European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137

www.elsevier.com/locate/ejpb

Research paper

Effect of cell media on polymer coated superparamagnetic

iron oxide nanoparticles (SPIONs):
Colloidal stability, cytotoxicity, and cellular uptake studies
a,*,1
Alke Petri-Fink , Benedikt Steitz a,b,1, Andrija Finka a,c,1
, Jatuporn Salaklang a,
Heinrich Hofmann a
a
Laboratory of Powder Technology, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland
b
Microsystems Laboratory 4, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland
c
Regenerative Medicine and Pharmacobiology Laboratory, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland

Received 15 November 2006; accepted in revised form 2 February 2007

Available online 13 July 2007

Abstract

The influence of the composition of the polymer coated polyvinyl alcohol (PVA), vinyl alcohol/vinyl amine copolymer (A-PVA) and
polyethylenimine (PEI) coated superparamagnetic iron oxide nanoparticles (SPIONs) on the colloidal stability, cytotoxicity and cellular
uptake of these particles in different cell media is reported in this paper. Although all examined polymer coated SPIONs were stable in
water and PBS buffer these colloidal systems had different stabilities in DMEM or RPMI media without and supplemented with fetal calf
serum (FCS). We found that A-PVA coating onto the surface of the SPIONs decreased the cytotoxicity of the polymer compared to the
same concentration of A-PVA alone. As well, polyplexes of PEI-SPIONs with DNA in concentration used for transfection experiments
showed no cytotoxicity compared to PEI and PEI-SPIONs. Our data show that the choice of medium largely influences the uptake of
these particles by HeLa cells. The optimal medium is different for the different examined polymer coated SPIONs and it should be deter-
mined in each case, individually.
Ó 2007 Elsevier B.V. All rights reserved.

Keywords: Superparamagnetism; Iron oxide; Nanoparticles; PVA; PEI; Colloidal stability; Cytotoxicity

1. Introduction fields of application. Beside the unique physical properties

induced by surface or quantum effects, the size of the pri-
The synthesis and characterization of nanoparticles have mary particles is with 2–30 nm comparable to the size of
been a focus of intensive research for more than 10 years biological building blocks and allows investigation of the
since they play an important role in electronics, catalysis, cellular functioning or direct interaction with biological
biology and medicine. Nanoparticles that can be biochem- targets. Well described examples are inorganic photolumi-
ically functionalized are potential medical device systems nescent particles (quantum dots) as markers or nanosized
that can be used in many different biological and medical superparamagnetic iron oxide particles. Superparamagnet-
ic particles with a diameter of around 10 nm have been
used for many years for e.g. non-viral gene delivery, as
*
Corresponding author. Laboratory of Powder Technology, Ecole MRI contrast agents, or for typical separation applications
Polytechnique Fédérale de Lausanne (EPFL), EPFL-STI-IMX-LTP, [1]. Those applications are relatively well established on the
Station 12, 1015 Lausanne, Switzerland. Tel.: +41 0 21 693 5107; fax:
+41 021 693 3089.
market, whereas targeted particle delivery is still in the
E-mail address: alke.fink@epfl.ch (A. Petri-Fink). development stage. It has been shown that magnetic parti-
1
These authors contributed equally to this work. cles are physiologically well tolerated and that the surface

0939-6411/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejpb.2007.02.024
130 A. Petri-Fink et al. / European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137

of the particles is responsible for the biocompatibility and 15 min (MowiolÒ 3-83) to 4 h (M12) at 90 °C, and filtering
stability to the reticulo-endothelial system [2]. On the other the hot solutions over paper filters (Schleicher & Schuell
hand, concerns have been raised about nanoparticle- AG). Ultra-pure deionized water (Seralpur delta UV/UF
derived adverse health effects and scientists currently aim setting, 0.055 lS/cm) was used in all synthesis steps. D-
at developing a rational, science-based approach to nano- 9527 Sigma cellulose membrane dialysis tubing with a
toxicology. Recent investigations were focused on rela- molecular weight cut-off at 12,000 was used for dialysis.
tively simple cytotoxic tests as readily available pre-
screening methods, on the effect of inhaled or instilled 2.2. Iron oxide nanoparticles
ambient nanoparticles that can induce oxidative stress or
pulmonary inflammation, or on possible consequences of Superparamagnetic iron oxide nanoparticles (SPIONs)
particle related dysfunction of the cardio-vascular system were prepared by alkaline co-precipitation of ferric and fer-
[3]. A key question is the assessment of nanoparticles’ rous chlorides in aqueous solution as described elsewhere
potential toxicity due to their nanosize nature whereas [9–11]. The obtained brown suspension was dialyzed
the type of particle does not seem to play an important role against 0.01 M nitric acid for two days, and stored at 4 °C.
[4,5]. The surface properties of the particles are particularly
investigated in more detail since the surface seems to be the 2.3. Polymer coated particles
determining factor for cell uptake and cytotoxicity.
Many applications and investigations use nanoparticles In order to obtain SPIONs coated with either polyvinyl
in colloidal suspensions. By tailoring interactions between alcohol (MowiolÒ 3-83) or with a mixture of polyvinyl
colloidal particles, one can design stable fluids, gels, or col- alcohol (MowiolÒ 3-83) and vinyl alcohol/vinyl amine
loidal crystals. Long range, attractive van der Waals forces copolymer (M12, vinyl alcohol/vinyl alcohol copolymer
are ubiquitous and must be balanced by Coulombic, steric, mass ratio = 45) the nanoparticle dispersion was mixed at
or other repulsive interactions to engineer the desired various ratios with the different polymer solutions. The
degree of colloidal stability [6]. The colloidal behaviour products will be referred to as PVA-SPION and A-PVA-
of nanoparticles in different cell media or body fluids is SPION in this work.
almost never considered or related to particle–cell interac- For PEI coating the iron oxide nanoparticles were
tions. Williams et al. investigated the impact of silica, sil- mixed at a PEI:Fe mass ratio of two (R = 2) with 25 kDa
ica/iron oxide, and gold nanoparticles on the growth and polyethylenimine (Aldrich). The samples will be referred
activity of Escherichia coli and correlated the results with to as PEI-SPION in this work. DNA-PEI-SPIONs were
dynamic light scattering experiments that were performed prepared at a N/P ratio (ratio of nitrogen-containing
in growth media [7]. It has been shown that submicron groups of the polymer to phosphate groups of the nucleic
polymeric particles coagulate in the cell medium, whereas acid) of 7.5, assuming the DNA was entirely complexed.
colloidal stability in aqueous solution was ensured for sev- The nitrogen content of PEI has been measured and calcu-
eral weeks [8]. lated according to Harpe et al. [12], and the phosphate con-
The aim of this work is to show in detail the influence of tent was calculated from the size of the plasmid (4.7 kb).
the composition of the polymer coating and surface charge The iron content of the suspensions was determined by
of nanoparticles on the colloidal stability of these particles redox-titration, essentially as described [13].
in different cell media. Furthermore, we tried to establish
correlations between cytotoxicity or uptake rates and 2.4. Transmission electron microscopy
agglomeration behaviour of the particles. This correlation
in particular is very important for the design of simplified Transmission electron microscopy (TEM) was per-
toxicity tests and for the further development of such par- formed using a Phillips CM-20 microscope operating at
ticles for in vivo applications. 200 kV. For sample preparation, dilute drops of suspen-
sions were allowed to dry slowly on carbon-coated copper
2. Materials and methods grids.

