In vitro studies with mammalian cell lines and gum arabic-coated magnetic nanoparticles


  • Ana Bicho,

    Corresponding author
    1. REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    • REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal.
    Search for more papers by this author
  • Ana Cecília A. Roque,

    Corresponding author
    1. REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    • REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal.
    Search for more papers by this author
  • Ana Sofia Cardoso,

    1. REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    Search for more papers by this author
  • Patrícia Domingos,

    1. REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    Search for more papers by this author
  • Íris Luz Batalha

    1. REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
    Search for more papers by this author

  • This article is published in Journal of Molecular Recognition as a special issue on Affinity 2009, edited by Gideon Fleminger, Tel-Aviv University, Tel-Aviv, Israel and George Ehrlich, Hoffmann-La Roche, Nutley, NJ, USA.


Iron oxide magnetic nanoparticles (MNPs) were synthesized by the chemical co-precipitation method and coated with gum arabic (GA) by physical adsorption and covalent attachment. Cultures of mammalian cell lines (HEK293, CHO and TE671) were grown in the presence of uncoated and GA-coated MNPs. Cellular growth was followed by optical microscopy in order to assess the proportion of cells with particles, alterations in cellular density and the presence of debris. The in vitro assays demonstrated that cells from different origins are affected differently by the presence of the nanoparticles. Also, the methods followed for GA coating of MNPs endow distinct surface characteristics that probably underlie the observed differences when in contact with the cells. In general, the nanoparticles to which the GA was adsorbed had a smaller ability to attach to the cells' surface and to compromise the viability of the cultures. Copyright © 2010 John Wiley & Sons, Ltd.


Over the past 10 years, the synthesis and modification of iron oxide superparamagnetic nanoparticles (MNPs) have been greatly improved due to their interesting new applications in biosensing, targeted drug delivery, contrast agents in magnetic resonance imaging (MRI), tissue repair, hyperthermia and cell separation, among others (Laurent et al., 2008). The surface modification of MNPs with various organic, inorganic and biological layers has proved to increase stabilization against aggregation in both biological medium and magnetic field, as well as to improve their biocompatibility and bioactivity for medical applications (Banerjee and Chen, 2008a). Natural polymer systems as dextran, alginate, pullulan and chitosan have shown to be particularly interesting for nanoparticle stabilization, increased biocompatibility and surface functionality (Häfeli and Pauer, 1999; Gupta and Gupta, 2005; Sasaki et al., 2008). Gum arabic (GA) is a naturally occurring exudate from Acacia senegal and Acacia seyal trees, and possesses a highly branched polysaccharide structure consisting of a complex mixture of potassium, calcium and magnesium salts from arabic acid, with residues of galactose, rhamnose, glucuronic acid and arabinose (Kattumuri et al., 2007). The most important applications of GA are in the food and pharmaceutical industries where it is used as an excellent emulsifier (Yadav et al., 2007), therefore being recognised as non-toxic for humans (Ali et al., 2009). Recently, GA has been regarded as an alternative biopolymer for the coating and stabilization of nanostructures, including carbon nanotubes (Bandyopadhyaya et al., 2002; Kumar et al., 2008), gold nanoparticles (Kattumuri et al., 2007) and iron oxide nanoparticles (Williams et al., 2006; Banerjee and Chen, 2007a, b; Banerjee and Chen, 2008a, b; Roque and Wilson, 2008; Wilson et al., 2008). In fact, GA showed to be an effective agent to improve colloid stability, enabling for example the production of a magnetic nanocarrier for targeted anticancer drug delivery by grafting cyclodextrin onto GA-modified MNPs (Banerjee and Chen, 2008a). However, there is not much evidence on how GA-coated MNPs interact with mammalian cells in vitro and in vivo. A recent article by Wilson Jr and co-workers (Wilson et al., 2008) reported preliminary studies on the cytotoxicity and cell interactions between GA-treated and untreated iron oxide MNPs and L929 fibroblast cells.

