[CANCER RESEARCH 61, 8194 – 8202, November 15, 2001] Molecular Basis for the Synergistic Interaction of Adriamycin with the Formaldehyde-releasing Prodrug Pivaloyloxymethyl Butyrate (AN-9) 1 Suzanne M. Cutts, Ada Rephaeli, Abraham Nudelman, Inesa Hmelnitsky, and Don R. Phillips 2 Department of Biochemistry, La Trobe University, Victoria 3086, Australia [S. M. C., D. R. P.]; Felsenstein Medical Research Center, Sackler School of Medicine, Tel Aviv University, Beilinson Campus, Petach Tikva 49100, Israel [A. R.]; and Chemistry Department, Bar Ilan University, Ramat Gan 52900, Israel [A. N., I. H.] ABSTRACT The interaction of Adriamycin and pivaloyloxymethyl butyrate (AN-9) was investigated in IMR-32 neuroblastoma and MCF-7 breast adenocar- cinoma cells. Adriamycin is a widely used anticancer drug, whereas AN-9 is an anticancer agent presently undergoing Phase II clinical trials. The anticancer activity of AN-9 has been attributed to its ability to act as a butyric acid prodrug, although it also releases formaldehyde and pivalic acid. Adriamycin and AN-9 in combination display synergy when exposed simultaneously to cells or when AN-9 treatment is up to 18 h after Adriamycin administration. However, the reverse order of addition re- sults in antagonism. These interactions have been established using cell viability assays and classical isobologram analysis. To understand the molecular basis of this synergy, the relative levels of Adriamycin-DNA adducts were determined using various treatment combinations. Levels of Adriamycin-DNA adducts were enhanced when treatment combinations known to be synergistic were used and were diminished using those treatments known to be antagonistic. The relative timing of the addition of Adriamycin and AN-9 was critical, with a 20-fold enhancement of Adria- mycin-DNA adducts occurring when AN-9 was administered 2 h after the exposure of cells to Adriamycin. The enhanced levels of these adducts and the accompanying decreased cell viability were directly related to the esterase-dependent release of formaldehyde from AN-9, providing evi- dence for the formaldehyde-mediated activation of Adriamycin. INTRODUCTION Adriamycin is a widely used drug in current chemotherapy regimes because it is effective against a broad range of neoplasms. It is used as a single agent but is more commonly used in combinations with other anticancer agents. The selection of these additional agents is not usually based on known synergistic interactions between the drugs but rather on complimenting mechanisms of action. The major drawbacks associated with the use of Adriamycin are its dose-dependent cardio- toxicity and the emergence of tumor resistance to the drug (1). Although Adriamycin is a known topoisomerase II inhibitor, this mechanism of action does not fully explain its broad-spectrum anti- cancer activity (1, 2). In recent years, it has been shown that Adria- mycin induces adducts with DNA, and these occur predominantly at 5'-GC sequences (3, 4). Chemical characterization of this structure has revealed that the 3' aminosugar of Adriamycin is covalently bound to the N2 of guanine via a formaldehyde-derived bridge (5, 6). Two-dimensional NMR 3 analysis of the structure showed that adducts at GC sequences are also virtual cross-links, because the Adriamycin monoadduct is stabilized by the complementary strand of DNA by intercalation and H-bonding (7). This structure of the virtual cross- link explains why the apparent Adriamycin cross-links are unstable. DNA cross-link formation by various anthracycline derivatives (in- cluding Adriamycin) has been correlated with cytotoxicity in HeLa cells (8), and more recently in MCF-7 cells, at sufficiently high levels to account for the cytotoxic response (9). A new drug, doxoform, has been designed recently to take advan- tage of the fact that Adriamycin can be activated by formaldehyde (10). This complex of Adriamycin with formaldehyde is dramatically (200-fold) more cytotoxic than Adriamycin, and this appears to be attributable to enhanced formation of DNA adducts. BA is an agent that induces differentiation primarily because of its ability to function as a histone deacetylase inhibitor (11). In human tumor cells in vitro, it displays growth arrest, decreased clonogenicity, and induction of morphological and biochemical changes resulting in antitu- mor activity (12, 13). However, BA is not clinically effective because of rapid metabolism and, to a lesser extent, excretion (14). To achieve a reduction in the clearance rate of BA, a panel of BA-releasing prodrugs were synthesized and screened for antitumor activity (15, 16). AN-9 is the best studied prodrug, and it affects cancer cells at 10-fold lower concentrations and at least 100-fold faster than BA. Moreover, it pene- trates 100-fold faster than BA into cancer cells in vitro (17). Derivatiza- tion of BA improves its permeability across cell membranes and enables efficient intracellular delivery of BA. AN-9 belongs to a well-established family of acyloxyalkyl ester prodrugs of carboxylic acids (18 –20) whose expected esterase-depen- dent intracellular hydrolytic degradation products are BA, pivalic acid, and formaldehyde (Fig. 1). Whereas pivalic acid does not con- tribute to the activity elicited by the prodrug, the role of the released formaldehyde remains unclear, and it also cannot be excluded that the intact AN-9 has some intrinsic activity. The pivaloyloxymethyl de- rivatives of propionic, valeric, and pivalic acids (analogues of AN-9 that lack a BA fragment) were found to have significantly lower antitumor activity in cancer cells (16). This suggests that the biolog- ical activity of AN-9 stems mostly from the released BA moiety. AN-9 was shown to inhibit the proliferation of a variety of cancer cell lines and primary human tumors (15, 16, 21). AN-9 displayed low toxicity in mice and was effective in prolonging survival of mice bearing melanoma, lung carcinoma, and monocytic leukemia (15, 16, 22). It induced transient hyperacetylation of histones (23), leading to relaxation of the chromatin structure, which allowed access of tran- scription factors to the DNA (24). This activity is likely to be an important mechanism by which AN-9 exerts its effect on gene mod- ulation. AN-9 modulates the expression of the early regulatory genes c-myc and c-jun and the tumor suppressor gene RB as well as the antiapoptotic gene bcl-2 in WEHI and HL-60 cells (20, 25–27). AN-9 induces differentiation and/or apoptosis depending on the concentra- tions and timing of the drug used (27). AN-9 formulated in lipid emulsion (PIVANEX), displayed low toxicity in a Phase I clinical study and was reported to have an estimated maximum tolerated dose of 2.7 g/m 2 /day (28). It is presently in Phase II clinical trials with non-small cell lung carcinoma and hepatoma patients. Synergistic effects between AN-9 and DNA-disrupting agents have been observed in murine monocytic leukemia cells. Furthermore, it 1 This work was carried out with the support of the Australian Research Council (to S. M. C. and D. R. P.), Grant 542/0 from the Israel Science Foundation, a project grant from the Israel Cancer Research Fund (to A. R. and A. N.), and the Marcus Center for Pharmaceutical and Medicinal Chemistry and the Bronia and Samuel Hacker Fund for Scientific Instrumentation at Bar Ilan University. Received 2/22/01; accepted 9/19/01. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 2 To whom requests for reprints should be addressed, at Department of Biochemistry, La Trobe University, Victoria 3086, Australia. 3 The abbreviations used are: NMR, nuclear magnetic resonance; BA, butyric acid; AN-9, pivaloyloxymethyl butyrate; AN-158, 1-pivaloyloxyethyl butyrate; DHFR, dihy- drofolate reductase; CI, combination index. 8194 Research. on October 10, 2015. © 2001 American Association for Cancer cancerres.aacrjournals.org Downloaded from