An

Ma

b

a

ARRAA

KACCMN

1

emDiaut

Maafdabtt

0d

Colloids and Surfaces B: Biointerfaces 76 (2010) 236–240

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces

journa l homepage: www.e lsev ier .com/ locate /co lsur fb

mphiphilic comblike polymers enhance the colloidal stability of Fe3O4

anoparticles

yeongjin Kima, Jaeyeon Junga, Jonghwan Leea, Kyunga Naa,b, Subeom Parka, Jinho Hyuna,∗

Department of Biosystems and Biomaterials Science and Engineering, Seoul National University, Seoul 151-742, Republic of KoreaResearch Institute for Agriculture and Life Sciences, Seoul National University, Seoul National University, Seoul 151-742, Republic of Korea

r t i c l e i n f o

rticle history:eceived 30 July 2009eceived in revised form 14 October 2009ccepted 27 October 2009vailable online 1 November 2009

eywords:mphiphilic

a b s t r a c t

Stable colloidal dispersions of magnetite (Fe3O4) nanoparticles (MNPs) were obtained with the inclusionof an amphiphilic comblike polyethylene glycol derivative (CL-PEG) as an amphiphilic polymeric surfac-tant. Both the size and morphology of the resulting CL-PEG-modified MNPs could be controlled and werecharacterized by transmission electron microscopy (TEM). The interaction between MNPs and CL-PEGwas confirmed by the presence of characteristic infrared absorption peaks, and the colloidal stability ofthe nanoparticle dispersion in water was evaluated by long-term observation of the dispersion usingUV-visible spectroscopy. SQUID measurements confirmed the magnetization of CL-PEG-modified MNPs.The zeta potential of the CL-PEG-modified MNPs showed a dramatic conversion from positive to negative

hemical coprecipitation

olloidal stabilityagnetiteanoparticles

in response to the pH of the surrounding aqueous medium due to the presence of carboxyl groups at thesurface. These carboxyl groups can be used to functionalize the MNPs with biomolecules for biotechno-logical applications. However, regardless of surface electrostatics, the flexible, hydrophilic side chainsof CL-PEG-modified MNPs prevented the approach of adjacent nanoparticles, thereby resisting aggrega-tion and resulting in a stable aqueous colloid. The cytotoxicity of MNPs and CL-PEG-modified MNPs was

.

evaluated by a MTT assay

. Introduction

Magnetite (Fe3O4) nanoparticles (MNPs) have recently beenvaluated for applications in biomedical imaging [1,2], immuno-agnetic separations [3–5], and as agents for drug delivery [6–8].ue to their susceptibility to aggregation [9,10], the colloidal stabil-

ty of MNPs is critical in these types of systems. This paper describesmethod for the preparation of well dispersed MNP suspensionssing a polymeric surfactant. This technique further allows long-erm immobilization of proteins on the MNP surface.

Polyethylene glycol (PEG) is a polymer commonly used to coatNPs for biomedical applications because of its biocompatibility

nd nonadhesive properties that arise from steric stabilizing effectst bio-interfaces, highly dynamic motion, and extended chain con-ormation [11,12]. Furthermore, MNPs modified with PEG haveemonstrated a relatively long circulation time in the bloodstream

nd are non-immunogenic, nonantigenic, and resistant to proteininding [13–15]. However, under biological conditions, a risk existshat the PEG coating may deteriorate due to relatively weak interac-ions with the particle [16]. Therefore, the development of a reliable

∗ Corresponding author. Tel.: +82 2 880 4624; fax: +82 2 873 2285.E-mail address: jhyun@snu.ac.kr (J. Hyun).

927-7765/$ – see front matter © 2009 Elsevier B.V. All rights reserved.oi:10.1016/j.colsurfb.2009.10.042

© 2009 Elsevier B.V. All rights reserved.

method for protecting MNPs against this deterioration is of greatimportance.

PEG-containing, comblike polymer (CL-PEG) consists of rel-atively hydrophilic side groups attached along a hydrophobicbackbone. This material is effective in coating hydrophobicnanoparticles because the backbone interacts with the hydropho-bic surface of the particle while the hydrophilic PEG side chainsprovide water solubility and enhance colloidal stability. The rel-ative hydrophilicity of the polymer can be tailored by controllingthe size of the PEG side chains or by chemical functionalization [17].The hydrophilic surface of CL-PEG-modified MNPs enables disper-sion in pure water and allows convenient one-step syntheses forchemical modification.

