Antiangiogenic and Anticancer Properties of Bifunctional Ruthenium(II)-p-Cymene Complexes: Influence of Pendant Perfluorous Chains.

Article pubs.acs.org/molecularpharmaceutics Antiangiogenic and Anticancer Properties of Bifunctional −p‑ fl Ruthenium(II) Cymene Complexes: In uence of Pendant fl Per uorous Chains Patrycja Nowak-Sliwinska,*,†,‡ Catherine M. Clavel,‡ Emilia Paŭ nescu,‡ Marije T. te Winkel,†,‡ Arjan W. Griffioen,† and Paul J. Dyson*,‡ † Angiogenesis Laboratory, Department of Medical Oncology, VU University Medical Center, Amsterdam, The Netherlands ‡ Institute of Chemical Sciences and Engineering, Swiss Federal Institute of Technology (EPFL), CH-1015, Lausanne, Switzerland ABSTRACT: Two bifunctional ruthenium(II)−p-cymene complexes with perfluorinated side chains, attached via pyridine ligands, have been evaluated in a series of in vitro and in vivo assays. Their effects on human endothelial (ECRF24 and HUVEC) cells, noncancerous human embry- onic kidney (HEK-293) cells, and various human tumor cells were investigated. The complex with the shorter chain, 1, inhibits the proliferation of the tumor cell lines and ECRF24, whereas 2 selectively inhibits ECRF24 and HUVEC proliferation. Neither inhibits the migration of ECRF24 cells whereasbothcompoundsinhibitsproutformationinHUVECcells.Usingthreepreclinicalmodels,i.e.,vasculatureformationin the chorioallantoic membrane (CAM) of the chicken embryo, human A2780 ovarian carcinoma tumors xenografted on the CAM,andhumanLS174Tcolorectaladenocarcinomatumorsgrowninathymicmice,theangiostaticandanticanceractivitiesof these two complexes were studied. Overall, 1 inhibited tumor growth predominantly through an anticancer effect whereas 2 inhibited tumor growth predominately via an antiangiogenic mechanism. KEYWORDS: antiangiogenesis, bioorganometallic chemistry, CAM model, colorectal adenocarcinoma, ovarian carcinoma, ruthenium(II)−arene complexes, fluorine chemistry ■ INTRODUCTION tends to be low and part of their activity may be attributed to interactions with extracellular proteins.20,21 Metal-based anticancer drugs used in the clinic are based on The development of new antiangiogenic drugs, which platinum, and, while there is continued interest in the developmentofnewplatinum-baseddrugs,1−3thereisgrowing actively target the tumor microenvironment, represents an attractivealternativetocytotoxicagentsforcancertherapy,and interest in the antitumor properties of ruthenium-based compounds.4−6 The latter research area has been stimulated c su o c m h b c i o n m at p io o n und w s it h h ave o b th ee e n r a t p r p e l a ie tm d e a n s t mo s n tr o a t t h e e g r i a e p s i . e 2 s 3 2 − 2 2 ,2 6 3o T r h in e bythe development of tworuthenium(III) complexes, namely, identification of the antiangiogenic properties of some KP1019 (indazolium trans-[tetrachlorobis(1H-indazole)- ruthenium complexes12,17,27 encouraged us to pursue more ruthenate(III)])7 and NAMI-A (imidazolium trans- detailed analysis of other compounds that present relevant [tetrachloro(dimethyl sulfoxide)(1H-imidazole)ruthenate- antiproliferative activities. (III)]).8,9 Both compounds are currently being developed in Recently, we reported the synthesis, characterization, and in clinicaltrials.10,11Inaddition,agrowingnumberofruthenium- vitroandinvivoevaluationoftheantiproliferative propertiesof (II) complexes show promising biological properties.12−14 For a series of ruthenium(II)−arene complexes designed to have example, organometallic ruthenium(II) complexes with the thermoactive properties by virtue of a pendant perfluorinated generalformula[Ru(η6-arene)Cl 2 (PTA)](arene=tolueneand chain.28,29 These derivatives are structurally related to the p-cymene, PTA = 1,3,5-triaza-7-phosphaadamantane), named RAPTA class of compounds presenting the same “piano-stool” RAPTA compounds, exhibit antimetastatic15,16 and antiangio- motif, with thePTA ligand beingreplaced byapyridine ligand genic17 properties, as well as some activity on primary human modified with a perfluorinated chain (Figure 1). We have tumor growth.18 Although the mechanism of action of these previously shown that 1 (Figure 1), when evaluated at 37 °C, compounds remains to be fully elucidated, recent studies indicate that RAPTA compounds bind preferably to Received: May27, 2015 proteins.19,20 Such protein targeting of the RAPTA family is Revised: June 28, 2015 markedly different from the mechanism of action of platinum Accepted: July9, 2015 compounds. Moreover, cellular uptake of RAPTA compounds ©XXXXAmericanChemicalSociety A DOI:10.1021/acs.molpharmaceut.5b00417 Mol.PharmaceuticsXXXX,XXX,XXX−XXX Molecular Pharmaceutics Article Proliferation Assay. Endothelial cells (5 × 103 cells/well) were seeded in gelatin-coated 96-well cell culture plates as described previously.50 Briefly, 24 h after seeding, culture medium containing or without the compounds was added and cells were grown for a further 72 h. Cell viability was assessed using the CellTiter-Glo Luminescent Cell Viability Assay Figure 1. Structures of 1 and 2 and RAPTA-C, [Ru(η6-p-cymene)- (Promega, Madison, WI, USA). Cl(PTA)](PTA=1,3,5-triaza-7-phosphaadamantane),usedhereinas Migration (Scratch) Assay. The migration capability of 2 areference compound. endothelial cells was determined using the wound assay.51 In