Confining Iron Oxide Nanocubes inside Submicrometric Cavities as a Key Strategy to Preserve Magnetic Heat Losses in an Intracellular Environment

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ "This document is the Accepted Manuscript version of a Published Work that appeared in final form in ACS Applied Materials and Interfaces, copyright © American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see https://doi.org/10.1021/acsami.9b15501 ." Confining iron oxide nanocubes inside submicrometric cavities as a key strategy to preserve magnetic heat losses in an intracellular environment Mikhail V. Zyuzin1,2, Marco Cassani1,3, Markus J. Barthel1, Helena Gavilan1, Niccolò Silvestri1,3, Alberto Escudero4,5, Alice Scarpellini1, Federica Lucchesi1,6,Francisco J. Teran7,8, Wolfgang J. Parak9*, Teresa Pellegrino1*. 1. Istituto Italiano di Tecnologia, Via Morego 30, 16163 Genova, Italy. 2. Faculty of Physics and Engineering, ITMO University, Lomonosova 9, 191023 St. Petersburg, Russia 3. Dipartimento di Chimica, Università di Genova , Via Dodecaneso 33, 16146 Genova, Italy 4. Leibniz Institute for New Materials. Campus D2 2. D-66123 Saarbrücken, Germany 5. Departamento de Química Inorgánica and Instituto de Investigaciones Químicas (IIQ), Universidad de Sevilla – CSIC, Calle Américo Vespucio 49, E-41092 Seville, Spain. 6. Dipartimento di Informatica, Bioingegneria, Robotica e Ingegneria dei Sistemi (DIBRIS), Via all'Opera Pia, 13, 16145, Genova 7. iMdea Nanociencia, Campus Universitario de Cantoblanco, 28049 Madrid, Spain 8. Nanobiotecnología (iMdea-Nanociencia), Unidad Asociada al Centro Nacional de Biotecnología (CSIC), 28049 Madrid, Spain 9. Faculty of Physics and Chemistry and CHyN, Universität Hamburg, 20146, Hamburg, Germany * corresponding authors: [email protected] and teresa.pellegrino@iit Abstract 1 The design of magnetic nanostructures whose magnetic heating efficiency remains unaffected at the tumor 2 site is a fundamental requirement to further advance magnetic hyperthermia in clinic. This work 3 demonstrates that the confinement of magnetic nanoparticles (NPs) into a submicrometric cavity is a key 4 strategy to enable a certain degree of nanoparticle motion and minimize aggregation effects, consequently 5 preserving the magnetic heat loss of iron oxide nanocubes (IONCs) under different conditions, including 6 intracellular environments. We fabricated magnetic Layer-by-Layer (LbL) self-assembled polyelectrolyte 7 submicrometric capsules using three different approaches, and we studied their heating efficiency as 8 obtained in aqueous dispersions and once internalized by tumor cells. First, IONCs were added to the hollow 9 cavities of LbL submicrocapsules, allowing the IONCs to move to a certain extent in the capsule cavities. 10 Second, IONCs were co-encapsulated into solid calcium carbonate cores coated with LbL polymer shells. 11 Third, IONCs were incorporated within the polymer layers of the LbL capsule walls. In aqueous solution, the 12 higher specific absorption rate (SAR) values were related to the ones of free IONCs, while lower SAR values 13 were recorded for capsule/core assemblies. However, after uptake by cancer cell lines (SKOV-3 cells), the 14 SAR values of the free IONCs were significantly lower than those observed for capsule/core assemblies, 15 especially after prolonged incubation periods (24 and 48 hours). These results show that IONCs packed into 16 submicrocavities preserve the magnetic losses, as SAR values remained almost invariable. Conversely, free 17 IONCs without the protective capsule shell agglomerated and their magnetic losses are strongly reduced. 18 Indeed, IONC loaded capsules and free IONCs reside inside endosomal and lysosomal compartments after 19 cellular uptake, show magnetic losses strongly reduced due to the immobilization and aggregation in 20 centrosymmetrical structures in the intracellular vesicles. The confinement of IONCs into submicrometric 21 cavities is a key strategy to provide a sustained and predictable heating dose inside biological matrices. 22 23 Introduction 24 The therapeutic benefits of heat have been well known for centuries. In oncothermia, it has been shown that 25 raising the temperature (i.e. inducing hyperthermia) can be a successful coadjuvant therapeutic approach 26 for treating solid tumors in combination with radioor chemotherapy.1,2 Today, nanotechnology provides 27 novel and minimally invasive ways to locally release heat inside a body by means of distinct remote activation 28 modalities.3–5 Recent studies underline the large benefits and efficiency of spatially controlled heat 29 deposition by magnetic nanoparticles to remove localized tumors.5–8 Among the suitable nanomaterials used 30 in hyperthermia studies, iron oxide nanoparticles (IONPs) are widely employed since their magnetic 31 properties can be precisely controlled.9–11 They benefit to have negligible toxicity drawbacks, and high 32 biodegradation and clearance capabilities after treatment.12,13 The exposure of iron oxide nanoparticles to 33 alternating magnetic fields and/or infrared radiation leads to heat generation, which has been successfully 1 employed to remove cancer cells.14 For this reason, enormous efforts have been devoted to the development 2 of nanostructures based on iron oxide nanoparticles with outstanding magnetic losses.15,16 Alongside, the 3 emerging technique of magnetic particle imaging (MPI)17,18 benefits from iron oxide nanoparticle tracers 4 whose magnetic characteristics are similar to the ones valid for magnetic hyperthermia.19 5 Subsequent to the first evidence of hyperthermia that was mediated by microparticles in the fifties,20 many 6 preclinical studies6,21–24 and clinical trials20,25 over the last 15 years have shown promising results with regard 7 to hyperthermia that is mediated by iron oxide nanoparticles for cancer treatment. Reports on the first 8 clinical trials using iron oxide nanoparticles for treating recurrent glioblastoma multiforme (DRKS00005476 9 clinical trial) and prostate carcinoma (clinical trial NCT02033447) in magnetic hyperthermia have provided 10 the first proof of concept on a human scale, rendering the introduction of magnetic hyperthermia to clinics 11 as the main research task today.25 For this purpose, it is necessary to control the heat dose inside tumors, a 12 highly relevant clinical requirement. Some efforts have been spent to produce magnetic nanoparticles that 13 possess outstanding heat efficiency in highly viscous environment.26,27 