Accepted Manuscript Nucleolin overexpression in breast cancer cell sub-populations with different stem-like phenotype enables targeted intracellular delivery of synergistic drug combination Nuno A. Fonseca, Ana S. Rodrigues, Paulo Rodrigues-Santos, Vera Alves, Ana C. Gregório, Ângela Valério-Fernandes, Lígia C. Gomes-da-Silva, Manuel Santos Rosa, Vera Moura, João Ramalho-Santos, Sérgio Simões, João Nuno Moreira, PharmD, MSc, PhD PII: S0142-9612(15)00659-6 DOI: 10.1016/j.biomaterials.2015.08.007 Reference: JBMT 17002 To appear in: Biomaterials Received Date: 17 March 2015 Revised Date: 2 August 2015 Accepted Date: 4 August 2015 Please cite this article as: Fonseca NA, Rodrigues AS, Rodrigues-Santos P, Alves V, Gregório AC, Valério-Fernandes Â, Gomes-da-Silva LC, Rosa MS, Moura V, Ramalho-Santos J, Simões S, Moreira JN, Nucleolin overexpression in breast cancer cell sub-populations with different stem-like phenotype enables targeted intracellular delivery of synergistic drug combination, Biomaterials (2015), doi: 10.1016/ j.biomaterials.2015.08.007. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT Nucleolin overexpression in breast cancer cell sub-populations with different stem-like phenotype enables targeted intracellular delivery of synergistic drug combination Short title: Targeting the stem-like phenotype in TNBC. CNC - Center for Neurosciences and Cell Biology, University of Coimbra, Faculty of Medicine (Polo I), Rua Larga, Coimbra, 3004-504, Portugal; 2 FFUC - Faculty of Pharmacy, University of Coimbra, Pólo das M AN US C 1 RI PT Nuno A. Fonseca1,2; Ana S. Rodrigues3,4; Paulo Rodrigues-Santos5,6; Vera Alves5; Ana C. Gregório1,3,7; Ângela Valério-Fernandes1,3,7; Lígia C. Gomes-da-Silva1,2,3; Manuel Santos Rosa5; Vera Moura1,8; João Ramalho-Santos4,9; Sérgio Simões1,2; João Nuno Moreira1,2,* Ciências da Saúde, Azinhaga de Santa Comba, Coimbra, 3000-548, Portugal; 3 PhD Program in Experimental Biology and Biomedicine (PDBEB), Center for Neuroscience and Cell Biology, University of 4 Coimbra, Faculty of Medicine (Polo I), Rua Larga, Coimbra, 3004-504, Portugal, Biology of Reproduction and Stem Cell Group, Center for Neuroscience and Cell Biology, University of Coimbra, Faculty of Medicine (Polo I), Rua Larga, Coimbra, 3004-504, Portugal; 5 Immunology Institute, Faculty of Medicine (Polo I), University of Coimbra, Rua Larga, Coimbra, 3004-504, Portugal; 6 Immunology and Oncology Laboratory, Center for Neuroscience and Cell Biology, University of Coimbra, Faculty of Medicine (Polo I), Rua Larga, Coimbra, 3004-504, Portugal; 7 IIIUC - Institute for Interdisciplinary Research, University of Coimbra, Casa Costa Alemão (Polo II), Rua Dom Francisco de Lemos, Coimbra, 3030-789, Portugal; 8 9 TREAT U, S.A., Parque Industrial de Taveiro, Lote 44, Coimbra, 3045-508, Portugal; Department of Life D Sciences, Faculty of Sciences and Technology, University of Coimbra, Calçada Martim de Freitas, TE Coimbra, 3000-456, Portugal. AC C EP *Corresponding author João Nuno Moreira, PharmD, MSc, PhD Center for Neuroscience and Cell Biology University of Coimbra Faculty of Medicine (Polo I), Rua Larga 3004-504 Coimbra, Portugal Tel: +351 916 885 272 E-mail: jmoreira@ff.uc.pt 1 ACCEPTED MANUSCRIPT Abstract Breast cancer stem cells (CSC) are thought responsible for tumor growth and relapse, metastization and active evasion to standard chemotherapy. The recognition that CSC may originate from non-stem cancer cells (non-SCC) through plastic epithelial-tomesenchymal transition turned these into relevant cell targets. Of crucial importance for successful therapeutic intervention is the identification of surface receptors RI PT overexpressed in both CSC and non-SCC. Cell surface nucleolin has been described as overexpressed in cancer cells as well as a tumor angiogenic marker. Herein we have addressed the questions on whether nucleolin was a common receptor among breast CSC and non-SCC and whether it could be exploited for targeting purposes. Liposomes functionalized with the nucleolin-binding F3 peptide, targeted M AN US C simultaneously, nucleolin-overexpressing putative breast CSC and non-SCC, which was paralleled by OCT4 and NANOG mRNA levels in cells from triple negative breast cancer (TNBC) origin. In murine embryonic stem cells, both nucleolin mRNA levels and F3 peptide-targeted liposomes cellular association were dependent on the stemness status. An in vivo tumorigenic assay suggested that surface nucleolin overexpression per se, could be associated with the identification of highly tumorigenic TNBC cells. This proposed link between nucleolin expression and the stem-like phenotype in TNBC, enabled 100% cell death mediated by F3 peptide-targeted synergistic drug combination, suggesting the potential to abrogate the plasticity and adaptability D associated with CSC and non-SCC. TE Ultimately, nucleolin-specific therapeutic tools capable of simultaneous debulk multiple cellular compartments of the tumor microenvironment may pave the way towards a EP specific treatment for TNBC patient care. Keywords: triple negative breast cancer, cancer stem cells, non-stem cancer cells, AC C nucleolin, targeting. 2 ACCEPTED MANUSCRIPT Introduction Breast cancer is a highly complex disease owing to intrinsic molecular and cellular heterogeneity associated with the tumor microenvironment [1]. The discovery of cancer stem cells (CSC) in solid tumors, as in breast [2], has greatly contributed to the establishment of the cancer stem cell model as a driver of tumor heterogeneity [3]. According to this model, tumor initiating cells (TIC) are a selected subset of CSC, with RI PT increased capacity to generate tumors in vivo [4]. Established in vivo by the limiting dilution assay, a given cell population, selected by any given marker(s), is considered to have a CSC phenotype when they are more tumorigenic (thus TIC-enriched) as compared to other cell sub-populations [4]. Several markers, including CD44, CD24 M AN US C and aldehyde dehydrogenase (ALDH), have successfully been used to identify highly tumorigenic putative CSC sub-populations in breast tumors [2, 5]. The sub-populations of breast cancer cells with stem-like characteristics, with increased tumorigenic capacity and the ability to recapitulate the tumor environment, have been associated with metastization, tumor relapse, poor disease prognosis and active evasion to standard chemotherapy [2, 3, 5, 6]. Overall, CSC represent a relevant therapeutic target aiming at successfully tackle tumor development and drug resistance. Currently, different drugs targeting developmental-associated pathways, such as Notch or Wnt signaling, known to control CSC self-renewal and maintenance D are in clinical development [7]. This includes, for example, inhibitors of γ-secretase (a Notch checkpoint activator), such as MK0752 or RO492909, for the treatment of (NCT00106145) or triple TE advanced negative breast cancer (NCT01238133), respectively [7]. In addition, canonical pathways, including PI3k/Akt signaling, are EP essential for CSC proliferation and survival [8]. A double PI3k/mTOR inhibitor, VS5584, is under clinical development against advanced non-hematologic malignancies and lymphoma (NCT01991938) [9]. However, single drug regimes, targeting AC C specifically cells with CSC phenotype, could be undermined by their plasticity and adaptability, enabling tumors to evade treatments and CSC enrichment [10]. In spite of combination chemotherapy is a widely adopted strategy to overcome drug resistance [11] its efficacy, upon systemic administration, can be limited owing to differences in pharmacokinetics, thus impairing tumor accumulation of the needed drug ratio, essential to hinder growth and proliferation of different cells within a solid tumor [12]. Provided the necessary accessibility to the CSC niche [13], nanotechnology-based strategies, enabling the simultaneous temporal and spatial delivery of drug combinations, targeting different signaling pathways activated in different tumor cells sub-populations, endows great potential to specifically overcome drug resistance. However, success is highly dependent on the identification of surface receptors [14], 3 ACCEPTED MANUSCRIPT preferentially overexpressed in both CSC and non-SCC (non-stem cancer cells) [3]. This is an aspect of primordial importance from a therapeutic standpoint, as it has been demonstrated that CSC can originate from non-SCC in an Epithelial-to-Mesenchymal Transition (EMT) dependent process, fuelling tumor growth [15]. Nucleolin, besides being overexpressed in cancer cells [16], is a marker of angiogenic blood vessels, mediating the anti-angiogenic and anti-tumoral activity of endostatin [17, RI PT 18]. Such features rendered nucleolin as an important target in cancer therapy, reinforced by the further development of several targeting moieties towards this protein [16, 19]. Accordingly, we have recently developed a F3 peptide-targeted liposomal strategy, targeting cell surface nucleolin, for the simultaneous delivery of a synergistic M AN US C combination of the pro-apoptotic C6-ceramide (C6-Cer), an inhibitor of PI3K/Akt signaling, and doxorubicin (DXR), a cornerstone topoisomerase II inhibitor for breast cancer treatment, aiming at promoting cancer cell death [20]. Building on current state-of-the-art, we recognize that identification of surface receptors enabling specific targeting of both CSC and non-SCC will be crucial to provide longterm disease free survival. Exploiting the described nucleolin role in the stemness maintenance of embryonic stem cells [21], as well as its increasing relevance in cancer development [22], the present work aims at assessing the potential of cell surface nucleolin as a target receptor in breast CSC (and non-SCC) for active intracellular D delivery