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

Article abstract of DOI:10.1016/j.jinorgbio.2020.111154

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-{CH₂CH₂O}_nR*, n = 1–4, R* = H, Me, or Et) substituent on the pyridyl ring para to the azo-bond, X is a monodentate halido ligand (Cl, Br or I), and Y is a counter-anion (PF₆^?, CF₃SO₃^? or IO₃^?). X-ray crystal structures of two complexes confirmed their ‘half-sandwich’ structures. Aqueous solubility depended on X, the AZPY substituents, arene, and Y. Iodido complexes are highly stable in water (X = I ? Br > Cl), and exhibit the highest antiproliferative activity against A2780 (ovarian), MCF-7 (breast), SUNE1 (nasopharyngeal), and OE19 (oesophageal) cancer cells, some attaining nanomolar potency and good cancer-cell selectivity. Their activity and distinctive mechanism of action is discussed in relation to hydrophobicity (RP-HPLC capacity factor and Log P_o/w), cellular accumulation, electrochemical reduction (activation of azo bond), cell cycle analysis, apoptosis and induction of reactive oxygen species (ROS). Two complexes show ca. 4× higher activity than cisplatin in the National Cancer Institute (NCI) 60-cell line five-dose screen. The COMPARE algorithm of their datasets reveals a strong correlation with one another, as well as anticancer agents olivomycin, phyllanthoside, bouvardin and gamitrinib, but only a weak correlation with cisplatin, indicative of a different mechanism of action.

Full text of DOI:10.1016/j.jinorgbio.2020.111154

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
JIB 111154
Reference:
To appear in:
Journal of Inorganic Biochemistry
Received date:
Revised date:
Accepted date:
9 April 2020
9 June 2020
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, errors may be discovered which could affect the content, and all legal disclaimers
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*.
Department of Chemistry, University of Warwick, Gibbet Hill Road, Coventry, CV4 7AL, U.K.
GRAPHIC:
Cell
Cycle
IC50
Solubility
Cell
Uptake
Cyclic
Voltammetry
Lipophilicity
ROS
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-{CH2CH2O}nR*, n = 1-4, R* = H, Me, or Et) substituent on
the pyridyl ring para to the azo-bond, X is a monodentate halido ligand (Cl, Br or I), and
-
Y is a counter-anion (PF6 , CF3SO3- or IO3-). X-ray crystal structures of two complexes
confirmed their ‘half-sandwich’ structures. Aqueous solubility depended on X, the
AZPY substituents, arene, and Y. Iodido complexes are highly stable in water (X = I
>>> Br > Cl), and exhibit the highest antiproliferative activity against A2780 (ovarian),
MCF-7 (breast), SUNE1 (nasopharyngeal), and OE19 (oesophageal) cancer cells, some
attaining nanomolar potency and good cancer-cell selectivity. Their activity and
distinctive mechanism of action is discussed in relation to hydrophobicity (RP-HPLC
capacity factor and Log Po/w), cellular accumulation, electrochemical reduction
(activation of azo bond), cell cycle analysis, apoptosis and induction of reactive oxygen
species (ROS). Two complexes show 4× higher activity than cisplatin in the National
Cancer Institute (NCI) 60-cell line five-dose screen. The COMPARE algorithm of their
datasets reveals a strong correlation with one another, as well as anticancer agents
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 to reduction of the azo-bond,
which contains a low lying π*-orbital capable of accepting two electrons. Previously-
reported piano-stool Os(II) arene AZPY complexes exhibit potent anti-proliferative
activity against a variety of cancer cell lines [8]. Their mechanism of action differs greatly
from earlier Os(II) arene complexes with σ-donor bidentate ligands, which rely primarily
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,
such as resistance towards hydrolysis and a mechanism of action involving increased
intracellular reactive oxygen species (ROS) levels [4-6, 14].
Current structure-activity relationships (SARs) reported for Os(II) arene AZPY
complexes include 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 Po/w is -0.4 to
+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
bioavailability and delivery [23]. There appear to be no reports of varying the chain
length and tail groups of glycolic substituents to modulate metallodrug properties,
however, they have been incorporated into homobinuclear Au(I) and Ru(II) anticancer
complexes to create variable length linkages between metal centres [24].
