Translesion Synthesis Inhibitors as a New Class of Cancer Chemotherapeutics

Abstract

Introduction:

Translesion synthesis (TLS) is a DNA damage tolerance mechanism that replaces the replicative DNA polymerase with a specialized, low-fidelity TLS DNA polymerase that can copy past DNA lesions during active replication. Recent studies have demonstrated a primary role for TLS in replicating past DNA lesions induced by first-line genotoxic agents, resulting in decreased efficacy and acquired chemoresistance. With this in mind, targeting TLS as a combination strategy with first-line genotoxic agents has emerged as a promising approach to develop a new class of anti-cancer adjuvant agents.

Areas covered:

In this review, we provide a brief background on TLS and its role in cancer. We also discuss the identification and development of inhibitors that target various TLS DNA polymerases or key protein-protein interactions (PPIs) in the TLS machinery.

Expert opinion:

TLS inhibitors have demonstrated initial promise; however, their continued study is essential to more fully understand the clinical potential of this emerging class of anti-cancer chemotherapeutics. It will be important to determine whether a specific protein involved in TLS is an optimal target. In addition, an expanded understanding of what current genotoxic chemotherapies synergize with TLS inhibitors will guide the clinical strategies for devising combination therapies.

1. Introduction

Human cells have evolved to coordinate complex pathways that aid in repairing DNA damage that occurs via both endogenous and exogenous mechanisms. These pathways are collectively termed the DNA damage response (DDR) and are efficient at recognizing various types of DNA damage and activating either specific repair pathways or apoptosis [1,2]. Many of these DDR mechanisms are well-characterized and include homologous recombination, non-homologous recombination, nucleotide excision repair, base excision repair, and mismatch repair [1,2]. Even with these varied repair mechanisms in place, the replication machinery can still encounter DNA lesions when cells are actively dividing. When this occurs, cells employ DNA damage tolerance (DDT) mechanisms to bypass the lesion, maintain genome stability, and prevent cell death, but this occurs at the expense of an increased cellular mutation rate [3,4].

Translesion synthesis (TLS) is the primary DDT mechanism that cells utilize to bypass DNA lesions during active replication [5-7]. Mammalian cells employ multiple error-prone, low-fidelity polymerases from the Y-family (REV1, POLη, POLι, and POLκ) and B-family (POLζ) to carry out REV1/POLζ-dependent TLS, which is inherently mutagenic. These TLS DNA polymerases are recruited to sites of DNA damage following monoubiquitination of the sliding clamp protein proliferating cell nuclear antigen (PCNA). PCNA encircles DNA where it serves as a processing factor. Monoubiquitination of PCNA at a DNA lesion serves as the initiation event that signals a switch from normal replication to error-prone REV1/POLζ-dependent TLS-mediated replication. PCNA and the TLS DNA polymerases form a heteroprotein complex mediated by multiple protein-protein interactions (PPIs) that allow for the insertion of nucleotides across a wide range of DNA lesions. The catalytic process of REV1/POLζ-dependent TLS-mediated replication is typically a two-step process in which one of the ‘inserter’ TLS polymerases (primarily POLη, POLι, or POLκ) incorporates a nucleotide across from the DNA lesion followed by extension of the DNA primer template by the ‘extender’ TLS polymerase, POLζ [8,9]. Other polymerases (POLs θ, λ, μ, ν, etc) have been implicated in TLS; however, their specific functions have not been firmly established. It is also important to note that both REV1 and POLζ have been identified as playing a role in interstrand crosslink (ICL) repair and homologous recombination repair in addition to their role in mutagenic REV1/POLζ-dependent TLS [10,11]. More detailed descriptions of the catalytic mechanisms and lesion specificity of TLS DNA polymerases as well as the coordinated steps involved in TLS can be found elsewhere [5-7].

TLS and other DDT mechanisms are essential for proper cell survival and genome maintenance; however, their activation in cancer cells has been implicated as a common mechanism through which tumors bypass DNA lesions induced by several first-line genotoxic agents [12,13]. Error-prone REV1/POLζ-dependent TLS increases the survival of cancer cells in the presence of genotoxic chemotherapy and contributes to enhanced mutagenesis in tumors, which can ultimately result in acquired resistance to the first-line agent. As such, TLS has emerged as a promising target for the development of a new class of anti-cancer agents that can enhance the efficacy of a wide range of drugs/treatments that exert their effects through DNA damage [12-15]. In this review, we will recap our knowledge of REV1/POLζ-dependent TLS in cancer and describe the current state of small molecule inhibitors of this promising DDT mechanism.

2. TLS and Cancer

Our primary understanding of TLS as a chemotherapeutic target is within the context of its role as a driver of resistance to genotoxic chemotherapies. A diverse set of DNA damaging agents including ionizing radiation, platinum drugs, alkylating agents and temozolomide (TMZ) are widely used as first-line agents for hematological and solid cancer. These drugs exert their anti-cancer effects by directly modifying the chemical structure of DNA, resulting in cancer cell apoptosis [16]. Active TLS can rescue cancer cells following administration of these DNA damaging agents, which ultimately leads to an increased mutation rate in tumor cells and acquired resistance to the first-line genotoxic therapy [12-15].

