ECM Mechano-Sensing Regulates Cytoskeleton Assembly and

Jun 8, 2017 - It is also well-known that cells are able to sense and respond to the local physical environment and that the substrate stiffness, and n...
0 downloads 0 Views 1MB Size
Subscriber access provided by Binghamton University | Libraries

Article

The ECM mechano-sensing regulates cytoskeleton assembly and receptor-mediated endocytosis of nanoparticles. Valeria Panzetta, Daniela Guarnieri, Antonio Paciello, Francesca Della Sala, Ornella Muscetti, Luca Raiola, Paolo Antonio Netti, and Sabato Fusco ACS Biomater. Sci. Eng., Just Accepted Manuscript • DOI: 10.1021/acsbiomaterials.7b00018 • Publication Date (Web): 08 Jun 2017 Downloaded from http://pubs.acs.org on June 13, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Biomaterials Science & Engineering is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

The ECM mechano-sensing regulates cytoskeleton assembly and receptormediated endocytosis of nanoparticles Valeria Panzettaa, Daniela Guarnieria, Antonio Pacielloa, Francesca Della Salaa, Ornella Muscettia, Luca Raiolaa, Paolo Nettia,b, Sabato Fusco*,a

a

Center for Advanced Biomaterials for Health Care IIT@CRIB, Istituto Italiano di Tecnologia,

L.go Barsanti e Matteucci 53, Naples, 80125, Italy f

Interdisciplinary Research Centre on Biomaterials, CRIB and Department of Chemical, Materials

& Industrial Production Engineering, University of Naples Federico II, 80125 Naples, Italy.

*Corresponding Author: [email protected], L.go Barsanti e Matteucci 53, Naples, 80125, Italy tel: +3908119933100; fax +390817682404

Abstract. It is possible to create sophisticated and target-specific devices for nanomedicine thanks to technological advances in the engineering of nano-materials. When on target, these nanocarriers often have to be internalized by cells in order to accomplish their diagnostic or therapeutic function. Therefore, the control of such uptake mechanism by active targeting strategy has today become the new challenge in nanoparticle designing. It is also well known that cells are able to sense and respond to the local physical environment and that the substrate stiffness, and not only the nanoparticle design, influences the cellular internalization mechanisms. In this frame, our work reports on the cyclic relationship among substrate stiffness, cell cytoskeleton assembly and internalization mechanism. Nanoparticles uptake has been investigated in terms of the mechanics of cell environment, the resulting cytoskeleton activity and the opportunity of activate molecular specific molecular pathways during the internalization process. To this aim the surface of 100 nm polystyrene nanoparticles was decorated with a tri-peptide (RGD and a scrambled version as a control), which was able to activate an internalization pathway directly correlated to the dynamics of the cell cytoskeleton, in turn, directly correlated to the elastic modulus of the substrates. We found that the substrate stiffness modulates the uptake of nanoparticles by regulating structural parameters of bEnd.3 cells as spreading, volume, focal adhesion and mechanics. In fact, the

ACS Paragon Plus Environment

1

ACS Biomaterials Science & Engineering

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 21

nanoparticles were internalized in larger amounts both when decorated with RGD, which activated an internalization pathway directly correlated to the cell cytoskeleton, and when cells resided on stiffer material that, in turn, promoted the formation of a more structured cytoskeleton. This evidence indicates the directive role of the mechanical environment on cellular uptake of nanoparticles, contributing new insights to the rational design and development of novel nanocarrier systems.

Keywords: Cell Mechanics; Nanoparticles Uptake; ECM Stiffness; Cytoskeleton Assembly; Receptor Mediated Endocytosis

1. Introduction.

The last decades have seen the growth of several nano-technological advances which have led to the development of much more advanced delivery systems. By engineering the materials, new nanoparticle (NP)-based systems have been created, making site-specific targeting possible for diagnostic aims or for the release of therapeutic agents. Thanks to the recent revelation of specific affinities between cellular receptors and their equivalent ligands and due to their notable exposure on the surface of second generation NPs, it is now possible to achieve enhanced accumulation in target sites such as organs, tissues and cells 1-3. When the NPs have reached their specific target and are carrying out their functions they need to be internalized by the cells. Among the diverse endocytic pathways, pynocytosis is the one common to all types of cells. It can also be classified in various sub-pathways: clathrin-mediated, caveolin-mediated, clathrin/caveolin independent and macropynocytosis. Changes in membrane shape, cell structure and organization are all associated with each of these mechanisms and, more generally, to endocytosis. NP internalization mechanisms are also strictly correlated to the cell cytoskeleton (specifically to the acto-myosin network architecture) and to cells ability to mechanosense the surrounding extracellular matrix (ECM). In fact, it is well known that the biophysical properties of ECM influence the assembly of the cell cytoskeleton and, consequently, such peculiar cellular functions as proliferation, migration, and differentiation 4-9. Recent research has also confirmed a correlation between the ECM stiffness and cell uptake of carboxylated polystyrene NPs

