PBMCs were collected by HemaTrate^® Blood Filtration System, following same protocol of clinical practice in limb rescue (De Angelis et al., 2015; Persiani et al., 2018), Whole peripheral blood (PB) obtained from three healthy volunteers (M/F, 2/1; age ranged from 38 to 45 years old) was concentrated using a HemaTrate^® Blood Filtration System (CH-WB110C, Patent n. EP 2602315A1; Pall Medistad B.V. (Medemblik; Netherlands) from Cook Regentec (Indianapolis, Indiana, United States). After filtration, PBMCs were isolated by Ficoll-Paque density gradient centrifugation (Cytiva, Marlborough, Massachusetts, United States), according to manufacturer’s instructions, for flow cytometer immunophenotyping. Conversely, filtrated PBMCs were directly used for 3D culture experiments, or were stored in 70% RPMI (Gibco™), 20% Fetal Bovine Serum (FBS, Gibco™), and 10% DMSO (Sigma-Aldrich, Milan, Italy) at -80°C until use.

BM specimens for hBM-MSC isolation were obtained from three healthy male donors (aged 26, 24, and 28 years old) after informed written consent in accordance with the Declaration of Helsinki and protocols approved by our Institutional Review Board (Ethic Committee “Campania Sud”, Brusciano, Naples, Italy; prot./SCCE n. 24,988). BM aspirate culture was described elsewhere (Giordano et al., 2014). BM-MSCs were subsequently seeded at 4,000 cells/cm² and expanded up to the third passage. Mesenchymal phenotype was confirmed according to the International Society of Cellular Therapy guidelines: i) ability to adhere to tissue culture plastics; ii) fibroblast-like spindle shape; and iii) characteristic immunophenotype by flow cytometry with positivity for CD90, CD105, and CD73, and negativity for CD34, CD14, CD45, and HLA-DR (Supplementary Figure S1) (Dominici et al., 2006). hBM-MSCs were co-cultured in a 3D system with hSkMs, with and without PBMCs. All experiments were performed in biological triplicates (N = 3).

hSkMs were purchased from (Gibco™). hSkMs were embedded in fibrin scaffolds to investigate their behavior in a 3D environment through qRT-PCR, as reported in the following section. Then, hSkMs were co-cultured with hBM-MSCs with and without PBMCs. Seeding ratio between hBM-MSCs:hSkMs was 2:1, as previously optimized in (Scala et al., 2022). All experiments were performed in biological triplicates (N = 3).

MSC and PBMC immunophenotype was investigated by flow cytometry. Briefly, for BM-MSCs, a minimum of 1 × 10⁵ cells at the third passage was stained with the following antibodies: 2.5 μl of fluorescein isothiocyanate (FITC) - conjugated anti-CD90 or 5 μl of FITC - conjugated anti-HLA-DR; 5 μl of allophycocyanin (APC) - conjugated anti-CD73; 10 μl of phycoerythrin (PE) - conjugated anti-CD105 or 10 μl of PE - conjugated anti-CD34; and 10 μl of phycoerythrin cyanin 7 (PC7) - conjugated anti-CD45 or 10 μl of PC7 - conjugated anti-CD14 (all antibodies from Beckman Coulter, Fullerton, CA, United States). Cells were incubated at room temperature (RT) for 20 min in the dark, washed with phosphate buffered saline (PBS, Gibco™), and resuspended in 300 μl of the same buffer for acquisition. For PBMC immunophenotyping, before and after filter and separation procedures, a minimum of 2 × 10⁵ cells were stained with the following antibodies: 5 μl of APC–conjugated anti-CD3; 5 μl PC7 - conjugated anti-CD14; 5 μl of FITC–conjugated CD34; 5 μl of phycoerythrin cyanin 5 (PC5)—conjugated CD8; and 5 μl of PE - conjugated anti-CD4. Cells were then incubated at RT for 20 min in the dark, washed with PBS and resuspended in 300 μl of the same buffer for acquisition. Sample acquisition was performed on a BD FACSVerse flow cytometer (Becton Dickinson, BD, NJ, United States) equipped with blue (488 nm) and red lasers (628 nm) and BD FACSuite software (BD Biosciences). PMT voltage setting, and compensation were carried out using single-color controls for each fluorochrome and an unstained sample as negative control. All samples were run with the same PMT voltages, and a minimum of 30,000 events were recorded. FlowJo software (v.10.7.1, LLC, BD Biosciences) was employed for post-acquisition compensation and analysis.

