Elsie M. Williams

Discovery and evaluation of Escherichia coli nitroreductases that activate the anti-cancer prodrug CB1954

Gene-directed enzyme prodrug therapy (GDEPT) aims to achieve highly selective tumor-cell killing through the use of tumor-tropic gene delivery vectors coupled with systemic administration of otherwise inert prodrugs. Nitroaromatic prodrugs such as CB1954 hold promise for GDEPT as they are readily reduced to potent DNA alkylating agents by bacterial nitroreductase enzymes (NTRs). Transfection with the nfsB gene from Escherichia coli can increase the sensitivity of tumor cells to CB1954 by greater than 1000-fold. However, poor catalytic efficiency limits the activation of CB1954 by NfsB at clinically relevant doses. A lack of flexible, high-throughput screening technology has hindered efforts to discover superior NTR candidates. Here we demonstrate how the SOS chromotest and complementary screening technologies can be used to evaluate novel enzymes that activate CB1954 and other bioreductive and/or genotoxic prodrugs. We identify the major E. coli NTR, NfsA, as 10-fold more efficient than NfsB in activating CB1954 as purified protein (k(cat)/K(m)) and when over-expressed in an E. coli nfsA(-)/nfsB(-) gene deleted strain. NfsA also confers sensitivity to CB1954 when expressed in HCT-116 human colon carcinoma cells, with similar efficiency to NfsB. In addition, we identify two novel E. coli NTRs, AzoR and NemA, that have not previously been characterized in the context of nitroaromatic prodrug activation.

Creation and screening of a multi-family bacterial oxidoreductase library to discover novel nitroreductases that efficiently activate the bioreductive prodrugs CB1954 and PR-104A

Two potentially complementary approaches to improve the anti-cancer strategy gene-directed enzyme prodrug therapy (GDEPT) are discovery of more efficient prodrug-activating enzymes, and development of more effective prodrugs. Here we demonstrate the utility of a flexible screening system based on the Escherichia coli SOS response to evaluate novel nitroreductase enzymes and prodrugs in concert. To achieve this, a library of 47 candidate genes representing 11 different oxidoreductase families was created and screened to identify the most efficient activators of two different nitroaromatic prodrugs, CB1954 and PR-104A. The most catalytically efficient nitroreductases were found in the NfsA and NfsB enzyme families, with NfsA homologues generally more active than NfsB. Some members of the AzoR, NemA and MdaB families also exhibited low-level activity with one or both prodrugs. The results of SOS screening in our optimised E. coli reporter strain SOS-R2 were generally predictive of the ability of nitroreductase candidates to sensitise E. coli to CB1954, and of the kcat/Km for each prodrug substrate at a purified protein level. However, we also found that not all nitroreductases express stably in human (HCT-116 colon carcinoma) cells, and that activity at a purified protein level did not necessarily predict activity in stably transfected HCT-116. These results highlight a need for all enzyme-prodrug partners for GDEPT to be assessed in the specific context of the vector and cell line that they are intended to target. Nonetheless, our oxidoreductase library and optimised screens provide valuable tools to identify preferred nitroreductase-prodrug combinations to advance to preclinical evaluation.

The Flavin Reductase MsuE Is a Novel Nitroreductase that Can Efficiently Activate Two Promising Next-Generation Prodrugs for Gene-Directed Enzyme Prodrug Therapy

