Synthetic lethal targeting of DNA double strand break repair deficient cells by human apurinic/apyrimidinic endonuclease (APE1) inhibitors

Abstract

An apurinic/apyrimidinic (AP) site is an obligatory cytotoxic intermediate in DNA Base Excision Repair (BER) that is processed by human AP endonuclease 1 (APE1). APE1 is essential for BER and an emerging drug target in cancer. We have isolated novel small molecule inhibitors of APE1. In the current study we have investigated the ability of APE1 inhibitors to induce synthetic lethality in a panel of DNA double strand break (DSB) repair deficient and proficient cells; a) Chinese hamster (CH) cells: BRCA2 deficient (V-C8), ATM deficient (V-E5), wild type (V79) and BRCA2 revertant (V-C8(Rev1)). b) Human cancer cells: BRCA1 deficient (MDA-MB-436), BRCA1 proficient (MCF-7), BRCA2 deficient (CAPAN-1 and HeLa SilenciX cells), BRCA2 proficient (PANC1 and control SilenciX cells). We also tested synthetic lethality (SL) in CH ovary cells expressing a dominant–negative form of APE1 (E8 cells) using ATM inhibitors and DNA-PKcs inhibitors (DSB inhibitors). APE1 inhibitors are synthetically lethal in BRCA and ATM deficient cells. APE1 inhibition resulted in accumulation of DNA DSBs and G2/M cell cycle arrest. Synthetic lethality was also demonstrated in CH cells expressing a dominant–negative form of APE1 treated with ATM or DNA-PKcs inhibitors. We conclude that APE1 is a promising synthetic lethality target in cancer.

Introduction

DNA base excision repair (BER) is critical for processing base damage induced by alkylating agents and radiation ^{1, 2}. Inhibitors that block BER, specifically those developed against PARP [poly-(ADP-ribose) polymerase], not only potentiate the cytotoxicity of chemotherapeutics and radiation, but also induce synthetic lethality in BRCA-deficient breast and ovarian cancers ^3-5. The BRCA genes encode BRCT repeat containing proteins that facilitate the efficient resolution of DNA double-strand breaks (DSBs) through a process called homologous recombination (HR). Cells lacking functional BRCA proteins are deficient in HR, and thus dependent on the more error-prone non-homologous end joining (NHEJ) pathway. This transition results in chromosomal instability, which could include oncogene activation and tumour-suppressor deletion that drives the malignant phenotype. Women carrying deleterious germline mutations in the BRCA1 and BRCA2 genes have a high risk of developing breast and ovarian cancers ⁶. It was recently demonstrated that HR impaired BRCA deficient cells are hypersensitive to PARP inhibitors that block single strand break (SSB) repair, a subpathway of BER ^{3, 4}. Although the precise mechanism for synthetic lethality is not fully known ⁷, SSB repair inhibition may result in the formation and accumulation of toxic DSBs at replication forks in BRCA deficient cells and induces synthetic lethality ^{3, 4}. Emerging data from clinical trials using PARP inhibitors in BRCA deficient breast and ovarian tumours has provided confirmatory evidence that synthetic lethality by targeting BER has the potential to improve patient outcomes ⁸.

Apurinic/apyrimidinic (AP) sites are obligatory repair intermediates in BER, and are formed spontaneously or as products of damage-induced or enzyme-catalyzed hydrolysis of the N-glycosylic bond. Unrepaired AP sites block replication fork progression and generate SSBs that eventually progress to toxic DSBs. Moreover, the ring opened aldehyde form of an AP site may be cytotoxic by virtue of its ability to react with nuclear proteins resulting in protein-bound DNA lesions that further interfere with DNA replication ^9-15. AP sites also affect topoisomerase activity and/or trap topoisomerase-DNA covalent complexes ^{16, 17}, contributing additional DNA strand breaks in genomic DNA. A recent study in yeast lacking AP endonucelase activity, accumulation of DSB was also demonstrated in G2 phase of the cell cycle ¹⁸. In human BER, AP sites are processed predominantly by AP endonuclease 1 (APE1), a multifunctional protein ¹. The DNA repair function is performed by the conserved C-terminal domain of the human enzyme. APE1 is also intimately involved in the coordination of BER and interacts with several factors within the pathway ¹. The N-terminal region of APE1 is involved in redox regulation of transcription factors, reducing an oxidized cysteine residue in the target protein to activate DNA binding and transcriptional activities ¹. The DNA repair and the redox functions of APE1 can operate independently from each other. In addition, APE1 is also involved in acetylation-mediated gene regulation ¹⁹ and RNA quality control ²⁰.

APE1 is essential for cell growth and survival, and is an emerging anticancer drug target. APE1 knockdown correlates with the accumulation of AP sites, induction of apoptosis and reduced cell proliferation. APE1 depletion sensitizes mammalian cells to a variety of DNA damaging agents ¹, and APE1 overexpression results in resistance to alkylating agents, bleomycin and radiation ¹. APE1 expression has prognostic and/or predictive significance in several human tumours including ovarian and breast cancers ¹. Nuclear expression of APE1 has been consistently observed in cervical, non–small cell lung cancer, rhabdomyosarcomas, and squamous cell head-and-neck cancer ¹. High APE1 expression correlates to poor survival in osteosarcoma. APE1 expression may also predict response to cytotoxic therapy in cervical and germ cell tumours ¹. We and others have initiated drug discovery programmes and isolated several small molecule inhibitor compounds of APE1 ^21-27. We have shown that APE1 inhibitors lead to accumulation of AP sites in vivo and potentiate the cytotoxicity of alkylating agents such as temozolomide in human cancer cell lines ^21-24.

