7-(2-Thienyl)-7-Deazaadenosine (AB61), a New Potent Nucleoside Cytostatic with a Complex Mode of Action

7-Deazapurine (IUPAC name: pyrrolo[2,3-d]pyrimidine) nucleosides have become a subject of research interest because of their powerful antibacterial, antifungal, and cytotoxic activities (1). Mechanisms of action of naturally occurring 7-deazapurine nucleosides tubercidin, toyocamycin, and sangivamycin (Fig. 1) have been extensively studied. Despite structural similarities, their biochemical and biologic properties show different features for each nucleoside. Tubercidin, toyocamycin, and sangivamycin are all phosphorylated in cells to their 5′-O-mono-, di-, and triphosphates which are subsequently incorporated into nucleic acids (2–4). In addition, tubercidin also inhibits nucleic acid and protein synthesis, interferes with mitochondrial respiration, rRNA processing, de novo purine synthesis (2), and it is a potent inhibitor of S-adenosylhomocysteine hydrolase (5). Toyocamycin showed inhibition of phosphatidylinositol kinase (6) and an inhibitory effect on rRNA synthesis and maturation (7, 8). On the other hand, sangivamycin acts mainly through potent and selective inhibition of protein kinase C (9).

In our previous research, nanomolar cytotoxic activities of hetaryl-7-deazapurine ribonucleosides were discovered. 6-Hetaryl-7-deazapurine ribonucleosides 1 showed strong cytotoxicity against a panel of cancer cell lines similar to that of conventional nucleoside cytostatics (gemcitabine, clofarabine; ref. 10). Similarly profound cytotoxic effects were also found in 7-hetaryl-7-deazaadenosines, that is, 7-hetaryltubercidins 2 (11). Structure–activity relationship studies revealed that modifications in the 2′ position of the ribose moiety lead to a decrease of cytotoxicity in 7-hetaryl-7-deazapurine ribonucleosides 2 (12, 13) and completely inactive compounds in 6-hetaryl-7-deazapurine ribonucleosides 1 (14, 15). In contrast, base-modified 7-hetaryl derivatives bearing modifications in positions 2 and 6 of 7-deazapurine showed submicromolar cytotoxic activities (16). The most promising derivative among the 7-hetaryl-7-adenosines is a 2-thienyl derivative AB61 (Fig. 1), which is not only highly cytotoxic against leukemic and solid tumor–derived cell lines but also noncytotoxic to normal human fibroblasts (11). In vivo antitumor activity of AB61 was demonstrated in a mouse leukemia survival model. The preliminary studies of its mechanism of action in CCRF-CEM lymphoblasts showed fast onset of the RNA synthesis inhibition and also inhibition of DNA synthesis at higher concentrations. AB61 is efficiently phosphorylated to its 5′-O-triphosphate (AB61-TP) but only very mild inhibition of RNA polymerase II by AB61-TP was observed (11). More recently, AB61 and its related derivatives were found (17) to be weak inhibitors of human adenosine kinase and AB61 itself was found to be a weak substrate for this enzyme (2%) as compared with adenosine.

The aims of this study were to elucidate the detailed mechanism of action of AB61 and understand why its cytotoxicity is selective for cancer cell lines. We hypothesized that AB61-TP may get incorporated into nucleic acids that have impaired biologic functions. Therefore, the first goal was to investigate phosphorylation of AB61 in normal and cancer-derived cell lines, incorporation of AB61 into nucleic acids both on enzymatic and cellular levels, and the effect of AB61 on gene transcription, translation, and DNA integrity. Furthermore, we have studied the in vivo antitumor activity of AB61 against xenotransplanted human solid tumors to translate promising in vitro data into preclinical efficacy.

Synthesis and characterization data of analytical standards AB61-DP, dAB61-MP, [³H]AB61 and tubercidin triphosphate (Tub-TP) are given in the Supplementary Data. Synthesis of AB61, AB61-MP and AB61-TP was published previously (11).

The CCRF-CEM (ATCC CCL-119), HCT116 (ATCC CCL-247), K-562 (ATCC CCL-243), BJ (ATCC CRL-2522), MRC-5 (ATCC CC-171), HT-29 (ATCC HTB-38), SK-OV-3 (ATCC HTB-77), U2OS (ATCC HTB-96), and BT-549 (ATCC HTB-122) cell lines were purchased from the American Tissue Culture Collection (ATCC) in 2010–2014. HCT116p53^−/− cell line was purchased from Horizon Discovery in 2011. The daunorubicin-resistant subline of CCRF-CEM cells (CEM-DNR-bulk) and paclitaxel-resistant subline K-562-tax overexpressing major drug resistance transporters were selected in our laboratory (18, 19). Mouse breast cancer 4T1-luc2 cell line stably transfected with firefly luciferase under cytomegalovirus promoter was a kind gift from Prof. Danuta Radzioch (McGill University, Montreal, Canada) in 2009. Cell line derived from human osteosarcoma U2OS stably transfected with 53BP1-GFP fusion gene (U2OS-53BP1-GFP) was prepared by Lipofectamine transfection of 53BP1 plasmid (ref. 20; a kind gift from Prof. Jiri Bartek, Danish Cancer Society Research Center, Copenhagen, Denmark) into parental cells and successful transfectants were selected by neomycin. Cell line authentication was performed using the Promega CELL ID TM System (8 STR markers + amelogenin) to verify that the genetic profile of the sample matches the known profile of the cell line. Mycoplasma contamination was tested using 1-My and 2-My primers by qPCR (21) for every batch of the frozen cells. The cell lines were cultured for maximum of 6 to 10 passages and tested for mycoplasma contamination on a weekly basis.

DNA oligonucleotides were purchased from Sigma-Aldrich and Generi Biotech. ON1: 5′-GCTAATACGACTCACTATAGGTGAGGTACTTGTAGTGATT (T7 promoter in italics); ON2: 5′-AATCACTACAAGTACCTCACCTATAGTGAGTCGTATTAGC (underlined nucleotides represent 2′-O-Me-RNA); Prim248short: 5′-CATGGGCGGCATGGG; Prb4basII: 5′-CTAGCATGAGCTCAGTCCCATGCCGCCCATG; OligoA^term: 5′—TCCCATGCCGCCCATG; (bio)-OligoA^term: 5′-biotin-TCCCATGCCGCCCATG.

The cells were maintained in Nunc/Corning 80 cm² plastic tissue culture flasks and cultured in cell culture medium (DMEM/RPMI1640 with 5 g/L glucose, 2 mmol/L glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 10% fetal calf serum, and NaHCO₃). Cell suspensions were prepared and diluted according to the particular cell type and the expected target cell density (25,000–30,000 cells/well based on cell growth characteristics). MTT assay was performed as described before and the IC₅₀ value, the drug concentration lethal to 50% of the cells, was calculated from appropriate dose–response curves (18).

