Effects of Fasting on Temporal Variations in Nephrotoxicity of Gentamicin in Rats

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Mar. 1996, p. 670 676 Vol. 40, No. 3 0066-4804/96/$04.00 0 Copyright 1996, American Society for Microbiology Effects of Fasting on Temporal Variations in Nephrotoxicity of Gentamicin in Rats DENIS BEAUCHAMP, 1,2 PIERRE COLLIN, 1 LOUIS GRENIER, 1,2 MICHEL LEBRUN, 1,2 MICHÈLE COUTURE, 1,2 LOUISE THIBAULT, 3 GASTON LABRECQUE, 4 AND MICHEL G. BERGERON 1,2 Laboratoire et Service d Infectiologie, Centre de Recherche du Centre Hospitalier de l Université Laval, 1 and Département de Microbiologie, Faculté demédecine, 2 and École de Pharmacie, 4 Université Laval, Sainte-Foy, Québec G1V 4G2, and School of Dietetics and Human Nutrition, Macdonald Campus of McGill University, Montreal, Quebec H9X 3V9, 3 Canada Received 2 February 1995/Returned for modification 25 June 1995/Accepted 14 December 1995 Evidence for temporal variations in the nephrotoxicity of low doses of aminoglycosides were recently shown by using specific and sensitive parameters of renal toxicity. The aim of the present study was to evaluate the effect of a short period of fasting on the temporal variations in the renal toxicity of gentamicin. Twenty-eight normally fed (i.e., food and water were available ad libitum throughout the experiment) female Sprague-Dawley rats (weight, 175 to 220 g) and 28 fasted rats (i.e., only water was available during a 12-h fast before and a 24-h fast after gentamicin injection) were used. The animals were synchronized on a 14-h light, 10-h dark cycle (lights on at 0600 h) for 1 week before gentamicin administration. In July 1993, each group of animals was treated with a single intraperitoneal injection of saline (NaCl, 0.9%) or gentamicin (150 mg/kg of body weight) at either the peak (1400 h) or the trough (0200 h) of the previously determined toxicity. On day 1, the 24-h urinary excretion of -galactosidase, N-acetyl- -D-glucosaminidase, and -glutamyltransferase was significantly higher in normally fed animals treated with gentamicin at 1400 h than in their time-matched controls and in normally fed animals treated at 0200 h (P < 0.01), which had normal levels of these enzymes. By contrast, the urinary excretion of these enzymes was significantly higher in both groups of gentamicin-treated, fasted rats than in their time-matched control groups (P < 0.01), reaching levels similar to those measured in normally fed rats treated at 1400 h. The accumulation of gentamicin was significantly lower in the renal cortex of normally fed rats treated at 0200 h than in rats treated at 1400 h (P < 0.05), but this time-dependent difference was not found in fasted rats treated at 0200 and 1400 h. Immunogold labeling done on ultrathin sections and observed by electron microscopy showed a similar subcellular localization of gentamicin in normally fed and fasted rats treated at either 1400 or 0200 h. These results suggest that the feeding period is of crucial importance in the temporal variations of the nephrotoxicity of gentamicin in rats. Aminoglycosides are still widely used as bactericidal agents for the treatment of severe gram-negative infections. Their antibacterial activity is mediated by an irreversible inhibition of protein synthesis in bacteria. However, the clinical use of aminoglycosides can be limited by their ototoxicity and nephrotoxicity. Several factors modulating aminoglycoside nephrotoxicity have been described over the last few years in both humans and experimental animals. Among these factors, the time of day of administration is being increasingly explored because of the recent interest in once-daily administration of aminoglycosides, which current evidence suggests to be as safe and effective as regimens that use multiple daily doses for the treatment of certain infections caused by gram-negative bacteria (for reviews, see references 2 and 14). Nakano and Ogawa (22) were the first to report a timedependent variation in the toxicity of gentamicin. They noted that the gentamicin 50% lethal dose was lower when the drug was injected in mice at 1400 h (middle of the rest period) than when it was injected at 0200 h (middle of the activity period). Several other investigations confirmed that the peak toxicity of * Corresponding author. Mailing address: Laboratoire et Service d Infectiologie, Room 9500, Centre de Recherche du CHUL, 2705 Boul. Laurier, Ste-Foy, Québec, Canada G1V 4G2. Phone: (418) 654-2705. Fax: (418) 654-2715. Electronic mail address: Denis.Beauchamp @crchul.ulaval.ca. aminoglycosides corresponds to the middle of the rest period, and the trough toxicity is measured during the activity period of animals, irrespective of the different doses and parameters of toxicity used (12, 24, 26, 32, 33). Recent studies done in our laboratory showed temporal variations in the nephrotoxicity of tobramycin (18, 19). Our data also showed