10

. This aspect is

particularly intriguing, considering that the organs and tissues present in the human body possess ECM with very different biophysical properties. The physiologic range of ECM stiffness, in fact, can span from 0.1 to several tens of kPa

11-13

abnormalities in the cytoskeleton mechanics

. Many diseases have also been associated with 14-23

. In particular, a less organized and rigid

ACS Paragon Plus Environment

2

Page 3 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

cytoskeleton has been observed in a large variety of cancers (breast, lung, prostate), where often a stiffer ECM is also linked with a higher grade of malignancy

18, 19, 24, 25

. In summary, the above

investigations highlight an active involvement of the cytoskeleton assembly in the cell uptake mechanism, but its level of participation in endocytosis is still unclear, especially in brain endothelial cells. Hence, unravelling the role of the interplay ECM/cytoskeleton and its contribute in the cell internalization process may give a solid contribution in the design and development of new nanocarrier systems based on this mechanisms. This report explains the interplay between substrate mechanical properties (mechanosensing), cell cytoskeleton assembly (mechanotransduction) and internalization mechanism of engineered NPs (uptake). To this purpose, the surface of the fluorescently-labelled, carboxyl-modified polystyrene nanoparticles with a 100 nm diameter was covalently modified with the RGD tri-peptide which, by the engagement of a specific ligand-receptor complex (RGD-integrin), was able to activate an internalization pathway directly correlated to the dynamics of the cell cytoskeleton 26. Mouse brain endothelial bEnd.3 cells were used as a model of brain endothelium to study NP uptake. Endothelium is one of the main barriers that nanoparticles meet, once injected into the human body, and represents one of the main goals of many nanoparticle-based therapeutic strategies 27-29. In this project, the mechanobiology of bEnd.3 cells was analyzed in response to soft and stiff substrates and related to the cell uptake of RGD-functionalized NPs. RDG scrambled peptide conjugated to nanoparticles was also used as a control. The amount of internalized NPs, uptake kinetics and mechanisms were also evaluated when the cells were cultured on polyacrylamide (PAAm) substrates which were mimicking different ECM mechanical conditions. In particular, we designed PAAm substrates with different values of Young’s modulus. The values of Young’s modulus were comprised between 3 kPa, which is inside the physiological range measured for intact brain tissue 30

, and 30 kPa, which could be representative of the mechanical rigidity of brain tumor, such as

glioblastoma multiforme tumor 7. The correspondences which were found between the NP cell uptake and the cell biophysical parameters (cell spreading, cell volume, focal adhesion length, cell apparent elastic modulus) modulated by the substrate stiffness were analysed in terms of NP functionalization and correlated to the involvement of the cell cytoskeleton in the internalization process. Results confirmed that, by the activation of cell structures (integrins) directly correlated to the cytoskeleton, NP uptake is highly influenced by ECM elastic modulus. This results open new routes in the design and development of carriers for nanomedicine suggesting the ECM stiffening (or softening) as additional parameter to be considered.

2. Materials and Methods ACS Paragon Plus Environment

3

ACS Biomaterials Science & Engineering

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 21

2.1 Materials Fluospheres®carboxylate (505/515) modified polystyrene microspheres 0.1 µm yellow-green fluorescent 2% solids (NPs) were purchased from Invitrogen™ Molecular Probes™ (F8803). N-(3dimethylaminopropyl)-N`-ethylcarbodiimide

hydrochloride

bio

extra

(EDAC,

03450),

Diisopropylcarbodiimide (DIC, D125407), Triethylamine (TEA, T0886), 4-Dimethylaminopyridine (DMAP, 107700), Diisopropylethylamine (DIPEA, 387649), Hydroxybenzotriazole (HOBt, 54802), QuantiPro™ BCA Assay Kit (QPBCA), Piperidine (411027), N-methylpirrolidine (NMP, M79204), Trifluoroacetic acid (TFA, T62200), thioanisole (TIS, 88470), Dimethylformamide (DMF, 437673), Diethyl ether (296082, Acetone (439126), Acetic anhydride (242845), Triisopropylsilane (233781), Ethylenediaminetetraacetic acid (EDTA, EDS), Phosphate Buffer Powder (P7994), Hydrochloric acid (435570), Sodium hydroxide (1.06469), Deuterium oxide (151882) and water purity solvent (W4502) were purchased from Sigma-Aldrich. Fmoc-Glycine (FAA1050), Fmoc-aspartic acid(trt) (HAA1073), Fmoc-arginine(pfb) (FAA1010), O-(Benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU) Oxima Pure (RL1030), Wang Resin (BR1420) were purchased from Iris Biotech German. Reagents and solvents were used without further purification unless otherwise specified. Spectra/Por® Dialysis membrane MWCO: 6-8,000 (132660) was purchased from Spectrum Laboratories, Inc.