Despite this, recent studies have 14 evidenced that the magnetic properties of nanoparticles are dramatically altered in different biological 15 environments for therapeutic or imaging applications.9,18,28,29 In particular, the magnetic heat losses from iron 16 oxide nanoparticles in cellular environments suffer significant reductions ranging from 70% to 90% 17 depending on nanoparticle size, chemical composition, and/or aggregation degree.9 Recent results show that 18 the enhancement of nanoparticle immobilization in an intracellular environment and, in particular, the 19 nanoparticle clustering during the intracellular transit alter the magnetic behavior under alternating 20 magnetic fields immediately after cell uptake.30 These results demonstrate that hyperthermia mediated by 21 IONPs behaves differently under in vitro and in vivo conditions than in colloidal dispersion. For this reason, 22 not only the synthesis protocols of iron oxide nanoparticles need to be improved, but new approaches to 23 preserve the magnetic losses and/or imaging (MPI/MRI) performance of magnetic nanostructures into 24 biological environments (i.e. cells and tissues) are also required.31 Recent works have shown that the random 25 confinement of IONP into capsules at fixed inter-particle distances freezes the magnetic dipolar interactions 26 in such a way that the agglomeration effects have no effect on the magnetic losses.32,33 Multilayer 27 polyelectrolyte (PE) encapsulation is a cheap and robust Layer-by-Layer (LbL) based approach which consists 28 of the subsequent assembly of oppositely charged polymers that form the capsule’s wall when they are 29 deposited onto a sacrificial template.34–37 The size of the capsules can be precisely tuned from tens of 30 nanometers to several microns according to the core template and, after the LbL procedure, the sacrificial 31 template can be dissolved so that hollow polyelectrolyte capsules can be obtained.34 The capsules can be 32 loaded with different compounds, like proteins, mRNA, biomolecules, or nanoparticles whose distribution 33 can be located at different places on the capsules (cavity or shells), thus, enabling them to protect the cargo 34 from enzymatic degradation.38,39 35 Here, we report on the quantification of magnetic heat losses of highly performing iron oxide nanocubes 1 (IONCs) with different edge lengths (8, 14, 16, 18 and 21 nm) when they are distinctly embedded into the 2 different capsules. The magnetic capsules developed here encapsulated IONCs in three different spatial 3 arrangements: i) in the hollow cavity; ii) in a solid CaCO3 core; iii) in-between the polymer layers of the capsule 4 walls. The magnetic heat losses observed for the different IONCs/capsule arrangements were compared to 5 those of free IONCs under different conditions: in water, in viscous media (water or glycerol dispersions), or 6 upon cellular internalization after different incubation times. Overall, our data suggest that, among the 7 different encapsulation studied, the spatial confinement of IONCs in the hollow submicrometric cavity is 8 optimal to preserve the magnetic heat losses of IONCs inside cells. Thus, the hollow capsules have sustained 9 and predictable heating efficiency for hyperthermia performances into biological environments. 10 Experimental section 11 Materials 12 Dopamine hydrochloride, sodium nitrite, sulphuric acid, α-ω-hydroxypoly(ethylene glycol) (PEG, Mw= 1400 13 g/mol), anhydrous dichloromethane (CH2Cl2), trimethylamine (TEA), 4-dimethylaminopyridine (DMAP), 14 succinic anhydride, N-hydroxysuccinimide (NHS),N,N'-dicyclohexylcarbodiimide (DCC), dimethylformamide 15 (DMF), toluene sodium carbonate (Na2CO3), calcium chloride dihydrate (CaCl2 x 2H2O), ethylene glycol 16 (EG),poly(sodium 4-styrenesulfonate) (PSS, Mw = 70000 g/mol), poly(allylaminehydrochloride) (PAH, Mw = 17 15000 g/mol), sodium chloride (NaCl), ethylenediaminetetraacetic acid (EDTA), glycerol, and McCoy’s 5A 18 medium were purchased from Aldrich and used as received. Ultrapure water or Milli-Q water with a 19 resistance greater than 18.2 MΩ cm−1 was used for all experiments. Dulbecco phosphate buffered saline 20 (DPBS), phosphate buffered saline (PBS), fetal bovine serum (FBS), trypsin-EDTA 1x in PBS, penicillin and 21 streptomycin (P/S), gluteraldehyde buffer, cacodylate buffer, osmium tetroxide, uranyl acetate, ethanol, and 22 epoxy Epon™ (TAAB) resin were purchased from Euro Clone and used as received. 23 24 Synthesis of Iron Oxide Nanocubes 25 IONCs with a cubic shape and different edge length sizes (8, 14, 16, 18, 21 nm) were synthesized according 26 to an already published protocol.15 Briefly, iron (III) acetylacetonate (1 mmol) was mixed in a three-neck flask 27 with decanoic acid (the added amount in mmol depends on the expected size) in a solution of dibenzyl ether 28 and squalane. The solution was then degassed at 65 °C for 2 hours and was later heated up to 200 °C under 29 nitrogen flow. After 2 hours at 200°C, the solution was heated up by refluxing until it reached a temperature 30 of 305 °C. It was kept at this temperature for 1 hour. The resulting dispersions containing the formed IONCs 31 were cooled down to room temperature (RT), centrifuged so as to remove the supernatants (8500 rpm for 32 10 min) and washed twice with acetone (i.e. acetone was added, and the supernatants were removed after 1 centrifugation). The IONCs were finally dispersed in chloroform (Figure S1). 2 Synthesis of α-nitrodopamine-ω-carboxypoly(ethylene glycol) (ND-PEG-COOH) as a water transfer ligand 3 α-nitrodopamine-ω-carboxy-poly(ethylene glycol) (ND-PEG-COOH) was chosen as a water transfer ligand for 4 the nanocubes. We adapted already existing procedures for the synthesis of the nitrodopamine modified 5 carboxyl terminated poly(ethylene glycol) (Figure 1 and Figures S2-S5 and Scheme S1 of the electronic 6 supporting information, ESI).40,41 The synthesis was carried out in four steps. First, nitrodopamine was 7 obtained by nitrating the aromatic ring of the dopamine molecule, asdescribed in literature but with minor 8 modifications.40 In parallel, a α-ω-hydroxy-poly(ethylene glycol) was reacted with succinic anhydride (1:2 9 ratio) in order to transform the hydroxyl into carboxy functions. Subsequently, one terminal carboxy group 10 of the PEG was activated using DCC/NHS chemistry, then nitrodopamine was coupled to the polymer 11 following a reported protocol.41 A detailed synthesis of the polymeric ligand and the characterization is 12 shown in the ESI. 13 14 Ligand exchange and water transfer of cubic iron oxide nanoparticles (IONCs) 15 A well-established protocol for the water transfer of IONCs was used with minor modifications. The 16 procedure is described for 18 nm edge size IONCs and is representative for all other samples in terms of the 17 stoichiometry.41,42 In detail, 1.88 mL of chloroform was added to 2.12 mL of the IONC dispersion in 18 chloroform, which contained 4 mg of Fe, yielding a final concentration of 1 g/L Fe. ND-PEG-COOH was used 19 as a hydrophilic ligand for the water transfer. For this, a ratio of 150 ligands/nm2 per IONC’s surface was 20 applied. 