Nanoparticle-induced unfolding of fibrinogen promotes Mac-1 receptor activation and inflammation.

Abstract

The chemical composition, size, shape and surface characteristics of nanoparticles affect the way proteins bind to these particles, and this in turn influences the way in which nanoparticles interact with cells and tissues. Nanomaterials bound with proteins can result in physiological and pathological changes, including macrophage uptake, blood coagulation, protein aggregation and complement activation, but the mechanisms that lead to these changes remain poorly understood. Here, we show that negatively charged poly(acrylic acid)-conjugated gold nanoparticles bind to and induce unfolding of fibrinogen, which promotes interaction with the integrin receptor, Mac-1. Activation of this receptor increases the NF-κB signalling pathway, resulting in the release of inflammatory cytokines. However, not all nanoparticles that bind to fibrinogen demonstrated this effect. Our results show that the binding of certain nanoparticles to fibrinogen in plasma offers an alternative mechanism to the more commonly described role of oxidative stress in the inflammatory response to nanomaterials.

Full text

Page 1

Nanoparticle-induced unfolding of fibrinogen
promotes Mac-1 receptor activation and
inflammation
Zhou J. Deng1, Mingtao Liang2,3, Michael Monteiro4, Istvan Toth2,3and Rodney F. Minchin1*
The chemical composition, size, shape and surface character-
istics of nanoparticles affect theway proteins bind to these par-
ticles, and this in turn influencestheway in which nanoparticles
interact with cells and tissues1–5. Nanomaterials bound with
proteins can result in physiological and pathological changes,
including macrophage uptake1,6, blood coagulation7, protein
aggregation8and complement activation7,9, but the mechan-
isms that lead to these changes remain poorly understood.
Here, we show that negatively charged poly(acrylic acid)-
conjugated gold nanoparticles bind to and induce unfolding of
fibrinogen, which promotes interaction with the integrin recep-
tor, Mac-1. Activation of this receptor increases the NF-kB sig-
nalling pathway, resulting in the release of inflammatory
cytokines. However, not all nanoparticles that bind to fibrino-
gen demonstrated this effect. Our results show that the
binding of certain nanoparticles to fibrinogen in plasma offers
an alternative mechanism to the more commonly described
role of oxidative stress in the inflammatory response to
nanomaterials.
Poly(acrylic acid)-coated gold nanoparticles (PAA–GNP)10with
hydrodynamic diameters of ?5, 10 and 20 nm were first syn-
thesized. Their interactions with human plasma proteins were
then examined using one- and two-dimensional gel electrophoresis
(Fig. 1). The negatively charged PAA coat prevented agglomeration
in solution (Supplementary Fig. S1a). Nanoparticle–protein binding
reached equilibrium within 5 min and remained constant over 4 h
(Supplementary Fig. S1b). Moreover, bound proteins remained
associated with the nanoparticles even after extensive washings,
indicating a high-affinity interaction (Supplementary Fig. S1c).
Sodiumdodecylsulphate–polyacrylamide
(SDS–PAGE) of human plasma proteins bound to PAA–GNP at
equilibrium showed three major protein bands with molecular
weights of ?65, 55 and 45 kDa (Fig. 1a). Two-dimensional gel elec-
trophoresis (Supplementary Fig. S1d) and mass spectroscopy
(Supplementary Table S1) confirmed that the three bands were a,
b and g chains of fibrinogen, respectively.
The binding capacity of the nanoparticles for the purified human
fibrinogen was determined by ultracentrifugation (Fig. 1b).
Consistent with Fig. 1a, the 20 nm PAA–GNP bound significantly
less fibrinogen than the 5 nm particles on a mass basis. However,
when adjusted for total surface area, the binding was found to be