2.2 Substrate Preparation and Mechanical Characterization PAAm gels were attached to fluorodish cell culture dish (World Precision Instruments, FD35-100). The fluorodishes were placed on a hot place (50-60 °C), covered with 500 µl of 20 mM NaOH. After evaporation, NaOH solution formed a thin semi-transparent film on the fluorodishes. 250 µl of 3-aminopropyltriethoxysilane (APTES, Sigma-Aldrich, 440140) was added to the fluorodishes in the fume hood for 10 min. The fluorodishes were rinsed with distilled H20 at least three times for 5 min each, until the unreacted APTES was completely removed. The fluorodishes were covered with 0.5% glutaraldehyde (Agar Scientific, R1020) solution in PBS for 30 min. The fluorodishes were extensively washed with distilled H20 and air dried. PAAm substrates were prepared by mixing acrylamide (Sigma-Aldrich, A4058), methylene-bis-acrylamide (Sigma-Aldrich, M1533), 1/100 total volume of 10% ammonium persulfate (Sigma-Aldrich, A3678) and 1/1000 total volume N,N,N’,N’-tetramethylethylenediamide

(TEMED,

Sigma-Aldrich,

T7024).

Two

different

combinations of acrylamide and bis-acrylamide were used to obtain 3 kPa (6 wt/vol% acrylamide and 0.06 wt/vol% bis-acrylamide) and 30 kPa (10 wt/vol% acrylamide and 0.3 wt/vol% bis-

ACS Paragon Plus Environment

4

Page 5 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

acrylamide) hydrogels. 20 µl of acrylamide/methylene-bis-acrylamide mixture were pipetted on the treated fluorodish and an untreated coverslip (15 × 15 mm #1, Thermo Scientific, BBAD01500150#1) was placed under the drop. The gel was allowed to polymerize for 30 min, then the untreated coverslip was removed and the hydrogel was rinsed twice for 10 min each time. The resulting gels were ̴ 100-120 µm thick as measured by microscopy. To allow for cell adhesion, substrates were functionalized with collagen, by using a bi-functional photo-linker, Nsulphosuccinimidyl-6-(4’-azido-2’-nitrophenylamino)hexanoate (sulpho-SANPAH, ThermoFisher Scientific, 22589) as a cross-linking agent to immobilize collagen. The freshly prepared sulphoSANPAH solution at a concentration of 0.2 mg/ml was placed onto PAAm substrates and exposed to 365 nm UV light (100 W, 0.8 A; Ted Pella, Inc.) for 10 min. After washing with PBS, the hydrogels were coated with 50 µg/ml of bovine type I collagen (Sigma-Aldrich, C4243) overnight at room temperature (RT). Samples were washed three times with PBS.

2.3 Cell Culture Mouse brain endothelial bEnd.3 cells (ATCC) were cultured with high glucose Dulbecco’s Modified Eagle’s medium (DMEM, Gibco, 21063045) supplemented with 10% fetal bovine serum (FBS, Gibco, 10270-106), 100 U/ml penicillin (Pen), 100 U/ml streptomycin (Strep, Pen-Strep solution, Gibco, 15070063) and 2× non-essential amino-acids and sodium pyruvate (SigmaAldrich). Cells were maintained in 100 mm diameter cell culture dishes in a humidified and controlled atmosphere at 37 °C and 5% CO2. Medium was changed every 3-4 days. For the experiments, collagen-functionalized PAAm substrates in glass-bottom 35 mm diameter cell culture dishes were first sterilized with a solution of Pen / Strep-PBS 1:1 for 24h at RT and then exposed to UV light emitted by germicidal lamp (Sankyo Denki, G30T8) for 1h. Then, samples were washed twice with PBS and incubated with cell culture medium for 1h at RT. After that, 1 × 104 cells were suspended in 1 ml of cell culture medium and seeded on each substrate.

2.4 Cell Adhesion Analysis For cell spreading and focal adhesion (FA) analyses, 24h after seeding onto the PAAm substrates, bEnd.3 cells were fixed with 4% paraformaldehyde (Sigma-Aldrich, 158127) for 20 min at RT. The fixed cells were permeabilized with 0.1 % Triton X-100 (Sigma-Aldrich, T8787) in PBS for 10 min. The actin filaments were stained with TRITC phalloidin (Sigma-Aldrich, P1951) in PBS for 30 min at RT. FAs were localized by rabbit monoclonal paxillin antibodies (Abcam, ab32084) and Alexa488 goat anti-rabbit secondary antibodies (Life Technologies, A11008). Finally, cell nuclei were stained with 4’,6-diamidino-2-phenylindole, DAPI, (Sigma-Aldrich, D9542). Images of the ACS Paragon Plus Environment

