Heteropentanuclear Oxalato-Bridged nd–4f (n=4, 5) Metal Complexes with NO Ligand: Synthesis, Crystal Structures, Aqueous Stability and Antiproliferative Activity

A series of heteropentanuclear oxalate-bridged Ru(NO)-Ln (4d–4f) metal complexes of the general formula (nBu4N)5[Ln{RuCl3(μ-ox)(NO)}4], where Ln=Y (2), Gd (3), Tb (4), Dy (5) and ox=oxalate anion, were obtained by treatment of (nBu4N)2[RuCl3(ox)(NO)] (1) with the respective lanthanide salt in 4:1 molar ratio. The compounds were characterized by elemental analysis, IR spectroscopy, electrospray ionization (ESI) mass spectrometry, while 1, 2, and 5 were in addition analyzed by X-ray crystallography, 1 by Ru K-edge XAS and 1 and 2 by 13C NMR spectroscopy. X-ray diffraction showed that in 2 and 5 four complex anions [RuCl3(ox)(NO)]2− are coordinated to YIII and DyIII, respectively, with formation of [Ln{RuCl3(μ-ox)(NO)}4]5− (Ln=Y, Dy). While YIII is eight-coordinate in 2, DyIII is nine-coordinate in 5, with an additional coordination of an EtOH molecule. The negative charge is counterbalanced by five nBu4N+ ions present in the crystal structure. The stability of complexes 2 and 5 in aqueous medium was monitored by UV/Vis spectroscopy. The antiproliferative activity of ruthenium-lanthanide complexes 2–5 were assayed in two human cancer cell lines (HeLa and A549) and in a noncancerous cell line (MRC-5) and compared with those obtained for the previously reported Os(NO)-Ln (5d–4f) analogues (nBu4N)5[Ln{OsCl3(ox)(NO)}4] (Ln=Y (6), Gd (7), Tb (8), Dy (9)). Complexes 2–5 were found to be slightly more active than 1 in inhibiting the proliferation of HeLa and A549 cells, and significantly more cytotoxic than 5d–4f metal complexes 6–9 in terms of IC50 values. The highest antiproliferative activity with IC50 values of 20.0 and 22.4 μM was found for 4 in HeLa and A549 cell lines, respectively. These cytotoxicity results are in accord with the presented ICP-MS data, indicating five- to eightfold greater accumulation of ruthenium versus osmium in human A549 cancer cells.


