J Nanopart Res (2015)17:351

DOI 10.1007/s11051-015-3148-5

RESEARCH PAPER

Iron oxide nanoparticles surface coating and cell uptake

affect biocompatibility and inflammatory responses
of endothelial cells and macrophages
Antonina Orlando . Miriam Colombo . Davide Prosperi .
Maria Gregori . Alice Panariti . Ilaria Rivolta . Massimo Masserini .
Emanuela Cazzaniga

Received: 24 June 2015 / Accepted: 7 August 2015

Ó Springer Science+Business Media Dordrecht 2015

Abstract Engineered iron oxide nanoparticles (IONP) cells exposed to NP after systemic administration, to
offer the possibility of a wide range of medical uses, assess the biocompatibility of PEG-IONP-PMA
from clinical imaging to magnetically based hyperther- (23.1 ± 1.4 nm) or IONP-PMA (15.6 ± 3.4 nm).
mia for tumor treatment. These applications require their PEG-IONP-PMA, tested at different concentrations as
systemic administration in vivo. An important property high as 20 lg mL-1, exhibited no cytotoxicity or
of nanoparticles is their stability in biological media. For inflammatory responses. By contrast, IONP-PMA
this purpose, a multicomponent nanoconstruct combin- showed a concentration-dependent increase of cytotox-
ing high colloidal stability and improved physical icity and of TNF-a production by macrophages and NO
properties was synthesized and characterized. IONP production by HUVECs. Cell uptake analysis suggested
were coated with an amphiphilic polymer (PMA), which that after PEGylation, IONP were less internalized either
confers colloidal stability, and were pegylated in order to by macrophages or by HUVEC. These results suggest
obtain the nanoconstruct PEG-IONP-PMA. The aim of that the choice of the polymer and the chemistry of
this study was to utilize cultured human endothelial cells surface functionalization are a crucial feature to confer
(HUVEC) and murine macrophages, taken as model of to IONP biocompatibility.

Keywords Iron oxide nanoparticles  Endothelial

Electronic supplementary material The online version of cells  Macrophages  Nanoparticles uptake 
this article (doi:10.1007/s11051-015-3148-5) contains supple-
mentary material, which is available to authorized users. Biocompatibility  Inflammation

A. Orlando  M. Gregori  A. Panariti 

I. Rivolta  M. Masserini  E. Cazzaniga (&)
Department of Health Sciences, University of Milano-
Bicocca, Via Cadore 48, 20900 Monza, Italy Introduction
e-mail: emanuela.cazzaniga@unimib.it
Nanotechnology is an exciting field of investigation
A. Orlando  M. Colombo  D. Prosperi 
for the development of new treatments for many
M. Gregori  I. Rivolta  M. Masserini  E. Cazzaniga
Nanomedicine Center NANOMIB, University of Milano- human diseases. Engineered nanoparticles (NP) of
Bicocca, Milan, Italy uniform size and shape are now available for a wide
range of biological applications such as in vivo
M. Colombo  D. Prosperi
imaging or drug delivery (Choi et al. 2009). However,
Department of Biotechnology and Biosciences, University
of Milano-Bicocca, piazza della Scienza 2, 20126 Milan, an understanding of the potential toxicity of NP is
Italy needed before considering clinical applications.

123
351 Page 2 of 13 J Nanopart Res (2015)17:351

Magnetic iron oxide nanoparticles (IONP) with a the high colloidal stability in biological media typical
diameter \50 nm are a novel class of magnetic of IONP generated by aqueous alkaline coprecipita-
resonance imaging (MRI) contrast agents and could tion with improved physical properties associated
be used for therapeutic purposes, using hyperthermia with engineered nanocrystals.
generated by an electromagnetic field (Shevtsov et al. In this context, in vitro biocompatibility testing of
2014). these NP is an essential step in order to indicate a
To date, a wide variety of magnetic NP have been concentration range, which could be used as a starting
produced, differing in size and type of coating point for in vivo experiments with the particles
material (Gupta and Gupta 2005; Shevtsov et al. described. It is also important to understand if the
2014). Particle size and surface modification lead to NP are able to enter the cell, this has implications not
different responses in terms of cell nonspecific or only for their fate but also for their impact on
receptor-mediated uptake of IONP (Ito et al. 2005; biological systems.
Simoni et al. 2008). It is well known that IONP can be A key issue is related to the observation that once
taken up into the reticuloendothelial system (RES) by systemically administered, NP designed to target
endocytosis or phagocytosis. IONP are also taken up pathological tissues interact either with the cells lining
by phagocytic cells such as monocytes, macrophages, the vascular bed or with cells circulating in the blood.
and oligodendroglial cells, and can also be found in In this study, we have utilized cultured human
the endothelial cells (Mahmoudi et al. 2009). This umbilical vein endothelial cells (HUVECs) and mouse
issue is particularly relevant when using engineered macrophages (RAW264.7) taken as model of cells
iron oxide nanocrystals, which are preferable to NP exposed to NP after systemic administration in order
obtained by aqueous coprecipitation methods due to to deepen the biocompatibility issue.
their improved chemical–physical properties, but are Although there are an increasing number of studies
usually highly immunogenic and subjected to aggres- on biological responses to a variety of nanomaterials
sive attack by biological agents. This feature is likely (Hardman 2006; Klostranec and Chan 2006; Colombo
caused by the presence of residual surfactants on their et al. 2012a, b), many questions remain unanswered.
surface originated from solvothermal decomposition What is the fate of the NP once they are presented to a
in organic media (Wu et al. 2010). The surface population of cells? If the NP cross the cell membrane,
modification of IONP with a polyethylene glycol what intracellular effects do they exert? (Shevtsov
(PEG) corona provides multiple benefits, including et al. 2014). Furthermore, there are few systematic
minimizing nonspecific protein adsorption and studies dealing with both cytotoxicity and inflamma-
reducing rapid degradation, aggregation and toxicity tory responses of cells exposed to NP (Shevtsov et al.
of NP (Piazza et al. 2011). However, the direct 2014).
immobilization of PEG molecules on IONP surface For these reasons, in this study, we have evaluated
does not confer sufficient protection to the NP, and a the cellular response induced by the exposure to a
residual surface area is still available for interaction colloidal dispersion of IONP consisting in a highly
with nonspecific molecular entities and for direct uniform and monodisperse iron oxide nanocrystal
access of water molecules assisting the release of core, stabilized by coating with the amphiphilic PMA
unsafe ionic species. For this reason, the design of polymer, functionalized or not with PEG. The
advanced colloidal nanocrystals for biological appli- responses were evaluated in terms of production either
cations may involve the exploitation of suitably of NO, one of the most important vasoactive sub-
designed multifunctional amphiphilic polymers stances released by the endothelium, or of various pro-
(PMA), e.g. the PMA polymer bearing several inflammatory mediators, secreted by macrophages in
carboxylate groups for further functionalization, response to harmful stimuli. These assays were carried
which have the advantage of conferring compelling out in combination with two different viability assays
colloidal stability in biological environment and of and with intracellular uptake studies. Moreover, we
reliably protecting the NP surface concomitantly have estimated the ability of NP to cross an in vitro
enabling proper outer functionalization with PEG model of blood–brain barrier (BBB) to investigate
and/or targeting molecules (Li et al. 2012). The their potential use as therapeutic agents for brain
resulting multicomponent nanoconstruct combines diseases.

