Tripodal O-N-O Bis-Phenolato Amine Titanium(IV) Complexes Show High in vitro Anti-Cancer Activity

Article abstract of DOI:10.1002/ejic.201900510

The octahedral titanium(IV) complexes trans,mer-[Ti{R³N(CH₂C₆H₂-2-O-4-R²-6-R¹)₂}₂] (R¹ = Me, OMe, Cl; R² = Me, OMe, F, Cl; R³ = Me, Et; not all combinations) are synthesised in two steps from simple phenols in 36–53 % overall yield. The highly crystalline (4 X-ray structures) complexes are active against MCF-7 (breast) and HCT-116 (colon) cancer cell lines showing widely varying GI₅₀ values in the range 1–100 μM depending on R¹–R³. Highest activities are realised when R¹ = OMe and R², R³ = Me (GI₅₀ ca. 1 μM for MCF-7 and 2–3 μM for HCT-116). These are respectively 8× and 3× times greater than the activities of cisplatin in the same cell lines. These titanium complexes show some significant selectivity for cancer cell lines; up to 7× higher in MCF-7 compared to non-cancer (MRC-5) fibroblast cells. Details of cellular mode of action indicators (cell cycle perturbation, Annexin V, γ-H2AX, and caspase studies) that point to an apoptosis mode for the most active compound (R¹ = OMe and R², R³ = Me) are also reported.||||||.

Full text of DOI:10.1002/ejic.201900510

A Journal of
Accepted Article
Title: Tripodal O-N-O bis-Phenolato Amine Titanium(IV) Complexes
Show High In Vitro Anti-Cancer Activity
Authors: Mohammed Abid, Ryan Nouch, Tracey D. Bradshaw, William
Lewis, and Simon Woodward
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201900510
Link to VoR: http://dx.doi.org/10.1002/ejic.201900510

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Tripodal O-N-O bis-Phenolato Amine Titanium(IV) Complexes Show High
In Vitro Anti-Cancer Activity
Mohammed Abid,[a,d] Ryan Nouch,[d] Tracey D. Bradshaw,[b] William Lewis,[c,d] and Simon
Woodward*[d]
[a] M. Abid, Department of Chemistry, College of Science, University of Anbar, Western side
of Ramadi City, Anbarshire (Republic of Iraq)
Emails: aboyazenabid@gmail.com and mohammed.abid@nottingham.ac.uk
[b] Associate Prof. Dr. T. D. Bradshaw, School of Pharmacy, Centre for Biomolecular Sciences,
University Park, Nottingham, NG7 2RD (United Kingdom)
Email: tracey.bradshaw@nottingham.ac.uk
[c] Present address: Dr. W. Lewis, School of Chemistry, The University of Sydney, Eastern
Avenue, Sydney, NSW 2006 (Australia)
Email: w.lewis@sydney.edu.au
[d] M. Abid, Dr. W. Lewis, R. Nouch, Prof. Dr. S. Woodward, GSK Carbon Neutral Laboratories
for Sustainable Chemistry, University of Nottingham, Triumph Road, Nottingham, NG7 2TU
(United Kingdom)
Email: simon.woodward@nottingham.ac.uk
URL: www.nottingham.ac.uk/~pczsw/SWGroup/
Supporting information for this article is available (end of manuscript)
Abstract: The octahedral titanium(IV) complexes trans,mer-[Ti{R3N(CH2C6H2-2-O-4-R2-6-R1)2}2]
(R1 = Me, OMe, Cl; R2 = Me, OMe, F, Cl; R3 = Me, Et; not all combinations) are synthesised in
two steps from simple phenols in 36-53% overall yield. The highly crystalline (4 X-ray
structures) complexes are active against MCF-7 (breast) and HCT-116 (colon) cancer cell lines
showing widely varying GI50 values in the range 1-100 µM depending on R1-R3. Highest
activities are realised when R1 = OMe and R2, R3 = Me (GI50 ~1 µM for MCF-7 and 2-3 µM for
HCT-116). These are respectively 8 and 3 times greater than the activities of cisplatin in the
same cell lines. These titanium complexes show some significant selectivity for cancer cell
lines; up to 7 higher in MCF-7 compared to non-cancer (MRC-5) fibroblast cells. Details of
cellular mode of action indicators (cell cycle perturbation, Annexin V, -H2AX, and caspase
studies) that point to an apoptosis mode for the most active compound (R1 = OMe and R2, R3
= Me) are also reported.