2.1. Materials 2.5. Colloidal stability

All chemicals were of analytical reagent grade and were The colloidal stability of coated particles in different
used without further purification. Polyvinyl alcohol PVA environments was investigated by turbidity measurements.
(MowiolÒ 3-83) with an average molecular weight (MW) Therefore, the particle dispersions were mixed with com-
of 14,000 g/mol and a hydrolysis degree of 83% was sup- monly used cell media such as RPMI and DMEM, both
plied by courtesy of CLARIANT. Vinyl alcohol/vinyl in presence and absence of 10% fetal calf serum (FCS)
amine copolymer M12, with an average MW of 80,000– thereby setting the iron concentration to 100 lg/ml. After
140,000 was supplied by courtesy of ERKOL. rapid homogenization, the turbidity was measured by light
Polymer solutions were prepared by dissolving the pow- absorption at a wavelength of 500 nm as a function of time
ders in water followed by rapidly heating the solutions for (t) [14].
 A. Petri-Fink et al. / European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137 131

2.6. Particle size lated. It is important to note that the different particles
were incubated in the corresponding media for 1 h prior
Light scattering measurements were carried out at 90° to cell exposure.
on a photon correlation spectrometer (PCS) from Brook-
haven equipped with a BI-9000AT digital autocorrelator. 3. Results
The CONTIN method was used for data processing. The
concentration of iron oxide nanoparticles was set to 3.1. Transmission electron microscopy
100 lg iron/ml for all measurements. The theoretical
refractive index of 2.42 of magnetite [15] was used to calcu- Transmission electron microscopy (TEM) was carried
late the number weighted distribution from the raw inten- out with individual and coated particles (Fig. 1). Primary
sity weighted data. Viscosity, refractive index and uncoated SPIONs showed ellipsoidal iron oxide particles
dielectric constant of pure water were used to characterize with an average size of 9 nm (Fig. 1a). Two perpendicular
the solvent. diameters were measured for 100 particles, and average axe
lengths of 10 ± 2 nm and 8 ± 2 nm were determined. An
2.7. Cell culture and treatments aspect ratio of 1.2 ± 0.2 was determined, corresponding
to the results obtained by other groups [15]. PVA-coated
HeLa (human cervix carcinoma cells) cells were grown iron oxide particles (PVA-SPION) were also observed by
in RPMI medium (Gibco-BRL), supplemented with 10% means of transmission electron microscopy (Fig. 1b).
fetal calf serum (FCS; Promochem) and 1% penicillin/ Although the polymer was responsible for image blurring
streptomycin. One day prior to experiments, the cells were (due to film formation) it seemed that the particle spatial
detached in trypsin–EDTA (Gibco-BRL) and grown in distribution was more homogeneous and that the minimum
complete medium in 48-well plates (Costar) at 104 cells distance between the particles was larger compared to the
per well. On the day of experiment, cells were washed with uncoated particles for equivalent concentrations. Iron
PBS and medium was changed. The dilutions of PVA- oxide particles embedded in a PEI polymer matrix (PEI-
SPION, A-PVA-SPION, PEI-SPION and DNA-PEI-SPION SPION) are shown in Fig. 1c. Each nanoparticle was asso-
were added for the concentration, time and temperature ciated with more than one strand of the PEI and, likewise,
indicated. After 22 h of incubation, the MTT assay was each strand of PEI attached to more than one nanoparticle,
performed to determine cell viability. resulting in a bridging aggregation. An average bead size of
27 ± 12 nm was obtained for sample R = 2 (PEI:Fe mass
2.8. Evaluation of cell viability ratio of 2) using 25 kDa PEI by measuring 360 beads by
transmission electron microscopy. Comparable PEI iron
MTT reduction was used to quantify metabolically oxide nanobeads using higher molecular weight
active cells. Briefly, medium was removed and cells were (800 kDa) PEI showed an average size of 200 nm [16].
exposed to 1 mg/ml MTT (3,4,5-dimethylthiazol-yl)-2,5-
diphenyl tetrazolium, Sigma in PBS for 2 h at 37 °C. 3.2. Turbidity measurements and photon correlation
After incubation, the cells were examined under an spectroscopy (PCS)
inverted microscope to ascertain the density of violet
spots corresponding to active mitochondria in order to The agglomeration of the different polymer coated
exclude a potential mitochondrial toxicity of the com- nanoparticles (PVA-SPION, A-PVA-SPION, and PEI-
pounds. The supernatant was then removed and the pre- SPION) was examined by turbidity measurement and pho-
cipitated formazan was dissolved in isopropanol and ton correlation spectroscopy (PCS). Uncoated SPIONs
quantified at 560 nm in a multiwell plate reader agglomerated in PBS and biological fluids immediately
(Saphire2, Tecan). Each experiment was repeated in trip- (data not shown). All investigated polymer coated SPIONs
licate wells at least three times. Means and standard devi- were absolutely stable in water and in PBS for months and
ations were calculated. over a pH range of 3–11 without showing agglomeration.
A-PVA-SPIONs and PVA-SPIONs showed no important
2.9. Iron determination agglomeration in none of the investigated media after
30 min as measured by PCS. However, a different behav-
The cell layer was dissolved in 6 N HCl (125 ll/well of a iour of the various examined polymer coated nanoparticles
48-well plate) for 1 h, then 125 ll of a 5% solution of with similar primary size in different media with respect to
K4[Fe(CN)6] (Merck) in H2O was added and the absor- their colloidal stability and hence their size was observed
bance was read after 10 min at 690 nm in a multiwell plate after 1 h. PVA-SPIONs showed high colloidal stability in
reader (Saphire2, Tecan). A standard curve using the differ- FCS supplemented media with respect to their size
ently coated iron oxide nanoparticles was recorded in the (Fig. 2a, PVA-SPIONs). In both FCS supplemented media,
same conditions to quantify the amount of cell-bound iron. we could not observe an increase in size within five days
Each experiment was repeated in triplicate wells at least (maximum time period observed) and the dispersions
three times. Means and standard deviations were calcu- remained transparent. Agglomeration of the particles in
132 A. Petri-Fink et al. / European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137