121 mg of the polymeric ligand (0.086 mmol) were dissolved in 1.61 mL of chloroform, yielding in a 21 polymer solution with a final concentration of 0.05 M. Both mixtures (the polymeric ligand and the 22 suspension containing the IONCs) were mixed in a glass vial, and 0.79 mL of TEA (70 eq. with regard to mmol 23 of ligands) was added. The mixture was vigorously shaken overnight at room temperature (RT) to allow the 24 initial hydrophobic ligands to be replaced with the hydrophilic ND-PEG-COOH, causing the nitrodopamine to 25 act as an anchor. Subsequently, 40 mL of toluene was added and the mixture was transferred into a 26 separatory funnel. 100 mL of MilliQ water was added, and the mixture was shaken to form an emulsion and 27 to induce the water transfer of the nanoparticles. The aqueous phase containing the first fraction of 28 nanoparticles was collected, 50 mL of water was added to the organic phase, and the procedure was repeated 29 so as to extract the remaining nanoparticles. Afterwards, both aqueous phases were combined, and possible 30 traces of remaining toluene were removed using a nitrogen flow. The sample was concentrated to a final 31 volume of approximately 10 mL by centrifugation using an amicon centrifuge filter (50000 g/mol cut-off, 1500 32 rpm). This dispersion was sonicated for 30 min at 65 °C with the purpose of redispersing the cubic IONCs. The 33 sample was transferred into a spectrum labs dialysis membrane bag (cut-off 100000 g/mol, 1700 rpm), and 34 the mixture was dialyzed against water for 2 days for purification. The purified sample was concentrated two 1 times using an amicon centrifuge filter until a final volume of approximately 2 mL was obtained. The sample 2 was then sonicated again for 30 min at 65 °C. The final product was analyzed by dynamic light scattering (DLS) 3 measurements (Figures S6-S9, Table S1). 4 Nanoparticle characterization 5 The colloidal stability of PEG-coated IONCs was studied by monitoring the hydrodynamic diameter in 6 different media with DLS. The nanoparticles were dispersed in MilliQ water, a 0.33 M calcium chloride 7 solution, a 0.33 M sodium carbonate solution, a 0.2 M ethylenediaminetetraacetic acid solution, and in full 8 cell growth medium. Both, the hydrodynamic diameter (Dh, which is given as the mean value of the number, 9 intensity and volume distribution) and the zeta potential (ζ) of the IONCs were measured using a Zetasizer 10 Nano ZS90 (Malvern, US) equipped with a 4.0 mW He−Ne laser operating at 633 nm and with an avalanche 11 photodiode detector, in the 173° backscattered mode. Before the measurements, the samples were highly 12 diluted in an aqueous solution. The measurements were performed at 25 °C. Measurements for each sample 13 were repeated three times, and each measurement included 10 runs for 10 s. Significant increase in the 14 hydrodynamic diameter was an indicator of nanoparticle aggregation.43 15 The effect of the possible aggregation of the nanoparticles on their magnetic properties was studied for both, 16 free and encapsulated IONCs. The different encapsulated IONCs were mixed at a ratio of 1:1 with 5x PBS until 17 a final concentration of 0.75 g(Fe)/L was reached. Magnetization curves were then measured at a frequency 18 (f) of 100 kHz and a magnetic field (H) of 24 kA/m (B = 30 mT) in water and 5x PBS, 0.05 M. 19 20 Synthesis of magnetic polyelectrolyte capsules 21 Synthesis of CaCO3 cores loaded with IONCs of different sizes 22 Iron oxide nanocubes (IONCs) with cube-edges of different lengths were encapsulated in submicrometric 23 calcium carbonate cores or embedded into the capsule walls using a modified protocol.44-46 Briefly, 0.33 M 24 calcium chloride and sodium carbonate solutions were prepared in an 5:1 in volume ofEG:H2O mixture. 50 25 µL of the water suspensions of the 8, 14, 16, 18 and 21 nm IONCs (c= 3 g(Fe)/L) was added to 2 mL of the 26 0.33 M CaCl2 solution under magnetic stirring. Then, 386 µL of 0.33 M Na2CO3 was added and mixed by stirring 27 vigorously at 1100 rpm for 20 min at room temperature. The formed submicrometric calcium carbonate cores 28 with encapsulated IONCs (8, 14, 16, 18, 21 nm) were then centrifuged (at 9000 rpm for 3 min), and the 29 supernatants were discarded. The cores were then washed twice with MilliQ water (i.e. addition of water, 30 centrifugation at 9000 rpm for 3 min and then removal of the supernatant). The CaCO3 cores that were loaded 31 with the different IONCs were then coated with non-biodegradable polyelectrolytes, namely poly(styrene 32 sulfonate) (PSS) and poly(allylamine hydrochloride) (PAH). Polyelectrolytes solutions with a concentration of 33 10 mg/mL were prepared in 0.05 M sodium chloride, adjusting the pH of the solution to 6.5. CaCO3 cores 34 were first dispersed in 1 mL of a 10 mg/mL PSS solution, then sonicated for 5 min, and shaken for 10 min. 1 The cores were then washed twice with MilliQ water at 9000 rmp for 3 min to remove any excess 2 polyelectrolytes. Then, 1 mL of a 10 mg/mL polycationic PAH solution was added to the cores. The dispersion 3 was sonicated again for 5 min, and shaken for 10 min. This procedure was repeated 8 times in order to obtain 4 8 monolayers (4 bilayers) of shell around the CaCO3 cores. Finally, the cores with immobilized IONCs in the 5 CaCO3 matrix coated with 4 bilayers of polyelectrolytes (CORES) were obtained. The sample was kept at an 6 iron concentration of 2 mg(Fe)/mL. 7 8 Synthesis of capsules loaded with different sized IONCs 9 Empty multilayer polyelectrolyte submicrometric capsules were obtained by treating the previously prepared 10 iron oxide CaCO3 cores with 1 mL of EDTA (0.2 M, pH 6). For this, as obtained cores were spun down at 9000 11 rpm for 3 min. Supernatants were then removed and EDTA was added and dispersions were kept overnight 12 in a shaker at room temperature. The polyelectrolyte capsules were then washed two times with MilliQ water 13 at 1500 rpm for 30 min. At the end of the procedure, capsules with IONCs in the cavity (CAPS) were obtained. 