brief,ECRF24weregrowntoconfluenceingelatin-coatedwells i.e., in the absence of a hyperthermia signal, displays a and “scratch wounds” (with an approximate width of 350 μm) remarkable cancer cell selectivity.30 Compound 1 was also were made in the monolayer by removing cells with a sterile found to inhibit A2780 tumor growth in a CAM preclinical scratch tool (Peira Scientific Instruments, Beerse, Belgium). model.28 Moreover, 2, which displays little activity at 37 °C Cellswere washed with PBS, andthe mediumwas replaced by toward all tumor cell types tested, showed thermoactive fresh medium containing 10 ng/mL bFGF (Tebu-Bio) and properties in vitro, i.e., inhibiting cell viability when combined incubated with or without tyrosine kinase inhibitors. Auto- with mild hyperthermia (41.5 °C).30 Based on these mated image acquisition was performed using a Leica observations further studies of the biological properties of 1 DMI3000B microscope in combination with UGR grabbing and 2 were undertaken, and, herein, we show that, at nontoxic software (DCI Laboratories, Peira Scientific Instruments). doses, 1 has direct anticancer properties whereas 2 exerts Bright field images taken at 5× magnification were processed predominately an antiangiogenic effect that leads to effective for computational analysis of scratch sizes using UGR Scratch t■umor growth inhibition. Assay 6.2 software (DCI Laboratories, Peira Scientific Instru- ments). EXPERIMENTAL SECTION Sprouting Assay. HUVEC spheroids were created using Compounds. RAPTA-C, 1, and 2 were prepared as the hanging drop method.17 HUVEC were suspended in previously described.30,47 The compounds were dissolved in medium containing 20% methocel (Sigma-Aldrich, St. Louis, DMSO immediately prior each experiment. USA)ataconcentrationof4×104cells/mL,and25μLdrops Log P Determination. Stock solutions of 1 and 2 was (containing 1000 cells) were aliquoted on the inside of the lid prepared in octanol saturated with PBS. A 1:1 ratio of octanol ofaPetridish.Thelidwassubsequently invertedtocreatethe andPBSwasusedfortheshake-flaskmethodwhileintroducing hanging drops and placed over the PBS containing Petri dish. thestocksolutionvolume.Thebiphasicsolutionwasshakenfor After 24 h the spheroids were harvested and embedded in a 24 h and centrifuged to separate the 2 phases. A colorimetric collagen gel (2 mg/mL) at 20 spheroids per well of a 96-well calibration curve was prepared for each complex in octanol plate.Aftersolidificationofthegel,mediumcontainingthetest saturated with PBS. Absorbance spectra were recorded in a 1 compounds was added and cells were allowed to sprout into cm quartz cuvette in a SpectraMax M5e reader (Molecular the collagen for 16 h. Sprouting spheroids (8 to 10 per Devices). condition) were pictured directly under a microscope (Leica Fluorescence Measurements. HSA and apotransferrin Microsystems GmbH, Wetzlar, Germany) under ×10 magni- were purchased from Sigma and used without further fication. Quantification of sprouting was performed using a purification. The stock solutions of proteins (0.4 mM) were semiautomatic macro based on the ImageJ platform.17 prepared inPBS,stored at 0−4°Cinthe darkfor aweek, and Flow Cytometry. Apoptosis was estimated by flow diluted to 2.4 μM when used. Stock solutions of RAPTA-C, 1, cytometric determination of subdiploid cells after DNA or 2 (1 mM) were prepared in DMSO. Fluorescence extraction and subsequent staining with PI as described measurements were performed using black 96-well plates on previously.13 Briefly, cells were harvested and subsequently a SpectraMax M5e (Molecular Devices). Fluorescence fixed in 70% ethanol at −20 °C. After 2 h the cells were quenching was measured after each successive addition of the resuspended in DNA extraction buffer (45 mM Na HPO , 2.5 2 4 complex stock solution in each measured well containing the mMcitricacidand1%TritonX-100,pH7.4)for20minat37 protein at 280 nm for the excitation wavelength and the °C.PIwasaddedtoafinalconcentrationof20μg/mL,andred emission spectrum recorded in the range of 300−450 nm. log-scale fluorescence was analyzed on a FACSCalibur (BD Cell Culture. The immortalized endothelial ECRF24 cell Biosciences, NJ, USA). Apoptosis was quantified as the line48 was maintained in DMEM/RPMI-1640 (1:1) (Invitro- percentage of subdiploid cells. gen, Carlsbad, CA, USA) supplemented with 10% FCS, 1% Cellular Ruthenium Uptake. Ruthenium cellular uptake penicillin, and streptomycin (BioConcept Ltd., Allschwil, wasperformedasdescribedpreviously.28Briefly,ECRF24cells Switzerland). Human umbilical vein endothelial cells were grown to ca. 70% confluency and incubated with 1 or 2 (HUVEC) were isolated from normal human umbilical cords (50μM)for5h.Attheendoftheincubationperiod,cellswere byperfusionwith0.125%trypsin/EDTAandculturedinRPMI rinsedtwicewith2.0mLofPBS,detachedbyadding0.5mLof culture medium supplemented as above and with 10% human enzyme free cell dissociation solution (Millipore, Switzerland), serum.49 HUVEC and ECRF24 cells were cultured as and collected by centrifugation. The samples were digested in previously described17 in 0.2% gelatin-coated tissue flasks. ICP-MS-grade concentrated hydrochloric acid (Sigma-Aldrich) Human colorectal adenocarcinoma LS174T cells (Cell Line for 3hat roomtemperature andfilledtoa total volumeof8.0 ServiceGmbH,Eppelheim,Germany)wereculturedinDMEM