Content uploaded by Axel Guenther
Author content
All content in this area was uploaded by Axel Guenther on Apr 17, 2018
Content may be subject to copyright.
Lab on a Chip
PAPER
Cite this: DOI: 10.1039/c7lc01236e
Received 20th November 2017,
Accepted 30th March 2018
DOI: 10.1039/c7lc01236e
rsc.li/loc
Handheld skin printer: in situ formation of planar
biomaterials and tissues†
Navid Hakimi,
a
Richard Cheng,
b
Lian Leng,
a
Mohammad Sotoudehfar,
a
Phoenix Qing Ba,
a
Nazihah Bakhtyar,
c
Saeid Amini-Nik, *
de
Marc G. Jeschke*
cdf
and Axel Günther *
ab
We present a handheld skin printer that enables the in situ formation of biomaterial and skin tissue sheets
of different homogeneous and architected compositions. When manually positioned above a target sur-
face, the compact instrument (weight <0.8 kg) conformally deposits a biomaterial or tissue sheet from a
microfluidic cartridge. Consistent sheet formation is achieved by coordinating the flow rates at which bio-
ink and cross-linker solution are delivered, with the speed at which a pair of rollers actively translate the
cartridge along the surface. We demonstrate compatibility with dermal and epidermal cells embedded in
ionically cross-linkable biomaterials (e.g., alginate), and enzymatically cross-linkable proteins (e.g., fibrin), as
well as their mixtures with collagen type I and hyaluronic acid. Upon rapid crosslinking, biomaterial and skin
cell-laden sheets of consistent thickness, width and composition were obtained. Sheets deposited onto
horizontal, agarose-coated surfaces were used for physical and in vitro characterization. Proof-of-principle
demonstrations for the in situ formation of biomaterial sheets in murine and porcine excisional wound
models illustrate the capacity of depositing onto inclined and compliant wound surfaces that are subject to
respiratory motion. We expect the presented work will enable the in situ delivery of a wide range of differ-
ent cells, biomaterials, and tissue adhesives, as well as the in situ fabrication of spatially organized biomate-
rials, tissues, and biohybrid structures.
Introduction
Skin is the largest organ that provides a barrier to the total
body surface area (TBSA), 1.6–2.0 m
2
in adults,
1
and possesses
a unique layered organization of cells and extracellular matrix
(ECM) constituents.
2
The viable epidermis
3
is the thin (thick-
ness δ∼95 μm, elastic modulus E∼1.5 MPa) outermost cellu-
lar layer that consists of densely packed keratinocytes and
provides a barrier against water loss and bacterial transport.
Beneath it are the dermis
3
(δ∼1.4 mm, E∼0.02 MPa), popu-
lated by fibroblasts along with various other cell types and an
extracellular matrix containing fibrous collagen;
4
and the
hypodermis
3
(δ∼0.8 mm, E∼0.002 MPa).
Patients with acute and complex full-thickness wounds are
particularly vulnerable to opportunistic infections and dehy-
dration. Full-thickness wounds where the dermis, epidermis,
and hypodermis are all destroyed do not heal or take a long
time to heal by reconstitution of the dermis followed by re-
epithelialization progressing from the wound edge.
5
The cur-
rent preferred treatment for full-thickness wounds is split-
thickness autografting.
5,6
In a first step of the procedure, an
acellular scaffold may be applied to aid the reconstitution of
the dermis and a temporary barrier is established against bac-
terial transport and water loss. In a second step that is often
performed at a later time point, the temporary barrier is re-
moved. Skin, usually 0.3 mm in thickness, containing the epi-
dermis and the upper part of the dermis is then harvested
from healthy regions of the body using a handheld instru-
ment, a dermatome, and redistributed onto the wound area
as a sheet or meshed graft. In large wounds, the available
healthy donor skin is often insufficient for autografting,
Lab ChipThis journal is © The Royal Society of Chemistry 2018
a
Department of Mechanical and Industrial Engineering, University of Toronto, 5
King's College Road, Toronto, Ontario M5S3G8, Canada
b
Institute of Biomaterials and Biomedical Engineering, University of Toronto, 164
College Street, Toronto, Ontario M5S 3G9, Canada.
