Structure-activity relationships for osmium(II) arene phenylazopyridine anticancer complexes functionalised with alkoxy and glycolic substituents.

Journal Pre-proof Structure-activity relationships for osmium(II) arene phenylazopyridine anticancer complexes functionalised with alkoxy and glycolic substituents Russell J. Needham, Hannah E. Bridgewater, Isolda Romero- Canelón, Abraha Habtemariam, Guy J. Clarkson, Peter J. Sadler PII: S0162-0134(20)30182-3 DOI: https://doi.org/10.1016/j.jinorgbio.2020.111154 Reference: JIB 111154 To appear in: Journal of Inorganic Biochemistry Received date: 9 April 2020 Revised date: 9 June 2020 Accepted date: 11 June 2020 Please cite this article as: R.J. Needham, H.E. Bridgewater, I. Romero-Canelón, et al., Structure-activity relationships for osmium(II) arene phenylazopyridine anticancer complexes functionalised with alkoxy and glycolic substituents, Journal of Inorganic Biochemistry (2020), https://doi.org/10.1016/j.jinorgbio.2020.111154 This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process,errorsmaybediscoveredwhichcouldaffectthecontent,andalllegaldisclaimers that apply to the journal pertain. © 2020 Published by Elsevier. Journal Pre-proof Structure-Activity Relationships for Osmium(II) Arene Phenylazopyridine Anticancer Complexes Functionalised with Alkoxy and Glycolic Substituents Russell J. Needham, Hannah E. Bridgewater, Isolda Romero-Canelón, Abraha Habtemariam, Guy J. Clarkson and Peter J. Sadler*. f Department of Chemistry, University of Warwick, Gibbet Hill Road, Coventry, CV4 7AL, U.K. o o r p GRAPHIC: - e r Cell P Cycle l Solubility a IC 50 n r u o Cyclic Cell J Voltammetry Uptake ROS Lipophilicity Apoptosis 1 Journal Pre-proof ABSTRACT: Twenty-four novel organometallic osmium(II) phenylazopyridine (AZPY) complexes have been synthesised and characterised; [Os(η6-arene)(5-RO-AZPY)X]Y, where arene = p-cym or bip, AZPY is functionalized with an alkoxyl (O-R, R = Me, Et, nPr, iPr, nBu) or glycolic (O-{CH CH O} R*, n = 1-4, R* = H, Me, or Et) substituent on 2 2 n the pyridyl ring para to the azo-bond, X is a monodentate halido ligand (Cl, Br or I), and Y is a counter-anion (PF -, CF SO - or IO -). X-ray crystal structures of two complexes 6 3 3 3 confirmed their ‘half-sandwich’ structures. Aqueous solubility depended on X, the f o AZPY substituents, arene, and Y. Iodido complexes are highly stable in water (X = I o >>> Br > Cl), and exhibit the highest antiproliferative activity against A2780 (ovarian), r p MCF-7 (breast), SUNE1 (nasopharyngeal), and OE19 (oesophageal) cancer cells, some - attaining nanomolar potency and goode cancer-cell selectivity. Their activity and r distinctive mechanism of action is discussed in relation to hydrophobicity (RP-HPLC P capacity factor and Log P ), cellular accumulation, electrochemical reduction o/w l a (activation of azo bond), cell cycle analysis, apoptosis and induction of reactive oxygen n species (ROS). Two complexes show 4× higher activity than cisplatin in the National r u Cancer Institute (NCI) 60-cell line five-dose screen. The COMPARE algorithm of their o datasets reveals a strong correlation with one another, as well as anticancer agents J olivomycin, phyllanthoside, bouvardin and gamitrinib, but only a weak correlation with cisplatin, indicative of a different mechanism of action. 2 Journal Pre-proof 1. Introduction Some of the first phenylazopyridine (AZPY) complexes of Ru(II) reported in 1979 exhibited strong dπ-pπ metal-ligand interactions, which endow them with high stability [1]. These effects were later utilised in both Ru(II) and Os(II) AZPY arene anti-cancer complexes [2-6]. Interestingly, Os(II) arene complexes of azobipyridine and Ru(II) arene complexes of AZPY have ligand-based redox properties [2, 7]. Their cyclic voltammogram reduction potentials can be attributed tfo reduction of the azo-bond, o which contains a low lying π*-orbital capable of aoccepting two electrons. Previously- r reported piano-stool Os(II) arene AZPY complexes exhibit potent anti-proliferative p activity against a variety of cancer cell lines [8]. Their mechanism of action differs greatly - e from earlier Os(II) arene complexes with σ-donor bidentate ligands, which rely primarily r P on dissociation of a monodentate ligand followed by DNA nucleobase binding [9-13]. In contrast, the strong π-acceptor character of AZPYs produces different characteristics, l a such as resistance towards hydrolysis and a mechanism of action involving increased n intracellular reactive oxygern species (ROS) levels [4-6, 14]. u Current structure-activity relationships (SARs) reported for Os(II) arene AZPY o complexes includJe the ability of monodentate iodido ligands to improve anti-cancer activity significantly over chlorido ligands, enhancement of activity by biphenyl (bip) over p-cymeme (p-cym) arene ligands, and ability of AZPY ligand substituents positioned on the pyridyl and phenyl moieties para to the azo-bond also to enhance activity [4, 5, 15]. Anticancer drug design concepts often make use of Lipinski's 'rule of five' (Ro5) to determine whether a drug candidate has pharmacological properties which make it suitable for absorption, distribution, metabolism and excretion within the human body [16, 17]. A variation to Ro5 suggests that the effective range of Log P is -0.4 to o/w +5.6 [18] and varying ligand substituents is an effective strategy for tuning metallodrug 3 Journal Pre-proof lipophilicity, cell uptake and activity. The inclusion of alkyl substituents generally increases drug lipophilicity [19] and in a class of Pt(IV) drugs, long chain alkyl substituents that mimic fatty acids improved cell uptake and human serum albumin interactions, resulting in enhanced activity [20]. In contrast, polyethylene glycol is a strongly solubilising