Toxicological Sciences

2′O(2methoxyethyl) nucleosides are not phosphorylated or incorporated into the genome of Human Lymphoblastoid TK6 Cells

Journal: Toxicological Sciences

Manuscript ID TOXSCI-17-0575.R2

Manuscript Type: Research Article

Date Submitted by the Author: 02-Jan-2018

Complete List of Authors: Saleh, Amer; AstraZeneca, Drug Safety and Metabolism Bachman, Martin; AstraZeneca , Discovery Sciences Priestley, Catherine; AstraZeneca, IMED operations Gooderham, Nigel; Imperial College London, Surgery and Cancer Andersson, Patrik; AstraZeneca, Drug Safety and Metabolism Henry, Scott; Ionis Pharmaceuticals Inc Edmunds, Nicholas; AstraZeneca, Drug Safety and Metabolism Fellows, Mick; AstraZeneca, Drug Safety and Metabolism

modified nucleosides, DNA incorporation, nucleoside phosphorylation, Key Words: nucleoside kinase

Page 1 of 30 Toxicological Sciences

1 2 3 2′O(2methoxyethyl) nucleosides are not phosphorylated or incorporated into 4 5 the genome of Human Lymphoblastoid TK6 Cells 6 7 Amer F Saleh 1*, Martin Bachman2, Catherine C Priestley3, Nigel J Gooderham4, 8 9 Patrik Andersson5, Scott P Henry6, Nick Edmunds1, and Mick D Fellows1 10 11 1 New Drug Modalities, Drug Safety and Metabolism, IMED, AstraZeneca, 12 13 14 Cambridge, United Kingdom 15 2 16 Discovery Sciences, AstraZeneca, Macclesfield, United Kingdom 17 18 3 IMED Operations, AstraZeneca, Cambridge, United Kingdom 19 20 4 Surgery and Cancer, Imperial College London, London, United Kingdom. 21 22 5 New Drug Modalities, Drug Safety and Metabolism, IMED, AstraZeneca, 23 24 Gothenburg, Sweden 25 26 6 27 Ionis Pharmaceuticals, Inc., Carlsbad, CA, USA. 28 29 *To whom correspondence should be addressed at New Drug Modalities, Drug 30 31 Safety and Metabolism, AstraZeneca, Cambridge, CB0 0WG, United Kingdom. Tel: 32 33 +441625513837. Email: [email protected] 34 35 36 37

38 39 40 41 42 Running title: 2′OMOE phosphorylation and incorporation into cellular DNA and 43 44 RNA 45 46 47 48 49 50 51 52 53 54 55 56 57 58 1 59 60 Toxicological Sciences Page 2 of 30

1 2 3 Abstract 4 5 Nucleoside analogues with 2′modified sugar moieties are often used to improve the 6 7 RNA target affinity and nuclease resistance of therapeutic oligonucleotides in 8 9 preclinical and clinical development. Despite their enhanced nuclease resistance, 10 11 oligonucleotides could slowly degrade releasing nucleoside analogues that have the 12 13 14 potential to become phosphorylated and incorporated into cellular DNA and RNA. 15 16 For the first time, the phosphorylation and DNA and RNA incorporation of 2′O(2 17 18 methoxyethyl) (2′OMOE) nucleoside analogues have been investigated. Using 19 20 LC/MS/MS, we showed that in the nucleotide salvage pathway including 21 22 deoxycytidine kinase (dCK) and thymidine kinase (TK1) displayed poor reactivity 23 24 toward 2′OMOE nucleoside analogues. On the other hand, 2′fluoro (F) 25 26 27 nucleosides, regardless of the nucleobase, were efficiently phosphorylated to their 28 29 monophosphate forms by dCK and TK1. Consistent with their efficient 30 31 phosphorylation by dCK and TK1, 2′F nucleosides analogues were incorporated into 32 33 cellular DNA and RNA while no incorporation was detected with 2′OMOE 34 35 nucleoside analogues. In conclusion, these data suggest that the inability of dCK and 36 37 TK1 to create the monophosphates of 2′OMOE nucleoside analogues reduces the 38 39 40 risk of their incorporation into cellular DNA and RNA. 41 42 43 44 Keywords: modified nucleosides, DNA incorporation, nucleoside phosphorylation, 45 46 nucleoside kinase 47 48 49 50 51 52 53 54 55 56 57 58 2 59 60 Page 3 of 30 Toxicological Sciences

1 2 3 Introduction 4 5 Oligonucleotidebased therapeutics have the potential to modulate expression 6 7 in mammalian cells to treat a wide range of diseases including cancer, inflammatory, 8 9 dyslipidemia and infectious diseases (Kole and Leppert, 2012; Shahbazi et al., 10 11 2016). Various chemical modifications to the base, sugar, and backbone have been 12 13 14 developed and applied to therapeutic oligonucleotides to improve their 15 16 pharmacokinetics and pharmacodynamics properties (Deleavey and Damha, 2012). 17 18 One of the most widely used modifications for antisense oligonucleotides is 19 20 phosphorothioate internucleotide linkages whereby one of the nonbridging oxygen 21 22 atoms in the phosphodiester linkage is replaced by sulphur (Eckstein, 2000). These 23 24 firstgeneration phosphorothioate oligonucleotides (PS ODN) have sufficient 25 26 27 nuclease resistance, support RNase H activity, and bind plasma proteins which is 28 29 critical to maintain favourable pharmacokinetics and tissue distribution properties 30 31 (Eckstein, 2000). Subsequently, the socalled second generation of chemical 32 33 modifications were developed by modifying the 2′position of the sugar moiety of the 34 35 ODN with either an alkyl group (e.g., 2′Omethyl, or 2′Omethoxyethyl) or a 2′fluoro 36 37 group (Monia, 1997). These modifications increased the hybridization affinity by 38 39 40 holding the ribose ring in a configuration that is optimal for hydrogen bonding with 41 42 the target RNA strand (Kawasaki et al., 1993). The 2’F modification has been used 43 44 extensively in siRNA constructs, whereas the singlestranded phosphorothioate 45 46 ODNs more commonly use the 2′alkyl modifications. The second generation 47 48 chemistry that is most well studied in development is the 2′OMOE modified PS 49 50 ODN (Geary et al., 2001a). 51 52 53 The 2′alkyl modifications have an added benefit in that they further enhance 54 55 nuclease resistance and modulate the pharmacokinetic properties of antisense 56 57 58 3 59 60 Toxicological Sciences Page 4 of 30