that renal clearance was significantly higher and the area under the concentrationtime curve in serum of tobramycin was significantly lower in animals treated at 0200 h than in animals treated at 1400 h (18). These changes were associated with a higher renal toxicity of tobramycin in animals treated at 1400 h than in those treated at 0200 h (18, 19). The aim of the present study was to investigate the effects of a short period of fasting on the temporal variations in gentamicin nephrotoxicity, because the trough of toxicity was consistently measured during the maximal food intake period of the animals. MATERIALS AND METHODS Animals and treatment. We used 56 female Sprague-Dawley rats (Charles River Inc., Montréal, Québec, Canada) weighing between 175 and 220 g. They were maintained on a 14-h light, 10-h dark cycle (lights on at 0600 h), which correspond to the conditions prevailing in our animal facilities. Room temperature was maintained at 23 2 C. The animals were divided into two groups of 28 animals each. The first group had free access to food and water throughout the experiment. The second group was fasted 12 h before and 24 h after gentamicin injection, but water was available ad libitum. Groups of six animals each received a single injection of either saline (NaCl, 0.9%) or gentamicin (kindly 670

VOL. 40, 1996 FASTING AND GENTAMICIN TOXICITY 671 donated by Schering Canada, Pointe-Claire, Québec, Canada) at 150 mg/kg of body weight intraperitoneally (i.p.). Injections were given at either 1400 h (middle of the rest period) or 0200 h (middle of the activity period). All animals were killed exactly 72 h after the injection. One hour before being killed, all animals were injected with a single dose of [ 3 H]thymidine (200 Ci; Amersham Canada, Oakville, Ontario, Canada) for measurement of cellular regeneration. The rats were killed by decapitation, blood was collected and centrifuged, and the serum was frozen for measurement of serum creatinine levels by using a Hitachi 737 analyzer. Both kidneys were removed and bisected; the renal cortex was dissected. Part of the tissue was immediately frozen on dry ice for further biochemical analysis. A piece of tissue was placed in a drop of 2% glutaraldehyde, cut into small blocks of 1 mm 3, and left in the same fixative overnight at 4 C. Urinary enzyme activities. All animals were individually housed in metabolic cages throughout the experiment. Urine was collected from 0 to 24 h and from 48 to 72 h after the injection of gentamicin; the volume was noted. The water intake as well as the weights of the animals were also monitored. The activities of -galactosidase and N-acetyl- -D-glucosaminidase were measured by following the methodology of Maruhn (21). The activity of -glutamyltransferase was assessed by the technique described by Persijn and van der Slik (28). The results are presented as the percentage of the enzyme activity measured in timematched control groups. All control groups (normally fed and fasted rats) given saline at either 1400 or 0200 h showed similar baseline data. Biochemical analysis. The measurement of [ 3 H]thymidine incorporation into the DNA of the renal cortex was performed as described by Laurent et al. (17). Evaluation of renal function was made by measuring serum creatinine levels in blood collected when the rats were killed. Results are presented as the percentage of the data measured in time-matched control groups. All control groups (normally fed and fasted rats) given saline at either 1400 or 0200 h showed similar baseline data. Gentamicin accumulation in the renal cortex was measured by a fluorescence polarized immunoassay (TDX System; Abbott Diagnostics, Mississauga, Ontario, Canada). Briefly, kidney samples were homogenized in distilled water with a Tissue-Tearor. The homogenates were also sonicated and diluted in the TDX buffer to obtain concentrations between 0 and 10 g/ml. The inferior limit of detection of the assay was 3 g/g of tissue. The percent recovery of gentamicin in renal homogenates was 78.0% 2.9%. The interday coefficients of variations were 3.44% at 1 g/ml and 2.7% at 8 g/ml. Histology. On the day after sacrifice, blocks of tissue previously fixed in 1% glutaraldehyde were washed in 0.1 M phosphate buffer (ph 7.4), dehydrated in ascending grades of ethanol, and embedded in Araldite 502 epoxy resin (J. B. Em Services Inc., Pointe-Claire, Québec, Canada). Ultrathin sections (silver to light gold) were cut with an ultramicrotome (Ultracut E; Leica Canada Inc., Québec, Québec, Canada) and were mounted on nickel grids before being processed for immunocytochemistry. Immunogold labeling. To determine the subcellular localization of gentamicin, we used the protein A-gold complex technique that we have described previously (3 5). Each grid was first floated on a drop of phosphate-buffered saline (PBS) containing 0.25% bovine serum albumin (Sigma Chemical