similar (Fig. 1b, inset). The maximum protein binding was 2 mg
for the 5 nm PAA–GNP, which represents one to two nanoparticles
perfibrinogenmolecule(nanoparticledensity,19.3 g cm23;diameter,
5.0–5.5 nm). For the 20 nm PAA–GNP, maximum binding was 8 mg
nanoparticles, or approximately seven molecules of fibrinogen
gelelectrophoresis
per nanoparticle. This variance was probably due to differences in
surface area (78 nm2and 1,040 nm2for the 5 and 20 nm PAA–
GNP, respectively), suggesting the configuration of bound fibrinogen
may vary for each nanoparticle.
Fibrinogen11, which has a length of ?45 nm and diameter of
5 nm, is much larger than the 5 nm PAA–GNP. The low stoichi-
ometry therefore suggests that the nanoparticles interact with
specific sites on the protein. Fibrinogen comprises three protein
chains (a, b and g) arranged as a dimer, with three domains: a
central E domain with two peripheral D domains11(Fig. 1c). The
protein is negatively charged at physiological pH (pI ¼ 5.5), so its
high-affinity interaction with the negatively charged 5 nm PAA–
GNP was unexpected. The C-terminus of each a chain may act as
a potential binding site12, as it is positively charged at pH 7.4.
Other than its role in clotting, fibrinogen also binds to surfaces,
unfolding to expose sequences normally embedded within
the protein13. One sequence, amino acids 377–395 in the C-termi-
nus of the g chain (g377–395), is located in the D domain and is
responsible for the recruitment of macrophages and leukocytes
(Fig. 1c, inset). This is achieved by the interaction of g377–395with
the Mac-1 receptor (CD11b/CD18 or aMb2)13. The positively
charged C-terminus of the a chain is located adjacent to the D
domain. We therefore posed the question of whether PAA–GNP
could induce unfolding of the fibrinogen molecule and Mac-1
receptor binding.
Far-ultraviolet circular dichroism was first used to monitor the
protein structure in the presence of increasing amounts of 5 nm
PAA–GNP. The spectrum for fibrinogen alone was typical for a
protein with both a-helix and b-sheet structure (Fig. 1d). The
addition of PAA–GNP resulted in a loss of protein secondary struc-
ture, as indicated by the progressive increase in ellipticity. Based on
these observations, we surmised that the PAA–GNP induces
changes in the fibrinogen structure that may expose the gC-termi-
nus, similar to previous reports for flat surfaces13. Figure 2a shows
thatthe 5 nm PAA–GNP significantlyincreased thebinding offibri-
nogen to Mac-1-receptor-positive THP-1 cells but not to Mac-1-
receptor-negative HL-60 cells, an effect not seen with albumin. To
further establishbinding specificity, Mac-1-receptor-negative
HEK293 cells were transfected with CD11b and CD18 (Fig. 2b,
upper panel). In empty vector-transfected cells, no change in fibri-
nogen binding was seen in the presence of PAA–GNP, but it
increased more than twofold in the CD11b/CD18-transfected
cells (Fig. 2b). These experiments show that fibrinogen bound to
5 nm PAA–GNP interacts with the Mac-1 receptor. However, the
larger nanoparticles (20 nm) did not promote cell interaction to
the same extent when adjusted for similar protein binding (Fig. 2c).
1School of Biomedical Sciences, University of Queensland, Brisbane 4072, Australia,
Queensland, Brisbane 4072, Australia,
Nanotechnology, University of Queensland, Brisbane 4072, Australia. *e-mail: r.minchin@uq.edu.au
2School of Chemistry and Molecular Biosciences, University of
4Australian Institute for Bioengineering and
3School of Pharmacy, University of Queensland, Brisbane 4072, Australia,
LETTERS
PUBLISHED ONLINE: 19 DECEMBER 2010 | DOI: 10.1038/NNANO.2010.250
NATURE NANOTECHNOLOGY | VOL 6 | JANUARY 2011 | www.nature.com/naturenanotechnology
39
© 2011 Macmillan Publishers Limited. All rights reserved.