1 2 3 ODNs (Prakash, 2011). As fully 2′ribose modified oligonucleotides do not support 4 5 RNase H to cleave the target RNA, socalled gapmers were developed whereby the 6 7 ribose sugar modifications are confined to both ends of ODN leaving a central ‘gap’ 8 9 of unmodified 2′deoxyphosphorothioate nucleotides (Monia et al., 1993). For 10 11 example, the FDAapproved antisense drug “Mipomersen”, used to treat familial 12 13 14 hypercholesterolemia, is a 20mer fully phosphorothioated oligo with 2′OMOE 15 16 modified ribose at both ends (Crooke and Geary, 2013). This makes the 3′ and 5′ 17 18 end of the antisense oligonucleotide highly resistant to exonucleasemediated 19 20 metabolism. 21 22 The genotoxic potential for singlestranded PS ODN has been thoroughly examined, 23 24 with and without 2′alkyl modifications, and have been found to be uniformly absent 25 26 27 of any genotoxicity potential, but the implications of the liberated mononucleotides 28 29 have never been directly assessed (Berman et al., 2016; Henry et al., 2002). The 30 31 released nucleotide analogues could have a genotoxic potential either via 32 33 perturbation of the endogenous nucleotide pool or by incorporation into newly 34 35 synthesised genomic DNA potentially leading to mutation or breakage 36 37 (Mattano et al., 1990; Phear et al., 1987). The likelihood of unbalancing nucleotide 38 39 40 pools through slow nuclease mediated metabolism seems an unlikely concern 41 42 because this would be addressed by the standard genotoxicity assays where uptake 43 44 and metabolism at extraordinarily high concentrations have been documented 45 46 without genotoxic impact (Henry et al., 2002). However, the European Medicines 47 48 Agency (EMA) raised a concern on the genotoxic potential of the liberated nucleotide 49 50 analogues as they could serve as substrates for various intracellular kinases which 51 52 53 would represent the first crucial step for their incorporation into newly synthesised 54 55 DNA 56 57 58 4 59 60 Page 5 of 30 Toxicological Sciences

1 2 3 http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/0 4 5 9/WC500003149.pdf. Indeed, using in vitro polymerase extension assays, previous 6 7 studies showed that 2′fluorodeoxynucleotides were substrates for human DNA 8 9 polymerases including polymerase alpha (pol α) and polymerase gamma (pol γ), and 10 11 could be incorporated into primed DNA (Richardson et al., 2000). Subsequently, the 12 13 14 same group demonstrated that 2′fluoropyrimidines (2′FC and 2′FU) could be 15 16 incorporated into DNA and RNA of various tissues of rats and woodchucks following 17 18 longterm treatment (90 days) (Richardson et al., 2002). These studies raise a 19 20 question as to whether other commonly employed 2′ribose modifications such as 2′ 21 22 OMOE could also permit DNA incorporation with the potential for mutagenic 23 24 consequences. Therefore, the purpose of this study was to evaluate if 2′OMOE 25 26 27 nucleoside analogues are phosphorylated by nucleotide salvage enzymes and 28 29 whether they are incorporated into cellular DNA and RNA. Here we confirm 30 31 phosphorylation and DNA/RNA incorporation of 2′F nucleosides but demonstrate 32 33 that these properties are not shared by 2′OMOE nucleosides. 34 35 Materials and Methods 36 37 Materials and Reagents 38 39 40 TK6 human lymphoblastoid cells were grown in Roswell Park Memorial Institute 41 42 (RPMI) 1640 medium supplemented with 10% heatinactivated horse serum, 2 mM l 43 44 glutamine, 100 U/ml of penicillin, and 100 g/ml of streptomycin. The cultures were 45 46 incubated at 37°C in a humidified atmosphere of 5% CO2. All cell culture reagents 47 48 were purchased from Invitrogen (Paisley, United Kingdom). Nucleosides: natural 49 50 nucleosides 2′deoxyguanosine, 2′deoxyadenosine, 2′deoxycytidine ,2′ 51 52 53 deoxythymidine and their corresponding 5′Omonophosphate derivatives were 54 55 purchased SigmaAldrich (Poole, United Kingdom).). 2′OMOEmodified 56 57 58 5 59 60 Toxicological Sciences Page 6 of 30

1 2 3 nucleosides including 2′O(2Methoxyethyl)5methylcytidine, 2′O(2methoxyethyl) 4 5 5methylcytidine5′Omonophosphate,2′O(2methoxyethyl)5methyluridine, 2′O 6 7 (2methoxyethyl)5methyluridine5′Omonophosphate,2′O(2methoxyethyl) 8 9 adenosine, 2′O(2methoxyethyl)adenosine5′Omonophosphate, 2′O(2 10 11 methoxyethyl)guanosine, 2′O(2methoxyethyl)guanosine5′Omonophosphate. 12 13 14 were obtained from Jena Bioscience. 2′F nucleosides were obtained from Ionis (CA, 15 16 USA). Enzymes: Human deoxycytidine kinase (DCK), human UMPCMPK Kinases 17 18 were obtained from Novocib (France). Thymidine Kinase 1 (TK1), human adenylate 19 20 kinase 1 (AK1), human guanylate kinase 1 (GK1), human deoxythymidylate kinase 21 22 (thymidylate kinase) (DTYMK) were obtained from ProSpec (Israel). 23 24 Enzymatic reaction 25 26 27 The phosphorylation reactivity of modified nucleosides by dCK was analysed by 28 29 incubating 200 M of the indicated nucleoside in 100 mM Tris, pH 7.5, 100 mM KCl, 30 31 10 mM MgCl2, 1 mM ATP and 1 g of in a 50 l reaction volume at 37 °C for 32

33 2 h. Phosphorylation by TK1 was analysed in 50 mM TrisHCl pH 7.5, 2.5 mM MgCl2, 34 35 10 mM dithiotreitol (DTT) and 3 mM NaF. AK1 and GK1 were assayed with 1 g of 36 37 the respective enzyme, 1 mM ATP and 200 M of substrate in 100 mM TrisHCl (pH 38 39 40 7.8), 100 mM KCl and 10 mM MgCl2. CMPK was assayed as above but in 100mM 41 42 TrisHCl (pH 9), 250mM KCl, 15mM Mgacetate and 5mM DTT. DTYMK was 43 44 assayed as above but in 50 mM TrisHCl pH 7.5, 2.5 mM MgCl2, 10 mM dithiotreitol 45 46 (DTT) and 3 mM NaF. 47 48 Nucleotide pool, RNA and genomic DNA extraction 49 50 TK6 cells (4 x 105 cell/ml) seeded in a 24well plate were treated with the indicated 51 52 53 concentration of the nucleotide for 24 h. Cells were treated with 100 M of isotope 54 55 labelled cytidine (15N3) for 4 h before nucleic acid extraction. To remove 56 57 58 6 59 60 Page 7 of 30 Toxicological Sciences

