NITRIC OXIDE SYNTHASE ACTIVITY IN RENAL ISCHEMIA-REPERFUSION INJURY IN THE RAT

Implications for Renal Transplantation¹

Abstract

Cadaveric kidney transplants with delayed graft function have poorer graft survival by an unknown mechanism. Nitric oxide, produced by nitric oxide synthase (NOS), has a proven role in both recovery of ischemia and promotion of rejection. We therefore wished to study the patterns of NOS activity in a model of renal ischemia. The left renal pedicle of Fisher rats was occluded for 1 hr. Both kidneys were removed at various times and frozen. Renal NOS activity was measured by conversion of [³H]arginine to[³H]citrulline and the content of endothelial NOS isoenzyme (eNOS) was compared by Western blot. NOS activity increased significantly in the left ischemic kidney over the first 24 hr, from a control of 33.8 pmol/min/mg to 79.8 at 2 hr and 56.8 at 24 hr. NOS activity then dropped below baseline, returning to near normal levels at day 21. eNOS content was stimulated over the entire time course, consistent with the presence of an eNOS inhibitor. Oral treatment with the NOS substrate L-arginine at 5 g/L significantly hastened the return of serum creatinine to baseline, if simultaneous contralateral nephrectomy was performed. The lazaroid U74389G given perioperatively also improved renal function and hastened recovery of NOS activity. Because nitric oxide plays an important role in maintaining blood flow during recovery from renal ischemia, the observed decrease in NOS activity may be prevented by perioperative treatment with oral L-arginine and corticosteroids. In addition, U74389G may provide a clinically useful method of minimizing and/or shortening DGF, thereby improving graft function and survival.

Ischemia-reperfusion injury is unavoidable in cadaveric renal transplantation and can contribute to transplant acute tubular necrosis(ATN^*) and delayed graft function (DGF). Clinically, DGF is associated with increased hospital costs and decreased allograft survival⁽¹⁾. The mechanism of decreased graft survival in DGF is unclear, but recent evidence suggests that the response to renal injury can increase the immunogenicity of the kidney ⁽²⁾, leading to increased acute rejection. In addition, cytokines and growth factors associated with chronic allograft nephropathy are released⁽³⁾. Consequently, graft survival could be improved if ischemia-reperfusion injury were reduced or if the mediators responsible for the link between DGF and graft loss were identified and modulated.

Nitric oxide (NO), produced by the nitric oxide synthase (NOS) enzymes, is a potentially key molecule in the link between ischemia-reperfusion injury and rejection in the kidney ⁽⁴⁾. NO causes relaxation of preglomerular arteries ⁽⁵⁾, improving renal blood flow and oxygenation ⁽⁶⁾. The constitutive enzyme, endothelial NOS(eNOS), is produced by endothelial cells in the kidney and modulates vascular tone. Because a renal transplant is denervated, a contributory role of neuronal NOS (nNOS) is much less likely. By contrast, NO produced in larger quantities by the inducible NOS (iNOS) enzyme, has been implicated as a mediator of alloimmune damage ^{(7, 8)}. Therefore, the activity of NOS enzymes after ischemia-reperfusion injury can influence the degree of ischemic damage and the rate of repair during the injury response and can interact with alloimmune molecules. In addition, because there are substrates and drugs that can stimulate or block NOS activity, modulation of NOS in ATN may be a therapeutic goal as well.

We studied the level of NOS activity in the rat kidney during the injury response in a model of native unilateral ischemic ATN, from 2 hr after injury through 3 weeks after injury. We then correlated these findings with changes in serum creatinine (with contralateral nephrectomy). Finally, we tested pharmacologic interventions in this model, designed to reduce reperfusion injury or modify NOS activity.

MATERIALS AND METHODS

Surgery. Unilateral ischemic ATN was produced in the left kidney, as previously reported ⁽²⁾. Male F344 rats, weighing from 150 to 250 g, were used. After anesthesia with ketamine (35 mg/kg i.m.) and xylazine (18 mg/kg i.m.), the animal was placed on a warming pad. The left renal pedicle was identified through a midline incision and occluded with a micro bulldog clamp for 60 min. Reperfusion was confirmed visually after clamp removal. Sixty minutes was selected as the optimal length of time for warm ischemia, to reliably produce histologic ATN without thrombosis⁽²⁾. In experiments designed to monitor the serum creatinine, the right kidney was removed, using mass ligature of the pedicle with 4-0 silk. The incision was closed with 3-0 chromic suture and the animals were allowed to recover from anesthesia. Rats were allowed free access to water and standard chow diet. At various time points, animals were killed and both kidneys were removed, snap-frozen in liquid nitrogen, and stored at-70°C. Assays were performed 2-6 weeks after freezing.