Here we explore the effects of changes in the charge polarisation of Os(II) arene
AZPY complexes on activity towards cancer cells, using a variety of alkoxy and glycolic
side-chains on AZPY ligands. A series of 24 new Os(II) arene AZPY complexes have
been synthesised and the SARs explored for activity against A2780 human ovarian
cancer cells, solubility in aqueous media, RP-HPLC capacity factors (lipophilicity), and
cellular uptake. We also explore variation of the counter-anion as a means of improving
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
OsCl3·3H2O was purchased from either Sigma Aldrich (UK) or Heraeus (South
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-
AZPY ligands were also purchased from Sigma Aldrich, with the exception of ethylene
glycol and diethylene glycol, which were obtained from Alfa Aesar (UK). All other
organic solvents and reagents for synthesis and analysis were purchased from
commercial suppliers and were used as received. Dimers; [Os(η6-p-cym)X2]2, where X =
Cl, Br or I, were prepared according to reported procedures [25-27] as was 2-
(phenylazo)-5-fluoropyridine [4]. Dulbecco’s Modified Eagle Media, Roswell Park
Memorial Institute-1640 cell culture medium, penicillin/streptomycin mixture, foetal
bovine serum, L-gutamine, phosphate buffered saline solution (PBS), trypsin, and
trypsin/ethylenediaminetetraacetic acid were all purchased from PAA Laboratories
GmbH. Cell culture media (500 mL) were supplemented with foetal bovine serum (50
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 CH2Cl2 (50 mL) and washed with water (3 x 50 mL).
5

Journal Pre-proof
The CH2Cl2 extract was dried over MgSO4, filtered, and the solvent was removed under
reduced pressure. Some reaction mixtures required vacuum distillation to remove excess
ROH. Where necessary, purification was performed by re-crystallisation or SiO2 column
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
[Os(η6-arene)X2]2 (where X = Cl, Br or I, and arene = p-cym or bip) was dissolved
in EtOH (10 mL), and a solution of RO-AZPY (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
NH4PF6, NH4CF3SO3 or KIO3 (10 mol. equiv.) was added. Complexes were isolated
using either method 1 or 2. Method 1 involves re-crystallisation. Some complexes proved
difficult to re-crystallise and method 2 was employed. In some cases, further purification
via automated reverse-phase chromatography was required. Method 1: The ethanolic
mixture was concentrated under reduced pressure to ~3 mL and placed in the freezer (-
20 °C) overnight. A dark crystalline precipitate formed which was collected via vacuum
filtration, washed with ice-cold EtOH (2 x 1 mL), then Et2O (2 x 5 mL), and dried
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 Et2O (3 x 5 mL). The residue
was redissolved in CH2Cl2 (5 mL), transferred to a pre-weighed vial, and dried overnight
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
Gemini four-circle system with a Ruby CCD area detector. All structures were refined by
full-matrix least squares against F2 using SHELXL 97 and were solved by direct methods
using SHELXS(TREF) with additional light atoms found by Fourier methods [28].
Hydrogen atoms were added at calculated positions and refined using a riding model.
Anisotropic displacement parameters were used for all non-H atoms; H-atoms were
given an isotropic displacement parameter equal to 1.2 (or 1.5 for methyl and NH H-
atoms) times the equivalent isotropic displacement parameter of the atom to which they
are attached. The data were processed by the modelling program Mercury 1.4.1. X-ray
crystallographic data for complexes 1 and 8 have been deposited in the Cambridge
Crystallographic Data Centre under the accession numbers CCDC1990713 and
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 H2O:MeCN (1:1, v/v, 50
mM NaCl) and the temperature of the column was maintained constant at 25 °C.
Samples (500 µM) were prepared in H2O:MeCN (1:1, v/v) and were analysed in
triplicates in three separate experiments (50 µL injection volume). Their capacity factors
were calculated using the following equations, where tR is the retention time of the
retained complex, t0 is the retention time of unretained compound (uracil), RF is the
retention factor, and K is the capacity factor.
