Myogenic commitment of human stem cells by myoblasts Co-culture: a static vs. a dynamic approach

Abstract An in-vitro model of human bone marrow mesenchymal stem cells (hBM-MSCs) myogenic commitment by synergic effect of a differentiation media coupled with human primary skeletal myoblasts (hSkMs) co-culture was developed adopting both conventional static co-seeding and perfused culture systems. Static co-seeding provided a notable outcome in terms of gene expression with a significant increase of Desmin (141-fold) and Myosin heavy chain II (MYH2, 32-fold) at day 21, clearly detected also by semi-quantitative immunofluorescence. Under perfusion conditions, myogenic induction ability of hSkMs on hBM-MSCs was exerted by paracrine effect with an excellent gene overexpression and immunofluorescence detection of MYH2 protein; furthermore, due to the dynamic cell culture in separate wells, western blot data were acquired confirming a successful cell commitment at day 14. A significant increase of anti-inflammatory cytokine gene expression, including IL-10 and IL-4 (15-fold and 11-fold, respectively) at day 14, with respect to the pro-inflammatory cytokines IL-12A (7-fold at day 21) and IL-1β (1.4-fold at day 7) was also detected during dynamic culture, confirming the immunomodulatory activity of hBM-MSCs along with commitment events. The present study opens interesting perspectives on the use of dynamic culture based on perfusion as a versatile tool to study myogenic events and paracrine cross-talk compared to the simple co-seeding static culture.


Introduction
Skeletal muscle tissue can heal spontaneously after an injury [1,2]. However, drug therapies with nonsteroidal anti-inflammatory drugs (NSAIDs), intramuscular corticosteroids, and operative management are often required; whereas, extended tissue damages need biological treatments, including cell therapy and tissue engineering [3].
In this field, in vitro models are extremely challenging and involve stem cells, such as bone marrow mesenchymal stem cells (BM-MSCs), and their ability to differentiate towards extremely diverse phenotypes under appropriate conditions [4][5][6][7][8]. Myogenic commitment of BM-MSCs involves key transcription factors and regulators as Myocyte Enhancer Factor-2 (MEF 2), Desmin [9,10], Myogenic Factor 5 (Myf 5), Myogenic Differentiation 1 (MyoD 1) [11], and Myogenin [12]. To promote myogenic commitment, culture medium can be supplemented with several growth factors, such as Hepatocyte growth factor (HGF) that has been described to promote myoblast proliferation and satellite cell activation [13,14], Insulin-like growth factor (IGF) that activate myoblast proliferation and differentiation [15], and Fibroblast growth factors (FGFs) that are involved in satellite cell proliferation and myoblast-to-myotube differentiation [16,17]. IGF, FGF, and HGF are usually added at a concentration of 10 ng/mL, while Epidermal growth factor (EGF) is used at 0.2 ng/mL [9,11,12]; however, the best mix and related concentration of growth factors for myogenic commitment are still unknown and under investigation. Myogenic commitment of BM-MSCs can also benefit by co-culture with skeletal muscle cells; however, seeding density ratios have not been well defined and standardised yet [9,10,12].
During tissue regeneration, MSCs and immune cells strongly interact, favouring the healing process. Immune cell phenotype is largely influenced by MSCs thanks to their immune-modulatory effect [18]. MSCs reduce T-lymphocyte proliferation, causing their unresponsiveness by decreasing interferon-c (IFN-c) secretion [19] and preventing their activation [20], can favour T-lymphocyte and natural killer cell proliferation and viability, without affecting their anticancer cytotoxicity [21]. Macrophage polarisation seems to be influenced by MSCs in-vitro, especially by secretion of IL-10, an anti-inflammatory cytokine [22]. Particularly, BM-MSCs can suppress the expression of CD4 and CD8 by mononuclear cells and induce IL-6, IL-10, and TGF-b1 secretion [23], inhibit proliferation of B-lymphocytes and impair the secretion of immunoglobulins (IgM, IgG, IgA) [24]. Moreover, BM-MSCs have been reported to prevent the differentiation of monocytes into dendritic cells and to affect their antigen-presenting capacity [25].
Perfusion bioreactor systems have been also described to promote in vitro models. These culture systems have been used to improve the spatial uniformity of cardiac myocyte distribution on the 3D scaffolds and enhance the expression of cardiac-specific markers [26]. A perfusion system has been adopted to commit hBM-MSCs towards osteogenic and chondrogenic phenotype, showing that medium perfusion allowed a better distribution of soluble factors produced by cells as well as a more efficient mass transfer of nutrients [27][28][29][30][31][32][33]. On the other hand, more organised perfusion culture systems were adopted to understand paracrine effects between different cells in the so-called "organ on chips" systems that may involve microfluidic circuits for culture medium and different cell phenotypes to investigate molecular communication and/or cell interaction [34,35].
Considering all the described concepts, the present work aimed to investigate the myogenic commitment of hBM-MSCs, adopting a growth factor (bFGF) supplemented medium and a co-culture with human primary skeletal myoblasts (hSkMs) both in static and perfused conditions. Cells were co-seeded in the same well in static culture, whereas hSkMs and hBM-MSCs were seeded in separated wells by adopting the perfused bioreactor system culture. hBM-MSCs commitment was defined based on the expression of genes associated with myogenic phenotype, i.e. Pax 3, MyoD 1, Myf 5, Myf 6, Desmin and Myosin heavy Chain II (MYH2), on Desmin and MYH2 protein expression levels by semi-quantitative immunofluorescence (qIF) and by western blotting in dynamic co-culture conditions. Our study promoted the perfused co-culture system as a more versatile in vitro model to study myogenic differentiation events and cell cross-talk.

