Musculoskeletal diseases are an increasing burden worldwide and a principal cause of persistent pain and disability. Among them, the incidence of degenerative joint diseases (DJD), mainly of osteochondral defects (OD) and osteoarthritis (OA) is on the rise, affecting hundreds of millions of people worldwide (Kloppenburg and Berenbaum, 2020). Ageing population, sedentary life style but as well increased occurrence of traumatic injuries due to traffic accidents, war casualties or natural disasters have been implicated in DJD. Their impact is even more dramatic due to the fact that such lesions do not heal and have as consequences permanent disabilities and reduced quality of life. In addition, DJD is getting “younger”, 16–17% of adults below 40 and up to 60% of adults of 40–65 years of age are reported to having various forms of the disease (Merkely et al., 2018). Current therapies are mainly symptomatic, failing to offer effective joint anatomical and functional rehabilitation. Total joint replacement still represents the ultimate method for treating DJD-induced joint deterioration. Regenerative therapies promise to deliver biological restauration of joint surface and to offer permanent and durable anatomical and functional recovery. Despite more than 30 years of basic and translational research in the field, current available products for joint resurfacing are failing to deliver consistent disease modifying therapeutic solutions. The most appropriate cell source for cartilage bioengineering is still under debate. ADSC are extensively investigated for their use as cell sources in regenerative medicine (RM) (Zizhen et al., 2019; Lima et al., 2015). ADSC are relatively easy to extract by means of enzymatic or mechanical methods from the adipose tissue derived from elective cosmetic procedures. ADSC are available as both autologous and allogeneic sources, can derive larger number of elements per tissue volume (Si et al., 2019) and possess superior proliferative and immunomodulatory capabilities compared with other mesenchymal stem cell sources (Ceccarelli et al., 2020) and could be an attractive modality to generate functional implantable tissue for the treatment of cartilage defects (Rahbar Saadat et al., 2020). However, the efficiency of chondrogenic conversion has proved not to equalize other ADSC sources (such as bone marrow or cord blood) (Wei et al., 2007). Derived from perinatal tissues (namely umbilical cord), Wharton Jelly is as a rich source of allogeneic MSCs with superior proliferative and immunomodulatory capabilities (Esposito et al., 2013). Chondrogenic conversion of Wharton jelly-derived mesenchymal stem cells (WJMSC) was shown to be superior to bone marrow mesenchymal stem cells (BM ADSC). Mechanic stimulation of ADSC was shown to increases differentiation to musculoskeletal lineages (bone, cartilage, skeletal muscle) and efficiently increased ADSC as well as WJMSC chondrogenic potential in vitro and in vivo (Zha et al., 2021). Previous studies have shown that different types of mechanical loading (such as compression, perfusion, vibration, stretching) are effective in increasing ADSC differentiation (Yong et al., 2020). Delivering the appropriate micro-mechanical stimulation that closely mimic the intensity, duration and orientation of mechanical cues in the developmental niche, at cellular level, has proved to be a challenge for musculoskeletal tissue engineering. Mechanical stem cell preconditioning using biophysical stimulation by applying various types of mechanical stress within dynamic bioreactors (compression, shear stress, and hydrostatic pressure) requires ancillary equipment and often a direct contact with the cells or cell media. This introduces supplementary steps in the process of cell manufacturing when intending clinical application. Moreover, at a cellular level, the distribution of applied forces might be uneven with consequences on quality and reproducibility of chondrogenic conversion (Fahy et al., 2018). Alongside with the perspective of introducing a method that waves the need of supplementary equipment, magneto-mechanical stimulation delivers biomechanical cues at a cellular level, more closely reproducing the natural biomechanics. In plus, the enhancement of the chondrogenic differentiation obtained within a magnetic field has also been demonstrated to having a synergistic effect with biochemical factors delivered by differentiation media resulting in an enhanced chondrogenic differentiation (Amin et al., 2014). The use of MNPs of various structures and coatings has been previously found appealing for regenerative medicine purposes when sought as drug and small molecule delivery vehicles as they have the potential to support regenerative processes in vitro and in vivo as well as for cell tracking and targeting purposes. Nanoparticle-based manipulation of cell and stem cell fate are recognized as the breakout technology capable to consistently contribute to advancement of biological joint resurfacing toward clinical application (Eftekhari et al., 2020). MNP are particularly appealing for manipulating cell fate due to their excellent biocompatibility, versatility and magnetic proprieties (Wimpenny et al., 2012). Of particular interest when using MNP is the ability to use magnetic actuation as a modality to deliver micromechanical stimulation to differentiating cells (Zhang et al., 2020). Mechanical stimulation consistently increase chondrogenic conversion in ADSC. However, the modality to deliver mechanical preconditioning are difficult to translate for potential clinical applications (O’Conor et al., 2013). Iron oxide MNP internalized by ADSC preserve their proliferative and differentiation capability while inducing cell magnetization. This is opening fascinating possibilities for remote cell manipulation under magnetic field (MF) aiming for MNP mediated cell actuation. Such particularity can be used as a modality to deliver remote micromechanical stimulation to stem cells differentiating to musculoskeletal lineages—osteoblasts and chondrocytes (Lima et al., 2015).