1
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Introduction
In 1969, cisplatin [cis-diamminedichloroplatinum(II)] first emerged as a cancer therapy,
subsequently becoming one of the most popular antineoplastic (cytotoxic) agents employed
against human ovarian, testicular, head, neck and bladder carcinomas. Although cisplatin
possesses potential against most carcinomas, its toxic side-effects (myelotoxicity and
nephrotoxicity) and problems with resistance soon became apparent.[1] Because of such
issues, alternative metal complexes have also been explored for anti-cancer activity.[2] These
related metal complexes can show mechanisms of action other than the DNA binding seen for
cisplatin depending on their metal acidity/oxidation state and their supporting ligands.[3] In
particular, titanium(IV) complexes have recently appeared showing activities higher or
comparable to cisplatin and its Pt-based derivatives and these are thought to be promising
anticancer drug candidates.[4] Unfortunately, progress towards potential clinical candidates
has been hindered by a lack of understanding of the mechanism(s) of action of titanium
agents.[5] Initial investigations in this area focused on titanocene dichloride Cp2TiCl2 (Cp =
η-C5H5) and this candidate reached clinical trial (although its low efficacy did not allow it to
progress beyond phase II). One of the problems in deriving the cellular mode of action of
Cp2TiCl2, and related titanium(IV) complexes, is that while partial ligand hydrolysis is
apparently vital in attaining the biological activity such reactions rapidly cascade providing
complex mixtures of species. Thus, correlation of the observed cellular outcomes to a
particular titanium species becomes difficult. With the aim of slowing down and controlling
such hydrolysis events Tshuva has introduced polydentate phenolato ligands of type LA and LB
Scheme 1).[6,7] We became intrigued as to why the related ligand cores LC, after coordination
to titanium(IV), had not been trialled in anti-cancer studies (even though examples of such
complexes have been known in the fields of coordination chemistry and catalysis since
1970s).[8] We were particularly interested to see if use of ligand of type LC (Scheme 1) could
lead to highly active titanium(IV) complexes suitable for subsequent mechanistic
investigations.
Scheme 1. Previously screened (LA, LB) and an untried class LC phenolato-amine ligand cores
(shown in bold) for anti-cancer trials of titanium(IV) complexes.
2
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Results and Discussion
Ligand and complex formation and stability
While over 5000 examples of entities containing the sub structure core of LC (shown in bold)
are known,[8] it is desirable for biological studies (e.g. for oxidative stability) to block sites ortho
and para to the phenol (OH) function and to restrict R to a simple methyl unit. Only 35
examples of such structures (2, Scheme 2) are known, where R1 and R2 are restricted to Br, Cl,
Me, tBu and adamantyl.[8] Importantly, no examples of 2 have been reported containing
electron-releasing OMe units, which we postulated would be the most biologically active when
R3 = Me. We thus prepared a small library of 4,6-substituted versions of LC 2a-f (Scheme 2)
using simple Mannich chemistry. Five moderate electron-rich ligands 2a-e containing methyl
(σp = -0.17) and methoxy (σp = -0.27) units were prepared of which only 2a was previously
known.[9] Non-commercial phenols (1d-e) are accessed by simple Bayer-Villiger oxidations of
the parent aldehydes (see Supporting Information). Crude yields for 2a-e were in the range 50
to 80+% and analytically pure materials were attained by layering pentane onto their ether
solutions. Attempts to prepare 4,6-substituted LC from nitrogen containing 1h were
unsuccessful (Supporting Information). For comparison electron deficient 2f (F, σp = +0.06) and
2g (Cl, σp = +0.23) were also targeted. While 2f is prepared in moderate yield 2g could not be
isolated pure in our hands, although it is described in the literature.[10] However, compound
2g is clearly present in the crude mixture and this could be used for subsequent complexation
studies (see later).
Scheme 2. 4,6-substituted ligand Lc library prepared in this study.
3
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Reaction of ligands 2 at a 2:1 stoichiometry with Ti(OiPr)4 in toluene at ambient temperature
leads to intense red coloured solutions from which complexes 3 can be isolated in good yield
as rhomboidal crystals or orange powders upon removal of the toluene and recrystallisation
from optimal solvents (see Supporting Information) (Scheme 3).
Scheme 3. Titanium complexes 3 library prepared in this study and comparison to the more
commonly targeted motifs (A).