2. Experimental

2.1. Synthesis of the PEG comblike polymer

CL-PEG was synthesized via a free radical polymerization

[17] by mixing 15.6 mL of 2-hydroxyethyl methacrylate (HEMA,100 mmol), 29.5 mL of hydroxy-poly(oxyethylene) methacry-late (HPOEM, 50 mmol), 0.87 mL of mercaptopropionic acid(MPA, 10 mmol), and 2.46 g of 2,2′-azo-bis-isobutyronitrile (AIBN,15 mmol) in 500 mL of tetrahydrofuran (THF). The solution was

es B: Biointerfaces 76 (2010) 236–240 237

dmwe

2

i[a52pocwfi

2

mT(tJtMsstsh(iEwm

SP

of the wells. The cells were then incubated for 24 h at 37 ◦C in ahumidified incubator maintained at 5% CO2/95% air. Cells were alsoincubated in the absence of MNPs as a negative control and in thepresence of 0.1% Triton X-100 as a positive control.

M. Kim et al. / Colloids and Surfac

egassed by bubbling nitrogen through it for 20 min and theixture was refluxed at 75 ◦C for 5 h. The resulting copolymeras purified by two precipitation steps in 8:1 (v/v) petroleum

ther:methanol and dried under vacuum at 25 ◦C for 24 h.

.2. Preparation of magnetite nanoparticles

A colloidal suspension of magnetite nanoparticles was preparedn accordance with a previously published coprecipitation method18]. First, 0.2 mol FeCl2·4H2O (99%; Sigma–Aldrich, St. Louis, MO)nd 0.4 mol FeCl3·6H2O (98%; Sigma–Aldrich) were dissolved inmL deionized water. A CL-PEG solution at concentrations of 1,, 3, and 4 mM was then added to prepare a precursor solution. Therecursor solution was added drop-wise to an aqueous solutionf ammonium hydroxide (28%, w/w (1.5 M); Duksan Pure Chemi-al Co., Ansan, Korea) with vigorous stirring. The precipitated MNPsere washed five times with deionized water by centrifugation andnally re-dispersed as a colloidal suspension in deionized water.

.3. Characterization of the magnetite nanoparticles

The morphology and size of MNPs were observed with a trans-ission electron microscope (TEM; JEM 1010; JEOL, Tokyo, Japan).

he hydrodynamic diameter of the MNP particles in suspension100 �g/mL) was measured at 90◦ by dynamic laser light scat-ering spectroscopy (DLS; DLS-7000; Otsuka Electronics, Osaka,apan). The surface chemistry of the nanoparticles was charac-erized by Fourier transform infrared spectroscopy (FTIR; M2000;

idac, Hamamatsu, Japan) from 4000 to 400 cm−1 using powderedamples pressed into KBr pellets. The long-term stability of MNPuspensions (10 �g/mL, pH 8.3) was characterized by measuringhe room-temperature optical density at 350 nm (OD350) of tightlyealed samples in a single cuvette over the course of 72 h at 1-intervals using an Optizen 2120UV UV-VIS spectrophotometer

Mecasys Co., Daejeon, Korea). The surface charge of MNPs was

nvestigated by measuring zeta potentials (ELS; ELS-8000; Otsukalectronics) as a function of pH from 4.0 to 12.0. Each measurementas done in triplicate. The degree of magnetization in MNPs waseasured with a superconducting quantum interference device

cheme 1. A schematic diagram shows magnetite nanoparticles modified with CL-EG. The inset presents the chemical structure of CL-PEG.

Fig. 1. FT-IR spectra of (A) magnetite nanoparticles, (B) magnetite nanoparticlesmodified with CL-PEG, and (C) CL-PEG.

(SQUID)-based magnetometer (MPMS-XL; Quantum Design, USA)at room temperature on MNP powders that had been freeze-driedovernight.

Cell viability was assessed by a 3-(4,5-dimethylthiazol)-2-diphenyltertrazolium bromide (MTT) assay. A 90-�L aliquot of NIH3T3 fibroblast cells (2.5 × 105 cell/mL) was seeded into 96-well tis-sue culture plates and 10 �L of MNP suspension was added to each

Fig. 2. (A) The TEM image shows magnetite nanoparticles after modification withCL-PEG; the scale bar corresponds to 20 nm. (B) The size distribution of the modifiedmagnetite nanoparticles was determined from the data in (A).

