UCLA
UCLA Previously Published Works
Title
A liver-on-a-chip platform with bioprinted hepatic spheroids.
Permalink
https://escholarship.org/uc/item/9f546248
Journal
Biofabrication, 8(1)
ISSN
1758-5082
Authors
Bhise, Nupura S
Manoharan, Vijayan
Massa, Solange
et al.
Publication Date
2016
DOI
10.1088/1758-5090/8/1/014101
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California
This content has been downloaded from IOPscience. Please scroll down to see the full text.
Download details:
IP Address: 129.10.9.42
This content was downloaded on 06/06/2016 at 21:34
Please note that terms and conditions apply.
A liver-on-a-chip platform with bioprinted hepatic spheroids
View the table of contents for this issue, or go to the journal homepage for more
Home Search Collections Journals About Contact us My IOPscience
Biofabrication 8 (2016) 014101 doi:10.1088/1758-5090/8/1/014101
PAPER
A liver-on-a-chip platform with bioprinted hepatic spheroids
Nupura S Bhise
1,2
, Vijayan Manoharan
1,2
, Solange Massa
1,2,3
, Ali Tamayol
1,2
, Masoumeh Ghaderi
1,2
,
Mario Miscuglio
1,2
, Qi Lang
1,2
, Yu Shrike Zhang
1,2
, Su Ryon Shin
1,2,4
, Giovanni Calzone
1,2
, Nasim Annabi
1,2,4
,
Thomas D Shupe
5
, Colin E Bishop
5
, Anthony Atala
5
, Mehmet R Dokmeci
1,2
and Ali Khademhosseini
1,2,4,6
1
Biomaterials Innovation Research Center, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston,
MA 02115, USA
2
Harvard-MIT Division of Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
3
Centro de Investigación Biomédica, Universidad de los Andes, Santiago 12445, Chile
4
Wyss Institute for Biologically Inspired Engineering, Harvard University, Boston, MA 02115, USA
5
Wake Forest Institute for Regenerative Medicine, Wake Forest University School of Medicine, Winston-Salem, NC 27101, USA
6
Department of Physics, King Abdulaziz University, Jeddah 21569, Saudi Arabia
E-mail: alik@rics.bwh.harvard.edu
Keywords: bioprinting, bioreactor, liver, drug toxicity, 3D culture, hepatocytes
Abstract
The inadequacy of animal models in correctly predicting drug and biothreat agent toxicity in humans
has resulted in a pressing need for in vitro models that can recreate the in vivo scenario. One of the
most important organs in the assessment of drug toxicity is liver. Here, we report the development of a
liver-on-a-chip platform for long-term culture of three-dimensional (3D) human HepG2/C3A
spheroids for drug toxicity assessment. The bioreactor design allowed for in situ monitoring of the
culture environment by enabling direct access to the hepatic construct during the experiment without
compromising the platform operation. The engineered bioreactor could be interfaced with a
bioprinter to fabricate 3D hepatic constructs of spheroids encapsulated within photocrosslinkable
gelatin methacryloyl (GelMA) hydrogel. The engineered hepatic construct remained functional
during the 30 days culture period as assessed by monitoring the secretion rates of albumin, alpha-1
antitrypsin, transferrin, and ceruloplasmin, as well as immunostaining for the hepatocyte markers,
cytokeratin 18, MRP2 bile canalicular protein and tight junction protein ZO-1. Treatment with
15 mM acetaminophen induced a toxic response in the hepatic construct that was similar to published
studies on animal and other in vitro models, thus providing a proof-of-concept demonstration of the
utility of this liver-on-a-chip platform for toxicity assessment.
Introduction
The current drug development process suffers from a
high failure rate during clinical trials, which is partly
due to inadequacy of animal models to accurately
predict the toxic side-effects of drugs in humans [1, 2].
There is a need to develop better predictive platforms
that can complement the existing approaches for drug
discovery based on preclinical animal models followed
by clinical trials [3, 4]. Engineered tissues are being
increasingly used to model diseases and for in vitro
drug testing [5–8]. Advances in the areas of biomater-
ials and tissue engineering have led to exciting devel-
opments towards the goal of generating tissue
constructs from human cells that can mimic the
endogenous cell–cell and cell–matrix interactions [6].
Moreover, the advances in microfluidic systems allow
the fabrication of microbioreactors for growing these
engineered tissues under continuous perfusion, and
providing cells with various physiological stimuli as
well as supplying cells with nutrients and oxygen
[9, 10]. Consequently, the concept of organ-on-a-chip
and body-on-a-chip has emerged as promising plat-
forms that can be used for studying the effect of drugs
and environmental toxins on the targeted and sur-
rounding tissues [11–15]. Organ-on-a-chip platforms
will also provide potential avenues for screening
biothreat and chemical warfare agents [3].
