Cultivating liver cells on printed arrays of hepatocyte growth factor
Caroline N. Jonesa, Nazgul Tuleuovaa,d, Ji Youn Leea, Erlan Ramanculovd, A. Hari Reddib,
Mark A. Zernc, Alexander Revzina,*
aDepartment of Biomedical Engineering, University of California, Davis, 451 East Health Sciences Dr. #2519, Davis, CA 95616, USA
bDepartment of Orthopaedic Surgery, Center for Tissue Regeneration and Repair, University of California, Sacramento, CA 95817, USA
cDepartment of Medicine, Transplant Research Institute, University of California, Sacramento, CA 95817, USA
dNational Center for Biotechnology, Astana, Republic of Kazakhstan
a r t i c l e i n f o
Received 10 January 2009
Accepted 21 March 2009
Available online 17 April 2009
Growth factor microarrays
a b s t r a c t
Growth factors are commonly present in soluble form during in vitro cell cultivation experiments in order
to provide signals for cellular proliferation or differentiation. In contrast to these traditional experiments,
we investigated solid-phase presentation of a hepatocyte growth factor (HGF), a protein important in
liver development and regeneration, on microarrays of extracellular matrix (ECM) proteins. In our
experiments, HGF was mixed in solution with ECM proteins (collagen (I), (IV) or laminin) and robotically
printed onto silane-modified glass slides. Primary rat hepatocytes were seeded onto HGF/ECM protein
microarrays and formed cellular clusters that corresponded in size to the dimensions of individual
protein spots (500 mm diameter). Analysis of liver-specific products, albumin and a1-antitrypsin,
revealed several fold higher levels of expression of these proteins in hepatocytes cultured on HGF/ECM
microarrays compared to cells cultivated on ECM proteins alone. In addition, cultivation of hepatocytes
on HGF/ECM protein spots led to spontaneous reorganization of cellular clusters from a monolayer into
three-dimensional spheroids. We also investigated the effects of surface-tethered HGF on hepatocytes
co-cultivated with stromal cells and observed a significantly higher level of albumin in co-cultures where
hepatocytes were stimulated by HGF/ECM spots compared to co-cultures created on ECM protein islands
without the growth factor. In summary, our study suggests that incorporation of HGF into ECM protein
microarrays has a profound and long-lasting effect on the morphology and phenotype of primary
hepatocytes. In the future, the number of growth factors printed on ECM microarrays will be expanded to
enable multiplexed and combinatorial screening of inducers of cellular differentiation or proliferation.
? 2009 Elsevier Ltd. All rights reserved.
Growth factors play an important role in regulating cellular
behavior, including stimulation of proliferation [1,2], migration [3,4]
hepatocytes, non-parenchymal cells and recruited inflammatory
cells. Hepatocyte growth factor (HGF), a mesenchyme-derived
regeneration after injury. HGF is a pleiotropic morphogen that has
been shown to have mitogenic, motogenic, and antiapoptotic effects
[6,7]. In addition, HGF is being explored as an anti-fibrotic agent and
may have applications for treatment of liver fibrosis [8,9].
Current in vitro cell cultivation strategies commonly rely on
providing growth factors in the soluble form. These traditional
approaches require significant amounts of expensive growth
factors (GFs) and are not optimal for stem cell differentiation or
primary cell maintenance studies where frequent media changes
are required. This presents a particularly challenging problem in
stem cell differentiation studies where a formulation of GFs
required for stem cell lineage selection is often unknown, requiring
extensive and expensive experiments involving combinations of
In contrast to in vitro experiments where GF molecules are
present in solution, invivo, GFs bind to ECM matrix proteins and are
dynamically released during matrix remodeling and protease
secretion by the surrounding cells [14–16]. To mimic solid-phase
presentation observed in vivo a number of reports have described
strategies for surface immobilization of GF molecules via covalent
tethering [17–20]. In addition, given that in vivo GFs form
secondary bonds with either glycosaminoglycans [2,21–24] or with
matrix proteins [15,25–27], non-covalent binding represents an
alternative route for surface immobilization of these molecules.
* Corresponding author. Tel.: þ1 530 752 2383; fax: þ1 530 754 5739.
E-mail address: email@example.com (A. Revzin).
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/biomaterials
0142-9612/$ – see front matter ? 2009 Elsevier Ltd. All rights reserved.
Biomaterials 30 (2009) 3733–3741
Association with ECM proteins may provide additional benefits of
stabilizing GF molecules against proteolytic degradation and
enhancing their activity .
Beyond offering a more physiological scenario, surface immo-
bilization of GF molecules makes it possible to design strategies for
high-throughput screening of cell function. Robotic microarraying,
a technology originally designed for high-throughput screening of
DNA hybridization events , is particularlyamenable for printing
arrays of biomolecules on cell culture surfaces. This robotic printing
technology has been previously employed for high-throughput
studies of cell–ECM , cell–biomaterial  and cell–small
molecule interactions . More recently Soen and Davis employed
arrays of ECM proteins and morphogens to investigate differenti-
ation of primary neuronal cells .
We have previously reported on the use of ECM protein
microarrays for cultivation of hepatocytes in mono- and co-cultures
. These microarray-based cell cultures were complemented
with methods for the analysis of gene expression and secreted
product signatures within the local microenvironment [35,36]. In
the present study, we describe GF presentation on printed ECM
microarrays as a means to deliver stimuli to cultured hepatocytes.
This study represents a step towards an integrated cell culture
platformwheresignals are delivered and cell function is detectedin
a high-throughput and location-specific manner.
2. Materials and methods
2.1. Chemicals and materials
Glass slides (75?25 mm2) were obtained from VWR (West Chester, PA). (3-
Acryloxypropyl)trichlorosilane was purchased from Gelest, Inc. (Morrisville, PA).