Ligands of type 2 (LC) complex readily to a wide variety of transition metals including: Zr,[11]
V,[12] Mo,[13] Mn,[14] Fe,[15] and Cu[16] However, for titanium only the preparation of 3a is
known[17] together with three Ti(substituted-LC)2 complexes using ligands related to 2 with R1,
R2 = Me (or its 3-isomer), tBu and R3 = Et, nPr, CH2CH2NMe2.[18] Other than that, the nearest
related titanium complexes we could identify (ca. 180 examples[8]) were those showing the
motifs (A) (Scheme 3), which have been prepared mostly for studies of alkene polymerisation
catalysis.[8] Complexes 3 are chiral (but racemic) and this causes observation of characteristic
4
diastereotopic aryl C-H (four J coupled doublets in the range H 6.61-6.44) and NCH2 (at H
4.85-3.30 showing 2J 12.8 Hz or second order behaviour) signals. Conveniently, complexes 3
~
are all highly crystalline so that they can be easily brought to analytical purity for both
biological and mechanistic studies (see Supporting Information for individual compounds). All
of the compounds 3 are stable in dmso-d6 solution (10 mM) for at least 5 days at ambient
conditions. Addition of excess D2O, simulating the serial dilutions carried out in the biological
assessments (see later), indicated that species 3 form colloidal suspensions and that hydrolysis
4
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
to afford ligands 2 as the major solution species is a slow process (>1 day, see Supporting
Information).
Crystallographic studies
The molecular structure of complex 3b was determined by X-ray studies and is shown in
(Figure 1). The same trans,mer geometry is also seen in related crystallographic studies of 3d,
3e, 3f (see Supporting Information) and for the literature structure of 3a.[17] The ligands in
these series 3 contain a common CH2NMeCH2 linker so that electronic effects on bonding to
TiO4N2 core induced by the aryl substituents can be examined. Comparing the bond distance
and angle data for these five complexes (Table 1) reveals a very similar range of geometries.
No significant structural correlation of the pro-drug structures of 3 to their cytotoxic activity
(see later) was observed, implying that the observed structure-activity-relationship is due only
to ligand electronic effects on the electron density in in vitro formed titanium species. At best
for precursors 3 a slight distortion of N-Ti-N angles correlated with the anti-cancer activity.
However, the latter may be due to crystal packing rather than substituent effects. Complexes
3d-f were found to crystallise as solvates, but these are readily lost in the absence of
supporting solvent and all subsequent biological studies were carried out on samples of 3
assayed as solvent free (NMR) and analytically pure (CHN).
Figure 1. Molecular structure of titanium complex 3b.
Table 1. Selected comparison of bond distance and angle data for titanium complexes 3a-3b
and 3d-f for their TiO4N2 cores.[a]
Complex
R1 (ortho)
R2 (para)
3a[b]
Me
Me
3b[c]
MeO
Me
3d[c]
Me
OMe
3e
OMe
OMe
3f
Me
F
Bond lengths (Å)
Ti(1)-O(1)
Ti(1)-O(2)
1.892
1.882
1.909
1.859
1.881
1.889
1.921
1.874
1.880
1.876
5
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Ti(1)-O(3)
Ti(1)-O(4)
Ti(1)-Oave
1.892
1.882
1.887
2.264
2.264
2.264
1.873
1.900
1.885
2.244
2.267
2.255
1.894
1.878
1.885
2.244
2.265
2.256
1.867
1.887
1.887
2.272
2.251
2.262
1.880
1.876
1.878
2.247
2.247
2.247
Ti(1)-N(1)
Ti(1)-N(2)
Ti(1)-Nave
Bond angles (o)
O(1)-Ti(1)-O(2) 168.6
O(3)-Ti(1)-O(4) 168.6
N(1)-Ti(1)-N(2) 179.7
167.5
167.6
174.7
167.1
167.9
173.8
165.2
166.4
173.6
167.9
167.9
179.8
[a]
The TiO4N2 donor set is numbered in all cases as for 3b in Figure 1 (see also Supporting
[b]
Information). Based on the published SEFZIN structure.[17] [c] Average of two independent
molecules within unit cell.
Growth inhibitory activity of 3b-e against MCF-7 and HCT-116
In vitro antitumor activity of all compounds (3) was assessed using the MTT assay against two
human-derived cancer cell lines: MCF-7 (breast) and HCT-116 (colon), with cisplatin as a
positive control for all tests, using standard protocols (see Supporting Information).
