In vitro study of the host responses to model biomaterials via a fibroblast/macrophage co-culture system

Abstract

Surface properties are believed to play important roles in initial inflammatory and subsequent wound healing/fibrotic responses after implantation of biomaterials. To investigate the surface property effect in mediating these host responses, we used an in vitro fibroblast/macrophage co-culture model established with a cell migration chamber, and a series of self-assembling monolayers (SAMs) bearing different terminal groups as model surfaces to study the effect of surface properties on macrophage fusion, fibroblast attachment, spreading morphology, proliferation, outgrowth, as well as pro-(interleukin-6) and anti-(interleukin-10) inflammatory cytokine production, expression of ED-A fibronectin (FN) and alpha-smooth muscle actin (α-SMA). The obtained results show that the hydrophobic CH₃ surface caused high levels of inflammatory but low levels of wound healing/fibrotic responses, while the hydrophilic/anionic COOH surface resulted in both low levels of inflammatory and wound healing/fibrotic responses. Interestingly, the hydrophilic OH surface was found to possess a low potential of inducing inflammatory responses but high potential of inducing wound healing/fibrotic responses. These results reveal that the extent of inflammation and wound healing/fibrosis might not be always related in vitro. However, more important is the observation of the macrophage contributions in facilitating the wound healing and fibrotic responses by up-regulation of fibroblast outgrowth, cytokine production as well as ED-A FN and α-SMA expression. Overall, by linking the surface properties to cell activities with our established fibroblast/macrophage co-culture system, we could provide an useful model system for in vitro studies to design more biocompatible biomaterials for various biomedical and tissue engineering applications.

---

1. Introduction

The implantation of biomaterials often triggers a series of host responses, among which chronic inflammation and fibrotic encapsulation at the end stage are important concerns for the design of biocompatible materials.¹ Macrophages and fibroblasts are two dominating effector cells involved in the cascade of host reactions.² Although macrophages are more dominant during the inflammatory responses while fibroblasts are more crucial during wound healing/fibrotic responses, the regulation of the whole host reactions needs the orchestration of both cell types by soluble signals such as cytokines, chemokines, and growth factors, but also fixed cues through direct cell–cell contacts.³ Additionally, we and many others have shown that macrophages, in the presence of foreign materials, can produce significant amounts of pro-inflammatory and pro-fibrotic cytokines, such as interleukin-1β (IL-1β), interleukin-6 (IL-6), tumour necrosis factor-α (TNF-α), and transforming growth factor-β (TGF-β), which are believed to affect the ensuing fibrotic reactions.^4–6 Acting on these signals, fibroblasts can be activated in terms of up-regulation of fibroblast proliferation, migration, expansion, and the synthesis of extracellular matrices (ECM) such as collagens, and finally leading to implant encapsulation.⁷ Thereupon, fibroblasts can differentiate into myofibroblasts, characterized by expression of an ED-A splice variant of fibronectin (ED-A FN) and α-smooth muscle actin (α-SMA), which are incorporated into stress fibres and thereby produce contractile forces to promote wound healing, but also scar formation.⁸ The normal wound healing processes are desired after implantation; however, the prolonged presence of myofibroblasts can lead to excessive collagen production and tissue contraction, which turns into fibrosis ultimately.⁹ Since fibrosis is one of the key factors leading to the failure of medical implants, it has become a ubiquitous problem and global burden.

In order to know more about fibrosis and improve the implant function and longevity, intensive research work has been done to study the host responses by using different in vitro and in vivo models.^10–13 Due to the key effects of macrophages and fibroblasts in the regulation of inflammatory and wound healing/fibrotic reactions, different in vitro macrophage/fibroblast co-culture models have been used to study inflammation and fibrosis.^14,15 However, most of the studies focused only on the pro-inflammatory responses while the studies of the fibrotic reactions are largely missing.¹⁶ On the other hand, self-assembling monolayers (SAMs) or micron-sized particles with different functional groups, were implanted in subcutaneous air pouches of a mice model to investigate the surface chemistry effect on fibrous capsule formation.¹⁷ But a systematic study to correlate the surface properties with a wide scale of inflammatory reactions and fibrotic processes is still lacking. Furthermore, macrophages and fibroblasts were used to correlate the effect of surface properties on cellular behaviour, during which conditioned medium from one cell type was used to challenge the other one.¹⁸ This leads to the lack of in situ signal exchanges, and also a more accurate method to mimic the complexity of the in vivo situations is missing.

To overcome the limitations of the previous models, we present here an in vitro fibroblast/macrophage co-culture system with the use of migration fence chambers, which possess an internal and an external compartment, to allow seeding of fibroblast and macrophages separately on model biomaterials. After 24 h incubation, the fence chambers were removed to allow the exchange of soluble signals, and at the same time, for fibroblast outgrowth, macrophage migration and interaction between the two cell types. THP-1 cells were applied and differentiated into macrophage-like cells by treatment with phorbol-12-myristate-13-acetate (PMA)¹⁹ due to their uniform genetic background with the advantage of no donor variation, and also these are more easily obtained from cell cultures than the monocytes isolated from blood.²⁰ Here, PMA-induced, but not M1 or M2 polarized macrophages were used due to the flexibility of the PMA-induced macrophage model, which can allow the switch from one to another phenotype when responding to different stimuli.²¹ On the other hand, SAMs possess a well-defined chemistry to tailor the material surfaces to obtain specific surface properties and are suitable to study the interactions with proteins and cells.^22,23 Therefore, SAMs with methyl (CH₃), amine (NH₂), hydroxyl (OH) and carboxyl (COOH) terminated groups having different wettabilities and surface potentials were used here as model surfaces to compare the surface chemistry effect on the host responses by fibroblasts and macrophages. The host reactions on different SAM surfaces were studied with respect to macrophage fusion, fibroblast attachment, spreading morphology, fibroblast proliferation, outgrowth, as well as pro-(IL-6) and anti-(IL-10) inflammatory cytokine production and expression of two markers for myofibroblasts, namely ED-A FN and α-SMA. The results are reported herein.

2. Materials and methods

2.1 Preparation of self-assembling monolayers (SAMs)

Round glass cover slides (∅ 15 mm, Menzel, Germany) were cleaned with 0.5 M NaOH in 96% ethanol (Roth, Germany) for 2 h. Subsequently, the slides were extensively rinsed with double-distilled water (10 × 5 min) and dried under nitrogen flow.

