Cells may feel a hard substrate even on a grafted layer of soft hydrogel

Shuhao Wang ab, Fei Zan ab, Yu Ke *c and Gang Wu *abd
aNational Engineering Research Center for Tissue Restoration and Reconstruction, Guangzhou, 510006, China
bSchool of Materials Science and Engineering, South China University of Technology, Guangzhou, 510641, China. E-mail: imwugang@scut.edu.cn
cDepartment of Biomedical Engineering, Jinan University, Guangzhou, 510632, China. E-mail: lisa6863@163.com
dGuangdong Province Key Laboratory of Biomedical Engineering, Guangzhou, 510641, China

Received 24th November 2017 , Accepted 15th February 2018

First published on 23rd February 2018

Introducing or grafting molecules onto biomaterial surfaces to regulate cell destination via biophysical cues is one of the important steps for biomaterial design in tissue engineering. Understanding how cells feel the substrate makes it easier to learn the mechanism behind cell–material interaction. In this study, on a glass substrate, we constructed poly-phenoxyethyl methacrylate (PHEMA) brushes having different lengths via a surface-induced atom transfer radical polymerization (SI-ATRP) method. FTIR-ATR and XPS tests of the formed polymer brushes indicate that these brushes have characteristic chemical structures of PHEMA; the polymer brush length revealed by the AFM tests increases linearly with reaction time. Cell lines of BMSCs, ATDC5, and human chondrocytes (HC) were cultured on these substrates to evaluate proliferation, adhesion, and differentiation. Our results demonstrated that the cells cultured on the substrates with short PHEMA brushes developed a spread morphology and organized actin fibers as compared to the cells cultured on those with long brushes. Different cell lines showed different responses depending on the PHEMA brush length. Cells cultured on long PHEMA brushes displayed a more rounded shape, higher gene expression of FAK and integrin, and lower gene expression of NCAM and N-cadherin as compared to those, especially ATDC5 cells, cultured on short PHEMA brushes. On PHEMA brushes with a long length, the cell lines express higher cartilage-specific genes including Sox9 and Col2 and GAG in ECM. The results suggest that polymer brushes having different lengths may interfere with the behavior of the cells cultured on them.


For tissue development, stem cells experience several stages to differentiate and develop into terminal cells in a tissue.1–4 The physical and chemical properties of the extracellular matrix (ECM) secreted by the cells dynamically change along with tissue regeneration; this provides a suitable environment for tissue development.5–10 Development of promising porous scaffolds, mimicking ECM, with better biological, biochemical, and biophysical properties that mimic ECM properties has been a challenge for tissue repair and regeneration.

To achieve this aim, a number of modification methods, including surface grafting or deposition,11–14 bulk blocking or tethering bioactive molecules,15,16 surface adsorption or embedding of medicine/growth factors,17–20 directly using biomolecule assemblies,21–23 and surface roughness modulation and morphological patterning,24–26 have been implemented to introduce positive cell–material interaction towards inducing the abovementioned cues; moreover, upon combining both the biophysical and biomolecule assembly cues, this result can be achieved. Mao et al. have first confirmed that a cooperative combination of ordered ridge/groove nanotopography and growth factor signal peptide cues can significantly promote the osteoblastic differentiation of iPSCs in one system.27 In another study reported by them, they prepared pure silk protein nanoridges by an ice-templating approach and found that decoration of the biomaterial surfaces with protein nanoridges could enhance bone tissue formation at the bone repair sites.28

Porous scaffolds fabricated by the synthesized degradable polymers are the major application types used in the tissue-engineering research. Grafting a molecule with biological activity onto the scaffold surface has been widely applied among these methods because it is easy to be implemented on various types of biomaterials without distinct alteration of the bulk properties. These bioactive molecules and their interacting cellular receptors mediate the bidirectional reaction between the extracellular and intracellular compartments to influence cell growth, survival, differentiation, and morphogenesis.29,30

Recently, a number of studies have demonstrated that modulation of the components of the grafting layer, even in a very subtle way, may impact the cell behavior and thus has been used to improve the biomaterial behavior; for example, a gradient scaffold that dynamically simulates ECM compositions has been reported, and the early chondrocyte ECM is more favorable for promoting cartilage differentiation.2 Besides biochemical signals, biophysical signals related to substrate stiffness may be transduced into the cells through the ECM–integrin–cytoskeleton axis system, resulting in a series of biological functions.31

These biomimetic molecules or structures conjugated onto a solid substrate surface demonstrate that a subtle deviation may possibly impact cell behavior in a very sensitive way. However, few previously reported studies reveal whether the signals from the substrate surfaces can be shielded by the grafting layer, the thickness variation of the introduced biomimetic layer impacts the behavior of the adhered cells, and these influences on the cells at different differentiation stages trigger same response or not. It is important to understand the dynamically changed effects in the process for designing scaffolds in the future.

