at SciVerse ScienceDirect

Biomaterials 34 (2013) 55e63

Contents lists available

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

The effect of molecular mobility of supramolecular polymer surfaces on fibroblastadhesion

Ji-Hun Seo a,b, Nobuhiko Yui a,b,*a Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Tokyo 101-0062, Japanb JST-CREST, Tokyo 102-0076, Japan

a r t i c l e i n f o

Article history:Received 22 August 2012Accepted 26 September 2012Available online 15 October 2012

Keywords:Molecular mobilityPolyrotaxaneCell adhesionFibronectin

* Corresponding author. Institute of BiomaterialsMedical and Dental University, Tokyo 101-0062, Japfax: þ81 3 5280 8027.

E-mail address: yui.org@tmd.ac.jp (N. Yui).

0142-9612/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.biomaterials.2012.09.063

a b s t r a c t

The effect of hydrated molecular mobility of polymer surfaces on cell adhesion behavior was investi-gated. ABA-type block copolymers composed of polyrotaxane (PRX) and hydrophobic anchoring terminalsegments were synthesized as a platform of molecularly mobile surfaces. The result of QCM-Dmeasurement in water revealed that the molecularly mobile PRX block copolymer surfaces werehigher in hydrated molecular mobility than the corresponding random copolymer surfaces with similarcontent of hydrophobic methoxy groups. The number of adhering fibroblasts depended on the amount offibronectin adsorbed from serum but was independent of the molecular mobility. However, themorphology of the adhering fibroblasts was strongly dependent on the extent of molecular mobility inwater. These results indicate that molecular mobility on polymer surfaces is one of the significantconsiderations for regulating cellular responses.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Regulation of cellular responses on artificial materials is a crit-ical issue in the fields of cell biology, biomaterials science, regen-erative medicine, etc. Because cells determine their physiologicalactivity by communicatingwith adhering substrates, awell-definedsurface property is essential for regulating cellular responses [1,2].Various studies have been carried out to find the critical factors thatare responsible for cellular responses, such as adhesionmorphology, directionality, differentiation rate, and even stem celldifferentiation [3e5]. All these studies have clarified that cellularresponses on artificial materials are affected by various physico-chemical factors such as polarity, roughness, and chemicalcomposition [6]. Recently, it was further clarified that cells cansensitively respond to not only the physicochemical properties butalso the mechanical properties of materials surfaces, such as bulkstiffness [7]. Although these studies have been successfully carriedout for clarifying the factors that affect cellular responses, one ofthe important factors, hydrated molecular mobility of the surfaces,still remains an unexplored research area. Living cells are known tocommunicate with their environments via their surfaces not

and Bioengineering, Tokyoan. Tel.: þ81 3 5280 8020;

All rights reserved.

statically but dynamically, thus inducing hectic and dynamicmotions such as pinocytosis [8]. This dynamic nature of the cellsurface allows the cell to continuously remodel the extracellularmatrix (ECM) and change its functionality [9]. Therefore, a mate-rials property in response to external signals, especially cell-induced dynamic signals, should be considered as a definitefactor in designing biomaterials. Recently, surface analysis usingquartz crystal microbalance-dissipation (QCM-D) equipment isgaining recognition as a useful method for measuring changes inhydrated molecular mobility in the microenvironment neara materials surface [10]. Highly dynamic molecular segments suchas grafted polymer chains or weakly cross-linked hydrogels inwater are known to have a high value of energy dissipation inresponse to microvibration (external signal) of the surfaces [11].Therefore, it may be hypothesized that the ability of materialssurfaces to respond to the dynamic motion of cell surfaces ina physiological environment could be indirectly obtained bymeasuring the dissipation signals in water in response to micro-vibration. If a meaningful relationship between the energy dissi-pation factor and cellular responses is obtained, the paradigm indefining the property of the biomaterials surface could be changedfrom static and stationary terms to dynamic and time-dependentterms for understanding dynamic communication between cellsand materials. In order to realize this hypothesis, a materials line-up possessing a wide range of hydrated molecular mobility isrequired. From this point of view, polyrotaxane (PRX) block

J.-H. Seo, N. Yui / Biomaterials 34 (2013) 55e6356

copolymer is considered a good model materials of extremelydynamic surfaces in water because of the mobile nature of a-cyclodextrin (a-CD) threaded onto a polyethylene glycol (PEG).Hydrated molecular mobility of the PRX block copolymers can beeasily modulated by changing the number of a-CDs threaded on thePEG backbone [12]. Previously, we have successfully synthesizedPRX block copolymers that can induce non-specific proteinadsorption on their surface [13]. However, the relationshipbetween hydrated molecular mobility and cell adhesion behaviorremains an unexplored area. In this study, we synthesized a varietyof PRX block copolymers with different numbers of CDs andmethylation degrees to modulate the hydrated molecular mobilityof their surfaces in an aqueous environment. The relationshipbetween adhesion behavior of fibroblasts andmolecular mobility inwater, estimated by QCM-D measurement, was investigated.