substituent, which improves the aqueous solubility of hydrophobic molecules [21, 22]. Moreover, polyethylene glycol can be utilised to encapsulate hydrophobic organic anticancer drugs as nano-micelles, drastically improving their f o bioavailability and delivery [23]. There appear to be no reports of varying the chain o length and tail groups of glycolic substituents to modulate metallodrug properties, r p however, they have been incorporated into homobinuclear Au(I) and Ru(II) anticancer - complexes to create variable length linkagees between metal centres [24]. r Here we explore the effects of changes in the charge polarisation of Os(II) arene P AZPY complexes on activity towards cancer cells, using a variety of alkoxy and glycolic l a side-chains on AZPY ligands. A series of 24 new Os(II) arene AZPY complexes have n been synthesised and the SARs explored for activity against A2780 human ovarian r cancer cells, solubility uin aqueous media, RP-HPLC capacity factors (lipophilicity), and o cellular uptake. We also explore variation of the counter-anion as a means of improving J aqueous solubility. Fig. 1 lists the synthesised complexes 1-24. Selected complexes were further screened against other cancer and normal cell lines (MCF-7 breast, SUNE1 nasopharyngeal, OE19 oesophageal and MRC-5 lung), and their ability to induce apoptosis, elevate ROS levels, and induce cell cycle arrest in A2780 cells were studied. 4 Journal Pre-proof 2. Experimental 2.1. Materials OsCl ·3H O was purchased from either Sigma Aldrich (UK) or Heraeus (South 3 2 Africa). Most chemicals were purchased from Sigma Aldrich (UK): α-terpinene, biphenyl, 2-amino-5-fluoropyridine, nitrosobenzene, ammonium hexafluorophosphate, ammonium triflate, potassium iodate. The alcohols/glycols used for synthesising RO- f AZPY ligands were also purchased from Sigma Aldricoh, with the exception of ethylene o glycol and diethylene glycol, which were obtained from Alfa Aesar (UK). All other r organic solvents and reagents for synthesips and analysis were purchased from - commercial suppliers and were used as received. Dimers; [Os(η6-p-cym)X ] , where X = e 2 2 Cl, Br or I, were prepared accordingr to reported procedures [25-27] as was 2- P (phenylazo)-5-fluoropyridine [4]. Dulbecco’s Modified Eagle Media, Roswell Park l Memorial Institute-1640 cell caulture medium, penicillin/streptomycin mixture, foetal n bovine serum, L-gutamine, phosphate buffered saline solution (PBS), trypsin, and r trypsin/ethylenediaminetetraacetic acid were all purchased from PAA Laboratories u GmbH. Cell cultureo media (500 mL) were supplemented with foetal bovine serum (50 J mL), penicillin/streptomycin mixture (5 mL) and L-glutamine (5 mL). Cancer cell lines were purchased from the European Collection of Cell Cultures and Public Health England. 2.2. Synthesis of ligands 2-(Phenylazo)-5-fluoropyridine (100.0 mg, 0.50 mmol) was dissolved in ROH (20 mL) and an aqueous KOH solution (5 mol equiv, 0.4 g/mL, 349 µL) was added. The mixture was heated under reflux or at 120 °C (depending on the b.p. of ROH) for 18 h. The product was extracted with CH Cl (50 mL) and washed with water (3 x 50 mL). 2 2 5 Journal Pre-proof The CH Cl extract was dried over MgSO , filtered, and the solvent was removed under 2 2 4 reduced pressure. Some reaction mixtures required vacuum distillation to remove excess ROH. Where necessary, purification was performed by re-crystallisation or SiO column 2 chromatography and the product was dried overnight on a vacuum line. Characterisation and purification methods used for ligands L1-L13 are reported in the SI. 2.3. Synthesis of complexes f o o [Os(η6-arene)X ] (where X = Cl, Br or I, and arene = p-cym or bip) was dissolved 2 2 r in EtOH (10 mL), and a solution of RO-AZPpY (2.1 mol. equiv., L1-L13) in EtOH (5 - mL) was added drop-wise. The mixture was stirred for 18 h at ambient temperature, and e NH PF , NH CF SO or KIO (10 molr. equiv.) was added. Complexes were isolated 4 6 4 3 3 3 P using either method 1 or 2. Method 1 involves re-crystallisation. Some complexes proved l difficult to re-crystallise and meathod 2 was employed. In some cases, further purification n via automated reverse-phase chromatography was required. Method 1: The ethanolic r mixture was concentratued under reduced pressure to ~3 mL and placed in the freezer (- 20 °C) overnight. Ao dark crystalline precipitate formed which was collected via vacuum J filtration, washed with ice-cold EtOH (2 x 1 mL), then Et O (2 x 5 mL), and dried 2 overnight in a vacuum desiccator. Method 2: The ethanolic mixture was concentrated under reduced pressure to ~3 mL and n-hexane was added (~1 mL). The mixture was placed in a freezer (-20 °C) overnight, forming a dark-brown amorphous residue. The solvents were decanted and the residue was washed with Et O (3 x 5 mL). The residue 2 was redissolved in CH Cl (5 mL), transferred to a pre-weighed vial, and dried overnight 2 2 on a vacuum line. Reagent quantities, characterisation and isolation methods employed for complexes 1-24 are reported in the SI. 6 Journal Pre-proof 2.4. X-ray crystallography Crystals of 1 were obtained by cooling a methanolic solution of 1-2 mg/mL in a freezer at -20 ºC. Crystals of 8 were obtained by slow evaporation of a methanol solution at ambient temperature. Diffraction data were collected on an Oxford Diffraction f o Gemini four-circle system with a Ruby CCD area detector. All structures were refined by o full-matrix least squares against F2 using SHELXL 97 and were solved by direct methods r p using SHELXS(TREF) with additional light atoms found by Fourier methods [28]. - e Hydrogen atoms were added at calculated positions and refined using a riding model. r Anisotropic displacement parameterPs were used for all non-H atoms; H-atoms were given an isotropic displacement pa rameter equal to 1.2 (or 1.5 for methyl and NH H- l a atoms) times the equivalent isotropic displacement parameter of the atom to which they n are attached. The data were processed by the modelling program Mercury 1.4.1. X-ray r u crystallographic data for complexes 1 and 8 have been deposited in the Cambridge o Crystallographic Data Centre under the accession numbers CCDC1990713 and J CCDC1990712, respectively. 