← Back

Cellular selectivity and biological impact of cytotoxic rhodium(III) and iridium(III) complexes containing methyl-substituted phenanthroline ligands.

PMID: 21337523
DOI:10.1002/cmdc.201000517 Cellular Selectivity and Biological Impact of Cytotoxic Rhodium(III) and Iridium(III) Complexes Containing Methyl-Substituted Phenanthroline Ligands Yvonne Geldmacher,[a] Igor Kitanovic,[b] Hamed Alborzinia,[b] Katharina Bergerhoff,[b] Riccardo Rubbiani,[c] Pascal Wefelmeier,[d] Aram Prokop,[d] Ronald Gust,[e] Ingo Ott,[c] Stefan Wçlfl,[b] and William S. Sheldrick*[a] The antiproliferative properties and biological impact of octa- line complexes 5 and 8 both cause significant inhibition of hedral iridium(III) complexes of the type fac-[IrCl(DMSO)(pp)] Jurkatleukemiacell proliferation andinvokeextensiveapopto- 3 containing pp=phenanthroline (1) and its 4- and 5-methyl (2, sis but negligible necrosis. The percentages of Jurkat cells ex- 3) and 4,7- and 5,6-dimethyl derivatives (4, 5) were investigat- hibiting high levels of reactive oxygen species correlate with edforbothadherentandnon-adherentcells.Aseriesofsimilar the percentages of cells undergoing apoptosis. The antiproli- rhodium(III) complexes were studied for comparison purposes. ferative activity of 5 and 8 is strongly selective toward MCF-7 TheantiproliferativeactivitytowardMCF-7cancercellsincreas- and HT-29 cancer cells over normal HFF-1 and immortalized eseightfoldfromIC =4.6 for1toIC =0.60mmfor5,andan HEK-293 cells. Complex 5 also exhibits high selectivity toward 50 50 even more pronounced 18-fold improvement was established BJAB lymphoma cells relative to healthy leukocytes. Both 5 for the analogous rhodium complexes 6 and 8, the respective and 8 invoke permanent decreases in the adhesion and respi- IC valuesforwhichare1.1and0.06mm.AnnexinV/propidium rationofMCF-7cells. 50 iodide assays demonstrated that the 5,6-dimethylphenanthro- Introduction Due to their prevailing chemical inertness, iridium complexes matic ligands. This led us to turn to facial trichlorido com- have been widely ignored in the search for novel metal-based pounds of the type fac-[IrCl(DMSO-kS)(pp)] in the search for 3 anticancer drugs. In earlier studies, biological inactivity was es- kinetically tunable Ir-based agents.[8,9] The related meridional tablished for various iridium(I)[1,2] and iridium(III)[3,4] complexes RhIII complexes mer-[RhCl(DMSO-kS)(pp)] are highly cytostat- 3 andthisledtothegeneralassumptionthatthelackofreactivi- ic[10] and exhibit IC values toward MCF-7/HT-29 cells varying 50 tyofsuch compounds willbereflected inaneffectiveabsence from 4.0/1.9 to 0.079/0.069mm depending on the polypyridyl ofcytotoxiceffectsonhumantumorcelllines,evenatrelative- ligandsize(IC :bpy > phen > dpq,dppzwheredpq=dipyr- 50 ly high concentrations. It should, however, be remembered ido [3,2-f:2’,3’-h] quinoxaline). At first sight, the facial IrIII com- that long-term stability and slow rates of ligand exchange are pounds would appear to be less promising on the basis of properties that have proved to be desirable in modern anti- cancer drug development. We therefore reasoned that the key [a] Dr.Y.Geldmacher,Prof.W.S.Sheldrick to the successful design of iridium-containing anticancer Fakult(cid:2)tf(cid:3)rChemieundBiochemie agentswouldlieintheabilitytotunetheirratesofligandsub- Ruhr-Universit(cid:2)tBochum,44780Bochum(Germany) Fax:(+49)234-3214420 stitution. E-mail:william.sheldrick@rub.de The lability of Ir(cid:2)X bonds in octahedral IrIII coordination [b] Dr.I.Kitanovic,H.Alborzinia,K.Bergerhoff,Prof.S.Wçlfl spheres is dependent on the trans effect of the opposite Institutf(cid:3)rPharmazieundMolekulareBiotechnologie ligand. For instance, the strong trans effect of the [CMe](cid:2) co- Ruprecht-Karls-Universit(cid:2)tHeidelberg 5 5 ligand in half-sandwich complexes of the type [(h5- ImNeuenheimerFeld364,69120Heidelberg(Germany) CMe)IrCl(pp)]+ (pp=polypyridyl ligand) induces rapid substi- [c] R.Rubbiani,Prof.I.Ott 5 5 Institutf(cid:3)rPharmazeutischeChemie tutionofthechlorideligandbysofternucleophilessuchasthe TechnischeUniversit(cid:2)tBraunschweig nucleobase N atoms of DNA.[5,6] An invitro antiproliferative ac- Beethovenstraße55,38106Braunschweig(Germany) tivity toward the human cell lines MCF-7 (breast cancer) and [d] P.Wefelmeier,Dr.A.Prokop HT-29(coloncancer)similartothatofcisplatinwasestablished DepartmentofPediatricOncology Children’sHospitalofCologne for the dppz complex (dppz=dipyrido[3,2-a:2’,2’-c]phenazine) AmsterdamerStraße59,50735Cologne(Germany) of this type.[7] However, the analogous phenanthroline (phen) [e] Prof.R.Gust complex is inactive, which suggests that the cytostatic proper- Institutf(cid:3)rPharmazie,Freie-Universit(cid:2)tBerlin ties may be dependent on specific interactions of larger aro- Kçnigin-Luise-Straße2-4,14195Berlin(Germany) ChemMedChem2011,6,429–439 (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim 429 MED W.S.Sheldricketal. their higher IC values, which lie in the low micromolar range 50 (e.g., 4.6/4.6mm for pp=phen and 0.8/1.5mm for pp=dppz). However, the kinetically more labile RhIII compounds have been found to exhibit two significant chemical disadvantages: 1)slowDMSO–HOexchangeinaqueoussolutionthatleadsto 2 inactiveaquaspecies,[11]and2)mer!facisomerizationinpolar solvents that also causes a time-dependent decrease in activi- Figure3.Rhodium(III)complexesfac/mer-[RhCl(DMSO)(pp)]10–11with ty.[9] In addition, their remarkably high levels of cellular uptake pp=dppzandmethyl-substituteddppzligands 3 . (e.g., 78mm in MCF-7 cells for a 1mm extracellular mer-[RhCl- 3 (DMSO)(phen)] solution)[10] may be expected to invoke a high level of general toxicity. In contrast, the more inert facial IrIII of DMSO in methanol in the absence of light. Their 1HNMR compounds are stable in aqueous solution and exhibit much spectra in CDCl were recorded in the dark and were in ac- 2 2 lower levels of cellular uptake. These findings motivated us to cordance with the expected patterns for the fac isomers. For concentrate on lead substance optimization for the fac-[IrCl- instance, only single resonances were present for each of the 3 (DMSO-kS)(pp)]species. H2/9, H3/8, and H5/6 or H4/7 proton pairs, respectively, of the Toward this goal, we investigated the antiproliferative prop- C-symmetrical4,7-Mephenand5,6-Mephencomplexes4and s 2 2 erties and cellular impact of complexes fac-[IrCl(DMSO- 5. Each of these protons would exhibit an individual signal in 3 kS)(pp)] 2–5containing the methyl-substituted phenanthroline the appropriate mer isomer, owing to the differing trans influ- ligands 4-Mephen, 5-Mephen, 4,7-Mephen, and 5,6-Mephen encesoftheoppositeDMSOandchlorideligands.