Introduction
Quite recently we becamei nterestedi nr uthenium and osmium nitrosyl complexes with the prospectt oc reate prodrugs able to release clinically effective levels of NO and metal complex within cancer cells. [1][2][3] This took into account the fact that several classes of, mostly mononuclear,r utheniuma nd osmium coordination compounds have demonstrated promising anticancerp otential both in vitro and in vivo. [4] Moreover, the role of nitric oxide in several biological processes is well es-tablisheda nd depends on its concentration in the cells. [5,6] It is beneficial at low level (< mm)b ut it may causec ell apoptosis at higher concentrationo fN O. [7] In addition, ruthenium nitrosyl complexes are known for their electron-transfer properties and/ or catalytic activity in organic synthesis, which are mainly based on the non-innocent character of the nitrosyl (NO) ligand. [8] These complexes may also photo-releaseN O. [9] Recently,w es ynthesized ruthenium and osmium nitrosylc omplexes with azole heterocycles that were shown to undergo the cis-trans isomerization through ad issociative mechanism. [3] We also evidenced much lower antiproliferative activity of the osmium complexes, in stark contrast to previous studies, where either smaller [10,11] or similarc ytotoxicity [12,13] has been observedf or related ruthenium and osmium complexes.F urther,w einvestigated the effect of incorporating oxygen donors in the coordinations phere of the metal-nitrosyl complexes,s tarting with as eries of osmium complexes with amino acids. [2] Theranostic agents combining at argeted therapeutic drug and ad iagnostic unit that fit the dose requirement for both the therapy and imaging would be ideally suited for cancer treatment allowing monitoring the therapy and response to therapy at the cellular level. [14] Several imaging techniques are well used clinically for cancer diagnosis now,s uch as magnetic resonance imaging (MRI), single photon emissionc omputed tomography (SPECT) or positrone mission tomography (PET). [15] Optical fluorescencei maging possesses an umber of advantages such as high sensitivity,a vailability,e xcellent spatial and very fast temporal resolution allowing visualization and monitoring of the tumor cell biology in real time. [16,17] For in vivo purposes the use of MRI or PETisp referred to overcomep roblems of background fluorescencea nd photobleaching typical for fluorescencei maging, and, high absorption (e.g.,h emoglobin) in the mid-visible range. [18] Exploitation of photophysical properties of lanthanides (luminescence), and, in particular,o f terbium ande uropium, characterized by long-lived (milliseconds timescale) excited states, is another way to avoid concerns relatedt of luorescencei maging. The long lifetimes provide an increase of as ignal-to-noise ratio, since time-resolved fluorescences pectroscopy and microscopy can be used.
With this in mind, we turned our attention to lanthanide-labeling of ruthenium and osmiumc omplexes with biologically active organic ligands. Theu se of luminescence emission from the lanthanide complexes in combinationw ith rutheniumbased anticancer therapeutic drugs is indeed appealing for the design and synthesis of new potentialc ancert heranostics. At the early stage of development lanthanide labeling may allow the monitoring of subcellular distribution of potentiald rugs. In addition, lanthanide(III) salts themselves were found to exhibit moderate antiproliferative activity in vitro [19,20] as well as in vivo. [21] These properties are