123
J Nanopart Res (2015)17:351 Page 3 of 13 351

Materials and methods removed by concentration at 3500 rpm through

100 kDa Amicon filters. The concentrated solution
Materials was washed twice in the same way and the NP were
then recovered as a stock solution in PBS.
Commercial chemicals were of the highest available For confocal microscopy experiments, fluores-
grade. 3-(4,5-dimetiltiazol-2-yl)-2,5-diphenyltetra- cently labeled NP were synthesized using fluorescent
zolium bromide (MTT), reagents for NO production PMA (Mazzucchelli et al. 2013). An aliquot of this
analysis, lipopolysaccharide (Escherichia coli solution (20 lL) was added to IONP (1.5 mg in
055:B5), resveratrol, rabbit anti-b-actin antibody CHCl3) and reacted as described above.
and bicinchoninic acid (BCA) assay were from For the synthesis of PEG-coated IONP (PEG-
Sigma–Aldrich (Milan, Italy). HUVECs were IONP), oleate-capped IONP (8.4 mg) in chloroform
obtained from Lonza (Walkersville, MD); the (2.5 mL) were mixed with an aqueous solution of
murine monocytic macrophage cell line RAW tetramethylammonium hydroxide (240 mg in 8 mL
264.7 was obtained from ATCC (LGC partner, deionized water). The mixture was vigorously stirred
Milan, Italy); EGM-2, SingleQuot kit and EBM-2 overnight at room temperature (Liong et al. 2010) and
were from Lonza (Walkersville, MD); all the stock then centrifuged at 1000 rpm for 3 min to separate the
solutions for cell cultures and Spin-X centrifuge phases. The organic solution was discarded. Next,
tube filters were from Euroclone (Milano, Italy); 20 lL of O-(2-Aminoethyl)-O0 -methyl-polyethylene
mouse anti-caspase3, mouse anti-eNOS, rabbit anti- glycol (PEG2000-NH2, MW 2000, 0.05 M in water)
PeNOS, rabbit anti-TNF-a, rabbit anti-IL-1b and was added and stirred for overnight at room temper-
rabbit anti-COX-2 were from Cell Signalling (Bev- ature. The excess of PEG was removed by centrifu-
erly, USA); secondary HRP-conjugated antibodies gation at 3000 rpm in 100 kDa Amicon. The
and ECL SuperSignal detection kit were from Pierce concentrated solution was washed, diluted with water
(Rockford, IL, USA); SDS-NuPAGE reagents and filtered through a 0.22-lm pore syringe filter. The
(4–12 % Bis–Tris gel; sample buffer; running NP were finally resuspended in PBS.
buffer) and Precision Plus Protein Standard were
from Invitrogen (Milan, Italy); complete protease Transmission electron microscopy
inhibitor Cocktail and Cytotoxicity Detection Kit
(LDH) were from Roche (Germany); nitrotyrosine For TEM analyses, NP were dispersed under sonica-
ELISA was from Millipore (Italy); TNF-a high tion in hexane (50 lg mL-1) and a drop of the
sensitivity immunoassay kits was from eBioscience resulting solution was placed on a Formvar/carbon-
(San Diego, CA); Texas Red phalloidin was from coated copper grid and air-dried. For ultrastructural
Molecular Probes (Lifetechnology, Monza, Italy). analyses of biological samples, cells were fixed in
Ultrapure and deionized water was obtained from 2.5 % glutaraldehyde in 0.1 M phosphate buffer, pH
Direct-Q5n system (Millipore, Italy). All other 7.2, for 2 h. After one rinsing with phosphate buffer,
chemicals were reagent grade. specimens were post-fixed in 1.5 % osmium tetroxide
for 2 h, dehydrated by 70, 90 and 100 % EtOH and
Synthesis and characterization of NP embedded in epoxy resin (PolyBed 812 Polysciences
Inc., USA). All samples were examined by means of a
Preparation of pegylated PMA-coated Fe3O4 Zeiss EM109 transmission electron microscope.
nanoparticles (PEG-IONP-PMA)
Dynamic light scattering and zeta potential
PMA-coated IONP (IONP-PMA) were synthesized measurements
according to a previously published procedure (Maz-
zucchelli et al. 2013). IONP-PMA (5 mg) were Dynamic light scattering (DLS) measurements were
reacted with 0.1 M EDC (10 lL) for 2 min, and then performed at 90° with a 90Plus Particle Size Analyzer
4 lL of O-(2-Aminoethyl)-O0 -methyl-polyethylene from Brookhaven Instruments Corporation (Holts-
glycol (PEG2000-NH2, MW 2000, 0.05 M in water) ville, NY), working at 15 mW of a solid state laser
was added and stirred for 2 h. The excess of PEG was (k = 661 nm). Viscosity and refractive index of pure