mL with ultrapure water. Indium was added as an internal medium supplemented with 10% fetal calf serum and 1% standardataconcentrationof0.5ppb.Determinationsoftotal antibiotics (as above). All cells were cultured in a highly metal contents were achieved on an Elan DRC II ICP-MS humidified atmosphere with 5% CO at 37 °C. instrument (PerkinElmer, Switzerland) equipped with a 2 B DOI:10.1021/acs.molpharmaceut.5b00417 Mol.PharmaceuticsXXXX,XXX,XXX−XXX Molecular Pharmaceutics Article Meinhardnebulizerandacyclonicspraychamber.TheICP-MS analyses were done using the Student’s t test, or ANOVA. *p- instrument was tuned daily using a solution provided by the ■values <0.05 were considered statistically significant. manufacturercontaining1ppbofeachMg,In,Ce,Ba,Pb,and RESULTS AND DISCUSSION U.Externalstandardswerepreparedgravimetricallyinamatrix identical to that of the samples (with regard to internal Activity of 1 and 2 against Endothelial Cells. standard and hydrochloric acid) with single element standards Compounds 1 and 2 were originally developed as potential (CPI International, The Netherlands). thermoactive drugs to be used in combination with hyper- TheChickenChorioallantoicMembraneModel(CAM) thermia,30however,bothcompoundsshowsomeactivitywhen and Quantification of the Vascular Effects. The appliedundernormalconditions,i.e.,at37°C.Thecytotoxicity antiangiogenic efficacy of 2 was tested in the physiologically of 1 and 2 on A2780, A2780cisR (human ovarian carcinoma developing chicken embryo chorioallantoic membrane (CAM) cisplatin sensitive and resistant, respectively), MCF-7 (human model52 between embryo development day (EDD) 11 and 15. breastadenocarcinoma),MDA-MB-231(invasivehumanbreast Treatments were performed via intravenous administration adenocarcinoma), A549 (human lung carcinoma), and non- (eachtime80μL),betweenEDD11and14.Thecontroleggs cancerousembryonickidneyHEK-293cells(Table1)hasbeen received 80 μL of DMSO diluted in 0.9% NaCl. At EDD 15, the CAMs were visualized in ovo by means of FITC−dextran Table 1. IC values of 1, 2, and RAPTA-C (Used as a 50 (20 kDa, 25 mg/mL, Sigma-Aldrich) epifluorescence angiog- Control)inHumanTumorCells(A2780,A2780cisR,MCF- raphy53 and subsequently analyzed by the image-processing 7, MDA-MB-231, A549, SW480, and LS174T), quantification method described previously.54 Briefly, on the Noncancerous Embryonic Kidney Cells (HEK-293), and basis of FITC−dextran fluorescence angiography the skeleton Human Endothelial Cells (Immortalized ECRF24 Cells and a of the vascular network is built and defined descriptor such as HUVEC) after 72 h Incubation branching points/mm2. 5−7 eggs were tested per condition. compound(μM) Tumor Implantation and Treatment on the CAM. A2780 cells (1 × 106) were prepared as a spheroid in a 25 μL celltypes 1 2 RAPTA-C hangingdropand3hlaterweretransplantedonthesurfaceof IC 50 (μM±SEM) the CAM.55 Vascularized tumors appeared approximately 3 A2780 44.2±0.1b >500b 230c days after inoculation on the surface of the CAM, and the A2780cisR 25.0±2.0b >500b 270c average tumor volume was 1.66 ± 0.09 mm3. At EDD 11 the MCF-7 37.9±1.7b >500b >50031 embryos were randomized and treated daily within the MDA-MB-23 36.4±1.9b >500b >500 A549 43.1±0.7b >500b >500d following groups via intravenous administration: 0.1% DMSO (CTRL),2(50μM;0.56mg/kg,80μL),orthemixtureof1+ SW480 41.4±2.6 >500 170d 2(premixed atratio1:1;25μMeach;combineddose0.5mg/ LS174T >200 >200 >500e kg, 80 μL). Tumor volumes (mm3) were calculated as follows: HEK-293 >500b >500b >500c volume = width2 × length/2. ECRF24 53.4±1.0 55.6±4.2 250e In Vivo Studies in a Xenografted LS174T Mouse HUVEC >100 43.0±5.0 300±8.0e IntracellularUptake(pmol/106cells) Model. Experiments in 8-week-old Swiss female athymic mice ECRF24 148 694 107 purchased from Charles River (Orleans, France) were carried LogP out according to a protocol approved by the Committee for −0.03 0.07 Animal Experiments for the Canton of Vaud, Switzerland aErrorsrepresentthestandarderrorofthemean.Intracellularuptake (license 2772). Mice were injected subcutaneously in the hind inECRF24cells(pMin106cells,5hincubation).bValuestakenfrom leg with 5 × 105 LS174T cells (Cell Line Service GmbH, ref30.cValues takenfromref32.dValues takenfromref33.eValues Eppelheim,Germany),whichwereculturedinDMEMmedium taken fromref 34. supplemented with 10% heat-inactivated fetal calf serum and 1% antibiotics. When tumors reached a size of 5−6 mm in previously reported.30 Here, we extended the study of 1 and 2 diameter, mice received 12.5 mg/kg/day of 2 (300 μL) by ip to SW480 (human colorectal adenocarcinoma) and LS174T injections. This treatment was repeated following the q4d3 (colorectal adenocarcinoma) cells, and also investigated their schedule, i.e., a total of 3 doses at 4-day intervals. Tumor antiproliferative activity against immortalized human endothe- dimensions and body weight were measured daily. Tumor lial cells (ECRF24), and freshly isolated human umbilical cord volumes(mm3)were calculated asfollows: volume=width2× endothelial cells (HUVEC) in order to assess potential length/2. Mice were euthanized when the control tumor size antiangiogenic properties, see Table 1. reached 1000 mm3. Tumors were resected and fixed for TheIC valuesof1rangefrom25μMinA2780cisRcellsto 50 immunohistochemical analysis. 