INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE PATHOGENESIS of progressive nephropathies involves the participation of a host of pathogenic factors, including glomerular hypertension (18), glomerular hypertrophy (7), formation of intracapillary microthrombi (19), and inflammation of the glomerulus and interstitium (15). Vasoactive compounds, such as cyclooxygenase derivatives (20, 23) and nitric oxide (NO; see Refs. 3 and 5), are also likely to participate in this process.

Increased renal synthesis of cyclooxygenase products, such as vasodilator prostaglandins (PG) and thromboxane (TX; see Refs. 17 and 30), is likely to contribute to the single nephron hyperfiltration and hyperperfusion associated with renal mass reduction, such as observed in the 5/6 ablation model (NX). Because prostanoids may also mediate chronic inflammatory processes (25), increased synthesis of these compounds may be maladaptive and may contribute to the pathogenesis of the progressive nephropathy associated with renal mass reduction. Chronic treatment with specific TX inhibitors ameliorates glomerular injury in NX rats (20, 30). However, the possible beneficial effect of inhibiting the entire cyclooxygenase pathway in progressive renal disease has not been evaluated, despite the current availability of powerful nonsteroidal anti-inflammatory drugs (NSAID). Two major untoward effects of NSAID severely restrict the chronic use of these compounds. First, NSAID can severely depress glomerular filtration rate (GFR) in patients with chronic renal failure by inhibiting the synthesis of vasodilator PG (21). Second, severe ischemia of the gastrointestinal mucosa, with consequent ulceration, can occur in association with chronic NSAID treatment, especially in the rat (26).

In recent years, the role of NO in the regulation of the circulation and in the pathogenesis of renal disease has been the focus of intensive investigation. Chronic NO inhibition leads to progressive arterial hypertension, glomerular ischemic injury, glomerulosclerosis (GS), and interstitial inflammation (22), suggesting that the hemodynamic and cellular effects of NO are essential to maintain renal integrity and circulatory homeostasis. The biosynthesis of NO may be impaired in NX rats (3), whereas chronic administration of exogenous NO may attenuate GS in this model (5). Similar effects have been reported after chronic administration of L-arginine, the biochemical precursor of NO (3).

Recently, a new class of NSAID has been developed by attachment of a nitroxybutyl moiety to a conventional NSAID molecule (26). These compounds are capable of donating NO in vivo, thus preventing the ischemia and consequent ulceration of the gastrointestinal mucosa associated with cyclooxygenase inhibition. These characteristics now make it feasible to assess a possible protective effect of chronic cyclooxygenase inhibition in the NX model. Additional benefit might derive from the NO-donating properties associated with these drugs.

In the present study, we examined the renal effects of chronic treatment with one of these new compounds, nitroflurbiprofen (NOF), in the NX model.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Male Munich-Wistar rats (202) obtained from our own breeding colony, with initial weights of 245-270 g, were used in this study. All rats received food (22% protein) and tap water ad libitum. Surgical reduction of renal mass (NX) was performed under anesthesia with sodium Nembutal, 50 mg/kg ip, by removal of the right kidney and infarction of approximately two-thirds of the left kidney. Sham-operated rats (Sham) underwent anesthesia, ventral laparotomy, and manipulation of the renal pedicles, without removal of renal mass. NOF was dissolved in a mixture of dimethyl sulfoxide and olive oil, with the final concentration of dimethyl sulfoxide being 5%. The compound was given by gavage two times daily (at 8:00 AM and 4:00 PM) in a volume of vehicle that never exceeded 0.3 ml. Untreated rats received vehicle only. All experimental procedures were conducted in accordance with our institutional guidelines.

Experimental groups. The following four experimental groups were studied: Sham (n = 23), sham-operated rats receiving vehicle; Sham + NOF (n = 23), sham-operated rats receiving flurbiprofen-4-nitroxybutylester, also designated as NOF, 30 mg · kg¹ · day¹ (41 µmol/kg) divided in two doses; NX (n = 31), rats subjected to 5/6 renal infarction and treated with vehicle only; and NX + NOF (n = 27), rats receiving NOF as in Sham + NOF group.

