Polyelectrolyte multilayer film modification for chemo-mechano-regulation of endothelial cell response

Abstract

The new multilayer polyelectrolyte films (PEMs) that are able to simulate the structure and functions of the extracellular matrix have become a powerful tool for tailoring biointerfaces of implants. In this study, bioactive PEM coatings have been investigated as a supportive system for efficient endothelialization of cardiovascular implants. The modern films were designed in a manner that allows one to potentially induce specific response from the tissues surrounding the biomaterial due to its chemical composition as well as mechanical properties. The PEM rigidity was regulated by the cross-linking chemistry as well as nanoparticle incorporation, while biochemical modification was performed by the VEGF adsorption within coatings. Obtained results have shown that PEM/VEGF films enhanced in vitro spreading and proliferation of endothelial cells, whereas VEGF presence inhibited IL-6 production and release. Since non-functionalized films also contributed to proliferation of endothelial cells and cytokine secretion, it may be supposed that PEM stiffness acts in synergy with the growth factor, but probably through a different pathway. Results clearly demonstrate the effectiveness of the proposed endothelialization strategy and confirm correlation between the chemical and mechanical properties of the PEMs in vitro.

---

1. Introduction

The design of a new multilayer polyelectrolyte film (PEM) biointerface aimed at inducing specific responses from the tissues surrounding the biomaterial has emerged as a new strategy in medical device and implant biocompatibility improvement.^1–3 Recently, thin coatings assembled with polypeptides and polysaccharides by the ’layer by layer’ technique have been shown to simulate extracellular matrix (ECM) structure and functions.^4–6 The natural cellular niche could be recreated by dealing with mechanical as well as biochemical properties of multilayer films. Both approaches have been found to enhance repopulation of grafted biomaterial surface by the host cells. The chemical cross-linking based on changes in polymers' functional groups interaction is commonly used rigidity control method.^7,8 An appropriate PEMs mechanical properties could be also achieved by nanoparticles incorporation.⁹ Simultaneously, this modification method may be applied to provide coating with antimicrobial properties.^10–12 Moreover, nanoparticles modified PEMs bring together features of tridimensional networks offered by polymers and the intrinsic functionalities of nanoparticles what opens the possibility to explore the application of such nanocomposites in localized, stabilized and controlled growth factors' release.^13–15 The growth factors are known to regulate multiple events as cell proliferation, migration and differentiation.¹⁶ The strategies for their immobilization in PEMs could be divided in two main categories: covalent and non-covalent binding to the polymers.¹⁷ The growth factors presentation on the material surface is a step toward recreating the natural cellular microenvironment by mimicking of ECM biochemical properties. The vascular endothelial growth factor (VEGF) is the one of major growth factors that has been highly investigated in neovascularization of implanted biomaterials mostly in bone or cardiovascular system regeneration.^18–20 Stimulation of implant vascularization represents an important strategy to allow fast tissue integration and to avoid graft rejection.²¹ Moreover, the VEGF specifically affects endothelial cells (ECs) through positive effects on migration, proliferation, differentiation and survival which allows to obtain functional endothelium monolayer on implanted material.²² Coating of internal surface of the natural bloodstream system with endothelial cells, prevent blood against coagulation.^23,24 This suggested the idea of creating hybrid polymeric materials with quasi-intima coated surfaces formed by ECs. The self-endothelialising synthetic graft is an attractive solution for cardiovascular regeneration. However, the difficulties experienced in achieving spontaneous endothelialization in humans led to the investigation of pre-implantation in vitro ECs seeding. The literature revealed that so far attempts to blood contacting materials designing have been sufficient to provide better internalization of implants due to their previous cellularization.^25,26 These elegant studies usually highlighted the effect of one material property, however the cellular response to a biomaterial comprises the net effect of all the material issues. The new generation of biomaterials still requires an improved understanding of how biochemical and mechanical properties coordinately regulate cellular response.

Therefore, our recent work was aimed at designing bioactive PEM coatings for an effective endothelialization of biomaterials for cardiovascular applications. Among natural polymers, the PEM systems composed of poly-L-lysine/sodium hyaluronate, poly-L-lysine/sodium alginate and chitosan/chondroitin sulphate were chosen as the most promising blood contacting materials due to their high biocompatibility. Mechanical properties of films were regulated by chemical cross-linking or nanoparticles incorporation. In this studies an effect of the commonly describedin the literature N-hydroxysulphosuccinimide/1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (NHS/EDC) cross-linking chemistry was compared with results of PEMs modification by an alternative, less cytotoxic genipin agent. Three different types of nanoparticles, i.e. silicon carbide (SiC), silver (Ag) or reduced graphene oxide (rGO) have been taken into consideration due to increasing interest in their many possible applications, especially in biosensors technology. Biochemical modification concerned the VEGF factor adsorption within multilayer that is well known to support endothelialization process. Endothelial cells' adhesion, proliferation and morphology was investigated in vitro in response to different content/release kinetics of adsorbed VEGF growth factor and changes in material stiffness. Finally, the idea of correlated chemo-mechano-regulation of endothelial cells response was applied to observe changes in IL-6 cytokine secretion profile. The tunable features of the materials developed here enables understanding of how the substrate mechanical properties and surface-adsorbed growth factors act in concert to affect a variety of cellular functions.

2. Materials and methods

The polyelectrolyte multilayer films were prepared from the low weight cationic chitosan (Chi) or poly-L-lysine (PLL) and the anionic chondroitin sulfate (CS) or sodium alginate (ALG) purchased from Sigma-Aldrich. The other type of anionic polymer such as sodium hyaluronate (HA) was provided by Lifecore Biomedical, LLC. Chemicals for cross-linking i.e. 1-ethyl-3-(3-dimethylamino-propyl)carbodiimide (EDC) and N-hydrosulfosuccinimide (NHS) were supplied by Sigma-Aldrich, whereas genipin was bought from Challenge Bioproducts Co. Ltd. The Human Umbilical Vein Endothelial Cell line was purchased from Lonza Group Ltd. Cellular response to various coating properties was assessed based on staining with Alexa Fluor®488 Phalloidin and DAPI supplied by Invitrogen. ELISA kits used in IL-6 and VEGF release kinetics assays were provided by R&D Systems, Inc.

2.1. Polyelectrolyte multilayer films' manufacturing

Polyelectrolyte multilayer films (PEMs) were deposited with a “layer-by-layer” (LbL) technique onto 1.5 cm × 1.0 cm glass substrate material. Substrates were activated by 10 M NaOH and washed with pure Milli-Q water. The Chi polymer was pre-dissolved in 0.1 M acetic acid. The other polymers (PLL, HA, ALG and CS) are quite well soluble in water, therefore pre-dissolution step was unnecessary. Final concentration of 0.5 mg ml⁻¹ Chi or PLL and 1 mg ml⁻¹ CS, HA or ALG was prepared in 0.15 M NaCl of pH 6.0. Films were manufactured with an automatic dipping machine by an alternately immersing substrate in proper solutions of cationic and anionic polyelectrolytes for 8 min each. After each deposition step, the substrate was rinsed in 0.15 M NaCl solution buffered at pH 6.0 to remove excess, non bonded polymer chains. The process was repeated until the desired number of 48 bilayers was obtained.

2.2. Polyelectrolyte multilayer films' chemical cross-linking

The chemical cross-linking of Polyelectrolyte Multilayer Films is essential for coatings stability and cellularization process control by topography, wettability, stiffness and other parameters optimization. Herein, modification process was performed as post-treatment after PEMs synthesis, in two different ways. In the first one 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and N-hydroxysulphosuccinimide (NHS) was applied according to the protocol described elsewhere.²⁷ Reagents were prepared in 0.15 M NaCl solution buffered at pH 5.5 and mixed together in 1 : 1 volume ratio immediately before application. Three different EDC concentrations (260 mM, 400 mM, 800 mM) were tested. In all reactions 100 mM NHS was used. Polyelectrolyte films were incubated in EDC/NHS for 18 hours at 4 °C. Then the cross-linker solution was removed, and films were rinsed several times with 400 mM HEPES/0.15 M NaCl solution buffered at pH 7.4 to wash out non-reacted amounts of reagents. Washing procedure was performed in two steps. At first coatings have been rinsing two times for an one hour under slow shaking at room temperature each time. Then in the second step coatings were conducted to three short (10 minutes each one) washes. The second cross-linking protocol concerned application of genipin. Reagent was prepared in three different concentrations (1 mM, 10 mM, 50 mM) by preliminary dissolution in ethanol and finally in 0.15 M NaCl solution buffered at pH 5.5. Polyelectrolyte films were incubated in genipin for 18 hours at RT and then rinsed with distilled water.

2.3. Nanoparticles immobilization inside PEMs

2.3.1. Graphene oxide nanoparticles introduction. Graphene oxide was prepared by modified Hummers' method.²⁸ The thermally expanded graphite (5 g, particle diameter: 300–425 mm) and potassium nitrate (6.5 g) were mixed in beaker containing 200 ml of concentrated sulfuric acid (96–98%). The solution temperature was slowly cooling down in the ice bath below 5 °C, while the 15 g of potassium permanganate was gradually added. Then beaker with solution was placed in water (25 °C) for 16 hours. Afterwards, it was once more put into the ice bath and 230 ml of deionized water (DI) was slowly poured into suspension. The mixture was then heating to 95 °C for 15 min. Further solution was diluted by addition of 280 ml of deionized (DI) water and cooled down to the room temperature. Finally, 5 ml of 30% hydrogen peroxide was added to the mixture. Graphite oxide suspension was washed with 3% hydrochloric acid solution and, after removal of sulphate ions, continuously washed with DI water, until none chloride ions were detected. The purified suspension was then ultrasonicated for 1 h in order to exfoliate oxidized graphene sheets inside graphite structure, which allows acquiring a stable suspension of graphene oxide in water. Reduced graphene oxide (rGO) flakes were introduced into PEMs by LbL method as the mixture with negatively charged polyelectrolyte (CS, HA or ALG) in the 1 : 100 volume ratio. The average number of rGO flakes per 100 nm² surface was calculated based on high resolution transmission electron microscopy images. The Tecnai G2 F20 (200 kV) FEG was used for analysis. Thin foils for TEM observations were prepared directly from the place of interest using focused ion beam technique. The Quanta 200 3D DualBeam was used for FIB preparation. Observations were performed in a bright field mode.