relatedt ot heir similarity to calci-um ions, whereby lanthanide ions are higherc harged and therefore show as trong affinity towards biological calcium bindings ites. [19,21,22] The anticancer activity could be furtherenhanced by complexation of lanthanide ions with various chelating and macrocyclic ligands such as chrysin (A), [23] texaphyrins (B) [24] or phenantroline derivatives (C). [25] La(phen) 3 (NCS) 3 (phen = 1,10-phenanthroline) was found to overcome drug resistance and has provedt ob eh ighly effective against the DLD-1 colon cancerm odel in vivo, [26][27][28] which is resistant to several chemotherapeutics amongst others due to oncogenem utations. [29] Moreover,a ttempts to generater esistant to La(phen) 3 (NCS) 3 cell models failed in contrast to many other investigated metallodrugs.
Herein we report on the syntheses of lanthanide-labeled ruthenium-nitrosyl complexes as potential anticancer drugs. The lanthanide has been coupled to the ruthenium by ab ridging oxalate, which is aw ell-known bioligand incorporated in the anticancer agent oxaliplatin. [31] The complexeso ft he general formula (nBu 4 N) 5 [Ln{RuCl 3 (ox)(NO)} 4 ], where Ln = Y( 2), Gd (3), Tb (4), Dy (5;S cheme 1), have been characterized by elemental analysis,E SI mass spectrometry and IR spectroscopy and, in case of 2 and 5 by X-ray diffraction. Their antiproliferative activity along with that of the precursor (nBu 4 N)[RuCl 3 (mox)(NO)] (1)h ave been investigated in the two humanc ancer cell lines HeLa (cervical cancer) and A549 (non-small cell lung cancer) and the noncancerousc ell line MRC-5 (lung fibroblasts) and compared with that fort he osmium analogues 6-9 (Scheme 1), which were reportedr ecently. [30] In addition, the effect of Ru versus Os as well as of the different lanthanide ions on the biological activity of the compounds is discussed and comparedt ot hat for the mononuclear ruthenium-nitrosyl complex 1.T he antiproliferative activi-ty of Ru(NO)-Ln (4d-4f) and Os(NO)-Ln (5d-4f) complexes has been correlatedw ith their accumulation in human A549 cancer cells.

Experimental Section
Materials Solvents were obtained from commercial sources and were used as received. The starting compound Na 2 [RuCl 5 (NO)]·6 H 2 Ow as prepared as described in the literature. [32] RuCl 3 ·3 H 2 Ow as purchased from Johnson Matthey and the lanthanide salts were from Sigma-Aldrich. All chemicals were used as received. The respective osmium-lanthanide complexes 6-9 were synthesized as reported recently. [30] Synthesis of complexes (nBu 4 N) 2 [RuCl 3 (ox)(NO)] (1):T oasolution of Na 2 [RuCl 5 (NO)]·6 H 2 O (0.3 g, 0.6 mmol) in water (2.5 mL) as olution of oxalic acid (0.12 g, 1.3 mmol) in water (2.0 mL) was added. The pH value was adjusted to 3u sing an aqueous solution of KOH. The reaction mixture was refluxed for 3hand the pH of the solution was kept at 3. nBu 4 NCl (0.36 g, 1.3 mmol) was added to the hot solution. The oil obtained was separated using as eparation funnel and dissolved in water (20 mL 5 [Y{RuCl 3 (m-ox)(NO)} 4 ]( 2):Y Cl 3 ·3 H 2 O( 13 mg, 0.043 mmol) was added to as olution of 1 (100 mg, 0.13 mmol) in acetonitrile (1.2 mL) and 2-propanol (0.65 mL) and the reaction mixture was refluxed for 1.5 h. The reaction mixture was cooled to room temperature and filtered. The solvent was removed under reduced pressure and the residue was dissolved in ethanol (2.0 mL). The product crystallized upon slow evaporation of the solvent at room temperature. Yield:4 0mg, 31 %; elemental analysis calcd (%) for C 88 H 180 Cl 12