Sulfuric acid, hydrogen peroxide, ethanol, collagenase, collagen from rat tail (type I),
collagen IV, laminin, hepatocyte growth factor (HGF), streptavidin-conjugated Alexa
546 and protease inhibitor cocktail were obtained from Sigma–Aldrich (St. Louis,
MO). Concentrated phosphate-buffered saline (10? PBS) was purchased from Lonza
(Walkersville, MD). Minimal essential medium (MEM), sodium pyruvate, nones-
sential amino acids, fetal bovine serum (FBS), Superscript III, RNaseOut (RNase
inhibitor), dNTPs and biotinylated anti-HGF antibodies were purchased from Invi-
trogen (Carlsbad, CA). 384-well polypropylene microarray plates were obtained
from Genetix (New Milton, Hampshire). Goat anti-rat cross-adsorbed albumin
antibody, reference serum, and HGF ELISA Quantitation Kit were obtained from
Bethyl Laboratories (Montgomery, TX). Goat anti-rat IgG Texas Red conjugate was
purchased from Santa Cruz Biotechnologies, Inc. Formalin was purchased from
Fisher (Pittsburgh, PA). Slide-A-Lyzer Mini Dialysis Units were purchased from
Pierce (Rockford, IL).
2.2. Preparation of glass substrates
Glass slides were cleaned by immersion in ‘‘piranha’’ solution consisting of 3:1
ratio of aqueous solutions of 50% v/v of sulfuric acid and 30% w/v of hydrogen
peroxide for 30 min (caution: this mixture reacts violently with organic materials and
must be handled with extreme care). The glass slides were thoroughly rinsed with
deionized water, dried under nitrogen, and kept in Class 10000 air prior to use. For
silane modification, the glass slides were exposed to oxygen plasma for 5 min at
300 W (YES3, Yield Engineering Systems, Livermore, CA) and then placed for 10 min
in a 2 mM solution of (3-acrylopropyl)trichlorosilane diluted in anhydrous toluene.
The reaction was performed in a glove box under a nitrogen blanket to avoid
exposure to atmospheric moisture. After silanization, the slides were rinsed with
fresh toluene, dried under nitrogen, and cured at 100?C for 4 h.
2.3. Printing ECM and growth factor microarrays
All ECM proteins employed for printing were dissolved in 1? PBSþ0.005%
Tween-20 at 0.2 mg/mL concentration. HGF was mixed with the ECM solution to
a final concentration of 500 ng/mL and allowed to bind to the ECM protein for
30 min at room temperature prior to printing. Protein microarrays were contact-
printed under ambient conditions on silane-modified 75 ?25 mm2glass slides
ECMþ500 ng/mL HGF in 1? PBS with 0.005% Tween) from a 382-well plate,
dispensing 20–70 nL of solution onto the glass slide and forming circular spots
w500 mm in diameter. Protein arrays were kept in a refrigerator before use and were
functional for at least one week.
2.4. Characterization of HGF retention on ECM microarrays
An array of ECM/HGF (6? 6) was printed onto silane-modified glass slides as
described before and incubated in 1? PBS at 37?C for 2 h. HGF molecules were
extracted from the silane surface using 50 mL of 4 M guanidine–HCl (pH 7.2) sup-
sulfonate (CHAPS), 10 mM EDTA, 0.05 M Tris and protease inhibitor cocktail. The
guanidine solution was incubated with the HGF microarray for 30 min. The super-
natant was buffer-exchanged with 6 M urea containing 0.05 M Tris (pH 7.4) using
Slide-A-Lyzer Mini Dialysis Units with 3500 MWCO for 1 h. The concentration of
HGF extracted from the printed microarray was determined according to the
manufacturer’s instructions using the HGF ELISA.
Immunofluorescent staining was used to determine retention of HGF on ECM
microarrays after 1, 3 and 5 days in media at cell culture conditions. Protein solution
containing 500 ng/mL HGF and 0.2 mg/mL collagen (I) was printed onto silanized
glass slides and in cell culture media at 37?C. At the desired time point, glass
substrates were removed from media and incubated with 1 mg/ml (in 1? PBS) of
anti-human HGF biotin conjugate at 37?for 2 h followed by incubation in 10 mg/ml
of streptavidin, Alexa Fluor?546 conjugate for 1 h at room temperature. Samples
were washed between each staining step with 1? PBS þ0.05% Tween-20. In order to
create a quantitative readout of fluorescence signal emanating from the array, the
laser microarray scanner (Agilent G2565BA fluorescent scanner, Expression Analysis
Facility, UC Davis Genome Center) was employed to scan the glass slides at a spot
pixel resolution of 5 mm. The fluorescence intensity of each array element was
determined using GenePix Pro 6.0 data analysis software (Molecular Devices,
Downingtown, PA). Fluorescent intensity was converted to a percentage of the
2.5. Cultivation of primary hepatocytes on HGF microarrays
Our studies employed primary rat hepatocytes. Cells were isolated from adult
female Lewis rats (Charles River Laboratories, Boston, MA) weighing 125–200 g,
using a two-step collagenase perfusion procedure as described previously .
Typically, 100–200 million hepatocytes were obtained with viability >90% as
determined by trypan blue exclusion. Primary hepatocytes were maintained in
DMEM supplemented with epidermal growth factor (EGF), glucagon, hydrocorti-
sone sodium succinate, recombinant human insulin, 200 units/mL penicillin,
200 mg/mL streptomycin and 10% FBS.
a concentration of 1?106cells/mL. After 1 h of incubation at 37?C, hepatocytes
becamelocalizedonECM/HGF domains,but didnot attachonthesurroundingsilane-
modified surface. The samples were then washed twice in PBS to remove unbound
hepatocytes and fresh media was added to the sample well.
