetil (MMF) (7,8) ameliorated lymphocyte and macrophage infiltration and reduced renal injury in this model, suggesting that inflammatory phenomena play an important role in the pathogenesis of progressive renal injury in this model.
Nitric oxide (NO) exerts a fundamental role in the regulation of cardiovascular and renal function, as indicated by the well-known finding that chronic NO inhibition with l-arginine analogues leads to progressive hypertension and severe renal injury (9,10). In addition, deficient NO production has been described in Nx rats and could contribute to renal injury in this model (11–13). Although these data are interpreted by several investigators as indicative that progressive renal disease is a state of NO deficiency (11–13), the role of NO derived from the inducible isoform, iNOS, remains uncertain. Several studies indicated that iNOS is expressed in large amounts in the normal renal tissue, localizing mainly in the tubules, and that pathologic conditions, such as clinical and experimental chronic renal insufficiency, are associated with marked iNOS downregulation (13–15). In other studies, however, the iNOS expression was found to be low or undetectable in normal kidneys, whereas several nephropathies were associated with substantial amounts of iNOS in the glomeruli and the renal interstitium (16–18). One possible reason for this disagreement is the wide heterogeneity of the experimental models studied so far. Additional discrepancy may arise from the fact that the primary antibodies used to detect iNOS come from several sources, because the behavior of different antibodies directed against NOS isoforms can vary dramatically according to type (monoclonal versus polyclonal), species in which the antibody was raised, and tissue in which the antibody is tested (19).
Recent evidence suggests that aminoguanidine (AG), an l-arginine analogue, attenuates renal injury in several instances of progressive renal disease, including experimental lupus nephropathy (20), nephrotoxic serum nephritis (17), allograft rejection (21), experimental diabetic nephropathy (22), and aging nephropathy (23). The exact mechanism underlying these protective effects is unclear. Although some investigators believe that AG works by inhibiting the formation of advanced glycation end products (AGE) (23,24), at least in diabetic and aging nephropathies, others contend that the beneficial effect of AG may be entirely due to an antiinflammatory effect, resulting from its inhibition of the inducible isoform of the nitric oxide synthase (iNOS) (25,26).
In this study, we investigated the renal distribution of the NOS isoforms in sham-operated and Nx rats. We were particularly interested in determining whether the discrepancy currently found in the literature could be reproduced by detecting renal iNOS with immunoblotting and immunohistochemistry techniques using primary anti-iNOS antibodies obtained from two different sources. We also studied whether chronic, selective inhibition of the iNOS isoform by AG treatment could mitigate renal injury in rats with Nx.
Materials and Methods
One hundred thirteen adult male Munich-Wistar rats (240 to 260 g) were obtained from a local colony. Fifty-seven rats underwent 5/6 nephrectomy (Nx) by right uninephrectomy and ligation of two branches of the left renal artery under anesthesia with sodium pentobarbital (50 mg/kg intraperitoneally). Fifty-six sham-operated rats served as controls. 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 22 ± 1°C under a 12/12 h light/dark cycle. All experimental procedures were conducted according to our institutional guidelines.
Experimental Groups
Twenty-four hours after Nx, rats were randomly assigned to four groups: S (n = 30), sham-operated rats; S+AG (n = 26), S rats treated with AG (Sigma, St. Louis, MO) in drinking water (0.33 g/L, corresponding to approximately 20 mg/rat per d); Nx, rats subjected to 5/6 nephrectomy (n = 29); Nx+AG (n = 28), Nx rats treated with AG in drinking water. Water intake in Nx rats was approximately twice as high as in S (see Results); therefore, the concentration of AG in drinking water to keep daily intake at approximately 20 mg/rat/d in Group Nx+AG was 0.17 g/L.
Renal Hemodynamic Studies
At 30 d of treatment, rats were prepared for renal functional studies under inactin anesthesia as described previously (7), to determine the GFR, the glomerular hydraulic pressure (PGC), and other hemodynamic parameters.