BM-MSCs were first identified using linear parameters (forward scatter area (FSC-A) vs. side scatter area (SSC-A), and double cells were excluded (area vs. height, FSC-A vs. FSC-H) (Supplementary Figure S1).

Cylindrical 3D scaffolds were prepared by mixing fibrinogen at 50 or 100 mg/ml (Sigma-Aldrich), α-aprotinin 15,600 U/ml (Sigma-Aldrich), α-MEM (Corning, NY, United States), and thrombin at 100 U/ml (Sigma-Aldrich), and then were incubated at 37°C for 30 min to allow fibrinogen polymerization (Supplementary Figure S2A). Samples were fixed in 4% paraformaldehyde (PFA; 4°C, overnight) and then dehydrated by multiple passages across ethanol: water solutions (10 min each) with increasing percentages of ethanol and dried using a dense carbon dioxide drying operating at 200 bar and 38°C for 4 h; other details on the dense gas drying protocol were reported elsewhere (Della Porta et al., 2013). System technology scheme is illustrated in Supplementary Figure S2B. Samples were immersed in liquid nitrogen and fractured with a needle, then placed on a double-sided adhesive carbon tape previously glued to an aluminium stub and coated with a gold film (250 A thickness) using a sputter coater (mod.108 A; Agar Scientific, Stansted, United Kingdom), before observation. Internal scaffold morphology was observed by field emission-scanning electron microscopy (FE-SEM, mod. LEO 1525; Carl Zeiss, Oberkochen, DE). Pore size frequency distribution was studied with ImageJ software (rel.1.52p National Institutes of Health, United States), and average perimeter and average Feret’s diameter values were reported (Figure 1).

Cylindrical 3D scaffolds were prepared by mixing fibrinogen at 50 mg/ml (Sigma-Aldrich), α-aprotinin 15,600 U/ml (Sigma-Aldrich), and α-MEM (Corning, NY, United States) containing 1 × 10⁶ cells. Two cell suspension solutions were prepared, composed by hBM-MSCs and hSkMs with PBMCs ([hBM-MSCs + hSkMs]: PBMCs, 1:1) and without PBMCs. This ratio was chosen based on previously published literature showing that lower PBMCs:co-cultured cells ratios are preferable (Sturlan et al., 2009; Zhang et al., 2015).

Suspensions were pipetted in different wells of a 96-well plate, thrombin at 100 U/ml (Sigma-Aldrich) was added, and samples incubated at 37°C for 30 min to allow fibrinogen polymerization. 3D scaffolds were then transferred to a culture plate (Supplementary Figure S2C).

3D scaffolds were placed in a perfusion bioreactor, formed by a custom multi-well plate, milled in poly (methyl methacrylate) (PMMA, Altuglas^® CN 100 10000, Altuglas International, La Garenne-Colombes Cedex, FR), a biocompatible material for biomedical application (Samavedi et al., 2014). This plate has two holes allowing the insertion of silicon tubes (Tygon^®, FR) for medium flowing at a constant flow rate of 1.0 ml/min maintained by peristaltic pumps (Pasini et al., 2019). This bioreactor system operates within a standard cell culture incubator (Supplementary Figure S2D).

Cell viability was detected by fluorescence live and dead assay after preparation (day 0) and at each time point (day 7, 14, and 21). Calcein AM solution (Cat. No C1359, Sigma-Aldrich) was used to stain live cells, while cell membrane-impermeable Ethidium homodimer I solution (Cat. No E1903, Sigma-Aldrich) for nuclei of dead cells. Cells were incubated for 1 h at 37°C, then washed in 1X PBS, and imaged using a fluorescence microscope (Eclipse Ti Nikon Corporation, Tokyo, Japan). Signal intensity was quantified using ImageJ software. Original RGB images were converted to 8-bit (grey scale) format, and tagged area intensities were expressed as mean value of pixel intensity within a range from 0 (dark) to 255 (white), as previously reported (Spaepen et al., 2011).

Expression of myogenic genes, including Pax3, Myo D1, Myf5, Myf6, Desmin, and MYH2 (Bio-Rad, Foster City, CA, United States), and cytokines, such as IL6, TNF, IL12A, IL1B, IFNG, IL10, IL4, TGFB1, and TGFB2 (sequences are reported in Table 1) were investigated by reverse transcription quantitative polymerase chain reaction (RT-qPCR).

