Colloids and Surfaces B: Biointerfaces

Colloids and Surfaces B: Biointerfaces 97 (2012) 175 181 Contents lists available at SciVerse ScienceDirect Colloids and Surfaces B: Biointerfaces jou rn al h om epage: www.elsevier.com/locate/colsurfb Synthesis and cellular uptake of scatteredly cyclic RGDfK-conjugated superparamagnetic iron oxide nanoparticles Myung-Ik Yoo a,b, Yeong-Ju Seo c, Kyu-Sil Choi c, Jeong Sook Ha b, Kyoungja Woo a, a Molecular Recognition Research Center, Korea Institute of Science and Technology, P.O. Box 131, Cheongryang, Seoul 130-650, Republic of Korea b Department of Chemical and Biological Engineering, Korea University, Seoul 136-701, Republic of Korea c Molecular & Cellular Imaging Center, Samsung Biomedical Research Institute, Seoul 135-710, Republic of Korea a r t i c l e i n f o Article history: Received 7 October 2011 Received in revised form 22 March 2012 Accepted 5 April 2012 Available online 13 April 2012 Keywords: Superparamagnetic iron oxide nanoparticles (SPIONs) Scatteredly crgdfk-conjugated Ultrasmall Water-dispersible Lipophilic Cellular uptake a b s t r a c t We prepared scatteredly cyclic RGDfK-conjugated water-dispersible superparamagnetic iron oxide nanoparticles (crgdfk WSPIONs) and investigated their cellular uptake to MS-1 cells (mouse endothelial cell lines, express integrin v 3 ) and MCF-7 cells (human breast cancer cells, express low level of integrin v 3 ) using in vitro MRI. The crgdfk WSPIONs were prepared from oleate-protected SPIONs (SPION OA) as follows. Some oleates (OAs) on the SPION OA were substituted by mercaptohexadecanoic acids (MHA) and crgdfks were conjugated to MHAs. Finally, the remaining OAs were substituted by mercaptopropionic acids without detaching preexisting crgdfk-conjugated MHA ligands from the SPION surface. The crgdfk WSPIONs showed drastically higher cellular uptake than its corresponding control WSPIONs to MS-1 cells and also, showed higher selectivity to MS-1 cells than to MCF-7 cells, both implicating integrin v 3 -mediated cellular uptake of scatteredly crgdfk-conjugated WSPIONs. 2012 Elsevier B.V. All rights reserved. 1. Introduction Surface engineering of superparamagnetic iron oxide nanoparticles (SPIONs) have attracted great attention since the SPIONs can be applied as a highly sensitive magnetic resonance imaging (MRI) agent utilizing T 2 contrast enhancement [1 7]. Commercial MRI agents using dextran-coated SPION clusters over 50 nm in hydrodynamic diameter (HD) are easily uptaken by the reticuloendothelial system (RES) and thus, have been used as a liver contrast agent. The clustered or aggregated nanoparticles greatly enhance their delivery to liver and spleen rather than the other target due to their much larger effective hydrodynamic size than the aggregation-free one [8,9]. Therefore, targeting and biocompatible SPIONs with HD below 10 nm in water are getting more important and challenging for noninvasive detection of tumors other than RES system at an early stage [7,10,11]. The core size below 10 nm is also required to ensure a superparamagnetism of the iron oxide nanoparticles without hysteresis, since the hysteresis induces magnetic agglomeration [2,12,13]. However, surface engineering of SPIONs (core size <10 nm) for targeting, biocompatible, aggregation-free, and water-soluble property has been a tough task, mostly because of Corresponding author. Tel.: +82 2 958 5265, fax: +82 2 958 5053. E-mail address: kjwoo@kist.re.kr (K. Woo). the multi-chelating hydrogen bonding between the nanoparticles during conjugation reaction in an aqueous solution [14 17]. It was a quite progress that Xie et al. [7] reported the targeting water-dispersible ultrasmall ( 8.4 nm) iron oxide nanoparticles based on the synthesis using 4-methylcatechol as a new surfactant and subsequent coupling of its aromatic ring to crgdyk via a Mannich reaction. Noteworthy is Tromsdorf et al. s [11] report on water-dispersible ultrasmall (HD < 10 nm) iron oxide nanoparticles based on PEGylation with