RF = (tR - t0)
(1)
(2)
K = RF/t0
2.7. Partition coefficients
Octanol/water partition coefficients (Log Po/w) of complexes were determined using
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 Po/w = Log ([Os]octanol/[Os]water
2.8. Electrochemistry
)
(3)
Cyclic voltammetry analyses were performed using a CH Instruments
Electrochemical Analyzer (CHI420C) and CH Instruments electrochemistry software.
n
Compounds were dissolved in acetonitrile (1 mg/mL) with Bu4NPF6 (0.1 M) as a
supporting electrolyte. Solutions were degassed under N2 and scanned between -2.0 V
and +2.0 V at scan rates of 0.1 or 0.5 V/s. A three-electrode system was used: a platinum
working electrode, Ag/Ag+ in AgNO3 (10 mM, MeCN) non-aqueous reference
electrode, and a platinum wire counter electrode.
2.9. In vitro growth 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. IC50
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 IC50 values.
2.10. Cell uptake.
Approximately 4×106 A2780 cells were seeded into P100 Petri dishes. After 24 h
pre-incubation in drug-free medium, complexes were added in equipotent concentrations
of IC50/3 and the cells incubated for 24 h. The cells were then treated with trypsin,
counted, and cell pellets were collected. Control samples were included with untreated
cells. Cell pellets were digested overnight in 72% distilled HNO3 at 80 °C in Wheaton V-
vials. The resulting solutions were diluted in a stabilisation solution (10 mM thiourea
and 0.1 g/L ascorbic acid), diluting them to 3.6% HNO3. Ascorbic acid reduces volatile
OsO4 generated during the digestion process and thiourea is an osmium binding agent
[31]. Osmium concentrations were determined by ICP-MS. Measurements were carried
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× IC50. After 24 h of
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
nm) for the oxidative stress and FL2 channel (excitation/emission: 550/620 nm) for
superoxide detection. Compensation adjustments were carried out using pyocyanin-
treated cells singly-stained with either fluorescent agent. Data were processed using
Flowjo software.
2.12. Cell cycle analysis.
Briefly, 1.5×106 A2780 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× or
2× IC50 concentrations. Control samples with untreated cells were included. After 24 h of
incubated drug exposure the supernatants were removed and the cells were washed with
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× IC50
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
flow cytometer, running Cell Quest software (20 000 events were collected from each
sample).
3.
Results and discussion
3.1. Synthesis
AZPY ligands functionalised with alkoxy or glycolic side-chains (RO-AZPY) were
prepared from a ligand we reported previously, 2-(phenylazo)-5-fluoropyridine (5-F-
AZPY, Scheme S1) [4]. The synthesis involves an addition-elimination nucleophilic
substitution mechanism, where the alcohol/glycol reagent (ROH) behaves as a solvent
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]PF6, was synthesised
12

Journal Pre-proof
previously and its X-ray crystal structure determined [32]. Their syntheses involve
stirring osmium dimer, [Os(η6-arene)X2]2 (where X = Cl, Br or I and arene = p-cym or
bip), in ethanolic solution with 2.1 mol. equiv. of ligand (L1-L13). The generated
-
cationic complex is then paired with a counter-anion; PF6 , CF3SO3-, or IO3-, by adding
either NH4PF6, NH4CF3SO3 or KIO3, respectively (Scheme S3). All complexes were
1
characterised by H NMR, electrospray ionisation mass spectrometry (ESI-MS) and
CHN analysis, and where satisfactory CHN analyses were not obtained, complexes were
additionally characterised by 13C NMR. Purity was determined by reverse-phase
analytical RP-HPLC and only complexes ≥96% pure were used for further studies.
Impure complexes were purified by recrystallisation or automated reverse-phase column
chromatography. ESI-MS analysis of complexes acquired in positive ion mode revealed
a m/z peak corresponding to the cationic species without counter-anion, [M-Y]. In
negative mode, a m/z peak corresponding to the anion was observed (145.0 and 149.0
for PF6- and CF3SO3-, respectively, see Fig. S1).
