Synthesis and characterisation of platinum(IV) polypyridyl complexes with halide axial ligands

Inorganica Chimica Acta 495 (2019) 118964

Contents lists available at ScienceDirect

Inorganica Chimica Acta journal homepage: www.elsevier.com/locate/ica

Synthesis and characterisation of platinum(IV) polypyridyl complexes with halide axial ligands Brondwyn S. McGhiea, Jennette Sakoffb, Jayne Gilbertb, Janice R. Aldrich-Wrighta,c,

T

⁎

a

School of Science and Health, Western Sydney University, Locked Bag 1797, Penrith South DC, 2751 NSW, Australia Calvary Mater Newcastle, Waratah, NSW 2298, Australia c School of Medicine, Western Sydney University, Locked Bag 1797, Penrith South DC, 2751 NSW, Australia b

ARTICLE INFO

ABSTRACT

Keywords: Platinum N-halogensuccinimide Cell growth inhibition SRCD

A series of complexes of the type [Pt(PL)(AL)(X)2]2+ (where PL = 1,10-phenanthroline, 5-methyl-1,10-phenanthroline or 5,6-dimethyl-1,10-phenanthroline; AL = 1S,2S-diaminocyclohexane and, X = Cl, Br or I) with anticancer potency, were synthesised in a one pot reaction using N-halogensuccinimide (NXS) as both the source of the ligand and the oxidizing reagent. It was determined that 2.4 equivalents of NXS resulted in 100% oxidation of Pt(II) in a solution of equal parts H2O, ethanol and 1 M HCl. This method of oxidation was 24 times faster than the established peroxide method resulting in the acceleration of the complete synthesis of [Pt(PL)(AL)(X)2]2+. All structures were confirmed using NMR, ESI-MS, CD and UV, while the purity was confirmed by microanalysis. The in vitro cytotoxicity assays revealed that they were more active than analogous complexes with hydroxido axial ligands, and share comparable activity with the corresponding Pt(II) complex.

1. Introduction Cancer causes the death of approximately 8 million people each year and that number is predicted to rise by 70% over the next two decades [1–3]. Unfortunately, current treatments for cancer patients have low success rates for many cancer types, and especially late stage cancers [3]. For example, brain cancer is treated with a combination of surgery, radiotherapy, and chemotherapy and yet the five year survival rate is just 22% [4,5]. One of the key factors influencing the lack of success of these treatments is their low specificity for cancer cells which can result in side effects, reduced dosages, and high incidence of resistance [6–10]. As these issues take a significant toll on both patient quality of life and the health care system, new treatment strategies are needed. The improvement of widely used conventional chemotherapy drugs, including platinum based drugs such as cisplatin, oxaliplatin and carboplatin, has been explored to overcome resistance [3,11–14]. The mechanism of action of cisplatin requires that the chloride ligands are substituted with water, allowing the platinum to bind covalently to DNA and form intra- and interstrand links [3,12,13]. This interaction distorts the DNA, prevents transcription and, ultimately, leads to apoptosis [15]. While this process is not specific to cancerous cells, the rapid division of malignant cancers makes them more susceptible than normally regulated cells [16]. Cisplatin, oxaliplatin and carboplatin are ⁎

predominantly cross resistant and operate via similar mechanisms of action. As such, some cancer cells exhibit both acquired and intrinsic resistance to all three commonly used compounds [17]. A more potent and cytotoxic platinum complex which can target cancer cells more effectively to result in a better prognosis is required. The approach used here to address this need has led to the development of platinum complexes of the type [Pt(PL)(AL)]2+ which, based on their unique conformation, may supersede conventional chemotherapy drugs. Currently, the most cytotoxic [Pt(PL)(AL)]2+ complex developed is [Pt(5,6-dimethyl-1,10-phenanthroline)(1S,2S-diaminocyclohexane)]2+ (56MESS). The mechanism of 56MESS cytotoxicity is not fully understood, however, these platinum complexes interact intracellularly and with DNA in ways that are distinct from cisplatin, and can therefore overcome resistance [18–21]. 56MESS is significantly more potent than cisplatin in all cell lines tested, indicating the potential of these unconventional platinum complexes as viable cancer treatment options [22]. Improved cytotoxicity reduces the dosage required, treatment time and reduces the chance of acquired resistance. Although the in vitro activity of these complexes is exceptional, in vivo activity has been disappointing due to poor pharmacokinetics [18]. A potential solution that is currently being explored is oxidizing the compound to platinum(IV) ([Pt(PL)(AL)(OH)2]2+) from the highly cytotoxic platinum(II) complexes ([Pt(PL)(AL)]2+) (Fig. 1) [23–25]. The change in oxidation state allows for two additional axial ligands to be

Corresponding author at: School of Science and Health, Western Sydney University, Locked Bag 1797, Penrith South DC, 2751 NSW, Australia. E-mail address: [email protected] (J.R. Aldrich-Wright).

https://doi.org/10.1016/j.ica.2019.118964 Received 1 May 2019; Received in revised form 10 June 2019; Accepted 12 June 2019 Available online 13 June 2019 0020-1693/ © 2019 Published by Elsevier B.V.

Inorganica Chimica Acta 495 (2019) 118964

B.S. McGhie, et al.

Fig. 3. General structure of the synthesised compounds 1–9 where X = Cl, Br or, I and R = H or CH3 and the stereo-centres (*) are S configuration.

in the axial positions of the platinum(IV) complex ([Pt(PL)(AL)(X)2]2+) and assessed any relative improvement to their cytotoxicity. Nine compounds were synthesised in order to evaluate the relative effects of chlorido, bromido, and iodido as axial ligands when attached to the platinum(IV) analogues of platinum(II) complexes [Pt(phen) (SSDACH)]2+ (PHENSS), [Pt(5-methyl-phen)(SSDACH)]2+ (5MESS) and 56MESS (Fig. 3).

Fig. 1. Cytotoxicity comparison of 56MESS, 56MESS(OH)2 with cisplatin, carboplatin, and oxaliplatin in multiple cell lines: HT29 colon, U87 glioblastoma, MCF-7 breast, A2780 ovarian, H460 lung, A431 skin, Du145 prostate, BE2-C neuroblastoma, SJ-G2 glioblastoma, MIA pancreas and murine glioblastoma (SMA).

bound to the complex, enabling additional synthetic modifications to fine tune the pharmacokinetics of the molecule without sacrificing its cytotoxicity [26]. Once inside the cell, it has been suggested that the complex will be reduced to the Pt(II) cytotoxic complex (Fig. 2) [27]. It is necessary to identify and develop ligands which, when coordinated to 56MESS, increase its affinity to cancerous cells so that systemic cytotoxicity is reduced. Carboxylic acids [22] coordinate effectively in the axial positions and are currently being used as a potential method of attaching other compounds to the basic platinum(IV) ([Pt(PL)(AL) (OH)2]2+) structure. It has been hypothesized that replacing the hydroxido group with an excellent leaving group like a halide, may have a positive effect on the overall efficacy of the drug [28]. This hypothesis was supported by evidence that the addition of halides to the axial ligands of cisplatin, oxaliplatin, and carboplatin significantly influenced their cytotoxicity [29,30]. Here, we investigated the influence of halide (X) coordination

2. Experimental 2.1. Materials and preparation Reagents were used as received unless otherwise specified. All solvents (supplied by Labserv, Chem-Supply or Merck Chemicals) used were of analytical grade or higher. Potassium tetrachloroplatinate (K2PtCl4) was purchased from Precious Metals Online. The chloride salts of platinum(II) complexes were synthesised using previously published methods [31]. N-chlorosuccinimide, N-bromosuccinimide, Niodosuccinimide and hydrochloric acid were purchased from SigmaAldrich. Methanol was obtained from Honeywell. Deuterated solvents d6-dimethylsulphoxide (DMSO-d6, 99.9%) and deuterium oxide (D2O, 99.9%) were purchased from Cambridge Isotope Laboratories.