44.2 μM in A2780 cells, and it is slightly less cytotoxic to Immunohistochemistry.5μmsectionswereblockedwith ECRF24 cells (IC 53.4 μM) and inactive on HEK-293 and 50 5%BSAinPBSfollowedbyincubationwithprimaryantibodies HUVEC cells. Complex 2 exhibits a markedly different against CD31 (1:200; clone SZ31, Dianova, Hamburg, behavior, being moderately active only on endothelial cells, Germany). CD31 stainings were performed using donkey with IC values of 55.6 μM and 43 μM for ECRF24 and anti-rat biotinylated secondary antibodies (Jackson, Suffolk, HUVEC 50 cells, respectively. Inaddition, intracellular ruthenium U.K.) and streptavidin-HRP (Dako, Glostrup, Denmark), and uptake of 1 and 2 was determined in ECRF24 cells (Table 1), visualized by 3,3′-diaminobenzidine (DAB). since 1 and 2 are equally cytotoxic to this cell line (53.4 and StatisticalAnalysis.Allvaluesaregivenasthemeanvalues 55.6 μM, respectively). Low internalization of ruthenium was ±SDorSEM.Dataarerepresentedasaveragesofindependent determined for 1 (148 pM), slightly higher than that of experiments, performed in duplicate or triplicate. Statistical RAPTA-C (107 pM), whereas an approximately 5-fold higher C DOI:10.1021/acs.molpharmaceut.5b00417 Mol.PharmaceuticsXXXX,XXX,XXX−XXX Molecular Pharmaceutics Article Figure2.MigrationinhibitionofECRF24(A)andMDA-MB-231(B)cellsafterexposureto1or2.WoundclosureinECRF24andMDA-MB-231 cultures after incubation of 1 (6−25 μM) and 2 (6−25 μM). DMSO 0.1% in DMEM culture medium and sunitinib (sut, 20 μM) were used as controls. **p <0.01 (for sunitinib). Error bars represent standard error of the mean. n= 6. Figure3.Effectof1and2onHUVECsproutformation.(A)Representativeimageofasproutingspheroidafterincubationwith2(50μM)for24h. Quantificationofsproutnumber(B)andtotalsproutlength/spheroid(C)followingexposureto1and2expressedasapercentageofthecontrol. Errors represent the standard error of the mean, **p <0.01. level of ruthenium internalization was determined for the Table 2. Cell-Cycle Changes in ECRF24 Cells Treated with a ECRF24cellstreatedwith2(694pM),i.e.,thecompoundwith 1 or 2 the longer hydrophobic chain. For a comparison, uptake of 1 compound(μM) G1/G0(±SEM) G2/M(±SEM) apoptosis(±SEM) (pmol/106 cells) in A2780 and HEK293 cells after a 24 h incubation (at a dose of 250 μM) is almost an order of CTRL 63.0±0.2 31.0±0.6 3.0±0.1 magnitude higher.14 1(6) 66.3±3.9 26.0±5.4 4.6±1.1 1(12) 68.0±0.0 22.0±0.0 6.1±0.3 In a 2D migration assay onECRF24 or MDA-MB-231 cells, 1(25) 61.5±0.7 27.0±0.0 8.0±1.3 1 and 2 did not inhibit the mobility of cells up to 1(50) 51.8±0.3 23.5±3.5 15.9±3.0 concentrations of 25 μM (Figures 2A and 2B). 2(6) 67.0±2.8 27.0±3.5 3.6±0.1 The activity of 1 and 2 on endothelial cell sprout formation 2(12) 64.0±2.8 29.5±2.1 2.4±1.3 was assessed using a bFGF-driven sprouting assay with 2(25) 62.5±2.8 28.5±0.7 5.4±0.6 HUVEC.17 Interestingly, incubation of HUVEC spheroids 2(50) 67.0±3.5 29.5±3.5 6.9±1.2 with1or2at50μMresultedinsignificantinhibitionofsprout aChanges to the cell-cycle were evaluated by FACS analysis of formation (p < 0.01, Figure 3). A representative image of a propidium iodide stained cells after 24 h incubation with 1 or 2 at sproutingHUVEC spheroid is provided in Figure 3A,together concentrations ranging from6to 50μM,n= 4.Errors represent the with image-based quantification of sprouting for the control standard error of the mean. cells and cells treated with 1 and 2. An inhibition of sprouting by 45% for 1 and 48% for 2 in the number of sprouts (**p < previously,1inducesapoptosisinMDA-MB-231cells,28witha 3.3-fold increase (at 50 μM) in apoptosis versus the control. 0.01, Figure 3B) was determined, and the mean sum of the sproutlengths/spheroidwasinhibitedby60%and61%(**p< Interactionsof1and2withSerumProteins.Following intravenous administration ruthenium compounds have been 0.01, Figure 3C) for 1 and 2, respectively. Since neither shown to bind to serum proteins, in particular albumin and compoundinhibitsmigration,thereductionofsproutinginEC transferrin, which are believed to play a crucial role in drug is presumably a result of the effect of 1 and 2 to inhibit transportandmetabolism.35−42Thebindingpropertiesof1,2, proliferation. and RAPTA-C with human serum albumin (HSA) and Complexes 1 and 2 induce apoptosis in ECRF24 cells in a apotransferrin were assessed using a fluorescence spectroscopy dose-dependent manner (Table 2), determined by FACS quenching assay (Table 3). HSA fluorescence results from analysis of subdiploid cells after exposure to a 50 μM dose. A tryptophan residues, one in the case of HSA (Trp214) and six 5.3-foldand3.4-foldincreaseinapoptoticcellscomparedtothe in the case of apotransferrin. All complexes led to a significant control was observed for 1 and 2, respectively. As reported fluorescencequenchingonHSAandtransferrinafter30h.The D DOI:10.1021/acs.molpharmaceut.5b00417 Mol.PharmaceuticsXXXX,XXX,XXX−XXX Molecular Pharmaceutics Article Table3.K andK valuesfor1,2,andRAPTA-CBindingto (CAM) of the chicken embryo following intravenous a SV Human Serum Albumin (HSA) and Apotransferrin administration(betweendevelopmentdays11and14)(Figure Determined by Fluorescence Quenching and Calculated 4). Morphological changes in the CAM organization recorded Using the Stern−Volmer Equation on development day 15 were imaged via epifluorescence microscopy