2.5. Hydrolysis studies Solutions of complexes (100 µM) were prepared in phosphate buffer solution (95 mol equiv, pH 7.4, 5% DMSO) with various concentrations of NaCl; 23 mM (intracellular conditions) and 103 mM (extracellular conditions). The samples were incubated at 37 ºC for 0, 2, 12 and 24 h and analysed via reverse-phase high-pressure 7 Journal Pre-proof liquid chromatography (RP-HPLC). No precautions were made to protect samples from air or light. 2.6. RP-HPLC capacity factors Isocratic RP-HPLC analysis of complexes and uracil was carried out (Instrument and column described in the SI). The mobile phase consisted of H O:MeCN (1:1, v/v, 50 2 mM NaCl) and the temperature of the column was maintained constant at 25 °C. f Samples (500 µM) were prepared in H O:MeCN (1o:1, v/v) and were analysed in 2 o triplicates in three separate experiments (50 µL injection volume). Their capacity factors r were calculated using the following equationps, where t is the retention time of the R - retained complex, t is the retention time of unretained compound (uracil), R is the 0 e F retention factor, and K is the capacity facrtor. P R = (t - t ) (1) F R 0 l a K = R /t (2) nF 0 r 2.7. Partition coefficients u o Octanol/water partition coefficients (Log P ) of complexes were determined using o/w J an adaptation of the shake-flask method [29, 30]. Octanol-saturated water (OSW) and water-saturated octanol (WSO) were prepared by stirring octanol (100 mL) and doubly- deionised water (100 mL) together for 24 h and separating the two layers. Prior to mixing, the water was supplemented with NaX (300 mM), where X = Cl, Br, or I, added to suppress the hydrolysis of chlorido, bromido, or iodido complexes, respectively. Complexes (0.5–3.0 mg) were dissolved in OSW (3 mL) with vigorous stirring/shaking for 24 h, then filtered. Saturated solutions of complexes in OSW (1 mL) were combined with WSO (1 mL) in falcon tubes and placed on a Vibrax VXB basic mechanical shaker 8 Journal Pre-proof for 24 h at 500 g/min. Samples of aqueous phases were collected before and after partitioning and their osmium concentrations were determined by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) or Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES). Partition coefficients were calculated via equation (3). Experiments were performed as duplicates of triplicates in two independent experiments at ambient temperature. Log P = Log ([Os] /[Os] ) f(3) o/w octanol water o o 2.8. Electrochemistry r p Cyclic voltammetry analyses were performed using a CH Instruments - e Electrochemical Analyzer (CHI420C) and CH Instruments electrochemistry software. r Compounds were dissolved in acetonitrile (1 mg/mL) with nBu NPF (0.1 M) as a P 4 6 supporting electrolyte. Solutions w ere degassed under N and scanned between -2.0 V 2 l a and +2.0 V at scan rates of 0.1 or 0.5 V/s. A three-electrode system was used: a platinum n working electrode, Ag/Ag+ in AgNO (10 mM, MeCN) non-aqueous reference 3 r u electrode, and a platinum wire counter electrode. o 2.9. In vitro growJth inhibition assay. Approximately 5000 cells (A2780, MCF-7, SUNE1, OE19 or MRC-5) were seeded per well in 96-well plates. The cells were pre-incubated in drug-free media for 48 h before adding different concentrations of complexes in medium. After 24 h of incubated drug exposure, the supernatants were removed and cells were washed with PBS. The cells were allowed to recover in drug-free medium for 72 h incubation and the sulforhodamine B assay was used to determine cell viability. Absorbance measurements of the solubilised dye (on a BioRad iMark microplate reader using a 470 nm filter) 9 Journal Pre-proof allowed the determination of viable treated cells compared to untreated controls. IC 50 values (concentration at which 50% cell death occurs) were determined as duplicates of triplicates for each complex. ICP-OES was used to determine the osmium concentrations of the initial stock solutions of complexes in medium, which were used for determination of IC values. 50 f 2.10. Cell uptake. o o Approximately 4×106 A2780 cells were seeded into P100 Petri dishes. After 24 h r p pre-incubation in drug-free medium, complexes were added in equipotent concentrations - of IC /3 and the cells incubated for 24 eh. The cells were then treated with trypsin, 50 r counted, and cell pellets were collected. Control samples were included with untreated P cells. Cell pellets were digested ove rnight in 72% distilled HNO 3 at 80 °C in Wheaton V- l a vials. The resulting solutions were diluted in a stabilisation solution (10 mM thiourea n and 0.1 g/L ascorbic acid), diluting them to 3.6% HNO . Ascorbic acid reduces volatile 3 r u OsO generated during the digestion process and thiourea is an osmium binding agent 4 o [31]. Osmium concentrations were determined by ICP-MS. Measurements were carried J out in triplicate for each complex and cellular accumulation was calculated in units of ng Os /106 cells. 