Design and activity of AP endonuclease-1 inhibitors

Abstract

Apurinic/apyrimidinic endonuclease-1/redox effector factor-1 (APE-1) is a critical component of base excision repair that excises abasic lesions created enzymatically by the action of DNA glycosylases on modified bases and non-enzymatically by hydrolytic depurination/depyrimidination of nucleobases. Many anticancer drugs generate DNA adducts that are processed by base excision repair, and tumor resistance is frequently associated with enhanced APE-1 expression. Accordingly, APE-1 is a potential therapeutic target to treat cancer. Using computational approaches and the high resolution structure of APE-1, we developed a 5-point pharmacophore model for APE-1 small molecule inhibitors. One of the nM APE-1 inhibitors (AJAY-4) that was identified based on this model exhibited an overall median growth inhibition (GI50) of 4.19 μM in the NCI-60 cell line panel. The mechanism of action is shown to be related to the buildup of abasic sites that cause PARP activation and PARP cleavage, and the activation of caspase-3 and caspase-7, which is consistent with cell death by apoptosis. In a drug combination growth inhibition screen conducted in 10 randomly selected NCI-60 cell lines and with 20 clinically used non-genotoxic anticancer drugs, a synergy was flagged in the SK-MEL-5 melanoma cell line exposed to combinations of vemurafenib, which targets melanoma cells with V600E mutated BRAF, and AJAY-4, our most potent APE-1 inhibitor. The synergy between AJAY-4 and vemurafenib was not observed in cell lines expressing wild-type B-Raf protein. This synergistic combination may provide a solution to the resistance that develops in tumors treated with B-Raf-targeting drugs.

Keywords: AP endonuclease; Abasic sites; DNA repair; Drug synergy; Toxicity.

Fig. 1

Structural details of APE-1. a The surface of crystal structure of apo-APE-1 (PDB/1BIX) [31], Arg177 was far away from Met270, so the potential binding pocket was open. b The surface of crystal structure of APE-1 bound with DNA (PDB/1DEW) [32], Arg177 and Met270 covered the potential binding pocket. c The interactions between APE-1 and abasic site fragment in the crystal structure of APE-1 bound with DNA (PDB/1DEW), several important residues including Tyr171, Asn174, Asn212, and His309 form strong ionic or H-bonding interactions with the negatively charged 5'-phosphate, a hydrophobic pocket surrounded by residues Phe266, Met270, Trp280, and Leu282. d Two-dimensional pharmacophore model hp3A2 with three hydrophobic or hydrophobic aromatic centers (hp) and two H-bond acceptors (A) was generated to represent APE-1 interactions with abasic DNA (PDB/1DEW). Distance restrictions of this pharmacophore model were listed in Table 1

Fig. 2

a Detailed interactions between APE-1 and AJAY-4, IC₅₀ 0.12 μM. b The pharmacophore model of AJAY-4. His309 and Arg177 formed strong hydrogen bonds with all our four compounds. Tyr171, Asn174, and Asn212 formed weak hydrophilic interactions with these compounds. Phe266 formed strong π-π interaction with all compounds. Trp280 mainly formed hydrophobic interactions with our inhibitors. Detailed interactions of our other compounds can be found in Fig. 2S

Fig. 3

Docking studies between APE-1 and active compounds with Arg177Ala mutation. a Compound 6-hydroxy-DL-DOPA, IC₅₀ 0.11 μM [22]. b AJAY-4, IC₅₀ 0.12 µM. Both the previously reported inhibitor and our inhibitor induced large conformational changes when interacting with Arg177Ala mutated. We suggest that Arg177 plays an important role for the recognition of the inhibitors. The conformations of compounds highlighted in pale-green were after Arg177Ala mutation, while the conformations highlighted in orange were before mutation

Fig. 4

MD simulations results for AJAY-4. a The alignments of AJAY-4 between before MD (orange) and after MD (blue). b The distances between AJAY-4 and APE-1, the H-bond distances between active compound AJAY-4 and Arg177/His309 remained stable. We also performed the MD simulations for inhibitor AJAY-15, as shown in Fig. 3S. However, the distances between inactive compound AJAY-15 and Arg177/His309 fluctuated greatly. More detail can be found in Fig. 3S. We suggest that Arg177 and His309 play key roles in the recognitions of inhibitors