In this work, we aimed to further explore the effect of GA-coated MNPs on cellular growth of mammalian cell lines of different origin and with distinct properties. MNPs were synthesized, functionalized with GA by adsorption and by covalent coupling and further used for in vitro assays with the mammalian cell lines HEK293, CHO and TE671. The cell cultures were grown in the presence of the MNPs for defined periods of time. The current study demonstrates that GA-coated MNPs prepared by distinct methods cause different effects on the cells. Moreover, the different cell lines respond differently to the presence of the particles.



Iron (III) chloride hexahydrate, Iron (II) chloride tetrahydrate and Ammonium hydroxide solution 25% were purchased from Fluka. (3-Aminopropyl)triethoxysilane (APTES), N′,N″-dimethylformamide (DMF), dimethyl-sulfoxide (DMSO), glutaraldehyde solution (50 wt% in H2O), GA and KBr were acquired from Sigma-Aldrich. Ethanol absolute was purchased from Panreac. Culture media, foetal bovine serum, antibiotics and fungizone were purchased from Gibco. All cell culture disposable materials were from Sarstedt. All chemicals were of the highest purity available.


The cell cultures were grown in a CO2 and temperature controlled Sanyo MCO-17AC incubator. All manipulations in sterility were performed inside a Sanyo Clean Bench laminar flow chamber. An Olympus BX51 microscope was used for monitoring cellular growth and the presence of debris (400× magnification).

Synthesis and modification of MNPs

The MNPs were synthesized by the co-precipitation method as to produce bare magnetite (MNP). GA (40 mg/ml) was adsorbed on to the surface of MNPs (MNP_GAADS) as described elsewhere (Roque et al., 2009). GA (40 mg/ml) was also covalently bound to the MNPs modified by two different routes: (i) glutaraldehyde-activated MNPs reacted with GA via their free amine groups (MNP_GAAPTS) and (ii) amine-modified MNPs reacted with GA via their free carboxyl groups through EDC reaction (MNP_GAEDC) (Roque et al., 2009). All particles were characterized by dynamic light scattering, Fourier transformed infrared spectroscopy and transmission electron microscopy.

Maintenance of cell cultures

HEK293 (Human Embryonic Kidney, ECACC No. 85120602) and TE671 (Human Caucasian rhabdomyosarcoma, ECACC No. 89071904) cells were grown in D-MEM culture medium (Dulbecco´s Modified Eagle Medium, Gibco) with 4500 mg/L glucose and GlutaMAX™ I, and without pyruvate. CHO cells (Hamster Chinese Ovary, ECACC No. 85050302) were grown in F-12 medium with glutamine. All culture media were enriched with 10% foetal bovine serum and supplemented with 50 IU/ml penicillin and 50 UG/ml streptomycin to prevent bacterial infections. Sub-culture was performed by trypsinization when cellular growth reached approximately 70% confluence. Cultures were grown in solid supports and kept at 37°C in a humidified atmosphere with 5% CO2 and 95% air.

Assays of MNPs with HEK293 cells at different incubation times

To prevent contamination of the cell cultures, the nanoparticles were washed three times with autoclaved PBS buffer supplemented with penicillin (50 IU/ml), streptomycin (50 UG/ml) and fungizone (2.5 µg/ml). Stock solutions of particles in buffer solution (1 mg/ml) were stored at 4°C. For the in vitro assays the HEK293 cell cultures were sub-cultured onto 13 mm diameter coverslips placed inside cell culture dishes. With this procedure, we ensured that all cell culture samples from one dish were grown in defined conditions. For each assay, the final concentration of particles in the culture medium was 0.017 mg/ml (or alternatively GA 0.67 mg/ml). The cells were allowed to attach and form a monolayer before adding the nanoparticles to the bathing media. At different times, ranging from 30 min up to 30 h, a coverslip sample was removed for further observation under the microscope. Prior to the observation, each coverslip was washed extensively in PBS by gentle shaking for 5 min to remove the excess of particles. Optical analysis of the samples was performed in Phase contrast under an Olympus BX51 microscope. The images were monitored with Cell F View Image System Software and photographs were taken from 6 to 10 random fields of each sample. The percentage of cells with particles was determined by direct cell count. The cellular density of the cultures was compared to that of the control cells. The assay was performed independently three times (n = 3).

