Influence of Ligand and Nuclearity on the Cytotoxicity of Cyclometallated C^N^C Platinum(II) Complexes

Abstract A series of cyclometallated mono‐ and di‐nuclear platinum(II) complexes and the parent organic ligand, 2,6‐diphenylpyridine 1 (HC^N^CH), have been synthesized and characterized. This library of compounds includes [(C^N^C)PtII(L)] (L=dimethylsulfoxide (DMSO) 2 and triphenylphosphine (PPh3) 3) and [((C^N^C)PtII)2(L‘)] (where L‘=N‐heterocycles (pyrazine (pyr) 4, 4,4‘‐bipyridine (4,4‘‐bipy) 5 or diphosphine (1,4‐bis(diphenylphosphino)butane (dppb) 6). Their cytotoxicity was assessed against four cancerous cell lines and one normal cell line, with results highlighting significantly increased antiproliferative activity for the dinuclear complexes (4–6), when compared to the mononucleated species (2 and 3). Complex 6 is the most promising candidate, displaying very high selectivity towards cancerous cells, with selectivity index (SI) values >29.5 (A2780) and >11.2 (A2780cisR), and outperforming cisplatin by >4‐fold and >18‐fold, respectively.


Introduction
Since the approval of the chemotherapeutic drug cisplatin, by the FDAi n1 978, platinum(II)-based complexes have become integral in the clinicalt reatmento farange of different cancers. [1] However,c linical platinum(II) anticancer drugs have several drawbacks associated with their use, including intrinsic and acquired resistance, lack of selectivity and neurotoxic side effects. [2] Consequently,t his has generated great interesti nt he development of alternative platinum(II) complexes whichh ave the potentialt oa ddress the significant disadvantages linked to current clinicalplatinum complexes. [3,4] Cyclometallated platinum(II) complexesh ave emerged as attractive alternatives to existing clinical antiproliferativep lati-num drugs, and several compounds were reported to possess moderate to potent cytotoxicity against cancerc ell lines that are resistant to current platinum(II) anticancerd rugs. [5,6] To date, cyclometallatedp latinum(II) compounds bearing ad iverse range of tridentate organic p-ligands scaffolds including, C^N^S, [8] C^N^N, [9] N^C^N [10] and N^N^N [11] have been reported to display up to sub-micromolar potencya gainst ar ange of human cell lines for example, breasta denocarcinoma( MCF-7) and colon carcinoma (HCT116). [5] Early reports from Lowe and co-workerso utlined the potent cytotoxicity of cyclometallated platinum(II)-complexes incorporating aNN^N terpyridine (terpy) ligand scaffold, and ar ange of N-heterocycles and thiolates in the fourth coordination site, against ap anel of human cancer cell lines including two human ovarian carcinomas, A2780 and A2780cisR. [12] Following on from this initial work, severalg roups have demonstrated the significant antiproliferative properties of platinum(II)-terpy systems,w ith reportso ft he biological activity rivalling that of cisplatin in ad iverse array of human cancer cell lines. [13][14][15][16][17][18] However,t he cytotoxicity of cyclometallated, C^N^C platinum(II) compounds has not been widely studied. [19,20] Klein and co-workers showed that as eries of cyclometallated complexesb ased on [(C^N^C)Pt II (L)],w herein C^N^C is at ridentate dianionicc yclometallating motif bearing ar ange of aryl groups;i ncluding phenyl, naphthyl and dibenzoacridine derivativesa nd L = DMSO or acetonitrile (MeCN), displayedg ood to moderate cytotoxicity againsth uman cell lines;c olorectala denocarcinoma (HT-29) and breast adenocarcinoma (MDA-MB-231). [21] Polynuclearp latinum(II) complexesr epresent an important growingc lass of anticancera gentsw ith potentialc linical