Targeted inhibition of RPA reveals cytotoxic activity, synergy with chemotherapeutic DNA damaging agents and insight into cellular function

Abstract

Targeting uncontrolled cell proliferation and resistance to DNA damaging chemotherapeutics with a single agent has significant potential in cancer treatment. RPA, the eukaryotic single-strand (ss) DNA binding protein, is essential for genomic maintenance and stability via roles in both DNA replication and repair. We have identified a novel small molecule that inhibits the in vitro and cellular ssDNA binding activity of RPA, prevents cell cycle progression, induces cytotoxicity and increases the efficacy of chemotherapeutic DNA damaging agents. These results provide new insight into the mechanism of RPA–ssDNA interactions in chromosome maintenance and stability. This represents the first molecularly targeted eukaryotic DNA binding inhibitor and reveals the utility of targeting a protein-DNA interaction as a therapeutic strategy for cancer treatment.

Introduction

Replication protein A (RPA) is a heterotrimeric single-stranded DNA (ssDNA) binding protein made up of 70, 34, and 14 kDa subunits (1). The ssDNA binding activity of RPA is required for several DNA metabolic pathways including DNA replication, recombination and repair. High affinity interactions with DNA are sustained by the numerous oligosaccharide/oligonucleotide binding (OB)-folds present on each of the three subunits (2;3). OB-folds in DNA binding domains A and B (DBD-A and DBD-B) in the central region of the p70 subunit contribute most of the binding energy for RPA-ssDNA interactions (2). Individual OB-folds are compact modular domains populated with hydrophobic and basic amino acids. These structural features make the OB-folds an attractive target for development of small molecule inhibitors (SMIs) of DNA binding activity.

Inhibiting RPA-DNA interactions has the potential to impact numerous DNA metabolic pathways that are differentially active in cancer cells. In DNA replication, RPA inhibition can be used to exploit the highly proliferative nature of cancer cells which is characterized by a large population of cells in S-phase. Consistent with this, a recent clinical trial correlated disease stage and metastasis in colon cancer with increased RPA p70 and p34 expression (4). RPA is also essential for several DNA repair pathways in the cell including nucleotide excision repair (NER). Cisplatin, a common chemotherapeutic used in the treatment of various cancers, induces its cytotoxic effect by forming intrastrand covalent DNA adducts that are repaired primarily by the NER pathway (5). Consistent with the role of NER in the repair of cisplatin-induced DNA damage, resistance to this treatment has been observed to be influenced by DNA repair capacity (6;7). Consequently, cisplatin treatment, in conjunction with decreased RPA ssDNA binding activity, would be expected to result in decreased efficiency of cellular repair of cisplatin-DNA adducts and increased cytotoxicity. Thus, targeting RPA has the potential not only for single agent activity but also to sensitize cancer cells to therapies that induce DNA damage and genetic instability, such as cisplatin, etoposide and ionizing radiation (IR).

We present the identification and development of the first small molecule that inhibits the ssDNA binding activity of RPA. Cellular RPA inhibition results in the inability to enter S phase, cytotoxicity and synergistic activity with the chemotherapeutic agents cisplatin and etoposide. This is the first characterization of a small molecule that is able to inhibit the ssDNA binding activity of RPA and presents a novel chemotherapeutic target both as a single agent and in conjunction with commonly used chemotherapeutics.

Materials and Methods

In vitro protein analysis

Small molecule inhibitors were obtained from ChemDiv and resuspended in DMSO. Compound 505 was independently synthesized and structure confirmed by mass spectrometry analysis. Human RPA was purified as previously described (8). Fluorescence Anisotropy based DNA binding assays were performed with 40 nM RPA and 20 nM 5’fluoroscein-labeled ss-dT₁₂ DNA as previously described (9). EMSAs were performed in 20 µL reactions containing 25 nM RPA, 25 nM 5’[³²P]-labeled 34-base pair DNA as previously described (8).

