Review Article

In vitro biocompatibility assessment of functionalized magnetitenanoparticles: Biological and cytotoxicological effects

D. A. Mbeh,1 R. Franca,1 Y. Merhi,2 X. F. Zhang,3 T. Veres,3 E. Sacher,1,4 L. Yahia1

1Laboratory for Innovation and Analysis of Bio-Performance, Ecole Polytechnique, C.P. 5079, Succursale Centre-ville,

Montr�eal, Qu�ebec, Canada H3C 3A72Montreal Heart Institute, Universit�e de Montr�eal, 5000 rue Belanger, Montr�eal, Qu�ebec, Canada H1T 1C83Institut des Mat�eriaux Industriels, 75 de Mortagne, Boucherville, Qu�ebec, Canada J4B 6Y44Regroupement Qu�ebecois de Mat�eriaux de Pointe, Ecole Polytechnique, C.P. 5079, Succursale Centre-ville, Montr�eal,

Qu�ebec, Canada H3C 3A7

Received 3 June 2011; revised 6 October 2011; accepted 6 January 2012

Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbm.a.34096

Abstract: In the biomedical field, nanomaterials have the

potential for use in the targeted delivery of drugs in the

human body and in the diagnosis and therapy of certain dis-

eases. In the category of targeted delivery, magnetite (Fe3O4)

nanoparticles have received much attention. As with any sim-

ilar new therapy, when such nanoparticles are functionalized

with chemical groups designed to permit the specific attach-

ment of drugs, cytotoxicological testing is necessary before

moving to animal models. Here, we consider several vari-

ously functionalized magnetite nanoparticles, including those

prepared with (1) a monolayer of oleic acid (Fe3O4@OA),

which is subsequently converted to (2) a shell of amine-con-

taining silane (Fe3O4@NH2), (3) a shell of silica (Fe3O4@SiO2),

and (4) a shell of amine-containing silane over a shell of

silica (Fe3O4@SiO2@NH2). These latter three functionalities

were evaluated for biocompatibility, cellular morphology,

mitochondrial function (MTT assay), lactate dehydrogenase

membrane leakage (LDH assay), and proinflammatory poten-

tial through enzyme linked immunosorbent assay (ELISA) for

interleukin 6 (IL-6). Controlled tests were performed over a

period of 72 h, with results showing LDH leakage and abnor-

mal Il-6 secretion at high concentrations (>50 lg/mL). The

tests showed that, in addition to the surface characteristics of

the nanoparticles, both the nutrient medium and the time of

suspension before exposure to cells also contribute to nano-

particle cytotoxicity. VC 2012 Wiley Periodicals, Inc. J Biomed Mater

Res Part A: 00A:000–000, 2012.

Key Words: biocompatibility, cell viability, cytotoxicological

effects, functionalized magnetite nanoparticles

How to cite this article: Mbeh DA, Franca R, Merhi Y, Zhang XF, Veres T, Sacher E, Yahia L. 2012. In vitro biocompatibilityassessment of functionalized magnetite nanoparticles: Biological and cytotoxicological effects. J Biomed Mater Res Part A2012:00A:000–000.

INTRODUCTION

In recent years, the growth of nanotechnology has animatedmany scientific debates, whose points of view may appearcontradictory. For example, research suggests many advan-tages, for both the wellbeing of man and assured future man-ufacturing growth, through the comprehension and control ofthe functionalization of nanoparticles (NPs).1–3 However, en-thusiasm for this nanotechnology raises the question of bothtoxicological and environmental impacts.4 In the biomedicalfield, several areas are represented, such as nanosurgery, inwhich NPs are used as nano-heaters to burn malignant cells,5

nano-vectors, in which NPs are used as drug delivery agentsto a target site, and nano-tracers, in which NPs are used fordiagnosis through medical imagery.6 Magnetite (Fe3O4) NPs

offer many potential applications to these fields: they aresuperparamagnetic (i.e., they exhibit magnetic behavior inthe presence of magnetic fields), they are MRI-detectable,with specific microfabrication increasing their magneticresonance imaging (MRI) sensitivities, their surfaces may befunctionalized to act as drug receptors, and their tempera-tures may be raised in alternating magnetic fields.7

As the use of NPs increases, with new nanotechnologi-cally applicable materials emerging, hazard identificationbecomes critical. In vitro studies will play an essential role,because they can provide high-throughput systems for therapid and cost-effective screening of hazards. Therefore, reli-able in vitro studies that generate comprehensible resultsare needed.

Correspondence to: D. A. Mbeh; e-mail: doris-antoinette.mbeh@polymtl.ca

Contract grant sponsors: Natural Sciences and Engineering Research Council of Canada; NanoQu�ebec and Nitrox Medical Device

VC 2012 WILEY PERIODICALS, INC. 1

Our ultimate purpose is to functionalize magnetite NPsfor magnetic targeting. This is because, with untargeted in-travenous administration, the quantity of potentially toxicdrug administered is necessarily high, so as to overcomedilution and losses through the liver, kidney, etc. The use ofmagnetic targeting is beneficial, in that it reduces both theamount of drug used and the resultant toxicity.

The NPs used in this study are composed of superpara-magnetic iron oxide nanoparticles (SPION) cores,surrounded by several different shells, including (1) amonolayer of oleic acid (Fe3O4@OA), (2) a shell of amine-containing silane (Fe3O4@NH2), (3) a shell of silica(Fe3O4@SiO2), and (4) a shell of amine-containing silaneover a shell of silica (Fe3O4@SiO2@NH2). The @OA is usedin the general preparation of the latter three NPs, while the@NH2, @SiO2,and @SiO2@NH2 are intended to provide teststructures for further functionalization with drugs.

