Endothelial nanoparticle binding kinetics are matrix and size dependent

This is the pre-peer reviewed version of the following article: Endothelial nanoparticle binding

kinetics are matrix and size dependent, Amber L. Doiron, Brendan Clark, Kristina D. Rinker;

Biotechnology and Bioengineering, 108(12), pages 2988–2998, December 2011, which has been

published in final form at http://onlinelibrary.wiley.com/doi/10.1002/bit.23253/abstract

Accepted 20 June 2011

Endothelial nanoparticle binding kinetics are matrix and size dependent

Amber L. Doiron1,2,4,6,7, Brendan Clark1, and Kristina D. Rinker1,2,3,5

1Cellular and Molecular Bioengineering Research Laboratory, University of Calgary, Calgary,

Canada

2Department of Chemical and Petroleum Engineering, University of Calgary, Calgary, Canada

3Department of Physiology and Pharmacology, University of Calgary, Calgary, Canada

4Department of Radiology, University of Calgary, Calgary, Canada

5Centre for Bioengineering Research and Education, University of Calgary, Calgary, Canada,

6Hotchkiss Brain Institute, University of Calgary, Calgary, Canada

7Seaman Family MR Research Centre, Foothills Medical Centre, Calgary, Canada

Corresponding Author: Kristina D. Rinker, University of Calgary, 2500 University Drive NW,

Calgary, Department of Chemical and Petroleum Engineering, AB T2N 1N4, 403-210-9733,

[email protected]

Running Title: Particle Binding Kinetics Dependence on Size

1

ABSTRACT

Nanoparticles are increasingly important in medical research for application to areas such as

drug delivery and imaging. Understanding the interactions of nanoparticles with cells in

physiologically relevant environments is vital for their acceptance, and cell/particle interactions

likely vary based on the design of the particle including its size, shape, and surface chemistry.

For this reason, the kinetic interactions of fluorescent nanoparticles of sizes 20, 100, 200, and

500 nm with human umbilical vein endothelial cells (HUVEC) were determined by 1) measuring

nanoparticles per cell at 37oC and 4oC (to inhibit endocytosis) and 2) modeling experimental

particle uptake data with equations describing particle attachment, detachment, and

internalization. Additionally, the influence of cell substrate compliance on nanoparticle

attachment and uptake was investigated. Results show that the number of binding sites per cell

decreased with increasing nanoparticle size, while the attachment coefficient increased. By

comparing HUVEC grown on either a thin coating of collagen or on top of three-dimensional

collagen hydrogel, nanoparticle attachment and internalization were shown to be influenced

significantly by the substrate on which the cells are cultured. This study concludes that both

particle size and cell culture substrate compliance appreciably influence the binding of

nanoparticles; important factors in translating in vitro studies of nanoparticle interactions to in

vivo studies focused on therapeutic or diagnostic applications.

Keywords: polystyrene nanoparticle uptake, nanotoxicity, 3D cell culture, compliance, HUVEC,

collagen, dextran

2

Introduction

Nanotechnology has rapidly progressed into an exciting field, and the push towards the nano-

scale has led to many medical advances. The use of nanoparticles to deliver an encapsulated or

attached agent to a specific site has promise to enhance efficacy, reduce toxicity, improve

imaging, and tune the biodistribution of a given therapeutic or imaging agent (Singh and Lillard,

2009). Nanoparticles of various shapes, sizes, and surface chemistries have been investigated for

therapy and imaging applications. While these properties are highly advantageous in tuning

nanoparticles for a particular application (Caldera-Moore et al., 2010), they also contribute to the

complexity of understanding how nanoparticles interact with their target cells.

Currently, it is customary for researchers to evaluate nanoparticles for toxicity, targetability,

and efficacy using in vitro cell cultures before proceeding to animal studies to evaluate the

nanoparticle in vivo. The two-dimensional (2D) nature of most adherent culture systems,

whereby cells are grown either on a bare or protein-coated tissue culture plate or glass substrate,

fails to translate into in vivo relevance, possibly because the 2D environment is a poor mimic of

the biological complexity of the three dimensional (3D) in vivo environment including the

extracellular matrix, matrix-cell interactions, transport limitations, and external mechanical cues

(Ng and Pun, 2008; Griffith and Swartz, 2006; Goldman et al., 2005). Two-dimensionally

cultured cells and in vivo cells differ in many aspects including gene expression (Griffith and

Swartz, 2006), biosynthesis of extracellular molecules (Albrecht et al., 2006) and cytoskeletal

components (Cukierman et al., 2001), and biological activity such as proliferation (Bhatia et al.,

1998; Heldin et al., 2004). For these reasons, the biological interactions of nanoparticles may be

more appropriately studied in vitro using a 3D culture environment. In both a 3D and 2D

environment in vitro, substrate compliance has been shown to greatly affect the compliance of

3

the endothelial cells themselves cultured atop the substrate resulting from a change in actin stress

fibers (Byfield et al., 2009). To the authors’ knowledge, no study to date explores the influence

of cell substrate compliance on the binding of nanoparticles, a potentially important factor in

translating in vitro studies of nanoparticle binding behavior to in vivo relevance. In this study, we

use cells cultured on a thin coating of collagen or on the surface of a collagen hydrogel of non-

negligible height to differentiate between the 2D and 3D extracellular environment of different

compliance in vitro.