13

Journal Pre-proof
Arene
p-cymene
biphenyl
Complex Arene
X
Y
R
Complex Arene
X
Y
R*
n
1
2
p-cym
p-cym
I
I
PF6
Me
13
14
15
16
17
18
19
20
21
22
23
24
p-cym Cl
p-cym Br
PF6
PF6
PF6
PF6
PF6
H
H
H
H
H
1
1
1
1
2
2
2
2
3
3
3
4
CF3SO3 Me
3
p-cym Cl
p-cym Br
PF6
PF6
Et
Et
p-cym
bip
I
I
I
I
I
I
I
I
I
I
4
5*
6
p-cym
p-cym
p-cym
bip
I
I
I
I
I
I
I
I
PF6
Et
p-cym
p-cym
p-cym
bip
CF3SO3
IO3
Et
PF6
PF6
Me
Et
Et
H
7
Et
8
PF6
Et
CF3SO3
PF6
9
bip
CF3SO3
PF6
Et
p-cym
p-cym
p-cym
p-cym
10
11
12
p-cym
p-cym
p-cym
nPr
iPr
nBu
PF6
Me
Et
H
PF6
PF6
PF6
PF6
*Complex previously synthesised and reported [32].
Fig. 1. Osmium(II) arene phenylazopyridine complexes synthesised, characterised and
studied in this work.
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 PF6 counter-anions in their x-
ray crystal structures.
(1-PF6)
(8-PF6)
Os1
Os1
O3
I1
I1
N8
N1
N7
O15
N8
N7
N1
Fig. 2. ORTEP diagrams for the X-ray crystal structures of complexes 1 and 8.
Ellipsoids are shown at the 50% probability level and all hydrogens and counter ions
have been omitted for clarity.
3.3. Hydrolysis studies
The extent of hydrolysis of chlorido, bromido, and iodido analogues, [Os(η6-p-
cym)(5-EtO-AZPY)Cl]PF6 (3), [Os(η6-p-cym)(5-EtO-AZPY)Br]PF6 (4) and [Os(η6-p-
cym)(5-EtO-AZPY)I]PF6 (5), was determined by RP-HPLC for 100 µM solutions after 0,
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 pKa (4.55), indicating that the more stable Os-OH species
15

Journal Pre-proof
predominates under physiological conditions over the more labile Os-OH2 species [32].
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%
and 2% formation of the chlorido species, respectively. Complexes 3 and 4 exhibit low
stability in comparison to previously reported chlorido AZPY complex, [Os(η6-p-
cym)(AZPY-NMe2)Cl]PF6, which remained inert towards hydrolysis even when heated
with 1 mol. equiv. AgNO3 [5]. Interestingly, this complex has sub-micromolar activity
against A2780 cells, which may in part be a result of its inherent stability.
3.4. Antiproliferative activity
The IC50 values for complexes 2-7 and 9-11, and 13-24 towards A2780 human
ovarian cancer cells were determined and are summarised in Table 1 and Fig. S3.
Complexes 1 and 12 were too insoluble in aqueous media to conduct biological assays.
-
However, replacing the PF6 counter-anion of 1 with CF3SO3- led to a 45-fold increase in
solubility (from 9 to 405 µM maximum solubility in 100 mM NaCl(aq) at 20 °C, Fig. S3),
-
enabling biological evaluation of 2. Moreover, changing the counter-anion from PF6 to
CF3SO3- has no major impact on antiproliferative activity, as shown by the similar
activities of 5 and 6 (A2780 IC50 = 0.92 and 0.9 µM, respectively), which exhibit
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 IC50(MRC-5)/IC50(A2780). With the exception of 19,
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
exhibiting low nanomolar potency (96 nM), whilst 11 exhibits the highest activity against
A2780 cells, comparable to other complexes reported in its class [4, 5]. In contrast, these
complexes are generally less active against MCF-7 cells, with the highest activity
observed for 5 (1.2 µM). Overall, complex 9 displays the broadest anticancer profile
amongst all the tested cell lines and the highest SF; comparable with that of cisplatin
[14].