of the F3 peptide-targeted liposomal synergistic DXR/C6-Cer combination, aiming at ablating both breast CSC and non-SCC, strong contributors for tumor TE heterogeneity and drug resistance. Materials EP Materials and methods AC C MCF-7, MDA-MB-231 and MDA-MB-435S cell lines were acquired from ATCC (Virginia, USA). Doxorubicin hydrochloride (DXR) was from IdisPharma (UK). Calcein, 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic Morpholino)ethanesulfonic dehydrate (EDTA), acid (MES), Trizma®Base, acid Disodium (HEPES), 2-(N- ethylenediaminetetraacetate 3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl-2H- tetrazolium bromide (MTT), sodium chloride (NaCl), 3β-hydroxy-5-cholestene-3hemisuccinate (CHEMS) and cholesterol (CHOL) were purchased from Sigma-Aldrich (USA). The lipids 2-dioleoyl-sn-glycero-3-phosphoethanolamine distearoyl-sn-glycero-3-phosphocholine (DSPC), phosphoethanolamine-N-[methoxy(polyethylene (DOPE), 1,2- 1,2-distearoyl-sn-glycero-3- glycol)-2000] (DSPE-PEG2k), 1,2- distearoyl-sn-glycero-3-phosphoethanolamine-N-[maleimide(polyethylene glycol)-2000] 4 ACCEPTED MANUSCRIPT (DSPE-PEG2k-maleimide), L-α-Phosphatidylethanolamine-N-(lissamine rhodamine B sulfonyl) (RhoB-PE), N-hexanoyl-D-erythro-sphingosine (C6-Ceramide) were acquired from Avanti Polar Lipids (USA). F3 (KDEPQRRSARLSAKPAPPKPEPKPKKAPAKK) and the non-specific (NS) peptides were custom synthesized by Genecust (Luxemburg). All other chemicals were of analytical grade purity. RI PT Cell culture Wild-type E14 mESC (E14-wt), derived by Dr. Martin Hooper from the mouse strain 129/Ola, or Oct4-GFP fusion protein-expressing E14 mESC (E14-GFP) [23, 24] were maintained in feed-layer free conditions in KnockOut-Dubelcco’s Modified Eagle’s M AN US C Medium (GIBCO Life Technologies, USA) supplemented with 15% KnockOut Serum Replacement (KSR - GIBCO Life Technologies, USA), 100 U/mL of penicillin and 100 µg/mL of streptomycin (GIBCO Life Technologies, USA), 1% Minimum Essential Medium non-essential aminoacids (Sigma-Aldrich, USA), 1% L-glutamine (2 mM) (GIBCO Life Technologies, USA), 0.1 mM β-Mercaptoethanol (Sigma-Aldrich, USA) in the presence of 10 U/mL of leukemia inhibitory factor (LIF) (Millipore, USA) at 37ºC in an atmosphere of 5% CO2. MCF-7 (luminal) and MDA-MB-231 (triple negative) breast cancer cells lines, and MDAMB-435S cell line were acquired from ATCC (Virginia, USA) and cultured in RPMI 1640 D (Sigma-Aldrich, USA) supplemented with 10% (v/v) of heat-inactivated Fetal Bovine Serum (FBS) (Invitrogen, USA), 100 U/mL penicillin, 100 µg/mL streptomycin (Lonza, TE Switzerland) and maintained at 37°C in a 5% CO2 atmosphere. Cells were routinely tested for mycoplasma contamination and morphology was assessed by microscopy. EP Nucleolin expression levels were verified based on cell-associated fluorescence following incubation with rhodamine-labelled F3 peptide-targeted liposomes, and AC C compared with positive controls our group has been collected in the last nine years. Preparation of Liposomes pH-sensitive liposomes without ceramide were composed of DOPE:CHEMS:DSPC:CHOL:DSPE-PEG2k (4:2:2:2:0.8 molar ratio) and pH-sensitive liposomes incorporating ceramide were composed of DOPE:CHEMS:DSPC:CHOL:DSPE-PEG2k:C6-ceramide (4:2:1:1:0.8:2 molar ratio). Liposomes were prepared by the ethanol injection procedure [25]. Ethanolic lipid mixtures were added to ammonium sulfate buffer (pH 8.5) at 60°C and the resulting liposomes were extruded through 80 nm pore size polycarbonate membranes using a LiposoFast Basic mini extruder (Avestin, Canada). The buffer was exchanged in a Sephadex G-50 gel column (Sigma-Aldrich, USA) equilibrated with Trizma®Base 5 ACCEPTED MANUSCRIPT sucrose (10%) buffer (pH 9.0). Encapsulation of DXR was carried out through ammonium gradient method, upon incubation with liposomes for 1.5 h at 60°C. Nonencapsulated DXR was removed using a Sephadex G-50 gel column equilibrated with 25 mM HEPES, 140 mM NaCl buffer (pH 7.4). Targeted liposomes were prepared by post-insertion of DSPE-PEG2k-F3 conjugate in a micellar form [26]. Briefly, thiolated derivative of F3 peptide was generated by reaction RI PT at room temperature with 2-iminothiolane (Sigma-Aldrich) in 25 mM HEPES, 140 mM NaCl, 1 mM EDTA buffer (pH 8.0) for 1 h in an inert N2 atmosphere. Thiolated derivatives were then incubated overnight at room temperature with DSPE-PEG2kmaleimide micelles in 25 mM HEPES, 25 mM MES, 140 mM NaCl, 1 mM EDTA (pH M AN US C 7.0). Micelles were then added to pre-formed liposomes, at 2 mol% relative to total lipid (TL), and DSPE-PEG2k-Peptide conjugates post-inserted onto the liposomal membrane upon incubation for 1 h at 50°C. For preparation of calcein-loaded liposomes, ammonium sulfate buffer was replaced by a 40 mM calcein solution, and the resulting liposomes were extruded as described above. Calcein excess was removed through a Sephadex-G50 column equilibrated with 25 mM HEPES, 140 mM NaCl buffer (pH 7.4), and the liposomes immediately submitted to the post-insertion procedure as previously described. Additionally, to prepare rhodamine B-tagged liposomes, RhoB-PE lipid was added to D the above lipid mixture (1 mol% of total lipid), and the ethanol solution was added to 25 mM HEPES, 140 mM NaCl buffer (pH 7.4). The resulting liposomes were extruded and TE preceded to post-insertion, as described above. The developed F3 peptide-targeted liposomes presented physico-chemical properties EP compatible with intravenous administration [27]. In fact, F3 peptide-targeted liposomes encapsulating the DXR:C6-ceramide combination, either at 1:1 (p[F3]DC11]) or 1:2 molar ratio (p[F3]DC12), presented a mean size of ≈150 nm, with a polydispersion AC C index of ≈0.1 (a measure of size homogeneity, with optimal values falling below 0.3). These values were within the same range of F3 peptide-targeted liposomes encapsulating only DXR (p[F3]SL) [20]. In addition, the DXR loading was 138.0 ± 6.8 nmol/µmol of lipid and 68.5 ± 5.1 nmol/µmol of lipid for p[F3]DC11 and p[F3]DC12 (containing half of the amount of DXR per liposome), respectively (loading efficiency ≈75%). In addition, liposomes encapsulating the drug combination retained more than 90% of DXR upon incubation in serum for 4 h at 37ºC [20], thus sharing similar stability with F3 peptide-targeted liposomes encapsulating only DXR [19]. PEGylated liposomes similar to the ones used herein have been previously characterized [28]. Electron microscopy and NMR studies demonstrated that they present a uniform lamellar structure with PEG-DSPE grafted onto the surface, and 6 ACCEPTED MANUSCRIPT doxorubicin sulfate crystals entrapped in the inner aqueous core, upon remote loading using ammonium sulfate gradient [28-30]. The presence of DSPE-PEG2k (at 7.5 mol%) renders a hydration layer on the liposomal surface, which contributes to a zetapotential close to neutrality [31]. DSPE-PEG2k also prevents phase separation that could result from the presence of ceramide (alone), thus contributing to the homogeneity of the lipid bilayer [32]. This effect, along with the additional hydration RI PT layer provided by the F3 peptide, could be responsible for the increased stability in terms of size and aggregation properties that we have previously observed [20]. Cellular association of F3 peptide-targeted nanoparticles with putative breast M AN US C cancer stem cells Half-million MCF-7 or MDA-MB-231 breast cancer cells, known to contain functional cancer stem cells [6, 33-35], were incubated with F3 peptide- or non-targeted rhodamine-labelled liposomes, or liposomes targeted by a non-specific peptide, at 0.4 mM of total lipid, for 1 h at 37ºC or 4ºC. After washing, cells were stained aiming at identifying cancer stem cells, as previously described [33]. Briefly, cells were first incubated with anti-CD44-PE/Cy5 antibody [rat IM7 clone] (Abcam, UK) or IgG2b isotype control (Biolegend, USA) for 30 min at 4ºC, in PBS buffer with 1% bovine serum albumin (BSA) and 0.1% sodium azide (PBS-BSA). Cells were then washed D with PBS-BSA and incubated with ALDEFLUOR® reagent (StemCell Technologies, Canada) for identification of aldehyde dehydrogenase (ALDH) activity, according to the TE manufacturer instructions. The cell-associated rhodamine signal was immediately analyzed by flow cytometry in a BD FACSCalibur system (BD Biosciences, USA) and a EP total of 30,000 events were collected. Appropriate controls were used to assure correct AC C compensation of fluorescence signals in each channel. Establishment of mammospheres from sorted sub-populations Mammosphere formation assay was used as a measure of stemness capability of subpopulations isolated from cell lines. Briefly, 2 x 106 MDA-MB-231 or MCF-7 cells were stained with CD44-PE/Cy5 and ALDEFLUOR® reagent as described above in PBS buffer with 1% BSA. Afterwards, sorting of ALDHhi/CD44hi and ALDHlow/-/CD44low/- cells was performed with a BD FACSAria III cell sorter (BD Biosciences, USA), collecting 515% and 15-20% of each selected sub-population, respectively, depending on the cell line tested [33]. Sorted cells were then seeded for mammosphere formation, as previously described [34, 36]. Briefly, 5000 single ALDHhi/CD44hi or ALDHlow/-/CD44low/cells were seeded in 2 mL Mammocult® Medium supplemented with 4 µg/mL of 7 ACCEPTED MANUSCRIPT heparin and 0.5 µg/mL of hydrocortisone (StemCell Technologies, Canada) per well, in low-adhesion 6-well plates (Greiner, Austria). For 1st generation sphere formation, cells were maintained for 10-21 days, depending on the cell line. To assess self-renewal, 1st generation spheres were collected by centrifugation at 115 g for 5 min, and then dissociated with 0.5% Trypsin (Sigma-Aldrich, USA). Five thousand mammospherederived single cells of each population were then seeded as described above for 7 