This shortened group of hits was tested in the cell-based potentiation assay for their ability to dosedependently enhance sensitivity of HeLa cells to the alkylating agent MMS. Of the 170 compounds tested, 12 enhanced the cytotoxicity of MMS significantly (the difference between the compound-plus-MMS and compound-alone responses was at least 30% and was maintained within at least a two-fold compound concentration range, Figure 5), while 16 compounds produced a modest response (the separation between the curves was greater than the errors of measurement but was less than 30% viability at any given compound concentration, Figure 6). The structures and assay data associated with the 12 compounds that exhibited significant effect in the MMS potentiation assay are presented in Figure 7.
AP Site Accumulation
As stated above, the application of the genotoxic agent MMS results in an increase in the level of AP sites within the genome. An effective APE1-targeted inhibitor applied under the conditions of increased genotoxic stress is expected to interfere with the repair of AP sites, while ideally having no effect on the AP site levels when applied as a single agent. Thus, the net effect of treating cells with a combination of MMS and inhibitor would be an increase in the number of AP site lesions. To determine the effect on AP site accumulation caused by treatment with MMS in combination with an APE1 inhibitor, as compared to treatment with MMS or inhibitor alone, we measured AP site levels in the chromosomal DNA from HeLa cells using an aldehyde-reactive probe-based colorimetric assay [40]. The concentrations of MMS and inhibitor (275 mM for MMS and between 5 and 30 mM for the inhibitor, respectively) were selected to fall below the onset of cytotoxicity for each agent based on the analysis of the MMS potentiation data shown in Figure 5. Of the 12 compounds that significantly enhanced the cytotoxicity of MMS (Figure 3), 11 were tested in the AP site accumulation experiments (MLS001217741 was not tested due to resupply shortage). These experiments revealed that in general the number of AP sites was only slightly increased in the cells exposed to the compounds alone relative to the vehicle control, but was significantly increased when cells were exposed to both MMS and most of the tested compounds (Figure 8). The exceptions to this trend were inhibitors where a noticeable increase in AP sites was observed upon treatment with the inhibitor alone (MLS001306480 (P = 0.0198), MLS001196838 (P = 0.0238), MLS001076862 (P = 0.0234)). Thus, compounds whose application as single agents resulted in an increase inAP sites may be of lesser utility as drug candidates, due to their potential to induce genotoxic damage by themselves. We note that because the AP site experiments are laborious in nature, and consistent with the need to provide initial profiling on a relatively large number of screening hits, the number of replicates incorporated in this experiment (two), as well as the combinations of inhibitor and MMS concentrations, were insufficient for a complete statistical analysis of the trends. In order for a true Combination Index to be derived [41], which in turn would allow for a definitive conclusion to be draw regarding the potential synergistic nature of the inhibitor effect on the MMS genotoxicity, a large matrix of inhibitor and MMS concentrations needs to be tested, making it impossible to provide such an exhaustive characterization for all top hits reported here. We further note that given the laborious nature of the AP site experiments, we have limited this initial study to only the 24-hour treatment condition: the longer-term effect (.3 days of treatment) of APE1 inhibitors on AP site accumulation is an important factor to consider during the development of one or more of the hits, particularly given the previously-highlighted vital role of APE1 as a resolver of abasic DNA damage accrued through natural causes [12]. With the exception of MLS001306480, an antimalarial pyronaridine used primarily in China [42], which is a quinacrine-like molecule that bears distant resemblance to the previously published APE1 inhibitors lucanthone [43], mitoxanthrone [21], and Reactive Blue 2 [21], none of the hit compounds that caused an increase in AP sites in cells under genotoxic stress were similar to previously reported APE1 inhibitors, including AR03 [23] or the pharmacophore model advanced by Zawahir and colleagues [22]. This overall result provides a support to the notion that screening of additional libraries to find novel APE1 inhibitory scaffolds is indeed a viable approach. Of the top hits, substituted isoxazole alkylamines like MLS001105846 have been patented as agri-horticultural fungicides [44]. Although outside this current profiling work, the MLS001105846 compound does not appear to have been tested in conjunction with human disease applications. MLS000419194 possesses no obvious liabilities and was recently taken through a medicinal chemistry optimization campaign by us [45]. The remaining hits that potentiate the MMS response and cause AP site accumulation belong to structural classes typically associated with promiscuity (that is, prone to bind to multiple protein targets in the cell) or carry potentially reactive or labile functional groups. The flavonoid derivative MLS000737267, also known as galloflavin (a product derived from gallic acid oxidation), has been shown to inhibit human immunodeficiency virus integrase at a low-micromolar IC50 [46]. Phenanthrenes derived from orchid plant species, bearing similar chemical features to the trimethoxyphenanthrene-diol hit MLS000863573, have recently been highlighted for their anti-inflammatory activity, presumably by inhibiting the lipopolysacharide-induced nitric oxide production in murine macrophages [47]. Both MLS000737267 and MLS000863573 are undesirable from the standpoint of further medicinal chemistry optimization because of the polyphenolic nature. MLS000594187, which bears a resemblance to the above two hits, in that it has a core consisting of three fused 6-membered rings in a phenanthrene-like configuration, belongs to a different chemical class, benzo(c)chromen-6-ones, which has been advanced as selective estrogen receptor b modulators [48]. MLS001196838 was recently characterized as a potent and selective inhibitor of the histone acetyltransferase (HAT) p300, and has been employed to probe HAT’s role in acute myelogenous leukemia [49,50]. A derivative of MLS001076862 has been reported as a modulator of the survival protein MCL-1, and displays cellular activity and thus potential utility for treating hyperproliferative, inflammatory, and other disorders [51]. Of note, MLS001196838 and MLS001076862 contain an exocyclic enone functionality, which may be reactive and would need to be modified prior to an extended optimization of these scaffolds. The pyrimidine-2,4,6trione central core unit within MLS000587064 derives from barbituric acid, although this compound itself does not belong to the group of typical barbiturates. Most recently, pyrimidine-2,4,6trione derivatives were shown to be highly effective in a protein aggregation protection assay model of amyotrophic lateral sclerosis and to possess good bioavailability [52]. Lastly, the furanyltetrahydromethyl-indole MLS000090966 was recently reported to be an effective inhibitor of tubulin assembly [53]. MLS000587064 and MLS000090966 contain a catechol diether moiety, a recognized chemical liability [54]. Taken together, our analyses have revealed a promising set of structurally diverse heretofore unreported APE1 inhibitors. As with all typical high-throughput screening and follow-up campaigns, it is plausible that additional APE1-inhibitory chemotypes of drug-like characteristics, which reside in our collection, may
have been missed due to the application of the series of stringent selection criteria in order to narrow down the list of hits being followed up. At present, it is hoped that the public availability of all screening and secondary assay data will lead to further exploration of this information-rich resource.
