Clin Invest Med 1996; 19 (6): 435-43.
[résumé]
(Original manuscript submitted Jan. 19, 1996; received in revised form Apr. 16, 1996; accepted June 10, 1996)
Paper reprints may be obtained from: Dr. Mortimer Levy, Department of Physiology, Room 1228, McIntyre Medical Bldg., McGill University, 3655 Drummond St., Montreal QC H3G 1Y6; tel 514 398-4344; fax 514 398-7452; LAVECCHI@RVHMED.LAN.MCGILL.CA
Design: Animal study.
Subjects: Twenty-nine dogs.
Intervention: Induction of acute pulmonary edema with intravenously administered alloxan and administration of ANF according to five protocols.
Outcome measures: Natriuretic effect of ANF before and after the induction of pulmonary edema during the protocols.
Results: In six control animals, induction of pulmonary edema was associated with diuresis (mean 0.38, standard error of the mean [SEM] 0.003 mL/min before alloxan administration v. mean 0.75, SEM 0.11 mL/min after); natriuresis (mean 60, SEM 8 µmol/min before v. mean 103, SEM 12 µmol/min after); and a decline in blood pressure (mean 114, SEM 7 mm Hg before v. mean 93, SEM 9 mm Hg after) and in the glomerular filtration rate (mean 52, SEM 3.3 mL/min before v. mean 36, SEM 3.7 mL/min after). When isoncotic dextran solution was infused in dogs with pulmonary edema, blood pressure was maintained but the glomerular filtration rate still declined by 42% and there was natriuresis. When the renal arteries were clamped for 5 minutes during the infusion of alloxan, diuresis and natriuresis were prevented, but the glomerular filtration rate still declined, although blood pressure was maintained. ANF administered intravenously during pulmonary edema induced a further significant natriuresis in all experimental protocols. Catalase, administered intravenously as a bolus just before the alloxan infusion, prevented pulmonary edema and the associated renal changes.
Conclusions: Although alloxan appears to be directly nephrotoxic, renal damage caused by this compound does not impair the natriuretic effect of ANF in acute pulmonary edema.
Devis : Étude chez l'animal.
Sujets : Vingt-neuf chiens.
Interventions : Induction de l'oedème pulmonaire aigu par l'administraiton intra-veineuse d'alloxan; administration de PNA selon cinq protocoles.
Variables mesurées : Effets natriurétiques du PNA avant et après l'induction de l'oedème pulmonaire durant les protocoles.
Résultats : Chez six témoins, une induction de l'oedème pulmonaire a été associée avec une diurèse (moyenne 0.38, erreur standard de la moyenne [ESM] 0.003 mL/min avant l'administration d'alloxan versus moyenne 0.75, ESM 0.11 mL/min après l'administration d'alloxan; l'oedème était également associée à une natriurèse (moyenne 60, ESM 8 µmol/min avant versus moyenne 103, ESM 12 µmol/min après) ainsi qu'une diminution de la pression artérielle (moyenne 114, ESM 7 mm Hg avant versus moyenne 93, ESM 9 mm Hg après) et dans le taux de filtration glomérulaire (TFG) (moyenne 52, ESM 3.3 mL/min avant versus moyenne 36, ESM 3.7 mL/min après). L'infusion d'une solution de dextran isooncotique chez les chiens avec oedème pulmonaire s'accompagna de maintien de la pression artérielle mais une natriurèse survint de même qu'une chute de 42% dans le TFG. Le clamp des artères rénales durant 5 minutes pendant l'infusion d'alloxan s'accompagna d'une prévention de la diurèse, d'un maintien de la pression artérielle et d'une chute du TFG. L'administration intra-veineuse de PNA pendant l'oedème pulmonaire induisit une natriurèse additionnelle et significative dans tous les protocoles expérimentaux. L'administration intra-veineuse de catalase en bolus immédiatement avant l'infusion d'alloxan prévint l'oedème pulmonaire et les modifications rénales associées.
Conclusion : Bien que l'alloxan semble directement néphrotoxique, les lésions rénales associées ne nuisent pas à l'effet natriurétique du PNA dans l'oedème pulmonaire aigu.