Many nanotherapies rely on vascular targeting or travel through the vasculature to reach an

intended target. Nanoparticles of several sizes have been shown to deposit preferentially in the

endothelium of vessels and largely avoid penetration into the media and adventitia (Chesler and

Enyinna, 2003). While the main strategy for delivery of nanoparticles to tissues up to now has

been the enhanced permeability and retention (EPR) effect, recent research in the area of

vascular biodiversity (Sergeeva et al., 2006) suggests wide utility for nanoparticle targeting to

the endothelium. Human endothelial cells are studied here in vitro as a model of the vessel wall

that nanomaterials encounter upon intravenous injection.

This study explores the kinetics of binding of polystyrene nanoparticles with diameters from

20 nm to 500 nm to elucidate how particles of varying size bind to the cell surface and are

internalized. Sedimentation varies over the range of nanoparticle diameters studied and is

accounted for in a system of equations based on a single cell model of attachment (Goodman et

al., 2008; Wilhelm et al., 2002). These equations are used in conjunction with data collected

using human umbilical vein endothelial cells (HUVEC) cultured on top of 2D or 3D collagen

exposed to polystyrene nanoparticles to obtain the kinetic coefficients.

Materials and Methods

4

Cell Culture

HUVEC were cultured according to the supplier’s directions with modifications described here.

HUVEC (Lonza, Walkersville, MD, USA) were seeded from cryopreservation at passage 4,

cultured on 0.1% gelatin (BD Difco, Mississauga ON, Canada) coated tissue culture flasks in

endothelial growth medium (EGM, Lonza) containing 2% serum. Well plates were prepared

prior to introduction of cells by coating with collagen or preparation of 3D collagen gels.

HUVEC were grown to confluence on 48 well plates before addition of nanoparticles. Studies

involving dextran were conducted using EGM adjusted to a viscosity of 3.0 cP by including

dextran (Spectrum Chemical Mfg, Gardena, CA, USA) at a concentration of 3.02 g per 100 mL

media. Culture medium was replaced with dextran-containing EGM just prior to introduction of

particles.

Preparation of Collagen Substrates

Thin collagen coatings (2D) were prepared by mixing rat tail type I collagen (5 mg/mL stock,

Gibco, Invitrogen, Grand Island, NY, USA) with room temperature acetic acid (0.02M, Fisher

Scientific, Ottawa ON, Canada) at 172.5 µg/mL final collagen concentration, and 150 µL of this

solution was evenly pipetted into each well. After three hours incubation at room temperature,

the solution was aspirated, and wells were gently rinsed with DPBS (Sigma-Aldrich, Oakville

ON, Canada) and placed under ultraviolet (UV) light for 15 minutes prior to seeding with

HUVEC at 7,500 cells/cm2.

Alternatively, to form three-dimensional 3.5 mg/mL collagen hydrogels in the wells of a 48

well plate, the following cold reagents were mixed: 1.36 mL deionized water, 130 µL 1M NaOH

(Fisher), 0.75 mL 10x DPBS (Sigma), and 5.25 mL collagen; 150 µL of the solution was added

to each well. Plates were incubated overnight at 37˚C before gels were washed three times (15

5

minutes each) with DPBS. Well plates were placed under a UV light for 15 minutes prior to

seeding with HUVEC at 10,000 cells/ cm2, a higher cell density than 2D collagen cultures due to

lower initial cell adhesion to the substrate. HUVEC were seeded on top of the pre-formed gels

and did not readily migrate into the gel during the time periods relevant to this study. Gels were

visually inspected to verify intactness during washes and media changes.

Nanoparticle Uptake Studies

Fluosphere® particles (Molecular Probes, Invitrogen, Eugene, OR, USA) were supplied as 2%

w/v suspensions of polystyrene spheres of diameters 0.02, 0.100, 0.200, and 0.500 µm loaded

with red fluorophore and with surface carboxylic acid moieties. Particles were used after

thorough mixing via sonication and vortexing. Standard curves were prepared for each

nanoparticle size sample to equate signal to concentration. All samples were run in triplicate. In

both 2D and 3D culture systems, care was taken to ensure the confluence of cells prior to

exposure to nanoparticles in order to avoid nanoparticles sticking to the cell substrate as opposed

to interacting with cells.

To determine concentration dependence, nanoparticles were added to cells at concentrations

of 1 µL of stock solution per mL EGM to 320 µL/mL and incubated for one hour at 4˚C to

inhibit endocytosis. Time dependence was studied at a nanoparticle concentration of 10 µL/mL

for 15 minutes to 48 hours at 37˚C. After addition of particles, plates were agitated on a shaker

(Lab-Line Instruments Inc., Melrose Park, IL, USA) for 30 seconds. After incubation, the media

was aspirated and wells were gently washed three times with DPBS. DPBS (0.5 mL) was added

to each well prior to reading on a plate reader (SpectraMax M2, Molecular Devices, Sunnyvale,

CA, USA) with an excitation wavelength of 580 nm and emission of 605 nm.