HCT116 and BJ cells were seeded into 6-well plates at 60% confluence (McCoy 5A and DMEM, respectively, supplemented with 5 g/L glucose, 2 mmol/L glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 10% fetal calf serum, and NaHCO₃). Next day, cells were treated with AB61 (10 μmol/L). After 3-hour incubation, cultures were washed 2 times with PBS containing 0.5% FBS and 3 times with PBS. The cell pellets were extracted with 70% cold MeOH (1 mL). Supernatants were collected, dried under vacuum, and samples were resuspended in DMSO-water (1:1, 100 μL) for analysis. The samples were analyzed using ultra-performance reversed phase chromatography coupled to triple tandem mass spectrometry UPLC-MS/MS. The UPLC chromatograph used in this study was Accela Thermo Scientific system consisting of gradient quaternary pump, thermostated autosampler, degasser, column oven and triple quadrupole mass spectrometer TSQ Quantum Access (Thermo Scientific). Xcalibur data system software was used for an instrument control and data analysis. The C18 column (XBridge BEH C18, 2.5 μm, 2.1 × 50 mm) and ammonium acetate (pH 5.04)/acetonitrile gradient elution were applied for chromatographic separation. The electrospray ionization and negative selected reaction mode MS/MS were used for analyte quantification. Standard curves and quality control samples were generated for all analytes using extracts from untreated cells.

A solution of template oligonucleotides ON1 and ON2 (50 μmol/L each) in annealing buffer [Tris (10 mmol/L), NaCl (50 mmol/L), EDTA (1 mmol/L), pH 7.8] was heated to 95°C for 5 minutes and then slowly cooled to 25 °C over a period of 45 minutes. The resulting dsDNA (50 μmol/L) was used as a template for transcription reactions. In vitro transcription reactions (10 μL; ApliScribe T7-Flash Transcription Kit, Epicentre) were performed in the presence of AB61-TP or Tub-TP (0.45 mmol/L, for synthesis see supporting information), CTP, GTP, UTP (4.5 mmol/L each), DTT (10 mmol/L), ApliScribe T7-Flash 10× Reaction Buffer (2 μL), template (2.5 μmol/L), [α-³²P]GTP (111 TBq/mmol, 370 MBq/mL, 0.4 μL), and ApliScribe T7-Flash enzyme solution (2 μL). In the negative control experiment, water was used instead of the solution of the tested compound, and in the positive control experiment, ATP (0.45 mmol/L) was used instead of the tested compound. The transcription reactions were performed at 37°C for 2 hours. RNA was purified on NucAway Spin Columns (Ambion, elution in DEPC-water). Samples (2 μL) were mixed with RNA loading dye (Fermentas; 2 μL), denatured at 90 °C for 10 minutes and cooled on ice. The samples were analyzed by gel electrophoresis on 12.5% denaturing polyacrylamide gel containing 1× TBE buffer (pH 8) and urea (7 mol/L) at 45 mA for 45 minutes. The gels were dried (85°C, 75 minutes), audioradiographed, and visualized by phosphorimager (Typhoon 9410, Amersham Biosciences).

The primer extension experiments were performed under following conditions: Klenow fragment: the reaction mixture (20 μL) contained DNA polymerase I, large (Klenow) fragment (New England BioLabs, 5 U/μL, 0.04 μL), natural dNTPs (10 mmol/L, 0.4 μL), AB61-TP or Tub-TP (10 mmol/L, 1 μL), primer Prim248short (3 μmol/L, 1 μL), 31-mer template Prb4basII (3 μmol/L, 1 μL), and NEBuffer 2 (2 μL) supplied by the manufacturer. Prim248short was labeled by the use of [γ-³²P]ATP according to standard techniques. In the positive and negative control experiments, dATP (10 mmol/L, 1 μL) and water, respectively, were used instead of the tested compound. The reaction mixtures were incubated for 15 minutes at 25°C. Human DNA polymerase β: Primer Prim248short was labeled by the use of [γ-³²P]ATP and annealed with template Oligo^Aterm (primer:template ratio 1:1.5) according to the standard techniques. The reaction mixture (10 μL) contained human DNA polymerase β (CHIMERx, 5 U/μL, 0.01 μL), AB61-TP (1 mmol/L, 1 μL), primer:template mixture (1 μmol/L primer, 1.5 μmol/L template; 1 μL), BSA (acetylated, 24 mg/mL, 0.167 μL), and glycerol (1.5 μL) and 10× buffer for DNA polymerase β (1 μL) supplied by the manufacturer. In the positive and negative control experiments, dATP and water, respectively, were used instead of AB61-TP. The reaction mixtures were incubated for 3 hours at 37°C. Human DNA polymerase γ: Primer Prim248short was labeled by the use of [γ-³²P]ATP and annealed with template Oligo^Aterm (primer:template ratio 1:1.5) according to the standard techniques. The reaction mixture (10 μL) contained human DNA polymerase γ (CHIMERx, 10 U/μL, 0.05 μL), AB61-TP (1 mmol/L, 1 μL), primer:template mixture (1 μmol/L primer, 1.5 μmol/L template; 1 μL), BSA (acetylated, 24 mg/mL, 0.25 μL), and 10× MnCl₂ solution (1 μL) supplied by the manufacturer and 10× buffer for DNA polymerase γ (1 μL) supplied by the manufacturer. In the positive and negative control experiments, dATP and water, respectively, were used instead of AB61-TP. The reaction mixtures were incubated for 3 hours at 37°C. The reactions were stopped by addition of the same volume of PAGE stop solution [80% (v/v) formamide, 20 mmol/L EDTA, 0.025% (w/v) bromophenol blue, 0.025% (w/v) xylene cyanol] and heated to 95°C for 5 minutes. Aliquots (2 μL) were analyzed by gel electrophoresis on 12.5% denaturing polyacrylamide gel containing 1× TBE buffer (pH 8) and urea (7 M) at 45 mA for 50 minutes. The gels were dried (85°C, 75 minutes), audioradiographed, and visualized by phosphorimager (Typhoon 9410, Amersham Biosciences). The incorporation of AB61-TP into DNA oligonucleotide was confirmed by MALDI-TOF analysis (see Supplementary information Fig. S1).