In this study we produced a canine model of localized edema not thought to be associated with urinary sodium retention. Acute pulmonary edema induced by an intravenously administered bolus of alloxan produces exudation of fluid into the alveolar spaces because of localized capillary changes and is not associated with a rise in pulmonary venous pressure.[6,7] This experimental model is a reasonable equivalent of adult respiratory distress syndrome, which involves pulmonary edema in the absence of central hemodynamic abnormalities.[68] Although there is considerable literature on the natriuretic effects of ANF in acute renal failure and generalized edema,[1,2] there is little on the efficacy of ANF in producing natriuresis in localized edema, in which urinary sodium retention is usually absent. Because this peptide is being increasingly considered for clinical use as an aid to diuretic therapy, we wished to determine the natriuretic effects of ANF in this canine equivalent of a commonly observed clinical state of localized edema. We also wished to determine whether this model would exhibit the profile of natriuretic response seen in other canine models of edema or urinary sodium retention: a 50:50 ratio between responders and nonresponders.
[Table of contents]
Methods
The studies were carried out in a total of 29 mongrel dogs of both sexes (but mainly female) weighing a mean of 15.2 kg (SEM 0.86 kg). All dogs were considered to be in good health and had been in the McGill Animal Centre for 2 to 3 days before the study. All protocols were approved by the University Committee on Animal Care, and the guidelines of the Canadian Council of Animal Care were followed throughout all experimental procedures.
Dogs were not given any food during the night before the experiment but were permitted free access to water. To that point, they had been fed a standard dog chow providing approximately 45 mmol of sodium per day. The dogs were anesthetized with an intravenous administration of sodium pentobarbital (25 mg/kg), with small supplemental doses administered as required throughout the study period. The animals were given mechanical ventilation with the use of a cuffed endotracheal tube and a Harvard Dog Mechanical Respirator (Harvard Apparatus, South Natick, Mass.). Polyethylene catheters (PE 205) were inserted into the femoral veins for infusion and PE 240 catheters into the femoral arteries for blood collection and recording of arterial blood pressure (ABP) by mercury manometry. Urine was collected directly through ureteral catheters (PE 190 or 205) placed through a low abdominal midline incision. A PE 50 catheter was inserted into an antecubital or jugular vein for delivery of inulin and para-amino hippurate (PAH) in water at 0.5 mL per minute by the constant-infusion technique, with the use of a Harvard peristaltic pump. This allowed us to measure the glomerular filtration rate (GFR) and effective renal plasma flow (RPF). The use of this technique in our laboratory has been previously described in detail.9 There were three clearance periods of at least 10 minutes during this process. Blood samples were collected at the midpoint of each clearance period and replaced with an equivalent volume of isotonic saline solution. Central venous pressure (CVP) was measured with a PE 190 catheter inserted into the right atrium through the right internal jugular vein and connected to a saline manometer. The fixed baseline was kept at the level of the anterior axilla, with the dog lying flat in the supine position. After each surgical procedure, the dogs were given 200 mL of isotonic saline solution to replace all losses and were given 30 minutes for equilibration before the clearance periods. Experiments were performed according to several different protocols.
(1) The first protocol, used in six dogs, examined the effects of alloxan. After samples were collected at baseline, alloxan (60 mg/kg in 20 mL of warm saline solution) was infused intravenously as a bolus for 5 to 10 minutes. After a 30-minute waiting period, samples were collected during the clearance period, for as long as 120 minutes after the bolus.
(2) In six dogs, immediately after the collection of clearance samples after the alloxan infusion, ANF (75 ng/kg per minute) was delivered intravenously at a constant rate. After a 10-minute waiting period, three samples were taken during the clearance periods. The ANF was then discontinued and, after a 30-minute recovery period, another set of clearance measurements were made.
(3) Because alloxan is thought to cause injury to the renal tubule8 we also studied a series of five dogs in which ANF was administered immediately after baseline measurements. After a 90-minute recovery period, alloxan was administered as in the first protocol, but during the 5-minute infusion the renal arteries were clamped with arterial bulldog clamps. After release and a 30-minute waiting period, three samples were collected, ANF was readministered, and a final set of three samples were taken.
(4) In four dogs, we studied the effect of volume infusion by administering intravenously 150 mL of 4% (isoncotic) dextran solution of mean molecular weight 70 KD with the alloxan. After a set of samples were taken during clearance periods, ANF was administered as in the second protocol and more clearance-period samples were taken. After a 30-minute recovery period, a final set of two clearance-period samples were taken.
