TISSUE ENGINEERING: Part A

Volume 18, Numbers 13 and 14, 2012

ª Mary Ann Liebert, Inc.
DOI: 10.1089/ten.tea.2011.0268

Decreased Platelet Adhesion and Enhanced

Endothelial Cell Functions on Nano
and Submicron-Rough Titanium Stents

Jing Lu, Ph.D., Chang Yao, Ph.D., Lei Yang, Ph.D., and Thomas J. Webster, Ph.D.

Endothelialization of a vascular stent is a critical step to prevent late stent thrombosis. In this study, electron
beam deposition was utilized to create different scales of roughness on titanium stents, including flat features (F-Ti),
a mixture of nanometer and submicron features (S-Ti), and nanometer features (N-Ti). The role of stent surface
roughness on initial protein adsorption, platelet adhesion, rat aortic endothelial cell adhesion, migration, and
nitric acid/endothelin-1 secretion was investigated in vitro. Results revealed the highest endothelial cell at-
tachment on S-Ti after 4 h. Moreover, under flow conditions, the endothelial cell layer remained the most intact
on S-Ti. These results were positively correlated with improved vitronectin adsorption on S-Ti. Endothelial cells
also showed the fastest migration on S-Ti of all the samples over a 96 h time period. Endothelial cells on S-Ti
exhibited the highest nitric acid/endothelin-1 ratio of all the samples, indicating potentially the best antith-
rombic endothelial cellular phenotype. This study also revealed the lowest platelet adhesion on S-Ti of all the
samples. In summary, without using pharmaceutical agents, significantly less platelet and greater endothelial
responses on nanometer to submicron rough titanium were observed in this study compared to flat titanium
and, thus, nanometer to submicron surface features on titanium should be further studied for vascular stent
applications.

Introduction cular stent nanometer and submicron surfaces compared to

T he endothelium is a single-cell layer maintaining
the health of normal arteries. Clinical observations
clearly show that the endothelium is resistant to thrombosis
formation.1 It has been observed in some cases that the stent
wall is not completely covered by endothelial cells from
months to years after stent insertion into an artery, and, thus,
acts as a reaction site for excessive inflammatory responses
and smooth muscle cell (SMC) proliferation. This delayed
endothelialization has been considered to be a major factor
for late stent thrombosis.
Therefore, the next generation of vascular stents should
enhance endothelialization upon implantation to inhibit im-
mune cell responses. Specifically, vascular stent surfaces
must be created, which diminish immune cell attachment
and increase endothelial cell recruitment, migration, and
growth. Along these lines, nanotechnology has been em-
ployed to modify stent topography, creating either periodic
or nanopatterned surfaces,2,3 nanotubular surfaces,4 or ran-
dom nanostructured surfaces.5–8 Even though it has been
previously reported that endothelial cell adhesion, prolifer-
ation, and collagen/elastin synthesis are improved on vas-
conventional micron-size surface features (i.e., rough at the
micron scale, but smooth/flat at the nanometer scale),2–8
other functions of endothelial cells remain to be tested.
To further such efforts, in the present study, the endo-
thelial cell density was characterized in both static and dy-
namic flow conditions. Moreover, attention was paid to
determine why the endothelial cell density is enhanced on
nanostructured and submicron structured stent surfaces, by
focusing on initial protein adsorption, focal adhesion, cell
migration, as well as cell nitric acid/endothelin-1 secretion
(bioactive molecules secreted by endothelial cells, which in-
dicate their thrombic or antithrombic state). In addition,
platelet adhesion on such materials was investigated for the
first time.
Thus, the aim of the present in vitro study was to further
understand the role that titanium submicron to nanos-
tructured surface features have on platelet and endothelial
cell responses. Specifically, electron beam deposition was
used in this study to create vascular stent surfaces with three
different scales of roughness: flat, nanometer to submicron,
and nanometer alone. Importantly, stent surface chemistry
was strictly controlled to isolate the influence of surface

School of Engineering, Brown University, Providence, Rhode Island.

