Flexible, Biocompatible PET Sheets: A Platform for Attachment, Proliferation and Differentiation of Eukaryotic Cells

Transparent, flexible, biaxially oriented polyethylene terephthalate (PET) sheets were modified by bioactive polymer-fibronectin top layers for the attachment of cells and growth of muscle fibers. Towards this end, PET sheets were grafted with 4-(dimethylamino)phenyl (DMA) groups from the in situ generated diazonium cation precursor. The arylated sheets served as macro-hydrogen donors for benzophenone and the growth of poly(2-hydroxy ethyl methacrylate) (PHEMA) top layer by surface-confined free radical photopolymerization. The PET-PHEMA sheets were further grafted with fibronectin (FBN) through the 1,1-carbonyldiimidazole coupling procedure. The bioactive PET-PHEMA-I-FBN was then employed as a platform for the attachment, proliferation and differentiation of eukaryotic cells which after a few days gave remarkable muscle fibers, of ~120 µm length and ~45 µm thickness. We demonstrate that PET-PHEMA yields a fast growth of cells followed by muscle fibers of excellent levels of differentiation compared to pristine PET or standard microscope glass slides. The positive effect is exacerbated by crosslinking PHEMA chains with ethylene glycol dimethacrylate at initial HEMA/EGDA concentration ratio = 9/1. This works conclusively shows that in situ generated diazonium salts provide aryl layers for the efficient UV-induced grafting of biocompatible coating that beneficially serve as platform for cell attachment and growth of muscle fibers.


Introduction
The development of flexible surfaces is essential to create soft spatial topologies that allow to study mechanical strains on cells, to create soft spatial topologies, or to graft cells into moving organs subjected to topological constrains, like muscles for example. 1,2,3 .Indeed, biomaterials play major role in modern medicine; they can be employed as part of reconstructive implants, 4 implanted sensing objects, 5 and system for site specific drug delivery. 6 . Advances in biomaterials have led to non-toxic implants as well as implants that are specifically designed to elicit particular functions within the host. 7,8 Extracellular matrix (ECM)-based tissue engineering strategies are already successfully employed clinically for the regeneration of a range of different tissues, 9 including heart valves, 10 trachea, 11 muscles, 12 tendons, 13 and abdominal walls. 14 . Successful clinical application of specifically designed implants has been thus reported in cardiovascular, gastrointestinal, and breast reconstructive surgery. 15 .
Eucaryotic cells that make up implants normally grow in multicellular organisms surrounded by a specific molecular environment, the extracellular matrix. 16 , 17 ECM is a dynamic and complex environment characterized by biophysical, mechanical and biochemical properties specific for each tissue. The ECM consists of a complex assembly of many proteins and polysaccharides whose precise composition varies from tissue to tissue. The primary components include insoluble fibrous structural proteins (i.e. collagens, laminins, fibronectin, vitronectin, and elastin), proteoglycans, and specialized proteins (i.e. growth factors, small matricellular proteins and small integrin-binding glycoproteins). 18,19 . The extracellular matrix is thus critically important for many cellular processes including growth, differentiation, survival, and morphogenesis. 20,21 .
Flexible substrates are not only of importance to biomedical research and development but find also many other applications pertaining to thermoelectric organic materials, 22 sensors, 23 dye-sensitized solar cells, 24 to name but a few. However, regardless the application, substrates usually require surface modifications particularly when they are intended to support reactive and functional compounds, 25 polymers, 26 biomacromolecules 27 and living cells. 28 Towards this end, several surface engineering strategies are currently investigated in view of designing robust devices. Particularly, over the recent years, time and efforts were spent on the surface modification of a range of substrates by ultrathin reactive and functional polymer films for a variety of purposes. 29 One of the elegant strategies to attach polymers to surfaces is through the modern approach employing diazonium coupling agents. 30 This strategy is now accepted and explored by many laboratories around the Globe. 25,31,32,33 However, whilst it is currently applied to numerous materials such as metals, carbon and semi-conductors, only a few reports considered the modification of plastic substrates by aryl diazonium salts for biomedical purposes. In this regard, Ben Slama et al. 34 prepared copolymer-silver grafts on ITO for antibacterial applications, whereas Mahjoubi et al. 35 have used diazonium salts for the surface phosphonation of polyetheretherketone (PEEK) in order to design new PEEK-based orthopedic implants.
In the present work, we have used polyethylene terephtalate sheets modified with N,Ndimethyl-p-phenylenediamine salts (PET-DMA) as flexible surfaces on which fibronectin, an important extracellular matrix component, has been covalently linked. To study the biocompatibility of this type of flexible surface, we coated it with myoblastic cells that can proliferate and then differentiate into myotubes, the constitutive components of muscular fibers. This type of muscular cells has two advantages: first, they grow rapidly, which allows measuring their well-being on the PET-DMA membranes by counting them along time and quantifying their rate of growth: second, because after reaching confluence, muscular cells spontaneously differentiate. The differentiation process is sensitive to the cellular environment, and particularly to the nature of the surface on which cells are growing 36 . We therefore used the capacity of differentiation into myotubes as essential criteria to quantify the quality of flexible surfaces studied in this work.
For this purpose, we have assessed the propensity of muscular cells to grow and differentiate on Fibronectin-coated, modified PET substrates. Herein, we show for the first time that this device combines flexibility as well as a good attachment for eukaryotic cells and enhancement of their proliferation and differentiation.
Solvents including acetonitrile (ACN), methanol (MeOH), Ethanol (EtOH), dimethyl sulphoxide (DMSO), dimethyl formamide (DMF) were purchased from VWR Prolabo. The organic solvents used were analytical grade and DI (de-ionized) water was used for various cleaning and solution preparation. The PET sheets (thickness ~ 100 µm) were purchased from DuPont. The PET sheets were cut into sizes of 30mm X 15 mm for use in our experiments.