Page 2

The peptide g377–395(P2 ¼ g377YSMKKTTMKIIPFNRLTIG395)
hasbeenshowntoinhibitMac-1-mediatedcelladhesiontoimmobi-
lized fibrinogen14. We used P2 together with a control sequence
(H19 ¼ g340HAGHLNGVYYQ GGTYSKA357)15to further investi-
gate nanoparticle-induced fibrinogen binding to the THP-1 cells.
At concentrations as low as 10 mM, P2 significantly inhibited
binding, but H19 had no effect (Fig. 2d). This result supports the
model of nanoparticle-induced unfolding of the D-domain of fibri-
nogen. However, the 20 nm nanoparticles did not appear to expose
the C-terminus of the g chain. This was not due to a lack of
proteinunfolding, asdemonstrated
(Supplementary Fig. S2). The apparent size of the 20 nm PAA–
GNP in the presence of saturating fibrinogen, determined by
dynamic light scattering, was 34.3+0.6 nm (n ¼ 5), suggesting that
the fibrinogen corona extended ?7 nm from the nanoparticle
surface. Witheachnanoparticlebinding?7moleculesoffibrinogen,
bycircular dichroism
crowding of the particle surface may result in steric hindrance, pre-
venting the exposed C-terminus of the g chain from interacting
with the Mac-1 receptor. To address this, 20 nm PAA–GNPs were
bound to decreasing amounts of fibrinogen (33–100% saturation)
and incubated with THP-1 or HL-60 cells (Fig. 2e). Binding to
THP-1 cells increased by more than a factor of six, but no change
wasseenwithHL-60cells,suggestingthatsterichindranceprevented
Mac-1 receptor interaction at high protein binding.
Mac-1 receptor activation increases NF-kB activity by promot-
ing the degradation of IkB, which then allows NF-kB to translocate
into the nucleus, and upregulates a number of pro-inflammatory
genes16. To examine this, electromobility shift assays were per-
formed using an NF-kB probe17. Fibrinogen bound to the 5 nm
PAA–GNP increased nuclear NF-kB (Fig. 2f, lane 4), which was
confirmed by anti-P65 antibody supershift (Fig. 2f, lane5).
Fibrinogen (Fig. 2f, lane 2) and PAA–GNP alone (Fig. 2f, lane 3)
also increased NF-kB in the nucleus, albeit to a lower extent.
Soluble fibrinogen is known to activate NF-kB in several cell
lines18via receptors other than Mac-1. These results show
that the fibrinogen/PAA–GNP complexes can increase nuclear
NF-kB in THP-1 cells.
NF-kB has been implicated as a key factor in many inflammatory
diseases19. To test whether fibrinogen bound to PAA–GNP could
increase secretion of inflammatory cytokines, THP-1 cells were
treated with either fibrinogen or 5 nm PAA–GNP, or both
(Fig. 3a,b). Neither fibrinogen norPAA–GNPalone altered cytokine
release, but the fibrinogen/PAA–GNP complex increased IL-8 and
TNF-a secretion by factors of 28 and 67, respectively (Fig. 3a,b).
This increase was attenuated by BAY 11-7082, an inhibitor of the
NF-kB pathway. Cells treated with fibrinogen and the 20 nm
PAA–GNP secreted cytokine only when non-saturating protein
levels were used (Fig. 3c–f). This is consistent with the results
shown in Fig. 2e.
We next investigated whether the surface charge density of
the 5 nm PAA–GNPs was affecting fibrinogen binding, by
using increasing proportions of the neutral polymer poly(2,3-
hydroxy-propylacrylamide) (PDHA) in the polymer coating
(Supplementary Table S2), which progressively increased the zeta
potential (Supplementary Fig. S3). Fibrinogen binding decreased
significantly with 20% PDHA, and was completely inhibited at
higher proportions of PDHA (Fig. 4a). Fibrinogen did not bind to
THP-1 cells in the presence of any PDHA-containing nanoparticles
(Fig. 4b). These results show that surface charge density is a critical
factor for fibrinogen binding and that small changes in charge
density can influence the ability of the fibrinogen–nanoparticle
complexes to interact with the Mac-1 receptor.
Previously, metal-oxide nanoparticles have been shown to bind
an array of plasma proteins including fibrinogen5. In this exper-
iment, nano-SiO2(7 nm), nano-TiO2(21 nm) and nano-ZnO
(30 nm) bound fibrinogen with approximately equal capacities
(Fig. 4c). However, their ability to promote fibrinogen binding to
THP-1 cells varied markedly. The smallest metal oxide (nano-
SiO2) increased fibrinogen binding by more than a factor of 20,
whereas the largest metal oxide (nano-ZnO) showed no significant
binding (Fig. 4d). Moreover, binding of nano-SiO2–fibrinogen to
THP-1 cells was inhibited by P2 (Fig. 4e) indicating that at least
part of the cell interaction involved Mac-1. These results suggest
that binding to fibrinogen alone does not predict Mac-1
receptor interaction.
A number of studies have shown that nanoparticle-bound pro-
teins undergo conformational changes20–22and it has been
suggested that unfolding of proteins could trigger the immune
system by exposing normally buried sequences3,4. It has been
shown that a high-affinity Mac-1 binding site exists in the gC-term-
inal tail region of fibrinogen. However, it is normally buried and
protected within the hydrophobic core of the protein15. This
a
D domainE domainD domain
~45 nm
C terminus of
α chain
C terminus of
α’ chain
95
72
55
43
Molecular weight (kDa)
c
200220240 260
−6
−4
−2
0
2
10 μg ml−1 fibrinogen + 10 μg ml−1 PAA-GNP
10 μg ml−1 fibrinogen
10 μg ml−1 fibrinogen + 30 μg ml−1 PAA-GNP
10 μg ml−1 fibrinogen + 20 μg ml−1 PAA-GNP
10 μg ml−1 fibrinogen + 40 μg ml−1 PAA-GNP
40 μg ml−1 PAA-GNP
Wavelength (nm)
Molar ellipticity
(deg cm2 dmol−1 × 103)
d
Size (nm)
51020
b
Unbound fibrinogen (%)
02468 10
0
50
100
150
PAA-GNP (μg)
0500 1,000 1,500
Surface area (mm2)
0
50
100
150
Unbound fibrinogen (%)
Figure 1 | Fibrinogen is the major human plasma protein bound by PAA–
GNP. a, SDS–PAGE of human plasma proteins bound to PAA–GNP with
diameters of 5, 10 and 20 nm. Three major protein bands were observed at
?65, 55 and 45 kDa. b, Unbound fibrinogen following pull-down with PAA–
GNP with diameters of ?5 nm (blue) or 20 nm (red). Purified fibrinogen
(0.6mg) was incubated with increasing amounts of PAA–GNP. Inset:
unbound fibrinogen is plotted against total surface area for the two
nanoparticles. c, Crystal structure of fibrinogen. The protein was drawn using
Swiss-PdbViewer and coordinates for PDB entry 3GHG. Common domains
are shown. Inset: the C-terminus of the g chain (purple) that interacts with
the Mac-1 receptor. d, Circular dichroism of fibrinogen in the absence and
presence of increasing concentrations of 5 nm PAA–GNP.
LETTERS
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.250
NATURE NANOTECHNOLOGY | VOL 6 | JANUARY 2011 | www.nature.com/naturenanotechnology
40
© 2011 Macmillan Publishers Limited. All rights reserved.

