Increased nephrotoxicity of gentamicin in pyelonephritic rats

Kidney International, Vol. 28 (198), pp. 16 113 LABORATORY INVESTIGATION Increased nephrotoxicity of gentamicin in pyelonephritic rats DENIS BEAUCHAMP, ANDRE PIRIER, and MICHEL G. BERGERON Infectious Diseases Service, Le Centre Hospitalier de l'université Laval, Québec, Québec, Canada Increased nephrotoxicity of gentamicin in pyelonephritic rats. Multiple factors may increase the nephrotoxic potential of aminoglycosides. We studied gentamicin susceptibility of kidneys infected with E. coli. Several parameters of renal function, histological changes on light and electron microscopy, and drug levels in renal parenchyma were compared in pyelonephritic and normal rats treated with low doses (1 mg/kg/q8 hr for 3 days), or high doses (6 mg/kg/day for 14 days), of gentamicin. A significant increase (P <.1) in 13-galactosidase and protein excreted in urine over a period of 17 days associated with severe changes in diuresis and osmolality was noted in the infected treated rats (low doses) compared with normal, treated, infected or control animals. Histological modifications compatible with gentamicin nephrotoxicity were more persistent in the infected treated animals. A significant decrease in '4C inulin (P <.1) and 3H-PAH clearance and secretion (P <.2) was observed in the infected treated rats receiving high doses of antibiotics. Cellular necrosis and tubular desquamation also were more severe in this group. Gentamicin levels in the cortex and medulla of infected animals were significantly higher than in the normals (P <.1) and might have been responsible for the increased toxicity noted in the pyelonephritic animals. Infected kidneys appeared to be more susceptible to the nephrotoxic potential of gentamicin. Néphrotoxicité accrue de Ia gentamicine cbez des rats pyélonéphritiques. De multiples facteurs peuvent accroitre le potentiel néphroxique des aminoglycosides. Nous avons étudie la susceptibilité a la gentamicine de reins infectés par E. coli. Diffdrents paramétres de Ia fonction rénale, les changements histologiques en microscopie optique et électronique, et les niveaux de médicament dans le parenchyme renal ont été compares chez des rats pyelonephritiques ou normaux, traités par de faibles doses (1 mg/kg/8h pendant 3 jours), ou de fortes doses (6 mg/kg/jour pendant 14 jours) de gentamicine. Une élévation significative (P <,1) de Ia 3-galactosidase et des protéines excrétées dans les urines sur une période de 17 jours, associées a des modifications sévéres de la diurése et de l'osmolalité ont été notées chez les rats infectés traités (faibles doses), par rapport aux animaux normaux infectés et traités, ou aux contrôles. Des modifications histologiques compatibles avec une ndphrotoxicité a Ia gentamicine ont été plus persistantes chez les animaux infectés traités. Une baisse significative de Ia clearance et Ia sécrétion de '4C inuline (P <,1) et de 3H-PAH (P <,2) a ete observée chez les rats infectés et traités recevant les fortes doses d'antibiotique. La nccrose cellulaire et Ia desquamation tubulaire étaient également plus sévères dans cc dernier groupe. Les niveaux de gentamicine dans le cortex et Ia médullaire des animaux infectés étaient significativement plus hauts que chez les normaux (P <,1) et pourraient avoir été responsables de l'augmentation de toxicité notee chez les animaux pyélonephritiques. Les reins infectés apparaissaient plus susceptibles au potentiel nephrotoxique de Ia gentamicine. Gentamicin is nephrotoxic. Following glomerular filtration, this aminoglycoside is reabsorbed partially by the proximal Received for publication June 29, 1984, and in revised form December 19, 1984 198 by the International Society of Nephrology tubular cells [1]. The first histologic alterations induced by gentamicin on kidney ultrastructure is the appearance of lamellar structure deposition in the lysosomes of these cells [21. In fact, gentamicin has been shown to interfere with the normal catabolism of phospholipids, leading to further cellular disturbance and necrosis [3}. Many risk factors modulating the susceptibility of the kidney to the toxic potential of aminoglycosides have been identified. These include an already-damaged kidney [4], metabolic acidosis [], potassium depletion [6], concomitant use of cephalosporins [7], diuretics [8], or nephrotoxic agents including glycerol [91, methoxyflurane [1], and mercuric chloride [11]. However, the question of whether kidney infection triggers the nephrotoxicity of aminoglycosides has not been solved. In fact, severe and chronic pyelonephritis, which may affect both proximal and distal tubules, cannot be excluded as a risk factor that may increase the susceptibility of the nephron to aminoglycosides. Our previous investigation has revealed a greater accumulation of gentamicin in the cortex and medulla of pyelonephritic animals than in the kidneys of non-infected animals [121. Moreover, this study has shown that the pyelonephritic rats treated with gentamicin had altered renal functions, while normal treated rats had no sign of toxicity [121. This investigation compares the nephrotoxic potential of gentamicin in both normal and pyelonephritic rats. Several sensitive parameters of renal injury [13], including urinary enzyme excretion, proteinuria, diuresis, osmolality, and histology, were evaluated as indices of early gentamicin