Physical measurements
Elemental analyses were performed by the microanalytical service of the Faculty of Chemistry of the University of Vienna on aP erkin-Elmer 2400 CHN Elemental Analyzer.U V/Vis spectra were recorded at 25 8Cu sing aP erkin-Elmer Lambda 650 spectrometer equipped with an optical cell of 1cmp ath-length in the wavelength range of 200 to 800 nm in combination with aP erkin-Elmer PTP-6 Peltier System. Electrospray ionization (ESI) mass spectrometry measurements were conducted on aB ruker HCT ion trap (Bruker Daltonics GmbH) by using methanol as asolvent. MIR spectra were recorded on aP erkin-Elmer 370 FTIR 2000 instrument using an ATR( attenuated total reflection) unit in the range of 4000-400 cm À1 .P hosphorescence emission spectra were recorded with aH oriba FluoroMax-4 spectrofluorimeter and the data were processed using the FluorEssence v3.5 software package.
X-ray crystallography X-ray diffraction measurements were performed on aB ruker X8 APEXII CCD diffractometer. Single crystals were positioned at 35, 40 and 40 mm from the detector,a nd 950, 1964 and 4113 frames were measured, each for 10, 30 and 10 so ver 18 scan width for 1, 2 and 5,r espectively.T he data were processed using SAINT software. [33] Crystal data, data collection parameters, and structure refinement details are given in Ta ble 1. The structures were solved by direct methods and refined by full-matrix least-squares techniques. Non-hydrogen atoms were refined with anisotropic displacement parameters. Hydrogen atoms were inserted in calculated positions and refined with ar iding model. The following computer programs and hardware were used:s tructure solution, SHELXS-97 and refinement, SHELXL-97; [34] molecular diagrams, ORTEP; [35] computer,I ntel CoreDuo. Disorder observed for tetrabu-tylammonium cation(s) in 2 and 5 was resolved by using SADI and EADP restraints and DFIX constraints implemented in SHELXL. CCDC 951636, 1402056 and 1402057 contain the supplementary crystallographic data for this paper.T hese data can be obtained free of charge from The Cambridge Crystallographic Data Centre.

XAS sample preparation
The Ru model compounds were diluted in BN (boron nitride, Sigma-Aldrich, CAS 10043-11-5, 99.5 %), filled into aluminum sample holders and sealed with Kapton foil. The BN samples were prepared for ac alculated theoretical absorption of about 1a bsorbance unit according to standards methods. [36] XAS data collectiona nd analysis The XAS experiment was carried out at beamline BM26A at the European Synchrotron Radiation Facility (ESRF) in Grenoble (France). [37] At beamline BM26A (ESRF,G renoble;F rance) three low noise ion chambers from Oxford Instruments were used for measurements in transmission mode. The absolute energy calibration was performed using ar uthenium powder (Sigma-Aldrich, CAS 7440-18-8, 99.9 %) BN preparation optimized for an absorption edge jump of 1abs, measured at the same time between ionization chambers two and three. The model compound was measured in transmission mode. An Oxford CCC 1204 cryostat provided as ample environment of 20 K. The ESRF storage ring was operated at 6G eV in the 7/8 + 1f illing mode. The beamline BM26A was equipped with ad ouble crystal Si(111)m onochromator and ab ending magnet source giving an energy range of 5-30 keV (flux of 11011phs À1 ). Higher harmonics were rejected using two mirrors with Pt and Si coatings. The XAS spectrum was measured at the Ru K-edge with ap reedge region from 21 869 to 22 083 eV with as tep size of 10 eV,a n edge region from 22 099 to 22 161 eV with as tep size of 1.3 eV. The k-space was measured from 3t o1 4A À1 with as tep size of 0.05 A À1 .T he scanning times per measurement point were 1s in the pre-edge, 5s in the edge and 5-25 s( 22 161-22 917 eV), increasing according to ap redefined curve, in the post-edge region. The spectrum of the model compound is the average of 2scans.

MTT assay
Antiproliferative activity of the ruthenium and osmium complexes was determined using 3-(4,5-dimethylthiazol-yl)-2,5-diphenyltetrazolium bromide (MTT,S igma-Aldrich) assay. [49] Cells were seeded into 96-well cell culture plates (Thermo Scientific Nunc), at ac ell density of 3000 cells per well (HeLa), 7000 cells per well (A549), and 5000 cells per well (MRC-5), in 100 mLo fc ulture medium. After 24 ho fg rowth, cells were exposed to the serial dilutions of the tested complexes. Complexes were dissolved in 1% DMSO:c omplex 1 at ac oncentration of 4mm,c omplexes 2-5 at ac oncentration of 1mm,c omplexes 6-9 at ac oncentration of 0.6 mm,a s stocks immediately prior use, and afterwards diluted with nutrient medium to desired final concentrations (in range up to 200 mm). Each concentration was tested in triplicate. After incubation periods of 48 h, 20 mLo fM TT solutions (5 mg mL À1 in phosphate buffer solution, pH 7.2) were added to each well. Samples were incubated for 4h at 37 8C, with 5% CO 2 in ah umidified atmosphere. Formazan crystals were dissolved in 100 mLo f1 0% sodium dodecyl sulfate (SDS). Absorbances were recorded after 24 h, on an ELISA reader (ThermoLabsystems Multiskan EX 200-240 V), at the wavelength of 570 nm. The IC 50 values, defined as the concentrations of the compound causing 50 %c ell growth inhibition, were estimated from the dose-response curves.