hBM-MSC isolation and harvesting
hBM-MSCs were obtained from two healthy donors after informed written consent in accordance with the Declaration of Helsinki and protocol approved by Our Institutional Review Board (Ethics Committee "Campania Sud", Brusciano, Naples, Italy; prot./SCCE n. 24988). Protocols and data related to hBM-MSC characterisation by flow cytometry are reported in the Supplementary Materials section [36].

Perfused medium bioreactor co-culture
Cells were cultured in a perfusion bioreactor formed by a custom multi-well plate milled in polymethyl methacrylate (PMMA, AltuglasV R CN 100 10000, Altuglas International, La Garenne-Colombes Cedex, FR), a biocompatible material for biomedical applications [37]. This plate has two holes for silicon tube (TygonV R , Charny, France) insertion that allows the creation of a circuit in which medium flow is provided by peristaltic pumps at a constant flow rate of 1.0 ml/min [38]. In this case, hBM-MSCs and hSkMs were seeded at a density of 4,000 cells/cm 2 in different wells always maintaining the ratios of 2:1 or 1:1. Medium ratios were adjusted accordingly, as described before.

Brightfield images
Brightfield images of hBM-MSC and hSkM cultures were captured using a LEICA DMIL LED microscope, at 10Â magnification, and acquired by LEICA DFC425 C camera.
RNA isolation and gene expression profiling mRNA expressions were analysed by Reverse Transcription quantitative polymerase chain reaction (RT-qPCR). Total RNA was extracted using RNeasy Micro Kit (Qiagen, Hilden, Germany). For each sample, 1 lg of total RNA was reverse transcribed using iScriptTM cDNA synthesis kit (Bio-Rad, Milan, Italy), and relative gene expression analysis was performed on a LightCyclerV R 480 Instrument (Roche, Basel, Switzerland), using SsoAdvanced TM Universal SYBRV R Green Supermix (Bio-Rad, Hercules, California, United States). Data were normalised to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) applying the geNorm method [39] with CFX Manager software (M < 0.5), and fold changes in gene expression were determined by 2 ÀDDCP method and presented as relative levels over day 0 ¼ 1.

Immunofluorescence assay
Cells were fixed with 4% paraformaldehyde for 30 min at RT followed by permeabilization with 0.1% Triton X-100 for 5 min and blocked with horse serum solution for 1 h. Cells were stained for Desmin (1:100; Abcam, Cambridge, UK) and for MYH2 (1:50, Thermo Fisher Sci., Waltham, MA, USA), and were incubated overnight at 4 C. Subsequently, cells were incubated for 1 h at RT with Alexa Fluor TM 488 goat anti-rabbit IgG (1:400; Thermo Fisher Sci.) and with VectaFluor TM anti-mouse IgG Dylight 594V R kit (Vector Laboratories, Burlingame, CA, United States) antibodies, and cell nuclei were counterstained using 4 0 ,6-diamidino-2-phenylindole (DAPI). All images were acquired by a fluorescence microscope with identical settings (Eclipse Ti Nikon Corporation, Tokyo, Japan). Signals intensity was quantified using Image J software (rel.1.52p National Institutes of Health, USA). Original images in RGB format were converted into a 16-bit (grey scale) format. Then, tagged areas were expressed as an average value of pixel intensity within a range from 0 (dark) to 255 (white). Data were normalised to the number of cells present in the whole field and reported as fold change relative to hBM-MSCs at day 0 [40].