CCRF-CEM cells (ATCC: CCL-119) were grown in RPMI1640 medium (Sigma) supplemented with 10% fetal calf serum, GlutaMAX (Gibco, 10 mL/L), penicillin (100 U/L), and streptomycin (100 mg/L). All experiments were done with exponentially growing cells. CCRF-CEM cells (7 × 10⁵ cells/mL, 10 mL) were incubated overnight in a humidified CO₂ incubator at 37°C. Solution of [³H]AB61 (for synthesis, see Supplementary Information “Synthesis of nucleosides and nucleotides”) in water (29 μL, 1 mCi/mL, 11.6 Ci/mmol) was added. The final concentration of [³H]AB61 was 250 nmol/L. Samples of the cell suspension (1.5 mL) were harvested immediately after the addition of [³H]AB61 or after incubation at 37°C (2.5 hours). Cells were washed twice with PBS and then either RNA was isolated using miRNeasy Mini Kit (Qiagen) according to manufacturer's protocol or DNA was isolated using using Dneasy Blood & Tissue Kit (Qiagen) according to manufacturer's protocol including RNA digestion with RNase A. Concentrations of RNA and DNA samples were measured by NanoDrop 1000 Spectrophotometer. Activity of the RNA (40 μL) or DNA (180 μL) was measured in AquaSafe 500 Plus LSC cocktail (4 mL) on a Liquid Scintillation Analyzer Tri-Carb 2900TR (Perkin Elmer).

RNA was isolated from the CCRF-CEM cells treated by [³H]AB61 (250 nmol/L) for 2.5 hours as described above. RNA (150 μL, 240 ng/μL) was digested by Nuclease P1 from Penicillium citrinum (Sigma, 1 mg/mL, 1 μL) in a digestion buffer (80 mmol/L Tris-HCl; 10 mmol/L NaCl; 1 mmol/L MgCl₂; 0.2 mmol/L ZnCl₂, pH 5.3). Final volume was 167.7 μL. Digestion was performed at 50°C for 1 hour. Then, ice-cold aqueous solution TCA (10% v/v, 167.7 μL) was added and samples were cooled on ice for 10 minutes. After centrifugation (14,000 × g, 5 minutes, 4°C) supernatant was transferred into a clear microtube. The supernatant was extracted with 1,1,2-trichlorotrifluoroethane-trioctylamine mixture (4:1, 335.4 μL) at 4°C. The aqueous phase was evaporated using a vacuum concentrator and dissolved in water (90 μL). The sample (50 μL) was analyzed on Waters high-performance liquid chromatography (HPLC) system (2996 PDA detector, 616 HPLC pump, 600S controller, PDA software Empower) with a Supelcosil LC-18-T 3 μm column (15 cm × 3 mm) with a flow rate of 0.75 mL/min. Solution A (50 mmol/L KH₂PO₄, 3 mmol/L tetrabutylammonium hydrogensulfate, pH 3.1) and solution B [50 mmol/L KH₂PO₄, 3 mmol/L tetrabutylammonium hydrogensulfate, pH 3.1, acetonitrile (50% v/v)] were used as a mobile phase. Elution gradient: 0–5 minutes: solution A; 5–6 minutes: linear gradient 0%–5% of solution B in solution A; 6–30 minutes: linear gradient 5%–100% of solution B in solution A. Fractions were collected for 15 seconds each. Activity of the fractions was measured in AquaSafe 500 Plus LSC cocktail on a Liquid Scintillation Analyzer Tri-Carb 2900TR (Perkin Elmer). Nucleotides eluted at the following retention times: t = 3.3 minutes (CMP), 8.6 minutes (AMP), 10.8 minutes (UMP), 11.3 minutes (GMP), 13.4 minutes (AB61), and 13.6 minutes (AB61-MP).

DNA was isolated from the CCRF-CEM cells treated by [³H]AB61 (250 nmol/L) for 2.5 hours as described above. DNA (80 μL, 500 ng/μL) was denatured at 100°C for 10 minutes and cooled on ice. DNA was digested by Nuclease P1 from Penicillium citrinum (Sigma, 1 mg/mL, 10 μL) in 10× digestion buffer (15 μL; 800 mmol/L Tris-HCl, pH 5.3; 100 mmol/L NaCl; 10 mmol/L MgCl₂; 2 mmol/L ZnCl₂). Final volume was 150 μL. Digestion was performed at 50°C for 16 hours. Then, ice-cold aqueous TCA solution (10% v/v, 150 μL) was added. Samples were cooled on ice for 10 minutes. After centrifugation (14,000 × g, 5 minutes, 4°C), supernatant was transferred into a clear microtube. The solution was extracted with 1,1,2-trichlorotrifluoroethane-trioctylamine mixture (4:1, 300 μL) at 4 °C. The aqueous phase was evaporated under reduced pressure and dissolved in water (150 μL). The sample (50 μL) was analyzed on Waters HPLC system as described for HPLC analysis after RNA digestion. Nucleotides eluted at the following retention times: t = 4.2 minutes (dCMP), 11.2 minutes (dAMP), 12.3 minutes (dGMP), 13.1 minutes (dTMP), 13.8 minutes (AB61-MP), and 14.8 minutes (dAB61-MP). (For synthesis of dAB61-MP, see Supplementary Information “Synthesis of nucleosides and nucleotides”).

RNA and DNA were isolated from the CCRF-CEM cells treated by AB61 (250 nmol/L) for 2.5 hours and digested by Nuclease P1 as described above. To detect individual AB61 metabolites in digested samples of DNA or RNA, respectively, individual samples were subjected to liquid chromatography/mass spectrometry (LC/MS) analysis using chromatograph Ultimate 3000 (Dionex) connected to mass spectrometer QTrap 5500 (AB Sciex). LC conditions were defined as: C18 column (Kinetex 2.6 μ, 100 × 3, Phenomenex), flow rate 400 μL/minute, column oven temperature: 30°C; separation was performed in a gradient mode of mobile phase A (10 mmol/L ammonium acetate, pH = 5.02) and B phase (100% acetonitrile): 0 minutes B 10%, 4 minutes 50% B, 4.4 minutes 50% B, 4.6 minutes 10% B, 6 minutes 10% B. Mass spectrometric analysis was performed in positive ESI mode with following conditions: G1 gas 60 psi, G2 gas 60 psi, ion source temperature 600°C, curtain gas 20 psi, ionization voltage 5.5 kV. MRM mode was used for detection and quantification of AB61 metabolites. For AB61-MP MRM transition 451-330.9 and for dAB61-MP 435-81 were applied. Estimated concentrations of individual compounds were expressed in fmoles per 1 μg of DNA and RNA, respectively.

U2OS cells were incubated on coverslips overnight in McCoy medium (with 10% FCSI, 1.5% of l-glutamine, 100 μg/mL streptomycin, and 100 U/mL of penicillin) at 37°C and 5% CO₂. The next day, cells were pretreated with medium or inhibitor of replicative DNA polymerases aphidicolin (ref. 22; 10 μmol/L) for 1 hour. Afterwards, cells were pulsed with 10 μmol/L of PNH404 (11) or medium for 3 hours. Coverslips with cells were washed in PBS, fixed in 4% formaldehyde, and permeabilized in 0.25% Triton X-100 in PBS for 15 minutes. Incorporated PNH404 was visualized via click chemistry using Click-iT Cell Reaction Buffer Kit (Thermo Fischer Scientific, cat.: C10269) and Azide-Fluor 568 fluorescent dye (Jena Bioscience, cat.: CLK-AZ106-5) as recommended by manufacturer. Nuclear DNA was counterstained with Hoechst 33342. Red fluorescence of incorporated PNH404 coupled to Azide-Fluor 568 and blue counterstained DNA were analyzed using confocal microscopy (Cell Observer SD, Zeiss).