toxicity in normal and infected animals. Less sensitive parameters of renal damage, including changes in the glomerular filtration rate, PAH clearance, and PAH secretion were evaluated in all groups, as indices of more severe gentaniicin toxicity. Gentamicin levels and colony counts in the renal parenchyma of these animals were done to see ii drug levels and the intensity of the infection could be correlated with signs of nephrotoxicity. Methods Pyelonephritis model Pyelonephritis was induced in 2 g female Sprague-Dawley rats by direct inoculation of the left kidney with.1 ml inoculum containing i7 to 18 cfu of E. co/i Yale strain (Dr. V. Andriole, Yale University, New Haven, Connecticut, USA) [14]. The minimal inhibitory concentration of gentamicin for E. co/i was 1.6 rg/ml. The bacteria were injected directly into the medulla through the upper and lower poles of the kidney [1]. This inoculation produced bilateral pyelonephritis in all animals; a severe pyelonephritis of the left kidney with extensive inflarn- 16

mation and abscess formation was induced by the inoculation, and a less severe infection of the right kidney with few foci of inflammation and slight swelling of the organ due to the reflux of infected urine into the right ureter and kidney was observed. Basal urine cultures done on all animals were sterile. Treatment Two groups of animals were controls. They included infected (I) and normal (N) rats that had their left kidneys injected with.1 ml sterile Mueller Hinton broth (MHB). The normal and infected rats were assigned to two regimens: short-term treatment where normal (G-3) and infected (IG-3) rats received a 1 mg/kg gentamicin (Schering Corp., Montreal, Canada) every 8 hr for 3 days, and long-term treatment where normal (G-6) and pyelonephritic (IG-6) animals were treated with 6 mg/kg/24 hr for 14 days. In all animals, treatment was started i.p. 24 hr after the induction of the pyelonephritis. Enzymuria, proteinuria, diuresis, and osmolalities Eight animals of each group were housed singly in metabolic cages and allowed 3 days to adapt. Rats had free access to food and water, and their weight was monitored daily. Both water and food intake were recorded. Twenty-four-hr urine collection was made under mineral oil on Days 3, 6, 8, 1, 14, and 17. Collected urines were centrifuged at 2 rpm for 1 mm. Urine volume was measured and osmolality was evaluated by freezing point depression. /3-galactosidase, a lysosomal enzyme, was measured by the method of Maruhn [161 using as substrate P-nitro-phenyl-/3-D-galactopyranoside dissolved in citrate buffer.1m, ph 4. Proteinuria was measured using Lowry's methodology [171. To preclude the possibility that the detected enzyme was not the product of bacterial destruction, 3-galactosidase from E. coli (molecular weight 4,) [181 was separated from 13- galactosidase originating from kidney cells (molecular weight 8,) [19] using a fine column (Sephadex type G-1, Pharmacia Canada Ltd., Montreal, Québec, Canada). Inulin and PAH clearance In the short-term treatment experiments, renal function evaluation was made on Day 1 when enzymuria and proteinuria were maximal and when histology was done, In the long-term treatment group, renal functions were done at the end of therapy. Six rats of each group were anesthetized with pentobarbital mg/kg (Nembutal, Abbott Laboratories, St-Laurent, Québec, Canada). The left carotid artery and the right jugular vein were canulated with PE- and PE-2 polyethylene tubing (Clay Adams, Div. of Becton, Dickinson, New Jersey, USA), for sampling blood and infusion, respectively. An initial bolus of 1 ml NaC1.8% containing 14C-inulin (.1 mci/dl, New England Nuclear, Montreal, Québec, Canada), 3H-PAH (. mciidl, New England Nuclear), unlabelled inulin (1 mg/mi), and PAH (1 mg/mi) was infused at a rate of.28 ml/min. This bolus was followed by a continuous, immediate infusion of this solution at.2 mi/mm. One hr later, urine was collected for three sequential 3-mm periods from the right and left ureters, canulated separately (PE-lO, Clay Adams), in order to evaluate the right and the left kidney function of each animal. Geniamicin toxicity in pyelonephritis 17 Blood samples were taken at the beginning and the end of each urine collection period. '4C-inulin and 3H-PAH were measured in a liquid scintillation counter (Model LS 7, Beckman, Montreal, Quebec, Canada), on double labelled program, with automatic quench correction. Glomerular filtration rate was determined by inulin clearance. PAH clearance and PAH secretion were evaluated also. Histology Five right kidneys from each group receiving high or low doses were processed for histological studies on the tenth day of the experiments. In the long-term treatment groups, as in the controls, histology was done on the last day of treatment, on Day 2, and up to Day 3 of the experiment. At the time of sacrifice, five animals of each group were anesthetized. A midline abdominal incision was made and a 21 gauge butterfly needle was inserted into the aorta above the renal arteries. Following a perfusion of 1 ml Krebs-Ringer solution (ph 7.4) at 7 mi/mm, 1 ml of glutaraldehyde 2% (ph 7.4) was infused into the live animals. After the in vivo glutaraldehyde fixation, the kidneys were excised and placed in the same fixative for 2 hr. Cubes of 1 cubic millimeter were taken from the cortex and outer medulla and left overnight in the same