2.11. ROS assay Flow cytometry analysis of ROS/superoxide generation in A2780 cells caused by exposure to osmium complexes, was carried out using the Total ROS/Superoxide detection kit (Enzo-Life sciences) according to the supplier’s instructions. Briefly, 1.5×106 cells per well were seeded in a 6-well plate. Cells were pre-incubated in drug-free 10 Journal Pre-proof media for 24 h, then complexes were added to triplicates at 1× and 2× IC . After 24 h of 50 incubated drug exposure the supernatants were removed and the cells were washed with PBS. Cells were harvested using trypsin and collected after centrifugation. Positive controls (cells treated with pyocyanin) and negative controls (untreated cells) were included in the study. Staining was achieved by re-suspending the cell pellets in buffer containing the orange/green fluorescent reagents. Cells were analysed on a Becton Dickinson FACScan Flow Cytometer using FL1 channel (excitation/emission: 490/525 f o nm) for the oxidative stress and FL2 channel (excitation/emission: 550/620 nm) for o superoxide detection. Compensation adjustments were carried out using pyocyanin- r p treated cells singly-stained with either fluorescent agent. Data were processed using - Flowjo software. e r P 2.12. Cell cycle analysis. l Briefly, 1.5×106 A2780 cells per well were seeded in a 6-well plate. Cells were pre- a incubated in drug-free median for 24 h, then complexes were added to triplicates at 1× or r 2× IC concentrations. Control samples with untreated cells were included. After 24 h of 50 u incubated drug expoosure the supernatants were removed and the cells were washed with J PBS, harvested using trypsin, then fixed with 70% ethanol and stored at -20 °C. DNA staining was achieved by re-suspending the cell pellets in PBS containing propidium iodide (PI) and RNase. Cell pellets were re-suspended in fresh PBS before being analysed in a Becton Dickinson FACScan flow cytometer, using excitation of DNA-bound PI at 536 nm with emission at 617 nm. Data were processed using Flowjo software. 2.13. Apoptosis. Flow cytometry analyses of apoptotic populations of A2780 cells caused by exposure to complexes were carried out using the Biovision Annexin V-FITC 11 Journal Pre-proof (fluorescein isothiocyanate) apoptosis detection kit according the supplier’s instructions. Briefly, 1.5×106 cells per well were seeded in a 6-well plate. Cells were pre-incubated in drug-free media for 24 h, then complexes were added to triplicates at 1× and 2× IC 50 concentrations. Control samples with untreated cells were included. After 24 h of incubated drug exposure, supernatants were removed and cells were washed with PBS. Cells were harvested using trypsin and collected after centrifugation. The cells were stained with Annexin V-FITC and PI and analysed using a Becton Dickinson FACScan f o flow cytometer, running Cell Quest software (20 000 events were collected from each o sample). r p - e 3. Results and discussion r P 3.1. Synthesis l a AZPY ligands functionalised with alkoxy or glycolic side-chains (RO-AZPY) were n prepared from a ligand we reported previously, 2-(phenylazo)-5-fluoropyridine (5-F- r u AZPY, Scheme S1) [4]. The synthesis involves an addition-elimination nucleophilic o substitution mechanism, where the alcohol/glycol reagent (ROH) behaves as a solvent J matrix, and nucleophile, RO-, is generated in situ under basic conditions, attacking the electropositive fluorinated carbon at the 5-position (Scheme S2). A total of 13 ligands were synthesised, L1-L13, utilising a variety of alcohols and glycols with different chain lengths. Table S1 summarises ligands L1-L13 and the reagents/solvents (ROH) used to generate them. Twenty-four novel Os(II) arene AZPY complexes were synthesised, where the arene, monodentate ligand (X), counter anion (Y), and ligand substituents (R) were varied (Fig 1). Complex 5, [Os(η6-p-cym)(5-EtO-AZPY)I]PF , was synthesised 6 12 Journal Pre-proof previously and its X-ray crystal structure determined [32]. Their syntheses involve stirring osmium dimer, [Os(η6-arene)X ] (where X = Cl, Br or I and arene = p-cym or 2 2 bip), in ethanolic solution with 2.1 mol. equiv. of ligand (L1-L13). The generated cationic complex is then paired with a counter-anion; PF -, CF SO -, or IO -, by adding 6 3 3 3 either NH PF , NH CF SO or KIO , respectively (Scheme S3). All complexes were 4 6 4 3 3 3 characterised by 1H NMR, electrospray ionisation mass spectrometry (ESI-MS) and CHN analysis, and where satisfactory CHN analyses were not obtained, complexes were f o additionally characterised by 13C NMR. Purity was determined by reverse-phase o analytical RP-HPLC and only complexes ≥96% pure were used for further studies. r p Impure complexes were purified by recrystallisation or automated reverse-phase column - chromatography. ESI-MS analysis of comeplexes acquired in positive ion mode revealed r a m/z peak corresponding to the cationic species without counter-anion, [M-Y]. In P negative mode, a m/z peak corresponding to the anion was observed (145.0 and 149.0 l a for PF - and CF SO -, respectively, see Fig. S1). 6 3 3 n r u o J 13 Journal Pre-proof Arene p-cymene biphenyl Complex Arene X Y R Complex Arefne X Y R* n o 1 p-cym I PF Me 13 p-cym Cl PF H 1 6 6 2 p-cym I CF SO Me 14 op-cym Br PF H 1 3 3 6 3 p-cym Cl PF Et 15 p-cym I PF H 1 6 r 6 4 p-cym Br PF 6 Et p16 bip I PF 6 H 1 5* p-cym I PF Et 17 p-cym I PF H 2 6 6 - 6 p-cym I CF SO Et 18 p-cym I PF Me 2 3 3 e 6 7 p-cym I IO Et 19 p-cym I PF Et 2 3 6 8 bip I PF Et r20 bip I CF SO Et 2 6 3 3 P 9 bip I CF SO Et 21 p-cym I PF H 3 3 3 6 10 p-cym I PF nPr 22 p-cym I PF Me 3 6 6 11 p-cym I PF iPr 23 p-cym I PF Et 3 6 l 6 12 p-cym I PFanBu 24 p-cym I PF H 4 6 6 n *Complex previously synthesised and reported. *Cormplex previously synthesised and reported [32]. u Fig. 1. Osmium(II) arene phenylazopyridine complexes synthesised, characterised and o studied in this work. J 3.2. X-ray crystal structures The molecular structures of compounds 1 and 8 were determined by X-ray crystallography (Fig. 2 and Table S2). The structures adopt the familiar pseudo- octahedral three-legged piano-stool geometry, where Os(II) is π-bonded to the arene