SAMs with different terminated groups were obtained by immersing the cleaned glass slides in different silane solutions. Chlorodimethyloctadecylsilane (ODS), 3-aminopropyltriethoxysilane (APTES), glycidoxypropyl trimethoxysilane (GPTMS) and triethoxysilylpropyl succinic anhydride (TESPSA) were all obtained from ABCR (Karlsruhe, Germany). The formation of SAMs was performed according to the previously described report²⁴ with slight modifications, and is also reported in our previous work.¹⁵ Briefly, CH₃-terminated SAMs were generated by immersion of clean glass in a 5% (v/v) solution of ODS in n-hexane for 1 h at room temperature. Then the surfaces were washed with n-hexane (2 × 5 min), ethanol (2 × 5 min), and double-distilled water (6 × 5 min) and dried under nitrogen flow. NH₂-, epoxy- and COOH-terminated surfaces were obtained by immersion of clean glass in 1% (v/v) solution of APTES, GPTMS and TESPSA, respectively, in ethanol for 16 h at room temperature. After immersion, the surfaces were rinsed extensively with ethanol, washed with double-distilled water (10 × 5 min) and dried with nitrogen. The OH-terminated surfaces were obtained by further treatment of epoxy-terminated surfaces in 100 mM HCl at 80 °C for 1 h.²⁴

2.2 Characterization of the surface properties of SAMs

2.2.1 Water contact angle measurements. Static water contact angle (WCA) measurements were carried out using the sessile drop method at room temperature with an OCA 15+ device from Dataphysics (Filderstadt, Germany). Five droplets of 2 μL ultrapure water were dropped onto each surface and the obtained values were used to calculate means and standard deviations. The experiments were conducted in triplicate and the means and standard deviations of two independent experiments were calculated.

2.2.2 Zeta potential measurements. The zeta potentials of different SAMs were measured with a SurPASS device (Anton Paar, Graz, Austria). Two SAM-coated glass slides were fixed and placed oppositely into the adjustable gap cell. The width of the flow cell was adjusted to a distance where a flow rate of 100 to 150 mL min⁻¹ was achieved at a maximum pressure of 300 mbar. A flow check was performed to achieve a constant flow in both directions. 1 mM potassium chloride (KCl) was used as a model electrolyte and 100 mM hydrochloric acid (HCl) was used for pH titration. The pH value of KCl was adjusted to pH 10.5 using 1 M sodium hydroxide (NaOH) before starting the measurement. Finally, the measurements were performed by using an automated titration program using titration steps of 0.03 μL from pH 10.5 to 5.0 and 0.25 μL from pH 5.0 to 3.0.

2.3 Cell experiments

2.3.1 Cell culture. Primary human dermal fibroblasts (HF, PromoCell, Heidelberg, Germany,) were grown in Dulbecco's modified Eagle's medium (DMEM, Biochrom AG, Berlin, Germany) supplemented with 10% fetal bovine serum (FBS, Biochrom AG), 1% antibiotic-antimycotic solution (AAS, Sigma-Aldrich, Germany) at 37 °C in a humidified 5% CO₂/95% air atmosphere using a NUAIRE® DH Autoflow incubator (NuAire Corp., Plymouth, Minnesota, USA). The cells were harvested after confluence by treatment with 0.25% trypsin/0.02% EDTA (Biochrom AG) for 5 min at 37 °C. The fibroblasts were used before passage 10 in the present study.

The cells of the human monocytic cell line THP-1 (DSMZ, Braunschweig, Germany) were cultured in RPMI-1640 medium (Biochrom AG, Germany) supplemented with 10% (v/v) FBS (Biochrom AG) and 1% (v/v) AAS (Sigma-Aldrich, Germany) at 37 °C in a humidified 5% CO₂/95% air atmosphere. The suspended cells were separated by centrifugation. The old medium was removed and the cell pellet was resuspended in fresh medium every second day in order to maintain a cell density of 0.5–1.0 × 10⁶ cells per mL. The macrophages were obtained by incubation of THP-1 cells with 200 nM phorbol-12-myristate-13-acetate (PMA, Sigma-Aldrich, Germany) for 48 h in T75 cell culture flasks (Greiner Bio-One, Frickenhausen, Germany).²⁵ Afterwards, the macrophages were detached by 0.25% trypsin/0.02% EDTA (Biochrom AG) and used for seeding in the co-culture systems. The characterization of macrophages by immunofluorescence and flow cytometry was performed in separate experiments and is outlined in more detail in the ESI.†

2.3.2 Establishment of the co-culture model. Cell migration fence chambers (Aix Scientifics, Aachen, Germany),²⁶ which possess internal and external compartments were used to separate fibroblasts and macrophages at the beginning. Fig. 1 shows the scheme for establishment of the fibroblast/macrophage co-culture model with cell migration fence chambers in a 24-well plate. 0.15 mL of a HF cell suspension with a cell density of 2.5 × 10⁴ cells per mL was seeded into the internal channel of the fence chamber, while 0.65 mL of a macrophage cell suspension with a cell density of 5.0 × 10⁵ cells per mL was seeded into the external compartment of the chamber for each well. The HF mono-culture contained the same conditions as HFs but with only 0.65 mL RPMI in the external compartment. After incubation for 24 h, the fence chambers were removed, resulting in a gap of approximately 500 μm between the two cell types, allowing for the exchange of soluble signals and interactions between the two cell types.

Fig. 1 Schematic description for the establishment of the fibroblast/macrophage co-culture model using the cell migration fence chamber.

2.3.3 HF adhesion studies. HF cells were seeded on SAM surfaces at a cell density of 2.5 × 10⁴ cells per mL in serum-containing DMEM medium. Cultures were incubated at 37 °C in humidified 5% CO₂/95% air for 4 h. Afterwards, the surfaces were gently washed with PBS once to remove the unbound HF cells. The attached cells were fixed and stained by 0.5% (w/v) crystal violet (Carl Roth, Karlsruhe, Germany) in methanol. The images of adherent HF cells were taken with a light microscope (Axiovert 100, Carl Zeiss MicroImaging GmbH, Germany) equipped with a CCD camera (Sony, MC-3254, AVT-Horn, Aalen, Germany). The cell number and area on different surfaces were evaluated by using ImageJ (version 1.46r).