Due to their magnetic responsiveness, cells loaded with MNP can be traced within living systems using clinically available MRI imagistic equipment or incoming magnetic particle imaging technologies. The magneto-mechanical effect adds to this already versatile stem cell-MNP functionality the ability to potentially control cell fate by improving differentiation to mechanosensitive lineages, especially those required for musculoskeletal regeneration. In our previous studies we have tested the effect of proprietary uncoated MNP in influencing osteogenesis and adipogenesis and found that magneto-mechanical stimulation increase osteogenesis and decrease adipogenesis in primary ADSC. Here we aimed to test comparatively the effect of magneto-mechanical stimulation in human primary ADSC and WJMSC chondrogenesis. The purpose was to test if MNP we are manufacturing can be used as a modality to deliver magneto-mechanical stimulation to stem cells undergoing chondrogenic differentiation in vitro and to identify the most appropriate cell source aiming future cartilage engineering strategies.

Remote magnetic actuation of ADSC and WJMSC, with or without MNP, in alternating magnetic field (MF) was investigated as a modality to deliver mechanical stimulation. Its effect on chondrogenic conversion was assessed quantitatively in order to detect the most appropriate cell source for cartilage engineering purposes.

The ability to produce bioengineered tissues and organs relies on abundant cell sources that have the capability to convert to the desired lineage under appropriate biochemical and mechanical cues. Mesenchymal stem cells have been intensively sought for tissue engineering of cartilage and a large number of research groups are reporting promising in vitro and in vivo results. The importance of mechanical stimulation for obtaining functional engineered bio equivalents especially in the case of musculoskeletal tissues has been stressed out (Athanasiou et al., 2015). Various methods for mechanical stimulation and dynamic bioreactors have been designed for this specific purpose. However, the ancillary equipment for delivering mechanical load is in many cases bulky, resource consuming and difficult to scale up for translational purposes. We have previously reported good results using magneto-mechanical stimulation for osteogenic and chondrogenic conversion of human primary ADSC using incorporated proprietary MNP and MF exposure. ADSC and WJMSC are considered efficient cell sources for tissue engineering due to their good proliferative and differentiation capabilities, easiness of access and lack of ethical issues with their procurement. WJMSC of allogeneic origin are reportedly non-immunogenic, can generate clinically relevant number of mesenchymal progenitors and are already the object of clinical trials (Liang et al., 2020). In this study we focused on comparing the two cell sources in respect to their capability to undergo chondrogenic lineage differentiation under magnetically delivered mechanical stimulation.