2.3.2. PACVD of SiC nanoparticles. SiC nanoparticles were introduced to PEM films by the plasma assisted chemical vapor deposition (PACVD) method.²⁹ Before nanoparticles incorporation, the polyelectrolyte-coated substrates were dried and introduced into the vacuum chamber.³⁰ The chamber was pumped to reach the vacuum down to at 5 Pa, while the precursors for the PACVD (plasma polymerization) are fed on the other side, controlled by flow meter and pressure. Hexamethyldisiloxane (HMDSO, kept at constant temperature of 32 °C) was used as precursor first to deposit 40 nm thin silicone plasma polymer in atmosphere of 20 Pa pressure, which was subsequently, in situ reactively post-treated to SiC nanoparticles. The average number of single SiC nanoparticles per 100 nm² surface was calculated based on high resolution transmission electron microscopy images.

2.3.3. In situ synthesis of Ag nanoparticles. For silver salt loading, films were immersed into 0.01 mM silver salt solution (pH ∼ 6) for 15 min, followed by three water rinses. Then samples were air dried in an incubator at 37 °C for 30 min. The silver-salt-loaded coatings were excited with an ultra violet lamp (365 nm, power = 36 W) at a distance of ∼1 cm for 24 h, similar as described by others.³¹ The average number of Ag nanoparticles per 100 nm² surface was calculated based on high resolution transmission electron microscopy images.

2.4. Surface topography

Atomic Force Microscopy measurements were performed on Innova commercial instrument. Topography pictures were obtained using tapping mode with tips FESPA (Bruker). All measurements were carried out at room temperature. Data treatment and presentation were realized with the help of Nanoscope 1.40 Software.

2.5. Mechanical properties

AFM nanoindentation measurements were performed in a 0.15 M NaCl at pH 7.4. Force-indentation profiles were recorded using borosilicate sphere-tipped cantilever (2.5 μm radius) with a nominal spring constant of 60 mN m⁻¹. Young modulus (E) was calculated using the Hertz model fit to the data over the indentation curve from 30 nm to a 300 nm indentation depth, a range over which the Hertz contact model still remain accurate (film thickness around 3 μm). The incompressibility of films was assumed due to their high water content (Poisson's ratio ν = 0.5), for each tested sample, two measurements were realized at six different positions on each sample. Young moduli were calculated by least-squares fitting the obtained force–indentation curves.

2.6. VEGF adsorption and release kinetics

Polyelectrolyte multilayer films were analyzed regarding to vascular endothelial growth factor (VEGF) adsorption efficiency and release kinetics. The VEGF is the key growth factor for the endothelialization, so its immobilization and controlled release could be crucial for bioactive surface preparation. At first each type of sample was incubated with 50 ng ml⁻¹ VEGF solution for 12 hours at 37 °C in order to adsorb growth factor within coating. Then VEGF was discarded and adsorbed protein was washed out from coatings by 0.2% sodium dodecyl sulfate (SDS) This step was performed at 37 °C for at least 2 hours. Concentration of protein in SDS solution was measured by enzyme-linked immunosorbent assay (ELISA) method for each type of sample and compared with the reference non-coated substrate. The release kinetics of VEGF from various coatings was determined also based on the ELISA assay. The growth factor was introduced by samples soaking for 12 hours at 37 °C in 50 ng ml⁻¹ concentration. Then samples were immersed in culture medium without VEGF and incubated at 37 °C for 12 days, In appropriate time points medium above coatings was collected and stored (−80 °C) for further ELISA investigations of growth factor time dependent concentration changes.

2.7. HUVECs proliferation and morphology

HUVECs (Human Umbilical Vein Endothelial Cells) were seeded with density of 4 × 10³ cells per cm² and cultured on all analyzed variants of scaffolds. The cell–material interactions were determined based on HUVECs morphology, adhesion and proliferation comparisons during 12 day culture. In appropriate time points cells were fixed (4% paraformaldehyde) and stained with Alexa Fluor 488 phalloidin in order to visualize cytoskeleton structure and with 4′,6-diamidino-2-phenylindole (DAPI) to show cell nucleus. Images were acquired using a confocal laser scanning microscopy (CLSM) Exciter5 AxioImager. Data was processed with the CLSM Zen 2008 software.

2.8. IL-6 secretion

HUVECs response to changes in surface chemical and mechanical properties was determined regarding to IL-6 factor secretion kinetics profile. The IL-6 cytokine is not only involved in inflammation and infection responses but also in the regulation of metabolic, regenerative processes, so regulation of its secretion by cells could be essential for bioactive surface design. The release kinetics by HUVECs growing on various coatings was determined based on ELISA assay. Cells were cultured at 37 °C for 12 days, when in appropriate time points culture medium was collected and kept frozen until ELISA investigations.

2.9. Statistical analysis

A statistical analysis (ANOVA and Tukey post hoc test, P value smaller than 0.05 was considered as significant – Statistica 10.0 PL) was performed on six replicates from each treatment in each type of described examination.

3. Results

3.1. Microstructure of modified PEMs

The investigated variants of PEMs prepared by “layer by layer” method, then modified by chemical cross-linking, nanoparticles incorporation or VEGF adsorption are presented in Table 1.

Table 1 Prepared and investigated PEM variants

Polymer type	Cross-linking	Nanoparticles incorporation	Growth factor adsorption
PLL/HA	Non	Non	Non
			VEGF
		rGO	Non
			VEGF
		SiC	Non
			VEGF
		Ag	Non
			VEGF
	260 mM NHS/EDC	Data not presented	Data not presented
	400 mM NHS/EDC	Non	Non
			VEGF
		rGO	Non
			VEGF
		SiC	Non
			VEGF
		Ag	Non
			VEGF
	800 mM NHS/EDC	Data not presented	Data not presented
	1 mM genipin	Data not presented	Data not presented
	10 mM genipin	Non	Non
			VEGF
		rGO	Non
			VEGF
		SiC	Non
			VEGF
		Ag	Non
			VEGF
	50 mM genipin	Data not presented	Data not presented
PLL/ALG	Non	Data not presented	Data not presented
	260 mM NHS/EDC
	400 mM NHS/EDC
	800 mM NHS/EDC
	1 mM genipin
	10 mM genipin
	50 mM genipin
Chi/CS	Non	Data not presented	Data not presented
	260 mM NHS/EDC
	400 mM NHS/EDC
	800 mM NHS/EDC
	1 mM genipin
	10 mM genipin
	50 mM genipin

---

The microstructure of multilayer coatings was amorphous (Fig. 1A–D). There was no significant difference in TEM cross-section images between films. After chemical cross-linking amorphous structure was preserved. There was no difference between microstructure of films modified by either NHS/EDC or genipin cross-linker as well as non-cross-linked films (Fig. 1A). PEMs amorphous microstructure was modified by immobilization of three different types of nanoparticles, i.e. graphene flakes, silicon carbide or silver nanoparticles.

Fig. 1 Microstructure of PLL/HA films: (A) 400 mM NHS/EDC cross-linked; (B) modified by graphene flakes; (C) modified by SiC nanoparticles; (D) modified by Ag nanoparticles.

Graphene flakes were introduced into PEMs based on electrostatic interactions as mixed with anionic polymers (HA, ALG or CS) during multilayer films deposition. The pure graphene occupies about 40 ± 5% of the coating volume which was calculated based on TEM images (Fig. 1B). The performed observations confirmed defects of typically symmetric graphene flakes structure. However, the HRTEM analysis indicated also areas of the well crystallized graphene of hexagonal lattice. The average size of the visible well-crystallized areas was in the range of 5–7 nm. The cross-linking process performed after PEM/graphene film deposition had no visible influence on films' microstructure. Beside attempts to PEMs microstructure changes by graphene incorporation, the addition of silicon carbide was investigated. TEM images revealed formation of three zones within polyelectrolyte multilayers, the SiC non-implanted zone, the highest density of SiC nanoparticles zone and the zone containing single SiC incorporated nanoparticles (Fig. 1C). Samples indicated variable homogeneity in aggregates distribution related to PACVD protocol specificity. There was no visible difference between microstructure of non-cross-linked and cross-linked samples containing SiC nanoparticles. HRTEM images indicated the average size of nanoparticles in the range of 5–10 nm and mean value of 17 single particles per 100 nm² in the nanoparticles dense zone and 8 ± 2 single particles per 100 nm² in the deeper coating layers, located closer to the substrate surface. The thickness of the mixed polymer with SiC layer was on the level of 150 ± 10 nm.

Silver nanoparticles were incorporated into PEMs by in situ nucleation method. It has been found that Ag nanoparticles did not form a separate layer within polyelectrolyte multilayer film, but were distributed within film volume in the form of aggregates (40.6 ± 4.5 nm diameter) (Fig. 1D). TEM images indicated no visible differences between microstructure of non-cross-linked and cross-linked samples containing silver nanoparticles. Moreover, it was measured that the average density of nanoparticles within the coating was equal to 28 ± 2 single particles per 100 nm², whereas the single silver nanoparticle diameter was about 5.0 ± 1.2 nm.

3.2. Surface topography

The surface morphology of non-cross-linked films was evaluated by atomic force microscopy measurements. It was found that coatings were deposited as continuous layer which completely covered substrate surface. The roughness R_a varies from 2.39 ± 0.13 nm for PLL/HA film to 4.12 ± 0.15 nm for PLL/ALG and to 16.38 ± 0.10 nm for Chi/CS films (Fig. 2A).