1 2 3 extracellular nucleosides for the analysis of intracellular exposure of modified 4 5 nucleosides, cells were first pelleted by centrifugation for 5 min at 300 x g in a 1.5 ml 6 7 microcentrifuge tube. The supernatant was completely removed and the cell pellet 8 9 was suspended in 1 ml of icecold PBS by gentle vortexing. Cell were then 10 11 centrifuged for 5 min at 300 x g before removing the PBS solution, and the washing 12 13 14 step was repeated twice. The intracellular nucleotide pool was extracted by adding 1 15 16 ml of icecold 80% methanol to the cell pellet. The samples were then vortexed, 17 18 centrifuged for 10 min at 20000 x g and the supernatant was transferred to a new 19 20 microfuge tube and dried by SpeedVac at 37 °C for 1.5 h. The samples were then 21 22 diluted in 50 l of H2O. For DNA and RNA extraction, the cells were washed 3 times 23 24 with PBS. DNA was then extracted using the DNaesy Kit (Qiagen) according to the 25 26 27 manufacturer protocol. In brief, the cell pellet was suspended in 200 l of PBS before 28 29 the addition of 20 l of Proteinase K, 4 l of RNase A (100 mg/ml) and 200 l of lysis 30 31 buffer provided with the kit. The samples were then mixed by vortexing and 32 33 incubated at 56°C for 10 minutes and 200 l ethanol (>99.6%) added to the sample 34 35 and mixed by vortexing. The mixture was then pipetted into DNeasy Mini spin 36 37 column (provided with the kit) and centrifuged at 6000 x g for 1 minute during which 38 39 40 the DNA is bound to the DNeasy membrane. This is followed by two wash steps to 41 42 remove contaminants. The DNA in the spin column was washed with 500 l of 43 44 ethanolbased wash buffer (provided with the kit) and centrifuged for 1 minute at 45 46 6000 x g. The spin column was placed in a new collection tube and washed with 500 47 48 l of ethanolbased wash buffer (provided with the kit) and centrifuged for 3 minutes 49 50 at 20000 x g. DNA was then eluted by pipetting 200 l buffer EB onto the DNeasy 51 52 53 membrane and incubation at room temperature for 1 minute before centrifugation for 54 55 1 minute at 6000 x g. Total RNA (longer than 200 nucleotides) was extracted and 56 57 58 7 59 60 Toxicological Sciences Page 8 of 30

1 2 3 purified using RNeasy Mini kit (Qiagen) according to the manufacturer protocol. In 4 5 brief, the cell pellet was suspended in 350 l of lysis buffer (provided with the kit) 6 7 followed by vortexing. The Lysate was pipetted into QIAshredder spin column and 8 9 centrifuged for 2 min at 20800 x g. One volume of 70% ethanol was added to the 10 11 homogenised lysate and mixed well by pipetting. The mixture was then transferred to 12 13 14 an RNeasy spin column and centrifuged for 15 seconds at 8000 x g. The column 15 16 was washed with 350 l wash buffer (provided with the kit) and centrifuged for 15 17 18 seconds at 8000 x g. DNase (10 l) was added to the spin column and incubated at 19 20 toom temperature for 15 minutes followed by the addition of 350 l wash buffer 21 22 (provided with the kit). Columns were then centrifuged for 15 seconds at 8000 x g 23 24 and washed twice with 500 l ethanol based wash buffer (provided with the kit). RNA 25 26 27 was eluted by addition of 50 l RNasefree water followed by centrifugation for 1 28 29 minute at 8000 x g. DNA and RNA and nucleotide pool extract were hydrolysed as 30 31 described before (Bachman et al., 2014). In brief, DNA or RNA (12 g) was 32 33 incubated with 5 U of DNA Degradase Plus (Zymo Research) in a total volume of 50 34 35 l for 3 h at 37°C. 36 37 LC/MS/MS analysis 38 39 40 Analysis was performed on Acquity UPLC Iclass system. The mobile phase solvent 41 42 A consisted of HPLC graded water containing 0.1% formic acid, and solvent B 43 44 consisted of acetonitrile containing 0.1% formic acid. Analysis was performed using 45 46 a mobile phase flow rate of 0.6 ml/min using a linear gradient starting from 2% 47 48 solvent B to 98% solvent B in 3.5 min and equilibrated as 2% solvent B for 1 min and 49 50 a Waters XSelect HSS T3, 3.5 m HPLC column (2.1 x 50 mm) (Waters). Samples 51 52 53 were analysed on atmospheric pressure ionization (API) 4000 triple quadrupole 54 55 mass spectrometer. Nucleosides were monitored in the selected reaction monitoring 56 57 58 8 59 60 Page 9 of 30 Toxicological Sciences

1 2 3 (SRM) mode using ion atmospheric pressure ionization at a source temperature of 4 5 450 °C and an IonSpray voltage of 4200 V and a collision energy of 30 eV. The dwell 6 7 time for each monitored transition was 30 ms. The SRM transitions were: dA 8 9 252136, dG 268152, dC 228112, T 243127, A 268136, G 284152, C 10 11 244112.1, U 245127, MOEA 326136, MOEG 342152, MOE5mC 12 13 14 316126, MOET 317127, 2’FA 270136, 2’FG 286152, 2’FC 246112, 2’ 15 16 FU 247113. The SRM transitions of nucleoside monophosphates and 17 18 diphosphates in the in the negative ion mode were: dAMP 33079, dADP 33079, 19 20 dGMP 34679, dGDP 33079, dCMP 30679, dCDP 33079, TMP 32179, 21 22 TDP 33079, MOEAMP 40479, MOEADP 33079, MOEGMP 42079, MOE 23 24 GDP 33079, MOE5mCMP 39479, MOE5mCDP 33079, MOETMP 39579, 25 26 27 MOETDP 33079, 2’FAMP 350136, 2’FGMP 366152, 2’FCMP 326112, 2’ 28 29 FUMP 327113Data analysis and peak areas were acquired using Analyst 30 31 Software, v3.3 (Applied BioSystems). The area under curve of the nucleoside 32 33 analogue and the corresponding native or isotopically labelled nucleosides was 34 35 obtained from the extracted ion chromatograms and expressed as a ratio. 36 37 Statistical analysis 38 39 40 The data were tested for statistical significant differences with oneway ANOVA. 41 42 Results 43 44 Phosphorylation of 2′OMOE and 2′F nucleosides 45 46 To be utilised by cellular DNA polymerases for DNA incorporation, nucleosides 47 48 analogues must be converted to their active triphosphate form via sequential 49 50 phosphorylation by various kinases in the nucleotide salvage pathway. Therefore, we 51 52 53 first examined the ability of salvage enzyme activity to convert of 2′OMOE 54 55 nucleosides to their monophosphate forms by liquid chromatography/tandem mass 56 57 58 9 59 60 Toxicological Sciences Page 10 of 30

1 2 2 3 spectrometry (LC/MS/MS). Calibration curves, correlation coefficient (r ), retention 4 5 times and limit of detection (LOD) are shown in Supplementary Figure S1. We 6 7 examined the substrate activity of the two main salvage enzymes, human 8 9 deoxycytidine kinase (dCK) and thymidine kinase 1 (TK1), on 2′OMOE nucleoside 10 11 analogues. We have chosen dCK and TK1 as both are located in the cytosol and 12 13 14 can phosphorylate natural purine and pyrimidine deoxyribonucleosides (Johansson 15 16 and Eriksson, 1996). In addition, comparisons were made with 2′F nucleosides 17 18 widely used in siRNA therapeutics (Watts et al., 2008). To the best of our knowledge, 19 20 there are no reports on the substrate activity of human dCK and TK1 towards 2′F 21 22 nucleosides except 2′FC (Kierdaszuk et al., 1999). Structures of 2′OMOE and 2′F 23 24 nucleosides are shown in Figure 1A. The level of conversion was calculated based 25 26 27 on relative peak area ratios of the reactant before and after the kinase reaction 28 29 (supplementary Fig. S2 for extracted ion chromatograms). As shown in Figure 1B, 30 31 dCK displayed poor reactivity towards 2′OMOE purine nucleosides and was only 32 33 able to convert about 5% of 2′OMOEG to monophosphate and was not able to 34 35 convert 2′OMOEA at all. Similarly, 2′OMOE5meC and 2′OMOET were poor 36 37 substrates for dCK and TK1 phosphorylating 5% of 2′OMOE5mC and 4% of 2′O 38 39 40 MOET respectively (Fig. 2, B). In contrast, 2′F nucleosides were remarkably good 41 42 substrates for both dCK and TK1, regardless of the nucleobase, exhibiting similar 43 44 substrate activities (> 95 % conversion) to their respective natural 2′deoxy 45 46 nucleosides (Fig. 1, B). 47 48 We next examined the capacity of human nucleoside monophosphate kinases 49 50 including cytidylate kinase 1 (CMPK1), thymidylate kinase (dTMPK), adenylate 51 52 53 kinase 1 (AK1) and guanylate kinase 1 (GK1) to phosphorylate 2′OMOE5mC, 2′ 54 55 OMOET, 2′OMOEA and 2′OMOEG monophosphates respectively. As shown 56 57 58 10 59 60 Page 11 of 30 Toxicological Sciences

