3D cytocompatible composites of PCL/Magnetite

materials Article 3D Cytocompatible Composites of PCL/Magnetite Esperanza Díaz 1,2,* , María Blanca Valle 3, Sylvie Ribeiro 4,5, Senentxu Lanceros-Mendez 2,4,6 and JoséManuel Barandiarán2,3 1Escuela de Ingeniería de Bilbao, Departamento de Ingeniería Minera, Metalúrgica y Ciencia de Materiales, Universidad del País Vasco (UPV/EHU), 48920 Portugalete, Spain 2BCMaterials, Basque Centre for Materials, Applications and Nanostructures, UPV/EHU Science Park, 48940 Leioa, Spain; [email protected] (S.L.-M.); [email protected] (J.M.B.) 3Facultad de Ciencia y Tecnología, Departamento Electricidad y Electrónica, Universidad del País Vasco (UPV/EHU), 48940 Leioa, Spain; [email protected] 4Centro de Física, Universidade do Minho, 4710-057 Braga, Portugal; sylvie.ribeir[email protected] 5Centre of Molecular and Environmental Biology (CBMA), Universidade do Minho, Campus de Gualtar, 4710-057 Braga, Portugal 6IKERBASQUE, Basque Foundation for Science, 48013 Bilbao, Spain *Correspondence: [email protected]; Tel.: +34-66-678-7663 Received: 10 October 2019; Accepted: 19 November 2019; Published: 21 November 2019   Abstract: A study of Magnetite (Fe 3 O 4 ) as a suitable matrix for the improved adhesion and proliferation of MC3T3-E1 pre-osteoblast cells in bone regeneration is presented. Biodegradable and magnetic polycaprolactone (PCL)/magnetite (Fe 3 O 4 ) scaffolds, which were fabricated by Thermally Induced Phase Separation, are likewise analyzed. Various techniques are used to investigate in vitro degradation at 37 ◦ C, over 104 weeks, in a phosphate buffered saline (PBS) solution. Magnetic measurements that were performed at physiological temperature (310 K) indicated that degradation neither modified the nature nor the distribution of the magnetite nanoparticles. The coercive field strength of the porous matrices demonstrated ferromagnetic behavior and the probable presence of particle interactions. The added nanoparticles facilitated the absorption of PBS, with no considerable increase in matrix degradation rates, as shown by the Gel Permeation Chromatography (GPC) results for Mw, Mn, and I. There was no collapse of the scaffold structures that maintained their structural integrity. Their suitability for bone regeneration was also supported by the absence of matrix cytotoxicity in assays, even after additions of up to 20% magnetite. Keywords: PCL; magnetite; scaffolds; magnetism; cytotoxicity; in vitro degradation 1. Introduction Nanomaterials have attracted increasing scientific interest, due to their numerous biomedical applications. Among such materials, Iron-Oxide NanoParticles (IONPs) awaken special interest, because of their magnetic properties, biocompatibility, and nanometric sizes that can interact at the cellular, subcellular, and even molecular levels [ 1 , 2 ]. As iron is a metal that is found in haemoglobin and myoglobin in the form of ferritin, the particles are non-toxic, although particle aggregation can be dangerous [ 3 – 5 ]. Mainly used in biomedicine, IONPs have FDA approval for use in cellular therapy, drug delivery, contrast agents, tumor hyperthermia, and tissue repair. At particle sizes of less than 30 nm, their diameters range between 5 and 100 nm. Each particle produces a nanoscale magnetic field and acts as a single magnetic domain, showing superparamagnetic behavior [2,6,7]. Biodegradable and biocompatible polymers, such as polycaprolactone (PCL), are widely used in biomedical regenerative treatments of hard tissue [ 8 – 10 ], where prolonged degradation times require Materials 2019,12, 3843; doi:10.3390/ma12233843 www.mdpi.com/journal/materials Materials 2019,12, 3843 2 of 15 lengthy curing phases [ 11 – 15 ]. PCL scaffolds with lower molecular weights and faster degradation rates are also used in the regenerative treatment of soft tissues, such as cartilage and muscle [12,14]. Therefore, scaffolds have become a key material in tissue engineering, in which