1389
1390
roughness alone on endothelial cells. Other stent surface
properties, which may influence cell responses, including
wettability and surface energy, were fully characterized in
this study. Lastly, initial protein adsorption events were
determined in an effort to elucidate why endothelial cell
functions were different on flat versus submicron and
nanometer titanium stent surface features.
Materials and Methods
Fabrication of titanium surfaces with different
nanometer to submicron roughness
Pure titanium (99.8% of purity: T-2069; Cerac, Inc.) was
deposited on glass coverslips with 50 nm of thickness using
an e-beam evaporator (10 - 7 torr) to generate flat titanium
surfaces (labeled as F-Ti). The default e-beam deposition rate
used in this experiment was 2Å/s with an e-beam current
density of 60–70 mA/cm2. In this study, a high deposition
rate (20Å/s with an e-beam current density of 130 mA/cm2)
was used to generate nanometer surface features with a de-
position of 50-nm thickness on F-Ti (labeled as N-Ti). In
addition, to generate a mixture of nanometer and submicron
surface features, the same e-beam current density was used
with a deposition of 1 mm thickness on F-Ti (labeled as S-Ti).
The e-beam energy was set to 7 keV during the experiments.
Surface characterization
The surface morphologies of all the titanium samples were
investigated by a field emission scanning electron micro-
scope (LEO 1530 VP) at 5–7 kV. Surface roughness mea-
surements of the samples were completed using a
multimode atomic force microscope (AFM, Dimension 3100;
Veeco) to compare flat, submicron, and nanometer surface
features. A drop-shaped analysis system (Easy drop-contact
angle system; Kruss) with analysis software (DSA1) was
used to determine the surface contact angles on the titanium
samples. Distilled water was used as a contacting solvent. All
data were obtained five seconds after placing the droplet on
the surfaces under ambient conditions. For the surface en-
ergy calculations, Es = Elv · cosh was used, where Elv = 72.8
mJ/m2 at 20C for pure water and h is the static contact
angle. Here, Es is the surface energy of the contacting surface,
and Elv is the surface energy between the water and air
under ambient conditions.
Protein adsorption
Cells attach to material surfaces via preadsorbed proteins,
thus, the specific kinds of proteins, their quantity and con-
firmation once adsorbed will determine cell responses
(which can be influenced by surface roughness). In this
study, the adsorption of two proteins involved in endothelial
cell adhesion and migration (vitronectin and fibronectin)
were studied on the different titanium stent surfaces.
For fibronectin and vitronectin adsorption, S-Ti, N-Ti, and
F-Ti substrates were exposed to 500 mL of fetal bovine serum
(FBS; Hyclone) for 4 h under standard cell culture conditions
(37C, humidified, 5% CO2/95% air environment). The
samples were then rinsed with phosphate-buffered saline
(PBS) three times and blocked with 2% bovine serum albu-
min (BSA) (in PBS) for 1 h. After rinsing twice with PBS, the
samples were incubated with an antibovine fibronectin an-
LU ET AL.
tibody (6 mg/mL in 1% BSA solution; Chemicon) or an an-
tibovine vitronectin antibody (6 mg/mL in 1% BSA solution;
Accurate Chemical) for 1 h under standard cell culture con-
ditions. After that, the samples were rinsed with 0.05%
Tween 20 (Sigma) in PBS three times and were incubated
with horseradish peroxidase conjugated with an anti-rabbit
secondary antibody (1:100 BSA solution; Biorad) for 1 h. The
samples were then rinsed with 0.05% Tween 20 three times
and were transferred to new cell culture plates before being
rinsed once with 0.05% Tween 20. Finally, an ABTS (2, 2¢-
azino-bis (3-ethylbenzthiazoline-6-sulfonic acid)) soluble
substrate kit (Vector Labs) was used to detect the secondary
antibodies using a spectrophotometer (SpectoMAX; Mole-
cular Devices) at 405 nm.
Endothelial cell adhesion
Rat aortic endothelial cells (RAEC) were used in this study
(VEC Technologies) at passage numbers between 2 and 10.
RAEC were cultured in the complete medium (MCDB-131;
VEC Technologies) under standard cell culture conditions
(37C, humidified, 5% CO2/95% air environment). For the
endothelial cell adhesion test, RAEC were seeded at a den-
sity of 3500 cells/cm2 on titanium surfaces. The cells were
incubated for 4 h in the MCDB-131 Complete Medium under
standard cell culture conditions. After 4 h, nonadherent
RAECs were removed by rinsing in PBS. Cells were then
fixed with 4% formaldehyde. Once fixed, the substrates were
washed three times with PBS. The samples were permeabi-
lized in 1% Triton X-100 in PBS for 10 min and blocked with
5% FBS/0.1% Triton X-100 in PBS solution for 20 min. After
rinsing twice with PBS, the cells were stained with Rhoda-
mine Phalloidin (R415, 1:40 in 5% FBS/0.1 Triton X-100 in
PBS; Molecular Probes) for 30 min to visualize F-actin fila-
ments as well as a Hoechst dye (33258, 1:30 in PBS; Sigma)
for 10 min to visualize the cell nucleus. The RAEC number
and aspect ratio were measured under a Leica DM5500
fluorescence microscope with Image-Pro software (Media-
Cybernetics).
To study the stability of adherent RAEC onto different
titanium surfaces, RAEC were seeded (500,000 cells/cm2) on
S-Ti, N-Ti, and F-Ti in the MCDB-131 Complete Medium for
36 h. After reaching confluence, the samples were rinsed with
PBS three times and placed on the flow deck of a home-made
parallel flow chamber system. The wall shear stress applied
to the cells was set at 10 dynes/cm2. The samples were ex-
posed to laminar flow of MCDB-131 Complete Medium for
24 h. At the end of 24 h, the substrates were removed from
the flow system and rinsed with PBS three times. They were
then fixed and stained with Rhodamine Phalloidin and DAPI
as previously described. The cell morphology on the samples
after exposure to the stimulated blood vessel model was
observed with a Leica DM5500 fluorescence microscope.
Endothelial cell migration
RAEC mobility on the substrates of interests was studied
by a time-dependent RAEC migration assay. A hemicycle
polydimethylsiloxane (PDMS; Sigma) substrate (diameter:
18 mm) was attached to S-Ti, N-Ti, and F-Ti surfaces to cover
half of the surface area. RAEC were placed on the substrates
covered with PDMS in the MCDB-131 Complete Medium at
a density of 500,000 cells/cm2 for 36 h until the RAEC grew
NANO AND SUBMICRON-ROUGH TITANIUM STENTS 1391

