3.1 Preparation of colloidal SPIONs suspensions

SPIONs were prepared by exploiting the method of thermal decomposition of iron (III) oleate as an iron-containing precursor [33]. After synthesis, structural characterisation of the oleic acid-stabilised nanoparticles (SPIONs-OA) was performed. The XRD pattern (Fig. S2A) showed multiple peaks consistent with the inverse spinel structure of magnetite [43] and TEM analysis (Fig. S2B) confirmed a uniform nanoparticle size, with dimensions of about 11 nm. The nanoparticles were further processed by exploiting a ligand exchange of the oleic acid surface layer with DMSA to obtain a stable suspension of nanoparticles in water (SPIONs-DMSA) [15, 33]. The FTIR spectrum (Fig. S4A) confirmed the success of the ligand exchange. Next, SPIONs-DMSA nanoparticles were covalently functionalised with an Arg-Gly-Asp (RGD) peptide via amide bonds to obtain the SPIONs-DMSA-RGD suspension (Fig. S3). For this purpose, the carboxylic acid groups present in the DMSA coating of the SPIONs were activated through EDC/sulfoNHS chemistry and reacted with the primary amine of the RGD sequence. The functionalisation was verified by FTIR, DLS (Fig. S4A-B) and chemical microanalysis (Table S1) [34]. The magnetization curves of the synthesised SPIONs showed absence of hysteresis loop, therefore confirming their superparamagnetic behaviour (Fig. S2C and S4C).

3.2 Scaffold design and fabrication

The combination of the FDM printing and thermal DOD inkjet of SPIONs was achieved thanks to the calibrated and accurate printing platform (see Supplementary Video S1.mp4 and Video S2.mp4). Scaffolds were 3D printed with a 3-polymer (3-P) blend filament composed of PLLA/PCL/PHBV (90/5/5 wt%). Biocompatibility and in vitro osteogenic potential of this polymer blend was previously reported [31, 32]. Each of the printed scaffolds (n = 300) was manually measured with a digital calliper in diameter and height, giving an average diameter and height of, respectively, 4.98 mm and 0.63 mm, and thus, a relative error of 0.4% and 5%. The accuracy (computed as the standard deviation of the difference between the correct diameter/height, 5 mm/0.6 mm), was ± 60 µm in diameter and ± 20 µm in height. Globally, it can be stated that the accuracy of 3D printed scaffolds was ± 40 µm whereas, for comparison, that of commercial FDM printers is ± 100 µm [44].

Two different SPIONs solutions were tested for the thermal DOD inkjet printing: (i) colloidal suspension of DMSA-coated SPIONs (SPIONs-DMSA) and (ii) colloidal suspension of SPIONs functionalised with an RGD peptide (SPIONs-DMSA-RGD). Currently, our research group have demonstrated the efficacy of these free-magnetic nanoparticles when exposed to the application of an external magnetic field for osteogenic purposes in hBM-MSCs. The results derived from this research indicates that remotely activated mechanotransduction using integrin-targeted magnetic colloidal SPIONs provokes a significant increase in runt-related transcription factor 2 (Runx2) and alkaline phosphatase (ALP) gene expression as well as ALP activity compared with non-targeted SPIONs (SPIONs-DMSA). In fact, a significant osteogenic differentiation is observed when the magnetic stimulus is applied when the nanoparticles encounter the cell membrane surface to initiate endocytic pathways. Then, in this manuscript both types of nanoparticles have been incorporated into the FDM devices to study the possible effect of integrin-targeting functionalisation on nanoparticles once immobilized in the scaffolds. The 2D pattern was repeatedly inkjet on the same scaffold outer surface 25 and 50 times, for each of the two different SPIONs-loaded solutions. From now on, polymer scaffolds without nanoparticles will be referred to as SC, scaffolds with 25 printing layers of SPIONs-DMSA as DMSA 25L, scaffolds with 25 printing layers of SPIONs-DMSA-RGD as RGD 25L, scaffolds with 50 printing layers of SPIONs-DMSA as DMSA 50L and scaffolds with 50 printing layers of SPIONs-DMSA-RGD as RGD 50L.