and quantified using an automated image- compound(μM) processing method with a descriptor of branching points/ parameters 1 2 RAPTA-C mm2 (Figure 4).43 BindingConstantK (M−1) Complex 1 administered at 50 μM (0.5 mg/kg) was a HSA 0.67×103 3.24×104 1.00×103 previously shown to induce an antivascular effect represented apotransferrin 3.32×103 1.84×106 1.00×106 by a 26.3% reduction in the number of branching points/ Stern−VolmerQuenchingConstantK (104M−1) mm2.28 The presence of avascular zones (black areas in SV HSA 0.5 2.4 0.8 fluorescence angiographies indicated by arrows, Figure 4) was apotransferrin 0.4 0.9 1.0 observed for 2 administered at 42 μM (0.47 mg/kg), with a 38%reductionofthenumberbranchingpoints/mm2versusthe most efficient fluorescence quenching was observed for control (***p = 2.8 × 10 −5), mainly along the larger vessels. apotransferrin in the presence of 1, with a fluorescence The avascular zones observed are most likely due to reduction of 63%, followed by 2 and RAPTA-C (36 and 32% antiangiogenic activity of 2. For comparison, NAMI-A in the reduction, respectively). HSA, in contrast, is essentially equally CAM-sponge model was shown to reduce the number of quenched by the three complexes, with a total fluorescence vessels in the dose range between 120 μM and 240 μM.44 A decrease between 49 and 54%. The binding constants of the similar antivascular effect in the CAM model was observed for complexes to HSA and apotransferrin were determined (Table RAPTA-C and RAPTA-T administered topically in the 3),withthehighestbindingconstants(K)estimatedfor2(K developmental CAM (EDD 7−9) at a dose of 154 μg/ =3.24×104M −1),i.e.,themorelipophil a iccomplex(notethat a embryo.17 these values are lower thanthe binding constant ofKP1019 to Antitumor Effect of 1 and 2 in Human Ovarian HSA).39 The calculated Stern−Volmer quenching constant CarcinomaandHumanColorectalAdenocarcinoma.The (K ) values for NAMI-A with apotransferrin and halotrans- antitumor activity of 1 was previously studied in vivo using the SV ferrinare,respectively,1.28×104M −1and1.28×104M −1.40,42 preclinicalmodelofhumanovarianA2780carcinomagrownon The K values in apotransferrin were lower for 1 (0.4 × 104 thechickenchorioallantoicmembraneCAM(Figure5A).28For SV M −1),2(0.9×104M −1),andRAPTA-C (1.0 ×104M −1),see comparison, 2 and a combination treatment of 1 + 2 were Table 3. Interestingly, the K for HSA with 2 was almost 10- evaluated using the same model and protocol. Embryos were SV fold higher than for 1 and 4-fold higher than for RAPTA-C. randomized and treated within the following groups via Influenceof1and2onVasculatureFormationinthe intravenous administration of 0.1% DMSO (CTRL), 2 (50 CAMModel.Aspreviouslyreportedfor1,28theantiangiogenic μM;0.56mg/kg,80μL),orthemixtureof1+2(premixedat activity of 2 was investigated in the chorioallantoic membrane ratio 1:1; 25 μM each; 0.5 mg/kg, 80 μL). To investigate the Figure4.Invivoactivityof2ontheCAMmodel.Complex2at8μMinducesonlyamildvasculareffectandat42μM2significantly(***p=2.8× 10−5)changesthevasculaturearchitecturecomparedtotheCTRL.Quantificationwasperformedbymeasurementofthebranchingpoints/mm2. Error barsrepresent the standard error of the mean, n = 5−7. E DOI:10.1021/acs.molpharmaceut.5b00417 Mol.PharmaceuticsXXXX,XXX,XXX−XXX Molecular Pharmaceutics Article Treatment with 1 induced significant tumor growth inhibition, by 88% versus the CTRL group, on the last day of theexperiment(***p=1.3×10 −5).28Administrationof2led to inhibition of tumor growth by 63% (*p = 0.03). In cell proliferation inhibition assays at 37 °C 2 was active exclusively on ECRF24 cells, and consequently tumor growth inhibition may be attributed to an antiangiogenic mechanism. The specificity of 2 toward endothelial cells contrasts with other ruthenium-based compounds that possess less specific anti- angiogenic characteristics, i.e., RAPTA-C, RAPTA-T, and DAPTA-C,17 triruthenium clusters,45 and NAMI-A.46 Thepremixedcombinationof1+2administratedivatdoses corresponding to half of the monotherapy doses of both compounds inhibited tumor growth efficiently, by 86% (*p = 0.047),thereductionbeingsimilartothatof1administratedat a 2-fold higher dose. Note that RAPTA-C administered in the same tumor model at 0.2 mg/kg (administered for 5 days) inhibitedA2780tumorgrowthby75%.18Embryobodyweight losswasnotobservedinanyofthetreatmentgroups,indicative of a lack of general toxicity. Figure 5. In vivo anticancer activity of complexes 1, 2, and 1 + 2 in Theabilityof2toreducetumorgrowthwasalsostudiedon humanovariancarcinomatumors(A2780)grownontheCAMusing humanadenocarcinomaLS174Tcarcinomaimplantedsubcuta- experimentalprotocolshownin(A).