Page 3

strongly agrees with our observations that fibrinogen unfolds on
nanoparticles, and nanoparticle-bound fibrinogen interacts with
Mac-1 receptors. However, mere binding to fibrinogen does not
predict Mac-1 receptor interaction. For particles as large as
20 nm, the degree of fibrinogen binding appears to influence the
accessibility of the C-terminus of the g chain. When completely
bound by fibrinogen, the 20 nm PAA–GNP did not induce cytokine
release. This is relevant to particles in the circulation where excess
fibrinogen is available, and may be important in designing
negatively charged nanoparticles for therapeutic use, such as drug
delivery23. Interestingly, the nano-metal oxides also demonstrated
size-dependent effects with fibrinogen. The metal oxides tend to
agglomerate in physiological solution, resulting in irregularly
shaped structures, often with single nanoparticles protruding from
their surfaces5. This may provide sufficient surface curvature for
the nano-SiO2to bind and unfold fibrinogen.
0
1,000
2,000
3,000
4,000
Fibrinogen + 5 nm PAA-GNP
Fibrinogen + 20 nm PAA-GNP
*
Fibrinogen
Bound protein (FU)
EV Mac-1
0
500
1,000
1,500
Fibrinogen
Fibrinogen + PAA-GNP
Bound protein (FU)
THP-1HL-60THP-1
0
500
1,000
1,500
2,000
Fibrinogen
Albumin
Protein
Protein + PAA-GNP
*
Bound protein (FU)
a

Fibrinogen −+−++
PAA-GNP −−+++
Anti-P65−−−−+
c
b
*

0
20
40
60
80
100
120
Bound protein (%control)
C
P2 (μM)H19 (μM)
*
*
d
FluorescenceFluorescence
Count
HL-60
THP-1
102
104
102
104
FluorescenceFluorescence
102
104
102
104
Count
EV
Mac-1
f
e
0
250
500
750
Bound protein (%control)
100
60
40
33
100
60
40
33
THP-1 HL-60
*
*
*
*
30 10 3010
Figure 2 | Selective binding of fibrinogen/PAA–GNP complexes to Mac-1 receptors. a, HL-60 cells are Mac-1-receptor-negative cells, as shown by flow
cytometry following labelling with fluorescent CD11b antibodies (upper panel, solid lines) and THP-1 cells are Mac-1-receptor-positive. This was confirmed by
plating both HL-60 and THP-1 cells onto immobilized fibrinogen (Supplementary Fig. S4). Lower panel shows binding of fibrinogen alone (open bars) or
fibrinogen with 5 nm PAA–GNP (filled bars) to THP-1 cells and HL-60 is shown. Addition of nanoparticles significantly increased the binding to THP-1 cells
but not to HL-60 cells. Replacing fibrinogen with albumin abrogated the effect of the nanoparticles. Results are mean+s.e.m, n¼3; asterisk indicates P,
0.05. b, HEK293 cells were transfected with empty vector (EV) or CD11b/CD18 constructs and Mac-1 receptor expression was determined by flow
cytometry (upper panels). Lower panel shows binding of fibrinogen alone (open bars) or fibrinogen with 5 nm PAA–GNP (filled bars) is shown. The
nanoparticles increased fibrinogen binding in Mac-1-receptor-positive cells only. Results are mean+s.e.m., n¼3; asterisk indicates P,0.05. c, PAA–GNP
(20 nm) did not induce fibrinogen binding to THP-1 cells to the same extent as the 5 nm. Binding was performed with an equal surface area (3,900 mm2)
for each of the nanoparticles to ensure similar protein binding. d, Pre-treatment of THP-1 cells with the P2 peptide significantly reduced the binding
of fibrinogen in the presence of 5 nm PAA–GNP (grey bars). Control peptide, H19, showed no inhibitory effects on binding (filled bars). Results are
mean+s.e.m., n¼3. e, PAA–GNP (20 nm) induced fibrinogen binding to THP-1 cells when the level of protein binding was reduced. However, no increase in
binding was seen under the same conditions in HL-60 cells. Percent binding saturation is shown on the x-axis and was determined from the data shown in
Fig. 1b. Results are mean+s.e.m., n¼3, expressed as a percentage of the control (100%). f, Electrophoretic mobility shift assay, showing the increased
nuclear localization of NF-kB in THP-1 cells when exposed to the fibrinogen/PAA–GNP complexes. The open arrow indicates the location of the P65 complex
of NF-kB, which was confirmed by a supershift (solid arrow) with anti-P65 antibody (lane 5).
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.250
LETTERS
NATURE NANOTECHNOLOGY | VOL 6 | JANUARY 2011 | www.nature.com/naturenanotechnology
41
© 2011 Macmillan Publishers Limited. All rights reserved.