2. Materials and methods

2.1. Materials

2-Methacryloyloxyethyl phosphorylcholine (MPC) was obtained from NOF Co.(Tokyo, Japan), and 4-(benzodithioyl)-4-cyanopentanoic acid (CTA) was synthesizedaccording to a previously reported method [14]. a-CD, n-butyl methacrylate (BMA),sodium hydride, iodomethane, a,a0-azobisisobutyronitrile (AIBN), and all theorganic solvents were purchased from Tokyo Kasei Co. (Tokyo, Japan) and used asreceived. PEG (average molecular weight of 20 kDa) (PEG 20k) was purchased fromSigmaeAldrich Chemical Co. (St. Louis, MO, USA), and the hydroxyl end group wassubstituted with amine groups by using a previously reported method [15].2-Methoxyethyl methacrylate (MEA) and 2-hydroxyethyl methacrylate (HEMA)were also purchased from SigmaeAldrich Chemical Co. (St. Louis, MO, USA) andwere distilled by passing them through a basic alumina column to remove aninhibitor prior to use.

Goat polyclonal antibody to mouse IgG conjugated with horseradish peroxidase(HRP) was purchased from Abcam Inc. (Cambridge, MA, USA), and mouse mono-clonal anti-fibronectin antibody (clone FN-15) was purchased from SigmaeAldrich(St. Louis, MO, USA). Cell Counting Kit #8 and Hoechst 33258 were purchased fromDojindo Lab (Kumamoto, Japan). Alexa Fluor 546 Phalloidin and other biologicalreagents were purchased from Gibco Invitrogen Corp. (Grand Island, NY, USA).

2.2. Synthetic process of polymer samples

PEG 20k macro CTA was synthesized as follows: 1 g of PEG 20k bis-amine(0.050 mmol) and 0.018 g of dimethylamino pyridine (0.15 mmol) were dissolvedin 5 mL of dichloromethane. To this, 0.14 g of CTA (0.50 mmol) and 0.082 g of watersoluble carbodiimide (0.50 mmol) were added and stirred for 12 h at roomtemperature. Next, fresh dichloromethane was added, and the mixture was re-precipitated in cold diethyl ether. The crude product was then dissolved in water,and dialysis was carried out (MWCO 3500) for a day, followed by lyophilization.

The obtained PEG macro CTA (0.35 g) was then mixed with 3.5 g of a-CD in25 mL of water at room temperature until a light pink and turbid precipitate wasformed. The precipitate was then isolated by centrifuging and washed again with25 mL of water, followed by a repeat of centrifuging. The pink-colored pseudo-PRXmacro CTA was finally obtained as an inclusion complex after lyophilization.

Synthesis of PRX block copolymers: The typical synthetic process of PRX blockcopolymer was as follows: 0.600 g of pseudo-PRX macro CTA was allowed to reactwith 0.354 g of MPC (1.20 mmol) and 0.633 g of BMA (4.45 mmol) monomer in 7 mLof ethanol/toluene (1/1) mixed solvent by using 0.820 mg of AIBN (5.00 mmol) as aninitiator. The heterogeneous solution was bubbled with an argon (Ar) atmospherefor 15 min prior to placement in a 60 �C oil bath. After 24 h, 15 mL of fresh mixedsolvent was added to the solution, and the precipitate was obtained by centrifuging.The obtained polymer was sequentially washed with ethanol, acetone, dimethylsulfoxide (DMSO), and acetone to remove residual monomers and a-CD. The finalprecipitate was then dried at 40 �C in vacuo, and the polymer was obtained asa white powder. Two different PRX block copolymers containing a different numberof a-CDs were synthesized by changing the in feed ratio of a-CD.

Methylation of PRX block copolymers: 200 mg of the synthesized PRX blockcopolymer was heterogeneously dissolved in 7 mL of dehydrated DMSO. To this,0.155 g of sodium hydride (6.3 mmol) was added under an Ar atmosphere andmixed for 30 min at room temperature. Next, 0.102 g of iodomethane (0.719 mmol)was slowly added to the mixture and stirred for 3 h at room temperature. After thepH was neutralized, the mixture was transferred to a dialysis tube (MWCO 20000),and the dialysis process was carried out for 3 days. The methylated PRX blockcopolymer was then obtained by lyophilizing. PRX block copolymers havingdifferent degrees of methylation were obtained by changing the methylation timefrom 1 h to 3 h.