Fig. 1. (a) High resolution transmission electron micrograph showing slightly faceted crystalline particles in the 10 nm range. (b) Bright field TEM picture
showing a homogeneous spatial distribution of the PVA coated iron oxide particles. (c) Bright field transmission electron micrograph showing one PEI-
SPION bead.

PVA-SPIONs
a b PVA-SPION RPMI
PVA-SPION RPMI 0.08PVA-SPION RPMI + FCS
PVA-SPION DMEM
120 PVA-SPION RPMI + FCS PVA-SPION DMEM + FCS
PVA-SPION DMEM 0.06
number weighted mean

PVA-SPION DMEM + FCS

Absorbance [a.u.]

80 water 0.04
size[nm]

0.02
40

0.00
0
0 20 40 60 80 100 120 0 30 60
time [min] time [min]

A-PVA-SPIONs

A-PVA-SPION RPMI A-PVA-SPION RPMI

A-PVA-SPION RPMI + FCS A-PVA-SPION RPMI + FCS
160
number weighted meanr

A-PVA-SPION DMEM A-PVA-SPION DMEM

A-PVA-SPION DMEM + FCS A-PVA-SPION DMEM + FCS
Absorbance [a.u.]

120 0.02
size [nm]

water
80

40 0.00

0
0 30 60 90 120 0 30 60
time [min] time [min]

PEI-SPIONs
PEI-SPION R=2 RPMI
300 PEI-SPION R=2 RPMI PEI-SPION R=2 RPMI + FCS
PEI-SPION R=2 RPMI+FCS
number weighted mean

* PEI-SPION R=2 DMEM

250 1.6
PEI-SPION R=2 DMEM PEI-SPION R=2 DMEM +FCS
Absorbance [ a.u.]

* PEI-SPION R=2 DMEM+FCS

200
*for longer times sizes not determinable
size [nm]

150 1.4
0.08
100
0.04
50
0.00
0
0 20 40 60 80 100 120 0 10 20 30 40 50 60

time [min] time [min]

Fig. 2. Stability of PVA-SPIONs, A-PVA-SPIONs, and PEI-SPIONs in different cell media (RPMI, RPMI+FCS, DMEM, and DMEM+FCS)
monitored by (a) PCS measurements and (b) UV measurements. The error is not indicated to simplify matters.
 A. Petri-Fink et al. / European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137 133