Bacterial nitroreductase enzymes that can efficiently catalyse the oxygen-independent reduction of prodrugs originally developed to target tumour hypoxia offer great potential for expanding the therapeutic range of these molecules to aerobic tumour regions, via the emerging cancer strategy of gene-directed enzyme prodrug therapy (GDEPT). Two promising hypoxia prodrugs for GDEPT are the dinitrobenzamide mustard PR-104A, and the nitrochloromethylbenzindoline prodrug nitro-CBI-DEI. We describe here use of a nitro-quenched fluorogenic probe to identify MsuE from Pseudomonas aeruginosa as a novel nitroreductase candidate for GDEPT. In SOS and bacteria-delivered enzyme prodrug cytotoxicity assays MsuE was less effective at activating CB1954 (a first-generation GDEPT prodrug) than the “gold standard” nitroreductases NfsA and NfsB from Escherichia coli. However, MsuE exhibited comparable levels of activity with PR-104A and nitro-CBI-DEI, and is the first nitroreductase outside of the NfsA and NfsB enzyme families to do so. These in vitro findings suggest that MsuE is worthy of further evaluation in in vivo models of GDEPT.

Site-Saturation Mutagenesis by Overlap Extension PCR

Site-saturation mutagenesis is a proven strategy for generating high-quality variant gene libraries of a defined size. Variation is introduced via incorporation of degenerate base combinations at specific codon locations, giving rise to a precise series of amino acid substitutions in the encoded protein. Here we describe a simple and efficient overlap PCR protocol for the introduction of degenerate bases at either single or multiple codon locations. The resulting libraries can then be directly screened for improved protein function as either an independent directed evolution study or an adjunct to random mutagenesis strategies (such as error-prone PCR) that are, in isolation, unlikely to access the full repertoire of possible amino acid substitutions at any given position.

Toward a high-throughput screening platform for directed evolution of enzymes that activate genotoxic prodrugs

Engineering of enzymes to more efficiently activate genotoxic prodrugs holds great potential for improving anticancer gene or antibody therapies. We report the development of a new, GFP-based, high-throughput screening platform to enable engineering of prodrug-activating enzymes by directed evolution. By fusing an inducible SOS promoter to an engineered GFP reporter gene, we were able to measure levels of DNA damage in intact Escherichia coli and separate cell populations by fluorescence activating cell sorting (FACS). In two FACS iterations, we were able to achieve a 90,000-fold enrichment of a functional prodrug-activating nitroreductase from a null library background.

Abstract B88: Discovery, characterization, and engineering of bacterial nitroreductases for gene-directed enzyme prodrug therapy.

Tumor-targeting viruses and bacteria hold great promise as anti-cancer agents. They kill cells by entirely different mechanisms to radio- and chemotherapies, and have potential to synergize with these treatments without overlapping toxicities. Furthermore, these agents can be ‘armed’ with genes that encode enzymes that activate prodrugs - compounds that are deactivated in their administered form, but become highly toxic upon metabolic activation. This not only improves killing of infected cells, but also neighboring non-infected cells, as the prodrug metabolites can diffuse locally and exert a bystander effect. A highly efficient activating enzyme in partnership with a prodrug that has a strong bystander effect can address some of the historical limitations of cancer gene therapy including the inability of biological vectors to reach every target cell. Phase I/II trials of the first-generation nitroaromatic prodrug CB1954 in conjunction with the prototype gene therapy nitroreductase, Escherichia coli NfsB, have been conducted in prostate and liver cancer. These trials provided some evidence of anti-tumor activity but, due to dose-limiting hepatotoxicity, the highest achievable plasma concentration of CB1954 was less than 1% of NfsB9s K m . As well as highlighting a need for more efficient nitroreductase enzymes, this has fuelled a search for superior nitroaromatic prodrugs. The next-generation dinitrobenzamide mustard prodrug PR-104A is not only 5–50 fold more dose-potent upon activation, but also better tolerated in humans (MTD 1100 mg/m 2 vs 24 mg/m 2 ; q3w, iv). However, E. coli NfsB also has relatively poor (millimolar) affinity for this substrate. To discover more efficient nitroreductases we developed screens for genotoxic prodrug activation, based on their ability to induce reporter genes linked to the E. coli SOS (DNA damage repair) response. We used these to screen a large library of candidate enzymes for DNBM activity, and selected E. coli NfsA as a top candidate for further improvement by random and targeted mutagenesis. High throughput screening of large error prone PCR libraries coupled with medium throughput screening of targeted mutagenesis libraries revealed 10 individual mutations that significantly increased NfsA activity. These mutations were then combined in a synthetic “smart” library, from which eight poly-mutant enzymes were selected for kinetic analysis. Relative to wild type the engineered variants display an 18–40 fold improvement in PR-104A K m with respect to E. coli NfsA, and are 860–1700 fold better than NfsB. Importantly they also retain, or are improved in, their ability to co-metabolize preferred 2-nitroimidazole probes with PET-imaging capabilities (see abstract: Patterson et al, “Molecular imaging using bacterial nitroreductase reporter genes by repurposing the clinical stage hypoxia PET probe EF5”). The enhanced prodrug activation and in vivo imaging potential of these enzymes is now being evaluated in human gene therapy models. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2011 Nov 12-16; San Francisco, CA. Philadelphia (PA): AACR; Mol Cancer Ther 2011;10(11 Suppl):Abstract nr B88.