Liver plays a critical role in drug metabolism and
detoxification of blood. Drug-induced hepatotoxicity
is one of the main reasons for drug withdrawal, thus
highlighting the need for developing a robust in vitro
RECEIVED
21 February 2015
REVISED
20 August 2015
ACCEPTED FOR PUBLICATION
28 August 2015
PUBLISHED
12 January 2016
© 2016 IOP Publishing Ltd
model for evaluating hepatotoxity [16, 17]. Recent
advancements in microfabrication technologies have
enabled the creation of architectures that can mimic
in vivo conditions. For instance, Ho et al micro-
fabricated centimeter scale morphology of liver
lobules and controlled the distribution of hepatocytes
and endothelial cells in two-dimensional (2D) culture.
They monitored the cellular activity for only 2 days
[18]. However, hepatic models that utilize conven-
tional 2D mono- and co-cultures of normal and
diseased human cells fail to mimic the complex three-
dimensional ( 3D) in vivo microenvironment and
extracellular matrix (ECM) support [16, 19, 20]. Thus,
it can prevent normal cellular function and response
to the tested drugs. To address this challenge, efforts
are focused on developing 3D models that can better
recapitulate in vivo cell–cell, cell-ECM interactions,
and the tissue architecture. Ex vivo studies using 3D
tissue biopsies or slices have been reported, but they
may be difficult to be used for high-throughput studies
and show rapid loss of functionality within a few days
in vitro, thus limiting their application for long-term
studies [21]. Part of recent approaches have thus
focused on encapsulation of cells within hydrogels
such as gelatin, poly(ethylene glycol) diacrylate
(PEGDA), and agarose, which mimic the physico-
chemical characteristics of native ECM [22–24].
Tsang et al encapsulated primary rat hepatocytes
within 3D PEGDA constructs doped with RGD and
cultured them in a flow bioreactor as a liver model for
12 days [24]. Li et al achieved the maintenance of liver-
specific functionality of rat hepatocyte microtissues
embedded wihtin PEGDA for about 50 days, but this
study was performed in static culture conditions [25
].
Moreover, the use of rat hepatocytes in these studies
limited their ability to predict human clinical out-
comes. Static cultures that are traditionally used for
drug testing lack the physiological cues important for
drug metabolism, which has been highlighted in mul-
tiple reports [13–15, 26]. Thus, the utilization of phy-
siologically relevant dynamic models, which provide
appropriate mass-transport for gas and metabolite
waste, becomes imperative for drug toxicity studies
[14, 15, 27]. In a study by He et al, HepG2 human
hepatocarcinoma cells were encapsulated within agar-
ose hydrogel containing planar perfusable channels
and continuosly perfused with unidirectional flow to
form a pre-vascularized liver-like structure [23]. How-
ever, they only monitored cellular viability for a 3 day
culture period and did not assess liver-specific func-
tionality. Interestingly, Esch et al reported that
bi-directional fluidic flow enhanced the metabolic
activity of multi-cellular 3D hepatic tissue cultured for
14 days as observed with unidirectional flow [27].
Another approach recently reported by Liu et al was to
apply magnetic field for controlled fabrication of 3D
microtissues [28]. Although this approach has shown
promising results for fabrication of simple 3D con-
structs, the fabrication of more complex biomimetic
constructs might be challenging. Limitations with
these approaches highlight the need for developing 3D
hydrogel-based human liver constructs in micro-
fluidic-based bioreactors amenable for long-term
functionality and drug toxicology studies. These plat-
forms reduce the consumption of expensive reagents,
allow for temporal control of drug dose, and provide
scalability and reproducibility [29].
Bioprinting approaches facilitate automated and
high-throughput fabrication of precisely controlled
3D architectures [30, 31]. Thus, combining them with
miniaturized bioreactors allows the realization of next
generation organ-on-a-chip platforms. Moreover, a
bioreactor that allows easy access to the hepatic con-
struct within the bioreactor without compromising
the construct is highly desirable for assessing tissue
behavior at multiple time points during the long-term
culture period. Among various 3D cell culture models,
multicellular spheroids, formed by aggregation of cells
compacted together by self-secreted ECM, have
recently been the focus of a number of studies [32
–37].
The oxygen tension in the core of the spheroids con-
trasts with that in the periphery, better mimicking the
in vivo oxygen gradient in the hepatic lobule [38]. As a
result, the use of cell spheroids instead of disperse cells
might improve the cellular functionality and response
for drug toxicity assessment. Moreover, the encapsula-
tion of cell spheroids within hydrogels that are compa-
tible with rapid fabrication techniques including
bioprinting, enables the fabrication of complex archi-
tectures similar to those observed in vivo.