Blood was collected for serum creatinine determination via the tail vein. Each experimental group represents six to eight animals.

NOS assay. NOS activity in renal tissue homogenates was determined by an enzymatic method ⁽⁹⁾. Briefly, the whole renal sample was weighed and homogenates were prepared in 6 volumes of cold buffer containing 0.32 M sucrose/20 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, pH 7.2/0.5 mM EDTA/1 mM dithiothreitol, and protease inhibitors (3 μM leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 1 μM pepstatin A), using the Polytron homogenizer (Brinkmann, Lucerne, Switzerland). The “cytosolic” and“particulate” fractions were separated by centrifugation at 12,500×g for 60 min ⁽¹⁰⁾. The cytosol fraction (containing microsomal membrane) was passed through Dowex AG5OWX-8(Na⁺) resin (pH 6.0) to remove endogenous arginine and 50-μl aliquots were incubated in triplicate for 45 min at 37°C, as indicated in the presence of 2 μCi/ml resin-purified [³H] L-arginine, 100 μM L-arginine, 0.45 mM Ca⁺⁺, and 2 mM NADPH. Cold L-arginine was included to replace the endogenous L-arginine removed from the cytosols. Cold L-arginine was needed to bring the substrate concentration (100 μM) well above the K_m (10-20 μM) and make possible the determination of V_max at the plateau.

After eliminating the residual [³H] L-arginine through the resin,[³H] L-citrulline was counted in the trichloroacetic acid ether-extracted supernatant. Determinations were in triplicate. All values were corrected by the radioactivity eluted in time zero incubations and expressed per milligram of protein in the sample.

The assays were repeated in the presence of N^G-nitro-L-arginine methyl ester (L-NAME) to demonstrate the specificity of the reaction for NOS. L-NAME inhibition was at a mean of 68.5±1.2%, which was consistent between groups. Therefore, the results are compared without a correction for L-NAME sensitivity. The precision of the assay was estimated with quadruplicate determinations of the same cytosol and is approximately±5%.

Western blotting for eNOS. Measurement of eNOS by Western blot was performed as previously described ⁽¹¹⁾. Briefly, equal amounts of protein from the rat kidney were run on 7.5% polyacrylamide sodium dodecyl sulfate gels and the proteins were transferred to a nitrocellulose membrane for 16 hr at 30 volts, followed by 1 hr at 100 volts. The transfer efficiency was controlled by gel staining with Coomassie blue. Prestained protein markers (48-199 kDa) were run in each gel. The immunodetection was carried out with an affinity chromatography-purified primary antibody consisting of the mouse monoclonal antibody against a 20.4-kDa protein fragment containing amino acids 1030-1209 (carboxy terminus) of human eNOS (1/500). This antibody has significant cross-reactivity with rat eNOS ⁽¹¹⁾, whereas the commercially available murine eNOS(Upstate Biotechnology, Lake Placid, NY) has not shown rat cross-reactivity in our hands. The secondary antibody was an anti-rabbit IgG (donkey) linked to horseradish peroxidase and the incubation (1/10,000 dilution) was carried out for 1 hr. The reactive bands were detected with a luminal-based kit. Densitometry was performed with standard computer software, after scanning the blot and subtracting background. The cytosol from rat cerebellum was used as a negative control, and positive controls were the cytosol from rat penile smooth muscle cells, induced with bacterial lipopolysaccharide and interferon-γ, in a commercial human endothelial cell lysate. In addition, reactions with control IgG or without primary antibody were performed and were consistently negative.

Treatments. L-arginine (Sigma, St. Louis, MO) was dissolved in tap water at 1.25 g/L or 5 g/L and supplied to the rats starting the day before surgery. Dexamethasone (Sigma) was dissolved in sterile water and injected 3 mg/kg i.v. via the penile vein 1 hr before surgery. The lazaroid U74389G (Pharmacia-Upjohn) was dissolved in citrate buffer and injected intravenously into the penile vein, at 5 mg/kg 1 hr before clamping the renal vessels and at 3 mg/kg 10 min after releasing the vessels. For animals studied beyond 6 hr, an additional dose of 10 mg/kg dissolved in a lipid vehicle was given intraperitoneally 6 hr after surgery. Control experiments were performed using only the citrate and lipid vehicles, which had no effect on NOS activity or serum creatinine at early or late time points, with or without renal pedicle clamping. For simplicity, these negative control are not included further in the Results section.

