Magnetic poly(ε-caprolactone)/iron-doped hydroxyapatite nanocomposite substrates for advanced bone tissue engineering

Abstract

In biomedicine, magnetic nanoparticles provide some attractive possibilities because they possess peculiar physical properties that permit their use in a wide range of applications. The concept of magnetic guidance basically spans from drug delivery and hyperthermia treatment of tumours, to tissue engineering, such as magneto-mechanical stimulation/activation of cell constructs and mechanosensitive ion channels, magnetic cell-seeding procedures, and controlled cell proliferation and differentiation. Accordingly, the aim of this study was to develop fully biodegradable and magnetic nanocomposite substrates for bone tissue engineering by embedding iron-doped hydroxyapatite (FeHA) nanoparticles in a poly(ε-caprolactone) (PCL) matrix. X-ray diffraction analyses enabled the demonstration that the phase composition and crystallinity of the magnetic FeHA were not affected by the process used to develop the nanocomposite substrates. The mechanical characterization performed through small punch tests has evidenced that inclusion of 10 per cent by weight of FeHA would represent an effective reinforcement. The inclusion of nanoparticles also improves the hydrophilicity of the substrates as evidenced by the lower values of water contact angle in comparison with those of neat PCL. The results from magnetic measurements confirmed the superparamagnetic character of the nanocomposite substrates, indicated by a very low coercive field, a saturation magnetization strictly proportional to the FeHA content and a strong history dependence in temperature sweeps. Regarding the biological performances, confocal laser scanning microscopy and AlamarBlue assay have provided qualitative and quantitative information on human mesenchymal stem cell adhesion and viability/proliferation, respectively, whereas the obtained ALP/DNA values have shown the ability of the nanocomposite substrates to support osteogenic differentiation.

1. Introduction

Tissue engineering is defined as a multi-disciplinary field, which benefits from principles of life sciences and engineering to develop biological substitutes that are able to restore, maintain or improve tissue function [1–4].

Taking into account the tissue engineering approach, it should be possible to overcome the limitations of conventional treatments that are mainly based on biomaterial implantation and organ transplantation [1,2,4]. Tissue engineering should potentially produce tissue substitutes that are able to ‘grow’ with the patient [1].

Many strategies, including cell-based therapies, tissue-inducing factors and biocompatible scaffolds, have been adopted to achieve tissue regeneration and studied separately or in combination.

Particularly, it is well known that scaffolds must satisfy several requirements such as controlled biodegradability and bioresorbability, interconnecting pores with specific shape and adequate size, suitable mechanical and mass transport properties, and surface topography and chemistry to favour cell adhesion, proliferation and differentiation.

In this context, different kinds of materials (natural, synthetic, semi-synthetic and hybrid) have been considered or properly synthesized to develop scaffolds according to the specific tissue requiring regeneration, also focusing the attention on the concept of cell guidance and the complex features of cell–material interaction [4,5]. Inorganic ceramic materials, such as hydroxyapatite and tricalcium phosphate, have been considered to manufacture scaffolds for bone tissue engineering, with appropriate performances [4–6].

Even though ceramic materials resemble the inorganic phase of natural bone and possess osteoconductive properties, they are brittle and do not match its mechanical performances [4,5,7]. On the other hand, ceramic scaffolds are also inadequate for soft-tissue engineering as they present a higher stiffness.

In the field of tissue regeneration, an interesting alternative should be represented by both natural (i.e. collagen, alginate, agarose, chitosan, fibrin and hyaluronic acid) and synthetic polymers [4]. In contrast to natural polymers, synthetic polymers show advantages such as more versatility and processability. Their degradation rate and mechanical performances, as well as their physical and chemical properties, can be suitably modulated by varying the chemical structure or composition of the macromolecule. Moreover, they should be surface-modified and bioactivated [4,5].

In this context, synthetic aliphatic polyesters such as polylactic acid and polyglycolic acid, and their copolymers such as poly(lactic-co-glycolic acid) and poly(ε-caprolactone) (PCL) have been widely studied [4,5].

According to the type of tissue being engineered (i.e. soft or hard tissue), scaffolds must possess suitable properties and structures made of either ceramics or polymers should be either too brittle or too flexible, respectively. For this reason, research attention has been focused on the development of polymer-based composite scaffolds consisting of polymers reinforced with inorganic ceramic micro-/nanofillers, especially for bone tissue regeneration [4,8–10]. Compared with polymeric structures, polymer-based composite scaffolds possess enhanced mechanical properties, while showing flexibility and structural integrity better than ceramic scaffolds. In this context, the mechanical performances clearly play a key role because hard tissues such as bone are stronger (higher strength) and stiffer (higher elastic modulus) than soft tissues [11].

As bone is a natural nanocomposite, nanocomposites consisting of a polymer matrix and inorganic-reinforcing nanofillers better reproduce its natural structure, and result in suitable candidates for bone tissue engineering. In comparison with conventional composites, they seem to induce cell response more efficiently and generally possess enhanced mechanical performances [4,12].

