Hybrid magnetic scaffolds: The role of scaffolds charge on the cell proliferation and Ca2+ ions permeation

Hybrid magnetic scaffolds: The role of scaffolds charge on the cell proliferation and Ca2+ ions permeation

Accepted Manuscript Title: Hybrid magnetic scaffolds: the role of scaffolds charge on the cell proliferation and Ca2+ ions permeation Authors: Pollyan...

1MB Sizes 0 Downloads 10 Views

Accepted Manuscript Title: Hybrid magnetic scaffolds: the role of scaffolds charge on the cell proliferation and Ca2+ ions permeation Authors: Pollyana S. Castro, Mauro Bertotti, Alliny F. Naves, Luiz Henrique Catalani, Daniel R. Cornejo, Georgia B. Delmilio, Denise F.S. Petri PII: DOI: Reference:

S0927-7765(17)30304-1 http://dx.doi.org/doi:10.1016/j.colsurfb.2017.05.046 COLSUB 8575

To appear in:

Colloids and Surfaces B: Biointerfaces

Received date: Revised date: Accepted date:

2-11-2016 9-5-2017 16-5-2017

Please cite this article as: Pollyana S.Castro, Mauro Bertotti, Alliny F.Naves, Luiz Henrique Catalani, Daniel R.Cornejo, Georgia B.Delmilio, Denise F.S.Petri, Hybrid magnetic scaffolds: the role of scaffolds charge on the cell proliferation and Ca2+ ions permeation, Colloids and Surfaces B: Biointerfaceshttp://dx.doi.org/10.1016/j.colsurfb.2017.05.046 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Hybrid magnetic scaffolds: the role of scaffolds charge on the cell proliferation and Ca2+ ions permeation

Pollyana S. Castro1, Mauro Bertotti1, Alliny F. Naves1, Luiz Henrique Catalani1, Daniel R. Cornejo2, Georgia B. Delmilio1 and Denise F. S. Petri1.*

1, Department of Fundamental Chemistry, Institute of Chemistry, University of São Paulo, Av. Prof. Lineu Prestes, 748, 05508-000, São Paulo, SP, Brazil

2. Institute of Physics, University of São Paulo, São Paulo, SP, Brazil

*corresponding author E-mail: [email protected], Tel.: 0055-11-30919154, Fax: 055-11-38155579

1

Graphical Abstract

Highlights 

Negatively charged scaffolds and magnetic particles favored cell proliferation.



The permeation of Ca2+ ions through the magnetic charged scaffolds was the fastest.



The magnetic particles stimulated mechanical vibrations in the charged scaffolds.



Mechanical vibrations favored the Ca2+ ions permeation through the scaffolds.



Mechanical vibrations might open the cell membranes mechano-sensitive ion channels.

2

Hybrid magnetic scaffolds: the role of scaffolds charge on the cell proliferation and Ca2+ ions permeation

Pollyana S. Castro1, Mauro Bertotti1, Alliny F. Naves1, Luiz Henrique Catalani1, Daniel R. Cornejo2, Georgia B. Delmilio1 and Denise F. S. Petri1.*

1, Department of Fundamental Chemistry, Institute of Chemistry, University of São Paulo, Av. Prof. Lineu Prestes, 748, 05508-000, São Paulo, SP, Brazil

2. Institute of Physics, University of São Paulo, São Paulo, SP, Brazil

*corresponding author E-mail: [email protected], Tel.: 0055-11-30919154, Fax: 055-11-38155579

3

Abstract Magnetic scaffolds with different charge density were prepared using magnetic nanoparticles (MNP) and xanthan gum (XG), a negatively charged polysaccharide, or hydroxypropyl methylcellulose (HPMC), an uncharged cellulose ether. XG chains were crosslinked with citric acid (cit), a triprotic acid, whereas HPMC chains were crosslinked either with cit or with oxalic acid (oxa), a diprotic acid. The scaffolds XGcit, HPMC-cit and HPMC-oxa were characterized by scanning electron microscopy (SEM), inductively coupled plasma atomic emission spectroscopy (ICP-AES), superconducting quantum interference device (SQUID) magnetometry, contact angle and zeta-potential measurements. In addition, the flux of Ca2+ ions through the scaffolds was monitored by using a potentiometric microsensor. The adhesion and proliferation of murine fibroblasts (NIH/3T3) on XG-cit, XG-cit-MNP, HPMC-cit, HPMC-cit-MNP, HPMC-oxa and HPMC-oxa-MNP were evaluated by MTT assay. The magnetic scaffolds presented low coercivity (< 25 Oe). The surface energy values determined for all scaffolds were similar, ranging from 43 mJ m-2 to 46 mJ m-2. However, the polar component decreased after MNP incorporation and the dispersive component of surface energy increased in average 1 mJ m-2 after MNP incorporation. The permeation of Ca2+ ions through XG-cit-MNP was significantly higher in comparison with that on XG-cit and HPMC-cit scaffolds, but through HPMC-cit-MNP, HPMC-oxa and HPMC-oxaMNP scaffolds it was negligible within the timescale of the experiment. The adhesion and proliferation of fibroblasts on the scaffolds followed the trend: XG-cit-MNP > XGcit > HPMC-cit, HPMC-cit-MNP, HPMC-oxa, HPMC-oxa-MNP. A model was proposed to explain the cell behavior stimulated by the scaffold charge, MNP and Ca2+ ions permeation.