to RI PT 21 days. Mammosphere formation efficiency was assessed upon image acquisition (9 random images per well) using either an Axiovert 200M microscope (5x objective) or an Axiovert 40C coupled to Canon Powershot G10 camera (10x objective), both controlled by Axiovision software (version 4.8.2) (Zeiss, Germany). Image analysis and M AN US C mammosphere counting was performed using Fiji software (US National Institutes of Health). Mammosphere formation efficiency (%) was calculated by the formula [(Number of spheres/ Number of total events)] x 100, where Total events are a sum of the number of mammospheres and single cells. Intracellular delivery to 2nd generation mammosphere-derived single cells Second generation mammospheres of each population from the triple negative MDAMB-231 and luminal MCF-7 breast cancer or MDA-MB-435S cell lines were dissociated as described above to obtain single-cell suspensions, and cellular condition was D evaluated by trypan-blue exclusion assay (a dye that only penetrates cells with damaged cellular membranes [37]). Fifty thousand cells were incubated with 50 µM of TE calcein-loaded liposomes for 1 h at 37°C. After was hing, cells were analyzed by flow cytometry and events were assessed using Cell Quest Pro software (BD Biosciences, EP USA). Forward and side scatter properties (measuring cell size and complexity, respectively) were monitored to ensure that cellular integrity did not change among the AC C tested liposomal formulations (Fig. S3A). Evaluation of mRNA levels of nucleolin and pluripotency transcription factors NANOG and OCT4 Nucleolin, NANOG and OCT4 mRNA levels in both mESC and breast CSC and nonSCC were evaluated. Briefly, E14 mESC were cultured for 72 h, as colonies, in medium either fully supplemented, maintaining pluripotency status (as described in Cell Culture section), in the absence of LIF or in the absence of both LIF and KSR, conditions under which pluripotency is lost. Additionally, 16 x 106 MDA-MB-231 or 24 x 106 MCF-7 cells were stained with CD44-PE/Cy5 and ALDEFLUOR® as described above, and both ALDHhi/CD44hi (CSC) and ALDHlow/-/CD44low/- (non-SCC) sub- 8 ACCEPTED MANUSCRIPT populations were sorted as described in the mammosphere assay. Upon cell collection, total RNA isolation was performed using the TRIzol® reagent (Invitrogen, Life Technologies, USA). A step of DNA cleanup was introduced using DNA-free™ kit (Ambion, Life Techologies, USA) as per manufacturer instructions. Afterwards RNA concentration and quality were determined using NanoDrop 2000 (Thermo Scientific, USA). Samples presenting a 260/280 ratio under 1.8 were discarded. Samples of total RI PT RNA were stored at -80°C until use [38]. cDNA was o btained using the iScript™ cDNA Synthesis kit (BioRad, USA) according to the protocol established from the manufacturer, using a S1000™ Thermal Cycler (BioRad, USA) programmed as follows: 5 min at 25°C; 30 min at 42°C; 5 min at 85°C and ho ld at 4°C for 1 h. Using species- M AN US C specific pairs of primers, nucleolin, NANOG and OCT4 gene expression was quantified by qRT-PCR using β-ACTIN as housekeeping gene for data normalization. The primers (see Table S1) were obtained (http://pga.mgh.harvard.edu/primerbank/) from and a primer acquired bank from data Integrated base DNA Technologies (IDT). SsoFast™ EvaGreen® Supermix (Bio-Rad, USA) was used to perform analysis of samples that were run in CFX96 Touch™ Real-Time PCR Detection System (BioRad, USA). mRNA fold change was calculated using the 2-∆∆Ct method [25]. D Cellular association of F3 peptide-targeted nanoparticles with embryonic stem cells TE For the cellular association studies, E14-wt or E14-GFP mESC cells were cultured for 72 h, as colonies, in medium either fully supplemented, maintaining pluripotency status EP (as described in Cell Culture section), or in the absence of LIF and KSR (conditions under which pluripotency status is lost). Cells were then incubated with 0.4 mM of rhodamine-labelled F3 peptide-targeted or non-targeted liposomes, or liposomes AC C targeted by a non-specific peptide for 1 h, at 4ºC or 37ºC, either as cell suspension or as colonies. Upon washing, cellular association was analyzed by flow cytometry. Nonviable cells were excluded from the analysis using 7-aminoactinomycin D (7-AAD) (Sigma-Aldrich, USA). Assessment of tumorigenic potential of sorted breast cancer cell subpopulations The tumor initiating capacity of sorted sub-populations from triple negative breast cancer cell line (either ALDH/CD44 or cell surface NCL-based selection), was evaluated. Briefly, 8 x 106 MDA-MB-231 breast cancer cells were stained with ALDEFLUOR® and CD44-PE/Cy5, and both ALDHhi/CD44hi (CSC) and ALDHlow/- 9 ACCEPTED MANUSCRIPT /CD44low/- (non-SCC) sub-populations were sorted as described in the mammosphere assay. Additionally, 90 x 106 cells were stained with anti-NCL-AlexaFluor®488 [mouse 364-5 clone] (Abcam, UK) or IgG1k isotype control (Affymetrix, USA) for 30 min at 4oC in PBS buffer with 1% BSA. Non-viable cells were excluded using 7-AAD to ensure that only viable, cell surface nucleolin positive (NCL+) and negative (NCLlow/-) cells were sorted. All cell sub-populations were further resuspended in 1:1 PBS:Extracellular orthotopically inoculated in both contralateral immunocompromised female mice (NOD.Cg-Prkdc scid RI PT Matrix (ECM) (Sigma-Aldrich, USA) mixture and 2000 or 20000 cells were mammary tm1Wjl Il2rg fat pads of /SzJ strain, a.k.a. NOD scid gamma (NSG)), Charles River, France), as previously described [35]. Mice were M AN US C monitored for tumor formation by palpation once-a-week post-inoculation by two independent researchers. Tumor-initiating cell (TIC) frequency was determined by limiting dilution analysis [39] using the L-Calc™ software package (v1.1) (StemCell Technologies, Canada). The animal experiments were approved by CNC ethical committee and Portuguese National Authority (Direcção Geral de Alimentação e Veterinária) and conducted according to accepted standards of animal care (2010/63/EU directive and Portuguese Act 113/2013). Cytotoxicity of F3 peptide-targeted doxorubicin (DXR):C6-ceramide (C6-Cer) D liposomal synergistic combinations against putative breast cancer stem cells The cytotoxic potential of F3 peptide targeted delivery of the synergistic DXR:C6-Cer TE drug combinations was further evaluated. First, to validate C6-Ceramide as valuable drug against putative CSC, impact on cell viability was assessed. Briefly, 0.25 x 106 EP MDA-MB-231, MCF-7 or MDA-MB-435S cells were incubated with indicated concentrations of C6-ceramide for 24 h at 37°C. Cel ls were then double-stained with ALDEFLUOR® reagent, as described by the manufacturer, and 7-AAD (Sigma-Aldrich, AC C USA), as an indicator of cell viability [40]. Cells were immediately analyzed by flow cytometry and the collected events evaluated with Cell Quest Pro software. The cytotoxic potential of F3 peptide-targeted doxorubicin (DXR):C6-ceramide (C6Cer) liposomal synergistic combinations [20] against putative breast cancer stem cells was then assessed. In brief, 2nd generation spheres from MDA-MB-231 cell line adhered to 96-well plates for 4 h, in RPMI 1640 supplemented with 10% FBS. Afterwards, cells were incubated with serial dilutions of DXR-encapsulating liposomes, for 24 h at 37°C/5% CO 2, after which cell culture medium was exchanged for fresh one and the experiment extended up to 96 h. Cell viability was assessed by the resazurin reduction assay, by monitoring absorbance at 570 nm and 600 nm (background) in a Spectramax Gemini EM (Molecular Devices, USA). Cell death was calculated by the 10 ACCEPTED MANUSCRIPT formula [100 – ((Test570-600 – CtrNeg570-600)/(Ctr570-600 – CtrNeg570-600)) x 100)], where Test570-600 is the corrected absorbance for treated cells, Ctr570-600 is the corrected absorbance for untreated controls and CtrNeg570-600 is the corrected absorbance for the negative control. Results RI PT Association of F3 peptide-targeted liposomes with putative breast cancer stem cells Identification of putative breast CSC in MCF-7 and triple negative MDA-MB-231 breast cancer cell lines was carried out using ALDEFLUOR® reagent and CD44 as previously described [5, 15] (Fig. 1A). Accordingly, in order to understand if one could actually M AN US C deliver a payload into identified putative breast CSC, we defined a gating strategy (Fig. 1B) enabling the evaluation of cellular association of F3 peptide-targeted fluorescently labelled liposomes with the different sub-populations expressing various levels of ALDH and CD44 (Fig. 1 C and D). The results clearly indicated that the F3 peptidetargeted liposomes (p[F3]SL) presented 3.2 (MCF-7, Fig. 1E) and 2.6-fold (MDA-MB231, Fig. 1F) higher cellular association with ALDHhi/CD44hi population (CSC) when compared to the ALDH-/low/CD44low/- population (non-SCC), an effect that was dependent on the presence of the F3 peptide. Additionally, F3 peptide-targeted liposomes (p[F3]SL) also associated with ALDHhi/CD44low/- and ALDHlow/-/CD44hi D populations, which might represent intermediate stages in the hierarchical organization of the cancer cell lines (Fig. 1 E and F). Similarly, the same observations were made TE for a MDA-MB-435S cell line (Fig. S1). Furthermore, at 4⁰C, a temperature not permissive to endocytosis, the F3 peptide-targeted liposomes presented lower cellular EP association with the different sub-populations in both cell line models, thus indicating that an energy-dependent internalization was taking place in all of them (Fig. 1 E and F AC C and Fig. S1). Assessment of drug delivery to mammosphere-derived breast cancer stem cells The putative CSC phenotype was validated through the assessment of the in vitro phenotypical characteristics of the selected ALDHhi/CD44hi and