Statistical analysis
Statistical analysis was conducted using Prism software (v.6.0, GraphPad Software, LLC, San Diego, California, United States). Results are presented as mean ± standard deviation (SD). Statistical analysis was performed using one-way analysis of variance (ANOVA) with Tukey's multiple comparison test, for comparison vs. hBM-MSCs at day 0, selected as control. Twoway analysis of variance (ANOVA) with Tukey's multiple comparison test was also performed for comparison of data obtained in perfusion system culture vs. hBM-MSCs at day 0. P values < .05 were accepted as significant [41].

hBM-MSC and hSkM characterisation
Mesenchymal phenotype was confirmed according to minimal criteria defined by the International Society of Cellular Therapy [42]. In particular, primary hBM-MSC were positive for mesenchymal markers, such as CD90, CD105, and CD73, while negative for CD34, CD14, CD45, and HLA-DR by flow cytometry analysis (see also Figure S1 in Supplementary materials).

Myogenic commitment strategies of hBM-MSCs
bFGF at a concentration of 10 ng/mL is largely recommended in literature as a supplement in the myogenic medium; however, when hBM-MSCs were cultured in static conditions within this environment, they exhibited poor changes in myogenic gene expression ( Figure 2). Indeed, Pax3 reached the maximum expression after 7 days of culture (34-fold), and MyoD 1, Myf 5 and Myf 6 sequentially picked the maximum of 30-fold, 23.6-fold, and 16.7-fold at day 7, respectively; then, these gene expressions were reduced. Desmin and MYH2 genes were not up-regulated throughout the whole culture time.
On the contrary, when hBM-MSCs were co-seeded with hSkMs at a ratio of 2:1, an excellent myogenic gene expression outcome was observed ( Figure 2): MyoD 1 expression was remarkable with the maximum statistically significant expression at day 14 with 3400-fold, while Myf 5 and Myf 6 reached the highest expression at day 7, with 2000-fold and 590-fold, respectively. Desmin was up-regulated at day 7 of 88-fold and at day 21 of 141-fold (p .001). MYH2 increased to 13-fold on day 14 and up to 32-fold (p .0001) on day 21. hSkM gene expression, normalised on hBM-MSCs at day 0, were added to Figure 2, as a positive control. From morphological analysis, hBM-MSC-hSkMs co-seeded showed a progressive elongated shape with an even more accurate orientation along the paralleled line (brightfield images in Figure 3a). This behaviour was particularly evident in zoomed areas at 150% inserted in each image of Figure 3a. IF images showed the presence of Desmin, as confirmed by q-IF with 1.5-fold (p .001) at day 7, and an increasing trend of MYH2 protein secretion, 1.1-fold (p .001) at day 21. From IF images, hSkMs nuclei clearly grouped appeared at day 14, forming the typically polynuclear syncytia, that can distinguish them from hBM-MSCs (Figure 3b). hBM-MSCs were significantly stained in green, the fluorochrome associated with MYH2. Poor hBM-MSC viability was observed when a 1:1 cell ratio was adopted (data not shown).
To study the immunomodulatory activity of hBM-MSCs along with myogenic commitment events, cytokine expressions were investigated ( Figure 4). hBM-MSCs culture with bFGF supplemented medium upregulated IL-12A at 67.5-fold at day 7; IL-1b reached 4-fold at day 7, then those expressions were significantly reduced along the further culture time. Among anti-inflammatory cytokines, IL-4 and IL-10 were the most overexpressed with the maximum fold reached by IL-4 of 44-fold and by IL-10 of 27-fold, both statistically significant, at day 7 ( Figure 4). In static co-seeding culture, TNF-a was the highest up-regulated pro-inflammatory cytokine, followed by IL-1b and IL-12A. The maximum peak at day 14 was for TNF-a (8-fold), IL-1b, (6-fold) and for IL-12A (4-fold). All expressions returned close to baseline levels at day 21 ( Figure 4).
The described data indicated that hBM-MSCs showed poor myogenic gene expression when cultured with bFGF Figure 1. Brightfield images and immunofluorescence assay of hSkMs. Brightfield images of hSkMs, seeded at a density of 120.000 cells/cm 2 , at days 7, 14 and 21 of culture. All images were captured using 10Â magnification; scale bar: 10 mm (a). IF assay at days 7, 14 and 21 of culture was performed by staining Desmin in red and MYH2 in green. All images were captured using 10Â magnification, scale bar: 200mm (b). Semi-quantitative IF of Desmin and MYH2 protein signals is shown as mean ± SD, and two-way ANOVA performed between time points, n ¼ 3, Ã p < .05, ÃÃ p < .01, ÃÃÃ p < .001 and ÃÃÃÃ p < .0001 (c). Figure 2. Gene expression profiling for myogenic markers by quantitative RT-PCR obtained from hBM-MSCs and hSkMs co-seeding culture in static condition with myogenic medium (10 ng/mL of bFGF). hBM-MSCs were co-seeded with hSkMs in ratio 2:1; hBM-MSCs and hSkMs were added, as controls. mRNA levels of myogenic markers: Pax 3, MyoD 1, Myf 5, Myf 6, Desmin and MYH2 were assayed by qRT-PCR at days 7, 14 and 21 of culture. The relative quantification of each mRNA gene expression normalised to endogenous GAPDH (internal control) was calculated using the 2 ÀDDCt method and presented as fold change over hBM-MSCs at day 0, selected as a control. All data were analysed by one-way ANOVA, n ¼ 3; Ã p < .05, ÃÃ p < .01, ÃÃÃ p < .001 and ÃÃÃÃ p < .0001. supplemented medium; whereas static hBM-MSCs and hSkMs (ratio 2:1) co-seeding induced an excellent expression of myogenic markers.