The ability of PARP inhibitors (that block single strand break repair) to induce synthetic lethality in BRCA deficient breast and ovarian cancers ^3-5 implies that other factors within BER are potential synthetic lethality targets. Given the essential role of APE1 in BER, we have investigated in the current study the ability of APE1 inhibitors to induce synthetic lethality in DSB repair deficient cells. This study using DNA repair deficient systems provides the first evidence that APE1 inhibition is a promising new synthetic lethality strategy in cancer.

Materials and Methods

Compounds and reagents

APE1 inhibitors were purchased from ChemDiv Inc. (CA, USA), Ukrorgsynthesis Ltd (Kiev, Ukraine) and Sigma-Aldrich (UK). E3330 and methoxyamine were purchased from Sigma-Aldrich (UK). NU1025, NU7441 and KU55933 were purchased from Tocris Bioscience, UK. Wortmannin was obtained from Calbiochem,UK. All compounds were dissolved in 100% DMSO and stored at -20⁰C. shRNA for APE1 knock down and transfection reagents were purchased from SA Biosciences, MD, USA.

Cell lines and culture

Previously well characterized CH lung fibroblast cells; V79 (Wild type), V-C8 (BRCA-2 deficient), V-C8(Rev1) (BRCA2 revertant), and V-E5 (ATM-like deficient) ^{28, 29} were grown in Ham's F-10 media (PAA, UK) [supplemented with 10% fetal bovine serum (FBS) (PAA,UK) and 1% penicillin/streptomycin]. A CH ovary cell line that allows tetracycline-regulated expression of a dominant–negative form of APE1 (E8 cells) and its comparative control line (T-REx) were grown in DMEM (InVitrogen, Carlsbad, CA, USA), supplemented with 10% FBS (tet-minus; Clontech Laboratories Inc., Mountain View, CA, USA), and 1% penicillin, streptomycin and glutamate ³⁰. The human breast cancer cell lines, MDA-MB-231 and MCF-7, were grown in RPMI1640 (Sigma, UK). MDA-MB-436 (BRCA1 deficient human breast cancer cell line) and PANC1 (human pancreatic cancer cell line) were grown in DMEM (Sigma, UK). CAPAN1 (BRCA2 deficient human pancreatic cancer cell line) was grown in IMDM (PAA, UK). All media used to culture human cancer cell lines were supplemented with 10% FBS (PAA, UK) and 1% penicillin/streptomycin. BRCA2 deficient HeLa SilenciX® cells and control BRCA2 proficient HeLa SilenciX® cells were purchased from Tebu-Bio (www.tebu-bio.com). HeLa SilenciX cells were grown in DMEM medium (with L-Glutamine 580mg/L, 4500 mg/L D-Glucose, with 110mg/L Sodium Pyruvate) supplemented with 10% FBS, 1% penicillin/streptomycin and 125 μg/ml Hygromycin B.

Clonogenic survival assay

For CH lung fibroblasts, two hundred cells per well were seeded in six-well plates. Cells were allowed to adhere for 4 hours. Compounds (APE1 inhibitors, E3330, methoxyamine, or APE1 non-inhibitors) were added at the indicated concentrations. The plates were left in the incubator for 10 days. After incubation, the media was discarded, fixed (with methanol and acetic acid mixture) and stained with crystal violet.

For T-REx CH control and E8 cell lines, cells were grown to confluence, then trypsinized and counted. One hundred fifty cells of each cell line were subsequently transferred to each well of a six-well plate. Cells were allowed to adhere for 2 hours before being treated with 1 μg/ml tetracycline ³⁰. At the end of 24 hour incubation, cells were treated for 1 hour at the indicated concentrations of NU7441, KU55933 or Wortmannin. Cells were then gently washed 2 times with 1× phosphate buffered saline and incubated for 10 days in fresh DMEM to allow colonies to form. Colonies were fixed with methanol, stained with methylene blue and counted. Surviving fraction = (No. of colonies formed/No. of colonies in untreated) ×100.

For human cancer cell lines, 200-400 cells per well were seeded in 6 well plates, and allowed to adhere for 4 hrs. APE1 inhibitor was then added at indicated concentrations. The plates were left in the incubator for 12-14 days. After 2 weeks of incubation, the media was discarded, fixed (with methanol and acetic acid mixture) and stained with crystal violet. Colonies were counted and survival fraction was calculated as follows: Surviving Fraction = [No. of colonies formed/(No. of cells seeded × Plating efficiency)] ×100. All clonogenic assays were done in triplicate.

AQ_ueous Non-Radioactive Cell Proliferation Assay (MTS assay)

To evaluate cytotoxic agents by APE1 inhibitors, MTS assays were performed as per the manufacturer's recommendation (Promega). Briefly, 2,000 cells per well (in 200 μL of medium) were seeded into a 96-well plate. Cells were incubated with varying concentrations of APE1 inhibitors and the MTS assay was performed on day 4.

Aldehyde Reactive Probe assay

AP sites were quantified as described previously ²². Untreated cells were compared to cells exposed to APE1 inhibitor. DNA was extracted pre-treatment, at 2 hours and 4 hours using the guanidine/detergent lysis method. AP site determinations were performed on the genomic DNA using an aldehyde reactive probe assay kit using the protocol provided by the manufacturer (BioVision Research Products, CA, USA). All experiments were performed in triplicate.

