Three-dimensional cell culture models for metallodrug testing: induction of apoptosis and phenotypic reversion of breast cancer cells by thetrans-[Ru(PPh₃)₂(N,N-dimethyl-N-thiophenylthioureato-k₂O,S)(bipy)]PF₆complex

INORGANIC CHEMISTRY FRONTIERS RESEARCH ARTICLE Three-dimensional cell culture models for metallodrug testing: induction of apoptosis and Citethis:DOI:10.1039/d0qi00502a phenotypic reversion of breast cancer cells by the trans N N N -[Ru(PPh ) ( , -dimethyl- -thiophe- 3 2 † nylthioureato-k O,S)(bipy)]PF complex 2 6 AmandaB.Becceneri, aAngelinaM.Fuzer, aAnaM.Plutin,bAlzirA.Batista, c SophieA.Lelièvre dandMarciaR.Cominetti *a Manystudieshaverevealedtheadvantagesofusingthree-dimensional(3D)cultureovertraditionaltwo- dimensional (2D) monolayer techniques. The 3D cell culture models represent biologically relevant approachestobettermimicthetumororganizationobservedinvivo.Thesemodelshavehighpotential for anticancer drug development and screening, for which challenges include tumor cell resistance to treatmentandtheriskoftoxicityforpatients.Manyanticancerdrugcandidatesdonotreachclinicaltrials astheyfailpreclinicaltestsinvivo,althoughtheywerepromisingin2Dcultures.Modelsfrom3Dcellcul- turesareproposedasintermediatescreeningfiltersbetween2Dandinvivoassays.Ithasbeensuggested thatrutheniumcomplexeshavegreatpotentialforbreastcancertreatment.Here,wetestedthetrans-[Ru (PPh ) (N,N-dimethyl-N-thiophenylthioureato-k O,S)(bipy)]PF complex in breast cancercell lines using 32 2 6 different3Dculturetechniques,includingtheembedded,‘ontop’anddisease-on-a-chip(DOC)models thatreproducephysicalaspectsofthetumormicroenvironment.Thecomplexhadpronouncedbutdis- tinctcytotoxiceffectsontumorsculturedincollagenIofappropriatestiffnessforthediseasestage;the extentofinducedapoptosisdependsonthepreinvasive(S2cells)andinvasive(T4-2andMDA-MB-231 cells) nature of triple-negative breast cancer models. Remarkably, in the DOC model, which simulates breastductalarchitecture,thecomplexwascytotoxicforT4-2tumorsbuthadnoremarkableeffecton thedifferentiatedluminalepithelialS1 cells,demonstratingselectivityagainst cancercells. Inaddition,a lowerconcentrationofthecomplexabrogatedthemalignantphenotypeoftheT4-2cellswithreduction Received1stMay2020, ofEGFR,p50NFκB,andβ1-integrinexpression.Tothebestofourknowledge,thisworkuniquelydemon- Accepted24thJune2020 stratestheeffectsofarutheniumcomplexontheinductionofapoptosisandonthephenotypicreversion DOI:10.1039/d0qi00502a of tumor cells in 3D cultures. These results warrant moving to in vivo evaluation of this ruthenium rsc.li/frontiers-inorganic complex. Introduction models for studying diseases in vitro by better mimicking the tumor microenvironment observed in vivo.1,2 The 3D cell New types of three-dimensional (3D) cell culture systems are cultureapproachhasthepotentialtoimprovethedrugscreen- being explored, since they represent biologically relevant ing pipeline, notably by reducing the numberof false-positive results,thuskeepingonlythemostpromisingdrugcandidates for costly and time-consuming clinical studies.2–4 Moreover, aDepartmentofGerontology,FederalUniversityofSãoCarlos,Rod.WashingtonLuís, animal models, even when humanized, like patient-derived Km235,SãoCarlos,SãoPaulo,13565-905,Brazil.E-mail:mcominetti@ufscar.br xenografts, have limitations such as the misrepresentation of bFacultaddeQuímica,UniversidaddelaHabana,Zapatas/nentreGyCarlitos Aguirre,10400Habana,Cuba thehumantumormicroenvironment.Thisgapmightbefilled cDepartmentofChemistry,FederalUniversityofSãoCarlos,Rod.WashingtonLuís, by relevant3D cell culture models.5,6 Standard 3D cell culture Km235,SãoCarlos,SãoPaulo,13565-905,Brazil of tumors now includes matrices of the defined stiffness dDepartmentofBasicMedicalSciencesandCenterforCancerResearch,Purdue degreestobettermimicthosespecifictodifferenttumordevel- University,625HarrisonStreet,WestLafayette,Indiana,47907,USA †Electronic supplementary information (ESI) available. See DOI: 10.1039/ opment stages.7 Sophisticated 3D cell culture models include d0qi00502a organ-on-a-chip8 and disease-on-a-chip (DOC)6,7 in which Thisjournalis©thePartnerOrganisations2020 Inorg.Chem.Front. .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online View Journal ResearchArticle InorganicChemistryFrontiers several cell types are present. Tissue chips usually require knowledgeontheefficacyofthiscomplexusingdifferenttypes engineeringofthecellcultureplatform.Forinstance,theDOC of 3D cell culture approaches, including the embedded, ‘on consists of acrylic-curved hemichannels to reproduce the geo- top’ and DOC models that reproduce physical aspects of the metry of the mammary duct in which the epithelial cells line tumor microenvironment. The complex exhibited selective the channels and the tumors grow in contact with the cytotoxiceffectsontumorcellscomparedtothedifferentiated epithelium.7,9 breastepithelium,withagradationinthedegreeofapoptosis The development of newanticancerdrugs hasbeen achal- depending on the breast cancer stage recapitulated in 3D cul- lenging endeavor.10 The use of chemotherapeutic agents tures.Itspowerfuleffectoncancercellswasfurtherillustrated usuallybringssevere adverseeffects andfastacquiredcellular by its capability to revert the malignant phenotype of invasive resistance to treatment (and no possibility to increase the ductalcarcinoma(IDC)cellsatlowconcentrations. dosing because of side effects); notably, the inability to eradi- cate all tumor cells facilitates the emergence of resistant cells.11,12 It is increasingly recognized that the tumor micro- Results and discussion environment contributes to tumor characteristics known to influence the response to cytotoxic drugs. The search forcan- The cytotoxicity of the Ru(ThySMet) complex was investigated didate compounds to develop new cytotoxic drugs for cancer in non-neoplastic epithelial and triple-negative IDC breast cells is hindered by limitations in the preclinical assays that cells. Initially, standard 2D culture and 3D ‘on top’ culture failtoreproducenecessarytumorcharacteristics. were used with the MTT method for comparison. Expectedly, Apoptosis is a type of cell death that may be specifically theresultsshowedthatthecellsculturedin2Daremoresensi- induced in response to the cytotoxicity of a compound. This tive tothe treatment compared to the3D cellculture; remark- typeofdeathresultsinregulatedcelldeaththatdoesnotgene- ably, the selectivity index was higher in 3D cultures (Table 1). rate local inflammation, unlike other types of cell death, such This is very interesting because avoidingside effects linked to as necrosis. Other types of regulated cell death are also being induced cell death in normal tissues is a quest foranticancer explored as targets for new anticancer drugs. In the case of drug development. The increased selectivity of Ru(ThySMet) apoptosis, inflammation does not occur because the cell frag- forcancercellsislikelyduetoitsabilitytobindtotransferrin mentsproducedattheendoftheprocess,knownasapoptotic receptors that are overexpressed in certain types of breast bodies, are rapidly phagocytosed by immune cells.13–16 The tumor cells.23 However, whether the expression or availability determination of the type of cell death induced byanticancer of these receptors is increased in the 3D cell culture remains candidates in appropriate