Fig. 2. Extra- and intracellular reduction of Pt(IV), resulting in the loss of the axial ligands, producing a Pt(II) compound with DNA binding activity. 2

Inorganica Chimica Acta 495 (2019) 118964

B.S. McGhie, et al.

spectrometer at 298 K, using 10 mM samples prepared in D2O. 1H NMR spectra were obtained using a spectral width of 8250 Hz and 65,536 data points, while 195Pt NMR spectra were acquired using a spectral width of 85,470 Hz and 674 data points. 1H-195Pt HMQC spectra were recorded using a spectral width of 214,436 Hz and 256 data points for the 195Pt nucleus (F1 dimension) and a spectral width of 4808 Hz with 2048 data points for the 1H nucleus (F2 dimension). Chemical shifts are reported in parts per million (ppm) with J coupling reported in Hz. Spin multiplicity is reported as: s (singlet), d (doublet), dd (doublet of doublet) and m (multiplet).

2.2. Synthesis Synthesis of c,c,t-[Pt(1,10-phenanthroline)(1S,2S-diaminocyclohexane)(X)2]2+, c,c,t-[Pt(5-methyl-1,10-phenanthroline)(1S,2S-diaminocyclohexane)(X)2]2+ and c,c,t-[Pt(5,6-dimethyl 1,10-phenanthroline)(1S,2S-diaminocyclohexane)(X)2]2+ [Pt(PL)(AL)](Cl)2 (80 μmol) was combined with N-halosuccinimide (2.4 equivalents) and dissolved in 9 mL of a 1:1:1 H2O, ethanol and 1 M HCl solution. The mixture was stirred at room temperature (RT) for 2 h before the volume was reduced under vacuum. A Sep-Pak (20 cc, 5 g) C-18 column connected to a pump apparatus with UV detector (Bio-Rad, EM-1 Econo™ UV Monitor) was used for initial purification. The column was activated with methanol (10 mL) and then flushed with water (~30 mL) until the UV absorbance equilibrated. The crude reaction was then loaded onto the column and eluted with water at a flow rate of 2 mL min−1. A single yellow coloured peak was detected and collected, and the resulting fraction was lyophilised. The dry compound was then weighed and loaded into a paper thimble. The thimble was placed in a soxhlet apparatus and washed continually with ~200 mL of heated toluene (170 °C) for two weeks, before being removed and allowed to dry. Any remaining impurities in the product were removed using flash chromatography. The pure fractions were combined and lyophilised to produce a pale yellow solid.

3.3. UV spectroscopy UV spectra were recorded on a Cary 1E spectrophotometer at RT in the 200–400 nm range, using a 10 mm quartz cell. All samples were automatically corrected for solvent baseline. A stock solution (1 mM, 10 µL, 10 aliquots) was titrated into known volume of water and the absorption measured after each addition to allow the extinction coefficient to be calculated. 3.4. ESI-MS

3. Biophysical characterisation

Electrospray ionisation mass spectroscopy (ESI-MS) experiments, were performed using a Waters TQ-MS triple quadrupole mass spectrometer in positive mode. Sample solutions (final concentration 0.5 mM in H2O) flowrate was 0.1 mL min−1. The desolvation temperature of 300 °C, and desolvation flow rate (nitrogen) of 500 L h−1 remained constant whilst the cone voltage and capillary voltage were varied for each sample. Spectra were collected over varied m/z ranges depending on the target mass.

3.1. Flash chromatography Samples were purified using the Reveleris X2 flash chromatography system fitted with a Reveleris reverse phase C-18 4 g column. The column was equilibrated with 3% MeOH for 2.4 min at 8 mL min−1, and the samples eluted for 9 min with the UV detector sensitivity on high and detecting at 230, 254 and 280 nm. Sample masses between 20 and 100 mg were prepared in ~1–2 mL H2O and injected with a flow rate of 8 mL min−1. The column was eluted with 3% methanol for 6 min then the methanol concentration was increased to 100% over 1 min, kept steady for 1 min before returning to 3% over another minute. In each case, only one peak was detected and this was collected in 10 mL fractions.

3.5. Circular dichroism (CD) Circular dichroism spectra were obtained using a Jasco-810 spectropolarimeter at RT. Spectra were obtained in a 10 mm quartz cell, and measured from 200 to 400 nm with a data pitch of 1 nm, bandwidth of 1 nm and response time of 1 s. For each sample solution (~28 µM), 40 accumulations were collected and a water baseline was subtracted.

3.2. NMR Spectral data were obtained using a 400 MHz Bruker Avance

Scheme 1. Reaction scheme for the formation of compounds 1–9. Where X is Br, Cl, or I. and where R1 and R2 are either H or CH3 to synthesise compounds 1–9 and succinimide bound by-products. 3

4

9.09 (d, 2H: CH, J = 5.56 Hz) –

8.33 (s,2H; CH) – – 3.35 (m, 2H; CH2) 2.35 (d, 2H; CH2, J = 15.56 Hz) 1.72 (m, 2H; CH2) 1.66 (d, 2H; CH2 J = 10.92 Hz) 1.30 (m, 2H; CH2) 1 H/195Pt −648

H2

H5 H6 CH3 H1′/2′ H3′/6′

Yield %

9.09/

H4′/5′

H4′/5′ H3′/6′

H9

H8

30

8.27 (dd, 2H: CH, J = 5.52, 8.35 Hz) –

H3

H7

9.07 (d, 2H: CH,k J = 8.31 Hz) –

PHENSS(Cl)2 (1)

Complex No.