Alkaline and neutral COMET assay

This assay was performed as described previously ³¹. Briefly, sub-confluent cells were exposed to APE1 inhibitor. At various time points after exposure (pretreatment, 2 hours, 4 hours, 8 hours, 24 hours and 48 hours) cells were extracted and comet assays were performed. Alkali electrophoresis buffer consisted of 200 mM NaOH, 1 mM EDTA and pH 13. Neutral electrophoresis buffer (TBE buffer) consisted of (89 mM Tris-base, 89 mM Boric acid, 2 mM EDTA and pH 8.3. The slides were then stained with SYBR® green (1:10,000 dilution) (Molecular Probes) for 10 minutes and images were visualized under a rhodamine filter with an Olympus BX40 microscope. The comets were analysed using Comet Assay III image analysis software (Perceptive Instruments, Suffolk, UK). A total of 200 comet images were evaluated for olive tail moment for each time point (pretreatment, 2 hours, 4 hours, 8 hours, 24 hours and 48 hours).

γH2AX immunocytochemistry

This assay was performed as described previously ³. Briefly, cells were incubated in medium containing APE1 inhibitor for 24 hours. After incubation, cells were washed, permeabilized and incubated with primary anti–phosphohistone H2AX (Ser139) antibody (clone JBW301, mouse monoclonal antibody; Upstate, Millipore Corp.) (1:200 dilution in blocking buffer) at room temperature for 1 hour. After incubation, the cells were washed and incubated with secondary anti-mouse antibody (polyclonal goat anti-mouse immunoglobulins, DAKO, dilution 1:200) at room temperature for 1 h in the dark, later washed, air dried at room temperature, stained with DAPI and stored overnight at 4°C before analyses. Images were obtained using Olympus BX40 microscope and the images captured by cellSens (Vers 1.4) Imaging Software and camera (Olympus). The frequencies of cells containing γH2AX foci were determined in 100 cells per slide in three separate experiments. Nuclei containing more than six γH2AX foci were considered positive.

Flow cytometric analyses (FACS)

Cells grown to sub-confluence were exposed to APE1 inhibitors and collected by trypsinization and centrifugation (1000 rpm for 5 minutes). The cell pellets were fixed in 70% ethanol in PBS, incubated at 4°C for atleast 12 hours to allow fixation and then stored under these conditions until FACS analyses. Prior to FACS analysis, fixed cells were harvested by centrifugation (1000 rpm for 5 mins) and the pellet was resuspended in PBS containing propidium iodide (2 μg/ml) and DNAse-free RNase A (10 μg/ml). After incubation at 37°C for 1 hr, the samples were analysed by flow cytometry using a Becton Dickinson FACS machine with a 488nm laser. Red fluorescence (DNA) was collected for 10,000 cells for each sample. Data were analysed using Flow Jo7.6 software.

APE1 knock down using shRNA

Transfection reagent was prepared according to manufacturers guidelines by adding 100 μl of Opti-MEM™ reduced –serum medium (Gibco), 0.40 μg of scrambled and APE1 knock out sequence containing shRNA plasmid (SA Bioscience) and 3 μl of Sure effect (SA Bioscience) into appropriate wells of a 24-well cell culture plate and mixed gently by rocking for 20 minutes. Cells at a density of 1.6 ×10⁶ cells/ml were prepared according to manufacturers indicated protocol. When the transfection complex formation was complete, 500μl of the cell suspension was added, mixed gently and incubated in an atmosphere of 5% CO₂ and 95% air for 24 to 48 hours. Cells were later passaged into 6 well plates, incubated and selected in media containing 800μg/ml of G418 (Gibco) for seven days (The concentration for G418 was selected after determining the dose response curve as recommended by the manufacturer and media replaced every 72 hrs). Stable APE1 knock cells were generated over a 12 weeks period and APE1 knockdown was confirmed by western blot analysis.

Western blot analysis

Protein samples were prepared by lysing cells in RIPA buffer (20mM Tris, 150mM Nacl, 1% Nonidet p-40, 0.5% sodium deoxycholate, 1mMEDTA, 0.1% SDS) containing protease inhibitor (Sigma) and phosphatase inhibitor cocktail 1 and 2 (Sigma). Western blot analyses was performed. Membranes were incubated with primary antibodies (4°C/overnight, APE-1, Novus Biologicals Inc, Littleton, CO 1:1000 dilution and Actin (Abcam) 1:10000 dilution) and infrared dye labeled secondary antibodies (Li-cor) (IRDye 800CW Donkey Anti-Rabbit IgG (H+L) and IRDye 680CW Donkey Anti-Mouse IgG (H+L) in the dilution of 1:15000 for 60 min. Protein expression was determined by scanning the membranes on Licor-Odyssey's Scanner at the predefined intensity fluorescence channel (700nm and 800nm).

Statistical analysis

Statistical significance of differences was determined using the student t-test in Microsoft Excel.