3.3 Characterisation of fabricated scaffolds

The morphology of the printed scaffolds was evaluated by SEM (Fig. 2A-E), which showed a generally smooth surface of the material with areas of increased roughness at the edge between one printed line and another. The filament width of the scaffolds was measured with ImageJ software [45] which corresponded to 0.58 ± 0.03 µm (average over n = 4 3D printed scaffolds). Filaments appeared well stacked with no visible pores on the surface, as expected (infill equal to 100%, Table 1). At low magnification (Fig. 2B-E), the achievement of the intended inkjet 2D tetragonal pattern was observed in the circular scaffold surface. At higher magnification in the SEM images (Fig. 2G-J), homogeneous deposits of SPIONs following the 2D pattern was observed, with neatly distributed spots over the entire circular surface. The "brighter" SEM intensity correlates with a higher average atomic number in the sample, and "dark" areas have a lower average atomic number. In this case it is easy to attribute the areas of higher brightness in the images to deposits of SPIONs due to the difference between the atomic number of iron and carbon. To determine the differences between the samples a study by high resolution scanning microscopy using a JSM7600 microscope coupled to an EDS detector was carried out (Fig. S6). The EDS-mapping revealed the homogeneous distribution of iron (in red) which is concentrated in the spots resulting from the DOD process. There were some iron satellites outside the spots nonetheless being almost negligible. HRSEM studies showed also a very homogeneous size of the iron spots with slight increase in size for 50L samples.

Fig. 2

A, F SEM images corresponding to SC; B, G SEM images corresponding to DMSA; C, H SEM images corresponding to DMSA 25L; D, I SEM images corresponding to DMSA 50L; and E, J SEM images corresponding to 50L, showing their surface. A–E Images at low magnification. F–J SEM images at higher magnification. These SEM studies have been carried out in a microscope field emission JSM6335 with a back scattering electron detector. K Stability of the different scaffolds with SPIONs showing the initial amount of Fe and the amount of Fe remaining in the scaffolds after 7 days in 1 × PBS at 37 °C. Asterisks indicate significant differences with respect to 25L (p value < 0.05). L Magnetisation curves of the different scaffolds. M Magnification of the curves at low fields

Figure 2K shows the amount of iron on the surface of the scaffolds after production (initial), and the amount of iron remaining in the scaffolds after 7 days in physiological media, measured by ICP-AES. The results obtained for DMSA 25L and RGD 25L were 2.32 and 2.25 µg iron, respectively. The resulting amounts for DMSA 50L and RGD 50L were 4.45 and 4.54 µg respectively. The amount of SPIONs deposited was doubled by doubling the number of printed layers, indicating a high accuracy in the production technique and versatility in producing new materials with different amount of SPIONs. Furthermore, no significant differences were observed in the amount of SPIONs deposited between the SPIONs-DMSA and SPIONs-DMSA-RGD suspensions for the same number of layers. Likewise, no significant differences were observed in the amount of iron on the surface of the scaffolds after 7 days in 1 × PBS at 37 °C with respect to the initial content, indicating that the SPIONs deposits are very stable after soaking and remain on the surface.