H4

Label

21

1.30 (m, 2H; CH2) 9.19/−645.8

8.33 (s,2H; CH) – – 3.39 (m, 2H; CH2) 3.34 (d, 2H; CH2 J = 11.00 Hz) 1.69 (m, 4H; CH2) –

9.07 (d, 2H: CH, J = 5.56 Hz) –

8.26 (dd, 2H: CH, J = 5.47, 8.17 Hz) –

9.03 (d, 2H: CH, J = 8.17 Hz) –

PHENSS(Br)2 (2)

16

8.31 (s,2H; CH) – – 3.35 (m, 2H; CH2) 2.36 (d, 2H; CH2 J = 12.13 Hz) 1.73 (m, 2H; CH2) 1.66 (d, 2H; CH2 J = 10.40 Hz) 1.30 (m, 2H; CH2) 9.12/−964.3

9.11 (d, 2H: CH, J = 5.55 Hz) –

8.27 (dd, 2H: CH, J = 5.41, 8.22 Hz) –

9.04 (d, 2H: CH, J = 8.43 Hz) –

PHENSS(I)2 (3)

53

8.90 (d, 1H: CH, J = 8.28 Hz) 9.13 (d, 1H: CH, J = 8.37 Hz) 8.20 (dd, 1H: CH, J = 5.57, 8.32 Hz) 8.29 (dd, 1H: CH, J = 5.57, 8.44 Hz) 9.01 (d, 1H: CH, J = 5.32 Hz) 9.10 (d, 1H: CH, J = 5.49 Hz) – 8.09 (s,1H; CH) 2.85 (s, 3H; CH3) 3.35 (m, 2H; CH2) 2.36 (d, 2H; CH2 J = 12.09 Hz) 1.72 (m, 2H; CH2) 1.66 (d, 2H; CH2 J = 10.90 Hz) 1.30 (m, 2H; CH2) 9.02/−635.2

5MESS(Cl)2 (4)

20

1.29 (m, 2H; CH2) 9.1/−643.1

8.89 (d, 1H: CH, J = 8.50 Hz) 9.11 (d, 1H: CH, J = 8.50 Hz) 8.17 (dd, 1H: CH, J = 5.83, 7.85 Hz) 8.25 (dd, 1H: CH, J = 5.60, 8.56 Hz) 8.97 (d, 1H: CH, J = 5.33 Hz) 9.05 (d, 1H: CH, J = 5.42 Hz) – 8.12 (s,1H; CH) 2.89 (s, 3H; CH3) 3.33 (m, 2H; CH2) 2.32 (d, 2H; CH2 J = 11.55 Hz) 1.67 (m, 4H; CH2) –

5MESS(Br)2 (5)

9

8.91 (d, 1H: CH, J = 8.44 Hz) 9.14 (d, 1H: CH, J = 8.72 Hz) 8.20 (dd, 1H: CH, J = 5.26, 8.21 Hz) 8.29 (dd, 1H: CH, J = 2.44, 5.62 Hz) 9.00 (d, 1H: CH, J = 5.58 Hz) 9.09 (d, 1H: CH, J = 5.58 Hz) – 8.12 (s,1H; CH) 2.86 (s, 3H; CH3) 3.34 (m, 2H; CH2) 2.36 (d, 2H; CH2 J = 12.05 Hz) 1.72 (m, 2H; CH2) 1.66 (d, 2H; CH2 J = 11.02 Hz) 1.30 (m, 2H; CH2) 9.10/−649.7

5MESS(I)2 (6)

35

– – 2.77 (s, 6H; CH3) 3.33 (m, 2H; CH2) 2.36 (d, 2H; CH2 J = 12.57 Hz) 1.72 (m, 2H; CH2) 1.66 (d, 2H; CH2 J = 10.71 Hz) 1.30 (m, 2H; CH2) 9.17/−640.7

9.02 (d, 1H: CH, J = 5.42 Hz) –

8.24 (dd, 1H: CH, J = 5.48, 8.56 Hz) –

9.14 (d, 1H: CH, J = 8.42 Hz) –

56MESS(Cl)2 (7)

46

1.30 (m, 2H; CH2) 9.15/−653.5

– – 2.79 (s, 6H; CH3) 3.35 (m, 2H; CH2) 2.33 (d, 2H; CH2 J = 11.53 Hz) 1.67 (m, 4H; CH2) –

8.99 (d, 1H: CH, J = 5.43 Hz) –

8.21 (dd, 1H: CH, J = 5.55, 8.55 Hz) –

9.13 (d, 1H: CH, J = 8.65 Hz) –

56MESS(Br)2 (8)

11

– – 2.67 (s, 6H; CH3) 3.30 (m, 2H; CH2) 2.33 (d, 2H; CH2 J = 12.08 Hz) 1.62(m, 2H; CH2) 1.69 (d, 2H; CH2 J = 9.50 Hz) 1.26 (m, 2H; CH2) 9.1/−969

9.00 (d, 1H: CH, J = 5.47 Hz) –

8.20 (dd, 1H: CH, J = 5.56, 8.56 Hz) –

9.07 (d, 1H: CH, J = 8.58 Hz) –

56MESS(I)2 (9

Table 1 Summary of NMR data of complexes 1–9 showing chemical shift (ppm) with integration, multiplicity and coupling constants. Experiments were performed in D2O, so amine resonances were not observed due to proton exchange.

B.S. McGhie, et al.

Inorganica Chimica Acta 495 (2019) 118964

Inorganica Chimica Acta 495 (2019) 118964

B.S. McGhie, et al.

5

209 (−6.29), 251 (1.63) 289 (281 ± 2.0), 211 (119 ± 1.3)

289 (226 ± 4.9), 210 (122 ± 1.5)

289 (238 ± 2.9), 211 (103 ± 1.5)

282 (336 ± 2.3), 209 (153 ± 1.3)

282 (165 ± 1.6), 228 (113 ± 1.3)

283 (188 ± 4.2), 210 (80 ± 2.4)

277 (376 ± 1.8), 208 (159 ± 0.7)

277 (308 ± 2.9), 207 (168 ± 1.1)

278 (579 ± 6.9), 208 (243 ± 3.3)

3.93 (37.47) 3.47 (3.39) 3.68 (3.72) 4.50 (4.27) 3.68 (3.57) 3.68 (3.57) 4.69 (4.61) 3.86 (3.70) 4.33 (4.51) 33.40 (33.36) 28.59 (28.85) 32.17 (32.76) 31.81 (31.94) 26.53 (26.83) 26.53 (26.83) 32.84 (32.71) 30.63 (30.26) 32.55 (32.53) 385.49 (385.89) 10.9 C20H26I2N4Pt

337.51 (337.07) 45.9 C20H26Br2N4Pt

293.56 (294.07) 34.7 C20H26Cl2N4Pt

9.16 C19H24I2N4Pt

56MESS(Cl)2 (7) 56MESS(Br)2 (8) 56MESS(I)2 (9)

19.9 C19H24Br2N4Pt

378.48(377.81)

52.9 C19H24Cl2N4Pt

330.50 (332.99)

16.3 C18H22I2N4Pt

286.55 (285.54)

21.2 C18H22Br2N4Pt

371.48 (372.10)

30.4 C18H22Cl2N4Pt

PHENSS(Cl)2 (1) PHENSS(Br)2 (2) PHENSS(I)2 (3) 5MESS(Cl)2 (4) 5MESS(Br)2 (5) 5MESS(I)2 (6)

323.49 (323.35)

C

Yield (%)

Table 2 Summary of the characterisation data of complexes 1–9.