phosphate functionalized PEG, though their materials do not include targeting molecules. Recently, starting from decanoate-protected Fe-rich SPIONs, we reported [18] the synthesis of scatteredly folate-conjugated lipophilic SPIONs (F LSPIONs) through scattered substitution of decanoate ligands for mercaptoundecanoic acid (MUA) and subsequent conjugation to folate. Here, we also reported [18] their folate receptor-mediated cellular uptake to KB cells (human epidermoid carcinoma cells, folate receptor positive) on a film made of F LSPIONs. This kind of targeting lipophilic SPIONs (LSPIONs) can be a nice candidate for T 2 contrast agent once they can be converted to aggregation-free water-dispersible ones. Meanwhile, we also reported [17] the preparation of scatteredly folate-conjugated quantum dot (QD) with HD below 10 nm by conversion of scatteredly folate-conjugated lipophilic QD to water-dispersible one. This water-dispersible conversion was possible through substitution of remaining decylamine ligands for mercaptopropionic acids 0927-7765/$ see front matter 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.colsurfb.2012.04.009

176 M.-I. Yoo et al. / Colloids and Surfaces B: Biointerfaces 97 (2012) 175 181 (MPAs) without detaching preexisting folate-conjugated MUA ligands from the QD surface. It was suggested that the targeting aggregation-free water-dispersible QDs in a physiological condition were obtainable by interrupting multi-chelating hydrogen bonding interactions between MPAs by using steric hindrance from the protruding targeting molecules [17]. This previous result encouraged us to expand the surface-engineering strategy of QDs to SPIONs and investigate their cellular interactions. In this report, PEG-bonded cyclic RGDfK (PEG crgdfk) was selected instead of folate as a targeting molecule to utilize its wellknown selectivity to integrin v 3, which is rich in the fibrinous cells and tumor tissues undergoing angiogenesis [19,20]. The PEG unit, which has been already bonded to crgdfk before conjugation to SPION, was employed for a biocompatibility [21,22]. However, its symbol PEG is frequently omitted in the abbreviated expression just for simplicity throughout this report. Our surface-engineering strategy worked nicely here with Fe-rich SPIONs like the case of Cd-rich QDs. The strategy is to substitute some oleates (OAs) of OA-protected Fe-rich SPION for mercaptohexadecanoic acids (MHAs) and subsequently conjugate crgdfk to MHA. Finally, the strategy converts scatteredly crgdfk-conjugated lipophilic SPIONs (crgdfk LSPIONs) to water-dispersible ones (crgdfk WSPIONs) by substitution of the remaining OAs for MPAs without detaching preexisting crgdfk-conjugated MHAs from the SPION surface. The crgdfk WSPIONs showed HD below 10 nm and integrin v 3 -mediated cellular uptake though the SPION was scatteredly crgdfk-conjugated. 2. Materials and methods 2.1. Synthesis of SPION( MPA) ex ( MHA apeg) 10 (WSPION) and SPION( MPA) ex ( MHA enpeg crgdfk) 10 (crgdfk WSPION) (material IV in Scheme 1) The PEG-conjugated and PEG crgdfk-conjugated LSPIONs, SPION( OA) ex ( MHA apeg) 10 and SPION( OA) ex ( MHA enpeg crgdfk) 10 (material III in Scheme 1, here ex is unknown excess amount [23]), were prepared according to the previous method [18] with some modifications using O- (2-aminoethyl)-O -methyl polyethylene glycol (apeg, M w = 2 kda, purchased from Fluka) and enpeg crgdfk (PEG M w = 600 Da, vide infra) and then, treated with MPA to transfer to aggregation-free WSPION and crgdfk WSPION. Briefly, Fe-rich SPIONs (8.5 nm by TEM image) protected with OA (SPION OA, I in Scheme 1) were prepared according to the published method [12,24] and saved as-prepared (about 22 ml) under Ar gas. A 3 ml portion of as-prepared Fe-rich SPION OA was washed with ethanol and acetonitrile consecutively, dispersed in toluene (4.5 10 6 M, 6 ml), and kept as a stock under Ar atmosphere to preserve Fe-rich surface until it is used. A 2 ml portion of the stock solution was diluted to 30 ml and