Molecular Modeling

Molecular modeling of compound 505 with the central DNA binding domain of RPA p70 (1FGU) was performed using Autodock 4.2 (10). Three independent grids established 60Å in each dimension to encompass either the interdomain region, DBD-A or DBD-B. Semi-flexible automated ligand docking was performed using the Monte Carlo based simulated annealing and locality search algorithms. The most stable complexes were selected based on binding energies from multiple analyses. Coordinates of the final docked complexes were displayed with PyMOL.

Flow Cytometry

H460 cells were analyzed for apoptosis using an Annexin V-FITC/Propidium iodide (PI) Vybrant Apoptosis Assay Kit (Invitrogen), according to manufacturer’s instructions. Cells were plated at a density of 1 × 10⁴ cells/cm² for 24 hours and then treated with compound 505 for 48 hours. Following plating and treatment of H460 cells, adherent and non-adherent cells were collected, processed, and analyzed on a BD FACScan flow cytometer. Data was analyzed using WinMDI software (The Scripps Research Institute, San Diego, CA). Cell cycle analysis was performed by PI staining. Briefly, cells were plated and treated with compound, collected and then washed twice with 1% BSA in PBS-EDTA. Cells were fixed in 70% EtOH at −20 °C followed by incubation on ice for 30 minutes. Cells were then collected and stained with 1 µg/mL PI and 25 µg/mL RNaseA for 1.5 hours and analyzed on a Becton Dickinson FACScan flow cytometer. Cells were gated and analyzed on a histogram with events plotted against the FL2-A parameter. Cell cycle distribution was analyzed using ModFit software. G2 arrest was induced by treatment with 0.8 µg/mL nocodazole for 12 hours (11). Cells were then washed with PBS and treated with either vehicle or compound 505 (100 µM). Cells were harvested and analyzed for cell cycle distribution as described above. To analyze BrdU incorporation, cells were labeled with 10 µM BrdU after which cells were collected, washed and, fixed in 70% EtOH. Following fixation, DNA was denatured with 2M HCl for 20 minutes at room temperature, washed and neutralized with 0.5 M sodium borate for 2 minutes. Cells were then incubated with anti-BrdU antibody (CalBiochem) (1:500) for 20 minutes at room temperature, washed and then incubated with AlexaFluor 488 goat anti-mouse secondary antibody (Invitrogen) (1:500) also for 20 minutes at room temperature. Following antibody incubation, cells were washed and incubated in 1 µg/mL PI in PBS-EDTA for at least 30 minutes at 4 °C. Cells were then analyzed as described above for cell cycle and BrdU staining using FLA and FL1 parameters, respectively.

Indirect Immunofluorescence

H460 cells were plated on chamber slides (LabTek) as described above. Cells were then treated for 3 hours with either 50 µM of compound 505 or vehicle as indicated and, following treatment, cells were fixed in 4% paraformaldehyde at 25 °C for 3 minutes followed by washing in 0.2% Triton X-100 for 2 minutes at 4 °C. The slides were then blocked in 15% FBS in PBS for 1 hour at 25 °C and then incubated with anti-RPA 34 primary antibody (Neomarkers) (1:500) in 15% FBS for 1 hour. Slides were then washed 3 × 10 minutes with 15% FBS and then incubated with Alexa Fluor-594 goat-anti-mouse secondary antibody (Invitrogen) (1:300) for 1 hour. Slides were again washed and stained with 300 nM DAPI diluted in PBS-EDTA for five minutes. Slides were then mounted and images captured using a Zeiss fluorescent microscope and images were captured using filters for Texas Red to visualize RPA staining and DAPI for visualizing DNA. Slides were visualized and images analyzed and quantified using ImageJ software.