various biological agents interact for their correct development [ 3 , 6 ]. The manufacture of scaffolds with magnetic fields that attract these biological agents has implied a significant advance. A similar concept was previously used for tumor hyperthermia and the administration of pharmaceuticals [ 7 , 14 – 16 ]. Therefore, magnetic nanoparticles and scaffolds are promising alternatives for bone regeneration. The magnetic fields of the scaffolds function by attracting and absorbing stem cells, in vivo growth factors, and other bioactive agents tied to magnetic particles, favoring bone regeneration and repair [17–23]. There are two methods for the manufacture of magnetic scaffolds: one consists of submerging a scaffold in water-based ferrofluids that contain IONPs through capillary adsorption, as described by other authors [ 6 , 19 ]. The other method involves the incorporation of IONPs during scaffold preparation, a method that requires effective ultrasonic nanoparticle dispersion, to prevent agglomeration [ 20 – 27 ]. A simple and efficient method of incorporating IONP in a scaffold that is manufactured by lyophilization will be presented in this study, so the scaffold magnetization process will not affect the external structure of the scaffold. An in vitro degradation study of the scaffolds manufactured with this method will also be performed. In addition, the utility of these magnetic scaffolds will be demonstrated in a cytotoxicity evaluation of samples with the MC3T3-E1 pre-osteoblast cell line for bone regeneration. 2. Experimental 2.1. Materials PCL polymer, which was supplied by Purac Biomaterials Purasorb PCL 12 (Gorinchem, Amsterdam, the Netherlands), was purified by dissolution in 1,4-dioxane (Panreac p.a. Barcelona, Spain). 1,4-Dioxane was used as a solvent. The values of Mw =134,418, Mn =92,103, and polydispersity Mw/Mn =1.459 were performed while using gel permeation chromatography (GPC, Perkin Elmer 200, Bridgeport, CT, USA) in tetrahydrofuran (THF). The PBS (phosphate buffer solution) used in the in vitro degradation was supplied by Fluka Analytical (Sigma-Aldrich, Darmstadt, Hessen, Germany) at a pH of 7.2. The magnetite nanoparticles were supplied by Sigma-Aldrich), CAS Number 1317-61-9, the particle size ≈50 nm. 2.2. Fabrication of Porous Matrices Pure PCL and PCL/Fe 3 O 4 composite scaffolds were made by Thermally-Induced Phase Separation (TIPS) and then by a freeze-drying technique. In brief, a solution of PLLA in 1,4-Dioxane (2.5% (w/v)). Subsequently, magnetite particles (in solid state) were homogeneously blended, in proportions of 5 and 10% of total polymer mass, by ultrasonic stirring. The solutions were frozen and freeze-dried (LyoQuest of Telstar, Barcelona, Spain) for seven days for the complete elimination of the 1,4 dioxane. 2.3. In vitro Degradation For the degradation test, the porous matrices were cut into 0.5 cm 2 rectangular section, weighed, immersed in tubes with 15 mL of PBS, and then incubated in a thermostated oven at 37 ◦ C. The pieces were recovered after 5, 8, 10, 16, 20, 25, and 28 weeks and then wiped. A pH meter PCE228 (PCE Instruments, Palm Beach, FL, USA) was used to determine the pH changes in the PBS. Water absorption (W a ) and mass loss (W L ) were respectively calculated by the following equations: Wa%= Ww −Wr Wr ×100% (1) WL%= Wo −Wr Wo ×100% (2) Materials 2019,12, 3843 3 of 15 where, W 0 is the original weight, W r is the residual weight after degradation and when completely dried, and Wwis the weight of the sample without surface water. 2.4. Magnetic Analysis A Vibrating Sample Magnetometer (VSM) that was developed at the University of the Basque Country was used for magnetic measurements at physiological temperature (310 K). Pure nickel was used to calibrate the VSM. The magnetic field range was ± 1.8 Tesla (18 kG) and the resolution ± 20 µTesla (0.2 G). 