treated as follows. (1) MHA (9 10 5 mmol in 0.15 ml toluene, roughly 10 eq. for SPION) was added to SPION OA (9 10 6 mmol in 30 ml toluene) at 100 C and the solution was gently refluxed for 1 h under inert atmosphere. (2) To this solution at room temperature, dicyclohexylcarbodiimide (DCC, 2.7 10 4 mmol in 0.15 ml toluene) was added and stirred for 1 h. Then, apeg or enpeg crgdfk (2.7 10 4 mmol in 0.5 ml dimethylsulfoxide) was added and stirred overnight with intermittent treatments within 1 s in a sonication bath, under inert atmosphere and dark conditions. (3) The solution was transferred to a conical tube and a mixture (0.05 M, 0.027 ml, 150 eq. for SPION) of MPA and NaOH in methanol was added and vortexed for 2 h with intermittent treatments within 1 s in a sonication bath. The short sonication within 1 s was helpful in steps (2) and (3) since it disturbs temporary weak agglomeration of nanoparticles Scheme 1. Surface modification process for aggregation-free, biocompatible, and targeting WSPION [I: SPION OA, II: SPION( OA) ex( MHA) 10, III: SPION( OA) ex( MHA enpeg crgdfk) 10 (crgdfk LSPION), IV: SPION( MPA) ex( MHA enpeg crgdfk) 10 (crgdfk-wspion)]. caused by poor solubility of conjugated crgdfk moiety and MPA in toluene. Here, 150 eq. of MPA was roughly chosen by considering the SPION surface area and the attachable MPA molecules, referring to the attached MPA molecules (80 100) to QD [25]. The resulting precipitate from step (3) was collected by centrifugation, rinsed with ethanol and then water, and finally dispersed in a ph 7.4 PBS solution to make WSPION and crgdfk WSPION solutions (20 ml each). The SPION samples were analyzed by FT-IR spectroscopy (Mattson IR 300) and TEM (Philips CM 30, 200 kv). The samples for FT-IR spectra were prepared as a pellet with KBr after fully dried. The HD was obtained from the Korea Basic Science Institute (Jeonju Center) using diluted samples (1 10 8 M in PBS) with a particle size analyzer (UPA-150). The magnetic hysteresis curve of SPIONs was obtained from the Korea Basic Science Institute (Busan Center) using vibrating sample method (Princeton Measurement Co., Model 2900 Micromag) by measuring the precisely weighed sample in a folded parafilm (5 mm 5 mm) at room temperature. Fe concentrations of WSPION and crgdfk WSPION solutions were obtained after pretreatment using aqua regia from the Advanced Analysis Center of Korea Institute of Science and Technology using atomic absorption spectrophotometer (Thermo Electron Corp.) as 160 and 190 ppm, respectively. The biocompatible and integrin v 3 -targeting molecule, enpeg crgdfk, was prepared by coupling polyethylene glycol (PEG) derivative with crgdfk and ethylenediamine (en) as follows. First, PEG diacid (0.0765 mmol, M w = 600, Fluka) was dissolved in dried and distilled dimethylformamide (DMF, 10 ml). In an ice bath, the diacid groups were activated by slow addition of DCC (0.168 mmol in 1 ml DMF) and stirring for 16 h. Since the solution became translucent, additional DMF was added until the solid was dissolved (solution 1). Meanwhile, the crgdfk 2TFA (70 mg, 0.0842 mmol, M w = 831.71, 99.5%, purchased from FUTURECHEM Co. Ltd.) was dissolved in dried and distilled DMF (2 ml) and

M.-I. Yoo et al. / Colloids and Surfaces B: Biointerfaces 97 (2012) 175 181 177 diisopropylamine (0.842 mmol, 99.95%, Aldrich) was added and stirred for 15 min to remove TFA from crgdfk. This solution was slowly added to solution 1 and stirred for 3 h to make an amide bond between crgdfk and one end of activated PEG diacid. Then, en was added and stirred for 1 h to make an additional amide bond between en and the other end of activated PEG diacid. Diethyl ether was added to collect enpeg crgdfk precipitate. The precipitate was rinsed with methylene chloride, dried with vacuum, filled with Ar gas, and saved in a freezer (87.5 mg, 90.6% yield). The amide bond formation at both ends of PEG with en and crgdfk was confirmed from the FT-IR spectrum (Supplementary data). 