1 2 3 in Fig.1C, AK1, GK1, CMPK1 and dTMPK were not able to create diphosphates of 4 5 2′OMOE modified nucleoside monophosphates regardless of the nitrogenous base. 6 7 In contrast, AK1 and GK1 were able to phosphorylate (about 95% conversion) dAMP 8 9 and dGMP controls respectively. Similarly, CMPK1 and DTMK were able to 10 11 phosphorylate >70% of CMP and TMP, respectively (Fig. 1, C). Extracted ion 12 13 14 chromatograms are shown in supplementary figure S3. 15 16 Cellular DNA and RNA incorporation of 2′OMOE nucleosides 17 18 The final substrate required by DNA and RNA polymerases for nucleoside 19 20 incorporation into DNA and RNA is the nucleoside triphosphate. Therefore, the 21 22 minimal activity of nucleoside kinases and nucleoside monophosphate kinases on 2′ 23 24 OMOE nucleoside analogues likely to present a major barrier for their incorporation 25 26 27 into cellular DNA and RNA. To confirm this, we asked whether the treatment with 2′ 28 29 OMOE or 2′Fmodified nucleosides leads to their DNA and RNA incorporation in 30 31 TK6 human lymphoblastoid cells. We used TK6 cells as they are humanderived and 32 33 widely used for genotoxicity testing including the in vitro micronucleus assay and 34 35 gene mutation assays (Lorge et al., 2016). It is important to add that TK6 cells are 36 37 p53competent and therefore mount a robust DNA damage response. In addition, 38 39 40 the growth characteristics of TK6 cells and their identity (karyotypes and genetic 41 42 status) are wellcharacterised. Accordingly, we incubated TK6 cells with different 43 44 concentrations (10, 250 and 1000 M) of 2′OMOE nucleosides or their 45 46 monophosphate forms for 24 hr. Comparisons were made to 2’F nucleosides as 47 48 previous results showed that 2’F pyrimidines can be incorporated into cellular DNA 49 50 and RNA (Richardson et al., 2002). Following treatment, DNA and RNA were 51 52 53 extracted and digested to nucleosides and analysed by LC/MS/MS as described 54 55 under the method section. 56 57 58 11 59 60 Toxicological Sciences Page 12 of 30

1 2 3 Effect of 2′OMOE and 2′F nucleosides on DNA and RNA synthesis 4 5 Using the DNA and RNA incorporation of isotopically labelled cytidine C [+3], we first 6 7 assessed the effect 2′OMOE and 2′F nucleoside treatment on DNA synthesis as a 8 9 measure of cell proliferation and overall DNA transcription (RNA synthesis) by 10 11 LC/MS/MS. Results were expressed as % of labelled C [+3] in total dC or C. As 12 13 14 shown in Fig. 2 B and C, treatment with 2′OMOE nucleosides or their 15 16 monophosphate forms had minimal effect on the overall DNA and RNA synthesis 17 18 even at the highest concentration (1000 M) with the level of synthesis remaining 19 20 similar to control untreated cells (pvalue > 0.05). However, there was a 21 22 concentration dependent decrease in DNA synthesis following treatment with all 2′F 23 24 nucleosides except 2′FU (Fig 2, B). For example, treatment with 250 M and 1000 25 26 27 M of 2′FA resulted in 50% and 92% reduction DNA synthesis compared to control 28 29 respectively. When cells were treated with 1000 M of 2′FG and 2′FC, a sharp 30 31 decrease (> 90%) in DNA synthesis compared to control was observed (Fig 2, B). 32 33 Likewise, treatment with all 2′F nucleoside analogous resulted in a concentration 34 35 dependent decrease in RNA synthesis (Fig 2, C). 36 37 Incorporation of 2′OMOE and 2′F nucleosides into cellular DNA and 38 39 40 RNA 41 42 Using the same LC/MS/MS conditions described above, we next evaluated the 43 44 genomic DNA and RNA incorporation of 2′OMOE and 2′F nucleosides expressed 45 46 as a ratio of the AUC of modified nucleosides compared to their respective natural 47 48 equivalents. A schematic of the LC/MS/MS method to evaluate the DNA/RNA 49 50 incorporation of modified nucleosides is shown in Figure 3, A. Figure 3 (B) shows 51 52 53 examples of extracted ion chromatograms derived from hydrolysed DNA and RNA 54 55 from cells treated with 250 M of 2′OMOEG and 2′FG nucleosides. Treating TK6 56 57 58 12 59 60 Page 13 of 30 Toxicological Sciences

1 2 3 cells for 24 hours with 2′OMOE nucleosides or their monophosphates, regardless 4 5 of the nucleobase, did not result in DNA or RNA incorporation using various 6 7 concentrations up to 1000 M. On the other hand, when cells were exposed to 2′F 8 9 nucleosides, we were able to detect incorporation of all nucleosides into DNA except 10 11 2′FU. For example, after treatment with 2′FG (250 M), 12% of total dG contained 12 13 14 2’FG modifications (Figure 3, C). 15 16 As might be expected even higher levels of incorporation of 2’F nucleosides were 17 18 seen for RNA. For example, 2’FC nucleosides accounted for as much as 46% of 19 20 total cytidine content after treatment with 1000 M of 2’FC nucleosides (Fig. 3, C). 21 22 It should be noted that when we analysed undigested DNA or RNA samples, we did 23 24 not observe a detectable signal for any of the natural or modified nucleosides 25 26 27 suggesting efficient removal of extracellular nonincorporated nucleosides from the 28 29 DNA or RNA samples before hydrolysis. 30 31 Intracellular exposure of 2′OMOE nucleosides and their 32 33 monophosphates 34 35 To become incorporated into cellular DNA and RNA, nucleoside analogues must 36 37 display efficient uptake across the plasma membrane. It is possible that the lack of 38 39 40 2′OMOE nucleosides incorporation into cellular DNA and RNA was due to poor 41 42 cellular uptake. Therefore, we determined the intracellular exposure of modified 43 44 nucleotides by measuring the ratio of the signal from the indicated nucleoside to 45 46 endogenous nucleosides in the nucleotide pool extracts. Incubation of TK6 cells with 47 48 various concentrations (10, 250 and 1000 M) of 2′OMOE for 24 hr resulted in a 49 50 significant intracellular uptake that was concentrationdependent (Fig. 4). However, 51 52 53 2′OMOEA and 2′OMOEG showed significantly lower cellular uptake than 2′O 54 55 MOE5mC or 2′OMOET (Fig. 4), the later of which accumulated to higher 56 57 58 13 59 60 Toxicological Sciences Page 14 of 30