Results

APE1 inhibitors

We have recently developed complementary drug screening strategies to isolate APE1 inhibitors ^21-24. Although detailed biochemical and cellular investigations have been presented in previous publications ^21-24, Table 1 summarises several key features of the three top prototypical APE1 inhibitors selected for the current study. Inhibitor-1 (N-(4-fluorophenyl)-2-[4-phenylsulfonyl-2-(p-tolyl)oxazol-5-yl]sulfanyl-acetamide) and inhibitor-2 (N-(9,10-dioxo-1-anthryl)-2-(1H-1,2,4-triazol-5-ylsulfanyl)acetamide) have an IC₅₀ for APE1 inhibition of 11.6 μM and 4 μM, respectively ²², while inhibitor-3 (3,5,7-Trihydroxy-2-(3,4,5-trihydroxyphenyl)- 4-chromenone, a.k.a. myricetin) was reported to have an IC₅₀ of 0.32 μM ²³. All three inhibitors are specific for APE1, in that they do not appear to interact with DNA and have no activity against E. coli endonuclease IV (a functional homolog with no sequence or structural homology to APE1). All these compounds have been found to be potent inhibitors of AP site cleavage activity in HeLa whole cell extract assays, lead to accumulation of AP sites in vivo in cancer cells (confirming target inhibition) and to potentiate the cytotoxicity of alkylating agents in cancer cell lines ^{22, 23}. Molecular modelling studies indicate that these compounds dock onto the active site of APE1 ^{22, 23}. Methoxyamine, a non-specific indirect inhibitor of APE1, binds irreversibly to AP sites in DNA ¹⁵ and prevents APE1 (and endonuclease IV) from processing the adducted AP lesion. E3330 is a APE1 redox domain inhibitor and does not inhibit the AP site cleavage activity of the protein ³². Three non-inhibitors of APE1 from a previous screen ²² were also chosen randomly as negative controls. NU1025, a well characterised PARP inhibitor, was chosen as a positive control. These compounds were employed collectively to explore synthetic lethality relationships in a panel of DNA repair deficient cells.

Table 1. APE1 Inhibitors.

Inhibitor	Structure	APE1 inhibition	Endo IV inhibition	IC50	BRCA2(−/−):SL	ATM(−/−):SL
N-(4-fluorophenyl)-2-[4-phenylsulfonyl-2-(p-tolyl)oxazol-5-yl]sulfanyl-acetamide		+	−	11.6 μM	+	+
N-(9,10-dioxo-1-anthryl)-2-(1H-1,2,4-triazol-5-ylsulfanyl)acetamide		+	−	4 μM	+	+
3,5,7-Trihydroxy-2-(3,4,5-trihydroxyphenyl)- 4-chromenone (Myricetin)		+	−	0.32 μM	+	+
Methoxyamine*		+	+	NA	+	+
(2E)-2-[(4,5-Dimethoxy-2-methyl-3,6-dioxo-1,4-cyclohexadien-1-yl)methylene]-undecanoic acid (E3330)**		−	−	−	−	−

^*= methoxyamine binds irreversibly to AP site in DNA. Methoxyamine bound AP sites are resistant to processing by APE1. Therefore, methoxyamine is an indirect inhibitor of APE1.

^**= E3330 blocks redox domain activity of APE1 but has no inhibitory activity on the DNA repair domain of APE1. IC ₅₀ = biochemical inhibition of AP site cleavage activity, SL= synthetic lethality.

APE1 inhibitors exhibit increased toxicity against BRCA2 deficient and ATM deficient CH cells

V79 (Wild type), V-C8 (BRCA-2 deficient) and V-C8(Rev1) (BRCA2 revertant) were investigated in clonogenic survival assays. Figure 1 demonstrates that inhibitor-1 (Figure 1A), inhibitor-2 (Figure 1B) and inhibitor-3 (Figure 1C) induce reduced survival of V-C8 cells in comparison to V79 and V-C8(Rev1) cells. We then tested APE1 inhibitors in V-E5 cells (ATM-like deficient). Figure 1D demonstrates that inhibitor-1 is more toxic to V-E5 cells than V-79 cells. A similar lethality profile was also seen when using inhibitors-2 and -3 (Table 1). Similar results were also seen with MTS assays. We then tested the effects of methoxyamine. Figure 1E demonstrates that methoxyamine is more lethal to V-C8 cells than V79 and V-C8(Rev1) cells. Similar lethality was also demonstrated with methoxyamine in V-E5 cells (data not shown). E3330, the redox inhibitor of APE1, did not show increased toxicity to V-C8 (Figure 1F) or V-E5 cells (data not shown). We also investigated the ability of NU1025 (PARP inhibitor) to induce lethality in our model systems. Figure 2A demonstrates the expected synthetic lethality of NU1025 in V-C8 cells compared to V79 and V-C8(Rev1) cells. Similar synthetic lethality was seen in V-E5 cells with NU1025 as well (data not shown). As an additional control, we tested three non-inhibitors of APE1 randomly chosen from a previous screen ²². Figure 2B shows that non-inhibitor-1 (tested at concentrations similar to the above APE1 inhibitors) did not induce lethality in V-C8 or V-E5 cells compared to V79 and V-C8(Rev1) cells. The other two non-inhibitors also did not induce lethality (data not shown).

Figure 1.

Clonogenic survival assays for CH lung cells. Inhibitors were added at the indicated concentrations (see methods for details): A. Inhibitor-1, B. Inhibitor-2, and C. Inhibitor-3 with V-C8, V79 and V-C8(Rev1) cells. D. Inhibitor-1 with V-E5 and V79 cells. E. Methoxyamine induces synthetic lethality in V-C8 cells compared to V-79 and V-C8 (Rev1) cells. F. E3330 does not induce synthetic lethality in V-C8 cells. Survival of V-C8 cells is similar to V79 and V-C8 (Rev1) cells.

Figure 2.

Clonogenic survival assays. Inhibitors were added at indicated concentrations (see methods for details). A. NU1025 induces synthetic lethality in V-C8 cells compared to V79 and V-C8 (Rev1) cells. B. Non-inhibitor-1 has no effect on V-C8 cells. Survival of V-C8 cells is similar to V79 and V-C8 (Rev1) cells. Clonogenic survival assays for T-REx CHO control cells and E8 cells. C. NU7441, D. KU55933, and E. Wortmannin. Results shown in panels C-E represent the average and standard deviation of six independent data points. F. Western blot (inset) confirming stable APE1 knock down (KO) in MDA-MB-231 treated with shRNA constructs compared to scramble and wild type cells. Clonogenic survival assays preformed with wortmannin at the indicated concentrations.