culture models is an important tobeinvestigated. aspectofthedrugevaluationprocess. Morphologically, the treatment of the multicellular struc- Transition metals have single characteristics that can be tureswiththeRu(ThySMet)complexpromoteddrasticchanges exploited for the design of new antitumor drugs,17–19 and in in IDC cultures, resulting in smaller and fragmented T4-2 thelastfewdecades,therehasbeengrowinginterestinmetal- structures, in contrast to S1 structures that appeared mostly lodrugs in cancer therapy.20 Specifically, ruthenium(II) com- unaffected by the treatment (Fig. 1A and B). For the highest plexes have potential for breast cancer treatment, since they drugconcentration(64µM),40%oftheT4-2cellsstainedwith havealreadydemonstratedantiproliferative,antiadhesive,anti- propidium iodide (PI), whereas this percentage was only 16% migratory, anti-invasive, and proapoptotic effects on different in S1 cells (Fig. 1C), indicating that the complex is more cyto- breast cancer cells in assays performed in 2D cultures. Also, toxictoIDCT4-2cellsthantonon-neoplasticS1cells.Only5% preclinical in vivo results are promising as ruthenium com- and4%ofuntreatedT4-2andS1cells(C−)weredead,respect- plexesinhibitbreasttumorgrowthandmetastasisindifferent ively. There was a reduction in the T4-2 nodule sizes at the animalmodels.21 highest drug concentrations compared to the untreated In previous work, we evaluated the cytotoxicity of new ruthenium complexes containing acylthiourea ligands against triple-negative malignant breast cells MDA-MB-231 and non- Table1 Cytotoxic activity of the Ru(ThySMet) complex in 2D and 3D neoplastic breast epithelial cells MCF-10A. The trans-[Ru cell cultures of breast non-neoplastic and cancer cells. Cells were (PPh 3 ) 2 (N,N-dimethyl-N-thiophenylthioureato-k 2 O,S)(bipy)]PF 6 treated with different concentrations [0–128 µM] of Ru(ThySMet) for complex (where bipy is 2,2′-bipyridine), hereafter called Ru 24 h. SEM = standard error of the mean; SI = selectivity index of the complex (ThySMet), was cytotoxic (IC 8.81 ± 0.81 µM), inhibited 50 migration, invasion, adhesion, changed morphology and IC (µM)±SEM inducedapoptosisthroughDNAdamageandnuclearfragmen- 50 tation of MDA-MB-231 cells at lower concentrations and MDA-MB-231 MCF-10A SI T4-2 S1 SI shorter incubation times compared to non-neoplastic 2D 8.81±0.81a 14.82± 1.68a 4.98± 19.19± 3.85 MCF-10Acells.Importantly,theRu(ThySMet)complexdidnot 2.50a 0.44 2.74 induceacutetoxicityinvivo.22 3D 33.40±1.93 >100 >2.99 21.08± >100 >4.74 3.74 According to the potential presented by the Ru(ThySMet) complex in preliminary studies, here we aimed to refine the aPreviouslypublished.22 Inorg.Chem.Front. Thisjournalis©thePartnerOrganisations2020 .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online InorganicChemistryFrontiers ResearchArticle progression series was established from fibrocystic breast tissuesand iscomposed ofdifferentcelllines:thenon-malig- nant S1, the triple-negative preinvasive (ductal carcinoma in situ-DCIS) S2, and the triple-negative invasive (IDC) T4-2 cells.31,32 There are different methods of 3D culture permitting the studyof the effect of the microenvironment on cell sensitivity to anticancer drugs; among them, ‘on top’ and embedded are the commonly used techniques. For the ‘on top’ method, the cells are cultured on the Matrigel®-coated surface to which a dripof10%Matrigel®isadded‘ontop’ofthecellsatthetime of seeding, whereas for the embedded method, the hydrogel Fig.1 The Ru(ThySMet) complex induces morphological alterations (Matrigel®, Collagen I, etc.) surrounds the cells.31,33 Collagen and cytotoxicity in multicellular structures formed by IDC T4-2 cells. Non-neoplasticS1andIDCT4-2cellswereculturedusingthe‘ontop’ matrices are increasingly used as the hydrogel for the method for four days for T4-2 cells and six days for S1 cells before a embedded method due to their ability to simulate the inter- 24-hour treatment with 4–64 µM of Ru(ThySMet). (A) Representative actionsthatoccurwiththeinterstitialmatrixinvivo34andthe imagesofthemulticellularstructurestreatedwithRu(ThySMet)for24h. possibility to tailor the matrix stiffness to cancerous microen- The arrows indicate drastic alterations in tumor morphology. (B) vironmental conditions since tumors have higher stiffness Representative immunofluorescence images of multicellular structures than healthy tissues.35 Interestingly, the increase of stiffness in which calcein-AM appears in green and PI in red. Conversion from greentoredfluorescenceindicatescelldeath.(C)Graphspresentingthe in a tissue can lead to an increase in cytoskeletal tension, percentages of total stained cells in green (live) and red (dead) fluor- hencestimulatingthedevelopmentofatumor,whichdemon- escence. Representative images were acquired at 200× magnification, strates the importance of the microenvironment in tumor scalebar:40µm.Bargraphsshowmean±SEM,andresultsarecom- progression.35 pared with the negative control (C−) (untreated). *p < 0.05 and **p < To further investigate the action of Ru(ThySMet) on 0.001. different stages of triple-negative breast cancer, we used a methodinwhichtheinvasiveIDCT4-2andMDA-MB-231cells were cultured embedded in collagen I with Young’s modulus control(C−)(Fig.1BandC).TheseresultsindicatethattheRu of 2020 Pa and the preinvasive (DCIS) S2 cells were cultured (ThySMet) complex can induce cell death in a concentration- with 133 µg ml −1 laminin-111 in collagen I with Young’s dependentmannerwithselectivityagainstneoplasticcells. modulusof1500Pa,tomimicextracellularstiffnessconditions Several studies evaluating the cytotoxicity of anticancer of these two stages of breast cancer progression.7 The cells drugs reported differences when comparing 3D to traditional were stained with caspase-3 and DAPI, and the percentage of 2D cell cultures, suggesting that 3D assays representing more the cells showing nuclear fragmentation and caspase-3 stain- physiologically relevant culture conditions might be essential ingwascalculated.Intheembeddedmethod,40%oftheDCIS inthesearchfornewdrugs.24–27Thesaturatedcyclometalated populationwasstainedforapoptosismarkercleavedcaspase-3 ruthenium(II) complexes exhibited cytotoxicity against cervical and/orpresentednuclearfragmentation(Fig.2AandD),while cancer HeLa cells both in 2D (0.8–3.6 μM) and 3D (2–5 μM), in the IDC population, the percentages were 88% of T4-2 and with slightly higher IC values in 3D cell cultures. Cisplatin 76% of MDA-MB-231 cells. The difference in percentages 50 was also cytotoxic to HeLacells in 2D cultures,whereas in 3D observedbetweenthetreatedIDCstructuresissignificant (p< the cells were more resistant with IC value, nearly ten-fold 0.005), demonstrating that T4-2 cells are more sensitive to the 50 higher.28 The cytotoxic effects of two ruthenium complexes Ru(ThySMet) complex than the MDA-MB-231 cells in the containing piplartine were also evaluated in human colon embedded method; moreover, the T4-2 tumors seemed to fall cancer HCT116 cells in 2D and 3D models. In 3D, the tumor apart and presented severe DNA damage (Fig. 2B, C and D). cells were less sensitive to [Ru(piplartine)(dppf)(bipy)](PF ) These cells are notoriously less aggressive (e.g., less invasive 62 and [Ru(piplartine)(dppb)(bipy)](PF ) complexes with IC of and without induction of