Proprietary bare magnetite MNP are prepared by co-precipitation of iron salts, method that enables production of relatively homogenous nanoparticles (Perlstein et al., 2010). MNP used in this study have average size of 20 nm and display ferromagnetic proprieties and characteristic magnetization curve under MF exposure. Due to their reduced size, sterilization by autoclave induce particle aggregation (data not shown). Therefore, we used for sterilization UV exposure in the biosafety hood. MNP, in fixed concentration of 40 μg/ml, were first diluted in complete culture media in order to allow formation of a protein corona. Since the protein corona is the major factor that determines MNP biological identity, “presentation” to cell membrane and their internalization (Nierenblerg et al., 2018), one single batch of FBS was used for all experiments in this study for consistency. Simple addition in culture media resulted in different behavior of MNP in relation to the two cell types investigated here. In the first 24 h of cell-particle interaction, MNP were observed to agglomerate around WJMCSs without being internalized. With increasing time in culture, particles were observed in the pericellular area as well as inside cytoplasm. Conversely, MNP could be detected inside ADSC cytoplasm as soon as 24–48 h and further on through all duration of observation (maximum 28 days).

The particularity of WJMSC-MNP interaction could be probably explained by the presence of a dense pericellular mesh (Wang et al., 2018) that interfere with initiation of endocytosis. At the initial contact with the cells, MNP are firstly interacting with pericellular matrix and plasma membrane and uploaded by endocytotic mechanisms that vary with their size, coating, and electric charge. MNP are included within intracellular vesicles and finally degraded in the acidic environment of the endo-lysosomal compartment (Kolosnjaj-Tabi, 2013). Bare and as well Fe₃O₄ MNP coated with biocompatible shells are known to be nontoxic, non-genotoxic and not to induce hemolytic processes in contact with blood (Fan et al., 2020). Previous reports have shown good viability in many stem cell types upon internalization of commercially available MNP with different coatings (Wimpenny et al., 2012). Our group has reported good viability in various human stem cells (bone marrow, trabecular bone, adipose derived) with proprietary bare, chitosan and palmitic acid coated Fe₃O₄ MNP. A particularity of MNP that distinguish them from other nanoparticles is the existence of a cellular mechanism for iron storage and metabolism. Ferritin-dependent iron homeostasis is an endogenous mechanism ubiquitous and evolutionary preserved in mammalian cells. Ferritin protein cages accommodate MNP degradation products while concomitantly slowing down the degradation process due to the colloidal behavior in acidic medium (Volatron et al., 2017). A dose dependent reduction in WJMSC viability and proliferation after exposure to commercially available MNP as assessed using MTT was reported (Ohki et al., 2020) suggesting that the maximum iron load the preserves phenotype is cell type dependent. The correlation between maximum iron load that preserves viability and specific cytosolic ferritin activity in various cell type would warrant further investigation. We found that viability of WJMSCs exposed to MNPs after 24 h was only 83,5% compared to non-loaded control as assessed by cell mitochondrial activity. While this is not obligatory a proof that MNPs are cytotoxic to WJMSCs, it is revealing limited mitochondrial activity compared to non-loaded. This could be possible explained by the fact than iron handling enzymatic equipment (especially mitochondrial ferritin) as well as iron overload threshold is cell specific. Mammalian cells response to similar amounts of iron traffic and load is known to not only dependent on cell type but as well on cell status and activity (Ward and Cloonan, 2018). Further investigatings are needed to explore this particularity.