E-mail: axel.guenther@utoronto.ca
c
Ross Tilley Burn Centre, Sunnybrook Health Sciences Centre and Sunnybrook
Research Institute, 2075 Bayview Ave, Room D704, Toronto, Ontario M4N 3M5,
Canada. E-mail: marc.jeschke@sunnybrook.ca
d
Department of Surgery, Department of Immunology, Division of Plastic Surgery
and General Surgery, University of Toronto, 149 College Street, Toronto, Ontario
M5T 1P5, Canada. E-mail: saeid.amininik@utoronto.ca
e
Department of Laboratory Medicine and Pathobiology, University of Toronto,
Medical Sciences Building, 6th Floor, 1 King's College Circle, Toronto, Ontario
M5S 1A8, Canada
f
Institute of Medical Science, University of Toronto, 1 King's College Circle, Room
2374, Toronto, Ontario M5S 1A8, Canada
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c7lc01236e
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
View Journal
Lab Chip This journal is © The Royal Society of Chemistry 2018
leaving a large portion of the wounded area either ungrafted,
allografted, or uncovered, and resulting in poor outcomes.
A large number of acellular skin substitutes based upon bio-
logical or synthetic biomaterials have been introduced to im-
prove wound healing in acute and chronic wounds.
6,7
An acel-
lular dressing based on chemically cross-linked and freeze-
dried collagen
8
remains the ‘gold standard’for clinical use.
When covered with a split-thickness autograft it reduces wound
contraction while promoting the migration of residual healthy
cells and leads to the reconstitution of a new dermal layer
within 2–3 weeks. Biomaterials options also include mats of
electrospun nanofibers. A wide range of tissue-engineered (i.e.,
cellular) skin substitutes have been introduced, including cul-
tured epithelial autografts.
5
In addition, spraying cells onto
partial-thickness
9
and full-thickness
10
acute wounds, as well as
onto chronic wounds
11
was reported to improve wound
healing. Accelerated healing in murine wounds was recently
reported for an injectable skin substitute composed of micro-
gel particles and cells.
12
In spite of the large number of avail-
able tissue-engineered skin substitutes, they are not yet widely
used in the clinic. Several barriers prevent the routine clinical
application of cell-based engineered skin substitutes in large
acute wounds, e.g., the long times required for expanding suffi-
cient cell numbers, high cost, and the inability of tailoring sub-
stitutes to specific wound and patient characteristics. We hy-
pothesize that some of these barriers can ultimately be
overcome by in situ skin printing and introduce the approach
along with its thorough characterization in this paper.
The spatial organization of biopolymers and cells is closely
associated with biological organ function in health and dis-
ease. Additive manufacturing methods aim to recapitulate as-
pects of this spatial organization in engineered tissues. A
shared motivation for many current studies is to elucidate the
extent to which the initially provided tissue organization con-
tributes to improving the functional characteristics of tissues
after in vitro culture or in vivo application, as well as to accel-
erating tissue regeneration. 3D printing approaches have been
extended to soft materials
13
and tissues.
14–19
Different extru-
sion or ink-jet based bioprinters as well as microfluidic ap-
proaches for the formation of biopolymer fibers
20–22
and
sheets
23
offer local control over the delivery of biopolymers
and cells. Synthetic and natural biopolymers with different
composition and cross-linking mechanisms serve as the ‘bio-
ink’.
24,25
Target bioink properties include shear thinnening
rheological behavior, rapid gelation, cell compatibility, and
upon gelation a chemical composition and stiffness mimick-
ing the microenvironment in the intact tissue of interest.
Protein-based biomaterials including collagen I, the most
abundant protein in the dermis, and fibrin, a protein involved
in wound healing, make for obvious candidates as they are
widely used in 3D cell culture and in clinical settings.
26,27
Their “printability”however is poor, as the low viscosity of
protein-based solutions and their long gelation times (>1
min) exceed characteristic time scales required for extrusion-
based bioprinting processes.
28
To improve printability, rap-
idly gelling biopolymers are often added.
Several recent examples have demonstrated bioprinted
skin tissues in vitro,
29–31
or in small animal models.
32,33
How-
ever, important requirements for the routine in vivo applica-
tion of bioprinted skin substitutes under conditions that are
compatible with use in large animal models and ultimately
the clinic are not yet met. They include the formation of rou-
tinely handleable larger skin tissues from composite mate-
rials while retaining a soft cellular microenvironment, or the
in situ formation of skin tissues from compact bioprinting
systems. A first strategy that was recently demonstrated in
other tissues
34
and can likely be applied to skin is a sequen-
tial multimaterial printing approach, where the deposition of
a backbone support structure from a biocompatible sacrifi-
cial material preceded the deposition of a bioink. A second
strategy consists of the in situ formation of organized bioma-
terial and skin tissue sheets conformal to wound surfaces.