DMEM and RPMI without FCS was fast and agglomerate A-PVA RPMI
mean sizes of over 80 nm were observed after 120 min. The a A-PVA RPMI + FCS
A-PVA DMEM
turbidity measurements were in good agreement with these 1.00 A-PVA DMEM + FCS

relative cell viability

observations. The agglomeration rate in media without
FCS was higher than in FCS supplemented media 0.75
(Fig. 2b; PVA-SPIONs). A-PVA-SPION nanoparticles
0.50
were relatively stable in DMEM with and without FCS
and comparatively less stable in RPMI with and without
0.25
FCS as obtained from turbidity measurements over 1 h
(Fig. 2a; A-PVA-SPIONs). This was confirmed by PCS 0.00
measurements (Fig. 2b; A-PVA-SPIONs). The size of the 0 200 1200
A-PVA-SPIONs in DMEM with and without FCS was polymer [μg/ml]
not increasing significantly within the first hour and mea-
sured sizes around 40 nm were obtained. Agglomeration b A-PVA-SPION RPMI
A-PVA-SPION RPMI + FCS
of the A-PVA-SPIONs was observed in DMEM after 1 h A-PVA-SPION DMEM
A-PVA-SPION DMEM + FCS
and the diameter increased to 110 nm after 2 h, whereas 1.00

relative cell viability

the particles in DMEM supplemented with FCS kept their
0.75
initial size up to nine days (maximum time observed).
In summary, we could not observe that the presence of
0.50
serum in the medium induced agglomeration of the PVA-
SPIONs and A-PVA-SPIONs. It seems that the combina- 0.25
tion of cell media and serum determines the colloidal
stability. The agglomeration behaviour of PEI-SPIONs in 0.00
these four cell media differed considerably from the results 0 200 1200
described above (Fig. 2a and b; PEI-SPIONs). PEI-SPION polymer [ μg/ml]
beads agglomerated immediately in the presence of fetal Fig. 3. (a) Relative cell viability of HeLa cells exposed to differently
calf serum whereas they remained stable in all media with- concentrated (a) aqueous solutions of A-PVA in different cell media
out FCS and no change in the turbidity and bead size could (RPMI, RPMI+FCS, DMEM, and DMEM+FCS) and (b) dispersions of
be observed within 2 h. However, the particles interacted A-PVA-SPIONs in different cell media (RPMI, RPMI+FCS, DMEM,
and DMEM+FCS).
differently with the media without FCS. In DMEM, high
turbidity and larger particle sizes (70 nm) of PEI-SPIONs
were observed after 24 h, whereas in RPMI the initial par- copolymer (M12, vinyl alcohol/vinyl alcohol copolymer)
ticle size remained unchanged during 2 days. at a mass ratio of 45. Low concentration of A-PVA-SPI-
ONs (20 lg Fe/ml) had no influence on cell viability in
3.3. Biological evaluations any of the media, whereas high concentrations
(120 lg Fe/ml) showed only pronounced cytotoxicity in
We assessed cytotoxicity of PVA-, A-PVA-, PEI-, and DMEM containing FCS. Fig. 4 shows the influence of
DNA-PEI-SPIONs after 24 h exposure to HeLa cells in PEI-SPIONs and DNA-PEI-SPIONs on cell viability at
RPMI and DMEM, both in presence and absence of 10% 24 h exposure to HeLa cells in RPMI and DMEM, both
FCS. Cytotoxic effects of the polymer coated SPIONS were in presence and absence of 10% FCS. PEI has been found
always compared to cytotoxic effects of the pure polymer to be a versatile polymeric vector for gene delivery that
solutions (Fig. 3). Four different iron concentrations (20, tightly condenses plasmid DNA [20]. Although the ability
40, 80, and 120 lg/ml) were assayed and the polymer/iron to deliver DNA to the nucleus enhances transgene expres-
ratio was maintained in all tests. The results are shown rel- sion, the presence of PEI in the nucleus may interfere with
ative to the cell viability in DMEM+FCS, since optimal transcriptional and translational processes and induce cell
growth of mammalian cells in tissue culture is routinely death [21,22]. Efforts have been undertaken to increase
obtained with medium supplemented with whole serum. transfection efficiency but at the same time decrease
The culturing of HeLa cells was reported in the literature adverse effects of PEI on cell viability [23]. Therefore, com-
with different media, such as RPMI 1640 [17,18] and paratively lower polymer concentrations were investigated
DMEM [19] all supplemented with 10% serum. PVA-SPI- here. The investigated concentration range as well as the
ONs, as well as PVA alone, showed no influence on cell used N/P ratio of 7.5 (molar ratio of nitrogen-containing
viability of HeLa cells at the incubated concentrations groups of the polymer to phosphate groups of the nucleic
ranging from 200 to 1200 lg/ml (results not shown). In acid) corresponded to the values needed for optimal mag-
all investigated cases, A-PVA-coated SPIONs were far less netically enhanced non-viral transfection in HeLa cells
toxic compared to the equal amount of the copolymer in [24]. PEI-SPIONs showed significant toxicity at concentra-
solution. Note that A-PVA is actually a mixture of polyvi- tions around 500 ng PEI/ml dispersion. A relation between
nyl alcohol (MowiolÒ 3-83) and vinyl alcohol/vinyl amine cell viability and medium in the presence or absence of
134 A. Petri-Fink et al. / European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137

PEI-SPION RPMI RPMI

PEI-SPION RPMI + FCS
1
a PEI-SPION DMEM RPMI+FCS

Cell bound iron [µg/ml]

1.00 PEI-SPION DMEM + FCS DMEM
0.8
relative cell viability

DMEM+FCS
0.75 0.6

0.50 0.4

0.2
0.25
0
0.00 1h1 3h2 1h3 3h4
0 100 500
A-PVA-SPION PVA-SPION
ng/ml PEI
Fig. 5. Uptake of A-PVA- and PVA-coated SPIONs (120 lg Fe/ml) after
b DNA-PEI-SPION RPMI 1 and 3 h in different cell media: RPMI, RPMI+FCS, DMEM, and
DNA-PEI-SPION RPMI + FCS DMEM+FCS.
DNA-PEI-SPION DMEM
DNA-PEI-SPION DMEM + FCS
1.00 iron increased up to 12 lg/ml in the case of A-PVA-
relative cell viability