2.5. SEM Analysis The morphology of the porous matrices was examined while using scanning electron microscopy (SEM) (HITACHI S-4800, Tokyo, Japan). Before the analysis, a JEL Ion Sputter JFC-1100 (Amiron Machinery, Oxnard, CA, USA) at 1200 V and 5 mA was used to coat with a cape of gold. 2.6. Differential Scanning Calorimetry (DSC) DSC assays were made with a Q200 TA Instruments. A nitrogen purge gas was used to avoid any oxidation of the scaffolds during the experiments. These were repeated three times. In the first scan, the scaffolds were heated from − 20 to 250 ◦ C at a rate of 10 ◦ C · min −1 to record the melting temperature, Tm, and the melting heat, ∆ H melt . Subsequently, cooled at 10 ◦ C · min −1 to obtain the glass transition, T g , and then finally heated at 10 ◦ C/min. Crystallinity, Xc, was calculated with the following equation: Xc(%) =(∆Hmelt −∆Hcrystallization)/∆H100% ×100% (3) ∆ H melt (J · g −1 of the crystalline polymer) is the enthalpy of fusion of the sample and ∆ H 100% is the enthalpy of fusion of a 100% crystalline polymer, which for PCL was 139.5 J · g −1 [ 24 ]. The crystallizable fraction (CF%) of the samples was calculated with the following equation: CF% =(∆Hc/∆Hm)×100% (4) 2.7. Fourier-Transform Infrared (FTIR) Spectroscopy A Thermonicolet Avatar 370 Fourier-transform infrared (FTIR, Thermo Electron Corporation, Waltham, MA, USA) spectrophotometer that was equipped with an attenuated total reflectance attachment with ZnSe crystal was used to make the infrared spectra. 2.8. Cytotoxicity Assay Membrane sterilization: For the in vitro assays, the membranes were cut with 0.1mg · mL −1 following an adaptation of the ISO 10993-5:2009 (third edition) standard test method (indirect cytotoxicity evaluation). The samples were sterilized by UV (1 h each side). After that, each sample was washed five times (5 min each wash) in phosphate buffered saline (PBS) solution. Cytotoxicity process: The samples were immersed in a 24-well tissue culture polystyrene plate with DMEM obtaining like this a conditioned medium. This medium contains 1 g · L −1 glucose (Gibco) that was supplemented with 10% FBS (Biochrom) and 1% P/S (Biochrom) (basal medium). The plate was incubated for 24 h at 37 ◦ C, in a 95% humidified atmosphere containing 5% CO 2 . In parallel, the MC3T3-E1 cells, Sigma-Aldrich (Darmstadt, Hessen, Germany) (passage number: 36, Riken bank) were seeded in the 96-well tissue culture polystyrene plate at a density of 3 × 10 4 cells · mL −1 and then incubated for cell plate attachment (24 h). Thereafter, the culture medium from the 96-well tissue culture polystyrene plate was removed and 100 µ L of conditioned medium was added to the wells incubating for 24 and 72 h. 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was used to quantify the viable cells. In this method, it was used a positive control (20% solution of dimethylsulfoxide Materials 2019,12, 3843 4 of 15 (DMSO, Sigma Aldrich)) and a negative control (cell culture fresh basal medium) to compare with the values that were obtained for the samples in study. Likewise, the conditioned medium was removed from each well at the end of each cycle (24 and 72 h) and the MTT solution (5 mg · mL −1 in PBS dissolved in DMEM in proportion of 10%) was added to the cells. The incubation was done for 2 h in the dark at 37 ◦ C. After the intended contact time, it was added 100 µ L of DMSO/well to dissolve the obtained MTT crystals after removing the medium. In the final, microplate reader at 570 nm was used to measure optical density All the quantitative results were obtained from four replicates samples and controls and they were analyzed as the average of viability ±standard deviation (SD). 3. Results and Discussion In this section, the results are presented of the in vitro degradation of the magnetic scaffolds that were manufactured with PLC incorporating magnetized particles of nFe3O4. 