View Article Online / Journal Homepage / Table of Contents for this issue

Lab on a Chip To our referees:

Published on 08 November 2010. Downloaded by Universidade Federal de Sao Paulo on 23/07/2013 08:59:02.

Micro- & nano- fluidic research for chemistry, physics, biology, & bioengineering

www.rsc.org/loc Volume 10 | Number 24 | 21 December 2010 | Pages 3309–3432

Dank u wel kiitos takk fyrir

aitäh děkuji D’akujem Благодаря
Спасибо Thank you Tak
grazie Takk Tack 唔該 Danke
Merci gracias Ευχαριστω

どうもありがとうございます。

As a result of your commitment and support, RSC journals have a reputation for
the highest quality content. Your expertise as a referee is invaluable – thank you.

ISSN 1473-0197

PAPER www.rsc.org/publishing
Registered Charity Number 207890
Yuen and Faris
Perfusion-based microfluidic device for three-dimensional dynamic primary
human hepatocyte cell culture in the absence of biological or synthetic
matrices or coagulants
 PAPER www.rsc.org/loc | Lab on a Chip

Perfusion-based microfluidic device for three-dimensional dynamic primary

human hepatocyte cell culture in the absence of biological or synthetic
Published on 08 November 2010. Downloaded by Universidade Federal de Sao Paulo on 23/07/2013 08:59:02.

matrices or coagulants†
Vasiliy N. Goral,a Yi-Cheng Hsieh,b Odessa N. Petzold,a Jeffery S. Clark,a Po Ki Yuen*a and Ronald A. Faris*a
Received 21st June 2010, Accepted 15th October 2010
DOI: 10.1039/c0lc00135j

We describe a perfusion-based microfluidic device for three-dimensional (3D) dynamic primary human
hepatocyte cell culture. The microfluidic device was used to promote and maintain 3D tissue-like
cellular morphology and cell-specific functionality of primary human hepatocytes by restoring
membrane polarity and hepatocyte transport function in vitro without the addition of biological or
synthetic matrices or coagulants. A unique feature of our dynamic cell culture device is the creation of
a microenvironment, without the addition of biological or synthetic matrices or coagulants, that
promotes the 3D organization of hepatocytes into cord-like structures that exhibit functional
membrane polarity as evidenced by the expression of gap junctions and the formation of an extended,
functionally active, bile canalicular network.

Introduction improvement of hepatocyte cell function, they do not restore

The utilization of cell-based assays for predicting hepatotoxicity
is hampered by the rapid loss of hepatic function in conventional
two-dimensional (2D) cell culture. This lack of function is in part
due to the loss of normal tissue architecture that is critical for
maintaining cell polarity. Indeed, it is well documented that both
cell–cell and cell–extracellular matrix (ECM) interactions are
critical for maintaining long-term viability and function in vitro.
In the past decade, progress has been made in the development
of cell culture configurations and systems to restore hepatocyte
membrane polarity and enhance hepatocyte function in vitro for
longer time periods. Some of the cell configurations and systems,
such as hepatocyte sandwich cultures of rats and humans
between two layers of collagen or Matrigel overlay,1 three-
dimensional (3D) primary rat hepatocyte culture and perfused
hepatocyte culture systems,2,3 indeed provide some degree of
enhancement of hepatocyte cell performance by restoring cell
polarity relative to conventional cultures in maintaining viable
cell cultures with some phenotypic relevance.
Various methodologies have also been developed and applied
to promote prolonged expression of drug metabolizing enzymes,
liver-specific functions and cell viability in vitro. These include (i)
the use of modified culture media,4,6 (ii) co-cultures3,5,6 and (iii)
the use of various ECM to promote 3D cellular organization.6,7
However, mimicking the complex in vivo liver architecture and
microenvironmental interactions that are critical for successful
long-term hepatocyte cultures remains a challenge. Also, while
a few of the approaches highlighted above show limited
a
Science & Technology, Corning Incorporated, Corning, New York,
14831-0001, USA. E-mail: yuenp@corning.com; farisra@corning.com;
Fax: +1 (607) 974-5957; Tel: +1 (607) 974-9680
b
Science and Technology, Corning Research Center, Taiwan
† Electronic supplementary information (ESI) available: Video of sliced
confocal images of MRP2 protein immunofluorescent staining of
cultured primary human hepatocytes inside microfluidic device. See
DOI: 10.1039/c0lc00135j
tissue-like liver architecture or the formation of an extended,
anastomosing bile canalicular network that would be needed for
the restoration of membrane polarity in 3D.
There are a variety of macro- and microfluidic devices inten-
ded for miniaturized chemical processes, biochemical assays, cell
culture and artificial liver devices reported in the literature.8–24
Most reports describe a system in which cells are seeded prior to
the complete assembly of the device9,10,15,16 and the culture of cells
either in a 2D format8,14–17 or inoculation of the device with pre-
formed spheroids2,20,21 or the use of coagulant chemistry as
a mechanism to aggregate cells in a 3D format.22,23 Often animal
derived ECM or synthetic matrices, such as coagulants and
intercellular linker of modified cells, are also used to adhere or
retain cells in the devices.8,9,22–24 Ong et al. reported a microfluidic
device that utilizes physical restraints (pillars) in addition to
a synthetic matrix to help retain the cells which were aggregated
by the matrix in the device.23 Also, previous literature does not
describe the formation of 3D tissue-like cellular structures that
promote restoration of membrane polarity in 3D as evidenced by
the formation of extended bile canalicular structures and trans-
port function in a microfluidic device without the addition of
biological or synthetic matrices or coagulants.
In this article, we describe a perfusion-based microfluidic
device that supports the dynamic 3D culture of primary human
hepatocytes via independent perfusion microchannels and virtual
suspension of cells on bottom patterned microstructured
supports. The perfusion-based microfluidic device, which is
similar to the one described by Toh et al.22 and Ong et al.,23
consisted of a cell culture chamber centered between two side
microchannels (Fig. 1). A series of retention micropillars was
designed around the cell culture chamber for cell culture media to
perfuse into the chamber via the two side microchannels. Unlike
other perfusion-based microfluidic devices,22,23 the bottom of the
cell culture chamber was patterned with microstructures. The
microstructural design or topographical feature at the bottom of
the cell culture chamber provided an additional control of