Statistical analyses. Descriptive data on samples from each time point is reported as a mean±SEM. The data was normally distributed with small variation between mean and median values, so parametric statistics were selected. NOS activity was compared between control and ischemic kidneys from the same animals using a paired t test, and between the ischemic kidney and time zero with an unpaired t test. Statistical significance was defined as P<0.05 (two-tailed).

RESULTS

The NOS activity in both the right (control) and left (ischemic) kidneys after unilateral ischemia-reperfusion injury is presented inFigure 1A. From a baseline of 33.7±2.3 pmol/min/mg, NOS activity increased in the ischemic kidney to 79.8±10 pmol/min/mg at 2 hr (left kidney vs. baseline, P=0.007). By 24 hr, the level decreased to 57±23 pmol/min/mg (P=NS) and by day 3 was significantly lower than baseline at 15.8±5.5 pmol/min/mg(P=0.001). Levels remained significantly depressed until a return to baseline was seen at day 21. In the right (control) kidney, NOS activity increased between 6 hr and 5 days to a range of 49-56 pmol/min/mg, but was not statistically significantly different from baseline (P=0.08 at day 3). To determine whether this increase in the right kidneys was due simply to a response to decreased functional renal mass, a group of animals underwent left nephrectomy, and the right kidney was removed for analysis 1 to 5 days later. These kidneys showed no increase in NOS activity above 40 pmol/min/mg at any time point.

To determine whether the early increase in NOS activity was due to iNOS, animals were injected with 3 mg/kg i.v. of dexamethasone before surgery, which at this dose and timing is a consistent inhibitor of iNOS transcription, due to its effects on NF-κB factor binding ^{(12, 13)}. At 2 and 6 hr after reperfusion, there was no change in the increase of NOS activity seen without pretreatment, suggesting that iNOS was not involved. Therefore, to look for eNOS protein expression, Western blotting was performed on the kidneys, using an antibody to human eNOS with known cross-reactivity with rat eNOS. As seen in Figure 1B, there was an increase in eNOS detected between the ischemic and control kidney, beginning at 2 hr and persisting even at 14 days. This was in contrast to the depression of NOS activity seen at these later time points. The second band in the endothelial lysate may either be a cross-reactive product or correspond to a splicing variant of eNOS. In any case, it is absent from the kidney.

The return of NOS activity to baseline over 21 days was mirrored by changes in serum creatinine if the left kidney was subjected to ischemia-reperfusion injury with simultaneous right nephrectomy (Fig. 2). After a peak serum creatinine of 5.6±0.4 mg/dl at day 3 (P=0.005 compared with 1.1 mg/dl at day 0), it returned to 1.4 mg/dl±0.5 mg/dl by day 21. In all of these experiments, only one rat died (at day 3) before the planned time of kidney removal.

To determine whether oral L-arginine would affect renal function after ischemia-reperfusion, animals were given L-arginine in the drinking water at 1.25 or 5 g/L, starting on the day before surgery. As seen inFigure 3, at the dose of 5 g/L of L-arginine, serum creatinine was significantly lower at 7 days (2.1 vs. 4.4 mg/dl,P=0.005) and lower at 14 days, but did not achieve the defined level of significance (1.4 vs. 2.2 mg/dl, P=0.06).

Because stimulation of NOS activity could improve recovery from ischemic damage, we wished to determine whether a treatment that decreased ischemic damage independent of NO would affect NOS activity. The experimental lazaroid U74389G decreases reperfusion injury by preventing lipid peroxidation. The effect of perioperative U74389G administration on NOS activity in unilateral ischemic ATN is seen in Figure 4A. The initial increase in NOS activity was completely blocked (39.3 vs. 79.8 pmol/min/mg,P=0.01). There was no effect at day 7, but by day 14, NOS activity had returned to baseline. Treatment with U74389G hastened the recovery of renal function, as measured by serum creatinine (Fig. 4B) at day 7 (2.0 vs. 4.4 mg/dl, P=0.01) and at day 14 (1.3 vs. 2.2 mg/dl, P=0.03).