Page 4

0
200
400
600
TNF-α (pg ml−1)
0
500
1,000
1,500
IL-8 (pg ml−1)
a
Fibrinogen
PAA-GNP
LPS
Bay 11-7082
400
800
TNF-α (pg ml−1)
b
Fibrinogen
5 nm PAA-GNP
20 nm PAA-GNP
LPS
c
d
200
400
600
800
IL-8 (pg ml−1)
*
*
**
**
*
*
**
**
Fibrinogen
PAA-GNP
LPS
Bay 11-7082
Fibrinogen
5 nm PAA-GNP
20 nm PAA-GNP
LPS
0
50
100
150
TNF-α (pg ml−1)
0
50
100
150
200
IL-8 (pg ml−1)
Fibrinogen
5 nm PAA-GNP
20 nm PAA-GNP
e
f
*
*
*
*
Fibrinogen
5 nm PAA-GNP
20 nm PAA-GNP
Figure 3 | Pro-inflammatory effects of fibrinogen/PAA–GNP complexes. a,b, Treatment of THP-1 cells with complexes of fibrinogen (33mg ml21) and 5 nm
PAA–GNP (100mg ml21) induced the secretion of IL-8 (a) and TNF-a (b). Neither fibrinogen alone nor PAA–GNP altered cytokine release. The concentration
effect of the complexes on cytokine release is shown in Supplementary Fig. S5. NF-kB pathway inhibitor, Bay 11-7082, inhibited the secretion of both
cytokines. Lipopolysaccharide-treated cells were used as a positive control. Results are mean+s.e.m., n¼3. Single asterisk indicates P,0.05 compared to
control (no treatment). Double asterisk indicates P,0.05 compared to respective treatments without Bay 11-7082. c,d, THP-1 cells treated with 20 nm
PAA–GNP bound to excess fibrinogen did not induce IL-8 (c) or TNF-a (d) release. Nanoparticles were adjusted to equal total surface area for similar protein
binding. e,f, Cells treated with 20 nm PAA–GNP bound to a reduced concentration of fibrinogen (10mg ml21), which resulted in only 33% binding saturation,
induced low but significant release of IL-8 (e) and a more substantial release of TNF-a (f). Nanoparticles (5 nm; 30mg ml21) bound to 10mg ml21fibrinogen
are shown for comparison. Asterisk indicates P,0.05 compared to nanoparticles alone (respective controls).
LETTERS
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.250
NATURE NANOTECHNOLOGY | VOL 6 | JANUARY 2011 | www.nature.com/naturenanotechnology
42
© 2011 Macmillan Publishers Limited. All rights reserved.