[9]Duetothe 2 2 (Figure1). A significantly enhanced activity has been reported asymmetric substitution patterns in the 4-Mephen and 5- Mephenligandsoffacialcomplexes2and3,respectively,sepa- rate resonances were present for each of the six aromatic pro- tons in their 1HNMR spectra. As two distinct mer isomers are possiblefor2and3,atotaloftwelvesignalswouldbeexpect- ed for the two alternative meridional arrangements of the li- gands. As previously reported for fac-[IrCl(DMSO)(phen)] 1, 3 the fac isomers 2–5 were stable over a period of 24h in DO 2 and CDOD but isomerized to fac/mer mixtures both in DMF 3 andDMSOsolutionsonexposuretoroomlight. Owingtothelowersolubilityofthesubstitutedphenanthro- Figure1.Facialiridium(III)complexesfac-[IrCl(DMSO)(pp)]1–5with 3 line ligands in polar solvents, the synthesis of the rhodium(III) pp=phenanthroline(phen)andmethyl-substitutedphenligands. complexes 7–10 was performed in CHCl rather than in a 3 CHOH/HO (1:1) mixture, as previously reported for mer- 3 2 for kinetically inert PtII complexes of type [Pt(en)(Mephen)]2+ [RhCl(DMSO)(phen)].[10] Treatment of the precursor mer,cis- x 3 (x=1, 2, 4)[12–14] in comparison with [Pt(en)(phen)]2+. The RhIII [RhCl(DMSO-kS)(DMSO-kO)(phen)][15] with one equivalent of 3 2 compounds fac/mer-[RhCl(DMSO-kS)(pp)] 7–10 containing the the appropriate polypyridyl ligand led toisolation of amixture 3 substituted ligands 4,7-Mephen, 5,6-Mephen, 4,7-(MeO)phen ofthefacialandmeridionalisomersineachcasefollowingsol- 2 2 2 (Figure2),andMedppz(Figure3)werestudiedforcomparison ventremoval.Thefac/merratiosvariedfromapproximately1:1 2 purposes. for the 4,7-Mephen complex 7 to about 1:3 for the 4,7- 2 (MeO)phen complex 9. Isolation of the mer isomers in the 2 solidstatewaspossibleowingtotheirmarkedlylowersolubili- ty in methanol in comparison with the alternative fac isomers. Figure4a depicts the 1HNMR spectrum of the fac isomer of 7 in [D]methanol in the dark following its separation from a 1:1 4 fac/mer mixture. The insoluble residue of the mer isomer was dissolved in CDCl in the dark to afford the 1HNMR spectrum 2 2 depicted in Figure4b. The observed downfield shift of the now separate H2 and H9 resonances (two doublets) is charac- Figure2.Rhodium(III)complexesfac/mer-[RhCl 3 (DMSO)(pp)]6–9with teristic for the mer isomer in comparison with the single dou- pp=unsubstitutedandmethyl-andmethoxy-substitutedphenligands. blet for the C-symmetric fac isomer.[10] Attempts to isolate the s fac isomers of 7–10 in the solid state were unsuccessful. Sol- Results and Discussion vent removal from the methanol solutions of the fac isomers led invariably to formation of fac/mer mixtures following iso- Synthesisandstructure merization prior to and/or during crystallization. Fac/mer mix- The iridium(III) compounds 2–5 were prepared by stepwise re- tures of 7–10 were therefore employed for the subsequent action of IrCl·3HO with an equimolar quantity of the appro- biological studies to allow a closer comparison with the 3 2 priatephenanthrolinederivativefollowedbyathreefoldexcess iridium(III) fac isomers 1–5.Therhodium(III) analogues of com- 430 www.chemmedchem.org (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim ChemMedChem2011,6,429–439 Rhodium(III)andIridium(III)Complexes Figure5.CDspectraforCT-DNAandmixturesoftherhodium(III)complexes 7(pp=4,7-Mephen)and8(pp=5,6-Mephen)withCTDNAat1:10molar 2 2 ratioina10mmphosphatebufferatpH7.2afteranincubationperiodof 2h. rhodium(III) complexes 7 and 8 invoked a similar effect (Figure5). Interestingly, though addition of the Medppz com- 2 plex 10 had effectively no influence on the DNA CD spectrum on initial mixing, marked changes were apparent after 24h (Figure6). The appearance of a negative CD band at about 295nm is characteristic for dppz intercalation.[5,6] Itis plausible that slow Cl(cid:2)/HO exchange will lead to cationic species, for 2 which intercalative binding will be supported by electrostatic interactionswiththephosphatebackbone. Figure4.1HNMRspectraofa)thefacisomerof[RhCl(DMSO)(4,7-Mephen)] 3 2 7inCDODinthedark,andb)themerisomerof7inCDCl inthedark. 3 2 2 plexes 2 and 3 were not included in the study owing to the fact that the asymmetric substitution pattern of their 4- Mephen and 5-Mephen ligands led invariably to the formation ofthree(twomeridionalandonefacial)isomericspecies. Changesinthethermaldenaturationtemperature(DT )and m inthecirculardichroism(CD)spectrumofDNAinthepresence of metal complexes can provide a means of gauging the strength and nature of possible binding interactions. Whereas the intercalation of polypyridyl ligands between the nucleo- Figure6.CDspectraforCT-DNAandmixturesoftherhodium(III)complexes base pairs generally leads to large increases in the DNA melt- 9(pp=4,7-(MeO) 2 phen)and10(pp=Me 2 dppz)withCTDNAat1:10molar ing temperature,[5,16] the formation of covalent metal–nucleo- ratioina10mmphosphatebufferatpH7.2afteranincubationperiodof 24h. basebondstypicallyinvokesonlyminorchangesinT andcan m evendestabilizetheBconformation.[17]Themodestchangesin DT of0–38CforcalfthymusDNA(CTDNA)inthepresenceof Bathochromic shifts and significant decreases in the [V] m complexes 1–10 (at 1:10 [complex]/[DNA] molar ratio, where value for thepositive CD band at 275nmare typical upon for- [DNA] is the DNA concentration in nucleotides) in a 10mm mation of covalent metal–nucleobase bonds in CT DNA.[7] The phosphatebufferatpH7.2(1–5:0,+1,+1,0,+28C;6–10:0, lack of a band shift and the increases in the molar ellipticity +3,0,+2,+18C)providednoevidenceforsignificantinterca- forthecomplex/DNAmixturesatthiswavelengthsuggestthat lation, even in the case of the Medppz complex 10. A nega- covalent binding is insignificant for the trichlorido complexes. 2 tiveCDbandat246nmduetothehelicalBconformationand Support for this conclusion was provided by time-dependent a positive band at 275nm caused by nucleobase stacking rep- 1HNMR studies on mixtures of the compounds with guano- resent the characteristic features in the CD spectrum of CT sine-5’-monophosphate (5’-GMP2(cid:2)). As previously reported for DNA.[18,19]Smallincreasesof10–20%inthemolarellipticity[V] 1 and 6,[8,10] no reaction with the nucleobase was observed of the positive band were observed on addition of 2–5 to CT after treatment of the 5,6-Mephen complexes 5 and 8 with a 2 DNA, which suggests that the presence of the iridium(III) com- twofold excess of 5’-GMP2(cid:2) at 608C for 24h. This lack of reac- plexes may have a stabilizing effect on the double helix. The tivityisinstarkcontrastto(h5-CMe)MIII(M=Rh,Ir)complexes 5 5 ChemMedChem2011,6,429–439 (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim www.chemmedchem.org 431 MED W.S.Sheldricketal. with labile chloride ligands, which readily bind to nucleobase the central metal atom (Ir!Rh), and 3)the trans effect Natoms.