Compound 3a has been previously reported in the literature,[17] but its anti-cancer properties
were not determined. Trials of 3a revealed unremarkable activity (GI50 values 35-40 μM)
against both these cell lines. Exchange of a single methyl for methoxy group in complex 3b
resulted in >10-fold increase in anti-cancer activity. The growth inhibitory activities of 3a and
3b are compared in Figure 2. The GI50 values of the electronically diverse complexes 3b-f were
similarly screened by MTT assay. Complex 3b was confirmed as the most active with GI50 values
of ~1 μM for MCF-7 and ~3.4 μM for HCT-116 (Table 2). Other compounds in the series showed
moderate to good activity anti-cancer performance (Figure 3 and Table 2). Typically, no
structural modifications were made to the compounds at the nitrogen donor (e.g. R3), as
preliminary studies changing the methyl group within 3b to the ethyl derivative 3c negatively
affected the compound class, reducing potency. The activity of 3c is similar to cisplatin, whose
control GI50 values in our studies were 6.6 μM for MCF-7 and 9.2 μM for HCT-116 (Figure 3A-
B). We suspect 3c of being less soluble in aqueous media leading to reduced activity. Our
results show that Ti(IV) complexes 3b-f are among the most active reported in the literature[4-
7]
and that this activity is strongly dependent on the ligand structure. While 3b is the most
potent compound, 3d showed good activity with GI50 values of 2 μM and 3 µM against MCF-7
and HCT-116 cell lines respectively (Figure 3A-B and Table 2). Complex 3e showed lower
potency than 3b and 3d against the same cell lines (GI50 values ~8.5 µM against both MCF-7
and HCT-116 cells; Figure 3A-B and Table 2). Complex 3f revealed reduced activity with GI50
values of 11.7 μM for MCF-7 and 22 μM for HCT-116 (Figure 3A-B and Table 2).
6
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Figure 2. Representative MTT graphs displaying the dose-dependent growth inhibitory
properties of 3a-b against MCF-7 (A) and HCT-116 (B) cell lines. Cells were seeded in 96-well
plates at a density of 3 x 103 cells/well. After allowing 24 h to adhere, cells were treated with
the specified compound and incubated for 72 h. Data points depict mean ±S.D. n = 8; MTT
assays were repeated >3 times. LineT0 is a control run: using just media and cancer cell line.
7
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Figure 3. Representative MTT graphs displaying the dose-dependent growth inhibitory
properties of 3b-f against MCF-7 (A) and HCT-116 (B) cell lines. Cells were seeded in 96-well
plates at a density of 3 x 103 cells/well. After allowing 24 h to adhere, cells were treated with
the specified compound and incubated for 72 h. Data points depict mean ±S.D. n = 8; MTT
assays were repeated >3 times. LineT0 is a control run: using just media and cancer cell line.
Tetrachloro analogue 3g showed the lowest activity amongst the new derivatives of 3
screened and was comparable to complex 3a. The 3g GI50 values were: 51.5 μM for MCF-7 and
64.8 μM for HCT-116 (Figure S4, see Support Information). The selectivity of 3b to cancerous
cells was evaluated by measuring its growth inhibitory activity in non-cancerous MRC-5 human
lung fibroblast (Figure S5-A, see Support Information). Complex 3b showed reduced activity
(cytotoxicity) toward these cells, implying a degree of cancer cell selectivity.[19] The GI50 value
of 3b in MRC-5 cells was higher at 7.3 μM (Table 2), indicating a ~7-fold selectivity when
compared to activity against MCF-7 breast cancer cells. The activities of all Ti(IV) complexes 3
were similarly assessed against MRC-5 and consistently, reduced activities were observed
against these non-tumourigenic fibroblasts; the GI50 values for 3e, 3f and 3g were ~16 μM ,
34 μM and >100 μM respectively. This was in complete contrast to cisplatin controls, which
were equi-active across all three cell lines (cancerous MCF-7 and HCT-116 and benign MRC-5
8
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
fibroblasts). Thus, the novel Ti(IV) complexes 3 demonstrate significant cancer cell selectivity
(Table 2 and Figure S5-B in Supporting Information).
The in vitro growth inhibitory activity of complexes 3a-g are summarised in Table 2. An
interesting correlation is seen between the electronic properties of the phenolic substituents
vs. the cellular cytotoxic activity. Moderate donation (R1, R2 = Me, MeO) into the phenolate
results in increased activity, but if this is either further increased (3e) or decreased (3a, 3f, 3g)
the biological activity falls. Bulky substitution of R3 (3b vs. 3c) also has a negative effect on
biological performance.
Table 2. Activity of complexes 3 (GI50 values in µM by MTT assay) against MCF-7 and HCT-116
cancer cell lines and normal MRC-5 fibroblast cells. Data were generated from ≥3 independent
trials; n = 8 per experimental condition per trial.
R1
R2
R3
σp(R1)[a] σp(R2)[a]
GI50 (MCF-7)
(µM)[b]
GI50 (HCT-
116) (µM)[b] (µM)[b]
GI50 (MRC-5)
3
3a
3b
3c
3d
3e
3f
Me
MeO Me
MeO Me
Me
MeO MeO Me -0.27
Me
Cl
-
Me
Me -0.17
Me -0.27
-0.17
-0.17
-0.17
-0.27
-0.27
+0.06
+0.23
-
36.3±0.1
1.0±0.04
6.6±0.07
2.2±0.06
8.4±0.1
11.7±0.05
51.5±0.1
7.8±0.04
38.6±0.1
3.4±0.07
9.2±0.09
3.0±0.05
8.6±0.15
22±0.031
64.8±0.07
8.4±0.06
>50
7.33±0.04
15.23±0.01
8.38±0.04
16.0±0.1
34±0.03
>100
Et
-0.27
MeO Me -0.17
F
Cl
-
Me -0.17
Me +0.23
3g
Cis-Pt[c]
-
-
7.6±0.09
[a]
[b]
Hammett parameter value for para substituent.[20]
As determined by MTT assay (3
duplicates). [c] Cisplatin.