DSB repair inhibitors induce lethality in CH cells expressing dominant-negative APE1 or human cancer cells depleted for APE1

A CH ovary cell line (E8) was recently generated to permit tetracycline-regulated expression of a dominant–negative form of APE1 ³⁰. Tetracycline induced E8 cells were shown to be BER deficient, in that they lack efficient AP site processing and display hypersensitive to DNA base damaging agents and certain anti-metabolites ³⁰. We have exploited this model system to investigate if DSB repair inhibitors such as KU55933 (ATM kinase inhibitor), NU7441 (DNA-PKcs inhibitor) and Wortmannin (broad spectrum DSB inhibitor) would induce lethality in BER defective CH cells. Each of these inhibitor compounds was found to be more toxic to the tetracycline induced E8 cell line in comparison to the T-REx CH ovary control cells (Figure 2C, Figure 2D and Figure 2E). We also confirmed this observation in E8 cells only that were either tetracycline induced (tet⁺) or un-induced (tet⁻). Western blot analysis for ED protein expression in the presence (+) or absence (−) of tetracycline is shown in supplemental Figure 2D. E8 tet⁺ cells are hypersensitive to NU7441 (supplemental Figure 2A), KU55933 (supplemental Figure 2B) and Wortmannin (supplemental Figure 2C).

To expand the studies using BER deficient CH cells, we developed a stable APE1 knockdown human breast cancer cell line using shRNA. Figure 2F (inset) demonstrates that compared to wild type and the scramble knockdown controls, cells transfected with APE1 shRNA had a near complete knockdown of APE1. We then tested lethality with the DSB repair inhibitors in this system. Figure 2F confirms that APE1 knockdown cells are hypersensitive to Wortmannin. Similar results were seen with KU55933 and NU7441 (supplemental Figures 1D and 1E). The results with human cancer cells concur with the data from the CH cells and confirm that in BER deficient systems, DSB repair inhibition by small molecule inhibitors likely leads to synthetic lethality.

Evidence for synthetic lethality upon APE1 inhibition in CH mutant cells

The clonogenic survival studies presented above provide the first evidence that modulation of APE1 is a promising new synthetic lethality strategy. To provide further evidence, we proceeded to investigate the functional consequence of APE1 inhibition in the DSB repair deficient and proficient CH cells.

To determine targeted inhibition of APE1 in vivo, the aldehyde reactive probe assay was performed to measure the level of unrepaired chromosomal AP sites. Figure 3A summarizes the results for V79, V-C8 and V-C8(Rev1) cells treated with 10 μM of inhibitor-1. In samples prior to inhibitor treatment, the background AP site level in genomic DNA was low (range 2-3 AP sites per 10⁵ base pairs). Conversely, within 2 hours of treatment with 10 μM of inhibitor-1, cells accumulated a significant number of AP sites (range 11-12 AP sites per 10⁵ base pairs). At 4 hours, AP site accumulation increased further to more than 15 AP sites per 10⁵ base pairs in genomic DNA (p<0.05). This data confirms APE1 specific inhibition in vivo. Similar accumulation of AP sites was also demonstrated using inhibitor-2 and inhibitor-3 (data not shown).

Figure 3.

A. Aldehyde Reactive Probe (ARP) assay. In pre-treatment samples, background AP site level was low in all CH cell lines. At 2 hours and 4 hours of treatment with 10 μM of inhibitor-1, cells accumulated a significantly (p<0.001) higher level of AP sites. See text for details. B. Alkaline COMET assay. V-C8 cells demonstrated a higher mean tail moment compared to V79 and V-C8(Rev1) cells at 24 hours (p= 0.0008). C. Neutral COMET assay was performed as described in methods at various time points. Mean tail moment was higher in V-C8 cells at 24 hours (p=0.002) and 48 hours (p=0.003) compared to V-79 and V-C8-Rev1 cells. D. γH2AX immunocytochemistry. After 24 treatment with 10μM of inhibitor-1, γH2AX positive cells are more (p=0.04) in V-C8 cells compared to, V-C8 (Rev1) and V79 cells (UT= untreated, T = treated). E. FACS analyses. At 24 hours, V-C8 cells were shown to be arrested in G2/M phase compared to V-79 cells. F. Quantification of each stage of cell cycle including standard deviations is shown here. See text for details

The alkaline COMET assay detects natural AP sites, as well as SSBs and DSBs. Figure 3B summarizes the results for V79, V-C8 and V-C8(Rev1) cells treated with 10 μM of inhibitor-1. Compared to pre-treatment samples, after 24 hours of exposure to APE1 inhibitor, V-C8 cells demonstrate a significantly higher mean tail moment compared to V-79 and V-C8(Rev1) cells (p=0.0008). Such data indicates that V-C8 cells accumulate alkaline sensitive sites and DNA strand breaks after exposure to an APE1 inhibitor compared to V-79 and V-C8(Rev1) cells. Similar results were seen with inhibitor-2 and inhibitor-3 (data not shown).

The neutral COMET assay specifically detects DSBs in DNA. Figure 3C summarizes the results for V79, V-C8 and V-C8(Rev1) cells treated with 10 μM of inhibitor-1. Compared to pre-treatment samples, after 24 hours of exposure to APE1 inhibitor, the mean tail moment was significantly higher in V-C8 cells at 24 hours (p=0.002) and at 48 hours (p=0.003) in comparison to V79 and V-C8(Rev1) cells. This data demonstrates that V-C8 cells accumulate DSBs after exposure to an APE1 inhibitor compared to V79 and V-C8(Rev1) cells. Similar results were seen with inhibitor-2 and inhibitor-3 (data not shown). We also demonstrated significant accumulation of DSBs in VE5 cells compared to V79 cells at 24 hours (p<0.006) and at 48 hours (p<0.001) (supplemental Figure 1A).