The initial method used to synthesise [Pt(1,10-phenanthroline) (1S,2S-diaminocyclohexane)Cl2]2+ was a modification of the procedure described previously [35]. While successful, this method required the in situ generation of the halide gas so that it could be passed through a Pt (II) compound suspension. Safety concerns made the N-halogensuccinimide reaction preferable. In the presence of ethanol, N-halogensuccinimides (NXS), produced halogen radicals that bound platinum and oxidized it to Pt(IV). In this reaction, approximately half of the NXS reacted with ethanol to produce HX (where X is Br, Cl, or I) which then reacted with the remaining NXS to produce halogen radicals. These radicals subsequently attacked the platinum(II), and oxidised it to Pt (IV) to produce [Pt(1,10-phenanthroline)(1S,2S-diaminocyclohexane) X2]2+ (Scheme 1). However, the introduction of the succinimide resulted in side products where one of the axial ligands was succinimide bound through the nitrogen. The presence of this side product was confirmed both by NMR and ESIMS. The side and intended products had the same solubility and could not be separated by filtration, precipitation or column chromatography. In order to eliminate the formation of PHENSS(X)(succinimide), the method

Molecular Formula

4.1. Synthesis and characterisation

Complex

4. Results and discussion

ESI-MS (m/z) [M−Cl]+ Calc. (Found)

Microanalysis Calc. (Found)

Cytotoxicity assay studies were performed at Calvary Mater Newcastle Hospital, Waratah, NSW Australia. In vitro studies were performed according to described methods [34]. Complexes 1–9 were prepared in DMSO as stock treatment (30 mM) solutions and stored at −20 °C. All cell lines were cultured in a humidified atmosphere with 5% CO2 at 37 °C and maintained in Dulbecco’s modified eagle’s medium (DMEM; Trace Biosciences, Australia) supplemented with 10% fetal bovine serum, sodium bicarbonate (10 mM), penicillin (100 IU mL−1), streptomycin (100 μg mL−1), and L-glutamine (4 mM). The non-cancer MCF10A cell line was cultured in DMEM.F12 (1:1) cell culture media (Composition in SI). Cytotoxicity was determined by plating cells in duplicate in 100 μL medium at a density of 2500–4000 cells per well in 96-well plates. After 24 h, when cells were in logarithmic growth, media (100 μL) with or without the test agent was added to each well (Day 0). After 72 h of exposure, growth inhibitory effects were evaluated by MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay, and absorbance was read at 540 nm. An eight-point dose-response curve was produced, from which the drug concentration at which cell growth is inhibited by 50% (GI50) was calculated. These calculations were based on the difference between the optical density values on day 0 and those at the end of drug exposure.

N

3.8. Assessing cytotoxicity

H

UV/λmax (nm) (ε/mol−1.dm3.cm−1) × 102

Microelemental analysis (C, H and N) was performed at the Chemical Analysis Facility, Department of Chemistry and Biomolecular Sciences, Macquarie University. An Elemental Analyser, Model PE2400 CHNS/O produced by PerkinElmer, USA, was used.

8.40 (8.40) 7.41 (7.45) 8.53 (8.55) 7.81 (7.72) 6.51 (6.41) 6.51 (6.41) 7.66 (7.52) 7.14 (6.98) 7.91 (7.43)

3.7. Elemental analysis

279.54 (280.03)

CD/λma (nm) (Θ) × 10−8

Experiments were performed at the AU-CD beamline on ASTRID2 [32,33] at ISA, Aarhus University. ASTRID2 operates in top-up mode with a current of 120 mA. The AU-CD beam-line operates in the wavelength range of 125–450 nm, with a bandwidth of 0.6 nm. The beam size on the sample is 2 (vert.) × 6 mm (horz.), with a sample to detector distance of 25 mm. All SRCD experiments were performed using a suprasil quartz cuvette with a 100 μm path-length. The SRCD data was processed using OriginPro8.5, and spectra were smoothed using 11 point smoothing.

268 (3.85), 216 (−25.5), 183 (−1.08) 279 (7.14), 226 (−20.2), 183 (−1.04) 270 (3.19), 217 (−24.3), 181 (−1.38) 255 (0.93), 215 (−3.60), 185 (−0.45) 282 (−0.12), 229 (−6.80), 185 (−2.31) 282 (2.19), 217 (−23.9), 304 (−1.18) 290 (4.66), 255 (8.56), 214 (−3.94) 292 (6.10), 224 (−20.8)

3.6. Synchrotron radiation circular dichroism (SRCD)

Inorganica Chimica Acta 495 (2019) 118964

B.S. McGhie, et al.

Fig. 4. The 1H–195Pt HMQC spectrum of PHENSS(Cl)2 (1) in D2O, displaying the correlations between the platinum centre and the protons from each ligand.

combination of 1H proton NMR spectra and 1H-195Pt heteronuclear multiple quantum correlation (HMQC) spectra (Figs. 4 and 5). The HMQC 195Pt peak was significantly different to other polypyridyl platinum complexes [18]. Typically a Pt(IV) complex of the type [Pt(PL) (AL)(OH)2]2+ will demonstrate a 195Pt resonance at 450 ppm, the complexes synthesised in this work had a 195Pt resonance at ~ 630 ppm (Supplementary Figs. A1, 3, 5, 7, 9, 11, 13, 15, 17). This indicated that the halide axial ligands were present [18]. An example of the HMQC spectra is shown in Fig. 4, the 195Pt chemical shift of −645.8 ppm is significantly different to that of the two starting Pt complexes, [Pt(SS-dach)Cl2] and K2PtCl4 (-3282 and −1650 ppm, respectively). The correlation between the Pt centre and the aromatic resonance (9.19 ppm) confirms the coordination of the 1,10-phenanthroline. The proton chemical shifts were almost identical to similar complexes in the literature, with very minor chemical shifts occurring due to stacking of the phen ligands in solution [36]. The aromatic region for PHENSS(Cl)2 (1) was assigned using the following rationale (Fig. 4): the singlet at 8.33 ppm was assigned to H5 and H6. The H5 and H6 protons shared the peak location due to the symmetry of the complex and the absence of protons on the adjacent carbons (Fig. 5). The doublet of doublets (dd) at 8.27 ppm was assigned to H3 and H8 as the presence of protons on the adjacent carbons resulted in a dd splitting pattern due to coupling. The remaining two doublet (d) peaks in the aromatic region were merged in the spectra. They could still be distinguished however, due to the smaller coupling constant of H2 and H9 resulting from the proximity of the nitrogen. The 9.03 ppm doublet J = 8.31 Hz was assigned to H4 and H7 and the doublet at 9.09 ppm, J = 5.56 Hz, was assigned to H2 and H9. The resonances for the aliphatic region were consistent with NMR spectra of the same ancillary ligands reported in the literature [18]. The amine proton resonances were not visible due to exchange with D2O. The NMR characterization of complexes 2–9 was achieved using the same rationale. There were some minor differences in the resonances for these complexes, particularly in the aromatic region where the relative concentration of each solution affected the stacking of the 1,10phenanthroline region and, thus, the resonance of the H2,9 and H4,7 peaks. The splitting in the aromatic region was significantly different in complexes 4–6 due to the asymmetry of those complexes. The aliphatic region for complexes 4–9 had an additional peak due to the presence of