Fig. 2 Surface topography of PEMs: (A) the roughness of non-cross-linked, NHS/EDC or genipin cross-linked PLL/HA, PLL/ALG and Chi/CS films. Data represents mean ± SD; (B) the roughness of PLL/HA modified by graphene flakes, silicon carbide or silver nanoparticles and coating without nanoparticles. Results obtained for non-cross-linked, 400 mM NHS/EDC and 10 mM genipin cross-linked samples. Data represent mean ± SD; *P < 0.05 vs. non-cross-linked sample with rGO; **P < 0.05 vs. non-cross-linked sample with SiC; ***P < 0.05 vs. non-cross-linked sample with Ag; ****P < 0.05 vs. non-cross-linked sample without nanoparticles.

Chemical cross-linking of multilayer films either by NHS/EDC or genipin chemistry caused a significant increase in surface roughness of PLL/HA and PLL/ALG films. It was found that application of NHS/EDC resulted in higher roughness than genipin treatment in case of all investigated PEMs. Gradual increase of roughness was observed with increasing concentration of applied cross-linker. The highest values were found for PLL/HA films, either in case of NHS/EDC (up to 30.39 ± 0.23 nm) or genipin cross-linking (15.30 ± 0.18 nm) (Fig. 2A). Contrary, the Chi/CS roughness decreased after either NHS/EDC or genipin treatment. Gradual decrease of roughness was observed with increasing concentration of applied cross-linker. It was found that application of genipin resulted in lower roughness decrease than NHS/EDC treatment.

The surface roughness of PEMs was determined after their structure stabilization by nanoparticles. It has been noticed that replacement of standard polyanions by their solution with graphene resulted in significant increase of roughness (PLL/HA non-cross-linked films – from 2.39 ± 0.13 nm to 21.10 ± 0.20 nm) (Fig. 2B). This tendency was found for all investigated non-cross-linked polyelectrolyte multilayer films. For further investigations the PLL/HA coating was chosen. Samples were cross-linked by 400 mM NHS/EDC or 10 mM genipin. Obtained results indicated no significant changes in surface roughness after 10 mM genipin cross-linking in comparison to non-cross-linked state (Fig. 2B and 3D and F). However, there was significant increase of R_a value after application of 400 mM NHS/EDC cross-linker (23.90 ± 0.12 nm) (Fig. 2B and 3E). The AFM analysis revealed inhomogeneous distribution of graphene within polyelectrolyte porous structure. Graphene flakes were located parallel with their longer axis to the film surface.

Fig. 3 The AFM surface morphology of PLL/HA coating: native ((A) non-cross-linked; (B) 400 mM NHS/EDC cross-linked; (C) 10 mM genipin cross-linked); containing rGO flakes ((D) non-cross-linked; (E) 400 mM NHS/EDC cross-linked; (F) 10 mM genipin cross-linked); containing SiC nanoparticles ((G) non-cross-linked; (H) 400 mM NHS/EDC cross-linked; (I) 10 mM genipin cross-linked); containing Ag nanoparticles ((J) non-cross-linked; (K) 400 mM NHS/EDC cross-linked; (L) 10 mM genipin cross-linked).

Surface topography of PEMs was also determined after silicon carbide nanoparticles immobilization by PACVD method. The performed studies have shown that SiC introduction into each type of investigated polyelectrolyte multilayer films caused a significant increase in surface roughness. More detailed investigations were performed based on non-cross-linked, 400 mM NHS/EDC cross-linked and 10 mM genipin cross-linked PLL/HA multilayer coatings (Fig. 2B and 3G–I). It has been shown that roughness of PLL/HA films increased from 2.39 ± 0.13 nm for non-cross-linked sample to 34.10 ± 0.23 nm for non-cross-linked film containing SiC nanoparticles. Cross-linking changed surface roughness significantly. There was difference between applied cross-linker type. The R_a of film cross-linked by 400 mM NHS/EDC reached 45.60 ± 0.24 nm, whereas the surface roughness after genipin cross-linking was equal to 41.40 ± 0.30 nm.

The surface morphology of PEMs containing silver nanoparticles was determined. For each investigated film type, silver nanoparticles immobilization increased roughness value compare to the native state PEMs. Therefore, detailed description was based on the PLL/HA film similar like for PEMs modified by other nanoparticles. Herein, non-cross-linked coatings containing Ag indicated significantly higher roughness (25.10 ± 0.20 nm) than already characterized non-cross-linked film without nanoparticles (Fig. 2B and 3J). There was significant R_a value difference between non-cross-linked and 10 mM cross-linked (25.90 ± 0.25 nm) films, both with silver nanoparticles. The significantly higher roughness was also obtained for 400 mM cross-linked Ag modified film (27.30 ± 0.25 nm). It was found that roughness of PEMs modified by nanoparticles changed according to the silicon carbide > silver > graphene relation.

3.3. Mechanical properties

Indentation test results are presented as mean values of the elastic modulus of investigated polyelectrolyte coatings. Performed studies have shown significant differences between elastic modulus of PLL/HA (10.7 ± 1.5 kPa), PLL/ALG (9.4 ± 0.7 kPa) and Chi/CS (8.5 ± 0.5 kPa) non-cross-linked films (Fig. 4A).

Fig. 4 Mechanical properties of PEMs: (A) mechanical properties of non-cross-linked, NHS/EDC or genipin cross-linked PLL/HA, PLL/ALG and Chi/CS films. Data represents mean ± SD; (B) the stiffness of PLL/HA modified by graphene flakes, silicon carbide or silver nanoparticles and coating without nanoparticles. Results obtained for non-cross-linked, 400 mM NHS/EDC and 10 mM genipin cross-linked samples. Data represent mean ± SD; *P < 0.05 vs. non-cross-linked sample with rGO; **P < 0.05 vs. non-cross-linked sample with SiC; ***P < 0.05 vs. non-cross-linked sample with Ag; ****P < 0.05 vs. non-cross-linked sample without nanoparticles.

Chemical cross-linking of multilayer films either by NHS/EDC or genipin chemistry caused a significant increase in their mechanical properties (Fig. 4A). It was found that application of NHS/EDC resulted in higher stiffness of all investigated PEMs than genipin treatment. Gradual increase of stiffness was observed with increasing concentration of applied cross-linker. The highest elastic modulus was found for Chi/CS films, either in case of NHS/EDC (up to 27.2 ± 2.1 kPa) or genipin cross-linking (20.5 ± 0.8 kPa). Contrary, the lowest elastic modulus after cross-linking was measured for PLL/ALG film. Non-cross-linked films revealed viscoelastic properties similar to obtained for soft thermoplastic polymers, while the cross-linked samples have lost tendency to creep.

PEMs mechanical properties were determined after their structure stabilization by nanoparticles. Graphene flakes were introduced into films as mixture with anionic polyelectrolyte (HA, ALG or CS) during LbL films deposition. It has been noticed that replacement of standard polyanions by their solution with graphene resulted in increase of elastic modulus value. This tendency was found for all investigated non-cross-linked polyelectrolyte multilayer films. For further investigations, only the PLL/HA coating was chosen. Samples were cross-linked by 400 mM NHS/EDC or 10 mM genipin. Obtained results indicated significant increase in elastic modulus after cross-linking (Fig. 4B). Surprisingly, there was no difference between applied cross-linkers. In both cases Young modulus increased from 158 ± 10 kPa for non-cross-linked to 195 ± 15 kPa for 10 mM genipin cross-linked and 197 ± 15 kPa for 400 mM NHS/EDC cross-linked films.

Mechanical properties of PEMs were determined after silicon carbide nanoparticles immobilization by PACVD method. The performed studies have shown that SiC introduction into each type of investigated polyelectrolyte multilayer caused a significant increase in film stiffness. More detailed investigations were performed based on non-cross-linked, 400 mM NHS/EDC cross-linked and 10 mM genipin cross-linked PLL/HA multilayer coatings (Fig. 4B). It has been shown that Young modulus value of PLL/HA films increased from 10.7 ± 1.5 kPa for non-cross-linked sample to 710 ± 13 kPa for non-cross-linked film containing SiC nanoparticles. The highest stiffness at the level of 828 ± 11 kPa was found for PLL/HA film modified by SiC nanoparticles and cross-linked by 400 mM NHS/EDC. The elastic modulus of film cross-linked by genipin was equal to 760 ± 16 kPa.

Finally, mechanical properties of PEMs with silver nanoparticles incorporated by in situ nucleation method were evaluated. It was found that for each type of film, nanoparticles immobilization led to increase of elastic modulus. Therefore, detailed description was based on the PLL/HA film similar like for PEMs modified by silver nanoparticles (Fig. 4B). Herein, non-cross-linked coatings containing Ag indicated significantly higher stiffness (213 ± 12 kPa) than already characterized non-cross-linked films without nanoparticles (10.7 ± 1.1 kPa). Moreover, there was no significant rigidity difference between non-cross-linked and 10 mM genipin cross-linked (214 ± 13 kPa) films, both with silver nanoparticles. The highest elastic value was obtained for 400 mM NHS/EDC cross-linked Ag modified film (237 ± 12 kPa).

3.4. VEGF adsorption and release kinetics

The present study was aimed at designing bioactive coatings to promote endothelialization. For this purpose, PEM films were functionalized by vascular endothelial growth factor (VEGF). The most favorable PEM-architecture for VEGF adsorption and bioactivity was selected among native, cross-linked and nanoparticles containing films. Finally, the influence of various films on in vitro endothelial cell adhesion and proliferation was determined. Performed ELISA analysis showed statistically significant differences between VEGF adsorption by PLL/HA (2400 ± 20 pg cm⁻²), PLL/ALG (2120 ± 28 pg cm⁻²) and Chi/CS (1950 ± 17 pg cm⁻²) non-cross-linked films (Fig. 5A). Chemical cross-linking either by NHS/EDC chemistry of each type of multilayer film caused a significant decrease of growth factor adsorption. However, it was found that application of genipin resulted in higher retention capacity of VEGF than native stage of films. Gradual decrease (NHS/EDC) or increase (genipin) of adsorption efficiency was observed with increasing concentration of applied cross-linker.