Total RNA from 3D dynamic cultures at each time point was extracted using QIAzol Lysis Reagent (Qiagen), chloroform (Sigma-Aldrich), and RNeasy Micro Kit (Qiagen). For each sample, 1 μg of total RNA was reverse transcribed using iScript™ cDNA synthesis kit (Bio-Rad), and relative gene expression analysis was performed on a LightCycler^® 480 Instrument (Roche, Basel, Switzerland) using SsoAdvanced™ universal SYBR^® Green Supermix (Bio-Rad). Specificity of formed products was assessed by melting curve analysis. Experiments were run in triplicate.

Data were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression (reference gene) applying the geNorm method (Hellemans et al., 2007) and using CFX Manager software (M < 0.5). Fold changes were determined by 2^−ΔΔCt method and presented as relative levels versus hSkMs and hBM-MSCs at day 0. All experiments were performed in biological triplicates (N = 3), and each experiment in technical triplicate.

Fibrin scaffolds were fixed in 4% PFA for 2 h at RT, cryo-protected in 30% sucrose (4 °C, overnight), included in optimal cutting temperature (OCT) compound, and cut in slices of 10 µm thickness using a cryostat (CM 1950, Leica, Wetzlar, Germany). Slices were permeabilized with 0.1% Triton X-100 for 10 min and blocked with horse serum solution for 1 h. Samples were then stained for Desmin (1:100; Abcam) and MYH2 (1:50, Thermo Fisher Sci.), incubated overnight at 4°C, and subsequently incubated for 1 h at RT with Alexa Fluor ™ 488 goat anti-rabbit IgG (1:400; Thermo Fisher Sci.), VectaFluor™ anti-mouse IgG Dylight 594^® kit (Vector laboratories), and DAPI. Slices were also stained for FITC-conjugated CD90 and PE-conjugated CD105 (Beckman Coulter), incubated overnight at 4°C, and the cell nuclei were counterstained using DAPI. Separate images were acquired using identical settings of light intensity, exposure time, and gain using a fluorescence microscope (Eclipse Ti Nikon Corporation, Tokyo, Japan).

For quantification of secreted cytokines in culture medium, an immunobead-based multiplex assay (Merck, Millipore) was employed for measurement of EGF, Eotaxin, GM-CSF, IFN-γ, IL-10, IL-12p70, IL-1RA, IL-1a, IL-1b, IL-2, IL-3, IL-4, IL-5, IL-7, MIP-1a, MIP-1b, TNF-α, bFGF, G-CSF, GRO, IL-6, IL-8, MCP-1, and VEGF, following manufacturer’s instructions.

Scaffold slices (15 μm thickness) were hydrated using a decreasing ethanol gradient, washed for 5 min in water and incubated with hematoxylin and eosin for 60 min, then dehydrated using an increasing ethanol gradient and cleared in xylene for 5 min. Sections were mounted using Eukitt (Sigma-Aldrich) mounting medium. Images were acquired using an Olympus microscope BX53 equipped with ProgRes SpeedXT ^core five camera.

Data were analyzed using Prism software (v.9.0, GraphPad software, LLC, San Diego, California, United States). Results are presented as mean ± standard deviation (SD). Statistical analysis was performed using two-tailed independent Student’s t-test for two group comparisons, or two-way analysis of variance (ANOVA) test for three or more group comparison with Tukey’s test for multiple comparisons between group. For flow cytometry data, results are presented as percentage of positive cells, and expression of each marker on single cells is also reported as histograms and using unstained samples as negative controls. For characterization of mesenchymal cells in the co-culture system, percentage of positive cells and median fluorescence intensity (MFI) values were calculated for each marker. Surface marker expression variations were calculated as fold change normalizing MFI values for each marker and from each time point to MFI obtained from hBM-MSC cultured alone (Tsai et al., 2020). A p-value < 0.05 was considered statistically significant (de Winter 2013).