Functionalization of PET substrate by in-situ grafting of N,N-dimethyl-pphenylenediamine (DMA):
The PET sheets were cleaned and activated prior to diazonium functionalization.
First, they were ultrasonically cleaned with chloroform and ethanol for 15 min and then dried with argon flush. Before diazonium functionalization, the PET sheets were hydroxylated in a dimethyl sulphoxide (DMSO) solution of potassium hydroxide (4 mg KOH dissolved in 30 ml DMSO; KOH concentration = 2x10 -3 mol.L -1 ) for 10 min maximum, then washed with copious amounts of de-ionized (DI) water and dried in argon flow. The surface chemical modification of PET sheets was carried out using the diazonium salt of N,N-dimethyl-p-phenylenediamine (DMA). Typically three necked round bottomed flask (vol: 100 ml) with a reflux condenser arrangement, was used for the functionalization reaction. The flask with the colorless, transparent substrate was kept on a preheated oil bath and DMA (15 mmol, 2.04 g) was introduced. Cautiously isopentyl nitrite (15 mmol, 2.01 mL) was added slowly via syringe and 10 ml dimethyl formamide (DMF) was also added for solubilization. The reaction was left to proceed for 5 hrs at 60 °C under continuous bubbling of argon. Finally, the yellow colored modified substrates were thoroughly washed with DMF, DI water and dried in argon flow.

Surface-confined photopolymerization.
A typical procedure of photopolymerization of HEMA on DMA modified PET surface was as follows. A homogeneous solution of HEMA monomer (20 mmol, 2600 mg) and benzophenone (208 mg, 8% wt percent relative to monomer), as photosensitizer, were prepared in chloroform (20 ml) in a glass bottle. The DMA-modified PET was dipped in the mixture and bubbled with argon gas for 15 minutes to degas. Radical photopolymerization was carried out using a Spectrolinker XL-1500 equipped with six lamps emitting light nominally at 365 nm with a power density of 5 mW/cm². The photopolymerization time was 1200 sec. To study the effect of crosslinker, in the present case poly-ethylene glycol diacrylate (PEGDA), we have added two different weight percent (1% and 10% with respect to total weight of monomer and crosslinker) of crosslinker with HEMA monomer. The use of crosslinker ensures better adhesion of the photopolymerized toplayer to the substrate. 37 The polymerized sample was then taken out of the glass bottle and washed with methanol in a Soxhlet extractor for 2 h to remove the unreacted monomer and then washed with copious amounts of dichloromethane to remove organic species. Finally polymer-coated sheets were dried under argon flow and used in next step of surface modification.