Murine 3T3 fibroblasts were maintained in DMEM supplemented with 10% FBS,
200 units/mL penicillin, and 200 mg/mL streptomycin at 37?C in a humidified 5% CO2
atmosphere. Cells were cultured until 90% confluence and then passaged. When
performing co-culture experiments, hepatocytes were allowed to spread out on the
protein spots overnight. The following day 3T3 fibroblasts were seeded on the
sample at 0.25?106cells/mL and were allowed to attach for 30 min. Unbound cells
were washed away and fresh media was added as previously described. In our
previous experiments [34,35] modification of glass substrates with acrylated silane
was found to render these surfaces partially non-fouling. The silane layer prevented
attachment of primary hepatocytes seeded first but supported adhesion of 3T3
fibroblasts that were seeded in the second step. Therefore, this micropatterning
strategy led to hepatocytes residing on protein islands and fibroblasts adhering on
the surrounding glass regions.
In our experiments, hepatocytes cultured on HGF/ECM microarrays were
compared with cells cultured on ECM protein arrays without HGF. Another control
experiment performed in parallel involved cultivation of hepatocytes on ECM
protein arrays with HGF present in solution at a concentration of 10 ng/mL.
Importantly, in contrast to HGF microarrays where no exchange or supplementation
of GF molecules was possible, soluble GF was changed daily along with the culture
media. In all three cases hepatic function was analyzed as described below.
2.6. Analysis of hepatic function
Expression of hepatic phenotype was assessed by intracellular staining of
albumin, ELISAof albumin and real-time RT-PCR of albumin and a1-antitrypsin gene
expression. For immunostaining, cells were fixed in 4% formalin in PBS for 20 min
and then permeabilized with 0.1% Triton X-100 for 5 min. The cells were then
incubated in blocking solution (1% bovine serum albumin (BSA) in 1X PBS) for 1 h at
room temperature and exposed to 1:250 diluted anti-rat serum albumin antibody
for 2 h at 37?C. Finally, cells were incubated in 1:100 diluted anti-mouse IgG
conjugated with Texas Red for visualization. Cells were washed between each step
with 1X PBS three times for 5 min. All incubations were performed at room
temperature if not specified. Stained cells were visualized and imaged using
C.N. Jones et al. / Biomaterials 30 (2009) 3733–37413734
a confocal microscope (Zeiss LSM Pascal). Cell culture media was collected everyday
and analyzed for secreted albumin content using standard protocols described
previously . Albumin concentration was estimated using a standard kit from
For real-time RT-PCR experiments, cells were collected from microarrays using
trypsin for 10 min at 37?. Extracted cells were stabilized in 100 mL of lysis buffer and
stored at ?20?C. Total RNA was extracted from the cell lysates using Absolute total
mRNA isolation microprep kit (Stratagene) according to the manufacturer’s
instructions. cDNA was synthesized using QuantiTect Reverse Transcription kit
(Qiagen) according to the manufacturer’s instructions using 12 mL of DNAse pre-
treated total mRNA. Quantitative real-time PCR was performed using SYBR Green
PCR Master Mix (Applied Biosystems). Primers for rat albumin, a1-antitrypsin and
GAPDH genes were selected from a database http://medgen.ugent.be/rtprimerdb.
Primer (Sigma Genosys) concentrations were optimized before use. SYBR Green
Master Mix (1?) was used with 1 mM of forward and reverse primers in a total
volume of 12 ml that also included 1 ml cDNA. All PCR reactions were done in
duplicate. PCR amplificationwas performed as follows: 95?C for 10 min, 40 cycles of
95?C for 15 s, 60?C for 10 s and 68?C for 1 min on Mastercycler Realplex (Eppen-
dorf). The comparative Ctvalue method, using housekeeping gene (GAPDH) as an
internal standard, was employed to determine relative levels of albumin and a1-AT
2.7. Characterization of cell morphology
Cell morphology was observed daily via brightfield microscopy. To document
changes in cellular conformation at higher magnification, images were obtained
using a Philips XL 30 scanning electron microscope (SEM) at 10 kV beamvoltage and
a tilt angle of 20?. Inpreparationfor SEM characterization, the cellular micropatterns
were washed in fresh media followed by 2? wash in 50% 0.2 M sodium phosphate
buffer. The patterns were then fixed in 2% glutaraldehyde dissolved in 0.2 M sodium
phosphate buffer for 15 min followed by 3? wash in the buffer solution. The cells
were then dehydrated by incubation for 10 min in 30%, 50%, and 70% ethanol
solutions. The cells were then washed three times in 95% ethanol for 5 min each and
finally two times in 100% ethanol for 5 min. In a final dehydration step, ethanol was
replaced with HMDS and heated to 60?C in order to evaporate the HMDS. The
dehydrated samples were coated with 6 nm layer of Au–Pd using a sputter coater
(Pelco SC7, Ted Pella, Inc., Redding, CA).
3. Results and discussion
In this study, primary rat hepatocytes were cultured on printed
arrays of three liver-related ECM proteins (collagen (I), (IV) and
laminin) with or without HGF. Solid-phase presentation of HGF on
ECM microarrays significantly stimulated transcription and trans-
lation of liver-specific proteins in hepatocytes and induced spon-
taneous formation of hepatic spheroids after 5 days in culture.
These data are significant as they demonstrate that one-time
presentation of HGF has a pronounced and long-lasting effect on
maintenance of differentiated hepatic phenotype in vitro.
3.1. Printing and characterization of ECM/HGF microarray
In a typical experiment, GF/ECM microarrays were fabricated
using a hand-held contact arrayer printing 500 mm diameter
protein spots. Prior to printing glass substrates were treated with
acrylated silane. The silanizationprocedure mayhave enhanced the
quality of the printed microarray spots by rendering the surface
more hydrophobic, with the water contact angle increasing from
near zero (after O2 plasma treatment) to 53 ?2?after silane
modification. More importantly, the silane layer did not support
attachment of primary hepatocytes and could be used to localize
these cells to printed islands of ECM proteins. In addition as
described in the following sections of this paper, silane coating
permitted fibroblast attachment and therefore allowed us to
co-cultures on protein microarrays.