3.1. Magnetism The hysteresis loops of both the Fe 3 O 4 nanoparticles and of the PCL/nFe 3 O 4 scaffolds are shown to be normalized to the nFe 3 O 4 content in Figure 1a for selected measurements. Magnetic parameters that cover magnetization measured at 1.5 T (close to saturation), the coercive field ( µ0 H c ), and the recalculated content of nanoparticles are summarized in Table 1, as the average of several measurements for each composition (typically three or four samples each). Errors have been estimated as a combination from the standard deviation of the measurements and the absolute accuracy of the instruments and they are always around 2% of the averaged value. As can be seen, the coercivity of the isolated particles and scaffolds is quite similar. Pure magnetite coercivity ≈ 12 mT ( ≈ 120 G) is high enough to discount classic superparamagnetic behavior. Therefore, the nanoparticle magnetic properties might be due to strong inter-particular interaction [ 27 ] or to the true ferromagnetic state, while considering that their sizes (diameters over 30 nm) are within the single domain region. Figure 1. ( a ) Magnetization curves of PCL-nFe 3 O 4 . ( b ) Magnetization curves of PCL-10%nFe 3 O 4 before and after 20 weeks degradation. Materials 2019,12, 3843 5 of 15 Table 1. Magnetization at 1.5 Tesla relative to the nominal content of nFe 3 O 4 , coercive field, and actual content of nFe 3 O 4 recalculated from the measurements. Estimated errors are displayed in brackets and refer to the variation of the last significative figure. % nFe3O4 Nominal Magnetization at 1.5T (Am2/kg nFe3O4) µ0Hc (mT) %nFe3O4 Recalculated 100 * 82 (2) 11.9 (2) - 5 78 (2) 12.0 (2) 4.8 (1) 10 81 (2) 11.9 (2) 9.9 (2) 15 75 (1) 11.1 (2) 13.8 (3) 20 69 (1) 11.4 (2) 16.8 (4) * Pure nFe3O4serving as a reference. Strong interaction within the scaffolds appears to be a less likely hypothesis, as the different Fe 3 O 4 concentrations caused no changes of coercivity. Particulate interactions can be short or long range: i.e., boundary exchanges or dipolar interactions, respectively. The latter can favor agglomerations, which leads to increased coercivity, by promoting exchange coupling within the agglomerates. Nevertheless, agglomerates are not seen in the microscopy studies, which suggest the good particulate dispersion at all concentrations, with low dispersion of inter-particle distances for each sample. Figure 1b shows the effect of 20 weeks degradation on the scaffolds PCL-10%nFe 3 O 4 . Both curves, before and after, are superimposed, which indicated that polymer degradation neither affected the characteristics nor the distribution of the magnetic nanoparticles. 3.2. FTIR The FTIR technique characterizes the samples that were subjected to degradation and those that have not been degraded. The characterization through FTIR of the porous substrates composed of PCL-nFe 3 O 4 (see Figure 2) indicated no interaction between both materials, so the magnetic nanoparticles must therefore be scattered throughout the scaffold structure. The typical vibration bands of C=O and the stretching vibration bands of C–O to 1725 cm −1 and to 1293 cm −1 , respectively, as well as the symmetric and asymmetric modes of CH 2 , at 2923 cm −1 and at 2940 cm−1, respectively, refer to the polymer (PCL). A band of H 2 O located above 3420 cm −1 was observed that grew in strength when increasing the amount of magnetite; a characteristic vibration band of the Fe–O bond, hardly perceived in the compositions of 5 and 10% nFe 3 O 4 , became clearer in the PCL-nFe 3 O 4 scaffolds containing a higher magnetite content; see Figure 2. The hydrophilic nature of the system was reflected by the appearance of broad peaks revealing the presence of hydroxyl groups around 3420 cm −1 (Figure 2) for all compositions. These bands indicated the presence of water in the system, despite the lyophilization process. In Figure 3a,b, the spectra of both PCL-5% nFe 3 O 4 and PCL-10% nFe 3 O 4 were observed after 0, 8, 20, and 25 weeks of degradation, in which no appreciable signs of degradation were noted, due to the existence of the broad band corresponding to the hydroxyl groups