significance. [22][23][24][25] Developing an understandingo ft he structure-activity relationships (SARs) within classes of biologically active complexesi si ntegralf or optimization of their per-formance. [12,[26][27][28][29] Che and co-workers probed the importance of nuclearity on the SAR of as eries of cyclometallated, tridentate C^N^Np latinum(II) complexes, establishing that dinuclear speciesdisplay more than one order of magnitude higher cytotoxicityt han their monomeric analogues against five human carcinomac ell lines. [15,30] Che and co-workers also established the significance of linker size on the antiproliferativea ctivity of ar elated series of [((C^N^N)Pt) 2 (m-NHC)] + platinum(II) complexes where HC^N^N = 6-phenyl-2,2'-bipyridyl and m-NHC = a bridging N-heterocyclicc arbene ligand. They showed that throughi ncreasing the length of the linker between the two metal centres, up to a2 -fold increase in cytotoxicity was induced against ap anel of human cell lines,i ncluding cervix epithelioid adenocarcinoma (HeLa), liver hepatocellular carcinoma (HepG2) and nasopharyngeal carcinoma (SUNE1). [24] In the same study,C he and co-workersi dentified that analogous mononuclear complexes, [(C^N^N)Pt(NHC)] + displayed significantly higher antiproliferativea ctivity,w ith nanomolar to submicromolar potency (IC 50 = 0.057-1.3 mm), compared to the dinuclear [((C^N^N)Pt) 2 (m-NHC)] + platinum(II) complexes, (IC 50 = 3.9-9.4 mm)a gainst three tested cancer cell lines (HeLa, HepG2 and SUNE1). [24] Lowe and co-workers investigated the influence of linker rigidity on the cytotoxicity of as eries of binuclear [((N^N^N)Pt) 2 (bipy)] (where N^N^N = terpy and bipy is ar ange of pyridine substituted derivatives connected by different linker groups)a gainst several humano varian cell lines including those resistantt oc isplatin and doxorubicin;C H1,C H1cis, CH1dox, A2780 and A2780cisR. Generally, they found that the presenceo fs hort rigid linkers, for example, 4,4'-bipyridine (4,4'-bipy), between the metal centres generatedc omplexes with potent cytotoxicity against the tested cancerc ell lines including nanomolar IC 50 values against the doxorubicin resistant ovarian cell line CH1dox. [12a] It is evident from the studies by Che and Lowe that the cytotoxicity of dinuclear cyclometallated platinum(II) complexesi sh ighly dependent on severalf actors, includingt he backboneo ft he structure and the nature of the linker ligand. However,d espite the increasing importance of cyclometallated platinum(II)-based complexes as potential anticancer therapeutics, reports on systematic studies to elucidate SARs for this class of compounds are limited.

Results and Discussion
Design and synthesis of library of compounds  Figure 1). Compounds 1-6 were prepared using known or modified literaturep rocedures. [31][32][33][34] Compounds 1-5 were characterized by 1 HNMR spectroscopy,m asss pectrometry, FTIR spectroscopy and elemental analysis. Compound 1 was additionally characterized by 13 C{ 1 H} NMR spectroscopy. [31][32][33][34] Compounds 2 and 3 were furtherc haracterized by single crystal X-ray diffraction and both complexes were confirmed to crystallize in the same crystal space group as the published literature structures. [31,34] Additionally, 6 was characterized by 1 H, 13 C{ 1 H} and 31 P{ 1 H} NMR spectroscopy,m ass spectrometry,m elting point analysis, FTIR spectroscopy and single crystal X-ray diffraction. As inglet resonance is observed in the 31 P{ 1 H} NMR spectrum of 6 at d 20.3 ppm (J PPt = 4960Hz) and is due to the two phosphines coupling with the platinum centresi nt his symmetric molecule.