In separate NX (n = 5) and Sham (n = 5) rats, the parent compound flurbiprofen was given by gavage at 14 mg · kg¹ · day¹ (29 µmol · kg¹ · day¹) divided in two doses of 7 mg/kg. This dosage of flurbiprofen was found to yield a similar pharmacokinetic profile as after NOF treatment (see below). As described previously (26), Sham rats receiving flurbiprofen lost an average of 7 g after only 1 mo of treatment, as opposed to a mean weight gain of 70 g in untreated rats. Similar results were obtained in NX rats receiving flurbiprofen. Both groups also exhibited diminished food intake, severe anemia, and poor general condition at this time. Comparing these groups with those enumerated above would be meaningless, since any renal protective effect of flurbiprofen would be indistinguishable from those derived from weight loss (24) or anemia (8). For these reasons, these groups were not analyzed further.

Kinetic studies. Sixteen Sham and 16 NX rats that had received no prior treatment were given a single dose of NOF, 15 mg/kg, 2 wk after the surgical procedure. An equal number of Sham and NX rats received a single dose of flurbiprofen, 7 mg/kg. Blood samples (600-800 µl) were obtained from a tail vein by a skilled operator 30 min (4 rats) and 1 (4 rats), 4 (4 rats), and 12 (4 rats) h after administration of NOF. An identical procedure was adopted for rats receiving flurbiprofen. The blood samples were collected into Eppendorf vials containing 1 mg of EDTA. The samples were immediately chilled, protected from light, and spun in a refrigerated centrifuge, and the plasma was stored at 20°C for future processing.

In rats given flurbiprofen, the plasma concentrations of this drug were determined by a reverse-phase chromatographic method with ultraviolet detection at 276 nm. Because NOF generates flurbiprofen in vivo (26), the plasma concentrations of both flurbiprofen and NOF were determined in the same fashion in rats receiving NOF. Before extraction, 100-µl plasma samples were mixed with 100 µl of a 10-µg/ml solution of flurbiprofen-butylester, used as an internal standard. This mixture was diluted in 1.5 ml of 2.5% acetic acid and then was loaded at a low flow rate by manual positive pressure into a solid-phase extraction cartridge (Waters C₁₈ Sep-Pak cartridges; Millipore, Milford, MA) previously washed with 10 ml of ultrapure water, conditioned with 5 ml of acetonitrile, and flushed with 10 ml of 2.5% acetic acid. After being flushed with 8 ml of 2.5% acetic acid, the compounds of interest were eluted in 2 ml of acetonitrile, which was evaporated to dryness under an N₂ stream at 37°C. The residue was dissolved in 200 µl of solution A (see below). A 100-µl aliquot of this new solution was used for chromatographic analysis. Freshly pooled rat plasma aliquots, spiked with flurbiprofen or NOF, were used as standards. The concentrations used to generate the calibration curves were 0.5, 1.0, 2.5, 5, 10, 20, and 50 µg/ml.

For chromatographic analysis, a linear gradient method, involving the admixture of two solutions, was employed. Solution A consisted of 0.1 M sodium acetate, pH 4.2, and acetonitrile, 60:40, vol/vol. Solution B consisted of 0.1 M sodium acetate, pH 4.2, and acetonitrile, 35:65, vol/vol. The gradient, expressed as the percentage of solution B in the mixture, had the following time profile: 0-100% in the first 20 min; 100% from 20 to 25 min; 100-0% from 25 to 30 min; and 0% from 30 to 35 min. The mobile phase was delivered through the chromatographic column (Spherisorb ODS-2, 5 µm, 250 × 4.6 mm ID; Sigma Aldrich) at a flow rate of 1.6 ml/min.