Synthesis of random copolymers containing methoxy and hydroxyl groups:Random copolymers containing the same weight composition of methoxy andhydroxyl groups as the PRX block copolymers were synthesized as follows: 0.633 gof BMA (4.45 mmol), 0.354 g of MPC (1.2 mmol), 0.637 g of HEMA (4.9 mmol), and1.106 g of MEA (8.51 mmol) were dissolved in 8 mL of toluene/ethanol (1/1 vol%)mixed solvent with 0.4mg of AIBN. After being bubbled with dry Ar, themixturewassealed and kept in a 60 �C oil bath for 24 h. The reaction mixture was precipitated incold diethyl ether, and the precipitant was transferred to a dialysis tube (MWCO10000) with water for 2 days of dialysis followed by lyophilization. Randomcopolymers with different compositions were synthesized by changing monomercompositions to produce the same weight compositions of methoxy and hydroxylgroups as those of the PRX block copolymers.

2.3. Surface characteristics

The synthesized polymer (5 mg) was initially dispensed in 5 mL ethanol. Next,5 mL of water was added to prepare 0.05 wt% of clear polymer solution. Eachpolymer solution (30 mL) was then uniformly cast on a Cell Desk� (SumitomoBakelite Co., Japan) and glass bottom dish for confocal laser microscope observation,and dried in a clean box at room temperature for a day. Each polymer surface wasstabilized in water for a day prior to surface characterization and other biologicalevaluations.

QCM-D monitoring on the polymer surfaces was carried out by using the Q-sense E1-HO device (Q-sense AB, Gothenburg, Sweden). The molecular mobility atthe hydrated surfaces was estimated as follows: The Au sensor was cleaned byapplying an O2 plasma treatment for 5 min and a sequential washing with acetoneand ethanol, followed by drying with an Ar blowing device. The sensor was placed inan open-type chamber equipped with the QCM-D apparatus at 25 �C. The resonancefrequency at 35 MHz (fgold, dry) and the dissipation energy (Dgold, dry) were thenmeasured. Subsequently, fgold, wet and Dgold, wet in a hydrated state were measuredwith the bare gold in contact with pure water. After the water was removed, 30 mL ofeach polymer solution was dropped on the surface. After the surface was dried, theresonance frequencies (fsample, dry and fsample, wet) and dissipation energies (Dsample,

dry and Dsample, wet) of the coated surface in both dry and hydrated states weremeasured using the same procedure as above.

2.4. Evaluation of fibronectin density on the block copolymer surface

Enzyme-linked immunosorbent assay (ELISA) was carried out to estimate thesurface density of fibronectin, representative cell adhesive protein. Initially, eachpolymer surface was brought in contact with 10% fetal bovine serum (FBS) for 1 h at37 �C. After 3 rinses with PBS, each sample was brought in contact with 2 mg/mL ofthe primary antibody (anti-fibronectin) solution for 1 h at room temperature. After 4rinses with PBS, the samples were allowed to react with 8 mg/mL of the secondaryantibody conjugated with HRP in bovine serum albumin (BSA)-pretreated 24-wellplates for 2 h. After 6 rinses with PBS, 0.5 mL of solution (mixture of 10 mL gua-nylic acid buffer [pH 3.3], 0.125 mL of 3,30 ,5,50-tetramethylbenzidine [44 mM], and0.018 mL of H2O2) was added to each sample surface in the BSA-pretreated well.After the reaction was quenched with 2N sulfuric acid, the absorbance at 450 nm ineach resulting solutionwasmeasured by a microplate reader (Multiskan FC; ThermoFisher Scientific, St. Herblain, France).

2.5. Evaluation of adhesion behavior of fibroblast

The cell adhesion test using NIH3T3 mouse fibroblast was performed on eachpolymer surface. Approximately 1.0 � 105 cells in 1.0 mL of minimum essentialmedium (Invitrogen Corp. Carlsbad, CA, USA) supplemented by 10% FBS was incu-bated on the polymer surfaces for 3 h. After a rinse with fresh medium, the surfaceadhering cells were observed using an optical microscope, and the number ofadhering cells was counted by a Cell Counting Kit #8 (Dojindo, Tokyo, Japan).

For fluorescent microscope observation, the cells on the polymer surfaces werestained as follows: Each substrate was carefully washed with fresh PBS and fixedwith 4.0% paraformaldehyde for 10 min at room temperature. After being washedwith fresh PBS, the cells were permeabilized with 2.5% Triton X-100 for 10 min andrinsed againwith PBS. Alexa Fluor 546 Phalloidin (diluted 1:200) and Hoechst 33258solution were allowed to sequentially react in the dark for 1 h at room temperatureafter gentle washing with PBS. The samples were then washed with PBS andobserved by a confocal laser microscope (FV10i; Olympus, Tokyo, Japan). The pro-jected cell area and best-fit ellipse aspect ratio of stained fibroblasts were calculatedusing ImageJ software. The aspect ratio was defined as the ratio of short axis dividedby long axis of the best-fit ellipse.