2.3.4 Immunofluorescence staining of HF cells. The HF cells were treated and incubated in the same way as described above. After 24 h incubation, the attached HF cells on different surfaces were fixed with 4% (w/v) paraformaldehyde (PFA, Sigma-Aldrich, Germany) in PBS for 15 min. Then the cells were permeabilized with 0.1% (v/v) Triton X-100 (Sigma, Germany) for 10 min. After two time rinsing with PBS, the non-specific binding sites were blocked by incubation with 1% (w/v) bovine serum albumin (BSA, Merck, Germany) for 1 h. Antibodies were diluted in 1% (w/v) BSA in PBS. The cells were firstly incubated with a mouse monoclonal antibody against vinculin (1 : 100, Sigma, Germany) for 30 min at room temperature. Subsequently, after washing with PBS twice, the secondary goat-anti-mouse antibody conjugated with CY2 (1 : 100, Dianova, Germany) were incubated with the cells for another 30 min. Actin and nuclei were stained by using Bodipy® Phalloidin (1 : 50, Invitrogen, Germany) and TO-PRO3 (1 : 500, Invitrogen, Germany) by 30 min incubation each. The samples were then washed, mounted with Mowiol (Calbiochem, Darmstadt, Germany) and examined with confocal laser scanning microscopy (CLSM, LSM 710, Carl Zeiss, Oberkochen, Germany) using a 63× oil immersion objective. Images were processed with the ZEN2011 software (Carl Zeiss, Oberkochen, Germany).

2.3.5 HF proliferation studies. Different SAM-coated glass surfaces were placed into 24-well tissue culture plates (Greiner Bio-One GmbH, Frickenhausen, Germany), sterilized with 70% ethanol for 10 min and rinsed with sterile PBS twice. 1 mL of a HF cell suspension was added to each surface at 2.0 × 10⁴ cells per mL in serum-containing DMEM medium. Cultures were incubated for 1 and 3 days respectively at 37 °C in humidified 5% CO₂/95% air. After the desired incubation time, a QBlue® cell viability assay kit (Biochain, Hayward, USA) was applied to quantify the amount of metabolic active cells. Briefly, the cells were firstly washed once with sterile PBS to remove the old medium. Then, 500 μL of pre-warmed colourless DMEM with the QBlue® assay reagent (10 : 1) were added to each well and incubated at 37 °C for 2 h. Thereafter, 100 μL of the supernatant from each well was transferred to a black 96-well plate, and the fluorescence intensities were measured at an excitation wavelength of 544 nm and emission wavelength of 590 nm with a plate reader (FLUOstar, BMG LabTech, Offenburg, Germany).

2.3.6 HF outgrowth studies. To measure the HF outgrowth on different SAM surfaces, 0.15 mL of a HF cell suspension with a cell density of 2.0 × 10⁵ cells per mL was seeded into the internal channel, while 0.65 mL of a macrophage cell suspension with a cell density of 5.0 × 10⁵ cells per mL was seeded into the external compartment of the chamber for each well in a 24-well tissue plate. The HF mono-culture was performed under the same conditions but with only 0.65 mL RPMI in the external compartment. After 24 h incubation, the fence chambers were removed and the surfaces were washed once with PBS. 1 mL of fresh serum-containing DMEM was added to each well. The fibroblasts were then left to outgrow for 3 days in both HF mono- and co-cultures. Afterwards, the cells were fixed and stained by 0.5% (w/v) crystal violet in methanol. HF outgrowth areas were photographed by using a stereo microscope (SMZ-168, Motic, Wetzlar, Germany) equipped with an attachable C-mount camera (Moticam 1000, Motic, Germany) to evaluate the outgrowth distances on different SAM surfaces. The macrophage regions were viewed by using an inverted contrasting microscope (Leica DMIL, Leica Mikrosysteme Vertrieb GmbH, Wetzlar, Germany) equipped with a high speed digital camera (Leica EC3, Leica Mikrosysteme Vertrieb GmbH, Germany) to evaluate the foreign body giant cell (FBGC) formation on different SAM surfaces.

2.3.7 Cytokine production assays. 0.15 mL of a HF cell suspension with a density of 2.0 × 10⁴ cells per mL in serum-free DMEM and 0.65 mL of a macrophage cell suspension with a density of 5.0 × 10⁵ cells per mL in 10% FBS-containing RPMI were added into the internal channel and external compartment of the chamber, respectively. The HF mono-cultures were grown under the same conditions but with only 0.65 mL RPMI in the external compartment. After 24 h incubation, the fence chambers were removed and the surfaces were washed once with PBS. 1 mL of 10 ng mL⁻¹ TGF-β1 (PeproTech, Hamburg, Germany) in serum-free DMEM was then added to each well and incubated with the cells for 3 days at 37 °C. After that, supernatants in both HF mono- and HF/macrophage co-cultures were collected for analysing the pro-inflammatory cytokine IL-6 and anti-inflammatory cytokine IL-10 production. The cytokine production on different SAM surfaces was detected using enzyme-linked immunosorbent assay (ELISA) according to the manufacturers’ instructions for the IL-6 Elisa kit (BD Biosciences Pharmingen, Heidelberg, Germany) and the IL-10 Elisa kit (Promokine, Heidelberg, Germany).

2.3.8 Expression of ED-A FN and α-SMA. The cells were treated and incubated as described in 2.3.7. After incubation with 10 ng mL⁻¹ TGF-β1 for 3 days, the cells were washed, fixed by using 4% PFA, permeabilized with 0.1% Triton X-100, and blocked with 1% BSA. Then, the cells were incubated against the primary antibody ED-A FN (1 : 100, Santa Cruz Biotechnology, Germany), secondary antibody CY2 (1 : 100, Dianova, Germany), CY3-conjugated anti-α-SMA (1 : 200, Sigma-Aldrich, Germany) and TO-PRO3 (1 : 500, Invitrogen, Germany) one after another, by 30 min incubation each. The samples were then washed, mounted with Mowiol and examined with CLSM (Carl Zeiss, Oberkochen, Germany) using a 20× objective. The images were processed with the ZEN2011 software (Carl Zeiss, Germany).

2.4 Statistics

All data are represented as mean values ± standard deviations (SD). Statistical examination was performed using one-way analysis of variance (ANOVA) followed by post-hoc Tukey testing. The significance level was set at p < 0.05 and indicated by an asterisk. The number of samples has been indicated in the respective figure caption.