Long-Term Studies
Ten to thirteen rats from each group were followed until 60 d after treatment, with monthly determination of tail-cuff pressure (TCP) and 24-h urinary albumin excretion rate. The concentration of total nitrate and nitrite (NOx) in the urine samples was determined at day 60 by chemoluminescence (27) using an NO analyzer (NOA 280; Sievers Instruments, Boulder, CO). At the end of the study, rats were anesthetized with sodium pentobarbital, 50 mg/kg intraperitoneally, and a blood sample was collected from the abdominal aorta to determine serum creatinine concentrations and glycated hemoglobin levels (28). The renal tissue was then prepared for morphologic analysis by either of two methods. In the paraffin fixation method, the kidneys were perfused in situ at the measured arterial pressure with Duboscq-Brazil solution after washout with saline. After fixation, the renal tissue was weighed and two midcoronal sections were postfixed in buffered 10% formaldehyde solution. The material was embedded in paraffin for assessment of glomerular and renal cortical interstitial injury and for immunohistochemical identification of macrophages and NOS expression. For the freezing method, the renal tissue of six rats of each group was excised, snap-frozen in liquid nitrogen, and stored at −70°C for later analysis by Western blot and to assess the expression of iNOS using a different primary antibody (see below).
Histomorphometric Analyses
Paraffin-embedded renal tissue was deparaffinized using standard sequential techniques, and 2- to 3-μm-thick sections were stained with periodic acid-Schiff (PAS) and by the Masson trichrome technique. All morphometric measurements were performed blindly by a single observer. The average glomerular tuft volume (VG), the extent of GS, and the fraction of renal cortex occupied by interstitial tissue were evaluated as described previously (7).
Western Blot Analyses
For Western blot analysis, 2- to 3-mm-thick midcoronal kidney slices were taken from frozen kidneys. The proportion of cortex to medulla in these slices was similar in sham and Nx rats, as estimated in preliminary experiments involving point counting in PAS-stained sections taken from similar slices. Kidney slices were homogenized (10% wt/vol) in the following buffer: 10 mM HEPES buffer (pH 7.6) containing 10% glycerol, 100 mM KCl, 3 mM MgCl2, 5 mM EDTA, 1 mM dithiothreitol, 0.1% sodium dodecyl sulfate (SDS), and 1% protease inhibitor cocktail (Sigma) under ice. The total kidney extract homogenates were centrifuged at 10,000 × g for 10 min at 4°C to remove tissue debris. The protein concentration of the supernatant was determined with the Bio-Rad protein assay kit (Bio-Rad, Hercules, CA). Thirty-microgram samples were size-fractionated by electrophoresis on a 12% denaturing (containing SDS) polyacrylamide gel at 200 V (constant voltage) for approximately 40 min. After electrophoresis, the proteins were transferred onto nitrocellulose membranes at 100 V for 1 h, using the mini-Trans-Blot Transfer Cell (Bio-Rad), under a chilled Tris-glycine buffer, pH 8.4. Broad-range prestained molecular weight markers (New England Biolabs, Beverly, MA) were used in each experiment to assure proper protein transfer. The membranes were then blocked overnight in Tris-buffered saline containing 0.2% Tween-20 (TBST) and 6% nonfat dry milk at room temperature. After washing in TBST, monoclonal (eNOS and iNOS) and polyclonal (bNOS) antibodies (Transduction Laboratories, Lexington, KY) were diluted to 1:1000 (eNOS and iNOS) and 1:2000 (bNOS), and incubation took place for 1 h and was followed by successive incubations with a secondary biotin-bound antibody (1 h) and a streptavidin-biotinylated horseradish peroxidase complex (Amersham Pharmacia Biotech, Piscataway, NJ) (45 min). After each step, washes of 40 min were undertaken, the wash buffer (TBST) being changed every 10 min. Finally, the membranes were analyzed for the bound antibody by chemoluminescence using the ECL detection reagents (Amersham Pharmacia Biotech) and subjected to autoluminography for 15 s. The films were scanned with a densitometer (Imagemaster 1D Elite, Pharmacia Biotech, Newcastle upon Tyne, NE, England) to determine the relative optical densities. All incubations were performed at room temperature.