Immobilization of fibronectin protein.
Before the fibronectin protein immobilization the polymer coated surface were activated by Imidazolyl carbamate. These moieties were attached to the polymer surface by the reaction of 1,1'-carbonyldiimidazole (CDI, 10 g/L in dioxane) with the hydroxyl group of PHEMA. The reaction was carried out for 6h at room temperature. Finally activated surface were thoroughly washed with dioxane and phosphate-buffered saline (PBS) solution to remove unreacted CDI. The covalent protein immobilization was performed by immersing the CDI activated surface in a fibronectin protein solution for 25 h under bidirectional stirring. The protein immobilization was tuned by changing the protein concentration ranging from 0.5 to 20 µg/mL. After the protein attachment the surfaces were carefully rinsed with 5% (v/v) aqueous solution of tween20 to remove unattached protein and then finally surface was cleaned with DI water.

XPS surface characterization
Surface chemical analysis was performed using a Thermo VG ESCALAB 250 X-ray photoelectron spectrometer (XPS) fitted with an Al monochromatic X-ray source (h = 1486.

Survey regions and surface elemental composition
Typical survey regions for clean and coated PET sheets are displayed in Figure 2.
Qualitatively, it is worth to note the efficiency of each step towards the preparation of bioactive platforms for cell adhesion. The main peaks C1s, O1s and N1s are centred at 285, 532 and 400 eV, respectively. The activated, clean PET displays only C1s and O1s peaks ( Figure 2a) while the diazonium attachment is testified by the appearance of an N1s peak from the grafted DMA aryl groups ( Figure 2b). With an initial concentration of the photosensitizer benzophenone of 8% (benzophenone to monomer molar ratio) it was possible to graft a PHEMA toplayer, thick enough (well above 10-12 nm) to screen the underlying DMA aryl layer; for this reason the N1s peak is no longer visible on the survey scan ( Figure 2c). However, after activation of polymer OH groups by CDI, hydroxyl groups convert to carbamates, hence the appearance of the N1s peak again (Figure 2d).
Immobilization of fibronectin was tuned by changing the initial concentration of fibronectin from 0.5 to 20µg/mL. Figure 2f shows that the relative intensity of the N1s peak is significantly higher than that exhibited by PET-PHEMA-FBN2 ( Figure 2e). The survey spectra permit thus to qualitatively diagnostic the sequential changes at the surfaces, simply by tracking the relative intensity of the N1s peak. The surface chemical composition (in atomic percents) is reported for bare and modified PET sheets in Table 1.

O1s
The main information that can be obtained from O1s regions is the change from PET to a PHEMA-rich surface (Figure 4). PET has two ester groups per repeat unit and thus one carbonyl and one alkoxy groups which are known to give a well resolved doublet as displayed in Figure 4a.

N1s and S2p
We discussed above about the usefulness of the N1s peaks in tracking changes in macromolecular species at the surface. The changes do not concern only the relative intensity visible on the survey regions and the contribution of nitrogen to the surface elemental composition in atomic percent ; herein, we demonstrate that the high resolution     In order to improve the chemical treatment of the PET surfaces, we modified the protocol of radical photopolymerization of HEMA on PET-DMA by adding PEGDA comonomer at 1% or crosslinked, biocompatible polymer grafts with improved polymer adhesion to diazoniummodified surfaces. 37 Investigations of cell growth were conducted in a similar manner as described above in Figure 7. Results are presented in Figure 8 and Table 2. The half-time defined as ln (2) Altogether, these results suggest that treatment with PEGDA considerably improves the efficiency of cell growth (on PET-DMA-PHEMA-PEGDAx-I-FBN surfaces).