3. Results

Two series of PRX block copolymers with different numbers ofa-CDmolecules (5�8% CDs threading determined by 1H NMR to themaximum stoichiometric value of a-CD/PEG inclusion complex,

J.-H. Seo, N. Yui / Biomaterials 34 (2013) 55e63 57

M1; 45�53%, M2) were synthesized. Hydrophobic methoxy groupswere then introduced to a-CD in each group of PRX block copoly-mers with 3 different compositions. Fig. 1 shows the molecularstructure of PRX and random copolymers used in this study, as wellas representative 1H NMR spectra of PRX and random copolymers.The characteristic signal of a-CD and its methylation peak areobserved at 4.8 and 3.2e3.4 ppm, respectively. Table 1 shows themolecular profile of the PRX and random copolymers. Each PRXblock copolymer group of 2 different numbers of a-CD (M1 andM2)was methylated to contain hydrophobic methoxy groups with 3different compositions, and the polymer surfaces were prepared bya simple solvent cast method (Scheme 1).

To estimate the dynamic nature of the prepared PRX blockcopolymers, the hydrated molecular mobility in an aqueous mediawas estimated by QCM-D measurement. To consider the effect ofthe amount of polymers, each DD value was normalized to itsadsorption mass related factor (Df) using the following equation[13,16]:

Surface mobility factor�Mf

�¼

�Dsample;wet � Dgold;wet

�.

��fgold;dry � fsample;dry

�

Fig. 2 shows the resulting value ofMf on each prepared polymersurface. In both PRX block copolymer groups, the significantincrease in Mf value was observed when the degree of methylationwas over 50%. In the case of random copolymers, the Mf valueslightly decreased within a reasonable error as the composition ofhydrophobic methoxy groupwas increased.Wettability and surfaceroughness of the copolymer surfaces was estimated by measuringair bubble contact angle inwater and AFM topological analysis, and

5.5 5.0

H1

4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ppm

MPC

2(-O

CH

2-)

MPC

(-CH

2N-)

BMA(

-CH

2-)

BMA(

-CH

2-)

MPC

+BM

A(-C

H3)

PEG O

Me3 OM

e6

S

SO

OO

OO

O

O O

nOH

NCO

O

Same wt %

OO

OO P

N +

a

OCH3

a NCNH

NH

mb

OO

+ N

P

a

OO

H O

H

OO

H

HO

6

123

4 56 CH

CH

3

3

CH3

Fig. 1. 1H NMR (DMSO-d6:MeOD ¼ 1:1) of (a) M2c-PRX and (b) M2c-Random and their moleinformation.

it was confirmed that all the copolymer surfaces showed hydro-philic nature (0e30�) in aqueous media and the RMS roughnesswas within 5 nm (data not shown) which indicates that thehydrophobic surface property was not induced by the surfaceroughness.

Prior to the investigation of cellular responses, the density offibronectin, a representative cell adhesive protein, on the copol-ymer surfaces was estimated by ELISA after contact with 10% FBS.Fig. 3 shows the results of relative fibronectin density on thecopolymer surfaces. In the case of M1-PRX, which contains less a-CDs, the fibronectin density increased with themethylation degree.In contrast, M2-PRX showed almost the same amount of fibro-nectin, being independent of the methylation degree. The amountof fibronectin adsorption on the random copolymers graduallyincreased with the wt% of the methoxy group.

The initial adhesion behavior (3 h) of NIH3T3 mouse fibroblastswas estimated by combining confocal laser microscopy with ImageJanalysis. The resulting cell morphologies are shown in Fig. 4. In thecase of M1 copolymers, significant fibroblast adhesion wasobserved only for M1c-PRX. Contrary to this, all the M2-PRXcopolymers showed strong adhesion behavior. In the case of M2-random copolymers, the adhesion area of fibroblasts increasedwith the composition of the methoxy group. Generally, the fibro-blasts on the M2-PRX surfaces were observed as an elongatedcytoskeleton, whereas radial shapes were generally observed onthe random copolymer surfaces.

The projected cell area and the aspect ratio of adhering fibro-blasts were calculated by ImageJ software. Fig. 5 (a) shows theresults of projected cell area of the adhering fibroblasts after 3-hincubation. Except for the case of M2c, fibroblasts on the PRXsurfaces showed larger projected cell area than that on the random

5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 ppm

1 3

4

2

OO m S

S

Same mol %

b

O

OO

OO

OO

O

OOH

Oc d

23 4

ba

1

OO

N +

O P

b

cular structures. 1H NMR of M2a and M2b (PRX and Random) are shown in supporting

Polyrotaxane segment

Anchoring group (PMB)

Stabilizing in water Cell adhesion behaviorCast

M1-PRX M2-PRX

0 % methylation (M1a)

13 % methylation (M1b)

90 % methylation (M1c)

43 % methylation (M2a)

61 % methylation (M2b)

70 % methylation (M2c)

Scheme 1. Schematic explanation of development of dynamic surface.