Company, St. Louis, Mo.) for 20 min. Each grid was then placed on a drop of sheep antigentamicin (Cortex Biochem, San Leandro, Calif.) diluted 1/100 to 1/500. These dilutions gave optimal results and low background activities. Incubations with the antisera were carried out for 60 min at room temperature in a moist chamber. The sections were then rinsed with PBS to remove unbound antibody and were incubated for 30 min on a drop of protein A-gold (15-nm diameter) complex diluted 1/10 with PBS-polyethylene glycol (0.02%). At the end of the incubation, the sections were successively washed with PBS and distilled water and dried. Staining of the grids was performed with uranyl acetate and lead citrate before they were examined with a JEOL JEM-1010 electron microscope (Jeol Canada, St-Hubert, Québec, Canada) at 60 kv. Immunogold labeling controls. Control experiments for assessing the specificity of the immunolabeling were performed as follows: (i) incubation with protein A-gold alone to identify its nonspecific absorption to the section; (ii) absorption of antibody with its specific antigen before performing the labeling protocol to verify the specificity of the antigen-antibody interaction; (iii) incubation with unlabeled protein A before applying the protein A-gold complex to verify the specificity of the immunoglobulin G-protein A interaction; (iv) replacement of the specific antibody with normal serum; and (v) immunogold labeling of the renal cortical tissue of untreated animals. Statistics. Statistical differences between groups were measured by variance analysis by using a least-squared methodology. For P values of less than 0.05, a group comparison was performed by using the Fisher protected least-significantdifference test. Analyses were performed on a personal computer with Statview Graphics. RESULTS Water intake, urine volume, and animal weight. The volume of water intake measured over a 24-h period corresponding to the collection of urine was similar in normally fed and fasted rats. The urine output was greater in fasted rats treated with FIG. 1. Twenty-four-hour urinary excretion of -galactosidase (percentage of the value for the control standard deviation) 0 to 24 h after a single injection of gentamicin (150 mg/kg, i.p.) given at either 1400 or 0200 h in normally fed and fasted rats. No significant difference was observed between all time-matched control groups., P 0.01 compared with time-matched control;, P 0.01 between groups. Open boxes, light period; closed boxes, dark period. saline at 1400 h (middle of the rest period) than in fasted rats treated with saline at 0200 h (middle of the activity period) and normally fed animals treated with saline at either 1400 and 0200 h. Treatment with gentamicin also increased the urine volume, as expected. Fasted rats lost approximately 25 g in weight during the fasting period. However, at the time of sacrifice, similar weights were measured between normally fed and fasted rats. Urinary enzymes excretion. Figure 1 shows the 24-h urinary excretion of -galactosidase measured 0 to 24 h after gentamicin injection. The results for each group are expressed as percentages of the values for their time-matched control groups. The -galactosidase activity in the urine of fed rats treated with gentamicin at 1400 h was significantly higher than those in matched controls and those treated with gentamicin at 0200 h (P 0.01). The -galactosidase excretion in the urine of the 0200-h gentamicin-treated group was not significantly different from that in their matched control group. In fasted rats, the urinary excretion of -galactosidase showed a different pattern. In fact, -galactosidase activity in urine was significantly higher in both gentamicin-treated groups (1400 and 0200 h) compared with that in their control groups, reaching excretion values similar to those measured in normally fed rats treated at 1400 h (P 0.01). From 48 to 78 h after the injection of gentamicin, no significant differences were seen in the excretion of -galactosidase between any of the groups (data not shown). A pattern of urinary excretion of N-acetyl- -D-glucosaminidase similar to that of -galactosidase was seen (data not shown). The urinary excretion of -glutamyltransferase is shown in Fig. 2. The results are expressed as percentages of the values for their time-matched control groups. In normally fed rats, a significant increase in the -glutamyltransferase excretion occurred within the first 24 h after the injection of gentamicin given at 1400 h compared with the excretion measured in the urine of rats treated at 0200 h. However, there was no difference between fasted rats treated at 1400 and 0200 h. In the fed group, the urinary excretion of -glutamyltransferase was significantly higher in rats treated at 1400 h than in controls (P 0.01). This difference between the treated groups and the controls was also significant at 1400 and 0200 h in the fasted rats.