Bright yellow crystals of 6 suitable for X-ray diffractionw ere obtained throught he slow evaporation of ac oncentrated chloroform solution at room temperature. The complex crystallizes in am onoclinic crystal system and solution refinement was performed in the space group P2 1 /c (Table S3 in Supporting Information).T he molecular structure of 6 is shown in Figure 2, with displacemente llipsoids placed at the 50 %p robability level and hydrogen atoms omitted for clarity.A ll of the bond lengths and angles are representative of a pseudo square-planar geometry expected of d 8 Pt II complex (Tables S4  and S5). [31] In the crystal packing of 6,t here are intermolecular, edge-to-face, p-p stacking interactions present between the phenylr ings of dppb and aromaticso ft he C^N^Cp incer ligand ( Figure S1). [35] Chemoselectivitys tudies Ligand 1,c omplexes 2-6,c isplatin (CDDP), oxaliplatin (OXA) and carboplatin (CARB)w ere all screened for their cytotoxicity against human cell lines:c isplatin-sensitive ovarian carcinoma (A2780), cisplatin-resistant ovarian carcinoma( A2780cisR) and breast adenocarcinomas (MCF-7 and MDA-MB-231). The IC 50 values were obtained via the MTT assay after a9 6hincubation period of each compound with the cells at 37 8Ca nd 5% CO 2 (Table 1a nd Figure 3). Ligand 1 was found to be moderate to non-toxica gainst all cell lines,w ith IC 50 values of 41 AE 2 mm to > 100 mm.T he mononuclear platinum(II) complexes, 2 and 3, have significant differences in their cytotoxicity.T he potencyo f complex 2 increases by up to 10-foldagainst MCF-7 in comparison to ligand 1 (p < 0.05), and exhibits similar potency( IC 50 = 4.4 AE 0.5 mm)t oCDDP (IC 50 = 3.07 AE 0.02 mm)a gainst MDA-MB-231 (p = 0.01). Importantly, 2 is the only compound in this library which has noticeable toxicity towards MCF-7 (IC 50 = 10 AE 1 mm), however,i tr emains > 6-fold less cytotoxic than CDDP (p < 0.05). Ar ecent reportb yK lein andc o-workers determined that 2 has aI C 50 value against MDA-MB-231o f1 2.12 AE 1.84 mm, which is within experimental error. [21] Furthermore, 2 is more active towards the cisplatin-resistant ovarian cell line A2780cisR than the parental cisplatin-sensitive A2780 cells. On replacing the monodentate DMSO ligand in 2 for PPh 3 in 3, the activity significantly decreases, and 3 is non-toxic against all cell lines tested (IC 50 > 100 mm)( Ta ble 1a nd Figure 3).
Complex 4 was designed to study the cytotoxicity effects of ad inuclears ystem with the short pyrazine linker.T he cytotoxicity of 4 increases by up to 27-folda nd 5-fold against A2780 (p < 0.05), in comparison to mononuclearc omplexes 2 and 3,   respectively.U pon extending the linker to 4,4'-bipy (5) [29] Whilst Lowe and co-workersr eported that the same complex had IC 50 values of > 25 mm against A2780 and 9.6 mm against A2780cisR. [12a] The differing IC 50 values of 5 and the [((N^N^N)Pt) 2 (4,4'-bipy)] (N^N^N = terpy) analogue, highlights that the nature of the pincer ligand is also an important factor on determining the cytotoxicity of thesec yclometallated platinum(II) complexes.