Studies have implicated several TLS DNA polymerases in the onset of resistance to cisplatin. Initial studies demonstrated that POLη is involved in bypass of cisplatin-induced DNA adducts in human fibroblasts [17]. Related studies showed that treatment of multiple cancer types with ICL-inducing agents induced POLη up-regulation in vitro and in vivo [18]. This POLη overexpression protected cancer cells in the presence of the anti-cancer agents and siRNA-knockdown of POLη restored sensitivity to the first-line agents. Additional studies demonstrated that reduced mRNA expression of REV3, which is the catalytic subunit of POLζ, significantly reduced both mutation rate and acquired resistance following cisplatin administration in human fibroblasts [19]. Knockdown of REV1 expression in ovarian cancer cell lines also reduced cisplatin-induced mutagenesis and resistance [20,21]. POLη expression is up-regulated in ovarian cancer stem cells (CSCs) and its knockdown re-sensitized CSCs to cisplatin in vitro and in vivo [22]. In a preclinical model of late-state non-small cell lung cancer (NSCLC), selective knockdown of REV3 increased sensitivity to cisplatin in vitro and in vivo and significantly increased overall survival [23]. Similar results were identified following REV3 inhibition in cervical cancer cells [24]. Depletion of either REV1 or REV3 in a model of B-cell lymphoma increased sensitivity to cisplatin in vitro and in vivo [25]. Decreased expression of REV7, a key accessory subunit of POLζ, in cisplatin-resistant testicular germ cell tumors (TGCT) restored cisplatin sensitivity in vitro and in vivo [26]. Microarray analysis of these TGCT cells demonstrated that REV7 depletion correlates with down-regulation of other DNA repair genes and up-regulation of apoptosis-associated genes. Finally, co-delivery of a REV1- or REV3-specific siRNA and a cisplatin prodrug encapsulated in a nanoparticle abrogated tumor growth in a xenograft model of prostate cancer [27].

Additional studies exploring TLS and chemoresistance have identified a potential role for inhibition of TLS as a combination therapy with several other DNA damaging agents. REV7 depletion in the TGCT models described above enhanced sensitivity to the alkylating agent mitomycin C and the DNA intercalator doxorubicin [26]. Reduced expression of REV1 in a murine model of Burkitt’s lymphoma decreased mutations and acquired resistance to the alkylating agent cyclophosphamide, which resulted in tumor regression and an increase in overall survival [25]. REV3^−/− cells demonstrated hypersensitivity to the alkylating agents temozolomide (TMZ) and fotemustine [28]. While combination studies with genotoxic agents other than cisplatin have received less interest, it will be important to continue exploring the potential synergism between inhibition of TLS and a wide range of first-line chemotherapeutics.

3. Small Molecule Inhibitors of TLS

There are two primary approaches to inhibiting TLS in cancer cells: inhibiting enzyme activity or disrupting key PPIs in the heteroprotein complex. Small molecule inhibitors of DNA polymerases are utilized for multiple human diseases, including cancer, viral infections, and autoimmune disorders [29,30]. The clinical efficacy of these compounds clearly validates inhibition of polymerase catalytic activity as a viable strategy for TLS inhibition. While there is overlap between normal DNA and TLS polymerase catalytic mechanisms, there are significant structural differences between TLS and replicative polymerases as well as between the various Y-family TLS polymerases, which provides the opportunity to develop selective inhibitors of TLS DNA polymerases [31,32]. On the other hand, there is redundancy across the TLS DNA polymerases with respect to the specific lesions bypassed [6,33]; therefore, an inhibitor of catalytic activity might need to inhibit multiple ‘inserter’ TLS polymerases to be most efficacious. The second approach to developing small molecule inhibitors of TLS is to target the essential PPIs that govern proper assembly of the TLS complex and mediate its access to DNA. In this section, we provide rationale for targeting specific PPIs in the TLS complex and highlight the current state of small molecule TLS inhibitors that function through both strategies.

3.1. Targeting TLS PPIs.

3.1.1. Essential protein-protein interactions in TLS.

Because TLS is inherently mutagenic, cells have evolved a complicated network of PPIs that regulate TLS function to ensure that the TLS heteroprotein complex is only recruited to sites of DNA damage. More specifically, these PPIs fine-tune assembly of the TLS complex, they localize the proper TLS DNA polymerase to the lesion site, and they regulate polymerase switching events between replicative and TLS polymerases [6-9]. These PPIs also allow the TLS complex to adopt different configurations depending on the type of DNA lesion, where the lesion occurs, and what specific step in TLS is currently underway. For this review, we have chosen to highlight several key PPIs that are both essential for proper REV1/POLζ-dependent TLS function and are potentially amenable to targeting with small molecules.