14 15 Synthesis of capsules with IONCs embedded in their walls 16 In order to obtain polyelectrolyte capsules with IONCs (8,14, 16, 18, 21 nm) embedded in the capsule wall, 17 bare calcium carbonate cores (with no IONCs) were prepared by directly mixing 2 mL of a 0.33 M calcium 18 chloride solution with 386 µL of a 0.33 M sodium carbonate solution, both of which were prepared in a 5:1 19 parts invol. EG:H2O mixture, in a similar way to the aforementioned procedure,38,39 but without the IONCs. 20 Two bilayers of polyelectrolytes were deposited onto the cores in the same way described above, and the 21 IONCs were incorporated after the second PAH layer by adding 50 µL of IONCs (3 g(Fe)/L) to the CaCO3 22 assemblies, then sonicating them (for 5 min), shaking them (for 10 min at room temperature) and washing 23 them. Afterwards, three extra monolayers of polyelectrolytes (PAH/PSS/PAH) were deposited by following 24 the same strategy that was mentioned above. After the deposition of the last polyelectrolyte layer (PAH, 1 25 mL, 10 mg/mL in 0.05 M NaCl) and the dissolution of the CaCO3 template (EDTA, 1 mL, 0.2 M, pH 6), capsules 26 with IONCs embedded in the walls (WALL) were obtained. 27 Structural characterization 28 Transmission electron microscopy (TEM) images were collected with a Jeol JEM 1011 (Jeol, Japan) electron 29 microscope (Electron Microscopy Facility – Fondazione Istituto Italiano di Tecnologia) operating at an 30 acceleration voltage of 100 kV, and recorded with a 11 Mp fiber optical charge-coupled device (CCD) camera 31 (GatanOrius SC-1000). For the sample preparation, 3 µL of the diluted sample was dropped onto a carbon32 coated copper grid, and the solvent was removed by evaporation at room temperature. 33 In order to perform a TEM analysis on SKOV-3 cells with internalized magnetic materials, previously seeded 1 cells in 18 mm diameter cell culture chambers (105 per well) were incubated with 14 nm free IONCs, capsules 2 that had IONCs in the cavity (CAPS), or capsules with IONCs embedded in the wall (WALL) for 24 hours. 3 Afterwards, the cells were fixed for 2 hours in a fixative solution (2% Glutaraldehyde in buffer Na-Cacodylate 4 0.1M) and were further post-fixed (2 h) in a solution of 1% OsO4 in a 0.1 M Na-cacodylate buffer. 5 Subsequently, samples were stained overnight in a 1% Uranyl acetate aqueous solution at 4 °C to better 6 highlight the polymer and cellular sub-compartments. After several washes in water, the samples were 7 completely dehydrated with Ethanol and then infiltrated with epoxy Epon™ (TAAB) resin. Once the resin was 8 hardened for 48 h in an oven at 65 °C, 70 nm thick sections were cut with a Leica EMU C6 ultra-microtome. 9 High-resolution scanning electron microscopy (SEM) imaging was carried out using a JEOL JSM 7500FA (Jeol, 10 Tokyo, Japan) equipped with a cold Field Emission Gun, operating at an acceleration voltage of 10 kV. When 11 necessary, the samples were carbon coated with a 10 nm thick film using an Emitech K950X high vacuum 12 turbo system (Quorum Technologies Ltd, East Sussex - UK). The samples were observed using both a 13 secondary electron detector, in order to enhance the topography, and a backscattered electron detector 14 (which is sensitive to differences in molecular weight) in order to detect and show the presence of the IONCs 15 within the capsules. 16 Representative bright field images were taken with a Nikon A1, CFI Plan Apo VC 60x Oil objective confocal 17 microscope. 18 To estimate the amount of iron and calcium in the synthesized samples, inductively coupled plasma optical 19 emission spectrometry (ICP-OES) measurements were performed using an iCAP 6000 Series spectrometer 20 (Thermo Scientific). Briefly, 10 µL of the stock core/capsule samples with IONCs was dissolved in 1 mL of aqua 21 regia overnight. After adding MilliQ water until a volume of 10 mL was reached, the solution was filtered with 22 a 0.45 µm PFTE filter and the sample was analyzed. In order to estimate the concentration of iron in the cell23 associated magnetic materials, 10 µL of the sample was dispersed in 375 µL of a HNO3:H2O2 (vol. 2:1) mixture, 24 which was then sonicated at 65 °C for 2 h. Subsequently, 750 µL of HCl was added, and the samples were 25 digested overnight. The next day, after adding MilliQ water until a volume of 10 mL was reached, the solution 26 was filtered with a 0.45 µm PFTE filter and elemental analysis was performed. = 27 Magnetic characterization 28 The magnetic characterization of the synthesized magnetic materials was performed under an alternating 29 magnetic field (AMF) in colloidal dispersions (50 μL at a concentration of 1.5 g(Fe)/L)at different frequencies 30 (50, 100, and 200 kHz) and constant magnetic field amplitude (24 kA/m). AC hysteresis loops were traced at 31 room temperature with a home-made inductive magnetometer built by the Advanced Instrumentation Unit 32 (iMdea Nanociencia, Madrid, Spain) based on the one described by Connord et al.47 . The magnetic field was 33 generated by a Litz wire solenoid, inside which two counterwise-wound coils with the same diameter and 34 1 Figure 3: SEM images of the capsules prepared using a LbL procedure. A,B) Backscattered electron signal (BES) 2 of the CAPS (left column) and the WALL (right column), showing the distribution of the IONCs (16±2 nm) inside 3 the capsules (COMPO mode). C,D) Secondary electron imaging (SEI) of the capsules, highlighting the different 4 surface roughness of the CAPS and WALLS (middle row). E,F) Magnifications of the capsules’ edges show an 5 irregular surface roughness for WALL capsules (white arrows, F) in comparison to CAPS, where the bright 6 spots of the IONCs are more distributed within the polymer (E). Scale bars correspond to 100 nm. 7 8 Hyperthermia measurements 9 The specific absorption rate (SAR) values of the IONCs and of all the synthesized capsules were measured by 10 calorimetry at different frequencies (185 kHz and 302 kHz) and field intensities (28 and 24 kA/m), 11 corresponding to field conditions that fits into Brezovich limit that actually sets the suitable clinical setting 12 for magnetic hyperthermia (𝐻∙𝑓 5x109 Am-1s-1, in which H is the amplitude of the magnetic field and f is 13 the field frequency51). Figure 4 shows the SAR values that were obtained at 302 kHz and 24 kA/m for the free 14 IONCs of different sizes (from 14 to 18nm) and the related capsules dispersed in water. 