metastasis) than MDA-MB-231 62 50 7.3 and 8.3 μM, respectively, compared to the IC of 1.7 and cells.33 50 5.5 μM in 2D cultures.29 Another study using the [Ru(5-FU) In the ‘on top’ method that uses Matrigel® (with normal (PPh ) (bipy)]PF complex and the same cell line also demon- stiffnesslevel;∼800Pa),67%ofT4-2and65%ofMDA-MB-231 32 6 strated that the treatment can affect the viability of the cells, cells were marked for cleaved caspase-3 and/or presented withIC of1.5μMin2Dand1.7μMin3Danddisruptthe3D nuclear fragmentation, a difference that was not significant 50 structures.30 Our results are in accordance with the literature, (Fig. 2D). However, statistical difference in the percentages of since the Ru(ThySMet) complex had higher IC in 3D com- apoptotic cells (caspase-3 positive and/or with nuclear frag- 50 paredtothe2Dcellcultureassays. mentation) between the two methods was found for the Ru A model to study breast cancer progression in 3D cultures (ThySMet)-treated IDC cells, with more pronounced effects in represented by the HMT-3522 series enabled us to further the ‘embedded’ method. Thus, it is possible that a matrix delineatetheactivityoftheRu(ThySMet)complex.Thiscancer stiffness appropriate with the cancer stage enables a clearer Thisjournalis©thePartnerOrganisations2020 Inorg.Chem.Front. .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online ResearchArticle InorganicChemistryFrontiers structures(Fig.2F)comparedtothecellsculturedwiththe‘on top’ method (Fig. 2E), suggesting that placing the cells in a propermicroenvironmentalstiffnessprovidesadditionalinfor- mationregardingtheeffectofthedrug. Previously,wedemonstratedthattheRu(ThySMet)complex was able to activate apoptosis in MDA-MB-231 cells in 2D cul- tures.22 Here, global DNA damage and apoptosis were also inducedbytheRu(ThySMet)complexintumorsformedin3D cultures, as observed by γ-H2AX and cleaved caspase-3 immu- nostainings, demonstrating the efficacy of this drug (even if the IC is higher than in 2D cultures) in the presence of 50 elements from the microenvironment that normally boost differentiation,stimulateDNArepairandprotectagainstapop- tosis induction, such as laminin-rich Matrigel used in the ‘on Fig.2 The Ru(ThySMet) complex is differentially cytotoxic depending top’method.36,37 ontheprogressionoftriple-negativebreastcancerandthematrixcon- To the best of our knowledge, there are no other studies ditions. DCIS (S2), IDC T4-2 (low aggressivity) and MDA-MB-231 (high evaluatingtheeffectofrutheniumcomplexesontheinduction aggressivity)cellsweretreatedwith8µMoftheRu(ThySMet)complex for24hoursina3DculturewithMatrigel®‘ontop’orwithincollagenI of apoptosis in 3D cultures. De Grandis et al. assessed the (embedded).Representativeimmunofluorescenceimagesofmulticellu- effectsofphosphine/diiminerutheniumcomplexescontaining larstructuresof(A)S2cellsculturedembeddedincollagenI(1500Pa) lawsoneontheprostatecancercelllineDU-145.38Theyverified foreightdaysandstainedwithanti-cleavedcaspase-3antibody,(B)T4- that one complex was cytotoxic and induced apoptosis in 2D 2andMDA-MB-231cellscultured‘ontop’withMatrigel®forfourdays cultures;however,in3Dcellcultures,theauthorsonlyinvesti- and stained with anti-cleaved caspase-3 antibody, (C) T4-2 and MDA-MB-231cellsculturedembeddedincollagenI(2020Pa)forfour gated cytotoxic effects. Silva et al. also evaluated the effects of days and stained with anti-cleaved caspase-3 antibody, (D) bar graph a ruthenium-based 5-fluorouracil complex on HCT116 cells with the percentages of S2, T4-2, and MDA-MB-231 cells stained for andverifiedcytotoxicityin3Dspheroidsbasedonthecytotoxic cleaved caspase-3 and/or showing nuclear fragmentation after a activity assay with alamarBlue after 72 h of incubation, but 24h-treatmentwith8µMofRu(ThySMet).Representativeimmunofluor- tests to verify the presence of apoptosis were only performed escenceimagesofmulticellularstructuresof(E)T4-2cellsstainedwith γ-H2AXinMatrigel®and(F)T4-2cellsstainedwithγ-H2AXincollagenI in 2D cultures.30 Nevertheless, several studies have demon- (2020Pa).(G)BargraphwiththepercentagesofT4-2cellsstainedwith strated that ruthenium is able to activate apoptosis in γ-H2AX after a 24 h-treatment with 8 µM of Ru(ThySMet). For each MDA-MB-231 cells and cause DNA damage in 2D experiment, 100 cells were analyzed per condition by fluorescence cultures.21,22,39 microscopy(*p<0.005and**p<0.0001).Thestructureswereselected fromrandomfieldsoftheimagesacquiredat400×magnification;scale Tissuechipsarearefinedformofa3Dcellcultureinwhich bar:40µm.NucleiarestainedwithDAPI(blue),cleavedcaspase-3stain- eachcharacteristicofthemodeliscarefullychosentoanswera ingisshowningreen,andγ-H2AXstainingisshowninmagenta.C−= particular scientificquestion bysimulating the aspects of real untreatedcontrol. tissue organization and functions.9,40 The curvature of the extracellular milieu that mimics the geometry of the breast duct in a disease-on-a-chip (DOC) was shown previously to identification of differences between cell lines corresponding influencetheeffectofchemotherapeuticdrugsonT4-2cells.41 to the diverse degrees of aggressiveness within the same We used the hemichannel settings of the DOC to explore the cancertype.Theimportanceofthematrixisfurtherillustrated effect of Ru(ThySMet) on IDC T4-2 cells and on the non-neo- bythefactthattheactivationofcaspase-3and/ornuclearfrag- plastic S1 cells cultured alone or in the presence of cancer mentation was moderate in DCIS S2 cells, those tumors are cells. For the coculture, the T4-2 tumors are in close contact somewhat protected by basal polarity32 compared to IDC withS1cells(ESIVideo1†).Briefly,inthissystemacrylichemi- MDA-MB-231 and T4-2 tumor cells. In this assay, we investi- channelsarecoatedwithlaminin-111beforeculturingS1cells gated the effects of Ru(ThySMet) on apoptosis by evaluating for 10 days to form a polarized epithelium on which tiny thecaspase-3levelsandnuclearfragmentation,whetherother tumorsmadebyT4-2cells(aftertwodaysintheMatrigel®‘on typesofcelldeathareinvolvedremainstobedemonstrated. top’ culture) are deposited and keep growing; here, the cocul- Todeterminethe extentof globalDNAdamageinducedby turewasperformedforthreedays. the Ru(ThySMet) complex, the cells were stainedwith ananti- After24hoursoftreatmentwithRu(ThySMet),therewasno bodyagainst γ-H2AX, a markerof DNA strand breaks. The Ru massiveactivationofcleavedcaspase-3inthedifferentiatedS1 (ThySMet) complex increased the percentage of IDC T4-2 cells epithelium cultured alone in the DOC. In contrast, tumors stained with γ-H2AX in the embedded (70%) and the ‘on top’ formed by T4-2 cells showed sensitivity to the drug in the (53%) culture methods when compared to the control group DOC;indeed,onaverage10%ofthetumornoduleshadapop- (1%)(Fig.2G).Specifically,inthestructuresembeddedincol- toticcellsfor 8µMof theRu(ThySMet)complex (comparedto lagen I therewas intense staining caused byavast numberof 0% in untreated control tumors), when there was no signifi- double-strand breaks, and also a reduction in the size of the cantdifferenceforthepercentagesofapoptoticcellsintheS1 Inorg.Chem.Front. Thisjournalis©thePartnerOrganisations2020 .