The evaluation of the magnetic properties of the materials showed a diamagnetic behaviour (negative slope) of the control scaffold (SC, scaffold without SPIONs), therefore, initially in the magnetisation curves of the scaffolds with SPIONs two components were observed, a diamagnetic one attributed to the polymer scaffold and a superparamagnetic one attributed to the SPIONs. The magnetisation curves of the DMSA 25L, RGD 25L, DMSA 50L and RGD 50L scaffolds (Fig. 2L) were plot after correction of the diamagnetic component. The curves corresponding to the scaffolds with 50 printed layers of SPIONs showed a saturation magnetisation value double that observed in the curves corresponding to the scaffolds with 25 printed layers. This increase is due to the amount of SPIONs present in the scaffold, in line with the ICP-AES result and SEM–EDS studies (Fig. 2K) where approximately twice as much iron is observed in the scaffolds with 50 printed layers. With respect to the type of nanoparticle incorporated (SPIONs-DMSA or SPIONs-DMSA-RGD), no major differences are detected. When the scaffolds have 50 printed layers the curves are similar (DMSA 50L and RGD 50L), and when they have 25 printed layers, the magnetisation of DMSA 25L is slightly higher than that of RGD 25L, which could be due to a slight difference in the amount of SPIONs between the samples, but still the values are very similar. Furthermore, the magnetisation curves of both SPIONs alone and incorporated on the scaffolds showed similar shape and coercivity. As shown in the graphs, in both cases coercivity values below 3 mT were recorded, which for most biomedical applications of SPIONs can be considered as a non-hysteretic response, i.e., coercivity and remanence close to zero; confirming their superparamagnetic behaviour (see Fig. 2M and insets of Fig. S2C, S4C and S5). Finally, the maximum magnetisation values normalised to the Fe content calculated for the SPIONs alone (Fig. S4C) and the scaffolds with SPIONs (Fig. S5) were consistent, obtaining a value around 90 emu/gFe in all cases. Therefore, the similarity between the different cycles suggests that the magnetic properties of the SPIONs were not affected by their incorporation on the surface of the polymer scaffolds.

Finally, to obtain more information at nanoscale, a deep AFM study was carried out. The most significant results are shown in Fig. 3 and Figure S7. Two different areas of the scaffolds have been studied (as indicated by the position of the AFM cantilever in the photographic images in Fig. 3 left), one inside a magnetic spot (Fig. 3A) and one outside (on the polymeric surface of the scaffold, Fig. 3B). A scan of 5 µm × 5 µm has been carried out in all areas studied showing notables differences between the different areas. While in the interior of the spots an agglomerate of nanoparticles is observed with an increase in z-height, in the area outside, a quasi-smooth zone is observed, typical of the polymeric surface of the FDM scaffold. The roughness value (Ra) for the pristine FDM scaffolds is of 5.2 nm in the measured area while Ra values within the spots is increased with a value of 29.9 nm and 52.5 nm for the DMSA 25L and RGD 25L samples, respectively. Note that the Ra values measured in the outside area are in agreement with those obtained for the control sample (8.3 nm and 7.9 nm for DMSA-25L and RGD-25L). This slight increase could be due to the incipient deposition of nanoparticles outside the spot area observed by EDS mapping (Fig. S6). The increase in the Ra value in the zone inside the spot is due to the deposition of the nanoparticles (which are perfectly visible in the AFM images), which is greatest with the degree of aggregation of the starting nanoparticles described above [34].

Fig. 3

AFM images of the DMSA 25L and RGD 25L samples showing their surfaces. For all samples, an area of 5 µm × 5 µm was scanned. The photographic images of the camera attached to the AFM equipment (images A and B on the left) indicate the area where the measurement was taken. Images A centre and right: analysis of an area with SPIONs (within the SPIONs spots). Images B centre and right: analysis of an area without SPIONs (outside the SPIONs spots, i.e. on the polymeric surface of the scaffold)

3.4 In vitro biocompatibility

Cell adhesion, viability and proliferation are important biological parameters to assess the biocompatibility or cytotoxicity of any biomaterial. Cell adhesion to biomaterials is crucial for downstream cellular processes to occur. Cell proliferation, the process by which a cell grows and divides into two daughter cells, reflects the growth capacity of the cells in contact with the materials to be tested. Cell viability is a parameter related to the integrity of the cell membrane. Thus, adequate cell adhesion accompanied by maintenance of cell viability and allowing cell proliferation is a desired outcome to validate a biomaterial with potential for tissue regeneration. The proliferation of hBM-MSCs in contact with the scaffolds surface was studied at 4, 7, 10 and 14 days (Fig. 4A). Results show that hBM-MSCs are able to proliferate on the surface of the different scaffolds over time. No differences in cell proliferation were observed between the control scaffold and the scaffolds with SPIONs at any of the times studied. Regarding the viability of hBM-MSCs (Fig. 4B), the results indicated very high viability percentages in all cases studied (SC, DMSA 25L, RGD 25L, DMSA 50L and RGD 50L), with no significant differences between them (≥ 95% viability).

