Antibiotic-related nephrotoxicity.

The toxicity of aminoglycosides is related to their concentrative uptake by proximal tubular cells and their capacity to interact with critical intracellular targets. Concentrative uptake is mediated by adsorptive endocytosis across the apical membrane followed by sequestration within lysosomes. The fundamental mechanism underlying the toxicity of these organic polycations is their capacity to interact electrostatically with and disrupt the metabolism of anionic phospholipids, especially the phosphoinositides. Polyaspartic acid, a polyanionic peptide, protects against aminoglycoside nephrotoxicity by forming electrostatic complexes with these drugs and inhibiting their interaction with critical intracellular targets. The selective toxicity of beta-lactams towards renal proximal tubular cells is related to their concentrative uptake via the organic anion transport system. Lipid peroxidation appears to play a major role in the toxicity of cephaloridine. Depressed mitochondrial respiration secondary to acylation of the mitochondrial transporter for succinate has been implicated in the pathogenesis of toxicity caused by other cephalosporins and carbapenems. The predilection of the kidney for amphotericin B toxicity is unclear as little drug is excreted by the kidneys. Toxicity is manifested by increased renal vascular resistance, depression of RBF and GFR, and altered tubular function that reflects the capacity of this drug to interact with cholesterol-containing membranes and increase membrane permeability to ions including potassium, hydrogen, calcium, and magnesium.

Mechanism of decreased arachidonic acid in the renal cortex of rats with diabetes mellitus.

The purpose of this study was to investigate the roles of decreased synthesis and increased consumption in the depression of arachidonic acid levels in renal cortex and glomeruli of rats with streptozotocin-induced diabetes mellitus. In diabetic rats, arachidonic acid was depressed 33.2% in renal cortex, 47.4% in liver and 66.1% in heart compared to values of control rats. delta 6 Desaturase activity was depressed in renal cortex, liver and heart of diabetic rats to 53.3, 55.5 and 63.7%, respectively, of control values. delta 5 Desaturase activity was also depressed 43.7, 55.5 and 47.6% in renal cortex, liver and heart of diabetic rats, respectively. In other rats the activities of five enzymes involved in the synthesis and esterification of arachidonic acid were measured in renal cortex and in isolated glomeruli. Both tissues from diabetic rats showed depressed activities of delta 5 and delta 6 desaturases, increased activities of long-chain acyl-CoA synthetase and 1-acyl-sn-glycero-3-phosphocholine acyltransferase and no change in the activity of elongase as compared to those in control tissues. Malondialdehyde, an end product of lipid peroxidation, was lower in the renal cortex of diabetic rats than in control rats, whereas beta-oxidation of linoleic acid and arachidonic acid were similar in diabetic and in control rats. Basal and stimulated prostaglandin E2 synthesis were significantly higher in isolated glomeruli from diabetic rats compared to those in control rats. In isolated tubules, prostaglandin E2 synthesis was similarly low in both groups.(ABSTRACT TRUNCATED AT 250 WORDS)

Biophysical and biochemical alterations of renal cortical membranes in diabetic rat.

The objective of this study was to determine whether streptozotocin-induced diabetes mellitus in the rat causes alterations in the lipid composition and fluidity of renal brush border membranes (BBM) and basolateral membranes (BLM). Compared to membranes of non-diabetic rats, BBM and BLM of diabetic rats contained 31% and 26%, respectively, less arachidonic acid and 36% and 46%, respectively, more linoleic acid esterfied in phospholipids. These changes were accompanied by a decrease in the average number of double bonds per mole of fatty acid, a measure of fatty acid unsaturation. In diabetic rats BLM had a higher total phospholipid/protein ratio (567 +/- 20 vs. 482 +/- 15 nmol/mg protein, P < 0.01), less cholesterol (369 +/- 30 vs. 512 +/- 34 nmol/mg protein, P < 0.01), more phosphatidylcholine (+72%) and less sphingomyelin (-22%) than did BBM. These differences were identical to those observed between BLM and BBM of non-diabetic rats. In control rats BLM was more fluid than BBM as assessed by the steady state fluorescence anisotrophy of diphenylhexatriene and by glycerol permeability. In diabetic rats the fluidity of BLM was not different from that of BBM as assessed by the steady state fluorescence anisotrophy of diphenylhexatriene whereas BLM was slightly more fluid than BBM as assessed by glycerol permeability. By both measures BLM and BBM from diabetic rats were significantly less fluid than BLM and BBM from control rats. Removal of proteins and cholesterol in sequence was accompanied by an increase in membrane fluidity in both groups. However, in no instance did the removal of proteins or cholesterol abolish the difference between the fluidity of diabetic membranes and that of control membranes. From these data we conclude that the reduction in fluidity of renal BLM and BBM in the diabetic rat is due to the change in the composition of fatty acids esterified in membrane phospholipids.