Table 1
IC50 concentrations (µM) for complexes 2-7 and 9-11, and 13-24 against A2780 human
ovarian cancer cells.
A2780 Ovarian cancer cells
17

Journal Pre-proof
Complex
IC50 (µM)
n.d.
(1) [Os(η6-p-cym)(5-MeO-AZPY)I]PF6
(2) [Os(η6-p-cym)(5-MeO-AZPY)I]CF3SO3
(3) [Os(η6-p-cym)(5-EtO-AZPY)Cl]PF6
(4) [Os(η6-p-cym)(5-EtO-AZPY)Br]PF6
(5) [Os(η6-p-cym)(5-EtO-AZPY)I]PF6
(6) [Os(η6-p-cym)(5-EtO-AZPY)I]CF3SO3
(7) [Os(η6-p-cym)(5-EtO-AZPY)I]IO3
0.5(±0.1)
15.1(±0.5)
14.1(±0.9)
0.92(±0.02)
0.9(±0.2)
1.6(±0.3)
n.d.
(8) [Os(η6-bip)(5-EtO-AZPY)I]PF6
(9) [Os(η6-bip)(5-EtO-AZPY)I]CF3SO3
(10) [Os(η6-p-cym)(5-nPrO-AZPY)I]PF6
(11) [Os(η6-p-cym)(5-iPrO-AZPY)I]PF6
(12) [Os(η6-p-cym)(5-nBuO-AZPY)I]PF6
0.51(±0.02)
0.33(±0.02)
0.30(±0.09)
n.d.
(13) [Os(η6-p-cym)(5-HOCH2CH2O-AZPY)Cl]PF6
(14) [Os(η6-p-cym)(5-HOCH2CH2O-AZPY)Br]PF6
(15) [Os(η6-p-cym)(5-HOCH2CH2O-AZPY)I]PF6
(16) [Os(η6-bip)(5-HOCH2CH2O-AZPY)I]PF6
(17) [Os(η6-p-cym)(5-HO{CH2CH2O}2-AZPY)I]PF6
(18) [Os(η6-p-cym)(5-MeO{CH2CH2O}2-AZPY)I]PF6
(19) [Os(η6-p-cym)(5-EtO{CH2CH2O}2-AZPY)I]PF6
(20) [Os(η6-bip)(5-EtO{CH2CH2O}2-AZPY)I]CF3SO3
(21) [Os(η6-p-cym)(5-HO{CH2CH2O}3-AZPY)I]PF6
(22) [Os(η6-p-cym)(5-MeO{CH2CH2O}3-AZPY)I]PF6
(23) [Os(η6-p-cym)(5-EtO{CH2CH2O}3-AZPY)I]PF6
(24) [Os(η6-p-cym)(5-HO{CH2CH2O}4-AZPY)I]PF6
>100
>100
20.7(±0.1)
2.37(±0.09)
2.04(±0.06)
1.80(±0.09)
0.68(±0.03)
1.9(±0.2)
12(±1)
2.1(±0.2)
2.6(±0.4)
7.6(±0.6)
Table 2
18

Journal Pre-proof
IC50 concentrations (µM) for complexes 5, 9, 11, 18, and 19, against various cancer cell
lines and normal MRC-5 cells. Selectivity factors (SF) are calculated as the ratio
IC50(MRC-5)/IC50(A2780).
Cell line
IC50 / µM
Complex
MCF-7
1.2(±0.2)
15(±2)
SUNE1
0.86(±0.06)
0.31(±0.02)
n.d.
OE19
0.20(±0.03)
0.096(±0.004)
n.d.
MRC-5
2.2(±0.2)
6.1(±0.1)
1.99(±0.07)
7.4(±0.8)
3.6(±0.3)
SF
2.4
12.0
6.6
4.1
5.3
5
9
11
18
19
7.1(±0.2)
10(±1)
n.d.
n.d.