672 BEAUCHAMP ET AL. ANTIMICROB. AGENTS CHEMOTHER. FIG. 2. Twenty-four-hour urinary excretion of -glutamyltransferase (percentage of the value for the control standard deviation) 0 to 24 and 48 to 72 h after a single injection of gentamicin (150 mg/kg, i.p.) given at either 1400 or 0200 h in normally fed and fasted rats. No significant difference was observed between all time-matched control groups., P 0.01 compared with timematched controls;, P 0.01 between groups. Open boxes, light period; closed boxes, dark period. These differences persisted for up to 72 h after gentamicin injection. Biochemical analysis. The cellular regeneration measured by the incorporation of [ 3 H]thymidine into the DNA of the renal cortex is shown in Fig. 3A. The results are expressed as percentages of the values for their time-matched controls. In normally fed animals, a significantly higher cellular regeneration was measured in rats treated at 1400 h compared with that in their matched controls (P 0.01), whereas those treated at 0200 h showed normal values. In fasted rats, a significant increase in cellular regeneration was noted in those treated at 1400 and 0200 h compared with that in their controls (P 0.01). Figure 3B shows the serum creatinine levels expressed as percentages of the levels in the sera of their time-matched controls. Significantly higher levels of creatinine were measured in the sera of fasted animals than in the sera of normally fed rats treated at either 1400 or 0200 h (P 0.05). However, no significant difference in serum creatinine levels was associated with the time of gentamicin administration. Gentamicin levels. Figure 4 shows the gentamicin accumulation in the renal cortex following a single injection given at either 1400 or 0200 h in normal and fasted rats. The levels of gentamicin were significantly lower in the renal cortices of normally fed animals treated at 0200 h than in the renal cortices of those treated at 1400 h (P 0.05). By contrast, the levels of gentamicin in the renal cortices of fasted rats treated at 1400 h did not differ significantly from those in the renal cortices of rats treated at 0200 h. Immunogold labeling. The subcellular distribution of gentamicin is illustrated in Fig. 5. In normally fed animals treated at either 1400 or 0200 h, gold particles were seen over the lysosomes of proximal tubular cells (Fig. 5A and B). However, fewer gold particles were found over the lysosomes of normally fed rats treated at 0200 h than in those treated at 1400 h. In fasted rats, the densities of the gold particles over the lysosomes of proximal tubular cells were similar in both gentamicin-treated groups and were also similar to the densities in normally fed rats treated at 1400 h (Fig. 5C and D). The control experiments proved the high specificity of the labeling. The absorption of the antibody with its antigen resulted in the abolition of the labeling. Finally, no labeling over any subcellular site of untreated control kidneys was observed. Each of these observations was similar for all experimental groups. FIG. 3. [ 3 H]thymidine incorporation in the renal cortex DNA (percentage of the value for the control standard deviation) (A) and serum creatinine levels (percentage of the value for the control standard deviation) (B) 72 h after a single injection of gentamicin (150 mg/kg, i.p.) given at either 1400 or 0200 h in normally fed and fasted rats. No significant difference was observed between all time-matched control groups., P 0.05 compared with time-matched controls;, P 0.01 compared with time-matched controls. Open boxes, light period; closed boxes, dark period.