SPIONs, whereas only a very low increase in iron content

0.75 could be detected for PVA-SPIONs (data not shown). This
observation was not surprising, since the preferential
0.50 uptake of cationic agents has been widely used in the field
of molecular biology, mainly for transfection purposes.
0.25
The uptake of A-PVA-SPIONS after 1 and 3 h of exposure
0.00 was strongly dependent on the cell medium, since uptake
0 100 500 was only detected in the absence of serum. This effect is less
ng/ml PEI pronounced for PVA-SPIONs, where merely a negative
Fig. 4. (a) Relative cell viability of HeLa cells exposed to differently influence of RPMI with FCS can be detected.
concentrated dispersions of (a) PEI-SPIONs in different cell media
(RPMI, RPMI+FCS, DMEM, and DMEM+FCS) and (b) DNA-PEI-
SPIONs in different cell media (RPMI, RPMI+FCS, DMEM, and 4. Discussion
DMEM+FCS).
4.1. Turbidity measurements and photon correlation
spectroscopy (PCS)
serum was not observed. It is interesting to note that
increasing concentrations of DNA-PEI-SPIONs had The study of particle aggregation is important since the
almost no effect on cell viability. aggregation of SPIONs reduces the superparamagnetic
In a second test series we determined if HeLa cells could properties of the nanoparticles as magnetic interactions
interact with PVA-, A-PVA-, PEI- or DNA-PEI-SPIONS. may be present between the particles during the agglomer-
Using the Prussian Blue reaction, the cellular iron content ate formation [28]. In addition, the size of the aggregates
was measured in DMEM and RPMI, both with and with- may influence the biological function of the particles, in
out FCS, at four different particle concentrations after 1, 3, in vitro [29,30] and in in vivo applications [31]. Although
and 24 h exposure to the differently coated nanoparticles. the exact determination of particle sizes by turbidity mea-
Earlier studies with dextran and silica coated iron oxide surements is difficult [32], the agglomeration behaviour of
particles had confirmed the reliability of the Prussian blue the particles is straightforward, more sensitive and easily
reaction in cell cultures [25]. The detection limit of this reproduced with turbidity measurements. This is advanta-
reaction was reported to be around one ppm [26]. The iron geous as most biological laboratories do not have precise
content was always below detection limits in cells not size measurement equipment but usually have access to
exposed to nanoparticles. UV/VIS spectrometers.
The uptake of A-PVA- and PVA-coated SPIONs In our study we found that A-PVA-SPIONs and PVA-
(120 lg Fe/ml) after 1 and 3 h in different cell media was SPIONs showed no important agglomeration in none of
investigated (Fig. 5). The cellular iron content increased the investigated media after 30 min. Therefore, biological
over time and was dependent on the polymer used to coat evaluations can only neglect the influence of the media
these nanoparticles. Earlier studies have shown that uptake on colloidal stability when working in this time frame.
was also dependent on the amount of incubated particles At least in the case of HeLa cells, 30 min of incubation
[27]. A-PVA-SPIONs were better absorbed by cells than at the investigated concentrations is not sufficient to
PVA-coated SPIONs. After 24 h of incubation, this effect achieve considerable uptake. This can be overcome by
becomes more obvious, when the amount of cell bound actually exploiting the superparamagnetic properties of
 A. Petri-Fink et al. / European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137 135