4

Keywords magnetic scaffolds; xanthan; HPMC; Ca2+ ions permeation; cell adhesion; cell proliferation

Introduction Polymeric scaffolds for cell proliferation and differentiation are biocompatible and biodegradable materials with high chemical and mechanical stability; they are applied either in vitro to create implantable tissues or in vivo to induce tissue regeneration within the organism [1,2]. Biodegradable polymers such as poly(lactic acid) (PLA), poly(glycolic acid) (PGA), poly(lactic-co-glycolic acid) (PLGA) [3] and polycaprolactone (PCL) [4] have been often used as implantable scaffolds. Pioneer studies reported that implantation of PGA [5] or PLGA-polylysine blends [6] scaffolds helped repairing injured brain. More recently, magnetic nanoparticles (MNP) have been added to polymeric scaffolds, creating hybrid magnetic scaffolds. The development of nontoxic biocompatible magnetic scaffolds should follow the acceptable limits for MNP size [7] and concentration [8,9], which might be cell type dependent [10]. In comparison to pure PCL, hybrid PCL/MNP scaffolds revealed a higher mineral induction, improved initial cell adhesion and osteoblasts proliferation [11]. Scaffolds made of hydroxyapatite, collagen and MNP [12] or Fe doped hydroxyapatite/collagen [13] stimulated in vitro adhesion and proliferation of human osteoblast-like. The combination of MNP and chitosan/poly(vinyl alcohol) [14] or silk fibroin protein [15] yielded successful scaffolds for adhesion and proliferation of osteoblasts. Hybrid

5

scaffolds of gelatin–siloxane/MNP presented better environment for spreading, proliferation and osteogenic differentiation of mesenchymal stem cells than the MNP free scaffolds [16]. Alginate/MNP hybrid scaffolds stimulated the proliferation and organization of aortic endothelial cells [8].

In comparison to neat xanthan (XG)

scaffolds, hybrid scaffolds of xanthan (XG) and MNP improved in vitro proliferation of fibroblasts [17] and stimulated in vitro neuronal differentiation of embryonic stem cells [9]. A relevant observation is that cell proliferation onto hybrid magnetic charged scaffolds such as gelatin-siloxane [16], alginate [8], collage-hyaluronic acid [18] and xanthan [9, 17] tends to be more pronounced than on the neat charged scaffolds. There are some hypotheses to explain the positive effects observed for the hybrid magnetic scaffolds in comparison to the corresponding neat scaffolds: (i) MNP might have ability to diminish intracellular H2O2 through intrinsic peroxidase-like activity [19], (ii) MNP can accelerate cell cycle progression, which may be mediated by the free iron (Fe) released from lysosomal degradation [20], (iii) MNP stimulate Ca2+ ions influx [17, 21] and (iv) small changes in the orientation and intensity of magnetic field would lead to changes in the MNP alignment, which would cause mechanical stress on the surrounding cellular structures opening the ion channels [22,23]. Considering that most of the cited scaffolds [8,9,12-18] are charged surfaces, another possible hypothesis is that the interactions between electrostatic charges and MNP modulate the cell mechanical behavior and the Ca2+ ions transport, which permeate through the scaffold. In order to gain insight about the role played by electrostatic charges and MNP on the cell behavior and Ca2+ ions diffusion through the scaffolds, magnetic scaffolds with different charge density were prepared using MNP and xanthan gum XG, a negatively charged polysaccharide, or hydroxypropyl methylcellulose (HPMC), an 6

uncharged polymer. XG chains were crosslinked with citric acid (cit), a triprotic acid, whereas HPMC chains were crosslinked either with cit or with oxalic acid (oxa), a diprotic acid. The esterification reaction between XG chains, which are composed by Dglucosyl, D-mannosyl, and D-glucuronyl acid residues in a 2:2:1 molar ratio and variable proportions of O-acetyl and pyruvyl residues, and side-chains of a trisaccharide composed of mannose (β-1,4) glucuronic acid (β-1,2) mannose, attached to alternate glucose residues in the backbone by α-1,3 linkages (Figure 1), and citric acid (Figure 1) did not necessarily consume all carboxylic acid groups, part of carboxylic acid groups, either from XG or from citric acid, might remain unreacted [24]. The unreacted carboxylic group is deprotonated at pH > 4.5, increasing the charge density in the scaffolds. Contrarily to XG, the HPMC chains carry no carboxylic acid group (Figure 1); HPMC has only hydroxyl groups involved in the esterification reaction with citric acid. Considering that citric acid is triprotic, some carboxylic groups remain unreacted, giving rise to negative charges at pH > 4.5 [25]. In the case of HPMC-oxa, the carboxylic acid groups are completely consumed in the esterification between HPMC chains and oxalic acid (Figure 1), a diprotic acid [26]. Thus, the order of the negative charge density in the scaffolds is XG-cit > HPMC-cit > HPMC-oxa. The scaffolds XG-cit, HPMC-cit and HPMC-oxa were characterized by scanning electron microscopy (SEM), inductively coupled plasma atomic emission spectroscopy (ICP-AES) analyses, superconducting quantum interference device (SQUID) magnetometry, contact angle and zeta-potential measurements. In addition, the permeation of Ca2+ ions through the scaffolds was evaluated by using a potentiometric microsensor (home-made) positioned at 300 µm above the surface. The adhesion and proliferation of murine fibroblasts (3T3 L1) on XG-cit, XG-cit-MNP, 7

HPMC-cit, HPMC-cit-MNP, HPMC-oxa and HPMC-oxa-MNP were evaluated by MTT assay, using plastic dishes as control experiment.

Experimental Section Scaffolds preparation and characterization Xanthan gum, XG, (Mv ~ 1.5 106 g mol-1, degree of pyruvyl = 0.38, degree of acetyl = 0.41) and hydroxypropyl methylcellulose, HPMC J5MS, (USP 1828, Mw 3 x 105 g mol-1, degree of substitution of methyl groups = 1.5, molar substitution of hydroxypropyl groups = 0.75) were kindly provided by CP Kelco USA and Dow Chemical Brazil Co., respectively. Citric acid (Labsynth, Brazil, 192.12 g mol-1) and oxalic acid (Labsynth, Brazil, 90.03 g mol-1) were used without further purification. Aqueous solutions of XG or HPMC prepared at 9 g L-1 in the presence of crosslinkers (citric or oxalic acid) at 0.45 g L-1 were cast in polystyrene dishes (85 mm diameter) and remained in oven at 45 oC overnight for water evaporation. The resulting films were heated at 165 C for 7 minutes to enable esterification reaction among the polymer chains and crosslinkers [24-26]. XG and HPMC crosslinked with citric acid were coded as XG-cit and HPMC-cit, respectively, and HPMC cross-linked with oxalic acid was coded as HPMC-oxa. All samples were extensively rinsed with MilliQ water (resistivity of ca. 18.2 M cm) to remove unreacted chemicals. Then, the samples were immersed in water for 20 min and the medium conductivity was measured; this procedure was repeated until the resistivity of ca. 18 M cm was achieved. After that, the scaffolds were freeze-dried and characterized by scanning electron microscopy (SEM) in a Jeol microscope FEG7401F equipped with a Field-Emission Gun. Samples were prepared by cryo-

8

fracturing freeze-dried samples. The cryo-fracture surfaces were analyzed after gold coating (~ 2 nm).