Immunohistochemical Analyses
To detect macrophages and NOS, 4-μm-thick sections obtained from paraffin-embedded tissue were dewaxed and mounted using conventional techniques. Sections were subjected to microwave irradiation in citrate buffer to enhance antigen retrieval and preincubated with 5% normal rabbit serum in Tris-buffered saline (TBS). The renal tissue was then incubated with monoclonal anti-rat ED-1 antibody (Serotec, Oxford, UK) for macrophage detection or with specific antibodies against eNOS (mouse monoclonal), iNOS (mouse polyclonal), or nNOS (rabbit polyclonal). Antibodies against eNOS (cat # N30020) and nNOS (cat # N31030) were obtained from Transduction Laboratories (TL) of Lexington, KY. The iNOS isoenzyme was detected using two distinct antibodies: (1) a rabbit polyclonal anti-iNOS antibody (cat # SC-650) purchased from Santa Cruz Biotechnology (SC) of Santa Cruz, CA; and (2) a mouse monoclonal antibody purchased from TL (cat. # N32020). All these antibodies were used with paraffin-embedded tissue. In addition, iNOS was detected in 4-μm-thick sections cut from frozen renal tissue, incubated with either SC or TL anti-iNOS antibody, and processed in the same manner as for paraffin-embedded tissue.
To complete the detection of macrophages, the alkaline phosphatase anti-alkaline phosphatase (APAAP) method was used as described previously (7). The slides were then developed with a fast-red dye solution, counterstained with Mayer’s hemalum solution (Merck, Darmstadt, Germany), and covered with Kaiser’s glycerin-gelatin (Merck). For NOS detection in either paraffin-embedded or frozen tissue, biotinylated secondary antibodies were employed. Sections were then incubated with a streptavidin-biotin/alkaline phosphatase complex (Dako Co., Glostrup, Denmark). For eNOS and nNOS, final development was performed with Nitro Blue Tetrazolium (NBT), whereas iNOS staining was performed in the same manner as for macrophage detection.
Negative control experiments were performed by (1) omitting the incubation with the primary antibody; (2) performing prior incubation of the primary antibody with purified iNOS protein (TL); (3) replacing the primary antibody with unspecific rat IgG (Sigma). All incubations were performed overnight at 4°C in a humidified chamber.
The extent of ED-1–positive cell infiltration was evaluated blindly under ×250 magnification and expressed as cells/mm2. For each section, 25 nonconsecutive microscopic fields, each covering an area of 0.06 mm2, were examined. For assessment of the renal cortical expression of iNOS detected by the TL antibody in frozen sections, the number of positively stained cells was counted in 25 microscopic fields under ×400 magnification, covering a total area of 1.5 mm2.
Statistical Analyses
One-way ANOVA with pairwise post test comparisons according to the Newmann-Keuls formulation was employed in this study (29). P ≤ 0.05 were significant. Glomerulosclerosis index (GSI) behaved as a continuous variable with non-normal distribution. An approximately Gaussian distribution for GSI was obtained in all groups by performing log transformation of the data. Albumin excretion rates behaved in the same manner, also requiring log transformation before statistical analysis.
Results
Renal Hemodynamic Studies
Neither the ablation procedure nor AG treatment altered food intake (expressed in g/24 h), which was similar among groups (23 ± 1 in S; 22 ± 1 in S+AG; 22 ± 1 in Nx; 21 ± 1 in Nx+AG; differences were NS). Water intakes (in ml) were 34 ± 1, 36 ± 1, 67 ± 3, and 68 ± 2, respectively (P < 0.05 Nx versus respective S; NS between treated and untreated). Renal hemodynamic parameters measured at 30 d after NX are presented in Table 1. AG treatment had no effect on hemodynamic parameters in sham-operated rats (S+AG group). In both Nx groups, rats exhibited lower body weight (BW) and left kidney weight (LKW) (both in absolute terms and factored by BW) when compared with S. Renal ablation was associated with marked elevation of MAP. Nx rats treated with AG showed hypertension of equivalent magnitude. Likewise, AG treatment had no significant effect on GFR or RPF. PGC was markedly elevated in group Nx compared with S. AG treatment promoted no change in PGC compared with untreated Nx. Both Nx groups exhibited larger glomerular volumes compared with S. AG treatment did not affect VG.
Renal functional and hemodynamic parameters at 30 d of treatmenta
Long-Term Studies
At 60 d of treatment, the urinary excretion of stable NO metabolites (UNOxV) was significantly lower in the Nx group than in S (3.6 ± 0.7 μmol/24 h versus 10.0 ± 0.7 in S; P < 0.05; Figure 1). AG treatment significantly reduced UNOxV in group S+AG (6.1 ± 1.2 μmol/24 h; P < 0.05 versus S) and promoted a numerical decrease in UNOxV in group Nx+AG (2.5 ± 0.6 μmol/24 h; NS versus Nx; P < 0.05 versus S).