Dynamic culture system
When hBM-MSCs were cultured under bFGF supplemented medium in a dynamic environment, a total absence of myogenic gene expression was observed (despite the control gene GADPH being monitored correctly); IF observation confirmed the total absence of myogenic proteins (data not shown).
When, hBM-MSC and hSkM dynamic co-culture was tested, the two different cell populations were spatially separated, since they were seeded in different wells, even if the same culture medium was flowing along with the whole culture ( Figure 5). In this dynamic arrangement, cell ratios explored were of 2:1 and 1:1 (hBM-MSCs:hSkMs); a poor cell survival was observed for the 1:1 ratio (data not shown); therefore, all following data are related to culture achieved adopting the 2:1 ratio.
In such cases, hBM-MSCs showed an excellent up-regulation of myogenic genes (Figure 6a). Pax 3 was up-regulated 11-fold at day 7; myogenic regulation factors (MRFs), at day 7 showed  . Gene expression profiling for pro-and anti-inflammatory cytokines by quantitative RT-PCR obtained from hBM-MSCs and hSkMs co-seeding culture in static condition with myogenic medium (10 ng/mL of bFGF). hBM-MSCs were co-seeded with hSkMs in ratio 2:1; hBM-MSCs under bFGF medium and hSkMs were added as controls. The mRNA levels of pro-inflammatory cytokines (TNF-a, IL-6, IL-12A and IL-1b) and anti-inflammatory cytokines (IL-10, IL-4, TGF-b1 and TGF-b2) were assayed by qRT-PCR at days 7, 14 and 21 of culture. The relative quantification of each mRNA gene expression normalised to endogenous GAPDH (internal control) was calculated using the 2 ÀDDCt method and presented as fold change over hBM-MSCs at day 0, chosen as control. All data were analysed by one-way ANOVA, n ¼ 3; Ã p < .05, ÃÃ p < .01, ÃÃÃ p < .001 and ÃÃÃÃ p < .0001. the highest up-regulation of 9-fold for MyoD 1, 38-fold for Myf 5 and 37-fold for Myf 6. Desmin showed a constant upregulation throughout the dynamic culture, with 10-fold at day 7, 16-fold at day 14 and 9.5-fold at day 21. MYH2 increase was 4.4-fold on day 7.
Cytokine expression by hBM-MSCs, normalised over day 0, was also investigated (Figure 6b).Overall anti-inflammatory cytokine upregulation with respect to pro-inflammatory ones was observed, as monitored in static co-seeding culture. More in detail, IL-12A showed an increasing trend with 2-fold, 5-fold, and 7-fold at the investigated time points, while IL-1b expression was almost constant along the culture (1-fold change). Other proinflammatory cytokines, such as IL-6, TNF-a and IFN-c, were not detected. Among anti-inflammatory cytokines, IL-10 and IL-4 were the most upregulated: i.e. IL-10 reached the maximum spike of expression at day 21 with 19.2-fold and IL-4 at day 14 with 11-fold. TGF-b expression was quite low, except for TGF-b1 at day 7 (4-fold).
IF staining performed on hBM-MSCs cultured within the perfusion system showed a significant signal for Desmin (stained in red) at earlier time points and a significantly more intense signal for MYH2 (stained in green) especially at day 21 ( Figure 7b); q-IF confirmed a statistically significant increase of 2.4-fold (p .001) at day 7 and 2.9-fold (p .01) for Desmin and of 2.25-fold (p .001) at day 7and 2.5-fold (p .01) at day 21 for MYH2 ( Figure 7c). Furthermore, hBM-MSCs showed an extremely ordinate alignment with a clearly elongate shape. When similar IF staining was performed on hBM-MSCs supplemented only with bFGF medium, neither Desmin nor MYH2 signals were observed (Figure 7a). In dynamic culture, hBM-MSCs and hSkMs were seeded independently, Desmin and MYH2 production was monitored also by western blot assay ( Figure  8). Both proteins showed similar expression trends with a significant increase at day 7 and the maximum expression at day 14.
The described data suggested that hBM-MSC co-culture in perfusion undergoes a successful myogenic commitment with significant upregulation of all gene markers and antiinflammatory cytokines with Desmin and MYH2 protein production. Figure 5. Schematic representation and image of perfused bioreactor system used for the dynamic co-culture of hBM-MSCs with hSkMs (2:1). hBM-MSCs were cultured in myogenic medium supplemented with 10 ng/mL of bFGF with perfused hSkMs medium up to 21 days. The number of the well were loaded to maintain the same cell ratio of 2:1, respectively, as studied in static culture. Each well has an internal diameter of 2 cm; the mean flow rate was 1 mL/min. Figure 6. Gene expression profiling of myogenic markers and cytokines expressed by hBM-MSCs within perfusion bioreactor culture. mRNA levels of myogenic markers (Pax 3, MyoD 1, Myf 5, Myf 6, Desmin and MYH2), pro-inflammatory cytokines (IL-6, IL-12A, IL-1b) and anti-inflammatory cytokines (IL-10, IL-4, TGF-b1, TGF-b2) were monitored. Relative quantification of each mRNA gene expression normalised to endogenous GAPDH (internal control) was calculated using the 2 ÀDDCt method and presented as fold change over hBM-MSCs and hSkMs collected at day 0 ¼ 1. Two-way analysis of variance (ANOVA) with Tukey's multiple comparison test was performed for comparison of data obtained at Day 0 in perfusion system culture. n ¼ 3. Ã p < .05, ÃÃ p < .01, ÃÃÃ p < .001.