DISCUSSION

Cadaveric donor kidneys are subjected to varying degrees of warm and cold ischemia before transplantation. Renal damage may occur from prolonged lack of oxygen and energy-producing substrates, or more commonly from reactive oxygen species generated after reperfusion ⁽¹⁴⁾. This injury produces a histologic picture of ATN and can clinically manifest as DGF. Although most kidneys with DGF will go on to acceptable function after a period of dialysis support, the long-term survival of these grafts is significantly worse than in those with early function. In fact, in one study, zero antigen-matched kidneys with early function had 6-yr graft survival that was equivalent to six antigen-matched kidneys that were shared nationally⁽¹⁵⁾. The mechanism of poor graft survival in DGF is unclear, however it seems to be related to immune (rejection) and nonimmune(ischemic) factors. The response to renal injury results in the release of chemokines, cytokines, and other mediators, which also play a role in allograft damage. Up-regulation of major histocompatibility complex antigen expression may increase the immunogenicity of the organ⁽²⁾. Maneuvers to improve early function and minimize the response to injury, therefore, have the potential to improve graft survival and decrease transplant costs.

NO is a potential candidate molecule in the link between DGF and poor allograft survival due to its proven role in ischemia-reperfusion injury⁽¹⁶⁾ and allograft rejection ⁽⁷⁾. In the normal kidney, NO has a proven role in renal vasodilation, tubuloglomerular feedback, sodium excretion, and angiotensin regulation. NO may also participate in the pathogenesis of glomerulonephritis and essential hypertension ⁽¹⁷⁾. In rodent models, blockade of NOS with competitive substrate inhibitors such as L-NAME and L-NMMA has been shown to cause arteriolar vasoconstriction ⁽¹⁸⁾, decrease medullary oxygenation ⁽⁶⁾, and impair recovery of renal function in ischemia-reperfusion injury after 3 hr ⁽¹⁹⁾ and 3 days ⁽²⁰⁾. Conversely, stimulation of NO production has been shown to decrease renal vascular resistance and improve recovery of renal function after ischemic damage. Bhardwaj and Moore showed, in a perfused rat kidney model, that L-arginine dilates resistance blood vessels, probably through NO release from vascular endothelial cells ⁽²¹⁾. Chintala et al. and Schramm et al. found that intravenous infusion of L-arginine improved renal perfusion pressure and glomerular filtration rate up to 3 hr after reperfusion in rodent models of renal ischemia^{(19, 22)}. Lopez-Neblina et al. showed improvement in renal function and decrease in neutrophil infiltration after infusion with sodium nitroprusside in the ischemic rat kidney ⁽²³⁾, presumably due to its function as an NO donor. Of note, other NO donors have been shown to decrease ischemic injury in nonrenal models⁽²⁴⁾.

For this study, we chose a model of native renal ischemic-reperfusion injury to focus on the response to ischemic injury alone, rather than in combination with alloimmune injury. Indeed, features of chronic allograft nephropathy can be seen after nonimmunologic ischemic manipulation of the native kidney ⁽²⁵⁾. The selection of a unilateral ischemic model provided a contralateral control kidney, which could control for variability of NOS stimulation by cytokines released systemically, as a response to surgical stress or anesthesia. By direct measurement of NOS activity in the renal tissue, we found a significant increase in the ischemic kidney for the first 6 hr after reperfusion, as was inferred indirectly by previous studies ^{(19, 20)}. Because this increase was unaffected by systemic steroids at a dose known to prevent the induction of iNOS ⁽²⁰⁾ and because nNOS is a minor NOS isoform in the kidney, we postulate that the endothelial NOS isoform of the kidney is predominantly involved in this effect ⁽²⁶⁾. A contributory role for eNOS is also supported by the increase in eNOS protein by 2 hr, seen with Western blotting. Likely stimuli for eNOS activity are the vascular shear stress of reperfusion injury, as well as the accumulated intracellular calcium of ischemia ^{(26, 27)}.

Unexpectedly, after 24 hr, the NOS activity in the ischemic kidney fell significantly below baseline and did not recover for 21 days, which mirrored the slow return to normal renal function, as estimated by serum creatinine. Reduced NOS activity could be deleterious to recovery from ischemia due to the resultant intrarenal vasoconstriction ⁽²⁸⁾. This drop could be due to (1) cell death with fewer viable cells to produce NOS, (2) an active inhibitor of the NOS enzymes, or (3) a lack of enzyme substrate, such as L-arginine. Because the ischemic kidney had increased enzyme production when compared with the contralateral control kidney by Western blotting, the mechanism of cell death is less likely than the presence of an NOS inhibitor. In a biologically complex model, such as ischemia-reperfusion, the generation of NOS inhibitors over the course of reperfusion is quite plausible. Proven inhibitors of NOS activity that might be active in this model include interleukin-10 ⁽²⁹⁾, oxygen free radicals⁽³⁰⁾, anionic phospholipids ⁽³¹⁾, and NO itself ⁽³²⁾.