5

ACS Biomaterials Science & Engineering

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

specimens were taken using an Olympus IX81 inverted microscope and a 10× objective to quantify cell spreading and a 100× objective to quantify FA length. Fluorescent images were imported into ImageJ software (NIH, Bethesda, MD, USA) for post-processing analysis and quantification of the cell area (cellular footprint) and FA length. Individual cells were thresholded manually on the basis of the phalloidin staining and the spreading area of their thresholded bodies was determined using a custom macro ImageJ. To quantify the focal adhesion average length, the paxillin images were assembled into a stack. First the stack was Gaussian-filtered using a radius of 30 pixels. This stack was then subtracted from the original in order to reduce diffuse the background signal. Adhesions were measured by thresholding the stacks and using an ellipse-fitting function in ImageJ. Objects with area ≤ 0.1 µm2 were discarded, in order to avoid possible errors due to background noise. Lengths of individual focal adhesions were determined on both substrates. For cell volume analysis, cells were stained with Cell Tracker green (Life Technologies, C7025) according to the manufacturer’s procedure before seeding on PAAm substrates. Then, 1 × 104 cells in 1 ml of cell culture medium were seeded on each substrate and incubated for 24 h at 37°C. Zsectioning images of the cells were acquired by a confocal microscope SP5 (Leica) equipped with a 40× water immersion objective. About 20-30 slices of 0.49 µm thickness and 1024×1024 pixel resolution were acquired for each z-stack. Cell volume was deconvoluted, rendered and analyzed by ImageJ analysis software with Deconvolutionlab 31 and Volumest 32 plugins. 2.5 Atomic Force Microscopy (AFM) The elastic moduli of hydrogels and cells cultured on PAAm substrates were probed with a commercial AFM (Nanowizard II, JPK Instruments, Germany). At least 10 square arrays of 8 × 8 indentations, covering (10 × 10) µm2 areas of cells or gels, were performed to quantify stiffness (Young’s modulus) parameters. We used a silicon nitride tip-less cantilever, V-shaped (MLCTO10, cantilever A, Bruker), with a spring constant of 0.07 N/m (calibrated by thermal noise method 33

). Polystyrene spheres (6 µm-diameter, Duke Scientific) were attached to the apex of AFM tips

using an optical adhesive (NOA61, Norland). To quantify the stiffness, the Hertz model gives the following relation (equation 1) between the indentation δ and the loading force F in the case of an infinitely hard sphere of radius R (AFM tip) touching a soft planar surface

F=

3 4 E rδ 2 3 1 −ν

(1)

where E is Young’s modulus and ν is the Poisson ratio, assumed equal to 0.457 34. For substrates and cells force curves (1 µm/s) were recorded with a maximum loading force of ~ 7 nN. ACS Paragon Plus Environment

6

Page 6 of 21

Page 7 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

Furthermore, to avoid contributions from the underlying substrates (glass in the case of hydrogels and PAAm substrates for the cells) to the extracted E, the maximum indentation depth was fixed to a few percent (~ 10%) of the sample thickness (~ 100 nm). For cells force curves were collected over cytoplasmic regions, avoiding nuclear and cell edge regions.

2.6 Synthesis and Characterization of RGD and RDG Peptides The RGD peptide and its scrambled version, RDG, were synthesized using standard solid-phase-9fluorenyl methoxy carbonyl (Fmoc) procedure. The syntheses were performed by Biotage Syro Wave peptide synthesizer and Wang resin was used as a solid-phase support. For both peptide, the C-terminal amino acids (10 eq) were manually attached to an equivalent amount of Wang resin using HOBt (10 eq), DIC (10 eq) and DMAP (0.4 eq). The Fmoc protecting group was removed using piperidine 40% (v/v) solution in DMF. Peptides were cleaved from resin using TFA/TIS/H2O (95/2.5/2.5) solution and precipitated in ice-cold diethyl ether. Un-protected peptides were dissolved in water and purified through multiple re-precipitation in cold acetone, then were filtered, lyophilized, and stored at -20 °C. Synthetized RGD and RDG peptides were characterized using LC-MS. The analyses were carried on injecting aqueous solutions of peptides into Agilent EclipsePlus C18 RRHD (1.8 µm, 2.1 × 50 mm) eluted with solvent solution of H2O and ACN containing 0.1% v/v formic acid. A linear gradient of 1-70 % of ACN over 10 min at a flow of 150 µl/min was used. We found the value of (M-H)+ 347 Da for both peptides. (Supplementary Fig. 4).

2.7 Synthesis of RGD-NPs and Scr-NPs RGD and RDG peptides were covalently attached to the carboxylic function on the surface of the modified polystyrene nanoparticles with a final formation of amide bond. Two round-bottom flasks containing 100 µl of nanoparticles were first diluted to a final volume of 3.0 ml using a conjugation buffer (CB) solution and then, the carboxylic functions were activated by adding 14.0 mg of EDAC (7.5 × 10-5 mol, 25 mM final concentration). The CB was prepared mixing 9 ml of 0.1 M phosphate solution and 1 ml of 0.05 M EDTA solution. After 10 min, 5 mg of RGD and RDG, corresponding to 1.4 × 10-5 mol, were dissolved in 0.2 ml of 0.01 M di TEA in CB to allow amino activation. The RGD and RDG solutions were then added to the two flasks separately and the reaction was allowed to proceed being stirred continually for 4h at RT. The two samples were then purified by dialysis/ ultrapure water for 36 h at + 4 °C to remove any unreacted peptides and then stored at + 4 °C. To

ACS Paragon Plus Environment

7

ACS Biomaterials Science & Engineering

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

avoid the formation of aggregates or the precipitation of nanoparticles, the pH of water was increased to 7.6 using 0.1 M of NaOH solution. To protect fluorescent probes contained inside the nanoparticles, the samples were not exposed to light during the entire time of the synthesis.

