Folate receptor targeted, carboxymethyl chitosan functionalized iron oxide nanoparticles: a novel ultradispersed nanoconjugates for bimodal imaging

Abstract

This article delineates the design and synthesis of a novel, bio-functionalized, magneto-fluorescent multifunctional nanoparticles suitable for cancer-specific targeting, detection and imaging. Biocompatible, hydrophilic, magneto-fluorescent nanoparticles with surface-pendant amine, carboxyl and aldehyde groups were designed using o-carboxymethyl chitosan (OCMC). The free amine groups of OCMC stabilized magnetite nanoparticles on the surface allow for the covalent attachment of a fluorescent dye such as rhodamine isothiocyanate (RITC) with the aim to develop a magneto-fluorescent nanoprobe for optical imaging. In order to impart specific cancer cell targeting properties, folic acid and its aminated derivative was conjugated onto these magneto-fluorescent nanoparticles using different pendant groups (–NH₂, –COOH, –CHO). These newly synthesized iron-oxide folate nanoconjugates (FA-RITC-OCMC-SPIONs) showed excellent dispersibility, biocompatibility and good hydrodynamic sizes under physiological conditions which were extensively studied by a variety of complementary techniques. The cellular internalization efficacy of these folate-targeted and its non-targeted counterparts were studied using a folate-overexpressed (HeLa) and a normal (L929 fibroblast) cells by fluorescence microscopy and magnetically activated cell sorting (MACS). Cell-uptake behaviors of nanoparticles clearly demonstrate that cancer cells over-expressing the human folate receptor internalized a higher level of these nanoparticle–folate conjugates than normal cells. These folate targeted nanoparticles possess specific magnetic properties in the presence of an external magnetic field and the potential of these nanoconjugates as T₂-weighted negative contrast MR imaging agent were evaluated in folate-overexpressed HeLa and normal L929 fibroblast cells.

---

1. Introduction

Superparamagnetic iron oxide nanoparticles (SPIONs) with multifunctional properties have been the subject for intensive research during the past three decades due to their excellent magnetism, improved biocompatibility, increased contrast enhancement and long circulation time in blood.^1–8 In the nano-realm, the unique properties of these multifunctional SPIONs endow them in great diagnostic and therapeutic applications in the fields of magnetic resonance imaging (MRI), tissue repair, immunoassay, cell separation, hyperthermia and magnetically controlled drug delivery.^9–14 Recently, more and more attention has been paid to fabricating a multifunctional nanoscaled particulate which possesses the combined properties of cellular targeting, optical and non invasive MR imaging in a single entity.

Many groups have published several research articles about the strategic fabrication of a multifunctional nanoparticulate system with combined targeting and detection properties.^15–17

For the development of a novel magnetic platform with multifunctional properties, appropriate surface engineering of SPIONs is of prime concern. For this purpose, different biocompatible, nonimmunogenic polymeric agents (polyethylene glycol, dextran, chitosan, synthetic polymers) are used which can not only prevent biofouling of magnetic nanoparticles in blood plasma via improving their blood circulation time, but also decorate the SPION surface with active functional groups for further bioconjugation.^18–25

In recent years, o-carboxymethyl chitosan has drawn significant attention as a surface coating agent due to its biocompatibility, biodegradability and amphiphilicity.²⁶ The active amino and carboxyl groups on the OCMC chains provide us great opportunities to intergrate OCMC stabilized magnetite nanoparticles with tailored physicochemical and biophysical properties.²⁷ Of note the synthesis of o-carboxymethyl chitosan stabilized iron oxide nanoparticles via physical adsorption and chemical ligation was previously reported by several groups.^28,29 However to the best of our knowledge, there are no reports related to the cancer cell targeting and in vitro MRI applications of such OCMC-stabilized magnetite nanoparticles (OCMC-SPIONs).

To fabricate a multifunctional OCMC-SPION based platforms with cancer cell specific targeting ability, targeting ligands should be attached on the surface of SPIONs to trigger active delivery to specific cancer tissues.³⁰ Among the different strategies for receptor mediated delivery of nanoparticles, folic acid (FA), a high affinity ligand to folate receptors (FRs), is known as a promising targeting agent for folate receptor mediated tumor cell specific nanoparticle delivery, because FRs are overexpressed in many human cancer cells providing a distinguisable marker from normal cells. FA has received promising consideration due to its nonimmunogenicity, high stability, low cost and its faster internalization kinetics through cellular membrane.^31–33

As for the fabrication of combined dual modality molecular imaging of SPIONs is concerned, no researchers to date have focused on the development of a series of folate-receptor targeted magnetic nanoparticles using OCMC as the surface anchoring agent. As the folic acid labeled magnetic nanoparticles could be internalized into FR bearing cells as free folic acid,³⁴ so it is a desirable strategy to modify the nanoparticulate system with folic acid to target specific cancer cells via receptor mediated endocytosis. The current research effort is therefore dedicated to the designed fabrication of a folate receptor targeted, highly hydrophilic, biocompatible, magneto-fluorescent nanoparticles with amine, carboxyl and aldehyde functionalities, to be later used as a bio-conjugating precursor or platform for eclectic in vitro or in vivo applications. Herein, we propose a synthetic approach for the development of multifunctional nanomedical platform for combined targeting and dual mode molecular imaging using OCMC-SPION-NH₂ as the bioconjugating precursor. Our established nanoparticle platform is composed of three components. (1) A fluorescent dye, RITC is coupled with amine groups of OCMC-SPION-NH₂, followed by surface modification with different functional groups (–COOH, –CHO, –NH₂) which permits fluorescence imaging. (2) These highly hydrophilic and biocompatible magneto-fluorescent nanoparticles with active functionalities (–COOH, –CHO, –NH₂) could be used as the core material to allow T₂-weighted magnetic resonance (MR) contrast enhancement. (3) These functionalized magneto-fluorescent supports are further conjugated to folic acid to prepare a series of highly stable cancer cell targeted nanoconjugates for cancer cell targeting applications. These nano-systems were analyzed for their size, surface charge, surface chemistry, composition, magnetic properties and colloidal stability by XRD, TEM, DLS, FT-IR, UV-Vis, XPS, Zeta and VSM. In vitro cellular uptake by magnetically activated cell sorting (MACS), fluorescence microscopy and MRI studies using HeLa and a normal L929 cell line demonstrate that these folate targeted nanoconjugates can preferentially target the HeLa cells which overexpressed folic acid receptors. These results clearly signify that our FA-RITC-OCMC-SPIONs could be utilized as a potential candidate for bimodal MRI and fluorescence imaging.

