A novel macromolecular prodrug concept exploiting endogenous serum albumin as a drug carrier for cancer chemotherapy.

Introduction. Serum proteins are potential drug carriers of antineoplastic agents due to their accumulation in tumor tissue.1 Uptake of these proteins in solid tumors is mediated by a number of factors including an increased metabolic activity of tumors, an enhanced vascular permeability of tumor blood vessels for circulating macromolecules, and a lack of a functional lymphatic drainage system in tumor tissue.2 Recently, a number of acid-sensitive albumin and transferrin conjugates with anthracyclines and the alkylating agent chlorambucil have shown promising in vitro activity.3-7 In addition, acid-sensitive doxorubicin conjugates with monoclonal antibodies and albumin doxorubicin conjugates prepared by glutaraldehyde cross-linking have shown promising antitumor efficacy in vivo.8,9 A selected acid-sensitive doxorubicin albumin conjugate that was developed in our group induced complete remissions of primary kidney tumors in murine renal carcinoma and prevented the formation of metastases in the lungs. In contrast, mice treated with doxorubicin at optimal dose manifested clearly visible kidney tumors at the end of the experiment and large numbers of lung metastases.10 This albumin doxorubicin conjugate was synthesized by coupling 1, a maleimide carboxylic hydrazone derivative of doxorubicin (see Figure 1), to thiolated albumin. 1 contains an acid-sensitive linker that allows the drug to be released at the low pH values present in lysosomes and endosomes of tumor cells. We have recently shown that 1 also binds covalently to the cysteine-34 of commercially available human serum albumin which is a mixture of mercaptalbumin and nonmercaptalbumin.11,12 Approximately 70% of circulating albumin in the blood stream is mercaptalbumin that contains an accessible cysteine-34 which is not blocked by endogenous sulfhydryl compounds such as cysteine or glutathione.12,13 Considering that free thiol groups are not found on the majority of circulating serum proteins except for albumin, cysteine-34 of endogenous albumin is a fairly unique amino acid on the surface of a circulating protein that could be exploited for developing a novel macromolecular prodrug concept. Since the maleimide group reacts specifically and selectively with thiol groups, we reasoned that it should be possible to preferentially bind maleimide drug derivatives to the HS group of the cysteine-34 position of albumin after intravenous application. In this way a macromolecular prodrug is formed after in situ coupling of a thiolbinding drug derivative to endogenous albumin in the blood circulation. Following this approach, it should be possible to avoid the ex vivo synthesis and characterization of drug albumin conjugates which are costly, are time-consuming, and rely on exogenous and possibly pathogenic albumin. The objective of the present work was to assess the feasibility and selectivity of our approach by carrying out in vitro and in vivo binding studies using the doxorubicin maleimide derivative 1. In addition, we wanted to obtain the first in vivo evidence that 1 is superior to free doxorubicin in an animal tumor model, i.e., in murine renal cell carcinoma (RENCA). In Vitro and in Vivo Binding Studies. To estimate the coupling rate and selectivity of 1 for endogenous albumin, 1 was incubated with human blood plasma at T ) 37 °C and the samples were subsequently analyzed by anion-exchange chromatography. (The ratio of drug to albumin was approximately 0.3:1; human blood plasma employed contained an albumin concentration of ∼700 μM as determined with a Vitros analyzer from Ortho-Clinical Diagnostics. 1 was synthesized previously;14 a 2000 μM solution of 1 was freshly prepared in 0.15 M NaCl, 0.004 M sodium phosphate buffer (pH 5.5), and 30 vol % of 1,2-propylene glycol for binding studies and animal experiments; doxorubicin was used as a 3400 μM stock solution in isotonic saline from Pharmacia & Upjohn.) Chromatograms after an incubation time of 5 min for 1 (Figure 2A) and for free doxorubicin (Figure 2B) under identical conditions are available as Supporting Information. Protein components were detected at 254 nm, and the anthracycline moiety was detected by simultaneous fluorescence excitation. In contrast to free doxorubicin, the major amount of 1 is associated with the albumin peak which elutes at around 23.5 min. Longer incubation periods led to a slight increase in albumin binding indicating that the coupling reaction proceeds rapidly within this short period. Incubation studies with total blood and subsequent HPLC analysis of the resulting blood plasma * To whom correspondence should be addressed: Dr. Felix Kratz, Tumor Biology Center, Department of Medical Oncology, Clinical Research, Breisacher Strasse 117, D-79106 Freiburg, FRG. Tel: 0049-761-2062176. Fax: 0049-761-2061899. E-mail: felix@ tumorbio.uni-freiburg.de. Figure 1. Structure of 1, a maleimide phenylacetylhydrazone derivative of doxorubicin. 1253 J. Med. Chem. 2000, 43, 1253-1256