238 M. Kim et al. / Colloids and Surfaces B: Biointerfaces 76 (2010) 236–240

Fi7

2

u0Wfct

3

giscaP

Ftpwhw

ig. 3. Normalized OD350 readings are shown for pure and CL-PEG-modified MNPsn water. The MNP suspensions were stored inside the UV spectrometer during the2-h course of the measurement.

.4. Atomic force microscopy (AFM)

The AFM topographic images were collected in contact modesing silicon nitride cantilevers (PSIA, Korea, spring constant.6 N/m; tip radius <10 nm) using an XE100 (PSIA, Korea) in air.e used AFM in force spectrometry mode to measure interaction

orces between either pure MNPs or CL-PEG-modified MNPs and aantilever tip that was functionalized with either CH3 thiol or COOHhiol.

. Results and discussion

Under biological conditions, MNPs coated with what areenerally considered biocompatible surfactants may becomencompatible due to weak anchoring of surface modifiers, non-

pecific adsorption, and colloidal instability. CL-PEG containsontinuous hydrophobic components that exhibit stronger inter-ctions with MNPs compared to short-chain surfactants or linearEGs. In addition, the biocompatibility and hydrophilicity of CL-

ig. 4. Surface characterization of MNPs by probing with surface modified cantileverips in air. Larger adhesive interaction forces were observed when pure MNPs wererobed with a hydrophobically modified AFM tip indicating the surface of pure MNPsas relatively hydrophobic. CL-PEG-modified MNPs showed larger interaction withydrophilically modified AFM tip indicating the surface of CL-PEG-modified MNPsas relatively hydrophilic.

Fig. 5. Variations in the zeta potentials of pure and CL-PEG-modified MNPs areshown as a function of the pH of the surrounding medium.

PEG-modified MNPs allow even dispersion in biological fluidsand relatively easy functionalization with established NHS/EDCchemistries due to the presence of surface carboxyl groups(Scheme 1).

The presence of organic molecules on the particle surface wasdetermined by comparing IR absorption peaks of modified anduncoated MNPs. For the uncoated MNPs, the characteristic absorp-tion peak of Fe–O vibrations in Fe3O4 was observed at 580.46 cm−1

(Fig. 1A). This band was blue-shifted to 595.89 cm−1 for CL-PEG-modified nanoparticles (Fig. 1B), suggesting the formation of newbonds between the MNP and CL-PEG [19,20]. The FT-IR spectra inFig. 1B and C confirm the presence of absorption bands represent-ing –C–O–C–, –CH2–, and C O groups and imply the formation ofCL-PEG-modified MNPs as presented in Scheme 1. The immobiliza-tion of CL-PEG on the MNP surface was robust and able to withstandultrasonication and ultracentrifugation at 10,000 rpm for 15 min inwater.

Fig. 2A shows a TEM image of CL-PEG-modified MNPs. The mod-

ified particles were spherical with a relatively uniform diameterof 16 ± 2 nm which was about 6 nm larger than uncoated MNPs(10 ± 2 nm). The size distribution was calculated according to alognormal function using the TEM images in Fig. 2B.

Fig. 6. Magnetic hysteresis curves of pure and CL-PEG-modified MNPs.

es B: B

suOTucaPsc

Pmmip

FmE

M. Kim et al. / Colloids and Surfac

Fig. 3 presents the time-dependent OD350 of a MNP suspen-ion that was allowed to settle for a total of 72 h. The OD350 of thencoated MNP suspension decreased about 45% after 72 h while theD350 of the modified MNP suspension decreased by only 11.21%.his difference is due to the inherent colloidal instability of thencoated particles in water and consequent aggregation and pre-ipitation. The CL-PEG coating prevented aggregation and enabledmore stable aqueous dispersion. Based on the structure of the CL-EG, the methyl methacrylate backbone is likely adsorbed to theurface of the nanoparticle while the relatively hydrophilic PEG-ontaining side chains extend out into the surrounding solution.