2. Materials and methods

2.1 Materials

FeCl₃ and FeSO₄ were obtained from Merck. Chitosan (Medium molecular weight), o-chloroacetic acid, succinic anhydride, 2,2′-(ethylenedioxy)-bis-(ethylamine) (EDBE), folic acid (FA), di-tert-butyldicarbonate (BoC₂O), dicyclohexyl carbodiimide (DCC), trifluoroacetic acid (TFA), N-hydroxysuccinimide (NHS), 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (EDC), sodium cyanoborohydride [NaBH₃CN], trinitrobenzene sulfonic acid (TNBS) and rhodamine isothiocyanate (RITC) were obtained from Aldrich Chemicals, USA. MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium], 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI), and agarose were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Commercially available dimethyl sulfoxide (DMSO) and N,N-dimethyl formamide (DMF) was purified by vacuum distillation. Pyridine was purified by distillation over KOH. 2-Propanol was used as received from Merck (Germany).

2.2 Synthesis of γ-N-{2-[2-(2-aminoethoxy) ethoxy]ethyl} folic acid (FA-EDBE)

FA-EDBE was synthesized according to our previously reported protocol.³⁵ The steps involved in the synthesis of FA-EDBE have been outlined in Scheme 1. In brief, to a stirred solution of folic acid (0.75 mmol) in 20 ml anhydrous DMSO and pyridine (8 ml), dicyclohexyl carbodiimide (2 mmol), followed by tert-butyl N-{2-[2-(2-aminoethoxy)ethoxy]ethyl}-carbamate (0.87 mmol) was added. The reaction mixture was stirred for about 18 h at room temperature under argon atmosphere and the resulting precipitate was filtered. The filtrate was poured dropwise into a vigorously stirred cold solution of Et₂O at 0 °C. The yellow precipitate thus obtained was collected and washed several times with cold Et₂O to remove traces of DMSO. The solvent was then removed under reduced pressure to afford 2.2 as a yellow solid. Trifluoroacetic acid (TFA) (2 ml) was added to 2.2 (0.65 mmol) at room temperature and allowed to stir. After 2 h, TFA was removed under reduced pressure and the resulting residue was dissolved in DMF. Then pyridine was added to initiate the formation of an orange-yellow precipitate. After complete precipitation, the resulting solid was washed with Et₂O and dried to afford 2.3. ¹H NMR (DMSO-d₆, 400 MHz): δ 2.5 (br s, β and γ –CH₂ groups superimposed, 4H), 2.9–3.5 (m, –CH₂– of EDBE, 12H), 4.3 (br s, NH₂, 2H), 4.5 (d, –NH–, 1H), 4 d at 6.5, 6.7, 7.6, 7.7 (aromatic ring), 6.9 (s, aromatic H of pteridine), 8.7 (s, OH, 1H).

Scheme 1 Synthesis of amine-, carboxyl- and aldehyde-functionalized magneto-fluorescent nanoparticles using anhydride and imine chemistries, respectively.

2.3 Synthesis of o-carboxymethyl chitosan (OCMC)

o-Carboxymethyl chitosan (OCMC) was prepared as reported previously by Chen and Park.³⁶ The degree of substitution and the yield are strongly dependent on the reaction time, temperature. Chitosan (2 g) was suspended in 40% (w/w) aqueous NaOH (15 ml) and kept at 0 °C overnight. The cold alkaline solution of chitosan was transferred to isopropanol (60 ml), and monochloroacetic acid (6 g) in isopropanol (2 ml) was slowly added to the solution over 30 min. This mixture was stirred at room temperature for 12 h. Finally HCl was added to the reaction mixture to adjust the pH to 7.0. The OCMC was filtered and washed with anhydrous ethanol and the product was vacuum-dried at room temperature. The products were dissolved in dilute ammonia and centrifuged to separate the unreacted chitosan. The OCMC was precipitated by ethanol from the water-soluble portion and vacuum-dried.

2.4 Synthesis of o-carboxymethyl chitosan stabilized iron oxide nanoparticles (OCMC-SPION-NH₂)

Superparamagnetic magnetite nanoparticles were prepared by controlled chemical co-precipitation of Fe²⁺ and Fe³⁺ (1 : 2 ratio) from ammoniacal medium at 80 °C under argon atmosphere. For surface modification of the as prepared magnetite nanoparticles by amine groups, 100 mg of the magnetic nanoparticles was added to 20 mL of phosphate buffer (0.2 mol/L Na₂HPO₄-NaH₂PO₄, pH 6.0) containing 100 mg of EDC, and then the reaction mixture was ultrasonicated for 1h. After that, 50 mL of carboxymethylated chitosan solution (2 mg/mL in phosphate buffer) was added, and the reaction mixture was ultrasonicated using a high intensity ultrasonic probe operated at 20 KHz with 100 W power for an hour. The resulting suspension was then stirred vigorously on a magnetic stirrer at 60 °C for 12h. The particles were then isolated with a rare-earth magnet, washed 3 times with deionized water and resuspended in PBS. These amine functionalized nanoparticles (1) were divided into two parts. Part A was labeled with RITC and subsequent functionalization steps as well as folate conjugation were performed using these magneto-fluorescent nanoparticles. Part B was kept unlabelled and further functionalization (including carboxyl modification as well as folate conjugation reactions) were performed with this unlabelled part, using the same protocol. The different nanoparticles preparation, obtained part B was freeze dried and preserved for various physicochemical characterizations.

2.5 Synthesis of amine functionalized magneto-fluorescent nanoparticles and its folate targeted counterpart (1 and 4)

1 mg of rhodamine isothiocyanate (RITC) dissolved in 1ml of DMSO-H₂O mixture was added dropwise to an aqueous suspension of amine functionalized magnetite nanoparticles OCMC-SPION-NH₂(1) at pH 8. The resulting suspension was sonicated for an hour in the dark. Particles were recovered by magnetic decantation and washed thoroughly with deionized water. In this way, OCMC-SPION-NH₂ magneto-fluorescent nanoparticles were prepared (Scheme 2). To conjugate the OCMC-SPION-NH₂ (1) with folic acid (FA), FA was dissolved in 10 ml of DMSO due to the sparingly solubility of folic acid in aqueous medium. The resulting solution was then mixed with a solution of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (75 mM) and NHS (15 mM). The pH of the solution was then adjusted to 8 by dropwise addition of pyridine. An aqueous dispersion of 200 mg of aminated magnetite (1) was added to it and the reaction mixture was stirred overnight at 37 °C in dark. These folate conjugated aminated magneto-fluorescent nanoparticles (FA-OCMC-SPION-NH₂) (4) were again isolated with a rare earth magnet, washed 5 times with de-ionized water and re-dispersed in PBS.