Drug-phospholipid interactions: role in aminoglycoside nephrotoxicity.

Aminoglycoside antibiotics are known to be transported and accumulated within lysosomes of renal proximal tubular cells and to cause proximal tubular cell injury and necrosis. The pathogenesis of aminoglycoside nephrotoxicity is postulated to be related to the capacity of these organic polycations to interact electrostatically with membrane anionic phospholipids and to disrupt membrane structure and function. Aminoglycoside antibiotics have been shown to bind to anionic phospholipids of model membranes and to alter membrane permeability and promote membrane aggregation. In vivo these drugs induce phospholipiduria and a renal cortical phospholipidosis. The latter reflects the accumulation of phospholipid-containing myeloid bodies within the lysosomal compartment consequent to aminoglycoside-induced inhibition of lysosomal phospholipases. The mechanism of drug-induced inhibition of phospholipases has been shown to be secondary to the binding of these cationic drugs to anionic phospholipids. As the lysosomes became progressively distended with myeloid bodies, they become unstable and eventually rupture, which results in the release of acid hydrolases as well as high concentrations of aminoglycosides into the cytoplasm where they interact with and disrupt the function of other membranes and organelles including mitochondria and microsomes. It is postulated that the redistribution of drug from the lysosomal compartment to organellar membranes is the critical event which triggers the irreversible injury cascade. Polyaspartic acid is a polyanionic peptide which when administered in vitro or in vivo forms electrostatic complexes with aminoglycoside antibiotics and prevents these drugs from interacting with anionic phospholipids, from perturbing phospholipid metabolism and from causing cell injury and necrosis.

Polyaspartic acid inhibits gentamicin-induced perturbations of phospholipid metabolism.

We investigated whether polyaspartic acid (PAA) can inhibit aminoglycoside-induced perturbations of phospholipid metabolism in cultured renal cells of opossum and rabbit and examined the mechanism involved. Cells incubated in medium containing gentamicin (10(-3) M) manifested a time-dependent increase in total phospholipid in association with the appearance of lysosomal myeloid bodies, impaired degradation of phospholipid, and disruption of the phosphatidylinositol (PI) cascade in response to bradykinin stimulation. These alterations of phospholipid metabolism were either completely or almost completely prevented in cells grown in medium containing gentamicin (10(-3) M) and PAA (3 x 10(-4) M, mol wt 11,000) even though PAA did not inhibit the cellular accumulation of gentamicin (40 +/- 1 vs. 42 +/- 1 micrograms/mg protein). In other in vitro studies, we demonstrated that gentamicin depressed the permeability of phosphatidylcholine (PC)/PI liposomes to glycerol and promoted liposomal aggregation. Both effects were blocked by prior addition of PAA. Methylene blue, a cationic dye, was shown to form an electrostatic complex with PAA; gentamicin competitively displaced methylene blue bound to PAA. Our results support the conclusion that the protective effect of PAA is related to its ability to serve as an anionic substrate that electrostatically binds aminoglycoside antibiotics and, thereby, prevents these polycationic drugs from interacting electrostatically with anionic phospholipid of cell membranes.

Polyaspartic acid protects against gentamicin nephrotoxicity in the rat.