to confluence. After the cell monolayer formed on the un- Endothelial cell nitric acid (nitric oxide)
covered region, the PDMS pieces were carefully removed, and endothelin-1 secretion
and the samples were rinsed with PBS three times. The cells
Lastly, as a measure of the thrombic state of the endothelial
were stained with 3.3 mM of calcein AM and were incubated
cells cultured on the substrates of interest, nitric oxide (NOx)
for 30 min before visualization using fluorescence micros-
and endothelin-1 assays were conducted. For this, RAEC were
copy at 530 nm. Through a fluorescence microscope, cell
seeded at 100,000 cells/cm2 in the MCDB-131 Complete
migration was determined by recording the migration dis-
Medium on S-Ti, N-Ti, F-Ti surfaces and etched glass cover-
tance into the open area created by removing the PDMS after
slips (as a reference) and were cultured for 24 h. After 24 h, the
0, 12, 24, and 96 h of culture.
Dulbecco’s Modified Eagle Medium (Invitrogen) supple-
To further study the migration of endothelial cells, the
mented with 1% penicillin/streptomycin (Hyclone) was added
focal adhesion protein vinculin was examined. For this,
to the cells for 4 h, followed by an incubation with the MCDB-
RAEC were seeded at a density of 50,000 cells/cm2 in the
131 Complete Medium for another 24 h. The culture medium
MCDB-131 Complete Medium on different titanium sub-
was collected after a select period of time and analyzed for
strates. After 24 h of culture, all the cells were fixed in 4%
NOx (Nitrate/Nitrite Colorimetric Assay Kit; Cayman Che-
formaldehyde (Fisher) for 10 min at room temperature. Once
mical Company) and endothelin-1 (Endothelin-1 Colorimetric
fixed, the substrates were washed three times with PBS and
Assay Kit; Cayman Chemical Company) following the man-
immersed with 1% Triton X-100 in PBS for 10 min to per-
ufacturer’s instructions. The final concentrations of NOx and
meabilize the cells. After being washed in PBS three times,
endothelin-1 were both normalized to the total cell numbers
the samples were incubated for 1 h at room temperature in
on the substrates, which were measured by a MTS cell pro-
1% BSA (in PBS) followed by treating the cells with an an-
liferation assay following the manufacturer’s instructions.
tivinculin antibody (1:50 in 1% BSA [in PBS], monoclonal
antivinculin FITC conjugated, and clone hVIN-1-purified
Platelet adhesion
mouse immunoglobulin; Sigma) for another 1 h. The cells
were washed three times, and their vinculin was visualized Whole blood was withdrawn from healthy domestic cross-
by a Leica DM5500 fluorescence microscope. bred swine and anticoagulated with heparin (500 units/mL).

FIG. 1. Scanning electron microscope

images of titanium surface roughness.
(a) Flat surface (F-Ti), scale bar = 2 mm;
(b) flat surface (F-Ti), scale
bar = 200 nm; (c) nanorough surfaces
(N-Ti), scale bar = 2 mm and (d)
nanorough surfaces (N-Ti), scale
bar = 200 nm; (e) submicron rough
surfaces, scale bar = 1 mm and (f)
submicron rough surfaces (S-Ti), scale
bar = 200 nm.
1392 LU ET AL.