ALDHlow/-/CD44low/populations, through evaluation of mammosphere formation and self-renewal (Fig. 1A). Results demonstrated that both isolated sub-populations from each of the cell lines tested were able to form 1st and 2nd mammospheres (Fig. S2 A and B). Strikingly though, while ALDHhi/CD44hi population maintained 2nd generation mammosphere formation efficiency, for both of MCF-7 and MDA-MB-231, ALDHlow/-/CD44low/- lost, in part, their capacity to generate spheres (Fig. S2C). Overall, these results indicated that both populations have different stem 11 and self-renewal potentials, therefore ACCEPTED MANUSCRIPT representing putatively different hierarchical clusters. Notably, F3 peptide-targeted liposomes enabled efficient delivery of encapsulated calcein to single cells derived from 2nd generation mammospheres of both sub-populations (Fig. 1 G and H, Fig. S3). These results reinforce the ability of liposomes functionalized with F3 peptide to target breast CSC and non-stem cancer cells, as well as nucleolin overexpression in RI PT mammospheres from both sub-populations. Nucleolin and pluripotency markers mRNA levels in breast CSC and mESC The results from previous sections suggested that nucleolin expression in breast CSC could be paralleled by the expression of pluripotency genes, also known to be M AN US C upregulated in cancer [41]. To test this hypothesis, we evaluated the pluripotency transcription factors NANOG and OCT4 and, concomitantly, the nucleolin mRNA levels in sorted sub-populations from breast cancer cell lines, as well as in mouse embryonic stem cells (mESC) used herein as phenotypic controls owing to the high conservation of nucleolin among species [42] (Fig. 2A). Indeed, when mESC were cultured in conditions favoring pluripotency loss, there was a decrease of NANOG and OCT4 mRNA levels that were paralleled by nucleolin (Fig. 2B and Fig. S4), in agreement with published data [21]. According to its role in cancer, one could think that nucleolin would be homogenously in cancer cells. Strikingly, MDA-MB-231 putative breast CSC D expressed (ALDHhi/CD44hi) presented 1.5-fold higher nucleolin mRNA level relative to non-SCC TE (ALDHlow/-/CD44low/-) (Fig. 2C). Moreover, the increased levels of nucleolin were paralleled by the overexpression of NANOG and OCT4 in breast CSC (Fig. 2C). In EP spite of following the same trend, the results obtained with MCF-7 cell line were highly variable (Fig. 3D). These results support the enhanced cellular association of F3 peptide-targeted liposomes with putative breast CSC, as well as the increased AC C mammosphere formation efficiency of those as compared to non-SCC (Fig. 1 E and F and Fig. S2C). Overall, the identified putative CSC populations are enriched for stem-like cells, as compared to non-SCC, indicating that nucleolin is in fact associated with the former phenotype. Cellular association of F3 peptide-targeted nanoparticles with embryonic stem cells Besides of its expression in different cellular compartments, nucleolin is also involved in embryonic stem cell self-renewal [18, 21]. Those facts, supported by the results from previous sections, raised the question on whether the nucleolin-mediated cellular 12 ACCEPTED MANUSCRIPT association of F3 peptide-targeted liposomes would be dependent on the stemness status. Accordingly, we evaluated the cellular association of F3 peptide-targeted liposomes after culturing mESC in conditions either impairing or favoring pluripotency (Fig. 3A). F3 peptide-targeted liposomes (p[F3]SL) associated with mESC in an extent significantly higher than non-targeted or non-specific targeted counterparts, and in a RI PT ligand-specific manner (Fig. 3B-D). Furthermore, the association of F3 peptide-targeted liposomes decreased upon incubation at 4ºC, a temperature non-permissive to endocytosis, suggesting that an active internalization through receptor-mediated endocytosis was taking place (Fig. 3 B and C). Strikingly, culturing E14-GFP mESC M AN US C cells without LIF and serum replacement (KSR) (thus inducing pluripotency loss, a condition supported by the decreased levels of the Oct4-GFP fusion protein, Fig. 3E) resulted in a significant reduction in cellular association, to levels close to the ones observed for non-targeted liposomes (Fig. 3C). Performing the experiment with cell colonies, F3 peptide-targeted liposomes associated with E14-GFP mESC cells, nonetheless in a lower extent (6.6-fold) (Fig. 3 B and D) than the one observed with cells in suspension (Fig. 3C). Such results could be explained by the lower accessibility of the targeted liposomes to E14 cells in colony, as well as by the increased surface area available for targeting when the experiment is performed with the cells in D suspension. This was reinforced by the 3.2-fold increase in cellular association obtained for cells grown in absence of LIF and KSR, as compared to standard growth TE conditions, since the resulting colonies were smaller thus facilitating the nanosystem access (Fig. 3D). EP Overall, these results strongly suggest that cell membrane nucleolin levels decrease according to cell pluripotency status, which is accompanied by reduction of Oct4 protein and therefore highly consistent with mRNA levels determination (Fig. 2B), thus AC C revealing that cellular association of F3 peptide-targeted liposomes is stemness statusdependent. Evaluation of the tumorigenic potential of cell surface nucleolin positive cells and putative breast CSC The above observations led one to question whether cell surface nucleolin overexpression could enable the identification of highly tumorigenic cells (Fig. 4A). Tumor development latency analysis revealed that NCL+ cells initiated tumors 1.43-fold faster that NCLlow/- cell sub-population at 2000 inoculated cells (not observed at 20000 inoculated cells) (Fig. 4B). Additionally, ALDHhi/CD44hi cells had increased tumor initiation capacity as compared with ALDHlow/-/CD44low/- (1.32 and 1.37-fold for 2000 13 ACCEPTED MANUSCRIPT and 20000 inoculated cells, respectively) (Fig. 4B), a feature consistent with literature [2, 5]. Strikingly, NCL+ and ALDHhi/CD44hi cell sub-populations demonstrated an increased capacity to generate orthotopic tumors as compared with NCLlow/- and ALDHlow/-/CD44low/- sub-populations, respectively (Table 1). This translated into a higher frequency of TIC within ALDHhi/CD44hi (putative breast CSC) and NCL+ subpopulations compared to the non-SCC (ALDHlow/-/CD44low/-) and NCLlow/- sub- RI PT populations (3.4 and 8-fold respectively, at 6 weeks) (Table 1). Noteworthy, overtime (until 10 weeks post cell inoculation), all sorted populations were able to seed the majority of new tumors (Table 1). Overall, these results suggest that overexpression of cell surface nucleolin per se could be useful for the identification of highly tumorigenic cells. The tumorigenic potential of M AN US C putative CSC, non-SCC and NCL+ tumor cells emphasizes the need to target simultaneously several populations within the tumor microenvironment, aiming at successful therapeutic intervention. In this respect, and based on the results previously presented, liposomes functionalized with the F3 peptide and targeting nucleolin are a drug carrier with great therapeutic potential. Cellular cytotoxicity mediated by F3 peptide-targeted combination of doxorubicin and C6-ceramide In order to overcome drug resistance, often associated with CSC, it has been D recognized that the successful application of small molecules in cancer therapy TE requires the identification of agents that, when combined, lead to synergistic tumor inhibition without significant systemic toxicity [3, 11, 43, 44]. Exploring the nanoparticlemediated simultaneous spatial and temporal delivery of combinations [12] and the EP enhanced specific intracellular delivery [19, 27], herein we evaluated the cytotoxic impact of a F3 peptide-targeted synergistic combination of doxorubicin (DXR) and C6- AC C ceramide, that we have previously developed [20], against putative breast CSC. At the highest concentration tested, C6-ceramide induced a 2.5- and 2.1-fold decrease in the number of viable ALDHhi cells (ALDHhi/7AAD-) from MCF-7 and MDA-MB-231 breast cancer cells, respectively, an effect apparently independent of C6-ceramide concentration (Fig. 5 A and B). CSC were shown to be resistant to DXR action compared to non-SCC [33]. By impairing ALDHhi cell viability, the aforementioned result supported the use of C6-Ceramide and DXR combinations. Accordingly, we evaluated the cytotoxicity of the DXR/C6-ceramide synergistic combination encapsulated in F3 peptide-targeted liposomes against mammospheres, known to better predict in vivo drug responses [36]. 