(B)Tumorgrowthcurvesofthe neously on the hind leg of athymic mice. Compound 1 was followingconditions:CTRL(100μL),1(50μM,80μL),282(50μM, evaluated on the same in vivo model injected intraperitoneally 80μL),or1+2(premixedratio1:1;25μMeach,80μL).Insetshows (ip)every4days.29Themediantumorvolumestreatedwith1 arepresentativetumorintheCTRLgroup.Inallthetreatmentgroups significant tumor growth inhibition compared to the CTRL was or 2 versus the control groups are shown in Figure 6. observed(***p=1.3×10−5for1,*p=0.03for2and*p=0.047for Significanttumorinhibitionwasobservedfor1(by61%versus 1+ 2). Error bars represent standard error of the mean. control,**p<0.01)andfor2(by66%vscontrol,**p<0.01). Tissue sections of treated tumors were subject to immunohis- tochemical analysis of CD31 to delineate the number of effect of 2 on tumor growth, A2780 ovarian carcinoma cells endothelial cells/blood vessels (Figure 6C). The number of were inoculated at EDD 7 of the CAM and monitored for 11 stained vessels wassignificantlylower following treatmentwith days.Establishedandvascularizedtumorsweredetected3days 2. A significant reduction of 32% (**p = 0.006) was observed. postimplantation (EDD 10). Treatment was performed by iv No inhibition of vessel density was observed after treatment injectionsonfourconsecutivedays(Figure5).Tumorsgrewto with 1. anaveragesizeofapproximately150mm3byEDD17whenleft The diameter of remaining vessels in tumors treated with 1 untreated. Tumor growth curves are shown in Figure 5B with was,however,smallerthanthosepresentinthecontroltumors. differences in tumor volume between the treatment groups Combined, these data indicate that treatment with 1 leads to noticeable from the fourth day of treatment. the inhibition of tumor growth via a direct anticancer effect, Figure6.LS174Tadenocarcinomagrowthinhibitionby1(A)and2(B)inathymicmice.Miceweretreatedwith1(75mg/kg,300μLip)or2 (12.5mg/kg,300μLip)givenevery4days(indicatedbyarrows).Thevehiclecontrolwastreatedwiththreedosesof10%DMSOinsterilesaline (0.9%NaCl)alsogivenevery4days.Valuesplottedaremedians±thestandarderrorofthemean,SEM(ANOVA,**p<0.01,***p<0.001).(C). Histochemical analysis of tumors stained for the presence of endothelial cells (CD31) and its quantification. Values plotted are medians ± the standard error of the mean. F DOI:10.1021/acs.molpharmaceut.5b00417 Mol.PharmaceuticsXXXX,XXX,XXX−XXX Molecular Pharmaceutics Article whereas tumor growth inhibition by 2 corresponds predom- Results of a Clinical Phase I Study in Tumor Patients. Chem. i■nantly to an antiangiogenic effect. Biodiversity 2008, 5,2140−2155. (8) Sava, G.; Gagliardi, R.; Bergamo, A.; Alessio, E.; Mestroni, G. CONCLUSIONS Treatment of metastases of solid mouse tumours by NAMI-A: Herein, we show that two closely related ruthenium(II)−p- c A o n m tic p a a n r c i e s r on Res w . 1 it 9 h 99 c , is 1 p 9 l , a 9 ti 6 n 9 , − c 9 y 7 c 2 l . ophosphamide and dacarbazine. cymene complexes, differing only in the length of a pendant (9) Rademaker-Lakhai, J. M.; van den Bongard, D.; Pluim, D.; fluorous chain attached to a pyridine ligand, inhibit tumor Beijnen, J. H.; Schellens, J. H. A Phase I and pharmacological study growth effectively. Interestingly, the mechanism of tumor with imidazolium-trans-DMSO-imidazole-tetrachlororuthenate, a growth inhibition of the two compounds is profoundly novel ruthenium anticancer agent. Clin. Cancer Res. 2004, 10, 3717− different. One compound, i.e., 1, with the shorter pendant 3727. fluorous chain, inhibits tumor growth via anticancer and (10)Hartinger,C.G.;Zorbas-Seifried,S.;Jakupec,M.A.;Kynast,B.; antiangiogenic mechanisms whereas 2 involves predominantly Zorbas, H.; Keppler, B. K. From bench to bedside–preclinical and an antiangiogenic mechanism. This research shows that fine- early clinical development of the anticancer agent indazolium trans- tuning of drug activities is possible through minor structural [tetrachlorobis(1H-indazole)ruthenate(III)] (KP1019 or FFC14A). J. changes,anapproachthatmayhaveanimpactonfuturecancer Inorg. Biochem. 2006,100,891−904. (11) Leijen, S.; Burgers, S. A.; Baas, P.; Pluim, D.; Tibben, M.; van ■therapy. Werkhoven,E.;Alessio,E.;Sava,G.;Beijnen,J.H.;Schellens,J.H.M. PhaseI/IIstudywithrutheniumcompoundNAMI-Aandgemcitabine AUTHOR INFORMATION in patients with non-small cell lung cancer after first line therapy. Corresponding Authors Invest. NewDrugs 2015,33, 201−214. *P.N.-S.: VU University Medical Center, De Boelelaan 1118, (12)Nazarov,A.A.;Baquie,M.;Nowak-Sliwinska,P.;Zava,O.;van PO Box 7057, 1007 MB Amsterdam, The Netherlands. Tel: + Beijnum,J.R.;Groessl,M.;Chisholm,D.M.;Ahmadi,Z.;McIndoe,J. 31 20 4443374. Fax: +31 20 4443844. E-mail: p.nowak- S.; Griffioen, A. W.; van den Bergh, H.; Dyson, P. J. Synthesis and sliwinska@vumc.nl. characterization of a new class of anti-angiogenic agents based on *P.J.D.: Swiss Federal Institute of Technology (EPFL), SB ruthenium clusters. Sci. Rep. 2013, 3,1485. ISIC LCOM, BCH 2402, Av. Forel 2, CH-1015 Lausanne, (13) Trondl, R.; Heffeter, P.; Kowol, C. R.; Jakupec, M. A.; Berger, Switzerland.Tel:+41216939854.Fax:+41216939780.E- W.; Keppler, B. K. NKP-1339, the first ruthenium-based anticancer mail: paul.dyson@epfl.ch. drug on the edge to clinical application. Chem. Sci. 2014, 5, 2925− 2932. Notes (14)Clavel,C.M.;Paunescu,E.;Nowak-Sliwinska, P.;Griffioen,A. T■he authors declare no competing financial interest. W.;Scopelliti,R.;Dyson,P.J.DiscoveryofaHighlyTumor-Selective Organometallic Ruthenium(II)-Arene Complex. J. Med. Chem. 2014, ACKNOWLEDGMENTS 57,3546−3558. We are grateful for technical assistance from Tse J. Wong (15)Scolaro,C.;Bergamo,A.;Brescacin,L.;Delfino,R.;Cocchietto, (VUMC, Amsterdam). We thank the Swiss National Science M.;Laurenczy,G.;Geldbach,T.J.;Sava,G.;Dyson,P.J.Invitroand F■oundation and EPFL for financial support. in vivo evaluation of ruthenium(II)-arene PTA complexes. J. Med. Chem. 2005,48,4161−4171. ABBREVIATIONS USED (16)Bergamo,A.;Masi,A.;Dyson,P.J.;Sava,G.Modulationofthe metastatic