On modification of the linker to af lexible butyl substituent, in the dinuclear diphosphine complex 6,t he cytotoxicity against MDA-MB-231d ecreases even further,w ith > 4-fold decrease when compared to 5 and > 52-fold decrease when compared to 4.M oreover,i ncreasing the nuclearity of the platinum(II)phosphine compounds from mononuclearity in 3 to dinuclearity in 6 resultsi nsignificantly increased cytotoxicity values against the ovarian carcinoma cell lines, with up to > 29-fold and > 11-fold increases observed against A2780 and A2780cisR, respectively.A st he IC 50 valueso f3 against both ovarian cancerc ell lines, A2780a nd A2780cisR, is greater than the tested threshold concentrationof100 mm,this observed increase in cytotoxicity between these two platinum(II)-phosphine ligandsc ould be greatert han reported here. As for 3, complex 6 is non-toxic against MCF-7 and MDA-MB-231 (IC 50 > 100 mm). On analysis of these results, no definitive SARs can be determined,h owever, some important structuralf eatures can be highlighted;i )the nature of the ligand plays as ignificant role in determining the cytotoxicity of the mononuclear complexes, with the presenceo ft he sulfoxidel igand in 2 increasing the potencyi na ll test cancer cellsl ines by up to 25-foldc ompared to the analogous phosphine-platinum(II) complex 3 (e.g., 2 versus 3 against MDA-MB-231), ii)the addition of as econd platinum centre and linker unit increases the potencyo ft he compounds by up to 29-fold (e.g. 3 versus 6 against A2780), iii)the shorter pyrazine linker in complex 4 is the optimal linker of the N-heterocyclic dinuclearspeciesexhibiting up to sub-micromolar potencya gainst MDA-MB-231( IC 50 = 1.9 AE 0.5 mm). A notable result is that of complex 6,w hich is non-toxic towards either of the two human breast cancerc ell lines (MCF-7 and MDA-MD-231, IC 50 > 100 mm), has significantly increased cytotoxicitya gainst the human ovarian cancer cell lines, > 29-fold against A2780 (IC 50 = 3.4 AE 0.4 mm, p < 0.05) and > 11-fold against A2780cisR (IC 50 = 8.9 AE 0.5 mm, p < 0.05).

Selectivity index (SI)
Due to the current issues of the clinicalp latinum compounds (CDDP, OXA and CARB), which exhibit high potency towards normalc ell types, this library of compounds was screened againstn ormalp rostate cell line, PNT2 (Table 1). As expected, the results highlight that the clinical platinum drugs are high to moderately cytotoxic towards this cell line, with IC 50 values, 1.3 AE 0.2 mm (OXA), 8.5 AE 0.4 mm (CDDP)a nd 27 AE 2 mm (CARB). Ligand 1 and complex 3 are non-toxic towards PNT2 (IC 50 > 100 mm). Complexes 2, 4 and 5 all remain moderately cytotoxic against the normalc ell line, and even though complex 4 is > 4-fold more cytotoxic against MDA-MB-231 than PNT2, it remains relatively cytotoxic against normal cells (IC 50 = 8.3 AE 0.6 mm). The most promisingr esult is observed for 6,w hich remains non-toxic against PNT2 (IC 50 > 100 mm)y et is cytotoxic towardshuman ovarian carcinomas.
The selectivity index (SI) values were calculated for all compounds, using the IC 50 values obtained against PNT2 and dividing by the IC 50 value against the cancerc ell line (Table 1a nd Figure4). Whereby aS Iv alue > 1i ndicates increases selectivity for the cancerous cell line over the normal cell line. Generally, compounds 1-4 do not have increased selectivityf or A2780, A2780cisR and MCF-7, however,t here are some notable increasesi ns electivity for the hormone independentb reast adenocarcinoma cell line, MDA-MB-231, where complexes 2 and 4 have SI of 4.0 and 4.4, respectively.C omplex 5 displays aS Io f 9.2 and 2.0 against the ovarian carcinoma celll ines, A2780 and A2780cisR respectively,w hich outperforms CARB by 5.8-fold (A2780) and 6.7-fold (A2780cisR) (Table1). Unlike 4,c omplex 5 does not have increased selectivity towards MDA-MB-231. The most promising result is observed for complex 6 against the ovarian cell lines, with SI values of > 29.5 and > 11.2 for A2780 and A2780cisR, respectively.T hese results are also minimum SI values, as the IC 50 value against PNT2 is greater than the tested threshold concentration of 100 mm,a nd so the SI could be greater than reported here. Cisplatin has ah ighS Ia gainst A2780 (SI = 6.4), however, 6 exhibits aS Iv alue of 29.5, which is 4-fold highert han that of CDDP (p < 0.05). Importantly,t he SI valueso bserved for complex 6 against A2780cisR show a higherd egree of selectivity than the clinical platinum compounds, with SI values > 18-fold( CDDP and OXA)a nd > 42fold (CARB).