When NPs are functionalized, changes occur in theirphysicochemical properties such as size, shape, and surfacereactivity, which can cause changes in biological responses.8

Currently, knowledge of the toxicity of manufactured NPsremains limited, with often contradictory results from dif-ferent laboratories, even when using the same NPs.9 Beyondbasic NP behavior, new parameters, such as the characteris-tics of the surrounding environment (e.g., the compositionof the culture medium, the pH value, the presence and typesof proteins) contribute to the ambiguity of cell interactionsin solution.10 For example, NPs suspended in culture mediaare known to adsorb a set of diverse proteins and form aprotein corona, whose composition depends on the proteinspresent, and their resultant epitopes on reaction with theNPs, as well as on the surface characteristics, size, and con-centration of the NPs.11 It is important to understand howthese engineered NPs interact with living systems, from themoment of first contact and the onset of interactionsinduced by their presence.

Traditionally, in vitro toxicity testing focuses on whetheror not exposure to a potentially toxic agent results in celldeath. However, even when no cell damage or death may beapparent after NP exposure, changes in cellular functionmay result. Therefore, it is important to verify that the endpoints chosen to signify cytotoxicity are appropriate. Weassess the toxicities of our NPs through changes in cellularmorphology, mitochondrial function, membrane leakage byLDH, and IL-6 production.12

Klaine et al.13 proposed that the activation of intracel-lular oxidative stress appears to provoke the toxicitymechanisms of many nanomaterials. As soon as the nano-material enters the cell, it may induce intracellular oxida-tive stress by disturbing the balance between oxidantand anti-oxidant processes. For biomedical applications,when choosing the nanomaterial, its surface chemistrymust be evaluated, so that both the cytotoxic effects andthe possible control of pro-inflammatory cytokines maybe taken into account. To identify systemic inflammatoryresponses, one may choose to measure immune systemproteins, such as pro-inflammatory cytokines (IL-1, IL-6,and tumor necrosis factor-alpha TNF-a), Th1-type

cytokines (IL-12 and Interferon-gamma IFN-c), and Th2-type cytokines (IL-4, IL-5, and IL-10) among others. Inour study, we chose to quantify IL-6 because we use cellline A549, which are epithelial cells that secrete IL-6 inresponse to stress.

MATERIALS AND METHODS

NanoparticlesMaterials. Oleic acid (OA, 90%), 1-hexaneol (99%), octylether (98%), ammonia solution (NH4OH, 28–30 wt % inwater), Triton X-100, hexane (95%), cyclohexane (99.5%),and tetraethoxysilane (TEOS, 99.999%) were purchasedfrom Sigma-Aldrich. Iron pentacarbonyl (99.5%) was pur-chased from Strem Chemicals. N-(2-aminoethyl)-3-amino-propyl trimethoxysilane (AEAP3, � 90%) was purchasedfrom Gelest.

Synthesis of Fe3O4@NH2 NPs. Oleic acid-coated Fe3O4 NPs(Fe3O4@OA) were synthesized using a thermal decomposi-tion process.14 Twenty milliliters of octyl ether was mixedwith 1.92 mL of oleylamine at room temperature for �10minutes, and subsequently heated to 100�C over 20 min,under an Ar flow. At 100�C, 0.4 mL of iron pentacarbonylwas injected, and the temperature was increased to refluxfor 2 hours, before cooling to room temperature by remov-ing the heating mantle. The product was precipitated byadding anhydrous ethanol, and separated by centrifugation(9000 rpm), followed by three washings with anhydrousethanol. The Fe3O4@NH2 NPs were obtained by mixing 10mL of 1 mg/mL Fe3O4@OA in hexane with 50 lL of AEAP3,and sonicating for 90 min. The resultant suspension wasleft overnight, and then separated by magnet. The separatedNPs were washed five times with 40 mL anhydrous ethanol.These NPs can then be dispersed in water or PBS buffer.

Synthesis of Fe3O4@SiO2 and Fe3O4@SiO2@NH2

NPs. Fe3O4@SiO2 NPs were formed in a water-in-oil microe-mulsion: 0.5 mL of 1 mg/mL Fe3O4@OA nanoparticles incyclohexane were injected into a mixture of 1.77 g of TritonX-100, 1.6 mL of anhydrous 1-hexanol and 7 mL of cyclo-hexane, and left under a strong vortex for 1 h, before adding0.5 mL of 6% ammonia solution. After 1 h, 25 lL of TEOSwas added and left for 24 h. The product, core/shellFe3O4@SiO2, was separated by centrifugation at 9000 rpm,and washed with ethanol; the centrifugation/wash proce-dure was repeated at least three times. The resultant NPswere directly dispersed in deionized water or PBS buffer foruse. For the formation of Fe3O4@SiO2@NH2 NPs, 25 lL ofAEAP3 was added into the reaction mixture for another 24h. The product was separated, purified, and stored as werethe Fe3O4@SiO2 NPs.