Figure 1. Urinary excretion of NO2−/NO3− at 60 d after 5/6 nephrectomy (Nx). S, sham-operated; S+AG, S rats treated with aminoguanidine (AG); Nx, rats subjected to 5/6 renal ablation; Nx+AG, Nx rats treated with AG. *P < 0.05 versus respective S; †P < 0.05 versus respective untreated.
BW, TCP, urinary albumin excretion rate (UalbV), plasma creatinine concentration (Pcreat), and glycated hemoglobin (HbA1c), measured at 60 d after Nx, are shown in Table 2. In both group Nx and group Nx+AG, BW was lower compared with S (P < 0.05). However, AG-treated rats exhibited slightly higher body growth than untreated rats (P < 0.05). TCP was markedly elevated in Nx rats (P < 0.05 versus S) and was unchanged after AG treatment. UalbV was increased in Nx compared with S (P < 0.05). AG treatment promoted no significant change in albuminuria compared with Nx. Pcreat was twice as high in group Nx as in S (P < 0.05). In rats receiving AG treatment, Pcreat was significantly reduced relative to Nx. HbA1c levels were similar among groups, suggesting that Nx rats had no propensity toward enhanced nonenzymatic glycation.
Long-term studies at 60 d of treatmenta
Histomorphometric Analyses
Sixty days after Nx, glomerular sclerotic lesions were evident in untreated Nx rats. Accordingly, GSI was negligible in S (Figure 2) and markedly elevated in Nx (113 ± 32 versus 1 ± 0.3 in S; P < 0.05). AG treatment dramatically reduced GSI to 21 ± 6 in group Nx+AG, a value numerically higher but not significantly different from the respective control. Interstitial expansion was also prominent in untreated rats (Figure 2) and was similarly limited by AG treatment.
Figure 2. The glomerulosclerosis index (GSI), the percent renal cortical area occupied by interstitium (%INT), and the extent of renal macrophage infiltration (Mφ) 60 d after Nx. The vast majority of Mφ located in the interstitial area. * P < 0.05 versus respective S; †P < 0.05 versus respective untreated.
Western Blot Analyses
All three NOS isoforms were expressed in the renal tissue of sham-operated rats (Figure 3A). In Nx rats, the abundance of all NOS isoforms was significantly reduced compared with S. AG treatment did not modify the abundance of any of the NOS isoforms in either sham or Nx rats. Identical results were obtained for the iNOS protein using either the TL or SC primary antibodies (Figure 3B). An approximately 70-kD additional band was systematically detected in all groups by both the TL and SC antibodies. This band was reported in a number of studies involving Western blot analysis of renal iNOS (16,17,30). Its nature has not been elucidated, although it seems reasonable to presume that it originates from partial degradation of the iNOS protein during processing of the renal tissue (16)
Figure 3. (A) Western blot analysis of renal NOS isoforms. Blots illustrate representative experiments for S (lanes 1 and 2), S+AG (lanes 3 and 4), Nx (lanes 5 and 6), and Nx+AG (lanes 7 and 8). *P < 0.05 versus respective S; †P < 0.05 versus respective untreated. (B) Complete Western blot analysis of renal iNOS using the Transduction Laboratories (TL) and Santa Cruz (SC) antibodies. Blots illustrate representative experiments for S (lanes 1 and 2), S+AG (lanes 3 and 4), Nx (lanes 5 and 6), and Nx+AG (lanes 7 and 8). Besides the 130-kD iNOS band, an approximately 70-kD band, which was not further investigated, was consistently detected by both antibodies. No difference was observed between blots obtained with the TL and SC antibodies.