Assays of MNPs with different cell lines (HEK293, CHO and TE671)

These assays were performed as described above using HEK293, CHO and TE671 cell lines (n = 5). The microscopy observations were made at 24 and 30 h.


Iron oxide MNPs were coated with GA by different techniques. One method involved the coating by adsorption at the surface of the bare magnetite, yielding MNP_GAADS particles. The mechanism of GA adsorption onto MNPs has been attributed to interactions between carboxylate groups in GA and the surface of bare magnetite (Roque and Wilson, 2008). The free functional amine and carboxylate groups from GA were also exploited as anchoring points for the covalent coupling of the polymer at the surface of MNPs, resulting in MNP_GAAPTS and MNP_GAEDC, respectively. The interest in coating MNPs with GA by covalent interactions resides in an attempt of creating a more stable magnetic shell. MNPs coated with GA by different methods presented distinct properties which are summarized in Table 1. The diameter of the magnetic core, as given by TEM analysis, was not affected by the surface modifications performed on bare magnetite. Also, the zeta potential values for all the particles tested were within the same range, demonstrating the existence of negatively charged surfaces with relative colloidal stability. However, the hydrodynamic diameter differed between samples. The particles to each GA was added after the co-precipitation of magnetite yielded lower hydrodynamic diameters, showing the dispersing effect of GA. FTIR analysis on the samples confirmed the presence of GA on the MNP surface. Previously analysis (Roque et al., 2009) showed that the amount of GA bound to the MNPs by the different coating methods yielded about 1 g GA/g support.

Table 1. Characterization of the MNPs prepared for the in vitro cellular studies
Particle typeDiametera (nm)Hydrodynamic diameterb (nm)Zeta potentialc (mV)FTIR bandsd (cm−1)
  • a

    Diameter of the magnetic core given by transmission electron microscopy.

  • b

    Hydrodynamic diameter given by dynamic light scattering analysis.

  • c

    Zeta potential given by dynamic light scattering analysis at neutral pH.

  • d

    Fourier transform infrared spectroscopy characteristic peaks (Roque et al., 2009).

MNP11 ± 3708 ± 192−22 ± 33443–3450
MNP_GAADS14 ± 1457 ± 66−22 ± 11032–1076
MNP_GAAPTS11 ± 3567 ± 139−25 ± 11032–1076
MNP_GAEDC12 ± 3319 ± 41−21 ± 11032–1076

HEK293 cell cultures were grown in the presence of non-coated (MNP) or coated particles (MNP_GAADS, MNP_GAAPTS and MNP_GAEDC). Cells showing MNPs at their surface could be observed after only 30 min of incubation (the shortest period tested), which confirms the great affinity of the magnetic particles towards the cells (Berry et al., 2004; Gupta and Curtis, 2004; Wilson et al., 2008). Our primary goal was to assess if the GA coating could change the way MNPs interfere with the cell cultures and affect their growth. We analysed samples taken at different incubation periods, from 30 min up to 30 h, and followed, within each sample, the proportion of cells presenting nanoparticles. Figure 1 shows typical phase contrast photographs of the cell cultures exposed to MNPs after 6 or 24 h. The percentage of cells with MNPs at their surface was determined and was represented as a function of the incubation time (Figure 2). The indicators chosen to assess cellular viability were the cellular density of each sample and the presence of cellular debris as compared to the control (cells grown in the absence of particles) and the results are presented in Table 2. Interestingly, we observed differences between the four types of particles tested. Bare magnetite (MNP) and GA covalently coupled particles displayed a similar behaviour: after 30 min of incubation about 10% of the cells presented MNPs, but between 1 and 3 h the number increased to 40–50%, and to more than 60% after 6 h of exposure. This was followed by a decrease in cellular density (Table 2). For these particles the reduction of the cellular density seems to correlate well to the growing number of cells with particles. In accordance to this, after the careful PBS buffer washing preceding the microscopic observations, we observed floating clusters of cells released from the coverslips and which were covered with particles, an additional indication of cellular damage for the longer periods of incubation tested. Nevertheless, on the samples with MNP_GAEDC we could detect the presence of cellular debris (which resembled tire marks on the surface of the coverslips) as early as 3 h of incubation while for bare magnetite this was only observed for the longest incubation of 30 h. Differently, MNPs coated with adsorbed GA (MNP_GAADS) showed the least appetence to interact with the cells, accounting for less than 50% of cells with the MNPs for periods of incubation up to 24 h. Also, no cellular debris were observed during the time course of the experiment (Table 2). This feature is distinctive from the other tested MNPs.