Scheme 2 Synthesis of iron-oxide folate conjugates 4–6: magneto-fluorescent nanoparticles bearing different reactive handles (–NH₂, –CO₂H, –CH=O) are covalently conjugated with folic acid utilizing diverse conjugation strategies.

2.6 Synthesis of carboxyl functionalized magneto-fluorescent nanoparticles and its folate targeted counterpart (2 and 5)

For the carboxyl modification of amine functionalized magnetite nanoparticles (2), a freshly prepared solution of 100 mg of succinic anhydride in DMSO (5 ml) were added dropwise to the suspension of aminated magneto fluorescent nanoparticles (50 mg nanoparticles dispersed in 10 ml DMSO) with ultrasonication. A catalytic amount of triethylamine (distilled) was added to the resulting suspension and the reaction was continued for 24 h in the dark. The particles were finally recovered by magnetic concentration and washed with distilled water. To prepare folic acid modified carboxyl functionalized magnetofluorescent nanoparticles (FA-OCMC-SPION-COOH) (5), 50 mg of succinylated nanoparticles (OCMC-SPION-COOH) (2) was dispersed in 10 ml sol of EDC (mM) and the same amount of NHS (mM). The pH of the resulting solution was adjusted to 8 and kept in the dark for 2h. To the resulting suspension, FA-EDBE (100 mg dissolved in 10 ml of millipure water) was added dropwise. The suspension was agitated overnight in dark at 37 °C.

2.7 Synthesis of iron oxide folate conjugates (6)

For the synthesis of iron-oxide folate conjugate FA-OCMC-SPION-CHO (6), at first aminated magnetofluorescent nanoparticles (1) were converted into aldehyde (–CH=O) groups via imine formation of glutaraldehyde crosslinking treatment as described in earlier reports.^37,38 Finally FA-EDBE was immobilized onto the aldehyde functionalized nanoparticles through reaction of the reactive aldehyde groups with amine groups of FA-EDBE, followed by reductive amination with NaCNBH₃. In a typical procedure, about 1 ml of 1% (w/v) glutaraldehyde solution in PBS (pH ∼ 7.4) was added to 10 ml aqueous dispersion of aminated magnetite (2.12 mg/ml) and the suspension was incubated for 3h at 20 °C under shaking. Nanoparticles were recovered using a magnetic concentrator and washed with PBS to remove any excess of glutaraldehyde. A higher ratio of primary amine to glutaraldehyde was maintained to evade the amine crosslinking possiblity due to the presence of excess amount of glutaraldehyde. 25 mg of FA-EDBE dissolved in 5 ml of millipure water and this solution was added to an aqueous dispersion of glutaraldehyde activated nanoparticles under overnight stirring. The folate-functionalized nanoparticles were then washed with PBS and placed in 5 ml of NaCNBH₃ for 30 min. The nanoparticles were washed with PBS and resuspended in 5 ml of the same solution.

2.8 Characterizations

The biofunctionalized nanoparticles (1–6) were studied using a Thermo Nicolet Nexux FTIR model 870 spectrometer. The X-ray photoelectron spectroscopic (XPS) data was collected using an Al Kα excitation source in an ESCA-2000 Multilab apparatus (VG microtech). The determination of amino groups on the surface of (1–6) nanoparticles were performed using the 2,4,6-trinitrobenzenesulfonic acid (TNBS) method according to the procedure adapted by Edwards-Levy et al.³⁹ To determine the extent of folate conjugation on 4, 5 and 6, UV-Vis spectra of 4, 5 and 6 were recorded by spectrophotometric analysis of absorbance at 286 nm only and also cross-checked by determination of residual amine concentration via the TNBS assay. The phase analysis of the synthesized magnetite nanopowder was performed on an X'pert Pro Phillips X-ray diffractometer. High-resolution transmission electron microscopy (JEOL 3010, Japan) was employed to characterize the microstructure of the different nanoparticle preparations. The hydrodynamic size (HD) of the particle aggregates was measured by laser light scattering using a Brookhaven 90 Plus particle size analyzer. The surface charge of the nanoparticles was investigated through zeta potential measurements (Zetasizer 4, Malvern Instruments, UK). The cells cultivated for in vitro experiments were human cervix adenocarcinoma, HeLa and a normal mouse fibroblast L929 cell line, obtained from the National Centre for Cell Sciences (NCCS), Pune, India. All cell lines were cultured on Dulbecco modified Eagle's medium (DMEM) and minimal essential medium (MEM), respectively, with 10% fetal calf serum, 100 units ml⁻¹ penicillin, 100 μg ml⁻¹ streptomycin, 4 mM L-glutamine at 37 °C in a 5% CO₂ and 95% air humidified atmosphere.

2.9 Nanoparticle mediated cytotoxicity assay by MTT

For the MTT assay, 4 × 10⁵ HeLa and a normal fibroblast L929 cells were seeded into 96-well tissue-culture plates in complete media (total volume 180 μl) and kept for 18 h. After that, different nanoparticle (1–6) preparations (20 μL) were added to the cells at different concentrations and incubated for 4 h at 37 °C in a humidified incubator (HERA cell) maintained at 5% CO₂, and the cell viability was assessed by the 3-(4,5 dimethylthiazol)-2-diphenyltetrazoliumbromide (MTT, sigma).

2.10 Intracellular uptake studies of the nanoconjugates

Nanoparticle uptake by a folate receptor overexpressed HeLa cancer cell (positive control) and a normal L929 fibroblast cell line was preliminarily studied by magnetically activated cell sorting (MACS) in absence of folic acid. The effect of surface coating on the uptake of nanoparticles by different cell lines was evaluated by culturing these cells with different nanoparticle preparations in the absence of folic acid in the culture medium. After 4h of incubation, the cells were washed to remove free nanoparticles and cell sorting was performed with MACS. For folate receptor targeted cancer imaging, the cellular uptake of iron-oxide folate conjugate (4) and its non-targeted counterpart (1) were studied using HeLa (cancer) and a L929 (normal) cell line by fluorescence microscopy. For this, folate receptor overexpressing HeLa cells (positive control) were treated with different nanoparticle preparations (1 mg mL⁻¹) followed by incubation for 24 h. After that, these cells were smeared on a clean glass slide, fixed with 3.7% formaldehyde for 15 min, permeabilized with 0.1% Triton X-100, and stained with DAPI (1 mg mL⁻¹) for 5 min at 37 °C. The cells were then washed with PBS and examined by fluorescence microscopy (OlympusIX 70). To further evaluate the specificity of receptor mediated cellular targeting of the nanoparticle conjugates, the cellular uptake of 4 was performed using the normal mouse fibroblast cell line L929 as negative control cultivated under similar conditions in their respective medium.

