Type I collagen mesh (Col) from bovine tendon (5% collagen gel in acetic acid) was produced by a solvent-casting method as previously described (Taraballi et al., 2013; Minardi et al., 2016c). Briefly, 1 g of type I collagen (Nitta Casings Inc.) was dissolved in an acetate buffer (pH 3.5) at the final concentration of 20 mg/ml. The collagen suspension was precipitated by the addition of sodium hydroxide (0.1 M) solution at pH 5.5. The resulting collagen slurry was mild crosslinked with 1.5 mM solution of 1,4-butanediol diglycidyl ether (Sigma-Aldrich) for 48 h. The crosslinked collagen was washed three times in water, and then the slurry was molden in metal racks, at a thickness of 2 mm and air dried under a fume hood for 3 days (final thickness: 0.1 mm). Covalent functionalization of heparan sulfate (HS) to collagen meshes was performed using the coupling reaction with 1-ethyl-3-(3-dimethyl aminopropyl)carbodiimide (EDC) and N-hydroxysuccinimide (NHS). After weight the collagen matrices, each sample was incubated in a solution of 50 mM 2-morpholinoethane sulfonic acid (pH 5.5) in ethanol. After 1 h, the samples were place in a solution of 5 mM EDC and 5 mM NHS and 0.2% (w/v) heparan sulfate (HS). After reaction for 24 h, the matrices were washed three times in PBS. For the stability study, collagen meshes have been incubated with the same concentration of HS in the absence of crosslinker.

Rat compact bone MSC were isolated as previously reported (Corradetti et al., 2015) and maintained in standard culture media represented by Dulbecco’s Modified Eagle’s Medium (Sigma) supplemented with 10% fetal bovine serum (Euroclone), 1% l-glutammine (Gibco), and 1% antibiotic/antimycotic solution (containing penicillin, streptomycin, and amphotericin B, Gibco) at 37°C in a humidified atmosphere with 5% CO₂. For maintenance of cultures, cells were plated at a cell density of 5 × 10³ cells/cm². Medium was changed twice per week thereafter or according to the experiment requirements. Adherent cells were detached and subcultured using TrypLE Express (Invitrogen, ThermoFisher Scientific) before reaching confluence (80%) and subsequently replated at the same density for culture maintenance (Corradetti et al., 2016). When seeded onto meshes, MSC were harvested, resuspended in standard cell culture medium, and seeded at the density of 1 × 10⁴ cells/well in 12-well plates (Corning). Inserts (Corning) were used to hang membranes (Col and HS-Col) and allow for cell adherence onto the collagen without reaching well surface. Standard culture medium was then added to each well. Animal studies were conducted following approved protocols (AUP-0115-0002) established by Houston Methodist Research Institute’s Institutional Animal Care and Use Committee in accordance with the guidelines of the Animal Welfare Act and the Guide for the Care and Use of Laboratory Animals.

Upon reaching confluence, cells were collected and characterized for the expression of MSC-associated markers (CD105, CD90, and CD105) with a FortessaTM cell analyzer (Becton Dickinson), under the assistance of the HMRI Flow Cytometry Core. Antibodies used (APC/Cy7-CD73, PE/Cy7-CD105, and Alexa Fluor 700-CD90) were purchased from BioLegend. Briefly, MSC were recovered by trypsin method and by spinning down at 500 g for 5 min. They were then washed with FACS buffer labeled with directly conjugated antibodies according to the manufacturer’s indications.

At passage 3, cells were also assessed for their capability to undergo osteogenic and chondrogenic differentiation following a previously reported procedure (Corradetti et al., 2015). Briefly, to induce osteogenesis, MSC were seeded at the density of 5,000 cells/cm² in 12-well plates. Osteogenic induction was performed over 14-day period using a StemPro Osteogenesis Differentiation Kit (Gibco). To confirm differentiation, conventional von Kossa was performed. For chondrogenesis, cells were seeded at the density of 5,000 cells/cm². Induction was performed using the StemPro Chondrogenesis Differentiation Kit (Gibco) for a 14-day period. ECM constituted by proteoglycans was visualized by conventional Alcian Blue staining.

To test the effect of HS in retaining fibroblast-growth factor (FGF) and improve cell proliferation, MSC were exposed to different conditions, and the growth rate recorded over a 7-day period was compared. Conditioned included the following: (1) MSC grown in 2D culture systems (CTRL) and treated with FGF (10 ng/ml, w/FGF; PeproTech), soluble HS (25 ng/ml, w/sHS; Carbosynth), and a combination of both (w/sHS and FGF), (2) MSC grown onto Col in presence of soluble HS (Col w/sHS), FGF (Col w/FGF), and a combination of the two (Col w/sHS and FGF), and (3) MSC grown onto HS-Col in presence in presence (HS-Col w/FGF) and absence (HS-Col) of FGF. Experiment was set in triplicate. MTT assay (Life Technologies) was performed according to the manufacturer’s procedures. The concentration of FGF to be used was previously defined in a range of 1–100 ng/ml (data not shown).