Figure 1.

Phase contrast photographs of HEK293 cells grown in the absence (control) and in the presence of bare magnetite (MNP) and GA modified particles (MNP_GAADS, MNP_GAAPTS and MNP_GAEDC). The incubation times were 6 and 24 h.

Figure 2.

Percentage of HEK293 cells with particles, at different incubation periods and for each particle type.

Table 2. Changes in cellular density (C) and occurrence of cellular debris (D) on HEK293 cell cultures at different incubation times, when compared to cells grown in the absence of particles
  Time (h)
  1. Key: (−), decrease in cellular density; (+), presence of debris.

 D    +
 D   + 

In order to determine if the MNPs could interact differently with cells of different proveniences we have grown three cell lines in the presence of the particles for a comparative study. The three cell types were chosen for their fast growing and easy maintenance. In addition to the HEK293 cells we have chosen the alternative CHO cell line which also shows epithelial morphology. For comparison we have also tested the MNPs on TE671 cells derived from a human rhabdomyosarcoma and which grow with a spindle shaped morphology. Samples were examined at 24 and 30 h of incubation and representative phase contrast microscopy photographs are shown in Figure 3. As with the HEK293 cells, the four types of particles tested also demonstrated the ability to deposit and attach onto the surface of the CHO and TE671 cells (Figures 3 and 4). The two epithelial cell lines had a similar behaviour in the presence of all the MNPs on what concerns the amount of cells with particles and the generalized decrease of the cellular density of the cultures (Table 3). CHO cultures showed a higher incidence of cellular debris than HEK293. Most likely HEK293 clusters of non-viable cells release from the coverslips, while CHO cells, even when completely covered with particles, remained attached to the growing substrate during the PBS washing procedures. As with HEK293 cells, MNP_GAADS demonstrated less ability to attach to CHO and TE671 cells, as compared to the other MNP types. In particular, less than 35% of TE671 cells presented particles at their surface after a period of 24 h of incubation. With the exception of MNP_GAEDC, all other tested particles are less effective on attaching to the surface of TE671 cells when compared to the results obtained with HEK293 and CHO cells. The observations made from the TE671 cultures at 24 and 30 h show that almost all cells present the MNP_GAEDC. Their negative influence on cell survival was also confirmed by the presence of cellular debris on all cultures and by a decrease on cellular density in comparison to the control (Table 3). These results show a variability of effects between distinct cellular types and between the different particles. Previous studies by other groups also suggested a possible variety of responses by different cell types. Häfeli and Pauer observed different toxicity levels caused by PLA, dextran and polystyrene-coated magnetic microspheres between two cell lines derived from human prostate cells and murine lymphoma (Häfeli and Pauer, 1999).

Figure 3.

Phase contrast photographs of HEK293, CHO and TE671 cells grown in the absence (control) and presence of bare magnetite (MNP) and GA modified particles (MNP_GAADS, MNP_GAAPTS and MNP_GAEDC). The incubation time was 24 h.

Figure 4.

Percentage of cells with particles after 24 h of incubation with the different types of particles. The cell cultures tested were (A) HEK293, (B) CHO and (C) TE671.

Table 3. Changes in cellular density (C) and occurrence of cellular debris (D) on three different cell cultures (HEK293, CHO and TE671), when compared to cells grown in the absence of particles
24 h30 h24 h30 h24 h30 h
  1. Key: (−), decrease in cellular density; (+), presence of debris.