progression of breast cancer with an organometallic CAM, chicken chorioallantoic membrane; ECRF24, immortal- ruthenium compound. Int. J. Oncol.2008,33,1281−1289. ized endothelial cell line; HUVEC, human umbilical vein (17)Nowak-Sliwinska,P.;vanBeijnum,J.R.;Casini,A.;Nazarov,A. endothelialcells;EDD,embryodevelopmentday;HSA,human A.; Wagnier̀es, G.; van den Bergh, H.; Dyson, P. J.; Griffioen, A. W. serum albumin; K, binding constant; bFGF, basic fibroblast Organometallic Ruthenium(II) Arene Compounds with Antiangio- a g■rowth factor genic Activity. J. Med.Chem. 2011, 54,3895−3902. (18)Weiss,A.;Berndsen,B.H.;Dubois,M.;Müller,M.;Schibli,R.; REFERENCES Griffioen,A.W.;Dyson,P.J.;Nowak-Sliwinska,P.Invivoanti-tumor activityoftheorganometallicruthenium(II)-arenecomplex[Ru(η6-p- (1)Lippert,B.,Ed.Cisplatin:ChemistryandBiochemistryofaLeading cymene)Cl2(pta)] (RAPTA-C) in human ovarian and colorectal Anticancer Drug;Verlag Helvetica Chimica Acta: 1999; 563 pp. carcinomas. Chem. Sci. 2014,5, 4742−4748. (2)Dhar,S.;Lippard,S.J.CurrentStatusandMechanismofAction (19) Adhireksan, Z.; Davey, G. E.; Campomanes, P.; Groessl, M.; of Platinum-Based Anticancer Drugs. In Bioinorganic Medicinal Chemistry;Wiley-VCH Verlag: 2011; pp 79−95. Clavel,C.M.;Yu,H.;Nazarov,A.A.;Yeo,H.F.;Ang,W.H.;Droge, P.; Rothlisberger, U.; Dyson, P. J.; Davey, C. A. Ligand substitutions (3)Wu,B.;Davey,G.E.;Nazarov,A.A.;Dyson,P.J.;Davey,C.A. between ruthenium−cymene compounds can control protein versus Specific DNA Structural Attributes Modulate Platinum Anticancer Drug Site Selection and Cross-Link Generation. Nucleic Acids Res. DNA targeting and anticancer activity. Nat. Commun. 2014,5, 3462. 2011,39,8200−8212. (20) Babak, M. V.; Meier, S. M.; Huber, K. V. M; Reynisson, J.; (4) Bergamo, A.; Sava, G. Ruthenium complexes can target Legin, A. A.; Jakupec, M.A.; Berger, W.; Dyson, P. J.; Superti-Furga, determinants of tumourmalignancy. Dalton T. 2007,1267−1272. G.; Keppler, B. H.; Hartinger, C. G. Target profiling of an (5) Bergamo, A.;Gaiddon, C.;Schellens, J.H.; Beijnen,J. H.; Sava, antimetastatic RAPTA agent by chemical proteomics: relevance to G.Approachingtumourtherapybeyondplatinumdrugs:statusofthe the mode of action. Chem. Sci. 2015,6,2449−2456. art and perspectives of ruthenium drug candidates. J. Inorg. Biochem. (21) Muramatsu, T. Midkine, a heparin-binding cytokine with 2012,106,90−99. multiplerolesindevelopment,repairanddiseases.Proc.Jpn.Acad.,Ser. (6) Singh, A. K.; Pandey, D. S.; Xu, Q.; Braunstein, P. Recent B 2010,86,410−425. advancesinsupramolecularandbiologicalaspectsofareneruthenium- (22) Carmeliet, P.; Jain, R. K. Molecular mechanisms and clinical (II)complexes. Coord. Chem. Rev. 2014,270−271, 31−56. applications of angiogenesis. Nature 2011, 473,298−307. (7)Hartinger,C.G.;Jakupec,M.A.;Zorbas-Seifried,S.;Groessl,M.; (23) Griffioen, A. W. Therapeutic approaches of angiogenesis Egger,A.;Berger,W.;Zorbas,H.;Dyson,P.J.;Keppler,B.K.KP1019, inhibition: are we tackling the problem at the right level? Trends ANewRedox-ActiveAnticancerAgent−PreclinicalDevelopmentand Cardiovasc. Med.2007, 17, 171−176. G DOI:10.1021/acs.molpharmaceut.5b00417 Mol.PharmaceuticsXXXX,XXX,XXX−XXX Molecular Pharmaceutics Article (24) Weiss, A.; van den Bergh, H.; Griffioen, A. W.; Nowak- (40) Spiewak, K.; Brindell, M. Impact of low- and high-molecular- Sliwinska, P. Angiogenesis inhibition for the improvement of masscomponentsofhumanserumonNAMI-Abindingtotransferrin. photodynamic therapy: the revival of a promising idea. Biochim. JBIC, J. Biol. Inorg. Chem. 2015,20,695. Biophys. Acta, Rev. Cancer 2012, 1826, 53−70. (41) Kratz, F.;Hartmann, M.;Keppler, B.; Messori, L.Thebinding (25)Huang,Y.;Yuan,J.;Righi,E.;Kamoun,W.S.;Ancukiewicz,M.; properties of two antitumor ruthenium(III) complexes to apotrans- Nezivar,J.;Santosuosso,M.;Martin,J.D.;Martin,M.R.;Vianello,F.; ferrin. J. Biol. Chem. 1994, 269,2581−2588. Leblanc,P.;Munn,L.L.;Huang,P.;Duda,D.G.;Fukumura,D.;Jain, (42)Mazuryk,O.;Kurpiewska,K.;Lewinski,K.;Stochel,G.;Brindell, R.K.;Poznansky,M.C.Vascularnormalizingdosesofantiangiogenic M.Interactionofapo-transferrinwithanticancerrutheniumcomplexes treatment reprogram the immunosuppressive tumor microenviron- NAMI-A and its reduced form. J. Inorg. Biochem.2012, 116,11−18. mentandenhanceimmunotherapy.Proc.Natl.Acad.Sci.U.S.A.2012, (43) Nowak-Sliwinska, P.; Ballini, J.-P.; Wagnier̀es, G.; van den 109,17561−17566. Bergh,H.Processingoffluorescenceangiogramsforthequantification (26) Griffioen, A.W.; Weiss, A.;Berndsen, R. H.; Abdul,U. K.; Te of vascular effects induced by anti-angiogenic agents in the CAM model. Microvasc. Res. 2010,79,21−28. Winkel, M. T.; Nowak-Sliwinska, P. The emerging quest for the (44) Vacca, A.; Bruno, M.; Boccarelli, A.; Coluccia, M.; Ribatti, D.; optimal angiostatic combination therapy. Biochem. Soc. Trans. 2014, 42, 1608−1615. Bergamo,A.;Garbisa,S.;Sartor,L.;Sava,G.Inhibitionofendothelial cellfunctionsandofangiogenesisbythemetastasisinhibitorNAMI-A. (27) Pintus, G.; Tadolini, B.; Posadino, A. M.; Sanna, B.; Debidda, Br. J. Cancer 2002, 86, 993−998. M.;Bennardini,F.;Sava,G.;Ventura,C.InhibitionoftheMEK/ERK (45)Nazarov,A.A.;Baquie,M.;Nowak-Sliwinska,P.;Zava,O.;van signaling pathway by the novel antimetastatic agent NAMI-A down Beijnum,J.R.;Groessl,M.;Chisholm,D.M.;Ahmadi,Z.;McIndoe,J. regulatesc-mycgeneexpressionandendothelialcellproliferation.Eur. J. Biochem. 