The chromatographic system consisted of two Shimadzu LC-10 AD V 3.1 pumps (Shimadzu) coupled to a photodiode array detector (1000S; Applied Biosystems). The eluent was monitored at 276 nm. The signal output was recorded on a computerized integrator (CI 4000; Milton Roy, Riviera Beach, FL). The samples were manually injected through an injector valve (Rheodyne 7125; Rheodyne, Cotati, CA) equipped with a 200-µl loop. The retention times for flurbiprofen, flurbiprofen-butylester, and NOF were 9.5, 14, and 24.5 min, respectively. Linearity was observed up to 50 µg/ml for flurbiprofen and up to 20 µg/ml for NOF. The sensitivity of the method (taken as 3 times baseline) was 0.5 and 1.0 µg/ml for flurbiprofen and NOF, respectively.

Short-term studies. To examine renal functional and structural parameters, eight rats of each group were subjected to micropuncture experiments 1 mo after renal ablation. Rats were anesthetized with Inactin (100 mg/kg of body weight ip) and placed on a temperature-regulated table. The femoral artery was catheterized for determination of baseline arterial hematocrit (Hct) and subsequent periodic blood sampling, as well as for continuous monitoring of mean arterial pressure (MAP) with a P23Db Statham pressure transducer connected to a chart recorder (model TA 240; Gould). Euvolemia was maintained by intravenous infusion of homologous rat plasma. Saline solution containing [¹⁴C]inulin (0.3 µCi/ml) was infused at 1.5 ml/h. After ~2.5 h of anesthesia, urine was collected from the left ureter for 20-30 min for the determination of flow rate and inulin concentration. Hydraulic pressures in superficial glomeruli (P_GC), tubules, and arterioles were determined with a servo-nulling device (model V; Instrumentation for Physiology and Medicine, San Diego, CA). Whole kidney filtration fraction (FF) was determined by the simultaneous collection of blood samples from the femoral artery and renal vein and the assessment of the respective ¹⁴C activities to calculate inulin extraction. Blood samples were obtained from the renal vein with a sharpened glass micropipette, 40-45 µm OD. Plasma and urine ¹⁴C activities were measured in a scintillation counter (Beckman Instruments, Shiller Park, IL). Renal plasma flow (RPF) was calculated as RPF = GFR/FF. Renal vascular resistance (RVR) was estimated by the expression RVR = MAP(1  Hct)/RPF. At the end of the experiment, the renal tissue was fixed by perfusion at the measured arterial pressure with 1.25% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4). After fixation, two midcoronal sections through the midportion of the remnant kidney were postfixed in buffered formaldehyde solution and processed for light microscopy through paraffin embedding. Sections 2-3 µm in thickness were stained with the periodic acid Schiff reaction and by the Masson trichrome technique.

The average glomerular tuft volume (V_G) for each rat was estimated by the method of Weibel (28), after light microscopic examination at a final magnification of ×320 under a 100-point ocular grid. The corresponding microscopic field covered an area of 133.932 µm². The mean glomerular cross-sectional area (A_G) was determined for each rat by averaging individual values for at least 50 randomly sampled glomerular tuft profiles. Individual glomerular values were calculated by counting points falling within the glomerular area. V_G was then calculated as V_G = 1.25(A_G)^3/2 (11).

Long-term studies. Fifteen rats from the Sham group, 15 rats from the Sham + NOF group, 23 rats from the NX group, and 19 rats from the NX + NOF group were monitored during 60 days of treatment. Urinary albumin excretion rate was measured at 30 and 60 days after ablation. At the end of the study period, awake systemic arterial pressure was evaluated by a tail-cuff method (29). Plasma creatinine concentration and the urinary excretion rate of / were also determined at this time. Rats were then anesthetized with pentobarbital sodium, 50 mg/kg ip. The remnant kidneys were perfusion fixed, weighed, and prepared for histological examination as described above. The extent of GS was evaluated by attributing a score to each glomerulus according to the extent of sclerotic injury (0, intact glomeruli; 1, lesions affecting 20% or less of the glomerular area; 2, lesions affecting 21-40% of the glomerular area; 3, lesions affecting 41-60% of the glomerular area; 4, lesions affecting 61-80% of the glomerular area; and 5, lesions exceeding 80% of the glomerular area). A glomerulosclerosis index (GSI) was calculated for each rat as the weighted average of all individual glomerular scores thus obtained. At least 150 glomeruli were examined for each rat. To assess the extent of interstitial expansion, the fraction of renal cortex occupied by interstitial tissue staining positively for collagen was quantitatively evaluated in Masson-stained sections by a point-counting technique (14) in 25 consecutive microscopic fields, at a final magnification of ×400 under a 100-point ocular grid.