3380 | Lab Chip, 2010, 10, 3380–3386 This journal is ª The Royal Society of Chemistry 2010

Published on 08 November 2010. Downloaded by Universidade Federal de Sao Paulo on 23/07/2013 08:59:02.
Fig. 1 Schematic diagram of perfusion-based microfluidic device.
hepatocyte polarity by minimizing cell spreading and cell–surface
interaction, and by allowing hepatocytes to be surrounded by
media. Also, cell culture media could be perfused into the cell
culture chamber via these bottom patterned microstructures. We
demonstrated that the device can be used to maintain 3D in vivo-
like cellular architecture while promoting cell–cell communica-
tion by the formation of gap junctions, restoring hepatocyte cell
polarity in 3D and enhancing cell-specific transporter function-
ality in vitro without the addition of biological or synthetic
matrices or coagulants.
Experimental
Device fabrication
Soft lithography was used to fabricate the microfluidic device
which was made from two poly(dimethylsiloxane) (PDMS)
replicas (Fig. 2).25,26 Briefly, features on two chrome masks were
transferred onto two silicon wafers, respectively, using standard
photolithographic process. The photoresist-defined silicon
wafers were then anisotropically etched to a desired depth in
a multiplex inductively coupled plasma (ICP) etching system
Fig. 2 Schematic diagram depicting the fabrication steps for the
perfusion-based microfluidic device.
This journal is ª The Royal Society of Chemistry 2010
(Surface Technology Systems, Newport, NJ, USA). After
photoresist removal and overnight silane (tri-
chloro(1H,1H,2H,2H-perfluorooctyl)silane) coating, a PDMS
prepolymer (10 : 1 w/w) (Sylgard 184, Dow Corning Corpo-
ration, Midland, MI, USA) was cast onto the etched silicon
wafers and cured at room temperature for at least 24 hours to
minimize shrinkage after curing. Next, the PDMS replicas were
carefully peeled away from the silicon wafers. Finally, after
punching inlet and outlet holes, the two PDMS replicas were
aligned and assembled under a microscope without any surface
treatment, i.e., the two PDMS replicas were reversibly bonded
together.
Flow characterization
Flow dynamics between the two side microchannels and the cell
culture chamber was experimentally characterized using car-
boxyfluorescein (4  105 M) and sulforhodamine B (SRB) (8.9
 105 M) fluorescent dyes in phosphate buffered saline (PBS)
solution. Carboxyfluorescein and SRB solutions were introduced
simultaneously into the empty cell culture chamber and the two
side microchannels, respectively, at a flow rate of 1 ml min1 per
inlet using a syringe pump (Model: SP230IW; World Precision
Instruments, Sarasota, FL, USA). Fluorescent flow distribution
inside the microfluidic device was monitored using the Zeiss
Axiovert 200 inverted fluorescence microscope (Carl Zeiss
MicroImaging, Inc., Thornwood, NY, USA) equipped with
standard fluorescein isothiocyanate (FITC) and Rhodamine
filters (Omega Optical, Brattleboro, VT, USA).
Flow dynamics of the bottom patterned microstructures was
experimentally characterized by introducing a fluorescent dye
(dextran–rhodamine conjugate (MW 10 000, 8 mg ml1) in
Hanks’ balanced salt solution (HBSS)) into the bottom patterned
microstructures at a flow rate of 0.1 ml min1 after primary
human hepatocytes were seeded and cultured for three days
inside the cell culture chamber. The three day cell culture was
necessary in order to ensure that primary human hepatocytes
were attached to the bottom microstructures and were not
washed away during the flow characterization experiments. The
low cell permeability of the dextran–rhodamine conjugate
allowed visualization of the flow inside the bottom patterned
microstructures and would significantly reduce the diffusion of
the dextran–rhodamine conjugate through the cultured cell body
and into the cell culture chamber and the two side microchannels.
Fluorescence flow distribution inside the microfluidic device was
monitored by using the Zeiss Axiovert 200 inverted fluorescence
microscope.
Primary human hepatocytes
Cryopreserved primary human hepatocytes were purchased from
XenoTech, LLC (Lenexa, KS, USA). Lots of hepatocytes were
prequalified for cell attachment and morphology by plating cells
on BD BioCoat Collagen I Cellware (BD Biosciences, San
Jose, CA, USA). Hepatocytes were routinely cultured in serum-
free MFE media commercially available from XenoTech, LLC.
Lab Chip, 2010, 10, 3380–3386 | 3381
 Cell seeding and culture transferred into 96-well microplate and chemiluminescent
intensity was measured on a plate reader.
Since cell–cell interactions play an important role in restoration
of hepatocyte specific functions, it was important to perform 2D
Published on 08 November 2010. Downloaded by Universidade Federal de Sao Paulo on 23/07/2013 08:59:02.