Polyamino acids including polyaspartic acid (PAA) have been reported to provide protection against the development of aminoglycoside-induced nephrotoxicity in the rat as assessed by histopathology scoring. We sought to confirm and extend these observations by determining whether PAA also prevented functional and biochemical lesions of gentamicin-nephrotoxicity in an animal model studied extensively in our laboratory. Rats were given injections of: 1) 0.9% NaCl at 2.5 ml/kg b.wt. per day; 2) PAA (mol.wt. 15,000) at 500 mg/kg per day; 3) gentamicin at 100 mg/kg per day or 4) gentamicin at 100 mg/kg per day and PAA at 500 mg/kg per day for 6 days. Rats injected with gentamicin exhibited: 1) increased urinary excretion of the brush border membrane enzyme alanine aminopeptidase and the lysosomal enzyme N-acetyl-beta-d-glucosaminidase after the first injection; 2) increased total phospholipid and malondialdehyde but decreased catalase activity in the renal cortex; 3) elevation of serum creatinine and depression of creatinine clearance and 4) extensive proximal tubular cell necrosis all determined 24 hr after the last injection of gentamicin. Rats injected with gentamicin plus PAA also exhibited increased urinary excretion of alanine aminopeptidase not different in magnitude from that of rats injected with gentamicin alone, whereas N-acetyl-beta-d-glucosaminidase rose more slowly and returned to base line by day 4. Total renal cortical phospholipid was elevated to the same extent in the two groups. Malondialdehyde was not different from control and catalase activity was significantly less depressed in rats injected with gentamicin plus PAA.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of gentamicin on phospholipid metabolism in cultured rabbit proximal tubular cells.

We examined the hypothesis that the accumulation of phospholipid in cells exposed to gentamicin is due to impaired degradation. Experiments were performed in rabbit proximal tubular cells grown in primary culture. Cells exposed to 10(-3) M gentamicin manifested myeloid body formation and a progressive increase in total phospholipid that by day 6 was 44% higher than that of control cells and reflected increases of phosphatidylinositol of 235%, phosphatidylcholine of 60%, phosphatidylethanolamine of 90%, and phosphatidylserine of 55% above control values. Gentamicin impaired the degradation of these phospholipids. The t1/2 of the phospholipid pool labeled with [3H]myoinositol increased 146% from 1.17 (control) to 2.88 days (gentamicin); the t1/2 of the [3H]choline pool increased 34% from 1.77 to 2.38 days; the t1/2 of the [3H]ethanolamine pool increased 57% from 3.14 to 4.93 days; the t1/2 of the [3H] serine pool increased 37% from 6.30 to 8.63 days. Exposure of cells to gentamicin for 2 days also stimulated increased incorporation of [3H]myoinositol (68%) and [3H]ethanolamine (59%) into phospholipid. The data are consistent with the hypothesis that gentamicin inhibits the activity of lysosomal phospholipases that results in the accumulation of phospholipid within the lysosome in the form of myeloid bodies. Increased phospholipid synthesis may represent a compensatory response to the impaired lysosomal degradation of phospholipid. We postulate that the preferential increase of phosphatidylinositol reflects the capacity of the polycationic gentamicin to interact electrostatically with the anionic phosphoinositides and inhibit their turnover.

Gentamicin inhibits agonist stimulation of the phosphatidylinositol cascade in primary cultures of rabbit proximal tubular cells and in rat renal cortex.