was modified with the addition of heat and acid. PHENSSXsuccinimide was synthesised by refluxing 1 mol equ. of PHENSS with 2.4 mol equ. of NXS in an ethanol and water solution (50:50, v:v) at 78 °C. NMR was used to confirm purity (Supplementary Figs. A17) and, the mass was confirmed by ESI-MS (Table 2). Interestingly, the addition of hydrochloric acid pushed the reaction to favour the production of the dihalogen product. Presumably, this was due to an interaction between the succinimide byproduct and the acid that prevented the coordination of the succinimide. Application of this method resulted in the successful synthesis of complexes 1–9 (Fig. 3, Spectra A1–16 in Supplementary data). NMR also revealed the presence of unbound succinimide in high quantities. The products and the succinimide could not be separated by filtration, precipitation or using a reverse phase C18 column. However, as free succinimide is slightly soluble in toluene the crude mixture was washed 5 times with toluene, reducing the amount of succinimide in the aqueous layer. A soxhlet reflux was used for a minimum of two weeks to remove the remaining succinimide, at which point the succinimide became undetectable. In some cases, the soxhlet reflux produced other small unidentified impurities and these were removed using flash chromatography. The soxhlet reflux and the flash chromatography contributed to product loss generating a yield of 9–45%. 4.2. Biophysical characterisation Each complex was characterised using a combination of NMR, CD and, UV spectroscopy. The NMR spectra produced peaks consistent with those seen in the literature for similar compounds with little to no impurities detected (Supplementary Figs. A1–18) [7]. The CD spectra confirmed that the chirality of the starting materials was retained during synthesis (Supplementary Figs. C1–15). Additionally, the synchrotron spectra revealed dramatic differences from similar, previously published Pt(IV) complexes (Supplementary Figs. D1–21). UV spectra was used to determine the extinction coefficient of each complex. ESI-MS allowed the correct mass peak to be identified for each sample, despite the appearance of some break down products. Table 1 shows a summary of the exact masses used and yields for compounds 1–9. Table 2 summarizes the ESI-MS, EA, extinction coefficient and SRCD data. Break down products were also observed in the HPLC so this technique was not used to characterise the complexes. The NMR characterization of complex 1 was achieved using a 6

Inorganica Chimica Acta 495 (2019) 118964

B.S. McGhie, et al.

Fig. 5. The 1H NMR spectrum of PHENSS(Cl)2 (1) in D2O, showing proton assignment.

the methyl group on the 1,10-phenanthroline but again, the remaining peaks were assigned using the rationale described above. The relative 195 Pt resonances of these complexes appear to be more dependent on the halide ligand than the variations in the PL. For example, bromido complexes tended to have a much lower ppm of ~-960 ppm while iodido complexes tended to have the highest 195Pt peak at ~-640 ppm. The results of the NMR spectra are summarised in Table 1. The UV absorption spectra of the nine complexes (minus a water blank) varied significantly, and the axial ligands had a more pronounced effect on this result than the polypyridyl ligands. Very little difference is observable between the band shapes at ~290 nm for each complex whereas for the bands below 250 nm the spectra were quite distinct (Supplementary Figs. B1–15). At these wavelengths, the spectra of complexes with chlorido or iodido axial ligands were similar to the spectra of complexes with hydroxido axial ligands [6]. Complexes with bromido axial ligands were easily distinguishable with characteristic broad bands with a shoulder between 220 and 250 nm. The transitions

of the phen analogue at 280–290 nm shifted further (blue) as the number of methyl groups increased (i.e. PHENSS(X)2 was the most bathochromic (red) shifted, 5MESS(X)2 was hypsochromic (blue) shifted and 56MESS(X)2 was the most hypsochromic shifted (Supplementary Figs. B3). A similar trend was observed for the weak bands > 300 nm. The 56MESS(X)2 compounds also displayed a shoulder at approximately 240 nm which was not present in the 5MESS (X)2 and PHENSS(X)2 spectra. However, this has also been observed previously with other polypyridyl complexes. Extinction coefficients were determined from seven data points; the averages and errors are reported in Table 2 and the spectra are available in Supplementary Figs. B9–15. The benchtop CD spectra of all nine complexes was compared to those of previously published PHENSS(OH)2 complexes [18]. The spectra of dihydroxido complexes are close to flat until ~270 nm, at which point there is a sharp drop off that continues for the remainder of the spectrum (Supplementary Figs. C1–15). While the spectra of

Fig. 6. SRCD spectra of 7 (black), 8 (red) and, 9 (blue); at RT in the 170–400 nm range, using a 100 µm cell, corrected for solvent baseline. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 7

Inorganica Chimica Acta 495 (2019) 118964

0.23 ± 0.0058 0.27 ± 0.023 0.24 ± 0.059 0.25 ± 0.18 0.044 ± 0.012 0.046 ± 0.010 0.037 ± 0.0067 0.035 ± 0.013 0.032 ± 0.017 0.030 ± 0.004 0.063 ± 0.016 1.0 ± 0.1 0.16 ± 0.0 9.2 ± 2.9

0.33 ± 0.0088 0.36 ± 0.029 0.32 ± 0.031 0.089 ± 0.018 0.070 ± 0.0091 0.065 ± 0.0061 0.048 ± 0.012 0.032 ± 0.0020 0.027 ± 0.0032 0.037 ± 0.009 0.053 ± 0.010 0.9 ± 0.2 1.6 ± 0.1 14 ± 1.0

0.44 ± 0.045 0.68 ± 0.023 0.44 ± 0.021 0.13 ± 0.028 0.11 ± 0.015 0.10 ± 0.0123 0.062 ± 0.013 0.061 ± 0.011 0.037 ± 0.0054 0.051 ± 0.021 0.10 ± 0.015 2.4 ± 0.3 4.1 ± 0.5 24.3 ± 2.2

0.11 ± 0.033 0.18 ± 0.010 0.11 ± 0.0033 0.023 ± 0.0030 0.025 ± 0.0053 0.027 ± 0.0027 0.012 ± 0.0017 0.011 ± 0.0031 0.025 ± 0.017 0.007 ± 0.002 0.009 ± 0.003 1.2 ± 0.1 2.9 ± 0.4 14.7 ± 1.2

0.44 ± 0.050 0.52 ± 0.025 0.37 ± 0.00000 0.25 ± 0.060 0.20 ± 0.010 0.20 ± 0.00000 0.12 ± 0.00000 0.34 ± 0.18 0.087 ± 0.063 0.10 ± 0.016 0.32 ± 0.061 1.9 ± 0.2 0.9 ± 0.2 18.7 ± 1.2

0.34 ± 0.063 0.42 ± 0.071 0.31 ± 0.031 0.18 ± 0.034 0.16 ± 0.044 0.16 ± 0.040 0.092 ± 0.039 0.074 ± 0.033 0.067 ± 0.028 0.074 ± 0.018 0.11 ± 0.009 0.4 ± 0.1 3.0 ± 1.2 5.7 ± 0.2

0.21 ± 0.029 0.25 ± 0.010 0.20 ± 0.013 0.044 ± 0.0045 0.037 ± 0.0046 0.032 ± 0.0022 0.028 ± 0.0021 0.024 ± 0.0026 0.022 ± 0.0022 0.015 ± 0.002 0.027 ± 0.002 7.5 ± 1.3 0.9 ± 0.2 > 50