In this study, we polymerized hydroxyethyl methacrylate (HEMA) on a glass substrate via the surface-initiated atomic transfer radical polymerization (SI-ATRP) method,32–34 allowing us to control the thickness and composition of the polymer brush accurately. Cell lines of undifferentiated mesenchymal stem cells (MSCs), semi-differentiated precursor chondrocyte of ATDC5 cells, and completely differentiated mature cells of human chondrocytes (HC) have been used as model cells to investigate the impacts of the grafting thickness on the cell behavior at specific evolution stages. This provided in vitro models for investigating cell response on different substrates during cartilage development, which would be very important for the design and preparation of tissue engineering scaffolds.

Materials and methods


2-Hydroxyethyl methacrylate (HEMA, 99%), 3-aminopropyltriethoxysilane (APTES, 99%), copper(I) chloride (99.995%), N,N,N′,N′′,N′′-pentamethyldiethylenetriamine (PMDET, 99%), α-bromoisobutyryl bromide (BIBB, 98%), and TRIzol reagent were obtained from Sigma-Aldrich (Shanghai, China). Sodium L-ascorbate (99%, J&K), triethylamine (99.5%, Aladdin), dichloromethane (99.9%), and ethanol anhydrous (99.5%) were purchased from Macklin (Shanghai, China). Phosphate buffered saline (PBS), 4′,6-diamidino-2-phenylindole (DAPI), All-in-One First-Strand cDNA synthesis kit, and All-in-One qPCR Mix (GeneCopoeia, USA) were bought from Qiyun biotech. Ltd. The CCK-8 assay kit was bought from Dojindo (Shanghai, China). The phalloidin conjugate solution was bought from Zhonghao Biotech. Co. Ltd (Beijing, China). The MSC growth medium (MSCGM) composed of low glucose Dulbecco's modified Eagle medium (DMEM), fetal bovine serum (FBS), 100 units per ml penicillin, and 100 μg ml−1 streptomycin were bought from thermo-fisher (Gibco, Shanghai, China). BMSC, ATDC5, and HC cells were provided by the National Engineering Research Center for Tissue Restoration and Reconstruction.

Preparation of the PHEMA brushes

PHEMA brushes were grafted on glass coverslips by the SI-ATRP method.34,35 Briefly, the glass coverslips (10 × 10 mm) were cleaned and hydroxylated in a piranha solution (98% H2SO4 and 30% H2O2, 7[thin space (1/6-em)]:[thin space (1/6-em)]3 v/v) at 90 °C for 30 min. Then, the mixture was rinsed copiously with deionized water and dried under nitrogen immediately. The glass coverslips were subsequently immersed in 0.4% (v/v) aqueous APTES for 20 min, rinsed with water, dried in a nitrogen stream, and cured at 120 °C for 3 h. The aminated coverslips were immersed in anhydrous DCM containing 2% (v/v) TEA and 0.04% (v/v) BIBB for over 12 h to introduce the brominated ATRP initiator, followed by successive rinsing with DCM, ethanol, and water and then drying with nitrogen.

The PHEMA brushes were grown from the ATRP-initiated layer on the glass coverslips via applying a catalytic system comprising CuCl (22 mg, 0.22 mmol), PMDETA (200 μl, 0.96 mmol), and sodium L-ascorbate (45 mg, 0.44 mmol) in 31 ml aqueous HEMA with a ratio of monomer to water of 1/30. The HEMA and sodium L-ascorbate mixed solution was deoxygenated by nitrogen bubbling for 30 min followed by the addition of CuCl. The mixture was subsequently transferred to glass bottles containing the ATRP initiator-functionalized glass coverslips. Polymerization was proceeded for different times at 50 °C and then stopped by overflowing the reaction mixture with pure water. The PHEMA brush-decorated coverslips were rinsed with ethanol and dried with nitrogen.

Surface characterization of PHEMA brushes

Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy. ATR-FTIR spectra were obtained using the Bruker Vector 33 spectrometer in the scan range of 600–4000 cm−1.
Water contact angle. The water contact angles were measured using a goniometer (OCA15, Data physics, Germany) by the sessile-drop method at 25 °C. Typically, 2 μl deionized water was dropped on each sample, and the surface contact angles were calculated based on the images obtained using the goniometer-equipped camera.
X-ray photoelectron spectroscopy (XPS). XPS analysis was performed using an X-ray photoelectron spectrometer (PHI X-tool, Ulavac-PHI Inc) equipped with a monochromatic Al Kα source to characterize the surface composition of the grafted PHEMA brushes. All data were obtained at the photoelectron take-off angle of 45° with respect to the sample surface and analyzed by the XPSpeak41 software. The spectra were peak fitted, and all the binding energies were referenced relative to the aliphatic hydrocarbon C1s signal at 285.0 eV. The atomic concentrations of the elements were calculated using integral peak intensities and sensitivity factors obtained from the manufacturer's handbook.
Atomic force microscopy (AFM). The thickness of the PHEMA brushes was examined by an AFM (Asylum research, American) via measuring the height of the cross-section profile from a scratched groove on the polymer-grafted surface. The measurements were performed using a silicon AFM probe (BudgetSensors) with a spring constant of 40 N m−1 determined via the thermal tune method in the tapping mode at room temperature. The thickness of the PHEMA brush in the wet state was measured using an AFM in a force mode according to the ref. 36 and 37. A silicon AFM probe (BudgetSensors) with a spring constant of 0.3 N m−1 was used for AFM nano-indentation. The trigger deflection was set to 60 nm such that the probe could be completely pressed on the glass surface. The thickness of the brush was calculated using the equation Z = Z1Z2, where Z is the brush thickness, Z2 is the position where the probe deflection reaches 30 nm, and Z1 is the position where the probe deflection increases (Fig. 1c).
image file: c7tb02967e-f1.tif
Fig. 1 Schematic of the PHEMA brush on glass (a) and the measurement of brush thickness in solution (b); mechanism of thickness determination by AFM probe indentation (c); comparison of brush thickness in the dry and wet state at various reaction times (d); AFM images and cross section profile acquired on PHEMA brush with the thickness of 38 nm (e) and 223 nm (f) in the dry state; brush thickness increases as a function of polymerization time (g).