While in situ bioprinting based on ink-jet approaches has re-
cently been reported in small and large animal wound
models
46
these approaches require laser scanning of wounds
as well as bulky setups and offer limited spatial control.
Here, we present a compact extrusion-based approach that
enables bioprinting of planar biomaterials and skin tissue
sheets compatible with in vivo application in large animals,
and, we expect, ultimately the clinic. Bioink solutions are spa-
tially organized using a microfabricated cartridge,
conformally deposited onto flat or curved surfaces, and rap-
idly cross-linked. Homogeneous and architected sheets are
consistently formed and characterized in vitro. Proof-of-
concept demonstrations for in situ bioprinting in murine and
porcine excisional wound models illustrate the compatibility
of the approach with compliant wound surfaces.
Materials and methods
Fig. 1a shows a schematic that introduces the concept of in
situ bioprinting using the handheld Skin Printer. Cells (autol-
ogous or allogeneic) are suspended in a hydrogel precursor
solution and loaded into one or several separate syringes. An-
other syringe contains a crosslinking solution that will aid
the gelation of the cell-laden biopolymer solution under mild
conditions (i.e., natural pH, physiological temperature) and
high cell viability. After loading the primed syringes into the
handheld bioprinter, the bioink is deposited as a biomaterial
or tissue sheet within a culture dish for in vitro studies, or di-
rectly onto a wound bed for in vivo studies. For example, bio-
ink containing human fibroblasts can be homogeneously dis-
tributed within the 0.1–0.6 mm thick dermal layer. Bioink
containing keratinocytes may be deposited within parallel
stripe patterns that are separated by cell-free stripes resem-
bling a meshed epithelial skin graft.
The handheld bioprinter is an integrated, lightweight in-
strument (weight <0.8 kg including loaded syringes) with a
high degree of portability. It is straightforward to operate
with only one hand and consists of several key parts that are
shown in Fig. 1b. A handle, ①, allows positioning the Skin
Printer manually above a flat surface or wound bed. During
Lab on a ChipPaper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab ChipThis journal is © The Royal Society of Chemistry 2018
deposition, a pair of actively driven rollers with rubber wheels, ②,
translates the instrument along the deposition surface at velocity V,
while conformally coating it with a bioink sheet covered with cross-
linker solution. Traction between the wheel and the deposition sur-
face, determined by the friction coefficient and contact pressure,
35
is important for smooth and definite translation speed and consis-
tent deposition. During in vitro experiments, we deposit sheets hori-
zontally, against agarose coated glass slides. During deposition in
large animal models, both wheels are in contact with inclined and
compliant surfaces of either intact skin or a wound. Two onboard
syringe pump modules,
36
③, deliver the bioink, ④,atvolumetric
flow rate Q
M
and the cross-linker solution, ⑤,atflowrateQ
C
,
irrespective of the bioink viscosity, the cell density and the instru-
ment orientation. The barrels of the bioink and cross-linker solu-
tions rest in a syringe holder, ⑤. Both solutions are individually
supplied to an exchangeable microfluidic cartridge, ⑦,via flexible
tube connections that have dead volumes less than 20 μl.
Microfluidic cartridge
The microfluidic cartridge, the central part of the instrument
(Fig. 1c, ESI†section S2), was 3D printed
37
using an optically
semi-transparent resin. Integrating standardized Luer lock
connectors and on-chip wells within one 3D printed part
allowed us to reduce unwanted dead volumes (∼65 μl). The
cartridge provides uniform lateral distribution of at least two
solutions within microchannel networks located in separate
planes. Cartridge designs had w
0
= 8 mm, 14 mm, and 20 mm
wide exit sections. During deposition, a H= 1 mm tall slit is
obtained between a short overhanging roof section on the top
and the deposition surface on the bottom. Fig. 1d schemati-
cally shows how a bioink layer of thickness δ<Hand width
wis deposited onto the surface while being covered by the
cross-linking solution, initiating rapid gelation into a sheet.
The position of the cartridge exit aligns in the lateral direction
with the contact points between the drive wheels and the de-
position surface. In this configuration the thickness of the
sheet is only weakly dependent on the operator-selected angle
at which the handheld Skin Printer is being held, relative to
the deposition surface. The formed sheet may be homoge-
neous or heterogeneous in composition. In the latter case, the
spatial organization is determined by the microchannel con-
figuration within the cartridge, along with the parameters se-
lected for V,Q
M
and Q
C
.