Discussion
bFGF supplemented medium showed low myogenic potential, whereas co-seeding hBM-MSCs with hSkMs promoted myogenic commitment events when a 2:1 ratio was adopted.Cell co-culture system was achieved by conventional static culture and by using a perfusion bioreactor to promote cell cross-talk by paracrine via. In both cases, the whole myogenic gene expression was progressively upregulated, probably due to a complex biochemical cell-to-cell cross-talk [44,45]. The expected commitment was checked by evaluating all MRF gene expressions, that act in a concerted manner and with a specific spatio-temporal expression during myogenesis. Indeed, the early factors (MyoD 1 and Myf 5) are involved in the commitment and proliferation of myogenic directed cells, and late factors (Myf 6) regulate terminal differentiation of committed cells, even if they can functionally overlap [46]. The hSkMs constitutively and constantly expressed these genes along the culture therefore when the same genes were upregulated in hBM-MSCs, an effective myogenic commitment of stem cells was assumed.
The co-seeding system in a static culture environment promoted myogenic events but has poor versatility, and the recovery of single-cell population for proper investigation is not even possible. On the contrary, co-culture adopting the perfusion system ensured successful outcomes and allowed cell collection independently. Additionally, it can be considered an excellent system to investigate in vitro the myogenic commitment by paracrine via. Indeed, in such a case, gene expression seemed to follow the same progression of commitment events previously observed in co-seeding experiments, suggesting that a paracrine effect could induce cell commitment versus a specific phenotype, probably thanks to complex cross-talk even in absence of direct cell-to-cell contact. In perfused systems, hBM-MSCs expressed and upregulated the same genes, such as Pax 3 and the whole MRFs. Pax 3 was overexpressed at day 7 (11.1-fold) and then its expression decreased; the same behaviour was observed for Myf 5 and Myf 6 that were significantly upregulated at day 7 of culture (38-fold, p < .01; 37-fold, respectively), followed by a decreasing trend observed at day 21. Desmin expression also significantly increased at day 14 (16-fold, p < .05), Figure 7. Immunofluorescence (IF) images of hBM-MSCs with myogenic medium and under co-culture within perfusion bioreactor. IF assay at days 7, 14 and 21 of culture was performed by staining Desmin in red and MYH2 in green. The images illustrated hBM-MSCs cultured in myogenic medium (a) and hBM-MSCs cultured in co-culture thanks to the perfusion system, in such case cells orientation and elongation and the presence of the myogenic protein are evident (b). All images were captured using 20Â magnification. Scale bar: 200 mm. Semi-quantitative IF of Desmin and MYH2 protein signals from hBM-MSCs is shown as mean ± SD, and two-way ANOVA performed between time points, n ¼ 3, Ã p < .05, ÃÃ p < .01, ÃÃÃ p < .001 and ÃÃÃÃ p < .0001 (c). whereas MYH2 started to be overexpressed significantly (4fold, p < .001) at day 7. Despite an apparently not largely overexpression of Desmin and MYH2 genes, with respect to all MRFs, when protein characterisation was performed through western blot assay, a clear presence of both Desmin and MYH2 protein bands was monitored confirming the proper myogenic commitment. Both proteins follow the same expression trend with the maximum peak at day 14, with Desmin more abundant than MYH2, because t Desmin constitutes intermediate filaments, and its expression is temporarily earlier than MYH2, that takes part of the contractile system.