The synthesis of magnetite was performed as described elsewhere [17]. Briefly, FeCl3.6H2O and FeCl2.4H2O (both from Labsynth, Brazil) were used to prepare 0.1 mol L-1 and 0.05 mol L-1 solutions, respectively, which were then mixed with NH4OH (25 % V/V, Labsynth, Brazil) under vigorous stirring until the solution achieved pH 9. N2 was bubbled directly into the media prior to reaction for removal of oxygen. Then, the system was placed in a water bath at (24 ± 1) oC, in which the sonotrode MS7 with acoustic power density of 130 W cm-2 coupled to the ultrasonic processor (Hielscher UP100H, Germany) was immersed. The sonotrode was kept outside of the reaction flask in order to avoid contamination by Ti particles stemming from the device. The sonotrode operated during 10 min. The temperature inside and outside of the reaction flask remained at (24 ± 1) oC. MNP were separated by centrifugation at 1200 g during 10 min. The supernatant was removed and the particles were re-dispersed in MilliQ water followed by centrifugation. This rinsing process was repeated three times in order to remove the excess of reactants. One should notice that no stabilizer was added to the MNP dispersions. The concentration of magnetite in the dispersion was determined by gravimetric measurements as (50 ± 2) g L-1. The MNP presented diameters of 17 ± 3 nm, isoelectric point of 6.5 ± 0.1 and nearly superparamagnetic behavior at room temperature, with coercivity less than 20 Oe in all samples [17]. XG-cit, HPMC-cit and HPMC-oxa membranes were immersed in the MNP dispersions at pH 7, which was adjusted by addition of HCl 0.1 mol L-1, and (24 ± 1) oC during 1 min. After that, they were removed and rinsed in MilliQ water for 20 s. This

9

process was repeated three times more in order to remove particles weakly attached to polymeric

matrix

(movie

at

www.youtube.com/watch?v=PhTn2M_NRF8&index=2&list=UUqVKEBUzLfSAoMIz F8r9VTA). XG-cit, HPMC-cit and HPMC-oxa membranes impregnated with magnetic nanoparticles were coded as XG-cit-MNP, HPMC-cit-MNP and HPMC-oxa-MNP. ICPAES analyses performed with a Spectro Smart Analyzer Vision equipment (SPECTRO Analytical Instruments GmbH, Germany) yielded the amount of iron in the XG-citMNP, HPMC-cit-MNP and HPMC-oxa-MNP samples. XG-cit-MNP, HPMC-cit-MNP and HPMC-oxa-MNP samples were gently dried with paper tissues and freeze-dried for analyses with a superconducting quantum interference device (SQUID) magnetometer (model MPMS, Quantum Design, USA, details in the Supplementary Information SI1). The surface energy (γS) of all scaffolds was determined by means of contact angles measured at 24 ± 1 °C, using sessile drops (10 µL) of ethylene glycol (γL = 48.8 x 10-3 J m-2; γpL = 19.8 x 10-3 J m-2 and γdL = 29 x 10-3 J m-2) and CH2I2 (γL = 50.8 x 103

J m-2; γdL = 50.8 x 10-3J m-2) [27], using a SEO (Surface Electro Optics, South Korea)

equipment. The polar (γpS) and dispersive (γdS) components of the surface energy were determined by Owens–Wendt's equation [28], also known as geometric mean equation. At least three scaffolds were analyzed. The zeta potential of scaffolds was determined at pH 5.7 in 1mmol L-1 KCl solution, at 299.8 ± 0.2 K, using a clamping cell (20 mm x 10 mm) mounted in the SurPASS 3 equipment (Anton Paar, Austria), using polypropylene film as reference. Cell culture proliferation studies Dulbecco’s modified Eagle’s medium (DMEM, 12100-46), fetal bovine serum (FBS, 16140-071), a 2.5% trypsin-EDTA solution (15090-046, diluted 10 fold for cell culture assays), penicillin and streptomycin (10000 units mL-1 and 10 μg mL-1, 10

respectively, 15140-122), amphotericin B (250 μg mL-1, 041-95780 D) were purchased from Gibco® Life Technologies. Phosphate buffered saline pH 7.4 (PBS, P3812) and 3(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, M5655) were supplied by Sigma-Aldrich. NIH/3T3 cells (American Type Culture Collection, ATCC® CRL-1658™) at passage 10 (P10) were used for cell culture proliferation assays. Culture medium (DMEM supplemented with 0.5% penicillin and 0.5% streptomycin) was used to swell the samples and complete culture medium (CCM, DMEM supplemented with 10% FBS, 0.5% penicillin, and 0.5% streptomycin) was used to culture the cell in the proliferation assays. Cell culture plates and flasks (Costar) were used as received. Ultrapure Milli-Q water (specific resistivity ~ 18.2 M cm at 25 °C) was used for dialysis and for preparing aqueous solutions. For in vitro experiments, dry scaffolds were cut into round shape (diameter 12 mm) and sterilized under UV light in the laminar flow for 15 min each side. The culture medium was removed; the cells were placed into 24-well plates and incubated with 1 mL culture medium for 24 h before plating the cells. NIH/3T3 fibroblasts were incubated with CCM in 75 cm2 tissue culture flasks under normal culture conditions (37°C, 5% CO2, controlled humidity) and sub-cultured up to P12. The cells were plated at 5 x 104 cells per well and placed into the incubator. After 3 h, 2 mL CCM were added to each well and the samples were incubated for up 21 days. CCM was replenished every 2 days, and MTT assays were performed to evaluate cell proliferation. For the MTT assays, the samples were gently removed from the well plates and placed in new 24-well plates. The cells adhered to the samples were incubated with a 1 mg mL-1 MTT solution for 3 h. Then, the culture medium was removed, and the formazan crystals were dissolved in 1 mL isopropanol. The absorbance was measured at 570 nm (iTecan, NanoQuant), and the cell proliferation was determined using a calibration curve. In vitro 11

cell proliferation assays were performed in triplicate. Cell proliferation assays on commercial plastic dishes without any scaffold were performed as control experiments.