Fig. 5 Surface concentration of VEGF adsorbed onto the different PEM architectures: (A) the growth factor adsorption onto non-cross-linked, NHS/EDC or genipin cross-linked PLL/HA, PLL/ALG and Chi/CS films. Data represents mean ± SD; (B) the VEGF adsorption for PLL/HA modified by graphene flakes, silicon carbide or silver nanoparticles and coating without nanoparticles. Results obtained for non-cross-linked, 400 mM NHS/EDC and 10 mM genipin cross-linked samples. Data represent mean ± SD; n = 6; *P < 0.05 vs. non-cross-linked sample with rGO; **P < 0.05 vs. non-cross-linked sample with SiC; ***P < 0.05 vs. non-cross-linked sample with Ag; ****P < 0.05 vs. non-cross-linked sample without nanoparticles.

The VEGF concentration for PEMs was determined after their structure stabilization by nanoparticles. It has been shown that graphene oxide flakes immobilization within PEMs resulted in significant increase of VEGF adsorption. This tendency was found for all preliminary investigated non-cross-linked polyelectrolyte multilayer films. The VEGF concentration was significantly higher for PLL/HA (2700 ± 15 pg cm⁻²) and PLL/ALG (2630 ± 20 pg cm⁻²) than for Chi/CS (2440 ± 25 pg cm⁻²) film. The PLL/HA coating was chosen for further considerations (Fig. 5B). Samples were cross-linked by 400 mM NHS/EDC or 10 mM genipin. Obtained results have shown significant decrease of adsorption rate for PLL/HA film after NHS/EDC chemical cross-linking and opposite response after genipin treatment. The growth factor concentration within 400 mM NHS/EDC cross-linked sample was 2300 ± 23 pg cm⁻², whereas protein retention within coating treated by 10 mM genipin reached 2780 ± 21 pg cm⁻².

The VEGF adsorption on PEMs was also examined after silicon carbide nanoparticles immobilization by PACVD method (Fig. 5B). Coatings with SiC indicated higher protein concentration than those without nanoparticles. The highest adsorption values were noticed for non-cross-linked multilayers. The VEGF concentration was again significantly higher for PLL/HA (2560 ± 24 pg cm⁻²) and PLL/ALG (2420 ± 16 pg cm⁻²) films than for Chi/CS (2100 ± 20 pg cm⁻²) multilayer. More detailed investigations were preformed based on 400 mM NHS/EDC cross-linked and 10 mM genipin cross-linked PLL/HA multilayer coatings. Cross-linking has changed efficiency of protein adsorption significantly. There was difference between applied cross-linker type. The growth factor concentration reached 2350 ± 25 pg cm⁻² for film cross-linked by 400 mM NHS/EDC, whereas after genipin cross-linking was equal to 2610 ± 20 pg cm⁻².

The growth factor concentration was also determined in PEMs with silver nanoparticles incorporated by in situ nucleation method (Fig. 5B). It was found that for each type of film, nanoparticles immobilization led to decrease of adsorption efficiency. Moreover, NHS/EDC cross-linking process enhanced this effect in all cases. Detailed description was based on PLL/HA films. The non-cross-linked coatings containing Ag indicated significantly lower VEGF concentration (1930 ± 18 pg cm⁻²) than already characterized non-cross-linked films without nanoparticles (2400 ± 20 pg cm⁻²). There was significant difference between non-cross-linked and 10 mM cross-linked (2110 ± 15 pg cm⁻²) films, both with silver nanoparticles. The lowest adsorption effect was noticed for 400 mM cross-linked Ag modified film (1670 ± 27 pg cm⁻²).

It was found that VEGF adsorption efficiency of PEMs modified by nanoparticles has changed according to the graphene > silicon carbide > silver relation.

The PLL/HA film with the highest protein retention capacity was chosen among characterized types of PEMs for VEGF release kinetics evaluation. The comparison of the VEGF release was performed for non-cross-linked, 400 mM NHS/EDC cross-linked and 10 mM genipin cross-linked films (Fig. 6A). It has been shown that unbound and/or loosely bound growth factor was mostly desorbed during first 48 hours. Herein, the highest release rate was noticed for 400 mM NHS/EDC cross-linked film, whereas the slowest VEGF desorption was observed for 10 mM genipin treated coating.

Fig. 6 The cumulative VEGF release kinetics from non-cross-linked, 400 mM NHS/EDC or 10 mM genipin cross-linked PLL/HA films: (A) without nanoparticles; (B) modified by graphene flakes; (C) modified by SiC nanoparticles; (D) modified by silver nanoparticles. Data represent mean ± SD; n = 6; P < 0.05.

Non-cross-linked films exhibited the highest level of protein release (800 ± 26 pg cm⁻²) after 24 hour period. Then, the 9 day decreasing release was observed. It was found that after 10 days VEGF was completely washed out from the coating. Generally, genipin cross-linked films showed a lower release rate than non-cross-linked coatings. The highest level of protein desorption (700 ± 24 pg cm⁻²) was noticed after 48 hour period. Then, growth factor have been releasing gradually on low concentration level over 12 days. After experimental period small amounts of the protein were still adsorbed within coating. The shortest 8 day release time was noticed for 400 mM NHS/EDC cross-linked coatings. The highest VEGF release (1150 ± 35 pg cm⁻²) was observed after 4 hours. Then, the 400 mM cross-linked coating promoted sustained low level growth factor desorption. The VEGF release curves had a classic power law shape.

The release kinetics was also determined for PEMs containing nanoparticles (Fig. 6B–D). Generally, it has been shown that graphene or SiC immobilization within PEMs resulted in significant decrease of VEGF release rate in comparison to their counterparts without nanoparticles. Contrary, coatings with Ag nanoparticles indicated the weakest VEGF binding strength and therefore the shortest release time of analyzed growth factor. Moreover, it was found that release time was shorter for 400 mM NHS/EDC cross-linked films than for 10 mM genipin treated coatings, no matter what type of nanoparticles was immobilized.

Non-cross-linked films containing graphene flakes exhibited the highest level of protein desorption (575 ± 33 pg cm⁻²) after 24 hour period (Fig. 6B). Then, coatings promoted sustained low level growth factor release over 12 days. After that time, small amounts of VEGF were still adsorbed within films. Genipin cross-linked films containing graphene flakes showed a lower release rate than non-cross-linked coatings. The highest release was observed within 48 hours of the experiment (after 24 hours – 440 ± 23 pg cm⁻²; after 48 hours – 460 ± 21 pg cm⁻²). Then, growth factor have been releasing gradually on low concentration level over next 10 days. The shortest 10 day release time was noticed for 400 mM NHS/EDC cross-linked coatings. After that time, there was no VEGF release observed, however it was found that protein still left within coating (400 ± 16 pg cm⁻²). Herein, the highest VEGF release (830 ± 34 pg cm⁻²) was observed within first 4 hours of the experiment.

The highest level of protein desorption (573 ± 27 pg cm⁻²) for non-cross-linked SiC nanoparticles containing coatings was noticed after 24 hours (Fig. 6C). The gradual growth factor release was observed over 12 days. After that time, small amounts of VEGF were still adsorbed within coating (252 ± 22 pg cm⁻²). Genipin cross-linked films with SiC nanoparticles showed lower release rate than non-cross-linked coatings. The highest level of protein desorption was noticed during 48 hours of release kinetics investigation (after 24 hours – 495 ± 15 pg cm⁻²; after 48 hours – 532 ± 21 pg cm⁻²). After 12 days, the VEGF protein was still adsorbed within coating (520 ± 24 pg cm⁻²). The highest VEGF release (890 ± 18 pg cm⁻²) for 400 mM NHS/EDC cross-linked coatings was observed within the first 4 hours of the experiment. Then, the 400 mM cross-linked coating promoted sustained low level growth factor desorption. It was found that the growth factor was not completely released during 12 day investigations (about 385 ± 14 pg cm⁻² of VEGF left).

Non-cross-linked films containing silver nanoparticles exhibited the highest level of growth factor desorption (820 ± 30 pg cm⁻²) after 24 hours (Fig. 6D). Then the VEGF was gradually released until complete washed out from the coating after 8 days. Genipin cross-linked films containing silver nanoparticles showed a slightly lower release rate than non-cross-linked coatings. The highest protein desorption peak (740 ± 27 pg cm⁻²) was noticed after 48 hours. Moreover, it has been shown that all adsorbed VEGF amount was washed out from the film during 10 day experimental period. The shortest 5 day release time was noticed for 400 mM NHS/EDC cross-linked coatings. The highest VEGF desorption (1340 ± 36 pg cm⁻²) was observed after 4 hours. Then, decreasing amounts of growth factor were releasing from the coating.

3.5. HUVECs proliferation and morphology

Results obtained in this study have shown that the VEGF adsorption and then release kinetics depends on film's modification type. Therefore, further experimental work was aimed at determination of growth factor release influence on coatings potential to endothelialization. Endothelial cells response on chemo-mechano-activation effect through VEGF immobilization and film structural modification was investigated. Endothelial cells' adhesion after 4 hours of culture and then 12 day proliferation was determined for PLL/HA film variants with adsorbed VEGF and compared with coatings without immobilized growth factor (Fig. 7A). Changes in HUVECs proliferation rate were analyzed on non-cross-linked, 10 mM genipin cross-linked and 400 mM NHS/EDC cross-linked multilayer. In case of VEGF modified samples, it has been found that cells adhesion was the most effective on genipin treated coatings, whereas the lowest number of cells settled non-cross-linked films. The same tendency was observed in proliferation process. For each analyzed coating type, the cells number reached plateau after 10 days of culture. Cells adhesion not supported by VEGF was significantly lower in comparison with growth factor containing surfaces. The 400 mM NHS/EDC cross-linked films were the most effectively settled by cells, whereas the lowest number of HUVECs was observed for non-cross-linked coatings. For each sample type cells proliferation plateau was obtained after 8 days.