Cell culture

The BMSCs and ATDC5 were cultured in 75 cm2 tissue culture flasks with an MSC growth medium (MSCGM) composed of low glucose DMEM supplemented with 10% FBS, 100 units per ml penicillin, and 100 μg ml−1 streptomycin at 37 °C and 5% CO2. HC were cultured in DMEM supplemented with 5% FBS, 100 units per ml penicillin, and 100 μg ml−1 streptomycin. Cells were subcultured until 80–90% confluenced and obtained at the passage 4–5 (P4–P5) for further use.

Cell staining and proliferation

Cells at 5000 cell per cm2 were seeded on the sample surfaces for 24 h, and the cell-seeded samples were rinsed with PBS and fixed with 4% paraformaldehyde for 30 min, followed by permeabilization with 0.1% Trition-X 100 for 5 min. After being washed with PBS, the samples were treated with a 5 μg ml−1 fluorescent phalloidin conjugate solution for 40 min at room temperature. Cell nuclei were counterstained with DAPI. Fluorescence images were acquired using confocal laser scanning microscopy. The cell spreading area and aspect ratio were calculated using the ImageJ software (USA) from 80 cell images randomly obtained on the substrate.

Cell proliferation was quantified by the CCK-8 assay kit following the manufacturer's instructions at the predetermined time intervals of 24 h, 72 h or 120 h. Since the proliferation of these cells selected in this research was very quick, after day 5, cell fusion was noticed for all the cells, and the growth would not follow the exponential rule; thus, we mainly used the data obtained before day 5 to compare the proliferation.

Herein, 12 images containing 80–100 cells were obtained from 12 specific areas evenly distributed on the substrate. The cell morphology was then processed by the image processing software Image J to obtain the spreading cell profile size i.e. perimeter, length, and width according to a predetermined threshold. Based on these parameters, the cell area and aspect ratio (length to width) were calculated by the software.

Cells adhesion force

The cell adhesion forces on the PHEMA brushes were measured using a centrifugation method according to the literature.14,38 In general, cells were cultured on the substrates for 12 h and washed two times using PBS to remove the floating cells. The numbers of cells on the substrates were then counted using a microscope. After this, the substrates were transferred into centrifuge tubes filled with PBS and centrifuged at 800 and 1500 rpm for 5 min, separately. The numbers of adherent cells were also counted. The adhesion force of the cells was calculated according to the equation reported in the literature.14,38

Real-time PCR analysis for gene expression

Total RNA extracted from the cells using 1 ml TRIzol reagent was dissolved in diethylpyrocarbonate water and quantified by Nanodrop spectrophotometry (Thermo, USA) at 260 nm. cDNA was synthesized from 1 μg total RNA through reverse transcription using the All-in-One First-Strand cDNA synthesis kit according to the manufacturer's protocol. The sequences for primers are shown in the ESI of Table S1 and S2. Quantitative real-time PCR was performed using the All-in-One qPCR Mix according to the manufacturer's instructions. Relative mRNA levels were calculated and normalized to GAPDH.

Statistical analysis

All statistical analyses were performed using the SPSS software. Statistical significance was determined by the analysis of variance (ANOVA) followed by Fisher's LSD with a significance level of P < 0.05. All experiments have been repeated three times, and the results are displayed as mean ± standard error.


The PHEMA brush on a substrate

Fig. 1a shows a schematic of PHEMA grafting on a glass substrate. The brush thicknesses in the dry and wet states were measured, as shown in Fig. 1. When the AFM probes are approaching, touching, and pressing the polymer brushes, the micro-cantilevers of the probe will bend and deform as a result of being hindered by the polymer brush. Thus, the laser signal of the photodiode array, which is reflected by the micro-cantilever, will be obtained as an AFM force curve (Fig. 1b). When the AFM probe is pressed into the samples with different stiffnesses (i.e. the glass surface and polymer brush have different stiffnesses), the force curve will change (Fig. 1c). The thickness of the polymer brush in the wet state is then calculated by the indentation of the probe in the Z axis direction.36,37 The results are shown in Fig. 1d, and the thickness of the polymer brush in the wet state is 1.3–2.5 times larger than that in the dry state; the AFM images (Fig. 1e and f) show the rough surface tethering of the PHEMA brush with a depth of 38 nm and 223 nm by measuring the profile of the scratched groove on the surface. A good linear relationship between the brush depth in the dry state and polymerization time can be observed in Fig. 1g. Longer reaction times resulted in more HEMA monomers combining into polymer chains; this led to a thicker stacking of the polymer layer on the glass. Because thickness in the dry state would be measured more precisely, we mainly used the thickness in the dry state to compare the difference, see the discussion section.