Ion-selective liquid-membrane microelectrode preparation and measurements protocol [29, 30] Potassium chloride (KCl), calcium chloride (CaCl2.2H2O), ethanol (C2H5OH), N-(trimethylsilyl) dimethylamine (TMSDMA, 226289 Sigma), and calcium ionophore I (cocktail A - ETH1001) were obtained from Sigma-Aldrich. Sulfuric acid (H2SO4) was purchased from Fischer Chemistry. Other chemicals used were reagent grade. All solutions were freshly prepared using Milli-Q water (Millipore Corp.) with resistivity of ca. 18.2 MΩ cm at 25°C. The ion-selective liquid-membrane microelectrode (potentiometric sensor) was fabricated using borosilicate capillaries (O.D. 1.5 mm, I.D. 0.86 mm, L. 150 mm, Sutter Instrument Co.). Prior to fabrication, the capillaries were cleaned by soaking them in a 1 mol L-1 H2SO4 solution for 24 h followed by a thorough washing with water and ethanol to completely remove traces of acid solution. The capillaries were then dried in an oven for 4 h at 180 ºC and stored in air tight containers until used. The cleaned capillaries were pulled into sharp pointed pipettes by using a laser puller (P-2000, Sutter Instruments) with a custom-developed program (Line 1: Heat 460, Fil 3 Vel 30, Del 150, Pul 25, Line 2: Heat 440, Fil 5 Vel 30, Del 135, Pul 40). A SEM (Jeol microscope FEG7401F) image of a typical pulled capillary with around 3 m of inner diameter was provided as Supplementary Information SI2.

12

In order to construct the potentiometric sensor, the pulled barrel was silanized by passing vapors of TMSDMA through the barrel for 5 minutes. This was accomplished by heating a small quantity of the silane compound at 100 °C and then collecting the vapor in a glass syringe (about 4 mL), followed by the injection of the vapor inside the pulled capillary by using a very long and flexible needle (MicrofilTM 37G). This procedure was repeated 3 times for each tip. The probes were then heated at 120 °C in an oven for 1 h to complete the silanization process. The silanized barrel was firstly back-filled with a calcium ionophore, followed by a 100 mmol L-1 CaCl2 solution (inner filling solution). Then, a Ag|AgCl quasi reference counter electrode (QRCE) was inserted into the CaCl2 solution to complete the electric circuit for the potentiometric measurements against a Ag|AgCl|KCl(sat) reference electrode. The optical image of a typical ion-selective liquid-membrane microelectrode can be shown in Supplementary Information SI2. The interface between the ionophore membrane and the inner filling solution (100 mmol L-1 CaCl2) can be clearly seen. After the fabrication, the potentiometric sensor was stored in a vertical position in a solution containing 50 mmol L-1 CaCl2 for 1 day prior to use. For the potentiometric investigation of Ca2+ ions transport through the scaffolds, a home-made cell of poly(methyl methacrylate), schematically represented in Supplementary Information SI3, was used. The scaffolds were mounted between the donor and receptor chambers, which were separated each other by using O-rings. The receptor chamber initially contained a 100 mmol L-1 KCl solution. The donor chamber was simply a reservoir containing a 1 mol L-1 CaCl2 solution connected to a plastic tube (inlet solution) with 10 cm height to control hydrostatically the pressure across the scaffold. This configuration allowed the permeation of Ca2+ ions from the donor to the receptor chamber only through the scaffold structure (outlet solution closed). The set up 13

included a reference electrode, which was placed in contact with the solution in the receptor chamber and was not represented in Supplementary Information SI3. All electrochemical measurements were performed inside a Faraday cage. The permeation experiments were carried out with a SECM instrument (Sensolytics GmbH Bochum, Germany) using a conventional two electrodes system with Nova 1.10 (Autolab). The sensor/scaffold separation was adjusted by using a video camera (The Imaging Source ® - Germany) to assist the approaching of the sensor to the scaffold surface until the tip almost touched the surface. Then, using step motors the tip was withdrawn from the surface to the distance of 300 m. All electrochemical measurements were then performed at this sensor/scaffold separation. All potentiometric sensors used herein were calibrated before and after each experiment, in a 100 mmol L-1 KCl solution (pH 6.8) containing CaCl2 in the concentration range between 0.1 μmol L-1 and 100 mmol L-1.

Statistical analysis Comparisons between experimental data were performed using one-way analysis of variance (ANOVA) with post hoc test in Excel 2013® for Windows® (Microsoft Office Home and Student®, 2013). Criteria for statistical significance were set at p < 0.05 (*) or p < 0.002(**).

RESULTS Characterization of Scaffolds XG-cit, HPMC-cit and HPMC-oxa scaffolds were 80 ± 10 µm thick and resisted over one year immersed in media at the pH range from 1 to 9 without dissolving or losing the form. SEM images of cryofracture surfaces of XG-cit, HPMC-cit and HPMC14