Confocal Microscopy

6

Cell samples were seeded on glass slides coated with 2D or 3D collagen. For collagen coating, a

central area on the slide was masked out using 0.0508 cm thick silicon gasket (Specialty

Manufacturing, Saginaw, MI, USA) with a 1.5 cm by 5 cm cutout, and 400 µl collagen solution

(2D or 3D) was added to this cutout area. Slides were incubated, washed, and seeded with cells

as described in previous section. Confluent cells were incubated with nanoparticles at 10 µl/mL

at 37˚C for 1 hour. Slides were washed three times with DPBS and incubated for 15 minutes at

room temperature with 1 µl/mL Hoechst 33258 (Invitrogen) and CellMask Deep Red

(Invitrogen) in DPBS. Slides were washed with DPBS, fixed for 10 minutes at 37˚C in 4%

paraformaldehyde (Sigma), and washed three times with DPBS prior to mounting with

VectaShield (Vector Laboratories, Burlingame, CA, USA) and coverslip. Imaging of slides was

carried out on an Olympus FluoView™ FV1000 confocal laser scanning microscope and

accompanying software, version 2.1b (Markham, Ontario, Canada) using a 60x objective.

Theoretical Model of Nanoparticle Binding Kinetics and Internalization

Particle interactions with cells are characterized by particle movement to the cell surface, particle

attachment and detachment to cells, and the internalization of particles, assuming homogeneity

of the system, as modeled similarly to previous studies (Goodman et al. 2008).

Sedimentation from the Bulk Fluid

Nanoparticles were assumed to travel to the cell surface via sedimentation, resulting in an

increase in concentration at the cell wall compared to the bulk concentration for all particle sizes

over time. Sedimentation was incorporated mathematically using an enhancement factor for the

concentration at the cell surface, Csurface:

bulksurface Ch

sttC

1

, (1)

7

where t is time, h is the height of media (1 cm) over which gravity is acting, Cbulk is the starting

bulk concentration at time zero, and s is the Einstein-Stokes sedimentation rate defined as:

9

2 2lpgr

s

, (2)

where r is the particle radius, g is the force of gravity, µ is the viscosity of the liquid with cell

media approximated here as water (µ = 0.79 cP at 37˚C and µ = 1.5 cP at 4̊ C) , ρp is the density

of the polystyrene particle (1.05 g/cm3), and ρl is the density of the liquid (water, ρl = 0.992

g/cm3at 37˚C and ρl 1.00 g/cm3= at 4˚C). The concentration of particles near the cell surface

changes over time as sedimentation occurs, as described by equations 1 and 2.

Single Cell Model of Reaction Kinetics

In order to determine the three kinetic rate coefficients, the single cell model of particle

interaction (Wilhelm et al. 2002) was utilized. Equations 3 – 7 describe particle interactions with

an individual cell in terms of the number of particles bound per cell, Nb, the particles internalized

per cell, Ni, the number of binding sites, Nbs, the number of binding sites available at time zero,

B, and the total number of particles bound and internalized, Nt, evolving over time t. The kinetic

rate coefficients for particle adsorption (binding), desorption, and internalization are ka (in M-1s-

1), kd (in s-1), and ki (in s-1), respectively.

The number of particles bound on a cell per unit time is proportional to the concentration of

nanoparticles in the reservoir at the cell surface (considered to change over time due to

sedimentation as given in equations 3 and 4) and the number of binding sites on the cell. Both

the number of particles detaching and particles internalized are proportional to the number of

bound particles, Nb(t). The number of bound nanoparticles changes over time following equation

3 and the number of internalized particles follows equation 4.

8

bibdbsab NkNkCNk

dt

dN (3)

bii Nk

dt

dN (4)

The binding capacity of the cell was assumed to remain the same over time (B=constant),

i.e., the cell surface was unobstructed by the surrounding cells and the number of binding sites

present (whether occupied by bound nanoparticles or not) was assumed constant due to

continuous turnover of the membrane. The number of binding sites available for binding (Nbs)

changes over time based on the number of particles already bound (Nb).

bbs NBtN (5)

Solutions to equations 3 - 5 give the number of particles bound to the cell surface (equation

6) and the number of bound and internalized particles (equation 7) evolving over time.

tb eN 1

CBka ida kkCk (6)

ti

iibt ek

tktNtNtN

1 (7)

The number of binding sites and the kinetic rate coefficients for particle attachment,

detachment, and internalization were determined using the above equations and experimental

data of the number of particles attached to the cell at various bulk concentrations over time.

Assumptions made in equations 3 – 7 included the following: exocytosis was represented as a

reduction in ki, endocytosis was assumed to account for all particle internalization based on

studies of latex nanoparticles of similar sizes to those used here (Rejman et al., 2004), and

binding sites were assumed to regenerate immediately upon endocytosis. To estimate the

parameter B, concentration-dependent experimental data at 4˚C were fitted to equation 5 using

9

non-linear least squares fitting in Matlab® version 7.8, where t was infinite due to saturated

binding sites and ki was zero due to the absence of endocytosis at 4˚C. For determination of

reaction parameters ka, kd, and ki, experimental time-dependence data at 37˚C were fitted to

equation 6 using the value obtained for B and non-linear least squares fitting in MatLab®.