Inductively coupled plasma mass spectrometry (ICP-MS)
Sample preparation for the measurement of intracellular Ru/Os accumulation using ICP-MS:R u/Os accumulation was analyzed in A549 cells with ICP-MS using Thermo Scientific iCAP Qc ICP-MS (Thermo Scientific, Bremen, Germany). [50] A549 cells were seeded into a2 5cm 2 dish (Thermo Scientific Nunc) and treated with the complexes 4 and 8 at concentrations equal to 0.5 IC 50 .A fter 6 and 24 h, cells were harvested by scraping, washed with ice-cold PBS and collected by centrifugation at 778 gfor 10 min. Sample preparation for the measurement of Ru/Os binding to DNA and proteins using ICP-MS:B inding of Ru/Os to cellular DNA and proteins was analyzed in A549 cells, using ICP-MS. A549 cells were prepared and collected using the same procedure as described above. To tal DNA and protein were isolated using TRI Re-  2 ]/(nÀp)} 1/2 ,w here n is the numbero fr eflections and p is the total number of parametersr efined. agent (Sigma-Aldrich) according to the manufacturer's procedure and concentrations were determined spectrophotometrically by measuring absorbances (Eppendorf BioPhotometer 6131).

Microwave digestion
The digestion of the samples for ICP-MS studies was performed on an advanced microwave digestion system (ETHOS 1, Milestone, Italy) using HPR-1000/10S high pressure segmented rotor.T he pressure-resistant PTFE vessels (volume 100 mL) used in this study consisted of fluoropolymer liner.B efore use, the PTFE vessels were acid cleaned and rinsed with deionized water.T his type of vessel permitted am aximum temperature of 240 8Ca nd am aximum pressure of 100 bar to be applied. Maximally ten PTFE vessels could simultaneously be mounted on the rotor.T he internal temperature was monitored only with one vessel equipped with as ensor unit, and this vessel had as ensor-protecting tube that directly contacted the digested solution, differing from the other common PTFE vessels. In the digestion, samples were mixed in each clean vessel with 4mLH NO 3 (65 %, Suprapure, Merck, Germany) and 4mLu ltrapure water and then heated with microwave energy for 10 min. The temperature was controlled with apredetermined power program. Digestion of the samples was carried out for 10 min at ac onstant temperature of 180 8C, with ap rior warmup linearly over 10 min to 180 8C. After cooling and without filtration, the solution was diluted to af ixed volume into a1 0mLv olumetric flask and made up to volume with ultrapure water.U ltrapure water was prepared by passing doubly deionized water from Milli-Q system (Millipore, Bedford) to aresistivity of 18.2 MW cm.

Instrumental analysis
ICP-MS measurements were performed using Thermo Scientific iCAP Qc ICP-MS (Thermo Scientific, Bremen, Germany) spectrometer with operational software Qtegra. For Ru determination the instrument was adjusted for optimum performance in He KED (kinetic energy discrimination) mode using the supplied autotune protocols. For Os determination the instrument was adjusted for optimum performance standard no gas mode using the supplied autotune protocols. The instrumental operating conditions for ICP-MS are shown in Ta ble 2.
Analytical blanks were run in the same way as the samples, and concentrations were determined using standard solutions prepared in the same acid matrix. The standard for the instrument calibration was prepared on the basis of ruthenium, plasma standard solution, Specpure, Ru 1000 mgmL À1 and osmium, plasma standard solution, Specpure, Os 1000 mgmL À1 certified reference solutions ICP standard purchased from Alfa Aesar GmbH &C oKG( Germany).

Abbreviations
XAS, X-ray absorption spectroscopy;X ANES, X-ray absorption near edge structure;E XAFS, extended X-ray absorption fine structure; ESRF,E uropean Synchrotron Radiation Facility;F T, Fourier transform.

