To evaluate a method for preventing the nephrotoxicity caused by the high renal accumulation of radiolabeled or toxin-conjugated small
immunoproteins used for
cancer therapy, we conjugated humanized anti-Tac
Fab fragments with various numbers of
glycolate molecules [glycolated
Fab fragments (glyco-Fabs)] and separated the conjugates by means of ion-exchange columns into three fractions, depending on their isoelectric points (pIs). We evaluated the biodistribution, pharmacokinetics, and catabolism in normal nude mice of nonglycolated Fab (pI > or = 9.3) and three different preparations of glyco-Fab, including strongly anionic glyco-Fab (sa-glyco-Fab: pI < or = 4.5), mildly anionic glyco-Fab (pI = 4.5-7), and mildly cationic glyco-Fab (pI = 7-9.3). In addition, the biodistributions of 125I-labeled sa-glyco-Fab and 131I-labeled nonglycolated Fab were evaluated in normal nude mice coinjected with 50 mg of
L-lysine and/or 1 microg of
furosemide and in a control group without coinjection. We then evaluated the serial biodistribution of 125I-labeled sa-glyco-Fab (4 microCi/1 microg) and 131I-labeled nonglycolated Fab (5 microCi/1 microg) in Tac
antigen-positive (ATAC4) and -negative (A431)
tumor-bearing nude mice with s.c.
tumor xenografts derived from Tac
antigen-positive ATAC4 cells and receptor-negative A431 cells. These animals were coinjected with 30 mg of
lysine i.v. and 30 mg of
lysine i.p. 15 min after the radiolabeled Fab injection. To evaluate the biodistribution data and study scintigraphic imaging, we performed serial scintigraphy on normal and
tumor-bearing mice with all four 131I-labeled preparations. 125I-labeled mildly cationic glyco-Fab and 131I-labeled nonglycolated Fab had similar distributions, except in the kidney. However, both 125I-labeled anionic glyco-Fab preparations showed significantly different distributions from both cationic Fabs in the blood, liver, lung, and spleen. Renal accumulation of all four radiolabeled Fab preparations increased significantly as the pI increased (P < 0.01). In addition, the intact fraction of Fab excreted into urine increased as pI decreased. Therefore, the glomerular filtration depended on whether the charge on the Fab was positive or negative. The proportion of Fab reabsorbed by the proximal tubules increased as pI increased. 125I-labeled sa-glyco-Fab and 125I-labeled mildly anionic glyco-Fab showed a similar distribution in the blood and all organs except the kidney.
Lysine led to an additional blocking effect on proximal tubular uptake of both sa-glyco-Fab and nonglycolated Fab. Addition of
furosemide yielded only a small effect when used with
lysine. With
lysine, the sa-glyco-Fab:nonglycolated Fab estimated integral radioactivity ratios were 4.7 and 0.7 in the ATAC4
tumor and in the kidney, respectively. The use of anionic fragments, which may be used in conjunction with
lysine, represents a promising approach that may help decrease the renal toxicity of other small fragments, the molecular weights of which range from Mr 40,000 to 70,000, and, thereby, allow higher doses of radiation to the
tumor.