Statistical Analysis

Error bars represent the standard error calculated from triplicate samples. Goodness of fit was

determined by finding the coefficient of determination (R2) between the fitted curve derived

from coefficients and equation 6 and the experimentally obtained data points in MatLab®.

Results

Confocal Imaging of Nanoparticles and HUVEC

HUVEC grown on top of either 2D or 3D collagen were incubated in the presence of

nanoparticles of various sizes for one hour prior to rinsing and subsequent staining of nuclei with

Hoechst 33258 and cell membrane with CellMask Deep Red. The absence of cell migration into

the gel as well as the absence of nanoparticles in the underlying collagen substrate was

confirmed with image stacks in the z-direction, as shown in Figure 1 with the compressed z-stack

as the central image and projections through the z-plane along the yellow lines in the x- and y-

directions given as the sidebars.

Additionally, z-stacks were taken and compressed into a single view as shown in Figures 2

and 3 of HUVEC grown on 2D and 3D collagen, respectively. For both 2D and 3D, images show

an increase in red fluorescence (particles) with decreasing size indicating an increase in the

binding and/or internalization of particles for smaller particles. Individual particles are below the

resolution of the microscope so that counting of particles is not possible, and smaller particles

cause a diffuse area of fluorescence rather than being particulate.

10

For the case of 20 nm particles and HUVEC grown on 2D collagen, red fluorescence appears

to overlay very closely with the white outline of the entire cell indicating almost no cell

membrane area present without being covered with nanoparticles. Conversely, with the other

sized particles and all 3D samples, the cell membrane is quite visible while particles are present

within the outline of the cell membrane. Qualitatively, it appears from the images that the 2D

samples have higher levels of red fluorescence arising from nanoparticles than 3D samples.

Further studies were then conducted to quantify differences in nanoparticle uptake between

endothelial cells grown on 2D or 3D surfaces for a range of nanoparticle sizes.

Endothelial Nanoparticle Adhesion/Uptake Differs with Particle Size

For endothelial cells cultured on thin coatings of collagen (2D), nanoparticle adhesion and

internalization were shown to depend on particle size and exposure time. Particle adhesion and

uptake increased with time until saturation was reached. While the 20 nm particles did not reach

saturation after 24 hours, the other particles saturated around 1-5 hours (100 nm and 200 nm) to

4-8 hours (500 nm). The highest levels of adhesion and uptake were seen for the smallest

particles (20 nm), which after 24 hours reached levels 300 fold higher than the next highest

adherent nanoparticle (100 nm; Figure 4).

Nanoparticle Adhesion/Uptake Is Affected by Extracellular Matrix Compliance

Overall, 20 nm nanoparticle uptake was less abundant on endothelial cells on 3D versus 2D

substrates as shown in microscopy images (Figures 2 and 3) and quantitative fluorescence data of

nanoparticles (Figures 4 and 5), while nanoparticle adhesion and uptake is greater for 3D than for

2D for all other sizes. In order to more accurately compare data from different nanoparticle sizes,

computational modeling was performed to correct for differences in sedimentation velocity and

endocytosis in determination of kinetic parameters.

11

Number of Binding Sites Decreases with Increasing Nanoparticle Size

Confluent cell cultures were exposed to particles at various concentrations at 4˚C to examine

attachment and detachment of particles in the absence of endocytosis. Nanoparticle adhesion to

and uptake by endothelial cells was found to be dose dependent at 4˚C (Figure 6). The number of

particles adherent to the cells increased significantly with decreasing particle size for HUVEC

grown on both collagen substrates. Data indicate a leveling off of nanoparticle adherence to cells

with increasing concentration of nanoparticles in the reservoir of media. The smaller particles

require larger concentrations of nanoparticles in order for this leveling off to occur, i.e., the cells

have a higher binding capacity for the smaller particles.

The fluorescence data from Figure 6 were fitted to equation 5 with ki equal to zero and the

number of HUVEC approximated as the expected cell counts from confluent 48 well plate wells.

The total number of binding sites per cell, B, was determined using non-linear least curve fitting

in Matlab (Table I and Figure 7). The number of available binding sites decreased four orders of

magnitude (from ~1 x 108 to 1 x 104) as the particle size increased; larger particles take up more

space on the cell surface, likely blocking binding of other particles. Also, HUVEC grown on 3D

collagen had more available binding sites compared to HUVEC grown on 2D collagen coatings,

particularly at low particle sizes.

Determining Reaction Kinetic Coefficients

The rate coefficients (Figure 8) for attachment (ka), detachment (kd), and internalization (ki)

versus size for cells grown on top of 2D and 3D collagen are summarized in Table I. The

adsorption rate coefficient, ka, is higher for the 2D culture system for each size except 500 nm

diameter particles which have a 3.1 fold change for 3D over 2D. For the 20 nm, 100 nm, and

200 nm particles, the adsorption rate coefficient is 6.0, 3.6, and 1.4-fold higher for 2D versus 3D-

12

grown HUVEC. The internalization rate coefficient is larger (closer to zero) in the 3D culture

system for all sizes except 20 nm particles which have a 11-fold higher ki than the same particles

in the 2D culture system. For the 100 nm, 200 nm, and 500 nm particles, the internalization rate

coefficient is 120, 140, and 6700-fold higher for 3D versus 2D-grown HUVEC.