(5) In eight dogs, catalase was administered before the alloxan injection, to provide an added control series. Catalase, which is reported to prevent the lung injury caused by alloxan,10 was administered as a bolus of 150 000 U/kg from bovine liver (Sigma Chemical Co., St. Louis, Mo.) in 10 mL of saline solution, injected directly into the catheter inserted in the right atrium. After the collection of samples during the clearance periods, ANF was administered as in the second protocol.
Upon completion of the studies, the dogs' chests were opened to inspect the condition of the lungs. In some dogs, the chest wall was opened before the protocol was followed to confirm that lung tissue was indeed normal. This manoeuvre did not influence the subsequent experimental data; therefore, all animals are considered together.
Inulin, PAH, sodium-ion concentrations, plasma protein concentrations and hematocrit were all analysed by techniques previously described for this laboratory.[9] Blood pH and partial pressure of carbon dioxide (pCO2) were measured with an IL 1302 pH/blood gas analyser from 0.5-mL samples collected anaerobically and kept on ice. Plasma levels of ANF were analysed by techniques previously described for this laboratory.[3,5] Statistical analysis was carried out with a paired Student's t-test or analysis of variance, as appropriate, and a p value of 0.05 was considered significant.
[Table of contents]
Results
Upon visual inspection after completion of the experiment, the lungs of most of the dogs appeared somewhat cyanotic, in a patchy fashion, with identifiable pockets of edema. In a few dogs some areas of the lungs appeared to be unventilated, and in a very few the lungs appeared normal. Since all dogs behaved the same way in regard to the alloxan infusion, the data from all of the dogs are considered together. In the dogs studied in the first protocol, the blood pH fell from 7.39 (SEM 0.02) to 7.27 (SEM 0.03, p < 0.05), whereas the pCO2 rose from 40.2 (SEM 2.28) mm Hg to 47.3 (SEM 3.2) mm Hg (p < 0.05).
Table 1 summarizes the data from the six dogs in the first protocol during the hour after the infusion of alloxan. Although there was a significant drop in the ABP, the CVP remained constant, and there was an increase in both urine flow and sodium excretion. Since the GFR declined significantly, by 30%, there was obviously a marked decline in fractional reabsorption of sodium ions after the alloxan infusion. In addition to the decline in the GFR, the RPF declined, so that the mean filtration fraction remained unchanged (0.35 before v. 0.34 after alloxin infusion). These changes in renal function persisted for at least 2 hours after the alloxan infusion.
The decline in the ABP and the rise in the hematocrit suggest that there had been some volume contraction due to fluid accumulating in the lung. The concomitant decline in plasma protein concentration suggests that this fluid was protein rich. The hematocrit rose by 8%, and the decline in plasma protein levels was 7.8%.
Table 2 summarizes the effect of an ANF infusion in acute pulmonary edema. Despite a natriuretic and diuretic state after the alloxan infusion, the intravenous administration of ANF further increased the urinary excretion of salt and water. As for the dogs described in Table 1, the GFR declined after the alloxan infusion and did not rise in response to the ANF infusion. The ABP declined by 10%, and the hematocrit rose in response to the alloxan infusion. All of the dogs responded to the ANF infusion with a significant natriuresis that exceeded a change in urinary sodium excretion of 30 µmol per minute, the level used in other studies in our laboratory as an arbitrary minimum natriuretic response to this peptide.[3]
Since alloxan may cause tubulointerstitial injury,[11] we conducted a series of experiments in which the renal arteries were clamped as the alloxan bolus was being infused. This manoeuvre was meant to protect the kidneys against possible acute toxic effects of the infusion. The clamping procedure attenuated the diuretic and natriuretic effect of the alloxan and the effect of ANF, compared with an ANF infusion given before the alloxan was administered (Table 3).
To prevent the volume contraction associated with pulmonary edema we infused the dogs with isoncotic dextran solution to counterbalance the protein and volume losses in the alveoli. This protocol augmented the natriuretic effect of the alloxan, possibly in part by reducing the filtration fraction. Although the GFR declined, the PAH clearance (a measure of renal plasma flow) remained intact, presumably because of the extra volume supplied. The ABP also remained intact (presumably for the same reasons), whereas the hematocrit declined, again because of the dilutional effects of the dextran solution infusion. This infusion was not associated with any rise in plasma immunoreactive ANF (iANF) levels beyond that usually observed after alloxan infusion (109 [SEM 16] pg/mL before dextran infusion v. 166 [SEM 27] pg/mL after infusion, p < 0.05). There are two possible reasons for this observation. First, the dextran solution may have preferentially extravasated into the lungs. Second, as we have previously demonstrated,[12] colloid infusions are unlikely to raise plasma levels of iANF unless the CVP and blood pressure become elevated. These two variables remained unchanged in the pulmonary edema and dextran phase (Table 4).