When creating GF microarrays, a solution of HGF (concentration
500 ng/ml) was mixed with either collagen (I) or collagen (IV) or
silanized glass slide. A question central to the success of the
proposed strategy of solid-phase presentation of HGF was the
Scenario 1: daily addition of HGF to culture media
Scenario 2: no HGF, just culture media
Scenario 3: solid-phase presentation of HGF
Fig. 1. Schematic representation of cell microarray experiments. Primary rat hepato-
cytes were cultured on protein microarrays. Three scenarios were compared: hepa-
tocytes on ECM protein spots with HGF added in solution (scenario 1), hepatocytes on
ECM spots without HGF (scenario 2) and hepatocytes on ECM spots with surface-
immobilized HGF (scenario 3). Immobilization of HGF on ECM protein spots had
a profound and long-lasting effect on morphology and phenotype of hepatocytes.
HGF concentration [pg/mm2]
Col(I)/HGF Col(IV)/HGFLn/HGF Col(I) w/o HGF
HGF Retention on Arrays (%)
Day 1Day 3Day 5
Fig. 2. Retention of HGF on ECM protein spots. HGF was mixed with one of three ECM
proteins (collagen (I), (IV) or laminin) and then printed onto silane-modified glass
substrates to create HGF/ECM protein spots (500 mm diameter). (A) The microarrays
were then exposed to 4 M guanidine–HCl extraction buffer to remove surface-bound
proteins. Presence of HGF in the extract was analyzed using ELISA. (B) HGF/col (I)
microarrays were analyzed using immunofluorescent staining after incubation in
culture media. For each time HGF spot replicates is n ¼20.
C.N. Jones et al. / Biomaterials 30 (2009) 3733–37413735
retention of this morphogen on printed ECM protein spots. To
answer this question we developed a protocol whereby arrays of
HGF co-printed with a chosen ECM protein onto a glass substrate
using guanidine-based protein extraction method. HGF extracted
from the surface along with carrier matrix proteins was then
analyzedusing ELISA.The analysisof HGF bindingonmicroarrays of
different ECM proteins showed that collagen (I) spots contained
0.8 pg of HGF/spot, whereas collagen IV and laminin contained
1.5 pg/spot, and 1.9 pg/spot respectively (Fig. 2A). These results
the amount of dispensed liquid per spot to be 20 nL (obtained from
manufacturer) the amount of HGF present on the spot assuming
100% retention will be w10 pg/spot. Therefore, ECM protein
microarrays retained w10–20% of HGF molecules. These results are
consistent with other studies that reported HGF to have moderate
affinity for ECM (Kd approximately 10?9mol/L) and demonstrated
that 15–20% of125I-labeled HGF bound to ECM proteins that had
been immobilized on polystyrene microtiter wells .
In addition to the amount of HGF captured on the spots
immediately after printing, we investigated retention of GF mole-
cules on the microarrays over the course of five days of incubation
in cell culture media at 37?C. The presence of HGF was revealed by
incubating microarrays with biotinylated anti-HGF antibodies and
fluorescently-labeled avidin. Immunofluorescence of the micro-
arrays was quantified with a laser scanner as described by us
elsewhere  and normalized by the fluorescence signal from the
initial time point. Fig. 2B shows that the HGF signal decreased as
Fig. 3. Primary hepatocytes cultured on HGF/ECM microarrays. (A–B) Hepatocytes on collagen (I) spots without HGF after 5 days in culture. (C) SEM micrograph of hepatocyte
cluster formed on HGF/collagen (I) spot after 4 days in culture. The edges of the spot are starting to roll up. (D) Hepatocytes cultivated on HGF/collagen for 5 days reorganized from
a monolayer culture into spheroids. (E) An array of hepatocyte spheroids formed on HGF/collagen (I). During this reorganization a 500 mm diameter hepatocyte cluster compacted
and rolled up into w100 mm diameter 3D spheroids. In the upper row of cell clusters one can see the boundary of 500 mm diameter cluster with a few remaining cells after 4 days in
culture. (F) Brightfield image shows the presence of typical hepatocytes with large nuclei and prominent cell borders in the center of the spheroid after 4 days in culture.
C.N. Jones et al. / Biomaterials 30 (2009) 3733–37413736
a function of time in culture, however, w50% of HGF molecules
were retained on col (I) microarrays after five days in culture. These
data further support the notion that HGF may be present on the
surface of ECM microarrays over extended periods of time to
stimulate adherent hepatocytes.
2 pg of HGF. When adding HGF in soluble form, 3 ml of 10 ng/ml
solution are added into a well of surface area of 9.6 cm2. Assuming
cell clusters are present in a well, individual cell cluster receives
w6 pg of HGF on daily basis. While this is a rough estimate, it shows
likely similar (same order of magnitude). In addition to interesting
bound HGF, solid-phase immobilization of this morphogen was
advantageous with regards to reagent conservation. When adding
HGF in soluble form, 30 ng/day of reagent was used per well of a 6-
well plate during a 10 day experiment, while the solid-phase
presentation required one-time use of 1.44 ng per well. Therefore,
3.2. Cultivation of hepatocytes on HGF microarrays
To test the ability of printed HGF to promote hepatic differen-
tiation, GF/ECM spots were printed onto silane-modified glass
substrates. When seeded on these glass substrates, hepatocytes
selectively attached on printed arrays forming w500 mm diameter
clusters containing w250 hepatocytes/spot. As shown in Fig. 3(A,B),
cell organization on protein arrays occurred with high fidelity and
minimal non-specific attachment of hepatocytes on silanized glass
regions. Cell number was consistent on arrays with and without
HGF (w280 cells/spot) with the exception of collagen (IV) based
microarrays where 110 cells/spot were observed.