Mesenchymal stem cells were seeded at the density of 100,000 cells/cm² onto HS-Col, Col meshes and cultured for 24 h to assess their immediate response to the material in terms of proliferation, morphology, adherence, and MSC-associated genes expression. Cell proliferation onto membranes was evaluated by Alamar Blue (ThermoFisher Scientific) according to manufacturer’s instructions during a 3-day period as previously reported (Corradetti et al., 2016). Optical density was measured at wavelengths of 570 and 600 nm. Data are shown as mean of three independent biological replicates. Values have been reported as %AB over time, which is associated with the presence of metabolically active cells. For comparison, data obtained from MSC grown in two-dimensional conditions (CTRL) were also acquired.

Seven days after seeding, cellularized patches and 2D controls were washed and fixed in 4% PFA for 20 min. Cell membranes were then permeabilized with 0.1% Triton X-100 for 10 min and then blocked by 10% goat serum in PBS for 1 h at RT. Cells’ actin cytoskeletons were stained with phalloidin (Alexa Fluor 555, Invitrogen by Life Technologies) and DAPI to identify actin cytoskeleton at a dilution of 1/100 in blocking solution at RT in darkness, and subsequently the nuclei stained with DAPI (ThermoFisher Scientific Inc.) according to the manufacturer’s instructions. Samples were imaged with a confocal laser microscope (A1 Nikon Confocal Microscope), and the images were analyzed through the NIS-Elements software (Nikon). Col and HS-Col sample were imaged by confocal laser microscopy, and Z-stacks were acquired.

To evaluate the efficacy of HS in supporting MSC retain their immunosuppressive potential, cells were seeded at the density of 100,000 cells/cm² onto HS-Col, Col meshes and cultured for 24 h at 37°C before stimulation with pro-inflammatory cytokines occurred. Stimulation was performed as previously reported (Hegyi et al., 2012; Corradetti et al., 2016). Briefly, a combination of rat recombinant IFN-γ and TNF-α (20 ng/ml; PeproTech) was used and exposure lasted for 48 and 72 h. Cells cultured in 2D conditions (CTRL), whether stimulated (Positive CTRL) or not, were used as negative and positive controls, respectively. The expression of immunosuppressive markers was determined in also on MSC grown onto meshes in absence of stimulation. At each time point, the levels of expression of genes associated with MSC immunosuppressive potential were analyzed as follows.

Peripheral blood mononuclear cells were obtained from heparinized whole blood samples using density gradient centrifugation (Lymphoprep; Axis-Shield, Oslo, Norway) following the manufacturer’s instructions. PBMC proliferation was induced as previously reported (Corradetti et al., 2016). The effect of human MSC grown onto meshes (either Col or HS-Col) on T lymphocyte proliferation was determined after 48 h in a cell–cell contact setting with MSC and PBMC plated at a 1:10 ratio by using CellTrace CFSE Cell Proliferation Kit (Life Technologies). Briefly, 24 h after seeding the MSC onto meshes, PBMC were stained with CFSE following the manufacturer’s indications and cocultured. The percentage of proliferating PBMC in each experimental condition was analyzed by flow cytometry (Millipore).

Total RNA was extracted from cells using TRI Reagent^® (Sigma) and complementary DNA (cDNA) was synthetized from RNA using iScript™ cDNA Synthesis Kit (BioRad) according to the manufacturer’s protocol. Gene expression analysis was performed using specific oligonucleotide primers, designed based on rat (Rattus norvegicus) specific sequences and reported in Table 1. Pro-inflammatory markers include interleukin-6 (Il-6), Tnf-α, and inducible nitric oxide synthase (iNos). Immunosuppression-associated markers include prostaglandin E2 synthase (Pges2), transforming growth factor-beta (Tgf-β), and cyclooxygenase (Ptgs2). Mesenchymal-associated markers include Cd90, Cd29, and Cd73. Quantitative (real-time) PCR was executed using StepOne™ Real-Time PCR System with SYBR^® green method. The presence of pro-inflammation markers was tested on each sample, while immunosuppression-associated markers analysis was performed only on non-inflamed samples, at different growth conditions. Glyceraldehyde 3-phosphate dehydrogenase (Gapdh) was used as reference gene. Each sample was analyzed in triplicate.