A growing body of evidence indicates that aminoglycoside antibiotics interact with phosphoinositides and this has led to the hypothesis that these drugs perturb the phosphatidylinositol (PI) cascade. To test this hypothesis we examined the effect of gentamicin on agonist stimulation of the PI cascade in primary culture of rabbit proximal tubular cells (RPTC) and in rat renal cortex. Parathyroid (PTH) (10(-6) M) stimulated a significant increase in total inositol phosphates, inositol monophosphate and inositol trisphosphate, but not inositol bisphosphate in RPTC with the peak effect at 2 min. This effect was completely inhibited in RPTC exposed to 10(-3) M gentamicin for 48 and 24 hr. In other experiments we demonstrated that angiotensin II, phenylephrine, bradykinin and arginine vasopressin (all at 10(-6) M) stimulated inositol trisphosphate generation in control RPTC but not in cells exposed to 10(-3) M gentamicin for 24 h. In contrast gentamicin did not block PTH-stimulation of cyclic AMP generation, which indicates that gentamicin did not prevent PTH from interacting with its plasma membrane receptor. PTH also stimulated redistribution of protein kinase C from the cytosolic to the membrane fraction of RPTC. This effect was completely abolished in RPTC exposed to 10(-3) M gentamicin for 2 days. PTH given i.p. to rats stimulated the redistribution of protein kinase C from the cytosolic to the membrane fraction of renal cortex. This effect was completely inhibited in rats injected with gentamicin, 100 mg/kg per day for 2 days. The

The effect of gentamicin on the biophysical properties of phosphatidic acid liposomes is influenced by the O-C = O group of the lipid.

We previously reported that gentamicin binds to liposomes composed of anionic phospholipids and depresses glycerol permeability and raises the activation energy for glycerol permeation in these liposomes. We postulated that these changes in the glycerol permeability and in the activation energy (Ea) for glycerol permeation were due to hydrogen bonding between O-C = O groups in the hydrogen belt and one or more amino groups of gentamicin. To test this hypothesis, we examined the effects of gentamicin on the membrane surface potential, the glycerol permeability coefficient (p), the Ea for glycerol permeation, and the aggregation of liposomes composed of 1:1 phosphatidylcholine (PC) and phosphatidic acid with the acyl chains of phosphatidic acid in either an ester (PA) or an ether (PA*) linkage. Gentamicin depressed the membrane surface electrostatic potential, measured by the partitioning of methylene blue between the bulk solution and the liposomal membrane, to an equivalent degree in PC-PA and PC-PA* liposomes, which indicates that substitution of the ether for the ester linkage did not interfere with the electrostatic interaction between the cationic drug and the negatively charged phosphate head group. Gentamicin caused a temperature-dependent decrease of p and raised Ea for glycerol permeation from 17.7 +/- 0.3 to 21.6 +/- 0.4 kcal/mol in PC-PA liposomes but had little or no effect on these parameters in PC-PA* liposomes. In contrast, gentamicin induced a significantly greater degree of aggregation of PC-PA* liposomes compared to that of PC-PA liposomes.(ABSTRACT TRUNCATED AT 250 WORDS)

Induction of nephrotoxicity by high doses of gentamicin in diabetic rats.

Rats with streptozotocin-induced diabetes mellitus (DM) are resistant to aminoglycoside (AG) nephrotoxicity presumably because of defective transport and accumulation of drug by proximal tubular cells. To test this hypothesis we injected DM rats with saline or with gentamicin, 100, 200, and 400 mg/kg per day for 6 days, to determine if the renal cortical concentration of gentamicin could be raised to toxic levels. Nephrotoxicity was assessed by monitoring for evidence of accelerated lipid peroxidation in the renal cortex, for elevation of the serum creatinine concentration, and for evidence of proximal tubular cell injury and necrosis by light and electron microscopy. At 100 mg/kg per day renal cortical gentamicin was 454 +/- 85 micrograms/g. Except for an increase in renal cortical phospholipids these rats manifested no evidence of accelerated lipid peroxidation or elevation of serum creatinine. At 200 mg/kg per day renal cortical gentamicin rose to 636 +/- 20 micrograms/g. These rats manifested mild functional and morphological evidence of toxicity. At 400 mg/kg renal cortical gentamicin rose to 741 +/- 43 micrograms/g. These rats developed severe nephrotoxic injury as manifested by a marked increase of lipid peroxidation evident by an increase of malondialdehyde from a control level of 0.48 +/- 0.02 to 1.72 +/- 0.12 nmole/mg protein, a shift from unsaturated to saturated fatty acids esterified in renal cortical phospholipids, depression of superoxide dismutase and catalase, and a shift from reduced to oxidized glutathione. The serum creatinine rose from a baseline level of 0.24 +/- 0.01 to 0.46 +/- 0.05 mg/dl. Light and electron microscopy revealed enlarged lysosomes distended with typical myeloid bodies and extensive proximal tubular cell necrosis. These observations provide compelling evidence in support of the view that the resistance of DM rats to AG nephrotoxicity is causally linked to the low rate of drug uptake by renal proximal tubular cells. When the renal cortical concentration reaches a critical level, it elicits a pattern of toxic injury indistinguishable from that of nondiabetic rats. Thus, there is nothing inherent to the diabetic state that prevents AGs from causing their usual adverse effects on the metabolism of renal proximal tubular cells once they gain access in sufficient quantity into these cells.