Analytic methods. Urinary albumin excretion rate was determined by radial immunodiffusion. Total plasma protein concentration was determined by refractometry. Serum creatinine concentration was assessed by a colorimetric method (10). The concentrations of and were assayed by first determining the concentration using the Griess reagent. Additional urine aliquots were incubated with commercially available Escherichia coli nitrate reductase and NADPH (Boehringher, Mannheim, Germany) to reduce all urinary to . The Griess reaction was then again employed to determine the concentration, now equivalent to the sum of the original urinary concentrations of and .

Platelet aggregation and TX generation. Thirty days after NX, six rats from the Sham group, six rats from the Sham + NOF group, six rats from the NX group, and six rats from the NX + NOF group were anesthetized with pentobarbital sodium, 50 mg/kg, and blood was collected from the abdominal aorta in a plastic tube containing 3.18% sodium citrate. Platelet-rich plasma was obtained by centrifugation at 200 g during 15 min. The remaining blood was centrifuged at 2000 g for an additional 15 min to obtain platelet-poor plasma.

Platelet aggregation was evaluated in platelet-rich plasma obtained by pooling individual plasma aliquots in each group. Aggregation was estimated by optical density variation in a two-channel aggregometer (Payton Scientific Instruments, Buffalo, NY), after stimulation with 5 µM of ADP. Five minutes after addition of ADP, aggregation was interrupted by addition of 100 µl of 100 mM EDTA. Platelet-rich plasma aliquots were then immediately chilled and centrifuged at 12,000 g for 2 min. Additional aliquots were preincubated with 10 µM indomethacin for 20 min as a negative control. The concentration of TXB₂, the major product of degradation of TXA₂, was measured by a radioimmunoassay technique.

Statistics. One-way analysis of variance with pairwise comparisons according to the Bonferroni method was employed in this study (27). Statistical significance was considered at P levels of 0.05 or less. Because GSI and albumin excretion rates were not normally distributed, log transformations were performed before statistical analysis of these parameters.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Kinetic studies. Results of the kinetic studies performed in Sham rats are shown in Fig. 1. NOF was detected at a concentration of 1.6 ± 0.8 µg/ml only at 30 min after oral administration of the drug. The plasma concentration of NOF-derived flurbiprofen at this time was nearly sixfold higher (9.5 ± 2.7 µg/ml), increasing to 15.7 ± 6.2 µg/ml at 1 h and returning to 9.6 ± 4.4 at 4 h. Flurbiprofen was still detected 12 h after intake of NOF (2.8 ± 0.8 µg/ml). Similar results were obtained in NX rats (data not shown). The time profile of flurbiprofen after oral administration of the compound is also shown in Fig. 1 for comparison. The peak concentration obtained in these animals occurred earlier than in NOF-treated rats, although its magnitude was almost two times as high (27.6 ± 11.1 µg/ml). Nevertheless, the areas under the two curves were similar (94.8 mg · h · l¹ after flurbiprofen vs. 96.2 after NOF).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Time-dependent plasma concentrations of nitroflurbiprofen () and nitroflurbiprofen-derived flurbiprofen () in sham rats receiving a single oral dose of nitroflurbiprofen (15 mg/kg). Time course of plasma flurbiprofen concentration after a single oral dose of flurbiprofen (7 mg/kg) is shown for comparison ().