Effect of 3b and 3d-f on MCF-7 and HCT-116 colony formation
Having established that Ti(IV) analogues 3 cause selective growth inhibitory activity in two
cancer cell lines, we sought to evaluate whether these cancer cells could survive a brief
challenge of 24 h exposure to 1 µM or 5 µM of representative Ti complexes 3 and retain their
proliferative capacity. To this end, clonogenic assays were set up. Clonogenic assays are a
recognised technique for determining the cytotoxic potential of a test agent and results are
typically represented as a survival fraction (%) vs control cell colony populations of non-treated
cells. Both 3b and 3d significantly inhibited colony formation after 24 h treatment of cells at
concentrations of 1 and 5 µM (Figure 4A-B). Complexes 3b and 3d inhibited MCF-7 colony
formation cell at 1 and 5 µM by 40-45% and 100% respectively. Despite the fact that MCF-7
and HCT-116 showed variable sensitivities to 3b and 3d (GI50 1.0 μM and 3.4 μM respectively),
similar inhibition of HCT-116 colony formation was observed, 51% and 100% by 1 and 5 µM
values respectively (Figure 4A-B). Complexes 3e and 3f also inhibited the colony formation in
MCF-7 cells by 25% and 29% (1µM), and 57% and 48% (5 µM) respectively. The same agents
inhibited HCT-116 colony formation by 16% and 22% (1 µM 3e and 3f) and 49% and 28% (5
µM 3e and 3f; Figure 6A-B). Complete (100%) inhibition of colony formation was observed in
HCT116 cells and MCF-7 exposed to 5 µM 3b and 3d, reflecting potency and cytotoxicity of
9
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
these complexes (Figure S6 C and D, see Support Information). These data indicate that the
cancer cells have lost their ability to form progeny colonies or have been killed (Figure S6-C,
see Support Information).[21] Representative colonies are shown in (Figure S6 C-D, see Support
Information). In contrast, at the concentrations adopted, 3e and 3f demonstrated reduced
potency, therefore exposure of MCF-7 or HCT-116 cells to 1 µM or 5 µM 3e and 3f failed to
inhibit colony formation by >50%.[22]
Figure 4. Effect 3b, 3d, 3e, and 3f on A) MCF-7, B) HCT-116 colony formation. A + B Mean
survival fraction of % control represented of mean ±SEM of 5 independent trials (n = 2 per
trial). Complexes 3b and 3d exhibited significant reduction in colony formation while 3e and
3f showed modest activity by comparison (p < 0.0001, n = 2).
Cell cycle analysis
Based on MTT and clonogenic assays, the complexes 3 compromise cancer cell viability and
growth. Led by these observations, we examined the effect of 3b, 3f (as efficient and modest
agents, i.e. GI50 values of ca. 7 and 34 µM against the HCT-116 cell line respectively), and
cisplatin (as a positive control) on cell cycle perturbation by flow cytometry. An HCT-116
control cell cycle distribution profile demonstrated: 53.4 %, 24.9 % and 15.3 % events for the
G1, S, G2/M phases[5] respectively. Negligible events can be detected in the preG1
compartment reflecting a healthy HCT-116 population (Figure S7, see Supporting Information).
In contrast, following 72 h exposure of these cells to 10 µM (~1.5 x GI50) cisplatin, a substantial
pre-G1 population is evident (25.5 % events) indicative of cells undergoing apoptosis. It is
known that cisplatin causes DNA inter- and intra-strand crosslinks,[23] blocking DNA replication;
alerted DNA repair mechanisms fail to repair cisplatin-induced DNA damage (initially) initiating
apoptosis. As DNA replication is blocked and repair is attempted, the cell cycle is halted during
S and G2 phases. After 72 h of continued treatment, DNA repair failure triggers programmed
10
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
cell death. No preG1 peak is evident in HCT-116 cells exposed to 10 µM (~3 x GI50) 3b; this
Ti(IV) analogue caused a profound accumulation of events in the G2/M phases, indicating a
different/delayed mechanism of action. If the arrested cell cycle occurs as a result of DNA
damage, its onset may be later than that caused by cisplatin.[24] In contrast, complex 3f (at 10
µM; <0.5 x GI50) caused negligible apparent cell cycle perturbation. MCF-7 cells exposed to
cisplatin, 3b or 3f demonstrated modest increase in late S/G2/M events (Supplementary Figure
S8).