Dextran Effects on Particle Uptake

The inclusion of dextran in the cell growth media during exposure to nanoparticles

significantly reduced uptake and adhesion of nanoparticles. For 2D culture conditions after an

hour of exposure to 20 nm, 100 nm, 200 nm, or 500 nm particles, signal from dextran samples

was only 15%, 22%, 11%, and 10%, respectively, of that from the corresponding experiments

conducted in the absence of dextran. For the 3D samples, these percentages of dextran signal to

no-dextran signal for each ascending particle size were as follows: 11%, 27%, 50%, and 32%.

Data were normalized to the particle concentration at the cell surface to account for viscosity

differences. Figure 9 shows nanoparticle adsorption/internalization at each particle size for both

culture conditions compared between dextran and non-dextran samples.

DISCUSSION

In this work, the influence of cell culture substrate on the binding kinetics of nanoparticles over a

range of particle sizes was studied using HUVEC and fluorescent nanoparticles. The effect of

cell substrate was examined by culturing endothelial cells on a thin coating of collagen or on the

surface of a collagen hydrogel, and a noteworthy difference in kinetic parameters resulted from

this variation. The number of nanoparticles per cell was determined for particles with diameters

from 20 nm to 500 nm over a range of concentrations and time. This data was fit to an adaptation

of a single cell model of nanoparticle-cell interactions (Goodman et al., 2008; Wilhelm et al.,

2002) incorporating sedimentation. Non-linear least squares fitting of the experimental data was

13

used to determine interaction parameters including the number of binding sites available and

kinetic rate coefficients of attachment, detachment, and internalization. The parameters vary over

the size range of nanoparticles used here (20 nm to 500 nm) and may hold predictive value for

negatively charged particles interacting with endothelial cells.

Researchers typically evaluate nanoparticles using in vitro cell cultures, yet the two-

dimensional nature and surface rigidity of most adherent culture systems does not accurately

represent the biological complexity of the three dimensional in vivo environment (Ng and Pun,

2008; Griffith and Swartz, 2006). In this study, we used HUVEC cultured either on a thin

coating of collagen or on the surface of a collagen hydrogel to differentiate between the 2D and

3D extracellular environment of different compliance in vitro. The two cell substrates were

seemingly similar in chemical composition and varied primarily in the mechanical properties

experienced by the cells. Nanoparticle adhesion and uptake varied significantly depending on

whether cells were grown on top of 2D or 3D collagen. This is a striking finding and may carry

important implications for in vitro studies mimicking in vivo responses to nanoparticles. It is

postulated here that the 3D culture environment more accurately mimics the in vivo environment

experienced by a cell. Our findings support those of others that show surface compliance alters

cell responses.

At a given concentration, the binding of nanoparticles to the cell surface can be understood as

an interplay between the number of available binding sites and the adsorption coefficient ka. The

rate at which particles bind to the surface is balanced by the number of available sites to bind.

Across the range of nanoparticle sizes studied here, ka increases with increasing particle

diameter. Larger particles bind at a higher rate per unit concentration than do smaller particles.

Conversely, as particle size increases, the number of available binding sites on the cell decreases.

14

This is postulated to occur due to steric hindrance of the plasma membrane from larger particles,

restricting the space available for binding. These statements hold true for both the 2D and 3D

collagen substrate cell growth conditions examined here.

For all nanoparticle sizes, the number of binding sites per cell is larger for the 3D-collagen

cultured HUVEC than for cells cultivated on the 2D substrate. The number of binding sites may

differ due to either upregulation at the protein (receptor) level or increased surface area due to

compliance of the 3D collagen gel. Additionally, in all cases except the 500 nm particles, the ka

values for 2D collagen grown HUVEC were larger than those for 3D culture conditions. The fold

change in ka between 2D and 3D conditions decreases with increasing nanoparticle size up to

200 nm, while 500 nm particles have an approximately three-fold change from 3D to 2D. The

number of nanoparticles per cell after 24 hours was higher for cells grown on the 3D substrate

for all particle sizes except 20nm. Therefore, it appears that the increase in number of binding

sites dominates over the decrease in attachment kinetics in determining the number of

nanoparticles per cell on 3D substrates for 100-200nm particles.

For cells grown on 2D collagen, the rate of internalization decreases with increasing particle

size. This result is somewhat intuitive in that the larger particles are slower to enter the cell. This

result has been shown previously with latex particles (Rejman et al., 2004), viral particles

(Matlin et al., 1982), and polyplexes (Godbey et al., 1999). However, this trend is not seen in our

work with HUVEC cultures on the surface of 3D collagen hydrogels. In fact, in examining the

two particle size extremes, 20 nm and 500 nm, the larger particles are internalized at a higher

rate than the small-sized particles. The explanation for this result is not clear here and requires

further examination into the mechanism of uptake.