NP characterization. The surface chemistry was character-ized by X-ray photoelectron spectroscopy (XPS), in a VGESCALab 3 Mk II, using non-monochromated Al Ka radiation(1486.6 eV), at a power setting of 300 W, giving an instru-ment resolution of 0.85 eV. Magnetic properties were eval-uated with a Quantum Design PPMS model 6000

2 MBEH ET AL. IN VITRO BIOCOMPATIBILITY OF FUNCTIONALIZED MAGNETITE NANOPARTICLES

magnetometer. Transmission electron microscopic (TEM)images were taken with a JEOL JSM-7600TFE, at a 200 keVaccelerating voltage. The TEM specimens were prepared bydropping the NP suspensions (Fe3O4@OA in hexane, and theothers in DI water) onto carbon-coated copper TEM grids,followed by evaporation under vacuum overnight.

NP preparation for cytotoxicological testing. Since it isintended that magnetic NPs accumulate at the target site,during drug delivery, at a locally elevated concentration, itis necessary to evaluate their cytotoxicity at such elevatedconcentrations. To do this, NPs were suspended in Ham’s F-12K culture medium (Sigma–Aldrich), supplemented with10% FBS (fetal bovine serum). In establishing the concen-trations to be used in our tests, we considered those ofFe3O4 used regularly in the literature for in vivo cytotoxicityassay. In order to obtain an idea of the toxicity of a CH3 NPsurface, which is chemically nonreactive under present con-ditions, we tested Fe3O4@OA, the only NP not intended forultimate medical use, at concentrations ranging from 0.15 to0.90 mg/mL. In order for Fe3O4@NH2, Fe3O4@SiO2 andFe3O4@SiO2@NH2 to be detectable by MRI, and given thatthe human dose concentration recommended for lymphog-raphy is < 45 lmol/kg,15 these latter three were addition-ally tested at concentrations from 0.05 to 200 lg/mL.

All the assays used controls. The background controlwas carried out with the medium only. The control cellswere those incubated with medium plus supplements, butwithout NPs. The control particles were in wells withoutcells, which underwent the viability assay.

Cell culture. The epithelium is the first barrier that con-fronts particles that deposit in the conducting airways or thealveolar region. Therefore, both bronchial and alveolar epi-thelial cells should be considered as target cells for in vitrostudies.16 The human alveolar epithelial cell line A549 (CCL-185) was obtained from the American Type Culture Collec-tion (ATCC); A549 is a human Caucasian lung carcinoma typeII epithelial cell. It was grown in a 75 cm3 glass flask usingHam’s F-12K, supplemented with 10% fetal bovine serum(FBS) and1% penicillin/streptomycin, to avoid bacterial con-tamination, and maintained at 37�C in a humidified atmos-phere of 5% CO2 and 95 % air. Cells were grown to 70%–80% confluence and then trypsinized to dissociate adherentcells from each other and from the container.

Cytotoxicity testsCurrently used in vitro cytotoxicity assays measure cell viabil-ity via colorimetric methods. These methods can be furtherclassified into tests that measure plasma membrane integrityand mitochondrial activity. Exposure to certain cytotoxicagents can compromise the cell membrane, which allows cel-lular contents to leak out. Viability tests based on this includethat for LDH. In this assay, LDH, released from damaged cells,oxidizes lactate to pyruvate. The amount of LDH released isproportional to the number of cells damaged or lysed.

Mitochondrial activity is another colorimetric cytotoxic-ity assay, which attempts to determine the mechanism

behind the induced cell death. Mitochondrial activity can betested by using tetrazolium salts, as mitochondrial dehydro-genase enzymes cleave the tetrazolium ring.

Inflammation is also a possible adverse effect of NP ex-posure. Commonly tested proinflammatory cytokines, orprotein signals, of inflammatory response include IL-1b, IL-6, and TNF-a, among others. These cytokines are detectedby using ELISA, and can be quantified by measuring the ab-sorbance from either alkaline phosphatase or streptavidin-horseradish peroxidase-labeled antibodies.

The MTT, LDH, and ELISA tests, described above, wereperformed with the same concentration of cells. According toour methodology, after incubation of the cells with the nano-particles, the supernatant was removed for LDH and ELISAtests, and the cells in the wells were used for the MTT test.

Assessment of cellular morphology by optical micro-scopy. The assessment of cellular morphology was carried outfor Fe3O4@OA, which has a chemically nonreactive17 CH3 sur-face (alkyl groups, such as CH3 do not react chemically underthe conditions of our experiment); this was done to determinewhether a larger cell morphology study was necessary, consid-ering factors other than the chemical groups present. To assessthe cellular morphology, cells, treated with 0.47 mg/mL ofFe3O4@OA in Ham’ S F-12k culture medium, were incubatedfor 24 and 72h at 37�C. Optical photomicrographs of the cellswere taken at an enlargement of 20�, and the morphology ofattached A549 cells was qualitatively evaluated.

Determination of mitochondrial activity, using the MTTtest. The MTT assay is based on the reduction of MTT ((3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide)by mitochondrial succinate dehydrogenase, to form darkblue formazan crystals. Only viable cells, with active mito-chondria, reduce significant amounts of MTT to formazan.18

For this assay, cells were seeded in a 96-well plate, using 1� 104 cells/well, and incubated for 24 h, before adding var-ious concentrations of Fe3O4@NH2, Fe3O4@SiO2 andFe3O4@SiO2@NH2 NPs to each well, and incubating the cellsfor a further 24, 48, and 72 h. After the incubation period,20 lL of MTT (0.5 mg/mL in phosphate buffer, at pH 7.4)was added to each well and the cells were incubated foranother 3 h. After discarding the media, the formazan crys-tals were solubilized by 100 lL of dimethyl sulfoxide(DMSO) per well and the absorbance was measured at 570nm using a BioTek Synergy 2 microplate reader.