15 A B CD EF 1 2 Figure 4: SAR values measured with the calorimetric method as obtained from free and encapsulated IONCs 3 (CORES CAPS and WALL) dispersed in water for different IONC sizes. 4 As is shown in Figure 4, the highest SAR values in aqueous media were obtained for free IONCs, increasing 5 with size. The related SAR values are comparable to those reported.9,15 However, SAR values obtained in the 6 capsules (i.e., CORES, CAPS and WALL) resulted non-significant differences (given the large error bars of the 7 calorimetric measurements) . Thus, SAR values were reduced between 30% and 70%, inside capsules and 8 depending on IONC size and encapsulation procedure albeit a little bit less for the CAPS. The reduction in 9 magnetic losses for IONCs of different sizes in the different capsulation modalities showed similar SAR 10 behavior, which became more pronounced when the IONC size is increased7,52 (Figure 4, S11, S18-S20). 11 However, immobilization was not the only effect reducing the heating mechanism for the 14 nm IONCs, which 12 were spatially confined into the CAPS, as also this sample showed a reduction in the SAR value (30%) with 13 respect to free IONCs. This reduction in the SAR value was lower than in the capsules with randomly 14 distributed IONPs, in which a drop of 50% was observed,32 underlining the relevance of the spatial control of 15 the IONCs in the employed encapsulation approach. This SAR reduction increased with an increase in the 16 IONCs' size, probably due to the prevalence of the Brownian relaxation that is significantly evident in the case 17 of the CORES and WALLS samples. It is well known that aggregation may also strongly shrink magnetic heat 18 losses when IONCs are embedded into capsules.32 Indeed, the interparticle distance and random spatial 19 distribution of IONCs in the distinct capsule arrangements (CORES, CAPS and WALL) cause different magnetic 20 dipolar interactions, which, may result in a reduction or increase of magnetic heat losses.53,54 Moreover, such 21 magnetic dipolar interactions increase when the IONC size is increased, which is in agreement with our 22 previous observations.16 23 IONCs CORES CAPS WALL 0 100 200 300 400 500 600 SAR [W/g] 14 nm 16 nm 18 nm 302 kHz, 24 kA/m In order to enlarge the range of sizes of IONCs, we tested the formation of capsules (CAPS and WALL types) 1 using also 8±1 nm and 21±2 nm IONCs (Figure S11). Based on previous results, due to the similar heating 2 abilities of CORES and WALL, we performed SAR measurements only of WALL sample modified with 8 and 21 3 nm IONCs. We observe the same trend in terms of structural properties, i.e. morphology and colloidal 4 parameters. Regarding the heating efficiency, indeed the 8 nm IONC sample was too small to provide 5 measurable SAR under the same measuring conditions used for the other IONC and capsules samples (300 6 kHz and 24 kA/m and 1 g(Fe)/L), nor any of the systems of CAPS and WALL made with IONCs of 8 nm. With 7 regard to the systems containing 21 nm IONCs, SAR values of IONCs in water are excellent (810±7 W/g) and 8 higher than any other system of CAPS or WALL samples presented in this work. Even after the encapsulation, 9 associated to a certain decrease in SAR again probably due to the prevalence of the Brownian relaxation, the 10 samples preserve high SAR values (CAPS: 380±25 W/g, WALL: 490±14 W/g). However, it is also worth to 11 mention that for 21 nm IONCs, the yield of capsules is lower than that for the other capsule preparation 12 (usually the yield is up to 30 % less than for the capsules samples made of IONCs with D < 21 nm). This may 13 be due to magnetic interaction during the preparations that promotes fast precipitations of nanoparticles 14 and less control on the protocol. 15 The common trend for encapsulated 14, 16 and 18 nm IONCs is clearly reflected in the dynamical hysteresis 16 loops that were obtained under similar HAC conditions (Figure 5, S25-S27). 17 18 Figure 5: AC hysteresis loops of the different capsules and free IONCs dispersed in water at given iron content 19 (1.5 g(Fe)/L), and field conditions (200 kHz and 24 kA/m) : (A) d(IONCs) = 14 nm; (B) d(IONCs) = 18 nm . 20 AC hysteresis loops are a direct and accurate method to probe and quantify the parameters that influence 21 the magnetic losses.26,30,55In agreement with calorimetry results, free 18 nm IONCs have the largest AC 22 hysteresis loop opening, followed by CAPS (blue curve), CORES (red curve), and WALL (pink curve) samples, 23 respectively. Confining the IONC within the capsules causes always a reduction on the related AC hysteresis 24 loop opening. This confirms the same behavior as the SAR calorimetric values, independent of the IONC size 25 AB -30 -20 -10 0 10 20 30 -100 -50 0 50 100 M [Am2/kg] H [kA/m] IONCs CAPS CORES WALL 200 kHz 14 nm -30 -20 -10 0 10 20 30 -100 -50 0 50 100 M [Am2/kg] H [kA/m] IONCs CAPS CORES WALL 200 kHz 18 nm (see, for instance, Figure 5A for 14 nm nanocubes). Moreover, as is indicated in Table S7, the area under the 1 hysteresis loops always increased when the edge length (14, 16, 18 nm) of the IONCs was increased. 2 It is also worth to mention that hysteresis Area and SAR values measured at clinical conditions used for the 3 treatment of glioblastoma multiforme (100 kHz and 24kA/m), are significant not only for the free nanocubes 4 but also for all the capsule’s samples (Table 1, in this latter cases the values are certainly reduced). 5 A(mJ/kg)/SAR(W/g) at 100 kHz and 24 kA/m IONCs, 14 nm 2329/233 CORES, 14 nm 1183/118 CAPS, 14 nm 1637/164 WALL, 14 nm 975/93 IONCs, 18 nm 3018/302 CORES, 18 nm 1249/125 CAPS, 18 nm 2004/200 WALL, 18 nm 1842/184 6 Table 1: Areas and SAR values obtained by AC magnetometry56 from the distinct magnetic nanomaterials at 7 clinical conditions (100 kHz and 24 kA/m). 