Different concentrations of fibrinogen (50 and 100 mg/ml) were tested to check the best concentration to simultaneously assure good scaffold integrity and void spaces with totally interconnected cells within 3D fibrin system. Fibrinogen concentrations of 5 mg/ml was too low to induce polymerization, whereas, scaffold obtained at 20 mg/ml were difficult to be managed and were discharged (unpublished data). To better observe internal scaffold morphology obtained at 50 and 100 mg/ml, hydrogels were converted to aerogels by dense gas drying process, a procedure already documented as capable to avoid the natural shrinkage and collapsing of the 3D hydrated system, normally observed in lyophilization, as previously optimized (Della Porta et al., 2013). Indeed, derived alcohol gels were processed with dense carbon dioxide for 4 h at 38°C and 200 bar with a flow rate of 1.2 kg/h (system technology scheme in Supplementary Figure S2B). Collected aerogels maintained the same volume with a shrinkage <2%. Thanks to this extremely low shrinkage scaffold morphology was investigated after their freeze-fracture and the internal void structure was observed by FE-SEM (Figure 1). Scaffolds obtained using 50 mg/ml of fibrinogen showed thinner and wider meshes than those obtained with 100 mg/ml of fibrinogen. Pores of the 50 mg/ml of fibrinogen scaffold showed a mean perimeter of 548 nm and a mean Feret’s diameter of 180 nm, while at 100 mg/ml of fibrinogen, pores had smaller perimeter (mean, 128 nm) and Feret’s diameter (mean, 40 nm). Therefore, scaffolds made with 50 mg/ml of fibrinogen were chosen for further experiments, because they displayed a more spacious microenvironment suitable for a better cell distribution and viability.

Next, a myogenic commitment model of hBM-MSCs-hSkMs co-cultured (ratio 2:1) with or without PBMCs was established using scaffolds composed by fibrinogen at 50 mg/ml as support for cell culture and placed in dynamic conditions using a custom-made perfusion bioreactor for maintaining a constant medium flow rate of 1 ml/min (Della Porta et al., 2015) (Supplementary Figures S2C,D). To confirm suitability of this in vitro system before performing further experiments, cell viability within fibrin scaffolds was assessed by live and dead assay over a culture period of 21 days. As expected, total cells were almost represented by live cells (>80%) throughout the culture (Figure 2). Furthermore, fibrin scaffolds also maintained structure and integrity, as cells were homogenously distributed and were in close contact throughout the culture period. This regular distribution was also documented by histological evaluation of 3D scaffold culture slices using hematoxylin and eosin staining (Figure 3). Furthermore, fibrin matrix behaves as inert matrix not really influencing myogenic marker expression of hBM-MSCs-hSkMs co-culture; indeed, when, those cells were co-seeded in 2D culture, almost similar gene expression profile has been observed (Scala et al., 2022).

First, mesenchymal phenotype of our primary hBM-MSCs was confirmed following International Society of Cellular Therapy criteria (Dominici et al., 2006): adherence to tissue culture plastics; fibroblast-like shape; positivity for CD90, CD105, and CD73 mesenchymal markers; and negativity for CD34, CD14, CD45, and HLA-DR by flow cytometry (Supplementary Figure S1). Then, PBMCs were collected by filtration of 120 ml of whole blood adopting a system conventionally used in clinical practice (HemaTrate^® system), and nucleated cell fraction was concentrated in 10 ml of physiological solution. PBMC immunophenotyping and whole blood composition were investigated by flow cytometry on cells separated by Ficoll-Paque density gradient; filter waste bag was considered as a negative control for each donor employed (Figure 4). No significant variations were described in blood frequency and composition before and after filtration or separation, while no PBMCs were observed in specimens obtained from filter waste bags (Figure 4). Moreover, no differences were described for total CD3⁺ T Cells and CD4⁺/CD8⁺ subpopulations (all p > 0.05), and for circulating CD34⁺ hematopoietic stem cells (p = 0.2875), confirming that HemaTrate^® filter was highly efficient in selective PBMC collection and concentration in a 10 ml final volume of filtrate.

Myogenic commitment potential of our 3D system was investigated by gene expression profiling (Figure 5). First, hSkMs behavior in 3D culture was explored, as shown in Figure 5A hSkMs showed a significant upregulation of Pax3 and MYH2 at day 21 (258-fold and 261-fold, respectively, p < 0.0001), as well as MyoD1, Myf5, Myf6, and Desmin genes with an increasing trend from day 0–21. This behavior was reported for comparison purpose.

Then, in Figure 5B gene expression profiling of hBM-MSCs-hSkMs in 3D co-culture with and without PBMCs is showed. Pax3 expression levels significantly differed between the two culture conditions over time. In details, Pax3 expression was upregulated at day 7 in the presence of PBMCs (20-fold), while decreased at day 14 (7-fold), and significantly upregulated at day 21 (27-fold, p < 0.05). A similar trend was described in the absence of PBMCs (from 8.6-fold at day 7 to 6.4-fold at day 14 and 11.3-fold at day 21).