at around 3420 cm −1 , which might hide the appearance of a band between 2800 and 3600 cm −1 attributable to a stretching vibration of the OH bands of the COOH and OH groups. Neither can an increase in the intensity be observed in the region of 1100–1050 cm −1 . Both of the changes would be due to the scission of the polymeric chains through the ester band, producing a carboxylic group and an alcohol group, as a consequence of the degradation of the PCL. Materials 2019,12, 3843 6 of 15 Figure 2. Fourier-Transform Infrared (FTIR) spectra of PCL-n Fe3O4. Figure 3. ( a ) Spectra of PCL-5% nFe 3 O 4 after various degradation times. ( b ) FTIR spectra of PCL-10% nFe3O4after various degradation times. Materials 2019,12, 3843 7 of 15 3.3. DSC Table 2shows the values of the characteristic glass-transition temperatures, T g , and fusion temperature (T m ), as well as the fusion enthalpies ( ∆ H m ), and crystallization ( ∆ H c ), and the degree of crystallinity. The T g temperature of the PCL-nFe 3 O 4 -porous scaffolds, − 59 ◦ C, was close to the values cited by other authors for PCL (~−60 ◦C). Table 2. Parameters obtained by DSC on the PCL-nFe 3 O 4 scaffolds: Tm =melting point ( ◦ C), ∆ H m = melting enthalpy (kJ/kg), Tc =crystallization temperature ( ◦ C), ∆ H c =crystallization enthalpy (kJ/kg), T g =glass transition temperature ( ◦ C), Xc =crystalline fraction (%), calculated as Xc% =[( ∆ H m1 − ∆Hc)/∆Hmo]×100% with ∆Hmo =139.5 J·g−1, CF% =(∆Hc/∆Hm1)×100%. Sample PCL dt (weeks) 1st Run 2nd Run 3rd Run Tg(◦C) Xc%CF% Tm1 (◦C) ∆Hm1 (J·g−1) Tc (◦C) ∆Hc (J·g−1) Tm2 (◦C) ∆Hm2 (J·g−1) 0%nFe3O460 107.4 26 64.7 57 83.2 −59 77 60.2 5%nFe3O40dt 60 94.6 30 60.2 58 74.1 −59 68 64 4dt 63 93.9 31 59.2 57 65.6 −58.5 67 63 8dt 64 88.7 30.5 58.6 58 65.4 −60 64 66 16dt 63 90 30 58.1 58 65.4 −63 65 65 20dt 67 80.5 31.5 53 56.5 72.9 −63 58 65 25dt 67 77.7 31 52.4 56 68.5 −63 56 67 104dt 65 90.2 31 57.4 58 59.7 −64 65 64 10%nFe3O40dt 59 92.2 31 59.9 56 73.6 −59 66 65 4dt 63 67 31 47.2 57.5 49 −58 48 70 8dt 64 91 31 59.3 58 66.9 −61 65 65 16dt 63 80.2 31 49.7 58 64.4 −58 57 62 20dt 66 77.5 30.5 49.4 55.5 57.6 −63 56 62 25dt 67 76.9 32 48 56 54.7 −63 55 62 104dt 65 94.3 31 55.4 57 62.2 −63 68 59 15%nFe3O40dt 59 83.1 30 52 59 69.7 −58 60 63 20%nFe3O40dt 58 81.3 30 51 58 65.9 −58 58 63 Even though PCL is a hydrophobic polymer, the presence of magnetite nanoparticles that were embedded in the polymeric matrix make the surface hydrophilic rather than hydrophobic, improving its biocompatibility through intramolecular bridging interactions, and hydrogen bonds. However, the addition of nanoparticles modified the polymer, reducing the degree of crystallinity (X c %) of the PCL, which passed from a value of 77% in the scaffolds of pure polymer, to 55% in PCL-20%nFe 3 O 4 , indicating that the magnetite nanoparticles played no role as nucleation agents [ 28 ]. The T c (crystallization temperature) notably increased with the addition of the nanoparticles, which entailed a much slower crystallization process of the PCL-nFe 3 O 4 scaffolds, thereby confirming that the nanoparticles were not acting as nucleating agents. ∆ H m clearly diminished with the increased concentration of nanoparticles, from 107 J · g −1 to 81 J · g −1 , in the pure PCL and in the PCL with a content of 20% nFe3O4, respectively. The samples with a higher content of nanoparticles presented lower crystallinity and fusion enthalpies, due to the increased rigidity of the polymeric chains [ 28 , 29 ], which was clearly evident from both the crystallinity percentages and the Tg. During the degradation process, the T g fell to temperatures lower than − 59 ◦ C in week 0 to − 63 ◦ C in week 104 for both series of compound materials. Other authors have observed this same diminishment of Tgin relation to the in-vitro degradation [30]. The degree of crystallinity of these samples hardly varied over the time of immersion. In addition, a stretching might be seen in the fusion temperature interval for