All the substrates/scaffolds that have been developed for hard- and soft-tissue regeneration, available on the market to date, present limitations related to the control of cell differentiation and angiogenesis processes [13]. It is well documented that the complete biological and morphological maturation of the tissue requires an extensive angiogenesis process as a crucial feature [13–15].

The aim of this study was to design and develop fully biodegradable and magnetic nanocomposite substrates for bone tissue engineering consisting of a PCL matrix reinforced with bioresorbable iron-doped hydroxyapatite (FeHA) nanoparticles.

The rationale for designing magnetic substrates/scaffolds is the possibility to obtain structures that can be manipulated in situ by applying external magnetic fields, which can also control specific processes at cell level by releasing biomolecules and bioactive factors, in turn, linked to magnetic nanocarriers [13,16–19].

The previously described concept of magnetic guidance basically spans from biomedicine to tissue engineering, involving drug delivery, hyperthermia treatment of tumours, magneto-mechanical stimulation/activation of cell constructs and mechanosensitive ion channels, magnetic cell-seeding procedures, and controlled cell proliferation and differentiation [16,20–37].

In the field of biomedicine, magnetic nanoparticles (MNPs) possess peculiar physical properties and provide some attractive possibilities because of their dimensions, ranging from a few nanometres up to tens of nanometres, that make them comparable to several biological entities. They show sizes that are close to or smaller than those of a protein, a virus, a cell or a gene. Basically, their magnetic features can be manipulated by applying an external magnetic field gradient. This suggests that it would be possible to immobilize and/or to transport the MNPs themselves and magnetically tagged biological units. Drugs or a group of radionuclide atoms could be suitably delivered to a targeted region of the body (i.e. a tumour). MNPs can resonantly respond to a time-varying magnetic field, and several advantages can be obtained by the energy transfer from the exciting field to the MNP [20]. As an example, MNPs may be heated up, allowing their use as hyperthermia agents able to deliver thermal energy to targeted bodies (i.e. tumours) or as elements capable of improving chemotherapy or radiotherapy by providing a degree of tissue warming appropriate for the destruction of malignant cells. Over the years, different kinds of magnetic micro-/nanoparticle carriers have been optimized and proposed for drug delivery applications. MNPs may be properly shielded from the surrounding environment or functionalized/bioactivated with specific molecules, biocompatible polymers (i.e. polyvinyl alcohol (PVA), polyvinylpyrrolidone (PVP) dextran or others). Inorganic silica-based materials have also been considered as coatings for magnetite (Fe₃O₄) or maghemite (γ-Fe₂O₃) MNPs [20].

Markaki & Clyne [25,26] studied the magneto-mechanical stimulation and actuation of a bonded array of ferromagnetic fibres (i.e. nickel-free ferritic stainless steel) in a non-magnetic matrix located in inter-fibre space, evidencing the possibility of changing the shape of the magnetized fibre array, straining the matrix and inducing a possible mechanism for bone growth stimulation by magnetic field application. Mannix et al. [27] focused their attention on the nanomagnetic actuation of receptor-mediated signal transduction, describing a magnetic nanotechnology that activates a biochemical signalling mechanism normally switched on by binding of multivalent chemical ligands. However, an interesting approach related to a selective activation of mechanosensitive ion channels using magnetic particles has also been reported by Hughes et al. [28]. This technique should permit the direct manipulation of ion channels in real time without the need for pharmacological drugs, and should be potentially considered as a tool for the treatment of human diseases ascribed to ion channel dysfunction [28].

Interestingly, Kanczler et al. [29] studied the controlled differentiation of human bone marrow stromal cells using remote magnetic field activation and MNPs. In the field of tissue engineering, novel methodologies, defined as ‘magnetic force-based tissue engineering’, and techniques for designing tissue-engineered tubular and sheet-like constructs using magnetite NPs and magnetic force were proposed and studied by Ito et al. [30,31]. It has also been reported that magnetic forces enable rapid endothelialization of a knitted Dacron graft externally covered by a magnetic sheet, benefiting from biophysical forces able to attach blood-derived endothelial outgrowth cells (EOCs) to the surface of prosthetic vascular grafts, because EOCs endocytose magnetic particles and as a result are attracted to the magnetized graft surfaces [32]. The basic principles in designing a novel magnetic force mechanical conditioning bioreactor for tissue engineering were proposed by Dobson et al. [33]. Furthermore, by binding MNPs to the surface of cells, the possibility to manipulate and control cell function through the application of an external magnetic field has been studied, providing information on cellular mechanics and ion channel activation [34]. This technique has been proposed as an investigative potential tool for actively controlling cellular function and processes with a special focus towards tissue engineering and regenerative medicine [34]. Magnetically actuable tubular scaffolds for smooth muscle tissue engineering were firstly fabricated from sheets of electrospun fibrils of biocompatible and biodegradable polymers containing uniform dispersions of Fe₂O₃ NPs and then wound into tubes for cell seeding from the inner layer to the outer one [35]. As a consequence of the magnetic field application, the induced deformation produces strains in the tube walls and fluid pumping through the walls which should promote cell proliferation and differentiation. The design of these tubular scaffolds was optimized through a model used to predict the deformation and fluid flow for specific magnetic field strength, material properties and geometrical parameters [35]. Furthermore, involving direct magnetic cell-seeding procedures and MNPs, novel strategies for vascular tissue engineering were proposed by Perea et al. [36] and Shimizu et al. [37]. First, the idea to develop magnetic scaffolds for additionally controlling angiogenesis in vivo has been considered by Bock et al. [16]. In that work, magnetic scaffolds were manufactured through dip coating of the scaffolds in aqueous ferrofluids that contained iron oxide nanoparticles coated with different polymers for biomedical applications [16]. From a magnetic point of view, nanoparticles below 30 nm in size of these materials present superparamagnetic behaviour, stressing their ability to be magnetized by applying a magnetic field without remanence once the field is turned off [16,20,38].