Table 2. Measurement of cell growth on glass and PET surfaces.
Curves of cell growth were fitted to exponential curves, and the exponential coefficient k, which indicates the speed of growth (in A = A0 e kt ) calculated. Another way to estimate cell growth is the doubling time (half-time) calculated (half-time = ln(2) / k) and shown in the right column. A second important biological parameter to investigate is the capacity of the cells to differentiate; it indicates how cells adequately interact with the support on which they grow.
Indeed, it is well known that differentiation is obtained only if cellular environmental parameters are optimal with regard to reduced growth factors in culture medium as well as attachment to extracellular molecules such as fibronectin. For that purpose, all myoblast cells were left to grow to confluence without changing the culture medium after day 7 to allow exhaustion of growth factors. In these conditions, cells switch form a growth program to a differentiation program: they align in a parallel fashion and start to fuse together to give multinucleated myotubes ( Figure 9A). At this step, cell culture medium is renewed, but the fetal calf serum providing growth factors is reduced to 2 % instead of 10 % to allow cell survival but not cell growth. To quantify cell differentiation, the number and diameter of the myotubes adhering to the PET surfaces were determined.
Differentiation was assessed on PET-DMA-PHEMA-I-FBNx (x= 5, 10 and 20 µg/ml of fibronectin), with the same conditions as used for Figure 7. After 8 days of cell plating, myotubes were already well developed, with a density between 6 and 12 per 0.6 mm 2 field (see Figure 9A). The best result is obtained with PET-DMA-PHEMA-I-FBN20 (12 myotubes per field), while lower concentrations (5 and 10 µg/ml fibronectin) remained at basal levels, equivalent to untreated PET surfaces (PET) that is 6 myotubes per field ( Figure 9B).
Interestingly, PET-DMA-PHEMA-FBN20 gives also a high level of differentiation, with 10 myotubes per field, slightly lower than PET-DMA-PHEMA-I-FBN20. With this condition, fibronectin is adsorbed on the surface, but not covalently linked to it. Similar results were obtained after 10 and 13 days of culture and differentiation (data not shown).
The number of myotubes counted and shown in Figure 9B does not take into account the evolution of differentiation once the myotubes are formed: when two cells fuse together, it is already considered as a myotube, but differentiation still goes on and other myoblasts fuse to this myotube, allowing it to grow in length, diameter, and number of nuclei. Therefore, in order to better quantify the progression of differentiation, the diameter of myotubes was measured. Indeed, the diameter is proportional to the number of cells that have fused into myotubes. Results are shown in Figure 9C, and are essentially similar to those displayed in Figure  In addition and as shown in Figure 8, we used conditions of preparation and grafting PET surfaces for myoblastic differentiation with addition of PEGDA crosslinker for PHEMA. Figure   10A shows  In conclusion, the results obtained so far show altogether that PET membranes grafted with fibronectin at the highest dose (20 µg/ml) through the DMA-PHEMA-I protocol, with addition of PEGDA, is very efficient in allowing proliferation and differentiation of muscular cells.

Discussion and Conclusion
Fibronectin, an important component of the extracellular matrix, was grafted on bioactive PET-DMA-PHEMA-I and derivatives in order to investigate proliferation and differentiation of muscular cells on these biomaterials. The general observation is that this type of treatment improves both proliferation and differentiation over untreated or partially treated PET surfaces. Several conclusions can be reached: (i) The untreated PET membrane ( Figure 7 and Table 2 With regard to cell differentiation, the situation appears more drastic: only the highest dose of fibronectin with PEGDA treatment (PET-DMA-PHEMA-PEGDA10-I-FBN20, Figure 10A) gives the best number of myotubes, as well as the largest diameter of myotubes, two criteria used to evaluate differentiation. The second best result is provided by PET-DMA-PHEMA-I-FBN20 ( Figure 9A). It appears that fibronectin is essential for a good differentiation, since only PET Altogether, these results suggest that either a threshold concentration of FBN is required for cells to fully differentiate, or that PET-DMA-PHEMA-I may inhibit cell differentiation, by providing inappropriate charges or signals to differentiating cells. This inhibition effect may be overcome by fibronectin when high concentration (20 µg/ml) is used. However, these two explanations are not necessarily mutually exclusive.
The diameter of myotubes may result from a complex situation: when muscular cells reach confluence, they may create with their neighboring cells a new myotube, or fuse with a preexisting myotube. It is possible that chemical treatment of PET membrane modifies this mechanism in a subtle fashion. The main result from this study is that fibronectin is important (PET-DMA-PHEMA-FBN20 and PET-DMA-PHEMA-I-FBN20, Figure 9B) but PEGDA improves the diameter, although PET-DMA-PHEMA gives a slightly higher value ( Figure 10B) possibly due to improved biocompatibility imparted by PHEMA.
To sum up, the full treatment, with PEGDA and fibronectin gives the best results in all cases, and stimulates cell proliferation as well as differentiation and growth of myotubes. This opens avenues for providing favourable environments for cells in order to graft them in muscular tissue, such as the heart to help for heart diseases.