2.11 In vitro cellular MR imaging studies of the nanoparticles

For the determination of iron content present in these nanoparticles, the iron present in the nanoparticle was completely extracted by dissolving the nanoparticles (10 mg) in HCl (10 mL, 30% v/v) for 2h at 50–60 °C. Ammonium persulfate (1.0 mg) was then added to oxidize the ferrous ions present in the above solution to ferric ions. Potassium thiocyanate (1 mL, 0.1 M) was added to this solution to form a blood-red iron thiocyanate complex. The iron concentration was determined by spectrophotometric measurements at 478 nm using a Shimadzu UV-1700 spectrophotometer. Samples for MR phantom imaging were prepared by suspending 10⁶ cells in low-melting 1% agarose gel (50 μL). Cell suspensions cultured with different nanoparticle concentrations (0.01–0.05 mg mL⁻¹) were loaded into 1.5 mL eppendorf tubes and allowed to solidify at 4 °C. Samples were then sealed with additional agarose to avoid air susceptibility artifacts. MRI was performed with a 1.5 T clinical MRI scanner (GE Medical systems, Milwaukee) using a prefabricated sample holder. A spin-echo multisection pulse sequence was selected from the GE Medical systems to acquire MR phantom images. A repetition time (TR) of 2100 ms and variable echo times (TE) of 42–110 ms were used. The spatial resolution parameters were set as follows: an acquisition matrix of 256 × 256, field of view of 240 × 240 mm², section thickness of 8 mm, and two averages. The MRI signal intensity (SI) was measured using the in-built software. T₂ values were obtained by plotting the SI of each sample over a range of TE values. T₂ relaxation times were then calculated by fitting a first-order exponential decay curve to the plot. The fitting equation can be expressed as SI = Ae^−TE/^T₂ + B, where A is the amplitude and B is the offset.

3. Results

3.1 Fabrication of folic acid modified magneto-fluorescent nanoparticles with surface-pendant amine, carboxyl and aldehyde groups

For the successful utilization of SPIONs as a targeted probe for magnetic resonance and fluorescence imaging, surface modification with targeting ligands and fluorescent dye on the nanoparticle system is of prime concern. Here OCMC was chosen as a surface modifying agent because of the abundance of amino and carboxyl groups in the polymer chain which lead us to develop an amine functionalized targeted nanoprobe. RITC was conjugated on the previously synthesized OCMC-SPION-NH₂ (1) via a thiourea linkage with these available amine functionalities of OCMC-SPION-NH₂ (1). These aminated magneto-fluorescent nanoparticles were further conjugated with a variety of small molecules (succinic anhydride and glutaraldehyde) through carboxyl, anhydride or imine chemistry to impart active carboxyl and aldehyde functionalities. Carboxylic acid modified magneto-fluorescent base OCMC-SPION-COOH (2) was produced by a simple ring-opening linker elongation reaction of the primary amine (–NH₂) functions of OCMC-SPION-NH₂(1) with succinic anhydride. It is a well recognized fact that an aldehyde functional group of glutaraldehyde reacts with an amino group of native OCMC to form a Schiff's base.⁴⁰ In our case, aldehyde (–CH=O) functionalized magnetite nanoparticles OCMC-SPION-CHO (3) were formed by treating the amine functionalized nanoparticles (1) with glutaraldehyde. The ratio of amine and aldehyde was maintained as 15 : 1 to avoid the possibility of crosslinking between these newly synthesized aldehyde functionalized nanoparticles before the addition of aminated derivative of folic acid.⁴¹

To synthesize FA-OCMC-SPION-NH₂ (4), the active NH₂ groups present on the surface of (1) were covalently conjugated with FA via nonselective activation of its carboxyl groups by the carbodiimide method.⁴² Between the possibility of generation of two structural isomers in which FA is linked either through the α-carboxyl or γ-carboxyl group of its glutamic acid moiety, the γ–carboxyl group isomer is precedented as the major isomer and retains a strong affinity towards its receptor, whereas its α-carboxyl derivatives are not recognized as readily by the cells.^43,44 For the grafting of FA on the surface of 2 and 3, an appropriate linker was necessary. We have made a bridge between the carboxyl groups of folate and OCMC-SPION-COOH (2) through an attachment of a hydrophilic spacer 2,2-(ethylenedioxy)-bis-ethylamine [EDBE] to folic acid using standard protocols.^35–45 FA-EDBE was immobilized onto the surface of OCMC-SPION-CHO (3) through Schiff's base formation, followed by the reductive amination with NaBH₃CN to afford folate-targeted aldehyde functionalized nanoparticles FA-OCMC-SPION-CHO (6).⁴⁶

3.2 Crystal structure, size, stability and magnetization of synthesized nanoparticles

The high resolution X-ray diffractogram pattern of 1 is shown in Fig. 1(A). It was observed that the d value corresponds to those of inverse spinel phase magnetite (Fe₃O₄) (JCPDS card no. 77-1545). The broadening of the diffraction bands represented the nanocrystalline nature of the synthesized powder. Crystallite size was evaluated from the XRD data using Debye-Scherrer equation, d = Kλ/βCosθ, where, d is the crystal thickness, K is Debye-Scherrer constant (0.89), λ is the X-ray wavelength (Co = 1.789 Å). The mean crystallite size was calculated to be around 10 nm. The morphology and size of the OCMC-SPION-NH₂ (1), and FA-OCMC-SPION-NH₂ (4) were investigated by TEM (Fig. 1(C) and (B)).

Fig. 1 A) XRD spectrum of 1. B) High resolution TEM image of 1. C) High resolution TEM image of 4. D) Particle size distribution histogram of 1. E) Single-particle HRTEM image and F) corresponding SAED pattern.

The TEM images of 1 and 4 displayed in Fig. 1(C) and (B) were found to be almost spherical in shape with a mean particle size of 12.5 and 13.7 nm with a standard deviation of 2.7 and 2.5 nm respectively. The selected-area diffraction (SAED) pattern indicated the polycrystalline nature of the embedded magnetite particles (1). The individual planes identified from the SAED pattern (Fig. 1(F)) correlated well with that of the XRD pattern of 1 (Fig. 1(A)). It was evident from the TEM micrographs that these SPIONs presented considerable dispersancy and remained non-aggregated for months.