2.8 RGD-NPs and Scr-NPs Characterization Nuclear Magnetic Resonance (NMR) analysis was performed by Agilent 600 MHz (14 T) spectrometer equipped with a DD2 console and a Magic Angle Spinning (MAS) gHX nanoprobe. 1

H 1D spectra of RGD-NPs and Scr-NPs (0.2% suspensions in 40 µl of 90% H2O /10% D2O) were

recorded at 300 K using 2048 scans to obtain a good signal-to-noise ratio for the bonded peptides. A saturation PRESAT pulse sequence was used to reduce residual peaks of water at 4.65 ppm. MAS experiments were performed at a spinning rate of 2.9 kHz. Spectra were transformed and analyzed using VNMRJ 4 software. Chemical shift scale was referenced to the solvent residual peak signal. The quantification of bond peptides to nanoparticles was performed using the standard procedure of QuantiProTM BCA Assay Kit (Sigma-Aldrich). After the purification procedure through dialysis for 48 h at RT to remove unbounded peptides, the absorbance measurements of the samples in the 560 nm region were performed by EnSpire Multimode Plate Readers spectrofluorometer (Perkin Elmer). Data were interpolated with the calibration curves obtained by measuring the absorbance of free peptide solutions with known concentrations. The hydrodynamic diameters and ζ-potentials of the RGD-NPs and Scr-NPs were measured using Zetasizer Nano ZS (Malvern Instruments, Malvern, UK), equipped with a He–Ne laser (λ = 633 nm) and a backscatter detector at a fixed angle of 173°. The size and ζ-potential measurements were performed in triplicate using a nanoparticle suspension of 1% (w/v) at RT. Data were reported in Table I.

2.9 Quantification of NP Cellular Uptake and Multiple Particle Tracking (MPT) Cells were incubated with NPs at the final concentration of 3.6 × 109 NPs ml-1 for 1 h and images of internalized NPs were acquired in time-lapse for 100 s with a 1 s sampling time by using a wide field fluorescence microscope (Olympus Cell-R) and a 60× water immersion objective with NA = 1.35, plus 1.6 × magnification of internal microscope lens. MPT analysis was performed by an algorithm fully described in the notes

35, 36

. After obtaining the trajectories of the nanoparticles,

mean squared displacements (MSDs) were calculated from the trajectories of the centroids of the NPs and correlated with the diffusion using equation 2

ACS Paragon Plus Environment

8

Page 8 of 21

Page 9 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

∆r 2 = 4Dtα

(2)

where D represents the diffusion coefficient and the exponent α provides an indication to the mode of motion. In the following analyses, we considered only those regression fits with a R2 greater than 0.85. In brief, we would describe as Brownian or sub-diffusive the NPs which have an MSD curve that undergoes a power law with exponent α = 1 and α < 1, respectively, whereas as actively transported, the particles with an MSD which has a scaling exponent 1 < α < 2. NPs showing an MSD curve with a time exponent α near zero were classified as entrapped (random caged) and not computed. The number of total tracked and transported NPs was calculated for both NP functionalizations.

2.10 RGD-NP Cellular Uptake Kinetics In order to evaluate RGD-NP uptake kinetics, cells were seeded in 1 ml of cell culture medium on polyacrylamide substrates in glass-bottom 35 mm-diameter cell culture dishes. Cells were incubated with the NPs at the final concentration of 3.6 × 109 NPs ml-1 for 1, 3, 6 and 24 h. After incubation, cells were rinsed five times with PBS to remove non internalized nanoparticles, fixed with 4% paraformaldehyde for 20 min, and the cell membrane was stained with wheat germ agglutinin (WGA), rhodamine conjugate (Molecular Probes, Invitrogen, W11261). Z-sectioning images of cells incubated with fluorescent NPs were collected with a confocal laser scanning microscope (Leica TCS SP5 MP) equipped with an Argon and He-Ne laser lines at the wavelengths of 488 and 543 nm, respectively, using a 25× water immersion objective (NA 1.00). The pinhole size was 90 µm and the Z-stack consisted of 5 images covering the total cell volume. Images resolution was fixed at 1024 × 1024 squared pixels (0.32 µm/px). Image analysis of fluorescent NPs internalized by cells was performed by NIH software (Fiji ImageJ). Briefly, a maximum projection image for both colour channels was constructed from the 5 consecutive focal planes. Red images were used to extract individual cell outlines using ImageJ ROI manager tool and RGD-NP internalization at each time-point was evaluated in terms of integrated fluorescence intensity of RGD-NPs within individual cell boundaries. Data of internalization kinetics were fitted using a typical saturation rate equation 3:

r=

Imax t km + t

(3)

where Imax represents the maximum uptake and km the uptake half-life. ACS Paragon Plus Environment

9

ACS Biomaterials Science & Engineering

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2.11 Statistical Analysis Data are reported as mean ± standard error (SE), unless otherwise indicated. Statistical comparisons were performed with a Student’s unpaired test. P values < 0.05 denote statistical significance.

