The sequence of research leading to a proposal for
anthracycline cross-linking of
DNA is presented.The clinical
anthracycline antitumor drugs are
anthraquinones, and as such are redox active. Their redoxchemistry leads to induction of oxidative stress and
drug metabolites. An intermediate in reductive glycosidiccleavage is a quinone methide, once proposed as an
alkylating agent of
DNA. Subsequent research nowimplicates
formaldehyde as a mediator of
anthracycline-
DNA cross-linking. The cross-link at 5'-GC-3'sites consists of a covalent linkage from the amino group of the
anthracycline to the 2-amino groupof the G-base through a methylene from
formaldehyde, hydrogen bonding from the 9-OH to the G-base onthe opposing strand, and hydrophobic interactions through intercalation of the
anthraquinone. The combinationof these interactions has been described as a virtual cross-linkof
DNA. The origin of the
formaldehyde in vivo remains a mystery. In vitro,
doxorubicin reacts withformaldehyde to give firstly a monomeric
oxazolidine,
doxazolidine, and secondly a dimeric
oxazolidine,
doxoform.
Doxorubicin reacts with
formaldehyde in the presence of
salicylamide to give the N-Mannich baseconjugate, doxsaliform. Doxsaliform is several fold more active in
tumor cell growth inhibition than
doxorubicin,but
doxazolidine and
doxoform are orders of magnitude more active than
doxorubicin. Exploratory researchon the potential for doxsaliform and
doxazolidine as targeted
cytotoxins is presented. A promisinglead design is
pentyl PABC-Doxaz, targeted to a carboxylesterase
enzyme overexpressed in liver cancercells and/or
colon cancer cells.