3. Results and discussion

3.1 Surface properties of SAMs

The wettability of the SAM surfaces was evaluated by static water contact angle (WCA) measurements (Fig. 2A). As expected, –CH₃ terminated SAMs provided an extremely hydrophobic substrate with WCA ∼ 101°, while the –NH₂ ²⁷ and –epoxy²⁸ terminated SAMs formed moderately wettable substrates with WCA ∼ 53° and 51°, respectively. By contrast, SAMs terminated with –COOH groups exhibited highly wettable surfaces with WCA ∼ 27°, which is in good agreement with the published values.²⁹ Furthermore, the WCA of epoxy-terminated surfaces decreased from 51° to 31° after modification with HCl, indicating the successful conversion of epoxy to OH groups according to previous protocols.²⁴ In general, the WCA values of the –CH₃, –NH₂, –epoxy, –OH and –COOH terminated SAMs corresponded well with previous studies.^22,30

Fig. 2 Physical characterization of the SAM surfaces. (A) Static water contact angle (WCA) measurement of CH₃, NH₂, epoxy, OH and COOH SAMs. The results are the means ± SD of two independent experiments, each performed in triplicate. (B) Zeta potential measurements of the SAMs with different terminating groups compared with bare glass. The zeta potentials were measured in the pH range of 3–10.

Fig. 2B shows the zeta potentials versus the pH value of a standard electrolyte solution (1 mM KCl) for bare glass and different SAM surfaces. The bare glass, OH, and COOH SAMs exhibited negative surface potentials throughout the whole measured pH range. By contrast, CH₃ and NH₂ SAMs were found to have a point of zero charge (PZC) at pH 3.4 and 4.8, respectively. The highest potential at pH 7.4 was found on the –NH₂ terminated SAM surface, which is attributed to the partial protonation of amino groups in the low pH region.³¹ The neutral –OH and –CH₃ terminated surfaces displayed negative zeta potentials at pH 7.4 because of a preferential anion adsorption on the non-charged surfaces.³² The curve for the –COOH terminated surface possessed a negative surface potential from pH 3–10 and a plateau in the basic pH region arising from the presence of dissociable acidic groups on this surface.³¹ Additionally, the zeta potential values of all SAM-modified surfaces were higher compared to bare glass at pH 7.4, which indicates the successful immobilization of the organosilanes and the formation of SAMs on the bare glass substrate.

Physical characterization studies confirmed that the SAMs possessed different wetting and surface charge properties, which made them applicable as model surfaces to study the effect of surface properties on fibroblast and macrophage reactions as described in the following sections.

3.2 HF adhesion, spreading and proliferation

As fibroblasts are one of the key effector cells during wound healing and fibrotic processes and also the initial inflammatory phase after a biomaterial is implanted, the initial adhesion, spreading as well as the following proliferation and outgrowth of fibroblasts crucial for the potential wound healing and potential development of fibrosis were studied here.

We firstly evaluated the HF morphology during initial cell adhesion on different SAMs by analysis of the focal adhesion complexes using vinculin (green) and actin (red) staining as shown in Fig. 3A. After 4 h incubation in the presence of serum, HFs adhering to NH₂ and COOH SAMs were well spread and contained short focal adhesion plaques at the cell periphery. Strong actin stress fibres were also formed. Additionally, a colocalization of actin and vinculin was observed, illustrated by the yellow colour. By contrast, adhesion and spreading of HFs on CH₃ and OH SAMs were largely suppressed. No focal adhesions or actin stress fibres were observed on these surfaces. Considering the presence of 10% serum in the medium, it can be anticipated that NH₂ and COOH SAMs are favourable to bind adhesion-promoting proteins such as fibronectin and vitronectin from serum,^33,34 which promoted fibroblast adhesion and spreading. By contrast, the OH surface was reported to inhibit protein adsorption, and thus impair cell adhesion.²⁹ On the other hand, although CH₃ surfaces can bind larger quantities of proteins, the conformation of the adsorbed proteins can easily change due to the hydrophobic interactions between the surfaces and the hydrophobic domains of proteins.³⁵ As a result, fibroblast adhesion and spreading were also supressed on this surface.

Fig. 3 Adhesion, spreading and proliferation of human fibroblasts (HFs) on CH₃, NH₂, OH and COOH SAMs. Representative CLSM images (A) of HFs after 4 h incubation with immune fluorescence staining of vinculin (green), actin (red) and nuclei (blue). Bar = 20 μm. Quantitative cell adhesion number (B) and spreading area (C) of HF adhesion on different SAMs after 4 h incubation in serum-containing medium. Data represent mean ± SD, n = 4, *p < 0.05. (D) HF proliferation after 24 h and 72 h culture on CH₃, NH₂, OH and COOH SAMs in serum-containing medium. Data represent mean ± SD, n = 4, *p < 0.05.

Apart from the HF morphology studies, a quantitative estimation of cell adhesion (cell count) and spreading (cell area) were conducted on different SAM surfaces after 4 h incubation in the presence of serum (Fig. 3B and C). The results show that there were significantly more cells adhering on NH₂ and COOH SAMs than on OH and CH₃ SAMs. The analysis of cell spreading data also revealed significantly a higher spreading area of HFs on NH₂ and COOH surfaces in comparison to OH and CH₃ surfaces. These findings are well in line with previous observations showing that cells prefer to adhere and spread more on moderately hydrophilic and hydrophilic/anionic surfaces compared to hydrophobic and non-ionic hydrophilic surfaces.^22,29,36

The proliferation of HFs on different SAM surfaces was studied after the incubation periods of 24 and 72 h. Fig. 3D shows the metabolic activity of HFs measured from the QBlue assay that represents an indication of the quantity of metabolically active cells on the different SAMs. The obtained data illustrate that cell growth was higher on NH₂ and COOH surfaces compared to OH and CH₃ surfaces after both 24 and 72 h, which is in line with the observed HF adhesion and spreading results. In general, the surface properties dictate the protein adsorption from serum-containing medium, which affects the ligand binding to adhesion receptors of cells like integrins.^33,37 Such a ligand–receptor binding mediates cell adhesion, signal transduction followed by cytoskeleton reorganization, protein synthesis and cell proliferation.³⁸