Fig. 4

A Proliferation at 4, 7, 10 and 14 days, B viability at 14 days and C intracellular ROS content at 14 days of hBM-MSCs cultured on SC, DMSA 25L, RGD 25L, DMSA 50L and RGD 50L. Asterisks indicate significant differences with respect to 4d (* p-value < 0.05). Hashtags indicate significant differences with respect to 7d (# p-value < 0.05). Ampersands indicate significant differences with respect to 10d (& p-value < 0.05 && p-value < 0.01). At symbols indicate significant differences with respect to SC (@ p-value < 0.05). D–M Confocal fluorescence microscopy images of hBM-MSCs cultured on the different scaffolds at 7 and 14 days. In red can be seen the actin filaments of the cytoskeleton labelled with phalloidin and in blue the nuclei labelled with DAPI

Several studies have proposed oxidative stress as a key mechanism involved in nanomaterials toxicity [46]. Overproduction of ROS leads to multiple adverse biological effects such as membrane lipid peroxidation, protein denaturation, mitochondrial dysfunction, lactate dehydrogenase leakage, and DNA and RNA damage, limiting the cellular ability to maintain normal physiological redox-regulated functions [47]. Moreover, several studies have highlighted the importance of measuring cellular ROS production in response to biomaterial surfaces to assess their in vitro biocompatibility [48]. In order to study possible oxidative stress in hBM-MSCs after interaction with the scaffolds surface, the intracellular ROS content was assessed after 14 days of culture (Fig. 4C). A decrease in intracellular ROS content was observed in cells cultured on the surface of scaffolds with SPIONs deposits compared to the control scaffold. These results are in agreement with other studies analysing the effect of SPIONs on hBM-MSCs and neural stem cells (NSCs) [14, 49], where contact with SPIONs also decreases the intracellular ROS content.

Adhesion, colonisation and cell morphology of hBM-MSCs cultured on the surface of the scaffolds were assessed by laser scanning confocal microscopy (Fig. 4D-M) and scanning electron microscopy (Fig. S8). Different parts of the scaffolds were scanned after 7 and 14 days of culture, using Phalloidin-Atto 565 as fluorescence probe for F-actin microfilaments and DAPI for nuclei. The images showed that hBM-MSCs presented their typical spindle-shaped morphology, adhered perfectly to the surface of the scaffolds and colonised the entire surface without leaving gaps. No differences were observed between the scaffolds with SPIONs and the control scaffold, neither between the type of SPIONs (DMSA or RGD) deposited, nor in the number of printed layers (25L or 50L). In addition, a higher number of cells is discernible in the 14-day images compared to the 7-day images, which is consistent with the cell proliferation results. The results showed a complete colonisation forming an adequate cellular lattice regardless of the presence of SPIONs. This result is confirmed by SEM images (Fig. S8).

In addition, another SEM analysis of the scaffolds colonised by the hBM-MSCs was carried out. In this case, prior to observation, several areas of the surface of the samples were punctured with tweezers to detach the cell monolayer and expose the surface of the scaffold. The images showed deposits of SPIONs under the cell monolayer (Figure S9), confirming their permanence on the scaffold surface after 14 days of culture with hBM-MSCs, thus ensuring long-term contact between the SPIONs and the cell surface.