To evaluate functionality of RNA with incorporated AB61-TP or Tub-TP, we first PCR-amplified template DNA coding EGFP-luciferase under T7 promoter. (See Supplementary Information “PCR amplification of EGFP-luciferase template for T7 polymerase transcription”). MEGAscript T7 kit (Ambion) was used for in vitro transcription of EGFP-luciferase template in the presence of hetaryl-7-deazapurine ribonucleoside triphosphates. In vitro transcription reactions were performed according to manufacturer's protocol. Each transcription reaction (20 μL) contained of EGFP-luciferase DNA template (2751 bp, 100 ng), ATP, GTP, CTP, and UTP (7.5 mmol/L each), enzyme mix (2 μL), and AB61-TP or Tub-TP (final concentration: 1 mmol/L). In the control experiment, water was used instead of the solution of the tested compound. The transcription reactions were performed at 37°C for 2 hours. Then, DNase I (2 U/μL, 1 μL) was added and the mixtures were incubated at 37°C for 15 minutes. RNA transcripts were purified on NucAway Spin Columns (Ambion). The quantity and purity of RNA transcripts was determined on Agilent 2100 Bioanalyzer (RNA 600 Nano Total RNA kit, Agilent) and by NanoDrop 1000 Spectrophotometer.

The EGFP-luciferase RNA transcripts with or without incorporated AB61-TP or Tub-TP were used as templates for in vitro translation using reticulocyte lysate system (Retic Lysate IVT Kit, Ambion) according to the manufacturer's protocol. The in vitro translation mixture (25 μL) contained EGFP-luciferase RNA transcript (500 ng), low salt mix-met (1 μL), high salt mix-met (0.25 μL), methionine (0.83 mmol/L, 1.5 μL), and the reticulocyte lysate (17 μL). In the positive control experiment, unmodified EGFP-luciferase RNA transcript was used, in the negative control experiment no RNA template was added. The translation reactions were incubated in water bath at 30°C for 1.5 hours. Efficacy of in vitro translation was evaluated via measurement of luciferase enzymatic activity (Luciferase Assay System, Promega) on EnVision Multilabel Reader (PerkinElmer) and by Western blot analysis using primary goat anti-Luciferase antibody (1:1,000, Promega) and secondary peroxidase-conjugated rabbit anti-goat IgG (1:10,000, Sigma-Aldrich). Blot was visualized by chemiluminescence (ECL Western Blotting Detection Reagent, Amersham).

CCRF-CEM cells (1 × 10⁶ cells/mL, 3 mL) were incubated for 3 hours in a humidified CO₂ incubator at 37°C with or without one of the following compounds at equitoxic concentrations: AB61 (250 nmol/L), Tub (550 nmol/L), or actinomycin D (1 nmol/L). Total RNA was purified from treated cell lines using TRI Reagent (Molecular Research Centre) according to the manufacturer's instructions. The concentration and purity of RNA was assessed by Nanodrop ND 1000 (ThermoScientific). RNA quality was measured using Agilent RNA 6000 Nano Kit (Agilent Technologies). Microarray analysis using the GeneChip Human Exon 1.0 ST Array and GeneChip Scanner 3000 7G (Affymetrix) was performed. The data were analyzed using R/Bioconductor freeware and additional Bioconductor packages. All arrays were preprocessed by RMA method (robust multi-array average) on the gene level using the oligo package. The processed gene expression levels were analyzed for their differential expression between the AB61-treated and control cells. Initially, the raw P values of moderated t test were calculated by the limma package, and the data were cut at P = 0.01 significance level which passed 732 genes. Genetic pathways and processes were evaluated using the MetaCore Analytical Suite v. 6.18 (GeneGo, Thomson Reuters). Enrichment analysis consisted of mapping gene IDs of the dataset onto IDs in entities of built-in functional ontologies represented in MetaCore by pathway maps and networks prioritized according to their statistical significance. The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (23) and are accessible through GEO Series accession number GSE62593 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE62593).

A sensoric cell line derived from human osteosarcoma U2OS stable transfected with 53BP1-GFP fusion gene (U2OS-53BP1-GFP) was used to visualize DNA damage foci in the nucleus of treated/control cells. U2OS-53BP1 cells were grown in McCoy medium supplemented with 10% FBS and 100 U/mL penicillin/streptomycin at 37°C and 5% CO₂. Cells were plated in density 10⁵ cells per well into a 96-well microplate in final volume of 150 μL. Following 24-hour incubation, etoposide, irinotecan, and AB61 were added to wells in final concentration of 1 μmol/L. After drug was added, cultivation of cells was continued in Operetta automated microscope (PerkinElmer) equipped with an incubation chamber. Using Operetta imaging system, 21 fields were acquired from each well with conditions of 475 nm excitation and 525 nm emission main wavelength, 40× objectives, and 200 ms exposure time. Images of cells were taken with the time-lapse fluorescence at 1-hour intervals for 12 hours. Number of 53BP1 foci was quantified using Columbus 2.4 imaging analysis system (PerkinElmer).

The validation of the DNA damage effects demonstrated using 53BP1-GFP sensoric cell line was performed on U2OS cells grown and treated under identical conditions, but fixed after 12 hours with 4% paraformaldehyde and stained for reference DNA damage marker, the phospho-γH2AX (24), antibody: Anti-phospho–Histone H2A.X (Ser139), clone JBW301, Millipore (cat. no. 05-638), and evaluated for nuclear foci formation.

4T1-luc2 cells were expanded in vitro in DMEM supplemented with 5 g/L glucose, 2 mmol/L glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, 10% fetal calf serum, and NaHCO₃. A total of 10⁵ cells were orthoptically transplanted into right inferior mouse breast pad of syngeneic Balb/c mice (Charles Rivers). Tumor bearing mice with palpable 3 to 5 mm tumors were injected intraperitoneally with single dose of d-luciferin (150 mg/kg) and baseline tumor luminescence corresponding to firefly luciferase protein expression was recorded 10 minutes later using Photon Imager (Biospacelab). Then, the animals were injected intraperitoneally with vehicle or AB61 at dose corresponding to the maximum tolerated dose (MTD, 60 mg/kg). Luciferase activity was again recorded 4 hours later after 30-minute pretreatment with d-luciferin (150 mg/kg, i.p.). Cumulative luminescence [cpm/mm²] was used as a surrogate indicator of luciferase protein expression in tumors.

Pharmacologic properties of AB61 and AB61-MP, for example, stability in plasma, microsomes, and penetration through artificial membrane, were tested under in vitro conditions.