Abstract B89: Molecular imaging using bacterial nitroreductase reporter genes by repurposing the clinical stage hypoxia PET probe EF5.

Oncolytic viruses and tumor-tropic bacteria offer promise as cancer therapeutics of the future. There is a need to develop technologies to monitor the spatio-temporal distribution of these live vectors in a manner that is predictive of normal tissue toxicity and antitumor efficacy. Positron emission tomography (PET) is the preferred non-invasive imaging modality (biomarker) but suitable advanced reporter gene/PET probe combinations are lacking. A range of 2-nitroimidazole (2-NI) probes are currently in clinical use for the detection of hypoxia (EF5, FMISO, HX4, FAZA) and thus have already attained a high level of clinical validation. We hypothesized that (2-NI) PET agents might be repurposed to monitor the biodistribution of replicating biological agents, thereby leveraging two decades of research efforts to optimise hypoxia PET probe pharmacology. Bacterial nitroreductases (NTRs; type I) are efficient O2-independent enzymes that provide the necessary catalytic flexibility to achieve this goal. Historically E. coli NfsB has been the focus of gene therapy applications. We cloned eleven candidate NTRs from E. coli namely; AzoR, KefF, MdaB, NemA, NfsA, NfsB, WrbA, YcaK, YcdI, YdjA and Yief. Using HCT116 cells engineered to express each NTR, we showed NfsA alone was able to metabolise a range of 2-NI probe molecules, including EF5 (pentafluoroetanidazole), leading to efficient cellular retention. We confirmed catalytic efficiency of EF5 reduction by recombinant NfsA (kcat/Km 97 s−1/mM). In contrast NfsB only weakly metabolised EF5 (kcat/Km 0.24 s−1/mM). We compared EF5 adduct retention in HCT116 cells, detected by mAb using flow cytometry, either under anoxia ( 100-fold greater in NfsA expressing cells than anoxic cells, whereas NfsB cells were negative. Mixed Wt : NfsA multicellular layers containing various ratios of cells demonstrated the capacity to detect, with precision, single NfsA-positive cells in mixed cell populations. HCT116 xenografts composed of increasing proportions of NfsA cells (0% to 25%) were established in nude mice and analysed by immunohistochemistry and flow cytometry. Ex-vivo pimonidazole-binding confirmed all NfsA positive cells were detected in vivo following administration of EF5. NfsA, like NfsB, can bioactivate the clinical-stage prodrug PR-104. We determined the relationship between EF5 retention and PR-104 cytotoxicity in HCT116 xenografts harbouring variable proportions of NfsA cells (0%-25%). A correlation was observed between total EF5 retention and global clonogenic cell kill (r2 = 0.828; p Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2011 Nov 12-16; San Francisco, CA. Philadelphia (PA): AACR; Mol Cancer Ther 2011;10(11 Suppl):Abstract nr B89.