fixative at 4 C. After washing with phosphate buffer. lm (ph 7.4), the cubes were further fixed in 2% osmium tetroxide for 3 hr at 4 C, dehydrated in ascending grades of alcohol, and embedded in araldite 2 resin. Thick sections (1 xm) were cut with a LKB ultratome III, stained with hot toluidine blue and examined to identify gross lesions. Thin sections double stained with alcoholic uranyl acetate and lead citrate were examined using a blinded code on a Phillips 3 electron microscope (Scarborough, Ontario, Canada). Histological damage induced by gentamicin was graded according to the scheme developed by Gilbert et al where the severity of changes observed were ranked on a to 4+ scale [21. Drug levels and colony counts Twenty-four rats were used for these experiments. Gentamicin levels were determined as described previously [21] in the serum, cortex, and medulla of normal and infected animals. Colony counts in whole kidney were estimated using the same technique [14]. Statistics Results, including enzymuria, proteinuria, osmolality, and physiological studies, were compared using an analysis of variance (ANOVA) for repeated measures. Comparison of group means was done using the Duncan's multiple range test with Kramer's adjustment for unequal frequencies [22]. Renal tissue gentamicin concentrations were analyzed by Student's t test. Results Enzymuria, proteinuria, diuresis, and osmolality Figure 1 shows the 24-hr urinary excretion of /3-galactosidase (/3-Gal) and proteinuria for all groups from 3 to 17 days after the beginning of short-term therapy. In contrast with normal rats (N), non-treated pyelonephritic rats (I), or normal treated animals (G-3), the pyelonephritic rats treated with gentamicin

18 Beauchamp et al 3 galactosidase Proteinuria 1 9 1 9 8 7 6 E 4 3 2 4 6 8 1 12 14 16 Days 18 L 8-7 6 E 4 3 2 4 6 8 1 12 14 16 18 Days Fig. 1. Urinary 24-hr excretion of 13-galactosidase and proreinuria in rats after beginning treatment. Symbols are: s., normal rats; o o, pyelonephritic rats; u u, normal rats treated with gentamicin 1 mg/kg Q8 hr for 3 days;, infected rats given G-3 treatment. N G-3 IG-3 Table 1. Diuresis and osmolality (mean 3 6 7.2 (O.)8b 12.4(3.2)" 4.9 (,8)8 8. (l.)8b 7 (7)8 6. (1.2)8.7 (l.)a 1.1 (2.)'-' SEM)8 of a 24-hr urine collected after induction of pyelonephritis Diuresis, ml/24-hr Day 8 1 14 17 8.1 (.6)8 8.1 (1.)8 7.7 (l.2) 1.7 (2.4) 6.3 (.6)8 6.9 (.6)8 13.2 (1,8)" 1.3 (2.)".6 (Ø,)a 6.3 (l.3)" 8.9 (2.3)" 11.2 (1.7)" 8.2 (1.6)8 7.9 (1.3)8 6.8 (1.1)8 1.4 (1.1)8 N -3 IG-3 3 6 181 (171)8 832 (11)b 126 (19)ab 928 (19)" 1746 (97)a 129 (44)c 169 (19)8b 119 (ll)c Osmolality, mosm/kgh2o Day 8 1 14 17 1632 (87)8 1463 (172)8b 1443 (l1)8b 16 (93)" 1982 (183)8 132 (121)" 1612 (71)ab 127 (13)b 1918 (138)8 164 (168)8I 16 (168)ab 131 (132)" 1488 (12)8 143 (96)8 199 (167)'-' 1422 (132)8 Abbreviations are: N, normal animals; I, infected; G-3, treated with 1 mg/kg/q8 hr of gentamicin for 3 days; IG-3, infected treated animals. a For a same day, means followed by a common letter are not significantly different at the 1% level, as shown by the Duncan's multiple range test. (IG-3) had a noticeable increase in their urinary /3-Gal and protein excretion. The excretion of/3-galactosidase of the IG-3 group was maximal from Days 6 to 1, and was significantly higher than in the other groups on Days 6, 8, and 1 (ANOVA P <.1). Enzyme differentiation studies showed that the /3-galactosidase was not the product of E. co/idestruction but of kidney damage. Protein excretion was maximal on the tenth day, and at that time was also significantly higher in the 1-3 animals than in the other groups (P <.1). Compared to the controls, infected, or treated rats, a decrease in the urine osmolality and an increase in diuresis were observed in the infected and treated animals (Table 1). These changes were maximal (P <.1) from Days 6 to 1. The diuresis in all groups was correlated directly with water intake. Contrasting the short-term experiments, there was no difference in /3-galactosidase or protein excretion, diuresis, and osmolality between the infected treated (IG-6) and treated rats (-6) groups. Levels of enzyme and protein and changes in diuresis and osmolality were maximal on Day 1 and were more striking in the long-term than in the short-term treatment groups. Table 2 summarizes the data on Day 1; the normal (N) and infected animals (I) had significantly different values than the -6 and IG-6 groups (P <.1). Daily weights were measured in all animals. Weight loss (< 4%) was observed in the first 3 days in the infected (I) and infected treated rats (IG-3, IG-6), but there was no difference in weight gain between the third and the 17th day of the experiment in all groups. Weight loss could be correlated with food intake, which diminished in the first 3 days in the infected and infected treated groups. mu/in and PAH c/earance In the short-term treatment experiments, there was no significant difference in inulin and PAH clearance between the groups. Figure 2 illustrates renal function tests for both left and right kidneys in the long-term treatment experiments. Compared to the other groups, a significant decrease (P <.2) in the CPAH and TPAH was observed in both kidneys of infected treated animals (1-6). A significant reduction of the CINU was noted in the left kidney of the IG-6 group (P <.1) when compared with either the treated, infected, or normal animals.