the core iron oxide nanoparticles. In this case, uptake can of the oxide [35]. The configuration of the adsorbed poly-
be significantly enhanced in the presence of a magnetic mer is difficult to determine [36], and is affected by the poly-
field [24]. After 1 h of incubation significant differences mer concentration, molecular weight, pH, ionic strength,
between A-PVA-SPIONs and PVA-SPIONs were detected and surface charge. To study the coating mechanism, it is
in depending on the medium. The high stability of PVA- very helpful to work with particles that are uniform in size
SPIONs in FCS supplemented media could be explained and shape. A number of such model systems have been
by depletion stabilization caused by the serum proteins studied in the last few years using inorganic particles as a
showing a similar size as the nanoparticles. Increased sta- core and various polymers as shells [37–39]. According to
bility of A-PVA-SPIONs in FCS supplemented DMEM the common literature, an increased block character of
(after 1 h) is attributed to a similar mechanism proposing the vinyl acetate units in the PVA is beneficial [40]. To
the additional adsorption of small negatively charged mol- our knowledge, real systems (i.e. particles that are not
ecules at the surface of the particles leading to a neutral monodisperse in size and are coated with a commercial
surface charge as observed with PVA-SPIONs. In the case polymer) have not been studied in detail so far, especially
of RPMI no such preferred adsorption could take place not for particles in that size range.
and the adsorption of FCS on the particles could be possi- The lower toxicity of the nanoparticle associated poly-
ble, since the surface of proteins is heterogeneous with mer indirectly suggested that the polymer is not desorbed
regard to charged and hydrophobic domains [29]. If one from the surface of the particles in cell media; otherwise
or more FCS proteins are attached to the particles, no comparably lower cell viabilities would have been mea-
depletion stabilization is possible and agglomeration sured. The cationic and therefore toxic charge density
occurs. However, the amine content is low (compared to might be diminished when the polymer is adsorbed on
e.g. PEI) and this ‘‘medium’’ effect is overcome by the the surface. The different mitochondria activity of A-
intrinsic particle properties after 2 h. The results between PVA-SPIONs in media was an artifact attributed to the dif-
the two systems were comparable already after 2 h. Only ferent cell viabilities in the media and could not be related
DMEM with FCS offered long term colloidal stability to the particles. As well, DNA-PEI-SPION complexes
for both systems. The medium and the surface of the par- showed low cytotoxicity towards HeLa cells compared to
ticles have to be evaluated fundamentally using statistical PEI coated SPIONs alone. It seemed that the density of
methods to gain an insight into the impact and participa- uncomplexed amino groups plays an important role
tion of both effects. The agglomeration behaviour of regarding cytotoxicity. The cytotoxicity of PEI-SPIONs
PEI-SPIONs was consistent with previous reports that was overcome only when DNA was added. It has been pos-
showed that PEI interacts unclearly with negatively tulated that during the spontaneous PEI-DNA complex
charged molecules in the serum as well as plasma proteins formation PEI wraps DNA via electrostatic interactions,
such as opsonins [33]. The difference in agglomeration thereby diminishing the high cationic and toxic charge den-
behaviour in the different media is hard to explain. The sity of PEI to a magnitude that promotes DNA delivery
two media RPMI 1640 and DMEM vary in their glucose but decreases the negative effects of PEI on viability [20].
content which is in RPMI 1640, 2000 mg/ml, and in The DNA-PEI-SPIONs used in this study were prepared
DMEM, 4500 mg/ml. Apart from differences in glucose at an N/P ratio of 7.5 with a PEI/SPION ratio of 2. At
content the main differences lie in the presence/absence these concentrations and ratios, respectively, the zeta
of e.g. sodium pyruvate pyridoxine, L-arginine, L-proline potential showed that there was only a slight net positive
and L-histidine. However, these variations could not charge of around 12 mV [41]. In addition, the uptake of
explain the different agglomeration behaviour. PVA-SPIONs and A-PVA-SPIONs was examined. We
already found in previous studies that PVA-coated SPI-
4.2. Biological evaluations ONs are well tolerated by cells and that uptake was negli-
gible in the absence of a magnetic field. On the other hand,
It was previously reported that uncoated iron oxide A-PVA-SPIONs (not PVA-SPIONs) were well taken up by
nanoparticles could have toxic effects on cells [34]. How- human melanoma tumor cells according to an active, satu-
ever, the cytotoxicity of polymer coated nanoparticles is rable and energy-dependent mechanism [27]. In our study
rather influenced by the outer polymer coating layer cover- we found that the presence of serum strongly inhibited
ing the SPIONs than by the iron oxide nanoparticles them- the uptake of A-PVA-SPIONs due to the previously men-
selves [4,5]. We found that PVA-SPIONs were not tioned adsorption of serum on the particle surface. This
cytotoxic for HeLa cells (results not shown). In general, is in agreement with the results from other groups who
A-PVA-coated SPIONs were far less toxic compared to showed that the presence of serum in the medium inhibits
the equal amount of the co-polymer in solution. This cellular uptake and binding of cationic carriers reduces
decrease in cytotoxicity might be due to the conformational their efficiency as nucleic acid carriers [42,43]. On the other
change of the polymer when adsorbing to the surface of the hand, the uptake of PVA-SPIONs is facilitated in DMEM
nanoparticles. The interaction with the surface results from supplemented with FCS. Serum protein adsorption on the
hydrogen bonding between polar functional groups of the PVA surface is very low especially on polymers with hydro-
polymer and hydroxylated and protonated surface sides lysis degrees higher than 88% [44]. At the same time these
136 A. Petri-Fink et al. / European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137