Dynamic light scattering studies of these targeted and non-targeted nanoconjugates are shown in Fig. 2(A). It was observed that the conjugate 1 gives a stable, non-aggregated suspension with a mean diameter of 70 nm at pH 4. These particles showed a little agglomeration in higher of range pH due to the successive deprotonation and aggregation tendency of surface exposed NH₂ groups. The final conjugate after folic acid modification (4–6) presented a mean diameter of 115 ± 0.7 (PDI =0.24), 108 ± 0.5 (PDI = 0.26), 109 ± 0.8 (PDI = 0.28) nm respectively at physiological pH and these folate conjugated nanoparticles were extremely stable over a period of three months (Fig. 2(B)).

Fig. 2 A) Variation of particle sizes of nanoconjugates (1–6) against pH. B) Variation of particle size of iron-oxide folate nanoconjugates (1–6) against time. C) Variation of zeta potential of nanoconjugates (1–6) against pH. D) Magnetization curve of both folate-functionalized (4) and non-functionalized nanoparticles (1) at 300 K.

Zeta potential measurement was performed on these functionalized magnetic nanoparticles (1–6) to validate the surface modification of the pristine nanoparticles (Fig. 2(C)). The hydrodynamic diameter, polydispersity index (PDI) and zeta potential of these functionalized nanoparticles are given in Table 1. It is known that with an increase of zeta potential, the surface charge of the nanoparticle increases. The zeta potential of these nanoconjugates (1–3) after surface modification with amino, carboxyl and aldehyde groups were found to be of +39.2 ± 0.8, 4.6 ± 0.5 and 17.8 ± 0.5 mV respectively at pH 4.

Table 1 Hydrodynamic (HD) size, PDI and zeta potential of the folate-functionalized (4–6) and non-functionalized nanoparticles (4–6) at physiological pH

SPIONs	Size (nm)	PDI	mV	FA-SPIONs	Size (nm)	PDI	mV
1	100	0.221	+18.9	4	115	0.24	−21.5
2	97	0.202	−31.6	5	108	0.26	−35.5
3	103	0.235	−1.5	6	110	0.28	−16.6

---

Results also showed that with an increase of pH, the surface charge of the 1 decreased which was probably due to the deprotonation tendency of the surface exposed amino groups at higher range of pH. The zeta potential for 2 was found to be −43.6 mV respectively at pH 9 but its zeta potential value was very low in lower range of pH, which was probably due to the presence of protonated carboxyl groups in lower range of pH. The zeta potential observed for the nanoconjugates (4–6) got increasingly negative with rising pH, which is probably because of the progressive deprotonation of the FA carboxylic groups.

Retention of their favorable magnetic properties after coating with polymers is a prerequisite for the successful application of these surface modified SPIONs as a targeted contrast agents for MRI probe. The magnetic property of 1 was evaluated by using vibration sample magnetometry (VSM) at 300 K (Fig. 2(D)). The saturation magnetization curve of 1 showed no hysteresis and was completely reversible at 300 K. Neither coercivity nor remanence was observed, indicating its superparamagnetic behavior. The saturated magnetizations (Ms) of 1 were found to be 45 emu/g which is much smaller than the theoretical value of bulk SPIONs (92 emu/g). This decrease in the Ms value might be due to the decrease in particle size effect accompanied by an increase of surface area.³⁵ After conjugation with FA, a further loss of value was observed in 4 which might be attributed from the accumulation of nonmagnetic mass on the nanoparticle surface.

3.3 Surface chemistry and composition

The FTIR spectrum (Fig. 3(A)) of the H-form of OCMC-SPION-NH₂ showed not only a peak of Fe–O vibration at 585 cm⁻¹ but also a characteristic peak of carboxyl groups of OCMC-SPION-NH₂ at 1722 cm⁻¹ (–COOH). A broad band was observed at 3440 cm⁻¹ corresponding to the stretching vibration of the N–H groups of OCMC-SPION. A broad peak was observed at 1600 cm⁻¹ corresponding to the asymmetric stretching of COO⁻ groups. The bands at 1655, 1540 and 1625 cm⁻¹ anticipated the presence of functional groups of amide I (ν C=O of amide I), amide II (free N–H of amide II) and primary amine groups (–NH₃⁺). For, OCMC-SPION-COOH, intense peaks appeared at 2922 cm⁻¹ (stretching of –CH₂–), 1637 cm⁻¹ (ν C=O of Amide I) and 1555 cm⁻¹ (free N–H of Amide II), and a peak with decreased intensity was observed around 3440 cm⁻¹ (amino group characteristics).²⁷ The typical band for –CO₂H group (1722 cm⁻¹) was intensified which evidences the successful modification of the surface amino groups with succinic anhydride.

Fig. 3 A) FT-IR spectra of a) SPION b) SPION-NH₂ (1) c) SPION-COOH (2). B) FT-IR spectra of a) pure folic acid b) FA-SPION-NH₂ (4) c) FA-SPION-COOH (5) c) FA-SPION-CHO(6).

The IR spectrum for pure folic acid was characterized by a number of characteristic bands occurring at 3543, 3416, 3324, 2959, 2924, 2844, 1694, 1640, 1605, 1484 and 1411 cm⁻¹.⁴⁷ These characteristic FT-IR absorption peaks of folic acid with a little shift were observed in the spectrum of folic acid-modified magnetite nanoconjugates (4–6) indicating successful surface modification of these nanoconjugates with folic acid. The most important characteristic FT-IR absorption peaks of folic acid at 1650 (–CONH amide II band), 1560 (–NH amide band II), and 1484 cm⁻¹ (–NH₂ of folic acid) were observed in these folate modified nanoconjugates.⁴⁸

The FTIR spectrum of 4 (Fig. 3(B)) showed increased absorbance at 1648 cm⁻¹ (bonded C=O of –CONH amide band II) and a new band appears at 1558 cm⁻¹ (–NH amine band II), generated from the amide bands within the FA structure as well as the amide bonding between FA and amine-functionalized nanoparticles. Similarly, following the immobilization of FA-EDBE on the carboxyl-terminated nanoparticles, the amide carbonyl bands and the C=C bands of folic acid at 1649, 1559 cm⁻¹ and 1620 cm⁻¹ were intensified in the spectrum of 5. Although characteristic bands of FA were present in the FTIR spectrum of 6, unlike 4 and 5 no absorption band could be visualized around 1550–1580 cm⁻¹ due to the absence of N–H amide bands in 6. Instead, a medium intensity band corresponding to –C–N stretching vibration appeared around 1278 cm⁻¹, indicating the immobilization of FA-EDBE on an OCMC-functionalized nanoparticle surface through a stable secondary amine linkage.