3. Results and discussions 3.1 bEnd.3 mechanobiology First we started by evaluating the mechanosensing activity of bEnd.3 cells. These cells are often investigated because they represent a very good model of endothelial barrier 37-42. In order to study the transport mechanisms of many molecules and nano-particles across this physiological barrier, bEnd3 were used to reproduce the blood brain barrier (BBB). We evaluated different biophysical parameters, such as cell spreading areas and FA lengths, cell elastic modulus and cell migration, in response to the stiffness modulation of the culture substrate. The latter was manufactured using PAAm hydrogels. Their Young’s moduli were modulated by adding different ratios of bisacrylimide crosslinker agent; mechanical AFM characterization verified that the elastic modulus (E) of the hydrogels increased with cross-linking density from ~ 3 to ~ 30 kPa (Supplementary Fig. 1). Soft and stiff gels presented average E that spanned over a range of values previously observed in brain in patho-physiological conditions 7, 9, 43-45. Each of the cell parameters evaluated on the PAAm substrates was dependent on the stiffness. We compared cell spreading areas and focal adhesion lengths using phalloidin and paxillin immunostaining, respectively. After 24 h, cell adhesion results showed that they were profoundly affected by substrate stiffness (Fig.1, Supplementary Fig. 2). In particular, cells cultured on 30 kPa substrates were well-spread compared to those on substrates with lower stiffnesses and the quantitative analysis showed the spreading area scaled by roughly a factor 3 passing from 3 to 30 kPa (200% increase of cell spreading) (Fig. 1a-c, Supplementary Fig. 2a-b). At the same time, the increase in PAAm stiffness promoted the formation of elongated paxillin-containing focal adhesions and prominent cytoskeletal stress fibers (Fig. 1d-e, Supplementary Fig. 2c-d). As for the spreading area, focal adhesion length increased by 20% on the 30 kPa (Fig. 1f, Supplementary Fig. 2d) and both parameters resulted in direct proportion to material stiffness substrate. Then, we used the AFM technique to investigate in which way changes in substrate stiffness affected the elastic modulus of individual bEnd.3 cells. As expected, a positive correlation between PAAm stiffness and cell Young’s modulus was found (Fig. 1l, Supplementary Fig. 2e). In addition, we have used the traction force microscopy (TFM) technique to measure the level of tension applied by bEnd.3 cells to the underlying substrates. Our data show that bEnd.3

ACS Paragon Plus Environment

10

Page 10 of 21

Page 11 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

cells are able to distinguish between 3 and 30 kPa substrates and, in particular, that 30 kPa substrates induced stronger traction stresses (Supplementary Fig. 10). Several works demonstrated that the intimate association between substrate and cell elastic moduli is mediated by cytoskeleton structure 46.. In particular, the focal adhesions work as tactile elements of cells and are the sites at which cytoskeleton forces originate through acto-myosin network. Our observations confirmed that stiff substrates regulated cell shape, focal adhesions elongation and cytoskeleton organization (Fig. 1). Specifically, the bEnd.3 cytoskeleton resulted in a more organized structure, as was also shown by actin immunostaining, which was justified by a larger adhesion area (cell spreading and FA size) and higher elastic moduli characterizing the cells cultured on stiffer substrates. The more structured actin structure of cells on 30 kPa PAAm was also evidenced from the more polarized shape and lower migration velocity on this kind of substrate (Supplementary Fig. 3). We used time-lapse imaging to record the random motion of sparsely cultured cells over 12 h. As had already been observed for other cell lines

46, 47

, the tension

generated by the actin-myosin machinery (Supplementary Fig. 10) induced a polarization of stress fibers, that orient in the direction of the applied force in response to substrate stiffness, and mean migration velocity decreased significantly with increasing substrate rigidity (Supplementary Fig. 3). In fact, cells needed more time (from a dynamic point of view) to assemble/disassemble focal adhesion and their internal cytoskeleton and to migrate along the stiffer substrates

48

. Conversely,

when cells resided on softer PAAm, their migration velocity increased thanks to the lower average time to polymerize/depolymerize their internal actin structures. Taken together, these observations indicate a good bEnd.3 mechanosensing and mechanoresponse to variations of substrate stiffness.