Resistance and sensitivity factors
As many cancers become resistance to drugs, including the clinicalp latinum drugs, there is an urgent need to address the issues of drug resistance, by designingn ew ande ffective drugs. To address the potential of compounds 1-6 to target the ovarian cisplatin-resistance cell line A2780cisR, the IC 50 values were compared with those from the ovarian cisplatinsensitivec ell line A2780.R esistance factors (RF)were obtained by dividing the IC 50 values of A2780 by A2780cisR,w here values > 1i ndicate ap reference for the cisplatin resistant cell line A2780cisR. The RF values could not be calculated for ligand 1 and complex 3,a sI C 50 values are > 100 mm against both ovarian cell lines. As with the clinicalp latinum drugs, complexes 4-6 are all more cytotoxic towards A2780. However, complex 2 has as lightly highers electivity for A2780cisR, with a RF value of 1.4( Figure S2).
To address the activity of compounds 1-6 in comparison with the clinicald rugs, the sensitivityf actors (SF) were calculated by dividing the IC 50 valueso ft he clinical drugs by the IC 50 values of compounds 1-6.A nS F> 1i ndicates as electivity for our library of compounds over the clinicald rugs ( Figure 5). When comparing the performance of our libraryofcompounds with CDDP ( Figure 5A), complexes 4 and 6 are 1.6x more cytotoxic against MDA-MB-231 (4)a nd A2780cisR (4 and 6)w hilst the rest of the libraryd isplays SF < 1a gainst the other tested cell lines. When the biological performance of 1-6 is compared to OXA (Figure 5B), only complex 4 is more cytotoxic, and is 1.3x more active against MDA-MB-231. The most promising results are observed when comparing the cytotoxicity values of the library with CARB,w herein complexes 2, 4-6 are markedly more cytotoxic than this clinicald rug against ar ange of the tested cell lines ( Figure 5C). In particular,c omplexes 2 and 4 have high selectivity for MDA-MB-231, with IC 50 values 7.4x and 17.0x higher than CARB,r espectively.T riple negative breast cancers (TNBC) are the most complex and aggressive types of breast cancer, [36] and are associated with high metastasis, patient relapse, poor prognosis and low survival rates. Therefore, designing and identifying drugs which are effective against these cancers is essential and urgent. Since severalo f our compounds have promising high activities against MDA-MB-231,t his warrants furtheri nvestigation into their use against TNBC. Complexes 2, 4-6 all have increases activity towards one or more of the ovarian carcinomas, when compared to CARB,w ith significantly higher activity against A2780 displaying SF ranging from 4.3 (5)t o5( 6)a nd, notably,a gainst the cisplatin-resistance ovarian cell line A2780cisR, with SF ranging from 5.4 (5)to1 1.4 (4). Additionally,complex 2 hassignificantly higher activity when compared to CARB against the breast adenocarcinoma MCF-7, displaying an IC 50 value at least 10x higher than this clinicaldrug. [37] To gether,t hese results highlight the significant cytotoxic potential for this library of compounds, especially for the dinuclear complexes 4 and 6.A lthough complex 4 is moderately cytotoxic towards the normal cell line, its increased selectivity towardst he TNBC line MDA-MB-231 is promising and should be investigated in further studies. Complex 6 is non-toxic to-wards the normalc ell line and has the highest cancerc ell selectivity for this library,w arranting further modificationsa nd in-depth in vitro studies. On analysis of these results, it is possible that thesec omplexes have mechanism of actions which differsf rom that of the clinical platinum drugs.A si th as previously been shown that cyclometallated N^N^N platinum(II) cationic complexes form strong intercalations with DNA, with binding constants from 0.8 10 5 m À1 to 2.0 10 7 m À1 , [38,39] it is possible that the compounds reported herein have the potential to exert their potency through as imilarm anner.I no rder to furtherd evelop this library of compounds, it is now neces-sary to obtained sufficient SARs and am ore in-depth in vitro screening, to underpin the cellular uptake and possible mechanisms of action of such diplatinumspecies.