Western Blot Analysis

H460 cells were plated and treated with either vehicle or increasing concentrations of compound 505 as indicated for 8 hours and then processed for western blot analysis using a RIPA lysis and extraction procedure. Protein concentrations were determined using the BioRad proteins assay with BSA standard and equal amounts of total protein loaded in each well. Following electrophoresis and transfer to PVDF, RPA was detected with an anti-RPA p34 (1:2000) or anti-p70 antibody (1:1000) (Neomarkers) and goat anti-mouse-HRP secondary (Santa Cruz) (1:5000 and 1:2500 respectively). Bands were visualized using chemiluminescence detection

Results

Identification of a small molecule inhibitor of RPA

Previous work from our laboratory has identified a series of small molecules from the NCI library (9) and a ChemDiv library (12) that inhibit the DNA binding activity of RPA. Hits obtained from the Chemdiv library were analyzed in a secondary assay using electrophoretic mobility shift assays (EMSAs) to confirm inhibitory activity (Figure 1A). Of those analyzed, two showed significant RPA inhibition in vitro (compounds 3 and 5). As in vivo inhibition of RPA using siRNA has been shown to be cytotoxic due to its critical role in DNA metabolic processes, we analyzed the effects of these two compounds on cellular viability (13). While minimal cellular effects were observed with compound 3 and 5, the analyses were somewhat hampered by poor solubility (data not shown). However, considering the in vitro activity of compound 3, we retained its core structure consisting of a substituted dihydropyrazole with a 4-oxo-butanoic acid at N1 and a phenyl substituent at C3 to initiate analysis of structure activity relationships (SAR) and identify compounds with improved cellular activity (Figure 1C). Eighty-one analogs were identified and obtained from the ChemDiv library with differing substitutions off the phenyl ring (R2) and varying substituents at position C5 on the dihydropyrazole ring (R1) (Figure 1C). These derivatives were analyzed for in vitro RPA inhibitory activity using EMSAs (Figure 1D and data not shown). A subset of these compounds that showed the ability to inhibit RPA’s DNA binding activity were further characterized to determine how the structure of each contributes to this inhibition. Using EMSA analysis, titrations of each compound were performed and the IC₅₀of each was determined. Among the compounds analyzed, compound 505 was the most potent inhibitor with an IC₅₀ of 13 µM while other compounds displayed varying capacities for RPA inhibition (Table 1 and Figure 2). In order to determine cellular activity of each of the compounds, cell viability was measured with Annexin V/PI analysis in the H460 NSCLC cell line and IC₅₀ values for each compound following a 48 hour exposure were determined. These data are presented in Table 1 and reveal a correlation between in vitro and cellular activity, consistent with cellular inhibition of RPA and indicating specificity for RPA inhibition. Interestingly, compound 523 which showed minimal inhibition of RPA-in vitro, did display modest cellular activity, which could be attributed to metabolism or other cellular effects. As compound 505 displayed the lowest in vitro IC₅₀ and was the most potent compound of those examined in cells, we selected this compound to further investigate its in vitro and cellular mechanisms of action.

Figure 1. Identification of Small Molecule Inhibitors (SMIs) of RPA.

1A. RPA was incubated with individual compounds (100 µm) and DNA binding analyzed by EMSA as described in Methods. The position of free DNA and the DNA–RPA complex is denoted in the figure. 1B. Structure of compound 3. 1C. Core structure of compound 3 where R1 and R2 were varied to compile a library of compounds for SAR analysis. 1D. EMSA analysis of compound 3 derivatives. EMSAs were performed as described in 1A. This gel is representative of those used to screen 86 compound 3 derivatives.

Table 1.