Fig. 1 Handheld skin printer. (a) Schematic diagram illustrating working principle of handheld bioprinter. One or several bio-ink solutions (green
color), containing premixed biomaterials and cells, and a cross-linker solution (blue color) are prepared. Handheld bioprinter converts bio-inks into
homogenous or architected biomaterial sheets or tissues directly within a culture dish or a wound site. (b) Rendered image of handheld bioprinter.
A handle ①enables positioning above target surface or wound. A stepper motor, pulley and drive mechanism ②define the deposition speed, V.
Two on-board syringe pump modules ③control the dispensing flow rates for bioink ④and cross-linker solution ⑤. Syringe holder ⑥. 3D printed
microfluidic cartridge ⑦for spatial organization of solutions and sheet formation. (c) Photograph of 3D printed microfluidic cartridge. Scale bar 10
mm. (d) Schematic side-view image showing sheet formation between moving microfluidic cartridge and deposition surface or wound. Inset indi-
cates fluid velocity profiles in bioink (green) and cross-linker layers (blue).
Lab on a Chip Paper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab Chip This journal is © The Royal Society of Chemistry 2018
Control box
A control box is connected to the handheld Skin Printer via a
pneumatic line and an electrical ribbon cable. The control box
powers the instrument, controls the deposition velocity V,the
flow rates Q
M
and Q
C
, and, in the case of pressure-controlled de-
livery, the head pressure and the duration of its application. For
a detailed description of the design we refer to ESI†section S2.
Preparation of Agarose substrate
A solution of 2% agarose (UltraPure Agarose, 16500100,
Invitrogen) in de-ionized (DI) water was prepared by micro-
wave heating. The solution was allowed to cool to 60 °C prior
to being poured into sterile square petri dishes (model
Z692344, Sigma Aldrich) and resulted in a 3 mm-thick gel
layer. The gel solidified while being kept at room temperature
for 30 min prior to use. For preparation of sodium alginate-
based sheets, 50 mM calcium chloride (CCL302, BioShop) was
added to the agarose solution prior to microwave treatment.
For printing of fibrin-based sheets, 2 ml of 50 IU thrombin
(T4648, Sigma Aldrich) in PBS (10010023, Gibco) was pipetted
to hydrate the agarose substrate prior to sheet deposition.
Bioink preparation
Bioinks with three different compositions were prepared. For
alginate-collagen sheets, sodium alginate (Pronva UPLVG,
Novamatrix) was dissolved in DMEM (11965-084, Gibco) and
20 mM HEPES (15630080, Gibco) and filtered using 0.1 μm
syringe microfilter (Millipore). Collagen type 1 (rat tail,
354249, Corning) was balanced to a pH of 7 using 1 M NaOH
in PBS. The two stock solutions were mixed to obtain a final
concentration of 5 mg ml
−1
collagen and 2% alginate. The so-
lution was kept on ice prior to use. To prepare the bioink for
the dermal layer 5% fibrinogen (F8630, Sigma) was dissolved
at 37 °C in PBS with mild agitation for 2 h. 1% HA (sodium
hyaluronate Pharma Grade 80, Novamatrix) was dissolved in
PBS. The solutions were mixed at a ratio of 1:1 and then fil-
tered. Collagen type 1 solution was balanced with NaOH to a
pH of 7 and mixed with the filtered fibrin/HA solution to ob-
tain a final concentration of 1.25% fibrinogen, 0.25% HA and
0.25% collagen. The solution was kept on ice prior to use.
The bioink for the epidermal layer was prepared with a final
concentration of 2.5% fibrinogen and 0.25% HA. ESI Table
S1 summarizes the composition of materials used for the
preparation of the bioinks.
For printing the fibrin-based sheets, a layer of 50 IU
thrombin was co-delivered above the fibrinogen based der-
mal and epidermal bioinks. The rapid enzymatic reaction be-
tween fibrinogen and thrombin is mass transfer limited in
the considered case. The selected approach allowed the for-
mation of sheets on the site of the deposition which solidi-
fied at time scales between tens of seconds and several mi-
nutes, depending on the selected thrombin concentration
and sheet thickness, δ. The gelation time is directly depen-
dent on the sheet thickness (Fig. 2e). For the dermal bioink
consisting of a mixture of collagen and fibrinogen, the gela-
tion of fibrinogen occurs first and is induced by the diffusion
of thrombin. As a result, the sheet thickness and composition
are maintained while the slower thermally induced gelation
of neutral pH collagen progresses.