[20] (N(imino)<Cl(cid:2))ofthedonoratomsituatedoppositetothekS- coordinated DMSO ligand. For instance, the significantly lower IC values of 4.6(0.5)/4.6(0.2)mm toward MCF-7/HT-29 cells[9] Antiproliferativepropertiesandstructure–activity 50 for fac-[IrCl(DMSO)(phen)] in comparison with values of relationships 20.3(0.5)/16.8 3 (0.1)mm for mer-[IrCl(DMSO)(phen)] correlate 3 Invitrostudiesontheantiproliferativepropertiesofcomplexes with the higher rate of DMSO substitution in the former com- 2–5 and 7–10 were performed using crystal violet assays for pound (DMSO trans to Cl(cid:2) vs. DMSO trans to N(phen)). In con- theadhesivehumancancercell linesMCF-7 and HT-29 andfor trast, the 70:30 fac/mer mixture [RhCl(DMSO)(phen)] 3 immortalized human embryonic kidney cells HEK-293. The re- (IC :1.1(0.2)/0.66(0.02)mm) is considerably less active toward 50 sulting IC values are listed in Table1 together with the litera- MCF-7/HT-29 cells than the pure mer isomer (IC :0.40(0.06)/ 50 50 ture values for fac-[IrCl(DMSO)(phen)], fac/mer-[RhCl(DMSO)- 0.19(0.05)mm). DMSO–HO exchange leads to relatively rapid 3 3 2 formation of the respective aqua speciesforthemuch morelabile Table1. IC values based on crystal violet assays for complexes of the types fac-[IrCl(DMSO)(pp)] 1–5 and rhodium(III) compounds and 50 3 fac/mer-[RhCl 3 (DMSO)(pp)]6–10towardtheadhesivehumancelllinesMCF-7,HT-29,andHEK-293. these are subsequently inactive toward the human cell lines.[11] IC [mm][a] 50 Complex ppLigand fac/mer MCF-7 HT-29 HEK-293 Due to the greater trans effect of the chloride ligand relative to 1 phen[8] 100:0 4.6(0.5) 4.6(0.2) n.d. 2 4-Mephen 100:0 1.79(0.14) 2.21(0.14) 2.85(0.53) the imino N atom of the phe- 3 5-Mephen 100:0 1.06(0.04) 1.88(0.41) 1.41(0.12) nanthroline ligand, DMSO/HO 2 4 4,7-Me 2 phen 100:0 2.31(0.01) 2.47(1.37) 1.00(0.18) substitution will proceed more 5 5,6-Mephen 100:0 0.60(0.27) 0.86(0.07) 3.64(0.77) 2 rapidly for the fac isomer. This 6[b] phen[9] 70:30 1.1(0.2) 0.66(0.02) n.d. 7 4,7-Mephen 48:52 0.31(0.005) 0.26(0.10) 0.99(0.01) leads to the observation of sig- 2 8 5,6-Mephen 35:65 0.06(0.02) 0.095(0.005) 0.97(0.01) nificantly higher IC values for 2 50 1 9 0 4,7- M (M e e d O p ) 2 p p z hen 3 2 9 4 : : 6 7 1 6 1 3 . . 0 4 4 5 ( ( 0 0 . . 0 5 7 1 ) ) 1 2 . . 8 5 4 0 ( ( 1 2 . . 4 0 ) 0) 1 1 . . 0 3 4 4 ( ( 0 0 . . 0 4 6 4 ) ) the fac/mer mixture of [RhCl 3 - 2 (DMSO)(phen)] relative to the 11[b] dppz[9] 50:50 0.093(0.05) 0.057(0.021) n.d. Cisplatin[8] 2.0(0.3) 7.0(2.0) 2.6(0.6) merisomer. Inspection of Table1 indicates [a]Values in brackets represent standard errors; n.d.=not determined. [b]IC values of 0.40(0.06)/ 0.19(0.05)mm and 0.095(0.020)/0.073(0.017)mm were reported for the mer complexe 50 s mer-[RhCl(DMSO)(pp)] that substitution of the phenan- 3 withpp=phenanddppz,respectively.[10] throline ligands in the iridium(III) complexes 2–5withelectron-do- natingmethylgroupssignificant- (phen)], and fac/mer-[RhCl(DMSO)(dppz)].[9] Invitro antiproli- ly improves their antiproliferative activity toward MCF-7 and 3 ferative activities of the methyl-substituted phenanthroline HT-29 cells in comparison with fac-[IrCl(DMSO)(phen)]. A simi- 3 anddppzligandsarepresentedinTable2. lar trend is apparent for the rhodium(III) compounds 7 and 8 Previousbiologicalstudies[8–10]oncompoundsofthegeneral butnotfortheMedppzcomplex10whencomparedwiththe 2 type [MCl(DMSO)(pp)] (M=Rh, Ir) have enabled the establish- dppz complex 11. The presence of methoxy substituents with 3 ment of a number of structure-activity trends. The cytostatic a (cid:2)I effect in fac/mer-[RhCl(DMSO){(MeO)phen}] 9 leads to 3 2 activity of such compounds toward MCF-7 and HT-29 cells is loss in activity when compared with phen complex 6. The dependent on: 1)the size of the polypyridyl ligand (bpy < greatestimprovementsinantiproliferativeactivitywererecord- phen, dpq < dppz), 2)the rate of substitution reactions for ed for the 5,6-Mephen complexes 5 and 8. For instance, the iridium(III) complex 5 is some 7.7-fold more active toward MCF-7 cells and 5.3-fold more active toward HT-29 cells than fac-[IrCl(DMSO)(phen)] 1. An interesting aspect is the much Table2. IC valuesbasedoncrystalvioletassaysforpolypyridylligands 3 50 higher activity of 5 and 8 toward the cancer cell lines than towardtheadhesivehumancelllinesMCF-7,HT-29,andHEK-293. toward immortalized HEK-293 cells. In the case of fac-[IrCl- 3 IC 50 [mm][a] (DMSO)(5,6-Mephen)] 5, activity ratios of 6.1 and 4.2 were re- ppLigand MCF-7 HT-29 HEK-293 2 corded for MCF-7 and HT-29 cells, respectively, versus HEK-293 phen 3.5(0.2) 2.7(0.5) 3.27(0.01) cells. 4-Mephen 0.61(0.24) 0.80(0.09) 0.58(0.15) These promising findings prompted us to investigate the 5-Mephen 0.72(0.09) 0.98(0.03) 0.81(0.45) 4,7-Me 2 phen 0.36(0.13) 0.33(0.16) 0.50(0.32) effect of the 5,6-Me 2 phen complexes on normal HFF-1 cells 5,6-Mephen 0.34(0.08) 0.28(0.16) 0.64(0.39) (human foreskin fibroblasts). Antiproliferative properties were 2 4,7-(MeO) 2 phen 1.33(0.39) 0.75(0.49) 0.60(0.15) determined in this case using the MTTassay after 48h incuba- Medppz n.d. 7.86(3.73) 7.24(2.46) 2 tion periods. Thus, the resulting IC values were compared dppz 0.8(0.6) 1.8(0.2) 1.46(0.27) 50 with those determined for MCF-7 (5, 1.80(0.14); 8, [a]Valuesinbracketsrepresentstandarderrors;n.d.=notdetermined. 0.54(0.12)mm) and HT-29 cells (5, 2.56(0.45); 8, 0.35(0.06)mm) 432 www.chemmedchem.org (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim ChemMedChem2011,6,429–439 Rhodium(III)andIridium(III)Complexes with the same assay. MTT is a tetrazolium salt that is reduced 63.4mm in MCF-7 cells. Similar intracellular concentrations of by mitochondrial dehydrogenase to dark-blue, water-insoluble 21.5 and 43.8mm were only obtained for fac-[IrCl(DMSO)- 3 formazan inviable cells. Following itsdissolutionin DMSO, the (phen)] 1 after exposure of cells to a 100-fold higher concen- quantity of formazan can be determined colorimetrically to tration(100mm)oftheiridium(III)complex.[8] give an estimate of the number of viable cells. The IC values 50 for 5 and 8 were higher than those obtained using the crystal violet assay, which measures the DNA content of living cells, Apoptosisinductioninnon-adhesivecells whiletheMTTassayreflectstheirmetabolicactivity.Significant- It was of interest to establish whether the 5,6-Mephen 5 and ly higher values are often observed for MTTassays because of: 2 8complexesinduceapoptosisinmalignantnon-adhesivecells, 1)residual enzymatic activity in dead/dying cells, 2)additional as previously reported for the compounds mer-[RhCl- washing steps before the final dissolution of the crystal violet 3 (DMSO)(pp)] (pp=dpq, dppz) with larger polypyridyl ligands.[9] dye (which may remove partially damaged cells that are not The observed dose-dependentloss of the mitochondrial mem- firmly attached), and 3)the t correction for the initial cell bio- 0 brane potential in lymphoma cells indicated that apoptosis is massinthecaseofthecrystalvioletassay.Theiridium(III)com- mediated via the intrinsic mitochondrial pathway. Staining plex 5 exhibits an IC value of 11.4(2.4)mm for HFF-1 cells and 50 with AnnexinV-FITC and propidium iodide (PI) provides a con- is therefore some 6.3- and 4.5-fold more active toward MCF-7 venientmethodofestimatingthenumberofviable,apoptotic, and HT-29 cells, respectively. Similar activity ratios of 4.9 and and necrotic cells following incubation with a cytotoxic agent. 