Porcine platelets were used due to their similarity to human structures were still less than 100 nm. The AFM depth anal-
platelets. Briefly, the blood was centrifuged at 100 g for ysis of S-Ti showed an average variation in vertical struc-
15 min with no brake at room temperature to isolate platelet- tures within 30–40 nm (in the vertical scale).
rich plasma (PRP) from the red blood cells. The PRP was The surface roughness, water contact angles, and surface
added to each substrate (S-Ti, N-Ti, F-Ti, and control glass) energy values for F-Ti, N-Ti, and S-Ti are summarized in
for 30 min at 37C. Then, the samples were rinsed three times Table 1. S-Ti was the most hydrophilic of all the samples and
with PBS to remove the nonadherent blood platelets. The had the highest surface energy (28% more than F-Ti). N-Ti
samples were fixed with 4% glutaraldehyde (Invitrogen) in also had a surface energy value 15% higher than that of F-Ti.
PBS for 30 min at room temperature. After rinsing three
times with PBS, the adherent platelets were dehydrated by Protein adsorption
soaking serially in 10%, 30%, 50%, 70%, and 90% ethanol for
Figure 2a demonstrates that both S-Ti and N-Ti promoted
30 min followed by soaking in 100% ethanol for 15 min three
vitronectin adsorption after 4 h compared to F-Ti. S-Ti and
times. The samples were examined by a scanning electron
N-Ti also exhibited significantly increased fibronectin ad-
microscope (Hitachi H4700) at 3 kV to quantify platelet ad-
sorption compared to F-Ti after 4 h, about 60% more ad-
hesion on the different substrates.
sorption on N-Ti compared to F-Ti (Fig. 2b).
Statistical analysis
RAEC adhesion
All experiments were completed at least in triplicate, three
Most importantly, RAEC densities increased on submi-
times each. Data were assessed for significance using the
cron (S-Ti) and nanosurface featured (N-Ti) Ti compared to
analysis of variance followed by Student’s t-tests.
flat titanium surfaces by about 216% and 58%, respectively
(Table 2). Magnified images of cell body staining (F-actin)
Results
Surface characterization
In this study, smooth cover glass provided for an under-
lying substrate with minimal (close to zero) roughness.
When titanium thin films were e-beam deposited onto cover
glass using slow deposition rates (2Å/s), the structure of the
deposited titanium was flat in both the micron and nano-
meter scales (F-Ti, Fig. 1a, b). The AFM sectional and depth
analysis indicated that the surface had negligible variations
in both the vertical and lateral scales. However, high e-beam
current densities (130 mA/cm2) and, thus, fast deposition
rates (20 Å/s) with a deposition thickness of 50 nm, gener-
ated some surface roughness at the submicron scale (Fig. 1c)
and sectional analyses showed that nanostructures were
clearly visible at the nanoscale (N-Ti: structures were 30–
40 nm in the lateral and 2–6 nm in vertical scales). Since both
vertical and lateral structures were less than 100 nm, all the
surface roughness that came from these features should be
considered as nanometer surface roughness. Figure 1e and f
demonstrate a drastic difference in surface structures at the
submicron and nanometer scale. In contrast to N-Ti with
increased surface roughness at the nanoscale than at the
submicron scale (Fig. 1c, d), high current density (130 mA/cm2)
with an increased deposition rate (20 Å/s) and e-beam
deposition thickness of 1 mm generated the roughest struc-
tures at the submicron scale (S-Ti: Fig. 1e). It was clear that
submicron scale surface roughness (larger than 100 nm in
lateral width) was dominant on S-Ti surfaces, but vertical

Table 1. Properties of the Various Titanium

Substrates Created by e-Beam Deposition

F-Ti N-Ti S-Ti

Roughness 0.38 – 0.08 1.15 – 0.11 13.77 – 0.83

(RMS, nm)
Contact angle () 41.7 – 0.11 31.4 – 0.64 18.0 – 2.14
Surface energy 54.75 – 0.16 62.16 – 0.83 69.24 – 1.79 FIG. 2. (a) Vitronectin and (b) fibronectin adsorption on S-
(mJ/cm2) Ti, N-Ti, and F-Ti substrates after 4 h of exposure.
Data = mean – standard error of the mean (SEM), n = 3.
RMS, root mean square. *p < 0.05, **p < 0.01.
NANO AND SUBMICRON-ROUGH TITANIUM STENTS 1393

Table 2. Rat Aortic Endothelial Cell Adhesion and RAEC mobility

Aspect Ratios on the Various Titanium Substrates
F-Ti N-Ti S-Ti
Adhering density 1269 – 268 1909 – 184 2681 – 328
(cells/cm2)
Aspect ratio 1.46 – 0.12 2.13 – 0.21 4.21 – 0.81
The aspect ratio was calculated by the length of an individual cell
divided by the width.
Data = mean – standard error of the mean; n = 3.
clearly showed different aspect ratios of RAEC on S-Ti and
N-Ti compared to F-Ti. The quantitative analysis showed
that the aspect ratios of RAEC on S-Ti improved by 100%
compared to those on N-Ti and F-Ti.
Figure 3 shows that S-Ti retained the endothelial cell layer
the best under flow conditions. In comparison, N-Ti gener-
ally kept the endothelial cell layer intact, but some of the cell
to cell contacts were disrupted after 24 h of laminar flow. F-Ti
was the worst and showed the least endothelial cells re-
maining on the surface after flow. The quantitative analysis
of the RAEC layer coverage under flow after 24 h showed
that more than 90% of the S-Ti area and more than 80% of the
N-Ti area were still covered by RAEC. In contrast, only 40%
of the F-Ti area was still covered by RAEC under the same
conditions.
Results further showed faster endothelial cell migration
on S-Ti (about 500 mm after 96 h) and on N-Ti (about 400 mm
after 96 h) compared to F-Ti (about 300 mm after 96 h; Figs. 4
and 5). Moreover, endothelial cells had more directionality
in their migration on S-Ti than on the N-Ti and F-Ti. A
much more pronounced protrusion of filopodia with a
significantly longer configuration and a higher degree of
contact was observed on the S-Ti surface than on the F-Ti
surface (Fig. 6).
Vinculin, a focal adhesion protein, plays an important
role in cell surface integrin receptors binding to extracellu-
lar matrix proteins and in directing cell migration. Fluor-
escence images of vinculin (Fig. 7) showed that vinculin
formed small dots when endothelial cells were cultured on
S-Ti (Fig. 7a). On the other hand, on F-Ti (Fig. 7c), a wider,
dash distribution of vinculin was observed. The morphol-
ogy of vinculin focal contacts on N-Ti (Fig. 7b) was in be-
tween that on S-Ti and F-Ti, where in some places they were
well distributed, while in the other places, small dots were
found.
RAEC NOx and endothelin-1 secretion
In the healthy endothelium, NOx is continuously secreted
by endothelial cells to maintain vascular homeostasis. NOx is
a vasodilator, and it can also inhibit platelet aggregation.