Tested compound was added to 1.3 mL of preheated human plasma (Transfusion Department, University Hospital Olomouc, Olomouc, Czech Republic) to yield final concentration of 3.3 μmol/L; DMSO did not exceed 0.1%. Procaine was used as a positive control of the assay. The assay was performed at 37°C. Plasma aliquots (75 μL) were sampled at 0, 15, 30, 60, and 120 minutes and the reactions were terminated adding 150 μL of acetonitrile–methanol mixture (2:1), centrifuged at 4°C for 10 minutes at 4,000 rpm. Supernatant was lyophilized, then dissolved in the 200 μL of the mobile phase and analyzed by the RapidFire RF300 system mass spectrometry (RF-MS). In vitro plasma half-life (t_1/2) was calculated using the equation t_1/2 = 0.693/k, where k is the slope of the natural logarithm of the percent compound remaining versus time curve (25).

The parallel artificial membrane permeability assay (PAMPA) was performed with the Millipore MultiScreen filter MultiScreen-IP Durapore 0.45-μm plates and receiver plates (Merck Millipore). The assay was performed according to the manufacturer's protocol (PC040EN00). Compounds were tested at the concentration of 20 μmol/L. Briefly, filters in each well were wetted with 10% lecithin (Sigma Aldrich) in dodecane mixture to create the artificial layer. The filter plate was placed on the top of the receiver plate, which was prefilled with donor solutions (tested compounds dissolved in PBS at final concentration 20 μmol/L). After incubation, 120-μL aliquots of acceptor and donor solutions were removed to the 96-well plate, lyophilized, and residues were dissolved in 200 μL of the mobile phase prior the RapidFire-MS analysis. The relative permeability logPe was calculated with equation: logPe = log{C×−ln(1−drug_A/drug_E)}; where C = (V_A × V_D)/{(V_D+V_A) × A × T}. V_D and V_A are the volumes of the donor and acceptor solutions, A is the active surface area in cm², T is time of the incubation in seconds, drug_A and drug_E is the mass of the compound in the acceptor and in the solution in theoretical equilibrium (as if the donor and acceptor were combined), respectively.

Reaction mixtures included tested compound (5 μmol/L), human liver microsomes (ThermoFisher Scientific, 0.5 mg/mL), NADPH generating system consisting of NADP⁺, isocitrate dehydrogenase, isocitric acid, and MgSO₄ in 0.1 mol/L K₃PO₄ buffer according to the protocol (26, 27). Assay was performed at 0, 15, 30, and 60 minutes. Reactions were terminated using acetonitrile–methanol (2:1) mixture. Prototype-stable versus nonstable drugs (propranolol and verapamil, respectively) were used as a reference (28).

The RapidFire RF300 system (Agilent Technologies) was interfaced with QTRAP 5500 (AB Sciex) mass spectrometer. For detailed description, see ref. (27). Samples were aspirated directly from 96-well plates into a 10-μL sample loop and passed through a C4 cartridge (Agilent Technologies) with solvent A (95% water/5% acetonitrile/0,1% formic acid) at a flow rate of 1.5 mL/minute for 3 seconds. After the desalting step, analyte retained on the cartridge was eluted with solvent B (95% acetonitrile/5% water/0.1% formic acid) to the mass spectrometer at a flow rate of 0.4 mL/minute for 7 seconds. The intrinsic clearance was calculated using the formula: CL_int = V*(0,693/t_1/2), where V is the volume of incubation in μL related to the weight of microsomal protein in mg per reaction. Half-life values (t_1/2) were calculated using the equation t_1/2 = 0.693/k, where k is the slope of the ln of the percent compound remaining versus time curve.

Human colorectal cancer cell line HT-29, ovarian cancer cells SK-OV-3, and breast tumors BT-549, respectively, were grown under in vitro conditions, harvested, and xenotransplanted (5 × 10⁶ cells/mice) subcutaneously on right flank of immunodeficient NOD/SCID mice (NOD.CB17-Prkdc^scid/NcrCrl, Charles Rivers). Tumor bearing mice (200–300 mm³) were treated with AB61 at dose corresponding to half MTD (MTD½, 30 mg/kg; ref. 11) for 3 or 5 consequent days per week in two to four cycles. Tumor length versus width was measured by caliperation three times per week and tumor volume was calculated. Animals were housed in specific pathogen–free conditions, 12-hour light/night regime, clinically examined on daily basis, water, and food ad libitum. Experiments were approved by Animal Ethics Committee of the Faculty of Medicine and Dentistry, Palacky University (Olomouc, Czech Republic).

To obtain analytical standards of AB61-derived metabolites, the nucleotides dAB61-MP and AB61-DP were prepared under standard conditions for Suzuki cross-coupling reaction (11, 29, 30) of the corresponding nucleotide or by phosphorylation of the nucleoside, respectively. [³H]-Labeled AB61 ([³H]AB61) was synthesized by Suzuki cross-coupling reaction of 5-iodotubercidin with 5-chlorothiophene-2-boronic acid followed by tritiation (Fig. 2). For detailed information on synthesis, see Supplementary Information “Synthesis of nucleosides and nucleotides.”

Cytotoxic activity of AB61 was determined using MTT assay following a 3-day incubation (Table 1). AB61 showed nanomolar IC₅₀ values against cancer cell lines (A549, CCRF-CEM, HCT116, and K-562). A strong cytotoxic effect of AB61 was also observed against paclitaxel- and daunorubicin-resistant cancer cell lines overexpressing multidrug resistance transporters (K-562-tax, CEM-DNR-bulk) and a colorectal cancer cell line with inactivated p53 gene (HCT116p53^−/−). IC₅₀ values of AB61 against normal lung and foreskin fibroblasts (MRC-5, BJ) were 2- to 5-orders of magnitude higher (micromolar) than the IC₅₀ values against the cancer cell lines. On the other hand, cytotoxic activities of AB61-5′-O-monophosphate (AB61-MP) against cancer cell lines were always at least one order of magnitude lower than those of AB61. Interestingly, AB61-MP also showed poor selectivity towards malignant versus nonmalignant cells. Its IC₅₀ values against normal fibroblasts were similar to those against cancer cell lines.