Taken together all these in vitro studies show that the incorporation of the SPIONs in the scaffolds by the DOD technique does not affect their biocompatibility in the two conditions studied and by increasing the number of SPIONs printed layers on the scaffold. The cytocompatibility results of this work in terms of viability, proliferation and cell adhesion agree with the results provided by other authors in different studies with other types of scaffolds also incorporating SPIONs in their composition [50,51,52].

3.5 Osteogenic effect

ALP glycoprotein on the cell surface hydrolyses the mineralisation inhibitor pyrophosphate to phosphate, which promotes mineral deposition in ECM collagen fibres [53,54,55]. As ALP potentiates ECM mineralisation, quantification of ALP at both the mRNA and protein level (enzymatic activity) has been used to describe osteoblast differentiation [56]. Also, as matrix mineralisation represents the last step of differentiation, the measurement of this marker is a very important outcome to determine the efficacy of osteoblast differentiation in bone research [56]. Here, the osteogenic differentiation of hBM-MSCs was assessed by measuring ALP and Runx2 expression level, ALP activity, mineralisation process and osteocalcin (OC) secretion. These markers were measured in the cells grown on the surface of the scaffolds with 25 printed layers of SPIONs, both in the presence and absence of the magnetic field. Our results revealed that 7 days of cell culture on the surface of both types of scaffolds with SPIONs deposits (DMSA 25L and RGD 25L) promoted a significant increase in ALP expression both in presence and absence of the magnetic field (MF), compared to cells cultured on the scaffolds without nanoparticles (SC and SC + MF respectively). After exposure of cells to MF, ALP gene expression was significantly increased in all conditions compared to unexposed cells (Fig. 5A). Thus, there was an increase of ALP expression associated exclusively with the material (SPIONs deposited on the scaffold surface), an increase associated exclusively with the magnetic field (SC + MF condition) and a larger increase associated with the combination of both (SPIONs deposited on the surface and + MF). Following the trend of gene expression at 7 days, a significant increase in ALP activity was observed in cells grown on DMSA 25L and RGD 25L scaffolds compared to cells grown on the control scaffolds (SC and SC + MF respectively), both applying and not applying MF. After exposure of cells to MF, ALP activity was significantly increased in all conditions compared to unexposed cells (Fig. 6A). Again, these results reveal one increase attributed to the presence of SPIONs, another related to MF, and the greatest increase when SPIONs are activated by MF. The results suggest that both types of scaffolds combined with the magnetic field promote efficient differentiation toward the osteogenic lineage of the human mesenchymal stem cells used in this study, as ALP expression has been described in preosteoblasts [56] and mature osteoblasts [57]. Our data agree with previous studies that have shown that SPIONs, static magnetic fields, and the combination of both promote osteogenic differentiation [34]. Since ALP potentiates extracellular matrix mineralisation [56, 58], which is another sign of osteogenic differentiation, we evaluated this process in our model. Figure 7A shows the results after 7 days of culture: despite being a very early time to evaluate a marker of osteogenic differentiation such as mineralisation, a significant increase in absorbance was observed for DMSA 25L + MF and RGD 25L + MF conditions compared to SC and SC + MF. This could suggest that in the DMSA 25L + MF and RGD 25L + MF scaffolds the cells start to mineralize earlier than in the other conditions, so that in a short time [7 days] the effect of osteogenic differentiation through the mechanotransduction process (MF-activated SPIONs) can be observed. Moreover, among the multiple transcription factors regulating osteoblast differentiation, RUNX2 is a master transcription factor in the differentiation of MSCs to osteoblasts [59]. Therefore, RUNX2 expression is commonly used as a marker of osteogenic differentiation of MSCs. RUNX2 is considered the master switch for the initiation of osteogenesis, as RUNX2 is expressed in MSCs and upregulated in pre-osteoblasts, while in mature osteoblasts, RUNX2 expression is downregulated [60, 61]. In this study we also measured the expression level of Runx2 at time 7 days. The results show a significant decrease in Runx2 expression in the DMSA 25L + MF and RGD 25L + MF conditions compared to the unstimulated conditions (DMSA 25L and RGD 25L respectively) (Fig. S10 in Supplementary Material). Furthermore, the lowest Runx2 expression values occur in the magnetic field stimulated conditions (+ MF) coinciding with the conditions with the highest values of ALP expression, ALP enzyme activity and mineralisation. These results could suggest that cells cultured on the surface of scaffolds with SPIONs deposits and subjected to magnetic stimuli (DMSA 25L + MF and RGD 25L + MF) differentiate into osteoblasts and start the mineralisation phase earlier than in the other conditions.