DSBs induce phosphorylation of H2AX at serine 139 (γH2AX), and accumulation of γH2AX foci in the nucleus is a marker of DSBs ³³. Therefore, γH2AX immunocytochemistry was performed in V-C8 cells, V-C8(Rev-1) and V-79 cells. Nuclei containing more than six γH2AX foci were considered positive. Cells were exposed to 10 μM inhibitor-1 for 24 hours and compared to control samples prior to compound treatment. In V79 cells, the mean γH2AX positive cells was 5 in pre-treatment cells and increased to 14 after the 24 hour treatment (Figure 3D). In V-C8(Rev-1) cells, the mean γH2AX positive cells was 11 in pre-treatment cells and increased to 20 after the 24 hour treatment. In V-C8 cells, the mean γH2AX positive cells was 9 in pre-treatment cells and increased to 30 after the 24 hour treatment (p=0.04). The data here provide additional evidence that V-C8 cells accumulate DSBs, consistent with the results obtained using the neutral COMET assay (Figure 3C). Similar γH2AX immunocytochemistry results were seen with inhibitor-2 and inhibitor-3 (data not shown). Similarly, in VE5 cells, the mean γH2AX positive cells were significantly high compared to wild type cells (supplemental Figure 1B) (p=0.03). Although γ H2AX foci may be indicative of other types of damage as well as DSBs, the data presented above along with the results presented for neutral COMET assays implies that there is DSB accumulation in BRCA deficient cells.

Accumulation of DSBs may delay cell cycle progression. FACS analyses were therefore performed in V-C8 cells and V79 cells exposed to inhibitor-1 for 24 hours, and cell cycle progression was evaluated and compared to control samples before treatment. Figures 3E and 3F summarise the data for cells treated with 10 μM of inhibitor-1. At 24 hours, V-C8 cells were shown to be arrested in G2/M phase of the cell cycle compared to V79 cells. This is consistent with a previous study where a PARP inhibitor was shown to induce G2/M arrest in BRCA deficient cells ⁴. Similar cell cycle results were seen with inhibitor-2 and inhibitor-3 (data not shown).

Evidence for synthetic lethality in BRCA deficient human cancer cell lines

Studies in the DNA repair deficient CH systems presented above provide clear mechanistic evidence that APE1 inhibition is a promising new synthetic lethality strategy. To confirm this observation in human systems, we employed a panel of human cancer cell lines known to be deficient in DSB repair. Figure 4A demonstrates that inhibitor-1 is lethal in BRCA2 deficient HeLa SilenciX cells in comparison to control BRCA2 proficient SilenciX cells. We then performed neutral comet assays to evaluate DSB accumulation (Figure 4B). The mean tail moment was higher in BRCA2 deficient HeLa SilenciX cells at 24 hours (p=0.01) and at 48 hours (p=0.01) post-treatment in comparison to BRCA2 proficient control SilenciX cells. γH2AX immunocytochemistry was performed as described above. Figure 4C summarises the results for cells treated with 10 μM of inhibitor-1. In BRCA2 proficient control SilenciX cells, the mean γH2AX positive cells was 14 in pre-treatment cells and increased to 28 after a 24 hour treatment with inhibitor-1. In BRCA2 deficient SilenciX cells, the mean γH2AX positive cells was 18 in pre-treatment cells and increased to 44 after a 24 hour treatment (p=0.02).

Figure 4.

A. Inhibitor-1 induces synthetic lethality in BRCA2 deficient SilenciX cells compared to BRCA2 proficient control SilenciX cells. B. Neutral COMET assay was performed as described in methods at various time points. Mean tail moment was higher in BRCA2 deficient SilenciX cells at 24 hours (p= 0.01) and 48 hours (p= 0.01) compared to BRCA2 proficient control SilenciX. C. γH2AX immunocytochemistry. After 24 hour treatment with 10μM of inhibitor-1, γH2AX positive cells are significantly more (p=0.02) in BRCA2 deficient SilenciX compared to BRCA2 proficient control SilenciX (UT= untreated, T = treated). D. Inhibitor-1 induces synthetic lethality in CAPAN1 cells compared to PANC1 cells. E. NU1025 induces similar synthetic lethality in CAPAN1 cells compared to PANC1 cells.

We then evaluated lethality in well characterised BRCA deficient human cancer cell lines. Figure 4D demonstrates that inhibitor-1 is more toxic to BRCA2 deficient CAPAN1 pancreatic cancer cells in comparison to BRCA2 proficient PANC1 pancreatic cancer cells. As a positive control we tested NU1025 (PARP inhibitors) in this system and observed the anticipated synthetic lethality (Figure 4D). We then performed neutral comet assays to evaluate DSB accumulation (Figure 5A). The mean tail moment was higher in CAPAN1 cells at 24 hours(p=0.03) and 48 hours (p=0.05) post-treatment (inhibitor-1) in comparison to PANC1 cells (Figure 5A). γH2AX immunocytochemistry was also performed, and Figure 5B summarises the results for cells treated with 10 μM of inhibitor-1. In PANC1 cells, the mean γH2AX positive cells was 7 in pre-treatment cells and increased to 11 after a 24 hour treatment with inhibitor-1. In CAPAN1 cells, the mean γH2AX positive cells was 9 in pre-treatment cells and increased to 17 after a 24 hour treatment with 10 μM of inhibitor-1 (p=0.05).