culture using highly confluent monolayers to maximize these Live/dead cell staining
interactions. For static 2D culture, we seeded 60 000 cells per well
The LIVE/DEAD Viability/Cytotoxicity Assay Kit for
in a 96-well microplate. This seeding density was chosen to
mammalian cells (Molecular Probes, Inc., Eugene, OR, USA)
obtain confluent monolayer of cells.
was used to determine the viability (live and dead) of the primary
For perfusion-based microfluidic device cell culture, the
human hepatocytes inside the microfluidic device. Fluorescent
microfluidic device, tubing (Tygon S-54-HL, Saint-Gobain
dye mixture was first perfused through the device via the two side
Performance Plastics, Akron, OH, USA) and syringes (500 ml,
microchannels for 30 minutes at a flow rate of 0.5 ml min1 fol-
gas tight 1700 series Hamilton, Reno, NV, USA) were sanitized
lowed by 15 minutes of PBS buffer wash at a flow rate of 0.5 ml
with 70% ethanol. The device was first filled with 70% ethanol
min1. Photographic images of viable cells were captured using
and soaked for 30 minutes. Then, it was rinsed with deionized
a Zeiss Axiovert 200 inverted fluorescence microscope equipped
water for 10 minutes at a flow rate of 1 ml min1. Next, the device
with an epifluorescence condenser and a camera system (Carl
was dried by applying a vacuum to the outlets. Tubing was
Zeiss MicroImaging, Inc., Thornwood, NY, USA).
sanitized by flowing 500 ml of 70% ethanol followed by 500 ml of
deionized water. After sanitization, droplets of 100 ml of cell
culture media were placed at the inlets and outlets of the device. Immunofluorescence staining of cultured hepatocytes
The device was then placed inside a vacuum chamber for 15
After seven day perfusion-based cell culture inside the micro-
minutes to eliminate any bubbles inside the device. A bright field
fluidic device, formaldehyde-fixed cultures were immunostained
microscope was used to inspect the primed device for any bubble
using antibodies against hepatocyte cell surface proteins, MRP2
entrapment. Once the primed device was free of bubbles, 5 ml of
and connexin 32. The choice of these two markers allowed
primary human hepatocytes suspension (2  106 cells ml1) were
documentation of the effects of dynamic culture on the forma-
introduced into the cell culture chamber using a 10 ml syringe via
tion of hepatocyte gap junctions and the impact of bottom
a syringe pump at a flow rate of 0.5 ml min1. Cell seeding was
patterned microstructures in restoring cell polarity. In order to
monitored under a bright field microscope and it was stopped
simplify the staining procedure, the top portion of the micro-
when the entire cell culture chamber was packed with cells.
fluidic device was removed to allow for easy access to the
After cell seeding, 300 ml of cell culture media were placed into
cultured primary human hepatocytes. Staining was performed
a well on top of each side microchannel inlet. Tubing was plug-
according to the protocol provided below. However, staining can
ged into the two side microchannel outlets and was connected to
be performed in the assembled device via the two side micro-
two syringes that were connected to a syringe pump. Next, the
channel perfusion of the device as well.
device was placed inside a 37  C cell culture incubator and
Briefly, after seven day perfusion culture inside the micro-
perfusion of cell culture media was started by withdrawing the
fluidic device, cultured primary human hepatocytes were first
two syringes at a flow rate of 5 ml h1. Cell culture media were
washed three times with 200 ml of PBS buffer per wash. Then,
added daily into each side microchannel inlet well during the
they were incubated with 200 ml of 3% formaldehyde for 15
seven day incubation period. As a control, primary human
minutes at room temperature. After another three times wash of
hepatocytes were also cultured in static condition in the micro-
200 ml of PBS buffer per wash, the cells were permeabilized with
fluidic device. In this case, cell culture media were changed daily
200 ml of 1% Triton X-100 solution for 15 minutes at room
inside the device without any perfusion by perfusing 10 ml of cell
temperature. Next, the cells were washed twice with 200 ml of
culture media through the two side microchannels.
washing buffer (0.1% Tween-20 in PBS) per wash. In order to
block nonspecific binding, the sample was incubated for 30
minutes in 200 ml of blocking buffer (0.1% Tween-20 and 0.5%
goat serum in PBS). After nonspecific binding blocking, the cells
Adenosine triphosphate (ATP) cytotoxicity assay
were incubated with a mixture of primary antibodies (1 : 100
The CellTiter-Glo Luminescent Cell Viability Assay Kit dilution by blocking buffer, 25 mg ml1 of mouse anti-connexin
(Promega Corporation, Madison, WI, USA) was used to deter- 32 unconjugated monoclonal antibody, 9 mg ml1 of Rabbit anti-
mine the number of viable cells in culture based on quantitation MRP2 polyclonal antibody (Abcam, Inc., Cambridge, MA,
of cellular ATP, which signals the presence of metabolically USA)) overnight at 4  C. After 4  C incubation, the sample was
active cells. First, calibration curve of chemiluminescent intensity washed three times with 200 ml of washing buffer per wash and
versus number of primary human hepatocytes seeded in BD incubated with secondary antibodies conjugated to FITC (494/
collagen coated 96-well microplate or microfluidic device was 518 nm) and Cy3 (550/570 nm) fluorescent labels in 1 : 100
determined. All assays in 96-well microplate format were per- dilution of blocking buffer for one hour at room temperature.
formed according to the manufacture protocol except the Unbound antibodies were washed away by washing the sample
microplate was allowed to incubate at room temperature for 60 three times with 200 ml of washing buffer per wash. Finally, 20 ml
minutes to stabilize the signal. In order to determine the cell of Vectashield mounting solution supplemented with DAPI stain
number in a microfluidic device, 1 : 1 mixture of CellTiter-Glo were added to the sample in order to stain the cells’ nuclei.
reagent and cell culture media was perfused through the device at Fluorescent images were acquired with a Zeiss Axiovert 200
a flow rate of 1.6 mm min1 for one hour. Then, perfusate was inverted fluorescence microscope.