 D +++  
 D  ++  
 D +++ +

To determine if GA by itself could be responsible for the observed results we have dissolved it directly in cell cultures medium at a final concentration of 0.67 mg/ml (over 10 times in excess to the maximum of GA present in the MNPs samples used for the cell assays) and tested it directly on the cell cultures. The samples observed after 24 h of exposure showed similar cellular density to the control (cultures grown in the absence of GA) and the total absence of cellular debris as expected (Al-Mosawi, 2006). Therefore, it seems unlikely that the differences we observed between particles are directly related to the polymer. Supporting evidence for this is the similarity of effects between the non-coated particles and the MNPs coated with GA by covalent interactions. Moreover, our results suggest that different methods for GA coupling at the surface of MNPs produce particles with distinct surface characteristics that probably underlie the observed differences when in contact with the cells.

Wilson and co-workers assessed the cytotoxicity of MNPs coated with GA using L929 fibroblast cells after 24 h. No differences were found between coated and non-coated particles, and a common survival rate of 90% using the live/dead assay (Wilson et al., 2008). Their results, based on light microscopy assays, indicated that the GA-treated particles were located at the cell membrane. However, some aspects of this report differ from the present study. Firstly, the MNPs were co-precipitated in the presence of GA yielding particles with about 1 µm (Wilson et al., 2008). In our work, we studied different cell lines and the GA functionalization was performed after MNP synthesis, yielding particles with smaller hydrodynamic diameters. We observed that the only particles which showed a clear distinct behaviour from the MNP were the MNP_GAADS. It is known that particle size influences the way cells respond to the presence of MNPs (Gupta and Gupta, 2005). It is generally acknowledged that particles with small diameters (<100 nm) may be easily taken in by different types of cells. All particles used here have a hydrodynamic diameter greater than 300 nm and most likely they do not enter the cells, but only adhere to the cell surface (Wilson et al., 2008). GA, being a highly branched polysaccharide which primarily consists of galactose, has the ability to interact with asialoglycoprotein receptors of hepatocytes (Groman et al., 1995). Although we do not exclude the possibility of receptor-mediated GA recognition, we have no knowledge of the existence of such receptors in any of the cells used in our study.


GA was used as a coating material for iron oxide MNPs following different coupling strategies, which ranged from physical adsorption to covalent coupling. Chemical modifications performed on the biopolymer lead to MNPs with distinct properties when contacted with cell cultures. The in vitro assays performed on three mammalian cell lines showed that all particles tested attach to the surface of the distinct cell types, but at different speeds and with different effects on cell survival. In general, the percentage of cells presenting particles at their surface was higher for those incubated in the presence of MNP and MNP_GAEDC, followed by MNP_GAAPTS and lastly by MNP_GAADS. However, in terms of decrease in cellular density and presence of cellular debris, the reversed order was observed. Also, the particles which required longer periods to attach to the surface of the cells (MNP_GAADS) caused less damage on the cell cultures for similar periods. GA per se does not appear to be the determining factor for the degree of toxicity of the particles. These preliminary studies are encouraging for developing improved cytotoxicity assays to better elucidate and quantify the effect of MNPs on several mammalian cell lines. Equally, TEM micrographs of cells incubated with MNPs will provide accurate information on the cellular localization of MNPs. With these data it will be possible to clarify which particle type will be more suited for different applications, ranging from targeted drug delivery to cell labelling, as the requirements for toxicity and MNP adhesion onto cells will be distinct.


The authors thank Dr Paulo Lemos (REQUIMTE, FCT/UNL) for advice in the microscopy assays and Dr Abid Hussain (REQUIMTE, FCT/UNL) for careful reading and commenting of the paper. They thank the financial support from Fundação para a Ciência e Tecnologia (FCT), through contract PTDC/BIO/65383/2006, and to British Council and Conselho de Reitores das Universidades Portuguesas through Treaty of Windsor Anglo-Portuguese Joint Research Programme (Acção Integrada Luso-Britânica B29/07 and B37/08).