0.30 ± 0.0033 0.36 ± 0.032 0.29 ± 0.032 0.094 ± 0.024 0.062 ± 0.0083 0.061 ± 0.0082 0.044 ± 0.0062 0.034 ± 0.0038 0.030 ± 0.0039 0.020 ± 0.005 0.030 ± 0.003 nd nd nd

0.23 ± 0.026 0.28 ± 0.017 0.25 ± 0.020 0.056 ± 0.0032 0.048 ± 0.0046 0.044 ± 0.0058 0.036 ± 0.0041 0.033 ± 0.0023 0.027 ± 0.0007 nd nd nd nd nd

dihalido complexes had a similar sharp drop off, they appeared to reach a minimum and start to rise before 200 nm. This suggested that further peaks would appear at shorter wavelengths. It was thus hypothesised that unlike the dihydroxido complexes, the SRCD spectra of these dihalido species would show additional characteristic peaks < 200 nm. The SRCD spectra of complexes 1–3 reveal the similarity in polarised light absorption between 1 and 3; the primary difference was the slightly more intense band at 216 nm for 1. The equivalent peak for 2 was red shifted, and appeared at 226 nm. The second peak at ~180 nm was similar for the compounds 1–3, although 1 had a large shoulder at 212 nm (Supplementary Fig. D1–3). For complexes 4–6 (Supplementary Fig. D4–6), the intensity of the (wavelength) band increased as the size of the ligand increased and as the electronegativity decreased; complex 6 had the most intense band followed by 5, and then 4. It can be hypothesised that either the size of the ligand or the electronegativity of the halide affect the conformation of the complex such that CD signal is altered. The spectra of 6 had the most intense absorption band as well as the sharpest, indicating that the iodido ligands had a stronger effect on the structure than the other two ligands. Differences in SRCD spectra were also observed for complexes 7–9 (Fig. 6 and Supplementary Fig. D1–3). The negative absorption band at 225 nm for 8 was much more intense than the blue shifted bands of 7 and 9. Complex 7 appears to have two peaks more than that of 8 with additional peaks at 184 and 302 nm. The opposite is true of 9 with only one true peak which is blue shifted compared to the equivalent peaks of the bromido and chlorido complexes. This indicated that, overall, this compound absorbed equivalent amounts of left and right polarised light. As noted earlier, the intensity of the peaks for complexes 7–9 did not follow a trend that correlated to either size or electronegativity and, thus, more complex factors are at play. When the SRCD spectra of chlorido complexes 1, 4 and 7 were compared, 1 and 7 were quite similar in terms of their intensity, whereas 4 had featureless-flat spectra (Supplementary Fig. D8). As mentioned previously, these spectral differences were most likely due to the asymmetry of 5MESS. Although the spectra for 7 and 1 are very similar, the bands between 275 and 325 nm form negative and positive maxima, respectively. The remaining peaks of 7 were red shifted compared to the peaks of 1 (Supplementary Fig. D8 & D9). Interestingly, the Cl2 complexes had the least similar SRCD spectra, which was unexpected as the chlorido ligand is the smallest halogen assessed here (Supplementary Fig. C4). The spectra of bromido complexes had 2–3 visible peaks; the peaks of 8 were the most red-shifted followed by 5 and, lastly, 2 (Supplementary Fig. C5). The iodide complexes, 3, 6 and 9, had the most similar spectra of any of the halogen complexes (Supplementary Fig. C6). In the instance of the 56MESS(X)2 complex, unlike the spectra of the dibromido and dichlorido complexes, the diiodido spectra was almost featureless. 4.3. Cytotoxicity The complexes 7–9, together with Pt(II) and Pt(IV) dihydroxido analogues (56MeSS and 56MESS(OH)2), cisplatin, carboplatin and oxaliplatin, were screened against 10 human cancer cell lines representative of colon (HT29), glioblastoma (U87 and SJ-G2), breast (MCF-7), ovarian (A2780), lung (H460), skin (A431), prostate (Du145), neuroblastoma (BE2-C), pancreas (MIA) along with normal breast (MCF10A) and cisplatin resistant ovarian (ADDP) cells. Cytotoxicity was evaluated by means of the MTT test after 72 h of treatment. The results, expressed as GI50 values calculated from dose-survival curves, are reported in Table 3. Complexes 1–9 have sub-micromolar GI50 values against all 10 cell lines making them significantly more potent than cisplatin, carboplatin and oxaliplatin. The average GI50 of 1–9 in the ten cancer cell lines was 0.35, 0.42, 0.33, 0.14, 0.10, 0.10, 0.06, 0.08, 0.04 μM, respectively compared to 56MESS (0.05 μM) and 56MESS(OH)2 (0.10 μM). Notably, the average GI50 value for 9 was 87, 40 and 348 fold more potent than

0.11 ± 0.028 0.16 ± 0.015 0.10 ± 0.0033 0.032 ± 0.0036 0.035 ± 0.0058 0.032 ± 0.0035 0.025 ± 0.0020 0.021 ± 0.0023 0.019 ± 0.0032 0.076 ± 0.061 0.022 ± 0.004 11.3 ± 1.9 0.9 ± 0.2 > 50 PHENSS(Cl)2 (1) PHENSS(Br)2 (2) PHENSS(I)2 (3) 5MESS(Cl)2 (4) 5MESS(Br)2 (5) 5MESS(I)2 (6) 56MESS(Cl)2 (7) 56MESS(Br)2 (8) 56MESS(I)2 (9) 56MESS 56MESS(OH)2 Cisplatin Carboplatin Oxaliplatin

0.81 ± 0.15 0.82 ± 0.090 0.70 ± 0.055 0.23 ± 0.033 0.20 ± 0.029 0.22 ± 0.030 0.12 ± 0.018 0.090 ± 0.012 0.074 ± 0.014 0.076 ± 0.014 0.14 ± 0.023 3.8 ± 1.1 1.8 ± 0.2 > 50

0.46 ± 0.069 0.53 ± 0.10 0.46 ± 0.10 0.22 ± 0.13 0.087 ± 0.032 0.091 ± 0.026 0.060 ± 0.010 0.11 ± 0.056 0.033 ± 0.0068 0.050 ± 0.020 0.14 ± 0.000 6.5 ± 0.8 0.5 ± 0.1 > 50

H460 Lung n = 3–4 A2780 Ovarian n = 3–4 MCF-7 Breast n = 3–4 U87 Glioblastoma n = 3–4 HT29 Colon n = 3–4 Complex

Table 3 Summary of cytotoxicity results in GI50 = Concentration (µM) that inhibits cell growth by 50%.

A431 Skin n = 3–4

Du145 Prostate n = 3–4

BE2-C Neuroblastoma n = 3–4

SJ-G2 Glioblastoma n = 3–4

MIA Pancreas n = 3–4

MCF10A Breast (Normal) n = 3–4

ADDP Ovarian n = 3–4

B.S. McGhie, et al.