We have further performed XPS (shown in Fig. 4) to understand the chemical bonding on the surface of nanoconjugates, as it is recognized as a quantitative surface elemental analysis and chemical state information. The high resolution N1S spectrum of OCMC-SPION-NH₂(1) revealed a broad shoulder at 398.6 eV, corresponding to the free amine groups on the surface. The O1S spectrum of 1 displayed three peaks at 531.1, 531.5 and 530.15 eV corresponding to oxygen being present in three different environments as –C–O, –O–H and Fe–O–C in compound 1. Taking 284.2 eV as standard for bulk C1S, the broad shoulder could be fitted into three peaks 285.18, 284.73 and 287.54 eV, which could be attributed to C–O, C–C and –NH–C=O groups in compound 1. The C 1S peak obtained after successful modification with succinic anhydride could be deconvoluted into four chemical environments such as –COOH, –NH–C=O, –C–C and –C–O respectively. The broad peak 401.3 eV was observed due to the amide linkage (–NH–C=O) between the succinic anhydride and OCMC-SPION-NH₂ (1). The high-resolution scans for C 1S of FA-OCMC-SPION-NH₂ (4) showed a broad shoulder in between 280 and 291 eV. Taking bulk C 1S at 285 eV as standard, the broad shoulder could be deconvoluted into three peaks at 284.2, 286.3 and 288.6 eV, which could be attributed to –C–C–, –NH–C=O and –CO₂H groups, respectively. The broad peak for N 1S could be deconvoluted into two peaks at 398.6 eV and 401.3 eV, corresponding to the free –NH₂ groups (native to the pteridine rings in FA) and amide (–NH–C=O) bonding within the FA structure and also the amide linkage between FA and nanoparticle surface. The Fe 2p doublet with binding energy values of 710 and 725 eV implied the presence of Fe–O bonds, typical for magnetite.⁴⁹ The unaltered Fe 2p spectrum, following FA immobilization, confirmed a clear proof that surface modification has no deleterious effect on the composition of the support material.

Fig. 4 High-resolution O 1s, C 1s, Fe 2p, and N 1s X-ray photoelectron spectra of 1, 2 and 4. A) O 1s spectrum of aminated magnetite 1. B) N 1s spectrum of aminated magnetite 1. C) Fe 2p spectrum of aminated magnetite 1. D) N 1s of 2 after modification with succinic anhydride. E) C 1s spectra of both 2 and 4. F) N 1s spectrum of FA-OCMC-SPION (4).

To further establish the successful derivatization by o-carboxymethyl chitosan, amine density present on the surface of incubation of 1–3 was checked by TNBS assay. The procedure consisted of the incubation of the material with an excess of TNBS and the back titration of the unreacted amount of the reagent. The amine densities calculated for 1, 2, 3 have been summarized in Table 2.

Table 2 Concentration and number of amino, carboxyl and aldehyde groups present on the nanoparticles.^a

1	2	3	4	5
a 1- SPION Modifications, 2- Functional groups, 3- Amine density (μmol/mg), 4- Concentrations, 5- No of groups.
Amine	–NH2	0.6	0.6	983
Carboxyl	–COOH	0.11	0.49	803
Aldehyde	–CHO	0.12	0.48	787

---

Since each particle contains magnetite (d = 5.214 g cm⁻³) with an average radius of 5 nm, the average mass of Fe₃O₄ particle is 2.721 × 10⁻¹⁸ respectively. Hence, 1 mg of Fe₃O₄ contained 3.675 × 10¹⁴ particles. From these values, the average numbers of –NH₂, –COOH, –CHO groups immobilized on 1, 2 and 3 were determined to be around 983.3, 803, and 737 respectively. The folate content can be conveniently obtained from the absorbance of 286 nm, which can be attributed to folate only. A calibration curve was developed by measuring the intensity of 286 nm absorbance as a function of folate concentration.⁵⁰ Using the calibration curve, the folate content on the surface of 4, 5 and 6 were estimated to be 0.57, 0.48 and 0.45 mM respectively.

3.4 Nanoparticle-mediated cytotoxicity

To further examine the biocompatibility of these functional nanoparticles as well as their folate-decorated counterparts for biomedical applications, an MTT assay was performed on the human cervical HeLa and a normal L929 cell line. Cells were incubated with various concentrations of the nanoparticles up to 200 μg ml⁻¹ of nanoparticles. After 4 h of incubation, no significant reduction in cellular viability was observed, the survival rate being higher than 85%, even at the higher nanoparticle concentrations (Fig. 5(A)). Hence these nanoparticles are safe as such and can be used for biomedical purposes.

Fig. 5 (A) Effect of folate-conjugated nanoparticles (4–6) and their non-targeted counterparts (1–3) on the viability of HeLa and L929 cells. (B) Comparison uptake of folate-targeted versus non-targeted nanoparticles by HeLa cells and a normal L929 cells (after 4 h incubation) as quantified by magnetically activated cell sorting (MACS).

3.5 Folate receptor mediated intracellular uptake of nanoconjugates

In order to evaluate the role of folic acid as a tumor cell targeting agent, cellular uptakes of the folate-targeted and non-targeted nanoconjugates were performed using human cervical carcinoma HeLa cells, a well known folate receptor overexpressed cancer cell line.⁵¹ Targeting agent triggered cellular internalization of these biofunctionalized nanoparticles inside the target cells (1–6) was studied using magnetically activated cell sorting (MACS) and fluorescence microscopy. According to the MACS based quantification data (Fig. 5(B)), the folate-conjugated nanoparticles (4–6) showed significant amount of uptake in HeLa cells after 4 h of incubation at 37 °C compared to their non-targeted counterparts (1–6). Cellular uptake of the folate targeted nanoconjugates (4–6) was significantly enhanced in comparison to their nontargeted control which confirms active cancer cell targeting of these nanoparticles to cancer cells. This was probably arising due to the interaction between the folate groups on the nanoparticle surface and folate receptors of HeLa cells.

The receptor mediated uptake of these nanoconjugates was confirmed by comparing the uptake of nanoconjugates (1–6) with a normal L929 fibroblast cells by MACS. It was clear from the data that for the folate receptor over-expressing HeLa cells, uptake of folate targeted nanoparticle 4 was much greater as compared to the same in normal fibroblast L929 cell line.