15

The desorption rate coefficient for 3D collagen-grown HUVEC increases with increasing

particle size. However, for 2D collagen grown cells, the trend does not hold, and in fact, three of

the four nanoparticle sizes have negative values for kd. While this result at first seems erroneous,

it is possible that the negative values are in fact a result of a more complex occurrence at the cell

membrane such as particle aggregation. If one particle binds to the cell surface and a second

particle subsequently aggregates to this particle without itself adsorbing to the cell surface, there

is no parameter in this single-cell attachment model to account for such an occurrence. The

particle would not accurately be described by the ka value or accounted for in the number of

binding sites. Therefore, these negative values for desorption may be explained by an

agglomeration process or more than one particle being attached per binding site.

Dextran largely blocks nanoparticle adsorption and uptake as shown in these results. Dextran

is a typical additive to media to increase viscosity for in vitro flow experiments (Blackman et al.,

2000; Rinker et al., 2001; Rouleau et al., 2010). Results suggest that the presence of dextran

blocks interactions between cells and particles, decreasing adherence and internalization.

Previous results support this finding as the dextran sugar molecules can be used to decrease

aggregation of particles by blocking charge-based interactions at the particle surface.

Nanoparticle design is a continually expanding area of research with the intended function

depending much on design factors such as size, surface chemistry, shape, surface charge, and

other factors. Size has been suggested by Morose as an important factor in determining the safety

of nanomaterials (2010). As shown in the results presented here, varying nanoparticle size has a

large influence on the attachment and internalization of particles. This information may be used

in the design of nanotherapies and other nanomaterials to result in a cell specific response. In

addition, these results may aide in the understanding of the possible toxic effects of

16

nanoparticles, i.e., nanotoxicity. While the uptake of nanoparticles has been shown to be

concentration, time, and energy-dependent (Davda and Labhasetwar, 2002; Panyam et al., 2003)

and uptake of polystyrene nanoparticles has been studied in several cell types (Clift et al., 2008;

Löhbach et al., 2006), the kinetics of uptake have not been determined over a range of sizes for

nanoparticles until the present study.

Conclusions

This study explored the adsorption and internalization kinetics of variously sized nanoparticles in

endothelial cells grown on two substrates of different compliance. Experimental data were fitted

to equations to extract information on the number of binding sites and kinetic coefficients for

adsorption, desorption, and internalization. Nanoparticle attachment and internalization were

influenced significantly by changing the substrate on which the cells were cultured. This study

elucidates the importance of particle size in nanoparticle-endothelial cell interactions, as the

number of binding sites per cell decreases with increasing nanoparticle size while the attachment

coefficient increases. Further, cell culture dimensionality and substrate compliance influenced

the binding of nanoparticles; this is a potentially important factor in translating in vitro studies of

nanoparticle binding to in vivo relevance.

Acknowledgements

The authors acknowledge the National Sciences and Engineering Research Council of Canada

(NSERC) Collaborative Health Research Projects (CHRP) grant for funding of the project, the T.

Chen Fong Postdoctoral Fellowship in Medical Imaging for funding of Amber Doiron, and the

Markin Undergraduate Student Research Program Scholarship for funding of Brendan Clark.

The authors additionally thank Robert Shepherd for helpful discussions and editorial comments.

17

References

Albrecht DR, Underhill GH, Wassermann TB, Sah RL, Bhatia SN. 2006. Probing the role of

multicellular organization in three-dimensional microenvironments. Nat Methods 3:369-

375.

Bhatia SN, Balis UJ, Yarmush ML, Toner M. 1998. Microfabrication of hepatocyte/fibroblast

co-cultures: role of homotypic cell interactions. Biotechnol Prog 14:378–387.

Blackman BR, Barbee KA, Thibault LE. 2000. In vitro cell shearing device to investigate the

dynamic response of cells in a controlled hydrodynamic environment. Ann Biomed Eng

28:363–372.

Byfield FJ, Reen RK, Shentu TP, Levitan I, Gooch KJ. 2009. Endothelial actin and cell stiffness

is modulated by substrate stiffness in 2D and 3D. J Biomech 42:1114-1119.

Caldorera-Moore M, Guimard N, Shi L, Roy K. 2010. Designer nanoparticles: incorporating

size, shape and triggered release into nanoscale drug carriers. Expert Opin Drug Delivery

7:479-495.

Chesler NC, Enyinna OC. 2003. Particle deposition in arteries ex vivo: effects of pressure, flow,

and waveform. J Biomech Eng 125:389.

Clift MJ, Rothen-Rutishauser B, Brown DM, Duffin R, Donaldson K, Proudfoot L, Guy K,

Stone V. 2008. The impact of different nanoparticle surface chemistry and size on uptake

and toxicity in a murine macrophage cell line. Toxicol Appl Pharmacol 232:418–427.

Cukierman E, Pankov R, Stevens DR, Yamada KM. 2001. Taking cell-matrix adhesions to the

third dimension. Science 294:1708-1712.

Davda J, Labhasetwar V. 2002. Characterization of nanoparticle uptake by endothelial cells. Int J

Pharm 233:51–59.