Table 1Molecular profile of synthesized copolymers (1H NMR).

PRX blockcopolymer

MPC (mol % in PMB, 1HNMR)

nBMA (mol % in PMB, 1HNMR)

aNum. of CD (threading %) (/PEG chain, 1HNMR)

Methylation (%, 1HNMR)

OMe (wt %, 1HNMR)

M1a-PRX 12.0 88.0 12 (5.3) 0 0M1b-PRX 25.0 75.0 18 (7.9) 13.0 0.41M1c-PRX 12.0 88.0 12 (5.3) >90 6.1M2a-PRX 20.9 79.1 123 (54.2) 43.3 9.8M2b-PRX 24.6 75.4 104 (45.8) 61.1 13.5M2c-PRX 26.0 74.0 103 (45.4) 70.0 15.4

Random copolymer MPC (mol %, 1H NMR) nBMA (mol %, 1H NMR) HEMA (mol %, 1H NMR) MEA (mol %, 1H NMR) OMe (wt %, 1H NMR)

M1c-Random 13.2 55.5 29.1 2.20 0.43M2a-Random 6.90 29.4 28.8 34.9 7.47M2b-Random 6.60 33.1 10.6 49.7 10.6M2c-Random 10.2 30.8 9.00 50.0 10.3

a Theoretically 100% of threading ratio requires 227 CDs per PEG.

1.4

1.2

1

0.8

0.6

0.4

0.2

0M2a M2b M2cM1bM1a

RandomPRX

Mobility factor (Mf) =

Mob

ility

fact

or (M

f,x10

-6)

M1cDsample,wet − Dgold,wetfgold,dry − fsample,dry

( )Fig. 2. Result of Mf estimated from QCM-D.

1.2RandomPRX

Rel

ativ

e am

ount

of 2

nd a

ntib

ody

1

0.8

0.4

0.2

0Cell Desk M1a M1b M1c M2a M2b M2c

0.6

Fig. 3. Relative amount of 2nd antibody binding to the anti-fibronectin 1st antibody(n ¼ 6). All values were normalized to the result of Cell Desk�.

J.-H. Seo, N. Yui / Biomaterials 34 (2013) 55e6358

Fig. 4. Confocal laser microscopy images of adhering fibroblasts after 3 h incubation (10% FBS, 37 �C). Nucleus and F-actin were stained with Hoechst 33258 and Alexa Fluor 546Phalloidin, respectively. Bar ¼ 50 mm.

J.-H. Seo, N. Yui / Biomaterials 34 (2013) 55e63 59

copolymer surfaces. The projected cell area on the random copol-ymer surfaces increased with the composition of the methoxygroup, while it was an almost constant value on the PRX surfaces.Morphology of adhering fibroblasts was numerically estimated bya best-fit ellipse method using ImageJ; the resulting aspect ratio ofadhering fibroblasts is shown in Fig. 5 (b). The aspect ratio offibroblasts adhering on the PRX block copolymers showed lowervalues (more ellipse-like), while the aspect ratio on the randomcopolymer surfaces was close to 1 (more radial shape). The aspectratio of fibroblasts on the PRX surfaces gradually decreased as the

degree of methylation increased, while it was an almost constantvalue on the random copolymer surfaces.

The number of adhering fibroblasts on the copolymer surfaces 1day after the seeding was estimated. Fig. 6 shows the images ofoptical microscopy taken after 1-day incubation. In the case of M1copolymers, only M1c-PRX showed a large amount of adheringfibroblasts on its surface. Contrary to this, the number of adheringfibroblasts was almost saturated on M2-PRX, being independent ofthedegreeofmethylation. In the caseof randomcopolymer surfaces,fibroblast adhesion was observed except for M2a-Random. The

500045004000350030002500200015001000500

0Glass

Proj

ecte

d ce

ll ar

ea (µ

m2 )

M1a M1b M1c M2a M2b M2c Glass

RandomPRX

RandomPRX

10.90.80.70.60.50.40.30.20.1

0

Aspe

ct R

atio

0.01>p**0.05>p*

0.001>p***0.001>p**0.01>p*

M1a M1b M1c M2a M2b M2c

** **

**

**

**

******

**

*

*

**

ba

*

Fig. 5. (a) Projected cell area (3-h incubation) estimated by ImageJ software. (b) Aspect ratio (short axis/long axis) of adhering fibroblast (3 h) estimated by a best-fit ellipse methodusing ImageJ. More than 10 images were used for statistical analysis.