Fig. 5

Evaluation of ALP gene expression in hBM-MSCs cultured on the surface of SC, DMSA 25L and RGD 25L scaffolds by applying (+ MF) and not applying (− MF) an external magnetic field of 1 Hz frequency for 1 h per day. A Expression was measured at 7 days and B 14 days. Results were statistically analysed by one-way ANOVA and Tukey's test

Fig. 6

Evaluation of ALP activity in hBM-MSCs cultured on the surface of SC, DMSA 25L and RGD 25L scaffolds by applying (+ MF) and not applying (− MF) an external magnetic field of 1 Hz frequency for 1 h per day. A Expression was measured at 7 days and B 14 days. Results were statistically analysed by one-way ANOVA and Tukey’s test

Fig. 7

Evaluation of the mineralisation process in hBM-MSCs cultured on the surface of SC, DMSA 25L and RGD 25L scaffolds by applying (+ MF) and not applying (− MF) an external magnetic field of 1 Hz frequency for 1 h per day. A Expression was measured at 7 days and B 14 days. Calcium deposits stained with alizarin red were quantified by measuring absorbance at 620 nm. Results were statistically analysed by one-way ANOVA and Tukey’s test

Considering 14 days of cell culture in the absence of MF, ALP gene expression was maintained in the DMSA 25L scaffold, i.e., no significant differences were detected compared to the SC condition. However, a significant decrease in gene expression was detected in the RGD 25L scaffold compared to cells cultured in SC. Again, in the presence of MF, gene expression was significantly increased in all conditions compared to unexposed cells. The results suggest that at short times (7 days) it is possible to detect the mechanotransduction effect, because the increase in ALP gene expression in the DMSA 25L + MF and RGD 25L + MF conditions is significantly higher than the increase in the SC + MF condition. Nevertheless, at long times (14 days) such effect is not so evident, since no differences are observed between the SC + MF and RGD 25L + MF conditions, even though a significant increase is still observed in the DMSA 25L + MF condition with respect to SC + MF. Results are shown in Fig. 5 A and B, respectively. At 14 days (Fig. 6B), no significant differences in ALP activity were observed between DMSA 25L and RGD 25L conditions with respect to SC, neither applying nor not applying MF. After exposure of cells to MF, ALP activity was significantly increased in all conditions compared to unexposed cells. Again, this result resembles that obtained in the 14 day gene expression analysis. An increase in ALP activity associated with the magnetic field is observed, but no differences are observed between the scaffolds with SPIONs and the control scaffolds. This could be due to the fact that both, the osteogenic effect produced by the differentiation medium (osteogenic factors) and the osteogenic effect produced by the differentiation medium plus the magnetic field, for a prolonged time (14 days), are much higher than the effect associated with the SPIONs and the effect associated with the SPIONs activated by the MF. Therefore, the effect of the differentiation medium at long time (14 days) might be masking the effect due to mechanotransduction, but at short time (7 days) it is evident. Finally, at 14 days (Fig. 7B), an increase in mineralisation was observed in cells cultured on the DMSA 25L scaffold compared to cells on the scaffold without SPIONs both with and without applying magnetic field. Furthermore, in the presence of MF, mineralisation was significantly increased in all conditions compared to unexposed cells. These mineralisation results are in agreement with those obtained for osteocalcin secretion (Fig. 8). After 7 days of culture (Fig. 8A), a significant increase in the amount of OC secreted was observed in the DMSA 25L + MF and RGD 25L + MF conditions compared to SC and SC + MF, as observed in the mineralisation process. At 14 days (Fig. 8B), a significant increase in OC secreted by cells cultured on scaffolds where MF was applied was observed compared to unexposed cells. Results were statistically analysed by one-way ANOVA (P < 0.001) and Tukey's test.