18

Godbey WT, Wu KK, Mikos AG. 1999. Tracking the intracellular path of poly

(ethylenimine)/DNA complexes for gene delivery. Proc Natl Acad Sci USA 96:5177-

5181.

Goldman J, Le TX, Skobe M, Swartz MA. 2005. Overexpression of VEGF-C causes transient

lymphatic hyperplasia but not increased lymphangiogenesis in regenerating skin. Circ

Res 96:1193-1199.

Goodman TT, Chen J, Matveev K, Pun SH. 2008. Spatio-temporal modeling of nanoparticle

delivery to multicellular tumor spheroids. Biotechnol Bioeng 101:388–399.

Griffith L, Swartz M. 2006. Capturing complex 3D tissue physiology in vitro. Nat Rev Mol Cell

Biol 7:211-224.

Heldin C, Rubin K, Pietras K, Ostman A. 2004. High interstitial fluid pressure - an obstacle in

cancer therapy. Nat Rev Cancer 4:806-813.

Löhbach C, Neumann D, Lehr C, Lamprecht A. 2006. Human vascular endothelial cells in

primary cell culture for the evaluation of nanoparticle bioadhesion. J Nanosci

Nanotechnol 6:3303-3309.

Matlin KS, Reggio H, Helenius A, Simons K. 1982. Pathway of vesicular stomatitis virus entry

leading to infection. J Mol Biol 156:609–631.

Morose G. 2010. The 5 principles of "design for safer nanotechnology". J Cleaner Prod 18:285–

289.

Ng CP, Pun SH. 2008. A perfusable 3D cell-matrix tissue culture chamber for in situ evaluation

of nanoparticle vehicle penetration and transport. Biotechnol Bioeng 99:1490-1501.

19

Panyam J, Sahoo S, Prabha S, Bargar T, Labhasetwar V. 2003. Fluorescence and electron

microscopy probes for cellular and tissue uptake of poly(D,L-lactide-co-glycolide)

nanoparticles. Int J Pharm 262:1-11.

Rejman J, Oberle V, Zuhorn IS, Hoekstra D. 2004. Size-dependent internalization of particles via

the pathways of clathrin-and caveolae-mediated endocytosis. Biochem J 377:159-169.

Rinker KD, Prabhakar V, Truskey GA. 2001. Effect of contact time and force on monocyte

adhesion to vascular endothelium. Biophys J 80:1722–1732.

Rouleau L, Rossi J, Leask RL. 2010. Concentration and time effects of dextran exposure on

endothelial cell viability, attachment, and inflammatory marker expression in vitro. Ann

Biomed Eng 38:1451-1462.

Sergeeva A, Kolonin MG, Molldrem JJ, Pasqualini R, Arap W. 2006. Display technologies:

application for the discovery of drug and gene delivery agents. Adv Drug Delivery Rev

58:1622–1654.

Singh R, Lillard Jr JW. 2009. Nanoparticle-based targeted drug delivery. Exp Mol Pathol

86:215–223.

Wilhelm C, Gazeau F, Roger J, Pons JN, Bacri J. 2002. Interaction of anionic superparamagnetic

nanoparticles with cells: kinetic analyses of membrane adsorption and subsequent

internalization. Langmuir 18:8148-8155.

20

Table Caption

Table I. Values for interaction parameters B (number of binding sites per cell) and kinetic rate

coefficients for attachment (ka), detachment (kd), and internalization (ki) for HUVEC grown on

2D and 3D collagen substrates with nanoparticles of various sizes. Obtained by fitting

experimental data to the single cell model of particle-cell interaction.

Figure Captions

signal shown in white is the cell membrane labeled with CellMask, blue is the cell nuclei, and

Figure 2. Confocal microscopy images of HUVEC grown on 2D collagen and exposed to no

nanoparticles (control), 20 nm, 100 nm, 200 nm, 500 nm nanoparticles for a period of one hour.

For each region of interest, z-stacks taken with a 60x objective were compressed into a single

view where fluorescence signal shown white is the cell membrane labeled with CellMask, blue is

the cell nuclei, and red is the nanoparticles. The final column is the overlay of all three. Scale

bar in the top left panel applies to all images and denotes 50 μm.

red is the nanoparticles. Compressed z-stack taken with 60x objective are shown as the central

image (A and D), and projections through the z-plane along the yellow lines in the y- (B and E)

and x-directions (C and F) are given as the sidebars. Green line along the sidebars denotes

location of glass slide and scale bar in A applies to all images and represents 50 μm.

(D-F) collagen exposed to 20 nm nanoparticles for a period of one hour where fluorescence

Figure 1. Confocal microscopy images of HUVEC grown on the surface of 2D (A-C) and 3D

21

Figure 3. Confocal microscopy images of HUVEC grown on 3D collagen and exposed to no

nanoparticles (control), 20 nm, 100 nm, 200 nm, 500 nm nanoparticles for a period of one hour.

For each region of interest, z-stacks taken with a 60x objective were compressed into a single

view where fluorescence signal shown white is the cell membrane labeled with CellMask, blue is

the cell nuclei, and red is the nanoparticles. The final column is the overlay of all three. Scale

bar in the top left panel applies to all images and denotes 50 μm.