Differentiation assays were used to establish the plastic potential of the cells used in the study as well as to understand whether the structural and chemical stimuli provided by the meshes could induce cell reprogramming toward two lineages of the mesodermic layer. MSC grown in 2D and/or in 3D (Col or HS-Col) conditions were harvested following 14 days from induction or culture in standard media. RNA was extracted using RNeasy mini kit (Qiagen). cDNA was reverse transcribed from 150 ng of total RNA with iScript retrotranscription kit (BioRad Laboratories, Hercules, CA, USA). Amplifications were carried on using TaqMan^® Fast Advanced Master Mix on a StepOne Plus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The following target probes (Applied Biosystems) were applied to evaluate the expression of: osteocalcin (Bglap) and alkaline phosphatase (Alp) for osteogenic differentiation. Transcription factors Sox-9 (Sox9) and Aggrecan (Acan) were evaluated for chondrogenic differentiation. The expression of each gene was first normalized to the level of housekeeping gene Beta Actin (Actb). Relative fold changes were normalized to the respective control. Technical triplicates for each biologic sample were evaluated, and results were reported as mean ± SD.

Statistical analysis was performed by using GraphPad Instat 3.00 for Windows (GraphPad Software, La Jolla, CA, USA, http://www.graphpad.com/). Three replicates to evaluate the metabolic state of the cells and gene expression were performed. The results are reported as mean ± SD, with p < 0.05 used as a threshold for significance. One-way analysis of variance for multiple comparisons by the Student–Newman–Keuls multiple comparison test was used.

Collagen type I from bovine tendon was used for the preparation of meshes by the solvent-casting method. Since collagen contains several free amines groups (32/1,000 amino acid residues), we decided to use that as graph point for the dock of heparan sulfate (HS) that is characterized by the free carboxyl groups to obtain collagen type meshes functionalized with HS (HS-Col) (Figure 2A). The topography of the surfaces has been analyzed by SEM. Figure 2B shows the topography of the HS-Col demonstrating that the crosslink does not affect the collagen nanostructure. The presence of HS is confirmed by FTIR [Figure 2C shows the averaged (n = 3) of FTIR spectra of Col and HS-Col]. Amide I (1,700–1,600 cm⁻¹) and Amide II (1,600–1,500 cm⁻¹) are related to the stretching vibration of C=O bonds, and C–N stretching and N–H bending vibration respectively and are present in both samples, as well as the Amide III region (≈1,200–1,300 cm⁻¹). In HS-Col, we found at ≈1,080 cm⁻¹ (C–O stretching of carbohydrate residues in collagen and proteoglycans), ≈845 and ≈1,120 cm⁻¹ (C–O–S stretching), and ≈1,397 cm⁻¹ (COO– stretching of amino side chains) vibrational modes of GAGs (Corradetti et al., 2016). We also performed a simple Alcian Blue staining to confirm the presence of HS on the collagen surface (Figure 2D). After mounting the membrane on the insets (Figure 2E), we performed the ELLA. Although the presence of HS at the first time point was not significant different, the stability of the HS adsorbed is significantly lower, in fact after 14 days we lost the half concentration of HS onto the surface (Figure 2F).

After ensuring the cell population was homogeneously positive for MSC-associated markers (Figure S1A in Supplementary Material) and confirming its differentiative potential toward osteogenic and chondrogenic lineages (Figure S1B in Supplementary Material), the efficacy of the proposed HS-Col meshes in acting as synthetic ECM and its capability to localize growth factors and improve MSC proliferation were assessed in the context of FGF over a 7-day period. MTT assay revealed that the presence of crosslinked HS allows for the fine control of FGF provided to cells (Figure 3). An overall decrease in cell proliferation was observed when cells were grown onto 3D structures, both Col and HS-Col compared to controls (CTRL), although the group HS-Col w/FGF showed a statistically significant (p < 0.01) increase compared to Col w/sHS and FGF at days 3 and 7. A block in cell growth was observed in the experimental group Col w/sHS and FGF, and no differences were shown between MSC grown in standard conditions supplemented with FGF (CTRL w/FGF) and a combination of FGF and sHS (CTRL w/sHS and FGF).

Alamar Blue assay demonstrated that both HS-Col and Col supported metabolically active cells growth during a period of 3 days (Figure 4A). As revealed by optical microscopy after 7 days of culture (Figure 4B), MSC spread across the meshes and reached confluence, as compared to 2D control (CTRL). Cells grown in presence of Col and HS-Col displayed a more spindle-shaped morphology compared to CTRL. To compare the cell structures formed by MSC adhering to meshes (both Col and HS-Col), antibodies to actin was used, and tridimensional images have been analyzed at confocal microscope (Figure 4C). Diffused cytoplasmic labeling was demonstrated when cells were seeded onto Col meshes, together with an intense staining confirming similar intracellular cytoskeleton organization. MSC grown onto HS-Col displayed a more elongated shape than their Col counterparts and maintained physical contact among them through extensions without completely spreading.