Fig. 2a shows the characteristic absorption bands, including O–H at 3500 cm−1, C–H at 2900 cm−1, and C[double bond, length as m-dash]O at 1720 cm−1, for PHEMA, suggesting that the brush has been successfully grafted onto the substrate. The increase in the intensity of the C[double bond, length as m-dash]O absorption band also reconfirms that the brush thickness increases with reaction time.

image file: c7tb02967e-f2.tif
Fig. 2 ATR-FTIR spectra of the PHEMA brush with different polymerization times (a), XPS survey spectra of PHEMA brush and elemental scan for the C1s (b), images of water contact angle on PHEMA with various thicknesses (c), static water contact angle (d), advancing water contact angle (e), and receding contact angle (f) on PHEMA brush with various thicknesses.

XPS survey spectra (Fig. 2b) of PHEMA polymerized for 30 min and 300 min show the binding energy of C1s and O1s at 287 eV and 525 eV, respectively. The resulting C1s band is resolved into four component peaks in the high resolution spectra at 289.1 eV, 287 eV, 285 eV, and 286.4 eV (Fig. 2b, inset), assigned to O–C[double bond, length as m-dash]O, C–O–C, C–C, and C–OH, respectively.39 The atomic ratio is similar to that displayed in the ESI of Table S3. XPS was used to check the chemical compositional variation after surface polymerization. As the results show, the chemical composition does not change as the polymer chain elongates; this avoids the possible interference of chemical compositional variations on the cell behavior.

The static, advancing, and receding water contact angles are displayed in Fig. 2c–f. The water contact angle decreases sharply after the introduction of the polymer brush into the bare glass surface and remains stable when the brush thickness exceeds 30 nm. The value of the static and advancing water contact angle leveled out at 40°–56°, whereas that of the receding water contact angle stabilized at 9°–26°.

Cell proliferation on the PHEMA brush

Fig. 3 shows an increase in the OD value of BMSCs, ATDC5, and HC cells on the PHEMA brushes with various depths within five days, indicating that all the cells proliferate well on these surfaces. For a specific cell line, cells show higher proliferation on the bare glass or on the substrate having a thinner brush depth; this indicates the interference of the brush depth on cell proliferation. Among the three cell lines, BMSC shows highest proliferation on all substrates possibly due to its high activity.
image file: c7tb02967e-f3.tif
Fig. 3 Proliferation of BMSCs, ATDC5, and HC cells on the surface with various thicknesses of the PHEMA brush at 1, 3 or 5 day.

Cell morphology on the PHEMA brush

Fig. 4a reveals the morphology of the three cell lines and their F-actin staining images on different surfaces obtained by confocal microscopy. On the glass and the surface with short brushes, all cell lines generally show a spindle-like shape, but on the surface with long brushes, they tend to show a round shape. It is also noticed that there is a minor variation in morphology for BMSC on different substrates, but obvious morphology change for ATDC5 and HC cells is observed from spindle to round shape as the PHEMA brush thickness increases. Semi-quantitative statistical data in terms of the cell spreading area and the aspect ratio on the substrates (Fig. 4b) are consistent with the images. All cell lines show the lowest cell spreading area and aspect ratio on the PHEMA brush layer having the highest height; however, on the brush with the lowest height and on the bare glass, cells show the largest area and aspect ratio.
image file: c7tb02967e-f4.tif
Fig. 4 Fluorescence micrograph (a) and spreading area and aspect ratio (b) of different cell lines on the surface of PHEMA brushes having various lengths. samples marked with (*, **, ***) are significantly different at the same time point (p < 0.1, <0.05, <0.01).

Hydrophilic PHEMA, in general, shows an antifouling effect. However, precise control over polymer length can modulate protein adsorption and cell adhesion/spreading. Ren et al. showed that cells migrated to short-length brushes on the polymer-modified surfaces with designed brush length gradients up to 30 nm; this was due to gradient cell adhesion forces generated by the water amount gradient related to the hydrophilic units along the gradient.34 Introduction of adhesive or bioactive molecules into this brush would significantly promote the cell adhesion efficiency and bioactivity of the brush.38,40 In this study, we used different cell lines of MSCs, ATDC5, and HC, and the cells displayed good adhesion and spreading on the polymer brush of 30 nm, but decreased adhesion and spreading on longer polymer brushes; on the contrary, the human vascular smooth muscle cells showed unsatisfied adhesion and spreading on the 30 nm polymer brushes, as reported in Ren's study.34 We speculate that the difference is mainly due to variation in the cell lines. As shown in our study, the cell lines in one differentiation cascade also display different sensitivities towards the same polymer brush length and subsequently demonstrate different behaviors; this can also be the reason the cells show good adhesion and spreading on the PHEMA brush as compared to those on the glass substrate. Other reasons for these cells showing better adhesion/spreading on the brush of 30 nm thickness than that on the glass surface could be the change in surface morphology after polymer brush grafting and the polymer brush being not thick enough to shield the cells from feeling the hard glass substrate.