Failure of inhibition of lipid peroxidation by vitamin E to protect against gentamicin nephrotoxicity in the rat.

We tested the hypothesis that accelerated lipid peroxidation, possibly at the level of the lysosome, is linked causally to the pathogenesis of aminoglycoside nephrotoxicity by investigating whether administration of vitamin E would inhibit lipid peroxidation and prevent or ameliorate gentamicin-induced proximal tubular cell injury. Five groups of rats were injected with either saline, vitamin E (600 mg/kg per day) for 6 days, gentamicin (100 mg/kg per day) for 6 days, vitamin E for 6 days plus gentamicin for 6 days or vitamin E for 12 days and gentamicin for the last 6 days. Gentamicin alone induced a 16% increase in renal cortical phospholipids; vitamin E had no significant effect on this change. Gentamicin alone caused accelerated lipid peroxidation evident by a doubling of renal cortical malondialdehyde to 1.23 nmol/mg protein, and a sharp decline of esterified polyunsaturated fatty acids, especially arachidonic acid which fell 43%. These changes were accompanied by depressions of superoxide dismutase, catalase, and total glutathione and a shift from reduced to oxidized glutathione. Concurrent treatment of rats with vitamin E plus gentamicin for 6 days had no significant effect on the gentamicin-induced alterations of malondialdehyde, superoxide dismutase, catalase or the glutathione cascade; however, the shift from polyunsaturated to saturated fatty acids was largely reversed. In rats pretreated with vitamin E for 6 days, gentamicin failed to raise renal cortical malondialdehyde above that of saline-treated rats. The changes in esterified fatty acids were prevented almost entirely, and there were no significant alterations from control of the glutathione cascade. The depressions of superoxide dismutase and of catalase, however, were not reversed. Vitamin E did not affect the amount of gentamicin accumulated in renal cortex nor did it prevent the gentamicin-induced rise of serum creatinine. Examination of renal cortex by light and electron microscopy revealed that vitamin E did not prevent or even reduce the severity of gentamicin-induced proximal tubular cell lesions and necrosis. These results confirm those we obtained in a previous study with the antioxidant diphenyl-phenylenediamine. The observation that inhibition of lipid peroxidation by two distinct antioxidants failed to prevent proximal tubular cell injury and renal dysfunction associated with gentamicin administration leads us to conclude that lipid peroxidation is a consequence and not a cause of gentamicin-induced nephrotoxicity.

Renal handling of netilmicin in the rat with streptozotocin-induced diabetes mellitus.