The significant effect of oral L-arginine with a positive dose response in our model suggests a role for the third mechanism of substrate deficiency. Although L-arginine is not an essential amino acid and its serum concentration is above the K_m for eNOS and nNOS, other model systems have shown the definite effects of oral L-arginine supplementation on NO-mediated pathways in rodents and humans ^{(33, 34)}. In our model, the renal supplies may be depleted after the initial increase of NOS activity, as well as general protein anabolism during renal repair. Theoretically, the L-arginine may have reduced ischemic damage through a mechanism unrelated to NO. We believe this is unlikely, given other models that show improvement of renal function with intravenous L-arginine after ischemia, but no improvement with D-arginine ⁽²¹⁾. Measurement of NOS activity without previous removal of endogenous L-arginine in kidney homogenates is probably not a good indicator of the situation in vivo. This is because of the equilibration of tissue L-arginine with the amino acid from blood contained in a highly vascular organ, such as the kidney.

In our model, the experimental lazaroid U74389G improved renal function, reduced the increased activity of NOS at 2 hr, and hastened the return of NOS activity to baseline over controls. This and other lazaroids have been shown to be effective in decreasing ischemic damage in other animal models of solid organ ischemia ⁽³⁵⁾. The presumed mechanism is prevention of iron-dependent lipid peroxidation caused by oxygen reactive species. Although U74389G did hasten a return of NOS activity to baseline by 14 days, there still was marked depression of NOS activity at 7 days compared with untreated controls, despite a significant improvement of serum creatinine in treated animals. Again, this could be due to the negative effects of oxygen free radicals on NOS expression ⁽³⁰⁾, because U74389G blocks their action on lipids, but not on NOS. This does show, however, that NOS activity alone cannot be used as an isolated indicator of recovering renal function.

In conclusion, we have shown a biphasic response of NOS activity to ischemia-reperfusion injury in the rat kidney with early stimulation, followed by a period of depressed activity, until renal function and NOS activity returned to baseline simultaneously. This occurred despite the presence of a persistently augmented content of eNOS protein. Putative substrate-induced stimulation of renal NOS activity with L-arginine, or reduction of lipid peroxidation with U74389G, improved renal function. Treatment of patients with established ATN in native or transplant kidneys is essentially supportive. Improvement of intrarenal blood flow by increased NO production using these pharmacologic manipulations has the clinical potential to hasten recovery of renal function in the face of established ischemia that would decrease renal NOS activity. Stimulation of NOS in the allograft setting would need to be accomplished selectively, so that the beneficial local effects of eNOS activity on the renal vasculature would not be offset by the detrimental cytotoxic effects of iNOS from infiltrating macrophages. Combination therapy with L-arginine and corticosteroids ⁽³⁶⁾ is one way to accomplish this selectivity, which we are currently testing in animal allograft models.

Figure 1: NOS in unilateral ischemic reperfusion injury. The left kidney was subjected to 1 hr ischemia followed by reperfusion, and both kidneys were assayed at the times specified. (A) NOS activity was measured enzymatically:*P<0.01, **P<0.001. -□-, right (control); - - -- -○- - - - -, left (ischemic). (B) Western blot using an antibody for eNOS (R, right control kidney; L, left ischemic kidney; HEL, human endothelial lysate; L/R ratio determined by densitometry with subtraction of background).

Figure 2: Effect of ischemia-reperfusion injury on serum creatinine. The left kidney was subjected to 1 hr ischemia-reperfusion with simultaneous right nephrectomy

Figure 3: Effect of oral L-arginine on serum creatinine after left renal ischemia-reperfusion and right nephrectomy. *P<0.01,**P<0.001. ▦, untreated; ▩, L-arginine at 1.25 g/L; ▨, L-arginine at 5 g/L.

Figure 4: Effect of U74389G on ischemia-reperfusion model. (A) Effect of U74389G on NOS activity in the ischemic left kidney. The horizontal line represents baseline NOS value of normal kidney. (B) Effect of U74389G on serum creatinine after left renal ischemia-reperfusion and right nephrectomy.*P<0.05, **P<0.01, ^# P=0.06. ▪, untreated; ▨, U74389G.