Table 2. Diuresis, osmolality, protein, and 3-galactosidase (mean Diuresis ml/24 hr Gentamicin toxicity in pyelonephritis 19 Osmolality mosm/kgh2o SEM)a excreted in a 24-hr urine on Day 1 after induction of pyelonephritis Protein mg/24 hr /3-galactosidase mpj24 hr N 8.1 (1.)a 1982 (183) 48.1 (.1)a 61.6 (.)a I 1.7 (2.4)a 132 (121)' 643 (99)ab.9 (7.6)a G-6 26.1 (4.9)" 61 (164)C 92.2 (l1)b 169.9 (38.4)" IG-6 26.3 (7)b 4 (11) 87.9 (77)b 183.4 (22.4)" Abbreviations are: N, normal animals; I, infected, G-6, normal animals treated with 6 mg/kg/day for 14 days; IG-6, infected animals treated with 6 mg/kg/24 hr for 14 days. a Means followed by a common letter are not significantly different at the 1% level, as shown by the Duncan's multiple range test. CINU mi/mm CPAH mi/mm TPAH moles/min.] -.6 -. -.3 -.2 I I I.4-.1- '-4 H. IRL RL R L R L N G-6 IG-6 H H RL R L RL R L RL RL R L RL N G-6 IG-6 N G-6 G-6 Fig. 2. Glomerular filtration rate (CINU), PAH clearance (CPAH), and PAH secretion (TPAH) of right (R) or left (L) kidneys in groups of rats defined previously. Values for CINU and CPAH are for 1 g body wt. ::.::.:::::::.:::::::::: Histology In the controls, except for the presence of inflammatory cells in the interstitial space, the non-treated right infected kidneys did not reveal any major histological change. No myelin figure nor enlarged lysosome could be seen at that time in this non-treated group, but severe histological signs compatible with pyelonephritis were present on Day 2 and thereafter in the right kidney infected by reflux. No histological modification on light and electron microscopy was noted on Day 1 in the kidneys of normal rats treated with gentamicin (G-3). In contrast, cells containing large and numerous lysosomes with myelin figures were observed in the kidneys of IG-3 rats. Cellular debris and numerous myelin figures were observed in the tubular lumen of these animals. No signs compatible with gentamicin therapy or pyelonephritis were observed in the sterile broth injected controls. Table 3 illustrates histological scores observed in the different groups based on the classification of Gilbert et al [21. Grade I changes were noted in the right kidney of all five infected and treated (IG-3) animals studied, while none of the other groups had any histological modifications compatible with gentamicin toxicity. Typical histological signs of gentamicin toxicity were seen on Table 3. Microscopic proximal tubule changes [2] in the right kidneys of normal and pyelonephritic rats given low and high doses of gentamicin on Day 1 N I G-3 IG-3 G-6 IG-6 No. Grade of proximal tubule changea Light microscopy 1 2 2 Electron microscopy Abbreviations are: No., number of animals studied; N, normal; I, infected; G-3, normal rats treated with 1 mg/kg Q8 hr for 3 days; IG-3, pyelonephritic rats given G-3 treatment; G-6, normal rats treated with 6 mg/kg Q24 hr for 14 days; IG-6, pyelonephritic rats given G-6 treatment. a Pathological changes described were observed in all animals in each group. Day 1 of the experiment on light microscopy in both the G-6 and IG-6 groups (Table 3). However, 1% (19/13 counted tubules) of the proximal tubules observed in the IG-6 kidneys 1 3 3

11 Beauchamp et a! Normal kidneys Table 4. Day 1 concentrations of gentamicin in cortex and medulla of normal and pyelonephritic kidneys of treated animals Pyelonephritic left kidneys Pyelonephritic right kidneys Cortex Medulla p-gig p-gig Low dose High dose Low dose High dose 46 (7)] 173 (24) 163 (32)J P <.1 Definitions are: Low dose, 1 mg/kg Q8 hr; high dose, 6 mglkg Q24 hr. a Standard deviation, NS, not significant. 16 (38)] 131 (36) F 86 (l4)j NS 9 (2) ] f4 (11) P <.1 69 (16) J 26 (6)1 16 (4) 1 ()J NS Table. Percentage of kidneys sterilized and log CFU at Day 1 (sem) in animals receiving short- or long-term therapy with gentamicin Left Right Sterility Log CFU Sterility Log CFU % g % g I 6.3 (.2) 4.6 (.2) IG-3.4 (.) 1 3. (.3) IG6a 1 1 Abbreviations are the same as previous tables. a In the high-dose experiments, 1% of the kidneys were still sterile at Days 1, 2, and 3. had epithelial necrosis and desquamation on Day 1 of treatment, while these changes were observed in only % (7/13 counted tubules) (P <.1) of the tubules of the G-6 kidneys. At the end of long-term treatment, changes were less severe than on Day 1, and we observed the presence of regenerative cells in both groups (G-6 and IG-6). The regeneration process was almost complete by Day 2 when less than 1% of the proximal tubule had epithelial necrosis and desquamation, while on Day 3 of the experiment, none of the tubule was necrotic. Gentamicin levels and colony counts Peak 1-hr serum levels following the last 1 mg/kg dose administered on Day 3 were 12.1 (±.8) p-g/ml in normal and 19. (± 2.3) in infected rats. After the last 6 mg/kg doses, the peak levels were 4.6 (± 3.1) p-g/ml in the normal and 9.4 (± 4.9) in the infected rats. Table 4 depicts the levels of gentamicin on Day 1 of the experiment in both cortex and medulla of animals receiving low and high doses of gentamicin. At the time where renal levels were done, we could not detect any aminoglycosides in serum (<.2 p-giml). When compared with the non-pyelonephritic animals, significantly higher levels of antibiotics were observed