Determination of lactate dehydrogenase. LDH leakage dueto membrane damage was assessed by measuring the quan-tity in the media after the incubation of cells and the NPs tobe analyzed.19 Using 2 � 104 cells per well in a 96-wellplate, incubation occurred for a period of 24 h, with NPs atconcentrations from 0.05 to 100 lg/mL. The release of LDHinto the supernatant was determined by an LDH activityassay kit (Roche, Mannheim, Germany) according to themanufacturer’s instructions. Incubation with 1% Triton-X100 served as a positive control, and cells without NPs, asa negative control. The results are given relative to the posi-tive control, in percent:

REVIEW ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | MONTH 2012 VOL 9999A, ISSUE 00 3

ððS� CÞ=ðT� CÞÞ � 100;

where S ¼ sample value, C ¼ negative control value, and T¼ 1% triton positive control value.

Quantification of cytokines by ELISA. Studies have shownthat the A549 human lung epithelial cell line can synthesizerelevant cytokines, including IL-6, which appear to play arole in the initiation of inflammation and, possibly, to con-tribute to the polarization of a Th2-dominated immuneresponse, which plays a role in the development of allergicrespiratory diseases.20 Concentrations of IL-6 in the super-natants obtained after exposure of NP suspensions weredetermined by sandwich ELISA with appropriate antibodies,according to the manufacturer’s instruction (R&D Systems,Minneapolis, USA).

RESULTS

Nanoparticle characterizationThe TEM image of as-synthesized Fe3O4@OA NPs, shown inFigure 1(a), indicates a narrow size distribution, with an av-erage diameter of �15 nm. Figure 1(b) shows theFe3O4@NH2 NPs; one finds that the self-assembledFe3O4@OA NPs in Figure 1(a) have been replaced by a dis-

ordered array, ascribed to the change of surface functionalgroups. Images of Fe3O4@SiO2 and Fe3O4@SiO2@NH2 areshown in Figure 1(c,d), respectively. From the statisticalanalysis of 200 particles, the particle diameters are �56and �65 nm, respectively, indicating shell thicknesses of�20 and �30 nm. Compared with the Fe3O4@SiO2 nanopar-ticles, the Fe3O4@SiO2@NH2 NPs present a mesoporoussilica shell with open pores extending to the surface. Thesemesoporous shells, containing both primary and secondaryamine groups, may be used for medical applications, as car-riers of biological species, such as drugs, antibodies andproteins. Compared with Fe3O4@NH2 NPs, we find that thesilica coatings result in better dispersibility and stability inaqueous solution, as shown in Figure 2. This may bebecause of hydrogen bonding with the solvent. Consideringthe different surface functional groups and microstructures,we were curious as to whether these features affect theircytotoxicity.

The XPS survey spectra for the nanoparticles are seen inFigure 3, and the relative concentrations of the elementsfound in Tables I and II. As Table I indicates, the shells aresufficiently thick to obviate the detection of Fe. Because theattenuation length of the Fe2p electron is �1.3 nm, theprobe depth (three times the attenuation length) is �4 nm;

FIGURE 1. TEM images of (a) Fe3O4@OA, (b) Fe3O4@NH2, (c) Fe3O4@SiO2, and (d) Fe3O4@SiO2@NH2 NPs. [Color figure can be viewed in the

online issue, which is available at wileyonlinelibrary.com.]

4 MBEH ET AL. IN VITRO BIOCOMPATIBILITY OF FUNCTIONALIZED MAGNETITE NANOPARTICLES

this means that all the shells are at least 4 nm thick. Thepresence of trace Si in the case of Fe3O4@OA is probablybecause of contamination.

The relative elemental concentrations of NPs as synthe-sized, on contact with the culture medium at a concentra-tion of 100 lg/mL and on cell contact, are found in Table II,and are given as histograms in Figure 4. Table II and Figure4 show that, on contact with both culture medium and cells,the NPs undergo changes in carbon, nitrogen, and oxygenconcentrations, indicating surface modification.

On exposure to culture medium, this probably occursthrough corona formation by one or more of the mediumcomponents. For the @NH2-terminated shells, the elementalconcentrations do not appear to change after subsequentcell contact; however, for @SiO2-terminated shell, there is aclear enrichment in both O and N, indicating coronamodification.

Viability of cells after incubation with nanoparticlesAssessment of cellular morphology on exposure toFe3O4@OA. The capacity of the A 549 cell line to maintainmorphology, on 24 and 72 h incubation with functionalizedSPIONs, was determined by optical microscopy, usingFe3O4@OA, and the results are found in Figure 5. The pho-tomicrographs show that the cell morphology was altered,compared with the controls. Growing the cells with SPIONsproduced a significant difference of the lateral extension ofthe cells, compared to that of the control cell population[Fig. 5(a–d)]. After 72 h, the morphology of the cells did notimprove with time; the A 549 cells shrank and becameirregular in shape [Fig. 5(d)]. This effect of Fe3O4@OA oncell morphology is surprising, and certainly not because ofchemical reaction with the NP surface CH3 groups.

Assessment of cell proliferation after 72 h of exposure toFe3O4@OA. Figure 6(a) shows a photomicrograph of the A549 cells cultured without particles (control). As seen, thecells are confluent. Figure 6(b–d) present photomicrographsof A 549 cells cultured with Fe3O4@OA. After 72 h, the cellsshow a significant dose-dependent reduction in their growthcapacity. The photomicrographs reveal that, with increasingconcentration, from 0.15 to 0.90 mg/mL, cells becomerounded and lose cell-to-cell and cell-to-dish contact. Theseconcentrations also lead to a detectable decrease in prolifer-ation and to a cell morphology alteration.