Cell adhesion force on the PHEMA brush-grafted surfaces

For a more quantitative analysis of the bindings between the substrates and cells, the cell adhesion forces of BMSC, ATDC5, and HC on the PHEMA brush surfaces with different thicknesses were characterized. As shown in Fig. 5, the cell adhesion force decreases with an increase in the thickness of the PHEMA brush. Moreover, the three cells have different cell adhesion forces on the same substrate. The cell adhesion force of ATDC5 was the biggest at 0.38–0.88 nN, and the cell adhesion force of HC was the lowest at 0.18–0.39 nN.
image file: c7tb02967e-f5.tif
Fig. 5 Cell adhesion force of BMSC, ATDC5, and HC on the PHEMA brush with different thickness. samples marked with (*, **, ***) are significantly different at the same time point (p < 0.1, <0.05, <0.01).

mRNA expression of various adhesive molecules on PHEMA brush-grafted surfaces

Fig. 6 displays the mRNA expression of two groups of adhesive molecules, i.e., integrinβ1 and focal adhesion kinase (FAK), which are responsible for cell–ECM interaction, and neural cell adhesion molecules (NCAM) and N-cadherin, which are responsible for cell–cell conjunction, from three cell lines on different surfaces. The expression level of FAK (Fig. 6a) shows a strong connection to that of integrinβ1 (Fig. 6b) for these cell lines on different surfaces. If we sum up the two gene expressions, it is observed that the total expression, except that on the surfaces with the shortest brushes, is lower on the brush grafting surface than that on the bare glass for all the cell lines (Fig. 6c). The ATDC5 cell shows the lowest expression of FAK and integrinβ1 among all the cells; this indicates that it is less sensitive to brush height.
image file: c7tb02967e-f6.tif
Fig. 6 Gene expression of FAK (a), integrinβ1 (b), NCAM (c), and n-cadherin (d) on the surface of PHEMA brushes having various lengths. samples marked with (*, **, ***) are significantly different at the same time point (p < 0.1, <0.05, <0.01).

The mRNA expression level of NCAM (Fig. 6d) and N-cadherin (Fig. 6e) shows a strong connection between them for the HC cell only, but not for the BMSC and ATDC5 cell lines. BMSC shows the highest expression level of the two mRNAs on all the PHEMA grafting surfaces, whereas ATDC5 shows the lowest expression level. The total expression level of NCAM and N-cadherin (Fig. 6f) shows that their expression is not interfered by the gene expression of adhesive molecules responsible for cell–ECM connection. The total mRNA expression of the two genes shows the largest difference for BMSC; this indicates that this cell line is most sensitive to brush height variation.

Cell differentiation on the PHEMA brush

As is known, cell morphology and cell adhesion tend to influence cell differentiation. Since we observed variations of these factors on different surfaces, to gain insight into the effects of brush height on chondrocyte phenotype maintenance, we measured the expression of cartilage-specific genes including Sox9, Collagen type II (Col2), as well as the chondrocyte hypertrophic-related gene of Collagen type I (Col1) and the amount of GAG secretion normalized to the total amount of DNA over two weeks.

The expression of the Sox9 gene is upregulated for all cell lines on the PHEMA brush as compared to that on the bare glass (Fig. 7a, d, and g). The Sox9 gene expression of BMSC and ATDC5 continues to increase with the culture time, but decreases for HC. A significant increase in the Sox9 gene expression is observed on the PHEMA brush as the height of the brush increases; this indicates that a thick PHEMA brush is favorable for maintaining chondrocyte phenotype. Col2, the downstream gene of Sox9, is expressed similarly as Sox9 (Fig. 7b, e, and h). GAG (the characteristic chondrocyte ECM molecule) secretion by the three cell lines normalized to the total DNA shows no significant difference among all the substrates and culturing times (Fig. 7c, f, and i), but the lowest GAG secretion is observed for ATDC5 (Fig. 7f) on the PHEMA surface with the lowest brush depth.

image file: c7tb02967e-f7.tif
Fig. 7 Gene expression of Sox9 (a,d, and e), Col2 (b, e, and h), Col1 (j, k, and l) and GAG secretion normalized to total DNA (e, f, and i) of BMSC, ATDC5, and HC cell lines on the surface of PHEMA brushes having various lengths. samples marked with (*, **, ***) are significantly different at the same time point (p < 0.1, <0.05, <0.01).

Col1, the chondrocyte hypertrophy-related gene, is upregulated for BMSC (Fig. 7j) and HC (Fig. 7l) on the brush surface as compared to that on the bare glass; this suggests that polymer brushes may not inhibit the hypertrophy of cell lines. As the culturing time increases, polymer brushes may promote the hypertrophy of the cells with the increasing expression of Col1, especially for the brush with the highest depth. However, the Col1 gene expression of ATDC5 (Fig. 7k) on the PHEMA brush with a lower brush depth is lowest, showing the inhibited hypertrophic capacity of ATDC5 by the surface.