The biochemistry behind how hBM-MSCs acquire myogenic phenotype is complex and many aspects remain still unclear. The unique commitment of hBM-MSCs towards myogenic phenotype occurs when Pax 3 signalling activates [46,47]; indeed, Pax 3 seems to act independently from Notch pathway, as Hes 1 and Hes 5 are not upregulated. Moreover, Pax 3 is the first transcription factor of myogenesis that regulates MRFs expression and blocks other potential mesenchymal differentiation events, such as adipogenesis, osteogenesis and chondrogenesis. In our study, in all experimental conditions, Pax 3 showed a good expression, reducing over culture time in favour of MRFs expression; thus, we could assume that hBM-MSCs were successfully committed towards myogenic phenotype. Some authors also described that the canonical Wnt pathway leads to b-catenin stabilisation, that, once in the nucleus, can induce the expression of target genes, such as MRF expression [47]. However, further investigations are required to better characterise the whole Wnt pathway. In this sense, our in vitro model may be also useful for these studies. In almost all experiments, best gene overexpression and protein production were monitored at day 14 of culture; this can be probably due to the fact that long-term culture can affect biological activity of MSCs, as largely reported in literature [48] Since hBM-MSCs show an immunomodulatory capacity, a broad range of cytokines has been investigated to understand which are expressed during the commitment process, considering that all of them play a pivotal role during myogenesis. For example, the role of TNF-a during myogenesis is still controversial and some authors reported TNF-a as an autocrine factor as it is constitutively expressed by myoblasts and largely released under differentiation stimuli [49,50]. On the other hand, TNF-a mostly contributes to myoblast proliferation by accelerating cell cycle progression [51,52]. In our study, hSkMs showed a low expression of TNF-a and its expression was not detected in perfusion system culture even if it increased 8-fold in co-seeded static culture ( Figure  4). IL-1 is involved in myogenesis and seems to favour myoblast mitosis, enhancing muscle protein synthesis [53]. However, conflicting data are reported on IL-1 role because prolonged exposure can allow a reduction in myotube width and actin levels [54]. Detected expression of IL-1b was very high in hSkMs up to 205-fold, while hBM-MSCs under bFGF medium expressed it only 4-fold (p < .01) at 7 days; in coseeded static culture, the expression of this cytokine increased at 6-fold (p < .01) at 14 days, and it was overexpressed 1-fold at day 7 in the dynamic system. IL-10 is an anti-inflammatory cytokine mostly involved in skeletal muscle regeneration, favouring newly regenerating myofibers [55]. The highest expression value of IL-10 was detected in hSkMs (89.9-fold p < .01) at day 14, followed by co-seeding static culture (38-fold p < .05), at the same time point. In the perfusion system, the expression of IL-10 was lower than static co-seeding culture, but one of the most upregulated (19.2-fold, p < .05) at day 21. The same trend was detected for IL-4. This cytokine promotes muscle differentiation [56,57]. In hBM-MSCs static co-seeding, the expression of IL-4 is high up to 55-fold and, in the perfusion system, IL-4 is one of the most upregulated of 11-fold (p < .05) at same day 14. TGFb family members do not exceed 4-fold, both in static and in perfusion system.
In static co-seeding culture, the overall expression of proinflammatory cytokines was balanced by the high expression of anti-inflammatory ones, such as IL-10 (38-fold, p < .05) and IL-4 (55-fold). In the perfusion system, pro-inflammatory cytokines showed an overall poor expression; conversely, anti-inflammatory ones were the most expressed. The overall data indicated that the perfusion system promoted myogenic commitment with a lower expression of pro-inflammatory cytokines and overexpression of anti-inflammatory ones.