The preferred binding of a hydrophobic backbone of CL-

EG molecules was investigated by measuring put-off forces ofolecules immobilized on the colloidal surface using surface-odified AFM tips. The tips were coated with gold and immersed

n either COOH-terminated or CH3-terminated thiols for thereparation of hydrophilically and hydrophobically modified tips,

ig. 7. Cell morphologies of the (A) negative (complete culture medium) and (B) positiveodified MNP suspensions. The concentration of the MNP suspensions was 20 �g/mL. (E)

rror bars indicate standard error (n = 10). *Significant (p < 0.05, ANOVA) difference in com

iointerfaces 76 (2010) 236–240 239

respectively. Hydrophobically modified tips showed adhesiveinteractions with the uncoated MNPs that were about four timesstronger than with the CL-PEG modified MNPs (Fig. 4). It providedthe possibility that the surface of MNPs was relatively hydropho-bic even though a certain amount of oxide groups might existon the surface of MNPs. Further experiments with hydrophilicallymodified tips showed relatively stronger adhesive interactionswith the CL-PEG modified MNPs than with the uncoated MNPsas shown in Fig. 4. It resulted from the preferred attraction ofhydrophobic backbone to the MNPs and extension of hydrophilicside chains out into the surrounding environment as shown inScheme 1.

The magnitude of the zeta potential is proportional to theamount of charge on the nanoparticle surface. Due to the presenceof surface-bound carboxyl groups, the zeta potential of modifiedMNPs changed drastically from a positive (26.05 mV) to a nega-tive value (−39.39 mV) as a function of the pH of the surrounding

control groups (Triton X-100), and after incubation with (C) pure and (D) CL-PEG-Cell viability was quantified with a MTT assay following incubation for 24 h at 37 ◦C.

parison to negative control or MNPs-CL-PEG.

2 es B: B

swcwnosttfblWrcitbbm

mmcurpdibw

ctrihITt(atcdtqiMuMo

[

[

[

[[[[

[

40 M. Kim et al. / Colloids and Surfac

olution (Fig. 5). In contrast, the zeta potential of uncoated MNPsas negative over the entire range of pH values. Moreover, the

olloidal stability of the modified MNPs was high even at pH 8.5,hen the zeta potential approached zero. While the uncoatedanoparticles aggregated regardless of zeta potential, suspensionsf coated MNPs were stable from pH 4.0 to 12.0 and showed noignificant signs of aggregation. Thus, the resistance to aggrega-ion was not due to electrostatic forces alone and was due in parto steric repulsion of PEG side chains. This phenomenon is mani-ested in the use of steric stabilizers, which commonly consist oflock copolymers with lyophilic and lyophobic components. The

yophobic component attaches to the particle surface via van deraals interactions while the lyophilic chain extends into the sur-

ounding medium. In an appropriate solvent system, the lyophilichains of adjacent particles do not interpenetrate but insteadnduce interparticle repulsion. However, in poor solvent systems,he interpenetration of stabilizer chains occurs until it is preventedy elastic repulsion. For these reasons, the observed colloidal sta-ility was achieved with CL-PEG-modified MNPs in a hydrophilicedium.To investigate the influence of the CL-PEG surface coating on the

agnetic behavior of the nanoparticles, magnetization measure-ents were performed using SQUID. Fig. 6 shows the hysteresis

urves obtained with uncoated and CL-PEG-modified MNPs. Sat-ration magnetization was attained at 87.32 and 56.36 emu/g,espectively, and confirmed the characteristic superparamagneticroperties of the magnetite nanoparticles. As the magnetic fieldecreased, the degree of magnetization decreased to zero. This

mplies that MNPs that have been separated from a suspensiony their superparamagnetic properties can be rapidly re-dispersedhen the magnetic field is removed.

The test of cytotoxicity is the initial phase in evaluating bio-ompatibility of biomaterials. Fig. 7 shows in vitro cytotoxicityests after incubation with uncoated and CL-PEG-modified MNPsespectively. After a 24 h incubation, the negative control groupncubated in the complete culture medium without MNPs grewealthily and attached well to the surface of the well plate (Fig. 7A).

n contrast, the positive controls where cells were exposed to toxicriton X-100 (0.1%) showed significant detachment of cells fromhe surface of the well plate inferring the high toxicity to cellsFig. 7B). Similar to the negative control group, cells incubated insuspension of CL-PEG-modified MNPs spread and attached well

o the plate surface even after 72 h incubation (Fig. 7D). However,ells incubated with uncoated MNPs did not spread properly andead cells detached from the surface of the well plate (Fig. 7C) ashe positive control group showed in Fig. 7B. Fig. 7E shows theuantitative viability of cells measured by the MTT assay reveal-

ng a significant difference between uncoated and CL-PEG-modifiedNPs. The level of MTT was lower for the medium containing

ncoated MNPs (54%, p < 0.05) compared with CL-PEG-modifiedNPs (94%, p < 0.05) after a 24 h incubation. Based on the results

f cell morphology and MTT assay, it could be determined that

[

[[

iointerfaces 76 (2010) 236–240

the modification of MNPs with CL-PEG significantly improved thebiocompatibility of a material.