123
351 Page 4 of 13 J Nanopart Res (2015)17:351

water were used to characterize the solvent. NP were multilabel counter, Perkin Elmer). Untreated cells
dispersed in the solvent and sonicated in a S15H were used as a negative control. Each sample was
Elmasonic apparatus (Elma, Singen, Germany) before analysed at least in triplicate.
analysis. Final sample concentration used for mea-
surements was typically of 5 lg mL-1. Zeta potential Membrane integrity
measurements are elaborated on the same instrument
equipped with AQ- 809 electrode. Data are processed After treatment with NP, LDH was analysed using the
by Zeta Plus Software. cytotoxicity detection kit according to the manufac-
turer’s instruction. The optical density was measured
HUVEC and RAW264.7 cell cultures with a microplate reader (Victor3 1420 multilabel
counter, Perkin Elmer) at 490 nm. The relative
HUVECs and the murine monocytic macrophage cell amount of released LDH was normalized to the total
line RAW264.7 were grown as previously described amount of LDH of control cells which were com-
(Orlando et al. 2013). pletely lysed with lysis buffer provided from the kit.
HUVECs and macrophages at 80–90 % of conflu- Each sample was analysed at least in triplicate.
ence were treated for 1 and 24 h with different NP.
Small aliquots of higher concentrated NP solution Measurement of NO production after NP exposure
were added to the culture medium in order to obtain
the required concentrations. NP were used at 20/50/ NO production was measured in conditioned media
100 lg mL-1. As a negative control, cells were of cells by using the Griess method (Hevel and
incubated with culture medium added with aliquots Marletta 1994; Orlando et al. 2013). To avoid the iron
of PBS in the same amount at which NP were added interference with the assay, after treatment, the NP
for the cell treatment. HUVECs were treated with were removed from the cell media using Spin-X
100 lg mL-1 PEG-IONPs-PMA also for 5/10/16 h. centrifuge tube filters (30K MWCO PES). Because
HUVECs and macrophages were also treated with NO in biological fluids is recovered as nitrate, the
100 lM resveratrol and 1 lg mL-1 lipopolysaccha- media were treated with nitrate reductase
ride (LPS) from E. Coli, respectively, as a positive (10 U mL-1) and NADPH (5 mM in Tris–HCl pH
controls for NO and cytokines production (Orlando 7.5) for 3 h at room temperature in order to reduce
et al. 2013). nitrate to nitrite. Finally, the samples were mixed
with an equal volume of freshly prepared Griess
Viability assays reagent (0.05 % N-[1-naphthyl] ethylenediamine
dihydrochloride and 0.5 % sulphanilamide in 2.5 %
Cells were treated with different doses of NP for 1 or ortho-phosphoric acid) for 20 min in the dark
24 h, and different aspects of cell viability were (Lovren et al. 2009). Sample nitrite levels were
evaluated by different assays: MTT ([3-(4,5-dimetil- calculated against a standard curve prepared with
tiazol-2-yl)-2,5-diphenyltetrazolium bromide]) assay 0.1–10 lM NaNO2.
to evaluate the mitochondrial activity (Orlando et al. The absorbance of each sample was measured at
2013) and the release of the cytoplasmic enzyme 540 nm using a microplate reader (Victor3 1420
lactate dehydrogenase (LDH) into the surrounding multilabel counter, Perkin Elmer).
medium to evaluate the leaking from membranes of
damaged or dead cells. Detection of eNOS phosphorylation

Mitochondrial activity Total cellular lysates were obtained resuspending

the cells in lysis buffer (containing 2 % SDS,
After treatment with NP, 0.5 lg mL-1 MTT solution 50 mM Tris–HCl, pH 6.8, 1 mM complete protease
was added to the cells for 2 h. After incubation, inhibitor Cocktail, and phosphatase inhibitors:
ethanol was added to each well to dissolve the formed 2 mM Na-orthovanadate, 1 mM Na-fluoride, 1 mM
formazan crystals. Absorbance at 550 nm was mea- Na-pyrophosphate) and harvested by scraping with a
sured with a microplate reader (Victor3 1420 rubber policeman. An aliquot of the total cellular

123
J Nanopart Res (2015)17:351 Page 5 of 13 351

lysates was analysed for the protein content by BCA actin bands for cell lysates or the Ponceau staining for
assay. After boiling for 5 min at 100 °C, equal medium aliquots (Romero-Calvo et al. 2010).
amounts of samples (15 lg of total proteins loaded) TNF-a was measured in supernatants from
were subjected to SDS-PAGE using a precast HUVECs treated with NP for 1 and 24 h by ELISA
NuPAGE 4–12 % gel, and proteins transferred to a carried out according to manufacturer’s instructions
nitrocellulose membrane. For b-actin detection and (eBioscience). Samples were measured in duplicate
caspase-3, the membrane was blocked for 30 min at against known amounts of murine TNF-a.
37 °C in TBS buffer containing 0.1 % Tween and
5 % non-fat milk and incubated over night at 4 °C Cellular NP uptake
with anti-b-actin (1:1500). The membrane was
blocked for 30 min at 37 °C in TBS buffer con- ICP-OES analysis
taining 0.1 % Tween and 5 % bovine serum albu-
min and then incubated over night at 4 °C with anti- Cells were incubated with NP at the concentration of
eNOS (1:1000) or anti-PeNOS (1:1000). 20 lg mL-1 for 1 h. For ICP-OES analysis, the cell
Immunoreactive proteins were revealed by pellets were suspended in 2 mL of PBS and 5 mL of
enhanced chemiluminescence (ECL) after incubation aqua regia was added. After 72 h, the samples were
for 2 h at RT with HRP-conjugated secondary anti- diluted with 10 mL of distilled water. All samples
bodies. The intensity of chemiluminescent spots was were measured in triplicate with Optima 7000 DV
semi-quantitatively estimated by ImageQuant ICP-OES (Perkin Elmer).
LAS4000 (GE Healthcare) and expressed as the ratio
between the intensity of the spot of interest and the Confocal microscopy
intensity of b-actin bands (Bulbarelli et al. 2012).
Cells were incubated with NP at the concentration of
Total nitrotyrosine measurement 20 lg mL-1. After 1 h, they were fixed with 4 %
paraformaldehyde in PBS at room temperature for
Total nitrotyrosine residues in the HUVECs lysates 20 min, washed three times with PBS, low salt buffer
were quantified by ELISA carried out according to (LS) and high salt buffer (HS), respectively, and
manufacturer’s instructions (Millipore). Samples were incubated for 90 min with GDB1X and digitonin
measured in duplicate against known standards, and 0.01 % and finally for 1 h with Texas Red conjugated
expression of nitrotyrosine was normalized to cellular phalloidin. Cells were then washed with HS and LS
protein content. three times, respectively. Coverslips were mounted
with a mounting medium (Sigma–Aldrich Milan,
Inflammation process activation Italy). Images were acquired by an LSM710 inverted
confocal laser scanning microscope equipped with a
After exposure, total cell lysates were obtained and Plan-Neofluar 63 9 1.4 oil objective (Carl Zeiss,
analysed using sodium dodecyl sulphate polyacrylamide Oberkochen, Germany) at k = 488 nm to detect NP,
gel electrophoresis/Western blotting as described above. k = 610 nm to detect actin filaments and k = 460 nm
For TNF-a, IL-1b and COX-2, the membrane was to detect the Hoechst labeled nuclei (Panariti et al.
blocked for 30 min at 37 °C in TBS buffer containing 2013).
0.1 % Tween and 5 % bovine serum albumin and then
incubated over night at 4 °C with different antibodies Assessment of NP ability to cross the in vitro BBB
(1:1000). model
Medium aliquots were loaded on a precast
NuPAGE 4–12 % gel, and electrophoresis/western For permeability experiments, immortalized hCMEC/
blotting was performed according to the procedure D3 were provided by Institut National de la Santé et de
described below (Ulivi et al. 2011). la Recherche Médicale ([INSERM] Paris, France) and
Immunoreactive proteins were revealed by ECL as cultured as previously described (Bana 2014) on
described above and expressed as the ratio between the 12-well transwell inserts coated with type I collagen in
intensity of the spot of interest and the intensity of b- a density of 7 9 104 cells/cm2, with 0.5 or 1 mL of