Firstly,c omparing the mononuclear complexes,p latinum(II)sulfoxide 2 and platinum(II)-phosphine 3,t he nature of the monodentate ligand at the fourth coordinations ite of the metal centre strongly influences the antiproliferative activity of the complex. The former displaying significantly increased potency against all tested cancerc ell lines (IC 50 = 4.4-19.7 mm) compared to 3 which is non-toxic (IC 50 > 100 mm).
Secondly,s hortening the length of the linker between the two metal centres in these cyclometallated platinum(II) complexes from pyr in 4 to 4,4'-bipy in 5,h as as ignificant effect on increasing the cytotoxicity of the complex, by up to 11-fold, against three of the tested cancerc ell lines (A2780cisR, MCF-7 and MDA-MB-231).
Thirdly,t here is ag eneral increase in potency observed for the dinuclear complexes 4-6 when compared to the mononuclear complexes 2 and 3.F or the platinum(II)-phosphine complexes 3 and 6,t he latter dinuclearc omplex showeds ignificantly increased cytotoxicity against the ovarian carcinomac ell lines with up to > 29-folda nd > 11-fold increase observed againstA 2780 and A2780cisR, respectively when comparedt o mononuclear complex 3.
Several of the studied complexes are more cytotoxic than the clinical drug, carboplatin (CARB), with cytotoxicity values up to 7.4x (2)a nd 17.0x (4)a gainst the TNBC line, MDA-MB-231. This is important in the development of breast cancer drugs, as this is the most complex and aggressive form of breast cancer. Generally,t hese cyclometallated complexes have increased activity towards ovarian carcinomas (compared to CARB)w ith significantly highera ctivity against the cisplatin-resistanceo varian cell line A2780cisR, with sensitivity factors (SF) ranging from 5.4 (5)t o1 1.4 (4). Notably,c omplex 6 was nontoxic towards the normalcell line (IC 50 > 100 mm)a nd has selectivity index (SI) values > 29 against A2780 and is up to 42-fold more selective than current clinicalp latinum drugs.
We have identified the dinuclearp latinum(II)-phosphine complex 6 as the lead candidate of the studied library,a si tr emains non-toxic towards the normal cell line, and our future work will now be aimed at underpinning the specific mechanisms of action of this complexes. We anticipate that the insights gained from this systematic study will continue to help inform the future design of novel cyclometallated diplatinum(II) complexes with high in vitro potential.

Experimental Section
General experimental details:A ll NMR spectroscopy was carried out on aB ruker Advance 400 FT NMR spectrometer using the residual solvent as the internal standard. All the chemical shifts (d) are quoted in ppm and coupling constants are given in Hz and are rounded to 0.1 Hz. Melting points were obtained on GallenKamp and are uncorrected. Single-crystal X-ray diffraction data on 2, 3 and 6 was collected using aB ruker X8 diffractometer with an APEX II detector and monochromated Mo Ka radiation (l = 0.7107 )a t 173 K. The data was processed using Bruker SAINT,t he structures determined with SHELXT [40] and subsequently refined with SHELXL [41] within the program olex2. [42] Crystal structures were visualized using Mercury. [43] Infrared spectroscopy was carried out on aP erkinElmer 100 FTIR instrument fitted with an ATRd etector. Mass spectrometry was recorded on aW aters Micromass Quattro Ultima quadrupole mass spectrometer at the Bradford Analytical Centre. All the reactions were conducted under nitrogen atmosphere. Potassium carbonate, sodium carbonate, concentrated aqueous ammonia and acetic acid were purchased from Camlab. Dimethylsulfoxide, toluene, methanol, dichloromethane, petroleum