Short-term studies. Renal and systemic functional and hemodynamic parameters at 30 days after renal ablation are given in Table 1. In contrast to the severe untoward effects of flurbiprofen observed in our preliminary studies and elsewhere (26), administration of NOF to Sham rats did not promote body growth stunting or anemia. All other functional and structural parameters were likewise identical between the Sham and Sham + NOF groups. Average food intake was similar among groups (21.1 ± 3.0 g/day in Sham, 20.1 ± 2.5 in Sham + NOF, 20.1 ± 1.1 in NX, and 19.5 ± 1.9 in NX + NOF).

Renal mass reduction was associated with slightly lower body weights and arterial Hct in all NX groups compared with Sham. NOF treatment promoted no further slowing of body growth or depression of Hct. Renal ablation was associated with elevation of MAP in the NX groups (131 ± 6 mmHg in NX, P < 0.05 vs. Sham and 124 ± 4 mmHg in NX + NOF, P > 0.05 vs. NX and P > 0.05 vs. Sham + NOF). Left kidney weight fell by only 18% after renal ablation (1.28 ± 0.06 g in NX group vs. 1.55 ± 0.03 in Sham, P < 0.05). This indicates the occurrence of marked hypertrophy of the remnant tissue, since renal mass was reduced by 83% in these rats. Treatment with NOF promoted no change in left kidney weight compared with NX (1.29 ± 0.08 g in NX + NOF group, P < 0.05 vs. Sham + NOF and P > 0.05 vs. NX). By a similar reasoning, remnant nephrons must have undergone marked hyperfiltration, since whole kidney GFR fell by only 41% after renal ablation (0.81 ± 0.11 ml/min in NX group vs. 1.37 ± 0.06 in Sham, P < 0.05). The administration of NOF promoted no further alteration in GFR (0.79 ± 0.10 in NX + NOF group, P > 0.05 vs. NX). The NX + NOF group exhibited a significant decrease in FF compared with the Sham + NOF group. P_GC was markedly elevated in the NX group (76 ± 3 vs. 53 ± 1 mmHg in Sham, P < 0.05). NOF treatment decreased P_GC by 7 mmHg (69 ± 2 mmHg, P < 0.05 vs. Sham + NOF and P > 0.05 vs. NX), although this difference failed to reach statistical significance. No differences in plasma protein concentrations were observed among groups. Predictably, RVR increased in the NX group (30.0 ± 4.6 mmHg · ml¹ · min¹ vs. 14.6 ± 0.5 in Sham, P < 0.05) compared with Sham. No additional change in RVR was induced by NOF treatment. Glomerular volumes were 33% larger in the NX group compared with sham-operated controls (1.32 ± 0.06 × 10⁶ µm³ vs. 1.01 ± 0.06 in Sham, P < 0.05). Glomerular hypertrophy was less pronounced in the NX + NOF group (1.23 ± 0.04 × 10⁶ µm³, P < 0.05 vs. Sham + NOF and P > 0.05 vs. NX).

Urine albumin excretion rate was markedly increased in NX rats (63.2 ± 6.8 vs. 1.6 ± 0.3 mg/24 h in Sham, P < 0.05). NOF treatment markedly reduced albuminuria (37.6 ± 5.8 in NX + NOF, P < 0.05 vs. Sham + NOF and NX). No alteration was induced by NOF treatment in Sham rats.

In pooled plasma obtained from the Sham + NOF group, platelet aggregation was reduced by 48% compared with Sham. Likewise, a 35% reduction was observed in the NX + NOF group compared with the NX group. Ex vivo platelet TXA₂ generation was nearly completely abolished in the Sham + NOF and NX + NOF groups compared with the respective untreated controls.

Long-term studies. The excretion of NO metabolites was increased by 58% in the Sham + NOF group, compared with the Sham group (Table 2). Urinary excretion of / was 31% lower in the NX group compared with Sham (P > 0.05). Compared with NX, NOF treatment increased / excretion by 248% in the NX + NOF group (P < 0.05).