1 2 3 concentrations than their native nucleoside counterparts. In general, the cellular 4 5 uptake of 2′OMOE nucleoside monophosphates was lower than their respective 6 7 nucleosides. Nonetheless, we were able to demonstrate intracellular uptake of 2′O 8 9 MOE nucleoside monophosphates that was concentration dependent (Fig. 4). 10 11 Similarly, treating TK6 cells with 2’F nucleosides resulted in a concentration 12 13 14 dependent increases in the cellular uptake up to 1000 M with lower accumulation of 15 16 2’FA, G and –U than 2’FC. 17 18 Discussion 19 20 In the current study, we demonstrated that 2′OMOE nucleosides are neither 21 22 effectively phosphorylated by cytosolic nucleoside kinases, nor are they incorporated 23 24 into cellular DNA or RNA. In contrast, 2′F modified nucleosides are readily 25 26 27 phosphorylated and under our experimental conditions are incorporated at relatively 28 29 high levels into cellular DNA and RNA, which in turn is accompanied by reductions in 30 31 DNA and RNA synthesis. To the best of our knowledge, this is the first report 32 33 evaluating the phosphorylation and DNA/RNA incorporation of 2′OMOE nucleoside 34 35 analogues and provide confidence in the safe application, with regard to liberated 36 37 nucleosides, of oligonucleotide therapeutics containing this modification. 38 39 40 Much of the understanding of nucleoside salvage pathways comes from the area of 41 42 nucleoside analogues used as antiviral and oncology drugs aimed at inhibiting DNA 43 44 synthesis and RNA transcription (Sofia, 2011; Van Rompay et al., 2000). These 45 46 drugs are often given as prodrugs that require phosphorylation by endogenous 47 48 kinases to their triphosphate forms to become incorporated into viral or cellular DNA. 49 50 In general, the formation of the monophosphate is the ratelimiting step in the 51 52 53 activation of nucleosides drugs. 54 55 56 57 58 14 59 60 Page 15 of 30 Toxicological Sciences

1 2 3 Our data suggests that the limited activity of nucleoside salvage kinases towards 2′ 4 5 OMOE nucleosides represents a major barrier for their cellular DNA and RNA 6 7 incorporation. The absence of DNA and RNA incorporation of 2′OMOE nucleosides 8 9 is unlikely to be due to the lack of cell exposure or inhibition of DNA and RNA 10 11 synthesis as we were able to detect 2′OMOE nucleosides in the intracellular 12 13 14 nucleotide pool and 2′OMOE treatments even at the highest concentrations used 15 16 (1000 M) had minimal effect on DNA and RNA synthesis. In addition, the 17 18 LC/MS/MS assay described in this paper to detect 2′OMOE and their respective 19 20 native nucleosides has a limit of detection in the femtomole range which is in line 21 22 with previous reports (Bachman et al., 2014). 23 24 With the exception of 2′FU, all 2′F nucleoside analogues were incorporated into 25 26 27 cellular DNA and RNA indirectly suggesting that 2′F nucleosides were 28 29 phosphorylated to their respective 5′ triphosphates and incorporated by cellular DNA 30 31 and RNA polymerases. This is consistent with previous reports showing that human 32 33 cellular polymerases (pol α and pol γ) could incorporate 2′F nucleosides into primed 34 35 DNA in vitro (Richardson et al., 2000). Later work from the same lab demonstrated 36 37 incorporation of 2′FC and 2′FU into DNA and RNA of various tissues of rats and 38 39 40 woodchucks following longterm in vivo treatment (Richardson et al., 2002). Although 41 42 showing almost complete phosphorylation by TK1, in our hands, 2′FU nucleosides 43 44 were not incorporated into cellular DNA and RNA. One potential reason for this lack 45 46 of incorporation could lie in the limited intracellular exposure observed with the 2’F 47 48 U nucleosides (Figure 4). Also, the calibrated sensitivity of our LC/MS/MS 49 50 methodology for detection of 2’FU was lower than for the other modified 51 52 53 nucleosides (Fig. S1), therefore incorporation might have been below our detectable 54 55 limits. Indeed, the LC/MS/MS assay used by Richardson et al. (2002) was able to 56 57 58 15 59 60 Toxicological Sciences Page 16 of 30

1 2 3 detect 2′FU as low as 400 pM while the limit of 2′FU detection described herein is 4 5 10 nM. In addition, it is possible that the difference in 2′FU incorporation into 6 7 DNA/RNA in vivo and the lack of incorporation in vitro as described in this paper was 8 9 due to variations in the cellular uptake or the phosphorylation capacity of salvage 10 11 enzymes. Variations in the phosphorylation and antiinfluenza activity of 2′F 12 13 14 nucleosides in different cell lines was previously reported (Tisdale et al., 1993), thus 15 16 it is possible that certain phosphorylating enzymes of the nucleotide salvage 17 18 pathway are absent or expressed differently in TK6 cells, potentially inhibiting the 19 20 incorporation of 2’FU into cellular DNA/RNA. 21 22 In contrast to 2′OMOE nucleosides, 2′F nucleosides treatment resulted in 23 24 concentration dependent inhibition of DNA synthesis which was evident with 2′FA, 25 26 27 2′FG and 2′FC. Several previous studies showed that 2′FC inhibits cell growth 28 29 and induces cytostasis due to an Sphase arrest (Brox et al., 1974; Stuyver et al., 30 31 2004). Similarly, data presented here showed that 2′FC reduced DNA synthesis by 32 33 about 30% at 250 M (Fig. 3, B). This inhibitory effect on cell growth could be 34 35 promoted by genomic DNA incorporation potentially leading to double strand DNA 36 37 breaks (Shen et al., 2015). Yet, we cannot exclude other mechanisms of toxicity 38 39 40 including disruption of the endogenous nucleotide pool balance or disruption of 41 42 mitochondrial DNA synthesis (Kunz et al., 1994; PanZhou et al., 2000). Further 43 44 studies are needed to understand the underlying mechanisms responsible for this 45 46 activity of the 2′F nucleosides. However, taken together, these data suggest that 47 48 caution should be applied when considering the use and application of 2′F 49 50 modifications in therapeutic oligonucleotides. 51 52 53 Although nonnatural 2′sugar modifications of therapeutic oligonucleotides can 54 55 enhance target binding affinity and nuclease stability, the potential of 2′nucleosides 56 57 58 16 59 60 Page 17 of 30 Toxicological Sciences