oxa scaffolds shown in Figures 2a, 2b and 2c, respectively, indicated that all scaffolds present micrometric pores. After MNP incorporation in the scaffolds, it was a hard task to visualize the MNP by means of SEM because the samples moved during the analyses, resulting in bad quality images. The movement was attributed to the attraction between magnetic scaffolds and electromagnetic lens. However, the MNP are expected to adsorb on the surface rather than to diffuse to the scaffold interior. As reported in a previous work [17], high-resolution SEM and STEM images evidenced the presence of very small particles (size of a few tens of nanometers) individually deposited along with particles and aggregates on many regions of the scaffold surface (Supplementary Information SI4). The amount of iron in XG-cit-MNP, HPMC-cit-MNP and HPMC-oxa scaffolds was found to be 0.2 wt%, 0.3 wt% and 0.2 wt%, respectively, hence they all lie below 1 wt%. Cytotoxicity was observed for iron content higher than 1 wt%, because the MNP start diffusing into the cells [8,9]. The magnetic properties of XG-cit-MNP, HPMC-citMNP and HPMC-oxa-MNP are presented in Figures 2a, 2b and 2c, respectively. They indicated very low coercivity (< 25 Oe) at 300 K, characteristic of superparamagnetic behavior, and coercivity in the range of 250 to 300 Oe at 5 K. The polar (γpS, blue columns) and dispersive (γdS, red columns) components of the surface energy and total surface energy (γS, grey columns) were determined for all scaffolds, as presented in Figure 3. The γS values for all scaffolds were similar, ranging from 43 mJ m-2 to 46 mJ m-2. Small but statistically significant differences were observed for the γpS, which decreased after MNP incorporation. On the other hand, the γdS increased in average 1 mJ m-2 after MNP incorporation. The zeta potential values determined at pH 5.7 for XG-cit and XG-cit-MNP scaffolds were found to be –(3.9 ± 0.5) mV and –(2.0 ± 0.3) mV, respectively, and followed the tendency observed in the decrease of γpS and 15

increase of γdS after MNP incorporation. These findings indicated that the presence of MNP (0.2 wt%) screened the negative charges on the xanthan scaffold surface. HPMCcit, HPMC-cit-MNP, HPMC-oxa and HPMC-oxa-MNP presented null zeta potential.

The adhesion, viability, and proliferation of murine fibroblasts (NIH/3T3) on XG-cit, XG-cit-MNP, HPMC-cit, HPMC-cit-MNP, HPMC-oxa and HPMC-oxa-MNP were evaluated by MTT assay, using plastic dishes as control experiment. Cells were seeded with the same density on all scaffolds. After 1 day of culture, none of the scaffolds presented cytotoxicity because the content of MNP in the scaffolds was low enough to avoid cytotoxic effects to the cells [8,9]. Figure 4a shows that cell adhesion (1 day) and proliferation after 7 days on XG-cit was similar to the control, but less pronounced than on XG-cit-MNP. After 21 days, fibroblasts proliferation on XG-citMNP was significantly (p < 0.05) larger than XG-cit or control, corroborating with previous report [17]. On the other hand, cells proliferated less on HPMC-cit and HPMC-cit-MNP than on the plastic dishes (control), even after 10 days, as shown in Figure 4b. On HPMC-oxa and HPMC-oxa-MNP the cells practically did not proliferate, in comparison to the plastic dish (Figure 4c). These findings evidenced that cell behavior is not only stimulated by the presence of MNP, but the scaffold charges also play an important role. The cells were seeded on the scaffolds swollen by the feeding culture medium, which contains Ca2+ ions (~ 0.1 mol L-1). Ca2+ ions are secondary messengers in many transduction pathways related to cell proliferation. Thus one can hypothesize that the scaffolds that favor the diffusion of Ca2+ ions would stimulate their proliferation. In order to test this hypothesis, the local concentration of Ca2+ ions at 300 µm above the 16

scaffolds was monitored with the potentiometric sensor by using the proposed setup (Supplementary Information SI3). Results shown in Figure 5a demonstrated a continuous increase in the local concentration of Ca2+ when the experiment was carried out with the XG-cit-MNP scaffold, hence it seems that the cation is easily transported through this membrane. On the other hand, a much less concentration increase was noticed when XG-cit and HPMC-cit scaffolds were employed, and no Ca2+ response was observed for HPMC-cit-MNP, HPMC-oxa and HPMC-oxa-MNP scaffolds within the measuring timescale. The slow permeation of Ca2+ ions through the HPMC based scaffolds also caused pressure against them, making the scaffolds to bend during the preparation of the permeation experiments. In order to examine the stability of the potentiometric sensor response, calibration curves were obtained before and after each permeation experiment (Figure 5b). CaCl2 solutions in a range of pCa2+ from 1 to 7 in 100 mmol L-1 KCl solution at pH 6.8 were used. A Nernstian response was obtained by correlating the open circuit potential and pCa2+, values being 31 mV pCa2+ unit-1 and 29 mV pCa2+ unit-1, before and after the permeation experiment, respectively. The stability of potentiometric response confirms the reliability of the results on the flux of Ca2+ ions across the scaffolds. The trend observed for the permeation of Ca2+ ions through the scaffolds in Figure 5a correlated well with that observed for cell proliferation in Figure 4 and raised the question why the presence of MNP in charged scaffolds is important for Ca2+ ions influx and, consequently, for cell proliferation? The answer to this question is presented in the Discussion section.

17

Discussion The physicochemical characteristics of scaffolds, such as surface energy, stiffness, porosity and surface anisotropy might favor interactions with key proteins (integrins, calmodulin among many others) responsible for cell attachment and proliferation. The surface energy values of fibronectin, collagen and polystyrene culture dish, which are reference surfaces for cell growth, amount to (38.9 ± 0.1) mJ m-2, (42.2 ± 4.7) mJ m-2 and (44 ± 4.5) mJ m-2, respectively [31]. XG-cit, HPMC-cit and HPMCoxa scaffolds presented surface energy values in the same range, namely (46.3 ± 0.6) mJ m-2, (44.8 ± 0.8) mJ m-2 and (43 ± 1) mJ m-2. The presence of MNP (0.2 wt% or 0.3 wt%) in the scaffolds slightly decreased the γpS values and increased the γdS values, indicating that the surface energy alone is not a decisive parameter and the presence of MNP stimulates the cell behavior in a particular way. Regarding the stiffness of XGcit, HMPC-cit, XPMC-oxa, XG-cit-MNP, HMPC-cit-MNP and XPMC-oxa-MNP, all of them behave as hydrogels, which are soft materials. Scaffolds should have high porosity to allow diffusion of nutrients to the cells and the diffusion of waste products from the scaffold. Although the optimal pore size depends on the cell type and on the tissue to be regenerated, it generally lies within the micrometer scale. For example, reports indicated pores of 5 μm for neovascularization [32], 30-50 μm for fibroblast [17] or neuronal cells [9], 100-325 μm for osteoblasts [33], and 250-500 μm for odontogenic differentiation [34]. The pore size of all scaffolds applied in the present study ranged from 5 μm to 50 μm. Therefore, in average the physicochemical properties of the scaffolds were similar, except for the fact the XG-cit has more negative charges than HPMC-cit or HPMC-oxa. The order of the negative charge density in the scaffolds is XG-cit > HPMC-cit > HPMC-oxa. The MNP have isoelectric point (pI) of (6.5 ±0.1). The incorporation of 18