123
351 Page 6 of 13 J Nanopart Res (2015)17:351

culture medium in the upper and in the lower chamber, with an PMA in alkaline environment (Colombo et al.
respectively. Cells were treated with NP when the 2012a, b). PEGylated IONP (PEG-IONP-PMA) were
transendothelial electrical resistance (TEER) value synthesized by reacting IONP-PMA with PEG-NH2
(measured by EVOMX meter, STX2 electrode; World (2000 Da) using EDC as coupling agent. IONP-PMA
Precision Instruments, Sarasota, FL, USA) was found and PEG-IONP-PMA were 15.6 ± 3.4 nm and
to be the highest. 23.1 ± 1.4 nm in size, respectively, as measured by
Both NP (20 lg mL-1) were added in the upper DLS in milliq water at a concentration of 5 lg mL-1,
chamber and incubated for 3 h at 37 °C, after which with a f potential of -45.32 ± 4.86 and -29.74 ±
the percentages of the total applied dose of iron in the 3.86 mV, respectively.
lower chamber were determined by means of ICP-
OES analysis. All samples were measured in triplicate Cell viability
with Optima 7000 DV ICP-OES (Perkin Elmer).
HUVECs and RAW264.7 cells were treated with
Statistical analysis different doses of NP, and different aspects of cell
viability were assessed.
All experiments were carried out at least in triplicate.
Statistical analyses were performed using OriginPro8 Mitochondrial activity
(OriginLab Corporation, Northampton, MA, USA).
Data were compared using the unpaired Student’s t The treatment of HUVECs with IONP-PMA signifi-
test and expressed as mean ± SEM (standard error of cantly decreased the mitochondrial activity already
the mean). after 1 h. We observed about 20 % activity after 24 h
at all concentration tested, whereas the treatment with
PEG-IONP-PMA induced a decrease only after 24 h
Results in a dose-dependent manner, reaching a 50 % at the
highest dose tested (Fig. 2a).
Synthesis and characterization of NP In macrophages, IONP-PMA treatment showed a
significant reduction of mitochondrial activity only
Surfactant-coated Fe3O4 NP were synthesized by after 24 h at 50 and 100 lg mL-1 concentration (80
solvothermal decomposition from iron salts in octade- and 40 %, respectively), and the PEG-IONP-PMA
cene using oleic acid as capping agent (IONP, Fig. 1) induced a reduction after 1 h at 50 and 100 lg mL-1
(Mazzucchelli et al. 2013). IONP were suspended in (65 and 40 %, respectively). After 24 h treatment at
chloroform and transferred to water phase by mixing the highest dose, the residual activity was only 10 %
(Fig. 2b).
PEG-IONP treatment showed similar results (see
Supplementary Materials).

Membrane integrity

HUVECs and RAW264.7 cells were treated with

different amounts of NP, and the loss of membrane
integrity was assessed using LDH assay (Soenen et al.
2011). Treatment of HUVECs with IONP-PMA
significantly increased the LDH leakage already after
1 h. PEG-IONP-PMA significantly increased the
enzyme leakage only after 24 h at 50 and
100 lg mL-1 (Fig. 2c). In macrophages, the increase
of LDH leakage is time- and dose-dependent with both
Fig. 1 Transmission electron micrographs of IONP. Surfac-
tant-coated Fe3O4 nanoparticles (IONP) were dispersed in types of NP, except for the PEG-IONP-PMA at
hexane and observed by TEM 20 lg mL-1 (Fig. 2d).

123
J Nanopart Res (2015)17:351 Page 7 of 13 351

Fig. 2 HUVECs and macrophages RAW264.7 viability after d). The results are reported as percentage respect to control
NP treatment. HUVECs (a, c) and macrophages RAW264.7 (b, (untreated cells). Data are mean ± S.E. of three separate
d) were incubated with different concentrations (20/50/ experiments performed in triplicate. The results were compared
100 lg mL-1) of IONP-PMA and PEG-IONP-PMA for 1 and by Student’s t test. **p \ 0.01; cnt untreated cells. See text for
24 h, the mitochondrial activity was determined by MTT assay abbreviations
(a, b), and membrane integrity was determined by LDH assay (c,