1 2 3 analogues to become incorporated into genomic DNA potentially leading to DNA 4 5 replication errors or mutations raises a safety concern for therapeutic 6 7 oligonucleotides. In general, 2′F nucleoside analogues are commonly used in 8 9 siRNAbased therapeutics to increase binding affinity and reduce immune activation 10 11 (Watts et al., 2008). However, only a limited number of nuclease resistant 12 13 14 phosphorothioate linkages can be tolerated while still providing siRNA efficacy. 15 16 Moreover, while 2′F modification confers increased hybridisation affinity, it does not 17 18 confer nuclease resistance (Manoharan, 1999). These two factors are likely to 19 20 increase the potential of 2′F catabolism and nucleoside release and therefore 21 22 subsequent conversion by cellular kinases to their triphosphate forms for 23 24 incorporation into newly synthesised DNA. This could have genotoxic consequences 25 26 27 as 2′F sugar modifications have been demonstrated to destabilise BDNA duplex by 28 29 inducing a more Alike conformation (Ikeda et al., 1998). Furthermore, once 30 31 incorporated into DNA, 2′F nucleosides appear resistant to normal mechanism of 32 33 DNA repair as a fluorine at C2′ blocks the activity of glycosylases, thus inhibiting 34 35 base excision repair (BER) for these analogues (Schroder et al., 2016; Su et al., 36 37 2016). Notwithstanding these considerations, chronic safety studies in rats and 38 39 40 woodchucks have shown that despite 2′FU and 2′FC incorporation into tissue 41 42 DNA and RNA, there are minimal toxicological consequences of the incorporation 43 44 (Richardson et al., 1999; Richardson et al., 2002). Furthermore, to our knowledge 45 46 the siRNA constructs that utilize the 2′F modified nucleosides have been negative in 47 48 genetic toxicity studies, although there are limited primary references from which to 49 50 draw (Berman et al., 2016; Janas et al., 2016). 51 52 53 In contrast to the more metabolically labile siRNA, single stranded antisense 54 55 oligonucleotides are usually designed with phosphorothioate backbone linkages 56 57 58 17 59 60 Toxicological Sciences Page 18 of 30

1 2 3 throughout the molecule with the modified nucleotides at terminal ends (e.g. 2′O 4 5 MOE) providing further nuclease resistance (Bennett and Swayze, 2010). Therefore, 6 7 the nucleosides concentrations used in this in vitro study are unlikely to be achieved 8 9 following in vivo administration due to the slow degradation of oligonucleotides 10 11 (Watanabe et al., 2006). Ultimately, the 2’OMOE modified wings are largely 12 13 14 excreted in urine with limited degradation of the modified nucleotides (Geary et al., 15 16 2001b). Although they are not incorporated into DNA and RNA, 2′OMOE 17 18 nucleosides could exert mutagenic consequences via a mechanism other than direct 19 20 DNA incorporation such as nucleotide pool balance disruption although this potential 21 22 mechanism is unlikely given the lack of phosphorylation, the slow rate of metabolism 23 24 and the consistent lack of genotoxicity that has been reported (Berman et al., 2016; 25 26 27 Henry et al., 2002). 28 29 The utility of nonnatural 2′sugar modification in oligonucleotidebased gene 30 31 silencing technologies has proven valuable for their clinical development. These 2′ 32 33 sugar modifications confer an RNAlike C3′endo conformation which enhances the 34 35 binding affinity to the target RNA (Kawasaki et al., 1993). The chemical structure of 36 37 2′OMOE and 2′F nucleosides analogous closely mimics the natural nucleosides 38 39 40 where the nucleobase component is unmodified. It is possible that the reduced steric 41 42 interactions of the less bulky fluorine atom combined with its higher hydrophobicity is 43 44 likely responsible for the higher ability of kinase salvage enzymes and cellular 45 46 polymerases to accept 2′F nucleoside analogue as a substrate resulting in their 47 48 incorporation into genomic DNA. Further understanding of structureactivity 49 50 relationship will help in the design of novel chemical modifications at the 2′postion of 51 52 53 the sugar moiety with improved chemical and biological stability. 54 55 56 57 58 18 59 60 Page 19 of 30 Toxicological Sciences

1 2 3 In conclusion, we found that 2′OMOE nucleoside analogues are poor substrates for 4 5 salvage enzymes largely preventing the formation of their corresponding active 5′ 6 7 triphosphates and subsequent incorporation into cellular DNA and RNA. To our 8 9 knowledge, this is the first study exploring the phosphorylation and DNA 10 11 incorporation of 2′OMOE modified nucleotides. It is clear from the current study that 12 13 14 there is a link between the substrate specificity of nucleotide salvage enzymes 15 16 towards 2′sugar modified nucleoside analogues and cellular DNA incorporation. We 17 18 believe the current work provides insight for the design of nucleoside analogues with 19 20 higher activity and improved safety profile when incorporated into therapeutic 21 22 oligonucleotides. 23 24 References 25 26 27 Bachman, M., UribeLewis, S., Yang, X., Williams, M., Murrell, A., and 28 29 Balasubramanian, S. (2014). 5Hydroxymethylcytosine is a predominantly stable 30 31 DNA modification. Nat Chem 6(12), 104955. 32 33 Bennett, C. F., and Swayze, E. E. (2010). RNA targeting therapeutics: molecular 34 35 mechanisms of antisense oligonucleotides as a therapeutic platform. Annu Rev 36 37 Pharmacol Toxicol 50, 25993. 38 39 40 Berman, C. L., Barros, S. A., Galloway, S. M., Kasper, P., Oleson, F. B., Priestley, C. 41 42 C., Sweder, K. S., Schlosser, M. J., and Sobol, Z. (2016). OSWG Recommendations 43 44 for Genotoxicity Testing of Novel OligonucleotideBased Therapeutics. Nucleic Acid 45 46 Ther 26(2), 7385. 47 48 Brox, L. W., LePage, G. A., Hendler, S. S., and Shannahoff, D. H. (1974). Studies on 49 50 the growth inhibition and metabolism of 2'deoxy2'fluorocytidine in cultured human 51 52 53 lymphoblasts. Cancer Res 34(8), 183842. 54 55 56 57 58 19 59 60 Toxicological Sciences Page 20 of 30