ligand, and coordinated to a monodentate iodido ligand and a bidentate RO-AZPY ligand, which constitute the three legs of the piano-stool. The complexes crystallised as 14 Journal Pre-proof racemates owing to their chiral metal centres, and contain PF - counter-anions in their x- 6 ray crystal structures. (1-PF ) (8-PF ) 6 6 Os1 Os1 O3 I1 I1 N1 N8 O15 fN1 N8 o N7 N7 o r p Fig. 2. ORTEP diagrams for the X-ray crystal structures of complexes 1 and 8. - Ellipsoids are shown at the 50% probabielity level and all hydrogens and counter ions r have been omitted for clarity. P l a 3.3. Hydrolysis studies n r The extent of hydrolysis of chlorido, bromido, and iodido analogues, [Os(η6-p- u cym)(5-EtO-AZPY)Col]PF (3), [Os(η6-p-cym)(5-EtO-AZPY)Br]PF (4) and [Os(η6-p- 6 6 J cym)(5-EtO-AZPY)I]PF (5), was determined by RP-HPLC for 100 µM solutions after 0, 6 2, 12 and 24 h at 37 °C. They were solubilised in phosphate buffer (95 mol equiv, pH 7.4, 5% v/v DMSO) and with two concentrations of added NaCl: 23 mM (intracellular conditions) and 103 mM (extracellular conditions). RP-HPLC chromatograms are shown in Fig. S2. Complexes 3 and 4 show significant levels of hydrolysis after 24 h incubation in 23 mM NaCl, with 85% and 84% of the hydroxido species, [Os(η6-p- cym)(5-EtO-AZPY)OH]+ (3OH), present, respectively. The previously reported aqua- species has a low pK (4.55), indicating that the more stable Os-OH species a 15 Journal Pre-proof predominates under physiological conditions over the more labile Os-OH species [32]. 2 Incubation in 103 mM NaCl partially suppresses hydrolysis of 3 and 4, with 61% and 58% hydrolysis occurring after 24 h, respectively. For bromido complex 4, formation of the chlorido species is also observed in the presence of NaCl, occurring at 8% and 27%, respectively for 23 mM and 103 mM NaCl. In contrast, the iodido complex 5 is significantly more inert towards hydrolysis, showing only 7% and 4% hydrolysis after 24 h incubation in the presence of 23 mM and 103 mM NaCl, respectively, and only 1% f o and 2% formation of the chlorido species, respectively. Complexes 3 and 4 exhibit low o stability in comparison to previously reported chlorido AZPY complex, [Os(η6-p- r p cym)(AZPY-NMe )Cl]PF , which remained inert towards hydrolysis even when heated 2 6 - e with 1 mol. equiv. AgNO [5]. Interestingly, this complex has sub-micromolar activity 3 r against A2780 cells, which may in paPrt be a result of its inherent stability. 3.4. Antiproliferative activity l a n The IC values for complexes 2-7 and 9-11, and 13-24 towards A2780 human 50 r ovarian cancer cells wuere determined and are summarised in Table 1 and Fig. S3. o Complexes 1 and 12 were too insoluble in aqueous media to conduct biological assays. J However, replacing the PF - counter-anion of 1 with CF SO - led to a 45-fold increase in 6 3 3 solubility (from 9 to 405 µM maximum solubility in 100 mM NaCl at 20 °C, Fig. S3), (aq) enabling biological evaluation of 2. Moreover, changing the counter-anion from PF - to 6 CF SO - has no major impact on antiproliferative activity, as shown by the similar 3 3 activities of 5 and 6 (A2780 IC = 0.92 and 0.9 µM, respectively), which exhibit 50 maximum aqueous solubilities of 52 and 185 µM, respectively. Further trends in aqueous solubility and antiproliferative activity are summarised in the SI. 16 Journal Pre-proof Seven complexes (2, 5, 6, 9, 10, 11 and 19) exhibited sub-micromolar activity against A2780 cells and five were selected for further evaluation against MCF-7 breast cancer, SUNE1 nasopharyngeal cancer, OE19 oesophageal cancer, and MRC-5 lung fibroblast cells (Table 2). Their selectivity factors (SF) towards cancer cells over normal cells were calculated as a ratio of IC (MRC-5)/IC (A2780). With the exception of 19, 50 50 complexes screened against SUNE1 cells show highly potent sub-micromolar activity. Complexes 5, 9 and 19 exhibit exceptionally potent activity against OE19 cells with 9 f o exhibiting low nanomolar potency (96 nM), whilst 11 exhibits the highest activity against o A2780 cells, comparable to other complexes reported in its class [4, 5]. In contrast, these r p complexes are generally less active against MCF-7 cells, with the highest activity - e observed for 5 (1.2 µM). Overall, complex 9 displays the broadest anticancer profile r amongst all the tested cell lines and the highest SF; comparable with that of cisplatin P [14]. l a n r u o J Table 1 IC concentrations (µM) for complexes 2-7 and 9-11, and 13-24 against A2780 human 50 ovarian cancer cells. A2780 Ovarian cancer cells 17 Journal Pre-proof Complex IC (µM) 50 (1) [Os(η6-p-cym)(5-MeO-AZPY)I]PF n.d. 6 (2) [Os(η6-p-cym)(5-MeO-AZPY)I]CF SO 0.5(±0.1) 3 3 (3) [Os(η6-p-cym)(5-EtO-AZPY)Cl]PF 15.1(±0.5) 6 (4) [Os(η6-p-cym)(5-EtO-AZPY)Br]PF 6 14.1(±0.9) (5) [Os(η6-p-cym)(5-EtO-AZPY)I]PF 0.92(±0.02) 6 (6) [Os(η6-p-cym)(5-EtO-AZPY)I]CF SO 0.9(±0.2) 3 3 (7) [Os(η6-p-cym)(5-EtO-AZPY)I]IO 1.6(±0.3) 3 (8) [Os(η6-bip)(5-EtO-AZPY)I]PF n.d. 6 (9) [Os(η6-bip)(5-EtO-AZPY)I]CF SO 0.51(±0.02) 3 3 f (10) [Os(η6-p-cym)(5-nPrO-AZPY)I]PF 0.33(±0.02) 6 o (11) [Os(η6-p-cym)(5-iPrO-AZPY)I]PF 0.30(±0.09) 6 o (12) [Os(η6-p-cym)(5-nBuO-AZPY)I]PF n.d. 6 r (13) [Os(η6-p-cym)(5-HOCH CH O-AZPY)Cl]PF >100 2 2 6 p (14) [Os(η6-p-cym)(5-HOCH CH O-AZPY)Br]PF >100 2 2 6 - (15) [Os(η6-p-cym)(5-HOCH CH O-AZPY)I]PF 20.7(±0.1) 2 2 e6 (16) [Os(η6-bip)(5-HOCH CH O-AZPY)I]PF 2.37(±0.09) 2 2 6 r (17) [Os(η6-p-cym)(5-HO{CH CH O} -AZPY)I]PF 2.04(±0.06) 2 P2 2 6 (18) [Os(η6-p-cym)(5-MeO{CH CH O} -AZPY)I]PF 1.80(±0.09) 2 2 2 6 (19) [Os(η6-p-cym)(5-EtO{CH CH O} -AZPY)I]PF 0.68(±0.03) l2 2 2 6 (20) [Os(η6-bip)(5-EtO{CaH CH O} -AZPY)I]CF SO 1.9(±0.2) 2 2 2 3 3 (21) [Os(η6-p-cym)(5-nHO{CH 2 CH 2 O} 3 -AZPY)I]PF 6 12(±1) (22) [Os(η6-p-cym)(5-MeO{CH CH O} -AZPY)I]PF 2.1(±0.2) r 2 2 3 6 (23) [Os(η6-p-cyum)(5-EtO{CH CH O} -AZPY)I]PF 2.6(±0.4) 2 2 3 6 (24) [Os(η6-p-cym)(5-HO{CH CH O} -AZPY)I]PF 7.6(±0.6) o 2 2 4 6 J Table 2 18 Journal Pre-proof IC concentrations (µM) for complexes 5, 9, 11, 18, and 19, against various cancer cell 50 lines and normal MRC-5 cells. Selectivity factors (SF) are calculated as the ratio IC (MRC-5)/IC (A2780). 