7.5 were observed for 8 (IC =2.62(1.52)mm for HFF-1 cells) 50 During apoptosis, the phospholipid phosphatidylserine is ex- toward these adherent cancer cell lines in comparison with posed to the outer leaflet of the plasma membrane.[23,24] An- normalcells. nexinV-FITC binds specifically to phosphatidylserine, causing Withtheexceptionofthe4,7-Mephencomplex4,theorder 2 anincreaseinfluorescencethatis dependent onthetotalper- of increasing antiproliferative activity in the iridium(III) com- centage of cells undergoing apoptosis. Propidium iodide, on plexes (1 < 4 < 2 < 3 < 5) toward MCF-7 and HT-29 cells the other hand, cannot enter viable cells or early apoptotic parallels that established for the methyl-substituted ligands on cells with still intact plasma membranes. PI positivity is there- their own (phen < 4-Mephen, 5-Mephen < 4,7-Mephen, 5,6- 2 fore an indicator for either cell necrosis or for the presence of Mephen). This suggests that ligand-based properties may be 2 late apoptotic cells, whose plasma membrane is no longer responsible for the improvements in activity. We postulated intact.[25] that enhancement of their cellular uptake might play a role. AnnexinV/PI assays were performed for leukemia (Jurkat, K- However, as depicted in Figure7, a quantitative AAS study for 562) and lymphoma cells (BJAB) as well as for ex vivo healthy complex 8 indicated that the intracellular rhodium concentra- leukocytes.Figures8and9illustratethecellstatedistributions for Jurkat and K-562 cells, respectively, after a 48h incubation Figure7.Cellularuptake(ngrhodiumpermgcellprotein)inMCF-7andHT- 29cellsfor1.0mmextracellularconcentrationsofthecomplexesfac/mer- Figure8.StateofJurkatcellsbasedonAnnexinV/PIbindingassayafter [RhCl(DMSO)(5,6-Mephen)]8andthemerisomerof[RhCl(DMSO)(phen)] 3 2 3 treatmentwithcomplexes5and8for48h.Cellswerealsotreatedwith 6[10]followinganincubationperiodof6h. 10mmcamptothecin(CMPT)andDMSOascontrols. tion in MCF-7 and HT-29 cells was, in fact, somewhat lower with the 5,6-Mephen complexes 5 and 8 at selected concen- 2 than for the mer isomer of [RhCl(DMSO)(phen)] 6,[10] following trations. Necrosis was negligible in both cases. The complexes 3 a 6h exposure to a 1.0mm solution of the compound. Taking caused significant inhibition of Jurkat cell proliferation at con- the mean cellular diameter and mean protein-to-volume ratio centrations in the 1.5–5.0mm range and invoked extensive for HT-29[21] and MCF-7[22] cells into account, we estimate that apoptosis.Effectivelynoviable cells were present after incuba- 1.0ngRhpermg proteincorresponds tomolarconcentrations tion with complex 8 at 2.5 and 5.0mm concentrations. A pro- of 1.9mm in HT-29 and 1.1mm in MCF-7 cells. This means that nounced cytostatic effect was also apparent for the rhodium thecellularmolarconcentrationof8was21.3mminHT-29and complex toward K-562 cells(Figure9), but in this case the ma- ChemMedChem2011,6,429–439 (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim www.chemmedchem.org 433 MED W.S.Sheldricketal. Figure11.PercentagesofviableBJABcellsandhealthyleukocytesafter treatmentwithcomplex8for48h.Valueswerenormalizedwithrespectto thecontrolvalues. Figure9.StateofK-562cellsbasedonAnnexinV/PIbindingassayafter treatmentwithcomplexes5and8for48h.Cellswerealsotreatedwith 10mmcamptothecin(CMPT)andDMSOascontrols. more potent rhodium complex 8 lacks this sensitivity for non- adherentcells. jority of the cells retained their viability even at the highest employed complex concentration of 5.0mm. In contrast to 8, Reactiveoxygenspecies the analogous iridium(III) complex 5 exhibited no anti-prolifer- ative effect toward the K-562 leukemia cells in the concentra- Reactive oxygen species (ROS) are by-products of the aerobic tionrange0.625–5.0mm. metabolism taking place mainly in the mitochondria. Only 1– AnnexinV/PI staining experiments also demonstrated that 3%ofallhealthycellstypicallyexhibithighROSlevels;[26]how- the rhodium complex 8 is more active toward lymphoma ever, high ROS concentrations have a destructive effect on (BJAB) cells than its iridium counterpart 5 (Figures10 and 11). DNA and many proteins and such oxidative stress can trigger Comparativemeasurementswithexvivohealthyleukocytesin- apoptosis.[27,28] Levels of ROS in Jurkat cells were estimated by dicated a high degree of cell selectivity for complex 5 toward fluorescence measurements at 564–606nm following treat- malignant leukemia (Jurkat) and lymphoma (BJAB) cells. As il- ment of cell suspensions with dihydroethidium in the dark for lustrated in Figure10, 93.3 and 88.1% of the leukocytes were 15min. Cells exhibiting fluorescence values more than 3 stan- still viable following 48h incubation with 5 and 7.5mm solu- darddeviations(s)abovethefluorescencevaluexforthemax- tions of the complex, respectively. These values were normal- imum of the Gaussian distribution registered for an untreated izedwithrespecttothevaluemeasuredfortheuntreatedcon- control were defined as possessing “high” ROS levels. Cells trol sample, which was set to 100%. In contrast to complex 5, with fluorescence values in the range between x(cid:2)3s and x+ only 45.5% of the leukocytes were still viable after treatment 3s were classified as exhibiting “low” ROS values. Significant with a 5mm solution of complex 8 (Figure11). Taken together percentages of cells exhibiting high ROS levels were observed with the results for normal HFF-1 and immortalized HEK-293 after an incubation period of 24h with complexes 5 and 8 cells(Table1),theseAnnexinV/PIstainingexperimentsdemon- (Figure12). These cell percentages were dose-dependent and stratedthattheiridiumcomplex5isselectivetowardbothad- increased markedly on prolonging the treatment for another herent cancer cells and malignant non-adherent cells. The 24h for cells in contact with higher concentrations of 5 (2.5 and5.0mm)andlowerconcentrationsof8(0.625and1.25mm). Following the 48h incubation period, correlations were appar- ent for each of the complexes between the percentages of cells exhibiting high ROS levels and both the total number of Jurkat cells andthepercentages thereof thatwere undergoing apoptosis (Figure8). Much lower percentages of cells with high ROS levels were recorded for K-562 cells following a simi- lar 48h treatment with complex 8 at concentrations in the range 0.625–5mm. These varied from 32% for the 0.625mm concentration up to 59% for the 5mm solution. Very low per- centagesofcellsexhibitinghighROSlevels(9–11%)weremea- suredfollowinga48hincubationwithcomplex5. Cellularmetabolism