Fig. 6. Optical microscopy images of adhering fibroblasts after 1-day incubation (10% FBS, 37 �C). Bar ¼ 200 mm.

J.-H. Seo, N. Yui / Biomaterials 34 (2013) 55e6360

RandomPRX

Num

ber o

f adh

erin

gce

lls (x

104

/cm

2 )9876543210

Cell Desk M1a M2aM1b M2b M2cM1c

Fig. 7. The number of adhering fibroblasts after 1-day incubation (10% FBS, 37 �C).

J.-H. Seo, N. Yui / Biomaterials 34 (2013) 55e63 61

quantitative results of adhering fibroblasts are shown in Fig. 7. Thenumber of adhering fibroblasts gradually increased with thecomposition of the methoxy group on the random copolymersurface. In the case of PRX surfaces, the number of adhering fibro-blasts showed an almost constant level independent of the degree ofmethylation.

4. Discussion

The ultimate purpose of this studywas to clarify the relationshipbetween hydrated molecular mobility and the cell adhesionbehavior. To this end, designing materials surfaces with variedhydrated molecular mobility is essential. Random copolymers

RandomPRX

RandomPRX

Relative amount of 2nd antibody

Num

ber o

f adh

erin

gce

lls (x

104 /

cm2 )

Proj

ecte

d ce

ll ar

ea (µ

m2 )

Projected cell area (µm2)

876543210

1400

1200

1000800

600400

200

00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

0 200 400 600 800 1000 1200 1400

a

c

Fig. 8. Plots of (a) projected cell area (3 h) vs. relative amount of 2nd antibody binding densidensity, (c) projected cell area (3 h) vs. the number of adhering cells (1 day), and (d) aspec

containing similar weight composition of methoxy and hydroxylgroups as that of the PRX block copolymers were prepared, asshown in Scheme 1, in order to estimate the effect of the compo-sition of the hydrophobic methoxy group. As a result, 2 groups ofPRX block copolymers with different numbers of a-CDs weresynthesized and successfully methylated to contain 3 differentcompositions of hydrophobic methoxy groups. Furthermore, it wasconfirmed that the random copolymers were successfully synthe-sized to contain similar weight compositions of the methoxy groupas the PRX block copolymers.

The hydrated molecular mobility of the polymer surfaces wasestimated by means of viscoelasticity measurement in water usingQCM-D. Generally, the energy dissipation value on a materialssurface (DD) is directly related to the viscoelasticity of the materialsadsorbed on an Au surface. When highly mobile surface elementssuch as tethering polymer chains or weakly cross-linked hydrogelexist in aqueous environment, DD values drastically increase. Thisviscoelasticity of the surface could be interpreted as one of theparameters that indicate hydrated molecular mobility of thematerials surfaces [17]. The interesting result that was gainedduring monitoring the mobility factors is its strong dependence onthe molecular structures rather than the compositions of hydro-phobic methoxy group. Most of the PRX block copolymer surfaces,which contain molecularly mobile a-CD, showed higher values ofMf, compared with random copolymer surfaces (Fig. 2). Moreover,the significant increase of Mf value was observed on the PRX blockcopolymer surfaces when the degree of methylation was over 50%.This result is not explained by the point of surface wettability,dependent on the compositions of hydrophobic groups, becausethe air bubble contact angle on the PRX block copolymer surfaces

RandomPRX

RandomPRX

Num

ber o

f adh

erin

gce

lls (x

104

/cm

2 )As

pect

Rat

io

Relative amount of 2nd antibody

Relative amount of 2nd antibody

876543210

1

0.8

0.6

0.4

0.2

00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

b

d

ty, (b) the number of adhering cells (1 day) vs. relative amount of 2nd antibody bindingt ratio (3 h) vs. relative amount of 2nd antibody binding density.

a b

140012001000800600400200

0Proj

ecte

d ce

ll ar

ea (µ

m2 )

0 0.2 0.4Mobility Factor (x10-6)

0.6 0.8 1 0 0.2 0.4 0.6 0.8 1Mobility Factor (x10-6)

1

0.8

0.6

0.4

0.2

0

Aspe

ct R

atio

RandomPRX

RandomPRX

Fig. 9. Plots of (a) projected cell area (3 h) and (b) aspect ratio (3 h) vs. Mf values.

J.-H. Seo, N. Yui / Biomaterials 34 (2013) 55e6362

was close to 0�30�. One of the possible reasons is the hindranceeffect of methylation to intermolecular hydrogen bonding inneighboring a-CD molecules in aqueous media on CD mobility,which may lead to higher Mf values compared with the non-methylated PRX segment. Although the value of Mf is reflected inthe viscoelastic nature of the whole cast polymer surface, themobile nature of a-CD in water is thought to be one of the factorsthat contribute to the increased hydrated molecular mobility,which is surmised in the concept of viscoelasticity measured byQCM-D.