ether,d i-ethyl ether and chloroform were purchased from Fisher. Te trakis(triphenylphosphine)palladium(0), 2,6-dibromopyridine, triethylamine, 4-phenylboronic acid, pyrazine, 4,4'-bipyridine, triphenylphosphine, 1,4-bis(diphenylphosphino)butane and potassium tetrachloroplatinate were purchased from Sigma Aldrich. Petrol refers to the fraction of light petroleum ether boiling between 40 and 60 8C. All chemicals were used as received unless otherwise stated. The following abbreviations are employed:4 ,4'-bipy = 4,4'bipyridine, aq. = aqueous, Ar = aromatic, br = broad, calc. = calculated, CARB = carboplatin, CDDP = cisplatin, d = doublet, DMSO = dimethylsulfoxide, dppb = 1,4-bis(diphenylphosphino)butane, eq. = equivalent(s), Et = ethyl, h = hour(s), Hz = Hertz, IC 50 = half maximal inhibitory concentration, IR = infrared, m = multiplet, Me = methyl, m.p. = melting point, MTT = 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, NMR = nuclear magnetic resonance, OXA = oxaplatin, pet. = petroleum, Ph = phenyl, pyr = pyrazine, s = singlet, SD = standard deviation, SI = selectivity index, terpy = terpyridine, TNBC = triple negative breast cancer and t = triplet. Compounds 1-5 were synthesized according to known literature procedures. [31][32][33] Cell culture: In vitro chemosensitivity tests were performed against cisplatin-sensitive human ovarian carcinoma (A2780), cisplatin-resistant human ovarian carcinoma (A2780cis) and human breast adenocarcinomas (MCF-7 and MDA-MB-231). Additionally,g rowth inhibitory effects were also tested against normal prostate cell line, PNT2. All cell lines were provided by the Institute of Cancer Therapeutics, University of Bradford and were routinely maintained as monolayer cultures in RPMI 1640 media supplemented with 10 % foetal calf serum, sodium pyruvate (1 mm)a nd l-glutamine (2 mm). All assays were conducted in 96-well round bottom plates, with control lanes for media and 100 %c ell growth. Cell concentrations of 1 10 4 cells mL À1 were used, and 100 mL(or 100 mLmedia in control lane 1) of cell suspension were incubated for 24 hours at 37 8C and 5% CO 2 prior to drug exposure. Ligand 1,c omplexes 2-6,c isplatin (CDDP), oxaliplatin (OXA)a nd carboplatin (CARB)w ere all freshly dissolved in DMSO to provide 100 mm stock solutions, which were further diluted with complete media to provide a range of final concentrations. After 24 hours incubation, 100 mLo f the drug/media solutions were added to the plates in columns 3-12 (100 mLm edia in lanes 1a nd 2f or controls), and then the plates incubated for 96 hours at 37 8Ca nd 5% CO 2 .D rug solutions were added to cells so that the final DMSO concentrations were less than 0.1 %( v/v) in all cases. After 96 hours, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (20 mL, 5mgmL À1 ) was added to each well and incubated for 3h ours at 37 8Ca nd 5% CO 2 .A ll solutions were then removed via pipette and 150 mL DMSO added to each well in order to dissolve the purple formazan crystals. AT hermo Scientific Multiscan EX microplate photometer was used to measure the absorbance of each well at 540 nm. Percentage cell viabilities were determined on al ogarithmic scale, and the half maximal inhibitory concentration (IC 50 )d etermined from a plot of %c ell survival versus concentration (mm). Each of the experiments was performed as duplicate technical repeats and triplicate experimental repeats, with mean values stated as IC 50 AE Standard Deviation (SD).
Statistical analysis:Atwo-tailed ANOVAt-test has been conducted using Graph Pad Prism 8, and used to compare all chemosensitivity data:p robability values p < 0.05 are considered significant.