3382 | Lab Chip, 2010, 10, 3380–3386 This journal is ª The Royal Society of Chemistry 2010
 MRP2 transport function assay
MRP2 is an important transporter protein that modulates
pharmacokinetics of many drugs. Its expression and activity in
Published on 08 November 2010. Downloaded by Universidade Federal de Sao Paulo on 23/07/2013 08:59:02.
primary human hepatocyte cells indicate the restoration of
phenotype specific functions of primary human hepatocyte cells.
In order to verify the functionality of MRP2 transporter protein
inside the perfusion-based cell cultured primary human hepato-
cytes, MPR2 substrate (5 mM 5-(6) carboxy-20 ,70 -dichloro-
fluorescein diacetate solution in cell culture media) was perfused
through the device for 10 minutes at a flow rate of 1 ml min1.
Carboxy-20 ,70 -dichlorofluorescein diacetate was absorbed by the
cells and metabolized. The metabolites were actively excreted by
MRP2 protein complexes into bile canalicular structures. Efflux
of fluorescent metabolites was monitored by a fluorescent
microscopy revealing the bile canalicular structures inside the cell
aggregates.
Results
Flow characterization
Fluorescent images of carboxyfluorescein and SRB fluorescent
dye showed good perfusion performance of the microfluidic
device between the cell culture chamber and the two side
microchannels (Fig. 3a and b). Also, fluorescent images of
dextran–rhodamine conjugate showed that independent flow can
be achieved through the bottom patterned microstructures of the
cell culture chamber (compare Fig. 3c with 3d) demonstrating
additional perfusion flow can be used with the bottom patterned
microstructures. In addition, we reasoned that perfusion flow in
the bottom channel can be used to mimic complex extracellular
fluid distribution that is observed in vivo, e.g., to provide
a gradient of bile salts as is characterized in liver tissue. However,
this bottom independent flow can only be achieved when the
entire cell culture chamber was densely packed with cells and
cells fuse together fluidically isolating bottom channel from two
Fig. 3 Fluorescent images during flow characterization experiments. (a)
Sulforhodamine B fluorescent dye perfusing from two side microchannels
into the empty cell culture chamber. (b) Carboxyfluorescein fluorescent
dye perfusing from empty cell culture chamber into two side micro-
channels. (c) Dextran–rhodamine conjugate flowing through bottom
patterned microstructures and it did not perfuse or leak into cell culture
chamber and two side microchannels. (d) Dextran–rhodamine conjugate
flowing through cell culture chamber and two side microchannels.
Primary human hepatocytes were seeded and cultured for three days
inside cell culture chamber before flow characterization experiments
depicted in (c) and (d).
This journal is ª The Royal Society of Chemistry 2010
side perfusion channels. Furthermore, the flow rate of the
bottom independent flow was an order of magnitude smaller
than the two side perfusion microchannels due to the high flow
resistance.
Primary human hepatocyte cell culture
Primary human hepatocytes were cultured in the microfluidic
devices in order to evaluate the impact of dynamic culture
conditions on hepatocyte viability, polarity and function for
prolonged periods of time. It was estimated by ATP assay that
approximately 10 000 primary human hepatocytes were seeded
inside the cell culture chamber once it was densely packed
(Fig. 4a). We monitored the hepatocytes inside the devices daily
and at different time points, we performed a live/dead cell
staining to monitor cell survival rate and morphology. After two
weeks of perfusion-based cell culture, primary human hepato-
cytes remained viable (>90%) (Fig. 4b and 5a). Also, after
dissembling the devices, it was found that hepatocytes adhered
tightly together to form cord-like structures that resembled 3D
tissue-like cellular architecture without addition of matrices or
coagulants (Fig. 5b). In contrast, after two weeks of static cell
culture in the device with daily media change, most of the
hepatocytes were dead (Fig. 5c) and readily dispersed when the