8 9 Probing the magnetic heating efficiency of magnetic materials inside cells 10 As previously mentioned, nanoparticle aggregation,30,57 immobilization30 and degradation28 occur when 11 magnetic nanostructures interact with cells, resulting in significant variations in the heating efficiency with 12 respect to the values that are obtained in colloidal dispersions. Their effects on the dynamical magnetic 13 response should be avoided in order to control the physicochemical properties of the synthesized 14 nanoparticles so that the heat dose supplied by the IONCs inside the cells or tissues can be preserved. To 15 investigate how the encapsulation of IONCs in polymeric capsules may preserve the magnetic losses once 16 associated to the cells, SAR values were measured during their interaction with the cells. For these 17 experiments, we used both, CAPS capsules with IONCs inside the hollow cavity and WALL capsules in which 18 the IONCs were randomly immobilized. In the latter case, since IONCs are immobilized at fixed at average 19 interparticle distances inside the CORES capsules, no significant differences were expected in terms of heat 20 efficacy. With regard to the capsules containing IONCs, we selected only the 14 and 18 nm IONCs. The SAR 21 values were monitored after different incubation times (0, 15, 45, 90, 180 min). In the case of free IONCs 22 (black curves), the SAR values drastically decreased to different extents depending on the nanoparticle size 1 when the incubation time was increased (about 54% for 14 nm IONCs and 60% for 18 nm IONCs) (Figure 6A 2 and 6B). Only slight reductions in the SAR values (ca. 20 %) were observed for CAPS (blue curves) and WALL 3 (pink curves) capsules, independent of the IONC size. Figure 6A and B show the variation in the SAR values 4 for different magnetic nanostructures over time (< 3 h) . 5 Interestingly, after 180 min of incubation, the cells that have received IONCs were attracted to the magnet 6 within 10 s, whereas pristine free IONCs (i.e. IONCs that were not incubated with cells) did not show any 7 visible response to the magnetic field (Figure S28). This suggests that, after 3 h of incubation, the IONCs were 8 associated with the tumour cells, which might be also promoted by electrostatic interactions and protein 9 serum absorption on capsules and IONCs.58,59 A significant and progressive reduction of more than 50% was 10 observed for free IONCs after just a few hours. During this period, the SAR values of capsules varied by less 11 than 30%. It is important to highlight that no variations in the SAR values were obtained for CAPS and WALL 12 capsules within the time frame of the study. However, for free IONCs, changes in the SAR were time 13 dependent (see Table S8 for summary of SAR values from calorimetric measurements in cells). This is in 14 agreement with the observations made by Di Corato et al.,9 who state that the SAR dynamics during the 15 incubation of nanocubes and other types of iron oxide nanoparticles show a significant reduction in SAR 16 values. Indeed, under this condition a fraction of the IONCs was associated with cell plasma membrane or 17 internalized while a fraction of IONCs was still not yet internalized by the cells contributing more to the 18 heating. The drop of SAR values over time is caused by the decrease of the fraction of free IONCs/capsules in 19 the media and the increase of fraction of IONCs/capsules that were internalized/associated particles able to 20 contribute to the heating efficiency. 60 21 1 Figure 6: Time-dependent SAR values of free and encapsulated IONCs during incubation with cells at different 2 time points at 302 kHz and 24 kA/m. A. d(IONCs) = 14 nm, B. d(IONCs) = 18 nm. Time-dependent iron 3 concentration of IONCs associated with cells. C. d(IONCs) = 14 nm, D. d(IONCs) = 18 nm. Time-dependent SAR 4 values of magnetic nanostructures during incubation with cells after 24 and 48 hours for E. d(IONCs) = 14 nm 5 and F. d(IONCs) = 18 nm. Time-dependent iron concentration of IONCs internalized by the cells G. d(IONCs) = 6 14 nm, H. d(IONCs) = 18 nm. 7 8 045 90 135 180 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 c [g(Fe)/L] t [min] IONCs CAPS WALL d=14 nm 045 90 135 180 0 50 100 150 200 250 300 350 400 SAR [W/g] t [min] IONCs CAPS WALL 302 kHz, 24 kA/m d=14 nm 045 90 135 180 0 50 100 150 200 250 300 350 400 SAR [W/g] t [min] IONCs CAPS WALL 302 kHz, 24 kA/m d=18 nm A B 24 48 0 50 100 150 200 250 300 350 400 SAR [W/g] t [h] free IONPs CAPS WALL 302 kHz, 24 kA/m d=14 nm 24 48 0 50 100 150 200 250 300 350 400 SAR [W/g] t [h] IONCs CAPS WALL 302 kHz, 24 kA/m d=18 nm C D E F 045 90 135 180 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 c [g(Fe)/L] t [min] IONCs CAPS WALL d=18 nm 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 48 c [g(Fe)/L] t [h] IONCs CAPS WALL 24 d=14 nm 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 48 c [g(Fe)/L] t [h] IONCs CAPS WALL 24 d=18 nm G H in suspension in adhesion At longer incubation times, the magnetic heat losses of the different magnetic nanostructures studied here 1 have similar trends: CAPS have higher SAR values, closely followed by WALL samples. For free 14 and 18 nm 2 IONCs incubated for 24 or 48 hours, there was a strong decrease in the SAR values as compared to the SAR 3 value of the capsule systems in water (about 68% for 14 nm IONCs and 76% for 18 nm IONCs). This effect was 4 independent of the IONC size. For the CAPS and WALL capsules, the SAR values at 24 and 48 hours were 5 similar and much larger than the free IONC values (Fig.6E and 6F). Nevertheless, a reduction in SAR values 6 over time, ranging from 10% up to 30%, was observed depending on the IONC size. These data show that the 7 magnetic losses of the CAPS and WALL capsules reduce more slowly inside cells and have higher values than 8 free IONCs. The preservation of magnetic losses by the polyelectrolyte encapsulation fails with time. After 48 9 h, SAR values were lower than they were after 24 h, suggesting that an internalization of IONCs/capsules 10 occurs, gradually causing sample aggregation. In the case of the capsules, which have a certain stiffness, their 11 deformation with time after internalization likely favors an enhancement of intracellular IONC clustering, 12 leading to a decrease in the SAR value.61 13 It is also worth noting that there is a difference in the SAR values for the free IONCs when the experiment 14 was performed with the cells in suspension (short incubation time, Fig. 6A and 6B) and with adherent cells 15 (long incubation, Figure 6E and 6F). Indeed, the SAR values for cells in suspensions were always higher than 16 those for adherent cells. It could be that, at a short incubation time, a fraction of the IONCs were associated 17 with cell plasma membrane while a large fraction of IONCs were still not yet internalized by the cells thus 18 having the free IONCs contributing more to the heating. In the case of adherent cells, considering the longer 19 incubation times, a larger fraction of IONCs is associated or internalized by the cells. This can explain why the 20 SAR values that were measured for the free IONCs in suspension are higher than those for adherent cells. 