Parameters evaluating renal and systemic hemodynamics, renal function, and renal structural injury after 60 days of ablation are shown in Table 2. Body weight was slightly but significantly lower in the NX group compared with Sham (307 ± 6 g vs. 333 ± 2 in Sham, P < 0.05). No further limitation of body growth occurred as a consequence of NOF treatment. Likewise, no difference in kidney weight among groups was observed at this time (data not shown). Tail-cuff pressure was markedly elevated in NX rats (152 ± 6 mmHg vs. 111 ± 2 in Sham, P < 0.05). TCP was numerically lowered by NOF treatment (137 ± 4 mmHg in NX + NOF group, P < 0.05 vs. Sham + NOF and P > 0.05 vs. NX). Plasma creatine concentration was markedly elevated in the NX group compared with Sham (1.09 ± 0.10 mg/dl vs. 0.62 ± 0.05 in Sham, P < 0.05). NOF treatment reduced plasma creatine concentration to 0.83 ± 0.03 mg/dl (P < 0.05 vs. NX and P > 0.05 vs. Sham + NOF). Urinary albumin excretion rate was markedly increased in NX relative to Sham (64.8 ± 7.5 mg/24 h vs. 1.3 ± 0.9 in Sham, P < 0.05). Albuminuria was attenuated but not normalized by treatment with NOF (38.1 ± 3.7 mg/24 h, P < 0.05 vs. Sham + NOF and NX). As expected, the GSI was markedly elevated in the NX group (22.1 ± 9.5, P < 0.05 vs. Sham). Treatment with NOF dramatically reduced GSI in the NX + NOF group (Fig. 2), although values remained significantly higher than in sham-operated rats (3.5 ± 0.7, P < 0.05 vs. Sham + NOF and NX). As shown in Fig. 3, renal ablation was also characterized by an increase in the fraction of renal parenchyma occupied by interstitial tissue (9.9 ± 1.2% in NX group vs. 0.9 ± 0.1 in Sham, P < 0.05). Treatment with NOF reduced interstitial area in the NX + NOF group (6.4 ± 0.8%, P < 0.05 vs. Sham + NOF and NX).

View larger version (9K): [in this window] [in a new window]   Fig. 2.   Bar graph representing the glomerulosclerosis index in the four groups studied. S, sham-operated rats receiving vehicle; S + NOF, sham-operated rats receiving flurbiprofen-4-nitroxybutylester, also designated as nitroflurbiprofen (NOF) (30 mg · kg1 · day1) (41 µmol/kg) divided in two doses; NX (n = 31), rats subjected to 5/6 renal infarction and treated with vehicle only; and NX + NOF (n = 27), rats receiving NOF as in S + NOF group. * P < 0.05 vs. respective sham;  P < 0.05 NX + NOF vs. NX.

View larger version (9K): [in this window] [in a new window]   Fig. 3.   Bar graph representing the percent cortical interstitial area in the four groups studied. * P < 0.05 vs. respective sham;  P < 0.05 NX + NOF vs. NX.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

As shown previously (26), NOF acted as an efficient cyclooxygenase inhibitor in both Sham and NX rats, nearly abolishing ex vivo platelet TX synthesis. Nevertheless, NOF-treated rats exhibited no anemia and grew at the same rate as untreated animals. By contrast, rats treated with the native compound flurbiprofen exhibited severe anemia and progressive weight loss, indicating the occurrence of intense and persistent gastrointestinal bleeding (26). The GFR depression often associated with NSAID treatment in subjects with chronic renal failure (17, 21) was also absent in NX rats receiving NOF.