FIG. 3. Double-stained
endothelial cells remaining
on (a) submicron rough (S-
Ti), (b) nanorough (N-Ti), (c)
flat (F-Ti) titanium surfaces,
and (d) glass after 24 h of
laminar flow. Shear stress
was maintained at 10 dynes/
cm2. All substrates were
precultured with endothelial
cells for 36 h to form a
complete layer of cells before
exposure to flow. Cells were
stained with Rhodamine
Phalloidin to visualize F-actin
filaments as well as a Hoechst
dye to visualize nuclei. Color
images available online at
www.liebertpub.com/tea
1394 LU ET AL.

FIG. 4. Rat aortic endothelial cell (RAEC) migration on S-Ti, N-Ti, and F-Ti substrates over 12, 24, and 96 h. Cells were
stained with Calcein AM. Scale bars = 200 mm. Dash lines indicate starting points at time zero. Color images available online at
www.liebertpub.com/tea

Therefore, NOx synthesis by endothelial cells is important Platelet adhesion

for the proper functioning of a vascular stent. The amount of
NOx synthesis by endothelial cells cultured on the different
titanium surfaces was investigated here, and results are re-
ported in Figure 8a. S-Ti surfaces caused the highest level of
NOx synthesis by RAEC, which increased by 70% compared
to that on F-Ti. There was no significant difference in NOx
secretion by RAEC on nanorough (N-Ti) and flat (F-Ti) tita-
nium surfaces. Endothelin-1, which is also synthesized and
secreted by endothelial cells as a vasoconstrictor and pro-
moter of platelet aggregation was also measured here. The
amounts of endothelin-1 secreted by endothelial cells on the
different titanium surfaces were investigated, and the results
are reported in Figure 8b. There was no significant difference
of endothelin-1 levels by endothelial cells cultured on any of
the titanium surfaces of interest here.
Figure 9 shows that there were about 57% and 44% less
platelets adherent to submicron (S-Ti)- and nanosurface-
featured (N-Ti) Ti than flat titanium surfaces (F-Ti), respec-
tively.
Discussion
Endothelial cell adhesion
The results from the contact angle analysis indicated that
the small nanosized surface features on N-Ti (root mean
square [RMS] values less than 2 nm, 40-nm lateral width, and
4-nm vertical depth achieved on 50-nm titanium thin films)
altered water contact angles by more than 10 compared to F-
Ti surfaces. This result is very intriguing, since small changes
NANO AND SUBMICRON-ROUGH TITANIUM STENTS
FIG. 5. RAEC migration on submicron (S-Ti), nano (N-Ti),
and flat (F-Ti) titanium surfaces after 12, 24, and 96 h,
showing that RAEC motility greatly increased on S-Ti com-
pared with that on N-Ti and F-Ti. All error bars are mean –
SEM; n = 3; *p < 0.05 (compared to F-Ti after 12 h), **p < 0.05
(compared to N-Ti and F-Ti after 24 and 96 h), and #p < 0.05
(compared to F-Ti after 12, 24, and 96 h).
in the surface area (a 0.7% increment between N-Ti and F-Ti)
led to a 15% increase in surface energy between N-Ti and F-Ti.
As for S-Ti, there was a 6% increase in the surface area, and
about a 30% increase in surface energy over F-Ti. AFM RMS
values of S-Ti features drastically increased (more than 10 nm)
due to an increased vertical depth compared to N-Ti.
RAEC adhesion results after 4 h showed that RAEC den-
sity increased more than 50% on N-Ti compared to F-Ti
surfaces and increased more than 100% on S-Ti compared to
F-Ti. Since in this study, the altered lateral width for N-Ti
was almost ten times higher than the altered vertical struc-
tures, an increase in lateral width was considered to be a
meaningful factor altering cell adhesion. This might be due
to altered titanium nanometer topography (increased lateral
width of surface features) inducing more cell spreading
(greater wettability or surface energy) with enhanced cell
adhesion than F-Ti surfaces of identical chemistry.9–12 It was
noticed that altered vertical structures (about 4 nm) did not
lead to a drastic increase in surface energy and cell adhesion
as observed by previous studies.13,14 In comparison, dra-
matic changes in lateral ( > 100 nm) and vertical (30–40 nm)
structures on S-Ti were considered to collectively contribute
to increased endothelial cell adhesion.
The increased surface area and surface energy of N-Ti and
S-Ti are hypothesized to promote critical protein adsorption
FIG. 6. Scanning electron microscope micrographs of RAEC filopodia extensions on (a) S-Ti, (b) N-Ti, and (c) F-Ti after 24 h
of culture.
1395
and subsequently influence endothelial cell adhesion, which
was confirmed in the present study. Results showed that
both S-Ti and N-Ti promoted vitronectin and fibronectin
adsorption from FBS than F-Ti. When comparing S-Ti and
N-Ti, vitronectin adsorption on S-Ti was significantly higher
than on N-Ti, while fibronectin adsorption on N-Ti was higher
than on S-Ti. It is known that FBS contains similar quantities
of fibronectin and vitronectin, but overall the adsorption of