14 ACCEPTED MANUSCRIPT The cytotoxicity results obtained with MDA-MB-231 2nd generation mammospheres indicated that ALDHhi/CD44hi cells were more resistant to F3 peptide-targeted doxorubicin (p[F3]SL(DXR)) than ALDHlow/-/CD44low/- cells. Notwithstanding, targeting these cell sub-populations through the nucleolin receptor, with the liposomal DXR, functionalized with the F3 peptide, enabled an IC90 lower than non-targeted formulation (pSL(DXR)) (Fig. 5 C and D). However, the co-encapsulation of DXR and C6-ceramide RI PT at 1:1 (p[F3]DC11) and 1:2 (p[F3]DC12) molar ratios in the F3 peptide-targeted nanoparticle enabled 100% cell death, while decreasing DXR IC90 (4-fold in case of ALDHhi/CD44hi cells) (Fig. 5 C and D), a condition not achievable with the single drug (DXR)-containing F3 peptide targeted nanoparticle. In addition, it was apparent that F3 peptide-targeted delivery of DXR:C6-Ceramide combination decreased the IC90 of DXR M AN US C to similar values in both sensitive (ALDHlow/-/CD44low/-) and more resistant (ALDHhi/CD44hi) cell sub-populations (Fig. 5D), thus seemingly overcoming putative CSC-associated DXR resistance. Discussion The CSC represent cellular populations with stem-like features responsible for tumor development and heterogeneity, drug resistance and disease relapse [3]. Notwithstanding, the acknowledgement that CSC may originate from non-SCC, interconverting through an EMT-mediated process [15] has turned these cell sub- D populations into two relevant therapeutic targets [3]. Therefore, to specifically tackle the TE disease at its roots, one has to find suitable molecular targets that enable simultaneous targeting of both CSC and non-SCC, provided the necessary accessibility to the CSC niche [3, 13]. EP Nucleolin, thought as homogenously overexpressed by cancer and angiogenic endothelial cells, has been exploited as a molecular target for drug delivery with AC C nanotechnology-based strategies [19]. It was demonstrated herein that a F3 peptidetargeted lipid-based nanoparticle was actively internalized by both breast non-SCC (ALDHlow/-/CD44low/-) and, in a higher extent, putative CSC (ALDHhi/CD44hi) (Fig. 1 E and F), enabling the delivery of the liposomal payload (Fig. 1H) into these subpopulations of cells with different stem-like phenotype (Fig. S2C). The extent of cellular association is consistent with our previous data on cancer cells [19], and therefore we hypothesize that receptor-mediated internalization followed by escape from endocytic route [19, 45], remains valid in CSC. In addition, simultaneous targeting of multiple cancer cell populations introduces a critical feature sought to be essential for next generation of cancer therapy [3]. Those results suggested that nucleolin could be expressed at different densities among those sub-populations. In addition, it is known 15 ACCEPTED MANUSCRIPT that nucleolin [19] and pluripotency markers [41] are expressed in both tumors and breast cancer cells. Nonetheless, the simultaneous upregulation in putative breast CSC has never been described. We have shown an upregulation of mRNA levels of the pluripotency markers NANOG and OCT4, which was paralleled by nucleolin, in triple negative putative breast CSC as compared to non-SCC (Fig. 2C), supporting the cellular association with both cells (Fig. RI PT 1F) and 2nd generation mammosphere-derived single MDA-MB-231 cells (Fig. 1H) as well as differences in stem-like phenotype (Fig. S2). To our best knowledge, this association was only described in mESC [21]. A similar trend in upregulation of both pluripotency markers and nucleolin was observed for MCF-7 breast cancer cell line, M AN US C though highly variable (Fig. 2D). It has been suggested that ALDH and CD44 may identify CSC with different degrees of differentiation according to the histological types of breast cancer (for example, luminal, MCF-7 versus the less differentiated triple negative type, MDA-MB-231) [46], which could account for these results (Fig. 2C vs Fig. 2D). We confirmed the aforementioned results using mESC as stemness gold-standard system, as nucleolin is an highly conserved protein among mammal species [42]. Culturing mESC in conditions favoring pluripotency loss (absence of LIF and KSR) led to a downregulation of NANOG, OCT4 and nucleolin mRNA levels (Fig. 2B), and, D consistently, a strong decrease in cellular association of F3 peptide-targeted liposomes (Fig. 3 B and C) and Oct4-GFP fusion protein (Fig. 3E). Nucleolin has been described TE to regulate self-renewal in mESC [21]. Yang et al. demonstrated that differentiation overtime led to a decrease in nucleolin and OCT4 expression [21], in agreement with EP our results (Fig. 2B, Fig. 3 and Fig. S4). Overall, the aforementioned results led to question whether cell surface nucleolin expression per se, would enable the identification of tumorigenic cells. Strikingly NCL+ triple negative breast cancer cells AC C presented increased tumorigenic capacity, paralleling ALDHhi/CD44hi cells from the same histological origin (Fig. 4B and Table 1), already described as highly tumorigenic [5, 33]. Besides nucleolin role in angiogenesis and targeted drug delivery [18, 19], it has been shown that AS1411 aptamer, targeting cell surface nucleolin, impairs cellular growth of cancer cells of different histological origins, including breast cancer [47], consequently establishing nucleolin as a disease driver. Our results reinforce the previous observation, suggesting that a small population of surface nucleolinoverexpressing triple negative breast cancer cells may contribute, at least in part, to tumor development. Interestingly, over time all tested cell sub-populations gave rise to, approximately, an equal number of tumors, especially at higher cell density, resulting in 16 ACCEPTED MANUSCRIPT similar TIC frequency estimation (Table 1). Consistently, it has been suggested that TIC frequency may increase with observation time length [48, 49]. Chaffer and colleagues demonstrated that notwithstanding basal-like CD44lo breast cancer cells generated tumors rather inefficiently, those tumors had high levels of CD44hi cells [15]. Once re-injected in NOD/SCID mice, these CD44hi cells readily formed new tumors as compared to inefficient CD44lo cells [15]. This established the EMT-mediated dynamic RI PT cell plasticity as fundamental for the spontaneous conversion of basal-like non-SCC (CD44lo) to CSC (CD44hi), a tumorigenicity-enhancing feature [15]. This is also consistent with less differentiated, thus more aggressive, nature of basal-like breast cancers [50]. Thus, at least in the case of basal-like breast cancer cells, as MDA-MB- M AN US C 231, cell plasticity, under a stimulus, as hypoxia [51], might enable the conversion from low into highly tumorigenic cells [15], an event that could support, in part, our observations (Fig. 4 and Table 1). Overall, our data reinforce the need to strategically target multiple cell sub-populations, including both non-SCC and CSC, within the tumor microenvironment. Efficient eradication of both CSC and non-SCC may reside in the identification of synergistic drug combinations that simultaneously tackle several deregulated signaling pathways [11]. Nonetheless, translation of the in vitro efficacy information to in vivo remains a bottleneck due to pharmacokinetic differences of the combined drugs [12]. D Ligand-mediated targeted nanotechnology has the advantage of enabling the simultaneous intracellular delivery of drug combinations, on a receptor-dependent TE manner, besides synchronizing the pharmacokinetic of the encapsulated drugs and subsequent tumor accumulation [12, 19, 27]. We have previously developed F3- EP targeted liposomes encapsulating defined ratios of C6-Ceramide and doxorubicin, which enabled synergistic cell cytotoxicity against the triple negative breast cancer cells [20]. AC C The above cellular association and delivery results (Fig. 1 F and H), led us to evaluate this innovative strategy against CSC-derived mammospheres, better predictors of in vivo drug responses [36]. It has been reasoned that CSC eradication may be dependent of targeting developmental-related pathways, such as Notch or Wnt [7]. However, tackling classical signaling pathways such as PI3k/Akt pathway is also noteworthy [8]. Indeed, free C6-ceramide, targeting PI3k/Akt pathway [20], was able to induce death of ALDHhi breast cancer cells (Fig. 5 A and B), known to be resistant to DXR [33], as well as in MDA-MB-435S cells (Fig. S5), thus confirming it as a valuable agent against putative CSC. Herein, the F3 peptide-targeted combination strategy enabled 100% cell death against mammospheres of both putative CSC and non-SCC, 17 ACCEPTED MANUSCRIPT even unattainable by targeted liposomal DXR (Fig. 5C), thus apparently overcoming DXR resistance [33]. Non-targeted nanotechnologies (namely lipid-based), containing either doxorubicin or C6-ceramide, have demonstrated to improve pharmacodynamics performance of encapsulated drugs against different cancers [27, 32, 52] . However, this type of approach, whose tumor accumulation is based on the Enhanced Permeability and RI PT Retention (EPR) effect, is devoided of the ability to target cell populations within the tumor microenvironment responsible for tumor resistance and relapse [1]. Engineering nanoparticles’ surface with internalizing ligands, targeting, for example, the HER2 or transferrin receptors, has shown increased specificity and efficacy against tumors of M AN US C breast or other solid malignancies relative to the non-targeted (without ligand) counterparts [14, 53]. However, many of those ligands target receptors not necessarily providing specific targeting to the tumor microenvironment, and particularly, to CSC. Addressing this issue, nanoparticle-mediated targeting exploring the overexpression of CSC putative markers, like CD44 or CD133, have been proposed for small drugs and siRNA delivery into CSC, mainly as single agents [54, 55]. Despite overexpressed in cancer cells, CD44 has also been described to regulate normal vascular endothelial barrier integrity [56]. Thus, a CD44-based targeting strategy would not be necessarily devoided of side effects. In contrast, a nucleolin-targeted approach would be less D prone to collateral toxicity owing to specific surface nucleolin overexpression by endothelial cells of tumor angiogenic blood vessels, as compared to its absence in TE normal tissues, such as liver or lung [17, 18]. Therefore, the described F3 peptidemediated targeting towards cell surface nucleolin, of both putative CSC and non-SCC EP combines with the specific targeting of endothelial cells of tumor angiogenic blood vessels [18, 19]. Moreover, the F3 peptide-mediated intracellular delivery of a synergistic cytotoxic combination of DXR:C6-Cer into those cells may represent a AC C suitable multitarget approach, tackling simultaneously