Immunohistochemical Analyses
Macrophage density in the renal cortex (Figure 2) was markedly increased in group Nx (151 ± 26 cells/mm2 versus 22 ± 2 in S; P < 0.05). There was strong linear correlation between the intensity of macrophage infiltration and the GSI, the interstitial area, or the Screat (P < 0.05). AG treatment significantly attenuated macrophage infiltration (94 ± 16 cells/mm2 in group Nx+AG; P < 0.05 versus Nx). The distribution of the three NOS isoforms in the renal tissue is illustrated in Figure 4. The eNOS isoform was widely expressed in sham-operated rats, appearing in the endothelial layer of the glomerular tufts, arteries, arterioles, and peritubular capillaries, as well as in tubular epithelial cells, especially in the medullary area. The nNOS isoform was detected almost exclusively in the cortical area, localizing mostly in the macula densa region. The pattern of distribution of the eNOS and nNOS isoforms in the renal tissue was similar in Nx rats. The renal expression and distribution of the iNOS isoform varied radically according to the primary anti-iNOS antibody that was used. With the SC antibody (in paraffin sections), the iNOS isoform was clearly expressed in tubular cells, in both the cortical and medullary areas. An identical distribution was observed in frozen sections (data not shown). When the TL antibody was used instead, no iNOS staining was obtained in paraffin-embedded tissue in any of the groups (data not shown). In frozen tissue, no iNOS protein was detected with the TL antibody in the S or S+AG groups (Figure 4). The results obtained with the SC and TL antibodies also diverged as to the effects of renal ablation; although the SC antibody revealed a decrease (relative to S) in the expression of (tubule-bound) iNOS in Nx rats, the TL antibody showed the appearance of iNOS in interstitial cells, mostly in inflamed areas. Figure 5 shows the estimated expression of this interstitial iNOS in the four groups studied. No glomerular iNOS expression was detected in any of the groups.
Figure 4. Representative micrographs illustrating the renal expression of the endothelial (eNOS), neuronal (nNOS), and inducible (iNOS) isoforms of the NO synthase, revealed by immunohistochemistry using the TL or SC antibodies.
Figure 5. Estimation of the renal expression of the inducible NO synthase isoform (iNOS) by immunohistochemistry, using the TL antibody. iNOS located almost exclusively in the interstitial area. *P < 0.05 versus respective S; †P < 0.05 versus respective untreated.
Treatment with AG had little impact on the immunohistochemical expression of eNOS or nNOS in either S or Nx rats (Figures 3 and 4). iNOS expression was also unchanged by AG treatment when the SC antibody was employed. However, when detection was performed with the TL antibody, AG treatment was shown to reduce significantly the expression of (interstitial) iNOS in NX+AG compared with Nx (Figures 4 and 5), in parallel with the amelioration of GS and interstitial inflammation.
Discussion
As expected, Nx rats exhibited progressive BP elevation and albuminuria. At 60 d after surgery, prominent glomerular sclerotic lesions and expansion of the renal cortical interstitial area were evident in Nx rats. Glomerular intracapillary hypertension and tuft hypertrophy, demonstrated in the Nx group 30 d after renal ablation, are likely to have initiated renal injury in these rats (1,2). In addition, the renal inflammatory process that developed in group Nx, indicated by the intense macrophage infiltration occurring in these animals at 60 d of ablation, likely contributed to propagate and perpetuate the initial insult (3,7,8,31).
In agreement with previous observations (13,15), UNOxV was decreased by more than 60% in group Nx compared with sham, suggesting that the overall production of NO was diminished in these rats. It should be noted that the UNOxV values obtained here may have overestimated the overall NO production because the diet was not controlled for NOx. However, because food intake was similar among groups, so must have been the respective dietary contributions to UNOxV; a fixed value would have to be subtracted from each UNOxV value to estimate NO production more precisely. Therefore, the differences in NO production among groups must have been even larger than inferred from Figure 1.
Given the vasodilator, antiproliferative, and antiplatelet actions of NO, basal, continuous NO production is believed to be central to circulatory regulation and to the maintenance of renal structural integrity (32). Accordingly, NO deficiency has been consistently shown in association with human (12) and experimental (13,15) renal insufficiency. Administration of an NO donor attenuated renal injury in the Nx model (33). Conversely, it is well established that chronic inhibition of NO synthesis promotes systemic hypertension and severe renal injury in intact rats (9,10) and aggravates renal damage in Nx rats (34).