Synthesis and characterization
As mentioned in the Introduction we were interested in labeling our ruthenium and osmium-nitrosyl complexes with lanthanideions. Asuitable strategy to accomplish such acombination wasinspired by aprevious report on aheteropentanuclear oxalate-bridged [Re IV 4 Gd III ]c omplex. [51] Quite recently we reported the synthesiso ft he osmium-nitrosyl analogues 6-9 (Scheme 1). [30] The ruthenium-nitrosyl analogues 2-5 (Scheme 1) were synthesized following as imilarp rocedure. An aqueous solution of Na 2 [RuCl 5 (NO)] was treated with2equiv oxalic acid to give rise to [RuCl 3 (ox)(NO)] 2À ,w hich was isolated as at etrabutylammoniums alt 1 in 47 %y ield by addition of nBu 4 NCl to the reactionm ixture. This complex proved to be suitable for the synthesis of pentanuclear heterometallic complexes 2-5 by treatment with 0.3 equiv of the respective lanthanide(III) or yttrium(III) salt in 4:1m olar ratio either in ethanol or in am ixture of acetonitrile and 2-propanol.A ll these 4d-4fm etal complexesw ere obtained as crystalline solids in 31-51 %y ield. The formationo fp entanuclear assemblies was confirmed by elemental analysis, ESI mass spectrometry and Xray diffraction of complexes 2 and 5.E SI mass spectra of complexes 1, 2,and 5 showedo nly as ignal for the ruthenium fragment [RuCl 3 (NO)(ox)] À (m/z 569) in the negative ion mode, while for 3 and 4 signals with ah igher m/z ratio were observed, which could be attributed to tetranuclear speciesc ontaining the respective lanthanide ion. 13 CNMR spectroscopic measurementso ft he diamagnetic compound 2 revealed no significant influenceo ft he Y III coordination on the chemical shifts of the oxalato carbon atoms.

Photophysical properties
The luminescence of the reported compounds was investigated using complex 4 as an example, since this property is most evident and often studied for europium or terbium ions. Emission spectra (l ex = 365 nm) of aqueous solutionso f4 at different concentrations ranging from 0.1 to 400 mm were measured between 450 and 700 nm using as litw idth of 10 nm. Even at the highest concentration level, only avery weakphosphorescence signal was found after ten flash counts (data not shown). To be able to observe the complete emission pattern typical for at erbium ion consisting of four peaks, up to 200 flash counts had to be applied (FigureS1i nt he Supporting Information). These attempts indicate that complexes of the www.chemeurj.org type 2-5 are not suitablef or furtherd evelopment as potential diagnostic agents.

X-ray diffraction analysis
The results of X-ray diffraction studies of complexes 1, 2 and 5 are shown in Figure 1, Figure 2a nd Figure 3. Note that the Xray diffractions tructure of [RuCl 3 (ox)(NO)] 2À was reported previously as ac esium salt and [NiL] 2 + complex, where L = 1,4,8,11-tetraazacyclotetradecane. [52] In the present work the complex( nBu 4 N) 2 [RuCl 3 (ox)(NO)] (1)w as studied since it served as as tarting materialf or furtherc omplexation reactions with lanthanide salts. In addition, we used this compound for determinationo ft he oxidations tate of ruthenium coordinated to an on-innocent NO ligand (vide infra). In 2 four complex anions [RuCl 3 (ox)(NO)] 2À are coordinated to yttrium(III) via unbound oxygen atoms of the oxalates which act as bridging ligands with formation of the complex [Y{RuCl 3 (m-ox)(NO)} 4 ] 5À . The negative charge is counterbalanced by five nBu 4 N + ions present in the crystal structure. Like (nBu 4 N) 5 [Y{OsCl 3 (mox)(NO)} 4 ] [30] the complex crystallizes in the tetragonal space group P4 2 1 c.T he yttrium atom and the nitrogen atom of one of the tetrabutylammonium cations lie on the fourfold rotation axis running along the c axis. So the asymmetric unit consists of one [RuCl 3 (ox)(NO)] 2À unit bound to Y III in as pecial position, one nBu 4 N + cation in ag eneral position and aq uarter of nBu 4 N + in as pecial position. The whole complex forms as phere the radius of which is of approximately 8.5 . The shortestR u···Ru separation is of 7.469 , while Y···Y distance is of 15.951 .
Selected bond lengths and bond angles in the coordination spheres of Ru and Yare quoted in the legend to Figure 2. The RuÀCl bonds in 2 are well-comparable to those in the precursor 1,w hile the RuÀOb onds are by approximately 0.03-0.04 longer in 2 compared to those in 1.T he RuÀN1 bond in 2 is very similart ot hat in 1,a sa lso are the N1ÀO1 bond and the Ru1-N1-O1 angle in both complexes.
The complex (nBu 4 N) 5 [Dy(EtOH){RuCl 3 (m-ox)(NO)} 4 ]·1.5 H 2 O crystallizes in the monoclinic non-centrosymmetric space group Cc.T he whole structure is severely affected by the disorder,w hich was mainly resolved in an isotropic model.T herefore, ac omparison of the metric parameters in 1 with those in 5 has not much sense. The asymmetricu nit consists of one complexa nion [Dy(EtOH){RuCl 3 (m-ox)(NO)} 4 ] 5À and five nBu 4 N + cations.W hile Y III is eight-coordinate in 2,t he Dy III ion is ninecoordinate in 5 with an additional coordination of an EtOH molecule, as was also found for the corresponding Os-Dy counterpart. [30] Complexes 3 and 4 were found to crystallize in   the monoclinic non-centrosymmetric space group Cc,a nd are affected by severe disorder.T herefore, only the parameters of the unit cells are given in Table S1 in the Supporting Information for the two compounds.I na ddition, crystallization of 4 from aC HCl 3 solution yielded crystalsi sostructural to that of 2 (see Table S1, entry 4' for the respective unit-cellparameters).
NO in complexes 1-5 acts as an on-innocent ligand [53] rendering the description of the exact electronic structure of the Ru(NO) entity difficult. According to Enemark and Felthamn otation [54] and taking into account the diamagnetism of 1,t he close to linearityR u-N-O angle and the IR n(NO) vibration (1842 cm À1 )i tc an be described in our case as {Ru(NO)} 6 .H owever,i td oes not revealt he actual physicala nd formal oxidation state [55] of the ruthenium and NO ligand. Therefore, XANES experiments were performed to determine the physical oxidation state of ruthenium in 1.