To further revalidate the receptor mediated internalization of nanoparticles by FA receptor overexpressing cancer cells, fluorescence microscopy was performed in the presence of HeLa in folate free medium using 1 and 4. After 4 h of incubation, HeLa cells treated with folate targeted 4 (Fig. 6(B)) were observed more fluorescent in color compared to their non-targeted counterpart (Fig. 6(A)), which significantly confirmed receptor mediated internalization of 4 to HeLa cells. The red colored folate conjugated nanoparticles were found to be distributed within the cytoplasm, leaving a clear zone of nucleus, suggesting cellular internalization through receptor-mediated endocytosis, instead of adhesion to the cell surface. The existence of the dispersed fluorescence in the cytoplasm was possibly induced by the unique uptake patterns for HeLa cells, in agreement with the results reported in literature.⁵² This study unequivocally established that all the iron-oxide folate nanoconjugates, developed in the course of our research, were preferentially targeted towards cancer cells and effectively internalized.

Fig. 6 (A) Hela cells treated with 1 after 4 h incubation. (B) Hela cells treated with 4 after 4 h incubation. (C) L929 normal fibroblast cells treated with 1. (D) L929 normal fibroblast cells treated with 4.

From the fluorescence image in Fig. 6(B), it was clearly observed that the HeLa cells expressed more folate receptors than the L929 cells and that the folic acid conjugated nanoparticles selectively accumulated on the surface of the FRs-positive HeLa cells as compared to the cellular surface of L929 cells (Fig. 6(D)). Hence, these results clearly confirmed that the binding and uptake of folate-conjugated nanoparticle (4) were internalized by folate receptor mediated endocytosis.

3.6 In vitro MR imaging studies of nanoconjugates

In vitro magnetic resonance phantom imaging (MRI) was further performed to evaluate the diagnostic potential of FA-OCMC-SPION-NH₂ (4) as a targeted MR probe to folate receptor positive HeLa cell that overexpresses folate receptor. Folate receptor positive HeLa and normal L929 cells cultured with 4 at various iron concentrations (0–50 μg/ml) were incubated for 2 h in agarose. These T₂-weighted MR images of the cells incubated with folate targeted and non-targeted nanoparticles at 1.5 T showed a clear contrast and a significantly reduced T₂ relaxation time when comparing with that of the control cells without treatment. The T₂-weighted phantom image of HeLa cells incubated with folate-targeted nanoparticles (4) exhibited a significant negative contrast enhancement (signal darkening) compared to the image obtained for 4 in L929 cells (Fig. 7(B)). It might be due to the rapid folate receptor mediated endocytosis to HeLa cells compared to L929 cells which leads to a distinguishable darkening of MR images with similar Fe concentrations. The signal intensity obtained for 4 in L929 cells was found to be comparatively lower than the control cells without nanoparticles. This could be due to the non-specific internalization of these small nanoparticles through an enhanced permeation and retention effect (EPR). The relative relaxation times of 4 in HeLa and L929 cells were quantified through T₂ weighted spin-echo MR images. Fig. 7(A) shows the T₂-relaxation time as a function of particle concentration in cell-culture media for the different nanoparticle preparations.⁵³ HeLa cells cultured with 4 possessed a shorter T₂-relaxation time (higher relaxivity) than the same in L929 cells arising due to enhanced magnetism. This case has resulted probably from the higher uptake of the 4 by HeLa cells in comparison to the same in L929 cells. All these results unequivocally established the potential of our as-prepared nanoformulations as a cancer-targeted, MRI probe.

Fig. 7 (A) T₂ relaxation analysis of HeLa and L929 cell suspensions labeled with nanoparticle designs 1 and 4. (B) T₂-weighted spin-echo MR phantom images of HeLa cells and L929 cells incubated with 0–50 μg/mL 1 and 4 for 2h.

4. Conclusions

In summary, a series of novel hydrophilic, biocompatible, magneto-fluorescent nanoparticles with surface pendant amine, carboxyl and aldehyde groups have been developed using o-carboxymethyl chitosan (OCMC) as a coupling agent. Amine functionalized magneto-fluorescent nanoparticles were developed by covalent immobilization of rhodamine isothiocyanate (RITC) on the amine-functionalized surface. These magneto-fluorescent nanoparticles were conjugated with reactive small molecules to fabricate magneto-fluorescent nanoparticles with diverse reactive functionalities (NH₂, COOH, CHO). Folic acid (FA), a widely used cancer cell targeting agent, was immobilized onto this synthesized magneto-fluorescent support, utilizing diverse conjugation strategies. Both the folate functionalized and folate-non-functionalized nanoparticles were extensively characterized in terms of size, charge, surface chemistry, composition and magnetic properties. The magnetite-nanoparticle-folate nanoconjugates showed excellent aqueous dispersion stability with reasonably good hydrodynamic size over a wide range of physiological conditions. Magnetically activated cell sorting (MACS), fluorescence microscopy and in vitro MRI performed with folate receptor positive HeLa cancer cells and a normal fibroblast L929 cell line, clearly established the targeting efficiencies as well as folate receptor mediated internalization of these non-cytotoxic iron-oxide folate nanoconjugates towards cancer cells. To the best of our knowledge, this is the first attempt to develop a series of folate receptor targeted OCMC-SPIONs combining the targeting agent and fluorophore into one magnetic probe for simultaneous targeting, detection and molecular imaging.

Acknowledgements

The authors gratefully acknowledge Dr K. R. Patil, Centre of Material Characterization, National Chemical Laboratory (NCL), Pune, and ECO MRI Scan Centre (Kolkata), for assistance with the XPS and MRI studies. Authors are grateful to the Council of Scientific and Industrial Research (CSIR), New Delhi, for providing financial support for this work.