Flow cytometry demonstrated a slight reduction in the percentage of MSC-associated markers at 3 days compared to controls (CTRL) following MSC culture onto HS-Col and Col (Figure 4D). To understand the effect of HS-Col on MSC at a genetic level, real-time PCR was used to evaluate the expression of mesenchymal-associated markers, including Cd29, Cd73, and Cd90 (Figure 4E). Molecular analysis showed that at 72 h the expression of integrin Cd29 was significantly (p < 0.01) higher in cells seeded onto Col than HS-Col and 2D controls (CTRL). A statistically significant increase compared to control (CTRL) in the expression levels of the same gene was found in HS-Col. The expression of Cd73 and Cd90 in cells treated with HS-Col was found lower than 2D control (0.475-fold and 0.155-fold, respectively). No significant differences were identified for the expression of these latter genes in Col.

Being cultured for 14 days onto meshes did not induce any increase in the expression of the osteogenesis-associated genes tested (Figure S2A in Supplementary Material), although the presence of collagen induced a statistically significant increase (p < 0.01) in the expression levels of the chondrogenic marker Acan compared to the 2D control. For this gene, values were assessed around 6.97 ± 1.71 and 10.43 ± 4.23 for Col and HS-Col, respectively. The presence of our meshes was associated with a great activation of both, osteogenic and chondrogenic genes, following 14 days of induction toward the two lineages (Figures S2B and S3 in Supplementary Material).

To understand whether the exposure of MSC to HS-Col and Col could represent a source of stress for the cells, their response to the material was tested in absence of any other pro-inflammatory molecule. At 48 and 72 h, the levels of expression of genes encoding for immunosuppressive factors (Tnf-α, iNos, and Il-6) were evaluated in comparison with untreated cells (CTRL) or inflamed cells used as positive controls. Figure 5A shows a great increase in the expression of Tnf-α and iNos in cells grown onto HS-Col for 48 h compared to CTRL (13.32 ± 0.23 and 19.02 ± 0.15, respectively). No significant differences in the expression levels of Il-6 were found in Col and HS-Col at the same time point or at 72 h. Interestingly, the expression of Tnf-α diminished overtime in presence of HS, showing lower values than Col (2.11 ± 0.11 vs 12.24 ± 1.3). The iNos expression was found even increased at 72 h in both Col and HS-Col. Concomitantly, a significant increase in the expression of markers mediating the reduction of inflammation (Tgf-β, Pges2, and Ptgs2) was observed at 48 h in comparison with the CTRL and Col, with values assessed around 3.83 ± 0.50 for Tgf-β, 4.78 ± 0.50 for Ptgs2, and 57.41 ± 8.60 for Pges2 (Figure 5B). Values were found even increased at 72 h in the case of Ptgs2 and Tgf-β (6.35 ± 0.29 and 9.46 ± 0.38), the same trend being observed also for cells exposed to Col. The expression levels of Pge2 were found reduced overtime when cells were grown onto HS-Col.

To test the efficacy of the proposed material to support MSC immunosuppressive potential, cells grown onto Col and HS-Col meshes were stimulated with the pro-inflammatory cytokines TNF-α and IFN-γ (at the concentration of 20 ng/ml) for 48 and 72 h. qPCR was used to evaluate the relative amount of genes responsible for the production of pro-inflammatory molecules, which are responsible for the activation of such process, including Tnf-α, iNos, and Il-6. Data were normalized to the respective untreated control (CTRL, value = 1). Consistent with the previous findings, in response to stimulation cells grown onto Col produced nearly identical amounts of immunosuppressive genes (iNos and Tnf-α) compared with those cultured in 2D and standard conditions, whereas cells grown onto HS-Col scaffolds produced significantly higher amounts. Figure 6A shows a significant increase in the expression of Tnf-α, iNos, and Il-6 in MSC grown onto inflamed HS-Col compared to those seeded onto inflamed Col at 48 h. Values were assessed around 4.82 ± 1.23, 6.03 ± 1.46, and 4.24 ± 0.41 for Tnf-a, iNos, and Il-6, respectively. At 72 h, a marked decrease in the expression of iNos was found in inflamed HS-Col compared to 48 h (6.03 ± 1.46 vs 1.88 ± 0.22), which was found even lower than the expression levels found in Col at the same time point (2.63 ± 0.52). Finally, when the immunosuppressive potential of MSC grown onto HS-Col was evaluated in a lymphocyte reaction assay, a marked reduction in the proliferation of PHA-stimulated PBMC was noticed after 2 days of coculture compared to their Col counterparts, with percentage of proliferative cells assessed around 44 and 72%, respectively (Figure 6B).