devices were dissembled (Fig. 5d). In addition, we were able to
dislodge the fused tissue-like cells from the bottom patterned
microstructures of the device (Fig. 5e).
The difference in cell viability in static and perfused-based cell
culture can be explained by the fact that in static cell culture,
delivery of nutrients and waste removal was performed cyclically
once a day, while in perfusion-based cell culture it was a contin-
uous process that better mimics in vivo liver perfusion with blood.
As a result, we observed high viability in the perfusion-based cell
culture and enhanced cell–cell fusion, due to the limited ability of
cell spreading that mimics sinusoid tissue organization of liver.
Evaluation of hepatocyte membrane polarity and gap junction
expression
It is known that hepatocytes are supported in a 3D configuration
by a combination of ECM and other non-parenchymal cells in
vivo. It is also known that in conventional 2D in vitro cell culture
format, primary human hepatocytes dedifferentiate rapidly
Fig. 4 Primary human hepatocyte cell viability study by ATP cytotox-
icity assay. (a) Chemiluminescent intensity as a function of number of
viable cells in both microfluidic device and 96-well microplate. (b)
Chemiluminescent intensity as a function of culture day in microfluidic
device at an original cell seeding density of approximately 10 000 cells per
device.
Lab Chip, 2010, 10, 3380–3386 | 3383
Published on 08 November 2010. Downloaded by Universidade Federal de Sao Paulo on 23/07/2013 08:59:02.
Fig. 5 Two week primary human hepatocyte cell culture inside micro-
fluidic devices. (a, b and e) Perfusion-based cell culture. (c and d) Static
cell culture. (a and c) Fluorescent images of live (green)/dead (red) cell
staining. (b, d and e) Microscopic images of hepatocytes after device
disassembly. (e) Fused tissue-like cells dislodged from the bottom pattern
microstructures of the device.
because of limited cell–cell interaction and the inability to restore
in vivo-like cellular organization. The maintenance of differen-
tiated functions of primary human hepatocytes is, in part,
dependent on the restoration of morphological structure and
membrane polarity. The metabolic functions of primary human
hepatocytes have been clearly correlated to the polarity of
hepatocytes induced by different culture configurations. There-
fore, restoration of hepatocyte polarity is important in the
maintenance of hepatocyte function.
Hepatocyte membrane polarity was evaluated by monitoring
the expression and function of the multidrug resistant protein 2,
a hepatobiliary transporter critical in the efflux of drug metab-
olites. In addition, the expression of connexin 32 was employed
to examine the impact of dynamic cell culture on the restoration
of hepatocyte gap junctions. In conventional 2D cell culture, the
expression of MRP2 protein was occasionally observed as tiny
unconnected dots between two adjacent hepatocytes illustrating
the limited formation of bile canalicular structures (Fig. 6a and
7a). Other investigators have previously attempted to improve
the formation of bile canalicular structures and hence the
expression of MRP2 protein. For example, Xia et al. described
a laminar-flow immediate-overlay hepatocyte sandwich perfu-
sion system for drug hepatotoxicity testing. However, only
limited expression of MRP2 protein in their perfusion culture
was shown.27 In contrast, culturing hepatocytes using our
uniquely designed perfusion-based microfluidic system promoted
the formation of tightly fused, 3D tissue-like cellular morphology
(Fig. 5b and e), restored cell membrane polarity (Fig. 6b and c),
and transport function (Fig. 7b) without the addition of animal
derived or synthetic matrices or coagulants. Cell membrane
3384 | Lab Chip, 2010, 10, 3380–3386
Fig. 6 Immunofluorescent staining of cultured primary human hepa-
tocytes. (a–c) MRP2 protein. (d) Connexin 32 protein. (a) Seven day 2D
static cell culture in a BD collagen coated 96-well microplate (60 000 cells
per well) showing polarity was not being restored. (b and c) Seven day
