Abstract
The ability of in vitro models to predict nephrotoxicity was studied by comparing in vivo and in vitro toxicities of parenteral antibiotics. Rabbits were treated with a single dose of carbapenem antibiotics. After 24 h, the toxicity ranking was imipenem > BMY-25174 > BMY-26225 based on blood urea nitrogen (BUN) and histopathology; previous studies showed the cephalosporin ranking to be cephaloridine > cefazolin > cephalothin. In vitro models used were from rabbit renal cortex and included primary proximal tubule cultures (RPTC), the cell line LLC-RK1, and cortical slices. Cell viability was determined after 24-h exposure of RPTC and LLC-RK1 monolayers to the drugs; in some cases the S9 fraction of cortical homogenate was added concomitantly. Based on the 50% effective concentration (EC50s), the toxicity ranking for cephalosporins in both culture models was cephalothin > cephaloridine > cefazolin > cephalothin + S9. The ranking of carbapenems in RPTC was imipenem > BMY-25174 > BMY-26225, but for LLC-RK1 was BMY-26225 > imipenem > BMY-25174 > BMY-26225 + S9. Slices were incubated with the drugs for 90 min, and effects on organic anion paraaminohippurate (PAH) and cation tetraethylammonium bromide (TEA) transport and oxygen consumption rate (Qo2) assessed. The drugs reduced PAH uptake with a ranking of cefazolin > cephalothin > cephaloridine, and imipenem > BMY-25174 > BMY-26225. Slice TEA accumulation was depressed in the order cephaloridine > cefazolin > cephalothin, and imipenem > BMY-25174 = BMY-26225. Reduction of Qo2 accurately reflected in vivo nephrotoxicity. Among the models tested, RPTC viability and slice respiration provided the most valuable nephrotoxicity screens for cephalosporin or carbapenem antibiotics.