The mechanisms by which the overall synthesis of NO is diminished in chronic nephropathies and in the Nx model have not been completely elucidated. Low NO production could reflect the renal mass reduction required by this model, although the contribution of the kidneys to the total NO synthesis may be small even in the normal condition when compared with the output of other organs (12). Decreased NOS abundance might promote additional depression of renal NO synthesis in the Nx model. In this study, it was possible to demonstrate decreased renal expression of eNOS and nNOS by both Western blot and immunohistochemical techniques. However, apparently inconsistent results were obtained regarding the renal iNOS expression. In the Western blot analysis, the abundance of iNOS in the renal tissue, which was evident in the sham group, fell by approximately 50% in Nx rats, irrespective of which primary antibody was used. These findings are in agreement with those reported by Aiello et al. (13), Ashab et al. (15), and Vaziri et al. (14). With the SC antibody, the immunohistochemical analysis showed a pattern that essentially confirmed that observed in the Western blots. We cannot establish on the basis of the present data whether this pattern reflects decreased cellular expression, tubular rarefaction, or both. However, the data obtained with the SC antibody seem to corroborate the concept that iNOS is constitutively expressed in the normal rat kidney and that 5/6 renal ablation reduces renal iNOS abundance, at least in relative terms. The physiologic importance of this tubular iNOS is unclear, because AG treatment decreased UNOx in both sham and Nx, presumably by inhibiting the synthesis of iNOS-derived NO, without inducing any apparent change in the renal or systemic hemodynamics. The iNOS expression may diminish in response to extracellular volume expansion (35); it is therefore conceivable that this tubular iNOS is involved in the physiologic regulation of sodium excretion, which might partly explain the low iNOS expression observed in Nx rats, in which expansion of the extracellular volume and increased fractional sodium excretion are expected (36).
With the TL antibody, the immunohistochemical iNOS pattern conflicted with both the Western blot results and the immunohistochemical profile obtained with the SC antibody: virtually no iNOS protein was seen in S kidneys, whereas the iNOS expression in Nx appeared exclusively in areas of interstitial inflammation, paralleling the intensity of macrophage infiltration and interstitial expansion. These findings, which confirm those reported by Bremer et al. (17), Bank et al. (16), and Heeringa et al. (37), are consistent with the notion that iNOS is absent from the normal rat kidney but plays a role in processes involving interstitial inflammation, such as observed in the remnant kidney. It should be stressed that enhanced iNOS expression has been reported in association with a number of clinical and experimental nephropathies (22,23,38) as well as in inflammatory disorders occurring in other tissues (39,40).
The reason for the striking discrepancy between the results obtained with the two primary antibodies, which nevertheless yielded identical data in the Western blot analysis, is unclear. One possibility is that, within the tubular cell, iNOS binds to other cytosolic molecules, such as calmodulin (41), which could conceal the epitope for the TL antibody, preventing its detection by immunohistochemistry, but not by Western blot analysis (which requires denaturation of the sample). This explanation is merely speculative in the absence of additional data. However, the present study does indicate that, given the variable properties of the available primary antibodies, caution should be taken in the interpretation of iNOS immunohistochemical data.
Taken as a whole, the immunohistochemical and Western blot data obtained in this study suggest that iNOS in the kidney may exert two distinct roles according to the location and the nature of the cells in which it originates. The dominant tubular fraction appears to be expressed in the normal condition and may be involved in the physiologic regulation of sodium excretion. Its functional importance is unclear, however, because chronic AG administration causes no hemodynamic effect in normal rats, as shown in this and in previous studies (42), although it may facilitate the development of salt-sensitive hypertension (43). The interstitial fraction appears to exert little function in the normal condition. Rather, its activity seems related to the pathogenesis of inflammation and tissue injury in such conditions as glomerulonephritis and chronic renal insufficiency. Thus the role of NO in the Nx model (and possibly in other forms of chronic nephropathy) may be more complex than initially thought. Although overall diminished NO synthesis in this model appears incontrovertible, interstitial NO generation via iNOS, possibly expressed by interstitial inflammatory cells, may represent an important mediator of injury and facilitate the progression to renal failure. The concept of two functionally distinct renal iNOS isoforms is consistent with that proposed by Mohaupt et al. (44), who showed that normal rat kidneys expressed a “constitutive” and an “inducible” iNOS, with distinct anatomical distributions and divergent responses to lipopolysaccharide stimulation.