Structure activity relationships of small molecule RPA inhibitors

Name	Structure	IC50 (µM)
		In vitro	cellular
TDRL-505		12.9 ± 1.3	30.8±1.7
TDRL-518		>100*	NA
TDRL-520		20.3±10.7	49.9±2.5
TDRL-521		71.7±33.9	56.9±6.7
TDRL-522		56.1±6.7	38±32***
TDRL-523		>100**	31.0±5.2

The in vitro IC₅₀ was determined by EMSA analysis as described in 1A. Cellular IC₅₀ was determined by treating H460 cells and analyzing annexinV/PI staining as described in Methods. The in vitro and cellular data was analyzed using standard 4 parameter logistic curve. The IC₅₀ values and standard error of the fit were determined from this analysis

^*inhibition at the highest concentration tested (100 µM) was 9%

^**inhibition at the highest concentration tested (100 µM) was 36%

^***Maximum observed cytotoxicity was 80% of control.

Figure 2. Analysis of in vitro activity of compound 505.

2A. Increasing concentrations of compound 505 were pre-incubated with RPA and DNA binding activity was assessed using EMSA analysis. 2B. Quantification of the gel presented in figure 2A. The average and standard deviation of each point is presented. The data was fit to a standard 4 parameter logistic curve with an N=4. 2C. Fluorescence anisotropy was performed with increasing concentrations of compound 505 with constant RPA and dT₁₂ substrate as described in Methods. The data was analyzed as described in figure 2B with N=3. 2D. Surface representation of RPA p70 181-422 (1FGU) with compound 505 docked in DBD-A (blue), DBA B (green) and the interdomain (ID) region (Red). Close up views of each interaction are presented in the outsets with compound 505 depicted using stick representations with CPK coloring.

In vitro inhibition of RPA’s DNA binding activity targeting DBD-A and B in the 70 kDa subunit of RPA

RPA binding to synthetic oligonucleotide substrates has been well characterized with respect to structural features and kinetics of binding (14–17). In order to determine the mechanism of action and the potential interactions of compound 505 with RPA, binding to oligonucleotides varying in length and sequence were employed. Compound 505 was first titrated with a constant amount of RPA in the presence of 5’-[³²P]-labeled 34-base purine–rich ssDNA. This length of DNA is capable of extending beyond the central OB-folds of the RPA p70 subunit to allow interactions with other OB-folds within the RPA heterotrimer. This analysis revealed a concentration dependent decrease in binding with an IC₅₀ of 13 µM (Figure 2A). We then extended the analysis to a ssDNA 12 bases in length which largely restricts binding to DBDs A and B, which has been shown in the co-crystal structure of RPA p70 (14). An FP assay was used to accommodate the dT₁₂ substrate and titration of compound 505 resulted in reduced RPA binding with an IC₅₀ value of 20.4 µM (Figure 2C). This data provides evidence for inhibition of RPA by blocking the interaction of the OB-folds in DBD-A and –B with DNA. In our initial screen of the ChemDiv library for RPA inhibitors, we counter-screened each compound against Xeroderma Pigmentosum Group A protein (XPA), an essential DNA binding protein in the NER pathway, and excluded any compound that inhibited both targets (12). To confirm this specificity for RPA, we examined the effect of 505 on XPA binding in an ELISA format, and no inhibitory activity was observed (data not shown).

To further define the interaction between compound 505 and DBD-A and –B, we undertook a molecular modeling approach. The crystal structure of RPA p70 181-422 was solved without DNA (18) and 3-dimensional coordinates obtained from the protein data bank (1FGU). Compound 505 was prepared as the ligand using Autodock tools 1.5.2 as described in Methods. Three potential binding sites were identified based on the lowest interaction energies. These sites include each of the OB-folds and the linker region between the two folds (Figure 2D). The most stable interaction was with DBD-B and the conformation of compound 505 while DBD-A domain ranked second in binding energy and the interdomain region showed the least stable interaction. The close-up of the interaction presented in the insets reveal that most of the stability is driven via hydrophobic interactions, although the conserved oxo-butyric acid is stabilized via interactions with basic amino acids in each of the three positions. To determine the relative importance of the interaction of compound 505 with each domain, docking analysis of compound 518 was performed which shows no inhibition of RPA’s DNA binding activity in vitro. Significantly increased ΔG values were obtained for interactions of 518 with DBD-B and the interdomain region while only modest increases were calculated for the interaction with DBD-A. These data suggest that the majority of the inhibitory effect of compound 505 is manifested through interaction with the DBD-B OB-fold and potentially the interdomain region.