0.2 - 0.3 wt% of MNP in the XG-cit scaffolds took place at pH ~ 6.0. At this pH, the MNP are positively charged and attach to the XG-cit scaffold surface by electrostatic interaction. The zeta potential decrease of approximately two units indicated clearly that the MNP attached predominantly on the XG scaffold surface, as evidenced by STEM images reported in a previous work [17], screening the negative charges on the surface. This effect probably favored the pronounced cell adhesion (after 1 day) on XG-cit-MNP in comparison to bare XG-cit or the control, which led to higher cell proliferation (after 7 days or more). Zhu and Fang observed that the adhesion of fibroblasts was much stronger

on

positively

charged

chitosan

than

on

negatively

charged

O-

carboxymethylchitosan due to the electrostatic interaction between cationic sites on chitosan and the negatively charged surface of cell membrane [35]. The binding of MNP to HPMC scaffolds is probably driven by van der Waals forces, which are weak, but not negligible [36]. Noteworthy, leaching of MNP from any scaffold was not observed. The iron content in the scaffolds determined by ICP-OES remained practically the same after cell assays. The best scaffold for fibroblasts proliferation was XG-cit-MNP and all other scaffolds presented here were similar or worse than the polystyrene dish (control), as shown in Figure 4. Also the permeation of Ca2+ ions through the XG-cit-MNP scaffold was the most pronounced (Figure 5a). Davila and co-workers [24] proposed a model to explain the orientation of pigeons flights through biogenic magnetite and the Earth's magnetic field. In the beak of pigeons, the trigeminal nerve is surrounded by biogenic MNP; any small change in the magnetic field is signaled by the magnetoreceptors of the nervous system. Depending on the orientation of the external magnetic field, the MNP can attract or repel each other, deforming the membrane and opening or closing the ions channels [25]. If the external field is parallel to the cell membrane, MNP attract each 19

other, compressing the membrane and closing the ion channels. If the external field is perpendicular to the cell membrane, causing the fields to align laterally, MNP repel each other, stretching the membrane and opening the ion channels. One important detail, which should be highlighted in the Davilla’s model, is that the cell membrane is negatively charged, as xanthan scaffolds. In the present study, the orientation of external magnetic field was not changed to cause attraction or repulsion among the MNP, but the addition of Ca2+ ions increased the medium ionic strength. The increase of ionic strength might weaken the electrostatic interaction between the negatively charged scaffold and MNP, allowing MNP movement, as schematically represented by the arrows in Figure 6a. Depending on the MNP dipole orientation in the scaffold, they can repel or attract each other, causing local compression and stretching. Such mechanical vibrations would speed up Ca2+ ions diffusion and could also stimulate ionic channels in the cell membrane to open and close. In the absence of MNP, the high charge density inside the XG-cit scaffold would attract the Ca2+ ions, regarding their diffusion through the scaffold. In the case of HPMC-cit or HPMC-oxa scaffolds, the charge density is very low, the binding between them and MNP is driven by ion-dipole or dipole-dipole interaction and for this reason, the increase of ionic strength due to the Ca2+ in the medium would have no significant effect on the binding of MNP to the scaffolds, as depicted in Figure 6b. Therefore, upon increasing the medium ionic strength, the MNP would not be compelled to move as in the case of MNP on negatively charged scaffolds. The very low charged HPMC-cit or uncharged HPMC-oxa scaffolds either in the absence or presence of MNP is not attractive for the Ca2+ ions, hence for this reason the Ca2+ ions diffusion is very slow, making pressure against the HPMC based scaffolds and bending them.

20

Thus the model proposed on the basis of Davilla’s model would explain the remarkable effects observed for the hybrid negatively charged scaffolds (XG-cit-MNP), but it also raises the question about the type of particles. Would similar behavior be expected if magnetic nanoparticles were replaced by diamagnetic nanoparticles? The present study refers to fibroblasts, but the answer to this question would require systematic studies with different diamagnetic nanoparticles and cell types. For instance, in the case of differentiation of neuronal embryonic stem cells, hydroxyapatite nanoparticles incorporated in XG-cit scaffolds were suitable only for glial differentiation, while the XG-cit-MNP scaffolds led to motor and sensory neurons. On the other hand, the presence of hydroxyapatite based nanoparticles in the scaffolds has successfully stimulated bone tissues regeneration and mineralization [37-41].

Conclusions The hypothesis that the presence of MNP in charged scaffolds stimulates cell proliferation and Ca2+ ions influx was tested by in vitro cell assays and potentiometric measurements with a microelectrode selective for Ca2+ ions, using scaffolds with high (XG-cit), low (HPMC-cit) and no (HPMC-oxa) charge density prepared in the absence and in the presence of MNP. The MNP are mainly on the surface of XG-cit scaffolds, as evidenced by surface energy and zeta potential data. The results showed that hybrid scaffolds of negatively charged polymer (XG-cit) and MNP were more efficient for the in vitro adhesion and proliferation of fibroblasts and for the permeation of Ca2+ ions than all other scaffolds, indicating that the hypothesis was right. The model proposed to explain the observed effects is based on the electrostatic interaction between MNP and negatively charged scaffolds, which is weakened upon increasing the medium ionic

21

strength. If the electrostatic interactions become weaker, the MNP are free to attract or repel each other, stimulating mechanical movements for the Ca2+ ions permeation through the scaffold and for opening the mechano-sensitive ion channels of cell membranes.