Page 5

The pro-inflammatory properties of different nanomaterials have
been reported previously, often linked with the ability to induce oxi-
dative stress24–26. However, recent studies suggest there is not always
a good correlation between free radical generation and inflam-
mation27,28. The current study proposes an alternative pathway for
nanoparticles to provoke an inflammation response. Intriguingly,
theinteractionoffibrinogenandMac-1iscriticalforhostimmunity29,
and can exacerbate inflammation in conditions such as Alzheimer’s30
and arthritis29. Whether fibrinogen–nanoparticle complexes could
aggravate those pathologies remains to be determined.
Insummary,thecurrentstudyshowsthatnegativelychargednano-
particles can unfold fibrinogen and promote activation of the Mac-1
receptor pathway. As fibrinogen has been reported to bind many
different types of nanomaterials in plasma, we propose that fibrino-
gen-bound nanoparticles are potentially pro-inflammatory.
Methods
Gold nanoparticles. PAA- and PDHA-coated spherical gold nanoparticles
were synthesized and characterized as reported elsewhere10. The sources and
characterization of the SiO2nanoparticles (nano-SiO2), TiO2nanoparticles
(nano-TiO2) and ZnO nanoparticles (nano-ZnO) have been reported elsewhere5.
Plasma protein binding. Human plasma samples were obtained from citrated blood
of eight healthy individuals according to institutional bioethics approval.
Nanoparticles in phosphate-buffered saline (150 mM sodium chloride and 10 mM
phosphate, pH 7.4, PBS) were incubated with 1% human plasma at 37 8C for 5 min
to 4 h and then separated from unbound proteins by centrifugation at 4 8C for
40 min (5 nm nanoparticles 50,000g; 10 nm nanoparticles 35,000g; 20 nm
nanoparticles 20,000g). Plasma without nanoparticles was used as a control. Pellets
were washed three times with PBS and reconstituted in SDS-containing buffer
(1 mg ml21nanoparticles, 2% SDS, 5% b-mercaptoethanol, 10% glycerol and
62.5 mM Tris-HCl), heated at 95 8C for 5 min (ref. 5) and then separated by
gel electrophoresis.
Protein separation and identification. Gel electrophoresis and mass spectrometry
were performed as described previously5. (Running conditions, protein digestion
and analysis procedures are provided in the Supplementary Information.)
Circular dichroism spectroscopy. Spectra were recorded using a J-815 circular
dichroism spectrometer (JASCO) using a 10-mm path length quartz cell over the
range 190–260 nm at ambient temperature. Data were collected every 0.2 nm with a
bandwidth of 1 nm, at 50 nm min21and averaging overeight scans. Protein samples
in PBS (10 mg ml21) were incubated with PAA–GNP for 5 min at ambient
temperature before each measurement. The final spectrawere baseline-corrected and
data presented as mean residue ellipiticity (u).
Protein labelling. Alexa Fluor 647 succinimidyl ester (Invitrogen) reconstituted in
dimethylsulphoxide (125 mg ml21) was used to fluorescently label protein
(2.5 mg ml21) in 150 mM sodium bicarbonate, pH 8.5 for 1 h. Free dye was
removed using PD-10 desalting columns (GE Health). The far-red fluorescent dye
was chosen to minimize interference from the PAA–GNP. The samples were stored
in aliquots at 220 8C.
Cultured cell experiments. HL-60, HEK293 and THP-1 were obtained from the
American Type Culture Collection (Manassas). Cells were cultured in RPMI 1640
with 5% fetal bovine serum and penicillin/streptomycin at 37 8C in 5% CO2in air.
For protein binding studies, cells (5 × 105per ml) were resuspended in the RPMI
1640 and incubated with 2.5 mg ml21fluorescently labelled protein in the presence
or absence of 10 mg ml21nanoparticles for 20 min at 37 8C unless otherwise
specified. For binding inhibition, cells were pre-treated with P2 or H19 peptides for
20 min at 37 8C. Following each binding study, cells were washed with RPMI 1640
and analysed by flow cytometry (BD FACSCanto, Becton Dickinson).
For transient transfection studies, HEK293 cells were seeded into six-well plates
(1× 106per well) for 24 h, then transfected with 2.5 mg each of CD11b and CD18
plasmids (courtesy Dr E. Plow, Cleveland Clinic Foundation) using
LipofectamineTM2000 (Invitrogen) according to the manufacturer’s protocol.
Cells were harvested 48 h later in 2 mM ethylenediaminetetraacetic acid and
0
20
40
60
80
100
120
Unbound fibrinogen (%)
*
*
0
500
1,000
1,500
Bound protein (FU)
a
b
*
0
20
40
60
80
100
120
Unbound fibrinogen (%)
ControlSiO2
TiO2
ZnO
*
*
*
0
5,000
10,000
15,000
Bound protein (FU)
*
*
ControlSiO2
TiO2
ZnO
% PAA −
% PDHA −
100 80
0
60
40
40
60
20
80 100
0
20
c
d
0
20
40
60
80
100
120
Bound protein (%control)
Control+P2 +H19
*
e
% PAA −
% PDHA −
100 80
0
60
40
40
60
20
80 100
0
20
Figure 4 | Effects of nanoparticle surface characteristics on fibrinogen binding. a, Unbound fibrinogen following pull-down with PAA–PDHA–GNP. Purified
fibrinogen (0.6mg) was incubated with 2mg of each nanoparticle. Only the 100% PAA/0% PDHA–GNP and 80% PAA/20% PDHA–GNP significantly bound
fibrinogen. b, Binding of fibrinogen to THP-1 cells was only significantly increased with the nanoparticles containing 100% PAA. Results are mean+s.e.m.,
n¼3. Asterisk indicates P,0.05 compared to control (no treatment). c, Binding of fibrinogen to nano-metal oxides was determined by ultracentrifugation
after mixing 0.6mg protein with 8mg nanoparticles. All the metal-oxide nanoparticles bound more than 70% of the fibrinogen. d, Binding of the fibrinogen–
metal oxide complexes to THP-1 cells was determined. Results are mean+s.e.m., n¼3. Asterisk indicates P,0.05 compared to control (no treatment).
e, Pre-treatment of THP-1 cells with the P2 peptide significantly reduced (P,0.05) the binding of fibrinogen in the presence of nano-SiO2. Control peptide,
H19, showed no inhibitory effect on the binding. Results are mean+s.e.m., n¼3.
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.250
LETTERS
NATURE NANOTECHNOLOGY | VOL 6 | JANUARY 2011 | www.nature.com/naturenanotechnology
43
© 2011 Macmillan Publishers Limited. All rights reserved.