perfusion-based cell culture inside microfluidic devices with microstruc-
tures at the bottom of cell culture chamber. FITC (green) fluorescent
showing extended bile canalicular structure formation and polarity
restoration. DAPI (blue) fluorescent showing cells’ nuclei. (c) Maximum
intensity projection image from sliced confocal images. (d) FITC (green)
showing expression of gap junction protein connexin 32. Cell nuclei were
stained with DAPI (blue).
polarity was evidenced by the extended formation of bile cana-
licular structures in 3D shown by the expression of the bile
canalicular marker MRP2 (Fig. 6b and c, and ESI: video of sliced
confocal images†) and gap junction protein connexin 32 (Fig. 6d)
via immunofluorescent staining. The bile canalicular MRP2
transporter is responsible for the efflux of drug metabolites into
bile. Transport function, often referred to as Phase III of drug
metabolism in the liver, is critical for removal of drug metabolites
or the active transport of drug compounds into cells. In this
study, human hepatocyte transporter function was demonstrated
in the microfluidic device using carboxyfluorescein diacetate,
a substrate for the MRP2 transporter. Carboxyfluorescein
diacetate is passively absorbed by the hepatocytes, metabolized
and fluorescein diacetate is actively effluxed via MRP2 transport
protein into extended bile canalicular structures (Fig. 7b). This
function is largely driven by restoration of cell polarity resulting
from the formation of tightly fused, 3D tissue-like cellular
structure. To the best of our knowledge, the formation of such
extended bile canalicular structures in 3D (restoration of
Fig. 7 MRP2 hepatocyte transport function assay. (a) Seven day 2D
static primary human hepatocyte cell culture in a BD collagen coated 96-
well microplate (60 000 cells per well) showing limited transport function.
(b) Seven day perfusion-based primary human hepatocyte cell culture
inside the microfluidic device with microstructures at the bottom of cell
culture chamber showing active transport of fluorescent dye by MRP2
protein into extended bile canalicular structures.
This journal is ª The Royal Society of Chemistry 2010
 membrane polarity) and the transport function of primary
human hepatocytes in a microfluidic device without the addition
of an ECM have not been demonstrated or reported before.
Published on 08 November 2010. Downloaded by Universidade Federal de Sao Paulo on 23/07/2013 08:59:02.
Discussion
We designed the feature dimensions of the perfusion-based
microfluidic device based on (a) mass transport of the perfused
fluids (e.g., cell culture media and reagents for culture and assay),
(b) the size of the primary human hepatocytes to promote cell–
cell interactions (signaling), and (c) the desired final cellular
organization. Thus, the width of the cell culture chamber was
chosen to be 100 mm to accommodate 3–5 cells diameter in width
while the dimensions of the side microchannels were designed to
ensure a ready supply and efficient transport of the cell culture
media and other reagents perfusing into the cell culture chamber.
Also, the two side microchannels had independent inlets and
outlets which could be used to perfuse two different fluids or fluid
with two different concentrations to provide a concentration
gradient across the packed cells inside the cell culture chamber.
In addition, the spacing between cell retention micropillars was
less than one cell diameter to provide efficient cell retention
during cell seeding. Various cross-sectional shapes (e.g., circular,
oval, square, triangular or V-shape) of retention micropillars
were investigated to increase the rigidity of cell retention
micropillars during device fabrication and to study the perfusion
efficiency. The V-shaped retention micropillar was found to be
the most rigid and robust to manipulate during fabrication and
use of the device as well as to provide a good perfusion perfor-
mance.
We incorporated microstructures in the form of 15 mm tall
micropillars (e.g., circular, oval or square) into the bottom of the
cell culture chamber to enhance cellular organization. Micro-