Intracellular phosphorylation is often a key step in activation of nucleoside drugs. Efficient phosphorylation of AB61 was previously shown in a Du145 prostate cancer cell line (11). Intracellular levels of AB61, its nucleoside monophosphate (AB61-MP), diphosphate (AB61-DP), and triphosphate (AB61-TP) were determined by HPLC analysis after 1- and 3-hour treatment with AB61 (1 μmol/L or 10 μmol/L) in colon cancer cell line (HCT116) and in normal foreskin BJ fibroblasts (Table 2). High levels of AB61-MP were observed in HCT116 cells after 3-hour treatment, whereas after 1-hour treatment no or only a minor amount of AB61-MP was detected after treatment with 1 or 10 μmol/L concentration of AB61, respectively. BJ fibroblasts showed good uptake of AB61 but no AB61-MP was detected, thus indicating ineffective intracellular phosphorylation of AB61 in BJ fibroblasts. AB61-DP and AB61-TP were not detected in any cell line, presumably due to short treatment time, fast incorporation into macromolecules and/or concentrations under the detection limit. Both AB61 and AB61-MP were shown to cross artificial cellular membrane in the PAMPA assay; however, the membrane permeability for AB61-MP was one order of magnitude lower than that of AB61 (Table 3). Furthermore, AB61 is stable in plasma and human microsomes, whereas AB61-MP is slowly degraded with plasma with half time around 52 minutes, presumably to the parent nucleoside AB61 due to serum phosphatase activity (Table 3).

Enzymatic incorporation of AB61-TP into RNA was tested using viral T7 RNA polymerase. The 40-mer dsDNA template contained two 2′-OMe-RNA nucleotides at the 5′-end of the noncoding strand to avoid n+1 activity of the T7 RNA polymerase (31). The template is transcribed into a 21-bp RNA transcript containing four ATP insertions. The transcription reactions performed in presence of either ATP, AB61-TP or Tub-TP and three natural NTPs (GTP, UTP, and CTP) afforded a full length transcript (Fig. 3A). The results show that both AB61-TP and Tub-TP are substrates of T7 RNA polymerase and are able to substitute for ATP during transcription and to be incorporated into RNA. Furthermore, these experiments provided evidence that T7 RNA polymerase is able to further elongate the transcript after incorporation of AB61-TP. Our studies also focused on enzymatic incorporation of AB61-TP into DNA in a primer extension experiment with bacterial DNA polymerase I (Klenow fragment) and a 31-bp DNA template allowing incorporation of four dATPs. A primer extension reaction with AB61-TP instead of dATP led to formation of a full-length product, confirming that AB61-TP was successfully incorporated into DNA without causing the polymerase to pause. In contrast, Tub-TP was not incorporated by Klenow fragment (Fig. 3B). Incorporation of AB61-TP by commercially available human DNA polymerases were also tested using a primer extension experiment. While AB61-TP is not a substrate for DNA polymerase β (Fig. 3C), AB61-TP was successfully incorporated by DNA polymerase γ (Fig. 3D). Incorporation of AB61-TP by DNA polymerase γ was also confirmed by MALDI-TOF analysis (See Supplementary Fig. S1).

Further studies focused on incorporation of AB61 into RNA and DNA in living cells. CCRF-CEM cells were incubated with ³H-labeled AB61 ([³H]AB61; 250 nmol/L), cellular RNA and DNA were isolated and activities of the RNA and DNA samples were measured using liquid scintillation counting (LSC; Table 4). Significant increases in activities were observed for both RNA and DNA samples after 2.5-hour treatment. To further confirm the incorporation of AB61, RNA, and DNA samples from cells treated with AB61 or [³H]AB61 (250 nmol/L) were digested with nuclease P1 affording nucleoside monophosphates. The hydrolyzed RNA and DNA samples were analyzed by HPLC-MS (Table 5) and HLPC with UV/VIS detection and radioactivity of HPLC fractions was determined by LSC (Fig. 4). Figure 4A and D show authentic analytical standards of ribo- or deoxyribonucleos(t)ides, respectively. After digestion of RNA, peaks of all natural NMPs were identified in the HPLC chromatogram (Fig. 4B). Three peaks were observed in HPLC-LSC analysis (Fig. 4C). One of the peaks corresponds with standard of AB61-MP confirming the incorporation of [³H]AB61 into RNA, while the two other peaks are presumably related to short [³H]AB61-containing oligonucleotides that were formed because of the incomplete digestion of the RNA sample. The presence of AB61-MP was also confirmed by LC-MS. Peaks of all natural dNMPs were found in the HPLC chromatogram of the digested DNA sample (Fig. 4E). A single peak was observed after digestion of DNA sample by the HPLC-LSC analysis (Fig. 4F). This peak clearly corresponds to ribonucleotide AB61-MP. In the LC-MS analysis of DNA isolated from AB61-treated cells, AB61-MP was detected as a major component, but a minor presence of the corresponding 2′-deoxyribonucleotide (dAB61-MP) was also observed. The molar ratio of AB61-MP and dAB61-MP was approximately 30:1 (Table 5). The results show that AB61 is mainly incorporated into DNA as a ribonucleotide although 2′-deoxyribonucleotide is also present in residual quantities. On the contrary, only AB61-MP was detected in the RNA of the treated cells. To localize incorporation of AB61 in cells and to further investigate the mechanisms of incorporation into the DNA, we have employed structural analogue of AB61, compound PNH404 (Fig. 1; ref. 11), which contains alkyne group available for azide-alkyne CuAAC cycloaddition of fluorescent dye Azide-Fluor 568 (“click chemistry”; ref. 32). In U2OS cells, PNH404 was incorporated predominantly into nuclei of treated cells with residual staining in the cytoplasm. Interestingly, the incorporation of the compound was not significantly inhibited by inhibitor of replicative DNA polymerases aphidicolin (ref. 22; Fig. 5).

To evaluate the biologic activity of the AB61 more broadly, we performed transcriptional microarray profiling coupled to the pathway analysis. On the basis of the differentially expressed genes the “DNA Damage_DBS Repair” network process was identified as the most significant process network following the AB61 treatment. Furthermore, other networks indicating significantly affected protein folding and translation, cell cycle, and angiogenesis were top ranked. Upregulation of BRIP1, RPA4, FANCF, and TOP1, and downregulation of TdT, FANCG, LIG4, LIG3, NRB54, Chk2, MGMT, RAD51B, Histone H4, Ubiquitin, and C1D genes which are well known to be involved in DNA damage processes was observed. This corresponds well to the cell biology data (vide infra). Reference compounds, tubercidin (Tub) and actinomycin D (ActD), with structural or mechanistic similarity to AB61, were used. The networks involved in “development_regulation of angiogenesis” and “development and blood vessel morphogenesis” were similarly affected; however, the DNA damage, cell-cycle alteration, and protein translation and folding were not among top listed networks in cells treated with ActD and Tub (Table 6).

To confirm AB61-induced DNA damage by an independent method, the U2OS human osteosarcoma cell line was stably transfected with the 53BP1-GFP (20) fusion gene. 53BP1 protein accumulates at DNA lesions, preferably at DNA double strand breaks, thus its GFP-conjugate enables visualization of DNA-damage sites in nucleus. Indeed, U2OS-53BP1-GFP cells treated with AB61 or control DNA–damaging agents like topoisomerase I or II inhibitors (irinotecan and etoposide, respectively) showed time-dependent accumulation of 53BP1 foci in the nuclei of treated cells (Fig. 6). Although the absolute number of 53BP1 foci in AB61-treated cells (Fig. 6B) was lower than in irinotecan- and etoposide-exposed cells (Fig. 6C–E), the foci were markedly larger compared with both irinotecan- and etoposide-treated versus untreated control cells (Fig. 6A). These results were independently validated using phospho-γH2AX staining in U2OS cells treated with AB61, etoposide or irinotecan, respectively (Fig. 6F–I).