Compound 505 decreases cell viability and nuclear RPA staining without induction of apoptosis and degradation of RPA

Having determined the in vitro activity of compound 505, we sought to determine the mechanism of cellular activity and the ramifications of inhibiting the central p70 OB-folds of RPA. To measure the induction of apoptosis, H460 NSCLC cells were treated with increasing concentrations of compound 505 for 48 hours, after which annexin V/Propidium Iodide (PI) staining was measured using flow cytometry (Figure 3). Minimal annexin V staining was observed, suggesting that classical apoptosis was not initiated (19). Interestingly, a concentration dependent increase in PI staining was observed, indicating a general loss of membrane integrity, suggestive of necrotic cell death (Figure 3A) (20). Quantification of the viable cells (annexin V negative/PI negative) resulted in an IC₅₀ of 30.8 µM (Figure 3B). As an independent measure of the effect of compound 505 on cell viability, we employed a crystal violet staining assay and obtained an IC₅₀ of 64 µM (data not shown). A similar result was also observed in treatment of the A549 NSCLC cell line with compound 505 while analysis using freshly isolated peripheral blood mononuclear cells (PBMCs) revealed minimal cytotoxic activity (data not shown). Therefore, compound 505 has shown significant cytotoxic effects in NSCLC cell lines while showing only modest activity in non-cancerous cells, supporting the possibility that a therapeutic treatment window may be achievable.

Figure 3. Effects of Compound 505 on cytotoxicity and cell cycle progression in a H460 NSCLC cell line.

3A. H460 cells were treated with compound 505 for 48 hours and analyzed for cell viability with AnnexinV/PI staining and presented as dot plots. 3B. Data from 3A was quantified and the percentage of viable Annexin-/PI- cells (lower left quadrant) was calculated. The average ± SD (N=4) are presented and the data was fit to a 4-parameter logistic curve. 3C. H460 cells were treated with 50 µM compound 505 or vehicle for 3 hours and analyzed for RPA expression and localization by indirect immunofluorescence using an Alexa Fluor594 secondary antibody (red). Slides were counter stained with DAPI (blue) and images merged. Magnification of the boxed cells is presented below the low magnification images. 3D. H460 cells were treated with increasing concentrations of TDRL-505 or vehicle for 8 hours and RPA expression was assessed via western blot analysis probing for the p70 and p34 subunit as indicated. Lane 1 is a vehicle treated control and Lanes 2–5 correspond to treatement with 25, 50, 75 and 100 µM TDRL-505, respectively. The positions of RPA p34 and p70 are indicated by the arrows.

The cellular effects seen following treatment with compound 505 are consistent with an inability of RPA to interact with DNA which could then result in numerous possibilities including degradation or redistribution of RPA within the cell. We therefore employed indirect immunofluorescence to assess how inhibition of RPA binding influences cellular localization. After 3 hours of treatment with compound, cells showed a decrease in the intensity of RPA staining compared to vehicle treated control without a change in overall sub-cellular localization (Figure 3C). Interestingly, western blot analysis revealed that both RPA p70 and p34 levels remain constant following treatment with increasing concentrations of compound 505 (Figure 3D). These results demonstrate that treatment with compound 505 induces differential distribution of RPA in the nucleus while the overall level of RPA is not dramatically affected.