50 50 Cell line IC / µM 50 Complex MCF-7 SUNE1 OE19 MRC-5 SF 5 1.2(±0.2) 0.86(±0.06) 0.20(±0.03) 2.2(±0.2) 2.4 9 15(±2) 0.31(±0.02) 0.096(±0.004) 6.1(±0.1) 12.0 f 11 7.1(±0.2) n.d. n.d. o1.99(±0.07) 6.6 18 10(±1) n.d. n.d. 7.4(±0.8) 4.1 o 19 12.9(±0.8) 1.7(±0.1) 0.42(±0.01) 3.6(±0.3) 5.3 r p - 3.5. Capacity factor and correlation with anteiproliferative activity r The capacity factor (K) of a coPmpound is related to its RP-HPLC retention time and is independent of column ge ometry and mobile phase flow rate. It serves as a l a measure of affinity towards a hydrophobic stationary phase versus an aqueous mobile n phase, hence providing a measure of relative lipophilicity [33, 34]. The higher the K- r u value of a compound, the greater its lipophilicity. The K-values of complexes 2-6 and 9- o 24 were determined by RP-HPLC using a C18 column and isocractic mobile phase of J MeCN:H O (1:1, v/v, 50 mM NaCl), and are shown in Fig. S3. Trends in the capacity 2 factors of these complexes are discussed in the SI, and correlations between antiproliferative activity towards A2780 cells and capacity factor are shown in Fig. 3. Inverse correlations are observed; i.e. when relative lipophilicity increases, the IC value 50 generally decreases (increased activity), which is overall likely due to the improved cellular accumulation of complexes exhibiting higher lipophilicity. This was investigated by examining correlations where specific substituents are varied. 19 Journal Pre-proof (A) (B) 1.0 16 (3) (5) (4) 0.9 ts 14 X = Cl ts R = Et n ia g a n o ita r tn e c n o C C I 05 n o ita r tn e c n o C C I )M µ ( s lle c 0 8 7 2 A0 5 1 1 2 4 6 8 0 2 X = B C r orrelati ( o 5) n = - X 0 . = 9 6 I n ia g a n o ita r tn e c n o C C I 05 n o ita r tn e c n o C C I )M µ ( s lle c 0 8 7 2 A 0 5 0 0 0 0 0 0 . . . . . . 3 4 5 6 7 8 (2 C ) or R r e = l M at e ion = -0.55 (11) R = iPr (10 R ) = nPr 0 0.2 1.5 2.0 2.5 3.0 3.5 4.0 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 Capacity Factor (K) Capacity Factor (K) f o (C) (D) o (15) 2.2 ts 20 n = 1 tsr2.0 (17) n ia g a n o ita r tn e c n o C C 05 n o ita r tn e c n o C C I 0 5 )M µ ( s lle c 0 8 7 2 A 1 1 5 0 5 Correla (1 ti 7 o ) ( ( n 2 2 4 1 = ) ) -0 n n .6 = = 9 4 3 P r e - p n ia g a n o ita r tn e c n o C C 05 n o ita r tn e c n o C C I )M µ ( s lle c 0 8 7 2 A0 5 0 1 1 1 1 1 . . . . . . 8 0 2 4 6 8 R* = H (18) R* = Me Cor ( r 1 e 9 l ) ation = -0.90 I I n = 2 0.6 R* = Et 0 1.00 1.02 1.04 1.06 1.08 1.10 1.12 1.14 1.0 1.5 2.0 2.5 3.0 3.5 4.0 l Capacity Factor (K)a Capacity Factor (K) n Fig. 3. Correlations between antiproliferative activity (IC towards A2780 cells, µM) r 50 u and capacity factor (K), where different substituents on the complexes are varied; (A) o monodentate halido ligand, X, (B) alkoxyl side-chain substituent, R, (C) length of J glycolic side-chain substituent, n, and (D) terminal group of glycolic side-chains, R*. When the monodentate halido ligand (X) is varied, a strong inverse correlation between activity and relative lipophilicity is observed (-0.96, Fig. 3a). Complexes with monodentate iodido ligands are considerably more active than their chlorido and bromido counterparts, with the observed trend: I >> Br ≥ Cl. The same trend was observed previously for other Os(II) AZPY complexes [4]. The improved stability of iodido complexes may account for their increased activity as they are less likely to be 20 Journal Pre-proof deactivated by extracellular side reactions. Using 131I radio-labelling, we showed previously that iodido complex 5 is stable in extracellular conditions and activated via monodentate ligand loss in MCF-7 breast cancer cells [32]. Likewise, the same trend in K is observed: I >Br > Cl, where iodido complexes exhibit the highest lipophilicity, owing to their lower polarisation and more covalent character of the Os-X bond. For complexes containing alkoxy substituents (R) on their AZPY ligand, a weaker correlation between activity and relative lipophilicity is observed (-0.55, Fig. 3b). f o Increasing the length of this hydrophobic substituent increases K (hence relative o lipophilicity) in the order: nPr>iPr>Et>Me. However, activity follows a different trend: r p iPr>nPr>Me>Et, that cannot be explained solely by capacity factors. - Moderate correlation was observed ebetween activity and relative lipophilicity for r complexes bearing glycolic side-chains, where the length of the glycol (n) was varied and P where R*=H (-0.69, Fig. 3c). Varying n does not appear to have a significant effect on l a the relative lipophilicity of the complex, possibly because each unit increase in the glycol n chain introduces an additional ethereal oxygen capable of H-bonding with water. r u Likewise, there is no clear trend in activity, which follows the pattern: n=2 > 4> 3 > 1. o In contrast, when the terminal group on the glycol side-chain (R*) is varied, a strong J correlation between activity and relative lipophilicity is observed (-0.90, Fig. 3d). As expected, the trend in K follows: Et > Me > H, where complexes with longer hydrophobic groups are more lipophilic. This trend is mirrored by activity, likely due to elevated cell uptake resulting from higher lipophilicity. 