Fig. 7 HUVECs adhesion and proliferation on non-cross-linked, 400 mM NHS/EDC or 10 mM genipin cross-linked PLL/HA films: (A) without nanoparticles; (B) modified by SiC nanoparticles; (C) modified by silver nanoparticles. Solid lines indicate coatings with adsorbed VEGF and dotted lines represent films without VEGF (mean ± SD; n = 6; P < 0.05).

The adhesion and proliferation was also determined for PLL/HA films supplemented by VEGF and containing nanoparticles (Fig. 7B and C). It has been found that graphene oxide flakes immobilization within PEMs inhibited endothelial cells adhesion and therefore proliferation, either on non-cross-linked or both type of cross-linked coatings. The same inhibitory effect was found for VEGF containing and unmodified by growth factor multilayer films. Contrary, the effective cellularization process was noticed for coatings with other nanoparticle types.

Non-cross-linked films with SiC (2.2 ± 0.1 × 10³ cm⁻²) indicated more efficient cells adhesion than those without nanoparticles (1.0 ± 0.03 × 10³ cm⁻²) either for VEGF or non-VEGF modified surfaces (Fig. 7B). The same tendency was observed in proliferation process. In case of VEGF containing films, cell adhesion and proliferation was the most effective for non-cross-linked coatings. The lowest number of cells settled 400 mM NHS/EDC cross-linked films. It was found that cross-linking has also changed significantly efficiency of cells adhesion and proliferation in case of multilayers without VEGF. HUVECs number was lower on each type of cross-linked coatings in comparison with their analogous without SiC nanoparticles. The plateau effect in cells number was observed after 8 day culture for all type of samples.

The non-cross-linked coatings containing Ag indicated better potential to endothelialization than already characterized non-cross-linked films without nanoparticles or with SiC nanoparticles (Fig. 7C). Cells adhesion and then proliferation was the most effective on 10 mM genipin cross-linked coatings with VEGF and on non-cross-linked films in case of samples without growth factor. The lowest number of cells settled 400 mM NHS/EDC cross-linked films either for VEGF modified or unmodified PEMs. For each analyzed coating type, the cells number reached plateau after 8 day culture.

Beside of cells number determination, analysis of HUVECs growth included observations of their morphology after DAPI and phalloidin staining. Changes in the cell morphology were assessed based on comparison of cytoskeleton structure and nucleus localization in the cellular body. The morphology of endothelial cells cultured on the surface of the PLL/HA polyelectrolyte multilayer exhibited a typical cobblestone shape as presented in Fig. 8 and 9. There was no significant difference in the cell morphology between cultures on the non-cross-linked, various cross-linked and polystyrene control during the whole experimental period both for VEGF modified and unmodified surfaces (Fig. 8A–C and 9A–C).

Fig. 8 HUVECs morphology on PLL/HA coating without VEGF: native ((A) non-cross-linked; (B) 400 mM NHS/EDC cross-linked; (C) 10 mM genipin cross-linked); containing SiC nanoparticles ((D) non-cross-linked; (E) 400 mM NHS/EDC cross-linked; (F) 10 mM genipin cross-linked); containing Ag nanoparticles ((G) non-cross-linked; (H) 400 mM NHS/EDC cross-linked; (I) 10 mM genipin cross-linked).

Fig. 9 HUVECs morphology on PLL/HA coating with VEGF: native ((A) non-cross-linked; (B) 400 mM NHS/EDC cross-linked; (C) 10 mM genipin cross-linked); containing SiC nanoparticles ((D) non-cross-linked; (E) 400 mM NHS/EDC cross-linked; (F) 10 mM genipin cross-linked); containing Ag nanoparticles ((G) non-cross-linked; (H) 400 mM NHS/EDC cross-linked; (I) 10 mM genipin cross-linked).

From the other side, there was clearly visible influence on variety in cells' shape and cytoskeleton architecture of investigated samples containing nanoparticles (Fig. 8D–I and 9E–I). A typical endothelial cells morphology was observed only on non-cross-linked films with SiC nanoparticles both for VEGF modified and unmodified samples. Cells on non-crossed-linked coatings with Ag nanoparticles possessed a spindle-like shape and formed long chains, no matter of VEGF presence. On the surface of cross-linked samples containing SiC or Ag nanoparticles cells were rounded with the most of cytoplasma condensed to thin layer around relatively big nucleus. This observations were found for either VEGF or without growth factor coatings. In order to quantify morphology similarities between cells on all investigated samples, the shape index (SI) was evaluated using image analysis.

The SI is a dimensionless measure of cell roundness and is defined as:

SI = 4πA/p² (1)

where A is the area of a cell, and P is the length of cell perimeter. Thus, SI values range from zero for a straight line to unity for a perfect circle. Herein, the obtained value of SI differ between samples and was in range of 0.60 ± 0.02 for endothelial cells on films with proper cobblestone shape to 0.90 ± 0.01 for cells on cross-linked coatings with SiC/Ag nanoparticles and 0.40 ± 0.01 for SiC modified non-cross-linked samples with cells' chains. There were no significant differences in shape index values between coatings containing VEGF or without growth factor. The SI for control cell culture on polystyrene was 0.60 ± 0.01.

3.6. IL-6 secretion

IL-6 is a multifunctional cytokine secreted by various cells, including endothelial cells which demonstrates both pro and anti-inflammatory properties in vivo and in vitro.³²

Herein, the influence of various films on secretion of IL-6 by endothelial cell was determined under in vitro conditions. The mechanism of HUVECs response on film structural modification and VEGF immobilization was investigated.

The PLL/HA film with the highest VEGF protein retention capacity was chosen among characterized types of PEMs for IL-6 secretion analysis. The comparison of the cytokine release profile was performed for non-cross-linked, 400 mM NHS/EDC cross-linked and 10 mM genipin cross-linked films for both VEGF modified and unmodified variants (Fig. 10A). It has been shown that the highest IL-6 secretion took place in first 4 hours after cells adhesion. Herein, HUVECs on the 400 mM NHS/EDC cross-linked film produced the highest amounts of IL-6, whereas the lowest secretion was observed for cells on the non-cross-linked coating. It was found that after 8 days cytokine was not released anymore by cells proliferating on non-cross-linked and 10 mM genipn cross-linked coatings. Contrary, HUVECs on 400 mM cross-linked films have released IL-6 gradually for 12 day culture. The release kinetics profile was similar for VEGF modified and unmodified samples. However, cytokine secretion by cells on sample with growth factor was significantly lower than by HUVECs on film without VEGF. This relation was noticed for all coating types.

Fig. 10 The cumulative IL-6 secretion kinetics by HUVECs on non-cross-linked, 400 mM NHS/EDC or 10 mM genipin cross-linked PLL/HA films: (A) without nanoparticles; (B) modified by SiC nanoparticles; (C) modified by Ag nanoparticles. Solid lines indicate coatings without VEGF and dotted lines represent films with VEGF (mean ± SD; n = 6; P < 0.05).

The IL-6 release kinetics was also determined for PEMs containing nanoparticles (Fig. 10B and C). In case of films with SiC nanoparticles, the highest level of cytokine was secreted by endothelial cells culture on non-cross-linked surface variant (Fig. 10B). In the same time the lowest IL-6 level was indicated for cells on the 400 mM NHS/EDC cross-linked coating. The IL-6 release profile for 10 mM genipin cross-linked and 400 mM NHS/EDC cross-linked multilayers overlapped. It was found that after 10 days cytokine was not released anymore by cells proliferating on all type of coatings. The release kinetics profile was similar for VEGF modified and unmodified samples. However, cytokine secretion by cells on sample with growth factor was significantly lower than by HUVECs on film without VEGF. This relation was noticed for all coating types and similarly to the films without SiC nanoparticles.

The 10 mM genipin cross-linked films containing silver nanoparticles exhibited the highest level of cytokine secretion among all investigated silver containing samples (Fig. 10C). In the same time, lower IL-6 release profile overlapped was observed for non-cross-linked and 400 mM NHS/EDC cross-linked multilayer. It was found that after 8 days cytokine was not released anymore by cells proliferating on all type of coatings. The release kinetics profile was similar for VEGF modified and unmodified samples. The cytokine secretion by cells on 10 mM genipin cross-linked sample with growth factor was significantly lower than by HUVECs on film without VEGF. However, there was no significant differences between other VEGF modified and unmodified coating types.

4. Discussion

4.1. Microstructure of modified PEMs

Microstructure plays a significant role in biomaterial mechanical properties as well as interactions with human cells and tissues. In this study the internal structure of 48 bilayer PLL/HA, PLL/ALG and Chi/CS films was modified by chemical cross-linking. Several research teams have used NHS/EDC chemistry to stabilized PEMs through the covalent coupling between carboxyl and amine functional groups.^33,34 Genipin cross-linker is well known as more biocompatible reagent and could be alternative for PEMs structure stabilization by NHS/EDC chemistry.^35–37 As a natural derived compound, genipin found application even for in vivo scaffolds cross-linking. However, it should be taken into consideration that genipin mechanism of action is still not well described. It was supposed that PEMs microstructure changes are based on polymers' amine groups binding by genipin.³⁵

The influence of nanoparticles incorporation on changes in microstructure was evaluated for native (non-cross-linked) and chemically cross-linked PEM films. The applied in this work graphene oxide flakes introduced into PEMs based on electrostatic interactions consisted about 40% of the each coating type volume which was stated on TEM images. Performed observations confirmed that beside of fully hexagonal symmetric structure, graphene flakes appeared in deformed variants. The cross-linking process performed after PEM/graphene film deposition had no visible influence on films' microstructure. However, it could be postulated that cross-linking by NHS/EDC or genipin, through the competition with nanoparticles for binding positive charged polymer groups (–NH³⁺) might lead to destabilization of graphene–polyelectrolyte chain interactions.