2.2. In vitro characterization of vˇ3-specific binding to WSPION and crgdfk WSPION In order to examine whether the scatteredly crgdfk-conjugated WSPIONs specifically bind to integrin v 3, MS-1 cells (mouse endothelial cell line) which express v 3 and MCF-7 cells (human breast cancer cell line) which express relatively low level of v 3 were chosen. Expression of integrin v 3 in MS-1 and MCF-7 was examined by western blot analysis using anti- v and 3 antibodies purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) and Cell Signaling Technology, Inc. (Danvers, MA), respectively. The expression of -actin as a control was also examined with its monoclonal antibody purchased from Santa Cruz Biotechnology, Inc (Santa Cruz, CA). The cells were cultured in DMEM supplemented with 10% fetal bovine serum until 90% confluence and then incubated with 20 g Fe/mL of WSPION and crgdfk WSPION for 2 h in the media. The cells were washed with PBS three times, scraped with rubber policeman and suspended in PBS (2 10 6 cells/ml). An aliquot of cells was analyzed for the content of iron bound to cells by ICP-MS (Agilent 7500A, Palo Alto, CA) and the rest of cells were imbedded in 1% agarose for MR imaging (Philips Healthcare, Achieva, 3.0 T). T 2 weighted MR imaging was performed in TR/TE = 5000/100 ms. The binding experiment was also done with Prussian blue staining to visualize the content of iron. MS-1 and MCF-7 cells were subconfluently cultured on the chamber slide and 20 g Fe/mL of WSPION and crgdfk WSPION for 2 h in the media. Cells were then washed three times with PBS, treated with 4% paraformaldehyde solution at 4 C for 30 min to fix the cells, and washed three times with PBS again. A 1:1 mixture of 5% potassium ferrocyanide (II) trihydrate solution and 5% HCl was added to each well and the cells were incubated at room temperature for 1 h before being counterstained with nuclear fast red for 5 min. Each well was then washed three times with PBS and analyzed by light microscopy. 2.3. Cytotoxicity of crgdfk WSPION The in vitro cytotoxic effect of crgdfk WSPION was tested in HT1080 (Human fibrosarcoma), MCF-7 and MS-1 cell lines using a 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay kit (R&D Systems, Minneapolis, MN, USA). Briefly, subconfluent cells in a 96-well plate were incubated with various concentrations (0.1 1 M) of crgdfk WSPION at 37 C for 12 h. As a cytotoxic control, cell lysis buffer (RIPA buffer) was also added with serial dilution. An MTT reagent was added to the culture media, and the cells were left at 37 C for 3 h. The cells were treated with detergent solution and were incubated at 20 C for 4 h in the dark. The absorbance of the well was measured at 570 nm with a reference wavelength of 650 nm in a plate reader (Spectra Max M2; Molecular Devices, Sunnyvale, CA, USA). Chart 1. Molecular structure of apeg and enpeg crgdfk. 3. Results and discussion Scheme 1 illustrates the stepwise surface modification process for WSPIONs and crgdfk WSPIONs. (1) The initial SPION OA (I) with Fe-rich surface was treated with 10 eq. of MHA, replacing OA at appropriately scattered places through Fe S linkage. (2) The next step was to make a simple conjugation to incoming functional molecule, apeg for WSPIONs and enpeg crgdfk for crgdfk WSPIONs (see Chart 1 for the structure), through amide linkage. Here, we assumed that most of the anchored MHA ligands were conjugated to apeg or enpeg crgdfk since the conjugation is an easy working reaction and we used 3 eq. of activating agent and apeg or enpeg crgdfk per MHA molecule. (3) Finally, the SPIONs were treated with MPA to substitute OA remaining on the LSPION surface, yielding aggregation-free WSPIONs and crgdfk WSPIONs. In this report, we have chosen cyclic RGD peptide unit as a targeting molecule instead of folate unit of the previous lipophilic SPION [18], since cyclic RGD peptide is well known as a specific marker of integrin v 3 which is rich in fibrinous cells and tumor tissues undergoing angiogenesis [19,20]. In step (1), roughly 10 eq. of