3.3 HF outgrowth studies

HF migration and growth are very important during the processes associated with wound healing and fibrosis and strongly affected by cytokine release from macrophages.³⁹ Therefore, we further evaluated the effects of surfaces on HF growth by a specific outgrowth study. Fig. 4A shows the crystal violet staining images of HF outgrowth after 3 days in the presence (co-culture) or absence (HF mono) of macrophages. The macrophages derived from the THP-1 monocyte cell line after induction with PMA used here represent an early phenotype positive for CD 68, a common macrophage marker, but not already committed to M1 or M2 phenotypes as found out by immunofluorescence studies and flow cytometry (see the ESI† for more details). During these experiments, the macrophages were seeded at the same time as that for HFs, but in the external compartment of the chamber, which separated them by a 500 μm gap after lifting up the chamber. Fig. 4B depicts the quantitative outgrowth distances on different SAMs after subtracting the original diameter of the HF region. It was observed that the CH₃ SAM induced the smallest outgrowth distance in both HF mono- and co-cultures. This can be explained by the impaired HF adhesion and proliferation due to the high hydrophobicity of this surface, leading to conformational changes of adsorbed adhesive proteins and impaired cell responses.^29,40 Contradictorily, the NH₂ SAM with the highest cell adhesion and proliferation in single cell cultures shown before, generated only an intermediate outgrowth, less than OH and COOH SAMs. This might be due to the strong interactions between the HFs and the NH₂ SAM, illustrated by the multiple, extended focal adhesion complexes (see Fig. 3A), which caused an impaired ability of cell migration,⁴¹ although with higher cell proliferation. The migration of HFs on NH₂ and COOH surfaces has been reported by Faucheux et al. previously, also revealing that HFs on the FN-coated COOH surface migrated faster than that on the NH₂ surface.²⁹ Additionally, it is notable that the OH SAM caused a comparable outgrowth distance like that of the COOH SAM. This might be probably due to the weak adhesion of HFs on the OH SAM, resulting in a higher migration ability.^30,39 Besides, the OH SAM provoked a moderate proliferation of HFs as observed in the proliferation studies shown here and by others.⁴² However, the most important finding of these experiments was that the outgrowth distance was significantly increased in co-cultures compared to fibroblast mono-cultures on all SAMs. During these experiments it was also observed that the effect of surface chemistry on HF outgrowth was the same as observed in the HF monoculture experiments. The increased outgrowth distance is obviously attributed to the presence of the surrounding macrophages. It was reported previously that cytokines such as IL-1β and IL-6 secreted by macrophages promote HF migration and proliferation,^43,44 which supports the findings made here.

Fig. 4 Representative crystal violet staining images (A) and quantitative outgrowth distances (B) of HF outgrowth in HF mono-cultures (HF mono, white bars) and fibroblast/macrophage co-cultures (Co-culture, black bars) on CH₃, NH₂, OH and COOH SAMs. The cells were seeded and cultured for 24 h before removal of the fence chambers. The fibroblasts were then left to grow for 3 days in both mono- and co-cultures. Afterwards, the cells were fixed and stained by using 0.5% (w/v) crystal violet in methanol. Bar = 2000 μm. Data represent mean ± SD, n = 6, *p < 0.05.

3.4 Macrophage fusion studies

In addition to the analysis of the HF outgrowth distance, the macrophage fusion on different SAM surfaces was also evaluated in the macrophage regions of the co-culture system (Fig. 5). It can be seen that large quantities of foreign body giant cells (FBGCs) were formed on the CH₃ SAM, illustrated by the large size of cells shown by crystal violet staining (red arrows). By contrast, fewer FBGCs with lower size were found on NH₂, OH and COOH SAMs. Macrophages fuse to form FBGCs in order to increase their phagocytosis ability when facing foreign materials of a larger size than themselves.⁴⁵ It was also found that the fusion extent represents a measure of biocompatibility of a given material.⁴⁶ Hence, the difference in FBGC formation classifies the CH₃ SAM as more pro-inflammatory than the other SAMs.

Fig. 5 Foreign body giant cell (FBGC) formation on CH₃, NH₂, OH and COOH SAMs in the macrophage region of co-cultures after 3 days of HF outgrowth. Afterwards, the cells were fixed and stained by using 0.5% (w/v) crystal violet in methanol. The red arrows show the FBGCs. Bar = 200 μm.

3.5 Pro- and anti-inflammatory cytokine production

The soluble signals including pro- and anti-inflammatory cytokines are important mediators during inflammatory and wound healing/fibrotic processes. Macrophages have a high degree of functional plasticity, which can act as both M1 inflammatory and M2 wound healing phenotype in response to different stimuli.²¹ M1 macrophages display more pro-inflammatory function by secretion of cytokines such as IL-1β, IL-6 and TNF-α, while M2 macrophages have anti-inflammatory function and promote wound healing by secretion of cytokines like IL-10 and TGF-β1.⁴⁷ Besides, fibroblasts not only can modulate the cytokine production by macrophages but also can generate both pro- and anti-inflammatory cytokines including IL-1β, IL-6, TGF-β and IL-10.^48,49 The balance between the pro- and anti-inflammatory mediators is critical for orchestrating the outcome of the inflammatory and wound healing/fibrotic responses. Therefore, a premature type of macrophage still not committed to the M1 or M2 phenotype was used here as generated by the use of PMA, only. We tested here both the pro-inflammatory cytokine IL-6 and anti-inflammatory cytokine IL-10 production on different SAMs.

Fig. 6A shows that there was no significant difference of IL-6 release among the various SAMs in HF mono-cultures. This might be related to the limited HF population due to the low seeding density in the internal channel of the chambers (0.15 mL of HFs at 2.5 × 10⁴ cells per mL) and limited proliferation of HFs in the absence of serum. By contrast, IL-6 released from cells adherent to different SAMs in co-cultures showed a surface-dependent behavior. It was found that the cells adhering on the CH₃ SAM produced significantly higher levels of IL-6 in comparison to the other SAMs. This result indicates that IL-6 production is provoked by the hydrophobic nature of the CH₃ SAM, which is in consistency with the previous findings that hydrophobic surfaces induced the highest levels of pro-inflammatory cytokines such as IL-1β, IL-6 and TNF-α.^15,18,50 On the other hand, IL-10 production showed an opposite surface-dependent effect compared to IL-6 in co-cultures (Fig. 6B). The results show that the cells cultured on the COOH SAM produced significantly higher amounts of IL-10 than on CH₃, NH₂ and OH SAMs. This is also in line with other reports documenting that the hydrophilic/anionic surfaces reduce pro-inflammatory responses and promote the wound healing processes.⁴ Additionally, the production of both IL-6 and IL-10 increased significantly in fibroblast/macrophage co-cultures compared to HF mono-cultures on all SAMs. This indicates that macrophages contribute largely to both pro- and anti- inflammatory signal production. Moreover, a much higher amount of IL-10 compared to IL-6 was observed for the same samples. This might imply a reduced inflammatory response and an increased wound healing response under the measuring conditions.