multiple pillars sustaining breast cancer. Conclusions Overall, our results suggested a clear link between nucleolin expression (including cell membrane nucleolin) and the stem cell-like phenotype in breast cancer, namely in the triple negative molecular subtype. It enabled increased cellular toxicity of F3 peptidetargeted drug combinations against both CSC and non-SCC, rendering 100% cell death. Combined with the established nucleolin-mediated targeting of tumor angiogenic blood vessels, the described strategy has the potential to simultaneously debulk 18 ACCEPTED MANUSCRIPT multiple cellular compartments in the tumor microenvironment, while decreasing tumor recurrence and systemic toxicity, ultimately enabling long-term disease free survival. Acknowledgements Nuno A. Fonseca was a student of the Pharmaceutical Sciences PhD program from the Faculty of Pharmacy, University of Coimbra and a recipient of the fellowship SFRH/BD/64243/2009 from the Portuguese Foundation for Science and Technology RI PT (FCT). Ana C. Gregório and Ângela Valério-Fernandes were students of PhD Program in Experimental Biology and Biomedicine (PDBEB), Center for Neuroscience and Cell Biology, University of Coimbra, and recipients of the FCT fellowships SFRH/BD/51190/2010 and SFRH/BD/51191/2010, respectively. The work was by the grants InovC/UC (2013), QREN/FEDER/COMPETE M AN US C supported (Ref. 30248/IN0617) and PEst-C/SAU/LA0001/2013-2014. Additional funding was granted by FEDER/COMPETE/FCT PTDC/EBB-EBI/120634/2010 and PDTC/QUI- BIQ/120652/2010 grants and QREN: CENTRO-01-0762-FEDER-00204. We would like to thank Rui Lopes for his support in the animal experiments. Conflict of interests V.M. is an employee of TREAT U, SA. All other authors declare no competing financial D interests. References AC C EP TE [1] Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646-74. [2] Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A. 2003;100:3983-8. [3] Visvader JE, Lindeman GJ. Cancer stem cells: current status and evolving complexities. Cell Stem Cell. 2012;10:717-28. [4] Scheel C, Weinberg RA. Cancer stem cells and epithelial-mesenchymal transition: concepts and molecular links. Seminars in cancer biology. 2012;22:396-403. [5] Ginestier C, Hur MH, Charafe-Jauffret E, Monville F, Dutcher J, Brown M, et al. ALDH1 is a marker of normal and malignant human mammary stem cells and a predictor of poor clinical outcome. Cell Stem Cell. 2007;1:555-67. [6] Marcato P, Dean CA, Pan D, Araslanova R, Gillis M, Joshi M, et al. Aldehyde dehydrogenase activity of breast cancer stem cells is primarily due to isoform ALDH1A3 and its expression is predictive of metastasis. Stem Cells. 2011;29:32-45. [7] Takebe N, Harris PJ, Warren RQ, Ivy SP. Targeting cancer stem cells by inhibiting Wnt, Notch, and Hedgehog pathways. Nat Rev Clin Oncol. 2011;8:97-106. [8] Dubrovska A, Kim S, Salamone RJ, Walker JR, Maira SM, Garcia-Echeverria C, et al. The role of PTEN/Akt/PI3K signaling in the maintenance and viability of prostate cancer stem-like cell populations. Proc Natl Acad Sci U S A. 2009;106:268-73. [9] Hart S, Novotny-Diermayr V, Goh KC, Williams M, Tan YC, Ong LC, et al. VS-5584, a Novel and Highly Selective PI3K/mTOR Kinase Inhibitor for the Treatment of Cancer. Molecular Cancer Therapeutics. 2012;12:151-61. 19 ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT [10] Badve S, Nakshatri H. Breast-cancer stem cells-beyond semantics. Lancet Oncol. 2012;13:e43-8. [11] Ramaswamy S. Rational design of cancer-drug combinations. N Engl J Med. 2007;357:299-300. [12] Dicko A, Mayer LD, Tardi PG. Use of nanoscale delivery systems to maintain synergistic drug ratios in vivo. Expert Opin Drug Deliv. 2010;7:1329-41. [13] Borovski T, De Sousa EMF, Vermeulen L, Medema JP. Cancer stem cell niche: the place to be. Cancer Res. 2011;71:634-9. [14] Wang AZ, Langer R, Farokhzad OC. Nanoparticle delivery of cancer drugs. Annu Rev Med. 2012;63:185-98. [15] Chaffer CL, Marjanovic ND, Lee T, Bell G, Kleer CG, Reinhardt F, et al. Poised Chromatin at the ZEB1 Promoter Enables Breast Cancer Cell Plasticity and Enhances Tumorigenicity. Cell. 2013;154:61-74. [16] Porkka K, Laakkonen P, Hoffman JA, Bernasconi M, Ruoslahti E. A fragment of the HMGN2 protein homes to the nuclei of tumor cells and tumor endothelial cells in vivo. Proc Natl Acad Sci U S A. 2002;99:7444-9. [17] Shi H, Huang Y, Zhou H, Song X, Yuan S, Fu Y, et al. Nucleolin is a receptor that mediates antiangiogenic and antitumor activity of endostatin. Blood. 2007;110:2899906. [18] Christian S, Pilch J, Akerman ME, Porkka K, Laakkonen P, Ruoslahti E. Nucleolin expressed at the cell surface is a marker of endothelial cells in angiogenic blood vessels. J Cell Biol. 2003;163:871-8. [19] Moura V, Lacerda M, Figueiredo P, Corvo ML, Cruz ME, Soares R, et al. Targeted and intracellular triggered delivery of therapeutics to cancer cells and the tumor microenvironment: impact on the treatment of breast cancer. Breast Cancer Res Treat. 2012;133:61-73. [20] Fonseca NA, Gomes-da-Silva LC, Moura V, Simoes S, Moreira JN. Simultaneous active intracellular delivery of doxorubicin and C6-ceramide shifts the additive/antagonistic drug interaction of non-encapsulated combination. J Control Release. 2014;196:122-31. [21] Yang A, Shi G, Zhou C, Lu R, Li H, Sun L, et al. Nucleolin maintains embryonic stem cell self-renewal by suppression of p53 protein-dependent pathway. J Biol Chem. 2011;286:43370-82. [22] Storck S, Shukla M, Dimitrov S, Bouvet P. Functions of the histone chaperone nucleolin in diseases. Subcell Biochem. 2007;41:125-44. [23] Feldman DE, Chen C, Punj V, Tsukamoto H, Machida K. Pluripotency factormediated expression of the leptin receptor (OB-R) links obesity to oncogenesis through tumor-initiating stem cells. Proc Natl Acad Sci U S A. 2012;109:829-34. [24] Wakayama T, Rodriguez I, Perry ACF, Yanagimachi R, Mombaerts P. Mice cloned from embryonic stem cells. Proceedings of the National Academy of Sciences. 1999;96:14984-9. [25] Gomes-da-Silva LC, Ramalho JS, Pedroso de Lima MC, Simoes S, Moreira JN. Impact of anti-PLK1 siRNA-containing F3-targeted liposomes on the viability of both cancer and endothelial cells. Eur J Pharm Biopharm. 2013;85:356-64. [26] Moreira JN, Ishida T, Gaspar R, Allen TM. Use of the post-insertion technique to insert peptide ligands into pre-formed stealth liposomes with retention of binding activity and cytotoxicity. Pharm Res. 2002;19:265-9. [27] Fonseca NA, Gregorio AC, Valerio-Fernandes A, Simoes S, Moreira JN. Bridging cancer biology and the patients' needs with nanotechnology-based approaches. Cancer Treat Rev. 2014;40:626-35. [28] Barenholz Y. Doxil(R)--the first FDA-approved nano-drug: lessons learned. J Control Release. 2012;160:117-34. [29] Fritze A, Hens F, Kimpfler A, Schubert R, Peschka-Suss R. Remote loading of doxorubicin into liposomes driven by a transmembrane phosphate gradient. Biochim Biophys Acta. 2006;1758:1633-40. 20 ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT [30] Garcia-Fuentes M, Torres D, Martin-Pastor M, Alonso MJ. Application of NMR spectroscopy to the characterization of PEG-stabilized lipid nanoparticles. Langmuir. 2004;20:8839-45. [31] Garbuzenko O, Zalipsky S, Qazen M, Barenholz Y. Electrostatics of PEGylated micelles and liposomes containing charged and neutral lipopolymers. Langmuir. 2005;21:2560-8. [32] Khazanov E, Priev A, Shillemans JP, Barenholz Y. Physicochemical and biological characterization of ceramide-containing liposomes: paving the way to ceramide therapeutic application. Langmuir. 2008;24:6965-80. [33] Croker AK, Allan AL. Inhibition of aldehyde dehydrogenase (ALDH) activity reduces chemotherapy and radiation resistance of stem-like ALDH(hi)CD44 (+) human breast cancer cells. Breast Cancer Res Treat. 2011. [34] Han YK, Lee JH, Park GY, Chun SH, Han JY, Kim SD, et al. A possible usage of a CDK4 inhibitor for breast cancer stem cell-targeted therapy. Biochem Biophys Res Commun. 2012. [35] Charafe-Jauffret E, Ginestier C, Iovino F, Wicinski J, Cervera N, Finetti P, et al. Breast cancer cell lines contain functional cancer stem cells with metastatic capacity and a distinct molecular signature. Cancer Res. 2009;69:1302-13. [36] Kim S, Alexander CM. Tumorsphere assay provides more accurate prediction of in vivo responses to chemotherapeutics. Biotechnology letters. 2013. [37] Strober W. Trypan blue exclusion test of cell viability. Current protocols in immunology / edited by John E Coligan [et al]. 2001;Appendix 3:Appendix 3B. [38] Varum S, Rodrigues AS, Moura MB, Momcilovic O, Easley CAt, Ramalho-Santos J, et al. Energy metabolism in human pluripotent stem cells and their differentiated counterparts. PLoS One. 2011;6:e20914. [39] Hu Y, Smyth GK. ELDA: extreme limiting dilution analysis for comparing depleted and enriched populations in stem cell and other assays. Journal of immunological methods. 2009;347:70-8. [40] Santos A, Sarmento-Ribeiro AB, de Lima MC, Simoes S, Moreira JN. Simultaneous evaluation of viability and Bcl-2 in small-cell lung cancer. Cytometry A. 2008;73A:1165-72. [41] Ling GQ, Chen DB, Wang BQ, Zhang LS. Expression of the pluripotency markers Oct3/4, Nanog and Sox2 in human breast cancer cell lines. Oncol Lett. 2012;4:1264-8. [42] Ginisty H, Sicard H, Roger B, Bouvet P. Structure and functions of nucleolin. J Cell Sci. 1999;112 ( Pt 6):761-72. [43] Mayer LD, Janoff AS. Optimizing combination chemotherapy by controlling drug ratios. Mol Interv. 2007;7:216-23. [44] Lee HE, Kim JH, Kim YJ, Choi SY, Kim SW, Kang E, et al. An increase in cancer stem cell population after primary systemic therapy is a poor prognostic factor in breast cancer. Br J Cancer. 2011;104:1730-8. [45] Gomes-da-Silva LC, Fernandez Y, Abasolo I, Schwartz S, Jr., Ramalho JS, Pedroso de Lima MC, et al. Efficient intracellular delivery of siRNA with a safe multitargeted lipid-based nanoplatform. Nanomedicine (Lond). 