A magnetic scaffold should modify the external magnetic flux distribution, thus causing a much higher concentration of magnetic flux in the vicinity of and inside the scaffold [16]. Consequently, the magnetic scaffolds may generate higher magnetic field gradients that are able to provide significant magnetic attractive forces.

Taking into consideration a superparamagnetic material, the resulting magnetic scaffold may be able to reach appropriate magnetization values (i.e. up to 15 emu g⁻¹ at 10 kOe) for ferrofluid or MNPs adhesion when applying an external magnetic field as reported by Bock et al. [16], but it may also be magnetically ‘turned off’ by removing the applied magnetic field. This point is crucial as these magnetization values can generate magnetic gradients and via magnetic driving scaffolds may attract and take up cells or other bioagents bound to MNPs and in vivo growth factors.

Benefiting from this approach and from rapid prototyping techniques, the possibility to design magnetic polymer-based nanocomposite scaffolds by embedding superparamagnetic PVP-coated Fe₃O₄ nanoparticles in a PCL matrix has already been shown [39]. Their biological and mechanical performances have been preliminarily assessed [39].

These structures can be seen as a fixed ‘station’, whose magnetization can be switched on and off by means of external magnetic fields, thus providing a programmed biofactor release.

Even though iron-oxide-based phases such as maghemite or magnetite have been widely considered, their long-term effects in the human body remain still unclear [40,41].

Clearly, the development of suitably surface-modified MNPs through the design of specific biocompatible layers consisting of polymers, inorganic phases or metals deposited on their surface should avoid the problems related to their eventual toxicity [38,42].

By stressing the importance of having non-toxic MNPs for all the above-mentioned applications, Tampieri et al. [38] synthesized a novel biocompatible and bioresorbable superparamagnetic-like phase (FeHA) by doping hydroxyapatite with Fe³⁺/Fe²⁺ ions, minimizing the formation of magnetite as secondary phase. Microstructural, physico-chemical and magnetic analyses were carried out on the nanoparticles, highlighting their intrinsic magnetization and suggesting new perspectives for devices for bone tissue engineering and for anti-cancer therapies based on hyperthermia.

In this study, the mechanical, biological and magnetic performances of the proposed PCL/FeHA nanocomposite substrates were properly evaluated.

2. Material and methods

2.1. Iron-doped hydroxyapatite nanoparticles

2.1.1. Synthesis method

FeHA nanoparticles were prepared according to the method previously described [38].

A basic suspension of calcium hydroxide Ca(OH)₂ (Aldrich, 95 wt% pure, 50 g in 400 ml of H₂O) was stirred and heated to 40°C. FeCl₂ · 4H₂O (Aldrich, greater than or equal to 99 wt% pure, 12.74 g in 75 ml of H₂O), and FeCl₃ · 6H₂O (Aldrich, 97 wt% pure, 17.86 g in 75 ml of H₂O) solutions were simultaneously added to the basic suspension as sources of Fe²⁺ and Fe³⁺ ions. The total amounts of iron ions with respect to calcium ions were adjusted so as to obtain: Fe/Ca = 20 mol%.

Soon after, a phosphoric acid (Aldrich, 85 wt% pure, 44.40 g in 300 ml of H₂O) solution was added drop-wise into the basic suspension of calcium hydroxide containing iron ions, over a period of 2 h, under constant heating and stirring.

The reaction products were kept in suspension by constant stirring and heating for 1 h, and then left ageing for 24 h at room temperature without further stirring. The precipitate was separated from mother liquor by centrifugation, and then washed with distilled water and centrifuged three times; finally, it was freeze-dried and sieved at 150 µm.

2.2. Magnetic poly(ε-caprolactone)/iron-doped hydroxyapatite nanocomposite substrates

2.2.1. Preparation

Magnetic nanocomposite substrates were produced by embedding FeHA nanoparticles into a PCL matrix.

PCL (M_w = 65 000, Aldrich) pellets were dissolved in tetrahydrofuran (THF) under stirring at room temperature.

FeHA nanoparticles having a diameter lower than 20 nm were added to the PCL/THF solution during stirring, and three different polymer-to-particle weight ratios (w/w) (90/10, 80/20 and 70/30 w/w) were used. An ultrasonic bath (Branson 1510 MT) was also used to optimize the nanoparticles dispersion in the polymer solution. Then, moulding and solvent-casting techniques were used to manufacture miniature disc-shaped specimens (figure 1) with a diameter of 6.4 mm and a thickness of 0.5 mm.