Within the TLS heteroprotein complex, PCNA exists as a homotrimer that can directly interact with DNA as well as all the other Y-family TLS DNA polymerases. A key initiation step of TLS is monoubiquitination of PCNA, which promotes PPIs between PCNA and the ubiquitin-binding motifs (UBMs) of REV1 and POLι or the ubiquitin-binding zinc fingers (UBZs) of POLη and POLκ [34]. Replicative polymerases do not contain UBM/UBZ domains; therefore, an inhibitor of the PPI between PCNA and the UBMs/UBZs could selectively inhibit TLS-mediated replication [35]. PCNA also interacts with POLη, POLι, and POLκ through non-canonical PCNA interacting protein box (PIP-box) motifs defined by the conserved sequence (M/K/R)xx(M/L/V)xx(F/Y)(F/Y) (Figure 1A) [36-39]. The PIP-box motif binding site on PCNA is clearly defined and can accommodate side chains from multiple amino acids, making it a promising target for developing small molecules that could disrupt this essential TLS PPI (Figure 1B-1D).

Figure 1.

Domains and structure of PCNA/PIP-Box PPIs. (A) Sequences of TLS polymerase PIP-box motifs. PCNA homotrimer complexed with POLκ (black, B, PDB ID 2ZVL), POLι (blue, C, PDB ID 2ZVM), POLη (red, D, PDB ID 2ZVK).

The primary role of REV1 in REV1/POLζ-dependent TLS is to serve as a scaffolding protein that mediates assembly of the TLS heteroprotein complex and recruits the ‘extender’ polymerase POLζ to the DNA lesion [6]. In addition to interacting with PCNA through its two UBM domains, REV1 also forms a PPI with PCNA through its BRCA1 C-terminus (BRCT) domain, which is localized to the N-terminus of REV1 [40]. The REV1-BRCT domain interacts with PCNA via the PIP-box motif-binding site on PCNA, suggesting that small molecule binders of this site on PCNA could inhibit PPIs between multiple essential TLS polymerases [41]. The second domain of REV1 that forms key PPIs to maintain proper TLS structure and function is the C-terminal domain (REV1-CT). REV1-CT is a four-helix bundle domain that mediates PPIs between REV1 and all other TLS polymerases. REV1-CT utilizes opposite interfaces to bind REV1-interacting regions (RIR) present in the ‘inserter’ TLS polymerases POLη, POLι, POLκ, and the accessory POLD3 subunit of POLζ (Figure 2) while simultaneously binding to the accessory REV7 subunit of POLζ (Figure 3) [42,43]. While the entire RIR motif is generally considered to be 16 amino acid residues, structural and mutational studies have repeatedly demonstrated the importance of a pair of phenylalanine residues flanked by an N-cap residue and four helix forming residues (Figure 2A). At the binding interface, one of these Phe residues penetrates a preformed binding site on the REV1-CT domain while the other Phe residue interacts with a surface groove located adjacent to the pocket [42-46]. On its opposite surface, REV1-CT interacts with the POLζ subunit REV7 through a combination of hydrophobic contacts and intermolecular hydrogen bonds between α-helices H2-H4 of REV1-CT and β-sheets 8 and 8” of Rev7 (Figure 3A and 3B) [42,43,47,48].

Figure 2.

Structures of REV1-CT/RIR complexes. (A) Sequences of RIR motifs in TLS DNA polymerases. Structures of REV1-CT in complex with key RIR recognition motifs from POLη (B, PDB ID 2LSK), POLκ (C, PDB ID 2LSI), and POLD3 (D, PDB ID 2N1G).

Figure 3.

Structure of the quaternary POLκ/REV1-CT/REV7/REV3 complex. (A) Full POLκ/REV1-CT/REV7/REV3 complex structure (PDB ID 4FJO). (B) and (C) Key amino acid residues at the REV7/REV3 interface.

Another essential PPI in the TLS complex is between the REV7 and REV3 subunits of POLζ (Figure 3). REV7 is classified as a HORMA domain, which interacts with other proteins through a ‘safety-belt’ mechanism in which the domain transitions from an open to closed state in the presence of a REV7 binding motif (RBM) in REV3. In the closed state, the REV3-RBM is locked underneath the ‘safety-belt’ of REV7 [49]. Interestingly, two distinct RBMs have been characterized in REV3 and recent studies demonstrated that these dual RBMs promote formation of a REV7 homodimer that is essential for proper TLS function [50,51].

3.1.2. Inhibitors of the PCNA/PIP-box PPI.

A screen of approximately 38K small molecules at St. Jude Children’s Hospital identified the thyroid hormone 3,3’,5-triiodothyronine (T3, 1, Figure 4) as a modest inhibitor (IC₅₀ ~3 μM) of the PCNA/PIP-box PPI [52]. A co-crystal structure of T3 complexed with the PIP-box interface of PCNA was utilized to design an improved analogue (T2AA, 2) that retained the ability to disrupt the PCNA/PIP-box PPI, but no longer exhibited thyroid hormone activity. Additional structural studies of T2AA complexed with PCNA demonstrated that this compound binds to PCNA in a bimolecular fashion with one molecule in the PIP-box binding region and the second molecule in a shallow cavity adjacent to Lys164 [53]. T2AA prevented interactions between PCNA and POLη/REV1, inhibited ICL repair, and promoted cisplatin-mediated double strand crossbreaks. Finally, T2AA (10 μM) reduced clonogenic survival of HeLa and U2OS cancer cells when combined with cisplatin [53].