Figure10.PercentagesofviableBJABcellsandhealthyleukocytesfollowing TheresponseofMCF-7cellstocomplexes5and8wasstudied incubationwithcomplex5for48h.Valueswerenormalizedwithrespectto thecontrolvalues. with a cell-based biosensor chip system that has the ability to 434 www.chemmedchem.org (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim ChemMedChem2011,6,429–439 Rhodium(III)andIridium(III)Complexes to that recorded for the 50mm solution. The experimental curvewasinaccordancewiththeinflictionofpermanentcellu- lar damage at the lower concentration. In contrast to 8, the analogous iridium complex 5 clearly had a less pronounced effect on the MCF-7 cells and caused only an approximately 40% decrease in cell impedance during the 37h incubation period at the higher 50mm concentration. A small degree of improvement of the cellular adhesion properties was also ap- parentduringthefinaldrug-freeperiod.Itshould,however,be noted that the IC values toward MCF-7 cells listed in Table1 50 were determined after a much longer incubation period of 96h. Linear extrapolation of the impedance curve from 38 to 96h for the 50mm solution of 5 leads to the prediction of a final level similar to that invoked by 8. As the molecular shape and size of 5 and 8 are effectively identical, it can be postulat- Figure12.Levelsofintracellularreactiveoxygenspecies(ROS)inJurkatcells edthatthesignificantlyslowersubstitutionrateoftheiridium- afterincubationwithcomplexes5and8fora)24handb)48h. (III) complex must be directly or indirectly responsible for its slowerrateofapoptosisinduction. Oxygen consumption is generally indicative of enhanced or monitor the biological impact of a cytotoxic agent by measur- decreased mitochondrial activity. Other oxygen-consuming ing three important parameters of cellular metabolism. These processes are much less efficient and thus unlikely to contrib- parametersarecellularmorphologyandadhesion,oxygencon- ute significantly to the signal. It is apparent from Figure14 sumption, and the extracellular acidification rate. The silicon that the oxygen consumption of the MCF-7 cells was dramati- chip contains interdigitated electrode structures (IDES) for measuringthecellularimpedance,[29]miniature Clark-typeelec- trodes for monitoring the cellular oxygen uptake,[30] and ion- sensitive field effect transistors to record extracellular pH changes.[31] IDES measurements of cellular impedance reflect the insulating properties of the cell membrane. Morphological changes and cellular adhesion properties, including cell–cell and cell–matrix contacts, are also monitored by cell impe- dance. As illustrated in Figure13, the cellular impedance of MCF-7 cells fell rapidly to a level of about only 30% on treat- ment with a 50mm solution of rhodium complex 8 (Figure13). This dramatic decrease corresponds to the induction of cell death.Amarkedlyslowerrateofimpedancedeclinewasregis- teredforthe5mmsolutionof8;however,thedecreasecontin- ued during the drug-free period to finally reach a level similar Figure14.StandardrespirationratesforMCF-7cellstreatedwith5and 50mmsolutionsofcompounds5and8overtheperiod6–43h.Theendof treatmentisindicatedbyaverticalline.Measurementswerecontinuedfor anadditional11hafterremovalofthesubstances.RM=runningmedium. callydecreasedduringthe37htreatmentperiodsatbothcon- centrations of 8. No recovery was observed. Complex 5 clearly had a lesser effect, but the decrease in cellular respiration was once again permanent. The observed lowering in oxygen con- sumptioncorrelatedwellwiththedose-dependentincreasesin intracellular ROS levels that were determined for Jurkat cells followingtreatment. Extracellular acidification isclosely linkedtocellular glycolyt- ic activity. This parameter is chiefly influenced by lactic acid production, which is the waste product of anaerobic metabo- lism. Both 5 and 50mm solutions of 8 invoked a steady de- Figure13.StandardcellimpedanceforMCF-7cellstreatedwith5and crease in the extracellular acidification rate for MCF-7 cells 50mmsolutionsofcompounds5and8overtheperiod6–43h.Theendof (Figure15). No recovery was apparent during the final drug- treatmentisindicatedbyaverticalline.Measurementswerecontinuedfor anadditional11hafterremovalofthesubstances.RM=runningmedium. free period, indicating once again persistent cellular damage. ChemMedChem2011,6,429–439 (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim www.chemmedchem.org 435 MED W.S.Sheldricketal. lined by its slower rate of apoptosis induction in MCF-7 cells (Figure13). Taken together with their long-term stability in aqueous solution, the combination of high selectivity for cancer cells versus normal cells and relatively slow apoptosis induction rates could give iridium complexes such as 5 a defi- niteadvantageaspotentialnovelanticanceragentsincompar- ison with their more toxic rhodium counterparts. We are cur- rently investigating the activity of iridium complexes with other types of phenanthroline substituents with the goal of further improving their selectivity, so as to provide suitable leadsubstancesforinvivotrials. Experimental Section Figure15.StandardextracellularacidificationratesforMCF-7cellstreated with5and50mmsolutionsofcompounds5and8overtheperiod6–43h. Materials and instrumentation: UV/Vis spectra were recorded on Theendoftreatmentisindicatedbyaverticalline.Measurementswerecon- an Analytik Jena SPECORD 200 spectrometer. A Jasco J-715 instru- tinuedforanadditional11hafterremovalofthesubstances.RM=running mentwasusedtomeasureCDspectrafor1:10complex/DNAmix- medium. tures [complex=20mm, DNA concentration in nucleotides= 200mm] in a 10mm phosphate buffer at pH7.2. Liquid secondary ion mass spectrometry (LSIMS) data were registered with a Fisons Although the 50mm solution of 5 also caused a significant de- VG Autospec mass spectrometer employing a cesium ion gun crease in the acidification rate during the treatment period, (17kV) and 3-nitrobenzyl alcohol as the liquid matrix. Bruker DPX 200 and DRX 400 spectrometers were used for 1HNMR spectro- the MCF-7 cells rapidly regained the non-treatment level once scopic characterizations, with chemical shifts being reported as d additionofthecomplexhadbeenstopped. valuesrelativetothesignalofthedeuteratedsolvent.Thesplitting ofprotonresonancesinthereportedspectraaredefinedass=sin- Conclusions glet, d=doublet, t=triplet and m=multiplet. Elemental analyses were performed on a Vario EL instrument (Elementar Analysensys- Therhodiumcomplexfac/mer-[RhCl(DMSO)(5,6-Mephen)]8is teme, Jena, Germany). RhCl·3HO and IrCl·3HO were purchased 3 2 3 2 3 2 significantly more potent toward MCF-7 and HT-29 cells than fromChempur,substitutedphenanthrolineligandsfromAcrosand its phenanthroline analogue. A similar effect is apparent for dimethyl sulfoxide (DMSO) from J.T. Baker. Deuterated solvents were obtained from DEUTERO GmbH; 7,8-dimethyl-dipyrido[3,2- fac-[IrCl(DMSO)(5,6-Mephen)] 5 in comparison to fac-[IrCl- 3 2 3 a:2’,3’-c]phenazine (Medppz) was prepared in accordance with (DMSO)(phen)]. As the order of increasing antiproliferative ac- 2 publishedprocedures.