ACS Paragon Plus Environment

11

ACS Biomaterials Science & Engineering

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. a-b, bEnd.3 cells cultured on collagen-conjugated PAAm gels with elastic moduli of 3 and 30 kPa were stained for F-actin (red) and nuclear DNA (blue). Bar, 50 µm. c, Stiffnessdependent changes in cell spreading area. ***, P100 for both substrate stiffness levels. de, High-magnification images of cytoskeletal and adhesion structures. bEnd.3 cells were stained for F-actin (red), the focal adhesion protein paxillin (green) and DNA (blue) for cell nuclei. Bar, 20 µm. f, Stiffness-dependent changes in focal adhesion length. ***, P10000 for both substrate stiffness levels. g-h, 3D reconstruction of cell volume. Cell volume was not affected by PAAm stiffness (i). n>400 for both substrate stiffness levels. j-k, AFM deflection images show that stress fibers become more organized and aligned increasing PAAm stiffness. Such changes in cytoskeleton organization correlated with cell elastic modulus. In fact we found a positive correlation between substrate stiffness and cell Young’s modulus. (l). ***, P < 0.001; n>400 for both substrate stiffness levels. ACS Paragon Plus Environment

12

Page 12 of 21

Page 13 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

3.2 NP surface decoration to activate cytoskeleton recognition Once the bEnd.3 cells capability to sense and to adapt their cytoskeleton assembly to the elastic modulus of the substrate was ascertained, the nanoparticles able to activate an endocytic pathway directly involving the cytoskeleton were designed and fabricated. Consequently, yellow-green fluorescent, carboxyl-modified polystyrene NPs were surface-conjugated with RGD peptides. Such peptides have been used extensively in surface biodecoration in order to promote cell adhesion of non-adhesive substrates

49, 50

, being RGD a characteristic sequence of the most abundant

constituents of the ECM, like fibronectin, vitronectin and fibrinogen. RGD is the minimal sequence recognized by some components of the integrin family (αV, β1, β3) and forms a direct engagement ligand-receptor during the cell adhesion process 51, 52. Integrins are trans-membrane proteins which connect the cell cytoskeleton to the surrounding ECM

53-55

and are actively involved in the

mechanisms of mechanosensing 56, 57. Therefore, in this study, the functionalization of polystyrene NPs with the RGD peptide sequence was made to study the effects of substrate stiffness on the cell uptake process of NPs (when these offer the cells a ligand which specifically activates communication - via integrins receptors - with the cytoskeleton). The RDG scrambled version was also used as a negative control for comparison. A detailed chemico-physical characterization of peptide conjugation, NP size and charge can be found in the Supplementary Information section.

3.3 Relation between NP cellular uptake, intracellular behavior and cytoskeleton assembly The functionalized NPs were incubated with bEnd.3 cells previously seeded on 3 and 30 kPa PAAm substrates at 37 °C for different time intervals. The effective NPs uptake was confirmed by the xz and yz cross-section of confocal image stacks acquisition of bEnd.3 cells cultured on 3 and 30 kPa substrates incubated (Supplementary Fig. 6). We did not observe NP accumulation on cell membrane. As reported in Fig. 2c, after 1 h of incubation, both RGD-NPs and Scr-NPs were internalized by the cells and the uptake process was dependent on substrate stiffness. These data were also confirmed by an alternative analysis which used different methodology and is reported in Supplementary Fig. 7. Also indicated by the uptake kinetic experiments (Fig. 2g), tracking analysis showed that the uptake of RGD-NPs after 1h was significantly higher in cells cultured on stiff PAAm substrates than on soft ones. However, Scr-NPs uptake results show a significant reduction when compared to RGD-NPs on both surfaces; we also found a significant positive correlation between uptake and substrate stiffness in this case (Fig. 2c). The substrate stiffness played a key role not only in the internalization of NPs, but also in their intracellular transport. In fact, the number of RGD-NPs, showing transported trajectories (NPs moving with not null average velocity ACS Paragon Plus Environment

13

ACS Biomaterials Science & Engineering

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 21

along a linear, and not random, pathway and characterized by an exponent α strictly greater than 1), resulted higher on 30 kPa (60% of total number of endocyted RGD-NPs) than on 3 kPa substrates (41% of total number of endocyted RGD-NPs, Fig. 2f). The RGD sequence increased the affinity of particles to the cytoskeleton components, in particular on stiffer hydrogels where the cell cytoskeleton appeared more structured and the cell Young’s modulus was higher (Fig. 1j-l). Actively transported particles underwent association from specific-binding to the cytoskeleton and/or motor protein-mediated transport along the cytoskeleton. Actively transported Scr-NPs were also more numerous in cells cultured on 30 kPa (61% of total number of endocyted Scr-NPs) than on 3 kPa substrates (37% of total number of endocyted Scr-NPs), presumably because a nonspecific cytoskeletal protein binding participated to their cytoplasmic transport (Fig. 2f). In fact, when in contact with cells, nanoparticles are covered by an absorbed layer (or multi-layers) of serum proteins present in the cell medium, the so-called protein corona