It was found that size and distribution of SiC nanoparticles formed in the PACVD process was independent of coating type. TEM observations indicated no significant differences in nanoparticles size and zonal distribution between non-cross-linked and cross-linked coatings. The observed character of nanoparticles arrangement could be the effect of the penetration mechanism through the porous coating. The nanoparticle deposition process was based on plasma-activated nucleation from the gas phase. At first, formed nanoparticles had reached the polymer coating surface and penetrated through the film's layers without incorporation. Then nanoparticles velocity decreased and the high SiC nanoparticles density zone was created. A small number of nanoparticles penetrated through the film and formed the single SiC nanoparticles dispersion zone within lower coating layers. It was supposed that charged nanoparticles interact with polymer films in varying degree depended on multilayer native or cross-linked state. The high surface energy made them liable to undergo chemical reactions with the environment and self aggregation. There was no difference in PEMs internal structure observed based on TEM images, therefore in our opinion further FTIR investigations would be desired to give an idea about nanoparticles – polyelectrolyte multilayer interactions.

In the presented studies, the PEMs were modified by the in situ nucleation and growth of silver nanoparticles. When films where immersed in AgNO₃ solution, functional groups from polyelectrolyte chains bind and stabilized silver cations from solution. Bound Ag ions were photochemically reduced to metal nanoparticles upon exposure to UV radiation, which generates the necessary electrons to catalyze the reaction. The process was also supported by natural biopolymers posses capacity to reduce silver ions. Importantly, not only aromatic, amine or amide groups play a role in the synthesis of silver NPs, but other groups such as C–O–C and carboxylate have been shown to exhibit reducing capacities. It has been proven that hyaluronic acid can help the formation of silver NPs under UV irradiation by oxidation of CH₂OH groups to CHO.³⁸ Thus, films internal structure could be changed as the result of silver nucleation process due to contribution of peptides to the reduction of precursor ions by acting as electron donors. The cross-linking process had no visible influence on Ag modified films' microstructure. It could be postulated that cross-linking by NHS/EDC, through the competition with nanoparticles for binding negatively charged polymer groups leading to destabilization of silver–polyelectrolyte chain interactions or lower cross-linking process efficiency. Then, the reverse effect of should be found in case of genipin treatment due to cross-linker interactions with positively charged –NH₃⁺ groups.

4.2. Surface topography

Many researchers had already described the importance of biomaterial surface roughness in cellular adhesion, growth and differentiation.^39,40 The present work indicated topography changes depend on used polyelectrolytes and broad spectrum of PEMs modification. It was found that among investigated coatings in native state, the Chi/CS films are characterized by the highest roughness. The R_a parameter decreased after cross-linking procedure. In contrast, the roughness measured for PLL/HA and PLL/ALG films increased after chemical cross-linking. This results agree with those presented by Schneider et al., that shown increase in the surface roughness with either genipin or NHS/EDC cross-linking of PEMs.⁴¹ Moreover, Hillberg et al. have reported that in the case of highly irregular, gel-like films with high surface roughness like Chi/ALG, observed R_a value decrease after cross-linking might be related to water removing from the system and rigidifying the film.⁴² The observed increase in surface roughness after nanoparticles incorporation was probably the effect of already described changes in PEMs microstructure.

4.3. Mechanical properties

PEMs structure modification leading to changes in mechanical properties is well described as regulator of cell–material interactions.⁴³ Stiffness of investigated PEMs in their native state differed just in a narrow elastic modulus range. However, it has been shown that even such slight rigidity differences could be essential for biomaterial – cell/tissue interactions.⁴⁴ This study revealed that stiffness of films increased either after chemical cross-linking or nanoparticles incorporation. Previous studies of Schneider et al.,⁴¹ Richert et al.⁴⁵ and Boura et al.⁴⁶ had also demonstrated that the mechanical modulus increased upon chemical cross-linking. Herein, it has been shown for the first time to our knowledge that genipin cross-linking allowed to lower elastic modulus increase than NHS/EDC chemistry. Moreover, it was found that the graphene oxide immobilization within coatings increased their stiffness slightly in comparison to native coatings. The observed effect could be a result of graphene oxide flakes immobilization method as a mixture with negatively charged polyelectrolytes. Herein, competition between anionic forms in respect to interaction sites with polycation polymer seemed to be shifted towards graphene oxide flakes mostly. This probably affected polycation–polyanion interactions between polymers' chains and thus mechanical properties of deposited films. Chemical cross-linking increased elastic modulus. Moreover, it was noticed that application of 10 mM genipin or 400 mM NHS/EDC resulted in the similar elastic modulus value. This could be again effect of graphene–polyanion competition for interaction sites within the polycation chains. Herein, probably only a small determined number of amino groups was available for chemical cross-linking in both cross-linker cases below tested concentration efficiency. Similar to results obtained for graphene oxide modified coatings, the mechanical testing showed that SiC nanoparticles allowed to prepare PEMs with significantly higher stiffness than it was possible with application of genipin or NHS/EDC cross-linking chemistry. The applied chemical cross-linking method modified PEMs structure due to replacement of weak electrostatic interaction between polymer chains by strong covalent bonds. In case of SiC nanoparticles PEMs stiffness increased as a result of more interactions between polymer chains linked by nanoparticles as well as due to the high Young modulus of silicon carbide material. Similar mechanism was probably responsible for films rigidity increase after silver nanoparticles in situ nucleation.

4.4. VEGF adsorption and release kinetics

The present work was aimed at designing bioactive coatings for controllable surface endothelialization of cardiovascular implants. For this purpose, PEM films were functionalized by VEGF and the most favorable PEM-architecture for growth factor adsorption as well as in vitro bioactivity was selected.

Obtained results indicated that the (PLL/HA)₄₈ architecture was the most appropriate to attain the highest adsorption of VEGF among three main investigated PEM variants. There was no correlation of surface roughness with proteins adsorption efficiency as well as release kinetics. Moreover, it has been found that adsorption on positive-ending polyelectrolyte layers (PLL) was not significant in comparison to values indicated for negative-ending films (HA). Therefore, the observed phenomenon might be explained by the theory of VEGF adsorption governed mainly by electrostatic interactions between the protein and the oppositely charged functional groups in multilayer films. This conclusion was postulated by Müller et al. They found that VEGF adsorption might be govern by electrostatic interactions between the protein and the polyelectrolyte chains previously deposited on the titanium surface.⁴⁷ In this study, the VEGF had an isoelectric point of 8.5, indicating that the growth factor carries a net positive charge under applied conditions. This also could explain difference between concentration of adsorbed VEGF within cross-linked films or those modified by nanoparticles. It was found that NHS/EDC cross-linking and SiC or Ag incorporation, decreased coating potential to growth factors adsorption probably due to decrease of polymers' charged free groups number consumed for covalent bonding formation or strong interactions with nanoparticles. Contrary, genipin treatment increased VEGF adsorption level due to different mechanism of cross-linking process, where only positively charged amino groups were used in reaction. This might result in higher availability of negatively charged binding sites for VEGF. An increase in adsorption level of films with graphene oxide flakes might be also an effect of significantly high contribution of negatively charged binding sites within PEMs. On the other hand, Wu et al. have found that the level of protein adsorption onto polyelectrolyte multilayers might be influenced by the chain stiffness of the polymers forming the multilayer. Moreover, they have noticed that when a pair of semi-flexible polyelectrolytes was used then only very small amount of protein was adsorbed, irrespective of which polyelectrolyte was used to terminate the film. When the film was formed by flexible polyelectrolytes, significant protein adsorption took place.⁴⁸ Therefore, it seems that an exact adsorption mechanism of the VEGF on different modified PEM variants should be found in further investigations.

Previous experiments with native PEMs have shown that surface erosion is a dominant mechanism of many immobilized molecules release, which results in a linear release profile.⁴⁹ Herein, the VEGF release curves have a power law shape, which suggests that diffusion occurs as a second release mechanism, which accompanies the surface erosion process. It has been found that cross-linked coatings and those modified by nanoparticles might be more hydrophobic and thus less degradable by ester hydrolysis, therefore it takes longer for that films to erode and release its contents. Similar observations were made by Macdonald et al. for FGF-2 release mechanism from hydrophobic polyelectrolyte multilayer films.⁵⁰

4.5. HUVECs proliferation and morphology

Cell processes are known to by partly governed by topography of scaffolds.^39,40 Although, the precise roughness range that triggers switches in cellular response is still under the discussion, surface topography should be always considered for a newly synthesized material types. In this study, the results obtained for different PEM variants indicated that HUVECs adhesion, proliferation and IL-6 production was not correlated with roughness parameter. Therefore, we hypothesized that other coating properties like stiffness and chemical composition are crucial for investigated coatings.

The mechanical strength and modulus of biomaterials are critical for the stability of medical implant applications.⁵¹ Furthermore, certain narrow ranges of stiffness have been implicated in regulation of cellular functions such as motility and differentiation.⁵² It has already been described by Schneider et al.⁴¹ and Boura et al.⁴⁶ that an effective cells adhesion process depends on optimal rigidity of a scaffold. Endothelial cells are able to sensing mechanical differences between substrates.^53,54 This led to the fact that cells suffer from apoptosis when they are seeded on top of soft coatings, especially in the case of natural polymers.⁵⁴ Moreover, it means that during the in vitro experiments, cells death is not necessarily associated with material cytotoxicity but rather with their response to negligible adhesion to the underlying scaffold.⁵³ Therefore, cellular density depends on initial cell attachment onto polyelectrolyte monolayers.⁵⁵ Crouzier et al. have already found that PLL/HA multilayers should be cross-linked, and thus made stiffer, to ensure a good myoblast adhesion.⁵⁶ Similar observations were made for the human mesenchymal stem cells by Semenov et al.⁵⁷ and Chaubaroux et al. for human umbilical vein endothelial cells.⁵⁸

In this study, the PEMs structure stabilization by cross-linking or nanoparticles immobilization was successfully applied to control endothelial cells settlement on the multilayer films. It has been presented that the cross-linked films indicated higher potential to endothelialization than non-cross-linked multilayers probably due to their higher stiffness. Results of the mechanical properties, cellular morphology, adhesion and proliferation analysis, allowed to find a relation between the applied concentration of chemical reagent and changes in film properties. Optimal scaffold parameters were obtained for cross-linked coatings by 400 mM EDC reagent. Higher concentrations limited cells proliferation and were found as a reason of coatings delamination from the surface. Also nanoparticles immobilization increased stiffness crucial for efficient endothelialization process. Herein, cells attachment and then proliferation was lower than for coatings modified in cross-linking process, however in the same time more efficient in comparison with cellularization of the native PEMs. The application of both stabilizers together resulted in stiffness over the value found for efficient cellular adhesion and proliferation.