Figure 1.

Schematic of moulding and solvent-casting techniques used to obtain polymeric and nanocomposite disc-shaped specimens. (Online version in colour.)

2.2.2. X-ray diffraction

PCL/FeHA nanocomposite substrates were analysed through X-ray diffraction (XRD), to determine the phase composition and crystallinity. To this aim, a D8 advance diffractometer (Bruker, Karlsruhe, Germany) was used, and the nanocomposite substrates were scanned from 2θ = 10° to 2θ = 60° using Cu-Kα radiation.

2.2.3. Scanning electron microscopy/energy dispersive spectroscopy analysis

Scanning electron microscopy (SEM) was used to evaluate the morphology of the PCL/FeHA nanocomposites by using an FEI Quanta FEG 200 scanning electron microscope (The Netherlands).

Energy dispersive X-ray spectroscopy (EDS) was used to determine whether the small particles observed in the nanocomposites were FeHA nanoparticles or not, and SEM–EDS P-, Ca- and Fe-mapping was applied to assess the dispersion of FeHA nanoparticles in the matrix of the nanocomposites.

2.2.4. Small punch tests

Small punch tests were carried out on PCL/FeHA disc-shaped specimens with a diameter of 6.4 mm and a thickness of 0.5 mm according to the ASTM F 2183 standard, in order to evaluate maximum load and displacement at maximum load. All the tests were performed using an INSTRON 5566 testing machine.

The experimental set-up consisted of a die, a hemispherical head punch and a guide for the punch. The specimen was loaded axisymmetrically in bending by the hemispherical head punch at a constant displacement rate of 0.5 mm min⁻¹ until failure occurred. The values of load and displacement of the punch were recorded continuously while performing the test.

2.2.5. Micro-computed tomography

A micro-computed tomography (micro-CT) was performed on disc-shaped PCL/FeHA specimens using a SkyScan 1072 system (Aartselaar, Belgium). A rotational step of 0.9° over an angle of 180° was used in order to visualize the morphological features and surface topography. Cross-sections and three-dimensional models of PCL/FeHA nanocomposite substrates were reconstructed using SkyScan's software package, ImageJ software, Materialise Mimics and Rapidform 2006 that allowed for analysing the results from a micro-CT system scan.

2.2.6. Contact angle measurements

Contact angle measurements were performed on PCL and PCL/FeHA substrates by using a DataPhysics OCA 20 apparatus. Briefly, distilled water was dropped onto at least five different sites on each specimen, and the static contact angle was measured. Results were reported as mean value ± s.d.

2.2.7. Magnetic analysis

Magnetization measurements were performed by using a superconducting quantum interference device (SQUID) magnetometer. This instrument measures the total magnetic moment of a sample, including all atomic and molecular magnetic contributions. Owing to size restrictions, a small part of the sample material was carefully cut away from the scaffolds, and then fixed in a specially designed sample holder, which allows for cancelling background contributions to the total magnetic moment. During the measurements, the magnetic field in the superconducting coil was either held constant at varying temperatures or was swept at constant temperature, whereas the samples were consistently moved through a pick-up coil system connected to the SQUID via a flux transformer. Magnetization data were taken at temperatures 5 K < T < 350 K using a liquid-helium cooled variable-temperature insert installed in the commercial SQUID-magnetometer set-up (MPMS, Quantum Design Inc., San Diego, CA, USA). In order to scale the measured magnetic moments to the amount of substance, the weight of the sample was determined with great care.

On the other hand, magnetic hyperthermia characterizations were performed applying a radiofrequency (RF) magnetic field to the samples, which were placed in a thermally isolated Teflon holder, whereas temperature was simultaneously recorded with an optical fibre thermometer. Both frequency and amplitude of the oscillating magnetic field (f = 260 kHz and 27 mT, respectively) were generated with a home-made alternating current source feeding a refrigerated copper coil. The hyperthermia device has been designed to be within the safety-specific absorption rate range for in vivo applications.

2.2.8. Cell adhesion study

Human mesenchymal stem cells (hMSCs, 1 × 10⁴ cells per sample) were seeded on PCL and PCL/FeHA substrates and grown in Dulbecco's modified Eagle's medium (DMEM) without foetal bovine serum (FBS).

The different kinds of cell constructs (PCL-hMSCs and PCL/FeHA-hMSCs) were analysed using confocal laser scanning microscopy (CLSM). They were fixed with 4 per cent paraformaldehyde, rinsed twice with phosphate-buffered saline (PBS) buffer and incubated with PBS–BSA (0.5%). Actin microfilaments were stained with phalloidin–tetramethylrhodamine B isothiocyanate (Sigma-Aldrich). Phalloidin was diluted in PBS–BSA (0.5%) and incubated at room temperature for a suitable time. The images of cell constructs were acquired by using a He–Ne excitation laser at the wavelength of 543 nm and an objective of 20 times.