3.6. Octanol/water partition coefficients Octanol/water partition coefficients (P ) of complexes were measured using a o/w modified version of the shake-flask method [6, 29, 30, 34], where OSW contained 300 21 Journal Pre-proof mM of NaX (X = Cl, Br or I) to suppress the hydrolysis of chlorido, bromido and iodido complexes, respectively, and complexes were prepared as saturated solutions in OSW prior to the addition of WSO. Without the addition of NaX, potential aqua-adduct formation would complicate the interpretation of data. We generally observed a decrease in complex solubility upon addition of NaX to aqueous media (salting-out effect) [35]. To gain a full understanding on the salt effect on Log P for our complexes, further o/w experiments at fixed concentrations would be required [36]. In general, increasing the f o ionic strength of OSW tends to increase Log P values for lipophilic compounds [37]. o/w o Log P was determined for selected complexes (4, 13, 15, 16, 18, 19 and 21), as it was o/w r p not possible to make measurements on complexes with very poor aqueous solubility due - to the limit of Os detection via ICP-MS afteer partitioning. r P An alternative RP-HPLC method for determining Log P was adopted; using the HPLC K-values, a calibration curve of Log P vs. Log K was plotted for complexes 4, 13, 15, lo/w a 16, 18, 19 and 21 [38] (Fig. S4), and Log P values for all complexes were derived n HPLC (Table 3). Due to relativerly weak correlation (+0.743), the Log P values have error HPLC u margins, but are suitable enough for these purposes of comparison. The Log K values of o 5, 6 and 10-12 faJll outside the range of the calibration plot and their Log P values HPLC were estimated by extrapolating the curve. All determined Log P values fall well HPLC within the Ro5 range (-0.4 to +5.6). Following an approach outlined by Sahu et al, Log P was plotted on the x-axis as a reference indicator against the expression Log (K/P ) o/w o/w to determine whether factors other than lipophilicity contribute towards partitioning, i.e. chemical interactions with solvents, such as hydrolysis [38]. The plot shows strong correlation (+0.935), indicating that partitioning of the complexes is largely determined by lipophilicity (Fig. S5). 22 Journal Pre-proof Table 3 Log K values for complexes 2-6 and 9-24, their derived Log P values, and Log P HPLC o/w values for complexes 4, 13, 15, 16, 18, 19 and 21 measured using a modified shake-flask method. Complex Log (K) Log (P ) Log (P ) o/w HPLC 2 0.42(±0.02) n.d. 2.1(±0.5) 3 0.29(±0.01) n.d. 1.8(±0.4) f 4 0.42(±0.01) 1.59(±0.03) 2.1(±0.5) o 5* 0.58(±0.01) n.d. 2.5(±0.6) o 6* 0.60(±0.01) n.d. 2.6(±0.6) r 9 0.56(±0.01) n.d. 2.5(±0.6) p 10* 0.779(±0.009) n.d. 3.0(±0.7) - 11* 0.733(±0.008) n.d. 2.9(±0.7) e 12* 0.972(±0.007) n.d. 3.5(±0.8) r 13 -0.25(±0.02) 0.11(±0.02) 0.6(±0.4) P 14 -0.13(±0.02) n.d. 0.8(±0.3) 15 0.02(±0.0 2) 1.87(±0.05) 1.2(±0.2) l 16 0.00(a±0.03) 1.10(±0.04) 1.1(±0.2) 17 0.038(±0.009) n.d. 1.2(±0.2) n 18 0.420(±0.005) 2.29(±0.04) 2.1(±0.5) r 19 0.574(±0.007) 2.6(±0.2) 2.5(±0.6) u 20 0.548(±0.007) n.d. 2.4(±0.5) o 21 0.044(±0.007) 1.41(±0.04) 1.3(±0.2) J 22 0.402(±0.006) n.d. 2.1(±0.5) 23 0.555(±0.006) n.d. 2.5(±0.6) 24 0.045(±0.007) n.d. 1.3(±0.2) * Log (K) value outside the calibration range 3.7. Cellular accumulation The cellular accumulation of Os in A2780 cells was determined for complexes 2-6, 9 and 13-15, by ICP-MS (Fig. 4). Trends between 3-5 and 13-15 show that iodido complexes promote greater accumulation of Os than their chlorido and bromido analogues, and the same 23 Journal Pre-proof trend is observed for anticancer activity: I > Br > Cl. The enhanced activity and accumulation of iodido complexes inside cells is likely to be due in part to increased lipophilicity, as well as their improved stability. The trend between 5 and 6 shows that anions have little effect on cellular accumulation, activity and lipophilicity, due to their lack of role once complexes are solvated in medium. Switching the arene from p-cym to bip decreases cellular accumulation and lipophilicity, as shown by the trend between 6 and 9. Interestingly, this change also led to an improvement in anticancer activity, which may be due to factors other than lipophilicity. f o When the R-substituent is varied, the trend in cell uptake follows the same order as o lipophilicity: Et > Me > HOCH CH , as observed for complexes 2, 5, 6 and 15. Overall, 2 2 r p reasonably strong correlation is observed between cellular accumulation and K-value for - complexes 2-6, 9 and 13-15 (+0.78), thus indeicating a linear trend between the lipophilicity of r this series of complexes and their ability to accumulate in cancer cells (Fig. S6). P l X = Cl, Y = PF, Ar= p-cym 6 a X = Br, Y = PF, Ar= p-cym 6 X = I, Y = PF, Ar= p-cym n 6 X = I, Y = CFSO, Ar= p-cym 3 3 r X = I, Y = CFSO, Ar= bip 3 3 u R = oMe R = Et J 35 s 30 lle ) c 0 s lle 25 )1 8 7 2 A n i m u im s O 6 0 1 / s O g c n ( 1 1 2 0 5 0 5 8( 2) (0 )8 .1 . 1 0 (9 1 10( 2) ( ) 3 9 9 0 . . 3 0 ( ) 0 8 0 0 . . 3 0 12( 2) 0 ((72)) ((83)) ((94)) ((150)) ((161)) ((193)) -- (14) -- ((1173)) ((1184)) ((1195)) Fig. 4. Trends in cellular Os accumulation for complexes 2-6, 9 and 13-15 in A2780 cells after treatment with IC /3 concentration of complex for 24 h. 50 24 Journal Pre-proof 3.8. Cyclic Voltammetry The electrochemistry of complexes 2, 3, 5, 9-11, 15, 18, 19, and ligand L2 in MeCN was investigated by cyclic voltammetry. Complexes 2, 5, 9, 11, and 19 were scanned between +2.0 to -2.0 V and notable redox activity was observed between +0.5 to -1.5 V (Fig. S7). All complexes exhibited two reductions between 0.0 to -2.0 V, assignable to irreversible reductions of the azo-bond [39] (Figs. 5, S8 and Scheme S4). f However, upon scanning from 0.0 to -1.0 V, chlorido complex 3 exhibited a reversible o first reduction potential, in contrast to its iodiodo anologues, all of which were r irreversible. Interestingly, the first reduction of 3 was only reversible when the azo-bond p did not undergo a second e- transfer, sugge-sting