Synthetic details
1:As olution of 2,6-dibromopyridine (2.0 g, 8.4 mmol) and Pd(PPh 3 ) 4 (0.39 g, 0.34 mmol) in toluene (28 mL) was treated with a solution of Na 2 CO 3 (3.6 g, 34 mmol) in H 2 O( 17 mL). As olution of 4phenylboronic acid (4.1 g, 34 mmol) in methanol (35 mL) was then added to this two-phase system and the reaction mixture stirred at 85 8Cf or 16 h. Upon cooling to room temperature, concentrated aqueous NH 3 (4 mL) and sat. Na 2 CO 3 (40 mL) were added and the aq. phase was extracted with CH 2 Cl 2 (3 30 mL). The organic washings were combined and washed with brine, dried over MgSO 4 and concentrated down under reduced pressure. The resulting residue was subjected to flash column chromatography (2:98; Et 2 O:pet. ether) and crystals were grown through recrystallization from methanol and 2:T oasolution of H 2 L (0.30 g, 1.3 mmol) in acetic acid (55 mL) was added as olution of K 2 PtCl 4 (0.54 g, 1.3 mmol) in H 2 O( 3mL) in a dropwise manner to generate al ight rose coloured solution. The mixture was refluxed for 18 h, after which time, the red colour of the Pt salt had disappeared. The yellow precipitate that had formed was filtered off and washed with H 2 O( 4mL), acetone (4 mL), Et 2 O( 4mL) and pet. ether (2 mL). The yellow solid was then dissolved in DMSO (3 mL) and K 2 CO 3 (0.75 g, 5.4 mmol) and H 2 O( 2mL) were added and the mixture heated to 90 8Cf or 1h. After this time, the reaction mixture was allowed to cool to room temperature and upon the addition of H 2 O( 10 mL), the product precipitated out as ab right yellow solid. This crude product was subjected to flash column chromatography (CH 2 Cl 2 ). Single bright yellow crystals suitable for X-ray diffraction were grown through the slow evaporation of chloroform from as aturated solution of 2. 3:T oasolution of 2 (0.050 g, 0.10 mmol) in CH 2 Cl 2 (2 mL) was added PPh 3 (0.026 g, 0.10 mmol) and the resultant pale-yellow solution was stirred at ambient temperature for 1min. After this time, the excess solution was removed under reduced pressure and the resulting yellow residue was subjected to flash column chromatography (1:1:0.01;C H 2 Cl 2 :pet. ether:NEt 3 )t oy ield ab right yellow solid. Single crystals suitable for X-ray crystallography were grown through the slow diffusiono fd i-ethyl ether into as aturated solution of 3 in chloroform. 4:T oasolution of 2 (0.050 g, 0.10 mmol) in CH 2 Cl 2 (2 mL) was added pyrazine (0.039 g, 0.050 mmol) and the resultant pale-yellow solution was stirred at ambient temperature overnight. After this time, the red precipitate that had formed was filtered off, washed with CH 2 Cl 2 (2 mL) to give the desired product as ab right red solid. 5:T oasolution of 2 (0.050 g, 0.10 mmol) in CH 2 Cl 2 (2 mL) was added 4,4'-bipyridine (0.0080 g, 0.050 mmol) and the resultant pale-yellow solution was stirred at ambient temperature overnight. After this time, the red precipitate that had formed was filtered off, washed with CH 2 Cl 2 (2 mL) to give the desired product as ab right orange solid. 6:T oasolution of 2 (0.050 g, 0.095 mmol) in CH 2 Cl 2 (2 mL) was added dppb (0.021 g, 0.048 mmol) and the resultant pale-yellow solution was stirred at ambient temperature for 1h.A fter this time, the pale-yellow precipitate that had formed was filtered off and yellow crystals suitable for X-ray diffraction were grown through the slow diffusion of pet. ether into as aturated solution of chloroform (0.015 g, 12 %). m.p. 310-311 8C; 1 HNMR (400 MHz, CD 2 Cl 2 ,2 98 K): d 7.70-7.65 (10 H, m, ArH), 7.44 (4 H, dd, 3 J = 8Hz,