Induction of apoptosis in cancer cells
Annexin-V/PI apoptosis assays[5] were performed to explore the apoptosis-inducing properties
of 3b and 3f; cisplatin being used as a positive control. MCF-7 and HCT-116 cells were exposed
to both complexes (5 µM and 10 µM; 72 h). Apoptotic populations were confirmed by dual
annexin V-FITC/PI staining (Figure 5). The majority (65%) of cells exposed to 10 µM cisplatin
(72 h) were undergoing apoptosis. Both 3b and 3f revealed profound apoptotic MCF-7 and
HCT-116 populations at the concentrations tested. However, while early and late apoptosis
were evident in HCT-116 cells treated with 3b, exposure to 3f for 72 h revealed a larger
proportion of cells undergoing early apoptosis – indicating later onset of programmed cell
death (Figure 5, and Figures S9, S10 Supporting Information).
Figure 5. Apoptotic effects of 3b, 3f , and cisplatin on A) MCF-7 and B) HCT-116 cells at 10 µM.
Cells were treated with 10 µM for 72 h. Annexin-V/PI apoptosis assay was adopted to
determine the percentage of apoptotic cells.
11
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Induction of DNA damage
To determine whether apoptosis is triggered by DNA double strand breaks (DSBs), flow
cytometric analyses of control, cisplatin-, 3b- and 3f-treated HCT-116 and MCF-7 populations
were performed following co-incubation with PI (propidium iodide) and a 1o Ab (primary
antibody) recognising -H2AX;[5] -H2AX foci appear at sites of DNA DSBs.[25] It is known that
cisplatin causes DNA damage: inter/intrastrand DNA crosslinks, and DNA monoadducts that
lead to DNA DSBs. These harbingers of apoptosis are largely responsible for Pt-drug antitumor
activity.[1,23] Thus, co-incubation of cell populations with PI and 1o Ab recognising -H2AX, prior
to flow cytometry analyses, allows cell-cycle-specific detection of sites of DNA DSBs.
Consistent with cell cycle analyses, dual PI/-H2AX labelling of HCT-116 cells supported 3b-
induced G2/M cell cycle arrest. Interestingly, HCT-116 G2/M events comprised a population
heavily positively stained -H2AX+ve (53 %) suggestive of the presence of DNA DSBs. Complex
3f, in contrast (consistent with HCT-116 cell cycle analysis and reduced potency) revealed
neither G2/M arrest, nor -H2AX-+ve events associated with this cell phase. Rather a minority
proportion of G1 events appeared to have incurred DNA DSB damage and were -H2AX+ve.
Validating these analyses, accumulation of G2/M cell cycle events following exposure of HCT-
116 cells to cisplatin (10 µM; 72 h was observed accompanied by -H2AX +ve events at sites of
DNA DSBs (Figure S11, Support Information). Obvious -H2AX +ve events were detected in G1
and G2 cell cycle phases following exposure of MCF-7 cells to cisplatin, 3b or 3f (Supporting
Information, Figure S12). Interestingly, chiral cyclopentadienyl complexes did not halt the cell
cycle in treated cells and DNA DSBs (-H2AX foci) were similarly not detected.[26] Instead,
antitumor activity of this compound was a consequence of paraptotic cell death, involving
MAP kinase signal transduction. In contrast, current Ti complexes 3b and 3f clearly evoke
apoptosis.
Caspase activation
Taken together, our data suggest that potent Ti-complex 3b causes later-onset apoptosis
compared to cisplatin. To irrefutably confirm an apoptotic cell fate for cells exposed to 3b, a
caspase 3/7 activation assay was performed following exposure of HCT-116 and MCF-7 cells
to escalating concentrations of 3b. Dose-dependent significant caspase activation (≥5 µM 3b
p< 0.0001; Figure 6) was detected following 72 h exposure of cells to 3b consistent with
apoptosis-induction. As a positive control, 50 µM cisplatin was to confirm the assay raised
caspase activation (e.g. by ~350% in HCT-116 populations; not shown). These data confirm
that after 72 h exposure, 3b triggers apoptotic cell death (Figure 6).[27]
12
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Figure 6. Dose-dependent elevation of caspase 3/7 activity in HCT-116 and MCF-7 cells
following 72 h exposure of cells to Ti-complex 3b. Data points are mean ±S.D; (p < 0.0001, n =
8).