Figure 5.

A. Neutral COMET assay was performed as described in methods at various time points. Mean tail moment was significantly higher in CAPAN1 cells at 24 hours (p=0.03) and 48 hours (p=0.05) compared to PANC1 cells. B. γH2AX immunocytochemistry. After 24 treatment with 10μM of inhibitor-1, γH2AX positive cells are more (p=0.05) in CAPAN1 cells compared to PANC1 cells(UT= untreated, T = treated). C. Inhibitor-1 induces synthetic lethality in MDA-MB-436 cells compared to MCF-7 cells. D. NU1025 induces similar synthetic lethality in MDA-MB-436 cells compared to MCF-7 cells. E. Neutral COMET assay was performed as described in methods at various time points. Mean tail moment was higher in MDA-MB-436 cells at 24 hours (p=0.002) and 48 hours (p=0.005) compared to MCF-7 cells. F. γH2AX immunocytochemistry. After 24 treatment with 10μM of inhibitor-1, γH2AX positive cells are more (p=0.05) in MDA-MB-436 cells compared to MCF-7 (UT= untreated, T = treated).

We then examined lethality in BRCA1 deficient human breast cancer cells. Figure 5C demonstrates that inhibitor-1 is more toxic to MDA-MB-436 cells (BRCA1 deficient) than MCF-7 cells (BRCA1 proficient). As a positive control, we tested NU1025 (PARP inhibitor) and demonstrated similar lethality (Figure 5D). We then performed neutral comet assays to evaluate DSB accumulation (Figure 5E). The mean tail moment was higher in MDA-MB-436 cells at 24 hours (p=0.002) and 48 hours (p=0.005) post-treatment (inhibitor-1) in comparison to MCF-7 cells. These results demonstrate that BRCA deficient human cancer cells accumulate DSBs after exposure to an APE1 inhibitor compared to BRCA proficient cells. γH2AX immunocytochemistry was performed as described previously. Figure 5F summarises the results for cells treated with 10 μM of inhibitor-1. In MCF-7 cells, the mean γH2AX positive cells was 5 in pre-treatment cells and increased to 7 after a 24 hours treatment with 10 μM of inhibitor-1. In MDA-MB-436 cells, the mean γH2AX positive cells was 4 in pre-treatment cells and increased to 11 after 24 hours treatment with 10 μM of inhibitor-1 (p=0.05).

The data presented in human cancer cell lines concurs with the results using CH cells. The findings in their entirety provide compelling evidence that APE1 inhibitors induce synthetic lethality in BRCA 1 deficient and BRCA 2 deficient cells by interfering with AP site processing and thereby driving DSB formation and cell death.

Discussion

BER is essential for processing base damage induced by alkylating agents and radiation. AP sites are obligatory cytotoxic repair intermediates in BER formed after excision by a DNA glycosylase and are subsequently processed by human APE1 ^{34, 35}. Preclinical and clinical studies have confirmed that APE1 is a promising target for drug development in cancer ¹. Multiple drug discovery programmes have been initiated and several novel small molecule inhibitor compounds of APE1 have been identified ^21-24. In previous studies, we have shown that APE1 inhibitors potentiate the cytotoxicity of alkylating agents, such as temozolomide, in cancer cell lines ^21-24.

Preclinical and clinical studies have confirmed the ability of PARP inhibitors, which block a subpathway of BER, to induce synthetic lethality in BRCA deficient breast and ovarian cancers ^3-5. Although a model has been proposed whereby inhibition of SSB repair leads to the excessive formation of DSBs upon replication fork collapse and cellular lethality, this model is far from complete ⁷. Recent studies have also challenged the targeting of PARP. For instance, the PARP superfamily of enzymes includes at least 17 members that have different structures and functions ^{36, 37}. Moreover, PARP1 appears to be essential for the repair of SSBs, yet has other functions including roles in nucleotide excision repair, telomere length maintenance and organisation of the spindle apparatus. PARP2 is also involved in DNA repair as well as gene transcription and T-cell development ³⁸ The functions of other PARP superfamily enzymes in DNA repair and other cellular pathways remains less clear ³⁷.

PARP enzymes utilize NAD+ (nicotinamide adenine dinucleotide) as a substrate to catalyze the formation of large branched chains of poly(ADP)-ribose on several acceptor proteins including those involved in DNA damage repair and chromatin remodeling. PARP inhibitors under clinical development compete with endogenous NAD+ for active site binding. However, the specificity of several PARP inhibitors currently undergoing clinical development remains unclear, and some inhibitors not only block PARP1 but also PARP2 and other members of the superfamily ^{36, 37}. Given the large PARP superfamily and the number of unrelated proteins that also utilize NAD+, concerns regarding non-specific activity and long term toxicity of PARP inhibitors has emerged ^{36, 38-40}. Nevertheless, evidence from studies of PARP inhibitors suggests that other targets specific to BER are likely to be promising candidates for drug development.

APE1 inhibition in cells leads to AP site accumulation. AP sites are cytotoxic by virtue of replication fork blockage, generation of SSBs and DSBs. Clearly in somatic cancer cells this will be of therapeutic value. Although there is possibility of ‘off-target’ activity, our data that APE1 inhibitors result in AP site accumulation (as assessed by aldehyde reactive probe assay) and previous studies of APE1 depletion by siRNA/shRNA/anti-sense oligonucleotide ¹ demonstrating AP site accumulation and reduced cell viability implies that our data is consistent with previous observations. AP site quantification is a robust assay for target inhibition in vivo and holds promise as a biomarker of target activity in vivo. We speculate that aldehyde reactive probe assays can be performed either in peripheral blood mononuclear cells (as a surrogate biomarker) or in tumour tissue sampled after exposure to APE1 inhibitors in xenograft studies and possibly in future early phase human trials of APE1 inhibitors.