Compound 505 affects cell cycle progression and DNA replication

The ability of compound 505 to reduce RPA nuclear staining and cell viability led us to assess the affect of treatment on cell cycle progression. siRNA knockdown of RPA has been demonstrated to induce a G1 cell cycle arrest consistent with the essential role of RPA in the initiation of S-phase DNA replication (21). Therefore we assessed the effect of 505 treatment on H460 cell cycle progression and an increase in the proportion of cells in G1-phase was observed in response to treatment (Figure 4A). To determine if entry into S-phase is inhibited in compound 505 treated cells, cells were synchronized in G2/M with nocodazole and then released from G2 arrest and re-fed complete medium supplemented with either vehicle or compound 505. Both control and treated cells rapidly progressed through mitosis into G1 after removal of nocodazole (Figure 4B). Cells that were treated with vehicle alone entered into G1, as seen at the 4 hour time point and progression into S-phase is apparent after 8 hours with progression into G2 evident at 12 hours (Figure 4B). Cells that were treated with compound 505 after release from nocodazole progressed into G1 but did not enter S-phase even after 12 hours post release. To increase the resolution of the biochemical steps in the transition from G1 to S phase, we assessed BrdU incorporation using the same treatment protocol. Following release of nocodazole arrested cells, 10 µM BrdU was added two hours prior to collection, and cells were then analyzed by flow cytometry for BrdU incorporation. In the presence of compound 505, no BrdU incorporation was observed while vehicle controls showed incorporation and progression through the cell cycle (Figure 4B). To determine the acute effect of 505 on DNA replication, asynchronous cells were treated with 505 for a short period of time (3 hours) and small but distinct differences in BrdU incorporation were observed (Figure 4C). Cells treated with TDRL-505 appear to display a lengthening of S-phase and fewer cells that have incorporated BrdU progressing through the cell cycle into G2 (Figure 4C).

Figure 4. Synchronized H460 cells show an inability to re-enter S-phase with 505.

4A. H460 NSCLC cells were treated with vehicle or compound 505 for 48 hours and then analyzed for cell cycle distribution. 4B. H460 cells were treated with 0.8 µg/mL nocodazole for 12 hours, washed then treated with either vehicle or 100 µM 505 for 4, 8 and 12 hours. BrdU (10 µM) was added during the final 2 hours of treatment. Cells were then harvested and analyzed for cell cycle distribution and BrdU incorporation. 4C. Asynchronous cells were treated for 3 hours with either vehicle or 505. BrdU (10 µM) was added during the final hour of treatment and representative dot plots of the flow cytometry data are presented.

The effect of RPA inhibition on DNA repair and the response to DNA damage

In addition to its essential role in DNA replication, RPA is involved in the DNA damage response and is required for the repair of bulky DNA adducts as well as DNA breaks induced by various types of exogenous and endogenous agents. The association of RPA with ssDNA is a critical feature of all of these pathways, indicating that inhibition of this activity would increase the cytotoxic effects induced by DNA damage. In order to determine the effect of RPA inhibition on cellular sensitivity to cisplatin, we evaluated the combination index using a concurrent treatement protocol (CI) (22). When cisplatin and compound 505 were used in combination, cell viability was decreased to a level that was greater than that induced by either agent alone, resulting in a synergy between the two compounds and CI of 0.4 at the highest fraction of cells affected (Figure 5). The interaction became additive and then antagonistic (revealed from CI values greater than one) at lower fractions of cells affected (Figure 5). These results demonstrate that compound 505 is able to potentiate the effect of cisplatin in H460 cells and is consistent with inhibition of the cellular activity of RPA in NER. The ability of compound 505 to synergize with etoposide was also examined. Etoposide induces replication fork arrest and DNA damage response, both cellular processes that require RPA (23). Using the same analysis as described above for cisplatin, compound 505 showed synergistic activity with etoposide at all fractions of cells affected (Figure 5). Interestingly, when the formation of etoposide-induced RPA foci (23) or phosphorylation of DNA-PKcs at serine 2056 was measured (24), the co-treatment with TDRL-505 resulted in a slight increase in number of cells with foci and DNA-PKcs phosphorylation (Supplemental Figures 1 and 2). These data demonstrate that TDRL-505 can influence the cellular response to etoposide consistent with the synergy observed between the two agents in the combination index analyses (Figure 5C).