XAS analysis
The XAS spectra of the Ru reference compounds have been published recently. [48] The method has been proven highly valuable for determiningt he oxidation state and coordination charge formetal complexes in vivo and in vitro. [56,57] In this study they formed the basis for the oxidation state assignment of ruthenium in 1.T he structural formulas are shown in Figure S2 in the Supporting Information. The prefix "R" has been added to the labeling and numbering of the reference compoundsi sp reserved to avoid overlaps.T he figure includes the followingm odel compounds:i ndazolium trans-[tetrachloridobis(1H-indazole)ruthenate(III)] R1 (with first coordination shell Ru III Cl 4 N 2 ), [58] tris(pentan-2,4-dionato)ruthenium(III) (R3,R u III O 6 ,S igma-Aldrich, CAS 14284-93-6, 97 %), [59] hexammineruthenium(III) trichloride (R4,R u III N 6 ,S igma-Aldrich, CAS 14282-91-8, 99 %), [60] mer,trans-aquatrichloridobis(indazole)ruthenium(III) (R5,R u III Cl 3 N 2 O), [61] trans,trans-dichloridotetrakis-(indazole)ruthenium(III) chloride (R6,R u III Cl 2 N 4 ), [62] mer-trichloridotris(indazole)-ruthenium(III) (R7,R u III Cl 3 N 3 ), [63] hexammineruthenium(II) dichloride (R8,R u II N 6 ,S igma-Aldrich, CAS 15305-72-3, 99.9 %), [64] mer,trans-trichlorido(dimethylsulfide)bis(indazole)-ruthenium(III) (R9,R u III Cl 3 N 2 S), [65] trans,transdichloridotetrakis(indazole)ruthenium(II) (R10,R u II Cl 2 N 4 ), [62] mer-trichloridotris(ethylphenylsulfide)ruthenium(III) (R11, Ru III Cl 3 S 3 ), [66] trans,trans,trans-dichloridobis(dimethylsulfide)bis(indazole)ruthenium(II) (R12,Ru II Cl 2 N 2 S 2 ). [62] The XANES spectra and their corresponding first derivatives for 1, R1 (first shell coordination:R u III Cl 4 N 2 ), R5 (Ru III Cl 3 N 2 O) and R3 (Ru III O 6 )e xhibiting am ixed chloride, nitrogen/oxygen first coordination sphere are shown in Figure 4. The edge position for 1 (RuCl 3 NO 2 )w as determined as 22 125.9 eV.I nc omparison to R7 (Ru 3 + Cl 3 N 3 )a nd R5 (Ru 3 + Cl 3 N 2 O), 1 displays an edge shift of + 2.2 and + 1.9 eV,r espectively.T he edge position for R3 (Ru III O 6 )h as been determined to be 22 126.6 eV and appears 0.7 eV above that for 1.I nt he previous study on ruthenium complexes it was shown that model compounds with the same first shells show an edge shifto fa bout + 2eVo ng oing from Ru II to Ru III . [48] Like models R1,R u III Cl 4 N 2 , R3,R u III O 6 ,a nd hexammine compounds R8,R u II N 6 and R4,R u III N 6 complex 1 (RuCl 3 NO 2 )e xhibits ac haracteristicedge shoulder. [48] In Figure 5t he calculated coordination charges versus the experimental determined Ru K-edge positions are shown. The edge energy of R1 Ru III Cl 4 N 2 in boron nitride (BN) was set as an arbitrary origin. Ar egression line with ac oefficient of determination R 2 = 0.95 could be aligned to the calculated coordination charges and the observed edge positions of Ru II andR u III model compounds with varying absorber-ligande nvironments, thereby proving the linear correlation between the coordination chargea nd the edge positions. [48] The compounds containing Ru II and/or Sa re on the left (lower energy) side, and the compounds containing Ru III ,N ,a nd Oa re on the right (highere nergy) side. The edge position of 22 125.9 eV for (nBu 4 N) 2 [RuCl 3 (ox)(NO)] clearly falls in the range of Ru 3 + compounds with am ixed nitrogen/oxygen/chloride coordination sphere.  The k 3 -weighted EXAFS spectraa nd the Fouriert ransforms (FT) of 1 are shown in Figure 6. For compounds with mixed N/ O/Cl first shells and increasing number of Na nd/or Ol igands as plitting of the first peak in the FT is observed. The backscattering amplitudes of the heavy scatterers like Sa nd Cl and the light scatterers, like Na nd Oa re out of phase. [67] For 1 this cancellation leads to an ode between 8a nd 11 À1 shown in Figure 6( top;b lack curve). The fitting analysisu sing FEFF [41,42] was restricted to the first coordination shell extracted from the first peak in the FT.T he identity and numbers of back-scatterers were fixed not to exceed the number of fitting parameters and the known crystallographic distances were taken as astarting point forthe fitting analysis.
The results of the first shell fit of 1 using theoretical amplitudes and phases provided by the FEFF code are presented in Ta ble 3, as well as the results for the DL-EXCURV fit (Table 4). [47] The distances are given as the average fitted distances for each atom type/shell (Cl, O/N and N/C). The curve fitting results obtainedb yF EFFa nd DL EXCURVE are both in good agreement with the crystallographic values.
The only other Ru-NO compound, the oxidation state of which has been investigated by XANES spectroscopy, is mer,trans-[RuCl 3 (1H-indazole) 2 (NO)]. [32] The authorsc omparedt he edge position of mer,trans-[RuCl 3 (1H-indazole) 2 (NO)] with the one of R5 (Ru 3 + Cl 3 N 2 O), and R10 (Ru 2 + Cl 2 N 4 )a nd concluded the oxidation state of 3.4(3) + for Ru in mer,trans-[RuCl 3 (1Hindazole) 2 (NO)].The oxidationstate of 3.4 + might be slightly overestimated owing to the lack of data for Ru IV reference compounds indicating that the physical oxidation number, [53,55] which is am easurable quantityd erived from ak nown d n configuration of am etal ion, in that case is 3 + (d 5 electron configuration). It differs from the formal oxidation state 4 + or 2 + of ruthenium in this mononuclear complex, which is an on-measurable integerd efined as the chargel eft on the metal after all ligands in [RuCl 3 (1H-indazole) 2 (NO)] have been removed in their normal, closed-shell configuration. [55] In the case of non-innocent ligand NO its closed shell configurations can be represented as NO À or NO + .