VOL. 40, 1996 FASTING AND GENTAMICIN TOXICITY 673 FIG. 4. Gentamicin levels (mean standard deviation) in the renal cortex 72 h after a single injection of gentamicin (150 mg/kg, i.p.) given at either 1400 or 0200 h in normally fed and fasted rats., P 0.05 between groups. Open boxes, light period; closed boxes, dark period. DISCUSSION The results of the present study show temporal variations in the nephrotoxicity of gentamicin in normally fed rats. In fact, the 24-h urinary excretion of -galactosidase, N-acetyl- -Dglucosaminidase, and -glutamyltransferase, the cellular regeneration, and the levels of gentamicin in the renal cortex were significantly higher in animals given a single injection of gentamicin at 1400 h than in those similarly treated at 0200 h. By contrast, no temporal variations in gentamicin toxicity were observed in fasted rats treated with the same doses of gentamicin at the same time of day. Previous studies done in our laboratory showed that the trough of tobramycin toxicity was repeatedly obtained when tobramycin was injected at 0200 h (middle of the activity period) (18, 19), and this was also seen in the present study in normally fed animals treated with gentamicin. The activity period of rodents also corresponds to the time of day of maximal food intake. The present study was thus designed to determine the influence of food intake on temporal variations in gentamicin toxicity. In fact, gentamicin toxicity was compared in normally fed and fasted rats treated at either the peak (1400 h) or the trough (0200 h) of toxicity, as previously determined in our laboratory (18, 19). To reduce stress to the animals, we used only a short period of fasting (12 h before and 24 h after gentamicin injection). A single high dose of gentamicin (150 mg/kg) was injected to allow detection of early and noninvasive parameters of toxicity. Measurements of -galactosidase, N-acetyl- -D-glucosaminidase, and -glutamyltransferase excretion in urine were used as parameters of cellular toxicity caused by gentamicin. The urinary excretion of these enzymes shows temporal variations in the toxicity of gentamicin in normally fed rats within the first 24 h. In fact, significantly higher enzyme activities were observed in the urine of rats given gentamicin at 1400 h than in those treated at 0200 h (Fig. 1 and 2). By contrast, the urinary excretion of these enzymes was similar in both gentamicintreated, fasted groups, reaching levels similar to those measured in normally fed rats treated at 1400 h. Similar results were obtained with tobramycin (unpublished data). Nakano and Ogawa (22), the first to show temporal variations in the toxicity of gentamicin, observed that gentamicin killed more mice when it was given at 1400 h (middle of the rest period) than when it was given at 0200 h (middle of the activity period). Similar results were obtained with dibekacin and netilmicin (8, 24). Several other studies were done with lower doses of aminoglycosides and more sensitive parameters of toxicity such as urinary enzyme excretion (8, 10, 12, 25, 26, 32, 33). More recently, Fujimura et al. (13) reported that the magnitude of the changes induced by amikacin on creatinine clearance and tubular function was maximal when the drug was injected at 1600 h (rest period) compared with that when the drug was injected at other times of day. The mechanisms associated with the temporal variations in aminoglycoside nephrotoxicity are unknown, but circadian rhythms in aminoglycoside pharmacokinetics were observed in experimental animals. In fact, Yoshiyama et al. (32, 33) measured lower levels of isepamicin in plasma 30 min following a 7-day treatment given at 0100 h (middle of the activity period) compared with those at 1300 h (middle of the rest period). Hosokawa et al. (16) reported a circadian variation in amikacin clearance, with a maximal value obtained when the drug was injected at 2100 h (activity period) and a minimum value obtained at 0900 h (rest period) following a single injection of 50 mg/kg. Studies done in our laboratory showed a significantly lower area under the curve in serum and a higher clearance of tobramycin from serum when the drug was injected at 0200 h (middle of the activity period) than when it was injected at 1400 h (middle of the rest period) (18). These pharmacokinetic changes can be explained by the