vitronectin on all the three substrates (S-Ti, N-Ti, and F-Ti)
was much larger than that of fibronectin. This may be because
vitronectin successfully adsorbs on substrates competitively in
the presence of other proteins in serum, while fibronectin
adsorption is inhibited due to the adsorption of other pro-
teins.15 Since it has been shown that vitronectin is a critical
protein in serum that mediates endothelial cell adhesion onto
various substrates,15–17 it is speculated that this is one of the
reasons why the highest endothelial cell adhesion was ob-
served on S-Ti compared to the other substrates.
The stability of endothelial cell monolayers on various ti-
tanium roughness surfaces under flow conditions also provi-
des valuable information to estimate stent efficacy in vivo. The
observed higher intactness of endothelial cell layers on S-Ti
and N-Ti surfaces than on F-Ti indicated stronger cell-sub-
strate adhesion and cell–cell interactions when endothelial cells
interacted with nanometer surface features even under flow.
This is supported by prior findings that endothelial cells pre-
ferred to attach to nanopitted PMMA surface features (100 nm)
created by e-beam lithography after 1-h exposure to flow.18
In summary, stronger endothelial cell adhesion under
static and flow conditions was found on S-Ti than on
smoother surfaces. However, it was also reported by our
group that SMC adhesion increased on S-Ti.5,8 Since exces-
sive smooth muscle adhesion and migration cause restenosis
or intimal hyperplasia, it is a concern whether endothelial
cells or SMC may be dominant in a competitive model. To
answer this question, our previous study used a coculture
model of RAEC and rat aortic smooth muscle cells (RASMC)
up to 5 days and revealed enhanced RAEC proliferation, but
simultaneously inhibited RASMC proliferation on S-Ti.8 The
mechanism of selective growth of endothelial cells over
SMCs on submicron titanium surfaces is unclear and needs
further investigation.
Endothelial cell mobility
Endothelial cell migration distance significantly increased
by 46% on S-Ti and 22% on N-Ti when compared to F-Ti
after 24 h. After 96 h, as the migration speed of RAEC on S-Ti
1396
FIG. 7. Immunofluorescent images of vinculin, which indicated focal adhesion of RAEC, on titanium surfaces of different
roughness after 24 h of incubation: (a) submicron rough titanium (S-Ti), (b) nanorough titanium (N-Ti), and (c) flat titanium
(F-Ti). All scale bars = 100 mm. Color images available online at www.liebertpub.com/tea
and N-Ti increased (as determined by the slope of the curves
between 24 and 96 h), incremental migration distances in-
creased by 56% on S-Ti and 25% on N-Ti compared to F-Ti.
Clearly, RAEC migrated the most quickly on S-Ti of all the
substrates studied here.
FIG. 8. (a) Nitric oxide (NOx) and (b) endothelin-1 syn-
thesized by RAEC on submicron rough (S-Ti), nanorough
(N-Ti) and flat (F-Ti) titanium substrates after 24 h of incu-
bation. Data = mean – SEM, n = 3. *p < 0.05.
LU ET AL.
It is interesting to compare these results with those of
RAEC adhesion. DiMilla et al. proposed that to trigger cell
locomotion on substrates, the cell-substratum adhesion
strength must be strongly considered.19 Previous studies
indicated that from both mathematical modeling and ex-
periments, cell migration rates have a biphasic dependence
on the concentration of adhesive ligands and cell adhesion
strength. In Phase I, cell migration speed increases as the
concentration of extracellular matrix proteins increases (such
as vitronectin).19,20 In this study, the effective adsorption of
vitronectin and the adhesion of RAEC were both enhanced
on S-Ti compared with the other substrates, meanwhile
RAEC migration was also the highest on S-Ti. These results
together clearly suggest a Phase I catogery behavior, that is,
S-Ti surfaces adsorbed a favorable amount of adhesive pro-
teins and in turn promoted both endothelial cell adhesion
and migration.
The nanometer to submicron features on S-Ti shown in
Figure 6a likely provided topographical cues that changed
surface wettability to promote vitronectin adsorption for
endothelial cell attachment and migration. To further inves-
tigate the migration of RAEC on different substrates, the
focal adhesion protein vinculin was analyzed in this study by
fluorescence microscopy. The protrusion of cell lamellipodia
is controlled by the formation of focal adhesions sites on the
leading edge of the cell and their disassembly on the rear
edge, thus, focal adhesion is one of the most important
FIG. 9. Platelet adhesion onto the different substrates (S-Ti,
N-Ti, and F-Ti) after 30 min of incubation with platelet-rich
plasma. Data = mean – SEM, n = 3. *p < 0.01. Color images
available online at www.liebertpub.com/tea
NANO AND SUBMICRON-ROUGH TITANIUM STENTS
regulators of cell migation. As demonstrated in Figure 7, the