MHA per SPION ( 8.5 nm) was chosen as a compromising point considering inter-nanoparticle magnetic interaction and our previous report [17] for QDs ( 6 nm), which showed the aggregation-free upper limit 15 eq. of MUA. The double bond in the middle of OA backbone has been known to be saturated during heating for SPION OA synthesis [26]. Thiolate anchoring occurred nicely when the SPION surface was Fe-rich [18] whereas there seemed to be no anchoring when its surface was oxidized. The surface oxidation can be indirectly judged from the following conjugation step since there is no conjugation to incoming amine-functionalized molecules in case of the oxidized SPIONs due to the lack of mediating MHA molecule. In step (3), like the cases of QDs [17] and gold nanoparticles [27], the mercapto-functionalized short-chain hydrophilic molecules readily substituted OA ligands, whereas they did not substitute the mercapto-functionalized long-chain hydrocarbon ligands on the Fe-rich SPION surface because of their (MPA here) weaker coordinating power. It was indirectly evidenced through a ligand exchange test as shown in Supplementary data. Here, for a comparison, we have used dodecanthiol (DDT) instead of MHA since fully MHA-protected SPION MHA was seriously aggregated and not dispersible in any solvent due to multi-chelating hydrogen bonding. The TEM images in Figure S2 indicate replacement of OA by DDT from slightly closer distances between SPION DDTs ( 2.0 nm) in b than SPION OAs ( 2.7 nm) in a, however, no replacement of DDT

178 M.-I. Yoo et al. / Colloids and Surfaces B: Biointerfaces 97 (2012) 175 181 Fig. 2. Hysteresis curve and photo image (inset) of crgdfk WSPION. Fig. 1. FT-IR spectra of SPIONs according to Scheme 1. (a) I: SPION OA, (b) II: SPION( OA) ex( MHA) 10, (c) III: SPION( OA) ex( MHA enpeg crgdfk) 10, (d) IV: SPION( MPA) ex( MHA enpeg crgdfk) 10, (e) enpeg crgdfk. by MPA from little change in inter-particle distances in c. The photo images in Figure S3 display the dispersion property of the SPIONs according to the ligand exchange test. The hydrophobic SPION OA and SPION DDT are dispersible in toluene whereas the hydrophilic SPION MPA should be dispersible in water if SPION DDT has been converted to SPION MPA. However, the SPION DDT has not transferred to water layer even after treatment with MPA. On the other hand, it has been known that Fe-rich SPION OA can be directly converted to SPION MPA by treatment with MPA [24]. Therefore, it is suggested that the mercapto-functionalized short-chain hydrophilic molecules readily substitute OA ligands whereas they cannot substitute the mercapto-functionalized long-chain hydrocarbon ligands on the Fe-rich SPION surface like the cases of QDs [17] and gold nanoparticles [27], since the longer chain provides the higher coordinating power to its terminal thiol group through higher electron density. Fig. 1 shows the FT-IR spectra of SPIONs according to the process of Scheme 1. Spectrum a for SPION OA (I) indicates the existence of iron oxide with oleate, which shows C H stretching bands just below 3000 cm 1, carbonyl band at 1702 cm 1, CH 2 and CH 3 wagging bands at 1465 cm 1 and 1376 cm 1, iron carboxylate bands around 1434 cm 1, C O stretching band at 1115 cm 1 and Fe O band around 603 cm 1 [24]. Spectrum b for scatteredly MHA substituted SPION OA (II) shows increased carboxylate bands around 1425 cm 1 and broad O H stretching band from H-bonded carboxylic acid around 3436 cm 1. The Fe O band at around 605 cm 1 has been relatively increased since more OA has been washed out during centrifugation for purification. Spectrum c displays the conjugation of enpeg crgdfk to SPION. The bands around 1026 cm 1 and 1658 cm 1 indicate C O stretching modes from PEG and amide C O bands from crgdfk, respectively. As a reference, the FT- IR spectrum of enpeg crgdfk is shown in e. Spectrum d shows Fig. 3. (a) TEM images of SPIONs according to Scheme 1 and (b) hydrodynamic diameter distribution of WSPION (A) and crgdfk WSPION (B).