21 However, SAR values obtained for CAPS and WALL capsules on adherent cells or on suspended cells are more 22 directly comparable. A possible explanation is that the polymer layers of the capsules act as a sort of barrier 23 to soften any changes in the environment, therefore the SAR is less dependent on the cellular association or 24 internalization. 25 26 It is worth mentioning that IONCs and polyelectrolyte capsules were internalized by cells with a different 27 rate. Given that the same amount of material was administered in terms of the amount of iron, it can be seen 28 that, in terms of iron uptake, free IONCs were always taken up to a lesser extent than when they were packed 29 in the capsules at 24 hours. This is reasonably due to the different charge of the free IONCs with respect to 30 the CAPS and WALLS. Indeed, free IONCs are negatively charged while CAPS and WALL are positively charged 31 (Figure S16).Previous works report that the positively charged nanoparticles are internalized by the cells at 32 higher extent compared to the positively charged ones,62-64 due to the higher interactions with the negatively 33 charged cell membrane. Moreover, the PEG coating of IONCs may seriously decrease protein corona 34 formation and, thus, the reduce the interaction with cell plasma membrane.65,66 However, the amount of iron 1 in the cells reaches the same amount for the 18 nm samples at 48 hours, while the amount of iron tends to 2 decrease progressively with time for the free 14 nm nanocubes. Indeed, the amount of iron was significantly 3 lower at 48 hours than that found in the same IONCs once they were incorporated into the cells (Figure 6C, 4 6D, 6G and 6H). 5 6 The intracellular localization of IONCs and capsules was studied using TEM images of the cells that had been 7 exposed to the magnetic materials for 24 h of incubation at 37°C.59 Figure 6 shows that the free IONCs 8 accumulated inside the intracellular vesicles following a typical endocytic pathway.67–69 Very closed packed 9 IONCs within the vesicles can be clearly seen, and these could be responsible for the observed decrease in 10 the SAR of free cubic IONCs. Instead, CAPS capsules show a broader spatial distribution of IONCs with higher 11 distances between them. In the case of WALL capsules, the darker elongated structures (black arrows) can 12 be attributed to the walls of the polyelectrolyte capsules, which are decorated with IONCs in a darker color. 13 With respect to the cells treated with the free IONCs, the lower density of the IONCs inside the hollow cavities 14 or in the polymeric walls suggests that IONCs have a certain degree of freedom and that aggregation has 15 been prevented to a certain extent. 16 All this evidence brings us to an important conclusion: the intracellular matrix does not alter the magnetic 17 losses of IONCs when they are packed with the right configuration. This proves that a polyelectrolyte capsule 18 shell can preserve IONCs from severe aggregation (i.e. the IONCs can still partly move, though there may be 19 attractive van der Waals forces between them) and immobilization effects. Consequently, this can 20 significantly alter the SAR values for free IONCs, as has often been reported in literature.9 21 22 1 Figure 7: Representative TEM micrograph sections of SKOV-3 cells incubated for 24 h with: free 14 nm IONCs 2 (Row A); CAPS (Row B); WALL (Row C.). The black arrows highlight the capsules’ wall structures. 3 4 Influence of viscosity and aggregation on the magnetic response of free IONCs and capsules 5 To better understand the cell internalization effects on the magnetic heat losses of the different magnetic 6 nanomaterials described above, we assessed the viscosity and aggregation effects on the SAR values of the 7 different magnetic nanostructures studies dispersed in aqueous media.30 The cellular association and further 8 internalization of the IONCs/capsules further suppresses Brownian relaxation, since the IONCs/capsules 9 cannot physically move. However, the aggregation of nanoparticles, which can take place on the cell plasma 10 membrane or inside the endocytic vesicles, may also influence the Néel relaxation. To elucidate the effect of 11 IONCs on membrane association, the first experiment that we set aimed to study the SAR behavior of the 12 capsules in viscous media, which suppresses Brownian rotation. The SAR values were measured in different 13 water-glycerol (W:Gly) mixtures set to 64:36 v% and 19:81 v%, corresponding to a mean η of 3.8 and 97.3 14 mPa·s respectively. As is shown in Figure 8and Figures S21-S24, S31A, for the 14 nm IONC sample, the SAR 15 values for the IONCs in water or in a glycerol solution are similar, indicating that the IONCs have a viscous 16 independent SAR behavior. We would expected that the SAR values for the capsule samples (CORES, CAPS 17 and WALL samples) would be similar to the SAR values of the initial IONCs, if no other immobilization effects 18 occurs. However, the SAR values of all the capsules prepared with 14 nm IONCs are always lower than the 19 SAR values of the 14 nm IONCs. This difference might be due to an additional immobilization or aggregation 20 effect, which could occur when the IONCs are associated/incorporated into the capsules. This SAR reduction 21 was also recorded for the other IONCs (16, 18 and 21 nm), but the dependency of the SAR on the viscosity in 22 these cases makes it more difficult to distinguish the effect due to the viscosity and the 1 immobilization/aggregation effects in the capsules. 2 Moreover, the SAR value of free 18 nm IONCs decreased when there was an increase in the viscosity of the 3 medium, whereas the SAR value remained constant at different viscosities for 14 nm IONCs. These results 4 are in agreement with previous studies, which claim that the Brownian relaxation magnetic mechanism starts 5 to have a significant contribution when the size of the IONCs is increased.16,26 Interestingly, for capsules with 6 IONCs in the cavity, the SAR remained constant in the different viscous media, while it dropped slightly for 7 capsules/cores with immobilized nanocubes (WALL and CORES). This suggests that IONCs in the cavity of the 8 capsule had a certain freedom to move, and they remained more isolated by the environment. We might 9 speculate that glycerol cannot enter through the capsules walls leaving the capsules mainly filled of water. 