Page 6

resuspended in RPMI 1640. Receptor expression was assessed by flow cytometry
after labelling cells with anti-CD11b antibody (ICRF44, Abcam) and Alexa Fluor
488-conjugated goat anti-mouse secondary antibody (Invitrogen). Mouse
immunoglobulins were used as an isotype control.
For cytokine determination, cells (1× 106per ml) in the RPMI 1640 were
treated with fibrinogen/PAA–GNP complexes (33 mg ml21and 100 mg ml21,
respectively, unless otherwise specified) for 2 h at 37 8C. Cells alone and cells treated
with either fibrinogen or nanoparticles were used as controls. All experiments were
performed in polypropylene tubes to minimize surface adsorption of fibrinogen.
After incubation, supernatants were harvested by centrifugation and cytokines were
determined using a cytometric bead array kit (Becton Dickinson) according to the
manufacturer’s instruction. For NF-kB inhibition, cells were pre-treated with 30 mM
Bay 11-7082 for 20 min at 37 8C before treatments.
Nuclear extraction and electrophoretic mobility shift assay (EMSA). Extraction
of THP-1 cell nuclei and procedures for electrophoretic mobility shift assays have
been described elsewhere31. Further details are provided in the Supplementary
Information.
Statistical analysis. All experiments were performed at least three times. Data are
presented as mean+s.e.m. Statistical comparisons between different treatments were
assessed by two-tailed t tests or one-way ANOVA assuming significance at P, 0.05.
Received 23 September 2010; accepted 16 November 2010;
published online 19 December 2010
References
1. Ehrenberg, M. S., Friedman, A. E., Finkelstein, J. N., Oberdorster, G. &
McGrath, J. L. The influence of protein adsorption on nanoparticle association
with cultured endothelial cells. Biomaterials 30, 603–610 (2009).
2. Fischer, H. C. & Chan, W. C. Nanotoxicity: the growing need for in vivo study.
Curr. Opin. Biotechnol. 18, 565–571 (2007).
3. Lynch, I. & Dawson, K. A. Protein–nanoparticle interactions. Nano Today 3,
40–47 (2008).
4. Lynch, I., Salvati, A. & Dawson, K. A. Protein–nanoparticle interactions: what
does the cell see? Nature Nanotech. 4, 546–547 (2009).
5. Deng, Z. J. et al. Differential plasma protein binding to metal oxide
nanoparticles. Nanotechnology 20, 455101 (2009).
6. Dutta, D. et al. Adsorbed proteins influence the biological activityand molecular
targeting of nanomaterials. Toxicol. Sci. 100, 303–315 (2007).
7. Dobrovolskaia, M. A., Aggarwal, P., Hall, J. B. & McNeil, S. E. Preclinical studies
to understand nanoparticle interaction with the immune system and its potential
effects on nanoparticle biodistribution. Mol. Pharm. 5, 487–495 (2008).
8. Linse, S. et al. Nucleation of protein fibrillation by nanoparticles. Proc. Natl
Acad. Sci. USA 104, 8691–8696 (2007).
9. Reddy, S. T. et al. Exploiting lymphatic transport and complement activation in
nanoparticle vaccines. Nature Biotechnol. 25, 1159–1164 (2007).
10. Liang, M. et al. Cellular uptake of highly packed polymer coated gold
nanoparticles. ACS Nano 4, 403–413 (2010).
11. Hall, C. E. & Slayter, H. S. The fibrinogen molecule: its size, shape, and mode of
polymerization. J. Biophys. Biochem. Cytol. 5, 11–16 (1959).
12. Lin, Y., Wang, J., Wan, L. J. & Fang, X. H. Study of fibrinogen adsorption on
self-assembled monolayers on Au(111) by atomic force microscopy.
Ultramicroscopy 105, 129–136 (2005).
13. Lishko, V. K., Kudryk, B., Yakubenko, V. P., Yee, V. C. & Ugarova, T. P.
Regulated unmasking of the cryptic binding site for integrin aMb2in the g
C-domain of fibrinogen. Biochemistry 41, 12942–12951 (2002).
14. Ugarova, T. P. et al. Identification of a novel recognition sequence for integrin
aMb2within the g-chain of fibrinogen. J. Biol. Chem. 273, 22519–22527 (1998).
15. Lishko, V. K., et al. Multiple binding sites in fibrinogen for integrin aMb2