circadian rhythms observed in the diuresis and glomerular filtration rate. In fact, Pons et al. (29) reported maximum water excretion between 0000 and 0400 h and minimum excretion between 1600 and 2000 h. A circadian rhythm in inulin clearance was also detected, with a nightly acrophase measured between 0000 and 0400 h and a bathyphase observed between 1200 and 1600 h (29). Temporal changes in the pharmacokinetics of aminoglycosides were also reported in humans. Nakano et al. (23) measured a lower clearance, a longer serum half-life, and a higher area under the curve following gentamicin injection given at midnight compared with the values obtained after the same injection given at noon. The mean and the trough concentrations of netilmicin in plasma were increased in patients treated at 0500 and 0900 h, respectively, suggesting a circadian variation with possible accumulation during the night (20). The role of fasting in aminoglycoside nephrotoxicity has never been investigated. By contrast, several studies have examined the effects of high- and low-protein diets on aminoglycoside nephrotoxicity, but contradictory data were generated. Grauer et al. (15) showed that dogs fed a high-protein diet (27.3%) had higher creatinine clearances, lower serum creatinine levels, lower fractional clearances of sodium, lower levels of urinary excretion of N-acetyl- -D-glucosaminidase, and lower trough gentamicin concentrations in serum compared with the values for dogs fed medium-protein (13.7%) and low-protein (9.4%) diets. No differences in the levels of gentamicin in the renal cortex were found among the groups. Andrews and Bates (1) reported that rats conditioned to high levels of dietary protein (60%) but switched to low levels of dietary protein (5%) during gentamicin injection had less nephrotoxicosis than rats maintained on high- or low-protein diets before and after gentamicin treatment. However, another study showed that rats fed 5% protein diets during gentamicin administration had less nephrotoxicosis than those fed 18% protein diets (31). The differences in the results of the previous studies might be due to the use of different animal species, doses of gentamicin (30 mg/kg to 120 to 150 mg/kg), and injection times. Pattyn et al. (27) showed a lower level of accumulation of gentamicin in the renal cortex associated with proteinuria, suggesting that high levels of protein in urine

674 BEAUCHAMP ET AL. ANTIMICROB. AGENTS CHEMOTHER. FIG. 5. Subcellular localization of gentamicin in proximal tubular cells after a single injection of gentamicin (150 mg/kg, i.p.) given at either 1400 or 0200 h in normally fed and fasted rats. (A) Normally fed rats treated at 1400 h. (B) Normally fed rats treated at 0200 h. (C) Fasted rats treated at 1400 h. (D) Fasted rats treated at 0200 h. Gold particles were found mostly in the lysosomes of proximal tubular cells in all treated groups. There is a noticeable difference in the density of the gold labeling between normally fed animals treated at 1400 and 0200 h. This difference was absent in fasted rats. Magnifications: A and B, 26,400; C and D, 19,800.

VOL. 40, 1996 FASTING AND GENTAMICIN TOXICITY 675 interfere with gentamicin reabsorption into proximal tubular cells. Finally, Song et al. (30) concluded from a study done in mice submitted to a time-restricted feeding schedule that the feeding schedule can modify the rhythm of the toxicity of gentamicin by changing the rhythm of its kinetics. Nutritional status was recently recognized as a risk factor associated with aminoglycoside nephrotoxicity in patients. A prospective study done in 1993 in 1,489 patients showed a significant increase in gentamicin toxicity (rise in serum creatinine level) in patients with low initial serum albumin levels (6). A decrease in serum albumin levels in patients has also been associated with a greater volume of distribution (11, 34) and a longer serum half-life (11) of aminoglycoside, probably resulting in more accumulation in tissue (6). On the other hand, protein supplementation (90 mg) induced a significantly shorter elimination half-life and a higher clearance of gentamicin in healthy adult men (9). Healthy subjects ingesting a high-protein diet had a significantly higher creatinine clearance rate than a comparable group of healthy subjects ingesting a vegetarian diet (7). An increase in