2.2.9. AlamarBlue assay

Bone-marrow-derived hMSCs (Clonetics, Italy) were maintained at 37°C and 5 per cent CO₂ in DMEM supplemented with 10 per cent FBS (BioWhittaker, Walkersville, MD, USA), 2 mM l-glutamine (Sigma, St Louis, MO, USA), 1000 U l⁻¹ penicillin (Sigma) and 100 mg l⁻¹ streptomycin (Sigma).

Disc-shaped PCL/FeHA substrates were prepared for cell seeding by soaking first in 70 per cent ethanol for 1 h, then in 1 per cent antibiotic/antimycotic in PBS for 2 h and prewetted in medium for 2 h.

Cells were statically seeded onto the PCL/FeHA nanocomposite substrates using a density of 1 × 10⁴ cells per sample.

Cell viability and proliferation were evaluated by using the AlamarBlue assay. It is based on a redox reaction that occurs in the mitochondria of the cells; the coloured product is transported out of the cell and can be measured through a spectrophotometer.

In particular, at 7, 14 and 21 days after seeding, the cell constructs were rinsed with PBS (Sigma-Aldrich, Italy), and for each sample, 200 μl of DMEM without Phenol Red (HyClone, UK) containing 10 per cent (v/v) AlamarBlue (AbD Serotec Ltd, UK) was added, followed by incubation in 5 per cent CO₂ diluted atmosphere for 4 h at 37°C.

A specific volumetric amount of solution was then removed from the wells and transferred to a new 96-well plate. The optical density was immediately measured using a spectrophotometer (Sunrise, Tecan; Männedorf, Zurich, Switzerland) at wavelengths of 570 and 595 nm. The number of viable cells correlates with the magnitude of dye reduction [11,43,44] and it is expressed as a percentage of AlamarBlue reduction according to the manufacturer's protocol. However, each test was repeated at least three times in triplicate.

2.2.10. Alkaline phosphatase/DNA assay

Samples were removed from the medium and washed twice with PBS on days 7, 14 and 21. The substrates were then submerged into 1 ml of lysis buffer. A cell density of 1 × 10⁴ cells per sample was used. ALP activity was measured using a specific biochemical assay.

The substrates were then centrifuged, and the supernatant was used to calculate the alkaline phosphatase (ALP) activity by the p-nitrophenylphosphate (p-NPP) method (SensoLyte pNPP alkaline phosphatase assay kit). ALP/DNA was then reported by using the Quant-iT PicoGreen assay kit that enables detection and quantification of DNA. Cultures were performed in both standard and osteogenic differentiation medium (+OM, Sigma-Aldrich).

3. Results and discussion

Magnetic activation results in an interesting strategy that has been previously proposed to answer the increasing need for assisted bone and vascular remodelling [38].

The idea to design biodevices that should be biologically manipulated or activated in situ by applying an external magnetic field results in a great challenge in tissue engineering.

As reported in the literature, a new class of magnetic hydroxyapatites has already been synthesized to avoid the problems of toxicity that are usually related to iron-oxide-based phases. Their synthesis involves a neutralization method used to synthesize HA nanopowders, and a partial substitution of Ca by Fe²⁺ and Fe³⁺ ions that can enter the HA lattice, as highlighted in a previous work [38]. In particular, by simultaneously adding both Fe species under specific and controlled synthesis conditions, it has been possible to obtain an FeHA that is characterized by a (Fe–Ca)/P ratio very similar to the theoretical ratio (Ca/P = 1.67), Fe³⁺/Fe²⁺ ratio of about three and a negligible amount of magnetite as secondary phase.

The biocompatibility and the intrinsic magnetism of these FeHA nanoparticles have been properly assessed by Tampieri et al. [13], proposing the possibility to potentially use magnetic hydroxyapatites for developing magnetic ceramic scaffolds with improved properties for bone regeneration, thus opening novel frontiers in the field of regenerative medicine [38]. A superparamagnetic-like behaviour of single-domain MNP has been evidenced by the magnetization curves as a function of the applied magnetic field [38].

Accordingly, trying to benefit from the biocompatibility/degradability of superparamagnetic FeHA nanoparticles that should overcome the side effects of long-term cytotoxicity, an interesting idea has been to develop nanocomposite substrates for bone tissue engineering by embedding FeHA nanoparticles into a PCL matrix.

XRD analysis has confirmed that the synthesis process used to realize such PCL/FeHA magnetic substrates does not modify the structure and the crystallinity of the magnetic FeHA. The XRD analysis performed on the PCL/FeHA (70/30) composite has evidenced the presence of peaks ascribed to the organic (PCL) phase and also to the inorganic (FeHA) phase that shows the same features of the starting FeHA powder (figure 2) [38].

Figure 2.

XRD spectra relative to PCL, FeHA and the composite containing the biggest amount of magnetic phase (PCL/FeHA 70/30).

With regard to the nanoparticle distribution, as an example, the SEM and EDS images of the PCL/FeHA (80/20) composite, as well as the SEM–EDS P-, Ca- and Fe-mapping photographs are presented in figures 3–5. It can be seen that the NPs and aggregates are observed in the nanocomposites and distributed uniformly in the matrix in general.

Figure 4.