Levels of NO after NP cells treatment The treatment of macrophages with both NP did not
induce an increase in NO production (data not shown).
HUVECs and RAW264.7 were treated with different
doses of NP, and the NO levels were measured. Total nitrotyrosine measurement
IONP-PMA induced in HUVECs a small increase of
NO production after 1 h (about ?25 %) at all doses ELISA quantification of total nitrotyrosine residues in
tested. The NO production after 24 h was dose- the HUVECs lysates after 100 lg mL-1 PEG-IONP-
dependent. PEG-IONP-PMA at the highest dose PMA treatment showed a transient increase starting
(100 lg mL-1) induced a significant increase of NO after 5 h of treatment, reaching a maximum at 10 h
production only after 24 h (about ?114 %) com- and then decreasing (Fig. 3c). We had observed the
pared to untreated cells (Fig. 3a). Resveratrol treat- caspase-3 activation in the cell lysates after 24 h of
ment (used as positive controls) induced in PEG-IONP-PMA treatment (Fig. 3d).
HUVECs an increase of 50 and 75 % of NO
production after 1 h and 24 h, respectively (data not Inflammation process activation
shown) (Orlando et al. 2013).
Given the obtained results on HUVECs with PEG- The release of the pro-inflammatory mediators TNF-a
IONP-PMA at the highest dose (100 lg mL-1), cells induced by NP in endothelial cells was investigated in
were treated with these NP also for different times to the cell media after 1 and 24 h. No differences were
understand whether NO production occurred during observed with respect to untreated cells (data not
the first hours of treatment. Our results show that NO shown).
production started from 5 h after PEG-IONP-PMA The production and the release of the pro-inflam-
treatment consequently at the activation of eNOS by matory mediators (TNF-a, IL-1b and COX2) induced
phosphorylation (Fig. 3b). by NP in macrophages were investigated in the cell

123
351 Page 8 of 13 J Nanopart Res (2015)17:351

Fig. 3 Nitric oxide production, total nitrotyrosine residues and evaluated by immunoblotting and ECL detection. d Cells were
caspase 3 activation in HUVECs after NP treatment. a Cells incubated with 100 lg mL-1 of PEG-IONP-PMA for 1–24 h
were incubated with different concentrations (20/50/ and total nitrotyrosine residues in the lysates were quantified by
100 lg mL-1) of IONP-PMA and PEG-IONP-PMA for 1 or the ELISA kit. e Cells were incubated with 100 lg mL-1 of
24 h, and the NO production was measured in conditioned PEG-IONP-PMA for 24 h, and caspase 3 activation was
media by Griess method. b Cells were incubated with analysed by immunoblotting followed by ECL detection (see
100 lg mL-1 of PEG-IONP-PMA for 1–24 h, and the NO text for details). Data are mean ± S.E. of three separate
production was measured in conditioned media by Griess experiments performed in triplicate. The results were compared
method. c Cells were incubated with 100 lg mL-1 of PEG- by Student’s t test. *p \ 0.05, **p \ 0.01; cnt untreated cells.
IONP-PMA for 1–24 h, and the eNOS phosphorylation was See text for abbreviations

lysate after 1 and 24 h and in the cell media after 24 h The ICP-OES analysis revealed an increased
of treatment with NP (Fig. 4). TNF-a and IL-1b were uptake of IONP-PMA, slightly by HUVECs and
produced already after 1 h of treatment with IONP- avidly by macrophages (Table 1) with respect to
PMA. PEG-IONP-PMA induced TNF-a production PEG-IONP-PMA.
after 1 h of treatment with 50 and 100 lg mL-1, Confocal images showed that after 1 h of incuba-
whereas the treatment with 20 lg mL-1 induced a late tion, NP appeared inside the cells (Fig. 5). As far as
production (Fig. 4a); no changes were observed in the concerned HUVEC cells, we could detect the green
IL-1b levels (Fig. 4b). fluorescence associated to NP in different region of the
COX-2 amount was no different in cells incubated cytoplasm, but preferentially clustered, as shown in
with IONP-PMA or PEG-IONP-PMA (data not Fig. 5b, while in the macrophages it appeared spread
shown). throughout the cytoplasm (Fig. 5e, f).

Cellular uptake of NP Transport of NP across the BBB model

HUVECs and RAW264.7 cells were treated with Transcytosis ability of NP across a BBB model was
20 lg mL-1 for 1 h. NP internalization was observed determined quantitatively with hCMEC/D3 cells.
by means of ICP-OES and confocal microscopy. Cells were incubated with different NP when TEER

123
J Nanopart Res (2015)17:351 Page 9 of 13 351

Fig. 4 Production of
inflammatory mediators
TNF-a and IL-1b by
macrophages RAW264.7
after treatment with NP.
Cells were incubated with
different concentrations (20/
50/100 lg mL-1) of IONP-
PMA and PEG-IONP-PMA
for 1 or 24 h, and a TNF-a
and b IL-1b were analysed
by immunoblotting
followed by ECL detection
(see text for details). Data
are mean ± S.E. of three
separate experiments
performed in triplicate. The
results were compared by
Student’s t test. **p \ 0.01;
cnt = untreated cells, cnt
pos positive control. See text
for abbreviations

Table 1 Levels of Cell-Associated NP attributable to IONP exposure. Available evidence

5
Cellular NP levels (10 NP/cell)
suggests that most metal NP that are being used have
cytotoxic effects (Wu et al. 2010; Van Tiel et al.
NP HUVEC Macrophages 2010). Recent studies have also suggested pro-
IONP-PMA 1.44 5 inflammatory effects of these NP (Zhu et al. 2011;
PEG-IONP-PMA 1 0.11 Li et al. 2012). The objective of this study was
therefore to determine the biocompatibility and
inflammatory potential in vitro of IONP-PMA,
value was the maximal (56 ± 6 X cm2). After 3 h of PEGylated or not, using human umbilical vein
incubation with 20 lg mL-1 of NP, the amount of iron endothelial cells (HUVECs) and mouse macrophages
in the lower chamber was measured by ICP-OES and (RAW264.7) as cellular models exposed to NP after
expressed as the percentages of the total applied dose. systemic administration. IONP were tested at con-
This amount was 4.8 ± 0.3 and 4.6 ± 0.3 % for centrations commonly used in vivo as potential agent
IONP-PMA and PEG-IONP-PMA, respectively. for MRI contrast (Shevtsov et al. 2014; Xiao et al.
2014; Wittenborn et al. 2014). Several endpoints
Discussion were studied: cell viability and cell toxicity, produc-
tion of NO, nitrotyrosine residues, TNF-a, COX-2
In view of the growing use in a broad range of and IL-1b.
medical applications, there is an urgent need to In HUVEC cultures, IONP-PMA reduced the cell
understand the potential adverse effects on cells viability already at very low concentrations after 1 h,