focal adhesion complexes of RAEC on S-Ti, which formed
small dots, were different in size from the large dash adhe-
sions observed on F-Ti. This is consistent with previous
studies showing smaller and less pronounced fibroblast focal
adhesions on nanoislands than on flat surfaces.21 In another
study, endotheial cells were found to form small dot focal
adhesions on titanium nanotubular structures compared to
large dash adhesions on flat control surfaces.22 In addtion,
focal adhesion distribution on S-Ti were more polarized,
primarily on the leading filopodia for migration, than on
flat titanium as shown in Figure 7. In a word, the observa-
tions of both endothelial cell small focal adhesions and im-
proved endothelial cell migration behavior in the present
study and in the reference22 indicate that endothelial cells
are more motile for spreading when small focal adhesions
are preferentially formed on the edges of the protruded
filopodia.
Endothelial cell NOx secretion
In this study, the secretion of NOx by RAEC on S-Ti after
24 h was higher than the other substrates (N-Ti and F-Ti). On
the other hand, the release of endothelin-1 remained similar
on all the substrates. Thus, the ratio of NOx/endothelin-1
synthesis was greatly enhanced on S-Ti compared with N-Ti
and F-Ti. The increased release of NOx from endothelial cells
on S-Ti could enhance the vasodilation bioactivity, keeping
the endothelium in a healthy state. In addition, the increased
balance between NOx/endothelin-1 might create a potential
antithrombotic environment on S-Ti.
It is well known that external stimuli like shear stress can
regulate NO production from endothelial nitric oxide syn-
thase (eNOS),23 and vascular endothelial growth factor
(VEGF) is considered a pivotal factor for eNOS expres-
sion.24,25 To this light, a detailed study in the intracellular
signal pathway of eNOS induced by VEGF would be helpful
to clarify the role of titanium surface topography/energy on
NO production by endothelial cells in the future.
Platelet adhesion
Platelet attachment onto stent surfaces is associated with
thrombosis and subsequent restenosis in the case of incom-
plete endothelialization of a stent. Porcine platelets were
used in the present study due to their similar response to
human platelets.26 Platelet adhesion results from this study
showed that nano and submicron rough titanium surfaces
induced less attachment than flat surfaces, possibly indicat-
ing better blood compatibility. This difference may be related
to the different degree of hydrophilicity of each surface,
which has also been revealed by other researchers. For ex-
ample, Rodrigues et al. found the highest number of adherent
and activated platelets on the most hydrophobic self-
assembled monolayer terminated with methyl groups.27 In
this light, the submicron rough titanium surfaces with low
platelet adhesion and high endothelial cell growth are sig-
nificant for stent applications, especially when considering
that no pharmaceutical agents were used in the present
study to make this happen. Of course, to further prove the
antithrombotic properties of such titanium surfaces, more in-
depth studies on platelet responses are needed in a separate
study.
1397
Conclusions
In summary, RAECs interacted more efficiently with
submicron and nanometer rough titanium stent surfaces
than flat surfaces. These interactions included enhanced en-
dothelial cell adhesion, migration, and antithrombogenic
behavior. The nanometer lateral features on N-Ti and the
submicron lateral and nanometer vertical features on S-Ti led
to an increase in the surface area and surface energy com-
pared to F-Ti, which in turn influenced critical initial protein
adsorption and cell adhesion (under static and flow condi-
tions). Small dotted vinculin on endothelial cells on S-Ti
showed that endothelial cells might be able to probe nano-
meter and submicron titanium features and that the cell la-
mellipodia more readily extended to promote cellular
migration. Last, S-Ti significantly increased endothelial cell
NOx/endothelin-1 ratios, which may be beneficial for an-
tithrombic purposes. Platelet adhesion was also reduced on
S-Ti. All these findings indicated that nanometer to submi-
cron rough titanium may be more favorable for vascular
stent applications, and, thus should be further studied.
Acknowledgment
The authors would like to thank the Hermann Foundation
for funding.
Disclosure Statement
No competing financial interests exist.
References
1. White, R. Atherosclerosis and Arteriosclerosis: Human Pa-
thology and Experimental Animal Methods and Models.
Boca Raton, FL: CRC Press, 1989.
2. Lu, J., Rao, M.P., Macdonald, N.C., Khang, D., and Webster,
T.J. Improved endothelial cell adhesion and proliferation on
patterned titanium surfaces with rationally designed, mi-
crometer to nanometer features. Acta Biomater 4, 192, 2008.