particles showed the best stability in DMEM supplemented References

with FCS.
Our experiments were not designed to formally prove [1] C.C. Berry, A.S.G. Curtis, Functionalisation of magnetic nanopar-
cell internalization of SPIONs, to determine their location ticles for applications in biomedicine, J. Phys. D Appl. Phys. 36 (2003)
R198–R206.
in the cellular compartments and/or to characterize the [2] C.C. Berry, Possible exploitation of magnetic nanoparticle–cell
uptake pathways involved, and it is presently too prema- interaction for biomedical applications, J. Mater. Chem. 15 (2005)
ture to postulate the exact mechanisms involved, either 543–547.
endocytosis/phagocytosis or receptor-mediated or chan- [3] P. Hoet, I. Bruske-Hohlfeld, O. Salata, Nanoparticles – known and
nel-mediated uptake. A logical next step involves the thor- unknown health risks, J. Nanobiotechnol. 2 (2004) 12.
[4] H. Lee, E. Lee, D.K. Kim, N.K. Jang, Y.Y. Jeong, S. Jon,
ough investigation of the particles’ surface after (a) Antibiofouling polymer-coated superparamagnetic iron oxide nano-
incubation in cell media or biological fluids and (b) cell particles as potential magnetic resonance contrast agents for in vivo
uptake to understand the processes that occur at the sur- cancer imaging, J. Am. Chem. Soc. 128 (2006) 7383–7389.
face of the particles. According to our preliminary results [5] M. Auffan, L. Decome, J. Rose, T. Orsiere, M. De Meo, V. Briois, C.
Chaneac, L. Olivi, J.L. Berge-Lefranc, A. Botta, M.R. Wiesner, J.Y.
it becomes obvious that each nanoparticle system has to
Bottero, In vitro interactions between DMSA-coated maghemite
be considered individually since cell uptake, and to certain nanoparticles and human fibroblasts: a physicochemical and cyto-
extent cytotoxicity, was not only dependent on the particle genotoxical study, Environ. Sci. Technol. 40 (2006) 4367–4373.
size and surface, but also highly dependent on the cell lines [6] D.F. Evans, The colloidal domain: where physics, chemistry, biology,
(unpublished data). There is an undeniable influence of the and technology meet, D. Fennell Evans, Håkan Wennerström, Wiley-
medium and the presence/absence of serum. However, the VCH, New York, 1999.
[7] D. Williams, S.H. Ehrman, T.R. P Poloman, Evaluation of the
correlation between the colloidal and biological data is pre- microbial growth response to inorganic nanoparticles, J. Nanobio-
mature and requires extensive and more sophisticated technol. 4 (2006), doi:10.1186/1477-3155-4-3, article 3.
investigations. The impact of the medium on colloidal sta- [8] M.R. Lorenz, V. Holzapfel, A. Musyanovych, K. Nothelfer, P.
bility and consequently cell interaction competes with the Walther, H. Frank, K. Landfester, H. Schrezenmeier, V. Mailander,
particles’ intrinsic properties (size and surface). Currently, Uptake of functionalized, fluorescent-labeled polymeric particles in
different cell lines and stem cells, Biomaterials 27 (2006) 2820–2828.
we can say that agglomeration behaviour in different media [9] M. Chastellain, A. Petri, H. Hofmann, Particle size investigations of a
cannot be correlated to cell uptake within the measured multistep synthesis of PVA coated superparamagnetic nanoparticles,
sizes of 40 to 130 nm and uptake times from 1 to 3 h. At J. Colloid Interface Sci. 278 (2004) 353–360.
current state-of-the-art, the stability of nanoparticle disper- [10] A. Bee, R. Massart, S. Neveu, Synthesis of very fine maghemite
particles, J. Magn. Magn. Mater. 149 (1995) 6–9.
sions in cell media should have to be checked routinely.
[11] G.A. van Ewijk, G.J. Vroege, A.P. Philipse, Convenient preparation
methods for magnetic colloids, J. Magn. Magn. Mater. 201 (1999)
5. Conclusion 31–33.
[12] A. von Harpe, H. Petersen, Y.X. Li, T. Kissel, Characterization of
The comparison of PVA-SPIONs with A-PVA-SPIONs commercially available and synthesized polyethylenimines for gene
delivery, J. Control. Release 69 (2000) 309–322.
showed that already minor modifications of the nanoparti-
[13] D.A. Skoog, Fundamentals of analytical chemistry, in: D.A. Skoog,
cles lead to an altered behaviour in stability, uptake, and D.M. West, F. James Holler (Eds.), Harcourt College, Fort Worth,
toxicity. Therefore, the stability of the particle dispersions London, 1996.
needs to be assessed after each surface modification such [14] R.M. Santos, J. Forcada, Acetal-functionalized polymer particles
as fluorescent labelling or chemical coupling of drug useful for immunoassays. II. Surface and colloidal characterization, J.
Polym. Sci. A: Polym. Chem. 37 (1999) 501–511.
molecules.
[15] R.M. Cornell, U. Schwertmann, The Iron Oxides, VCH Verlagsge-
Although we answered some of the currently discussed sellshaft GMBH, Weinheim, Germany, 1996.
questions on colloidal nanoparticles and their behaviour [16] Y. Zhang, N. Kohler, M.Q. Zhang, Surface modification of super-
in vitro, more fundamental studies will have to be carried paramagnetic magnetite nanoparticles and their intracellular uptake,
out in the future. Those must focus on colloidal stability Biomaterials 23 (2002) 1553–1561.
[17] N. Kohler, C. Sun, J. Wang, M.Q. Zhang, Methotrexate-modified
and its influence on biological properties and will provide
superparamagnetic nanoparticles and their intracellular uptake into
a profound base for future discussions on toxicity and human cancer cells, Langmuir 21 (2005) 8858–8864.
potential application of nanoparticles in the field of [18] C. Sun, R. Sze, M.Q. Zhang, Folic acid-PEG conjugated superpara-
biomedicine. magnetic nanoparticles for targeted cellular uptake and detection by
MRI, J. Biomed. Mater. Res. A 78A (2006) 550–557.
[19] A.M. Koch, F. Reynolds, M.F. Kircher, H.P. Merkle, R. Weissleder,
Acknowledgements L. Josephson, Uptake and metabolism of a dual fluorochrome tat-
nanoparticle in HeLa cells, Bioconjug. Chem. 14 (2003) 1115–1121.
This work was supported by the Swiss National Science [20] O. Boussif, F. Lezoualch, M.A. Zanta, M.D. Mergny, D. Scherman,
Foundation (SNF 200020-109490 and SFN 205321- B. Demeneix, J.P. Behr, A versatile vector for gene and oligonucleo-
tide transfer into cells in culture and in-vivo – polyethylenimine, Proc.
111908) and the Competence Centre of Materials Science
Natl. Acad. Sci. USA 92 (1995) 7297–7301.
of the Swiss Federal Institute of Technology, Project [21] W.T. Godbey, K.K. Wu, A.G. Mikos, Poly(ethylenimine)-mediated
PAPAMOD. We thank Prof. J. Hubbell for the access to gene delivery affects endothelial cell function and viability, Biomate-
cell culture facilities. rials 22 (2001) 471–480.
 A. Petri-Fink et al. / European Journal of Pharmaceutics and Biopharmaceutics 68 (2008) 129–137 137