123
351 Page 10 of 13 J Nanopart Res (2015)17:351

Fig. 5 Representative confocal images of HUVECs and PMA (c and f) for 1 h. In red is the cytoskeleton actin staining
macrophages RAW264.7 exposed to NP. HUVECs (a, b and and in green the fluorescence associated to the NP. (Color figure
c) and macrophages (d, e and f) were incubated or not (a and online)
d) with 20 lg mL-1 of IONP-PMA (b and e) or PEG-IONP-

whereas PEG-IONP-PMA had a lower impact; indeed, We have observed that the reduction of HUVECs
we have observed a reduction of mitochondrial viability after IONP-PMA treatment was accompanied
enzyme succinate dehydrogenase activity by MTT by NO production, and instead, the NO production
assay and a loss of plasma membrane permeability, induced by PEG-IONP-PMA was apparent only at the
measured by LDH assay, only after 24 h. higher concentrations after 24 h.
Endothelial cells maintain basal vascular tone and Recent evidence indicates that most of the cyto-
actively regulate vascular reactivity under physiolog- toxicity attributed to NO is rather due to peroxynitrite,
ical and pathological conditions. They respond to produced by the reaction between NO and the
mechanical forces and neurohumoral mediators with superoxide anion (Pacher et al. 2007). Previous studies
the release of a variety of relaxing and constricting showed that increments in peroxynitrite formation are
factors. Nitric oxide (NO) is one of the most important involved in the signalling pathway of programmed cell
vasoactive substances released by the endothelium, death leading to vascular cell loss (Sennlaub et al.
not only acting as a vasodilator, but also inhibiting 2002; Beauchamp et al. 2004; Kowluru and Odenbach
inflammation of cells (Galley and Webster 2004). 2004, Kowluru et al. 2007; Abdelsaid et al. 2010).
Inflammatory cytokines, such as tumor necrosis factor We have investigated these phenomena on endothe-
a (TNF-a), play a critical role in atherogenesis and lial cells treated with PEG-IONP-PMA, which are the
cause endothelial dysfunction (Rosenkranz-Weiss best candidates for the in vivo experiments. The time-
et al. 1994; Anderson 2003; Jantzen et al. 2007). course experiments revealed that the NO production,

123
J Nanopart Res (2015)17:351 Page 11 of 13 351

as a result of eNOS phosphorylation, started after 5 h inflammatory response (Cohen-Sela et al. 2006; Afer-
of NP treatment, as previously reported in another gan et al. 2010).
endothelial cell line (Zhu et al. 2011), and this Indeed, regarding the production of pro-inflamma-
phenomenon was accompanied by a significant tory mediators, IONP-PMA exposure stimulated the
increase in total nitrotyrosine residues leading to production of TNF-a and IL-1b, whereas PEG-IONP-
apoptosis, as shown by caspase-3 activation. Since NO PMA stimulated a late TNF-a and no IL-1b production,
plays a role in attenuating the inflammation responses both IONP not induced COX-2 and NO production by
of endothelial cells (Galley and Webster 2004), we macrophages.
evaluated the inflammatory status of HUVECs mea- Uptake analysis demonstrated that IONP-PMA
suring if the pro-inflammatory mediator TNF-a was were strongly internalized by macrophages, and
present in the cell-conditioned media after NP treat- therefore, NP may be recognized as ‘foreign’ and
ment. The levels of this cytokine were not different trigger the inflammatory response. The uptake of
from control ones. PEG-IONP-PMA by HUVECs is lower than no-
We found that the cellular uptake of the NP was PEGylated NP; this could explain the greater biocom-
followed by an enhanced NO release in HUVECs. patibility of these NP.
Since it is known that the internalization of NP by hCMEC/D3 cells, displaying characteristics and
endothelial cells leads the induction of NO release functionality mimicking the basic features of the BBB
(Nishikawa et al. 2009), we suppose that NO produc- (Weksler et al. 2013), were used to study the
tion following eNOS activation could be regulated by transcytosis ability of NP.
the PEG-IONP-PMA caveolae-mediated internaliza- Results showed that both IONP-PMA and PEG-
tion. Further studies will be needed to deepen the IONP-PMA crossed the BBB model more than other
internalization pathway of these NP. reported iron oxide NP (Dan et al. 2015). This innate
In our study, we have used also mouse macro- ability could be exploited and increased in future
phages RAW264.7 cells, which represent a primary experiments of IONP functionalization with brain-
line of defence to foreign materials since they specific ligands.
generate various pro-inflammatory mediators includ-
ing TNF-a, interleukin-1b (IL-1b), IL-6, COX-2 and
NO (Lin and Karin 2007). When these mediators are Conclusion
overproduced, they cause excessive inflammatory
responses (Lawrence et al. 2002; Choi et al. 2009; For the correct evaluation of cell function and
Choi et al. 2011). Concerning the interaction with behaviour, any effects of the NP on the cells must be
macrophages, it has already been demonstrated that completely ruled out. In our study, two different
NP are taken up by these circulating cells both in vitro methods for analysing cell viability are presented and
and in vivo (Dobrovolskaia et al. 2008; Walkey et al. several important factors in studying cytotoxic effects
2012) and that macrophages play an important role in are addressed.
the inflammatory processes in response to particles In conclusion, we has demonstrated that PEG-
(Lucarelli et al. 2004; Mitchell et al. 2007; Waldman IONP-PMA were slightly internalized by cultured
et al. 2007). endothelial and macrophages cells and exhibited no
Treatment of macrophages with IONP-PMA sig- cytotoxicity and pro-inflammatory response at the
nificantly reduced the cell viability only after 24 h, lower concentration tested. These biocompatible
and the PEG-IONP-PMA induced a cell impaired characteristics and their ability to pass the BBB make
metabolism after 1 h, as showed by MTT assay results, PEG-IONP-PMA promising candidate for biomedical
but a loss of plasma membrane permeability, mea- applications in biological diagnostics, cell imaging
sured with the LDH assay, was observed only after and drug delivery.
24 h. These data suggest that PEG-IONP-PMA cause
mild metabolic stress in the cell, which could Acknowledgments The authors report no conflict of interest.
The authors are the sole responsible for the content and writing
transiently inactivate the macrophages compromising
of the paper. This work was supported by grants from FAR
their function and leading to a diminished 2010, FAR 2011, and ‘‘The MULAN Project’’ from Cariplo