12.9(±0.8)
1.7(±0.1)
0.42(±0.01)
3.5. Capacity factor and correlation with antiproliferative activity
The capacity factor (K) of a compound is related to its RP-HPLC retention time
and is independent of column geometry and mobile phase flow rate. It serves as a
measure of affinity towards a hydrophobic stationary phase versus an aqueous mobile
phase, hence providing a measure of relative lipophilicity [33, 34]. The higher the K-
value of a compound, the greater its lipophilicity. The K-values of complexes 2-6 and 9-
24 were determined by RP-HPLC using a C18 column and isocractic mobile phase of
MeCN:H2O (1:1, v/v, 50 mM NaCl), and are shown in Fig. S3. Trends in the capacity
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 IC50 value
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
(5)
R = Et
(3)
X = Cl
16
14
12
10
8
(4)
0.9
0.8
0.7
X = Br
Correlation = -0.96
0.6
0.5
0.4
0.3
0.2
(2)
6
R = Me
(10)
R = nPr
R = iPr
4
(11)
2
Correlation = -0.55
(5)
3.5
X = I
0
1.5
2.0
2.5
3.0
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)
(C)
(D)
(15)
n = 1
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
20
15
10
5
(17)
R* = H
(18)
Correlation = -0.69
R* = Me
(21)
n = 3
Correlation = -0.90
n = 4
(24)
(17)
(19)
n = 2
R* = Et
0
1.0
1.5
2.0
2.5
3.0
3.5
4.0
1.00 1.02 1.04 1.06 1.08 1.10 1.12 1.14
Capacity Factor (K)
Capacity Factor (K)
Fig. 3. Correlations between antiproliferative activity (IC50 towards A2780 cells, µM)
and capacity factor (K), where different substituents on the complexes are varied; (A)
monodentate halido ligand, X, (B) alkoxyl side-chain substituent, R, (C) length of
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).
Increasing the length of this hydrophobic substituent increases K (hence relative
n
lipophilicity) in the order: Pr>iPr>Et>Me. However, activity follows a different trend:
iPr>nPr>Me>Et, that cannot be explained solely by capacity factors.
Moderate correlation was observed between activity and relative lipophilicity for
complexes bearing glycolic side-chains, where the length of the glycol (n) was varied and
where R*=H (-0.69, Fig. 3c). Varying n does not appear to have a significant effect on
the relative lipophilicity of the complex, possibly because each unit increase in the glycol
chain introduces an additional ethereal oxygen capable of H-bonding with water.
Likewise, there is no clear trend in activity, which follows the pattern: n=2 > 4> 3 > 1.
In contrast, when the terminal group on the glycol side-chain (R*) is varied, a strong
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 (Po/w) of complexes were measured using a
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 Po/w for our complexes, further
experiments at fixed concentrations would be required [36]. In general, increasing the
ionic strength of OSW tends to increase Log Po/w values for lipophilic compounds [37].
Log Po/w was determined for selected complexes (4, 13, 15, 16, 18, 19 and 21), as it was
not possible to make measurements on complexes with very poor aqueous solubility due
to the limit of Os detection via ICP-MS after partitioning.
An alternative RP-HPLC method for determining Log PHPLC was adopted; using the
K-values, a calibration curve of Log Po/w vs. Log K was plotted for complexes 4, 13, 15,
16, 18, 19 and 21 [38] (Fig. S4), and Log PHPLC values for all complexes were derived
(Table 3). Due to relatively weak correlation (+0.743), the Log PHPLC values have error
margins, but are suitable enough for these purposes of comparison. The Log K values of
5, 6 and 10-12 fall outside the range of the calibration plot and their Log PHPLC values
were estimated by extrapolating the curve. All determined Log PHPLC values fall well
within the Ro5 range (-0.4 to +5.6). Following an approach outlined by Sahu et al, Log
Po/w was plotted on the x-axis as a reference indicator against the expression Log (K/Po/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 PHPLC values, and Log Po/w
values for complexes 4, 13, 15, 16, 18, 19 and 21 measured using a modified shake-flask
method.