There is evidence in the literature that the loading of physical forces on certain cells leads to physiologic and morphologic changes.⁵⁹ It has been described that mechanical stimulation of endothelial cells activates tyrosine phosphorylations and augments the expression of cell adhesion molecules.⁶⁰ This study results also indicted significant influence of coating stiffness on endothelial cells morphology. There was similarity between the striated F-actin and polygonal shape of HUVECs on the cross-linked films and in cardiovascular tissues due to corresponding range of stiffness for both type of scaffolds. An increase in surface rigidity of non-cross-linked films after nanoparticles immobilization resulted in elongated, spindle like cells morphology. While the highest stiffness values obtained for SiC/Ag containing and cross-linked PEMs was probably responsible for rounded HUVECs shape. Cells clustering visible on films cross-linked by 10 mM genipin and modified by silver nanoparticles might be an effect of deficient cell-coating interactions. Similar as in the studies of Gaudière et al. cells might aggregate through cadherin-mediated interactions. This mechanism is responsible for cells protection against anoikis, which is type of apoptosis triggered by poor adhesion.⁶¹ The second and more optimistic concept assumes angiogenesis and tubes formation process. Anyway, both of theories need further investigations which will deliver clear explanations of observed phenomenon.

Beside mechanical properties, cellular response is modulated by surface chemistry.⁶² One approach to overcome the limitation of material composition and its effect on cellular functions is to bind ECM components/growth factors onto material. The VEGF is one of the major factors that affects endothelial cells through effects on migration, proliferation, differentiation and survival.⁶³ It has a key role in angiogenesis regulation. Native VEGF at concentrations of 150 to 300 pg ml⁻¹ has been reported to stimulate endothelial cells in vitro.⁶⁴ This cellular response is mainly mediated by the specific membrane receptor VEGFR2 which is activated after binding of VEGF.⁶⁵ In this study, HUVECs proliferation indicated that all type of prepared coatings with VEGF enhanced cells adhesion and proliferation in comparison to films without growth factor. It seems that higher adsorption and retention of growth factor in coating at the first four hours is crucial for more efficient cells adhesion and further for proliferation rate. Contrary, there was no significant difference in endothelial cells morphology between VEGF modified and unmodified samples.

The positive effect of the PEMs/VEGF films on HUVEC adhesion and proliferation seems to be the result of the joint action of the growth factor and the PEM stiffness. It seems that observed additive effect is more crucial for scaffolds with lower stiffness. Derricks et al. have reported that extracellular matrix stiffness can differentially drive endothelial response to VEGF. They found that maximal response (2–3.5 fold increases) took place at low VEGF concentrations at moderate stiffness scaffold, whereas at high concentrations of VEGF the softest substrates showed the most intense response, but the response dissipated the fastest.⁶⁶ Müller et al. have reported that PEM films act in synergy with the growth factor, but probably through an additional way different from the VEGFR2-induced pathway.⁴⁷ This is in agreement with presented in this study, the theory about additive regulatory effect of film stiffness and VEGF immobilization on endothelial cells response.

4.6. IL-6 secretion

Interleukin-6 (IL-6) secretion from endothelial cells in response to mechanical stimuli plays an important role in the regenerative and inflammatory responses. Li et al. have reported that cytokine provides immunoprivilege of cells after their implantation in biomaterial scaffold.⁶⁷ Moreover, it has been suggested that IL-6 may induce angiogenesis through direct stimulation of the motility of endothelial cells.⁶⁸ Up to date, just a little information has been provided concerning the expression of cytokines in response to scaffold stiffness. Kobayashi et al. have observed that continuous stretching enhanced IL-6 production, most likely through sequential activation of IKKs and NF-jB.⁶⁹ It has been considered that mechanical stress mediated IL-6 secretion is associated with integrin receptors. Moreover, Ebrahem et al. have demonstrated that IL-6-induced angiogenesis is VEGF dependent. In addition, directly or indirectly, both cytokines are likely to be involved in the neovascularization. Therefore, the cooperation between VEGF and IL-6 may be part of a complex regulatory mechanism.⁷⁰

In this study, the theory about additive effect of VEGF and coating stiffness influence on the IL-6 secretion by endothelial cells was investigated. The release of IL-6 from both type of surfaces with and without VEGF is positively correlated with HUVECs proliferation rate found for samples unmodified by VEGF. This observation may suggest that cytokine secretion mainly depends on mechanical properties of coating. However, the VEGF presence in the coating resulted in decrease of cytokine amounts secreted by cells. Therefore, it seems that film stiffness acts in synergy with the VEGF through different intracellular signaling pathways, which should be checked in further molecular investigations.

5. Conclusions

Chemo-mechano-sensing determined by the cell's reaction to chemical factors and forces applied at its surface seems to be an essential components of biomaterial internalization and function in tissues. Cells contact scaffolds via numerous protein complexes that engage signaling pathways in distinct ways. The mechanical stresses generated at any certain site depend on the mechanical properties of the biomaterial as well as the other signals that cell receives from chemical stimuli. Unfortunately, it is not well known how all those signals regulate each others. Therefore, this work is one of the first attempts to determined correlation of different biomaterial issues and their potential influence on regeneration of cardiovascular system.

In the current study, chemical composition and mechanical properties of films were established and described to find an optimal parameter for endothelial cells attachment, proliferation and to control IL-6 secretion. The cross-linking chemistry and nanoparticles introduction was applied in order to moderate PEMs rigidity. Biochemical modification concerned the VEGF adsorption within multilayer. We showed that PEMs/VEGF films enhance in vitro spreading and proliferation of endothelial cells, whereas VEGF presence slightly inhibited IL-6 production and release. Since non-functionalized films also contributed to adhesion and proliferation of endothelial cells and cytokine secretion, we can suppose that PEM films mechanical properties act in synergy with the growth factor, but probably through a different pathways. In conclusion, we clearly demonstrated in vitro the effectiveness of the proposed endothelialization strategy and confirmed additive effect of chemical and mechanical properties of PEMs. Moreover, it seems that performed investigations indicated PLL/HA films cross-linked by 400 mM NHS/EDC or the same coatings containing SiC nanoparticles, both types modified by VEGF, as promising biomaterials for cardiovascular applications.

Further studies should be focused on the molecular signaling pathways to prove the positive effect of VEGF-functionalized PEM films of various stiffness on the reported cellular response. Herein, the most interesting seems to be identification of the mechanism responsible for cells clustering visible on films cross-linked by 10 mM genipin and modified by silver nanoparticles. The other challenge is to identify not only the proteins and signals involved in chemo-mechano-sensing but also the principles that enable cells to measure stiffness. Then such information might be used to control cellular fate by designing scaffolds with highly precise properties.

Acknowledgements

The research was financially supported by the project No. PBS3/A7/17/2015 “Development of innovative bioactive prosthetic heart valve” of the Polish National Centre for Research and Development and project No. 2014/15/N/ST8/02601 of the Polish National Center of Science.