in both kidneys of infected animals treated with 1 mg/kg doses (P <.1). In the 6 mg/kg treated rats, the renal levels were identical in both normal and pyelonephritic rats and were lower than in the low-dose treated rats. While 1% of the left kidneys and 9% of the right kidneys were still infected following short-term therapy (Table ), all of the left and right kidneys were sterilized on Days 1, 1, 2, and 3 in the long-term treatment group. Discussion These results support the hypothesis that pyelonephritic kidneys are more susceptible than normal kidneys to the nephrotoxic potential of gentamicin. The use of two therapeutic regimens allowed us to study both very sensitive markers of tubular damage and late manifestations of renal injury. Depending on the more or less sensitive criteria used to evaluate nephrotoxicity, infected treated animals showed signs of renal damage that were more striking than those observed in normal treated rats. In the short-term treatment experiments, major changes in urine osmolality, diuresis, /3-galactosidase and protein excretion associated with concomitant histopathological changes were detected in the pyelonephritic treated animals, but the dosage used was insufficient to induce changes in GFR, a late manifestation of the nephrotoxocity of aminoglycosides [23, 24]. In the long-term treatment experiments, inulin, PAH clearance, and PAH secretion were affected more severely in the infected treated animals, but sensitive markers of renal injury could not discriminate between changes occurring in normal and infected treated animals. These observations are consistent with others that show high doses of gentamicin given for a long period are necessary to decrease GFR and modify PAH transport [24, 2]. Although both normal and infected animals treated with high doses of gentamicin showed similar histopathological changes, the percentage of tubules where these changes occurred was significantly higher in the pyelonephritic rats, suggesting that the infected kidney may be more susceptible to the nephrotoxic potential of aminoglycosides. The possibility that renal infection might have delayed the regeneration process was questioned, but, at least in the high doses experiments where we did observe regenerative cells, we could not follow the speed of regeneration in normal treated and pyelonephritic treated kidneys using conventional histology. Others have shown that pathological changes were very severe by Day 1 [26, 27], and the regeneration process started thereafter [27, 28]. By Day 2, the regeneration was almost complete [27, 29], and all the tubules were normal on Day 3. The mechanism by which aminoglycosides did induce renal toxicity in our study is not well understood [3, 3]. Factors other than infection should be considered. Although this study was not designed to compare the frequency of administration on nephrotoxicity, the use of TID and QD administration of gentamicin should be considered as another factor modulating nephrotoxicity. As shown by Bennett et al [31], multiple doses of a specific daily amount of aminoglycosides were generally more toxic than a single daily administration of the same quantity. In our short- and long-term experiments, both daily

Gentamicin toxicity in pyelonephritis 111 doses and duration of therapy were different, so no conclusion can be drawn on whether an intermittent daily administration is more toxic than a single daily dose. Gender is another risk factor that has to be considered in both human experimentation and animal nephrotoxicity models. Data from Bennett et a! [321 have shown a significantly greater incidence of acute tubular necrosis in male than in female rats. On the other hand, Kourilsky et al [331 have observed a higher incidence of interstitial inflammation in female rats treated with gentamicin. Our study used female rats exclusively, so the decrease in renal function noted in animals treated with high doses of gentamicin was due more likely to tubular damage than to interstitial inflammation. Moreover, we have not observed spontaneous pyelitis in untreated female rats. This and previous studies [12, 141 have used female Sprague-Dawley pyelonephritis models. In fact, it was in this model that we first discovered evidence that infection might potentiate nephrotoxicity of gentamicin [34]. Moreover, two recent reports by Kourilsky et al [33] and Moore et al [3] seem to suggest that human females have a higher incidence of aminoglycosides toxicity. Although there is not necessarily an association between levels of antibiotics in the renal parenchyma and renal toxicity, most investigators consider high renal levels of antibiotics as potentially nephrotoxic [31. The previous observations of Bergeron et a! [121 and this study, which demonstrates again that the infected renal parenchyma accumulated more gentamicin than the normal kidney, suggest that these high levels of aminoglycosides in both the cortex and medulla of pyelonephritic rats might have contributed to the renal impairment observed in these animals. The increase in aminoglycoside levels was very striking in the low-dose experiments (IG-3), a drug regimen that could not sterilize the kidneys, supporting the idea that infection modifies the intrarenal distribution of gentamicin