(Mac-1). J. Biol. Chem. 279, 44897–44906 (2004).
16. Pasparakis, M. Regulation of tissue homeostasis by NF-kB signalling:
implications for inflammatory diseases. Nat. Rev. Immunol. 9, 778–788 (2009).
17. Sitrin, R. G., Pan, P. M., Srikanth, S. & Todd, R. F. 3rd. Fibrinogen activates
NF-kB transcription factors in mononuclear phagocytes. J. Immunol. 161,
1462–1470 (1998).
18. Masamune, A. et al. Fibrinogen induces cytokine and collagen production in
pancreatic stellate cells. Gut 58, 550–559 (2009).
19. Li, Q. & Verma, I. M. NF-kB regulation in the immune system. Nat. Rev.
Immunol. 2, 725–734 (2002).
20. Teichroeb, J. H., Forrest, J. A. & Jones, L. W. Size-dependent denaturing kinetics
of bovine serum albumin adsorbed onto gold nanospheres. Eur. Phys. J. E 26,
411–415 (2008).
21. Shang, W., Nuffer, J. H., Dordick, J. S. & Siegel, R. W. Unfolding of ribonuclease
A on silica nanoparticle surfaces. Nano Lett. 7, 1991–1995 (2007).
22. Lundqvist, M., Sethson, I. & Jonsson, B. H. Protein adsorption onto silica
nanoparticles: conformational changes depend on the particles’ curvature and
the protein stability. Langmuir 20, 10639–10647 (2004).
23. Kim, B. et al. Tuning payload delivery in tumour cylindroids using gold
nanoparticles. Nature Nanotech. 5, 465–472 (2010).
24. Park, E. J. & Park, K. Oxidative stress and pro-inflammatory responses induced
by silica nanoparticles in vivo and in vitro. Toxicol. Lett. 184, 18–25 (2009).
25. Xia, T. et al. Comparison of the abilities of ambient and manufactured
nanoparticles to induce cellular toxicity according to an oxidative stress
paradigm. Nano Lett. 6, 1794–1807 (2006).
26. Yang, H., Liu, C., Yang, D., Zhang, H. & Xi, Z. Comparative studyof cytotoxicity,
oxidative stress and genotoxicity induced by four typical nanomaterials: the role
of particle size, shape and composition. J. Appl. Toxicol. 29, 69–78 (2009).
27. Diaz, B. et al. Assessing methods for blood cell cytotoxic responses to inorganic
nanoparticles and nanoparticle aggregates. Small 4, 2025–2034 (2008).
28. Lu, S. et al. Efficacy of simple short-term in vitro assays for predicting the
potential of metal oxide nanoparticles to cause pulmonary inflammation.
Environ. Health Perspect. 117, 241–247 (2009).
29. Flick, M. J. et al. Leukocyte engagement of fibrin(ogen) via the integrin receptor
aMb2/Mac-1 is critical for host inflammatory response in vivo. J. Clin. Invest.
113, 1596–1606 (2004).
30. Ryu, J. K. & McLarnon, J. G. A leaky blood–brain barrier, fibrinogen infiltration
and microglial reactivity in inflamed Alzheimer’s disease brain. J. Cell. Mol. Med.
13, 2911–2925 (2008).
31. Butcher, N. J. & Minchin, R. F. Arylamine N-acetyltransferase 1 gene regulation
by androgens requires a conserved heat shock element for heat shock factor-1.
Carcinogenesis 31, 820–826 (2010).
Acknowledgements
This work was supported by grants from the Australian Research Council (DP8787331)
and the National Health and Medical Research Council (569694).
Author contributions
Z.J.D. performed all the biological experiments, assisted in designing the biological
experiments and co-wrote the manuscript. M.L. synthesized and characterized the
nanoparticles. M.M. and I.T. designed the nanoparticle synthesis procedure. R.F.M.
conceived and designed the biological studies and co-wrote the manuscript.
Additional information
The authors declare no competing financial interests. Supplementary information
accompanies this paper at www.nature.com/naturenanotechnology. Reprints and
permission information is available online at http://npg.nature.com/reprintsandpermissions/.
Correspondence and requests for materials should be addressed to R.F.M.
LETTERS
NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2010.250
NATURE NANOTECHNOLOGY | VOL 6 | JANUARY 2011 | www.nature.com/naturenanotechnology
44
© 2011 Macmillan Publishers Limited. All rights reserved.