[22] B.I. Florea, C. Meaney, H.E. Junginger, G. Borchard, Transfection [33] M. Neu, D. Fischer, T. Kissel, Recent advances in rational gene
efficiency and toxicity of polyethylenimine in differentiated Calu-3 transfer vector design based on poly(ethylene imine) and its deriva-
and nondifferentiated COS-1 cell cultures, AAPS PharmSci. 4 (2002), tives, J. Gene. Med. 7 (2005) 992–1009.
doi:10.1208/ps040212, article 12. [34] T.J. Brunner, P. Wick, P. Manser, P. Spohn, R.N. Grass, L.K.
[23] E. Wagner, M. Zenke, M. Cotten, H. Beug, M.L. Birnstiel, Limbach, A. Bruinink, W.J. Stark, In vitro cytotoxicity of oxide
Transferrin-polycation conjugates as carriers for DNA uptake into nanoparticles: comparison to asbestos, silica, and the effect of particle
cells, Proc. Natl. Acad. Sci. USA 87 (1990) 3410–3414. solubility, Environ. Sci. Technol. 40 (2006) 4374–4381.
[24] S.W. Kamau, P.O. Hassa, B. Steitz, A. Petri-Fink, H. Hofmann, M. [35] L.T. Lee, P. Somasundaran, Adsorption of polyacrylamide on oxide
Hofmann-Amtenbrink, B. von Rechenberg, M.O. Hottiger, Enhance- minerals, Langmuir 5 (1989) 854–860.
ment of the efficiency of non-viral gene delivery by application of [36] J.M.H.M. Scheutjens, G.J. Fleer, M.A.C. Stuart, End effects in
pulsed magnetic field, Nucleic Acids Res. 34 (2006) e40. polymer adsorption – a tale of tails, Colloids Surf. 21 (1986) 285–306.
[25] B. Schöpf, The effect, behavior and distribution of supraparamagnetic [37] K. Furusawa, Y. Kimura, T. Tagawa, Preparation of composite
nanoparticles (SPIONs) in the musculoskeletal system, Veterinary polymer lattices including inorganic particles, Kobunshi Ronbunshu
Faculty, University of Zürich, Zürich, 2005. 40 (1983) 697–702.
[26] F. Timm, Der Histochemische Eisennachweis, Histochemie 2 (1960) [38] R. Partch, S.G. Gangolli, E. Matijevic, W. Cai, S. Arajs, Conducting
143–149. polymer composites 1. Surface-induced polymerization of pyrrole on
[27] A. Petri-Fink, M. Chastellain, L. Juillerat-Jeanneret, A. Ferrari, H. iron(III) and cerium(IV) oxide particles, J. Colloid Interface Sci. 144
Hofmann, Development of functionalized superparamagnetic iron (1991) 27–35.
oxide nanoparticles for interaction with human cancer cells, Bioma- [39] C.L. Huang, E. Matijevic, Coating of uniform inorganic particles
terials 26 (2005) 2685–2694. with polymers 3. Polypyrrole on different metal-oxides, J. Mater. Res.
[28] J.M. Vargas, L.M. Socolovsky, M. Knobel, D. Zanchet, Dipolar 10 (1995) 1327–1336.
interaction and size effects in powder samples of colloidal iron oxide [40] C.A. Finch, Polyvinyl Alcohol – Properties and Applications, Wiley-
nanoparticles, Nanotechnology 16 (2005) S285–S290. Interscience, Chichester, 1973.
[29] A. Vonarbourg, C. Passirani, P. Saulnier, J.P. Benoit, Parameters [41] B. Steitz, H. Hofmann, A. Petri-Fink, Characterization of PEI coated
influencing the stealthiness of colloidal drug delivery systems, superparamagnetic iron oxide nanoparticles for transfection: size
Biomaterials 27 (2006) 4356–4373. distribution, colloidal properties and dna interaction, J. Magn. Magn.
[30] L.K. Limbach, Y.C. Li, R.N. Grass, T.J. Brunner, M.A. Hintermann, Mater. (2006), doi:10.1016/j.jmmm.2006.10.1194.
M. Muller, D. Gunther, W.J. Stark, Oxide nanoparticle uptake in [42] A.G. Tkachenko, H. Xie, Y.L. Liu, D. Coleman, J. Ryan, W.R.
human lung fibroblasts: effects of particle size, agglomeration, and Glomm, M.K. Shipton, S. Franzen, D.L. Feldheim, Cellular trajec-
diffusion at low concentrations, Environ. Sci. Technol. 39 (2005) tories of peptide-modified gold particle complexes: comparison of
9370–9376. nuclear localization signals and peptide transduction domains,
[31] L.Q. Galuppo, S.W. Kamau, B. Steitz, P.O. Hassa, M. Hilbe, L. Bioconjug. Chem. 15 (2004) 482–490.
Vaughan, S. Koch, A. Fink-Petri, M. Hofman, H. Hofman, M.O. [43] J.M. Colet, Y. Vanhaverbeke, R.N. Muller, Evidence for attachment
Hottiger, B. von Rechenberg, Gene expression in synovial membrane of magnetic starch microspheres to Kupffer cells receptors in excised
cells after intraarticular delivery of plasmid-linked superparamagnetic and perfused-pat-liver, Invest. Radiol. 29 (1994) S223–S225.
iron oxide particles - A preliminary study in sheep, J. Nanosci. [44] D.A. Barrett, M.S. Hartshorne, M.A. Hussain, P.N. Shaw, M.C.
Nanotechnol. 6 (2006) 2841–2852. Davies, Resistance to nonspecific protein adsorption by poly(vinyl
[32] G. Crawley, M. Cournil, D. DiBenedetto, Size analysis of fine particle alcohol) thin films adsorbed to a poly(styrene) support matrix
suspensions by spectral turbidimetry: potential and limits, Powder studied using surface plasmon resonance, Anal. Chem. 73 (2001)
Technol. 91 (1997) 197–208. 5232–5239.