As resulting from population doublings, both cell types, loaded and non-loaded, continue to proliferate up to 14 days when they reach a plateau, most probably due to culture conditions (reaching over confluence). ADSC-MNP display slightly higher PD compared to non-loaded while ADSC-MNP slight decrease proliferation upon MNP exposure. Iron is known to play a crucial role in energy metabolism, being constituent of mitochondrial enzymes of the electron transport chain, involved in ATP production, oxygen consumption, cell cycle progression and proliferation. Excess of iron, however, generate Fenton reaction producing reactive oxygen species (ROS), especially hydroxyl radicals which can have destructive effects on cellular lipids and DNA (Wessling-Resnick, 2017). The delicate balance between promoting mitochondrial activity and cell proliferation and inducing ROS formation that arrest cell cycle is dependent on complex factors of which free iron ions formation, and availability of intracellular ferritin. Expression of iron transport and metabolism proteins in ADSC, as well as activation of ROS-stimulated pathways are likely to be dependent on tissue of origin, donor age, presence or absence of associated disease (diabetes mellitus, obesity) (Mehta, 2021). Here we found that even if the amount of iron accumulated per cell in the first days in WJMSC is significant lower compared to ADSC-MNP, their growth rate is non-significantly decreased. WJMSC are isolated from WJ, a tissue originating from extraembryonic mesoderm that functions as the very first hematopoietic niche in the developing embryo. WJMSC were shown to support hematopoiesis in vitro (Lo Iacono et al., 2018) and to possess a particular transcriptome compared to other ADSC (Jeong et al., 2021), possibly being less equipped for iron internalization. Adenoviral-mediated ferritin overexpression for WJMSC performed as a method to induce their traceability using magnetic resonance imaging (MRI) equipment was found to significantly decrease cell proliferation. (Dai et al., 2017). Further investigation of WJMSC particularities regarding iron transport and metabolism would be necessary.

Bare MNP could be uploaded by ADSC, but not by WJMSC within first 24 h. The average amount of 14.9 pg iron load per cell unit induced efficient cell magnetization in ADSC-MNP suspension as demonstrated by VSM. In WJMSC, the amount of up-loaded iron was significantly lower; however, cells still displayed a magnetic behavior, demonstrating they could respond to an applied MF. This finding was further supported by the configuration of cytoskeletal fibers in loaded and non-loaded cells exposed to MF. After brief exposure (10 min) to a relatively low intensity MF (6 mT), F actin fibers showed the tendency to disorganize and to distort cell contour. Static field exposure (1 mT) of dental pulp stem cells (DPSCs) was shown to rearrange cytoskeleton and to increase their migration proliferation and differentiation by recruiting YAP/TAZ to the nucleus, inhibiting its phosphorylation and upregulating two YAP/TAZ-regulated genes, CTGF and ANKRD1. MF induced YAP/TAZ nuclear localization was inhibited by cytoskeleton inhibitor, cytochalasin D (Zheng et al., 2018). In our study ADSC-MNP and, at a lower extent, WJMSC-MNP displayed increased cytoskeletal fiber rearrangement under MF (Figure 4). MNP are known to be trafficked inside cell cytoplasm along cytoskeletal fibbers. It is possible their proximity to actin fibbers during MF exposure to generate micro movements that act as internal mechanical stimulation, influencing YAP/TAZ nuclear translocation and possible other mechano-responsive pathways.

An interesting observation is the particularity of iron load dynamic on the duration of 28 days in ADSC-MNP. After 24 h cell-MNP interaction, cells were found to accumulate the maximum amount of iron/cell (14.9 pg/cell), quantity that was found to decrease to almost one third (6.64 pg/cell) at 7 days, then increased again to reach approximatively the initial amount at 28 days (14.71 pg/cell). The initial decrease in iron content at 7 days can be partly explained by proliferation and consequent distribution of MNP in the dividing cells. Other possible explanations such as particle biodegradation and re-synthesis could be involved. Using magnetism as a modality to demonstrate the MNP presence inside cells, a possible mechanism of biodegradation and “re-biosynthesis” was demonstrated to occur in human ADSC. Cells were found to be “re-magnetized” after degradation of nanoparticles that took place at 7–10 days after particle internalization (Van de Walle et al., 2019). This finding was interpreted as a direct demonstration of a possible neo synthesis of magnetic nanoparticles inside endosomes, a process H-ferritin dependent that could explain, at least in part, the gradual increase in iron content/cell at 14 and 28 days for ADSC-MNP. WJMSC increased iron/cell gradually up to 28 days probably as a result of slow internalization of the MNP due to rich fibronectin pericellular matrix and a possible low expression of transferrin or other iron transporters, fact that needs to be further investigated.