We examined the hypothesis that the reduced accumulation of aminoglycoside in renal cortex of rats with streptozotocin-induced diabetes mellitus (DM) is secondary to lower rates of tubular transport of drug compared with non-DM rats. Using whole kidney clearance techniques we found that the fractional excretion of [3H]netilmicin in DM rats rose from an initial value of 92.4 +/- 1.3% to 101 +/- 4.5% (N = 10) after eight 20-min periods. These values were not significantly different from those of non-DM rats (96.4 +/- 1.8 and 106.7 +/- 1.4%, respectively, N = 7). The plasma concentration of drug was similar in the two groups whereas inulin clearance and the filtered load of drug were higher in DM rats. In both groups microinjection experiments revealed the presence of an absorptive flux of [3H]netilmicin along the proximal tubule and loop of Henle, but no absorptive flux was detected along the distal nephron. In free-flow micropuncture experiments in DM rats a net secretory flux of netilmicin was detected in the early proximal tubule and a net absorptive flux was detected along the loop of Henle, presumably the pars recta. No net flux occurred along the distal tubule. These findings are similar to those previously reported by us for non-DM rats. At the end of the clearance experiments the concentration of netilmicin in renal cortex of DM rats (56 +/- 5 micrograms/g wet wt.) was significantly less (P less than .01) than that in the renal cortex of non-DM rats (122 +/- 6 micrograms/g wet wt).(ABSTRACT TRUNCATED AT 250 WORDS)

Effects of diphenyl-phenylenediamine on gentamicin-induced lipid peroxidation and toxicity in rat renal cortex.

The hypothesis that lipid peroxidation is linked causally to the pathogenesis of aminoglycoside nephrotoxicity was tested by determining whether administration of the antioxidant, diphenyl-phenylenediamine (DPPD) would inhibit lipid peroxidation and ameliorate gentamicin-induced proximal tubular cell injury. Rats were injected with saline, gentamicin or gentamicin plus DPPD for 4 days and were sacrificed 48 hr later. Gentamicin increased malondialdehyde in renal cortex from a control level of 0.65 +/- 0.04 to 1.01 +/- 0.03 nmol/mg of protein, P less than .01; it was reduced to 0.20 +/- 0.03 by DPPD, P less than .01 compared to control. Arachidonic acid comprised 27.6 +/- 0.5% of the fatty acid in renal cortical phospholipid of control rats. Gentamicin lowered arachidonic acid to 16.7 +/- 0.9%, P less than .01, and promoted a shift toward saturated fatty acids. DPPD reversed these changes. Gentamicin depressed catalase activity from a control value of 0.211 k/min to 0.154 +/- 0.008 k/min, P less than .01. DPPD depressed catalase further to 0.095 +/- 0.066 k/min, P less than .01. Total glutathione and reduced glutathione were depressed whereas the fraction of total glutathione in the oxidized state was augmented by gentamicin. These changes were prevented by DPPD. The renal cortical phospholipidosis induced by gentamicin was not altered by DPPD. The increased urinary excretions of alanine aminopeptidase and N-acetyl-beta-glucosaminidase induced by gentamicin were augmented further by DPPD. In DPPD rats serum creatinine (0.45 +/- 0.04 mg/dl) was higher (P less than .01) than that of gentamicin rats (0.35 +/- 0.01 mg/dl), which was higher (P less than .01) than that of control rats (0.26 +/- 0.01 gm/dl).(ABSTRACT TRUNCATED AT 250 WORDS)

Comparative effects of aminoglycosides on renal cortical and urinary phospholipids in the rat.

We examined the relationship between the nephrotoxicity potential of four aminoglycosides and the capacity of the drugs to induce a renal cortical phospholipidosis. Sprague-Dawley rats were injected subcutaneously with neomycin, gentamicin, tobramycin, or netilmicin, 100 mg/kg per day, for 1 to 4 days, and phospholipid accumulation in the renal cortex and phospholipid excretion in the urine were measured. The rank order of the drug-induced renal cortical phospholipidosis was netilmicin greater than tobramycin greater than gentamicin greater than neomycin. This order is the reverse of the previously established nephrotoxicity potentials of these drugs. Conversely, the rank order according to peak urinary excretion of phospholipids was gentamicin greater than neomycin greater than tobramycin greater than netilmicin. The rank order of the total urinary phospholipid excretion during the 4 days of the study was neomycin greater than or equal to gentamicin greater than tobramycin greater than or equal to netilmicin. Urinary phospholipid excretion may prove to be a sensitive indicator of aminoglycoside nephrotoxicity.

Effect of netilmicin on the phospholipid composition of subcellular fractions of rat renal cortex.