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online InorganicChemistryFrontiers ResearchArticle Fig.3 The DOC permits the coculture of non-neoplastic epithelium andtumorsinhemichannelswithabreastterminalduct-likecurvature toperformadditionaltestingofRu(ThySMet)cytotoxicity.S1cellswere culturedfor10daysonlaminin-coatedhemichannelsintheDOC;then, tinyT4-2tumors(4–5cells)stainedwithDilwereseededontopofthe S1 epithelium. The coculturewas pursued for three days before treat- ment of half of the DOCs for 24 h with 8 µM of the complex Ru (ThySMet). Representative immunofluorescence images of (A) S1 cells stainedwithanti-cleavedcaspase-3antibody(green)incontrolandRu (ThySMet) treated cells, (B) S1 and T4-2 cells in coculture; T4-2 are stainedwithDil(red),(C)S1+T4-2cellsincocultureincontrolandRu Fig.4 TheRu(ThySMet)complexreorganizesT4-2multicellularstruc- (ThySMet) treated conditions and stained with anti-cleaved caspase-3 ture phenotype. (A) Representative immunofluorescence images of antibody (green; see arrow) in hemichannels. Nuclei are stained with labeling with α6-integrin in T4-2 tumors cultured with the ‘on top’ DAPI (blue). The structures were selected from random fields of the methodwithMatrigel®forfourdays,anduntreated(C−)ortreatedwith imagesacquiredat400×magnification;scalebar:40µm. 8µMofRu(ThySMet)for24hours.(B)Bargraphsofareaandcircularity corresponding to the conditions described in (A). (C) Representative immunofluorescence images of labeling with α6-integrin of T4-2 tumorsculturedwiththeembeddedmethodincollagenI(2020Pa)for fourdaysanduntreated(C−)ortreatedwith8µMofRu(ThySMet).(D) epithelium compared to control (2.66 ± 0.33 for control vs. Bar graphs of area and circularity corresponding to the conditions 4.33±0.33for8µMRu(ThySMet))(Fig.3A–C). described in (C). (E) Representative images of T4-2 (left panel) and S1 Importantly, when cultured on laminin-111 on acrylic on a (rightpanel)structurestreatedwithRu(ThySMet)atthetimeofseeding flatsurface(locatedintheDOCnexttoahemichannel),theS1 forthe‘ontop’methodwithMatrigel®andculturedforfourdays(T4-2) cellsshowedsensitivitytoconcentrationsaslowas2µMofthe and six days (S1) in the presence of increasing concentrations of Ru (ThySMet). (F) Immunofluorescence staining of T4-2 cells (left panel) drug.Thepercentagesofthecellspositiveforcleavedcaspase- andS1cells(rightpanel)forα6-integrin(green)andβ-catenin(reddish), 3 were 1.66 ± 1.20 for control and 15 ± 2.88 for 2 µM Ru and nucleus staining with DAPI (blue). PD98059 (2 µM) was used as a (ThySMet) (data not shown), suggesting that the geometry of reversion control (C+). This figure part is a composite of randomly the hemichannels is an important factor in exquisitely deter- selected multicellular structures. The white arrows show basal (sur- miningunwantedtoxicityofanticancerdrugsonneighboring, rounding the tumor) and lateral (at the cell–cell junction) signs of polarity,instructuresformedbyT4-2cells.Thestructureswereselected phenotypicallynormalepithelialtissues. fromrandomfieldsoftheimagesacquiredat400×magnification;scale Thus, the DOC device that integrates the curved geometry bar:40µm.Barsshowmean±SEM,andresultsarecomparedwiththe of a terminal duct in the mammary epithelium in which the negativecontrol(C−)(untreated).*p<0.001and**p<0.0001. tumorthrivesallowedustodemonstratestrikingdifferencesin sensitivitytoRu(ThySMet)oftumorscomparedtothephenoty- picallynormalepitheliuminthesameculture. The taming effect of antiproliferative drugs might have an increasingthecircularityonlyfortumorsculturedwiththe‘on impact on the morphology of tumors, measurable notably via top’method.Thesealterationsinmorphometrybeingsignsof area (size) and shape (circularity) that are indicators of poten- potentially tamed phenotype for the tumors, we decided to tialchangesinphenotype.TheT4-2tumorsafterthe24h-treat- analyzetheabilityofRu(ThySMet)torevertthemalignantphe- ment with the Ru(ThySMet) complex were stained with α6- notypeoftheT4-2cellsatnon-cytotoxicconcentrations. integrin to delineate the multicellular structures. In the ‘on Phenotypic reversion occurs when disrupted metabolic top’ and embedded methods, the Ru(ThysMet) complex sig- pathwaysarerestoredinmalignantcells.Thisphenomenonis nificantly decreased the size of the T4-2 tumors (Fig. 4A–D). observable only in 3D cultures and, most particularly, in the Interestingly, the circularity of the tumors in the ‘on top’ presence of basement membrane components (like those culturesignificantlyincreasedafterthetreatment (Fig.4Aand included in Matrigel®) that provide differentiation and survi- B), whereas no such effect was detected in the embedded val cues, leading to basolaterallypolarized multicellular struc- method (Fig. 4C and D), suggesting additional taming of the tures in part similar to those established by non-neoplastic tumorphenotypeunderMatrigel®‘ontop’cultureconditions. cells.42,43 This method used with the ‘on top’ culture, with a The complex [Ru(phen)(bzq)(tbtfpip)](PF ) was shown to microenvironmentclosetothatofthenormalbreast (∼900Pa 6 decrease the tumor size and to cause alterations in lung for matrix stiffness and the presence of basement membrane tumors formed by A549Rcells in vitro, following three days of components), was shown to reveal potential for chemicals to treatment (1 and 2 μM).27 The treatment with 8 μM of Ru revert the phenotype of cancer cells towards lower (ThySMet) for one day after the formation of multicellular aggressiveness.42 structures for four days also caused a decrease in size of the To further investigate the effect of Ru(ThySMet) on tumor tumors both for ‘on top’ and embedded cultures, while phenotype,wetreatedT4-2andS1cellsatthetimeofseeding, Thisjournalis©thePartnerOrganisations2020 Inorg.Chem.Front. .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online ResearchArticle InorganicChemistryFrontiers withlowconcentrations(0.5,1and2μM)ofthedruginorder to avoid cytotoxicity, and cultured them for four (T4-2) or six (S1) days. The Ru(ThySMet) complex reverted the malignant phenotypeoftheT4-2cellswiththeformationofbasolaterally polarized structures (as shown with α6-integrin and β-catenin stainings) similar to the features observed with S1 cells. The T4-2 structures appeared smaller and better organized as the concentrations of the drug increased (Fig. 4E and F). In our investigation of the Ru(ThySMet) complex, so far we have observed differences in response to Ru(ThySMet) treatment in IDCT4-2cellsdependingonwhethertheywereinthe‘ontop’ soft Matrigel® culture or embedded in stiffer collagen I, with more sensitivity to the drug in terms of DNA damage and apoptosis for T4-2 tumors under physiologically relevant col- lagen I conditions. Moreover, tailored ECM for DCIS and IDC cellshaverevealedthelesssensitivenatureofDCIScellstothe Fig.5 Ru(ThySMet) treatment decreases the expression of proteins drug, most likely due to their basal polarity32 enabled by involvedinsignaling pathwaysthatdirecttumor phenotype.Non-neo- laminin mixed with collagen I. Whether the higher sensitivity plasticS1andIDCT4-2cellswereculturedusingthe‘ontop’method to Ru(ThySMet) of IDC T4-2 tumors in collagen compared to andtreatedwith0.5–2µMofRu(ThySMet)orPD98059atthetimeof the Matrigel® culture is linked to a protective effect of seedingandculturedforfour(T4-2)orsix(S1)days.