the glomerular filtration rate associated with feeding probably resulted in increased gentamicin clearance. In fact, our data showed that rats fed a normal diet and given gentamicin at 0200 h (maximal food intake period) had a significantly higher clearance rate of gentamicin compared with the rate for those treated at 1400 h (minimal food intake period) (18). These pharmacokinetic changes were associated with a significant modification of the gentamicin levels in the renal cortex (18). In the present study, gentamicin levels in tissue were significantly higher in normally fed rats treated at 1400 h than in those treated at 0200 h, as previously observed in our laboratory. By contrast, gentamicin levels in the renal cortex were similar in fasted rats treated at 1400 and 0200 h, reaching levels similar to those in normally fed rats treated at 1400 h. The mechanisms by which fasting increased gentamicin toxicity in the present study are unknown, but we cannot exclude the possibility that fasting was associated with a lower clearance of gentamicin, as suggested by the higher creatinine levels in serum compared with the levels in normally fed rats and the higher levels of gentamicin measured in the renal cortices of fasted rats treated at 1400 h compared with the levels in normally fed rats treated at 0200 h. A possible shift in the acrophase and bathyphase of gentamicin toxicity might also have been induced by fasting and might have not been detected in the present study, because data from only two injection times were evaluated. It is difficult to extrapolate to the human situation the results of the present study. Further investigations should be done to better understand the role of the nutritional status of patients on aminoglycoside pharmacokinetics and nephrotoxicity. A better understanding of the relationship between the diet, the nutritional status of patients, and the temporal variations in the nephrotoxicities of aminoglycosides associated with the once-daily dosing of these drugs might lead to the definition of innovative approaches of reducing significantly the nephrotoxicities of these drugs. ACKNOWLEDGMENTS This study was supported by The Kidney Foundation of Canada and by grant MA-12749 from the Medical Research Council of Canada. Pierre Collin was the recipient of a studentship from the PMAC Health Research Foundation. REFERENCES 1. Andrews, P. W., and S. B. Bates. 1987 88. Dietary protein as a risk factor in gentamicin nephrotoxicity. Renal Fail. 10:153 159. 2. Bates, R. D., and M. C. Nahata. 1994. Once-daily administration of aminoglycosides. Ann. Pharmacother. 28:757 765. 3. Beauchamp, D., P. Gourde, and M. G. Bergeron. 1991. Subcellular distribution of gentamicin in proximal tubular cells, determined by immunogold labeling. Antimicrob. Agents Chemother. 35:2173 2179. 4. Beauchamp, D., P. Gourde, M. Simard, and M. G. Bergeron. 1992. Subcellular localization of tobramycin and vancomycin given alone and in combination in proximal tubular cells, determined by immunogold labeling. Antimicrob. Agents Chemother. 36:2204 2210. 5. Beauchamp, D., P. Gourde, M. Simard, and M. G. Bergeron. 1994. Subcellular distribution of daptomycin given alone or with tobramycin in renal proximal tubular cells. Antimicrob. Agents Chemother. 38:189 194. 6. Bertino, J. S., L. A. Booker, P. A. Franck, P. L. Jenkins, K. R. Franck, and A. N. Nafziger. 1993. Incidence and significant risk factors for aminoglycoside-associated nephrotoxicity in patients dosed by using individualized pharmacokinetic monitoring. J. Infect. Dis. 167:173 179. 7. Bosch, J. P., A. Saccaggi, A. Lauer, C. Ronco, M. Belledonne, and S. Glabman. 1983. Renal functional reserve in humans. Effect of protein intake on glomerular filtration rate. Am. J. Med. 75:943 950. 8. Cambar, J., C. Dorian, and J. C. Cal. 1987. Chronobiologie et physiopathologie rénale. Pathol. Biol. 35:977 984. 9. Dickson, C. J., M. S. Schwartzman, and J. S. Bertino. 1986. Factors affecting aminoglycoside disposition: effects of circadian rhythm and dietary protein intake on gentamicin pharmacokinetics. Clin. Pharmacol. Ther. 39:325 328. 10. Dorian, C., P. Catroux, and J. Cambar. 1987. Étude de la chrononéphrotoxicité de l amikacine après intoxication chronique de 7 jours chez le rat. Pathol. Biol. 35:735 738. 11. Etzel, J. V., A. N. Nafziger, and J. S. Bertino. 