Fig. 6 IL-6 (A) and IL-10 (B) release in human fibroblast mono-cultures (HF mono, white bars) and fibroblast/macrophage co-cultures (Co-culture, black bars) on CH₃, NH₂, OH and COOH SAMs after 3 days of incubation with 10 ng mL⁻¹ TGF-β1. Data represent mean ± SD, n = 6, *p < 0.05.

3.6 ED-A FN and α-SMA expression

Myofibroblast phenotype transformation is an essential step during wound healing and fibrotic processes, which is normally accompanied by abundant ED-A FN and α-SMA expression in these cells.⁸ TGF-β1 can stimulate fibroblast differentiation into myofibroblasts and is known as an important contributor to fibrosis both in vitro and in vivo.⁵¹ Hence, the levels of ED-A FN and α-SMA expression on different SAMs were studied in the presence of TGF-β1 as also done in other studies^52,53 to mimic the in vivo situation to some extent.

Fig. 7A presents the ED-A FN expression on different SAM surfaces in both HF mono- and co-cultures. A comparison of the staining revealed obvious surface chemistry-dependent differences with the NH₂ SAM expressing the highest level of ED-A FN in both mono- and co-cultures. By contrast, on the CH₃ surface considerably lower levels of ED-A FN were detected, while COOH and OH surfaces exhibited intermediated expression compared to NH₂ and CH₃ SAMs. Overall, the ED-A FN expression results corresponded well to the HF adhesion, proliferation and outgrowth experiments and showed an order of NH₂ > COOH ∼ OH > CH₃. Besides, the markedly increased expression levels of ED-A FN in all co-cultures revealed that macrophages amplified the ED-A FN matrix synthesis by HFs. Indeed, a large body of research studies have confirmed that macrophage contributes to wound healing and fibrosis by promoting fibroblast activation, proliferation and differentiation into myofibroblasts.^43,44

Fig. 7 Immunofluorescence staining of (A) ED-A FN (green), (B) α-SMA (red) and nuclei (blue) after 3 day incubation with 10 ng mL⁻¹ TGF-β in HF mono-cultures (HF mono) and fibroblast/macrophage co-cultures (Co-culture) on CH₃, NH₂, OH and COOH SAMs. Bar = 100 μm.

Furthermore, the expression of α-SMA, a key feature of their transformation into myofibroblast during wound healing and fibrosis, was also analysed on the different SAMs (Fig. 7B). Notably more α-SMA positively stained HFs (α-SMA-positive HFs) were observed on the NH₂, OH and COOH surfaces in comparison to the CH₃ surface. Moreover, the α-SMA-positive HFs on all co-culture surfaces were much more elongated than those in the corresponding mono-cultures. Fig. 8A depicts the quantitative evaluation of the cell aspect ratio of the α-SMA-positive HFs on the different SAMs. First, no significant difference of the aspect ratio was observed among NH₂, OH and COOH SAMs in both HF mono- and co-cultures. By contrast, the cells on the CH₃ SAM exhibited a significantly lower aspect ratio than NH₂, OH and COOH SAMs in co-cultures. The elongation extent of the cells represents their contractile capability, which plays important roles in wound healing and fibrotic processes. Therefore, the in vitro data obtained here may suggest that the hydrophobic CH₃ SAM had lower potential of promoting wound healing and fibrosis. Furthermore, the aspect ratio of the α-SMA-positive HFs in all co-cultures significantly increased compared to the corresponding mono-cultures. This can be attributed to the pro-wound healing and pro-fibrotic activities of the macrophages in co-cultures.²¹

Fig. 8 Surface chemistry-dependent differences in α-SMA expression after 3 day incubation in the presence of 10 ng mL⁻¹ TGF-β. Aspect ratio (A) and percentage (B) of α-SMA positively stained cells on CH₃, NH₂, OH and COOH SAMs in both HF mono-cultures (HF mono) and fibroblast/macrophage co-cultures (Co-culture). Data represent mean ± SD, n ≥ 20, *p < 0.05.

The determination of the percentage of α-SMA-positive cells shown in Fig. 8B reveals that the OH surface caused the highest percentage of α-SMA-positive cells, while the COOH surface induced the lowest percentage, following the order OH > NH₂ ∼ CH₃ > COOH. Here, the higher adhesion and outgrowth of the fibroblasts on the hydrophilic/anionic COOH SAM resulted in the lowest percentage of α-SMA-positive cells, indicating a lower potential of inducing wound healing/fibrosis. This might also be related to the highest IL-10 cytokine production on this surface, which was reported to have both anti-inflammatory and anti-fibrotic effects.⁵⁴ By contrast, the hydrophilic OH surface provoked a lower extent of pro-inflammatory response in terms of FBGC formation and pro-inflammatory cytokine IL-6 production, but more myofibroblasts than the other surfaces. This indicates that OH-functionalized surfaces have a higher potential of inducing wound healing/fibrosis. These findings are generally consistent with previous work using polypropylene microspheres with different functional groups as also used here to study the extent of fibrotic tissue reactions elicited by the different functionalities, and their data show that the OH surface triggers the strongest capsule formation while the COOH surface promotes the least capsule formation.⁵⁵ It was suggested that the complement activation can affect the recruitment of inflammatory cells and the following fibrotic tissue reactions.⁵⁶ The OH groups are reported to bind with the C3b component and activate the alternative complement pathway to a much greater extent than COOH groups,⁵⁷ which may partly explain the findings here.