pillars were spaced at a distance less than one cell diameter to
prevent cells from falling into the bottom during cell seeding.
This spacing also allowed cells to attach to the top of the
microstructures during cell culture (incubation). It further
restricted the spreading of cells by limiting accessible surface area
within the plane for cell spreading. In addition, we designed the
microstructures to contain or entrap cells in the cell culture
chamber with minimized or limited cell interaction with the walls
or surrounding structures that the device provided. Therefore,
the primary human hepatocytes were retained or even encapsu-
lated by retention micropillars separating cells from the two side
perfusion microchannels and the bottom patterned microstruc-
tures. Tight packaging of cells inside the device in the absence of
ECM materials induced extensive 3D cell–cell interactions
during perfusion-based culture.
Conclusions
We presented perfusion-based microfluidic devices for functional
maintenance of primary human hepatocytes in highly differen-
tiated state in vitro. The ability of primary human hepatocytes to
support their in vivo functions while being cultured in vitro has
high importance in tissue engineering applications and evalu-
ating therapeutic candidates. The device design features promote
and maintain 3D in vivo-like cellular structure that promotes
cell–cell interactions as evidenced by the formation of gap
This journal is ª The Royal Society of Chemistry 2010
junctions and the formation of extended bile canalicular struc-
tures in 3D. For the first time, we demonstrated a microfluidic
device that enables the restoration of membrane polarity and
transport function in vitro without the addition of biological or
synthetic matrices or coagulants. This cell culture system may be
used to support long term cell cultures, to promote restoration of
in vivo-like cellular organization that increases phenotype specific
activity of cultured cells thus providing physiologically relevant
information for cell-based assays.
Acknowledgements
We would like to thank Dr Joydeep Lahiri for his support of this
project, Dr Jin Liu and Wendy Baker for their invaluable tech-
nical advice for culturing primary human hepatocytes and Todd
L. Heck for taking the confocal images.
References
1 A. Kern, A. Bader, R. Pichlmayr and K.-F. Sewing, Drug metabolism
in hepatocyte sandwich cultures of rats and humans, Biochem.
Pharmacol., 1997, 54, 761–772.
2 M. J. Powers, D. M. Janigian, K. E. Wack, C. S. Baker, D. Beer Stolz
and L. G. Griffith, Functional behavior of primary rat liver cells in
a three-dimensional perfused microarray bioreactor, Tissue Eng.,
2002, 8, 499–513.
3 R. Gebhardt, J. G. Hengstler, D. M€ uller, R. Gl€
ockner, P. Buenning,
B. Laube, E. Schmelzer, M. Ullrich, D. Utesch, N. Hewitt, M. Ringel,
B. R. Hilz, A. Bader, A. Langsch, T. Koose, H. J. Burger, J. Maas and
F. Oesch, New hepatocyte in vitro systems for drug metabolism:
metabolic capacity and recommendations for application in basic
research and drug development, standard operation procedures,
Drug Metab. Rev., 2003, 35, 145–213.
4 N. Katsura, I. Ikai, T. Mitaka, T. Shiotani, S. Yamanokuchi,
S. Sugimoto, A. Kanazawa, H. Terajima, Y. Mochizuki and
Y. Yamaoka, Long-term culture of primary human hepatocytes
with preservation of proliferative capacity and differentiated
functions, J. Surg. Res., 2002, 106, 115–123.
5 D.-H. Lee, H.-H. Yoon, J.-H. Lee, K.-W. Lee, S.-K. Lee, S.-K. Kim,
J.-E. Choi, Y.-J. Kim and J.-K. Park, Enhanced liver-specific
functions of endothelial cell-covered hepatocyte hetero-spheroids,
Biochem. Eng. J., 2004, 20, 181–187.
6 E. L. LeCluyse, P. L. Bullock and A. Parkinson, Strategies for
restoration and maintenance of normal hepatic structure and
function in long-term cultures of rat hepatocytes, Adv. Drug
Delivery Rev., 1996, 22, 133–186.
7 J. Z. Tong, S. Sarrazin, D. Cassio, F. Gauthie and F. Alvarez,
Application of spheroid culture to human hepatocytes and
maintenance of their differentiation, Biol. Cell, 1994, 81, 77–81.
8 J. Park, M. Toner, M. L. Yarmush and A. W. Tilles, Microchannel