Determination of LDH leakage. The rupture of the cellmembrane induces the release of LDH into the extracellularmedium, with the quantity of LDH that is salted out in theculture medium being proportional to cellular death. Thepercentage of LDH was calculated by dividing the quantity

FIGURE 2. Optical photographs of Fe3O4@OA NPs, which can be easily

dispersed in hexane (left); after silica coating, the Fe3O4@SiO2NPs can

be dispersed in aqueous solution (right). [Color figure can be viewed in

the online issue, which is available at wileyonlinelibrary.com.]

FIGURE 3. XPS survey spectra of (a) Fe3O4@OA, (b) Fe3O4@NH2, (c)

Fe3O4@SiO2 and (d) Fe3O4@SiO2@NH2 NPs.

TABLE I. Elemental Composition on NPs Surface

Carbon(%)

Oxygen(%)

Nitrogen(%)

Silicium(%)

Fe3O4@OA 88.80 10.17 – 0.42Fe3O4@NH2 87.54 8.14 2.48 1.38Fe3O4@SiO2 45.35 40.19 – 14.46Fe3O4@SiO2@NH2 54.24 28.36 8.20 9.20

REVIEW ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | MONTH 2012 VOL 9999A, ISSUE 00 5

in the medium (sample tested) by the positive control(lysed cells). The negative control consisted of cells unex-posed to NPs, in their culture medium. The results repre-sent the average of three independent experiments. For thethree types of NPs, we find that the quantity of LDH beginsto increase, corresponding to lower viability, beginningbetween 50 and 100 lg/mL for both Fe3O4@NH2 andFe3O4@SiO2@NH2 (Fig. 7). Both have the same surface func-tionalities and amine concentrations after suspension in cul-ture medium (see Table II, column CM). However, as men-tioned earlier, they differ in size (Fig. 1). Given the sizedifferences, we believe that the rate difference involves thedifference in size, and are pursuing this in our present stud-ies. This view is supported by previously publishedstudies.21,22

Quantification of cytokines by ELISA. Our aim was to seewhether NPs, when exposed to the cells, induced the secre-tion of pro-inflammatory cytokines, such as IL-6. The A549cells were exposed to several concentrations of Fe3O4@NH2,Fe3O4@SiO2, and Fe3O4@SiO2@NH2 NPs, as with the LDHtest. The quantitative evaluation of IL-6 was carried out as

described in the Methods section. The control consists ofcells in the culture medium without NPs. The results con-firm those of the LDH test: large quantities of IL-6 are pro-duced by cells exposed to high concentrations of all thefunctionalized NPs. Cells show a significant difference fromthe control, at concentrations of 50 and 100 lg/mL (Fig. 8),although lower concentrations (10 lg/mL and less) do notshow such a difference.

Assessment of cell viability by MTT test. The results of theMTT assays show that exposure to both Fe3O4@NH2 andFe3O4@SiO2@NH2 NPs caused similar reductions in cell via-bility as a function of time, compared with Fe3O4@SiO2,which is different (Fig. 9). However, all the NPs are seen tobecome more toxic with time, especially at low concentra-tions (Fig. 10).

DISCUSSION

As indicated in the Introduction, the biocompatibility ofSPIONs is a prerequisite for any clinical use. Up to now, theoverwhelming majority of reports have claimed that SPIONare biosafe because they are presumably bio-inert and/orbecause their metabolite, Fe, is essential for living organ-isms and thus biocompatible. Many studies have also shownno adverse effects on potential cytotoxicity and cell functionat low doses (�10 pg Fe/cell) for optimal MRI cell track-ing.23 The conclusion of biosafety, however, may be prema-ture, given the lack of criteria for evaluating the safety andtoxicity of these nanomaterials. Indeed, growing evidencehas demonstrated that SPION labeling more or less altersseveral cellular events of mesenchymal stem cells.24 There-fore, the effects of SPIONs on diverse aspects of cellularactivities should be carefully evaluated even though suchNPs are generally considered biocompatible at present.

Cell morphologyA change in cell morphology is one of the first signs of itsexposure to toxic products.25 After 24 and 72 h ofFe3O4@OA NP exposure, the cell morphology was found tobe altered (Fig. 5), characterized by a reduction of the lat-eral extension of the cells. This indicates a reduced adhesionof the cells to the substrate [Fig. 5(b,d)], indicating a

TABLE II. Elemental Composition of NPs Surface Before and

After Suspension in the Ham’s F-12K

E AS CM C

Fe3O4@NH2

Silicium 2.63 0 0Carbon 48.8 63.18 59.55Nitrogen 7.75 10.27 12.47Oxygen 36.1 23.08 23.55Fe3O4@SiO2

Silicium 19.51 0 0Carbon 26.14 65.98 58.72Nitrogen 0.71 6.36 11.93Oxygen 52.21 14.44 25.55Fe3O4@SiO2@NH2

Silicium 20.73 0 0Carbon 22.45 61.14 57.85Nitrogen 5.34 11.29 11.95Oxygen 51.48 23.32 25.53

E, elements; AS, NPs after synthesis; CM, NPs after suspension in

the culture media; C, NPs suspended in the culture media and after

contact with cells.