While hepatocytes cultured on ECM arrays behaved unremark-
ably, de-differentiating rapidly (within days), hepatocytes residing
on HGF/ECM spots underwent dramatic changes in morphology,
spontaneously forming three-dimensional spheroids after 4–5 days
in culture. As highlighted in Fig. 3C, hepatocyte clusters started to
‘‘roll-up’’ after 4 days of culture on HGF/ECM spots so that by day 5
at least 75% of the clusters were in a spheroid formation (Fig. 3D).
Looking at an image of an array of hepatocyte clusters (Fig. 3E) one
can see the boundaries of 500 mm diameter spots occupied by
hepatocytes prior to spheroid formation. Interestingly, while the
edges of the spot rolled up, prototypical hepatocytes with prom-
inent nuclei and cuboidal morphology were observed in the center
of the spheroid (Fig. 3F). Organization of hepatocyte clusters into
three-dimensional constructs was not observed when cells were
cultured on the three ECM proteins tested (collagen (I), (IV), lam-
inin) without HGF. Similarly, no changes in morphology of the
hepatocyte spots were observed when cells cultured on ECM
protein spots were supplemented with soluble HGF at 10 ng/ml
concentration. Conversely, when cultured on HGF containing spots,
hepatocytes formed spheroids on all three types of ECM proteins at
approximately the same time – 5 days of culture. This leads us to
conclude that surface-immobilized HGF provided the signals to
induce changes in cell morphology. HGF is a known mitogenic
factor and DNA synthesis within the hepatocyte spheroids was
investigated using BrdU assay. However, DNA synthesis was not
observed suggesting that HGF signalling did not induce prolifera-
tion in hepatocytes – not surprising given their terminally differ-
entiated state. It is therefore more likely that reorganization of
hepatocytes into three-dimensional constructs was connected to
motogenic properties of HGF and involved cytoskeleton and cell–
cell adhesion molecules. Similar effects of HGF on hepatocellular
reorganization without proliferation have recently been reported
by Hoshiba et al. . A number of recent studies suggested that
hepatocytes cultured in 3D configuration such as a spheroid are
more functional compared to standard monolayer cultures [38–42].
Therefore, the organization of hepatocytes into 3D spheroids
observed in Fig. 3 may also be connected to the enhancement of
ECM w/soluble HGF
ECM + printed HGF
Fig. 4. Albumin synthesis in hepatocytes cultured on protein microarrays. (A) ELISA
results for hepatocytes cultured on collagen (I) spots w/HGF, w/o HGF and with soluble
HGF. HGF/collagen (I) spots induce significantly higher levels of albumin production
compared to collagen (I) spots. (N ¼3 samples, p-value < 0.005). (B–C) Albumin ELISA
analysis of primary rat hepatocytes cultured on collagen (IV) (B) and laminin (C)
suggests that solid-phase presentation of HGF significantly upregulates albumin
production. (N ¼ 3 samples, p-value <0.005).
C.N. Jones et al. / Biomaterials 30 (2009) 3733–3741 3737
hepatic function induced by HGF microarrays and discussed in the
It should be noted that spheroid formation comparable to that
described in Fig. 3 also occurred when hepatocytes werestimulated
with HGF on larger, w3 mm diameter, ECM spots (data not shown).
However, in this latter case, spheroids were distributed on the
surface randomly and not in a periodic pattern as shown in Fig. 3E.
3.2.1. Functional analysis of hepatocytes cultivated on HGF arrays
In order to assess function of hepatocytes cultured on HGF
microarrays we monitored production of albumin – a liver-specific
serum protein. Primary rat hepatocytes werecultured foraweek on
three ECM proteins – collagen I, collagen IV, and laminin – with and
without HGF. An additional control performed in our studies,
involved cultivation of hepatocytes on ECM microarrays with HGF
added into culture media at concentration of 10 ng/ml. Thus we
compared hepatic function under three scenarios: ECM micro-
arrays alone, ECM microarraysþsoluble HGFand ECM microarrays/
solid-phase presented HGF (see Fig. 1 for pictorial illustration).
Fig. 4 compiles the results of this study. As shown in Fig. 4A,
hepatocytes cultivated on collagen (I) arrays with and without HGF
produced comparable levels of albumin at the beginning of the
experiment (Day 2); however, albumin production by hepatocytes
on collagen (I) arrays decreased rapidly pointing to de-differenti-
ation of these cells. This was in starkcontrast tothe HGF-stimulated
hepatocytes that exhibited high levels of albumin synthesis (5.5-
fold higher than collagen (I) controls). We also tested two other
ECM proteins found in the liver – collagen IV (Fig. 4B) and laminin
(Fig. 4C) – and found a similar trend in upregulation of albumin
production in HGF-stimulated cells.
Relative Expression Level
Col (I) +
Relative Expression Level
Col (I) +
Fig. 5. Immunostaining and RT-PCR. (A) Intracellular albumin staining of hepatocytes after 4 days of cultivation on collagen (I) spots. (B) Intracellular albumin staining of hepa-
tocytes after 4 days of cultivation on HGF/collagen (I) spots reveals significantly higher level of albumin synthesis compared to (A). (C–D) Higher magnification (63?) fluorescence
(D) and merged brightfield/fluorescence image (63?) of intracellular albumin staining after 4 days of culture on HGF/collagen (I) shows localization of albumin to cytoplasm of
hepatocytes. (E–F) Expression level of two liver-specific genes (albumin and a-1 antitrypsin) relative to b-actin housekeeping gene.