and may potentiate the nephrotoxic capacity of this agent. Serum levels were undetectable at the time of kidney extractions for tissue levels, but one cannot exclude the possibility that higher serum levels, which were observed on the last day of therapy (Day 3), could have been associated with greater tubular reabsorption with resultant higher cortical concentrations of aminoglycosides, a key factor in inducing greater toxicity in the pyelonephritic model. The reason for the noticeable differences between the infected (19. xgiml) and noninfected (12.1 xg/ml) is not clear, but recent data from our laboratory [36] and from Wilson, Moore, and Eakle [37] in horses suggest that endotoxin may modify the pharmacokinetic of antibiotics and increase serum levels. For the low-dose experiments where infection was still active at the time when serum levels were taken, we cannot exclude the possibility that endotoxin might have been liberated from the infected site. Volume contractions in the infected animals cannot be excluded as an explanation for the higher serum levels. Further, we cannot eliminate the possibility that in the acute phase of infection there might have been a transient diminution in the glomerular filtration rate that could have contributed to the elevated aminoglycoside levels in serum. Following high doses of gentamicin, there was no difference in the levels of drugs in the kidneys of normal and sterilized pyelonephritic kidneys, suggesting again that in the low-dose experiments, active infection has modified the intrarenal pharmacokinetic of gentamicin [12]. The combined action of high levels of gentamicin and of infection in the renal parenchyma might have been more toxic to renal cells than either of those factors taken individually. Even though most of the toxicity of aminoglycosides has been attributed to proximal tubular cell damage [1], disturbed concentrating ability of the kidney, also a function of the distal nephron, has been shown as one of the early signs of nephrotoxicity associated with gentamicin [2]. In fact, levels of gentamicin in the medulla may persist for long periods following cessation of therapy [21] and, as demonstrated again in this study, the medullary levels were much higher in the infected than in the normal rats. In addition to the possible gentamicin effect on the concentrating ability of the kidney [38], active infection [39], and tissue damage secondary to pyelonephritis or even endotoxin [4], with or without the presence of bacteria in the renal parenchyma, could have contributed to the concentrating defects observed in pyelonephritis. Aminoglycosides then might diminish further the already disturbed urinary concentrating ability of the pyelonephritic kidneys. Moreover, water losses, presumably associated with potassium and sodium losses, cannot be excluded as potential factors in increased nephrotoxicity. Sodium and potassium depletion has been shown to increase the toxic potential of gentamicin [6, 41]. Nor can changes in extracellular volume secondary to acute pyelonephritis be overlooked as one of several variables that might have modulated the nephrotoxicity of the aminoglycosides. Proteinuria, which can be observed in pyelonephritis, is an early manifestation of gentamicin nephrotoxicity [26, 42]. In the infected treated animals, it may be postulated that proteins in the urine might have resulted from the combined action of gentamicin and infection on proximal tubule [26], and infection in the distal elements of the nephron [43]. Protein leakage through the glomerula, although possible [26], seems unlikely to most investigators who believe that following aminoglycosides proteins leak through the tubular cells [26, 42]. Primary in these findings of the infected treated rats is the significant modification in renal function and enzyme excretion, associated with concomitant histological modifications compatible with an increased gentamicin nephrotoxicity. We have shown, as have Patel et al [44], that there is a good correlation between the level of /3-galactosidase excreted in the urine and damage to the lysosomes. These enzymes were not the product of bacterial destruction [181, but most likely originated from lysosomal damage in tubular cells [19]. As more gentamicin was given [4], the excretion of J3-galactosidase increased. The mere increase in the amount of aminoglycosides in the cortex of the infected treated animals is probably sufficient to explain the anomalies observed, but one should consider the possibility that endotoxin or other bacterial products liberated during antibiotic therapy [46] might act concomitantly on tubular cells to potentiate the toxicity of gentamicin. It might be speculated that endotoxin, which is known to affect lysosomes [33, 47] and which can, as does gentamicin, inhibit mitochondrial enzymes [48], block receptors [47, 49], and decrease the activity of Na K ATPase [, 1], could act synergistically with the aminoglycoside to induce tubular changes. We are studying now whether kidneys infected with enterococci, a pathogen without endotoxiri, are as susceptible to aminoglycosides as those infected by E. co/i. In addition to the combined action of high levels of