Adhesion behavior of NIH3T3 fibroblasts on the polymersurfaces was analyzed at the point of cell adhesive proteinadsorption. Cell adhesion to ECM protein is mediated by interactionbetween integrin of transmembrane and binding motif of celladhesion proteins such as fibronectin, vitronectin, and collagen. Inparticular, interaction induced by a5b1 integrinwith a bindingmotifis a major concern, because its activation is directly related withproliferation and differentiation of fibroblast [18]. This a5b1 integrinis known to specifically bind to the RGD motif of fibronectin in thepresence of the proline-histidine-serine-arginine-asparagine(PHSRN) domain [19]. Therefore, the amount of surface fibro-nectin is one of the important considerations in discussing celladhesion behavior on PRX block copolymer surfaces. Fig. 8(a)shows the relationship between the amount of fibronectinadsorption (Fig. 3) and the projected cell area (Fig. 5(a)) at the earlystage of cell culture (3 h). Obviously, the projected cell area linearlyincreased with the amount of the surface fibronectin. Becauseinteraction between materials and cells is induced by the surfaceproteins [20], this result reaffirms the fact that proteinematerialsinteraction is the primary factor inducing cellular responses onthe materials surfaces. The number of adhering fibroblasts aftera long-term adhesion test (1 day) also showed a similar relation-ship with the fibronectin density on the surfaces (Fig. 8(b)). Asa result, it could be confirmed that the number of adhering fibro-blasts 1 day after seeding is strongly dependent on the projectedcell area at the early stage of cell adhesion (Fig. 8(c)). These resultsindicate that the fibroblasts’ adhesion is mainly dominated by theamount of cell adhesive protein adsorbed on the surfaces. However,no significant relationship between adhesion morphology (aspectratio) and the density of surface fibronectin was found, as shown inFig. 8(d). It is known that morphology of the adhering cells is notalways directly related with the amount of surface protein. Instead,physical or geometrical characteristics such as stiffness or surfacemorphology are known to contribute to the fate of adhering cells[21,22]. The projected cell area and the aspect ratio of the adheringfibroblasts were plotted with the Mf value, and the results areshown in Fig. 9. Although the projected cell area showed strong

relationships with the amount of protein adsorption and celladhesion (Fig. 8(a)e(c)), no significant relationship was found withthe Mf value on the PRX block copolymers (Fig. 9 (a)). This resultindicates that the hydrated molecular mobility on the PRX surfaceis not directly related to the amount of protein adsorption and thenumber of adhering fibroblasts. In the case of random copolymers,it is thought that the increased Mf value is mainly due to theincreased polarity induced by decreased composition of hydro-phobic unit. Therefore, the decreased hydrophobic interactionbetween proteins and the polymer surfaces (at higherMf surface) isthe possible reason of decreased projected cell area as increasedMf

value. Contrary to this, a strong linear relationship between theaspect ratio and the Mf value was observed, as shown in Fig. 9(b).Although the mobility factor was evaluated on the Au coated QCM-D sensor which is different substrate with glass that the biologicalresponses were evaluated, it was confirmed that the adsorptiontendency of cell adhesive protein such as collagen and fibronectinon polymer coated Au substrate was same with those on glasssubstrates (to be reported in another journal). The exact reasonwhythis relationship was found is still unclear, and the topic is beingpursued further by applying for various cell lines. Nevertheless, thefollowing hypothesis could be applied to explain the presentresults. Because cell surfaces dynamically change and reorder theirsurfaces, the materials surfaces that dynamically respond to themicrovibration given by QCM-D can actively communicate withsuch cell surfaces, thus inducing cytoskeletal reorganization viaintracellular signal transduction to form more elongated adhesionmorphology.

5. Conclusions

Non-specific protein adsorption and the following fibroblastadhesion on highly dynamic PRX block copolymers were investi-gated to clarify the relationship between hydrated molecularmobility and cell adhesion behavior. By designing various PRXderivatives, we could develop polymer surfaces having a broadrange of hydrated molecular mobility in an aqueous environment.The number of adhering fibroblasts and the projected cell areawerenot related with the hydrated molecular mobility but were directlyrelated with the amount of surface fibronectin. Contrary to this, themorphology of adhering fibroblasts was directly related with thedynamic nature of polymer surfaces in hydrated states. Throughoutthis study, we confirmed the possibility that the fate of fibroblastadhesion could be modulated by the hydrated molecular mobilityand that PRX derivatives are appropriate materials for developinga wide range of hydrated molecular mobilities.

J.-H. Seo, N. Yui / Biomaterials 34 (2013) 55e63 63

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.biomaterials.2012.09.063.