(A)Representative westernblottingimagesofβ1-integrin,EGFR,p38MAPK,andNFκBpro- laminin-rich Matrigel® for a portion of the cells in the popu- teins. (B) Bar graph of β1-integrin, EGFR, p38 MAPK, and NFκB levels lation remains to be established, but the strong influence of (mean ± SEM); results are compared with the negative control (C−) thedrugonphenotypicreversioninthepresenceofMatrigel® (untreated).PD98059(2µM)wasusedasapositivecontrolofphenoty- supportsthishypothesis. pic reversion (C+). (C) Inverted image of MMP9 zymography of T4-2 ToconfirmaphenotypiceffectoftheRu(ThySMet)complex cells.S1cellsMMP9gelatinaseactivitywasusedasacontrolreference (C− = untreated; C+ = PD98059; and 2 µM = Ru(ThySMet)). (D) Bar onT4-2cells,weinvestigatedthelevelsofproteinsinvolvedin graphofMMP9foldactivityrelativetoS1cells.Barsshowmean±SEM, the reversion of the malignant phenotype. Indeed, decreased andresultsarecomparedwiththenegativecontrol(C−)(untreated)or levels of β1-integrin and p50 NFκB were quantified after a thepositivecontrol(C+)(treatedwithPD98059).*p<0.05and***p< 24-hourtreatmentwith2µMofthedrug,whiletheexpression 0.001. ofp38-MAPKandEGFRwasslightlydecreased(Fig.5AandB). Phenotypic reversion in the ‘on top’ culture was further vali- dated by zymography assay that revealed a significant complex (Δ-[Ru(bpy) 2 (HPIP)](ClO 4 ) 2 ) alone or in combination reductionof MMP9gelatinaseactivityincellstreatedwiththe with doxorubicin in 2D cultures leads to a decrease of NFκB Ru(ThySMet)complex,bringingthisactivitytothelevelofthat expressionrelatedtoapoptosisinduction.47Moreover,therole measured in S1 cells. This reduction was significantly greater of NFκB in the control of phenotypic reversion was recently whencomparedto T4-2 cellstreatedwithPD98059, aninhibi- highlighted.48,49 The reduction in MMP9 activity by Ru tor of MEK 1, which is upstream of p38 MAPK in the MAPK/ (ThySMet) corroborates this hypothesis because it is known ERKpathway44(Fig.5CandD). thatimpairingNFκBsignalingalsodecreasesthetranscription The literature isscarce regarding the effects of compounds ofMMP9,hencecontributingtothestabilizationoflaminin5, on the reversion of the malignant phenotype in 3D cultures. oneofthemostcriticalstepsinphenotypicreversion.48 One of the few studies available demonstrated that [10]-gin- gerol,aginger-derivedmolecule,canrevertthemalignantphe- notype of T4-2 cells with the downregulation of EGFR and β1- Experimental integrin, and induced apoptosis of T4-2 and MDA-MB-231 cells.45Arecentstudyperformedbyourgroupshowedthatthe Celllinesandmedia limonoid cedrelone also reverted the malignant phenotype of ThebreastcelllinesusedwereS1(humanmammaryepithelial breast cancer cells in 3D cultures through downregulation of non-neoplastic), S2 (human mammary epithelial triple-nega- NFkB and reduction of MMP-9 levels.46 In the present study, tive preinvasive) and T4-2 (human mammary epithelial inva- theRu(ThySMet)complexwasalsoabletorevertthemalignant sive triple-negative cancer) cells that belong to the HMT-3522 phenotypeoftheT4-2cells,downregulatingthesameproteins. breast cancer progression series;32,50,51 and MDA-MB-231 Moreover, the complex induced apoptosis in the malignant (triple-negative breast cancer) and MCF-10A (human populationofT4-2andMDA-MB-231cells.Thedemonstration mammary epithelial non-neoplastic) cells.52 The S2 and T4-2 of apoptosis induction, together with the downregulation of cells were cultured in an appropriate serum-free H14 culture NFκBlevelsduringphenotypicreversionwiththeRu(ThySMet) mediumcontainingDulbecco’sModifiedEagle’sMedium,and complex, suggests that NFκB signaling might regulate both NutrientMixtureF-12(DMEM-F12)supplementedwith250ng phenomena. Indeed, it was reported that the ruthenium(II) ml −1insulin,10µgml −1transferrin,2.6ngml −1sodiumsele- Inorg.Chem.Front. Thisjournalis©thePartnerOrganisations2020 .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online InorganicChemistryFrontiers ResearchArticle nite, 10 nM β-estradiol, 1.4 µM hydrocortisone and 5 µg ml −1 and 5% CO for 20 min and then the medium was carefully 2 prolactin. For S1 cells, the same medium was used with the added on the gel containing the cells. The IDC cells were cul- addition of 5–10 ng ml −1 epidermal growth factor (EGF). The tured forfour days,and the DCIScellswere cultured foreight MDA-MB-231cellsweremaintainedinDMEMcontaining10% days to allow the formation of tumors before treatment with of fetal bovine serum (FBS), L-glutamine (2 mM), penicillin theRu(ThySMet)complexfor24h. (100 UI ml −1) and streptomycin (100 mg ml −1), and the Forphenotypicreversionexperiments,thecellsweretreated MCF-10A cells were cultured in DMEM/F12 containing 5% with the appropriate drugs the Ru(ThySMet) complex or MEK horse serum, EGF (0.02 mg ml −1), hydrocortisone (0.05 mg 1inhibitorPD98059atthetimeofseedingin3Dcultures. ml −1), insulin (0.01 mg ml −1), L-glutamine (2 mM), penicillin (100 UI ml −1) and streptomycin (100 mg ml −1). All the cells 3Dcultureassayusingacrylichemichannels wereincubatedat37°Cwith5%CO inahumidifiedincuba- 2 The 2 × 2 cm chips with laser-carved acrylic hemichannels, tor.TheMEK1inhibitorPD98059(2µM)(Abcam)wasusedas calleddisease-on-a-chip(DOC)7,9fabricatedatthemanufactur- apositivecontrolforphenotypicreversion(C+). ing facility at the Birck Nanotechnology Center were coated Rutheniumcomplex with laminin-111 (133 µg ml −1) and left to dry overnight at 37°C.Then,theS1cellswereseeded(0.69×106)andcultured The Ru(ThySMet) complex was provided by the Laboratory of for 10 days to allow coverage of the cell culture surface and StructureandReactivityofInorganicCompounds,Department differentiation into a polarized epithelium.18 On day seven, of Chemistry, Federal University of São Carlos. The synthesis EGFwasomittedfromtheculturemediumtoallowfulldiffer- and characterization of the complex were previously entiation. Forcoculturewith S1 cells, IDC T4-2 cells were first described.21 cultured separately using Matrigel® with the ‘on top’ method 3Dcellculture in 35 mm cell culture dishes for two days to induce the for- mation of tiny tumors (encompassing few cells). After two Twodifferentstandard3Dculturemethodswereused,the‘on days, the nodules were detached with 2 mg ml −1 of dispase top’ and the embedded methods. Most of the tests were per- (104 μl cm −2), stained with 1 mg ml −1 of Dil (1,1′-dioctadecyl- formedwiththe‘ontop’method,sincethismethodhasbeen 3,3,3′,3′-tetramethylindocarbocyanine perchlorate (‘DiI’; widelyusedforthecultureoftumorsforcytotoxicitytestsand DiIC18(3))), and the nodules (equivalent to 0.16 × 106 cells) it facilitates analysis and quantification from fluorescence were added ‘on top’ of the epithelium formed by S1 cells, as images with multicellular structures in a single plane; also previously described.7 Coculturewas performed for three days multicellular structures may be easily removed from the beforedrugtreatment. matrix.53 Importantly, it is the culture method typically used for phenotypic reversion assays that require basement mem- Cytotoxicityassays branecomponents.Theembeddedmethodwasusedwithcol- lagenItoexploretheeffectofmatrixstiffnessonthetreatment Briefly,5×103cellswereculturedin96-wellplatesusinga2D withthecytotoxicdrugofinterest. approach or the 3D ‘on top’ method. The cells were treated The3D‘ontop’cellculturewasaccomplishedaccordingto with different concentrations of the complex Ru(ThySMet) the previously described protocol.45 Briefly, the plates were (0–128 µM) for 24 h. Afterwards, the cells in the ‘on top’ coated with chilled Matrigel® and incubated for polymeriz- culture were placed on ice and carefully washed with phos- ationfor20minat37°Cand5%ofCO .Then,thecellswere phate buffered saline (PBS). The plates were centrifuged and 2 separated by trypsinization at the time of use, counted and PBS was removed. The medium was simply removed for the seededontothecoatedwells,allowedtositfor15minintothe cells cultured in 2D. The [3-(4,5-dimethylthiozol-2-yl)-2,5- incubator before the addition, drop-by-drop, of medium with diphenyltetrazoliumbromide](MTT)solution(1mgml −1)was 10% Matrigel® ‘on top’ of the cells. The non-neoplastic cells addedtothewellsandtheplateswerekeptat37°Cand5%of were kept in the culture for six days and the neoplastic cells CO for 4 h. Subsequently, the plates were centrifuged again, 2 for four days before treatment with the Ru(ThySMet) complex thepurplecrystalsofformazanweredissolvedbytheaddition for24h. of 100% DMSO, and the absorbance was measured using a The3Dembeddedcellculturewasaccomplishedaccording plate spectrophotometer (Labtech LT4000) at a wavelength of to the protocol previously described,7,9,41 in the presence of 540nm. tunable collagen I (Advanced BioMatrix). A Young’s modulus Live/deadassaywasalsoperformedforthecellsculturedin of 2020 Pa was used for the IDC cells, and 1500 Pa was used 3D.Non-neoplasticS1andIDCT4-2nodulesweretreatedwith fortheductalcarcinomainsitu(DCIS)S2cellsthatweremixed theRu(ThySMet)complex(8,16or32μM)for24handstained withlaminin-111(133µgml −1)beforetheywereseededincol- using calcein-acetoxymethyl diacetylester (calcein AM) (1 µg lagen I. Briefly, the wells were coated with chilled collagen I ml −1) and propidium iodide (PI) (5 µM ml −1) for 1 h in an (14 μl cm −2) and the cells were separated by trypsinization at incubator (37 °C, 5% CO ). Then, the cells were immediately 2 the time of use, counted and seeded on the coated wells of a analyzedusinganautomatedmicroscopeImageXpress®Micro 4-well chambered slide at 0.86 × 106. The 4-well chambered XLS System (Molecular Devices). Nine pictures were taken of slides were maintained in a humidified incubator at 37 °C, randompartsofthewellstoproceedwithquantification. Thisjournalis©thePartnerOrganisations2020 Inorg.Chem.Front. .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online ResearchArticle InorganicChemistryFrontiers Cellmorphologyassay Biosciences), anti-p38 MAPK (1 µg ml −1) and anti-NFκB p105/ p50 (1 µg ml −1) (Abcam) primary antibodies. Secondary anti- Thenon-neoplasticS1andIDCT4-2cellswereculturedforsix andfourdays,respectively,in24-wellplates(0.5×105cellsper bodies were HRP conjugated (anti-mouse or rabbit) (Abcam). well)usingthe‘ontop’method,andtreatedwithdifferentcon- The endogenous loading control was β-catenin. The Clarity™ centrations of the Ru(ThySMet) complex (8, 16 or 32 μM) for Western ECL substrate (BioRad) was used, and membranes 24h. Picturesweretakenusingacamera(Moticam1000– 1.3 were photographed using a ChemiDoc MP imager (BioRad). ThebandswerequantifiedusingImageJsoftware. MP) attached to an inverted microscope (Nikon Eclipse T100), andmorphologywasimagedusingphase-contrastmicroscopy. Zymographyassay Immunostainingandmorphometryanalysis The supernatants of the cell culture plates used for the Fluorescence immunostaining was performed as described westernblotswerecollectedbeforetheremovalofthemulticel- previously.41,53Briefly,thecellswerefixedwith4%paraformal- lularstructures,and15μgofsupernatantproteinsweresepar- dehyde solution and then washed three times with PBS and atedinTGXFastCast10%acrylamidegelscontaining5mgof glycine (50 mM). Afterwards, the slides were blocked in an bovine gelatin under non-reducing conditions. The gels were immunofluorescence solution (IF) with goat serum (10%). washed twice in Triton-X-100 2.5% solution and then incu- Subsequently, immunostaining was carried out using anti- bated for 20 h at 37 °C in buffer (Tris 50 mM, CaCl 5 mM, 2 cleavedcaspase-3(0.17µgml −1)(CellSignalingTechnologies), NaN 0.02%, and ZnCl 1 μM). The gels were stained with a 3 2 anti-α6-integrin (0.1 mg ml −1), anti-β-catenin (0.83 µg ml −1) Coomassie Brilliant Blue solution (0.25% Coomassie brilliant (BD Biosciences) and anti-γ-H2AX (0.83 µg ml −1) (Abcam) blue, 50% isopropanol, and 10% acetic acid) and destained primaryantibodies,followedbythecorrespondingAlexaFluor with an aqueous solution containing 10% methanol and 10% 488-green(40µgml −1)(ThermoScientific),AlexaFluor568-red acetic acid until clear bands were visible. Then, the gels were (40µgml −1)(ThermoScientific),fluorescein(FITC)-conjugated rehydrated in distilled water and photographed using the AffiniPure F(ab′) Fragment Donkey Anti-Rabbit IgG (H+L) ChemiDocMPimager(BioRad),andthebandscorresponding 2 (2 µg ml −1) (Jackson ImmunoResearch) or Goat F(ab′) Anti- to the matrix metalloproteinase 9 (MMP9) were quantified 2 Mouse IgG-F(ab′) , pre-adsorbed (2 µg ml −1) (Abcam) second- usingImageJsoftware. 2 aryantibodies.Nucleiwerecounterstainedwith0.5µgml −1of DAPI (4′,6-diamidino-2-phenylindole) (Thermo Scientific). The Statisticalanalysis slides were photographed using an inverted microscope AllthestatisticalanalyseswereconductedusingPrism6.0soft- (Fluorescence Olympus BX61-TRF5, camera DP73), or a con- ware(GraphPad,SanDiego,CA,USA).Thedifferencesbetween focal multiphoton LSM 780 (Carl Zeiss), and the morpho- groups were investigated by one-way analysis of variance metric analysis of tumors was performed using ImageJ soft- (ANOVA), followed by Tukey’s post hoc test. The IC calcu- ware (https://imagej.nih.gov/ij/download.html). The size and 50 lationsweremadewiththeequationlog(inhibitor)vs.response circularity of at least 50 tumors in random sites were measured.Forcleavedcaspase-3,∼100nucleiundereachcon- –variableslopeusingGraphPadPrism6.0software. ditionperexperimentwereanalyzed,exceptintheDOCmodel withcoculturewhere10noduleswereanalyzed. Conclusions Proteinelectrophoresisandwesternblotting The structuresformed in‘on top’culturesthatwere treatedat TheRu(ThySMet)complexisnotonlycytotoxicfortriple-nega- thetimeofseedingwith0.5–2μMoftheRu(ThySMet)complex tive breast cancer cells organized into tumors in 3D cultures, but it does so more selectively for IDC than for DCIS, while were removed with the addition of cold PBS containing EDTA sparing the non-neoplastic epithelium, as shown with models (5 mM), NaVO (1 mM), NaF (1.5 mM) and protease inhibitor 4 cocktail (Sigma-Aldrich), as described earlier,52 for 20 min at that mimic important characteristics of the mammary tumor microenvironment. Remarkably, the Ru(ThySMet) complex 4 °C. The collected structures were agitated at 4 °C for 5 min revertsthemalignantphenotypeofIDCT4-2cells,asshownby and centrifuged at 850 rpm (Eppendorf® Centrifuge 5804, its action on the morphometryof multicellular structures and radius of rotor 9.55 cm). The pellets formed were lysed with the expression levels of EGFR, p50 NFκB, and β1-integrin that Cell Lytic M (Sigma-Aldrich) and centrifuged for 10 min at contribute to the pathways known to help tumors thrive. 