Following the evidence of AB61-TP and Tub-TP incorporation into mRNA, we wanted to evaluate functional properties of the RNA harboring ribonucleoside analogues in terms of translational efficacy. A template with a model EGFP-luciferase fusion gene under T7 promoter was transcribed into RNA. The transcription experiments were performed in the presence of all four natural NTPs with/without AB61-TP or Tub-TP so that AB61-TP or Tub-TP could be randomly incorporated into the RNA transcripts. After the in vitro transcription, the template DNA was digested by DNase I and the RNA transcripts were purified. In the next step the RNA transcripts were used as templates for in vitro translation using a commercial reticulocyte lysate in vitro translation system. The resulting EGFP-luciferase protein in the translation reactions was detected on the basis of its enzymatic luciferase activity and Western blotting (Fig. 7). The results showed that the translation reactions with unmodified RNA template and RNA template produced in the presence of Tub-TP afforded similar amounts of EGFP-luciferase. In contrast, synthesis of EGFP-luciferase protein from RNA template produced in the presence of AB61-TP was completely blocked, as evidenced from both the measurement of luciferase enzymatic activity and protein content (Fig. 7). This result shows that AB61-TP can be incorporated into RNA by T7 RNA polymerase in the presence of ATP. Furthermore, the presence of AB61 in the RNA hampers its function as a template for the translation. This effect was not observed for Tub-TP and to our knowledge it is the first description of such an effect in the class of cytostatic nucleosides.

To evaluate the effect of AB61 on luciferase activity/expression in vivo, Balb/c mice were transplanted with the 4T1-Luc2 breast cancer cells stably transfected with luciferase under strong CMV promoter and bioluminescence imaging technique was employed to visualize and quantify the luciferase (Fig. 8). Indeed, the tumor-bearing mice treated with AB61 (MTD, 60 mg/kg) intraperitoneally demonstrated 4 hours later significant decrease of luciferase activity in vivo as compared with vehicle-treated mice. In vivo anticancer activity of AB61 was determined using human ovarian (SK-OV-3), breast (BT-549), and colorectal (HT-29) tumors xenografted to the NOD/SCID mice (Fig. 9). Animals were treated with MTD½ (30 mg/kg) of AB61 in 3 (low-intensity) or 5 consequent days per week (high-intensity scheme) in two to four cycles. In all cancer models employed, highly significant reduction of tumor volume was observed in treated mice. However, significant reduction of tumor volume in ovarian cancer model SK-OV-3 was only observed with high intensity scheme suggesting dose dependent effect. AB61 also significantly prolonged survivals in BT-549 and HT-29 models, but not in the SK-OV-3 due to no efficacy in the low-intensity administration schedule and toxicity in the high-intensity regimen.

In this study, we focused on elucidation of the mechanism of action of a novel 7-deazaadenine analogue, AB61. Despite the very strong cytotoxic effect of AB61 against cancer cell lines we found out that cytotoxic activity of AB61-MP against cancer cell lines was always lower (typically by 2 orders of magnitude) than that of AB61. The main reason is probably in lower (by one order of magnitude) permeability of the monophosphate AB61-MP through cell membrane compared with nucleoside AB61 (as confirmed by PAMPA assay) but we can also hypothesize on involvement of membrane efflux and/or complex phosphorylation/dephosphorylation dynamics. On the other hand, AB61-MP showed higher cytotoxicity against normal fibroblasts compared with AB61, which was comparable with cytotoxicity in cancer cell lines. The PAMPA test confirmed that AB61-MP can still relatively well penetrate through membrane (though less efficiently than nucleoside AB61) causing the nonspecific cytotoxicity. Apparently, intracellular availability of AB61-MP increases the cytotoxic effect. Our further results demonstrated that AB61 is phosphorylated in HCT116 colon cancer cell line (cytotoxicity IC₅₀ = 1.9 nmol/L) but the phosphorylation does not proceed in normal BJ fibroblasts (cytotoxicity IC₅₀ = 8 μmol/L). These data are in accordance with the fact that expression of adenosine kinase, the enzyme that might be involved in the phosphorylation of AB61, is increased in colorectal tumors compared with the normal tissue (33–35). These findings reveal the importance of intracellular phosphorylation for both the efficiency and cancer selectivity of AB61 and suggests for potentially synergistic combinations with therapies increasing nucleoside phosphorylation rates in tumors, for example, radiation (36). Furthermore, we showed that AB61-TP is effectively incorporated into RNA by T7 RNA polymerase in an enzymatic assay. Although AB61-TP is not accepted as a substrate by human DNA polymerase α(13) and β, both human DNA polymerase γ and Klenow fragment successfully incorporated AB61-TP into DNA, even though it is not a 2′-deoxyribonucleotide but a ribonucleotide. However, our experimental design for AB61-TP incorporation by DNA polymerase β did not contain a single nucleotide gap substrate, which is a typical site of polymerase β action (37), nor accessory proteins involved with polymerase β activity (PARP, XRCC1; ref. 38), and therefore the incorporation of AB61-TP might not have been a suitable substrate in the cell-free system we have used. In contrary, Tub-TP is accepted neither by the human DNA polymerase α (13) nor the Klenow fragment. Both DNA polymerase γ and Klenow fragment represent A family DNA polymerases. Despite the fact that DNA polymerase γ is a mitochondrial polymerase and Klenow fragment is a bacterial enzyme, we hypothesize that AB61-TP might be also incorporated by nuclear DNA polymerases from the A family, that is, DNA polymerase θ and DNA polymerase ν which are both involved in DNA repair processes and are known for lower fidelity (39). Also treatment of CCRF-CEM cells with AB61 or [³H]AB61 resulted in its incorporation into both RNA and DNA. HPLC-MS analysis of DNA lysates revealed that AB61 is incorporated into DNA predominantly as a ribonucleotide and only rarely as a 2′-deoxyribonucleotide (30:1 ratio) which means that AB61-TP is a substrate for human DNA polymerases. Despite this finding might seem surprising, ribonucleotide incorporation into cellular DNA is quite common (40–42). In contrast, tubercidin (2) and toyocamycin (3) are incorporated into DNA only after reduction by ribonucleotide reductase as 2′-deoxyribonucleotides. Consistently with mass spectrometry data, PNH404, the ethynyl-analogue of AB61, was detected in nuclei and cytoplasm of treated U2OS cells, which is in agreement with incorporation of AB61 into DNA and RNA, respectively. Interestingly, the incorporation of AB61 analogue into nuclei was not affected by the inhibitor of replicative DNA polymerases aphidicolin, indicating (in accord with previous data) that incorporation is rather dependent on nonreplicative, for example, reparative DNA polymerases (39). To see a broader picture of the AB61-induced changes, we performed microarray gene expression profiling, which, in principle, suggested DNA damage repair networks and translational or folding machinery are primarily affected in cellular response to the AB61. Several genes involved in DNA damage and/or repair mechanisms were altered. However, the affected networks are only partially overlapping with reference compounds tubercidin and actinomycin D, suggesting a different mechanism of action. To validate DNA repair pathway at the cellular level, we used the 53BP1-GFP reporter U2OS cell line. Protein 53BP1 is known to rapidly accumulate at DNA double strand breaks leading to activation of DNA repair mechanisms (43). After treatment with AB61, formation of large 53BP1-GFP foci was observed in a time-dependent manner indicating that AB61 induces DNA damage. However, the pattern of 53BP1-GFP nuclear foci is different from the DNA damage patterns induced by topoisomerase inhibitors irinotecan, etoposide, or p53 activator 7-iodotubercidin (44), which suggests that DNA damage caused by AB61 is achieved, maintained, or (un)repaired via different mechanisms. In addition to the 53BP1-GFP accumulation in the DNA damage sites, similar staining pattern was independently confirmed for phospho-γH2AX, one of the most robust surrogate markers of DNA double strand breaks. As (i) AB61-TP is not a substrate for human polymerase α(13) or polymerase β, (ii) incorporation of AB61 analogue into cell nuclei was not decreased upon inhibition of replicative but not reparative DNA polymerases by aphidicolin, (iii) gene expression study revealed activation of DNA damage pathway, and (iv) formation of unusually large 53BP1 and/or phospho-γH2AX DNA damage foci in treated tumor cells was observed, we hypothesize that AB61 is incorporated into DNA by polymerases involved in the repair mechanisms, thus further aggravating the local DNA damage in a positive back-loop manner and thereby inducing stronger signal for 53BP1 and/or phospho-γH2AX to accumulate.