References

[1] Chen H, Yuan L, Song W, Wu Z, Li D. Biocompatible polymer materials: role ofprotein-surface interaction. Prog Polym Sci 2008;33:1059e87.

[2] Horbett TA, Brash JL. Proteins at interfaces II: fundamentals and applications.In: ACS symposium series, vol. 602. Washington, D.C.: American ChemicalSociety; 1995.

[3] GuvendirenM Burdick JA. The control of stem cell morphology and differen-tiation by hydrogel surface wrinkles. Biomaterials 2010;31:6511e8.

[4] Lipski AM, Pino CH, Haselton FT, Chen IW, Shastri VP. The effect of silicananoparticle-modified surfaces on cell morphology, cytoskeletal organizationand function. Biomaterials 2008;29:3836e46.

[5] Seo J-H, Matsuno R, Takai M, Ishihara K. Cell adhesion on phase-separatedsurface of block copolymer composed of poly(2-methacryloyloxyethyl phos-phorylcholine) and poly(dimethylsiloxane). Biomaterials 2009;30:5330e40.

[6] Aizawa Y, Owen SC, Shoichet MS. Polymers used to influence cell fate in 3Dgeometry: new trends. Prog Polym Sci 2012;37:645e58.

[7] Khoutorsky M, Lichtenstein A, Krishnan R, Rajendran K, Mayo A, Kam Z, et al.Fibroblast polarization is a matrix-rigidity-dependent process controlled byfocal adhesion mechanosensing. Nat Cell Biol 2011;13:1457e65.

[8] Mager MD, LaPointe V, Stevens MM. Exploring and exploiting chemistry at thecell surface. Nat Chem 2011;3:582e9.

[9] Kennedy L, Shi-Wen X, Carter DE, Abraham DJ, Leask A. Fibroblast adhesionresults in the induction of a matrix remodeling gene expression program.Matrix Biol 2008;27:274e81.

[10] Meyers SR, Khoo X, Huang X, Walsh EB, Grinstaff MW, Kenan DJ. The develop-ment of peptide-based interfacial biomaterials for generating biological func-tionality on the surface of bioinert materials. Biomaterials 2009;30:277e86.

[11] Ishida N, Biggs S. Effect of grafting density on phase transition behavior forpoly(N-isopropylacrylamide) brushes in aqueous solutions studied by AFMand QCM-D. Macromolecules 2010;43:7269e76.

[12] Ooya T, Eguchi M, Yui N. Supramolecular design for multivalent interaction:maltose mobility along polyrotaxane enhanced binding with concanavalin A.J Am Chem Soc 2003;125:13016e7.

[13] Seo J-H, Kakinoki S, Inoue Y, Yamaoka T, Ishihara K, Yui N. Designingdynamic surfaces for regulation of biological responses. Soft Matter 2012;8:5477e85.

[14] Mitsukami Y, Donovan MS, Lowe AB, McCormick CL. Water-soluble poly-mers. 81. Direct synthesis of hydrophilic styrenic-based homopolymers andblock copolymers in aqueous solution via RAFT. Macromolecules 2001;34:2248e56.

[15] Zalipsky S, Gilon C, Zilkha A. Attachment of drugs to polyethylene glycols. EurPolym J 1983;19:1177e83.

[16] Inoue Y, Ye L, Ishihara K, Yui N. Preparation and surface properties ofpolyrotaxane-coating tri-block copolymers as a design for dynamic bioma-terials surfaces. Colloids Surf B Biointerfaces 2012;89:223e7.

[17] Eisele NB, Andersson FI, Frey S, Richter RP. Viscoelasticity of thin biomolecularfilms: a case study on nucleoporin phenylalanine-glycine repeats grafted toa histidine-tag capturing QCM-D sensor. Biomacromolecules 2012. http://dx.doi.org/10.1021/bm300577s.

[18] Moursi AM, Globus RK, Damsky CH. Interactions between integrin recpetorsand fibronectin are required for calvarial osteoblast differentiation in vitro.J Cell Sci 1997;110:2187e96.

[19] Cutler SM, Garcia AJ. Engineering cell adhesive surfaces that direct integrina5b1 binding using a recombinant fragment of fibronectin. Biomaterials 2003;24:1759e70.

[20] Goddard JM, Hotchkiss JH. Polymer surface modification for the attachment ofbioactive compounds. Prog Polym Sci 2007;32:698e725.

[21] Roach P, Parker T, Gadegaard N, Alexander MR. Surface strategies for controlof neuronal cell adhesion: a review. Surf Sci Rep 2010;65:145e73.

[22] Smith KE, Hyzy SL, Sunwoo MH, Gall KA, Schwartz Z, Boyan BD. The depen-dence of MG63 osteoblast responses to (meth)acrylate-based networks onchemical structure and stiffness. Biomaterials 2010;31:6131e41.