In this study, we fabricated a perfusable bioreactor
that allowed for direct access to cells during the long-
term culture period. We interfaced the designed bior-
eactor with a direct write printer to create HepG2/
C3A hepatic spheroid-laden hydrogel constructs. The
bioprinted tissue-like constructs were incubated
under continuous perfusion, and cellular functionality
was assessed over a period of 4 weeks. We also assessed
cellular response to acute acetaminophen (APAP)
exposure in this bioreactor platform as a model for
predictive drug toxicity comparable with in vivo
conditions.
Materials and methods
Materials
2-Hydroxy-1-(4-(hydroxyethoxy)phenyl)-2-methyl-
1-propanone (Irgacure 2959, CIBA Chemicals) was
used as photoinitiator (PI) for crosslinking photocros-
slinkable gelatin (GelMA ) hydrogel. Dulbecco’s mod-
ified Eagle medium (DMEM), fetal bovine serum
(FBS), LIVE/DEAD® Viability/Cytotoxicity Kit for
mammalian cells, penicillin/streptomycin antibiotics,
0.05% trypsin-EDTA (1X), and PrestoBlue® cell
viability reagent were all purchased from Invitrogen
(Life Technologies, Carlsbad, CA, USA). All antibodies
and enzyme-linked immunosorbent assay (ELISA) kits
2
Biofabrication 8 (2016) 014101 N S Bhise et al
were purchased from Abcam ( Cambridge, MA, USA).
Other reagents and materials were purchased from
Sigma Aldrich (St. Louis, MO, USA) unless mentioned
otherwise.
GelMA was synthesized as previously described by
functionalizing 10% (w/v) gelatin (G2500, Sigma-
Aldrich) in Dulbecco ’s phosphate buffered saline
(DPBS) by the dropwise addition of methacrylic anhy-
dride (276 685, Sigma-Aldrich) [7, 39]. The concentra-
tion of methacrylic anhydride in the reaction mixture
was 8% (v/v). This reaction was carried out for 2 h at
50 °C with constant stirring at 200 rpm. The reaction
was then stopped by dilution with DPBS and dialyzed
for 7 days at 40 °C under constant stirring at 500 rpm.
The dialyzed solution was sterile filtered and freeze
dried for 5 days. The lyophilized foam was stored at
room temperature for any future use.
Bioreactor design and fabrication
Figure 1(a) shows the schematic of the overall bior-
eactor platform utilized in the present study. The
designed bioreactor consisted of multilayers of poly-
dimethylsiloxane (PDMS) and poly(methyl methacry-
late)(PMMA) and included three chambers connected
by fluidic channels as shown in figures S1(a) and (b) in
ESI. The chambers and channels were all fabricated
using PDMS by casting PDMS around laser cut
PMMA molds. A 3-(trimethoxysilyl)propyl methacry-
late coated glass slide was used as the bottom layer and
the hydrogel constructs were directly bioprinted on
top of it (
figure 1(b)). The multilayer design was
selected in a way that any introduced bubble would be
washed away without getting trapped within the
chambers. The PDMS layers and glass slide were
sandwiched between two layers of PDMS covered
PMMA sheets. Eight sets of screws and nuts were used
to compress the layer together to prevent potential
leakage (figure 1(c)). Unlike the majority of PDMS-
based on-chip systems, which are permanently plasma
bonded to seal the compartments, the bioreactor could
be unscrewed for disassembly at any time for allowing
direct access to the hepatic construct and then
resealed.
The central cell culture chamber size was selected
to allow it to be interfaced with the employed direct
write bioprinter. The dimensions of the chambers and
the fluidic channels are all listed in figure S1. The total
volume of the system was estimated to be 2.4 mL.
Upon the completion of the bioprinting process
(explained later), the bioreactor was connected to a
syringe pump (Harvard Apparatus PhD 2000, Cam-
bridge, MA, USA) and was perfused with media
Figure 1. (a) Schematic of the hepatic bioreactor culture platform integrated with a bioprinter and biomarker analysis module.
(b) Bioprinting photocrosslinkable GelMA hydrogel-based hepatic construct within the bioreactor as a dot array. (c ) Top-view
(i) and side-view (ii) of the assembled bioreactor with the inlet and outlet fluidic ports as indicated. Scale bar=1 mm. (d) Oxygen
concentration gradient in the bioreactor, considering the oxygen uptake of, case A: 400 000 hepatocytes on day 1 (16 000 cells per dot),
and case B: 4000 000 hepatocytes on day 30 (160 000 cells per dot).
3
Biofabrication 8 (2016) 014101 N S Bhise et al