Physical properties of the biomaterial surfaces are important factors in regulating the cell behavior. The SI-ATRP method can modify the surface chemical composition and surface physical properties as well. It can be performed on smooth 2D surfaces and irregular 3D surfaces regardless of whether these surfaces are exposed outside directly or not.32,33,41–44 A polymer brush could therefore be tethered to metal, inorganic, and/or organic surfaces with controlled grafting density, thickness, and stiffness via the SI-ATRP method.35 The PHEMA polymer brushes on glass were used as a model to investigate their influence on the three cell lines having different chondrogenesis capabilities. These brushes had similar chemical compositions and surface hydrophilicities (Fig. 2), but different thicknesses via controlling the reaction time of grafting polymerization (Fig. 1).

Chondrogenesis involves a cascade of strictly regulated events, including condensation of mesenchymal stem cells (MSCs), chondrogenic differentiation, and maturation of chondrocytes. In fact, accompanying the progression of chondrogenesis, ECM compositions and structures are remodeling and changing. This change would have impacts on the differentiation of the cells at different stages. Conversely, cells at various differentiated stages would also feel dynamical environmental changes. But few works have reported these cells have divers sensibility to feel environment changing. In this study, we have used cell lines of undifferentiated mesenchymal stem cells (MSCs), semi-differentiated precursor chondrocyte of ATDC5 cells, and completely differentiated mature cells of human chondrocytes (HC) as model cells to investigate the impacts of the grafting thickness on cell behaviors, and the diverse sensibility of the cells in the various differentiation cascade stages in the feeling surroundings has been compared. It may not only provide good in vitro models for investigation of cell–matrix interactions during cartilage tissue development but also provide important information for the design and preparation of tissue engineering scaffolds.

In this research, the three cell lines showed different proliferation rates on the same PHEMA brush. However, it may not be clear that the cell lines must have proliferation priority towards a specific substrate since these cell lines also show different proliferation rates on the bare glass. The different rate should be resulted from the inherent proliferation difference among these cells.

Cell morphology, adhesion, and aggregation have been reported to rely on the expression of adhesion molecules5 including integrins,45 FAK,46 NCAM, and the N-cadherin.47–49 Cells on low-depth brushes developed spread morphology as compared to those on high-depth brushes that developed a round-shaped morphology, and the round cells were easily detached from the surfaces once attached (observed from experiments); this suggested the impacts of brush length on cell attachment.

FAK and integrinβ1 are the major structures for cell–ECM combinations as well as for cell spreading on the surfaces. Downregulation of these gene expressions was observed with round-shaped cells on high-depth brushes (Fig. 4). Functioned as the integrinβ1 binding adhesive motif, the adsorption of proteins from culture media on the substrate is critical for this step. Shorter brushes with a lower rotation ability did not favor repelling of proteins approaching the surface; therefore, protein adsorption increased, resulting in more adhesive motifs on the surface. This could be the possible reason why on the PHEMA brush of thin height, high FAK and integrinβ1 expression was observed. On the other hand, the cell adhesion force on the PHEMA brush of thin height was bigger than that on the thicker surface. It demonstrates that the higher expression of FAK and integrinβ1 enforces the cell adhesion force on the surface.

Intracellular conjugation is also an important factor that influences the chondrocyte relevant gene expression because cell aggregation is believed to benefit chondrocyte differentiation. These gene expressions for intracellular conjunction were remarkably up-regulated for BMSC although the cell–ECM interaction decreased, but were opposite that of the ATDC cells. No significant relationship was observed for HC cells. These adhesive molecules that control cell–ECM and cell–cell conjugation might induce various complicated reactions in different cell lines. Generally, chondrocyte differentiation was maintained well on the high-depth brush.

The mechanism of the impact of brush length on cell response is not clear. The proposed hypothesis could relate to the rotation of the polymer chain. Brushes with a shorter chain length (low-depth brush) possess fewer conformations via chain rotation; this suggests that there exists rigid polymer properties. On the contrary, brushes with a longer chain length (high-depth brush) possess soft properties. The mechanical properties of the brushes influence the cytoskeleton development of cells adhered to the surface and determine the various morphologies of the cells.

According to the tensional integrity theory,50 cells adjust their own tension to match the elastic modulus of ECM to reach equilibrium. These physical signals are transduced into the cells through the ECM–integrin–cytoskeleton axis system to induce cell proliferation, adhesion, differentiation, and a series of biological functions.13 This can be the reason that these cell lines maintain the chondrocyte phenotype on the surfaces having PHEMA brushes with long length (soft nano-stiffness) as these cells show similar behavior on soft substrates.

We have also noticed that three cell lines have different sensitivities to the substrate nano-mechanic; this suggests that one promising scaffold should not be universal for all kinds of cells. Although the mechanism has not been well studied, possible reasons for this phenomenon can be hypothesized as follows. ECM components that are secreted by cells dynamically change as the trapped cells receive endogenous biomolecular signals such as growth factors or hormones, resulting in variations in the physical properties in the tissue development process; cells trapped in the ECM develop inconsistent cytoskeletal tension to adapt to the diverse ECM environments to maintain homeostatic balance; this results in different sensitivities of the cells based on the surrounding environment when they evolve from progenitor cells to mature tissue cells. This provides useful information for us to design cartilage tissue engineering scaffolds when different types of cells are used.