Figure 4.

Small molecule inhibitors of TLS PPIs.

These same researchers followed up their initial studies by synthesizing and evaluating a series of T2AA analogues to generate SAR for the scaffold with respect to its ability to inhibit TLS [54]. In general, removal of the 3’-iodine and replacement of the primary alcohol with substituted amides provided the most active compounds (represented by 3 and 4) with IC₅₀ values for disruption of the PCNA-PIP box PPI of between 1.3 and 1.5 μM. Both of these compounds inhibited TLS-dependent replication and increased cisplatin-mediated DNA damage. Finally, treatment of U2OS cells with 3 or 4 did not reduce cellular growth; however, the combination of T2AA analogue (15 μM) and cisplatin significantly reduced cell growth in this cell line [54]. These researchers continued their exploration of this scaffold by preparing an irreversible inhibitor (T2Pt, 5) through structure-based design [55]. This compound combined the 3/4 scaffold and a platinating moiety to specifically target two residues on PCNA (Met40 and His44) that are essential for the PPI. When combined with cisplatin, T2Pt was able to sensitize cells to cisplatin at a modest concentration (30 μM).

3.1.3. Inhibitors of the PCNA/REV1 PPI.

Based on the ability of T2AA to bind the cavity on PCNA in close proximity to Lys164, the same researchers at St. Jude Children’s Hospital exploring PCNA/PIP-box inhibitors utilized a related AlphaScreen approach to evaluate analogues of T2AA for their ability to disrupt the PPI between monoubiquitinated PCNA (UbPCNA) and REV1 (via a REV1/REV7/REV3 complex) [56]. Two compounds (6 and 7, Figure 4) demonstrated modest inhibition of the UbPCNA/REV1 PPI with IC₅₀ values of 3.4 and 9.7 μM, respectively. Interestingly, NMR studies showed that compound 7 bound to the UBM of REV1, but studies exploring binding to PCNA were not reported. To further explore the anti-TLS activity of these compounds, 6 was evaluated for its ability to affect TLS-mediated replication in the presence of UV-induced cyclopyrimidine dimers (CPD). Compound 6 delayed removal of UV-induced DNA crosslinks and reduced UV-induced mutations of the HPRT gene. Finally, compound 6 (75 μM) decreased clonogenic survival of U2OS cells when combined with either cisplatin or 4-hydroxycyclophosphamide [56], suggesting compound 6 enhances the efficacy of chemotherapy drugs, albeit at high concentrations.

3.1.4. Inhibitors of REV1-CT/RIR PPIs.

Collaborative researchers at the University of Connecticut developed and implemented a fluorescence polarization (FP)-based biochemical screen (~10K compounds) to identify small molecules capable of disrupting the REV1-CT/RIR PPI. This screen led to the identification of multiple scaffolds with moderate activity in the initial FP assay (IC₅₀ values range 1-15 μM) with several compounds demonstrating the ability to synergistically enhance the efficacy of cisplatin in human cancer cell lines [57,58]. Of these, the best-characterized compound is 8 (Figure 4), which is moderately active in the FP assay (IC₅₀ = 4.1 μM), but has demonstrated a wide range of anti-TLS activity across multiple different cell lines [57,59,60]. Depending on the cell type, compound 8 alone or in combination with cisplatin reduces the clonogenic survival of human cancer cells derived from multiple different tissue types, including primary ovarian cancer ascites. In addition, 8 reduces cisplatin-induced mutagenesis in a human fibrosarcoma cell line [57]. NMR studies demonstrated clear binding interactions between compound 8 and REV1-CT and detailed analysis of the chemical shift perturbations (CSPs) for REV1-CT localized these binding interactions to the RIR-interface on REV1-CT. Compound 8 has also been utilized as a chemical probe to explore the mechanism(s) through which inhibition of REV1/POLζ-dependent TLS inhibits cancer cell growth and progression [59,60]. These studies revealed that activation of TLS subverts the replication stress response in cancer cells by promoting global replication during stress and restricting ssDNA gaps. Within this context, compound 8 re-sensitized cells to treatment with small molecule inhibitors of ATR and WEE1, which function primarily by inducing replication gaps [59].

As noted above, two Phe residues flanked by an n-cap residue and four helix forming residues (-nFFhhhh-) is an essential RIR motif for binding REV1-CT. The spatial and conformational arrangement of the two Phe residues from the POLη RIR were utilized as a structural model to identify the phenazopyridine (PAP, compound 9, Figure 4) scaffold as a mimic of the FF motif that could disrupt the REV1-CT/RIR PPI (IC₅₀ = 0.99 μM) [61]. Initial structure-activity relationships (SAR) for the PAP scaffold focused on incorporating modifications to the phenyl ring to reduce rotation of the diazo linker while also appending small amino acids to the amines on the pyridine ring. These studies ultimately resulted in the identification of PAP analogue 10, which demonstrated enhanced potency for disrupting the PPI (IC₅₀ = 0.39 μM) and also bound directly to REV1-CT as determined through fluorine-19 NMR experiments.