123
351 Page 12 of 13 J Nanopart Res (2015)17:351

Foundation (Grant No. 2011-2096). We thank Pierre-Olivier Galley HF, Webster NR (2004) Physiology of the endothelium.
Couraud for providing the hCMEC/D3 cells. Br J Anaesth 93:105–113
Gupta AK, Gupta M (2005) Synthesis and surface engineering
of iron oxide nanoparticles for biomedical applications.
Biomaterials 26(18):3995–4021
Hardman R (2006) A toxicologic review of quantum dots:
References toxicity depends on physicochemical and environmental
factors. Environ Heath Perspect 114:165–172
Abdelsaid MA, Pillai BA, Matragoon S, Prakash R, Al-Shab- Hevel JM, Marletta MA (1994) Nitric-oxide synthase assays.
rawey M et al (2010) Early intervention of tyrosine nitra- Methods Enzymol 233:250–258
tion prevents vaso-obliteration and neovascularization in Ito A, Shinkai M, Honda H, Kobayashi T (2005) Medical
ischemic retinopathy. J Pharmacol Exp Ther 332(1): application of functionalized magnetic nanoparticles.
125–134 J Biosci Bioeng 100(1):1–11
Afergan E, Ben David M, Epstein H, Koroukhov N, Gilhar D Jantzen F, Könemann S, Wolff B, Barth S, Staudt A et al (2007)
et al (2010) Liposomal simvastatin attenuates neointimal Isoprenoid depletion by statins antagonizes cytokine-in-
hyperplasia in rats. AAPS J 12(2):181–187 duced down-regulation of endothelial nitric oxide expres-
Anderson TJ (2003) Nitric oxide, atherosclerosis and the clinical sion and increases NO synthase activity in human
relevance of endothelial dysfunction. Heart Fail Rev umbilical vein endothelial cells. J Physiol Pharmacol
8:71–86 58(3):503–514
Bana L, Minniti S, Salvati E, Sesana S, Zambelli V et al (2014) Klostranec JM, Chan WCW (2006) Quantum dots in biological
Liposomes bi-functionalized with phosphatidic acid and an and biomedical research: recent progress and present
ApoE-derived peptide affect Ab aggregation features and challenges. Adv Mater 18:1953–1964
cross the blood–brain-barrier: implications for therapy of Kowluru RA, Odenbach S (2004) Effect of long-term admin-
Alzheimer disease. Nanomedicine 10(7):1583–1590 istration of alphalipoic acid on retinal capillary cell death
Beauchamp MH, Sennlaub F, Speranza G, Gobeil F Jr, Checchin and the development of retinopathy in diabetic rats. Dia-
D et al (2004) Redoxdependent effects of nitric oxide on betes 53:3233–3238
microvascular integrity in oxygen-induced retinopathy. Kowluru RA, Kanwar M, Kennedy A (2007) Metabolic memory
Free Radic Biol Med 37:1885–1894 phenomenon and accumulation of peroxynitrite in retinal
Bulbarelli A, Lonati E, Brambilla A, Orlando A, Cazzaniga E capillaries. Exp Diabetes Res 2007:21976
et al (2012) Ab42 production in brain capillary endothelial Lawrence T, Willoughby DA, Gilroy DW (2002) Anti-inflam-
cells after oxygen and glucose deprivation. Mol Cell matory lipid mediators and insights into the resolution of
Neurosci 49(4):415–422 inflammation. Nat Rev Immunol 2(10):787–795
Choi J, Zhang Q, Reipa V, Wang NS, Stratmeyer ME et al Li M, Kim HS, Tian L, Yu MK, Jon S et al (2012) Comparison of
(2009) Comparison of cytotoxic and inflammatory two ultrasmall superparamagnetic iron oxides on cyto-
responses of photoluminescent silicon nanoparticles with toxicity and MR imaging of tumors. Theranostics
silicon micron-sized particles in RAW 264.7 macrophages. 2(1):76–85
J Appl Toxicol 29(1):52–60 Lin WW, Karin M (2007) A cytokine-mediated link between
Choi YH, Jin GY, Li GZ, Yan GH (2011) Cornuside suppresses innate immunity, inflammation, and cancer. J Clin Investig
lipopolysaccharide-induced inflammatory mediators by 117(5):1175–1183
inhibiting nuclear factor-kappa B activation in RAW 264.7 Liong M, Shao H, Haun JB, Lee H, Weissleder R (2010) Car-
macrophages. Biol Pharm Bull 34(7):959–966 boxymethylated polyvinyl alcohol stabilizes doped fer-
Cohen-Sela E, Dangoor D, Epstein H, Gati I, Danenberg HD rofluids for biological applications. Adv Mater
et al (2006) Nanospheres of a bisphosphonate attenuate 22:5168–5172
intimal hyperplasia. J Nanosci Nanotechnol 6(9–10): Lovren F, Pan Y, Shuklap P, Quan A, Teoh H et al (2009)
3226–3234 Visfatin activates eNOS via Akt and MAP kinases and
Colombo M, Mazzucchelli S, Montenegro JM, Galbiati E, Corsi improves endothelial cell function and angiogenesis
F et al (2012a) Protein oriented ligation on nanoparticles in vitro and in vivo: translational implications for
exploiting O6-alkylguanine-DNA transferase (SNAP) atherosclerosis. Am J Physiol Endocrinol Metab
genetically encoded fusion. Small 8(10):1492–1497 296(6):E1440–E1449
Colombo M, Sommaruga S, Mazzucchelli S, Polito L, Verderio Lucarelli M, Gatti AM, Savarino G, Quattroni P, Martinelli L
P et al (2012b) Site-specific conjugation of scFv antibodies et al (2004) Innate defence functions of macrophages can
to nanoparticles by bioorthogonal strain-promoted alkyne- be biased by nano-sized ceramic and metallic particles. Eur
nitrone cycloaddition. Angew Chem Int Ed 51:496–499 Cytokine Netw 15:339–346
Dan M, Bae Y, Pittman TA, Yokel RA (2015) Alternating Mahmoudi M, Simchi A, Milani AS, Stroeve P (2009) Cell
magnetic field-induced hyperthermia increases iron oxide toxicity of superparamagnetic iron oxide nanoparticles.
nanoparticle cell association/uptake and flux in blood-brain J Colloid Interface Sci 336(2):510–518
barrier models. Pharm Res 32(5):1615–1625 Mazzucchelli S, Colombo M, Verderio P, Rozek E, Andreata F
Dobrovolskaia MA, Aggarwal P, Hall JB, McNeil SE (2008) et al (2013) Orientation-controlled conjugation of
Preclinical studies to understand nanoparticle interaction haloalkane dehalogenase fused homing peptides to multi-
with the immune system and its potential effects on functional nanoparticles for the specific recognition of
nanoparticle biodistribution. Mol Pharm 5(4):487–495 cancer cells. Angew Chem Int Ed Engl 52(11):3121–3125