12000 rpm (Eppendorf® Centrifuge 5804, radius of rotor Together, these results support pursuing the investigation of 9.55 cm). Then, thewestern blots were obtained as previously described.22Briefly,15μgofproteinswereseparatedby4–20% Ru(ThySMet)asapotentialanticancerdruginvivo. PAGEMini-PROTEANTGX™Precastgels(BioRad),transferred ontonitrocellulosemembranes(BioRad)andblockedwith1% Casein in Tris-buffered saline (TBS) (BioRad). Afterwards, the Conflicts of interest membraneswereincubatedwithanti-β-catenin(0.25µgml −1), anti-β1 integrin (0.25 µg ml −1), anti-EGFR (0.25 µg ml −1) (BD Theauthorsdeclarethattheyhavenoconflictsofinterest. Inorg.Chem.Front. Thisjournalis©thePartnerOrganisations2020 .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online InorganicChemistryFrontiers ResearchArticle Acknowledgements 12 B. S. Chakraborty, Cancer Drug Development - Key Regulatory Considerations, Health Administrator, 2007, The authors are thankful for the financial support from 20(1–2),29–36. FAPESP (Fundação de Amparo à Pesquisa do Estado de São 13 P.M.Henson,D.L.BrattonandV.A.Fadok,Apoptoticcell Paulo, grants #2017/20055-5, #2014/25121-8, #2018/19342-2, removal,Curr.Biol.,2001,11,R795–R805. #2013/07600-3, #2015/24940-8 and #2017/01287-2), from 14 G.Kroemer,B.DallaportaandM.Resche-Rigon,Themito- CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível chondrial death/life regulator in apoptosis and necrosis, Superior, grant #001) and from Conselho Nacional de Annu.Rev.Physiol.,1998,60,619–642. Desenvolvimento Científico e Tecnológico. We thank the 15 S. Kasibhatla and B. Tseng, Why Target Apoptosis in Bissel Laboratory at the National Berkeley Laboratory – Cancer Treatment?, Mol. Cancer Ther., 2003, 2, 573– California, USA, and the E-signal Laboratory – University of 580. São Paulo, Brazil, for providing the HMT-3522 cells, the 16 L. Galluzzi, I. Vitale, S. A. Aaronson, J. M. Abrams, RahimiLaboratoryattheBirckNanotechnologyCenter-Purdue D. Adam, P. Agostinis, E. S. Alnemri, L. Altucci, I. Amelio, University, West Lafayette, Indiana, USA, for providing the D.W.Andrews,M.Annicchiarico-Petruzzelli,A.V.Antonov, DOC platforms and the Lelièvre Laboratory at Purdue E. Arama, E. H. Baehrecke, N. A. Barlev, N. G. Bazan, Universitywherepartofthestudywasdeveloped. F. Bernassola, M. J. M. Bertrand, K. Bianchi, M. V. Blagosklonny, K. Blomgren, C. Borner, P. Boya, C. Brenner, M. Campanella, E. Candi, D. Carmona- References Gutierrez, F. Cecconi, F. K. M. Chan, N. S. Chandel, E. H. Cheng, J. E. Chipuk, J. A. Cidlowski, A. Ciechanover, 1 D.L.HollidayandV.Speirs,Choosingtherightcelllinefor G. M. Cohen, M. Conrad, J. R. Cubillos-Ruiz, breastcancerresearch,BreastCancerRes.,2011,13,215. P.E.Czabotar,V.D’Angiolella,T.M.Dawson,V.L.Dawson, 2 B. Weigelt, C. M. Ghajar and M. J. Bissell, The need for V. De Laurenzi, R. De Maria, K.-M. Debatin, complex 3D culture models to unravel novel pathways and R. J. DeBerardinis, M. Deshmukh, N. Di Daniele, F. Di identify accurate biomarkers in breast cancer, Adv. Drug Virgilio, V. M. Dixit, S. J. Dixon, C. S. Duckett, DeliveryRev.,2014,69–70,42–51. B. D. Dynlacht, W. S. El-Deiry, J. W. Elrod, G. M. Fimia, 3 Y. Imamura, T. Mukohara, Y. Shimono, Y. Funakoshi, S. Fulda, A. J. García-Sáez, A. D. Garg, C. Garrido, N. Chayahara, M. Toyoda, N. Kiyota, S. Takao, S. Kono, E. Gavathiotis, P. Golstein, E. Gottlieb, D. R. Green, T. Nakatsura and H. Minami, Comparison of 2D- and 3D- L. A. Greene, H. Gronemeyer, A. Gross, G. Hajnoczky, culture models as drug-testing platforms in breast cancer, J. M. Hardwick, I. S. Harris, M. O. Hengartner, C. Hetz, Oncol.Rep.,2015,33,1837–1843. H. Ichijo, M. Jäättelä, B. Joseph, P. J. Jost, P. P. Juin, 4 Y. Fang and R. M. Eglen, Three-Dimensional Cell Cultures W. J. Kaiser, M. Karin, T. Kaufmann, O. Kepp, A. Kimchi, in Drug Discovery and Development, SLAS Discovery, 2017, R. N. Kitsis, D. J. Klionsky, R. A. Knight, S. Kumar, 22,456–472. S. W. Lee, J. J. Lemasters, B. Levine, A. Linkermann, 5 C.G.BegleyandL.M.Ellis,Drugdevelopment:Raisestan- S. A. Lipton, R. A. Lockshin, C. López-Otín, S. W. Lowe, dards for preclinical cancer research, Nature, 2012, 483, T. Luedde, E. Lugli, M. MacFarlane, F. Madeo, 531–533. M. Malewicz, W. Malorni, G. Manic, J.-C. Marine, 6 T. Denayer, T. Stöhr and M. Van Roy, Animal models in S. J. Martin, J.-C. Martinou, J. P. Medema, P. Mehlen, translational medicine: validation and prediction, New P.Meier,S.Melino,E.A.Miao,J.D.Molkentin,U.M.Moll, Horiz.Transl.Med.,2014,2,5–11. C.Muñoz-Pinedo,S.Nagata,G.Nuñez,A.Oberst,M.Oren, 7 A. Chhetri, S. Chittiboyina, F. Atrian, Y. Bai, D. A. Delisi, M. Overholtzer, M. Pagano, T. Panaretakis, M. Pasparakis, R. Rahimi, J. Garner, Y. Efremov, K. Park, R. Talhouk and J. M. Penninger, D. M. Pereira, S. Pervaiz, M. E. Peter, S. A. Lelievre, Cell Culture and Coculture for Oncological M. Piacentini, P. Pinton, J. H. M. Prehn, H. Puthalakath, Research in Appropriate Microenvironments, Curr. Protoc. G.A.Rabinovich,M.Rehm,R.Rizzuto,C.M.P.Rodrigues, Chem.Biol.,2019,11,e65. D. C. Rubinsztein, T. Rudel, K. M. Ryan, E. Sayan, 8 D. Huh, G. A. Hamilton and D. E. Ingber, From 3D cell L.Scorrano,F.Shao,Y.Shi,J.Silke,H.-U.Simon,A.Sistigu, culturetoorgans-on-chips,TrendsCellBiol.,2011,21,745– B. R. Stockwell, A. Strasser, G. Szabadkai, S. W. G. Tait, 754. D. Tang, N. Tavernarakis, A. Thorburn, Y. Tsujimoto, 9 P. A. Vidi, T. Maleki, M. Ochoa, L. Wang, S. M. Clark, B.Turk,T.VandenBerghe,P.Vandenabeele,M.G.Vander J.F.LearyandS.A.Lelievre,Disease-on-a-chip:mimicryof Heiden, A. Villunger, H. W. Virgin, K. H. Vousden, tumorgrowthinmammaryducts,LabChip,2014,14,172– D. Vucic, E. F. Wagner, H. Walczak, D. Wallach, Y. Wang, 177. J. A. Wells, W. Wood, J. Yuan, Z. Zakeri, B. Zhivotovsky, 10 W. N. Hait, Anticancer drug development: the grand chal- L. Zitvogel, G. Melino and G. Kroemer, Molecular mecha- lenges,Nat.Rev.DrugDiscovery,2010,9,253–254. nisms of cell death: recommendations of the 11 R. V. Chari, Targeted cancer therapy: conferring specificity Nomenclature Committee on Cell Death 2018, Cell Death tocytotoxicdrugs,Acc.Chem.Res.,2008,41,98–107. Differ.,2018,25,486–541. Thisjournalis©thePartnerOrganisations2020 Inorg.Chem.Front. .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online ResearchArticle InorganicChemistryFrontiers 17 L. Ronconi and P. J. Sadler, Using coordination chemistry A.A. Batistaand D.P.Bezerra,Novelpiplartine-containing to design new medicines, Coord. Chem. Rev., 2007, 251, ruthenium complexes: synthesis, cell growth inhibition, 1633–1648. apoptosisinductionandROSproductiononHCT116cells, 18 M. Frezza, S. Hindo, D. Chen, A. Davenport, S. Schmitt, Oncotarget,2017,8,104367–104392. D. Tomco and Q. P. Dou, Novel metals and metal com- 30 V. R. Silva, R. S. Correa, L. S. Santos, M. B. P. Soares, plexes as platforms for cancer therapy, Curr. Pharm. Des., A. A. Batista and D. P. Bezerra, Aruthenium-based 5-fluor- 2010,16,1813–1825. ouracil complex with enhanced cytotoxicity and 19 U.Ndagi,N.MhlongoandM.E.Soliman,Metalcomplexes apoptosisinductionactioninHCT116cells,Sci.Rep.,2018, incancertherapy-anupdatefromdrugdesignperspective, 8,288. DrugDes.,Dev.Ther.,2017,11,599–616. 