Finally, we attempted to completely attenuate the alloxan effect by pretreating the dogs with catalase. Table 5 indicates that, although the natriuretic effect was prevented, there was still a small increase in urine flow for reasons that are unclear. The lungs of all dogs appeared normal, and the usual hypotensive response was not observed. ANF induced the usual natriuretic and diuretic effects, as well as a small increase in the GFR. The pCO2 after the alloxan infusion remained normal (39.5 [SEM 2.1] mm Hg before the infusion v. 39.9 [SEM 2.9] mm Hg after).
[Table of contents]
Discussion
Several investigators have given dogs alloxan (mesoxalylurea), in boluses of 60 to 100 mg/kg administered intravenously, to produce acute pulmonary edema.[68,10] Alloxan attacks the endothelial barrier directly and is thought to alter normal permeability by producing hydrogen peroxide and hydroxyl radicals, which then serve as major mediators for pulmonary edema.[10] Pulmonary edema is of a normal pressure type, and a transient rise in pulmonary pressure, at best, has been reported.[7] Therefore, this model seems a reasonable equivalent of localized edema and the so-called "shock-lung" or adult respiratory distress syndrome.
Although we did not systematically examine all vascular territories after the infusion of the alloxan, we are confident that the permeability disruptions were confined to the lungs. Examination of the urine with a commercial urinary dipstick at the end of the experiment did not reveal proteinuria, and careful inspection of the legs and the abdomen, through incision down to the muscle layer, did not reveal any edematous tissue. In this model, widespread edema would be unlikely, since it is thought that the alloxan is very quickly taken up by the pancreas, renal cortex and lungs.[13] Moreover, the size of the decrease in plasma volume also suggested that the extravasation of fluid from the vascular compartment was localized. The rise in the hematocrit after the alloxan infusion was 8%, suggesting a fall in plasma volume of this degree. In our laboratory, measurements of plasma volume in large numbers of dogs with the use of the Evans Blue T1824 dye dilution technique[14] have shown that plasma accounts for 4.5% of body weight. Since the mean weight of the dogs used in this study was 15.2 kg, plasma volume would have averaged 690 mL. An 8% decrease implies that approximately 55 mL of fluid had left the vascular space. This value is physiologically reasonable, given the appearance of the lung tissue.
By 30 minutes after the infusion of the alloxan bolus, the dogs' lungs were clearly edematous and there was evidence of some carbon dioxide retention. To our surprise, during the initial collections of samples during clearance periods after the alloxan infusion, the kidneys exhibited a significant diuresis and natriuresis, although the ABP and GFR had declined. We did not conduct a biopsy of the kidney tissue, but, since there is a great deal of presumptive evidence that the plasma volume declined, it seems reasonable to believe that the diuresis and natriuresis reflected a toxic effect on the tubules. The resulting natriuresis would be similar to that observed after the infusion of bile salts, which are initially natriuretic although ultimately very toxic to the renal tubule.[11] In addition to a toxic effect, the natriuresis after alloxan infusion is likely explained, in part, by hypercapnia.[15] Hypercapnia was seen in all of the experimental groups of dogs, with an increase in pCO2 ranging from 6.5 to 8.0 mm Hg in the dogs that underwent protocols 2, 3 and 4 (p < 0.05).