123
J Nanopart Res (2015)17:351 Page 13 of 13 351

Mitchell LA, Gao J, Vander Wal R, Gigliotti A et al (2007) maghemite nanoparticle surface-coated with meso-2-3-
Pulmonary and systemic immune response to inhaled dimercaptosuccinic acid to serum albumin. J Nanosci
multiwalled carbon nanotubes. Toxicol Sci 100:203–214 Nanotechnol 8(11):5813–5817
Nishikawa T, Iwakiri N, Kaneko Y, Taguchi A, Fukushima K et al Soenen SJ, Rivera-Gil P, Montenegro JM, Parak WJ, De Smedt
(2009) Nitric oxide release in human aortic endothelial cells SC et al (2011) Cellular toxicity of inorganic nanoparticles:
mediated by delivery of amphiphilic polysiloxane nanopar- common aspects and guidelines for improved nanotoxicity
ticles to caveolae. Biomacromolecules 10:2074–2085 evaluation. Nano Today 6:446–465
Orlando A, Re F, Sesana S, Rivolta I, Panariti A et al (2013) Ulivi V, Lenti M, Gentili C, Marcolongo G, Cancedda R et al
Effect of nanoparticles binding b-amyloid peptide on nitric (2011) Anti-inflammatory activity of monogalactosyldia-
oxide production by cultured endothelial cells and mac- cylglycerol in human articular cartilage in vitro: activation
rophages. Int J Nanomed 8:1335–1347 of an anti-inflammatory cyclooxygenase-2 (COX-2) path-
Pacher P, Beckman JS, Liaudet L (2007) Nitric oxide and per- way. Arthritis Res Ther 13(3):R92
oxynitrite in health and disease. Physiol Rev Van Tiel ST, Wielopolski PA, Houston GC, Krestin GP, Bern-
87(1):315–424 sen MR (2010) Variations in labeling protocol influence
Panariti A, Lettiero B, Alexandrescu R, Collini M, Sironi L et al incorporation, distribution and retention of iron oxide
(2013) Dynamic investigation of interaction of biocom- nanoparticles into human umbilical vein endothelial cells.
patible iron oxide nanoparticles with epithelial cells for Contrast Media Mol Imaging 5(5):247–257
biomedical applications. J Biomed Nanotechnnol Waldman WJ, Kristovich R, Knight DA, Dutta PK (2007)
9(9):1556–1569 Inflammatory properties of iron-containing carbon
Piazza M, Colombo M, Zanoni I, Granucci F, Tortora P et al nanoparticles. Chem Res Toxicol 20:1149–1154
(2011) Uniform LPS-loaded magnetic nanoparticles for the Walkey CD, Olsen JB, Guo H, Emili A, Chan WC (2012)
investigation of LPS/TLR4 signaling. Angew Chem Int Nanoparticle size and surface chemistry determine serum
50:622–626 protein adsorption and macrophage uptake. J Am Chem
Romero-Calvo I, Ocón B, Martı́nez-Moya P, Suárez MD, Zar- Soc 134(4):2139–2147
zuelo A et al (2010) Reversible Ponceau staining as a Weksler B, Romero IA, Couraud PO (2013) The hCMEC/D3
loading control alternative to actin in Western blots. Anal cell line as a model of the human blood brain barrier. Fluids
Biochem 401(2):318–320 Barriers CNS 10(1):16
Rosenkranz-Weiss P, Sessa WC, Milstien S, Kaufman S, Wat- Wittenborn TR, Larsen EK, Nielsen T, Rydtoft LM, Hansen L
son CA et al (1994) Regulation of nitric oxide synthesis by et al (2014) Accumulation of nano-sized particles in a
proinflammatory cytokines in human umbilical vein murine model of angiogenesis. Biochem Biophys Res
endothelial cells. Elevations in tetrahydrobiopterin levels Commun 443(2):470–476
enhance endothelial nitric oxide synthase specific activity. Wu X, Tan Y, Mao H, Zhang M (2010) Toxic effects of iron
J Clin Invest 93:2236–2243 oxide nanoparticles on human umbilical vein endothelial
Sennlaub F, Courtois Y, Goureau O (2002) Inducible nitric cells. Int J Nanomed 9(5):385–399
oxide synthase mediates retinal apoptosis in ischemic Xiao N, Gu W, Wang H, Deng Y, Shi X et al (2014) T1-T2 dual-
proliferative retinopathy. J Neurosci 22:3987–3993 modal MRI of brain gliomas using PEGylated Gd-doped
Shevtsov MA, Nikolaev BP, Yakovleva LY, Marchenko YY, iron oxide nanoparticles. J Colloid Interface Sci
Dobrodumov AV et al (2014) Superparamagnetic iron 417:159–165
oxide nanoparticles conjugated with epidermal growth Zhu MT, Wang B, Wang Y, Yuan L, Wang HJ et al (2011)
factor (SPION-EGF) for targeting brain tumors. Int J Endothelial dysfunction and inflammation induced by iron
Nanomed 9:273–287 oxide nanoparticle exposure: risk factors for early
Simoni AR, Garcia MP, Azevedo RB, Chaves SB, Lacava ZG atherosclerosis. Toxicol Lett 203(2):162–171
et al (2008) Evaluation of the binding properties of

123