1 2 3 Crooke, S. T., and Geary, R. S. (2013). Clinical pharmacological properties of 4 5 mipomersen (Kynamro), a second generation antisense inhibitor of apolipoprotein B. 6 7 Br J Clin Pharmacol 76(2), 26976. 8 9 Deleavey, G. F., and Damha, M. J. (2012). Designing chemically modified 10 11 oligonucleotides for targeted gene silencing. Chem Biol 19(8), 93754. 12 13 14 Eckstein, F. (2000). Phosphorothioate oligodeoxynucleotides: what is their origin and 15 16 what is unique about them? Antisense Nucleic Acid Drug Dev 10(2), 11721. 17 18 Geary, R. S., Khatsenko, O., Bunker, K., Crooke, R., Moore, M., Burckin, T., Truong, 19 20 L., Sasmor, H., and Levin, A. A. (2001a). Absolute bioavailability of 2'O(2 21 22 methoxyethyl)modified antisense oligonucleotides following intraduodenal instillation 23 24 in rats. J Pharmacol Exp Ther 296(3), 898904. 25 26 27 Geary, R. S., Watanabe, T. A., Truong, L., Freier, S., Lesnik, E. A., Sioufi, N. B., 28 29 Sasmor, H., Manoharan, M., and Levin, A. A. (2001b). Pharmacokinetic properties of 30 31 2'O(2methoxyethyl)modified oligonucleotide analogs in rats. J Pharmacol Exp 32 33 Ther 296(3), 8907. 34 35 Henry, S. P., Monteith, D. K., Matson, J. E., Mathison, B. H., Loveday, K. S., 36 37 Winegar, R. A., Matson, J. E., Lee, P. S., Riccio, E. S., Bakke, J. P., et al. (2002). 38 39 40 Assessment of the genotoxic potential of ISIS 2302: a phosphorothioate 41 42 oligodeoxynucleotide. Mutagenesis 17(3), 2019. 43 44 Ikeda, H., Fernandez, R., Wilk, A., Barchi, J. J., Jr., Huang, X., and Marquez, V. E. 45 46 (1998). The effect of two antipodal fluorineinduced sugar puckers on the 47 48 conformation and stability of the DickersonDrew dodecamer duplex 49 50 [d(CGCGAATTCGCG)]2. Nucleic Acids Res 26(9), 223744. 51 52 53 Janas, M. M., Jiang, Y., Duncan, R. G., Hayes, A. N., Liu, J., Kasperkovitz, P. V., 54 55 Placke, M. E., and Barros, S. A. (2016). Exposure to siRNAGalNAc Conjugates in 56 57 58 20 59 60 Page 21 of 30 Toxicological Sciences

1 2 3 Systems of the Standard Test Battery for Genotoxicity. Nucleic Acid Ther 26(6), 363 4 5 371. 6 7 Johansson, N. G., and Eriksson, S. (1996). Structureactivity relationships for 8 9 phosphorylation of nucleoside analogs to monophosphates by nucleoside kinases. 10 11 Acta Biochim Pol 43(1), 14360. 12 13 14 Kawasaki, A. M., Casper, M. D., Freier, S. M., Lesnik, E. A., Zounes, M. C., 15 16 Cummins, L. L., Gonzalez, C., and Cook, P. D. (1993). Uniformly modified 2'deoxy 17 18 2'fluoro phosphorothioate oligonucleotides as nucleaseresistant antisense 19 20 compounds with high affinity and specificity for RNA targets. J Med Chem 36(7), 21 22 83141. 23 24 Kierdaszuk, B., Krawiec, K., Kazimierczuk, Z., Jacobsson, U., Johansson, N. G., 25 26 27 MunchPetersen, B., Eriksson, S., and Shugar, D. (1999). Substrate/inhibitor 28 29 properties of human deoxycytidine kinase (dCK) and thymidine kinases (TK1 and 30 31 TK2) towards the sugar moiety of nucleosides, including O'alkyl analogues. 32 33 Nucleosides Nucleotides 18(8), 1883903. 34 35 Kole, R., and Leppert, B. J. (2012). Targeting mRNA splicing as a potential treatment 36 37 for Duchenne muscular dystrophy. Discov Med 14(74), 5969. 38 39 40 Kunz, B. A., Kohalmi, S. E., Kunkel, T. A., Mathews, C. K., McIntosh, E. M., and 41 42 Reidy, J. A. (1994). International Commission for Protection Against Environmental 43 44 Mutagens and Carcinogens. Deoxyribonucleoside triphosphate levels: a critical 45 46 factor in the maintenance of genetic stability. Mutat Res 318(1), 164. 47 48 Lorge, E., Moore, M. M., Clements, J., O'Donovan, M., Fellows, M. D., Honma, M., 49 50 Kohara, A., Galloway, S., Armstrong, M. J., Thybaud, V., et al. (2016). Standardized 51 52 53 cell sources and recommendations for good cell culture practices in genotoxicity 54 55 testing. Mutat Res 809, 115. 56 57 58 21 59 60 Toxicological Sciences Page 22 of 30

1 2 3 Manoharan, M. (1999). 2'carbohydrate modifications in antisense oligonucleotide 4 5 therapy: importance of conformation, configuration and conjugation. Biochim Biophys 6 7 Acta 1489(1), 11730. 8 9 Mattano, S. S., Palella, T. D., and Mitchell, B. S. (1990). Mutations induced at the 10 11 hypoxanthineguanine phosphoribosyltransferase locus of human Tlymphoblasts by 12 13 14 perturbations of purine deoxyribonucleoside triphosphate pools. Cancer Res 50(15), 15 16 456671. 17 18 Monia, B. P. (1997). First and secondgeneration antisense oligonucleotide 19 20 inhibitors targeted against human craf kinase. Ciba Found Symp 209, 10719; 21 22 discussion 11923. 23 24 Monia, B. P., Lesnik, E. A., Gonzalez, C., Lima, W. F., McGee, D., Guinosso, C. J., 25 26 27 Kawasaki, A. M., Cook, P. D., and Freier, S. M. (1993). Evaluation of 2'modified 28 29 oligonucleotides containing 2'deoxy gaps as antisense inhibitors of gene 30 31 expression. J Biol Chem 268(19), 1451422. 32 33 PanZhou, X. R., Cui, L., Zhou, X. J., Sommadossi, J. P., and DarleyUsmar, V. M. 34 35 (2000). Differential effects of antiretroviral nucleoside analogs on mitochondrial 36 37 function in HepG2 cells. Antimicrob Agents Chemother 44(3), 496503. 38 39 40 Phear, G., Nalbantoglu, J., and Meuth, M. (1987). Nextnucleotide effects in 41 42 mutations driven by DNA precursor pool imbalances at the aprt locus of Chinese 43 44 hamster ovary cells. Proc Natl Acad Sci U S A 84(13), 44504. 45 46 Prakash, T. P. (2011). An overview of sugarmodified oligonucleotides for antisense 47 48 therapeutics. Chem Biodivers 8(9), 161641. 49 50 Richardson, F. C., Kuchta, R. D., Mazurkiewicz, A., and Richardson, K. A. (2000). 51 52 53 Polymerization of 2'fluoro and 2'OmethyldNTPs by human DNA polymerase 54 55 alpha, polymerase gamma, and primase. Biochem Pharmacol 59(9), 104552. 56 57 58 22 59 60 Page 23 of 30 Toxicological Sciences