PHEMA brushes with different thickness were successfully grafted on glass surfaces by a simple and reproducible Si-ATRP method. The thickness of the PHEMA brush increased linearly with polymerization time, showing good accuracy over the thickness from 10 nm to more than 200 nm. When the thickness exceeds 30 nm, the chemical compositions and surface hydrophilicities of the different brushes do not show any significant difference.

Cell behaviors of BMSC, ATDC5, and HC in different cartilage-differentiated stages were studied on PHEMA brush surfaces with different thicknesses. The three cell lines show different behaviors in terms of proliferation, cell morphology, and differentiation on the same brush. Each cell line tends to show a round shape and maintain chondrocyte phenotype on the long-length brush with more intercellular adhesion molecular expression. The possible mechanism might be related to the mechanical properties of the substrates via changing the conformations of the PHEMA brush that could be controlled by the polymer chain length or brush thickness through graft polymerization.

Conflicts of interest

There are no conflicts to declare.


This research was financially supported by the NSFC Project (31470934, 51572110), the Guangdong Science and Technology Program Key project (2014B010105007), the Guangdong Science and Technology Project (2015A020212025), and the Guangdong Natural Science Foundation Project (2016A030313085).


  1. S. Kumar, Nat. Mater., 2014, 13, 918–920 CrossRef CAS PubMed.
  2. R. Cai, T. Nakamoto, N. Kawazoe and G. Chen, Biomaterials, 2015, 52, 199–207 CrossRef CAS PubMed.
  3. I. G. Kim, J. Ko, H. R. Lee, S. H. Do and K. Park, Biomaterials, 2016, 85, 18–29 CrossRef CAS PubMed.
  4. A. Higuchi, Q. D. Ling, S. S. Kumar, Y. Chang, A. A. Alarfaj, M. A. Munusamy, K. Murugan, S. T. Hsu and A. Umezawa, J. Mater. Chem. B, 2015, 3, 8032–8058 RSC.
  5. T. A. Mahmood, R. de Jong, J. Riesle, R. Langer and C. A. van Blitterswijk, Exp. Cell Res., 2004, 301, 179–188 CrossRef CAS PubMed.
  6. A. J. Engler, S. Sen, H. L. Sweeney and D. E. Discher, Cell, 2006, 126, 677–689 CrossRef CAS PubMed.
  7. Y. S. Pek, A. C. Wan and J. Y. Ying, Biomaterials, 2010, 31, 385–391 CrossRef CAS PubMed.
  8. P. C. Georges, W. J. Miller, D. F. Meaney, E. S. Sawyer and P. A. Janmey, Biophys. J., 2006, 90, 3012–3018 CrossRef CAS PubMed.
  9. N. P. Murphy and K. J. Lampe, J. Mater. Chem. B, 2015, 3, 7867–7880 RSC.
  10. S. H. Wu, R. J. Xu, B. Duan and P. Jiang, J. Mater. Chem. B, 2017, 5, 3870–3878 RSC.
  11. M. Y. Yang, G. S. Zhou, H. Castano-Izquierdo, Y. Zhu and C. B. Mao, J. Biomed. Nanotechnol., 2015, 11, 447–456 CrossRef CAS PubMed.
  12. T. G. Kim, H. Shin and D. W. Lim, Adv. Funct. Mater., 2012, 22, 2446–2468 CrossRef CAS.
  13. K. von der Mark, J. Park, S. Bauer and P. Schmuki, Cell Tissue Res., 2010, 339, 131–153 CrossRef CAS PubMed.
  14. S. Yu, Y. Gao, X. Mei, T. C. Ren, S. Liang, Z. W. Mao and C. Y. Gao, ACS Appl. Mater. Interfaces, 2016, 8, 29280–29288 CAS.
  15. G. Wu, M. Xiao, J. Xiao, L. Guo, Y. Ke, H. Li, L. Fang, C. Deng and H. Liao, Mater. Sci. Eng., C, 2018, 83, 67–77 CrossRef CAS PubMed.
  16. J. Chen, Y. Li, B. Wang, J. Yang, B. C. Heng, Z. Yang, Z. Ge and J. Lin, J. Mater. Chem. B, 2018, 6, 675–687 RSC.
  17. C. Chung and J. A. Burdick, Adv. Drug Delivery Rev., 2008, 60, 243–262 CrossRef CAS PubMed.
  18. D. Shi, J. Xiao, R. Gu, G. Wu and H. Liao, J. Biomed. Mater. Res., Part A, 2017, 105, 1156–1165 CrossRef CAS PubMed.
  19. Q. Wang, L. Y. Cao, Y. Liu, A. Zheng, J. N. Wu, X. Q. Jiang and P. Ji, Int. J. Nanomed., 2017, 12, 7721–7735 CrossRef PubMed.