The second important finding of this study is the clear protective effect of AG. Albuminuria was significantly reduced by AG treatment, as were both the extent and the frequency of glomerular sclerotic lesions, leading to a dramatic decrease of the GSI. Likewise, the expansion of the cortical interstitial area was strongly attenuated, as was the estimated expression of interstitial iNOS. AG also tended to correct the imbalance between the two iNOS fractions, numerically increasing the expression of tubular iNOS and reducing that of interstitial iNOS. Accordingly, Pcreat was lowered in group Nx+AG compared with Nx, indicating amelioration of the renal function.
The beneficial effect of AG in the present study was not associated with a hemodynamic effect, because neither BP nor PGC were changed by treatment. Glomerular tuft volume was also unaffected; the protective effect exerted by AG cannot therefore be ascribed to lessening of the mechanical strain to the glomerular wall. Limitation by AG of a toxic effect of filtered protein (45) is also unlikely to explain its protective effect, because proteinuria was not significantly reduced by AG treatment. Therefore, renal preservation by AG treatment must have resulted from some direct effect of the compound.
The mechanisms by which AG might prevent inflammation and progression of the renal injury have not been elucidated. One major effect of AG is to inhibit the formation of AGE, which results from nonenzymatic glycation of proteins, especially those taking part in the structure of tissues. This reaction has been postulated as a central mechanism in the pathogenesis of progressive nephropathies secondary to diabetes mellitus and aging, and could explain the beneficial effects of AG in these conditions, although this hypothesis has been disputed (25,26). In this study, however, the protective effect of AG can hardly be attributed to the prevention of AGE formation: the Nx model is not associated with a tendency toward increased nonenzymatic glycation of proteins, as indicated by the similar levels of HbA1c observed in the sham, Nx, and Nx+AG groups.
AG administration to either S or Nx rats is unlikely to have significantly inhibited the constitutive NOS, because such effect would have been expected to promote perceptible alterations in renal and/or systemic hemodynamics. Therefore, AG must have exerted a relatively specific inhibition on the iNOS isoform. Given the proinflammatory action postulated for iNOS and its anomalous presence in the renal interstitium of Nx rats, it seems reasonable to ascribe the renal protective effect of AG to its iNOS inhibition, although we cannot exclude other effects of AG, such as inhibition of diamine oxidase and stabilization of S-adenosylmethionine decarboxylase (46). The mechanisms by which iNOS-derived NO might favor the development of progressive renal injury are presently unclear. Increased production of NO may exert a cytotoxic effect either directly or, alternatively, by reacting with superoxide to form the much more toxic peroxynitrite anion, especially when it is massively produced by macrophages (47). Additional tissue injury may result from NO-induced cell apoptosis (48) and from enhanced prostanoid synthesis consequent to cyclooxygenase activation (49). It should also be noted that iNOS can produce superoxide anions rather than NO under conditions of absolute or relative l-arginine deficiency (50,51) and could therefore promote NO-independent renal injury.
In summary, this study provides evidence of the existence of two functionally distinct iNOS fractions in the rat kidney: a tubular fraction, which is present in the normal condition and possibly related to sodium handling; and an interstitial fraction, associated with interstitial inflammation. AG treatment greatly attenuated renal injury in Nx rats by a direct antiinflammatory effect, likely related to iNOS inhibition. Further investigation is needed to establish the exact function of these iNOS fractions and to ascertain whether AG may help to prevent the progression of human renal diseases not associated with diabetes or aging.
Acknowledgments
This work was supported by grants 95/4710–2 and 98/09569–4 from the São Paulo Foundation for Research Support (FAPESP). Preliminary results of this study were presented at the 30th Congress of the American Society of Nephrology, San Antonio, TX, November, 1997, and published in abstract form (J Am Soc Nephrol 8: 615A, 1997). During these studies, RZ was the recipient of Research Award 326.429/81 from the Brazilian Council of Scientific and Technologic Development (CNPq). We are grateful to Ms. Gláucia R. Antunes, Ms. Cláudia R. Sena, Ms. Cristiane Y. N. de Oliveira, Ms. Marinete M. dos Santos, and Ms. Márcia Regina Soares Correia for expert technical assistance and to Dr. Francisco R. M. Laurindo for help with the NO analyzer.
- © 2002 American Society of Nephrology
References