Not only was NOF devoid of nephrotoxic effects, it also substantially mitigated albuminuria, GS, interstitial injury, creatinine retention, and systemic hypertension in NX rats. The mechanisms underlying these protective effects are unclear. NOF may have lessened GS by limiting platelet aggregation, hence preventing the formation of intracapillary microthrombi (19). In addition, vasodilator prostanoids may mediate the glomerular hyperperfusion and hyperfiltration associated with NX (17). The biosynthesis of vasodilator PG and TX in NX rats might also increase in such proportion as to promote predominant afferent vasodilation, therefore favoring glomerular hypertension (2). Because the latter may initiate progressive GS (18), the beneficial effect of NOF may have resulted in part from the mild lowering of P_GC observed in treated rats. Additional protection may have derived from the observed attenuation of glomerular hypertrophy (7). In particular, concomitant lowering of glomerular hypertension and hypertrophy may have reduced considerably the capillary wall strain (6), helping to prevent mechanical injury to the glomerulus (9). However, the effects of NOF on P_GC (9%) and V_G (7%) were modest and may be insufficient to explain the substantial reduction of glomerular injury afforded by the drug. In addition, amelioration of cortical interstitial expansion, another prominent effect of NOF treatment, is less readily attributable to the attenuation of glomerular hypertension or hypertrophy. However, the possibility that NOF prevented further elevations of glomerular pressure after 30 days of NX cannot be excluded. In addition, because only superficial glomeruli were evaluated, the possibility remains that NOF effectively ameliorated glomerular hypertension in deeper glomeruli.

The protective effect of NOF treatment at the glomerulus and interstitium may have resulted directly from its anticyclooxygenase activity. Cyclooxygenase activation may mediate chronic inflammatory processes as disparate as rheumatoid arthritis (12) and chronic granulomatous inflammation (25). The pathogenesis of progressive renal disease may incorporate several elements of chronic inflammatory processes, such as macrophage activation, fibroblast proliferation, and extracellular matrix overproduction (15). Vasodilator PG have also been shown to influence cellular events, such as mesangial cell migration (13) and proliferation (16), which may impact on the development of progressive renal disease. Chronic cyclooxygenase inhibition may have limited all of these processes in the present study. Additionally, NOF may have directly ameliorated the glomerular barrier function (21), thus limiting the interstitial damage that might result from massive proteinuria (1). Inhibition of TX synthesis may also have contributed to the protective effect of NOF. In addition to its effect on platelet aggregation, TX enhances cell proliferation (23) and extracellular matrix production (4) and might therefore facilitate renal injury in NX. Accordingly, specific TX inhibition attenuated renal damage in NX rats (20, 30). The present observations constitute the first evidence that chronic treatment with cyclooxygenase inhibitors may ameliorate progressive renal disease. Unfortunately, the hypothesis that suppression of TX and prostanoids underlay this protective effect could not be directly tested due to the severe untoward effects elicited by the parent compound.

NX rats exhibited a numerical decrease in the urinary excretion of NO metabolites compared with Sham rats, consistent with recent studies suggesting deficient NO synthesis in the remnant kidney (3). Chronic administration of L-arginine may prevent glomerular injury (3), although preliminary evidence suggested that chronic treatment with the NO donor molsidomine may attenuate GS in NX rats (5). NOF acted as an efficient NO donor, markedly increasing the urinary excretion of / in treated rats. This property may well explain the favorable effects of the drug at the renal tissue, a hypothesis that cannot be excluded on the basis of the present data. However, the kinetic measurements performed in the present study showed that the plasma levels of NOF were always far lower than those attained by its derivative, flurbiprofen. This observation indicates that, at least in the rat, NOF is a short-lived compound that decomposes almost immediately after entering the circulation, generating flurbiprofen and NO. It appears unlikely that this surge of NO, known to be an extremely labile compound, can promote a lasting pharmacodynamic effect such as observed with molsidomine and other NO donors (5). Rather, flurbiprofen generation (hence cyclooxygenase inhibition) may account for the therapeutic effects of NOF, whereas the NO-donating capability of this new compound is likely to explain its low gastrointestinal toxicity (26).

In summary, the novel NSAID NOF ameliorates glomerular and interstitial injury in the remnant kidney, without promoting anemia, weight loss, or renal functional depression. The mechanisms underlying this effect could not be directly attributed to NO donation and may result entirely from the attending inhibition of cyclooxygenase derivatives.