The rise in plasma levels of iANF observed in the dogs (Table 1) after the infusion of alloxan raises the possibility that the observed natriuresis was due to an increase in the endogenous levels of this peptide. We do not believe, however, that this is the correct explanation. First, in the renal arterial clamping studies (Table 3), no natriuresis was observed during the pulmonary edema phase, although plasma levels of iANF rose by the same degree (mean 82.6 [SEM 11.9] pg/mL during recovery phase v. mean 181 [SEM 21.6] pg/mL after alloxan administration, p < 0.05), as shown in Table 1. Second, in earlier dose-response studies carried out in normal dogs with ANF,16 the level of plasma iANF was associated with an infusion rate for this peptide of less than 25 ng/kg per minute -- a rate not associated with a natriuretic effect. Third, an infusion of isotonic saline solution equivalent to that used in our studies given to three control dogs did not elevate their plasma iANF levels (82.6 pg/mL before infusion v. 98.7 pg/mL after infusion, not significant). Although catalase prevented the rise in plasma iANF levels usually observed with alloxan infusion (80.6 pg/mL at baseline v. 89.3 pg/mL after administration of catalase and alloxan), as well as the natriuresis, it also prevented pulmonary edema and any consequences that could account for the elevation of plasma iANF levels, e.g., small increases in right atrial stretch.
That alloxan could be exerting a direct, natriuretic effect is supported by two lines of evidence. First, we infused catalase into the dogs before the bolus administration of alloxan. Catalase has been reported to prevent the toxic effects of alloxan by acting as an enzymatic scavenger for hydrogen peroxide.[10] In our study, this enzyme prevented both pulmonary edema and most of the renal effects as well as the decline in the ABP and plasma protein concentration and the rise in the hematocrit. Second, when the renal arteries were clamped during the bolus infusion of alloxan, the usual effects of this compound on renal function were not observed (except for a slight decline in the GFR), although pulmonary edema occurred as usual. Wigness and associates8 have documented that obstructing the renal arteries during the intravenous infusion of alloxan prevents chronic renal failure, which develops in these animals when alloxan is employed as a diabetogenic agent. Since beta-cell lesions within the pancreas occur within 5 minutes of an infusion, and afterward alloxan is no longer found in the blood,[8] we were clearly able to prevent adequate amounts of alloxan from reaching the renal parenchyma. The slight fall in the GFR may have been caused by the clamping procedure. In these dogs, ANF was effective, but the effect was attenuated in comparison with the ANF-induced natriuresis before alloxan infusion. This attenuated effect was probably due to the clamping procedure.
There is no evidence that the pulmonary edema produced in these studies was associated with a rise in the CVP or pulmonary arterial pressure. The decline in the ABP after alloxan infusion was inconsistent and did not appear in every group of dogs.
Despite the acute pulmonary edema and the modest volume contraction, an infusion of ANF still produced a significant natriuretic effect. This response did not depend on a "sensitization" of the nephron by toxic effects of the alloxan, as can be deduced from the experiments in which the renal arteries were clamped during the alloxan bolus and the kidneys thus spared the toxic changes.
A rather surprising result was that, after alloxan administration, plasma iANF levels rose, despite apparent plasma volume contraction. It is uncertain why this occurred. It may have been a direct stimulatory effect of the alloxan infusion or a transient rise in right atrial pressure not recorded by our CVP manometer. Alternatively, given the decrease in the GFR usually observed in response to this agent as well as the toxic effect on the renal tubules, the rise in plasma iANF levels could have resulted from a combination of decreased renal clearance and decreased degradation by the endopeptidase enzymes of the proximal tubule.
[Table of contents]
Conclusion
The intravenous infusion of alloxan produces acute pulmonary edema in the absence of heart failure and has a "toxic" effect on the kidneys associated with a natriuresis. Whether this alloxan-induced natriuresis is present or not, ANF exerts a further significant natriuretic effect which does not depend on prior renal tubular retention of sodium. Heterogeneity of natriuretic response to ANF was not observed in the dogs treated with alloxan; all were natriuretic responders. This observation suggests that ANF may be useful as an aid to therapy in noncardiogenic pulmonary edema. Our observations extend and confirm those previously recorded by our laboratory for dogs with acute pancreatitis;[17] that is, in an acute tubular injury, ANF may still be significantly natriuretic, presumably because of its extreme distal action within the nephron. This persistence of a potent distal natriuretic effect in various conditions associated with acute renal failure,[5,17] as opposed to intense sodium-retaining states (generalized edema), suggests that ANF can induce salt wasting in localized edema, although edema may coexist with tubular injury.
[Table of contents]
Acknowledgements
We thank Alison Taylor, MSc, Emma Resurreccion and Luigi Franchi for careful and expert technical assistance, and Christine Pamplin for expert secretarial assistance. This research was generously funded by the Medical Research Council of Canada.
[Table of contents]
References