Notes and references

1. D. G. Yu, W. C. Lin, C. H. Lin, Y. H. Yeh and M. C. Yang, J. Biomed. Mater. Res., Part B, 2007, 83(1), 105–113.
2. A. Mzyk, R. Major, M. Kot, J. Gostek, P. Wilczek and B. Major, Archives of Civil and Mechanical Engineering, 2014, 14(2), 262–268.
3. F. Gaudière, I. Masson, S. Morin-Grognet, O. Thoumire, J. P. Vannier, H. Atmani, G. Ladam and B. Labat, Soft Matter, 2012, 8, 8327–8337.
4. M. P. Lutolf and J. A. Hubbell, Nat. Biotechnol., 2005, 23, 47–55.
5. M. Sailer, K. L. W. Sun, O. Mermut, T. E. Kennedy and C. J. Barrett, Biomaterials, 2012, 33(24), 5841–5847.
6. D. Volodkin, R. von Klitzing and H. Moehwald, Polymers, 2014, 6(5), 1502–1527.
7. G. Francius, J. Hemmerlé, J. Ohayon, P. Schaaf, J. C. Voegel, C. Picart and B. Senger, Microsc. Res. Tech., 2006, 69(2), 84–92.
8. A. Mzyk, R. Major, J. M. Lackner, F. Bruckert and B. Major, RSC Adv., 2014, 4, 31948–31954.
9. N. E. Vrana, O. Erdemli, G. Francius, A. Fahs, A. Rabineau, C. Debry, A. Tezcaner, D. Keskin and P. Lavalle, J. Mater. Chem. B, 2014, 2(8), 999–1008.
10. T. Kruk, K. Szczepanowicz, D. Kręgiel, L. Szyk-Warszyńska and P. Warszyński, Colloids Surf., B, 2016, 137, 158–166.
11. F. Boulmedais, B. Frisch, O. Etienne, P. Lavalle, C. Picart, J. Ogier, J. C. Voegel, P. Schaaf and C. Egles, Biomaterials, 2004, 25, 2003.
12. X. Zan and Z. Su, Thin Solid Films, 2010, 518(19), 5478–5482.
13. M. Matsusaki, H. Sakaguchi, T. Serizawa and M. Akashi, J. Biomater. Sci., Polym. Ed., 2007, 18(6), 775–783.
14. M. L. Macdonald, R. E. Samuel, N. J. Shah, R. F. Padera, Y. M. Beben and P. T. Hammond, Biomaterials, 2011, 32(5), 1446–1453.
15. N. J. Shah, M. N. Hyder, M. A. Quadir, N. M. D. Courchesne, H. J. Seeherman, M. Nevins and P. T. Hammond, Proc. Natl. Acad. Sci. U. S. A., 2014, 111(35), 12847–12852.
16. L. Liu, B. D. Ratner, E. H. Sage and S. Jiang, Langmuir, 2007, 23, 11168e73.
17. G. A. Hudalla and W. L. Murphy, Adv. Funct. Mater., 2011, 21, 1754e68.
18. K. S. Masters, Macromol. Biosci., 2011, 11, 1149e63.
19. M. Crombez, P. Chevallier, R. C. Gaudreault, E. Petitclerc, D. Mantovani and G. Laroche, Biomaterials, 2005, 26, 7402.
20. C. Fischbach and D. J. Mooney, Biomaterials, 2007, 28, 2069.
21. H. C. Grillo, Ann. Thorac. Surg., 2002, 73, 1995.
22. N. Ferrara, Endocr. Rev., 2004, 25, 581.
23. L. E. Freed, F. Guilak, X. E. Guo, M. L. Gray, R. Tranquillo and J. W. Holmes, J. Tissue Eng., 2006, 12, 3285–3305.
24. J. Li, M. Ding, Q. Fu, H. Tan, X. Xie and Y. Zhong, J. Mater. Sci.: Mater. Med., 2008, 19(7), 2595–2603.
25. N. Salmon, E. Paternotte, V. Decot, J. F. Stoltz, P. Menu and P. Labrude, Bio-Med. Mater. Eng., 2008, 19(4–5), 349–354.
26. V. Moby, P. Labrude, A. Kadi, L. Bordenave, J. F. Stoltz and P. Menu, J. Biomed. Mater. Res., Part A, 2011, 96(2), 313–319.
27. A. Schneider, G. Francius, R. Obeid, P. Schwinté, B. Frisch, P. Schaaf, J. C. Voegel, B. Senger and C. Picart, Langmuir, 2006, 22, 1193–1200.
28. W. S. Hummers Jr and R. Offeman, J. Am. Chem. Soc., 1958, 80(6), 1339.
29. J. M. Lackner, W. Waldhauser, R. Major, L. Major and P. Hartmann, Surf. Coat. Technol., 2013, 215, 192–198.
30. J. M. Lackner, M. Wiesinger, R. Kaindl, W. Waldhauser, D. Heim and P. Hartmann, Plasma Chem. Plasma Process., 2014, 34(2), 259–269.
31. L. Shen, L. Rapenne, P. Chaudouet, J. Ji and C. Picart, J. Colloid Interface Sci., 2012, 388(1), 56–66.
32. J. Scheller, A. Chalaris, D. Schmidt-Arras and S. Rose-John, Biochim. Biophys. Acta, Mol. Cell Res., 2011, 1813(5), 878–888.
33. T. Boudou, T. Crouzier, K. F. Ren, G. Blin and C. Picart, Adv. Mater., 2010, 2, 441–467.
34. T. Crouzier, T. Boudou and C. Picart, Curr. Opin. Colloid Interface Sci., 2010, 15, 417–426.
35. H. W. Sung, R. N. Huang, L. L. Huang and C. C. Tsai, J. Biomater. Sci., Polym. Ed., 1999, 10, 63–78.
36. V. Chiono, E. Pulieri, G. Vozzi, G. Ciardelli, A. Ahluwalia and P. Giusti, J. Mater. Sci.: Mater. Med., 2008, 19, 889–898.
37. L. P. Yan, Y. J. Wang, L. Ren, G. Wu, S. G. Caridade, J. B. Fan, L. Y. Wang, P. H. Ji, J. M. Oliveira, J. T. Oliveira, J. F. Mano and R. L. Reis, J. Biomed. Mater. Res., Part A, 2010, 95, 465–475.
38. X. Q. Cui, C. M. Li, H. F. Bao, X. T. Zheng, J. F. Zang, C. P. Ooi and J. Guo, J. Phys. Chem. C, 2008, 112, 10730–10734.
39. C. M. Lo, H. B. Wang, M. Dembo and Y. L. Wang, Biophys. J., 2000, 79, 144–152.
40. K. Anselme, A. Ponche and M. Bigerelle, Proc.–Inst. Mech. Eng., 2010, 224, 1487–1507.
41. A. Schneider, L. Richert, G. Francius, J. V. Voegel and C. Picart, Biomed. Mater., 2007, 2(1), 45–51.
42. A. L. Hillberg, C. A. Holmes and M. Tabrizian, Biomaterials, 2009, 30(27), 4463–4470.
43. V. Gribova, R. Auzely-Velty and C. Picart, Chem. Mater., 2012, 24, 854–869.
44. S. Miserere, S. Ledru, N. Ruille, S. Griveau, M. Boujtita and F. Bedioui, Electrochem. Commun., 2006, 8, 238–244.
45. L. Richert, F. Boulmedais, P. Lavalle, J. Mutterer, E. Ferreux, G. Decher, P. Schaaf, J. Voegel and C. Picart, Biomacromolecules, 2004, 5(2), 284–294.
46. C. Boura, P. Menu, E. Payan, C. Picart, J. C. Voegel, S. Muller and J. F. Stoltz, Biomaterials, 2003, 24, 3521–3530.
47. S. Müller, G. Koenig, A. Charpiot, C. Debry, J. C. Voegel, P. Lavalle and D. Vautier, Adv. Funct. Mater., 2008, 18(12), 1767–1775.
48. B. Wu, G. Liu, G. Zhang and V. S. Craig, Soft Matter, 2014, 10(21), 3806–3816.
49. N. J. Shah, M. L. Macdonald, Y. M. Beben, R. F. Padera, R. E. Samuel and P. T. Hammond, Biomaterials, 2011, 32(26), 6183–6193.
50. M. L. Macdonald, N. M. Rodriguez, N. J. Shah and P. T. Hammond, Biomacromolecules, 2010, 11(8), 2053–2059.
51. J. P. Winer, A. Chopra, J. Y. Kresh and P. A. Janmey, Mechanobiology of Cell-Cell and Cell-Matrix Interactions, Springer US, 2011, pp. 11–22.
52. A. Joy, D. M. Cohen, A. Luk, E. Anim-Danso, C. Chen and J. Kohn, Langmuir, 2011, 27(5), 1891–1899.
53. C. P. Vazquez, T. Boudou, V. Dulong, C. Nicolas, C. Picart and K. Glinel, Langmuir, 2009, 25(6), 3556–3563.
54. C. Picart, Curr. Med. Chem., 2008, 15(7), 685–697.
55. S. Chien, Am. J. Physiol.: Heart Circ. Physiol., 2007, 292(3), H1209–H1224.
56. T. Crouzier, K. Ren, C. Nicolas, C. Roy and C. Picart, Small, 2009, 5(5), 598–608.
57. O. V. Semenov, A. Malek, A. G. Bittermann, J. Vörös and A. H. Zisch, Tissue Eng., Part A, 2009, 15(10), 2977–2990.
58. C. Chaubaroux, E. Vrana, C. Debry, P. Schaaf, B. Senger, J. C. Voegel, Y. Haikel, C. Ringwald, J. Hemmerlé, P. Lavalle and F. Boulmedais, Biomacromolecules, 2012, 13(7), 2128–2135.
59. J. C. Kohn, D. W. Zhou, F. Bordeleau, A. L. Zhou, B. N. Mason, M. J. Mitchell, M. R. King and C. A. Reinhart-King, Biophys. J., 2015, 108(3), 471–478.
60. Z. Noureddine, O. Dubrovskyi and K. G. Birukov, Microvasc. Res., 2012, 83(1), 71–81.
61. F. Gaudière, S. Morin-Grognet, L. Bidault, P. Lembré, E. Pauthe, J.-P. Vannier, H. Atmani, G. Ladam and B. Labat, Biomacromolecules, 2014, 15(5), 1602–1611.
62. B. N. Mason, A. Starchenko, R. M. Williams, L. J. Bonassar and C. A. Reinhart- King, Acta Biomater., 2013, 9(1), 4635–4644.
63. M. L. Chrétien, M. Zhang, M. R. Jackson, A. Kapus and B. Langille, J. Cell. Physiol., 2010, 224(2), 352–361.
64. T. A. Brock, H. F. Dvorak and D. R. Senger, Am. J. Pathol., 1991, 138, 213.
65. S. Cebe-Suarez, A. Zehnder-Fjallman and K. Ballmer-Hofer, Cell. Mol. Life Sci., 2006, 63, 601.
66. K. E. Derricks, V. Trinkaus-Randall and M. A. Nugent, FASEB J., 2013, 27(1), 646–711.
67. P. Li, S. H. Li, J. Wu, W. F. Zang, S. Dhingra, L. Sun, R. D. Weisel and R. K. Li, J. Cell. Mol. Med., 2013, 17(9), 1136–1145.
68. C. K. Poh, Z. Shi, T. Y. Lim, K. G. Neoh and W. Wang, Biomaterials, 2010, 31(7), 1578–1585.
69. S. Kobayashi, M. Nagino, S. Komatsu, K. Naruse, Y. Nimura, M. Nakanishi and M. Sokabe, Biochem. Biophys. Res. Commun., 2003, 308(2), 306–312.
70. Q. Ebrahem, A. Minamoto, G. Hoppe, B. Anand-Apte and J. Sears, Invest. Ophthalmol. Visual Sci., 2006, 47(11), 4935.

---