31 P. A. Vidi, M. J. Bissell and S. A. Lelievre, Three-dimen- 20 E. Alessio and L. Messori, The Deceptively Similar sional culture of human breast epithelial cells: the how Ruthenium(III)DrugCandidatesKP1019andNAMI-AHave andthewhy,MethodsMol.Biol.,2013,945,193–219. Different Actions. What Did We Learn in the Past 30 32 A. Rizki, V. M. Weaver, S. Y. Lee, G. I. Rozenberg, K. Chin, Years?,Met.IonsLifeSci.,2018,18,141–170. C. A. Myers, J. L. Bascom, J. D. Mott, J. R. Semeiks, 21 C.P.Popolin,J.P.B.Reis,A.B.Becceneri,A.E.Graminha, L. R. Grate, I. S. Mian, A. D. Borowsky, R. A. Jensen, M.A.P.Almeida,R.S.Correa,L.A.Colina-Vegas,J.Ellena, M.O.Idowu,F.Chen,D.J.Chen,O.W.Petersen,J.W.Gray A. A. Batista and M. R. Cominetti, Cytotoxicity and anti- andM.J.Bissell,Ahumanbreastcellmodelofpreinvasive tumoreffects of new ruthenium complexes on triple nega- toinvasivetransition,CancerRes.,2008,68,1378–1387. tivebreastcancercells,PLoSOne,2017,12,e0183275. 33 O. W. Petersen, L. Ronnov-Jessen, A. R. Howlett and 22 A. B. Becceneri, C. P. Popolin, A. M. Plutin, E. L. Maistro, M. J. Bissell, Interaction with basement membrane serves E. E. Castellano, A. A. Batista and M. R. Cominetti, The torapidlydistinguishgrowthanddifferentiationpatternof trans-[Ru(PPh3)2(N,N-dimethyl-N’-thiophenylthioureato-k normal and malignant human breast epithelial cells, Proc. (2)O,S)(bipy)]PF6 complex has pro-apoptotic effects on Natl.Acad.Sci.U.S.A.,1992,89,9064–9068. triplenegative breastcancercellsand presents lowtoxicity 34 V. V. Artym and K. Matsumoto, Imaging cells in three- invivo,J.Inorg.Biochem.,2018,186,70–84. dimensional collagen matrix, Curr. Protoc. Cell Biol., 2010, 23 M. A. Naves, A. E. Graminha, L. C. Vegas, L. Luna-Dulcey, 1–20. J. Honorato, A. C. S. Menezes, A. A. Batista and 35 M. J. Paszek, N. Zahir, K. R. Johnson, J. N. Lakins, M.R.Cominetti,TransportoftheRutheniumComplex[Ru G. I. Rozenberg, A. Gefen, C. A. Reinhart-King, (GA)(dppe)2]PF6intoTriple-NegativeBreastCancerCellsIs S. S. Margulies, M. Dembo, D. Boettiger, D. A. Hammer FacilitatedbyTransferrinReceptors,Mol.Pharm.,2019,16, and V. M. Weaver, Tensional homeostasis and the malig- 1167–1183. nantphenotype,CancerCell,2005,8,241–254. 24 H. K. Dhiman, A. R. Ray and A. K. Panda, Three-dimen- 36 V. M. Weaver and M. J. Bissell, Functional culture models sional chitosan scaffold-based MCF-7 cell culture for the to study mechanisms governing apoptosis in normal and determination of the cytotoxicity of tamoxifen, malignant mammary epithelial cells, J. Mammary Gland Biomaterials,2005,26,979–986. Biol.Neoplasia,1999,4,193–201. 25 T. Magdeldin, V. Lopez-Davila, C. Villemant, G. Cameron, 37 V. M. Weaver, S. Lelievre, J. N. Lakins, M. A. Chrenek, R. Drake, U. Cheema and M. Loizidou, The efficacy of J. C. Jones, F. Giancotti, Z. Werb and M. J. Bissell, beta4 cetuximab in a tissue-engineered three-dimensional integrin-dependent formation of polarized three-dimen- in vitro model of colorectal cancer, J. Tissue Eng., 2014, 5, sional architecture confers resistance to apoptosis in 1–9. normal and malignant mammary epithelium, Cancer Cell, 26 F. Salehi, H. Behboudi, G. Kavoosi and S. K. Ardestani, 2002,2,205–216. MonitoringZEOapoptoticpotentialin 2Dand3Dcellcul- 38 R. A. De Grandis, P. Santos, K. M. Oliveira, tures and associated spectroscopic evidence on mode of A. R. T. Machado, A. F. Aissa, A. A. Batista, interactionwithDNA,Sci.Rep.,2017,7,2553. L.M.G.AntunesandF.R.Pavan,Novellawsone-containing 27 L. Zeng, Y. Chen, J. Liu, H. Huang, R. Guan, L. Ji and ruthenium(II) complexes: Synthesis, characterization and H. Chao, Ruthenium(II) Complexes with 2-Phenylimidazo anticanceractivityon 2D and3D spheroid modelsof pros- [4,5-f][1,10]phenanthroline Derivatives that Strongly tatecancercells,Bioorg.Chem.,2019,85,455–468. Combat Cisplatin-Resistant Tumor Cells, Sci. Rep., 2016, 6, 39 C.P.PopolinandM.R.Cominetti,AReviewofRuthenium 19449. ComplexesActivitiesonBreastCancerCells,Mini-Rev.Med. 28 H. Huang, P. Zhang, H. Chen, L. Ji and H. Chao, Chem.,2017,17,1435–1441. Comparisonbetweenpolypyridylandcyclometalatedruthe- 40 S. A. Lelievre, T. Kwok and S. Chittiboyina, Architecture in nium(II)complexes:anticanceractivitiesagainst2Dand3D 3D cell culture: An essential feature for in vitro toxicology, cancermodels,Chemistry,2015,21,715–725. Toxicol.InVitro,2017,45(Pt3),287–295. 29 C. O. D’Sousa Costa, J. H. Araujo Neto, I. R. S. Baliza, 41 S. Chittiboyina, R. Rahimi, F. Atrian, M. Ochoa, B. Ziaie R. B. Dias, L. F. Valverde, M. T. A. Vidal, C. B. S. Sales, andS.A.Lelievre,Gradient-on-a-ChipwithReactiveOxygen C. A. G. Rocha, D. R. M. Moreira, M. B. P. Soares, Species Reveals Thresholds in the Nucleus Response of Inorg.Chem.Front. Thisjournalis©thePartnerOrganisations2020 .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online InorganicChemistryFrontiers ResearchArticle Cancer Cells Depending on the Matrix Environment, ACS Down-Regulating PI3 K/AKT Signaling Pathway, Front. Biomater.Sci.Eng.,2017,4(2),432–445. Oncol.,2020,10,141. 42 V. M. Weaver, O. W. Petersen, F. Wang, C. A. Larabell, 48 S. Furuta, G. Ren, J. H. Mao and M. J. Bissell, Laminin P. Briand, C. Damsky and M. J. Bissell, Reversion of the signals initiate the reciprocal loop that informs breast- malignant phenotype of human breast cells in three- specificgeneexpressionandhomeostasisbyactivatingNO, dimensional culture and in vivo by integrin blocking anti- p53andmicroRNAs,eLife,2018,7,1–40. bodies,J.CellBiol.,1997,137,231–245. 49 S. Becker-Weimann, G. Xiong, S. Furuta, J. Han, I. Kuhn, 43 F. Wang, R. K. Hansen, D. Radisky, T. Yoneda, U. D. Akavia, D. Pe’er, M. J. Bissell and R. Xu, NFkB dis- M. H. Barcellos-Hoff, O. W. Petersen, E. A. Turley and rupts tissue polarity in 3D by preventing integration of M. J. Bissell, Phenotypic reversion or death of cancer cells microenvironmental signals, Oncotarget, 2013, 4, 2010– by altering signaling pathways in three-dimensional con- 2020. texts,J.Natl.CancerInst.,2002,94,1494–1503. 50 P.Briand,O.W.PetersenandB.VanDeurs,Anewdiploid 44 G. Chen, M. Hitomi, J. Han and D. W. Stacey, The p38 nontumorigenic human breast epithelial cell line isolated Pathway Provides Negative Feedback for Ras Proliferative and propagated in chemically defined medium, In Vitro Signaling,J.Biol.Chem.,2000,275,38973–38980. Cell.Dev.Biol.,1987,23,181–188. 45 A.M.Fuzer,S.Y.Lee,J.D.MottandM.R.Cominetti,[10]- 51 P.Briand,K.V.Nielsen,M.W.MadsenandO.W.Petersen, Gingerol Reverts Malignant Phenotype of Breast Trisomy7pandmalignanttransformationofhumanbreast Cancer Cells in 3D Culture, J. Cell. Biochem., 2017, 118, epithelial cells following epidermal growth factor withdra- 2693–2699. wal,CancerRes.,1996,56,2039–2044. 46 A. B. Becceneri, A. M. Fuzer, C. P. Popolin, C. d. M. Cazal, 52 P. A. Kenny, G. Y. Lee, C. A. Myers, R. M. Neve, V. d. C. Domingues, J. B. Fernandes, P. C. Vieira and J. R. Semeiks, P. T. Spellman, K. Lorenz, E. H. Lee, M.R.Cominetti,Acetylationofcedreloneincreasesitscyto- M. H. Barcellos-Hoff, O. W. Petersen, J. W. Gray and toxicactivityandrevertsthemalignantphenotypeofbreast M. J. Bissell, The morphologies of breast cancer cell lines cancer cells in 3D culture, Chem.-Biol. Interact., 2020, 316, in three-dimensional assays correlatewith their profiles of 108920. geneexpression,Mol.Oncol.,2007,1,84–96. 47 K. Lin, Y. Rong, D. Chen, Z. Zhao, H. Bo, A. Qiao, X. Hao 53 G. Y. Lee, P. A. Kenny, E. H. Lee and M. J. Bissell, Three- and J. Wang, Combination of Ruthenium Complex and dimensional culture models of normal and malignant DoxorubicinSynergisticallyInhibitsCancerCellGrowthby breastepithelialcells,Nat.Methods,2007,4,359–365. Thisjournalis©thePartnerOrganisations2020 Inorg.Chem.Front. .MP 71:40:9 0202/3/7 no dnalgnE weN fo ytisrevinU yb dedaolnwoD .0202 yluJ 30 no dehsilbuP View Article Online