10 Thus, the heat dissipation processes remained constant, and the heat efficiency of the capsules depended 11 less on the viscous environment in which they were dispersed. 12 13 Figure 8: SAR values of different magnetic nanostructures dispersed in different viscous media at 320 kHz and 14 24 kA/m for (A) 14 nm IONCs and (B) 18 nm IONCs. 15 On the other hand, aggregation of the magnetic materials can occur during cell uptake, resulting in a 16 reduction in SAR values. In order to study the aggregation effect, we intentionally induced 17 nanoparticle/capsule aggregation by dispersing the samples in a concentrated phosphate buffer solution 18 (PBS 5, 0.05 M). The ions in the PBS solution contribute to screening the surface charge of colloids, so that 19 van der Waals forces start to dominate and IONCs or capsules lose their colloidal stability and tend to 20 aggregate.70 The effect of the aggregation of the IONCs/capsules was determined from the SAR values that 21 were obtained from the AC hysteresis loops (Figure S29, 30, 31B). As expected, a significant decrease in the 22 SAR value of free IONCs was observed, with the aggregation in PBS (Figure 9) being more pronounced for 14 23 nm than for 18nm IONCs. In contrast, the SAR values of capsules showed just slight variations when they 24 were dispersed in PBS rather than in water (Figure 9). These findings strongly suggest that the capsule provide 25 IONCs CORES CAPS WALL 0 100 200 300 400 500 600 SAR [W/g] water glycerol 36% glycerol 81% 302 kHz, 24 kA/m d=14 nm IONCs CORES CAPS WALL 0 100 200 300 400 500 600 SAR [W/g] water glycerol 36% glycerol 81% 302 kHz, 24 kA/m d=18 nm A B (52) Lévy, M.; Wilhelm, C.; Siaugue, J.-M.; Horner, O.; Bacri, J.-C.; Gazeau, F. Magnetically Induced 1 Hyperthermia: Size-Dependent Heating Power of γ-Fe2O3 Nanoparticles. J. Phys. Condens. Matter 2 2008, 20, 204133. 3 (53) Niculaes, D.; Lak, A.; Anyfantis, G. C.; Marras, S.; Laslett, O.; Avugadda, S. K.; Cassani, M.; Serantes, D.; 4 Hovorka, O.; Chantrell, R.; Pellegrino, T. Asymmetric Assembling of Iron Oxide Nanocubes for 5 Improving Magnetic Hyperthermia Performance. ACS Nano 2017, 11, 12121–12133. 6 (54) Branquinho, L. C.; Carrião, M. S.; Costa, A. S.; Zufelato, N.; Sousa, M. H.; Miotto, R.; Ivkov, R.; Bakuzis, 7 A. F. Effect of Magnetic Dipolar Interactions on Nanoparticle Heating Efficiency: Implications for 8 Cancer Hyperthermia. Sci. Rep. 2013, 3, 1-10. 9 (55) Ovejero, J. G.; Cabrera, D.; Carrey, J.; Valdivielso, T.; Salas, G.; Teran, F. J. Effects of Interand Intra10 Aggregate Magnetic Dipolar Interactions on the Magnetic Heating Efficiency of Iron Oxide 11 Nanoparticles. Phys. Chem. Chem. Phys. 2016, 18, 10954–10963. 12 (56) Mehdaoui, B.; Carrey, J.; Stadler, M.; Cornejo, A.; Nayral, C.; Delpech, F.; Chaudret, B.; Respaud, M. 13 Influence of a Transverse Static Magnetic Field on the Magnetic Hyperthermia Properties and High14 Frequency Hysteresis Loops of Ferromagnetic FeCo Nanoparticles. Appl. Phys. Lett. 2012, 100, 052403. 15 (57) Levy, M.; Wilhelm, C.; Luciani, N.; Deveaux, V.; Gendron, F.; Luciani, A.; Devaud, M.; Gazeau, F. 16 Nanomagnetism Reveals the Intracellular Clustering of Iron Oxide Nanoparticles in the Organism. 17 Nanoscale 2011, 3, 4402. 18 (58) Zyuzin, M. V.; Yan, Y.; Hartmann, R.; Gause, K. T.; Nazarenus, M.; Cui, J.; Caruso, F.; Parak, W. J. Role 19 of the Protein Corona Derived from Human Plasma in Cellular Interactions Between Nanoporous 20 Human Serum Albumin Particles and Endothelial Cells. Bioconjug. Chem.2017, 28, 2062–2068. 21 (59) Kastl, L.; Sasse, D.; Wulf, V.; Hartmann, R.; Mircheski, J.; Ranke, C.; Carregal-Romero, S.; Martínez22 López, J. A.; Fernández-Chacón, R.; Parak, W. J.; Elsasser, H.P.; Rivera Gil, P. Multiple Internalization 23 Pathways of Polyelectrolyte Multilayer Capsules into Mammalian Cells. ACS Nano 2013, 7, 6605–6618. 24 (60) Lak, A.; Cassani, M.; Mai, B. T.; Winckelmans, N.; Cabrera, D.; Sadrollahi, E.; Marras, S.; Remmer, H.; 25 Fiorito, S.; Cremades-Jimeno, L.; et al. Fe2+ Deficiencies, FeO Subdomains, and Structural Defects Favor 26 Magnetic Hyperthermia Performance of Iron Oxide Nanocubes into Intracellular Environment. Nano 27 Lett. 2018, 18, 6856–6866. 28 (61) Bédard, M. F.; Munoz-Javier, A.; Mueller, R.; del Pino, P.; Fery, A.; Parak, W. J.; Skirtach, A. G.; 29 Sukhorukov, G. B. On the Mechanical Stability of Polymeric Microcontainers Functionalized with 30 Nanoparticles. Soft Matter 2009, 5, 148–155. 31 (62) Kim, B.; Han, G.; Toley, B. J.; Kim, C.; Rotello, V. M.; Forbes, N. S. Tuning Payload Delivery in Tumour 32 Cylindroids Using Gold Nanoparticles. Nat. Nanotechnol. 2010, 5, 465–472. 33 (63) Bannunah, A. M.; Vllasaliu, D.; Lord, J.; Stolnik, S. Mechanisms of Nanoparticle Internalization and 1 Transport Across an Intestinal Epithelial Cell Model: Effect of Size and Surface Charge. Mol. Pharm. 2 2014, 11, 4363–4373. 3 (64) Oh, N.; Park, J.-H. Surface Chemistry of Gold Nanoparticles Mediates Their Exocytosis in Macrophages. 4 ACS Nano 2014, 8, 6232–6241. 5 (65) Walkey, C. D.; Chan, W. C. W. Understanding and Controlling the Interaction of Nanomaterials with 6 Proteins in a Physiological Environment. Chem. Soc. Rev. 2012, 41, 2780–2799. 7 (66) Gref, R.; Lück, M.; Quellec, P.; Marchand, M.; Dellacherie, E.; Harnisch, S.; Blunk, T.; Müller, R. . 8 ‘Stealth’ Corona-Core Nanoparticles Surface Modified by Polyethylene Glycol (PEG): Influences of the 9 Corona (PEG Chain Length and Surface Density) and of the Core Composition on Phagocytic Uptake 10 and Plasma Protein Adsorption. Colloids Surf. B Biointerfaces 2000, 18, 301–313. 11 (67) Feliu, N.; Hühn, J.; Zyuzin, M. V.; Ashraf, S.; Valdeperez, D.; Masood, A.; Said, A. H.; Escudero, A.; Pelaz, 12 B.; Gonzalez, E.; Duarte, M.A.C.; Roy, S.; Chakraborty, I.; Lim, M.L.; Sjöqvist, S.; Jungebluth, P.; Parak, 13 W.J. Quantitative Uptake of Colloidal Particles by Cell Cultures. Sci. Total Environ. 2016, 568, 819–828. 14 (68) Rahmani, S.; Ashraf, S.; Hartmann, R.; Dishman, A. F.; Zyuzin, M. V.; Yu, C. K. J.; Parak, W. J.; Lahann, J. 15 Engineering of Nanoparticle Size Via Electrohydrodynamic Jetting. Bioeng. Transl. Med. 2016, 1, 82– 16 93. 17 (69) Iversen, T.-G.; Skotland, T.; Sandvig, K. Endocytosis and Intracellular Transport of Nanoparticles: 18 Present Knowledge and Need for Future Studies. Nano Today 2011, 6, 176–185. 19 (70) Zyuzin, M. V.; Honold, T.; Carregal-Romero, S.; Kantner, K.; Karg, M.; Parak, W. J. Influence of 20 Temperature on the Colloidal Stability of Polymer-Coated Gold Nanoparticles in Cell Culture Media. 21 Small 2016, 12, 1723–1731. 22 23