Acknowledgments Authors gratefully acknowledge the financial support from Brazilian Funding Agencies Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP Grant 2015/251032) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq Grants 448497/2014-0 and 305178/2013-0) and Dr. T. Luxbacher, Anton Paar (Austria) for the zeta potential measurements.

References 1. W. L. Murphy, T. C. McDevitt and A. J. Engler, Nat. Mater., 13 (2014) 547-557. 2. A. T. Semeano, T. Glaser, H. Ulrich, H. and D. F. S. Petri, in P. D. Negraes and H. Ulrich, H. (Eds.), Working with Stem Cells, Springer International Publishing, New York, NY, 2016, Chapter 20. 3. P. Gentile, V. Chiono, I. Carmagnola, and P. V. Hatton, Int. J. Mol. Sci., 15 (2014) 3640-3659. 4. J. M. Williams, A. Adewunmi, R. M. Schek, C L. Flanagan, P. H. Krebsbach, S. E. Feinberg, S. J. and S, Das, Biomaterials, 23 (2005) 4817-4827. 5. E. Y. Snyder and K. I. Park, Nat. Med., 8 (2002) 928-930. 6. Y. D. Teng, E. B. Lavik, X. Qu, K. I. Park, J. Ourednik, D. Zurakowski, R. Langer, and E. Y. Snyder, Proc. Natl. Acad. Sci. U S A, 99 (2002) 3024-3029. 7. F. Y. Chou, J. Y. Lai, C. M. Shih, M. C. Tsai and S. J. Lue, Colloids Surf. B: Biointerfaces, 104 (2013) 66– 74. 22

8. Y. Sapir, S. Cohen, G. Friedman and B. Polyak, Biomaterials, 33 (2012) 4100– 4109. 9. T. Glaser, V. B. Bueno, D. R. Cornejo, D. F. S. Petri, H. Ulrich, Biomed. Mater., 10 (2015) 045002. 10. L. Lei, J. Ling-Ling and Z. Yun, Chin. Phys. B, 22 (2013) 127503. 11. J.J. Kim, R. K. Singh, S. J. Seo, T. H. Kim, J. H. Kim, E. J. Lee and H. W. Kim, RSC Adv,. 4 (2014) 17325-17336. 12. N. Bock, A. Riminucci, C. Dionigi, A. Russo, A. Tampieri, E. Landi, V. A. Goranov, M. Marcacci, and V. A. Dediu, Acta Biomater, 6 (2010) 786-796. 13. A. Tampieri, M. Iafisco, M. Sandri, S. Panseri, C. Cunha, S. Sprio, E. Savini, M. Uhlarz and T. Herrmannsdörfer, ACS Appl. Mater. Interfaces 6 (2014) 6 15697– 15707. 14. Y. Wei, X. Zhang, Y. Song, B. Han, X. Hu, X. Wang, Y. Lin and X. Deng. Biomed. Mater., 6 (2011) 055008. 15. S. K. Samal, M. Dash, T. Shelyakova, H. A. Declercq, M. Uhlarz, M. BañobreLópez, P. Dubruel, M. Cornelissen, T. Herrmannsdörfer, J. Rivas, G. Padeletti, S. De Smedt, K. Braeckmans, D. L. Kaplan and V. A. Dediu, ACS Appl. Mater. Interfaces, 7 (2015) 6282–6292. 16. K. Dashnyam, R. A. Perez, R. K. Singh, E. J. Lee and H. W. Kim, RSC Adv, 4 (2014) 40841-40851. 17. V. B. Bueno, A. M. Silva, L. R. S. Barbosa, L. H. Catalani, E. Teixeira-Neto, D. R. Cornejo and D. F. S. Petri, Chem Commun, 49 (2013) 9911-9913. 18. N. Zhang, J. Lock, A, Salleea and H. Liu, ACS Appl. Mater. Interfaces, 7 (2015) 20987−20998.

23

19. L. Gao, L.Zhuang, L. Nie, J. Zhang, Y. Zhang, N. Gu, T. Wang, J. Feng, D. Yang, S. Perrett and X. Yan, Nat Nanotechnol, 2 (2007) 577–583. 20. D. M. Huang, J. K. Hsiao, Y. C. Chen, L. Y. Chien, M. Yao, Y. K. Chen, B. S. Ko, S. C. Hsu, L. A. Tai, H. Y. Cheng, S. W. Wang, C. S. Yang and Y. C. Chen, Biomaterials, 30 (2009) 3645-51. 21. A. Tay, A. Kunze, C. Murray and D. Di Carlo, ACS Nano, 10, (2016) 2331–2341. 22. A. F. Davila, G. Fleissner, M. Winklhofer and N. Petersen, Phys. Chem. Earth, 28 (2003) 647–652. 23. K. J. Lohmann, Nature, 46 (2010) 1140-1142. 24. V. B. Bueno, R. Bentini, L. H. Catalani and D. F. S. Petri, Carbohydrate Polymers, 92 (2013) 1091-1099. 25. P. L. Marani, G. D. Bloisi and Petri, D. F. S. Cellulose, 22 (2015) 3907-3918. 26. B. F. Martins, P. V. Toledo, P. V. O. and D. F. S. Petri, Carbohydrate Polymers, 155 (2017) 173-181. 27. M. K. Chaudhury, Mater. Sci. Eng., 16 (1996) 97–159. 28. D. K. Owens and R. C. Wendt, J. Appl.Polym. Sci., 13 (1969) 1741–1747. 29. C. Wei, A. J. Bard, G. Nagy and K. Toth, Anal. Chem., 67 (1995) 1346-1 356. 30. S.-K. Lee, W. F. Boron and M. D. Parker, Sensors, 13 (2013) 984-1003. 31. E. M. Harnett, J. Alderman and T. Wood, Colloids Surf B: Biointerfaces, 55 (2007) 90-97. 32. J .H. Brauker, V. E. Carr-Brendel, L. A. Martinson, J. Crudele, W. D. Johnston and R. C. Johnson, J. Biomed. Mater. Res., 29 (1995) 1517-1527. 33. V. Karageorgiou and D. Kaplan, Biomaterials, 26 (2005) 5474–5491. 34. C. M. Murphy, M. G. Haugh and F. J. O'Brien, Biomaterials, 31 (2010) 461–466. 35. A. P. Zhu and N. Fang, Biomacromolecules, 6 (2005) 2607-2614.