Log (K)
Log (Po/w
)
Log (PHPLC)
Complex
2
3
0.42(±0.02)
0.29(±0.01)
n.d.
n.d.
2.1(±0.5)
1.8(±0.4)
2.1(±0.5)
2.5(±0.6)
2.6(±0.6)
2.5(±0.6)
3.0(±0.7)
2.9(±0.7)
3.5(±0.8)
0.6(±0.4)
0.8(±0.3)
1.2(±0.2)
1.1(±0.2)
1.2(±0.2)
2.1(±0.5)
2.5(±0.6)
2.4(±0.5)
1.3(±0.2)
2.1(±0.5)
2.5(±0.6)
1.3(±0.2)
4
0.42(±0.01)
1.59(±0.03)
n.d.
5*
6*
9
0.58(±0.01)
0.60(±0.01)
n.d.
0.56(±0.01)
n.d.
10*
11*
12*
13
14
15
16
17
18
19
20
21
22
23
24
0.779(±0.009)
0.733(±0.008)
0.972(±0.007)
-0.25(±0.02)
-0.13(±0.02)
0.02(±0.02)
n.d.
n.d.
n.d.
0.11(±0.02)
n.d.
1.87(±0.05)
1.10(±0.04)
n.d.
0.00(±0.03)
0.038(±0.009)
0.420(±0.005)
0.574(±0.007)
0.548(±0.007)
0.044(±0.007)
0.402(±0.006)
0.555(±0.006)
0.045(±0.007)
2.29(±0.04)
2.6(±0.2)
n.d.
1.41(±0.04)
n.d.
n.d.
n.d.
* 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.
When the R-substituent is varied, the trend in cell uptake follows the same order as
lipophilicity: Et > Me > HOCH2CH2, as observed for complexes 2, 5, 6 and 15. Overall,
reasonably strong correlation is observed between cellular accumulation and K-value for
complexes 2-6, 9 and 13-15 (+0.78), thus indicating a linear trend between the lipophilicity of
this series of complexes and their ability to accumulate in cancer cells (Fig. S6).
X = Cl, Y = PF6, Ar = p-cym
X = Br,Y = PF6, Ar = p-cym
X = I, Y = PF6, Ar = p-cym
X = I, Y = CF3SO3, Ar = p-cym
X = I, Y = CF3SO3, Ar = bip
R = Et
R = Me
35
30
25
20
15
10
5
12( 2)
10( 2)
8( 2)
0
(
(7
2
)
)
(
3
)
)
(
(9
4
)
)
(
(
1
5
0
)
)
(
(
1
6
1
)
)
(
(
1
9
3
)
)
--
(14)
--
(13) ((114)
(15)
(17) 8) (19)
Fig. 4. Trends in cellular Os accumulation for complexes 2-6, 9 and 13-15 in A2780 cells
after treatment with IC50/3 concentration of complex for 24 h.
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).
However, upon scanning from 0.0 to -1.0 V, chlorido complex 3 exhibited a reversible
first reduction potential, in contrast to its iodido anologues, all of which were
irreversible. Interestingly, the first reduction of 3 was only reversible when the azo-bond
did not undergo a second e- transfer, suggesting irreversible chemical changes to the
structure occur once the azo-bond is fully reduced. Furthermore, second segments in the
voltammograms for 5 and 9 show oxidation potentials (-0.35 and -0.31 V, respectively)
not corresponding to reversal of their first azo-bond reductions. Hence implying that
irreversible chemical changes to iodido complexes can occur after the first azo-bond
reduction.