Conclusion
A series of Ti(IV) complexes (3) has been synthesised and their structures confirmed by X-ray
crystallography. Complex 3b evoked potent cancer-selective growth inhibitory and cytotoxic
activity. Cell cycle analyses indicated late-onset G2/M cell cycle arrest, which at 10 µM
(representing ~3 x GI50) triggered an apoptotic cell fate. Experiments are underway to
determine modes of cell death and molecular targets of this promising class of titanium
anticancer agents.
Acknowledgments
MA is grateful to the University of Anbar (Iraq) for PhD study leave (award IDB600032695). RN
is grateful to Aesica and the School of Chemistry for support of a studentship. We thank Dr
Huw E. Williams, Dr Thomas Macinally, Mr Mustapha Musa and the reviewers for helpful
suggestions.
Graphical Abstract
13
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
Keywords: • Antiproliferation • Cytotoxicity • Titanium • Phenolato • MCF-7 • HCT-116 • Cell
cycle • Annexin V • -H2AX • Caspase • DNA DSBs
TOC Text: ‘Clothes make the man’, but substituents make the activity in titanium anti-cancer
agents! Appropriately placed OMe groups yield a >100-fold reduction in GI50 value for the MCF-
7 cancer cell line, suggesting highly active complexes for this motif.
References
[1] For leading overviews of Pt-agents see: (a) E. R. Jamieson, S. J. Lippard, Chem. Rev. 1999,
99, 2467–2498. (b) J. S. Saad, M. Benedetti, G. Natile, L. G. Marzilli, Inorg. Chem. 2010,
49, 5573-5583 and citations thereof.
[2] For introductions to metal (including non-Pt) anti-cancer agents see: (a) Metal-based
Anticancer Agents, Eds. A. Casini, A. Vessières, S. M. Meier-Menches, Royal Society of
Chemistry, Cambridge, UK, 2019, pp 500. (b) U. Ndagi, N. Mhlongo, M. E. Soliman, Drug
Des. Devel. Ther. 2017, 11,599-616. (c) M. J. Clarke, F. Zhu, D. R. Frasca, Chem. Rev. 1999,
99, 2511-2534.
[3] D. R. Williams, Chem. Rev. 1972, 72, 203-213.
[4] For overviews of titanium anti-cancer agents see: (a) E. Y. Tshuva, J. A. Ashenhurst,
Eur. J. Inorg. Chem. 2009, 2203-2218. (b) K. Strohfeldt, M. Tacke, Chem. Soc. Rev. 2008,
37, 1174-1187. (c) K. M. Buettner, A. M. Valentine, Chem. Rev. 2012, 112, 1863-1881.
(d) S. A. Loza-Rosas, M. Saxena, Y. Delgado, K. Gaur, M. Pandrala, A. D. Tinoco,
Metallomics 2017, 9, 346-356.
[5] M. Cini, T. D. Bradshaw, S. Woodward, Chem. Soc. Rev. 2017, 46, 1040-1051.
[6] (a) M. Shavit, D. Peri, C. M. Manna, J. S. Alexander, E. Y. Tshuva, J. Am. Chem. Soc. 2007,
129, 12098-12099. (b) D. Peri, S. Meker, M. Shavit, E. Y. Tshuva, Chem. Eur. J. 2009, 15,
2403-2415. (c) S. Meker, O. Braitbard, M. D. Hall, J. Hochman, E. Y. Tshuva, Chem. Eur.
J. 2016, 22, 9986-9995.
[7] For recent relevant developments in the titanium anti-cancer agent area see: (a) R.
Manne, M. Miller, A. Duth, M. F. C. G. da Silva, E. Y. Tshuva, T. S. B. Baul, Dalton Trans.
2019, 48, 304-314. (b) N. Ganot, O. Briaitbard, A. Gammal, J. Tam, J. Hochman, E. Y.
Tshuva, ChemMedChem 2018, 13, 2290-2296. (c) M. Miller, E. Y. Tshuva, Nature Sci. Rep.
2018, 8, 9705. (d) S. A. Loza-Rosas, A. M. Vazquez-Salgado, K. I. Rivero, L. J. Negron, Y.
Delgado, J. A. Benjamin-Rivera, A. L. Vazquez-Maldonado, T. B. Parks, C. Munet-Colon,
A. D. Tinoco, Inorg. Chem. 2017, 56, 7788-7802. (e) S. Barroso, A. M. Coelho, S. Gómez-
Ruiz, M. J. Calhorda, Z. Zizak, G. N. Kaluðerović, A. M. Martins, Dalton Trans. 2014, 43,
17422-17433.
[8] Substructures LC were identified through Scifinder (May 2019). Prepared by simple
Mannich chemistry, use in attaining catalysts for the polymerisation of alkenes and
lactones is common; representative examples: (a) A. Lehtonen, R. Sillanpaeae, Inorg.
