Mitoxantrone, 1,4-dihydroxy-5,8-bis[[2-[(2-hydroxyethyl)amino]ethyl]amino]-9,10-anthracenedione (Fig. 1 A), is an approved chemotherapeutic drug, with demonstrated efficacy in treating breast and prostate cancers, lymphomas, and leukemias [1]. In addition to applications in cancer treatment, mitoxantrone is also approved in immunotherapy for multiple sclerosis [2]. The therapeutic efficacy of mitoxantrone has been attributed to its function as a topoisomerase II poison, by stabilizing the enzyme-DNA reaction intermediate following double-strand break formation, thus preventing DNA re-ligation [3], [4], [5].

However, the clinical use of mitoxantrone is limited by side effects that include the development of life-threatening heart conditions [6], [7], [8]. The molecular mechanisms responsible for its toxicity in both cardiomyocytes and other cells have not been fully elucidated, in part because of its complex interactions with biomolecules and various cellular structural components [1]. In this regard, the binding of mitoxantrone to DNA has been extensively studied, with the consensus model showing interactions of the anthraquinone moiety with DNA bases through π-π stacking and the hydroxyethylaminoethyl-amino tails with the sugar phosphate backbone through hydrogen bonding [5], [9], [10], [11], [12], [13], [14], [15], [16]. In addition, it has recently been shown that mitoxantrone forms a Schiff base conjugate with apurinic/apyrimidinic (AP) sites in DNA [17]. Given that AP sites (Fig. 1B) are continuously generated within cells by a variety of mechanisms including spontaneous loss of both normal and alkylated bases [18], [19], [20], [21], [22], [23] as well as the glycosylase activities within base excision repair (BER) [24], [25], such interactions have the potential to modulate biological processes. Although rates of AP site formation via depurination are estimated to be as high as 12,000/cell generation [18], efficient repair processes considerably limit their intracellular levels [25], [26]. Highlighting the importance of these repair mechanisms, recent steady-state measurements of the number of AP sites in tissues estimate only 1-3 AP sites per 107 nucleotides [21], [22], [23].

Germane to the efficiency of the AP site repair, prior investigations identified mitoxantrone as a potential inhibitor of human AP endonuclease 1, APE1 [27]. APE1 is a key component of BER [26]. APE1 cleaves the DNA phosphodiester backbone on the 5′ side of an AP site, generating a nick in the DNA with 3′ hydroxyl and 5′ deoxyribose 5-phosphate (5′ dRP) termini. Inhibition of APE1 endonuclease activity by mitoxantrone was originally identified through a screen of the Library of Pharmacologically Active Compounds (LOPAC1280) [27]. The experimental design utilized a fluorescent-based strategy for the detection of APE1 activity on synthetic oligodeoxynucleotides containing a tetrahydrofuran (THF) moiety (Fig. 1B). Although mitoxantrone showed a favorable IC50 of 2.0 μM, it also showed a comparable IC50 for a functionally similar, but structurally distinct AP endonuclease, Escherichia coli endonuclease IV, with the inhibition attributed to non-specific mitoxantrone-DNA interactions [27].

Here, we address if the biological fate of AP sites in mitoxantrone-treated cells may be modulated by its capacity to form AP site conjugates and an ability to inhibit APE1 activity within the BER pathway. Given these potentially competing roles for mitoxantrone in the processing and subsequent repair of AP sites, we investigated the mechanisms of mitoxantrone interactions with AP site-containing DNAs. Our experimental design measured the inhibition of APE1 using fluorescently labelled site-specifically modified oligodeoxynucleotides. In addition to inhibition of nuclease activity of APE1 via specific DNA binding, these in vitro studies also revealed the ability of mitoxantrone to nick DNA at AP sites. The biological significance of these properties of mitoxantrone was examined using cellular toxicity as an outcome.