M.-I. Yoo et al. / Colloids and Surfaces B: Biointerfaces 97 (2012) 175 181 179 Fig. 4. (a) Western blot analysis, (b) T 2-weighed MR images, (c) relative intensity of T 2 image, (d) SPION uptake, (e) Prussian blue staining of MS-1 and MCF-7 cells after culture with no SPION ( ), WSPION (A) and crgdfk WSPION (B). *P < 0.05, n = 3 for (c) and (d). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.) decrease of C H stretching band just below 3000 cm 1 and increase of H-bonded carboxylic acid O H bands around 3440 cm 1 and H- bonded C O bands at 1644 cm 1, due to substitution of remaining OA by MPA. Fig. 2 shows the hysteresis curve of crgdfk WSPION, including a photo image of its solution as an inset. The coercivity and remanence values are not discernible at room temperature, showing superparamagnetic behavior. The saturation magnetization value 40 emu/g was slightly lower than that (42 emu/g) [12] for 11 nm maghemite nanoparticles with OA surfactants as expected from their size 8.5 nm. This lower value compared to the bulk maghemite (73.5 emu/g) [12] is likely due to the mass of organic molecules bonded to the SPION surface, whereas the bulk maghemite does not have any organic molecules. It implicates that the magnetic property has been preserved after all the surface engineering process. The crgdfk WSPION solution was a dark brown homogeneous solution and the crgdfk WSPION solid was collectable by an external magnet. Although it precipitated after long time, it recovered its original dispersible property by simple shaking.

180 M.-I. Yoo et al. / Colloids and Surfaces B: Biointerfaces 97 (2012) 175 181 Fig. 3(a) shows the TEM images of each SPION according to the process of Scheme 1. The inter-particle distances in the TEM image I indicate the spaces occupied by OA on the SPION. When some of the OA was substituted by MHA, there could be weak H- bonding attractions between the particles through carboxylic acid groups. However, the SPIONs are still separated by the organic molecules in all the solutions up to III and IV, which correspond to enpeg crgdfk conjugated lipophilic and water-dispersible SPIONs, respectively. Fig. 3(b) shows the hydrodynamic diameter distribution of WSPION (A) and crgdfk WSPION (B) solution. They show quite narrow size distribution and the average hydrodynamic diameter of A and B were 9.9 and 9.4 nm, respectively, implicating that they are aggregation-free in aqueous solution. The zeta potential of WSPION was 19.60 ± 0.92 mev and explains its waterdispersible property. MS-1 and MCF-7 cells were chosen for cellular uptake test since MS-1 cells highly express v 3 whereas MCF-7 cells express relatively low level of v 3. Fig. 4(a) shows western blot analysis using anti- v and 3 antibodies, and -actin as a control. Both cells showed relatively low level of v expression, although MCF-7 cells showed slightly higher v level than MS-1 cells. On the other hand, MS-1 cells showed high level of 3 expression whereas MCF-7 cells showed almost undetectable level of 3 expression. Fig. 4(b) shows the T 2 -weighed MR images of both cells cultured in the SPION-free (, control), A, and B solutions. As expected, the darkness of T 2 image contrast was B > A > control in both MS-1 and MCF-7 cells. However, MS-1 cells showed much lower T 2 intensity than MCF-7 cells as expected from the western blot analysis. The relative intensity of T 2 image is shown in Fig. 4(c). The relative intensity of T 2 image was analyzed in the software, Image J. The intensity of signal decrease was B > A > control in both MS-1 and MCF-7 cells and the difference of signal decrease was much higher in MS-1 cells than MCF-7 cells as expected. The SPION uptake by the cells was analyzed by the Fe content per cell and displayed in Fig. 4(d). The SPION uptake from B was more than twice of that from A in case of MS-1 cells, whereas it showed only slight difference in case of MCF-7 cells. The SPION uptake from B was much higher in MS-1 cells than MCF-7 cells. The Fe content of cells cultured without SPION was almost undetectable in both cells. The cells cultured in SPION-free, A, and B solutions were visualized by Prussian blue staining and shown in Fig. 4(e). The blue spots were rarely observed in the both cells cultured without SPION. In case of MS-1 cells, the prominent blue spots were observed when the cells were cultured with B whereas the blue spots were greatly reduced when cultured with A. In case of MCF-7 cells, there was only a slight difference in the blue spots with a slightly higher intensity in B than A. All these results confirm that the cellular uptake of B was mediated by integrin v 3 binding. The cytotoxic effect of crgdfk WSPION was investigated with MTT assay in HT1080 (human fibrosarcoma cells), MCF-7, and MS- 1 cells. The cell lysis buffer (RIPA) was used as a control of cell killing. The cytotoxic effect of crgdfk WSPION on the tested cells was minimal as shown in Fig. 5. The SPION did not show any significant cytotoxicity to the three kinds of cells in the tested range of concentration. In summary, we have prepared ultrasmall (HD < 10 nm) nontoxic aggregation-free water-dispersible scatteredly crgdfk (specific marker of integrin v 3 )-conjugated SPIONs starting from oleate-protected SPIONs. The preparation was done stepwise by substitution of some oleates of SPION OA for mercaptohexadecanoic acids and then, conjugation to crgdfk and finally, conversion to water-dispersible ones by substitution of remaining oleate surfactants for mercaptopropionic acids. The PEGconjugated SPIONs without crgdfk have been prepared for a control experiment. The crgdfk-conjugated SPIONs showed higher cellular uptake to MS-1 cells than MCF-7 cells and also, showed Fig. 5. MTT assay of HT1080 (, ), MS-1 (, ), and MCF-7 (, ) cells with RIPA (,, ) and crgdfk WSPION (,, ). drastically higher uptake to MS-1 cells than the control SPIONs, confirming integrin v 3 -mediated cellular uptake of the scatteredly crgdfk-conjugated SPIONs. Acknowledgment This research was supported by Future Key Technology Program of Korea Institute of Science and Technology and by Future-based Technology Development Program (Green Nano Technology Development Program) through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (grant number 2011-0030204). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.colsurfb.2012.04.009. References [1] Y. Zhang, N. Kohler, M. Zhang, Biomaterials 23 (2002) 1553. [2] J.H. Lee, Y.M. Huh, Y. Jun, J. Seo, J. Jang, H.T. Song, S. Kim, E.J. Cho, H.G. Yoon, J.S. Suh, J. Cheon, Nat. Med. 13 (2007) 95. [3] Y.W. Jun, Y.M. Huh, J.S. Choi, J.H. Lee, H.T. Song, S. Kim, S. Yoon, K.S. Kim, J.S. Shin, J.S. Suh, J. Cheon, J. Am. Chem. Soc. 127 (2005) 5732. [4] J. Qin, S. Laurent, Y.S. Jo, A. Roch, M. Mikhaylova, Z.M. Bhujwalla, R.N. Muller, M. Muhammed, Adv. Mater. 19 (2007) 1874. [5] S. Rasaneh, H. Rajabi, M.H. Babaei, S. Akhlaghpoor, J. Nanopart. Res. 13 (2011) 2285. [6] H.J. Chung, H. Lee, K.H. Bae, Y. Lee, J. Park, S.-W. Cho, J.Y. Hwang, H. Park, R. Langer, D. Anderson, T.G. Park, ACS Nano 5 (2011) 4329. [7] J. Xie, K. Chen, H.-Y. Lee, C. Xu, A.R. Hsu, S. Peng, X. Chen, S. Sun, J. Am. Chem. Soc. 130 (2008) 7542. [8] S.M. Moghimi, A.C. Hunter, J.C. Murray, Pharm. Rev. 53 (2001) 283. [9] T. Neuberger, B. Schopf, H. Hofmann, M. Hofmann, B. von Rechenberg, J. Magn. Magn. Mater. 293 (2005) 483. [10] H.S. Choi, W. Liu, P. Misra, E. Tanaka, J.P. Zimmer, B.I. Ipe, M.G. Bawendi, J.V. Frangioni, Nat. Biotechnol. 25 (2007) 1165. [11] U.I. Tromsdorf, O.T. Bruns, S.C. Salmen, U. Beisiegel, H. Weller, Nano Lett. 9 (2009) 4434. [12] K. Woo, J. Hong, S. Choi, H.-W. Lee, J.-P. Ahn, C.S. Kim, S.W. Lee, Chem. Mater. 16 (2004) 2814. [13] E. Schellenberger, J. Schnorr, C. Reutelingsperger, L. Ungethum, W. Meyer, M. Taupitz, B. Hamm, Small 4 (2008) 225. [14] J. Ge, Y. Hu, M. Biasini, W.P. Beyermann, Y. Yin, Angew. Chem. Int. Ed. 46 (2007) 4342. [15] H. Ai, C. Flask, B. Weinverg, X.-T. Shuai, M.D. Pagel, D. Farrell, J. Duerk, J. Gao, Adv. Mater. 17 (2005) 1949. [16] H.-B. Xia, J. Yi, P.-S. Foo, B. Liu, Chem. Mater. 19 (2007) 4087. [17] J. Moon, K.-S. Choi, B. Kim, K.-H. Yoon, T.-Y. Seong, K. Woo, J. Phys. Chem. C 113 (2009) 7114.

M.-I. Yoo et al. / Colloids and Surfaces B: Biointerfaces 97 (2012) 175 181 181 [18] K. Woo, J. Moon, K.-S. Choi, T.-Y. Seong, K.-H. Yoon, J. Magn. Magn. Mater. 321 (2009) 1610. [19] Z.-F. Su, G. Liu, S. Gupta, Z. Zhu, M. Rusckowski, D. Hnatowich, Bioconjug. Chem. 13 (2002) 561. [20] Z.B. Li, W.B. Cai, Q.Z. Cao, K. Chen, Z.H. Wu, L.N. He, X.Y. Chen, J. Nucl. Med. 48 (2007) 1162. [21] M.D. Bentley, M.J. Bossard, K.W. Burton, T.S. Viegas, Mod. Biopharm. 4 (2005) 1393. [22] A.M. Derfus, W.C. Chan, S.N. Bhatia, Adv. Mater. 16 (2004) 961. [23] Since the number of OA molecules on SPION-OA is quite variable according to the washing condition, it is not certain how many OAs are remained after ten MHAs replace ten OAs on SPION I. Knowing that there are large excess amount of OA molecules per SPION-OA by thermal analysis, we assumed that there are still large excess amount of OAs on the SPION II in Scheme 1. So, ten MHA-anchored and then, enpeg crgdfk-conjugated or apeg-conjugated lipophilic SPION III was designated respectively as SPION( OA)ex( MHA enpeg crgdfk)10 and SPION( OA)ex( MHA apeg)10. [24] Y.T. Lee, K. Woo, K.-S. Choi, IEEE Trans. Nanotechnol. 7 (2008) 111. [25] W. Cai, D.-W. Shin, K. Chen, O. Gheysens, Q. Cao, S.X. Wang, S.S. Gambhir, X. Shen, Nano Lett. 6 (2006) 669. [26] A.L. Willis, N.J. Turro, S. O Brien, Chem. Mater. 17 (2005) 5970. [27] R.S. Ingram, M.J. Hostetler, R.W. Murray, J. Am. Chem. Soc. 119 (1997) 9175.