1992. Variation in the pharmacokinetics of gentamicin and tobramycin in patients with pleural effusions and hypoalbuminemia. Antimicrob. Agents Chemother. 36:679 681. 12. Fauconneau, B., E. DeLemos, C. Pariat, S. Bouquet, P. Courtois, and A. Piriou. 1992. Chrononephrotoxicity in rats of a vancomycin and gentamicin combination. Pharmacol. Toxicol. 71:31 36. 13. Fujimura, A., T. Sudoh, and A. Ebihara. 1994. Time-dependent change in the toxic effects of amikacin on renal functions. Life Sci. 55:367 372. 14. Gilbert, D. N. 1991. Once-daily aminoglycoside therapy. Antimicrob. Agents Chemother. 35:399 405. 15. Grauer, G. F., D. S. Greco, E. N. Behrend, M. J. Fettman, R. S. Jaenke, and T. A. Allen. 1994. Effects of dietary protein conditioning on gentamicininduced nephrotoxicosis in healthy male dogs. Am. J. Vet. Res. 55:90 97. 16. Hosokawa, H., S. Nyu, K. Nakamura, K. Mifune, and S. Nakano. 1993. Circadian variation in amikacin clearance and its effects on efficacy and toxicity in mice with and without immunosuppression. Chronobiol. Int. 10:259 270. 17. Laurent, G., P. Maldague, M. B. Carlier, and P. M. Tulkens. 1983. Increased renal DNA synthesis in vivo after administration of low doses of gentamicin to rats. Antimicrob. Agents Chemother. 24:586 593. 18. Lin, L., L. Grenier, Y. Bergeron, M. Simard, M. G. Bergeron, G. Labrecque, and D. Beauchamp. 1994. Temporal changes of pharmacokinetics, nephrotoxicity, and subcellular distribution of tobramycin in rats. Antimicrob. Agents Chemother. 38:54 60. 19. Lin, L., L. Grenier, G. Thériault, P. Gourde, M. G. Bergeron, G. Labrecque, and D. Beauchamp. 1994. Chrononephrotoxicity of low doses of tobramycin in rats. Life Sci. 55:169 177. 20. Lutcht, F., S. Tigaud, G. Esposito, J. Cougnard, M. P. Fargier, and D. Peyramond. 1990. Chronokinetic study of netilmicin in man. Eur. J. Clin. Pharmacol. 39:199 201. 21. Maruhn, D. 1976. Rapid colorimetric assay of -galactosidase and N-acetyl- -glucosaminidase in human urine. Clin. Chim. Acta 73:453 461. 22. Nakano, S., and N. Ogawa. 1982. Chronotoxicity of gentamicin in mice. IRCS J. Med. Sci. 10:592 593. 23. Nakano, S., J. Song, and N. Ogawa. 1990. Chronopharmacokinetics of gentamicin: comparison between man and mice. Annu. Rev. Chronopharmacol. 7:277 280. 24. Pariat, C., J. Cambar, and P. Courtois. 1984. Circadian variations in the acute toxicity of three aminoglycosides: gentamicin, dibekacin, and netilmicin in mice. Annu. Rev. Chronopharmacol. 1:381 384. 25. Pariat, C., P. Courtois, J. Cambar, A. Piriou, and S. Bouquet. 1988. Variations in gentamicin nephrotoxicity in rats. Annu. Rev. Chronopharmacol. 5: 461 463. 26. Pariat, C., P. Courtois, J. Cambar, A. Piriou, and S. Bouquet. 1988. Circadian variations in the renal toxicology of gentamicin in rats. Toxicol. Lett. 40: 175 182. 27. Pattyn, V. M., G. A. Verpooten, R. A. Giuliano, F. Zheng, and M. E. DeBroe. 1988. Effect of hyperfiltration, proteinuria, and diabetes mellitus on the uptake kinetics of gentamicin in the kidney cortex of rats. J. Pharmacol. Exp. Ther. 244:694 698. 28. Persijn, J. P., and W. van der Slik. 1976. A new method for the determination of -glutamyltransferase in serum. J. Clin. Chem. Clin. Biochem. 14: 421 427. 29. Pons, M., J. Tranchot, B. L Azou, and J. Cambar. 1984. Circadian rhythms of renal hemodynamics in unanesthetized, unrestrained rats. Chronobiol. Int. 11:301 308. 30. Song, J., S. Ohdo, N. Ogawa, and S. Nakano. 1993. Influence of feeding

676 BEAUCHAMP ET AL. ANTIMICROB. AGENTS CHEMOTHER. schedule on chronopharmacological aspects of gentamicin in mice. Chronobiol. Int. 10:338 348. 31. Whiting, P. H., D. A. Power, and J. Peterson. 1988. The effect of dietary protein restriction on high dose gentamicin nephrotoxicity in rats. Br. J. Exp. Pathol. 69:35 41. 32. Yoshiyama, Y., T. Kobayashi, F. Tomonaga, and S. Nakano. 1992. Chronotoxical study of gentamicin induced nephrotoxicity in rats. J. Antibiot. 45:806 808. 33. Yoshihama, Y., S. Nishikawa, T. Sugiyama, T. Kobayashi, H. Shimada, F. Tomonaga, S. Ohdo, N. Ogawa, and S. Nakano. 1993. Influence of circadian stage-dependent dosing schedule on nephrotoxicity and pharmacokinetics of isepamicin in rats. Antimicrob. Agents Chemother. 37:2042 2043. 34. Zarowitz, B. J., A. M. Pilla, and J. Popovich. 1990. Expanded gentamicin volume of distribution in patients with indicators of malnutrition. Clin. Pharm. 9:40 44.