Fig. 5

Growth inhibition by the APE-1 inhibitor AJAY-4 in the NCI 60 cell lines. Cells from each of the NCI 60 cell lines were harvested, counted, and seeded into the wells of T0 and T72 384-well assay plates at seeding densities that allowed for continuous proliferation throughout 96 h of incubation at 37 °C, 5 % CO₂, and 95 % humidity. After 24 h in culture, Cell Titer Glo detection reagent was added to the wells of the T0 assay plate and the cellular ATP-dependent luminescent signal was captured on an M5e micotiter plate reader. Also, after 24 h, ALAY-4 that had been serially diluted in tissue culture medium to provide a 5-point 10-fold dilution series starting at maximum concentration of 10 μM (final in well) was transferred into the wells of the T72 assay plate that was then returned to the incubator. After an additional 72 h of incubation, Cell Titer Glo detection reagent was added to the wells; the T72 assay plate and the cellular ATP-dependent luminescent signal was captured on an M5e micotiter plate reader. The percent of growth in the compound-treated wells was normalized relative to the growth observed in the corresponding T0 and T72 assay plate control wells, and the data was fit to curves using the Sigmoidal dose response variable slope equation Y = Bottom + [Top − Bottom]/[1 + 10^{(LogEC50 − X)} × HillSlope] of the GraphPad Prism 5 software. The AJAY-4 concentration that inhibited the growth of each of the 60 cells lines by 50 % (GI₅₀) is presented

Fig. 6

Mechanism of cytotoxicity associated with inhibition of APE-1 in T98G cells exposed to various combinations of AJAY-4 and MeLex for 2, 6, 12, 18, 24, and 26 h and assayed for a caspase-3 and caspase-7 activation; b PARP activation; c PARP cleavage: white diamond, untreated; black diamond, 0.2 % DMSO; green square, 80 μM MeLex + 0.25 μM AJAY-4; blue circle, 0.96 μM AJAY-4 (GI₅₀); open circle, 0.25 μM AJAY-4 (LD₁₀); red solid triangle, 250 μM MeLex (LD₅₀); red open triangle, 80 μM MeLex (LD₁₀); black vertical bar, 20 µM etoposide (used in PARP activation and PARP cleavage experiments

Fig. 7

Pilot screen-flagged synergy for cytotoxicity of the APE-1 inhibitor in combination with vemurafenib in SK-MEL-5 melanoma cells carrying a V600E mutation in BRAF. In a drug combination matrix screening assay using 10 randomly selected cell lines from the NCI 60 panel, we assayed the effect of APE1 inhibitor AJAY-4 on the growth inhibition responses of 20 recently approved non-genotoxic anticancer drugs. In this pilot screen, an apparent synergistic (>additive) growth inhibition response was identified in the SK-MEL-5 melanoma line for the combination of AJAY-4 with vemurafenib (VMFB). Individual compound controls run in singlet were included in each drug combination matrix and compared to replicate (n = 10) controls for VMFB and AJAY-4. The combination of the two compounds at different concentrations are shown: green (synergistic) means that the combination of compounds enhanced toxicity by at least three standard deviations from the combined mean of the individual compounds1. A1, A2, and A3 = 2, 0.2, and 0.02 μM AJAY-4 and B1, B2, and B3 = 5, 0.5, and 0.05 μM VMFB, respectively. The percent of growth in the compound-treated wells was normalized relative to the growth observed in the corresponding T0 and T72 assay plate control wells that was determined using the Cell Titer Glo cellular ATP detection reagent as described above

Fig. 8

Individual GI₅₀ determinations for a vemurafenib and b AJAY-4 in the SK-MEL-5 (BRAF-V600E) and SK-MEL-2 (WT BRAF) melanoma cell lines. SK-MEL-5 and SK-MEL-2 cells were exposed to the indicated concentrations of AJAY-4 or VMFB for 72 h and the corresponding Cell Titer Glo (ATP content) signals relative to those of DMSO controls were used to determine their respective GI₅₀s. The percent of growth in the compound-treated wells was normalized relative to the growth observed in the T72 assay plate DMSO control wells and the data was fit to curves using the Sigmoidal dose response variable slope equation Y = Bottom + [Top − Bottom]/[1 + 10^{(LogEC50 − X)} × HillSlope] of the GraphPad Prism 5 software. The mean ± sd (n = 3) growth inhibition data from triplicate wells for each concentration of AJAY-4 are presented as the percent of the DMSO plate controls; SK-MEL-5 (black circle) and SK-MEL-2 (white circle). Representative experimental data from one of three independent experiments are shown

Fig. 9

Drug combination index analysis for selected vemurafenib and AJAY-4 ratios in the SK-MEL-5 (BRAF-V600E) and SK-MEL-2 (WT BRAF) melanoma cell lines. For drug combination testing, SK-MEL-5 and SK-MEL-2 cells were exposed to the indicated drug combination ratios of AJAY-4 plus VMFB for 72 h, and the corresponding Cell Titer Glo signals relative to those of cells exposed to the individual drugs at the same concentrations were used to determine the fraction of cells affected/killed (Fa) and to calculate a combination index (CI) using the Compusyn software; SK-MEL-5 (black circle) and SK-MEL-2 (white circle). Fa = 1/(1 + (D_m/D)m) where D = dose, D _m = median-effect dose, and m = slope, hill-type coefficient, and CI = (D comb)1/(D alone)1 + (D comb)2/(D alone)2; CI = 1 indicates summation, CI > 1 indicates antagonism, and CI < 1 indicates synergism. Representative experimental data from one of three independent experiments are shown