The PAP scaffold was utilized to generate three- and four-point pharmacophore models that were used in a structure-based virtual screen (~150K compounds) to identify additional small molecule inhibitors capable of disrupting the REV1-CT/RIR PPI [62]. An initial screen to determine which compounds satisfied the key pharmacophoric features of the model was followed by a docking-based virtual screen to identify an initial series of 33 hit compounds that maintained the key pharmacophoric features of the reference PAP structure and were predicted to bind with high affinity to the RIR interface on REV1-CT. From these 33 hit compounds, 8 compounds exhibited IC₅₀ values in the FP displacement assay less than 15 μM. The three most potent compounds (compounds 11-13) enhanced cisplatin-mediated cell killing of mouse embryonic fibroblasts (MEFs) at low micromolar concentrations [62].

A series of stapled peptides derived from the POLκ RIR (residues 566-575, i, and i + 4) was recently designed and synthesized at MIT to evaluate their ability to bind REV1-CT and inhibit TLS [63]. Stapled peptides bound with greater affinity to REV1-CT (K_d ~13 μM) than the corresponding non-stapled versions. Replacement of the hot spot residue Phe568 with either cyclohexylalanine or naphthylalanine resulted in an approximated 3-fold increase in binding affinity for REV1-CT (K_d values ~4 – 6 μM). The stapled peptide containing the cyclohexylalanine residue disrupted the REV1-CT/RIR PPI (IC₅₀ = 6.6 μM) and reduced cisplatin-induced mutagenesis in the HT1080 human fibrosarcoma cell line [63].

3.1.5. Inhibitors of REV1-CT/REV7 PPIs

A research team at Duke University screened ~10K small molecules in an ELISA-based assay to identify small molecule inhibitors of the REV1-CT/REV7 PPI [64]. From this screen, compound 14 (Figure 4) was identified as a promising lead compound that disrupted the PPI in a dose-dependent manner (IC₅₀ = 0.78 μM) and directly bound to REV1-CT (K_d = 0.42 μM). Structural analysis of the co-crystal complex between 14 and REV1-CT demonstrated that the compound binds at the REV7 interface of two distinct REV1-CT monomers, resulting in a REV1-CT dimer, which forms a binding pocket that completely encloses the compound. This binding model ultimately conceals the REV7 interface on REV1-CT, preventing the PPI and blocking recruitment of POLζ to the TLS complex. Compound 14 enhanced the cytotoxicity of cisplatin in a panel of human cancer cell lines in a REV1-dependent manner and reduced cisplatin-mediated mutagenesis. Direct intratumoral injection of the combination regimen of 14 (1.6 mg/kg) and cisplatin (1.0 mg/kg) decreased tumor volume and increased survival compared to either single agent regimen [64].

3.1.6. Inhibitors of REV7/REV3 PPIs

A collaborative research group at St. Jude Children’s Hospital developed an AlphaScreen assay to identify compounds that inhibit the REV7/REV3 PPI [65]. For this assay, ~8.4K compounds were evaluated for their ability to disrupt a complex between histidine-labeled REV7 (His-REV7) and biotin-labeled Rev3 [REV3(1875–1895)-biotin]. This screen and follow-up studies to remove false positives identified compound 15 (Figure 4) as a modest inhibitor of the REV7/REV3 PPI. To establish SAR for the scaffold, several analogues of the hit compound were synthesized and evaluated for their ability to disrupt the PPI in a related FP-based assay. Replacement of the furan generally resulted in inactive compounds. The addition of a 5-methyl group on the furan ring (16) or modifications to the piperidine were well-tolerated. Compound 16 (IC₅₀ = 78 μM) was chosen for follow-up NMR studies that showed the compound bound directly to REV7. In addition, a cell survival assay was performed to validate the chemotherapeutic potential of compound 16. Co-administration of cisplatin (300 nM) and 16 (78 μM) reduced the clonogenic growth of HeLa cells [65]. In order to better understand the intermolecular binding interactions between this scaffold and REV7 a series of computational studies (docking, molecular dynamics, and MM/PBSA free energy calculations) were undertaken for the previously synthesized compounds [66]. These studies suggest that the 5-methylfuran moiety is essential for binding and that hydrophobic van der Waals interactions with multiple binding site residues at the REV3 interface of REV7 (Leu149, Trp171, and Leu173) contribute strongly to the free binding energy for the most potent compounds.