For MRF genes, MyoD1 expression showed lower fold-change values than those without PBMCs at day 7 (28.35-fold vs. 70-fold); however, at day 14 in the presence of PBMCs a significant upregulation occurred (118-fold, p < 0.05) as well as at day 21 (81-fold vs. 51-fold). In cultures with PBMCs, Myf5 expression was increased about 220-fold and without PBMCs about 329-fold (p < 0.05) at day 7, while at day 14 slightly decreased with PBMCs (152-fold with vs. 186-fold without PBMCs), remaining stable at day 21 (136-fold with vs. 99-fold without PBMCs). Generally, Myf6 expression decreased along the culture with PBMCs, from 215-fold (p < 0.5) at day 7, to 75.2-fold at day 14, and 31-fold at day 21. Without PBMCs, the expression was of 165-fold at day 7, about 105-fold at day 14, and decreased at 64-fold at day 21. Desmin expression displayed an increasing trend with PBMCs about 3-fold at day 7, 5-fold at day 14, and 6-fold at day 21. When cultured without PBMCs, Desmin expression was slightly upregulated at day 7 and 14 (10-fold and 9.5-fold, respectively; p < 0.01), and at day 21 of 8-fold (p < 0.05). MYH2 was also investigated but not detected at any time point along the culture. All these profiling were also different from those observed for the hSkMs alone in 3D culture, suggesting a fundamental role of the different cell population cross-talk within the 3D system.

To further confirm myogenic commitment of hBM-MSCs with or without PBMCs in a 3D culture system, myogenic-related proteins, such as Desmin and MYH2, were assayed by IF (Figures 6, 7). In the presence of PBMCs, Desmin was detectable from day 7 with a signal that became brighter throughout the culture. Despite not well detected by qRT-PCR, MYH2 protein was highlighted by IF assay at day 21 days but only in the culture with PBMCs. Furthermore, almost all 3D cultures of hBM-MSCs-hSkMs without PBMCs displayed IF signals less intense at each time point compared to cultures with PBMCs (Figure 6). CD90 and CD105 protein expression was also investigated to confirm mesenchymal phenotype at baseline. Signals were clearly detected at day 0, confirming the mesenchymal phenotype of cultured hBM-MSCs, while very low or undetectable expression was observed at day 21 (Figure 7).

We next explored paracrine and cell-to-cell contact effects of PBMCs on cytokine expression in hBM-MSCs-hSkMs cultured in a 3D in vitro perfused model (Figure 8). hSkMs showed a significant upregulation of pro-inflammatory cytokine as TNF, 478-fold (p < 0.0001) and IL12A, 14-fold (p < 0.5) (Figure 8A). In co-culture system, pro-inflammatory cytokines were downregulated with PBMCs, especially IL6 (0.04-fold at day 7, 0.1-fold at day 14, and 0.05-fold at day 21; p < 0.01). At day 14 and 21, IL12A increased of 6-fold and 17-fold with PBMCs, and of 75-fold and 13-fold without PBMCs; similarly, IL1B increased at 6-fold and 15-fold with PBMCs, while 20.5-fold and 8-fold without PBMCs. Conversely, among investigated anti-inflammatory cytokines, IL10 was the most expressed and was higher in the presence of PBMCs (23.3-fold vs. 11.5-fold at day 7, or 9-fold vs. 5-fold at day 14, with vs. without PBMCs, respectively). TGFB1 showed an increasing trend in the presence of PBMCs from 1.8-fold at day 7 to 9.5-fold at day 14 and 12-fold at day 21. TGFB2 displayed the maximum expression at day 14, about 7-fold without PBMCs. TNF, IL4, and IFNG were not detected (Figure 8B).

Cytokine secretion was also investigated by analyzing culture medium supernatants at 7, 14 and 21 days by multiplex bead-based immunoassay (Figure 8C). Among 24 cytokines studied, growth-regulated alpha protein (GRO; 2.5 ng/ml), IL-6, granulocyte-colony stimulating factor (G-CSF), IL-8, monocyte chemoattractant protein-1 (MCP-1), and vascular epidermal growth factor (VEGF; 0.1 ng/ml) were detected in culture medium at day 7. Remaining cytokines were not present in culture medium supernatants at day 7, and no cytokines were detected at day 14 and 21.