4. Conclusions

CL-PEG-modified MNPs were synthesized by a chemical copre-cipitation method from a ferrous/ferric salt solution and an aqueousammonium hydroxide solution containing CL-PEG at room tem-perature. The presence of CL-PEG on the surface of the MNPs wasconfirmed by characteristic peaks in the FT-IR spectra of the mod-ified particles, and a stable colloidal dispersion of the modifiedMNPs was indicated by optical density analyses. Surface chargemeasurements revealed zeta potentials for both the uncoated andCL-PEG-modified particles as a function of the pH of the surround-ing medium. The CL-PEG modification resulted in a well dispersedsuspension of particles throughout an aqueous solution due to thepresence of hydrophilic side chains at the surface of the MNPs. Thebiocompatibility of MNPs modified with CL-PEG was confirmed byrelative cell morphology observations and MTT assays.

Acknowledgment

This work was supported by Grant R01-2006-000-10217 fromthe Basic Research Program of the Korea Science & EngineeringFoundation.

References

[1] R.Y. Hong, B. Feng, L.L. Chen, G.H. Liu, H.Z. Li, Y. Zheng, D.G. Wei, Biochem. Eng.J. 42 (2008) 290.

[2] D.K. Kim, M. Mikhaylova, Y. Zhang, M. Muhammed, Chem. Mater. 15 (2003)1617.

[3] M. Takahashi, Y. Akiyama, J. Ikezumi, T. Nagata, T. Yoshino, A. Lizuka, K. Yam-aguchi, T. Matsunaga, Bioconjugate Chem. 20 (2009) 304.

[4] Y.C. Liu, Y.H. Che, Y.B. Li, Sens. Actuators B: Chem. 72 (2001) 214.[5] K.S. Lee, S.S. Ryu, C. Kim, B.K. Ju, S.K. Lee, J.Y. Kang, Biochip J. 1 (2007) 165.[6] S.W. Cao, Y.J. Zhu, J. Phys. Chem. C 112 (2008) 12149.[7] S.S. Guo, C.C. Zuo, W.H. Huang, C. Peroz, Y. Chen, Microelectron. Eng. 83 (2006)

1655.[8] Y.N. Xia, P.D. Yang, Y.G. Sun, Y.Y. Wu, B. Mayers, B. Gates, Y.D. Yin, F. Kim, Y.Q.

Yan, Adv. Mater. 15 (2003) 353.[9] A.H. Lu, E.L. Salabas, F. Schuth, Angew. Chem. Int. Ed. 46 (2007) 1222.10] M.A. Morales, T.K. Jain, V. Labhasetwar, D.L. Leslie-Pelecky, J. Appl. Phys. 97 (10)

(2005) Q905.11] S.X. Wang, Y. Zhou, S.C. Yang, B.J. Ding, Colloid Surf. B: Biointerfaces 67 (2008)

122.12] R.R. Seigel, P. Harder, R. Dahint, M. Grunze, F. Josse, M. Mrksich, G.M. Whitesides,

Anal. Chem. 69 (1997) 3321.13] F.X. Hu, K.G. Neoh, L. Cen, E.T. Kang, Biomacromolecules 7 (2006) 809.14] A.K. Gupta, M. Gupta, Biomaterials 26 (2005) 3995.15] M.Q. Zhang, T. Desai, M. Ferrari, Biomaterials 19 (1998) 953.16] W.W. Yu, E. Chang, J.C. Falkner, J.Y. Zhang, A.M. Al-Somali, C.M. Sayes, J. Johns,

R. Drezek, V.L. Colvin, J. Am. Chem. Soc. 129 (2007) 2871.17] J. Jung, K. Na, B. Shin, O. Kim, J. Lee, K. Yun, J. Hyun, J. Biomater. Sci. Polym. Ed.

19 (2008) 161.18] J. Sun, S.B. Zhou, P. Hou, Y. Yang, J. Weng, X.H. Li, M.Y. Li, J. Biomed. Mater. Res.

Part A 80 (2007) 333.19] Z.M. Gao, T.H. Wu, S.Y. Peng, Acta Phys. Chim. Sin. 11 (1995) 395.20] M. Ma, Y. Zhang, W. Yu, H.Y. Shen, H.Q. Zhang, N. Gu, Colloid Surf. A: Physic-

ochem. Eng. Asp. 212 (2003) 219.