A summary of the measured parameters involved in the inflammatory and wound healing/fibrotic responses on CH₃, NH₂, OH and COOH SAMs is shown in Table 1. Firstly, we observed that the hydrophilic/anionic COOH SAM caused lower inflammatory reactions in comparison to the CH₃ SAM. Furthermore, although the highest fibroblast adhesion, proliferation, outgrowth, as well as strong expression of ED-A FN and α-SMA were found on the COOH SAM, the percentage of the α-SMA positive cells was the lowest, indicating a lower extent of inducing wound healing/fibrosis. Therefore, it can be concluded that the hydrophilic/anionic COOH SAM has both anti-inflammatory and anti-fibrotic potential. Secondly, it was observed that the hydrophobic CH₃ SAM with the highest potential of inducing inflammatory responses illustrated by the highest level of FBGC formation and IL-6 production revealed a low potential of inducing wound healing/fibrotic responses through the observation of fibroblast activations. Here, the low wound healing/fibrotic potential of the CH₃ SAM seems to be contradictory to some in vivo findings that hydrophobic surfaces induced thicker fibrous capsules.⁵⁸ However, we wanted to clarify that we have seeded fibroblasts at a very low density here, which is different from the in vivo situations that a full coverage of fibroblasts on the surfaces can be achieved. Thus we suggest that the impaired fibroblast adhesion, spreading, proliferation and outgrowth on the hydrophobic CH₃ surface here should be responsible for the reduced potential of inducing the wound healing/fibrotic reactions. On the other hand, the hydrophilic OH SAM triggered a low level of inflammatory responses but indicated a higher potential of inducing wound healing/fibrotic reactions, which is consistent with the poor biocompatibility of this surface reported also by others.⁵⁵ Furthermore, the results also reveal that the inflammatory and wound healing/fibrotic responses might not always be consistent.⁵⁹ Further in-depth inflammation–wound healing/fibrosis link studies will help to design biomaterials with appropriate fibrotic tissue responses.

Table 1 Surface chemistry differentially affects the parameters of the inflammatory (dark grey color) and wound healing/fibrotic responses (light grey color) on CH₃, NH₂, OH and COOH SAMs (+ weak, ++ strong, +++ very strong)

---

4. Conclusion

In the present study, we have established an in vitro fibroblast/macrophage co-culture system to study the host responses ranging from inflammatory to wound healing/fibrotic responses of a series of SAMs with different terminating groups as model materials. It was found that the hydrophilic/anionic COOH SAM caused low levels of both inflammatory and wound healing/fibrotic responses while the hydrophobic CH₃ surfaces possess the highest potential of inducing inflammatory responses but induced a low level of wound healing/fibrotic responses. By contrast, the OH SAM which evoked a lower extent of pro-inflammatory responses, but revealed a higher potential of inducing wound healing/fibrosis. These observations can provide useful clues for the design of biomaterials by triggering appropriate host responses for different biomedical and clinical applications. More important however is the obvious advantage of the presented in vitro fibroblast/macrophage co-culture system, which can generate a spatially controlled interaction between the macrophages and the fibroblasts to study the host responses of biomaterials in vitro. The results showed that macrophages can promote wound healing/fibrotic responses by up-regulation of fibroblast outgrowth, cytokine production as well as ED-A FN and α-SMA expression in co-cultures. Lastly but importantly, one can also study here the in vitro effects of additional factors like cytokines and other bioactive molecules on the interplay between the two cell types using our co-culture system.

Disclosure

No conflicts of interests are declared.

Acknowledgements

G. Zhou thanks the China Scholarship Council (CSC) and the Prorectorate for Structur and strategic development of Martin Luther University Halle-Wittenberg together with the Federal Ministry of Education and Research (BMBF) for offering the scholarships to work in Germany. Mrs M. Porobin is gratefully acknowledged for the assistance of zeta potential measurement. We are also very grateful to Dr A. Navarrete Santos from the University Medical Centre, Core Facility Centre of Martin Luther University Halle-Wittenberg for the kind support to perform the flow cytometry measurements for cell analysis and cell sorting.