We used assessment of B-Gal activity as a modality to detect culture induced senescence. Not surprisingly, B-Gal expression was found to increase at 28 days in all cell types, loaded and non-loaded, compared to 14 days. However, ADSC-MNP showed less B-Gal activity compared to non-loaded at both time points whereas WJMSC-MNP were found to express increased enzyme activity compared to non-loaded. The presumptive lower activity of cytosolic ferritin and increased ROS formation in WJMSC could be involved. Culture-induced senescence is a well-known phenomenon that impedes in vitro cell expansion and negatively influences MSC therapeutic potential (Liu et al., 2020). No significant ROS increase and cell ageing could be observed in human myoblasts loaded with different amounts of superparamagnetic iron oxide nanoparticles (SPIOS) that retained their phenotype in vitro and in vivo (Wierzbinski et al., 2018). Here we found that non coated MNP internalization decreased B-Gal activity during extended culture in ADSC but not in WJMSC. Further studies focusing on ROS expression as well as on investigating senescence associated phenotype are needed to determine if MNP loading could function as a modality of preventing culture induced senescence in ADSC populations.

Based on quantification of GAG deposition per cell, we found that condrogenic conversion of MF-exposed and non-exposed ADSC-MNP was significantly increased compared to non-loaded. Furthermore, MF exposure increased GAG deposition in ADSC-MNP compared to non-exposed ones, albeit non-significantly. Iron is an essential factor for cartilage formation and its internalization is required for ECM synthesis. Reduced iron is a catalyst of proline and lysine hydroxylation reactions that are essential to the formation of mature collagen molecules (Pacifici and Iozzo, 1988). MNP can provide slow release of reduced iron ions supporting chondrogenic ECM deposition. Stem cells were found to handle iron transport differently depending on differentiation pathways even in the absence of MNP. In MSC chondrogenesis, transferrin is down regulated while ferroportin is significantly higher expressed compared to osteo and adipogenesis. Moreover, ADSC loaded with MNP, overexpressed L-Ferritin (involved in iron storage as nonmagnetic ferrihydrite in the ferritin cage) during chondrogenesis. Conversely, H-Ferritin, possible involved in de novo MNP biosynthesis, was found to be higher expressed during osteogenesis and adipogenesis (Van de Walle et al., 2019). Iron availability and supplementation has been recently confirmed as a modality to increase chondrogenesis in vitro and in vivo in Guineea pigs. Ascorbic acid and iron supplementation in the form of commercially available ferumoxytol was found to significantly increase chondrogenesis in bone marrow ADSC in vitro as well as in vivo in Guinea minipig model of cartilage defect (Theruvath et al., 2021). Ferrumoxytol, is used as iron supplement for intravenous administration and occasionally off label as a MRI contrast agent for vascular and nodal metastasis. It consists of carbohydrate-coated ultra-small SPIOs (Vasanawala et al., 2016). Their hydrodynamic size (average 25 nm) and iron oxide composition are similar with particles used in our study. Here we used non-coated MNP, fact that could increase intracellular iron availability and decrease the time of MNP processing by the cells. The fact that WJMSC iron handling mechanisms cannot provide similar amounts of intracellular iron compared to ADSC (as resulting from the quantitative assessment of intracellular iron content) can possible interfere with chondrogenic conversion. Profiling comparatively the cellular iron enzymatic equipment in loaded and non-loaded ADSC and WJMSC during chondrogenic conversion could reveal very useful insights potentially offering molecular targets to enabling future cartilage engineering procedures. ADSC chondrogenesis was previously found inferior to other stem cell sources (Fellows et al., 2016). In this study, significant higher amounts of GAG/cell obtained from non-loaded WJMSC compared to non-loaded ADSC, confirming the fact that non-modified ADSC are not the best source when considering cartilage engineering. However, MNP presence clearly increased GAG deposition in ADSC, especially after MF exposure. Actin cytoskeleton was shown to modulate COL2/aggrecan fragment generation; the factors that control actin dynamic decisively control the chondrocyte phenotype (Lauer et al., 2021). MNP presence, in close relation to actin fibbers (Orynbayeva et al., 2015), can induce the necessary actin dynamics to initiate and promote chondrogenesis. In this study, we showed that MF exposure influences the cytoskeletal fiber organization. GAG deposition in magnetically stimulated ADSC-MNP was almost 5 times higher compared to non-loaded WJMSC, pointing to a possible modality for manipulating ADSC for improved chondrogenic conversion in vitro and, possibly, future methods for improving cartilage-repair pulsed magnetic field exposure (PEMF) of BM-ADSC in combination with magnetic hydrogels which were shown to increase repair in rabbit model of cartilage defects (Huang et al., 2019). Further studies are needed to more precisely clarify the molecular mechanism governing the correlation between MNP positioning, MF exposure and increased chondrogenesis in ADSC. Chondrogenic induction of WJMSC was previously shown to result in increased upregulation of cartilage-specific type II collagen synthesis compared to BM-ADSC and decreased osteoinductive Runx2 and collagen X in similar culture conditions (Reppel et al., 2015), suggesting their use could avoid hypertrophic cartilage formation and graft ossification. Previous studies have used human BM ADSC loaded with homogenous sized MNP derived from bacteria (Magnetospirillium sp) for inducing cell aggregation in pellets and for delivering magnetic mediated shear stress during initial stages of chondrogenic conversion. Increased levels of GAGs and collagen production as well as the absence of hypertrophic conversion was observed in the MNP loaded cells (Son et al., 2015).