8

Inorganica Chimica Acta 495 (2019) 118964

B.S. McGhie, et al.

cisplatin (3.69 μM) carboplatin (1.68 μM) and oxaliplatin (14.43 μM) in the human cancer cell lines (Table 4), they are also more active than 56MESS or 56MESS(OH)2. While all the complexes are very potent, the average GI50 values of 56MeSS(X)2 complexes 7–9 are ordered Br > Cl > I. This same ranking is also evident for PhenSS(X)2 while for 5MESS(X)2 the order is Cl > Br = I. Changes in both the polyaromatic and axial ligands affected the cytotoxicity of these complexes. When compared based on polyaromatic ligand, cytotoxicity increased with methylation, such that PHENSS < 5MESS < 56MESS. However, there are some exceptions to this overall trend expressed in individual cell lines; for example, amongst the chloride coordinated complexes against the ovarian (A2780) cell line, 56MESS (0.037 ± 0.0067 µM) was the most cytotoxic, followed by PHENSS (0.23 ± 0.0058 µM) and 5MESS (0.25 ± 0.18 µM). While for [Pt(PL)(AL)(Br)2]2+ complexes, the trend against breast (MCF-7) cell lines from least to most cytotoxic was PHENSS > 56MESS > 5MESS. There was no deviation from the general trend for the [Pt(PL)(AL)(I)2]2+ complexes as changing the axial ligand appeared to have less effect on the relative cytotoxicity. As complexes 7–9 were the most potent, their cytotoxicity was compared with that of 56MESS and 56MESS(OH)2, (Fig. 7A–I) against the various cell lines. Against colon (HT29) cells 56MESS was the least cytotoxic while the GI50 values for 7–9 and 56MESS(OH)2 were

Table 4 Average of GI50 values (Concentration µM) and fold potency against cisplatin, carboplatin and oxaliplatin. Complex Carboplatin PHENSS(Cl)2 (1) PHENSS(Br)2 (2) PHENSS(I)2 (3) 5MESS(Cl)2 (4) 5MESS(Br)2 (5) 5MESS(I)2 (6) 56MESS(Cl)2 (7) 56MESS(Br)2 (8) 56MESS(I)2 (9) 56MESS 56MESS(OH)2 Cisplatin Carboplatin Oxaliplatin

Average* GI50 values (µM) Oxaliplatin 0.35 0.42 0.33 0.14 0.10 0.10 0.06 0.08 0.04 0.05 0.10 3.69 1.68 14.43

Cisplatin

11 9 11 25 38 38 61 46 87 72 38 1 2 0.3

5 4 5 12 17 17 28 21 40 32 17 0.5 1 0.1

41 34 44 100 149 148 239 181 341 280 147 4 9 1

*Average of HT29, U87, MCF-7, A2780, H460, A431, Du145, BE2-C, SJ-G2, MIA cell lines.

Fig. 7. GI50 values of 7, 8, 9, 56MESS, and 56MESS(OH)2 in multiple cell lines: HT29 colon, U87 and SJ-G2 glioblastoma, H460 lung, A431 skin, BE2-C neuroblastoma, MIA pancreas, Du145 prostate, A2780 ovarian, MCF-7 breast and MCF10A breast (normal).

9

Inorganica Chimica Acta 495 (2019) 118964

B.S. McGhie, et al.

comparable (Fig. 7A). In the U87 glioblastoma cell line the GI50 values of 56MESS(I)2 (0.074 ± 0.014 µM) and 56MESS (0.076 ± 0.014 µM) were comparable and significantly lower than 56MESS(Cl)2 (0.12 ± 0.018 µM) and 56MESS(OH)2 (0.14 ± 0.023 µM), demonstrating the contribution of the axial halides. This contribution is also evident in the SJ-G2 glioblastoma cell line as 56MESS(I)2 (0.067 ± 0.028 µM) was the most cytotoxic of all complexes (Fig. 7B). The impact of axial halides was also evident against lung (H460), skin (A431) and neuroblastoma (BE2-C) (Figs. C, D and E) cell lines where 56MESS(I)2 was more potent than 56MESS. This suggests that despite the theory that 56MESS(I)2 would have reduced to 56MESS within the cell; the dissociated axial ligand and or the reduction potential must in some way be contributing to the observed potency. Prostate and ovarian cancer cell lines were instances where 56MESS(I)2 was not the most potent complex. We evaluated the cytotoxic activity of 7–9, 56MESS and 56MESS (OH)2 against non-tumor human breast MCF10A and cancerous breast MCF-7 cells (Fig. 7H). Although 9 (0.033 ± 0.0068 µM) was more cytotoxic than 56MESS (0.050 ± 0.020 µM), the selectivity index (SI = ratio between average GI50 of non-tumor cells and GI50 of malignant cells) of 9 (1.1) is lower than that of 56MESS (2.5) or 56MESS (OH)2 (7.0). This suggested that while the cytotoxicity of 9 was promising, additional modifications to improve selectivity would maximise the potential of these complexes. Overall when examined based on polyaromatic ligand, there was no identifiable general cytotoxicity trend for the response of individual cell lines to the complexes 1–3. However, against all cell lines, cytotoxicity increased based on axial halide Br > Cl > I (Supplementary Fig. E7). Both ovarian cell lines were the exception to this trend and while the bromido. The group with the most varied trends was that of complexes 4–6; against Du145 prostate cell line, these complexes were ranked I > Cl > Br and this ranking was reversed when assessed against colon (HT29) cells (Br > Cl > I) (Table 3). Against ovarian (A2780), breast (MCF-7), and neuroblastoma (BE2-C) cell lines, the chlorido complexes were the most cytotoxic and the bromido, the least (Br > I > Cl) (Table 3). All data and further figures are provided in the Supplementary Information.