week 25 in both compositions. Materials 2019,12, 3843 8 of 15 In Figure 4, T m1 can be seen to gain a much higher temperature of 65 ◦ C for both compositions for week 104 of degradation and, at the same time, has a broad endothermic peak. This behavior is indicative of crystallites of different sizes, a form of degradation in which polymeric chains of different lengths are produced that are characterized by their different fusion temperatures. The shortening of the polymeric chains was unrelated to increased crystallinity. In any case, the presence of polymeric chains of different lengths makes it clear that heterogeneous degradation took place at this stage [31]. Figure 4. ( a ) Thermograms of PCL-5%nFe 3 O 4 after various degradation times. ( b ) Thermograms of PCL-10%nFe3O4after various degradation times. 3.4. SEM In Figure 5, we can observe the magnetite nanoparticles and the PCL scaffold both without mixing. Magnetite nanoparticles were analyzed while using TEM and SEM techniques Figure 1a,b, they are agglomerated and it was very difficult to identify single nanoparticles of ~50 nm, as indicated by the provide. The matrix manufacturing technique (TIPs) with low percentages of polymer and a volume of solvent created highly porous scaffolds with random pore structures, as may be seen in Figures 5c and 6. A typical nucleation structure in all of the scaffolds, due to an excessively rapid cooling speed, was obtained, as the solid-liquid phase separation process to take place [32]. We can also see in Figure 6that the increased content of nFe 3 O 4 in the composition of the scaffold produced larger-sized pores and thinner walls, acquiring a more scattered appearance and more irregular pore distribution. Measuring the exact sizes of the pores was not simple, due to the anisotropy of the porosity. In general, the macropores appeared to have average diameters of ~100 µ m for the composition with 5% magnetite and ~130 µ m for the composition with 10%. The diameter of the macropores fluctuated between 25–210 µ m and the micropore diameter was between 2–10 µ m, for the composition with 5% magnetite and between 45–250 µ m, and between 4–15 µ m, respectively, for the composition with 10%. Both of the compositions presented the typical topographic features of the nucleation mechanism and growth, with large perceptible increases (Figure 6a,d). The surface of the pore walls contained hill-style structures, with small peaks that gave the walls an appearance of micro-nano ruggedness (marked with green circles in Figure 6a,d) [ 33 ]. As the process of in-vitro degradation advanced, the pores both shrank and lengthened (Figure 6b,e); both of the compositions presented greater microporosity. In the PCL-5%nFe 3 O 4 composition, the gradual appearance of micropores can start to be appreciated in a generalized way on the walls of the scaffold skeletons in week eight of degradation (Figure 6b), Materials 2019,12, 3843 9 of 15 while in the composition with 10% nFe 3 O 4 , they were appreciable from week 16 onwards (Figure 6e). The pore walls, rather than not clean and smooth, were rugged, characteristic of the liquid-liquid separation through the nucleation mechanism and growth. Figure 5. ( a ) Scanning electron microscopy (SEM) micrograph of magnetite nanoparticles. ( b ) TEM micrograph of magnetite nanoparticles. (c) SEM micrograph of poly(ε-caprolactone) (PCL) scaffold. The magnetite particles were visible at higher magnifications after the degradation process at week 25, as can be clearly seen in Figure 6c (blue circles) for the composition with fewer particles. While observing Figure 6c,e, it may be seen that there were more broken fibers and slightly deteriorated pores in week 25 of degradation: no degradation products were observed. This deterioration process provided the clearest morphological evidence of scaffold degradation, although the structural integrity of the scaffolds was maintained and no collapse of the structure occurred.