3. Miller, D.C., Thapa, A., Haberstroh, K.M., and Webster, T.J.
Endothelial and vascular smooth muscle cell function on
poly(lactic-co-glycolic acid) with nano-structured surface
features. Biomaterials 25, 53, 2004.
4. Peng, L., Eltgroth, M.L., Latempa, T.J., Grimes, C.A., and
Desai, T.A. The effect of TiO2 nanotubes on endothelial
function and smooth muscle proliferation. Biomaterials 30,
1268, 2009.
5. Choudhary, S., Berhe, M., Haberstroh, K.M., and Webster,
T.J. Increased endothelial and vascular smooth muscle cell
adhesion on nanostructured titanium and CoCrMo. Int J
Nanomed 1, 41, 2006.
6. Choudhary, S., Haberstroh, K.M., and Webster, T.J. En-
hanced functions of vascular cells on nanostructured Ti for
improved stent applications. Tissue Eng 13, 1421, 2007.
7. Samaroo, H.D., Lu, J., and Webster, T.J. Enhanced endo-
thelial cell density on NiTi surfaces with sub-micron to
nanometer roughness. Int J Nanomed 3, 75, 2008.
8. Lu, J., Khang, D.W., and Wesbter, T.J. Greater endothelial
cell responses on submicron and nanometer rough titanium
surfaces. J Biomed Mater Res A 94A, 1042, 2010.
9. Chow, T. Wetting of rough surfaces. J Phys Condensed
Matter 10, L445, 1998.
10. McHale, G., Shirtcliffe, N., Aqil, S., Perry, C., and Newton,
M. Topography driven spreading. Phys Rev Lett 93, 36102,
2004.
1398
11. Liu, X., Lim, J.Y., Donahue, H.J., Dhurjati, R., Mastro, A.M.,
and Vogler, E.A. Influence of substratum surface chemistry/
energy and topography on the human fetal osteoblastic cell
line hFOB 1.19: phenotypic and genotypic responses ob-
served in vitro. Biomaterials 28, 4535, 2007.
12. Pesakova, V., Kubies, D., Hulejova, H., and Himmlova, L.
The influence of implant surface properties on cell adhesion
and proliferation. J Mater Sci Materi Med 18, 465, 2007.
13. Kunzler, T., Drobek, T., Schuler, M., and Spencer, N. Sys-
tematic study of osteoblast and fibroblast response to
roughness by means of surface-morphology gradients. Bio-
materials 28, 2175, 2007.
14. Washburn, N., Yamada, K., Simon, C., Kennedy, S., and
Amis, E. High-throughput investigation of osteoblast re-
sponse to polymer crystallinity: influence of nanometer-scale
roughness on proliferation. Biomaterials 25, 1215, 2004.
15. Underwood, P., and Bennett, F. A comparison of the bio-
logical activities of the cell-adhesive proteins vitronectin and
fibronectin. J Cell Sci 93, 641, 1989.
16. Hayman, E., Pierschbacher, M., Suzuki, S., and Ruoslahti, E.
Vitronectin—a major cell attachment-promoting protein in
fetal bovine serum. Exp Cell Res 160, 245, 1985.
17. Bale, M., Wohlfahrt, L., Mosher, D., Tomasini, B., and Sut-
ton, R. Identification of vitronectin as a major plasma protein
adsorbed on polymer surfaces of different copolymer com-
position. Blood 74, 2698, 1989.
18. Martines, E., McGhee, K., Wilkinson, C., and Curtis, A. A
parallel-plate flow chamber to study initial cell adhesion on
a nanofeatured surface. IEEE Transact Nanobiosc 3, 90, 2004.
19. DiMilla, P., Stone, J., Quinn, J., Albelda, S., and Lauffen-
burger, D. Maximal migration of human smooth muscle
cells on fibronectin and type IV collagen occurs at an inter-
mediate attachment strength. J Cell Biol 122, 729, 1993.
20. DiMilla, P., Barbee, K., and Lauffenburger, D. Mathematical
model for the effects of adhesion and mechanics on cell
migration speed. Biophys J 60, 15, 1991.
LU ET AL.
21. Dalby, M., Riehle, M., Johnstone, H., Affrossman, S., and
Curtis, A. In vitro reaction of endothelial cells to polymer
demixed nanotopography. Biomaterials 23, 2945, 2002.
22. Brammer, K., Oh, S., Gallagher, J., and Jin, S. Enhanced
cellular mobility guided by TiO2 nanotube surfaces. Nano
Lett 8, 786, 2008.
23. Boo, Y.C., and Jo, H. Flow-dependent regulation of endo-
thelial nitric oxide synthase: role of protein kinases. Am J
Physiol Cell Physiol 285, 499, 2003.
24. Feliers, D., Chen, X., Akis, N., Choudhury, G.G., Madaio,
M., and Kasinath, B.S. VEGF regulation of endothelial nitric
oxide synthase in glomerular endothelial cells. Kidney Int
68, 1648, 2005.
25. Park, J.S., Hong, G.R., Baek, S.W., Shin, D.G., Kim, Y.J., and
Shim, B.S. Expression and regulation of endothelial nitric
oxide synthase by vascular endothelial growth factor in ECV
304 cells. J Korean Med Sci 17, 161, 2002.
26. Goodman, S.L. Sheep, pig, and human platelet-material in-
teractions with model cardiovascular biomaterials. J Biomed
Mater Res 45, 240, 1999.
27. Rodrigues, S.N., Goncalves, I.C., Martins, M.C.L., Barbosa,
M.A., and Ratner, B.D. Fibronogen adsorption, platelet
adhesion and activation on mixed hydroxyl-/methyl-
terminated self-assembled monolayers. Biomaterials 27,
5357, 2006.
Address correspondence to:
Thomas J. Webster, Ph.D.
School of Engineering
Brown University
Providence, RI 02912
E-mail: thomas_webster@brown.edu
Received: May 10, 2011
Accepted: March 9, 2012
Online Publication Date: June 29, 2012