The purpose of this study was to determine the subcellular site(s) of the renal cortical phospholipidosis induced by aminoglycosides. For this purpose we injected male Sprague-Dawley rats s.c. with netilmicin, containing tracer quantities of [3H]netilmicin, at 100 mg/kg/day for 2 days; control rats were injected with saline. Twenty-four hours after the second injection of drug the rats were sacrificed and the renal cortex was fractionated by differential ultracentrifugation and Percoll gradient density techniques to obtain purified lysosomes, mitochondria, microsomes, brush border membranes and basolateral membranes. The total phospholipid content of the renal cortex was 300 +/- 5 nmol/mg of protein in control rats and 340 +/- 5 nmol/mg of protein in netilmicin-injected rats. The total phospholipid content of the lysosomal fraction of netilmicin rats, which was enriched in myeloid bodies and [3H]netilmicin, was 91% greater than that of control rats and reflected significant increases of phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine and phosphatidylinositol. This pattern is identical to that reported previously for the rat renal cortical phospholipidosis induced by aminoglycosides. The total phospholipid contents of the mitochondrial, microsomal, brush border membrane and basolateral membrane fractions of netilmicin-injected rats were higher by approximately 10% than the respective fractions of control rats and each fraction exhibited a significant increase of one or more of the four phospholipids elevated in the renal cortical homogenate and in the lysosomal fraction. The data indicate that the myeloid body is the primary source of the lysosomal phospholipidosis induced by netilmicin which provides support for the hypothesis that the lysosomal phospholipidosis is secondary to aminoglycoside-induced inhibition of phospholipid degradation. In addition the findings of increased phospholipid content and altered phospholipid composition of the other subcellular fractions raise the possibility that aminoglycoside antibiotics cause a more generalized disturbance of phospholipid metabolism characterized by altered synthesis as well as degradation in renal proximal tubular cells.

Effect of gentamicin on lipid peroxidation in rat renal cortex.

We examined the hypothesis that lipid peroxidation participates in the pathogenesis of aminoglycoside-induced nephrotoxicity. Male Sprague-Dawley rats were injected subcutaneously with gentamicin, 100 mg/kg per day, for 1-4 days. Twenty-four or forty-eight hours after the last injection the rats were killed and the renal cortex was processed for total phospholipids, malondialdehyde (MDA), phospholipid fatty acid composition, superoxide dismutase, catalase and glutathione. Gentamicin induced a significant increase in total renal cortical phospholipids which was evident after a single injection and by the third injection reached a plateau 17% above the baseline level. MDA, an end product of lipid peroxidation, increased from 0.674 +/- 0.021 nmole/mg protein in the control group to 0.931 +/- 0.053 nmole/mg protein (P less than 0.001) 48 hr after the fourth injection. As another index of lipid peroxidation, we determined the shift from polyunsaturated to saturated fatty acids of renal cortical phospholipids. By the second injection of gentamicin we detected a significant decline of arachidonic acid (20:4) present in phospholipid. By the fourth injection, arachidonic acid had fallen 48% below control and was accompanied by reciprocal increases of more saturated fatty acids including linoleic (18:2), oleic (18:1) and palmitic (16:0) acids. The number of double bonds per mole of fatty acid declined from a baseline value of 1.62 +/- 0.01 to 1.20 +/- 0.02 (P less than 0.001) by the fourth injection of drug. Superoxide dismutase showed no consistent alteration, whereas catalase activity (k) fell from the control value of 0.221 +/- 0.007 min to 0.155 +/- 0.009 min (P less than 0.01) by the third injection, where k is the first-order rate constant. Total and reduced glutathione declined after the fourth injection of gentamicin accompanied by a shift to oxidized glutathione with an increase in the ratio of oxidized to total glutathione. These data support the conclusion that accelerated lipid peroxidation occurs early in the course of gentamicin administration and raise the possibility that lipid peroxidation is a proximal event in the injury cascade of gentamicin nephrotoxicity.