FIGURE 4. Elemental composition of NP surfaces before and after suspension in the Ham’s F-12K, as determined by XPS. E, elements; AS, as

synthesized; CM, after suspension in the culture media; C, after contact with cells. [Color figure can be viewed in the online issue, which is avail-

able at wileyonlinelibrary.com.]

6 MBEH ET AL. IN VITRO BIOCOMPATIBILITY OF FUNCTIONALIZED MAGNETITE NANOPARTICLES

disturbance of the structure and function of the cytoskele-ton, and suggesting that the CH3-functionalized NPs havecaused deterioration in elements of the matrix (probably fi-bronectin and biopolymers, as was found for NH2 function-alization26) that influence cellular proliferation.

Fe3O4@OA is found to be toxic for A549 cells. Similarresults have been reported by Yin and his collaborators,who demonstrated that the Fe3O4@OA (CH3-exposed) weremore cytotoxic than Fe NPs with two layers of oleic acid

(COOH-exposed).27 Their study indicated that oleic acid, perse, is not cytotoxic; however, when spatially aligned, onbeing coated onto the NPs, cytotoxicity was observed. Ourviability results suggest that the cellular responses dependon several parameters, of which NP functionalization is onlyone. As a result of this finding, we have initiated a testingprogram to determine these parameters.

Effect of nanoparticle functionalizationBesides the NP size, the chemical nature of its surfaceappears to determine the intensity of the biological effectswith significant differences between different types of coat-ing because of their differing surface chemistries that reactin different fashions with the components of the culture me-dium before reacting with the cells. That appears to explainwhy cells exposed to NPs with an amine coating(Fe3O4@NH2 and Fe3O4@SiO2@NH2) react with the sametendency, compared to one with a silica coating (Fe3O4@-SiO2) at a concentration of 200 lg/mL. We are presently

FIGURE 5. Morphological differences of the A549 cells before and af-

ter exposure to Fe3O4@OA. (a) Control: no treatment during 24 h, (b)

exposure for 24 h, (c) control: no treatment for 72 h, (d) exposure for

72 h. [Color figure can be viewed in the online issue, which is avail-

able at wileyonlinelibrary.com.]

FIGURE 6. Reduction of cellular proliferation with increasing concen-

tration of NPs: (a) Control, (b) 0.15, (c) 0.47, and (d) 0.90 mg/mL.

[Color figure can be viewed in the online issue, which is available at

wileyonlinelibrary.com.]

FIGURE 7. Effect of nanoparticles on the secretion of LDH by A549

cells. [Color figure can be viewed in the online issue, which is avail-

able at wileyonlinelibrary.com.]

FIGURE 8. Quantity of IL-6 secreted by A549 cells exposed to NPs at

various concentrations. [Color figure can be viewed in the online

issue, which is available at wileyonlinelibrary.com.]

REVIEW ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | MONTH 2012 VOL 9999A, ISSUE 00 7

exploring this with other functionalized nanoparticles, andwill report our findings in a future paper.

Effect of coronaWe can observe the initial effect of a coating when NPs aresuspended in solution. Some coating reduces the interactionbetween NP themselves, which reduces aggregation, whileother coatings facilitate dispersion. Figure 2 shows the @OAcoating that does not mix with water, and the @SiO2 coat-ing, which is more dispersible. The state of agglomeration ofFe3O4@OA in water is a parameter that may be a bias inour cytotoxic analysis, as well as that of Yin et al.,28 dis-cussed above. Fe3O4@NH2 and Fe3O4@SiO2@NH2 NPs con-tain exposed amine groups, and Fe3O4@SiO2 NPs containexposed hydroxyl groups, both groups assuring good aque-ous dispersibility. The TEM images in Figure 1(b–d) confirmthis, and no particle stacking is observed.

Cell viability assays indicate that the presence of theNPs reduces cell viability significantly (p < 0.05), as com-pared to the cells that were not exposed. One possible ex-planation for this may be that these NPs are taken up bythe cells as a result of endocytosis, promoting apoptosis.29

Despite the fact that exposure to NPs reduces viability, wenote that those with silica coatings are less toxic than thosewith amine group coating, as shown in Figure 9. NP interac-tions with cells depend on their surface chemical properties,which may be explained by their interactions with compo-nents of the cell medium and the resultant modification ofthe NP surface.30 Thus, the NP coating has an importantrole to play in cellular functions, including cell growth,migration, differentiation, survival, and tissue organization.

Regarding safety assessments, it is increasingly evidentthat nanomaterials cannot be treated in the same manneras chemical compounds. Interactions between NPs and nu-trient medium components occur, and must have an impacton any subsequent cyto- and genotoxicity test results.

The growth factors, amino acids, and nutrients necessaryto support cell growth, contained in nutrient media, aredemonstrably adsorbed onto the NP surfaces, forming coro-nas.31 The variations of relative elemental concentrationswe determined by XPS (Table II), on NP contact with theculture medium, indicates corona formation prior to cellcontact. The influence of this interaction on cellularresponse varies with time, as illustrated in Figure 10. Theresults of such interaction remain to be identified by morespecific analysis, a study necessary to determine the specificchemical groups that influence cell viability.