  20. B. J. Niu, B. Li, Y. Gu, X. F. Shen, Y. Liu and L. Chen, J. Biomater. Sci., Polym. Ed., 2017, 28, 257–270 CrossRef CAS PubMed.
  21. J. R. Vetsch, D. C. Betts, R. Muller and S. Hofmann, PLoS One, 2017, 12 Search PubMed.
  22. B. R. Cao, M. Y. Yang and C. B. Mao, Acc. Chem. Res., 2016, 49, 1111–1120 CrossRef CAS PubMed.
  23. J. L. Wang, Q. Yang, C. B. Mao and S. M. Zhang, J. Biomed. Mater. Res., Part A, 2012, 100a, 2929–2938 CrossRef CAS PubMed.
  24. S. Chen, N. Kawazoe and G. Chen, Tissue Eng., Part C, 2017, 23, 367–376 CrossRef CAS PubMed.
  25. X. L. Wang, T. Nakamoto, I. Dulinska-Molak, N. Kawazoe and G. P. Chen, J. Mater. Chem. B, 2016, 4, 37–45 RSC.
  26. S. W. Chen, T. Nakamoto, N. Kawazoe and G. P. Chen, Biomaterials, 2015, 73, 23–31 CrossRef CAS PubMed.
  27. J. L. Wang, L. Wang, M. Y. Yang, Y. Zhu, A. Tomsia and C. B. Mao, Nano Lett., 2014, 14, 6850–6856 CrossRef CAS PubMed.
  28. M. Y. Yang, Y. J. Shuai, K. S. Sunderland and C. B. Mao, Adv. Funct. Mater., 2017, 27 Search PubMed.
  29. T. Rozario and D. W. DeSimone, Dev. Biol., 2010, 341, 126–140 CrossRef CAS PubMed.
  30. N. D. Evans and E. Gentleman, J. Mater. Chem. B, 2014, 2, 2345–2356 RSC.
  31. F. X. Han, C. H. Zhu, Q. P. Guo, H. L. Yang and B. Li, J. Mater. Chem. B, 2016, 4, 9–26 RSC.
  32. A. E. Rodda, F. Ercole, V. Glattauer, D. R. Nisbet, K. E. Healy, A. P. Dove, L. Meagher and J. S. Forsythe, J. Mater. Chem. B, 2016, 4, 7314–7322 RSC.
  33. Q. Guo, X. Cai, X. Wang and J. Yang, J. Mater. Chem. B, 2013, 1, 6644–6649 RSC.
  34. T. C. Ren, Z. W. Mao, J. Guo and C. Y. Gao, Langmuir, 2013, 29, 6386–6395 CrossRef CAS PubMed.
  35. D. Tranchida, E. Sperotto, T. Staedler, X. Jiang and H. Schönherr, Adv. Eng. Mater., 2011, 13, B369–B376 CrossRef.
  36. S. Yamamoto, M. Ejaz, Y. Tsujii and T. Fukuda, Macromolecules, 2000, 33, 5608–5612 CrossRef CAS.
  37. T. W. Kelley, P. A. Schorr, K. D. Johnson, A. Matthew Tirrell and C. D. Frisbie, Macromolecules, 1998, 31, 4297–4300 CrossRef CAS.
  38. T. Ren, Z. Mao, S. E. Moya and C. Gao, Chem. – Asian J., 2014, 9, 2132–2139 CrossRef CAS PubMed.
  39. T. W. Ying Mei, Chang Xu, Kurt J. Langenbach, John T. Elliott, K. L. B. Bryan, D. Vogt, Eric J. Amis and Newell R. Washburn, Langmuir, 2005, 21, 12309–12314 CrossRef PubMed.
  40. T. Ren, S. Yu, Z. Mao, S. E. Moya, L. Han and C. Gao, Biomacromolecules, 2014, 15, 2256–2264 CrossRef CAS PubMed.
  41. G. T. Knight, T. Klann, J. D. McNulty and R. S. Ashton, J. Visualized Exp., 2014, e52186,  DOI:10.3791/52186.
  42. S. Tugulu, P. Silacci, N. Stergiopulos and H. A. Klok, Biomaterials, 2007, 28, 2536–2546 CrossRef CAS PubMed.
  43. A. A. Stefano Tugulu, I. Sielaff, K. Johnsson and H.-A. Klok, Biomacromolecules, 2005, 6, 1602–1607 CrossRef PubMed.
  44. L. Moroni, M. Klein Gunnewiek and E. M. Benetti, Acta Biomater., 2014, 10, 2367–2378 CrossRef CAS PubMed.
  45. X. Chen, J. Du and X. Liang, et al. , Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 9466–9471 CrossRef PubMed.
  46. P. J. R. Shyamali Roy and S. K. Hanks, J. Cell. Biochem., 2002, 84, 377–388 CrossRef PubMed.
  47. P. Raffo, S. Tavella, C. Tacchetti, R. Cancedda and P. Castagnola, Exp. Cell Res., 1994, 215, 354–362 CrossRef PubMed.
  48. T. R. Tyndall WA, Mol. Biol. Cell, 1994, 5, 103 Search PubMed.
  49. T. R. Oberlander S, Development, 1994, 120, 177–187 Search PubMed.
  50. D. E. Ingber, Prog. Biophys. Mol. Biol., 2008, 97, 163–179 CrossRef PubMed.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c7tb02967e

This journal is © The Royal Society of Chemistry 2018