C.N. Jones et al. / Biomaterials 30 (2009) 3733–37413738
The function of hepatocytes in contact with solid-phase pre-
sented HGF was also compared to daily additions of soluble HGF
into culture media. The concentration of soluble GF was chosen
based on the conditions typically used in conjunction with this GF
molecule [43,44]. The direct comparison of soluble vs. solid-phase
GF presentation is difficult to accomplish because it is unclear what
fraction of soluble and matrix-bound GF molecules stimulates the
hepatocytes in each of these two scenarios. Nevertheless, albumin
ELISA results pointed to similar levels of protein synthesis in
hepatocytes cultured on HGF/collagen (I) and HGF/laminin arrays
compared to cells cultured on the same ECM proteins but with
soluble HGF. In the case of collagen (IV) spots, a 4-fold higher
albumin production was observed on solid-phase presented HGF
compared to soluble HGF. In addition, we compared function of
hepatocytes cultured on 500 mm and 3 mm diameter HGF/ECM
spots and found hepatic albumin production to be similarly
elevated in both cases (data not shown). Our results are consistent
with previous studies by Bhatia and colleagues where function of
hepatocytes was shown to be independent of micropattern
dimension for spots larger than 500 mm in diameter .
Intracellular immunostaining of albumin also showed a signif-
icantly stronger signal on printed HGF/collagen I spots than on
collagen I control spots (Fig. 5A,B) thereby corroborating ELISA
data. A higher resolution image (63?) of albumin immunostaining
shows specific cytoplasmic staining of albumin (Fig. 5C,D). The
ELISA and immunostaining results were also corroborated by RT-
PCR data by clearly pointing to a 20-fold upregulation in albumin
gene expression in the case of printed HGF (high) and soluble HGF
(Fig. 5E). RT-PCR analysis of another liver-specific protein a1-
antitrypsin revealed a 23-fold higher level of expression of this
gene in hepatocytes cultured on HGF/collagen (I) compared to
hepatocytes on collagen (I) spots (control experiment) (see
Hepatic function results presented in this section are significant
as they demonstrate that surface immobilization of HGF at the
beginning of the experiment is at least as effective in maintaining
differentiated hepatic phenotype as daily addition of soluble HGF
into culture media. Beyond considerations of the cost of reagent,
which is conserved by using HGF microarrays, our results suggest
that some surface-bound HGF molecules may remain potent and
active for several days under cell culture conditions, inducing
lasting changes in hepatic phenotype expression.
3.2.2. Effects of printed HGF on hepatocytes in co-cultures
A number of studies have shown that co-cultivation of
hepatocytes with non-parenchymal cells leads to the enhance-
ment and maintenance of differentiated hepatic phenotype [46–
48]. The co-culture format was further enhanced by Bhatia and
co-workers who incorporated micropatterning strategies and
showed placement of different cell types into precise locations
on the same surface. In the subsequent studies, micropatterned
co-cultures were used to dissect the effects of homotypic and
heterotypic interactions between the hepatocytes and non-
parenchymal cells . In our approach, hepatocytes were
seeded on the silane-modified glass substrates containing prin-
ted microarrays and became adherent exclusively on the protein
islands (500 mm diameter) (Fig. 6A). The non-parenchymal cells
(3T3 fibroblasts) were added in the subsequent cell seeding step,
attaching on the silane-modified glass substrate around the
hepatic islands (Fig. 6B). This micropatterning strategy allowed
us to selectively stimulate one cell type (hepatocytes) in a co-
culture format and query the effects of HGF on the hepatocytes
cultured in the presence of growth factor-producing non-
parenchymal cells. As illustrated in Fig. 6C, solid-phase presen-
tation of HGF-induced formation of hepatic spheroids even in
the context of a co-culture where hepatocytes residing on HGF/
ECM islands (500 mm diameter) were in contact with the
surrounding fibroblasts. Based on the previous reports, hepato-
cytes co-cultivated with 3T3 fibroblasts were expected to
express higher function compared to hepatocytes cultured alone
. This trend was corroborated by our albumin ELISA experi-
ments (Fig. 6D). Interestingly, hepatocyte–fibroblast co-cultures
250 µm250 µm250 µm
Monoculture on col(I)
Co-culture on col(I)
Co-culture on col(I)/HGF
Fig. 6. Hepatocyte–fibroblast co-cultures created on HGF microarrays. (A) Monoculture; Col I control (B) Co-culture; Col IþSoluble HGF (C) Co-culture; Col Iþ printed HGF (D)
ELISA data show enhanced maintenance in albumin secretion in ECM-bound as well as soluble HGF compared to ECM control w/o growth factor even in the presence of fibroblast
C.N. Jones et al. / Biomaterials 30 (2009) 3733–3741 3739
created on HGF/collagen (I) arrays exhibited significantly higher
levels (1.8-fold) of albumin production compared to co-cultures
with hepatocytes sitting on collagen (I) arrays. These data are
interesting as they demonstrate that solid-phase presentation of
HGF had an observable and long-lasting effect even in the
presence of fibroblasts – non-parenchymal cells that are likely
secreting endogenous growth factors.
While there have been reports of surface-immobilized GF
molecules stimulating cell function, [18,28,33] we are not aware of
studies involving primary hepatocytes. These cells de-differentiate
rapidly in vitro and require special cultivation approaches such as
collagen double gel [49,50] or co-cultures  to rescue hepatic
function and to ensure long-term maintenance. Our results are
intriguing as they suggest that one-time presentation of matrix-
bound HGF at the beginning of the experiment without additional
supplementation is sufficient to enhance and maintain phenotype
expression of primary hepatocytes after 10 days in culture. The
mechanism of the phenotype induction is unclear at the moment. It
has been suggested previously that GF interactions with ECM
proteins may protect morphogens against proteolytic degradation
, therefore, HGF may be stabilized by the association with
matrix proteins and may remain potent and active throughout the
experiment. In addition, HGF-induced organization of hepatocytes
into 3D spheroids may also contribute to the enhanced phenotype
The present paper investigated the use of HGF microarrays for
cultivation of primary hepatocytes. Analysis of hepatic phenotype
with ELISA, RT-PCR, and immunostaining techniques revealed that
solid-phase presentation of HGF induced significant enhancement
in synthesis and transcription of liver-specific proteins, albumin
and a1-antitrypsin. In addition, hepatocytes cultured on HGF
microarrays spontaneously reorganized from a monolayer to
spheroid configuration – a behavior that was not observed other-
wise. Solid-phase presentation of HGF was also seen to affect
morphology and phenotype of hepatocytes in co-cultures with
Overall, the strategy of using robotic printing to create GF
arrays has a number of advantages over traditional methods of
cell stimulation with GF molecules. Solid-phase presentation of
GF molecules associated with matrix proteins may be more
physiological. It also dramatically decreases the cost of per-
forming experiments and makes the microarray format particu-
larly well-suited for high-throughput
interactions where different combinations and/or concentrations
of GF molecules may be tested in a multiplexed and combina-
We thank Mr. Sunny Shah for the assistance with collection of
SEM micrographs. NT was supported by a fellowship from the
National Center for Biotechnology, Republic of Kazakhstan. Finan-
cial support for this work was provided by NIH grant (DK073901)
awarded to AR.