Stability in aqueous media
The stabilityo ft he heteropentanuclear complexes 5 and 9 in aqueous solution has been investigated by UV/Vis spectroscopy over 96 h. There was no change in the optical spectra observed ( Figures S3 and S4 in the Supporting Information), which indicates ahigh stability of the complexes under applied conditions.
To exclude immediate dissociation of the pentanucleara ssembly with releaseo fam etal-oxalate fragment an aqueous solution of 5 was evaporated to dryness after standing for 24 h in air and the IR spectra of the residue and freshly prepared compound 5 were compared. There were no differences in the spectra observed.

Cytotoxic activity
The antiproliferativea ctivity of the ruthenium and osmium lanthanidec omplexes 1-9 was evaluated for 48 ho fc ontinuous drug action, using colorimetric MTT assay.T he study was performed in two human neoplastic cell lines (HeLa,A 549), and human fetal lung fibroblast cell line (MRC-5), which was used  [a] N fix is the fixed coordinationn umber, R is the average distance, R cryst is the crystallographicv alue, DR is the difference between R and R cryst , s 2 is the Debye-Waller factor, E 0 is the residual shift of the edge energy. [a] N fix is the fixed coordination number, R is the average distance, R cryst is the crystallographicv alue, DR is the difference between R and R cryst , s 2 is the Debye-Waller factor, E 0 is the residuals hift of the edge energy, R fit is the quality of the fit, fit index is the sum of the square of the residuals. The results showed that all tested ruthenium compounds exhibitedd ose-dependentc ytotoxicity in the range of concentrationsu pt o2 00 mm,b eing up to 10-times more active than their osmium analogues, especially in A549 cells, where osmium compounds did not reach their IC 50 values in the examined range of concentrations. Concentration-effect curves for each cell line are depicted in Figure S5 in the Supporting Information, illustrating the pronouncedd ifferences in activity between complexes containing ruthenium and osmium.T hese differences are of special note, since osmium compounds were reportedt ob ee ither as active as or even more potent than their ruthenium analogues. [10,12,[68][69][70] Exceptions have only been reportedf or ap air of Ru III /Os III tetrazole complexes [13] and as eries of ruthenium and osmium complexes containing azole and NO ligands, [1] where the ruthenium complexes were found to be significantly more active.
Higher IC 50 values obtained after treatment of A549 compared to HeLa cells were expected, because of decreased sensitivity and slower response to treatment of A549 cells. Complex 4 exhibited the highest antiproliferative activity in general, with that against A549 cells in the range of activity obtained in HeLa cells (20.0(AE 2.4) vs. 22.4(AE 3.1) mm,r espectively), whichi s ap romisingr esult.
Cell-types electivityi sa lso noted in the analysis of the effect of tested compounds (ruthenium ando smium analogues) in MRC-5 normalc ell line. While osmium compounds exhibited cytotoxic activity in MRC-5 comparable to cytotoxicity in HeLa cells, ruthenium complexes showed high cytotoxic potential in vitro in MRC-5 cell line, whichm ay be considered as the major drawback in the preliminary studies of these complexes.
Comparison of antitumora ctivity of 2-5 with that of 1 indicates slightly highera ctivity of the former species. Dissociation of 2-5 with releaseofLn III would generate 4equivof1 and approximately af ourfold increaseo fcytotoxicity would be expected. Indirectly,t hesed ata provide further evidence about the stability of the complexes under the conditions used for MTT assays. Enhancementofa ntiproliferativeactivity by coordination of organic biologically active speciest ot he Ln III ion is well-documented in the literature. [71] Intracellular distribution/accumulation of investigated complexes Discovery and development of new metal-baseda nticancer agentsi sl argely based on cell viabilitya ssays (IC 50 values), intracellulara ccumulation and distribution studies. Considering the obtained IC 50 values, complexes 4 and 8 werec hosen for the ICP-MSanalysis in order to investigate intracellular distribution and accumulationofRu/Os in A549 cells.
In particular,w es eparately analyzed metal (Ru/Os) distribution among the DNA and protein fractions, as well as total intracellulara ccumulation, using ICP-MSa nalysis, after 6a nd 24 ht reatment with 0.5 IC 50 of the investigated complexes. Each analyzed metal compound was found in the cells, althoughe xhibiting different levelso fa ccumulation and various affinities for protein and DNA binding. Ruthenium exhibited greater (five-to eightfold) total intracellular accumulation than osmium,w ith time-dependent increaseo fa ccumulation characteristic for both metals( Figure 7A). The results also show that osmium complex 8 was bound to cellular DNA more efficiently than Ru complex following 6h treatment (139.5(AE 9.4) vs. 63.4(AE 1.8) pg metal mg À1 DNA, respectively). With prolonged incubation time, complex 4 inducedm ore DNA binding, while complex 8 exhibited the opposite behavior, although at the lesser extent( Figure 7B). Analyses of metal content in protein cell fractions indicated that after 6htreatment both complexes induce as imilar level of metal bindingt oc ellular proteins, but with the treatment prolongation the level of Os-protein binding decreased twofold, while Ru-protein binding increased to ag reat extent( Figure 7C).
Low Os intracellulara ccumulation compared to Ru, as well as time-dependent decrease of protein/DNAb inding,i ndicate ar eversible nature of interactions of complex 8,w hich all may be ar easonf or its low cytotoxic action. Higher cytotoxicity of 4 comparedt o8 may be attributed to its ability to bind DNA and proteins more efficiently, to increaset he number of interactions with these biomolecules over time, and possibly to form different DNA and protein conformational distortions and lesions. As ar esult of these differences in cellular accumulation and DNA/protein binding, potentiald ifferences in cellular response to Ru/Os treatment arise.
Detailed analysis of the in vitro antitumora ctivity concerning complexes 2-5 showedacytotoxicity enhancement of the synthesized compounds compared to the starting compound (1). Lower activity of the previously reported osmium analogues 6-9 in all tested cell lines leads to the conclusion that the presence of the ruthenium centeri nh eteronuclear nd-4f metal complexes enhances their cytotoxicity.M oreover,t he cytotoxic potential of investigated complexes 2-5 is in accord with af ive-to eightfold greater cellular accumulation of ruthenium comparedt oo smium obtainedf rom ICP-MS investigations of complexes 4 and 8.F urther studies on elucidating mechanisms underlying different anticancer activity andc ellular accumulation of these compoundsa re required in order to ascertain their potential for the development as anticancer agents.