[32] tivity 1 < 2 < 3 < 5 for the iridium complexes parallels that established for the methyl-substituted ligands on their own fac-[IrCl 3 (DMSO)(4-Mephen)] (2): IrCl 3 ·H 2 O (105.8mg, 0.3mmol) (phen < 4-Mephen, 5-Mephen < 5,6-Mephen), it can be and 4-methyl-1,10-phenanthroline (58.3mg, 0.3mmol) were dis- 2 solvedin10mLCHOHandheatedfor2hinthedarkat808C.Fol- postulated that ligand-based properties may be responsible 3 lowing addition of DMSO (64.8mL, 0.9mmol) and further heating fortheobservedimprovementsinactivity.Theantiproliferative for 3h at 808C, the solvent was removed under vacuum. The re- activity of 5 and 8 is much higher toward MCF-7 and HT-29 sultingsolidwasdissolvedin2mLCHOHand,followingprecipita- 3 cells than toward normal HFF-1 or immortalized HEK-293 cells. tion with diethyl ether, the product was dried in vacuum. Yield: As both complexes invoked extensive apoptosis levels that 106.9mg (62.4%); 1HNMR (200MHz, CDCl): d=3.03 (s, 3H, 4- 2 2 were directly correlated to the ROS concentrations in Jurkat Me),3.53(s,6H,DMSO-Me),7.93(d,1H,H3),7.99(d,1H,H8),8.18 leukemia cells after 48h, it can be concluded that oxidative (d, 2H, H5/6), 8.57 (d, 1H, H7), 9.41 (d, 1H, H2), 9.85 (d, 1H, H9); stress must play a central role in initiating the programmed LSIMS: m/z (%): 535 (27) [M(cid:2)Cl]+, 515 (31) [M(cid:2)DMSO+Na]+, 492 (30) [M+H(cid:2)DMSO]+, 457 (100) [M(cid:2)DSMO(cid:2)Cl]+, 422 (81) cell death. Cellular metabolism studies demonstrated that the [M(cid:2)DMSO(cid:2)HCl(cid:2)Cl]+;Anal.calcdforC H ClIrNOS(M=570.9):C inductionofMCF-7celldeathistime-dependentandaccompa- 15 16 3 2 r 31.6%,H2.8%,N4.9%,found:C31.1%,H3.2%,N5.3%. nied by an immediate and steady decrease in cellular oxygen consumption. This suggests that apoptosis is mediated via the fac-[IrCl 3 (DMSO)(5-Mephen)] (3): Preparation as for 2 with 5- methyl-1,10-phenanthroline (58.3mg, 0.3mmol). Yield: 100.0mg intrinsic mitochondrial pathway as previously established for (58.4%); 1HNMR (200MHz, CDCl): d=2.88 (s, 3H, 5-Me), 3.53 (s, othermembersofthisclassofpolypyridylcomplexes.[9] 2 2 6H, DMSO-Me), 7.89 (s, 1H, H6), 8.05 (m, 2H, H3/8), 8.65 (d, 1H, ComparativeAnnexinV/PI stainingexperimentsforleukemia H4), 8.89(d,1H, H7),9.66 (2d,2H, H2/9); LSIMS: m/z (%): 535(14) (Jurkat) and lymphoma (BJAB) cells versus ex vivo healthy leu- [M(cid:2)Cl]+, 516 (16) [M(cid:2)DMSO+Na]+, 493 (16) [M+H(cid:2)DMSO]+, 457 kocytes indicated a high degree of selectivity for the iridium (100) [M(cid:2)DMSO(cid:2)Cl]+, 422 (80) [M(cid:2)DMSO(cid:2)HCl(cid:2)Cl]+; Anal. calcd complex 5 toward the malignant cells. Though considerably for C H ClIrNOS (M=570.9): C 31.6%, H 2.8%, N 4.9%, found: 15 16 3 2 r less potent than its rhodium analogue 8, the iridium complex 31.1%,H3.2%,N4.9%. 5 appears to possess the distinct clinical advantage of a lower fac-[IrCl(DMSO)(4,7-Mephen)] (4): Preparation as for 2 with 4,7- 3 2 levelofgeneraltoxicitytowardnormalcells.Apossibleexplan- dimethyl-1,10-phenanthroline (62.5mg, 0.3mmol). Yield: 89.9mg ation could lie in its relative kinetic inertness, which is under- (51.2%); 1HNMR (200MHz, CDCl): d=2.97 (s, 6H, 4/7-Me), 3.47 2 2 436 www.chemmedchem.org (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim ChemMedChem2011,6,429–439 Rhodium(III)andIridium(III)Complexes (s, 6H, DMSO-Me), 7.86 (dd, 2H, H3/8), 8.26 (s, 2H, H5/6), 9.46 (d, Biologicalinvestigations 2H, H2/9); LSIMS: m/z (%): 471 (100) [M(cid:2)DMSO(cid:2)Cl]+, 436 (84) [M(cid:2)DMSO(cid:2)HCl(cid:2)Cl]+;Anal.calcdforC H ClIrNOS(M=585.0):C Cell cultures: MCF-7 human breast cancer and HT-29 human colon 16 18 3 2 r 32.9%,H3.1%,N4.8%,found:C33.1%,H3.0%,N4.9%. carcinomacellsweremaintainedinDMEMHighGlucose(PAALab- oratories)supplemented withgentamycin (50mgL(cid:2)1)and10% (v/ fac-[IrCl(DMSO)(5,6-Mephen)] (5): Preparation as for 2 with 5,6- 3 2 v)fetalcalfserum(FCS)at378Cunder5%CO andpassagedtwice dimethyl-1,10-phenanthroline (62.5mg, 0.3mmol). Yield: 99.1mg 2 a week according to standard procedures. HFF-1 cells were culti- (56.4%); 1HNMR (200MHz, CDCl): d=2.87 (s, 6H, 5/6-Me), 3.53 2 2 vated in DMEM supplemented with FCS (both PAA) under similar (s, 6H, DMSO-Me), 8.04 (dd, 2H, H3/8), 8.71 (d, 2H, H4/7), 9.65 (d, conditions. For the growth and maintenance of HEK-293 human 2H, H2/9); LSIMS: m/z (%): 549 (28) [M(cid:2)Cl]+, 471 (100) embryonic kidney cells, Dulbecco’s modified Eagle’s medium (Invi- [M(cid:2)DMSO(cid:2)Cl]+, 436 (72) [M(cid:2)DMSO(cid:2)HCl(cid:2)Cl]+; Anal. calcd for trogen, Karlsruhe, Germany) supplemented with 10% fetal bovine C H ClIrNOS·1.5HO (M=612.0): C 31.4%, H 3.5%, N 4.6%, 16 18 3 2 2 r serum and 100UmL(cid:2)1 penicillin and streptomycin was employed found:C31.1%,H3.5%,N5.0%. at 378C under 5% CO. The cells were split and aliquots thereof 2 fac/mer-[RhCl(DMSO)(4,7-Mephen)] (7): mer/cis-[RhCl(DMSO- wereseededin35mmculturedishesthreetimesaweek.Forstud- 3 2 3 kS)(DMSO-kO)] (133.1mg, 0.3mmol) and 4,7-Mephen (62.5mg, iesofnon-adhesivecelllines,humanT-cellleukemiaJurkatcells,K- 2 2 0.3mmol)wereheatedin10mLCHCl for24hat658C.Following 562(humanchronicmyeloidleukemiacellsinblastcrisis)andBur- 3 solvent removal the product was washed with diethyl ether and kitt-like lymphoma (BJAB) cells were employed. The Jurkat and K- dried in vacuum. Yield: 105.5mg (70.9%); 1HNMR (400MHz, 562 cells were purchased from the German Collection of Microor- CDCl): d=2.89 (s, 6H, 4/7-Me, fac), 2.97, 3.01 (2s, 6H, 4/7-Me, ganisms and Cell Cultures (DSMZ, Braunschweig) and maintained 2 2 mer), 3.52, 3.56 (2s, 6H, DMSO, mer), 3.70 (s, 6H, DMSO, fac), 7.68 at 378C in Roswell Park Memorial Institute (RPMI) 1640 medium (d, 2H, H3/8, fac), 7.85 (2d, 2H, H3/8, mer), 8.18 (s, 2H, H5/6, fac), with 10% heat-inactivated FCS (both from PAA). The BJAB cells 8.25(dd,2H,H5/6,mer),9.28(d,2H,H2/9,fac),9.94,10.03(2d,2H, were likewise held at 378C in RPMI 1640 (GIBCO, Invitrogen) sup- H2/9, mer), fac/mer ratio 48:52; LSIMS: m/z (%): 494 (7) [M+H]+, plemented with 10% FCS, penicillin (100UmL(cid:2)1), streptomycin 459(79)[M(cid:2)Cl]+,424(28)[M(cid:2)HCl(cid:2)Cl]+,381(36)[M(cid:2)DMSO(cid:2)Cl]+, (0.1gL(cid:2)1) and l-glutamine (0.56gL(cid:2)1). Twenty-four hours before 346 (100) [M(cid:2)DMSO(cid:2)HCl(cid:2)Cl]+; Anal. calcd for C H ClNORhS thestartoftheassay,cellswereculturedtoaconcentrationof3(cid:3) 16 18 3 2 (M=495.7): C38.8%,H3.7%,N5.6%,found:C38.8%,H4.1%,N 105cellsmL(cid:2)1toensurestandardizedgrowthconditions. r 5.2%. Antiproliferative activity measurements: The antiproliferative effects fac/mer-[RhCl 3 (DMSO)(5,6-Me 2 phen)](8):Preparationasfor7with of the compounds 1–10 on the adhesive cell lines MCF-7, HT-29, 5,6-dimethyl-1,10-phenanthroline (62.5mg, 0.3mmol). Yield: and HEK-293 were determined by established procedures using a 98.4mg (66.2%); 1HNMR (400MHz, CD 2 Cl 2 ): d=2.80 (s, 6H, 5/6- crystal violet assay.