Figure 5. Compound 505 acts synergistically with both cisplatin and etoposide in H460 cells.

H460 cells were treated with increasing fractions of the IC₅₀ concentration of either cisplatin or etoposide with compound 505 for 48 hours. After treatment, cells were harvested and analyzed for annexin V/PI staining. Open circles indicate CI analysis of cisplatin with 505 and closed circles represent etoposide with 505. The combination index analysis was performed as previously described (22). The data are presented as the average ± SD from (N=3).

Discussion

Advances in high throughput screening of chemical libraries have resulted in an explosion of putative cancer targets and their inhibitors. To date, the majority of these target enzymatic activity associated with a specific protein. Our results targeting the non-enzymatic DNA binding activity of RPA opens up an entire new class of putative interactions for therapeutic development. One such compound, 505 inhibits RPA-DNA interactions in vitro, blocks cell entry into S-phase and results in a cytotoxic/cytostatic response. Each of these responses is consistent with inhibiting RPA’s role in the initiation of DNA replication, which involves a complex series of interactions, one of which is the loading of RPA at replication origins in a S-phase CDK dependent process (21;25). Data presented demonstrate that TDRL-505 also abrogates BrdU incorporation into DNA in cells released from G2 arrest without an overall decreased in asynchronous cells, consistent with the importance of RPA in replication fork firing in the initial stages of DNA replication. Interestingly, treatment of asynchronous cells with TDRL-505 did display a prolongation of S-phase evidenced by fewer cells completing S-phase and progressing into G2 in the pulse-labeling experiments. Therefore, replication forks that are established prior to the addition of TDRL-505 appear not to be disrupted however the initiation from additional, late firing replication origins is decreased in the presence of TDRL-505. The potential exists that cells that have already began replicating their DNA are not affected by the inhibition of RPA binding to DNA, which correlates with observations that TDRL-505 is unable to compete RPA away from DNA once it is already bound (data not shown). Therefore, cells dependent on entry into S phase for continuous cell proliferation would be expected to be negatively impacted by treatment with TDRL-505. The inability of cells to enter S-phase is a potential mechanism contributing to cytotoxicity. This allows for the possibility of a therapeutic window for specifically targeting actively dividing cells in the context of cancer treatment using SMIs to block the cellular activity of RPA. This rationale is further bolstered by the clinical observation that high levels of RPA expression correlated with disease stage in a colon cancers (4).

The role of RPA in DNA repair also allows for inhibition of its activity to increase the efficacy of current chemotherapeutics that induce DNA damage in the context of combination therapy. The inhibition of DNA repair is anticipated to result in persistent DNA damage which would increase cytotoxicity. The indispensible role of RPA in the recognition and verification steps of NER is well characterized and, in addition, RPA participates in the re-synthesis step following excision of the damaged oligonucleotide (26). Previous studies have shown that cells with decreased levels of NER proteins demonstrate increased sensitivity to cisplatin treatment (6). Consistent with this, our data reveal a synergistic interaction between compound 505 and cisplatin at high fractions of cells affected. Interestingly, at low fractions of cells affected, an antagonistic interaction is observed with combination indices greater than one. This is likely the result of interactions not at the level of repair but at the level of signaling. As cisplatin leads to activation of a G2 checkpoint and induces apoptosis from an extended G2 arrest, the finding that compound 505 blocks cells in G1 indicates that fewer cells would be subject to cisplatin induced G2 arrest. Likewise, if compound 505 toxicity stems from an extended G1 arrest, the G2 checkpoint induced by cisplatin would result in less cell death as a result of treatment. At high concentrations, this effect is mitigated by the interaction at the level of DNA repair with RPA inhibition increasing cisplatin toxicity and overcoming the antagonistic signaling interaction.