Effect of culture mediumDuring viability assay, the SPIONs are added to Ham’sF-12K, which contains amino acids, inorganic salts, vitamins,etc., as described by the manufacturer.32 There are interac-tions reported between chemical groups on the NPs andvarious components of the medium in which they are sus-pended33,34; these interactions appear to influence the stateof agglomeration of the NPs and the materials adsorbed ontheir surfaces. The effective unit of interest in the cell-nano-particle interaction is, then, not the NP but the corona-cov-ered NP. It is the nature of the organization of the absorbedbiomolecules on the NP surface, and any subsequent changein their epitopes, that determines the initial biologicalresponse to the presence of NPs. Because nutrient media

FIGURE 9. Cellular viability as a function of duration, at a NP concen-

tration of 200 lg/mL. [Color figure can be viewed in the online issue,

which is available at wileyonlinelibrary.com.]

FIGURE 10. Cellular viability after 72 h of exposure of NPs left in the culture medium for 2 (N) and 24 h (NT) before contact with the cells. [Color

figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

8 MBEH ET AL. IN VITRO BIOCOMPATIBILITY OF FUNCTIONALIZED MAGNETITE NANOPARTICLES

differ, the interactions we identified must be specific toHam’s F-12K33; this is presently being explored.

In the present study, we examined the effect of a widerange of NP concentrations on A549 cells. LDH and IL-6essays show that our NPs have low toxicity overall, becom-ing toxic to A549 cells at concentrations above 50 lg/mL.The MTT assay shows that, at a concentration of 200 lg/mL, no difference is observed after 24 h exposure of thethree types of NPs, but after 48 and 72, the amine-coatedNPs (Fe3O4@NH2 and Fe3O4@SiO2@NH2) show a decreasingviability while the silica-coated NPs (Fe3O4@SiO2) showsbetter viability. In addition, the MTT assay shows that allthe NPs suspended in culture medium for 2 h before expo-sure to A-549 cells are less toxic than when they are sus-pended for 24 hours, suggesting a time-dependent surfacechemical change at the NP surface.

As mentioned earlier, a good correlation already existsbetween the ELISA and LDH results (Figs. 7 and 8, bothcontaining results for 100 lg/mL for 24 h). We note thatthe percentages of cell viability obtained with the MTT testshowed a remarkably similar significant decrease in viabilityafter 24 hours in contact with all three types of NPs, at aconcentration of 200 lg/mL (Fig. 9). We have every reasonto believe that the trends indicated in Figures 7 and 8 willpersist to a concentration of 200 lg/mL.

CONCLUSIONS

We have determined the cytotoxicities of several functional-ized magnetite NPs, including Fe3O4@NH2, Fe3O4@SiO2, andFe3O4@SiO2@NH2. All show cytotoxic effects at concentra-tions above 50 lg/mL. The first, and most readily noticea-ble, effect is the alteration of cell shape, but there are alsomitochondrial dysfunction, an increase of LDH leakage, andIL-6 secretion. These cellular responses depend on severalparameters, including:

1. Surface functionalization: the MTT assay reveals function-alization-dependent cytotoxicity of NPs towards epithelialcells.

2. Corona formation: XPS analysis has shown corona forma-tion on the NP surface on suspension in culture media.

3. A dynamic interaction between NP and culture medium:the time of suspension has an impact on cell response.

Further studies are in progress to identify the specificbimolecular interaction with the functionalized surfaces andtheir effect on subsequent cell contact.

REFERENCES1. Sahoo SK, Parveen S, Panda JJ. The present and future of nano-

technology in human health care. Nanomed Nanotechnol Biol

Med 2007;3:20–31.

2. Linazasoro G. Potential applications of nanotechnologies to Par-

kinson’s disease therapy. Parkinsonism and Relat Disord 2008;14:

383–392.

3. Silva GA. Neuroscience nanotechnology: Progress, opportunities

and challenges. Nat Rev Neurosci 2006;7:65–74.

4. Turk V, Kaisera C, Schalle S. Invisible but tangible? Societal

opportunities and risks of nanotechnologies. J Cleaner Prod 2008;

16:1006–1009.

5. Singha N, Jenkinsa GJS, Asadib R, Doak SH. Potential toxicity of

superparamagnetic iron oxide nanoparticles (SPION). Nano Rev

2010;1:5358–6032.

6. Na HB, Song IC, Hyeon T. Inorganic nanoparticles for MRI con-

trast agents. Adv Mater 2009;21:2133–2148.

7. Zabow G, Dodd S, Moreland J, Koretsky A. Micro-engineered

local field control for high-sensitivity multispectral MRI. Nature

2008;2008:1058–1063.

8. Park MV, Annema W, Salvati A, Lesniak A, Elsaesser A, Barnes C,

McKerr G, Howard CV, Lynch I, Dawson KA and others. In vitro

developmental toxicity test detects inhibition of stem cell differen-

tiation by silica nanoparticles. Toxicol Appl Pharmacol 2009;240:

108–116.

9. Suh WH, Suslick KS, Stucky GD, Suh YH. Nanotechnology, nano-

toxicology and neuroscience. Prog Neurobiol 2009;87:133–170.

10. Maiorano G, Sabella S, Sorce B, Brunetti V, Malvindi MA, Cingo-

lani R, Pompa PP. Effects of cell culture media on the dynamic

formation of protein_nanoparticle complexes and influence on

the cellular response. ACS Nano 2010;4:7481–7491.

11. Cedervall T, Lynch I, Lindman S, Berggard T, Thulin E, Nilsson H,

Dawson KA, Linse S. Understanding the Nanoparticle protein co-

rona using methods to quantify exchange rates and affinities of

proteins for nanoparticles. Proc Natl Acad Sci 2007;104:2050–2055.

12. Unfried K, Albrecht C, Klotz L-O, Von Mikecz A, Grether-Beck S,

Schins RPF. Cellular responses to nanoparticles: Target structures

and mechanisms. Nanotoxicology 2007;1:52–71.