irreversible chemical changes to the e structure occur once the azo-bond is fully reduced. Furthermore, second segments in the r P voltammograms for 5 and 9 show oxidation potentials (-0.35 and -0.31 V, respectively) not corresponding to reversal of ltheir first azo-bond reductions. Hence implying that a irreversible chemical changes to iodido complexes can occur after the first azo-bond n reduction. r u In summary, the first and second azo-bond reductions of these complexes range o between -0.75 to J-0.79 and -1.27 to -1.63 V, respectively (Table S3). Ligand L2, which is present in complexes 3, 5, and 9, exhibits only one reversible reduction at -1.73 V. In comparison to previously reported complexes, this series has lower potentials for their first azo-bond reductions. The reductions of parent complex [Os(η6-p-cym)(AZPY- NMe )I]PF (FY026), and its Ru analogue, [Ru(η6-p-cym)(AZPY-NMe )I]PF , were 2 6 2 6 previously determined as -0.64 and -0.40 V, respectively. Both are notably higher and the latter falling within the biologically relevant region [40]. 25 Journal Pre-proof (A) 15.0 15.0 RN121 RN121 -1.63V -0.77V -0.78V 10.0 10.0 5.0 S2 S1 )A 5.0 )A -2 V .0 oltage (V) -1.5 -1.0 -0.69V -0.5 0.0 - 0 5 .0 .0 µ ( tn e r r u C -1 S .0 2 -0.8 -0.6 -0.4 S1 -0.2 0.0 0.0 µ ( tn e r r u C Voltage (V) -0.34V -5.0 -10.0 -0.30V -0.67V -15.0 -10.0 ( R B N30 ) 4 -0.78V 15.0 RN304 -0.79oV f 15.0 -1.50V 10.0 10.0 o 5.0 5.0 S2 S1 )A r S1 )A µ µ 0.0 ( tn p 0.0 ( tn -2.0 -1.5 -1.0 -0.5 0.0 e -1.0 -0.8 -0.6 -0.4 -0.2 0.0 e Voltage (V) r r u-Voltage (V) r r u -5.0 C S2 -5.0 C e -10.0 -10.0 -0.36V r -0.35V P -15.0 -15.0 ( R C N31 ) 8 l 20.0 RN318 20.0 -0.75V a -0.75V 15.0 15.0 -1.25V n 10.0 -2.0 S2 -1.5 o -1.0 u r -0.5 S1 0.0 0 5 . . 0 0 )A µ ( tn e r r u C S2 S1 0 5 1 . . 0 0 0 .0)A µ ( tn e r r u C Voltage (V) -5.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 Voltage (V) J -5.0 -10.0 -0.29V -15.0 -0.31V -10.0 Fig. 5. Cyclic voltammograms for complexes 3 (A), 5 (B), and 9 (C) in MeCN with 0.1 M Bu NPF as supporting electrolyte. Complexes were scanned between 0.0 to -2.0 V 4 6 (left hand side) and 0.0 to -1.0 V (right hand side), and back at 0.1 V/s. The blue and red lines represent the first and second segments, respectively. 26 Journal Pre-proof 3.9. Induction of ROS Complexes 2, 5, 9, 11, and 19 were screened for ROS induction in A2780 human ovarian cancer cells using flow cytometry. The protocol utilises a total ROS/Superoxide detection kit capable of distinguishing O ·- from other ROS. The positive control was 2 treated with pyocyanin and showed the majority of cells expressing high levels of ROS/O ⸱-, and the negative control was untreated cells. Figs. 6 and S9 show 2 f quantitatively the elevation of ROS after incubation with 2, 5, 9, 11, and 19 at 1× and 2× o IC . In every case, significantly elevated levels of oROS and O ⸱- were observed, thus 50 2 r further highlighting that the mechanism of action of this class of Os(II) complex involves p generating oxidative stress in cancer cells [5-, 6, 14, 40]. Interestingly, 5 and 19 both e elevated cellular levels of O ⸱- when their concentrations were doubled, as shown by 2 r P significant increases in the populations of cells in quadrant Q2. Complexes 3, 5 and their intracellularly generated hydroxidlo analogue, [Os(η6-p-cym)(5-EtO-AZPY)OH]+ (3OH), a were also previously reported to generate the highly reactive hydroxyl radical (OH⸱) n upon catabolising H O [3r2]; a comparatively weak ROS that is over-produced in cancer 2 2 u cells [41-43]. Hydroxyl radicals cause damage to cellular lipids, proteins and o compartments, leJading to cell death [44] and the ability of the complexes studied here to produce both O ⸱- and OH⸱ under physiological conditions likely plays a role in 2 inducing oxidative stress in cancer cells. 27 Journal Pre-proof (B) (C) (A) Positive Control Negative Control Q1: Q2: Cells with Cells with 0.0% 99.5% 0.0% 0.0% elevated elevated O - ROS + O - 2 2 Q4: Q3: Healthy Cells with Cells elevated ROS 0.0% 0.4% 100% 0.0% 1X IC 2X IC 50 50 (D) 0.9% 15.0% 0.5% 63.7% f o o 3.4% r80.7% 0.0% 35.8% p 0.5% -34.0% 1.8% 53.3% (E) e r P 0.1% 65.4% 0.1% 44.9% l a n Fig. 6. (A) Diagram showing the four quadrants in the cell population plots; Q1 = cells r with elevated levels ouf superoxide, Q2 = cells with elevated levels of ROS and o superoxide, Q3 = cells with elevated levels of ROS, and Q4 = healthy cells. The J populations of A2780 cells after 24 h incubation at 37 °C with; (B) pyocyanin (positive control), (C) no drug added (negative control), (D) various concentrations of 5 (1× and 2× IC ), and (E) various concentrations of 19. 50 28 Journal Pre-proof 3.10. Cell cycle analysis Changes in the cell cycle phase distribution for A2780 cells after incubation with sub-micromolar active complex 5, and less active 15 were monitored using flow cytometry (Figs. 7 and S10). Complex 15 is analogous to 5, whereby its terminal OH group situated on the ethoxy substituent is deactivating towards anticancer activity but has a strong solubilising effect (Fig. S3a). At 1× IC , complex 5 causes S-phase cell cycle 50 f arrest with an observable decline in the proportion of cells in G1 phase, and at 2× IC , a o 50 rise in the proportion of sub-diploid cells is present.o These are damaged non-viable cells r with compromised membranes and leakage of their DNA content. In contrast, 15 causes p slight elevations in both the G1 and S-phase- populations at 1× IC alongside a small 50 e observable decline in the G2/M-phase population. Doubling the concentration induces a r P large presence of sub-diploid cells, which come primarily from the G1-phase. Such variations in cell cycle effects may arise from differences in AZPY substituents and their l a subsequent electronic properties. n r u o J 29