112 Beauchamp et a! gentamicin and of bacteria on kidney cells, one cannot exclude the influence of the pathological process of pyelonephritis on the dynamic interaction between aminoglycosides and kidney cells [43]. Even in the absence of active bacterial infection [2] observed in the high-dose experiments, where kidneys were sterilized by therapy, renal inflammation might affect membrane permeability, cause local ischemia, or induce the liberation of several mediators of inflammation that may make kidney cells more susceptible to aminoglycoside toxicity [341. Whether synergism exists between pyelonephritis and aminoglycosides that would produce nephrotoxicity is diflicult to determine. Recent work by Moore et al [3] could not identify specific infections as a risk factor for aminoglycosides-induced nephrotoxicity, although a correlation with septicemia was observed for ototoxicity [3]. Future analyses of kidney infections might define more specifically the interaction between aminoglycosides and infected kidney cells. Whatever the mechanism involved, renal infection should be considered an additional risk factor involved in the increased susceptibility of the kidney to aminoglycosides. Acknowledgments Dr. D. Beauchamp won the Student Award of the Canadian Society for Clinical Investigation, 1982, for this work, supported by Medical Research Council of Canada Grant MA-27. These data were presented in part at the International Congress of Chemotherapy, Vienna, Austria, August, 1983, and at the Canadian Society for Clinical Investigation, Québec City, Québec, Canada, September, 1982. The authors thank Dr. A. Whelton, Johns Hopkins University, for advice and review of the manuscript, and Mrs. C. Giguère and Mrs. L. Villeneuve for secretarial assistance. Reprint requests to Dr. M. G. Bergeron, Chief, Infectious Disease Service, Le Centre Hospitalier de l'université Lava!, 27 Laurier Blvd., Quebec, P.Q., GI V 4G2, Canada References 1. SILVERBLATT FJ, KUEHN C: Autoradiography of gentamicin uptake by the rat proximal tubule cell. Kidney In! 1:33 34, 1979 2. KOSEK JC, MAZZE RI, COUSINS MJ: Nephrotoxicity of gentamicin. Lab Invest 3:48 7, 1974 3. LAURENT G, CARLIER MB, ROLLMAN B, VANHOFF F, TULKENS P: Mechanism of aminoglucoside-induced lysosomal phospholipidosis: in vitro and in viva studies with gentamicin and amikacin. Biochem Pharm 3 1:3861 387, 1982 4. DEROSA F, BEcONcRITIANI V, CAPITANNUCI V, FRONGELLO RF: Tobramycin: Toxicological and pharmacological studies in animals and pharmacokinetics research in patients with varying degrees of renal impairment. J mt Med Res 2:1 114, 1974. Hsu CH, KURTZ TW, EASTERLING RE, WELLER JM: Potentiation of gentamicin nephrotoxicity by metabolic acidosis. Proc Soc Exp Ther Med 146:894 897, 1974 6. BRINKER KR, BULGER RE, DOBYAN DC, STACEY TR, SOUTHERN PM, HENRICH WL, CRONIN RE: Effect of potassium depletion on gentamicin nephrotoxicity. J Lab Gun Med 98:292 31, 1981 7. GREEN WH: Cephalosporin and aminoglycoside interaction. Ann Intern Med 9:43 436, 1979 8. ADELMAN RD, SPANGLER WL, BEASON F, ISHIZAKI G, CONZELMAN GM: Furosemide enhancement of experimental gentamicin nephrotoxicity: comparison of functional and morphological changes with activities of urinary enzymes. J Infect Dis 14:342 32, 1979 9. Hiisscis GH: Enhancement of gentamicin nephrotoxicity by glycerol. Toxicol App! Pharmacol 29:27 276, 1974 1. BARR CA, MAZZE RI, COUSINS MJ, KOSEK JC: An animal model for combined methoxyflurane and gentamicin nephrotoxicity. Br J Anaesth 4:36 312, 1973 11. LUFT FC, YUM MN, KLEIT SA: The effect of concomitant mercunc chloride and gentamicin on kidney function and structure in the rat. flab Clin Med 89:622 631, 1977 12. BERGERON MG, TROTTIER S, LESSARD C, BEAUCHAMP D, GAGNON PM: Disturbed intrarenal distribution of gentamicin in experimental pyelonephritis due to E. coli. J Infect Dis 146:436 439, 1982 13. KALOYANIDES GJ, PASTORIZA-MUNOZ E: Aminoglycosides nephrotoxicity. Kidney mt 18:71 82, 198 14. BERGERON MG, BASTILLE A, LESARD C, GAGNON PM: Significance of intrarenal concentration of gentamicin for the outcome of experimental pyelonephritis in rats. Jinfect Dis 146:91 96, 1982 1. KAYE D: The effect of water diuresis on spread of bacteria through the urinary tract. J Infect Dis 124:297 3, 1971 16. MARUHN D: Rapid colorimetric assay of /3-galactosidase and N- acetyl-/3-glucosidase in human urine. Clin Chim Acta 73:43 461, 1976 17. LOWRY OH, ROSEBROUGH NJ, FARE AL, RANDALL RI: Protein measurement with the folin phenol reagent. J Biol Chem 193:26 27, 191 18. CRAVEN GR, STEERS E, ANFINSEN CB: Purification composition and molecular weight of the /3-galactosidase of E. co!i K12. J Biol Chem 24:2468 2477, 196 19. ROBINSON P, PRICE RG, DANCE N: Separation and properties of 13-galactosidase, /3-glucosidase, /3-glucuronidase and N-acetyl-3- glucosaminidase from rat kidney. Biochem J 12:2 32, 1967 2. GILBERT DN, PLAMP C, STAW P, BENNETT WM, HOUGI-ITON DC, PORTER G: Comparative nephrotoxicity of gentamicin and tobramycin in rats. Antimicrob Agents Che,nother 13:34 4, 1978 21. BERGERON MG, TROTTIER S: Influence of single or multiple doses of gentamicin and netilmicin on their cortical, medullary and papillary distribution. Antimicrob Agents Chemoiher 1:63 641, 1979 22. KRAMER CY: Extension of multiple range tests to group means with unequal numbers of replication. Biometrics 