2013;8:1397-413. [46] Ricardo S, Vieira AF, Gerhard R, Leitao D, Pinto R, Cameselle-Teijeiro JF, et al. Breast cancer stem cell markers CD44, CD24 and ALDH1: expression distribution within intrinsic molecular subtype. J Clin Pathol. 2011;64:7. [47] Reyes-Reyes EM, Teng Y, Bates PJ. A new paradigm for aptamer therapeutic AS1411 action: uptake by macropinocytosis and its stimulation by a nucleolindependent mechanism. Cancer Res. 2010;70:8617-29. [48] Ishizawa K, Rasheed ZA, Karisch R, Wang Q, Kowalski J, Susky E, et al. Tumorinitiating cells are rare in many human tumors. Cell Stem Cell. 2010;7:279-82. [49] Quintana E, Shackleton M, Sabel MS, Fullen DR, Johnson TM, Morrison SJ. Efficient tumour formation by single human melanoma cells. Nature. 2008;456:593-8. 21 ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT [50] Schmitt F, Ricardo S, Vieira AF, Dionisio MR, Paredes J. Cancer stem cell markers in breast neoplasias: their relevance and distribution in distinct molecular subtypes. Virchows Archiv : an international journal of pathology. 2012;460:545-53. [51] Conley SJ, Gheordunescu E, Kakarala P, Newman B, Korkaya H, Heath AN, et al. Antiangiogenic agents increase breast cancer stem cells via the generation of tumor hypoxia. Proc Natl Acad Sci U S A. 2012. [52] Shabbits JA, Mayer LD. Intracellular delivery of ceramide lipids via liposomes enhances apoptosis in vitro. Biochim Biophys Acta. 2003;1612:98-106. [53] Reynolds JG, Geretti E, Hendriks BS, Lee H, Leonard SC, Klinz SG, et al. HER2targeted liposomal doxorubicin displays enhanced anti-tumorigenic effects without associated cardiotoxicity. Toxicol Appl Pharmacol. 2012;262:1-10. [54] Swaminathan SK, Roger E, Toti U, Niu L, Ohlfest JR, Panyam J. CD133-targeted paclitaxel delivery inhibits local tumor recurrence in a mouse model of breast cancer. J Control Release. 2013;171:280-7. [55] Ganesh S, Iyer AK, Morrissey DV, Amiji MM. Hyaluronic acid based selfassembling nanosystems for CD44 target mediated siRNA delivery to solid tumors. Biomaterials. 2013;34:3489-502. [56] Flynn KM, Michaud M, Canosa S, Madri JA. CD44 regulates vascular endothelial barrier integrity via a PECAM-1 dependent mechanism. Angiogenesis. 2013;16:689705. 22 ACCEPTED MANUSCRIPT Figure captions M AN US C RI PT Figure 1 – Cellular association of F3 peptide-targeted liposomes with putative breast cancer stem cells. (A) Schematic representation of the cellular association experiments. Cellular association (B,C,D,E,F) was performed in MCF-7 and triple negative MDA-MB-231 breast cancer cells whereas payload delivery (G,H) was assessed in mammosphere-derived single triple negative breast cancer cells. (B) Representative region criteria for the identification of CSC-enriched (R3) and non-SCC (R4) sub-populations based on CD44 and ALDH activity measurement (DEAB – diethylaminobenzaldehyde, a specific inhibitor of ALDH activity). (C, D) Represent the rhodamine side scatter dot-plots reflecting the signal distribution in each identified sub-population for the MCF-7 and MDA-MB-231 cell lines, respectively, following incubation with 0.4 mM of Rhod-labelled F3 peptide-targeted (p[F3]SL), non-specific peptide targeted (p[NS]SL) or non-targeted (pSL) liposomes for 1 h at 4 or 37ºC. (E, F) Represent the rhodamine geometric mean fluorescence of each sub-population for MCF-7 and MDA-MB-231 cell line, respectively (light-blue: putative cancer stem cells; orange: non-stem cancer cells). (2-Way ANOVA p<0.0001 for formulations tested and cell sub-populations assessed; ns p>0.05 and ***p<0.001 Bonferroni’s post test, N=3). (G) Representative dot-plots of calcein signal and nd (H) corresponding mean signal from 2 generation mammosphere-derived single cells obtained from MDA-MB-231 cells, upon incubation with 0.05 mM of calcein-loaded non-targeted (pSL), non-specific peptide- (p[NS]SL) and F3 peptide-targeted (p[F3]SL) liposomes at 37ºC for 1 h. Data represent the mean ± SEM. EP TE D Figure 2 – Comparative analysis of pluripotency genes and nucleolin mRNA levels in putative breast cancer stem cells (CSC) and mouse embryonic stem cells (mESC). (A) MCF-7 and MDA-MB-231 breast cancer cells were stained with CD44-PE/Cy5 and ® hi hi ALDEFLUOR reagent, and immediately sorted for isolation of ALDH /CD44 (putative CSC) low/low/(non-SCC) similarly as in Mammosphere assay. mESC were cultured and ALDH /CD44 for 72 h either in medium without LIF and Serum replacement [KSR] (inducing loss of pluripotency) or in fully supplemented medium containing LIF (Control). (B) Effect on NANOG, OCT4 and NCL mRNA levels from mESC culture in conditions inducing pluripotency loss. (C, D) Represent the relative mRNA fold-change of NANOG, OCT4 and nucleolin (NCL) of hi hi low/low/ALDH /CD44 relative to ALDH CD44 cells for both MDA-MB-231 and MCF-7 breast cancer cells lines, respectively. Data represent the mean ± SEM (N=2-3; p value was calculated using t test). AC C Figure 3 – Cellular association of F3 peptide-targeted liposomes with mouse embryonic stem cells. (A) E14-GFP mouse embryonic stem cells (mESC) or the corresponding colonies, were incubated with 0.4 mM total lipid of F3 peptide-targeted (p[F3]SL), non-specific peptide targeted (p[NS]SL) or non-targeted (pSL) liposomes incorporating 1 mol% of Rhodamine-PE, for 1 h at 4 or 37ºC and analyzed by flow cytometry, after 72 h in culture either in the presence of LIF (pluripotency maintenance) or in the absence of LIF and serum replacement (KSR). (B) Represents the rhodamine-side scatter dot-plots reflecting the signal distribution. (C) and (D) represent the geometric mean of rhodamine fluorescence for each system normalized against the corresponding signal of the untreated E14 mESC cells, in suspension and in colony, respectively (2-Way ANOVA p<0.016 for both culture conditions and liposome formulation variables; ***p<0.001 and *p<0.05 Bonferroni’s post-test, N=2-3). (E) Represents the Oct4-GFP levels of E14-GFP mESC cells according to culture conditions used (1-Way Anova p<0.0024; ns **p<0.01 and p>0.05 Tukey’s post-test, N=2-3). Non-viable cells were excluded from the analysis using 7-AAD. E14-wt mESC were used as controls to correct auto-fluorescence. Data represent the mean ± SEM. 23 ACCEPTED MANUSCRIPT RI PT Figure 4 – Tumor development latency of sorted cell populations upon inoculation in NOD scid gamma mice. (A) Sorting strategy for isolation of cell surface nucleolin positive + low/(NCL ) and nucleolin low/negative (NCL ) cells from single-cell suspensions of triple negative MDA-MB-231 breast cancer cells; 7-AAD was used to exclude non-viable cells. (B) Following ® staining of MDA-MB-231 cells with ALDEFLUOR /CD44-PE/Cy5 or with anti-NCL® AlexaFluor 488 antibody, sorted populations, as presented in the x-axis, were orthotopically inoculated in the mammary fat pad of NOD scid gamma (NSG) mice (6 mice/group). Dots and triangles represent the lapsed time for first palpation after inoculation, for a cell density of 2000 or 20000 cells, respectively. Bars represent the mean latency time for first palpation (2-way ANOVA: p = 0.0014 and p = 0.0428, for cell density and sorted population variables, respectively). AC C EP TE D M AN US C Figure 5 – Cellular cytotoxicity of doxorubicin (DXR):C6-ceramide (C6-Cer) combinations delivered by F3 peptide-targeted liposomes. (A) and (B) Represent the effect of free C6hi ceramide on viable ALDH cell sub-population from MCF-7 and MDA-MB-231 breast cancer cells, respectively (data represent mean ± SEM; *p<0.05 and **p<0.01 Tukey’s test, compared hi hi to untreated, N=3). (C) Representative dose-response curves of MDA-MB-231 (ALDH /CD44 low/low/nd and ALDH /CD44 ) cells derived from 2 generation mammospheres, incubated with F3 peptide-targeted liposomes either encapsulating DXR (p[F3]SL) or a combination of DXR and C6-Cer at 1:1 (p[F3]DC11) or 1:2 molar ratio (p[F3]DC12) or non-targeted liposomal DXR (pSL) (D) Inserted Table - IC90 values calculated from representative dose-response experiment by linear interpolation of dose values immediately above or below 90% effect). 24 ACCEPTED MANUSCRIPT Table 1 – Tumorigenic potential of different cell sub-populations sorted from the triple negative breast cancer cell line MDA-MB-231. Time (weeks after cell inoculation) MDA-MB-231 6 weeks Number of Sorted subpopulation low/- ALDH low/- /CD44 hi hi ALDH /CD44 low/- NCL NCL+ Number of Tumors 2000 20000 cells cells 0/6 2/6 3/6* 10 weeks Tumors TIC -1 frequency Number of Tumors TIC -1 frequency 2000 20000 cells cells 55400 1/6 6/6 6141 6/6** 2848 4/6* 6/6 1820 2/6 4/6 13391 2/6 5/6 8957 3/6 5/6 7310 5/6* 6/6 1116 2000 20000 cells cells TIC -1 frequency RI PT 4 weeks 5/6 6/6 1116 5/6 6/6 1116 5/6 6/6 1116 5/6 6/6 1116 AC C EP TE D M AN US C Data represent the number of tumors generated per sorted population injected (as presented in the table) 2 in NOD scid gamma mice (*p<0.05, **p<0.01 by χ test versus respective low/negative control at same cell density). Tumor initiating cell (TIC) frequency was calculated by the limiting dilution analysis using the Lcalc™ software. 