Chem. 2004, 43, 6501-6506. (b) C.-T. Chen, C.-A. Huang, B.-H. Huang, Macromolecules
2004, 37, 7968-7973. (c) C. M. Silvernail, L. J. Yao, L. M. R. Hill, M. A. Hillmyer, W. B.
Tolman, William B. Inorg. Chem. 2007, 46, 6565-6574. (d) A. Lehtonen, H. Balcar, J.
Sedlacek, R. Sillanpaeae, J. Organomet. Chem. 2008, 693, 1171-1176.
[9] W. J. Burke, J. L. Bishop, E. L. Mortensen Glennie, W. N. Bauer, J. Org. Chem. 1965, 30,
3423-3427.
[10] W. J. Burke, E. L. M. Glennie, C. Weatherbee, J. Org. Chem. 1964, 29, 909-912.
[11] For example: A. J. Chmura, M. G. Davidson, M. D. Jones, M. D. Lunn, M. F. Mahonc,
Dalton Trans. 2006, 887-889.
14
This article is protected by copyright. All rights reserved.

10.1002/ejic.201900510
European Journal of Inorganic Chemistry
Mohammed Abid, Ryan Nouch, Tracey Bradshaw, William Lewis, Simon Woodward – Full paper for EJIC
[12] For example: D. Maity, J. Marek, M.-F. Sheldrick; M. Ali, J. Mol. Catal. A 2007, 270, 153-
159.
[13] For example: A. Peuronen, A. Lehtonen, Topics Catal. 2016, 59, 1132-1137.
[14] For example: T. Weyhermüller, T. K. Paine, E. Bothe, E. Eberhard; E. Bill, P. Chaudhuri,
Inorg. Chimica Acta 2002, 337, 344-356.
[15] For example: T. Weyhermüller, R. Wagner, P. Chaudhuri, Eur. J. Inorg. Chem. 2011, 2547-
2557.
[16] S. Phongtamrug, K. Tashiro, M. Miyata, S. Chirachanchai, J. Phys. Chem. B, 2006, 110,
21365-21370.
[17] (a) A. J. Chmura, M. G. Davidson, M. D. Jones, M. D. Lunn, M. F. Mahon, Dalton Trans
2006, 887-889. (b) Y. Chung, Y. Kim, Acta Cryst. Sec. E: 2013, 69, m222. (c) The
Cambridge Crystallographic Data Centre (www.ccdc.cam.ac.uk) codes for these
published structures of 3a are SEFZIN and WHIMAD respectively.
[18] D. Peri, S. Meker, C. M. Manna, E. Y. Tshuva, Inorg. Chem. 2011, 50, 1030-1038. This
includes the CCDC structures REZRET: Ti{LC(2)}2 R1 = R2 = Me; R3 = Et; REZRIX: Ti{LC(2)
LC(2’)} first ligand with R1 = R2 = Me; R3 = Et and second ligand with R1 = R2 = tBu; R3 = Et
and REZRUJ: Ti{LC(2) LC(2’)} first ligand with R1 = Me (but in 3-position) R2 = Me; R3 =
CH2CH2NMe2 and second ligand with R1 = R2 = tBu; R3 = Et.
[19] For a related selectivity study see ref. 6c.
[20] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165-195.
[21] For a related example see: M. E. Qazzaz, V. J. Raja, K.-H. Lim, T.-S. Kam, J. B. Lee, P.
Gershkovich, T. D. Bradshaw, Cancer Lett. 2016, 370, 185-197.
[22] For a related example see: V. J. Raja, K.-H. Lim, C.-O. Leong, T.-S. Kam, T. D. Bradshaw,
Invest New Drugs 2014, 32, 838-850.
[23] For an overview of DNA-drug interactions see: D. Yang, J. Reedijk, G. A. van der Marel,
A. H. J. Wang, J. Biomol. Struct. Dyn. 1996, 13, 989-998.
[24] For a recent related G2 block observation see: T. Pesch, H. Schuhwerk, P. Wyrsch, T.
Immel, W. Dirks, A. Bürkle, T. Huhn, S. Beneke, BMC Cancer 2016, 16, 469.
[25] L. J. Kuo, L.-X. Yang, In Vivo 2008, 22, 305-309.
[26] (a) M. Cini, T. D. Bradshaw, S. Woodward, W. Lewis, Angew. Chem. Int. Ed. 2015, 54,
14179-14182. (b) M. Cini, H. Williams, M. W. Fay, M. S. Searle, S. Woodward, T. D.
Bradshaw, Metallomics 2016, 8, 286-297.
[27] R. U. Jänicke, Breast Cancer Res. Treat. 2009, 117, 219-221.
15
This article is protected by copyright. All rights reserved.