authors also wish to thank Dr Benjamin Pages, Dr Elisé Wright and Mr Dale Ang for constructive editorial suggestions. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ica.2019.118964. References [1] B.W. Stewart, C.P. Wild, World Cancer Report 2014, in: B.W. Stewart, C.P. Wild (Eds.), World Cancer Report, World Health Organisation: Lyon Cedex, France, 2014, pp. 77–619. [2] F. Bray, et al., Global cancer transitions according to the Human Development Index (2008–2030): a population-based study, Lancet Oncol. 13 (8) (2012) 790–801. [3] N.J. Wheate, et al., The status of platinum anticancer drugs in the clinic and in clinical trials, Dalton Trans. 39 (35) (2010) 8113–8127. [4] F. Sitas, et al., Cancer incidence and mortality in people aged less than 75 years: changes in Australia over the period 1987–2007, Cancer Epidemiol. 37 (6) (2013) 780–787. [5] E.E. Calle, et al., Overweight, obesity, and mortality from cancer in a prospectively studied cohort of US adults, New England J. Med. 348 (17) (2003) 1625–1638. [6] Z.H. Siddik, Cisplatin: mode of cytotoxic action and molecular basis of resistance, Oncogene 22 (47) (2003) 7265–7269. [7] T.W. Hambley, The influence of structure on the activity and toxicity of Pt anticancer drugs, Coord. Chem. Rev. 166 (1997) 181–223. [8] M. Kartalou, J.M. Essigmann, Mechanisms of resistance to cisplatin, Mutation Res./ Fundamental Mole. Mech. Mutagen. 478 (1–2) (2001) 23–43. [9] A. Shikanov, et al., Poly (sebacic acid-co-ricinoleic acid) biodegradable carrier for paclitaxel: In vitro release and in vivo toxicity, J. Biomed. Mater. Res. Part A 69 (1) (2004) 47–54. [10] N.J. Wheate, et al., Novel platinum (II)-based anticancer complexes and molecular hosts as their drug delivery vehicles, Dalton Trans. 43 (2007) 5055–5064. [11] M. Pavelka, M.F.A. Lucas, N. Russo, On the hydrolysis mechanism of the secondgeneration anticancer drug carboplatin, Chem.-A Eur. J. 13 (36) (2007) 10108–10116. [12] L. Kelland, The resurgence of platinum-based cancer chemotherapy, Nat. Rev. Cancer 7 (8) (2007) 573–584. [13] M. Treskes, W.J. van der Vijgh, WR2721 as a modulator of cisplatin-and carboplatin-induced side effects in comparison with other chemoprotective agents: a molecular approach, Cancer Chemother. Pharmacol. 33 (2) (1993) 93–106. [14] E. Wong, C.M. Giandomenico, Current status of platinum-based antitumor drugs, Chem. Rev. 99 (9) (1999) 2451–2466. [15] S. Dasari, P.B. Tchounwou, Cisplatin in cancer therapy: molecular mechanisms of action, Eur. J. Pharmacol. 740 (2014) 364–378. [16] M.A. Fuertes, et al., Cisplatin biochemical mechanism of action: from cytotoxicity to induction of cell death through interconnections between apoptotic and necrotic pathways, Curr. Med. Chem. 10 (3) (2003) 257–266. [17] M. Kartalou, J.M. Essigmann, Recognition of cisplatin adducts by cellular proteins, Mutat. Res-Fund. Mol. M. 478 (1) (2001) 1–21. [18] F.J. Macias, et al., Synthesis and analysis of the structure, diffusion and cytotoxicity of heterocyclic Platinum(IV) complexes, Chem. Eur. J. 21 (47) (2015) 16990–17001. [19] B.J. Pages, K.B. Garbutcheon-Singh, J.R. Aldrich-Wright, Platinum intercalators of DNA as anticancer agents, Eur. J. Inorg. Chem. 2017 (12) (2017) 1613–1624. [20] V.V. Kostjukov, et al., Calculation of the electrostatic charges and energies for intercalation of aromatic drug molecules with DNA, Int. J. Quantum Chem. 111 (3) (2011) 711–721. [21] S. Wang, et al., Identification of the molecular mechanisms underlying the cytotoxic action of a potent platinum metallointercalator, J. Chem. Biol. 5 (2) (2012) 51–61. [22] K.B. Garbutcheon-Singh, et al., Cytotoxic platinum(ii) intercalators that incorporate 1R,2R-diaminocyclopentane, Dalton Trans. 42 (4) (2013) 918–926. [23] M.D. Hall, T.W. Hambley, Platinum(IV) antitumour compounds: their bioinorganic chemistry, Coord. Chem. Rev. 232 (1–2) (2002) 49–67. [24] F.J. Macias, et al., Synthesis and analysis of the structure, diffusion and cytotoxicity of heterocyclic platinum(IV) complexes, Chem. – A Eur. J. 21 (47) (2015) 16990–17001. [25] B.W.J. Harper, et al., Synthesis, characterization and in vitro and in vivo anticancer activity of Pt(iv) derivatives of [Pt(1S,2S-DACH)(5,6-dimethyl-1,10-phenanthroline)], Dalton Trans. 46 (21) (2017) 7005–7019. [26] E. Wexselblatt, D. Gibson, What do we know about the reduction of Pt(IV) prodrugs? J. Inorg. Biochem. 117 (2012) 220–229. [27] Y.-R. Zheng, et al., Pt(IV) prodrugs designed to bind non-covalently to human serum albumin for drug delivery, J. Am. Chem. Soc. 136 (24) (2014) 8790–8798. [28] M.D. Hall, et al., Basis for design and development of platinum(IV) anticancer complexes, J. Med. Chem. 50 (15) (2007) 3403–3411. [29] Z. Xu, et al., Halogenated PtIV complexes from N-halosuccinimide oxidation of PtII antitumor drugs: synthesis, mechanistic investigation, and cytotoxicity, Eur. J. Inorg. Chem. 2017 (12) (2017) 1706–1712. [30] B. McGhie, N-halogen succinimides: alternative oxidants of platinum anticancer agents, Aust. J. Chem. 71 (5) (2018) 397–398. [31] N.J. Wheate, et al., Novel platinum(ii)-based anticancer complexes and molecular hosts as their drug delivery vehicles, Dalton Trans. 43 (2007) 5055–5064.

5. Conclusion Nine novel Pt(IV) polyaromatic complexes were synthesised and characterised, using techniques that produced Pt(IV) complexes using efficient and rapid symmetrical oxidation. The cytotoxicity of these compounds was determined against several cell lines and their potency was significantly improved compared to conventional anticancer agents. In most cell lines, the cytotoxicity of the compounds described here was also better than previously synthesised Pt(IV) polypyridyl complexes. Additionally, not only are these complexes potential chemotherapeutic options themselves, they also offer access to new synthetic pathways as [Pt(AL)(PL)(X)2]2+ can be used as intermediates to create targeted Pt(IV) species. The influence of the difference in redox potential between these and other Pt(IV) complexes must also be explored. The novel and unknown mechanism and efficacy of these compounds warrants further investigation against cancers that are difficult to treat and have been associated with increased tumorigenicity and poor prognosis, like KRAS mutated cell lines. The techniques and compounds described in this work will continue to add to the tools at our disposal for effective and inventive cancer treatment. Acknowledgements We thank Western Sydney University for providing financial support through internal research grants. Collection of SRCD data was possible through granting of beam time from the ISA, Department of Physics & Astronomy, Aarhus University and an International Synchrotron Access Program from the Australian synchrotron, AS/IA182/14195. The 10

Inorganica Chimica Acta 495 (2019) 118964

B.S. McGhie, et al. [32] A.J. Miles, et al., Light flux density threshold at which protein denaturation is induced by synchrotron radiation circular dichroism beamlines, J. Synchrotr. Radiat. 15 (4) (2008) 420–422. [33] A.J. Miles, et al., Synchrotron radiation circular dichroism (SRCD) spectroscopy: new beamlines and new applications in biology, J. Spectr. 21 (5–6) (2007) 245–255. [34] M. Tarleton, et al., Library synthesis and cytotoxicity of a family of 2-

phenylacrylonitriles and discovery of an estrogen dependent breast cancer lead compound, Med. Chem. Comm. 2 (1) (2011) 31–37. [35] A. Syamal, R.C. Johnson, Solvent effects in platinum(II)-catalyzed substitution reactions of platinum(IV) complexes, Inorg. Chem. 9 (2) (1970) 265–268. [36] A.M. Krause-Heuer, et al., Diffusion-based studies on the self-stacking and nanorod formation of platinum(ii) intercalators, Chem. Commun. 10 (2009) 1210–1212.

11