24

36. S. V. Aradhya, M. Frei, M. S. Hybertsen, L. Venkataraman, Nat. Mater., 11 (2012) 872-876. 37. R. Schröder, H. Pohlit, T. Schüler, M. Panthöfer, R. E. Unger, H. Frey and W. Tremel, J. Mater. Chem. B, 3 (2015) 7079-7089. 38. S. V. Dorozhkin, J. Funct. Biomater., 6 (2015) 708-832. 39. G. X. Ni, K. Y. Chiu, W. W. Lu, Y. Wang, Y. G. Zhang, L. B. Hao, Z. Y. Li, W. M. Lam, S. B. Lu and K. D. K. Luk. Biomaterials, 27 (2006) 4348–4355. 40. C. T. Wong, W. W. Lu, K. M. C. Chan, K. M. C,; Cheung, K. M .C.; Luk, K. D. K.; D. S. Lu, A. B. M. Rabie, L. F. Deng and J. C. Y. Leong, J. Biomed. Mater. Res. A, 68A (2004) 513–521. 41. V. B. Bueno, R. Bentini, L. H. Catalani, L. R. S. Barbosa and D. F.S. Petri, Mater. Sci. Eng. C, 37 (2014) 195 - 203.

25

Citric acid

Oxalic acid

xanthan

HPMC

Figure 1. Representation of the chemical structure of xanthan repeating unit, citric and oxalic acid, and HPMC repeating unit, with R = -H, -CH3, -CH2-CHOH-CH3.

26

0.04

-0.02

-2

-1

0

1

H (kOe)

2

3

5K 300 K

0.00

(a)

0.00

-0.04

-0.06

-0.12 -3.0

5K 300 K

0.04

(emu/g)

0.06

0.00

-0.04 -3

HPMC-oxa-MNP

HPMC-cit-MNP

5K 300 K

 (emu/g)

 (emu/g)

0.02

0.08

0.12 XG-cit-MNP

-2.0

-1.0

0.0

1.0

H (kOe)

(b)

2.0

3.0

-0.08 -3.0

-2.0

-1.0

0.0

1.0

2.0

3.0

H (kOe)

(c)

Figure 2. SEM images for cryofracture surface of (a) XG-cit, (b) HPMC-cit and (c) HPMC-oxa. The scale bars correspond to 20 μm, 10 μm and 20 μm, respectively. Hysteresis loops determined for (a) XG-cit-MNP, (b) HPMC-cit-MNP and (c) HPMCoxa-MNP at 5 K and 300 K.

27

 (mJ/m²)

60

total

dispersive

polar

40

20

* **

*

** -o xa

-c C

PM

PM

C

HPMC-citMNP

it

XG-citMNP

HPMC-oxaMNP

H

H

XG

-c

it

0

Figure 3. Polar (γpS, blue columns) and dispersive (γdS, red columns) components of the total surface energy (γS, grey columns) determined for all scaffolds. Statistical

0.8

1d 7d 14 d 21 d

*

0.6 0.4 0.2 0.0

control -c

it

(a) P N

XG-cit-MNP

M

XG-cit

XG

O.D. (570 nm)

significance was set at p < 0.05 (*) or p < 0.002(**).

28

O.D. (570 nm)

1.0

3d 7d 10 d

0.5

0.0

HPMC-cit

HPMC-cit-MNP

Control

(b)

O.D. (570 nm)

1.0 1d 3d 7d 10 d

0.5

0.0

HPMC-oxa

HPMC-oxa-MNP

Control

(c) Figure 4. MTT assay of fibroblasts on (a) XG-cit, XG-cit-MNP and control, (b) HPMC-cit, HPMC-cit-MNP and control and (c) HPMC-oxa, HPMC-oxa-MNP and control. Experiments were performed as three independent experiments (* p < 0.05, versus control).

29

18

2+

[Ca ] (mM)

15 12 9 6 3 0 0

250

500

750

1000

1250

90 60

(sat)

120

E vs. Ag/AgCl/KCl

E vs. Ag/AgCl/KCl(sat) (mV)

150

(mV)

time (s)

120 90 60 30 0 -30 1

30

(a)

2

3

4

pCa2+

5

6

7

0 -30 0

200

400

600

Time (s)

800

1000

(b)

Figure 5. (a) Calcium ion concentration measured with a potentiometric sensor positioned at 300 μm above (──) HPMC-cit, (──) XG-cit and (──) XG-cit MNP scaffolds using the setup shown in Supplementary Information SI3. (b) Calibration plots obtained with a Ca2+-potentiometric sensor before (black line) and after (red line) a permeation experiment.

30

(a)

(b)

Figure 6. (a) Schematic representation of how the increase of ionic strength caused by the addition of Ca2+ ions (blue spheres) to the medium might affect the scaffolds. (a) XG-cit has many charges available inside the network and the diffusion of Ca2+ ions is slow because they are attracted by the free carboxylate groups. In the presence of MNP (red spheres), the electrostatic interaction between the negatively charged XG-cit scaffold and MNP decreases due to charge screening, allowing MNP movement; 31

depending on the orientation of MNP magnetic dipole in the scaffold, the MNP can repel or attract each other, causing local compression and stretching. Such movements stimulate the cells and propel the Ca2+ ions through the scaffold. (b) The very low charged HPMC-cit scaffolds either in the absence or presence of MNP would be less attractive for the Ca2+ ions than XG-cit and for this reason the Ca2+ ions diffusion is much slower. The interaction between HPMC and MNP is not electrostatic, therefore the increase of ionic strength due to the Ca2+ ions in the medium would have no significant effect on the binding of MNP either to HPMC-cit or HPMC-oxa scaffolds.

32