Notes and references

1. Y.-W. Jun, J.-W. Seo and J. Cheon, Acc. Chem. Res., 2008, 41, 179–189.
2. E. I. Sega and P. S. Low, Cancer Metastasis Rev., 2008, 27, 655–664.
3. J. Zhang, S. Rana, R. S. Srivastava and R. D. K. Misra, Acta Biomater., 2008, 4, 40–48.
4. C. Sun, R. Sze and M. Zhang, J Biomed Mater Res A, 2006, 78, 550–557.
5. Y. Jun, J. Lee and J. Cheon, Angew. Chem., Int. Ed., 2008, 47, 5122–5135.
6. J. R. McCarthy and R. Weissleder, Adv. Drug Delivery Rev., 2008, 60, 1241–1251.
7. J. Kim, Y. Piao and T. Hyeon, Chem. Soc. Rev., 2009, 38, 372–390.
8. H. B. Na, I. C. Song and T. Hyeon, Adv. Mater., 2009, 21, 2133–2148.
9. M. Knauth, T. Egelhof, S. U. Roth, C. R. Wirtz and K. Sartor, AJNR Am. J. Neuroradiol., 2001, 22, 99–102.
10. L. M. Lacava, Biophys. J., 2001, 80, 2483–2486.
11. R. Weissleder, A. Bogdanov, E. A. Neuwelt and M. Papisov, Adv. Drug Delivery Rev., 1995, 16, 321–324.
12. Y. Zhang, N. Kohler and M. Zhang, Biomaterials, 2002, 23, 1553–1561.
13. P. K. Gupta and C. T. Hung, Life Sci., 1989, 44, 175–186.
14. N. Kohler, C. Sun, J. Wang and M. Zhang, Langmuir, 2005, 21, 8858–8864.
15. J. Xie, K. Chen, H.-Y. Lee, C. Xu, A. R. Hsu and S. Peng, J. Am. Chem. Soc., 2008, 130, 7542–3.
16. E. V. Groman, J. C. Bouchard, C. P. Reinhardt and D. E. Vaccaro, Bioconjugate Chem., 2007, 18, 1763.
17. M. Tsotsalas, M. Busby, E. Gianolio, S. Aime and L. De Cola, Chem. Mater., 2008, 20, 588.
18. S. Liu, X. Wei, M. Chu, J. Peng and Y. Xu, Colloids Surf., B, 2006, 51, 101–106.
19. S. Mornet, J. Portier and E. Duguet, J. Magn. Magn. Mater., 2005, 293, 127–134.
20. J. Lee, T. Isobe and M. Senna, J. Colloid Interface Sci., 1996, 177, 490–494.
21. Y. Zhang and J. Zhang, J. Colloid Interface Sci., 2005, 283, 352–357.
22. A. K. Gupta and M. Gupta, Biomaterials, 2005, 26, 3995–4021.
23. M. Mahmoudi, A. Simchi, M. Imani and U. O. Hafeli, J. Phys. Chem. C, 2009, 113, 8124–8131.
24. S. R. Bhattarai, K. C. R. Bahadur, S. Aryal, M. S. Khil and H. Y. Kim, Carbohydr. Polym., 2007, 69, 467–477.
25. P. Chan, M. Kurisawa, J. E. Chung and Y. Y. Yang, Biomaterials, 2007, 28, 540–549.
26. A. P. Zhu, M. B. Chan-Park, S. Dai and L. Li, Colloids Surf., B, 2005, 43, 143–149.
27. A. Zhu, L. Yuan and Y. Lu, Colloid Polym. Sci., 2007, 285, 1535–1541.
28. Y. Y. Liang and L. M. Zhang, Biomacromolecules, 2007, 8, 1480–1486.
29. A. Zhua, L. Yuan and T. Liao, Int. J. Pharm., 2008, 350, 361–368.
30. X. H. Peng, X. Qian, H. Mao, A. Y. Wang, Z. G. Chen and S. Nie, Int J Nanomedicine, 2008, 3, 311–321.
31. J. J. Lin, J. S. Chen, S. J. Huang, J. H. Ko, Y. M. Wang and T. L. Chen, Biomaterials, 2009, 30, 5114–5124.
32. J. H. Ke, J. J. Lin, J. R. Carey, J. S. Chen, C. Y. Chen and L. F. Wang, Biomaterials, 2010, 31, 1707–1715.
33. B. Stella, S. Arpicco, M. T. Peracchia, D. Desmaele, J. Hoebeke, M. Renoir, D. J. Angelo, L. Cattel and P. Couveur, J. Pharm. Sci., 2000, 11, 1452–1464.
34. P. S. Low and A. C. Antony, Adv. Drug Delivery Rev., 2004, 56, 1055–1058.
35. S. Mohapatra, S. K. Mallick, T. K. Maiti, S. K. Ghosh and P. Pramanik, Nanotechnology, 2007, 18, 385102–385111.
36. X. G. Chen and H. J. Park, Carbohydr. Polym., 2003, 53, 355–359.
37. I. J. Bruce and T. Sen, Langmuir, 2005, 21, 7029–7035.
38. A.d. Campo, T. Sen, J. P. Lellouche and I. J. Bruce, J. Magn. Magn. Mater., 2005, 293, 33–40.
39. F. E. Levy, M. C. Andry and M. C. Levy, Int. J. Pharm., 1993, 96, 85–90.
40. A. J. Dong, M. H. Feng, H. Y. Qi, S. S. Li and D. L. Deng, J Mater Sci: Mater Med, 2008, 19, 869–876.
41. O. A. C. Monteiro Jr and C. Airoldi, Int. J. Biol. Macromol., 1999, 26, 119–128.
42. H. Yang, Y. Zhuang, H. Hu, X. Du, C. Zhang and X. Shi, Adv. Funct. Mater, 2010, 20, 1–9.
43. S. Wang, J. Luo, D. A. Lantrip, D. J. Waters, C. J. Mathias, M. A. Green, P. P. Fuchs and S. Low, Bioconjugate Chem., 1997, 8, 673–679.
44. A. Gabizon, A. T. Horowitz, D. Goren, D. Tzemach, F. Mandelbaum-Shavit, M. M. Qazen and S. Zalipsky, Bioconjugate Chem., 1999, 10, 289–298.
45. R. Schneider, F. Schmitt, C. Frochot, Y. Fort and N. Lourette, Bioorg. Med. Chem., 2005, 13, 2799–2808.
46. A. D. Campo, T. Sen, J.-P. Lellouchec and I. J. Bruce, J. Magn. Magn. Mater., 2005, 293, 33–40.
47. J. Zhang, S. Rana, R. S. Srivastava and R. D. K. Misra, Acta Biomater., 2008, 4, 40–48.
48. Q. Yuan, S. Hein and R. D. K. Misra, Acta Biomater., 2010, 6, 2732–2739.
49. S. Liana, Z. Kanga, E. Wanga, M. Jianga, C. Hua and L. Xua, Solid State Commun., 2003, 12, 605–608.
50. H. Choi, S. R. Choi, R. Zhou, H. F. Kung and I. W. Chen, Acad. Radiol., 2004, 11, 996–1004.
51. M. S. Justin, A. Annapragada, J. V. Natarajan and R. V. Bellamkonda, J. Controlled Release, 2003, 92, 49–67.
52. G. Destito, R. Yeh, C. S. Rae, M. G. Finn and M. Manchester, Chem. Biol., 2007, 14, 1152–1162.
53. J. Chen, I. R. Corbin, H. Li, W. Cao, J. D. Glickson and G. Zheng, J. Am. Chem. Soc., 2007, 129, 5798–5799.

---