3.2. Inhibitors that Target TLS Polymerase Activity.

A fluorescent reporter strand displacement assay (~16K compounds) and an orthogonal detection method were used to identify 60 hits as inhibitors of POLκ [67]. The orthogonal assay was a gel-based primer extension assay using an acrolein-derived ring-opened reduced form of the γ-HOPdG lesion that is typically bypassed by POLκ. This assay identified three compounds as the most potent inhibitors; candesarten cilexetil (17, IC₅₀ = 11 μM), manoalide (18, IC₅₀ = 5.6 μM), and MK-886 (19, IC₅₀ = 14 μM) (Figure 5). These IC₅₀ values correlated well with the ability of the compounds to inhibit POLκ in the initial screening assay. Candesartan cilexetil enhanced UV-induced cellular toxicity and primer extension reactions concluded that this compound does not selectively inhibit POLκ by intercalating into the DNA. Candesarten cilexetil also demonstrated comparable activity against TLS polymerases POLη and POLι. Manoalide and MK-886 inhibited POLκ on a non-damaged DNA template and a lesion-induced DNA damage template, but were unable to enhance UV toxicity in cultured cells [67]. Follow-up studies with MK-886 demonstrated that it was 6- to 8-fold more potent against POLι compared to the other Y-family TLS polymerases [68]. Docking studies aimed at better probing the intermolecular interactions that govern its improved activity against POLι identified two potential binding pockets present in all three polymerases and a third potential binding pocket distinct to POLι.

Figure 5.

Small molecule inhibitors of TLS DNA polymerases.

A small library of compounds (380) targeted against nucleic acid interacting proteins was evaluated for inhibitory activity against POLη by researchers at the University for Arkansas Medical Sciences [69]. A robust and quantitative assay that has been used extensively to identify small-molecule inhibitors of Y-family and other DNA polymerase families was utilized to evaluate polymerase activity over a given period of time. This assay quantifies polymerase-catalyzed displacement of a fluorescently-labeled (TAMRA) oligonucleotide. Initially, 28 potential hits were identified in this screen, but follow-up assays narrowed this pool of compounds for further evaluation down to the indole thiobarbituric acid derivative ITBA-12 (20, Figure 5), which inhibited POLη activity with an IC₅₀ value of 29.8 μM. To determine the specificity of ITBA-12 for POLη, the compound was evaluated against a panel of seven additional polymerases. Of these, inhibition of POLβ, POLγ and POLκ was comparable with IC₅₀ values of 41, 49, and 59 μM, respectively while ITBA-12 was significantly less active against the remaining polymerases. Mechanistic studies on the mode of polymerase inhibition by ITBA-12 demonstrated that it acts as a partial allosteric competitive inhibitor of dNTP binding. Preliminary SAR on a small series of ITBA-12 analogues identified ITBA-16 and ITBA-19 (21 and 22, respectively), which incorporate an N-napththyl substituent in place of the bromophenyl, as improved inhibitors of POLη (IC₅₀ values = 16 and 17 μM, respectively) [69].

Additional SAR for the ITBA scaffold identified compound 23 (Figure 5) as a slightly more potent inhibitor of POLη compared with the previously described compounds (IC₅₀ = 8 μM) [70]. Compound 23 inhibited the catalytic activity of REV1 and the X-family DNA polymerase POLλ with equal potency (IC₅₀ values = 8 μM for both polymerases). Compound 23 enhanced sensitivity of myeloid leukemia cells to cisplatin in a POLη-dependent fashion. Finally, computational and fingerprinting assays strongly suggest that 23 interacts with a binding site on the little finger domain of POLη and prevents the template DNA from orienting properly in the TLS complex. Continued SAR on this scaffold focused on replacing the thiobarbuturic acid moiety with an aminoguanidine identified IAG-10 (24) as a selective inhibitor of POLκ (IC₅₀ value = 7.2 μM) [71]. Chemical footprinting suggested that inhibition of POLκ by 24 is based on its ability to disrupt the interaction between the N-clasp and TLS POL core, which is distinct from the mechanism through which the structurally related ITBA compounds inhibited POLη. Finally, compound 24 decreased clonogenic survival and increased temozolomide-induced DNA damage in a POLκ-dependent fashion in a cellular model of myeloid leukemia [71].

Multiple natural products primarily derived from fungal species have been characterized as non-selective inhibitors of replicative and TLS DNA polymerases [72-76]. In general, these compounds are moderate to weak inhibitors of catalytic polymerase activity and cellular anti-TLS or anti-cancer activity has not been described. With this in mind, detailed descriptions of these compounds and their activities are not included herein.

3.3. Small molecules that inhibit TLS through other mechanisms.

Researchers from Cleveland State University have recently reported on studies to evaluate the ability of synthetic nucleosides to inhibit TLS and synergize with the DNA damaging agent temozolomide (TMZ) [77,78]. Both 5-nitroindolyl-2’-deoxyriboside (5-NidR) and 3-ethynyl-5-nitroindolyl-2’-deoxyriboside (3-Eth-5-NidR) (25 and 26, respectively; Figure 6) synergized with TMZ to decrease the number of viable cells in several immortalized cancer cell lines. Interestingly, 3-Eth-5-NIdR did not potentiate the cell killing effects of other anti-cancer drugs that form DNA crosslinks (cisplatin, chlorambucil, and carmustine) or single-strand DNA (hydroxyurea), suggesting that the non-natural nucleoside selectively inhibits TLS in the presence of abasic sites or double strand breaks [77]. Combination therapy of 5-NIdR (100 mg/kg) and TMZ (40 mg/kg) in a murine flank allograft model of glioblastoma significantly decreased tumor volume and increased survival compared to a monotherapy regimen with either single agent [78].

Figure 6.

TLS inhibitors that function through other mechanisms.