EDS image of PCL/FeHA (80/20). (Online version in colour.)

Figure 3.

SEM image of PCL/FeHA (80/20).

Figure 5.

SEM–EDS P-, Ca- and Fe-mapping photographs of PCL/FeHA (80/20).

As a first step towards the mechanical characterization, the small punch test has been chosen to assess the performances of the proposed disc-shaped PCL/FeHA nanocomposite substrates, as it may be considered as a reproducible miniature specimen test method. This test method has already been taken into consideration to evaluate the mechanical properties of retrieved acrylic bone cement, and PCL reinforced with sol–gel synthesized organic–inorganic hybrid fillers [11,45].

Results from small punch tests on PCL and PCL/FeHA substrates have shown load–displacement curves generally characterized by an initial linear trend, followed by a decrease in the curve slope until a maximum load is reached. Finally, it is evident that a decrease in the load until failure has occurred for all specimens (figure 6). Maximum load and displacement at maximum load obtained from PCL and PCL/FeHA substrates are reported as mean value ± s.d. in table 1.

Figure 6.

Load–displacement curves obtained from small punch tests performed on neat PCL and PCL reinforced with FeHA nanoparticles.

Table 1.

Results from small punch tests: maximum load and displacement at maximum load are reported as mean value ± s.d.

materials	maximum load (N)	displacement at maximum load (mm)
PCL	15.30 ± 1.30	1.60 ± 0.20
PCL/FeHA (90/10)	22.51 ± 0.60	2.40 ± 0.45
PCL/FeHA (80/20)	12.19 ± 0.57	1.32 ± 0.12
PCL/FeHA (70/30)	10.27 ± 1.05	1.12 ± 0.18

Figure 6 and table 1 highlight that the inclusion of 10 per cent by weight of FeHA nanoparticles represents an effective reinforcement in terms of higher maximum load, providing mechanical performances that are better than those obtained for the neat PCL substrates and the other compositions of nanocomposite.

In particular, PCL/FeHA (90/10 w/w) substrates have provided higher values of maximum load (22.51 ± 0.60 N) and displacement at maximum load (2.40 ± 0.45 mm) in comparison with neat PCL and other nanocomposites, thus resulting in substrates that are stronger but flexible and tough at the same time. However, it is worth noting that PCL substrates show values of maximum load (15.30 ± 1.30 N) that are greater than those achieved by PCL/FeHA (80/20 and 70/30 w/w).

It is well known that weakness in the structure may be clearly due to discontinuities in the stress transfer and generation of stress concentration at the nanoparticle/matrix interface, which may be ascribed to the difference in ductility between the polymeric matrix and the inorganic nanofillers.

Our experiments suggest that beyond a specific limit of nanoparticle amount, by further increasing the nanoparticle concentration, the mechanical performances of the nanocomposite substrates decrease because the nanoparticles act as ‘weak points’ instead of reinforcement for the polymeric matrix.

Micro-CT analysis has enabled three-dimensional reconstructions of the PCL and PCL/FeHA substrates, qualitatively evidencing the morphological features, surface topography and the presence of eventual defects (i.e. clusters, voids and similar defects; figure 7).

Figure 7.

Results obtained from micro-CT analysis: reconstruction of (a) PCL/FeHA (90/10 w/w) and (b) PCL/FeHA (70/30 w/w) obtained by integrating Skyscan's software package, ImageJ software, Materialise Mimics and Rapidform 2006. (Online version in colour.)

The hydrophilicity of the manufactured PCL/FeHA substrates was examined by measuring the contact angle, and the results were compared with those obtained from the neat PCL ones (figure 8 and table 2).

Figure 8.

Typical image qualitatively representing the water contact angle. (Online version in colour.)

Table 2.

Water contact angles reported as mean value ± s.d. for PCL and PCL/FeHA substrates.

materials	water contact angle, θ (°)
PCL	81.4 ± 4.4
PCL/FeHA (90/10)	75.7 ± 4.6
PCL/FeHA (80/20)	74.8 ± 2.6
PCL/FeHA (70/30)	64.9 ± 8.2

The water contact angle of PCL/FeHA nanocomposite substrates is lower than that of neat PCL ones, thus indicating that the presence of FeHA nanoparticles embedded in the polymeric matrix makes the surface more hydrophilic. In particular, the higher the amount of the FeHA nanoparticles, the lower the water contact angle; its values basically span from 81.4° for PCL substrates to 64.9° for PCL/FeHA (70/30 w/w).

All of the earlier-mentioned results might be ascribed to the synergetic contribution of both surface chemistry and topography that can be clearly varied by including the FeHA nanoparticles.

However, it is well known that PCL is a hydrophobic polymer and the water contact angle measured for PCL substrates should be greater than 90°. In contrast to this, with regard to PCL substrates, table 2 reports a water contact angle of 81.4 ± 4.4°. This result may be related to the specific techniques (moulding and solvent-casting) used to manufacture the substrates that obviously alter the surface topography and roughness, thus reducing the value of the water contact angle.

The approach to design magnetic scaffolds for tissue engineering clearly rises from the challenging idea of guiding tissue-regeneration process benefiting from a magnetic field.