Conclusions and perspectives
The present work described myogenic commitment of hBM-MSCs by co-culture with hSkMs in a perfused system. The proposed culture seemed more advantageous than the conventional static one, especially owing to easy cell collection. The described system opens perspectives for further studies on more complex cross-talk mechanisms that may involve multiple cell phenotypes. Indeed, the perfused system will easily allow a multiple cell culture with several cell phenotypes to better investigate their specific cross-talk and how they participate to activate molecular signals or cytokines that can exert a fine regulation for muscle regeneration and healing. As an example, the proposed in vitro model can be improved with Peripheral Blood Mononuclear Cells (PBMCs) to better understand the role of those cells in myogenic events, alternatively, the perfusion may allow the circulation of biomimetic vesicles fabricated for the targeted delivery of specific signals, in both physiological and pathological simulated conditions.
A deep understanding of the whole biology involved behind stem cell differentiation is still a challenge, and the described model may open perspectives for a comprehensive exploration of stem cell behaviour along with myogenic commitment or for studying a more complex cross-talk between stem cell and myoblast phenotype in health and disease.

Disclosure statement
The authors declare that there is no conflict of interest regarding the publication of this paper. The funders had no role in the design of the study, in the collection, analyses, or interpretation of data, in the writing of the manuscript, or in the decision to publish the results.

Funding
The authors gratefully acknowledge financial support from Athena Srl 139, Viale Europa À 50126 Firenze (IT) under the Research Contract entitled "The design and implementation of a 3 D bioengineered model for study on skeletal muscle regeneration and inflammation processes. The year 2020-2021. Principal Investigators: Prof. Nicola Maffulli and Giovanna Della Porta. PS is the recipient of a PhD grant Cycle XXXV in Translational Medicine at Dept. of Medicine, Surgery and Dentistry, the University of Salerno entitled: "Skeletal muscle regeneration modulated by inflammation: an in vitro study on a 3 D bioengineered system" funded by Athena Srl 139, Viale Europa À 50126 Firenze (IT).

Author contributions
Pasqualina Scala developed the experimental activity and optimised the protocols and methodology; she was responsible for the paper draft preparation and revision; Joseph Lovecchio designed the bioreactor and related operative protocols; Erwin Lamparelli contributed to the experimental activity and qRT-PCR data acquisition; Rossella Vitolo and Valentina Giudice isolated the stem cells and characterised them with formal analysis and validated methodology; Emanuele Giordano supervised the bioreactor protocols; Carmine Selleri provided the methodology for hBM-MSCs cultivation; Laura Rehak provided contribution in supervision and paper writing; Nicola Maffulli helped in the interpretation of the data, reviewed the manuscript and was responsible for funding acquisition; Giovanna Della Porta was responsible for experimental data design, production, curation and supervision, paper writing and editing, funding acquisition and research project administration. Data availability statement All data achieved and described in the present work are available at the following link: https://drive.google.com/drive/folders/1Z-Iufq9eekpTHC wvWFNtXoGrJ4o6AXaW?usp=sharing.