Notes and references

1. J. M. Anderson, A. Rodriguez and D. T. Chang, Semin. Immunol., 2008, 20, 86–100.
2. T. Glaros, M. Larsen and L. W. Li, Front. Biosci., 2009, 14, 3988–3993.
3. S. Franz, S. Rammelt, D. Scharnweber and J. C. Simon, Biomaterials, 2011, 32, 6692–6709.
4. W. G. Brodbeck, Y. Nakayama, T. Matsuda, E. Colton, N. P. Ziats and J. M. Anderson, Cytokine, 2002, 18, 311–319.
5. G. Zhou, M. S. Niepel, S. Saretia and T. Groth, J. Biomed. Mater. Res., Part A, 2016, 104, 493–502.
6. G. Zhou, H. Al-Khoury and T. Groth, Int. J. Artif. Organs, 2016, 39, 37–44.
7. T. A. Wynn, J. Pathol., 2008, 214, 199–210.
8. G. Serini, M. L. Bochaton-Piallat, P. Ropraz, A. Geinoz, L. Borsi, L. Zardi and G. Gabbiani, J. Cell Biol., 1998, 142, 873–881.
9. S. E. Mutsaers, J. E. Bishop, G. McGrouther and G. J. Laurent, Int. J. Biochem. Cell Biol., 1997, 29, 5–17.
10. Z. Yuan, J. Zhao, W. Zhu, Z. Yang, B. Li, H. Yang, Q. Zheng and W. Cui, Biomater. Sci., 2014, 2, 502–511.
11. C. Helary, A. Abed, G. Mosser, L. Louedec, D. Letourneur, T. Coradin, M. M. Giraud-Guille and A. Meddahi-Pellé, Biomater. Sci., 2015, 3, 373–382.
12. W. J. Hu, J. W. Eaton and L. P. Tang, Blood, 2001, 98, 1231–1238.
13. A. P. Marques, R. L. Reis and J. A. Hunt, Macromol. Biosci., 2005, 5, 775–785.
14. K. G. Battiston, R. S. Labow and J. P. Santerre, Biomaterials, 2012, 33, 8316–8328.
15. G. Zhou, H. Loppnow and T. Groth, Acta Biomater., 2015, 26, 54–63.
16. M. R. MacEwan, W. G. Brodbeck, T. Matsuda and J. M. Anderson, J. Biomed. Mater. Res., Part A, 2005, 74A, 285–293.
17. J. N. Barbosa, P. Madureira, M. A. Barbosa and A. P. Aguas, J. Biomed. Mater. Res., Part A, 2006, 76A, 737–743.
18. F. F. R. Damanik, T. C. Rothuizen, C. van Blitterswijk, J. I. Rotmans and L. Moroni, Sci. Rep., 2014, 4, 6325.
19. E. K. Park, H. S. Jung, H. I. Yang, M. C. Yoo, C. Kim and K. S. Kim, Inflammation Res., 2007, 56, 45–50.
20. A. Schildberger, E. Rossmanith, T. Eichhorn, K. Strassl and V. Weber, Mediators Inflammation, 2013, 2013, 697972.
21. P. J. Murray and T. A. Wynn, Nat. Rev. Immunol., 2011, 11, 723–737.
22. G. Altankov, F. Grinnell and T. Groth, J. Biomed. Mater. Res., 1996, 30, 385–391.
23. V. Llopis-Hernandez, P. Rico, J. Ballester-Beltran, D. Moratal and M. Salmeron-Sanchez, PLoS One, 2011, 6, e19610.
24. G. K. Toworfe, R. J. Composto, I. M. Shapiro and P. Ducheyne, Biomaterials, 2006, 27, 631–642.
25. J.-W. Tjiu, J.-S. Chen, C.-T. Shun, S.-J. Lin, Y.-H. Liao, C.-Y. Chu, T.-F. Tsai, H.-C. Chiu, Y.-S. Dai, H. Inoue, P.-C. Yang, M.-L. Kuo and S.-H. Jee, J. Invest. Dermatol., 2009, 129, 1016–1025.
26. E. G. Fischer, A. Stingl and C. J. Kirkpatrick, J. Immunol. Methods, 1990, 128, 235–239.
27. V. Katzur, M. Eichler, E. Deigele, C. Stage, P. Karageorgiev, J. Geis-Gerstorfer, G. Schmalz, S. Ruhl, F. Rupp and R. Muller, J. Colloid Interface Sci., 2012, 366, 179–190.
28. I. Luzinov, D. Julthongpiput, A. Liebmann-Vinson, T. Cregger, M. D. Foster and V. V. Tsukruk, Langmuir, 2000, 16, 504–516.
29. N. Faucheux, R. Schweiss, K. Lutzow, C. Werner and T. Groth, Biomaterials, 2004, 25, 2721–2730.
30. J. Benesch, J. F. Mano and R. L. Reis, Acta Biomater., 2010, 6, 3499–3505.
31. J. J. Shyue and M. R. De Guire, Langmuir, 2004, 20, 8693–8698.
32. C. Werner, H. Korber, R. Zimmermann, S. Dukhin and H. J. Jacobasch, J. Colloid Interface Sci., 1998, 208, 329–346.
33. B. G. Keselowsky, D. M. Collard and A. J. Garcia, Biomaterials, 2004, 25, 5947–5954.
34. F. Grinnell and M. K. Feld, J. Biol. Chem., 1982, 257, 4888–4893.
35. B. G. Keselowsky, D. M. Collard and A. J. Garcia, J. Biomed. Mater. Res., Part A, 2003, 66A, 247–259.
36. Y. Arima and H. Iwata, Biomaterials, 2007, 28, 3074–3082.
37. N. M. Alves, I. Pashkuleva, R. L. Reis and J. F. Mano, Small, 2010, 6, 2208–2220.
38. S. Miyamoto, H. Teramoto, O. A. Coso, J. S. Gutkind, P. D. Burbelo, S. K. Akiyama and K. M. Yamada, J. Cell Biol., 1995, 131, 791–805.
39. J. J. Zhao, L. Hu, N. Y. Gong, Q. M. Tang, L. Y. Du and L. L. Chen, Tissue Eng., Part A, 2015, 21, 982–991.
40. G. Hummer, S. Garde, A. E. Garcia, A. Pohorille and L. R. Pratt, Proc. Natl. Acad. Sci. U. S. A., 1996, 93, 8951–8955.
41. A. Huttenlocher and A. R. Horwitz, Cold Spring Harbor Perspect. Biol., 2011, 3, a005074.
42. G. Altankov, K. Richau and T. Groth, Materialwiss. Werkstofftech., 2003, 34, 1120–1128.
43. F. Y. Chow, D. J. Nikolic-Paterson, R. C. Atkins and G. H. Tesch, Nephrol., Dial., Transplant., 2004, 19, 2987–2996.
44. S. Van Linthout, K. Miteva and C. Tschope, Cardiovasc. Res., 2014, 102, 258–269.
45. S. MacLauchlan, E. A. Skokos, N. Meznarich, D. H. Zhu, S. Raoof, J. M. Shipley, R. M. Senior, P. Bornstein and T. R. Kyriakides, J. Leukocyte Biol., 2009, 85, 617–626.
46. J. M. Anderson and K. M. Miller, Biomaterials, 1984, 5, 5–10.
47. A. Mantovani, S. Sozzani, M. Locati, P. Allavena and A. Sica, Trends Immunol., 2002, 23, 549–555.
48. H. Pan, H. L. Jiang, S. Kantharia and W. L. Chen, Biomed. Mater., 2011, 6, 065002.
49. K. P. Sundararaj, D. J. Samuvel, Y. Li, J. J. Sanders, M. F. Lopes-Virella and Y. Huang, J. Biol. Chem., 2009, 284, 13714–13724.
50. S. Hamlet, M. Alfarsi, R. George and S. Ivanovski, Clin. Oral Implants Res., 2012, 23, 584–590.
51. M. B. Vaughan, E. W. Howard and J. J. Tomasek, Exp. Cell Res., 2000, 257, 180–189.
52. A. Watarai, L. Schirmer, S. Thoenes, U. Freudenberg, C. Werner, J. C. Simon and U. Anderegg, Acta Biomater., 2015, 25, 65–75.
53. A. van der Smissen, S. Samsonov, V. Hintze, D. Scharnweber, S. Moeller, M. Schnabelrauch, M. T. Pisabarro and U. Anderegg, Acta Biomater., 2013, 9, 7775–7786.
54. J.-H. Shi, H. Guan, S. Shi, W.-X. Cai, X.-Z. Bai, X.-L. Hu, X.-B. Fang, J.-Q. Liu, K. Tao, X.-X. Zhu, C.-W. Tang and D.-H. Hu, Arch. Dermatol. Res., 2013, 305, 341–352.
55. A. Nair, L. Zou, D. Bhattacharyya, R. B. Timmons and L. Tang, Langmuir, 2008, 24, 2015–2024.
56. L. Liu and H. Elwing, J. Biomed. Mater. Res., 1994, 28, 767–773.
57. J. N. Barbosa, M. A. Barbosa and A. P. Aguas, Biomaterials, 2004, 25, 2557–2563.
58. J. N. Barbosa, P. Madureira, M. A. Barbosa and A. P. Águas, J. Biomed. Mater. Res., Part A, 2006, 76A, 737–743.
59. L. P. Tang, Y. L. Wu and R. B. Timmons, J. Biomed. Mater. Res., 1998, 42, 156–163.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6bm00247a

---