As already specified, even though the magnetic guiding process is already well known in nanomedicine (i.e. drug delivery and hyperthermia treatment of tumours), this concept is not yet used in the field of scaffolds for tissue engineering. For this reason, as a first step, preliminary magnetic measurements have been carried out in terms of magnetization and susceptibility. In particular, three different measurement modes have been used to perform a magnetic characterization of the samples: measurements of the field dependence of the magnetization M (H, T) at body temperature (T = 310 K), measurements of its temperature dependence in a small field (H = 50 Oe) and measurements of the frequency dependence of the magnetic susceptibility χ(f) at T = 310 K. The results of these measurements confirmed the superparamagnetic character of the FeHA nanoparticles in the samples indicated by a very low coercive field, a defined saturation magnetization and a strong history dependence in temperature sweeps.

The results of the field-dependent magnetization measurements are depicted in figure 9. Independent of their FeHA content, the coercive field of all samples took a value of 15 Oe at body temperature, indicating vanishingly small interactions between the MNPs. The saturation magnetization values were found to be strictly proportional to the FeHA content. The magnetization curves are superimposed by a diamagnetic background originating from the PCL and FeHA content of the samples. This diamagnetic background can be roughly ascribed to a value of χ_PCL, which is equal to −1.46 × 10⁻⁷ emu (g %w_PCL Oe)⁻¹ at 310 K.

Figure 9.

Field-dependent magnetization of the three PCL/FeHA compositions investigated in the article. Each curve was taken at T = 310 K (human body temperature). The coercive field is approximately 15 Oe, independent of the composition. Details are given in the text.

The temperature dependence of the magnetization of one sample (PCL/FeHA 90/10) is shown in figure 10. The results for the other compositions investigated for this study are similar to those given in figure 10. The data were taken in a zero-field-cooling–field-heating–field-cooling sequence in order to trace the polarization process. Within the temperature range shown, the field-heating and field-cooling curves do not seem to overlap, making it difficult to determine the blocking temperature unambiguously. From the temperature value at the maximum of the field-heated curve, we derive an effective MNP-core diameter of d_eff = 13 nm [46].

Figure 10.

Temperature dependence of the magnetization of the PCL/FeHA (90/10) sample. The sample was cooled in zero-field, then heated in a small field (50 Oe—fh branch) and again cooled in field (fc branch). Details are given in the text.

The frequency-dependent susceptibility is an important indicator for the applicability of these materials to hyperthermia treatment methods. All measurements discussed here have been taken at a background field much smaller than 5 Oe. The results for the sample of PCL/FeHA (90/10) are shown in figure 11, again representing the other compositions that showed similar results. At a fixed excitation and 310 K, χ′(f), which is the real part of χ(f), takes a constant value of 6 × 10⁻³ emu (g Oe)⁻¹ within the frequency range 0 Hz < f < 1500 Hz, increasing linearly with the excitation amplitude. The imaginary part χ″(f) is approximately zero below f = 1200 Hz. For frequencies above 1200 Hz, increases approximately linearly with frequency, reaching a value of 7 × 10⁻⁴ emu (g Oe)⁻¹ at 1500 Hz.

Figure 11.

Frequency dependence of the (a) real (b) and imaginary parts of the magnetic susceptibility of PCL/FeHA (90/10). Details are given in the text.

This increase in χ″(f) can be interpreted as an onset of a maximum of dissipation, which is due to thermally agitated directional fluctuations rather than to ferromagnetic resonance [47].

Regarding the heating properties of the samples, figure 12 shows the temperature increase in the PCL/FeHA magnetic composites under an alternating magnetic field of 27 mT at a frequency of f = 260 kHz. First of all, it can be observed that all the samples show a magnetically induced thermal response, owing to the energy released through the Néel relaxation process, which is the only mechanism contributing for superparamagnetic nanoparticles embedded in a rigid solid. Significant temperature increases between 2 and 10 K were achieved after 5 min of exposure to an external magnetic field for all the PCL/FeHA compositions. A progressive increase in the heating rate is observed, as the amount of magnetic FeHA nanoparticles increases in the composite.

Figure 12.

Hyperthermia curves of the PCL/FeHA magnetic scaffolds under application of a RF magnetic field of f = 260 kHz and H = 27 mT, suitable for in vivo applications. (Online version in colour.)

These results provide PCL/FeHA scaffolds with unique properties to be used in in vivo applications, because their functionality can be remotely fine-tuned by controlling both the amount of MNP concentration and the time of the magnetic field exposure.

The biocompatibility of the PCL/FeHA nanocomposite substrates has been studied in vitro by using hMSCs. The results of the in vitro study are reported in figures 13–15.

Figure 13.

Cell adhesion study: CLSM images at different times after cell seeding. (i) From top to bottom, PCL, PCL/FeHA 90/10, PCL/FeHA 80/20 and PCL/FeHA 70/30 at 7 days after cell seeding. (ii) From top to bottom, PCL, PCL/FeHA 90/10, PCL/FeHA 80/20 and PCL/FeHA 70/30 at 14 days after cell seeding. (iii) From top to bottom, PCL, PCL/FeHA 90/10, PCL/FeHA 80/20 and PCL/FeHA 70/30 at 21 days after cell seeding. Scale bar, 100 µm.