In summary, the first and second azo-bond reductions of these complexes range
between -0.75 to -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-
NMe2)I]PF6 (FY026), and its Ru analogue, [Ru(η6-p-cym)(AZPY-NMe2)I]PF6, were
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
15.0
15.0
10.0
5.0
RN121
RN121
(A)
-0.77 V
-0.78 V
-1.63 V
10.0
5.0
S2
S1
S1
S2
0.0
-2.0
-1.5
-1.0
-0.5
0.0
0.0
Voltage (V)
-1.0
-0.8
-0.6
-0.4
-0.34 V
-0.2
0.0
-5.0
-10.0
-15.0
-0.69 V
Voltage (V)
-5.0
-10.0
-0.30 V
-0.67 V
15.0
10.0
5.0
15.0
10.0
5.0
RN304
RN304
(B)
-0.79 V
-0.78 V
-1.50 V
S1
S1
S2
0.0
0.0
-2.0
-1.5
-1.0
-0.5
0.0
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
Voltage (V)
Voltage (V)
-5.0
-10.0
-15.0
-5.0
-10.0
-15.0
20.0
15.0
10.0
5.0
S2
-0.36 V
-0.35 V
20.0
15.0
10.0
5.0
RN318
RN318
(C)
-0.75 V
-0.75 V
-1.25 V
S1
S2
S2
S1
0.0
-2.0
-1.5
-1.0
-0.5
0.0
0.0
Voltage (V)
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
-5.0
-10.0
-15.0
Voltage (V)
-5.0
-10.0
-0.29 V
-0.31 V
Fig. 5. Cyclic voltammograms for complexes 3 (A), 5 (B), and 9 (C) in MeCN with 0.1
M Bu4NPF6 as supporting electrolyte. Complexes were scanned between 0.0 to -2.0 V
(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 O2·- from other ROS. The positive control was
treated with pyocyanin and showed the majority of cells expressing high levels of
ROS/O2⸱-, and the negative control was untreated cells. Figs. 6 and S9 show
quantitatively the elevation of ROS after incubation with 2, 5, 9, 11, and 19 at 1× and 2×
IC50. In every case, significantly elevated levels of ROS and O2⸱- were observed, thus
further highlighting that the mechanism of action of this class of Os(II) complex involves
generating oxidative stress in cancer cells [5, 6, 14, 40]. Interestingly, 5 and 19 both
elevated cellular levels of O2⸱- when their concentrations were doubled, as shown by
significant increases in the populations of cells in quadrant Q2. Complexes 3, 5 and their
intracellularly generated hydroxido analogue, [Os(η6-p-cym)(5-EtO-AZPY)OH]+ (3OH),
were also previously reported to generate the highly reactive hydroxyl radical (OH⸱)
upon catabolising H2O2 [32]; a comparatively weak ROS that is over-produced in cancer
cells [41-43]. Hydroxyl radicals cause damage to cellular lipids, proteins and
compartments, leading to cell death [44] and the ability of the complexes studied here to
produce both O2⸱- and OH⸱ under physiological conditions likely plays a role in
inducing oxidative stress in cancer cells.
27

Journal Pre-proof
(B)
Positive Control
(C)
Negative Control
Q1:
Q2:
Cells with
elevated
(A)
0.0%
0.0%
0.0%
99.5%
Cells with
elevated
-
-
O2
ROS + O2
Q4:
Healthy
Cells
Q3:
Cells with
elevated
ROS
100%
0.0%
0.4%
0.0%
2X IC50
1X IC50
0.5%
63.7%
15.0%
0.9%
(D)
(E)
3.4%
0.5%
80.7%
34.0%
0.0%
1.8%
35.8%
53.3%
44.9%
0.1%
0.1%
65.4%
Fig. 6. (A) Diagram showing the four quadrants in the cell population plots; Q1 = cells
with elevated levels of superoxide, Q2 = cells with elevated levels of ROS and
superoxide, Q3 = cells with elevated levels of ROS, and Q4 = healthy cells. The
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× IC50), and (E) various concentrations of 19.
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× IC50, complex 5 causes S-phase cell cycle
arrest with an observable decline in the proportion of cells in G1 phase, and at 2× IC50, a
rise in the proportion of sub-diploid cells is present. These are damaged non-viable cells
with compromised membranes and leakage of their DNA content. In contrast, 15 causes
slight elevations in both the G1 and S-phase populations at 1× IC50 alongside a small
observable decline in the G2/M-phase population. Doubling the concentration induces a
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
subsequent electronic properties.
29