As the DNA repair domain of APE1 is highly specific to BER, we hypothesised that APE1 is a promising alternative synthetic lethality target and could potentially bypass many of the challenges concerning the development of PARP inhibitors. We demonstrate herein that APE1 inhibitors are synthetically lethal in BRCA deficient and ATM deficient cells. We have concluded synthetic lethality for the following reasons. First, in a CH cell system that expresses a dominant-negative form of APE1 we observed increased sensitivity to DSB repair inhibitors, a phenomenon also seen in a human cancer cell line deficient in APE1. Second, functional analyses in DSB repair deficient CH and human cells confirmed that inhibitors against APE1 led to an accumulation of AP sites, elevated DNA DSBs, and/or arrest of G2/M cell cycle progression. Weigent et al ²⁸ reported the generation of BRCA2 revertants. V-C8 cells display non-sense mutation in Brca2, one in exon 15 and another in exon-16, both resulting in a truncated Brca-2 protein. Mitomycin-c resistant clones were generated from V-C8 cells in that study ²⁸. V-C8(Rev1) clones were isolated that had restoration of one of the brca2 alleles. Although the Brca2 heterozygote [V-C8(Rev1)] did not gain the entire wild-type phenotype in that study, the inability of APE1 inhibitors to induce selective toxicity in the Brca2 heterozygote [V-C8(Rev1)] cells compared to VC8 in our study implies that endogenous BRCA2 in this heterozygote is sufficient to allow DSB repair and to rescue the synthetic lethality phenotype. This data is consistent with the preclinical studies that employed PARP inhibitors using BRCA deficient cell systems ^{3, 4}. VC8 cells were BRCA2 corrected using a BAC with the murine Brca2 gene in the Bryant et al study ³. VC8-B2 cell which was BRCA2 proficient was compared to VC8 cells. In contrast to VC8-B2 cells, BRCA2 deficient VC8 cells were hypersensitive to PARP inhibitors in that study ³. Thus, our study provides not only confirmation that BER inhibition is responsible for the synthetic lethality observed with DSB repair deficient cells treated with PARP inhibitor, but also evidence that APE1 inhibitors must be evaluated further as a synthetic lethality strategy in in vivo models . We present a working model for APE1 inhibition as a synthetic lethality strategy in DSB deficient cells. In brief, APE1 inhibition leads to AP site accumulation, which results in the indirect generation of SSBs that are eventually converted to toxic DSBs at replication forks. In cells deficient in DSB repair, DSBs would persist and lead to the observed synthetic lethality. In cells that are proficient in DSB repair, DSBs would be repaired and cells would survive.

Aberrant vasculature and oxygen starvation results in acute/chronic tumour hypoxia. Tumour hypoxia is a feature of common solid tumours such as breast, lung and colorectal cancers. Hypoxia in the tumour microenvironment promotes anaerobic glycolysis and lactic acid accumulation leading to a low extracellular pH and an acidic microenvironment. A recent study highlighted the critical role of BER in cancer cell survival in acidic tumour microenvironments ⁴¹. An acidic microenvironment enhances oxidative stress which results in oxidative base damage that leads to the up-regulation of BER factors such as APE1. An APE1 inhibitor was shown to be more cytotoxic to cancer cells in the acidic microenvironment in that study ⁴¹. Interestingly, this BER up-regulation in a hypoxic tumour microenvironment may be coincident with a decrease in DSB repair, including HR ^42-44. Thus, BER up-regulation and HR depletion may present an opportunity in the more common sporadic solid tumour microenvironment for synthetic lethality. This has potential for wider clinical application as current synthetic lethality strategies are only applicable to tumours from patients with germ-line deficiency in the BRCA 1 and BRCA 2 genes. The mechanistic study presented here provides the first preclinical proof that APE1 is a promising new synthetic lethality target and that APE1 inhibitors could have significant translational clinical applications in patients.

Supplementary Material

Supp Fig S1

Supplemental Figure 1. A. Neutral COMET assay was performed as described in methods at various time points. Mean tail moment was higher in V-E5 cells at 24 hours (p=0.006) and 48 hours (p=0.001) compared to V-79 cells. B. γH2AX immunocytochemistry. After 24 treatment with 10μM of inhibitor-1, γH2AX positive cells are more (p=0.03) in V-E5 cells compared to, V79 cells (UT= untreated, T = treated). C. MTS proliferation assay showing that V-C8 cells have reduced viability and proliferation compared to V-C8(Rev1) and V-79 cells. D. Clonogenic survival assays preformed with KU55933 at the indicated concentrations. E. Clonogenic survival assays preformed with NU7441 at the indicated concentrations.

Supp Fig S2

Supplemental Figure 2. ED expression induces increased cellular lethality in the presence of DNA DSB repair inhibitors. E8 T-REx CH ovary cells were treated with (tet +) or without (tet -) tetracycline for 24 hours (see Materials and Methods); tetracycline induces ED protein expression (see panel D). Subsequently, NU7441 (A), KU55933 (B), or Wortmannin (C) was added and incubated at the indicated concentrations as described in Materials and Methods. Shown are the averages and standard deviations of six independent data points from a single colony formation experiment. (D) Western blot analysis for ED protein expression in the presence (+) or absence (-) of tetracycline. Included is a positive human APE1 control protein.