Figure 15.

Results obtained from ALP/DNA assay at 7, 14 and 21 days after cell seeding. Error bar represents the s.d.

First, CLSM analyses performed on all the cell constructs have qualitatively provided interesting results in terms of hMSCs adhesion and spreading at 7,14 and 21 days after seeding (figure 13).

Figure 13 highlights the cell cytoskeleton organization over the time. In particular, hMSCs were well spread and better adhered on PCL/FeHA nanocomposites in comparison with cells seeded on PCL substrates, as qualitatively suggested by actin cytoskeleton staining. In addition, an increase in the adhered number of hMSCs is evident in the case of FeHA nanocomposites.

On the other hand, the AlamarBlue assay has provided information on cell proliferation and viability over the culture time through a quantitative evaluation of the percentage of AlamarBlue reduction for the substrates. The results are graphically shown in figure 14, where they are reported as mean value and error bars represent standard deviation.

Figure 14.

Results obtained from AlamarBlue assay at 7, 14 and 21 days after seeding. Error bar represents the s.d. *p < 0.05; **p < 0.01; ***p < 0.001, indicate statistically significant differences between nanocomposite and PCL substrates, at the same time from cell seeding (one-way ANOVA followed by Tukey's post hoc test).

The results obtained from the AlamarBlue assay have evidenced that hMSCs were viable on both PCL and PCL/FeHA substrates over time, as the percentage of AlamarBlue reduction increases with time. In particular, even though at 7 days after seeding, there is no great difference among the PCL and PCL/FeHA substrates, some differences may be noticed at 14 and 21 days.

In particular, PCL/FeHA (70/30 and 80/20 w/w) substrates have provided higher values of the percentage of AlamarBlue reduction at 21 days after seeding.

It seems that especially the results obtained at 21 days are correlated with the values of water contact angle. The inclusion of FeHA nanoparticles intrinsically enhances the hydrophilicity and modifies the substrate topography, thus favouring cell viability and proliferation.

As CLSM analyses and AlamarBlue assay have allowed for obtaining qualitative and quantitative information on cell adhesion and viability/proliferation, respectively, ALP (ALP/DNA) measurements were used to assess the osteogenic expression of hMSCs as ALP is an early marker for the osteogenic differentiation of cells. As previously stated, ALP activity was measured using a specific biochemical assay. The osteogenic differentiation of the hMSCs cells has been assessed by normalized ALP activity and reported as mean value and standard deviation over time (figure 15).

The significant increase in the ALP activity from day 7 to 14 clearly suggests that cells were induced to differentiate during this period, thus evidencing the ability of both polymeric and nanocomposite substrates to support the osteogenic differentiation of hMSCs. The decrease in the ALP activity from day 14 to 21 is also consistent with that reported for the MG63 cells seeded on hydroxyapatite/collagen nanocomposite sponges and ascribed to the expression of later-stage osteogenic markers and to the begin of calcium deposition [48].

However, especially for PCL substrates, the effect of the osteogenic medium is well evident; ALP activity of the osteogenically induced hMSCs results higher than that obtained from cultures under standard conditions. Furthermore, taking into account the values of ALP activity reported in figure 15, it may be noticed how, under standard culture conditions, the nanocomposite PCL/FeHA substrates seem to better promote the osteogenic differentiation than the PCL ones. These results are also consistent with those previously obtained from quantitative biological analyses (CLSM analysis and AlamarBlue assay), stressing the importance of how the chemistry of FeHA nanoparticles and/or the different surface topography and roughness of the nanocomposites can enhance cell adhesion.

4. Conclusions

This study may be considered as a first step of a future complex work with the aim of designing three-dimensional magnetic nanocomposite scaffolds for bone tissue engineering, benefiting from the eventual effect of the magnetic field on the tissue-regeneration process.

For this reason, nanocomposite substrates were first designed by embedding FeHA nanoparticles into a PCL matrix.

XRD analysis has been used as a tool to demonstrate that the process used to make the substrates does not affect the structure and crystallinity of the magnetic FeHA, whereas the effect of nanoparticle inclusion on the mechanical performances of the substrates has been evaluated through small punch tests.

On the other hand, a magnetic characterization has been carried out to assess the magnetization as a function of field and temperature, as well as the frequency dependence of the susceptibility. In addition to the magnetic performances of the nanocomposites, the thermal activity obtained by magnetic hyperthermia experiments reveals this material to be extremely useful for remotely controlled in vivo applications.

The biological performances of the nanocomposite substrates have been properly evaluated through AlamarBlue and ALP/DNA measurements as well as through CLSM. In particular, CLSM analyses and AlamarBlue assay have provided qualitative and quantitative information on hMSC adhesion and viability/proliferation, respectively, whereas the obtained ALP/DNA values have shown the ability of the nanocomposite substrates to support the osteogenic differentiation. Furthermore, it is worth noting how the results obtained from biological analyses are correlated with the values of water contact angle.