We had previously reported poor inhibition of RNA polymerase II by AB61-TP (11). Having the evidence that AB61 is incorporated into RNA of exposed cells, the RNA polymerase II is not inhibited, but the RNA synthesis is down regulated, we focused on the possible consequences of AB61 incorporation into mRNA. We studied in vitro translation of RNA templates containing randomly incorporated AB61. We discovered that the incorporation of AB61 into mRNA template completely prevents formation of a protein. This result was further confirmed in mice by the observation of significant and rapid inhibition of luciferase protein expression in AB61-treated 4T1-luc tumors. In contrast, even though Tub-TP is incorporated into RNA by T7 RNA polymerase, no effect on in vitro translation was observed. In vivo studies of AB61 in human colorectal (HT-29), ovarian (SK-OV-3), and breast (BT-549) cancer xenografts revealed highly significant reduction of tumor mass in treated animals, which was reflected with increased survival preferentially in low-density regimen. Despite better tumor control of the high-intensity schedule, it was not always translated into increased survival due to cumulative therapeutic toxicity with leading symptoms of body weight lost and anorexia. Interestingly, the highest in vivo activity of the compound was demonstrated in slowly growing HT-29 tumors, suggesting preferential anticancer activity in slowly proliferating tumor cells. Therefore, the in vivo anticancer effect of AB61 is very promising and warrants further studies.

Most of the nucleosides used in cancer therapy are incorporated into DNA leading to termination of nascent DNA strand extension and DNA damage response activation (45). In this point of view, the mechanism of action of AB61 is more complex. The study clearly showed that the cytotoxic effect of AB61 is tightly linked with its intracellular phosphorylation and the inefficient phosphorylation in nonmalignant cells is a key factor of AB61 selectivity towards cancer cell lines. AB61 is incorporated both into DNA and RNA as a ribonucleotide (and only in trace amounts as a 2′-deoxyribonucleotide into DNA) where it leads to DNA damage and protein translation arrest, respectively. Rapid inhibition of translational machinery may also explain previously observed downregulation of RNA and DNA synthesis as a secondary effect (11). In spite of the structural similarity of AB61 with tubercidin, its mode of action seems to be unique and different from the one of gemcitabine, the only clinically used nucleoside cytostatic for treatment of solid tumors (46). The mechanism targeting at the same time DNA damage/repair and inhibition of protein translation machinery might be highly relevant for elimination of slowly proliferating, dormant, and senescent tumor cells, which are not cycling enough thus highly resistant to conventional chemotherapy and radiation. In summary, we showed that AB61 is a potent and selective cytostatic with exceptional mechanisms of action. Further pharmacologic studies of AB61 are currently under way.

No potential conflicts of interest were disclosed.

Conception and design: P. Perlíková, A. Bourderioux, M. Hajdúch, M. Hocek

Development of methodology: P. Perlíková, G. Rylová, P. Nauš, E. Tloušťová, K. Motyka, P. Znojek, A. Nová, M. Harvanová

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P. Perlíková, P. Nauš, P. Znojek, A. Nová, M. Harvanová, M. Šiller

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P. Perlíková, E. Tloušťová, L.P. Slavětínská, K. Motyka, A. Nová, P. Džubák, J. Hlaváč, M. Hajdúch, M. Hocek

Writing, review, and/or revision of the manuscript: P. Perlíková, G. Rylová, P. Nauš, A. Nová, P. Džubák, M. Hajdúch, M. Hocek

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): G. Rylová, P. Džubák

Study supervision: M. Hajdúch, M. Hocek

Other (prepared the tritium-labeled AB61 by performing the tritiation reaction followed purification and characterization of [³H]AB61by radio-HPLC and 3-H NMR): T. Elbert

Other (in vivo study): D. Doležal

The authors thank Lenka Radova and Josef Srovnal for the help with the microarray experiments and data analysis and Natalie Taborska for the help with in situ metabolic labeling.

This work was supported by the Czech Science Foundation P207/11/0344 (to M. Hajdúch, P. Perlíková, M. Hocek, and P. Nauš), Technology Agency of the Czech Republic TE01020028 (to M. Hajdúch, P. Perlíková, M. Hocek, P. Džubák, and G. Rylová), Ministry of Education of the Czech Republic LO1304 (to M. Hajdúch, J. Hlaváč, P. Džubák, D. Doležal, P. Znojek, A. Nová, M. Harvanová, and M. Šiller), CZ.1.07/2.3.00/30.0041 (J. Hlaváč, K. Motyka, and M. Šiller), Academy of Sciences of the Czech republic RVO:61388963 (to M. Hocek) and the European Social Fund and the state budget of the Czech Republic, project no. CZ.1.07/2.3.00/30.0041 (to P. Znojek), and by Gilead Sciences, Inc. (to M. Hocek, P. Perlíková, P. Nauš, and A. Bourderioux).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.