2002, 269,5861−70. S.; Griffioen, A. W.; van den Bergh, H.; Dyson, P. J. Synthesis and characterization of a new class of anti-angiogenic agents based on (28) Clavel,C.M.;Paunescu, E.;Nowak-Sliwinska,P.;Griffioen,A. ruthenium clusters. Sci. Rep. 2013, 3,1485. W.;Scopelliti,R.;Dyson,P.J.DiscoveryofaHighlyTumor-Selective (46) Vacca, A.; Bruno, M.; Boccarelli, A.; Coluccia, M.; Ribatti, D.; Organometallic Ruthenium(II)-Arene Complex. J. Med. Chem. 2014, Bergamo,A.;Garbisa,S.;Sartor,L.;Sava,G.Inhibitionofendothelial 57, 3546−3558. cellfunctionsandofangiogenesisbythemetastasisinhibitorNAMI-A. (29) Clavel,C.M.;Nowak-Sliwinska, P.;Paunescu,E.;Griffioen,A. Br. J. Cancer 2002, 86, 993−8. W.; Dyson, P. J. In vivo evaluation of small-molecule thermores- (47) Allardyce, C. S.; Dyson, P. J.; Ellis, D. J.; Heath, S. L. ponsive anticancer drugs potentiated by hyperthermia. Chem. Sci. [Ru(eta(6)-p-cymene)Cl-2(pta)] (pta=1,3,5-triaza-7-phosphatricyclo- 2015,6, 2795−2801. [3.3.1.1]decane): a water soluble compound that exhibits pH (30) Clavel, C. M.; Paunescu, E.; Nowak-Sliwinska, P.; Dyson, P. J. dependent DNA binding providing selectivity for diseased cells. Thermoresponsive organometallic arene ruthenium complexes for Chem. Commun. 2001, 1396−1397. tumourtargeting. Chem. Sci. 2014, 5,1097−1101. (48)Fontijn,R.;Hop,C.;Brinkman,H.J.;Slater,R.;Westerveld,A.; (31) Ang, W. H. Novel strategies for overcoming drug resistance in vanMourik,J.A.;Pannekoek,H.Maintenanceofvascularendothelial transition metal-based anticancer compounds; Swiss Federal Institute of cell-specific properties after immortalization with an amphotrophic Technology: Lausanne, 2007. replication-deficient retrovirus containing human papilloma virus 16 (32) Kilpin, K. J.; Clavel, C. M.; Edafe, F.; Dyson, P. J. E6/E7 DNA. Exp. Cell Res.1995, 216,199−207. Naphthalimide-Tagged Ruthenium-Arene Anticancer Complexes: (49)vanBeijnum,J.R.;Rousch,M.;Castermans,K.;vanderLinden, Combining Coordination with Intercalation. Organometallics 2012, E.; Griffioen, A. W. Isolation of endothelial cells from fresh tissues. 31, 7031−7039. Nat. Protoc. 2008, 3,1085−1091. (33)Hanif,M.;Nazarov,A.A.;Legin,A.;Groessl,M.;Arion,V.B.; (50) Brandwijk, R. J.; Nesmelova, I.; Dings, R. P.; Mayo, K. H.; Jakupec,M.A.;Tsybin,Y.O.;Dyson,P.J.;Keppler,B.K.;Hartinger, Thijssen, V. L.; Griffioen, A. W. Cloning an artificial gene encoding C. G. Maleimide-functionalised organoruthenium anticancer agents angiostaticanginex:Fromdesignedpeptidetofunctionalrecombinant and their binding to thiol-containing biomolecules. Chem. Commun. protein. Biochem.Biophys. Res. Commun. 2005,333, 1261−1268. 2012,48,1475−1477. (51)vanderSchaft,D.W.;Dings,R.P.;deLussanet,Q.G.;vanEijk, (34)Nowak-Sliwinska,P.;vanBeijnum,J.R.;Casini,A.;Nazarov,A. L.I.;Nap,A.W.;Beets-Tan,R.G.;Bouma-TerSteege,J.C.;Wagstaff, A.; Wagnier̀es, G.; van den Bergh, H.; Dyson, P. J.; Griffioen, A. W. J.;Mayo,K.H.;Griffioen,A.W.Thedesigneranti-angiogenicpeptide Organometallic Ruthenium(II) Arene Compounds with Antiangio- anginex targets tumor endothelial cells and inhibits tumor growth in genic Activity. J. Med.Chem. 2011,54,3895−3902. animal models. FASEB J. 2002, 16,1991−1993. (35)Guo,W.;Zheng,W.;Luo,Q.;Li,X.;Zhao,Y.;Xiong,S.;Wang, (52) Nowak-Sliwinska, P.; Segura, T.; Iruela-Arispe, M. L. The F. Transferrin serves as a mediator to deliver organometallic chicken chorioallantoic membrane model in biology, medicine and ruthenium(II) anticancer complexes into cells. Inorg. Chem. 2013, bioengineering. Angiogenesis 2014,17,779−804. 52, 5328−5338. (53) Nowak-Sliwinska, P.; Storto, M.; Cataudella, T.; Ballini, J. P.; (36)Timerbaev,A.R.;Hartinger,C.G.;Aleksenko,S.S.;Keppler,B. Gatz, R.; Giorgio, M.; van den Bergh, H.; Plyte, S.; Wagnier̀es, G. K. Interactions of antitumor metallodrugs with serum proteins: Angiogenesisinhibitionbythemaleimide-basedsmallmoleculeGNX- advances in characterization using modern analytical methodology. 686. Microvasc. Res. 2012,83,105−110. Chem. Rev. 2006,106, 2224−2248. (54) Nowak-Sliwinska, P.; Ballini, J. P.; Wagnieres, G.; van den (37)Novohradsky,V.;Bergamo,A.;Cocchietto,M.;Zajac,J.;Brabec, Bergh,H.Processingoffluorescenceangiogramsforthequantification of vascular effects induced by anti-angiogenic agents in the CAM V.;Mestroni,G.;Sava,G.InfluenceofthebindingofreducedNAMI- model. Microvasc. Res. 2010,79,21−28. A to human serum albumin on the pharmacokinetics and biological activity. Dalton T. 2015,44,1905−1913. (55) Adar, Y.;Stark, M.;Bram, E. E.;Nowak-Sliwinska, P.; vanden Bergh,H.;Szewczyk,G.;Sarna,T.;Skladanowski,A.;Griffioen,A.W.; (38) Bergamo, A.;Messori, L.;Piccioli,F.; Cocchietto,M.;Sava, G. Assaraf, Y. G. Imidazoacridinone-dependent lysosomal photodestruc- Biological role of adduct formation of the ruthenium(III) complex tion:apharmacologicalTrojanhorseapproachtoeradicatemultidrug- NAMI-Awithserumalbuminandserumtransferrin.Invest.NewDrugs 2003,21,401−411. resistant cancers. Cell Death Dis. 2012,3, e293. (39)Domotor,O.;Hartinger,C.G.;Bytzek,A.K.;Kiss,T.;Keppler, B. K.; Enyedy, E. A. Characterization of the binding sites of the anticancerruthenium(III)complexesKP1019andKP1339onhuman serum albumin via competition studies. JBIC, J. Biol. Inorg. Chem. 2013,18,9−17. H DOI:10.1021/acs.molpharmaceut.5b00417 Mol.PharmaceuticsXXXX,XXX,XXX−XXX