As noted above, PCNA forms a homotrimer prior to its incorporation in the TLS DNA complex. To identify compounds that could inhibit DNA replication associated with PCNA, a docking-based virtual screen (~300K compounds) to identify scaffolds predicted to bind with high affinity at the interface between PCNA monomers was performed [79]. This screen suggested PCNA-I1 and PCNA-I2 (27 and 28, respectively, Figure 6) would bind with enhanced affinity to PCNA and these compounds, along with several structurally related analogues available in the ZINC database, were evaluated in a series of in vitro assays. Interestingly, PCNA-I1 bound with high affinity to the PCNA trimer (K_d values = 0.14 - 0.41 μM), strongly suggesting that its subsequent ability to inhibit PCNA activity was not a function of disrupting the homotrimer complex. PCNA-I1 reduced the association of PCNA with chromatin in a dose-dependent fashion and selectively inhibited the growth of tumor cells from various tissue types compared to normal cells (GI₅₀ values ~0.2 and ~2 μM, respectively) [79,80]. Finally, PCNA-I1 reduced clonogenicity of prostate cancer cell lines (1 μM) and reduced tumor volume in a mouse model of prostate cancer following intravenous injection (10 mg/kg) [80].

Based on the promising results of PCNA-I1, mutational and SAR studies were performed to identify compounds with improved activity and more fully characterize the mechanisms through which the scaffold inhibits PCNA activity [81,82]. PCNA-I1 did not affect the ability of PCNA containing point mutations at the predicted binding site to associate with chromatin, supporting the previous docking studies localizing the binding pocket to the interface of two individual PCNA molecules. Extensive SAR studies established key features for this class of compounds. The Z-configuration around the central hydrazone was preferred and the hydroxynaphthyl moiety was essential for activity. On the ‘right-side’ of the scaffold, phenyl and pyrazole moieties were preferred as were functional groups in the ortho-position capable of hydrogen bonding. Two compounds representing this overall SAR trend (29 and 30) inhibited cell growth in multiple cell lines with GI₅₀ values of 0.09 – 0.38 μM and stabilized PCNA trimerization [81,82]. It is important to note that the ability of these compounds to specifically regulate TLS or enhance the sensitivity of cancer cells to first-line anti-cancer agents have not been reported.

4. CONCLUSION

The identification that REV1/POLζ-dependent TLS plays an essential role in allowing cancer cells to survive genotoxic chemotherapies has opened the door to developing small molecule TLS inhibitors as adjuvant agents that can potentially enhance the short- and long-term efficacy of these compounds. Multiple research groups have identified and characterized compounds that inhibit TLS function by targeting a variety of different proteins in the TLS heteroprotein complex. Many of these compounds have demonstrated initial anti-cancer effects in vitro against a wide range of cancer subtypes, particularly when combined with cisplatin. The continued development of TLS inhibitors during the coming years will serve to fully define their clinical potential.

5. EXPERT OPINION

Drug resistance is a well-known challenge for many first-line anti-cancer therapeutics. It is common for tumors to respond initially to the chemotherapy, but ultimately relapse to be more aggressive and drug-resistant, which can limit the long-term effectiveness of these agents. This is especially common for first-line genotoxic agents that function by directly damaging the DNA in rapidly dividing cancer cells. Cancer cells that survive first-line genotoxic drugs exhibit increased mutation rates, which ultimately drives the acquired resistance and tumor relapse often seen following initial treatment. Recent studies into the cellular mechanisms that allow cancers to survive DNA damaging agents has identified an essential role for TLS in this process [4,6,11]. With this in mind, developing small molecule inhibitors of the TLS polymerase machinery has emerged as a promising therapeutic strategy for enhancing the efficacy of first-line chemotherapy and mitigating acquired cellular resistance to DNA damaging agents. Multiple studies have clearly demonstrated that inhibition of REV1/POLζ-dependent TLS can sensitize several types of cancer to first-line genotoxic agents while reducing the mutagenesis that can lead to resistance and relapse [20-21,23-26]. While recent years have seen a steady increase in the number of small molecule TLS inhibitors undergoing preclinical evaluation, several considerations need to be addressed before the full clinical scope of this class of compounds can be realized.

First, there are multiple distinct proteins within the TLS heteroprotein complex that are being explored as viable targets for small molecule development. To date, the primary focus has been on targeting either (1) the catalytic activity of an ‘inserter’ TLS polymerase (POLι, POLη, or POLκ) or (2) an essential TLS PPI. Both of these strategies have demonstrated initial promise; however, further studies are needed to determine whether a particular polymerase or specific PPI is most promising. This is of particular importance with respect to direct inhibition of several ‘inserter’ polymerases, which are known to play important roles in nucleotide excision repair and base excision repair independent of their roles in REV1/POLζ-dependent TLS [83,84]. There is also the possibility of selectivity and off-target effects for compounds that directly target the catalytic activity of the ‘inserter’ polymerases. The majority of polymerase inhibitors described above demonstrate broad, non-selective inhibition against multiple polymerases (TLS and non-TLS) suggesting that a significant hurdle for these compounds will be selective activity against the desired TLS ‘inserter’ polymerase. In addition, while inhibitors of the catalytic activity of the ‘extender’ polymerase POLζ may hold promise as TLS inhibitors, compounds targeting its catalytic subunit REV3 have not been disclosed. Much is still unknown about TLS; new polymerases that aid in the process of DNA damage bypass are still being characterized. A key to determining whether a specific protein/PPI is optimal will be understanding how inhibition of these individual targets affects REV1/POLζ-dependent TLS in normal cells and whether a specific approach can more selectively target cancer cells.