1 2 3 Richardson, F. C., Tennant, B. C., Meyer, D. J., Richardson, K. A., Mann, P. C., 4 5 McGinty, G. R., Wolf, J. L., Zack, P. M., and Bendele, R. A. (1999). An evaluation of 6 7 the toxicities of 2'fluorouridine and 2'fluorocytidineHCl in F344 rats and 8 9 woodchucks (Marmota monax). Toxicol Pathol 27(6), 60717. 10 11 Richardson, F. C., Zhang, C., Lehrman, S. R., Koc, H., Swenberg, J. A., Richardson, 12 13 14 K. A., and Bendele, R. A. (2002). Quantification of 2'fluoro2'deoxyuridine and 2' 15 16 fluoro2'deoxycytidine in DNA and RNA isolated from rats and woodchucks using 17 18 LC/MS/MS. Chem Res Toxicol 15(7), 9226. 19 20 Schroder, A. S., Kotljarova, O., Parsa, E., Iwan, K., Raddaoui, N., and Carell, T. 21 22 (2016). Synthesis of (R)Configured 2'Fluorinated mC, hmC, fC, and caC 23 24 Phosphoramidites and Oligonucleotides. Org Lett 18(17), 436871. 25 26 27 Shahbazi, R., Ozpolat, B., and Ulubayram, K. (2016). Oligonucleotidebased 28 29 theranostic nanoparticles in cancer therapy. Nanomedicine (Lond) 11(10), 1287308. 30 31 Shen, W., Liang, X. H., Sun, H., and Crooke, S. T. (2015). 2'Fluoromodified 32 33 phosphorothioate oligonucleotide can cause rapid degradation of P54nrb and PSF. 34 35 Nucleic Acids Res 43(9), 456978. 36 37 Sofia, M. J. (2011). Nucleotide prodrugs for HCV therapy. Antivir Chem Chemother 38 39 40 22(1), 2349. 41 42 Stuyver, L. J., McBrayer, T. R., Whitaker, T., Tharnish, P. M., Ramesh, M., Lostia, 43 44 S., Cartee, L., Shi, J., Hobbs, A., Schinazi, R. F., et al. (2004). Inhibition of the 45 46 subgenomic hepatitis C virus replicon in huh7 cells by 2'deoxy2'fluorocytidine. 47 48 Antimicrob Agents Chemother 48(2), 6514. 49 50 Su, M., Kirchner, A., Stazzoni, S., Muller, M., Wagner, M., Schroder, A., and Carell, 51 52 53 T. (2016). 5Formylcytosine Could Be a Semipermanent Base in Specific Genome 54 55 Sites. Angew Chem Int Ed Engl 55(39), 11797800. 56 57 58 23 59 60 Toxicological Sciences Page 24 of 30

1 2 3 Tisdale, M., Appleyard, G., Tuttle, J. V., Nelson, D. J., NusinoffLehrman, S., Al 4 5 Nakib, W., Stables, J. N., Purifoy, D. J. M., and Darby, G. (1993). Inhibition of 6 7 influenza A and B viruses by 2′deoxy2′fluororibosides. Antivir. Chem. Chemother. 8 9 4(5):281287. 10 11 Van Rompay, A. R., Johansson, M., and Karlsson, A. (2000). Phosphorylation of 12 13 14 nucleosides and nucleoside analogs by mammalian nucleoside monophosphate 15 16 kinases. Pharmacol Ther 87(23), 18998. 17 18 Watanabe, T. A., Geary, R. S., and Levin, A. A. (2006). Plasma protein binding of an 19 20 antisense oligonucleotide targeting human ICAM1 (ISIS 2302). Oligonucleotides 21 22 16(2), 16980. 23 24 Watts, J. K., Deleavey, G. F., and Damha, M. J. (2008). Chemically modified siRNA: 25 26 27 tools and applications. Drug Discov Today 13(1920), 84255. 28 29 30 31 32 33 Figure legends 34 35 Figure 1. Phosphorylation of 2′modified nucleoside analogues. A) Molecular 36 37 structures of deoxyribonucleosdies, 2′OMOE and 2′F nucleoside analogues, B) 38 39 40 The ability of dCK and TK1 to phosphorylate was analysed by incubating 200 M of 41 42 the indicated nucleoside with 1 mM ATP and 1 g of dCK (for dA, dG, dC and their 43 44 2′OMOE analogues) or TK1 (for T and 2′OMOET) in a 50 l reaction volume at 45 46 37 °C for 2 h. The samples were then then analysed by LC/M/SMS and the % 47 48 conversion was calculated based on the peak area of the substrate before and after 49 50 the addition of the enzyme. C) The ability of cytidylate kinase 1 (CMPK1), 51 52 53 thymidylate kinase (dTMPK), adenylate kinase 1 (AK1) and guanylate kinase 1 54 55 (GK1) to phosphorylate natural and 2′OMOEmodified CMP, TMP, dAMP, dGMP 56 57 58 24 59 60 Page 25 of 30 Toxicological Sciences

1 2 3 respectively was analysed by as indicated above. N.D, not detected. Means ± S.D. 4 5 of three independent experiments are indicated. 6 7 Figure 2. Effect of 2′OMOE and 2′F nucleosides on cellular DNA and RNA 8 9 synthesis.. A) DNA synthesis: TK6 cells were incubated with various concentrations 10 11 of 2′OMOE, 2′OMOE monophosphate and 2′F nucleosides for 24 h. The 12 13 14 extracted DNA from cells were hydrolysed to nucleoside and the base composition 15 16 (dC, dC+3) was then analysed by LC/MS/MS. Data show the abundance of dC+3 17 18 relative to dC. B) RNA synthesis: TK6 cells were incubated with various 19 20 concentrations of 2′OMOE, 2′OMOE monophosphate and 2′F nucleosides for 24 21 22 h. The extracted RNA from cells were hydrolysed to nucleoside and the base 23 24 composition (C and C+3) was then analysed by LC/MS/MS. Data show the 25 26 27 abundance of C+3 relative to C. Means ± S.D. of three independent experiments are 28 29 indicated. **p < 0.01, ***p < 0.001 and ****p < 0.0001 as compared with control 30 31 untreated cells. 32 33 Figure 3. A) Schematic of the LC/MS/MS method to evaluate nucleoside analogue 34 35 incorporation into cellular DNA and RNA. B) Extracted ion chromatograms of 36 37 nucleosides from digested DNA and RNA of TK6 cells treated with 250 M of 2′O 38 39 40 MOEG and 2′FG nucleosides for 24 hr analysed by LC/MS/MS. C) Incorporation of 41 42 2′F nucleosides into cellular DNA and RNA. TK6 cells were treated with various 43 44 concentrations of 2′F for 24h. The DNA and RNA was then extracted and 45 46 hydrolysed to nucleosides. Incorporation of 2′F nucleosides was expressed as the 47 48 ratio of AUC of 2′F nucleosides relative to their natural counterpart. Means ± S.D. of 49 50 three independent experiments are indicated. **p < 0.01, ***p < 0.001 and ****p < 51 52 53 0.0001. ns, not significant. 54 55 56 57 58 25 59 60 Toxicological Sciences Page 26 of 30

1 2 3 Figure 4. Cellular exposure to 2′OMOE and 2′F nucleosides analogous. TK6 cells 4 5 were treated for 24 hr with various concentrations of 2′OMOE, 2′OMOE 6 7 monophosphate and 2′F nucleosides. The intracellular nucleotide pool was then 8 9 extracted and digested to nucleosides. The cellular uptake was expressed as the 10 11 ratio of AUC of modified nucleosides relative to their natural counterpart. Means ± 12 13 14 S.D. of three independent experiments are indicated. *p < 0.05, **p < 0.01, ***p < 15 16 0.001 and ****p < 0.0001 as compared with control untreated cells. 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 26 59 60 Page 27 of 30 Toxicological Sciences

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Figure 1. 46 47 282x382mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Toxicological Sciences Page 28 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Figure 2 28 29 161x99mm (300 x 300 DPI) 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 29 of 30 Toxicological Sciences

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Figure 3 46 47 271x346mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Toxicological Sciences Page 30 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Figure 4 23 24 110x50mm (300 x 300 DPI) 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60