[33,34] Cells were suspended in cell culture Me, fac), 2.83, 2.85 (2s, 6H, 5/6-Me, mer), 3.52, 3.56 (2s, 6H, medium (MCF-7: 104 cellsmL(cid:2)1, HT-29: 2850 cellsmL(cid:2)1, HEK-293: DMSO, mer), 3.71 (s, 6H, DMSO, fac), 7.92 (m, 2H, H3/8, mer), 8.09 2000cellsmL(cid:2)1),and100mLaliquotsthereofwereplatedin96-well (dd, 2H, H3/8, fac), 8.71 (d, 1H, H4, mer), 8.79 (d, 2H, H4/7, fac), platesandincubatedat378C/5%CO for48h(HT-29orMCF-7)or 2 8.81 (d, 1H, H7, mer), 9.43 (d, 2H, H2/9, fac), 10.07, 10.16 (2d, 2H, 72h (MCF-7). One plate (t) was subsequently used for determina- H2/9,mer),fac/merratio35:65;LSIMS:m/z(%):459(100)[M(cid:2)Cl]+, tion of the initial cell bio 0 mass.[33,34] Stock solutions of the com- 424(38)[M(cid:2)HCl(cid:2)Cl]+,346(100)[M(cid:2)DMSO(cid:2)HCl(cid:2)Cl]+;Anal.calcd poundsinDMFwerefreshlypreparedanddilutedwithcellculture for C 16 H 18 Cl 3 N 2 ORhS (M r =495.7): 38.8%, H 3.7%, N 5.6%; found: C medium to the desired concentrations (final DMF concentration: 38.9%,H3.6%,N5.4%. 0.1% v/v). The medium in the plates was replaced with medium containing the compounds in graded concentrations (six repli- fac/mer-[RhCl(DMSO)(4,7-(MeO)phen)] (9): Preparation as for 7 3 2 cates).Afterfurtherincubationfor72h(HT-29)or96h(MCF-7and with4,7-dimethoxy-1,10-phenanthroline(72.1mg,0.3mmol).Yield: 128.9mg (80.1%); 1HNMR (200MHz, CDCl): d=3.51 (s, 6H, HEK-293), the cell biomass was determined by crystal violet stain- 2 2 ing. The mean absorption of the initial biomass t plate was sub- DMSO,fac),3.69(s,6H,DMSO,mer),4.22,4.26(2s,6H,OMe,mer), 0 tractedfromthemeanabsorptionofeachexperimentandcontrol. 4.25 (s, 6H, OMe, fac), 7.28, 7.38 (2d, 2H, H3/8, mer), 8.30 (d, 2H, IC values were established as those concentrations causing 50% H3/8,fac),8.31(2s,4H,H5/6,fac/mer),9.42(d,2H,H2/9,fac),9.83, 50 inhibitionofcellproliferation.Resultswerecalculatedfrom2–3in- 9.93 (2d, 2H, H2/9, mer), fac/mer ratio 24:76; LSIMS: m/z (%): 491 dependentexperiments. (48) [M(cid:2)Cl]+, 456 (13) [M(cid:2)HCl(cid:2)Cl]+, 413 (28) [M(cid:2)DMSO(cid:2)Cl]+, 378 (100) [M(cid:2)DMSO(cid:2)HCl(cid:2)Cl]+, 343 (44) [M(cid:2)DMSO(cid:2)2HCl(cid:2)Cl]+; AdditionalMTTassayswereperformedforMCF-7,HT-29andHFF-1 Anal. calcd for C H ClNORhS·0.5HO (M=536.7): C 35.8%, H 16 18 3 2 3 2 r cells following treatment with complexes 5 and 8. The cells were 3.6%,N5.2%;found:C36.0%,H4.2%,N4.8%. plated into 96-well plates (Granier) at 3000 (for MCF-7 and HT-29) fac/mer-[RhCl(DMSO)(Medppz)] (10): Preparation as for 7 with and 5000 cells (for HFF-1) per 200mL of medium in the well, re- 3 2 Medppz (93.1mg, 0.3mmol). Yield: 171.2mg (84.4%); 1HNMR spectively. The cells were cultivated under standard conditions 2 (200MHz, CDCl): d=2.55 (s, 6H, dppz-Me, fac), 2.64, 2.67 (2s, (378C/5% CO, maximum humidity) for 48h prior to treatment 2 2 2 6H,dppz-Me,mer),3.53(s,6H,DMSO-Me,fac),3.75(s,6H,DMSO- with desired concentrations of the compounds under similar con- Me, mer), 7.96 (dd, 2H, H3/8, fac), 8.14, 8.19 (2dd, 2H, H3/8, mer), ditions. After 47h of treatment, MTT reagent {3-(4,5-dimethylthia- 8.14 (s, 2H, H11/14, fac), 8.20, 8.23 (2s, 2H, H11/14, mer), 9.43 (d, zol-2-yl)-2,5-diphenyltetrazolium bromide} (Sigma–Aldrich) was 2H, H2/9, fac), 9.78 (d, 2H, H4/7, fac), 9.85, 9.88 (2d, 2H, H4/7, added at a concentration of 0.5mgmL(cid:2)1 to each sample and the mer), 10.17, 10.27 (2d, 2H, H2/9, mer), fac/mer ratio 39:61; LSIMS: cells were incubated for a further 1h. The medium was removed m/z (%): 561 (48) [M(cid:2)Cl]+, 526 (24) [M(cid:2)HCl(cid:2)Cl]+, 483 (27) and 200mL DMSO per well were added to dissolve the resulting [M(cid:2)DMSO(cid:2)Cl]+, 448 (100) [M(cid:2)DMSO(cid:2)HCl(cid:2)Cl]+, 413 (60) dark-blue formazan crystals at RTwith shaking at 400rpm for 10– [M(cid:2)DMSO(cid:2)2HCl(cid:2)Cl]+; Anal. calcd for C H ClNORhS·(CH)SO 15min in a thermomixer (Eppendorf). The formazan absorbance 22 20 3 4 32 (M=675.8) C 42.7%, H 3.9%, N 8.3%; found: C 43.2%, H 3.9%, N wasmeasuredwithaTecanUltraplatereaderat590nm(reference r 8.3%. filter620nm). ChemMedChem2011,6,429–439 (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim www.chemmedchem.org 437 MED W.S.Sheldricketal. AnnexinV–PI binding assay: For this assay, up to 4(cid:3)105 cells per DMSO was freshly prepared and diluted with cell culture medium well were incubated with 5 and 8 at different concentrations for to 1.0mm with a DMSO concentration of 0.1% (v/v). The cell cul- 48h. Aliquots of cell suspension (300mL) were taken from each turemediumofacellcultureflaskwasreplacedwith10mLofthis well and the cells were resuspended in 50mL AnnexinV binding mediumcontaining8andtheflaskwasincubatedfor6hat378C/ buffer. Then 2.5mL AnnexinV-FITC (Bioscience, San Diego, CA) 5% CO. The cell pellets were isolated, resuspended in 1–5mL 2 were added and the samples were vortexed and incubated for twice-distilled water, lysed by sonication using a sonotrode, and 10min at RT. Following addition of 450mL of binding buffer, the appropriatelydilutedusingtwice-distilledwater.Therhodiumcon- suspensions were transferred to FACS tubes, mixed with 1.5mL PI tentofthesamplewasdeterminedbyatomicabsorptionspectros- (1mm) and further incubated for 5–10min at RT. FACS analyses copy (AAS) using a Vario G graphite furnace AASinstrument (Ana- were performed on a FACSCalibur flow cytometer (Becton Dickin- lytikJena)atawavelengthof343.5nmandabandpassof0.5nm. son,Heidelberg,Germany)usingtheCellQuestTMProanalysissoft- A deuterium lamp was employed for background correction. The ware and an excitation wavelength of 488nm and emission of protein content of separate aliquots was determined by the Brad- 540–606nmforthedetectionofPI. ford method. To correct for matrix effects in AAS measurements, samples and standards were adjusted to the same concentration MeasurementofintracellularROS:JurkatandK-562cellswereincu- by dilution with twice-distilled water (matrix-matched calibration). bated at different concentrations of the compounds 5 and 8. The PriortoAASanalysis,20mLTritonX-100(1%)and20mLnitricacid cellswerecollectedafter24h/48htreatment,centrifugedat200g (13%) were added to each 200mL sample of the cell suspension. (1500rpm) and re-suspended in FACS buffer (D-PBS, Gibco, +1% Matrixcontainingstandardswereobtainingbyadditionofarhodi- BSA, PAA Laboratories). Cell suspensions were treated with DHE umstocksolution(1mgmL(cid:2)1Rhin10%HCl)tountreatedcellsus- (dihydroethidium, Sigma, 5mL of 5mm stock solution per 1mL of pensions.AASprobeswereinjectedatavolumeof20mLintostan- cell suspension containing 106 cells) at RT in the dark for 15min, dard graphite tubes. Drying, pyrolysis, and atomization in the washed once with FACS buffer and immediately analyzed on a graphite furnace were performed according to theconditions pre- FACSCaliburflowcytometerusingtheCellQuestTMProanalysissoft- viously reported for cellular Rh determinations.