13. Klaine SJ, Alvarez PJJ, Batley GE, Fernandes TF, Handy RD, Lyon

DY, Mahendra S, McLaughlin MJ, Lead JR. Nanomaterials in the

environment: Behaviour, fate, bioavailability and effects. Environ

Toxicol Chem 2008;27:1825–1851.

14. Zhang X, Clime L, Roberge H, Normandin F, Yahia L’H, Sacher E,

Veres T. pH-triggered doxorubicin delivery based on hollow nano-

porous silica nanoparticles with free-standing superparamagnetic

Fe3O4 cores. J Phys Chem 2011;115:1436–1443.

15. Woo K, Hong J, Choi S, Lee HW, Ahn JP, Kim CS, Lee SW. Easy

synthesis and magnetic properties of iron oxide nanoparticles.

Chem Mater 2004;16:2814–2818.

16. Moret S. ‘‘Synthese et modification chimique de la surface des

nanoparticules de magh�emite �a des fins d’applications biom�edi-

cales ‘‘, PH.D. dissertation, -France,. Bordeaux: Universit�e de Bor-

deaux I; 2002. 221 p.

17. Goffin AL, Duquesne E, Raquez JM, Miltner HE, Ke X, Alexandre

M, Van Tendeloo G, Van Mele B, Dubois P. From polyester graft-

ing onto POSS nanocage by ring-opening polymerization to high

performance polyester/POSS nanocomposites. J Mater Chem

2010;20:9415–9422.

18. Oberdorster G, Maynard A, Donaldson K, Castranova V, Fitzpa-

trick J, Ausman K, Carter J, Karn B, Kreyling W, Lai D, Olin S,

Monteiro-Riviere N, Warheit D, Yang H. Principles for characteriz-

ing the potential human health effects from exposure to nanoma-

terials: Elements of a screening strategy. Part Fibre Toxicol 2005;

2:8.

19. Beasley AM, Auffarth GU, Von Recum AF. Intraocular lens

implants: A biocompatibility review. J Invest Surg 1996;9:399–413.

20. Hulander M, Hong J, Andersson M, Gerven F, Ohrlander M, Teng-

vall P, Elwing H. Blood interactions with noble metals: Coagula-

tion and immune complement activation. ACS Appl Mater Interf

2009;1:1053–1062.

21. Thorek DLJ, Tsourkas A. Size, charge and concentration depend-

ent uptake of iron oxide particles by non-phagocytic cells. Bioma-

terials 2008;29:3583–3590.

22. Oh WK, Kim S, Choi M, Kim C, Jeong SY, Cho BR, Hahn JS, Jang

J. Cellular uptake, cytotoxicity, and innate immune response of

silica titania hollow nanoparticles based on size and surface func-

tionality. ACS Nano 2010;4:5301–5313.

23. Kroll A, Pillukat MH, Hahn D, Schnekenburger J. Current in vitro

methods in nanoparticle risk assessment: Limitations and chal-

lenges. Eur J Pharm Biopharm 2009;72:370–377.

24. Bulte JWM. In vivo MRI cell tracking: Clinical studies. Am J

Roentgenol 2009;193:314–325.

25. Kostura L, Kraitchman DL, Mackay AM, Pittenger MF, Bulte JW.

Feridex labeling of mesenchymal stem cells inhibits

REVIEW ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | MONTH 2012 VOL 9999A, ISSUE 00 9

chondrogenesis but not adipogenesis or osteogenesis. NMR

Biomed 2004;17:513–517.

26. Asharani PV, Mun GLK HM, Valiyaveettil S. Cytotoxicity and geno-

toxicity of silver nanoparticles in human cells. ACS Nano 2009;3:

279–290.

27. Christopher RA, Judge SR, Vincent PA, Higgins PJ, McKeown-Lon

PJ. The amino-terminal matrix assembly domain of fibronectin

stabilizes cell shape and prevents cell cycle progression. J Cell

Sci 1999;112:3225–3235.

28. Yin H, Too HP, Chow GM. The effects of particle size and surface

coating on the cytotoxicity of nickel ferrite. Biomaterials 2005;26:

5818–5826.

29. Gupta A.K, Gupta M. Cytotoxicity suppression and cellular uptake

enhancement of surface modified magnetic nanoparticles. Bioma-

terials 2004;26:1565–1573.

30. Gupta AK, Curtis ASG. Lactoferrin and ceruloplasmin derivatized

superparamagnetic iron oxide nanoparticles for targeting cell sur-

face receptors. Biomaterials 2004;25:3029–3040.

31. Lynch I, Dawson KA. Protein-nanoparticle interactions. Nano-

Today 2008;3:40–47.

32. ATCC. F-12K Medium Formulation. In: Collection ATC, editor.

http://www.atcc.org/Attachments/4892.pdf. Manassas USA;

2004.

33. Buford MC, Hamilton RF Jr, Holian A. A comparison of dispersing

media for various engineered carbon Nanoparticles. Part Fibre

Toxicol 2007;4:6.

34. Allouni ZE, Cimpan MR, Hol PJ, Skodvin T, Gjerdet NR. Agglom-

eration and sedimentation of TiO2 nanoparticles in cell culture

medium Colloids Surf B Biointerfaces 2009;68:83–87.

10 MBEH ET AL. IN VITRO BIOCOMPATIBILITY OF FUNCTIONALIZED MAGNETITE NANOPARTICLES