Figure 4. Number of adherent and internalized particles over time per HUVEC (n=3). Cells

cultured on 2D collagen were incubated at 37˚C with particles of various sizes: 20 nm (A), 100

nm (B), 200 nm (C), and 500 nm (D), and fluorescence intensity was measured. The goodness

of fit (R2) for each fitted curve is 0.961 (A), 0.958 (B), 0.911 (C), and 0.951 (D).

Figure 5. Number of adherent and internalized particles over time per HUVEC (n=3). Cells

cultured on 3D collagen were incubated at 37˚C with particles of various sizes: 20 nm (A), 100

nm (B), 200 nm (C), and 500 nm (D), and fluorescence intensity was measured. The goodness

of fit (R2) for each fitted curve is 0.989 (A), 0.987 (B), 0.978 (C), and 0.982 (D).

Figure 6. Number of adherent particles per HUVEC. Cells cultured on either 2D (A-D) or 3D

(E-H) collagen were incubated at 4˚C with nanoparticles of various sizes: 20 nm (A and E), 100

nm (B and F), 200 nm (C and G), and 500 nm (D and H) over a range of concentrations for one

hour. The goodness of fit (R2) for each fitted curve is 0.992 (A), 0.975 (B), 0.991 (C), 0.964 (D),

0.976 (E), 0.925 (F), 0.956 (G), and 0.976 (H).

22

Figure 7. The number of nanoparticle binding sites per cell grown on either 2D or 3D collagen

substrate over the range of particle sizes. Data from Figure 6 were used to determine the number

of binding sites.

Figure 8. (A) Rate coefficient for attachment (ka) of nanoparticles to the cell surface for HUVEC

grown on either 2D collagen thin coatings or 3D collagen hydrogel. (B) Rate coefficient for

internalization (ki) of nanoparticles by HUVEC grown on either 2D collagen thin coatings or 3D

collagen hydrogel. Data from Figures 4 and 5 were used to determine kinetic coefficients.

Figure 9. Normalized number of nanoparticles bound per cell grown on either 2D or 3D

collagen, with sedimentation accounted for, after 1 hour of exposure in media with or without

dextran.

y

xz

z

A B

C

y

xz

z

D E

F

Control

20 nm

100 nm

200 nm

500 nm

Control

20 nm

100 nm

200 nm

500 nm

0 3 6 9 12 15 18 21 240

0 3 6 9 12 15 18 21 240

0 3 6 9 12 15 18 21 240

0 3 6 9 12 15 18 21 240

4 x 104

4 x 108

3 x 108

2 x 108

1 x 108

1 x 104

2 x 104

3 x 104

4 x 104

2 x 104

8 x 105

6 x 105

4 x 105

2 x 105

1 x 106

1.2 x 106

6 x 104

8 x 104

A

D

B

C

Time (hours)

0 3 6 9 12 15 18 21 240

0 3 6 9 12 15 18 21 240

0 3 6 9 12 15 18 21 240

0 3 6 9 12 15 18 21 240

1 x 108

1.5 x 108

5 x 107

4 x 104

3 x 104

2 x 104

1 x 104

2 x 106

3 x 106

1 x 106

4 x 105

3 x 105

2 x 105

1 x 105

Time (hours)

A

D

B

C5 x 104

0 0.5 1.0 1.5 2.0 2.50

Nanomolar

0 2 4 6 8 10 12 14 16 18 200

Nanomolar0 0.5 1.0 1.5 2.0 2.5

0

Micromolar

0 20 40 60 80 100 120 140 1600

Picomolar

2 x 107

4 x 107

6 x 107

8 x 107

2 x 105

4 x 105

6 x 105

1 x 105

8 x 104

6 x 104

4 x 104

2 x 104

1 x 104

8 x 103

6 x 103

4 x 103

2 x 103

0 0.5 1.0 1.5 2.0 2.50

Micromolar0 2 4 6 8 10 12 14 16 18 20

0

Nanomolar

0 0.5 1.0 1.5 2.0 2.50

Nanomolar0 20 40 60 80 100 120 140 160

0

Picomolar

5 x 105

1 x 106

1.5 x 106

2 x 106

2.5 x 106

1 x 105

5 x 104

1 x 108

5 x 107

2 x 108

1.5 x 108

2 x 105

1.5 x 105

2.5 x 105 1.5 x 104

1.25 x 10 4

1 x 104

7.5 x 103

5 x 103

2.5 x 103

3 x 1062.5 x 108

A

F

D

B

HG

E

C

Concentration of Nanoparticles in Media

20 nm 100 nm 200 nm 500 nm

3D collagen

2D collagen

Nanoparticle Size

1

1x102

1x104

1x103

1x105

1x106

1x107

1x108

1x109

10

20 nm 100 nm 200 nm 500 nm

3D collagen

2D collagen

1

1x102

1x104

1x103

1x10 5

1x10 6

1x107

1x108

1x109

10

A B

20 nm 100 nm 200 nm 500 nm

3D collagen1

1x10-8

1x10-6

1x10-7

1x10-5

1x10-4

1x10-3

1x10-2

1x10-1

Nanoparticle Size Nanoparticle Size