[17] During the tem- ware. Excitation and emission settings were 488 and 564–606nm perature program, the graphite tube was purged with a constant (FL2 filter), respectively. For analyses of the FACS data, histograms argon gas flow, which was only heated during the zeroing and of cell counts versus fluorescence signal values were plotted for atomization steps. Pyrolysis and atomization temperatures were the treated cells and compared with the Gaussian distribution ob- optimizedpriortotheexperiments.Therecoveryratesofthemet- tained for an untreated control. The fluorescence value x for the allodrugs were determined initially and used for calculation of the distribution maximum and its standard deviation s were deter- finalexperimentalvalues.Themeanintegratedabsorbancesofdu- mined for the control histogram. “High” ROS levels were assigned plicate injections were used throughout the study. Cellular uptake to treated cells with fluorescence values above x+3s and “low” was expressed as ng rhodium per mg cell protein for data ob- ROSlevelstothosewithfluorescencevaluesintherangex(cid:2)3sto tainedfromtwoindependentexperiments. x+3s. Cellularmetabolism:Changesincellularmetabolismandmorpholo- gywereanalyzedusingaBionas2500Sensorchipsystem(Bionas, Acknowledgements Rostock, Germany). Its sensor chip (SC1000) enables continuous measurementofoxygenconsumptionusingoxygen-sensitiveelec- Financial support for this work by the Deutsche Forschungsge- trodes, pH changes of the medium by employing ion-sensitive meinschaft (DFG) within the research unit FOR 630 “Biological field effect transistors, and the impedance between two interdigi- function of organometallic compounds” is gratefully acknowl- tated electrode structures to register the impedance under and acrossthecelllayeronthechipsurface.[29–31]Beforemeasurement, edged. We are grateful to Heike Scheffler for excellent technical cells were seeded on the sensor chip in DMEM (PAA Laboratories, support. E15-883)withpenicillin/streptomycinand10%(v/v)FCS(PAALab- oratories), and incubated in a standard tissue culture incubator at Keywords: anticancer activity·apoptosis·iridium·leukemia· 378C/5% CO and 95% humidity for 24h until 90% confluency 2 rhodium was reached. Sensor chips with cells were then transferred to the Bionas 2500 analyzer in which medium is continuously exchanged in 10min cycles (3min exchange of medium and 7min without [1] G.Sava,T.Giraldi,G.Mestroni,G.Zassinovich,Chem.-Biol.Interact.1983, flow) during which the parameters were measured. The running 45,1–6. medium used during the analysis was DMEM without carbonated [2] N. Falzone, L. Bçhm, J.C. Swarts, C.E.J. VanRensburg, Anticancer Res. buffer (PAN-Biotech, P03-0010) and only weakly buffered with 2006,26,147–152. 1mmHEPESandreducedFCS(0.1%).Fordrugactivitytesting,the [3] P. Mura, A. Casini, G. Marcon, L. Messori, Inorg. Chim. Acta 2001, 312, following steps were followed: (a) 6h equilibration with running 74–80. medium with 3 and 7min stop/flow incubation intervals, (b) 37h [4] L. Messori, G. Marcon, P. Orioli, M. Fontani, P. Zanello, A. Bergamo, G. drug incubation period with substances freshly dissolved in Sava,P.Mura,J.Inorg.Biochem.2003,95,37–46. [5] D.Herebian,W.S.Sheldrick,J.Chem.Soc.DaltonTrans.2002,966–974. medium at indicated concentrations (same stop/flow intervals), [6] S.Sch(cid:4)fer,W.S.Sheldrick,J.Organomet.Chem. 2007,692,1300–1309. and(c)an11hstepinwhichthecellswereagainfedwithrunning [7] S. Sch(cid:4)fer, I. Ott, R. Gust, W.S. Sheldrick, Eur. J. Inorg. Chem. 2007, medium without substances. At the end of each experiment, cell 3034–3046. layers were removed by additionof 0.2% Triton X-100 toobtain a [8] M.A.Scharwitz, I.Ott, R.Gust, A.Kromm,W.S.Sheldrick,J.Inorg.Bio- basicsignalwithoutlivingcellsonthesensorsurfaceasanegative chem.2008,102,1623–1630. control. [9] M.Dobroschke,Y.Geldmacher,I.Ott,M.Harlos,L.Kater,L.Wagner,R. Gust,W.S.Sheldrick,A.Prokop,ChemMedChem2009,4,177–187. Cellularuptake:MCF-7cellsweregrowntoatleast70%confluency [10] M. Harlos, I. Ott, R. Gust, H. Alborzinia, S. Wçlfl, A. Kromm, W.S. Shel- in 175cm2 cell culture flasks. A stock solution of complex 8 in drick,J.Med.Chem.2008,51,3924–3933. 438 www.chemmedchem.org (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim ChemMedChem2011,6,429–439 Rhodium(III)andIridium(III)Complexes [11] R.Bieda,I.Ott,R.Gust,W.S.Sheldrick,Eur.J.Inorg.Chem.2009,3821– [24] R.A.Schlegel,P.Williamson,CellDeathDiffer.2001,8,551–563. 3831. [25] I.Vermes,C.Haanen,H.Steffens-Nakken,C.Reutelingsperger,J.Immu- [12] W.D. McFadyen, P.G.W. Lawrence, I.A.G. Roos, V.A. Leopold, J. Med. nol.Methods1995,184,39–51. Chem.1985,28,1113–1116. [26] C.Batandier,E.Fontaine,C.Keriel,X.M.Leverve,J.Cell.Mol.Med.2002, [13] C.R.Brodie,J.G.Collins,J.R.Aldrich-Wright,DaltonTrans.2004,1145– 6,175–187. 1152. [27] H.-U. Simon, A. Haj-Yehia, F. Levi-Sch(cid:4)fer, Apoptosis 2000, 5, 415–418. [14] A.M. Krause-Heuer, R. Gr(cid:5)nert, S. K(cid:5)hne, M. Buczkowski, N.J. Wheate, [28] P.A.Sandstrom,M.D. Mannie, T.M. Buttke, J.LeukocyteBiol. 1994,55, D.D. LePevelen, L.R. Boag, D.M. Fisher, J. Kasparkova, J. Malina, P.J. 221–226. Bednarski,V.Brabec,J.R.Aldrich-Wright,J.Med.Chem.2009,52,5474– [29] R. Ehret, W. Baumann, M. Brischwein, A. Schwinde, B. Wolf, Med. Biol. 5484. Eng.Comput.1998,36,365–370. [15] B.R.James,R.H.Morris,Can.J.Chem.1980,58,399–408. [30] B.Wolf,M.Brischwein,W.Baumann,R.Ehret,M.Kraus,Biosens.Bioelec- [16] A.Frodl,D.Herebian,W.S.Sheldrick,J.Chem.Soc.DaltonTrans.2002, tron.1998,13,501–509. 3664–3673. [31] M. Lehmann, W. Baumann, M. Brischwein, R. Ehret, M. Kraus, A. [17] M.A.Scharwitz,I.Ott,Y.Geldmacher,R.Gust,W.S.Sheldrick,J.Organo- Schwinde,M.Bitzehofer,I.Freud,B.Wolf,Biosens.Bioelectron.2000,15, met.Chem.2008,693,2299–2309. 117–124. [18] V. Brabec, V. Kleinw(cid:4)chter, J.-L. Butow, N.P. Johnson, Biophys. Chem. [32] R.E.Holmlin,J.A.Yao,J.K.Barton,Inorg.Chem.1999,38,174–189. 1990,35,129–141. [33] I. Ott, K. Schmidt, B. Kirchner, P. Schuhmacher, T. Wiglenda, R. Gust, J. [19] M.Eriksson,B.Nord(cid:6)n,MethodsEnzymol.2001,340,68–98. Med.Chem.2005,48,622–629. [20] P. Annen, S. Schildberg, W.S. Sheldrick, Inorg. Chim. Acta 2000, 307, [34] R.Bieda, M. Dobroschke, A.Triller,I. Ott,M. Spehr,R.Gust, A.Prokop, 115–124. W.S.Sheldrick,ChemMedChem2010,5,1123–1133. [21] I.Ott,H.Scheffler,R.Gust,ChemMedChem2007,2,702–707. [22] R. Gust, B. Schnurr, R. Krauser, G. Bernhardt, M. Koch, B. Schmid, E. Hummel,H.Schçnenberger,J.CancerRes.Clin.Oncol.1998,124,585– Received:November30,2010 597. Revised:January11,2011 [23] V.A.Fadok,D.Xue,P.Henson,CellDeathDiffer.2001,8,582–589. PublishedonlineonFebruary17,2011 ChemMedChem2011,6,429–439 (cid:2)2011Wiley-VCHVerlagGmbH&Co.KGaA,Weinheim www.chemmedchem.org 439