bioreactors for bioartificial liver support, Microfluid. Nanofluid.,
2006, 2, 525–535.
9 L. De Bartolo, G. Jarosch-Von Schweder, A. Haverich and A. Bader,
A novel full-scale flat membrane bioreactor utilizing porcine
hepatocytes: cell viability and tissue-specific functions, Biotechnol.
Prog., 2000, 16, 102–108.
10 L. De Bartolo, S. Salerno, S. Morelli, L. Giorno, M. Rende,
B. Memoli, A. Procino, V. E. Andreucci, A. Bader and E. Drioli,
Long-term maintenance of human hepatocytes in oxygen-permeable
membrane bioreactor, Biomaterials, 2006, 27, 4794–4803.
11 P. J. Lee, P. J. Hung and L. P. Lee, An artificial liver sinusoid with
a microfluidic endothelial-like barrier for primary hepatocyte
culture, Biotechnol. Bioeng., 2007, 97, 1340–1346.
12 P. J. Lee, T. A. Gaige, N. Ghorashian and P. J. Hung, Microfluidic
tissue model for live cell screening, Biotechnol. Prog., 2007, 23, 946–
951.
13 M. Y. Zhang, P. J. Lee, P. J. Hung, T. Johnson, L. P. Lee and
M. R. K. Mofrad, Microfluidic environment for high density
hepatocyte culture, Biomed. Microdevices, 2008, 10, 117–121.
Lab Chip, 2010, 10, 3380–3386 | 3385
 14 P. J. Hung, P. J. Lee, P. Sabounchi, N. Aghdam, R. Lin and L. P. Lee,
A novel high aspect ratio microfluidic design to provide a stable and
uniform microenvironment for cell growth in a high throughput
mammalian cell culture array, Lab Chip, 2005, 5, 44–48.
Published on 08 November 2010. Downloaded by Universidade Federal de Sao Paulo on 23/07/2013 08:59:02.
15 A. Sin, K. C. Chin, M. F. Jamil, Y. Lostov, G. Rao and M. L. Shuler,
The design and fabrication of three-chamber microscale cell culture
analog devices with integrated dissolved oxygen sensors, Biotechnol.
Prog., 2004, 20, 338–345.
16 G. T. Baxter and R. Freedman, A dynamic in vivo surrogate assay
platform for cell-based studies, Am. Biotechnol. Lab., 2004, 28–29.
17 B. G. Chung, L. A. Flanagan, S. W. Rhee, P. H. Schwartz, A. P. Lee,
E. S. Monuki and N. L. Jeon, Human neural stem cell growth and
differentiation in a gradient-generating microfluidic device, Lab
Chip, 2005, 5, 401–406.
18 Y. Tanaka, K. Sato, M. Yamato, T. Okano and T. Kimatori, Cell
culture and life support system for microbioreactor and bioassay, J.
Chromatogr., A, 2006, 1111, 233–237.
19 S. Ostrovidov, J. Jiang, Y. Sakai and T. Fujii, Membrane-based
PDMS microbioreactor for perfused 3D primary rat hepatocyte
cultures, Biomed. Microdevices, 2005, 6, 279–287.
20 M. J. Powers, K. Domansky, M. R. Kaazempur-Mofrad, A. Kalezi,
A. Capitano, A. Upadhyaya, P. Kurzawski, K. E. Wack, D. Beer
Stolz, R. Kamm and L. G. Griffith, A microfabricated array bioreactor
for perfused 3D liver culture, Biotechnol. Bioeng., 2002, 78, 257–269.
3386 | Lab Chip, 2010, 10, 3380–3386
21 A. Sivaraman, J. K. Leach, S. Townsend, T. Iida, B. J. Hogan,
D. Beer Stolz, R. Fry, L. D. Samson, S. R. Tannenbaum and
L. G. Griffith, A microscale in vitro physiological model of the
liver: predictive screens for drug metabolism and enzyme induction,
Curr. Drug Metab., 2005, 6, 569–591.
22 Y.-C. Toh, C. Zhang, J. Zhang, Y. M. Khong, S. Chang,
V. D. Samper, D. van Noort, D. W. Hutmacher and H. Yu, A
novel 3D mammalian cell perfusion-culture system in microfluidic
channels, Lab Chip, 2007, 7, 302–309.
23 S.-M. Ong, C. Zhang, Y.-C. Toh, S. H. Kim, H. L. Foo, C. H. Tan,
D. van Noort, S. Park and H. Yu, A gel-free 3D microfluidic cell
culture system, Biomaterials, 2008, 29, 3237–3244.
24 M. S. Kim, J. H. Yeon and J.-K. Park, A microfluidic platform for 3-
dimensional cell culture and cell-based assays, Biomed. Microdevices,
2007, 9, 25–34.
25 Y. Xia and G. M. Whitesides, Soft lithography, Annu. Rev. Mater.
Sci., 1998, 28, 153–184.
26 D. C. Duffy, J. C. McDonald, O. J. A. Schueller and
G. M. Whitesides, Rapid prototyping of microfluidic systems in
poly(dimethylsiloxane), Anal. Chem., 1998, 70, 4974–4984.
27 L. Xia, S. Ng, R. Han, X. Tuo, G. Xiao, H. L. Leo, T. Cheng and
H. Yu, Laminar-flow immediate-overlay hepatocyte sandwich
perfusion system for drug hepatotoxicity testing, Biomaterials, 2009,
30, 5927–5936.
This journal is ª The Royal Society of Chemistry 2010