Figures with essential colour discrimination. Parts of Figs. 1,3
and 5 in this article may be difficult to interpret in black and white.
The full colour images can be found in the on-line version, at doi:
 Masumoto A, Yamamoto N. Stimulation of DNA synthesis in hepatocytes by
hepatocyte growth factor bound to extracellular matrix. Biochem Biophys Res
 Kato S, Ishii T, Hara H, Sugiura N, Kimata K, Akamatsu N. Hepatocyte growth
factor immobilized onto culture substrates through heparin and matrigel
enhances DNA synthesis in primary rat hepatocytes. Exp Cell Res 1994;
 Chan GK, Lutterbach BA, Pan BS, Kariv I, Szewczak AA. High-throughput
analysis of HGF-stimulated cell scattering. J Biomol Screen 2008;13(9):847–54.
 Gherardi E, Gray J, Stoker M, Perryman M, Furlong R. Purification of scatter
factor, a fibroblast-derived basic protein that modulates epithelial interactions
and movement. Proc Natl Acad Sci U S A 1989;86(15):5844–8.
 Snykers S, De Kock J, Vanhaecke T, Rogiers V. Differentiation of neonatal rat
epithelial cells from biliary origin into immature hepatic cells by sequential
exposure to hepatogenic cytokines and growth factors reflecting liver devel-
opment. Toxicol In Vitro 2007;21(7):1325–31.
 Schmidt C, Bladt F, Goedecke S, Brinkmann V, Zschiesche W, Sharpe M, et al.
Scatter factor/hepatocyte growth factor is essential for liver development.
 Matsumoto K, Nakamura T. Emerging multipotent aspects of hepatocyte
growth factor. J Biochem 1996;119(4):591–600.
 Breitkopf K, Haas S, Wiercinska E, Singer MV, Dooley S. Anti-TGF-beta strat-
egies for the treatment of chronic liver disease. Alcohol Clin Exp Res 2005;
 Ueki T, Kaneda Y, Tsutsui H, Nakanishi K, Sawa Y, Morishita R, et al. Hepatocyte
growth factor gene therapy of liver cirrhosis in rats. Nat Med 1999;5(2):
 Gouon-Evans V, Boussemart L, Gadue P, Nierhoff D, Koehler CI, Kubo A, et al.
BMP-4 is required for hepatic specification of mouse embryonic stem cell-
derived definitive endoderm. Nature Biotechnol 2006;24(11):1402–11.
 Duan Y, Catana A, Meng Y, He S, Gupta S, Gambhir SS, et al. Differentiation
and enrichment of hepatocyte-like cells from human embryonic stem cells
in vitro and in vivo using lentivirus systems. Stem Cells 2007;25(12):
 Shirahashi H, Wu J, Yamamoto N, Catana A, Wege H, Wager B, et al. Differ-
entiation of human and mouse embryonic stem cells along a hepatocyte
lineage. Cell Transplant 2004;13(3):197–211.
 D’Amour KA, Agulnick AD, Eliazer S, Kelly OG, Kroon E, Baetge EE. Efficient
differentiation of human embryonic stem cells to definitive endoderm. Nat
 Nishino M, Iimuro Y, Ueki T, Hirano T, Fujimoto J. Hepatocyte growth factor
improves survival after partial hepatectomy in cirrhotic rats suppressing
apoptosis of hepatocytes. Surgery 2008;144(3):374–84.
 Schuppan D, Schmid M, Somasundaram R, Ackermann R, Ruehl M,
Nakamura T, et al. Collagens in the liver extracellular matrix bind hepatocyte
growth factor. Gastroenterology 1998;114(1):139–52.
 Kim TH, Mars WM, Stolz DB, Petersen BE, Michalopoulos GK. Extracellular
matrix remodeling at the early stages of liver regeneration in the rat. Hep-
 Fan VH, Tamama K, Au A, Littrell R, Richardson LB, Wright JW, et al. Tethered
epidermal growth factor provides a survival advantage to mesenchymal stem
cells. Stem Cells 2007;25(5):1241–51.
 Alberti K, Davey RE, Onishi K, George S, Salchert K, Seib FP, et al. Functional
immobilization of signaling proteins enables control of stem cell fate. Nat
 Mann BK, Schmedlen RH, West JL. Tethered-TGF-beta increases extracellular
matrix production of vascular smooth muscle cells. Biomaterials 2001;
 Kuhl PR, Griffith-Cima LG. Tethered epidermal growth factor as a paradigm for
 Bashkin P, Doctrow S, Klagsbrun M, Svahn CM, Folkman J, Vlodavsky I. Basic
fibroblast growth factor binds to subendothelial extracellular matrix and is
 Ashikari-Hada S, Habuchi H, Kariya Y, Itoh N, Reddi AH, Kimata K.
Characterization of growth factor-binding structures in heparin/heparan
 Sampath TK, Muthukumaran N, Reddi AH. Isolation of osteogenin, an extra-
cellular matrix-associated, bone-inductive protein, by heparin affinity chro-
matography. Proc Natl Acad Sci U S A 1987;84(20):7109–13.