Fig. 8

Evaluation of the amount of OC secreted by hBM-MSCs cultured on the surface of SC, DMSA 25L and RGD 25L scaffolds by applying (+ MF) and not applying (− MF) an external magnetic field of 1 Hz frequency for 1 h per day. A Amount of OC measured at 7 days and B at 14 days. Results were statistically analysed by one-way ANOVA and Tukey’s test

Hu et al. [62] demonstrated enhanced bone regeneration in vivo in rats using a gelatine sponge (GS) loaded with SPIONs as a scaffold (SPIONs-GS). The results ensured that SPIONs induced active osteogenesis without using an external magnetic field. Jia et al. [63] coated 3D printed scaffolds with SPIONs and used for palate bone regeneration in a rat model. The results demonstrated that scaffolds coated with SPIONs can be used to treat defects of the palate. Moreover, several studies have revealed that static magnetic fields can regulate the proliferation, differentiation and function of bone tissue cells, including hBM-MSCs, osteoblasts, osteoclasts and osteocytes. As well as a large number of animal experiments and clinical studies have shown that static magnetic fields have effective therapeutic effects on bone-related diseases such as non-healing fractures, non-union of bone implants, osteoporosis and osteoarthritis [64]. In recent years, research on the combination of static magnetic fields with bone regenerative materials, especially magnetic materials, has increased [58]. Zhang et al. have also shown that the magnetic field stimulates osteogenic differentiation of MSCs on magnetic scaffolds [65]. However, unlike our results, they do not detect the osteogenic effect in the short term (7 days) and do detect it in the 14-day culture. This suggests that in their materials, magnetic particles coupled to the magnetic field could promote osteogenesis, but a longer exposure time is needed.

Regarding the presence or absence of RGD peptide in SPIONs deposited on the surface of the scaffolds, after 7 days of culture no significant differences were observed in either ALP activity or mineralisation, regardless of whether or not a magnetic field was applied (DMSA 25L vs. RGD 25L and DMSA 25L + MF vs. RGD 25L + MF). After 14 days of culture, there was also no difference between the presence or absence of RGD peptide on the ALP activity of the cells. Functionalisation of SPIONs with the RGD peptide was carried out with the aim of favouring interactions between SPIONs and cellular integrins, which can act as cellular mechanosensors [66,67,68]. The recent study by our group [34], demonstrates that functionalisation of this type of SPIONs with RGD peptide enhances osteogenic differentiation of hBM-MSCs when added in aqueous solution on cells and magnetic field is applied after the first 30 min upon incorporation. However, in the present strategy SPIONs are in the form of deposits on the scaffold surface, "trapped" between the scaffold surface and the cell monolayer surface. So, the SPIONs are in contact with the cell surface whether or not the cells are functionalized with the RGD peptide (as can be seen in the confocal microscopy images). Also, in addition to integrins, cells present other surface proteins and membrane structures such as piezo-type calcium channels, gap junctions, or primary cilia, which play a crucial role in mechanosensing of mechanical stimuli [69, 70]. So when the magnetic field is applied to generate a mechanical stimulus through SPIONs, the cells possess several pathways to detect that mechanical stimulus. Therefore, this fact could explain why no increase in differentiation is observed in the condition of scaffolds with SPIONs functionalized with RGD with respect to the condition of scaffolds with SPIONs without RGD.

Nevertheless, after 7 days of culture, an increase in ALP expression was observed in the DMSA 25L condition compared to the RGD 25L condition in both cases applying and not applying magnetic field. This fact was also observed after 14 days of culture in the results of ALP expression without magnetic field and mineralisation after 14 days of culture.