In our experience, MNP mediated aggregation in spheroid culture under MF was not efficient in increasing the amount of GAG production compared to simple pellet procedure, probably due to increase compaction of three dimensional structure and decreased nutrient availability (Labusca et al., 2021).

A limitation of this study is that we have not investigated hypertrophic markers. However, previous evidence supports the fact that significant GAG increase in ADSC-MNP at protein level is an indicator of normal articular cartilage, being correlated with collagen type expression II and absence of hypertrophy (Kamisan et al., 2013). Further studies will be needed as well to address donor variability especially for the case of ADSC. WJMSC donor population is more likely to being homogenous as it involves processing umbilical cord sample from women of fertile age and with healthy pregnancies. ADSC donors, however, could be more heterogeneous, especially if the use of autologous cells is envisaged for a future therapeutic purpose. A comparative study including a larger number of donors stratified based on age, gender and BMI as well as profiling ADSC characteristics relative to their chondrogenic conversion capability will to enable further donor selection for cartilage regeneration purposes.

ADSC upload proprietary MNP by simple addition in culture media after 24 h in quantities that endows the cells with magnetic properties. ADSC-MNP, submitted to low-frequency MF during initial stages of chondrogenic induction, generated the highest amount of chondrogenic conversion at protein level as assessed by GAG/cell deposition after 21 days pellet culture. WJSC displayed low MNP internalization and significant lower amount of GAG deposition in similar conditions. B-Gal activity, after extended culture in vitro, was significantly decreased in ADSC-MNP compared to non-loaded, but not in WJMSC, suggesting a possible senescent protective mechanism that needs further investigation. Magneto-mechanical stimulation of MNP-loaded cells could be an efficient modality to increase the efficiency of ADSC, but not of WJMSC for cartilage engineering. Further studies are necessary to confirm the feasibility of this method in animal models of cartilage defects and to detect the molecular basis of MNP mediated magneto-mechanical stimulation.