The role of RPA in DNA replication restart and processing of collapsed replication forks also presents opportunities for combination therapy (27;28). Interestingly, combination index analysis of the activity of etoposide with compound 505 showed synergistic activity at all fractions of cells affected. Etoposide inhibits the enzymatic activity of topoisomerase II (topo II) resulting in persistent covalent-cleavage complexes on DNA which lead to replication fork arrest and both single and double strand breaks (29). RPA has been demonstrated to respond to and repair these types of lesions and reduction of the levels of RPA p70 using siRNA results in increased formation of DNA double strand breaks in response to etoposide treatment (23). These data are consistent with our observation that inhibiting RPA’s DNA binding activity potentiates the effects seen by inhibiting topo II. Additional data demonstrate that the formation of etoposide-induced RPA foci measured at 4 hours treatment was not inhibited by concurrent treatment with TDRL-505 though increased cytotoxicity was observed. Clearly it will be of interest to assess foci dynamics as a function of the timing of RPA inhibition with respect to the induction of DNA damage. In addition, in asynchronous cultures, at any given time a cell undergoing replication would be expected to be in various stages of origin firing. RPA is required in early replication firing, while topo II has been shown to be required for later stage replication events (30). Therefore, inhibition of both stages of replication progression would be expected to show a greater effect than inhibiting either one of the steps individually, which is consistent with the synergistic relationship between compound 505 and etoposide. Inhibition of RPA activity and abrogation of pathway function has the potential for widespread utility in cancer treatment. While we focused on a subset of pathways, the role of RPA in several other repair pathways opens up other opportunities for combination therapy. Specifically, combining molecularly targeted RPA inhibition with radiation therapy could lead to increased cytotoxicity in tumor cells via inhibition of DNA double strand break repair via non-homologous DNA end joining or homologous recombination, both of which have been shown to require RPA (31–33).

As in the case of any novel small molecule inhibitor, specificity for the proposed target is a concern. We have demonstrated specificity of TDRL-505 for RPA in in vitro analysis and have shown a correlation between the in vitro IC₅₀ and cellular IC₅₀ for all compounds analyzed except one. TDRL-523 did show robust cellular activity with modest in vitro RPA inhibition. The potential exists that TDRL-523 undergoes extra-and/or intracellular modification that results in a higher affinity for RPA in a cellular model compared to in vitro conditions. It is also possible that TDRL-523 does not show a high specificity for RPA and is interacting with other proteins that contain structurally similar OB-fold regions, such telomere end protection proteins (34). Although OB-folds adopt a similar global structure, the lack of sequence similarity between these domains suggests that specific interactions between amino acids and nucleic acids allows for the differential binding to nucleic acids (35). We are currently elucidating the interaction between TDRL-505 and specific amino acids in the A and B DBDs of RPA p70 in order to determine if TDRL-505 can potentially interact with other OB-fold containing proteins. It is also likely that compounds specific for a given OB-fold within RPA will have different effects on DNA metabolism dependent on the role each OB-fold plays in a given pathway.

While targeting the enzymatic activity of proteins with small molecules is well accepted, the research presented in this paper demonstrates the feasibility and utility of targeting a non-enzymatic protein-DNA interaction. These compounds therefore represent the first SMIs of RPA which display both in vitro and cellular activity. The approach of targeting RPA for cancer chemotherapy has several unique advantages including the lack of redundancy resulting from no back-up systems to counteract a loss of RPA activity. Also, inhibition of RPA is expected to have broad spectrum utility as the reliance on RPA for increased cell proliferation and repair of chemotherapeutic DNA damaging agents is not unique to any single cancer. Our targeting of the DNA binding activity of RPA with a small, drug-like molecule sets the precedent for targeting this class of proteins and thus alters the current drug discovery paradigm to open up an entire new class of targets with potential broad spectrum utility.