A related consideration for the continued optimization of these compounds is the development of assays that can be utilized to determine whether selective inhibition of TLS is responsible for the anti-cancer effects of these compounds. The majority of cellular assays used for these compounds explore their ability to enhance tumor cell killing or reduce mutation rates in the presence of first-line agents. While both of these phenotypic responses have been linked to inhibition of TLS, they could also result from these compounds targeting other cellular proteins. This is of particular importance for compounds that are designed to disrupt a specific TLS PPI. The ability to directly correlate cellular inhibition of a TLS PPI with enhanced efficacy of a first-line genotoxic agent will be essential for determining on-target vs. off-target effects for these small molecules.

Finally, expanding our understanding of which DNA damaging agents will synergize with TLS inhibitors is important to the preclinical progression of this class of compounds. As noted above, the majority of small molecule TLS inhibitors have been evaluated as combination therapies with cisplatin. Combination studies with alkylating agents have shown initial promise, but more data is needed before determining whether synergism between TLS inhibitors and alkylating agents will be clinically efficacious. To date, combination regimens consisting of a TLS inhibitor and ionizing radiation have not been explored, but synergism between these two classes of compounds might also prove effective.

With all this information in mind, we have collated key data in Table 1 for what we believe to be the most promising TLS inhibitors reported to date. These compounds target a wide range of proteins involved in REV1/POLζ-dependent TLS and each of them demonstrate inhibitory activity against their respective target in the low or sub-micromolar range. Several of the compounds described above and several classes of compounds not included in this review are less active against their target and their path to a potential clinical trial is unlikely or will include significant medicinal chemistry optimization of the scaffold. The majority of compounds in Table 1 have also demonstrated the ability to synergize with one or more first-line genotoxic agents in cancer cells. Two of these TLS inhibitors, 14 and 25, have demonstrated anti-cancer activity in animal models of human cancer; however, caveats exist for the further development of both of these compounds. First, compound 14 inhibits the REV1-CT/REV7 PPI through an indirect, non-conventional mechanism (dimerization of REV1) and its in vivo activity was a result of direct intratumoral injection, not the more rigorous intraperitoneal or oral administration. The anti-cancer activity of compound 25 was dependent on the type of DNA lesion present (abasic sites or double strand breaks), suggesting that this compound may not be as broadly active as other TLS inhibitors in development. Overall, the development of small molecule inhibitors of TLS has emerged as a promising strategy to enhance the efficacy of first-line genotoxic agents and the continued identification and development of these compounds will enhance our understanding of TLS as a viable anti-cancer drug target.

Table 1.

In Vitro and in vivo activity for key TLS inhibitors identified to date.^a

Cmpd	TLS Target	Activity against target (IC50, μM)b	Activity in cellular assays (μM range)c	Evaluated in vivo?	Key References
T2AA	PCNA/PIP-box PPI	1.3	10 – 15	No	52-54
6	UbPCNA/REV1 PPI	3.4	~75	No	56
8	REV1-CT/RIR PPI	4.1	1.5 – 20	No	57,59-60
10	REV1-CT/RIR PPI	0.39	NDd	No	61
14	REV1-CT/REV7 PPI	0.78	1.5 – 10	Yes, active	64
24	POLκ	7.2	0.5	ND	71
25	DNA replication	Not applicable	10	Yes, active	77,78
PCNA-I1	PCNA Trimer	0.14 – 0.41	ND	No	79,80

^aThis table summarizes key data for the most promising inhibitors of TLS reported to date. Each of these compounds demonstrate moderate to potent activity against their individual target, and the majority synergize with first-line genotoxic agents in cell culture. Two compounds, 14 and 25, have also demonstrate anti-cancer activity in vivo.

^bData in this column represent the most appropriate biochemical assay to measure direct inhibition of the compound target. See the text for a more complete description.

^cData in this column represent concentrations at which the compound inhibited cellular TLS or enhanced the effects of a first-line agent in cancer cells.

^dND = Not determined or not reported to date.

Article Highlights:

• Translesion synthesis (TLS) is a mechanism through which actively replicating cells bypass DNA damage to prevent cell death.

• TLS has been linked to the onset of chemotherapy resistance in multiple forms of cancer.

• The development of small molecule inhibitors of TLS is emerging as a promising field to enhance tumor sensitivity to first-line genotoxic drugs.

• Inhibitors of both TLS polymerase activity and key protein-protein interactions involved in TLS initiation and progression are in various stages of preclinical development.

• Small molecule TLS inhibitors have demonstrated anti-cancer activity as single agents or in combination regimens with approved first-line genotoxic agents.