23:37 31, 196 23. Hsu CH, KURTZ TW, WELLER JM: In vitro uptake of gentamicin by rat renal cortical tissue. Antimicrob Agents Chemot her 12:192 194, 1977 24, COHEN L, LAPKIN R, KALOYANIDES GJ: Effect of gentamicin on renal function in the rat. J Pharmacol Exp Ther 193:264 273, 197 2. LAPKIN R, BOWMAN R, KALOYANIDES GJ: Effect of gentamicin on P-amino-hippurate metabolism and transport in rat kidney slices. J Pharmacol Exp Ther 21:233 242, 1977 26. CUPPAGE FE, SETER K, SULLIVAN LP, REITZES EJ, MELNYKO- VYcH AD: Gentamicin nephrotoxicity. II. Physiological, biochemical and morphological effects of prolonged administration to rats. Virchows Arch [Ce!! Pathol] 24:121 138, 1977 27. GILBERT DN, HOUGHTON DC, BENNETT WM, PLAMP CE, REGER K, PORTER GA: Reversibility of gentamicin nephrotoxicity in rats: Recovery during continuous drug administration. Proc Soc Exp Biol Med 16:99 13, 1979 28. LUFT FC, RANKIN LI, SLOAN RS, YUM NY: Recovery from aminoglycoside nephrotoxicity with continued drug administration. Antimicrob Agents Chemother 14:284 287, 1978 29. KAHN T, BOSCH J, WIENER P, DIKMAN S: Course of gentamicin nephrotoxicity. Toxicology 16:49 7, 198 3. SOBERON L, BOWMAN RL, PASTORIZA-MUNOZ E, KALOYANIDES GH: Comparative nephrotoxicities of gentamicin, netilmicin and tobramycin in the rat. J Pharmacol Exp Ther 21:334 343, 1979 31. BENNETT WM, PLAMP CE, GILBERT DN, PARKER RA, PORTER GA: The Influence of Dosage Regimen on Experimental Gentamicin Nephrotoxicity: Dissociation of Peak Serum Levels from Renal Failure. J Infect Dis 14:76 79, 1979 32. BENNETT WM, PARKER RA, ELLIOTT WC, GILBERT DN, HOUGHTON DC. Sex-Related Differences in the Susceptibility of Rats to Gentamicin Nephrotoxicity. J Infect Dis 14:37 373, 1982 33. KOURILSKY, SOLEY K, MOREL-MAROGER L, WHELTON A, DUHOUX P, SEAER J. The Pathology of Acute Renal Failure due to Interstitial Nephritis in Man with Comments on the Role of Interstitial Inflammation and Sex on Gentamicin Nephrotoxicity. Medicine 61:28 268, 1982 34. BEAUCHAMP D, PIRIER A, BERGERON MG: Increased Nephrotoxicity of Gentamicin in the Presence of Infection, in

Gentamicin toxicity in pyelonephritis 113 Current Chemotherapy and Immunotherapy, edited by PERITI P, GRAssi GG, Washington, D.C., American Society for Microbiology, 1982, pp 811 813 3. Mooit RD, SMITH CR, LEPSKY JJ, MELLETS ED, LEITMAN PS: Risk Factors for Nephrotoxicity in Patients Treated with Aminoglycosides. Ann Intern Med 1:32 37, 1984 36. BEAUCI-IAMP D, BERGERON Y, BERGERON MG. Influence of Endotoxin on the Intrarenal Distribution of Gentamicin (abstract) 22nd Interscience Conference on Antimicrobial Agents and Chemotherapy, Florida, USA, 1982. 37. Wnso RC, MooRE JN, EAKLE N. Gentamicin Pharmacokinetics in Horses Given Small Doses of Escherichia coil Endotoxin. Am J Vet Res 44:1746 1749, 1983 38. MCNEILL JS, JACKSON B, NELSON L, BATKUS DE: The Role of Prostaglandins in Gentamicin-Induced Nephrotoxicity in the Dog. Nephron 33:22 27, 1983 39, KAYE D, ROCHA H: Urinary Concentrating Ability in Early Experimental Pyelonephritis. J Clin Invest 49:1427 1437, 197 4. MANNY J, LivNi N, SCHILLER M, GUTTMAN A, Boss J, RABINOVICI N: Structural changes in the perfused canine kidney exposed to the direct action of endotoxin. Isr J Med Sd 16:13 161, 198 41. BENNETT' WM, HARTNETT MN, GILBERT D, HOUGHTON D, POR- TER GA: Effect of sodium intake on gentamicin nephrotoxicity in the rat (39296). Proc Soc Exp Biol Med 1 1:736 738, 1976 42. LUFT FC: Experimental aminoglycoside nephrotoxicity. J Lab Clin Med 86:213 22, 197 43. SHIMAMURA T: Mechanism of renal tissue destruction in an experimental acute pyelonephritis. Exp Mo! Pathol 34:34 42, 1981 44. PATEL V, LUFT FC, YUM MN, PATEL B, ZEMAN W, KLEIT SA: Enzymuria in gentamicin induced kidney damage. Antimicrob Agents Chemother 7:364 369, 197 4. WELLWOOD JM, LOVELL D, THOMPSON AE, TIGLE JR: Renal Damage Caused by Gentamicin: A Study of the Effects on Renal Morphology and Urinary Enzyme Excretion. J Pathol 118:171 182, 1976 46. GoTo H, NAKAMURA S: Liberation of endotoxin from Escherichia coli by addition of antibiotics. Jpn J Exp Med :3 43, 198 47. SHANDS JW: Affinity of endotoxin for membranes. J Infect Dis 128:197 21, 1973 48. MCGIVNEY A, BRADLEY SG: Action of bacterial endotoxin and lipid A on mitochondrial enzymes activities of cells in culture and subcellular fractions. Infect Immun 2:664 671, 1979 49. SASTRASJNH M, KNAUSS TC, WEINBERG JM, HUMES HD: identification of the gentamicin receptor of renal brush border membranes (abstract) Kidney Int 19:213, 1981. CHAHWALA SB, HARPAR ES: An investigation of the effects of aminoglycoside antibiotics on Na-K-ATPase as a possible mechanism of toxicity. Res Commun Chem Pathol Pharmaco! 3:63 78, 1982 1. UTTLI R, ABERNATHY CO, ZIMMERMAN HG. Inhibition of Na- KtAdenosinetriphosphatase by Endotoxin. A Possible Mechanism for Endotoxin-Induced Cholestasis. J Infect Dis 136:83 86, 1977 2. BRAUN JP, RICO AG, BENARD P, BURGAT-SOCAZE V, EGHBALI B, GODFRAIN JC: La Gamma-Glutamyl Transferase urinaire en toxicologie rénale, bases de son utilisation-intérét lors de néphrite aiguë mercuielle. Toxicology 11:73 82, 1978 3. MOORE RD, SMITH CR, LIETMAN PS: Risk-Factors for the Development of Auditory Toxicity in Patients Receiving Aminoglycosides. J Infect Dis 149:23 3, 1984