25 ACCEPTED MANUSCRIPT RI PT Supplementary Data AC C EP TE D M AN US C Figure S1 - Cellular association of F3 peptide-targeted liposomes with putative cancer stem cells. Half million MDA-MB-435S cells were incubated with 0.4 mM of Rhod-labelled F3 peptidetargeted (p[F3]SL), non-specific peptide targeted (p[NS]SL) or non-targeted (pSL) liposomes for 1 h at 4 or 37ºC and subsequently stained with anti-CD44-PE/Cy5 antibody and with ® ALDEFLUOR reagent, and immediately analyzed through flow cytometry system. (A) Represents the rhodamine-side scatter dot-plots reflecting the signal distribution of each identified sub-population. (B) Represents the rhodamine geometric mean fluorescence of each sub-population (light-blue: putative cancer stem cells; orange: non-stem cancer cells). Data represent mean ± SEM (2-Way ANOVA p<0.0001 for formulations tested and cell subns populations assessed; p>0.05 and **p<0.01, Bonferroni’s post test, N=3). Figure S2 - Evaluation of mammosphere formation of sorted putative breast cancer stem cells. Two-million MCF-7 and MDA-MB-231 cells were stained with CD44-PE/Cy5 and ALDEFLUOR® hi hi reagent, and immediately sorted for isolation of ALDH /CD44 (putative cancer stem cells) and low/low/ALDH /CD44 (non-stem cancer cells). Sorted cells were cultured using fully supplemented 26 ACCEPTED MANUSCRIPT AC C EP TE D M AN US C RI PT Mammocult® Medium. (A) Representative sorting criteria for all cell lines tested, where P1 is hi hi the gate to exclude debris and death cells from cell sorting. Gating-criteria for ALDH /CD44 low/low/and ALDH /CD44 cell populations enabled the collection of 5-15% (P2) and 15-20% (P3) of total events depending on the assessed sub-population. (B) and (C) Representative images of 1st and 2nd generation (self-renewal) mammospheres and mammosphere formation hi hi low/low/efficiency data of ALDH /CD44 and ALDH /CD44 sub-populations, from MCF-7 MDAMB-231 cell lines (bar = 200 µm). Figure S3 - F3 peptide-mediated intracellular delivery to mammosphere-derived singles nd cells. Single cells derived from 2 generation mammospheres obtained from luminal MCF-7 breast cancer cells and MDA-MB-435S cancer cells were incubated with 0.05 mM of calceinloaded non-specific peptide- (p[NS]SL) and F3 peptide-targeted (p[F3]SL) liposomes at 37ºC for 1 h (A) Representative Forward Scatter/Side Scatter and Calcein/Side Scatter dot-plots from MDA-MB-435S cells. (B) Representation of the cell-associated mean calcein fluorescence. Data represent the mean ± SEM. 27 ACCEPTED MANUSCRIPT M AN US C RI PT Figure S4 - Comparative analysis of pluripotency genes and nucleolin mRNA levels in mouse embryonic stem cells (mESC). E14 mESC were cultured in fully supplemented medium in the presence of LIF (Control), in absence of LIF (w/o LIF) for 72 h. Figure represents the fold-change in mRNA levels of pluripotency markers NANOG, OCT4 and nucleolin (NCL) relative to control (data represents mean ± SEM, N=3, p value was calculated with t test). hi AC C EP TE D Figure S5 – Cytotoxicity of C6-ceramide against ALDH sub-population from MDA-MB435S cells. (Data represent mean ± SEM; *p<0.05 and **p<0.01 Tukey’s test, compared to untreated, N=3). 28 ACCEPTED MANUSCRIPT Table S1 – List of primer nucleotide sequences for qRT-PCR. Primer Bank Nucleotide Sequence 5 -3 ID 31543315a1 OCT4 356995852c3 NANOG 31338864a1 β-ACTIN 6671509a1 NUCLEOLIN 55956787c2 OCT4 4505967a1 NANOG 153945815c3 β-ACTIN 4501885a1 FW RV FW RV FW RV FW RV FW RV FW RV FW RV FW RV AAAGGCAAAAAGGCTACCACA GGAATGACTTTGGCTGGTGTAA CGGAAGAGAAAGCGAACTAGC ATTGGCGATGTGAGTGATCTG TCTTCCTGGTCCCCACAGTTT GCAAGAATAGTTCTCGGGATGAA GGCTGTATTCCCCTCCATCG CCAGTTGGTAACAATGCCATGT GCACCTGGAAAACGAAAGAAGG GAAAGCCGTAGTCGGTTCTGT CTTGAATCCCGAATGGAAAGGG GTGTATATCCCAGGGTGATCCTC CCCCAGCCTTTACTCTTCCTA CCAGGTTGAATTGTTCCAGGTC CATGTACGTTGCTATCCAGGC FW CTCCTTAATGTCACGCACGAT – RI PT NUCLEOLIN M AN US C Human Mouse Gene AC C EP TE D Forward; RV – Reverse; 29 ACCEPTED MANUSCRIPT A 1st Gen. mammospheres CSC 2st Gen. mammospheres Breast cancer cell lines Rhodamine-labeled F3 peptide-targeted liposomes 7-21 days 10-21 days Suspension Culture (Mammocult® media) Cell Sorting 1 h, 37ºC 4ºC Cellular association Calcein-loaded F3 peptide-targeted liposomes Mammosphere-derived single cells (ALDHhi/CD44hi) Mammosphere dissociation (single cell suspension) Suspension Culture (Mammocult® media) Mammosphere dissociation 1 h, 37ºC Payload delivery assessment 7-21 days 10-21 days non-SCC 101 R5 102 CD44 103 104 1000 100 101 103 104 800 101 102 RhoD 103 104 1000 TE 1000 1000 101 102 Calcein 103 104 100 100 104 800 800 Side Scatter 400 600 Side Scatter 400 600 103 104 100 102 RhoD 103 104 1000 101 800 Side Scatter 400 600 104 p[F3]SL 4ºC 102 RhoD 103 104 ns ns ns Ctr pSL p[NS]SL 102 Calcein 103 104 100 101 102 Calcein 103 101 102 Calcein 103 104 100 101 102 Calcein 103 p[F3]SL p[F3]SL 4ºC H p[F3]SL 104 0 0 101 103 ns 200 200 800 Calcein 102 Calcein 100 100 Side Scatter 400 600 103 Side Scatter 400 600 200 0 101 104 200 200 102 Calcein 800 100 101 103 *** 0 101 Side Scatter 400 600 1000 100 100 102 RhoD 300 1000 104 p[NS]SL 800 800 Side Scatter 400 600 103 p[F3]SL 800 pSL 0 102 Calcein 200 Side Scatter ALDHlow/-/CD44low/- 102 RhoD 101 MDA-MB-231 p[F3]SL 4ºC 200 Side Scatter 400 600 0 101 1000 100 101 p[NS]SL pSL 100 104 0 p[F3]SL 1000 Untreated 800 1000 p[NS]SL 200 ALDH /CD44 hi 400 ns Side Scatter 400 600 pSL ** 200 Ctr Mammosphere-derived single cells hi F 0 ns Ctr 103 Rhodamine 1000 ns 0 G 100 EP MCF-7 AC C Geomean (RFU) *** 0 102 RhoD 1000 102 RhoD E 500 101 0 p[F3]SL 4ºC Geomean (RFU) 104 100 D 103 104 Side Scatter Side Scatter 400 600 200 102 RhoD 0 p[F3]SL 101 Rhodamine 1500 103 200 Side Scatter 400 600 100 1000 102 RhoD 200 200 p[NS]SL 0 101 800 800 Side Scatter 400 600 200 100 100 4 200 10 104 Mean Calcein Signal (RFU) 3 0 10 800 2 Side Scatter 400 600 10 RhoD 200 1 M AN U Side Scatter 400 600 Side Scatter 400 600 200 pSL 0 0 10 0 Side scatter 800 800 1000 800 600 Side Scatter 400 200 Ctr 0 1000 10 1000 D 1000 C SC CD44 1000 100 800 104 Side Scatter 400 600 103 200 102 CD44 0 101 RI PT 10 1 R4 - DEAB 100 1000 104 800 Forward scatter 103 Side Scatter 400 600 102 CD44 0 101 ALDHhi/CD44low/ALDHhi/CD44hi ALDHlow/-/CD44low/ALDHlow/-/CD44hi 0 + DEAB 100 1000 ALDH 10 2 10 3 10 1 10 1 ALDH 800 10 0 600 10 0 400 ALDH 10 2 ALDH 10 2 200 10 0 0 R3 R2 10 3 10 4 10 4 10 3 800 600 400 200 R1 0 Side scatter 1000 B 10 4 (ALDHlow/-/CD44low/-) 800 Mammosphere-derived single cells 600 400 200 ALDHhi/CD44hi ALDHlow/-/CD44low/- 0 Ctr pSL p[NS]SL p[F3]SL ACCEPTED MANUSCRIPT A C qRT-PCR CSC MDA-MB-231 (ALDHhi/CD44hi) p=0.121 Cell Sorting 72 h w/o and LIF KS R Pluripotency maintenance Pluripotency loss RNA Extraction 0.0 E14 mESC 1.0 p<0.05 p<0.05 AC C mRNA Fold-Change p<0.05 EP 1.5 0.5 0.0 NANOG OCT4 NCL OCT4 NCL D TE D B 0.5 M AN U IF +L 1.0 RI PT non-SCC (ALDHlow/-/CD44low/-) 1.5 p=0.193 SC NCL NANOG OCT4 mRNA Fold-Change RNA Extraction mESC p=0.085 2.0 6 mRNA Fold-Change Breast cancer cell lines MCF-7 5 4 3 2 1 0 NANOG Control OCT4 NCL w/o LIF/Serum replacement [KSR] NANOG ALDHlow/-/CD44low/- ALDHhi/CD44hi ACCEPTED MANUSCRIPT A C mESC single cell suspension Normalized GeoMean (vs untreated) 200 + 72 h 1000 800 Side Scatter 400 600 100 102 103 100 104 200 0 0 101 101 RhoD 102 103 104 100 1000 1000 1000 1000 800 800 800 800 104 100 TE Side Scatter 400 600 102 RhoD 103 104 100 103 104 100 50 pSL p[NS]SL p[F3]SL ns 101 102 RhoD 103 104 800 103 *** 100 200 200 102 RhoD E14 cells in colony E 0 101 p[F3]SL 150 104 1000 102 RhoD Side Scatter 400 600 800 Side Scatter 400 600 102 RhoD 100 104 p[NS]SL 0 101 0 103 0 101 200 200 101 200 Side Scatter 400 600 103 100 100 1000 104 104 50 D 0 Side Scatter 400 600 103 200 102 RhoD 104 200 102 RhoD 0 101 103 1000 101 800 1000 800 Side Scatter 400 600 200 100 100 104 102 RhoD 0 0 103 1000 102 RhoD 101 AC C Side Scatter 400 600 200 Side Scatter 400 600 200 0 101 103 104 800 100 102 RhoD 103 Side Scatter 400 600 104 102 RhoD 200 103 1000 102 RhoD 800 101 1000 100 EP 104 800 103 0 0 200 200 Side Scatter 400 600 Side Scatter 400 600 Side Scatter 400 600 0 102 RhoD 101 RhoD 1000 101 101 1000 800 200 Side Scatter 400 600 1000 100 100 pSL Normalized GeoMean (vs untreated) 104 150 0 SC Side Scatter 400 600 200 103 800 200 0 104 0 37ºC Side Scatter w/o serum 102 RhoD 101 102 RhoD 103 104 Colony 100 -LIF 101 Side Scatter 400 600 103 800 100 +LIF 37ºC 100 800 102 200 37ºC 104 Side Scatter 400 600 0 101 RhoD w/o serum 103 Suspension 100 -LIF 102 RhoD D 101 0 0 100 104 1000 103 800 800 200 200 0 102 RhoD 200 +LIF 4ºC Side Scatter 400 600 Side Scatter 400 600 0 101 1000 100 p[F3]SL 1000 1000 800 1000 800 Side Scatter 400 600 200 +LIF 37ºC p[NS]SL M AN U SR /K pSL Control *** * E14 cells in suspension RI PT IF -L B mESC colonies Pluripotency loss Oct4-GFP GeomMean (RFU) LI F E14 mESC colonies Cellular association with F3-targeted liposomes Pluripotency maintenance 150 ** ** 100 50 0 Rhodamine 37ºC (+LIF) Suspension 4ºC (+LIF) Colony 37ºC (-LIF; w/o serum replacement [KSR]) ACCEPTED MANUSCRIPT A B Isotype RI PT NCLlow/- SC 600 0 0 NCL+ 0 200 400 600 800 10 0 1000 10 1 10 2 10 3 NCL Forward Scatter TE AC C 800 600 200 10 1 10 2 7-AAD 10 3 10 4 10 0 10 1 10 2 NCL 10 8 6 4 2 /- low 2000 Cells 10 3 hi 44 D44 hi /CD /low /C H H ALD ALD 0 0 12 0 ≈ 0.8% 400 Side Scatter 600 400 200 R3 NCL+ Viable Cells 0 Side Scatter 800 R2 Anti-NCL-AlexaFluor488 NCLlow/- EP 1000 1000 Viability Analysis 10 10 4 D R1+R2 M AN U 200 400 Side Scatter 600 400 R1 200 Side Scatter 800 800 R3 First palpation (weeks after cell inoculation) 1000 1000 Debris Exclusion 10 4 /- low NCL + NCL 20000 Cells ACCEPTED MANUSCRIPT B 20 2 * ** 1 0 20 μM Untreated 15 5 0 40 μM M AN U 100 90 80 10 1 AC C D ALDHlow/-/CD44low/- 100 90 80 60 0.1 100 1 10 [DXR] (μM) EP [DXR] (μM) pSL(DXR) p[F3]SL(DXR) p[F3]DC11 p[F3]DC12 Citotoxicity (IC90) of liposomal formulations of DXR or DXR/C6-Cer against mammospheres derived from ALDHhi/CD44hi and ALDHlow/-/CD44low/sub-populations sorted from MDA-MB-231 triple breast cancer cells. Formulation ALDHhi/CD44hi IC90 (μM) SEM ALDHlow/-/CD44low/IC90 (μM) N/D pSL(DXR) 14 μM C6-Ceramide 70 TE 70 Cell Death (% of Control) 110 D Cell Death (% of Control) ALDHhi/CD44hi ** 7 μM Untreated C 60 0.1 * 10 C6-Ceramide 110 MDA-MB-231 RI PT MCF-7 SC ALDH hi /7AAD (%of cells) 3 ALDH hi /7AAD (%of cells) A SEM N/D p[F3]SL(DXR) 8.08 2.82 2.98 p[F3]DC1 1 1.52 0.03 1.91 0.25 0.05 p[F3]DC12 1.48 0.07 1.27 0.05 100