58, 59

. In the case of RGD-

NPs and in comparison to the scrambled peptide, cell uptake resulted enhanced on soft and stiff PAAm substrates, although we did not find a different level of NP accumulation on the cell membrane (with respect to pristine and Scr-NPs, data not shown). This leads to the indication that, when cells may find the RGD peptide, among the component of the protein corona which covers NPs surfaces, the internalization process gets a boost. As shown in Fig. 2g, this result was also confirmed at other intervals of time. In particular, the uptake values registered 24 h after contact demonstrated how the cells internalized a different amount of NPs depending on the cytoskeleton assembly. Indeed 24 hours may be considered as a plateau time for NP uptake. However, the kinetics of internalization (Fig. 2g) shows how cells, when sensing a stiffer substrate and are characterized by a greater structuration of the cytoskeleton and a higher stiffness, internalized a greater amount of NPs at every time investigated. Recently, Huang et al. have demonstrated the influence of substrate stiffness on NP uptake. Interestingly, the authors showed how this mechanism is regulated by membrane tension that is directly correlated to the spreading cell area. The uptake of NPs actually lowered when normalized with respect to the cell area on stiffer substrates (5.71 kPa) where cells presented higher membrane tension. In this work we confirmed the same data trend (Fig. 2h) normalizing the uptake quantification to the cell area. Surprisingly, we found a completely opposite trend when the normalization was performed with respect to the cell volume (Fig. 2j). Then, depending on the way data are presented, NP uptake resulted to increase (cell volume normalization) or to decrease (cell area normalization) with substrate stiffness. We reported both the results, although we are inclined to find the correlation between the amounts of internalized NPs and the cell volume more logical for two main reasons: i) the uptake quantification is generally expressed as the amount of internalized NPs per cell volume; ii) while the cell spreading area is

ACS Paragon Plus Environment

14

Page 15 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

strongly influenced by the substrates stiffness (increasing linearly with E, Fig. 1c), the cell volume remains constant (Fig. 1i), behaving as an invariant in the processes of cell mechanosensing and mechanotransduction, and representing an optimal parameter for uptake normalization. We also evaluated the auto-diffusion coefficient of the lipids constituting the membrane layer of bEnd.3 cells to obtain a qualitative indication of the membrane tension values. As reported in Supplementary Fig. 8, no statistical difference was found for cells residing on 3 and 30 kPa substrates. Such results indicated that, for these types of cells, membrane tension is not a parameter able to modulate NP uptake. The modulation with substrate Young’s modulus we found is therefore ascribable to the different mechanical state of the cells and to the different organization of the cell cytoskeleton. Furthermore, considering that cells increased their spreading area maintaining the volume constant, it is easy to figure a more compact cytoskeleton and explain the enhancement of cell elastic modulus. Finally, to confirm the hypothesis that the cytoskeleton assembly is one of the major regulators of the internalization mechanism, we further explored how pharmacological inhibitors of actin cytoskeleton dynamics and myosin activity affected RGD-NP uptake. To this aim, bEnd.3 cells seeded on PAAm substrates were pre-treated with Cytochalasin-D (30 µM) to perturb the actin cytoskeleton and ML7 (50 µM) to inhibit myosin light chain kinease for 30 min before cells were incubated with RGD-NPs

60, 61

. Cytochalasin-D treatment induced a sensitive reduction of the cell

spreading which continued to be related to the substrate stiffness (Supplementary Fig. 9). However, the hard disruption of the cytoskeleton network severely inhibited RGD-NP entrance via endocytosis, especially in cells cultured on 30 kPa substrate. In particular, RGD-NP uptake was reduced by 64% on the 3 kPa substrate (P < 0.001) and by 78% on the 30 kPa substrate (P < 0.001). Interestingly, after cytoskeleton disassembly the amount of endocytosed NPs were uncorrelated with PAAm gel stiffness (Supplementary Fig. 9). ML7 induced a substantial reduction of cell spreading areas, which became uncorrelated with the substrate rigidities, and decreased RGD-NP uptake by 92% on the 3 kPa substrate (P < 0.001) and by 94% on the 30 kPa substrate (P < 0.001) (Supplementary Fig. 9). The effect of myosin inhibitor was naturally strongly dependent on its concentration. This effect was profoundly dependent on ML7 concentration; in fact, at a low concentration of ML7 (10 µM), when the activity of myosin was only partially inhibited, the spreading areas and the RGD-NPS were only slightly reduced (data not shown). These data confirm the pivotal role of the cytoskeleton network and, in particular, of actin and myosin in NP uptake.

ACS Paragon Plus Environment

15

ACS Biomaterials Science & Engineering

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2. bEnd3 cells cultured on 3 (a-d) and 30 (b-e) kPa substrates and incubated for 1 h with RGD-NPs (a-b) and Scr-NPs (d-e). NP uptake (c) and intracellular behavior - trajectories of endocytosed NPs analyzed by MPT - (f) resulted to be strongly affected by substrate stiffness. ***, P