 Wissink MJ, Beernink R, Pieper JS, Poot AA, Engbers GH, Beugeling T, et al.
Binding and release of basic fibroblast growth factor from heparinized
collagen matrices. Biomaterials 2001;22(16):2291–9.
 Taipale J, Keski-Oja J. Growth factors in the extracellular matrix. FASEB J
 Macri L, Silverstein D, Clark RA. Growth factor binding to the pericellular
matrix and its importance in tissue engineering. Adv Drug Deliv Rev 2007;
 Wang X, Harris RE, Bayston LJ, Ashe HL. Type IV collagens regulate BMP sig-
nalling in Drosophila. Nature 2008;455(7209):72–7.
from thesolidphase.Nat Med
C.N. Jones et al. / Biomaterials 30 (2009) 3733–37413740
 Paralkar VM, Vukicevic S, Reddi AH. Transforming growth factor-beta binds Download full-text
collagen IV of basement membrane matrix: implications for development.
Dev Biol 1991;143(2):303–8.
 Schena M, Shalon D, Davis RW, Brown PO. Quantitative monitoring of gene
expression patterns with a complementary DNA microarray. Science 1995;
 Flaim CJ, Chien S, Bhatia SN. An extracellular matrix microarray for probing
cellular differentiation. Nat Methods 2005;2(2):119–25.
 Anderson DG, Levenberg S, Langer R. Nanoliter-scale synthesis of arrayed
biomaterials and application to human embryonic stem cells. Nat Biotechnol
 Ding S, Schultz PG. A role of chemistry in stem cell biology. Nat Biotechnol
 Soen Y, Mori A, Palmer TD, Brown PO. Exploring the regulation of human
neural precursor cell differentiation using arrays of signaling microenviron-
ments. Mol Syst Biol 2006;2(37):1–14.
 Revzin A, Rajagopalan P, Tilles A, Berthiaume F, Yarmush M, Toner M.
Designing a hepatocellular microenvironment with protein microarraying
 Lee JY, Jones CN, Zern MA, Revzin A. Analysis of local tissue-specific gene
expression in cellular micropatterns. Anal Chem 2006;78(24):8305–12.
 Jones CN, Lee JY, Stybayeva GS, Zern MA, Revzin A. Multifunctional protein
microarrays for cultivation of cells and immunodetection of secreted cellular
products. Anal Chem 2008;80(16):6351–7.
 Dunn JCY, Yarmush ML, Koebe HG, Tompkins RG. Hepatocyte function and
extracellular matrix geometry: long-term culture in a sandwich configuration.
FASEB J 1989;3(2):174–7.
 Hoshiba T, Wakejima M, Cho CS, Shiota G, Akaike T. Different regulation of
hepatocyte behaviors between natural extracellular matrices and synthetic
extracellular matrices by hepatocyte growth factor. J Biomed Mater Res A
 Chang TT, Hughes-Fulford M. Monolayer and spheroid culture of human liver
hepatocellular carcinoma cell line cells demonstrate distinct global gene
expression patterns and functional phenotypes. Tissue Eng Part A 2009;
 Nishikawa M, Yamamoto T, Kojima N, Kikuo K, Fujii T, Sakai Y. Stable immo-
bilization of rat hepatocytes as hemispheroids onto collagen-conjugated
poly-dimethylsiloxane (PDMS) surfaces: importance of direct oxygenation
through PDMS for both formation and function. Biotechnol Bioeng 2008;
 Curcio E, Salerno S, Barbieri G, De Bartolo L, Drioli E, Bader A. Mass transfer
and metabolic reactions in hepatocyte spheroids cultured in rotating wall gas-
permeable membrane system. Biomaterials 2007;28(36):5487–97.
 Nakazawa K, Lee SW, Fukuda J, Yang DH, Kunitake T. Hepatocyte spheroid
formation on a titanium dioxide gel surface and hepatocyte long-term culture.
J Mater Sci Mater Med 2006;17(4):359–64.
 Kimura M, Okamoto H, Ogihara M. Activation of mitogen-activated protein
kinase by hepatocyte growth factor is stimulated by both alpha1- and beta2-
adrenergic agonists in primary cultures of adult rat hepatocytes. J Pharm Sci
 Yang H, Magilnick N, Xia M, Lu SC. Effects of hepatocyte growth factor on
glutathione synthesis, growth, and apoptosis is cell density-dependent. Exp
Cell Res 2008;314(2):398–412.
 Bhatia SN, Balis UJ, Yarmush ML, Toner M. Effect of cell–cell interactions in
preservation of cellular phenotype: cocultivation of hepatocytes and non-
parenchymal cells. FASEB J 1999;13:1883–900.
 Corlu A, Ilyin G, Cariou S, Lamy I, Loyer P, Guguen-Guillouzo C. The coculture:
a system for studying the regulation of liver differentiation/proliferation and
its control. Cell Biol Toxicol 1997;13(4–5):235–42.
 Bhatia SN, Yarmush ML, Toner M. Controlling cell interactions by micro-
patterning in co-cultures: hepatocytes and 3T3 fibroblasts. J Biomed Mater Res
 Guguen-Guillouzo C, Clement B, Lescoat G, Glaise D, Guillouzo A. Modulation
of human fetal hepatocyte survival and differentiation by interactions with
a rat liver epithelial cell line. Dev Biol 1984;105(1):211–20.
 Moghe PV, Berthiaume F, Ezzell RM, Toner M, Tompkins RG, Yarmush ML.
Culture matrix configuration and composition in the maintenance of hepa-
tocyte polarity and function. Biomaterials 1996;17:373–86.
 Berthiaume F, Moghe PV, Toner M, Yarmush ML. Effect of extracellular matrix
topology on cell structure, function, and physiological responsiveness: hepa-
tocytes cultured in a sandwich configuration. FASEB J 1996;10(13):1471–84.
C.N. Jones et al. / Biomaterials 30 (2009) 3733–37413741