The alginate and alginate-collagen based sheets were pre-
pared by co-delivery of 10 mM calcium chloride above the
biopolymer layer. Similarly, rapid ionic cross-linking of algi-
nate preceded the slower thermal gelation of neutral pH
collagen.
In vivo experiments
The animal experiments were reviewed and approved by and
performed in accordance with the guidelines and regulations
set forth by the Sunnybrook Research Institute and
Sunnybrook Health Sciences Animal Policy and Welfare Com-
mittee of the University of Toronto, Ontario Canada. All pro-
cedures using animals were approved by the Sunnybrook ani-
mal care committee, approval #17-503 for murine
experiments and #17-600 for the porcine experiments under
the auspices of the Canadian Council on Animal Care. For a
more detailed description of experimental and characteriza-
tion methods see ESI†section S1.
Results and discussion
Single and multilayered biomaterial sheets
Fig. 2a shows a bright field image of a sheet with the uniform
thickness of δ= 300 μm produced with the handheld Skin
Printer (right) in comparison with a non-uniform thickness
pattern obtained by manually pipetting a comparable amount
of the same hydrogel precursor (left). Both images were taken
at an angle of 4°with respect to the flat deposition surface
consisting of an agarose layer that was hydrated with the
cross-linker solution. In the latter case, the pipetted hydrogel
assumed a dome shape. Uniform spreading of the bio-
polymer solution was prevented by gelation starting along the
contact line and then radially progressing inward. In the for-
mer case of the printed hydrogel sheet, however, gelation uni-
formly progressed in the direction normal to the deposition
surface. This was achieved by first delivering the hydrogel
precursor solution through a bifurcating microchannel net-
work. At each bifurcation point within the 3D printed micro-
channel network, the hydraulic diameter of the daughter
branches decreased in accordance with Murray's law. For
more detail see ESI†section S2. The bifurcated channel net-
work ensures the flow resistance in the smallest daughter
channels at the device exit to be high and leads to a constant
thickness biopolymer layer to be deposited. A cross-linker
layer was co-delivered atop to initiate gelation. Fig. 2b shows
a profilometer reading that confirms a consistent thickness
along the >0.9 wwide center portion of the obtained sheet.
Next, we discuss how the different experimental parame-
ters affected the sheet thickness δ. We consider the laminar
flow of a layered fluid between the two surfaces and apply the
lubrication approximations.
38
Because w
0
/H>10 we approxi-
mate the hydraulic diameter as 2H. We neglect the pressure
Lab on a ChipPaper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab ChipThis journal is © The Royal Society of Chemistry 2018
gradient in the z-direction as well as inertia effects. The conti-
nuity equation and the simplified momentum balance result
in a single elliptic differential equation, the Reynolds equa-
tion, for the pressure gradient along the film. An analytical
model is derived and presented in ESI†section S3. The model
allows predicting δbased on the fluid viscosities μ
M
and μ
C
,
and the flow rates of the biopolymer and the cross-linker solu-
tions, Q
C
and Q
M
(Fig. 2c). The corresponding dimensionless
quantities are and .
For given values of the dimensionless sheet thickness, μ*, the
dimensionless flow rates and are selected in such a
way that the pressure is invariant along the deposition direc-
tion, i.e., . At the selected condition, unwanted
bioink back flow or leakage at the sides of the cartridge are
avoided during deposition. The dimensionless bioink and
cross-linker flow rates required to obtain a target sheet thick-
ness δ*are
Fig. 2 In situ deposition of homogeneous and layered hydrogel sheets. (a) Comparison of side view images after manual deposition of 100 μl
fibrin/HA bioink droplet (left) and sheet deposited using handheld bioprinter (right). Agarose substrates hydrated with cross-linker solution used in
both cases. Images acquired at 4°angle. (b) Representative optical profilometry image and cross-sectional view of δ= 300 μm sheet printed with
14 mm wide microfluidic cartridge. (c) Measurement and model predictions for dimensionless sheet thickness, δ*=δ/H, as function of dimension-
less bioink flow rate, . (d) Analytical solution for velocity distribution within confinement. (e) Characterization of gelation kinet-
ics based on measurement of time-dependent changes in normalized turbidity of fibrin-based sheets with different thicknesses. (f) Microstructure
characterization using scanning electron microscopy, (g) Young's modulus and (h) elongation at break, for sheets consisting of fibrin-HA/collagen
I, fibrin-HA, collagen I-alginate and alginate. (i) Confocal image of bi-layer sheet prepared by subsequent deposition of 200 μm thickness alginate
layer containing FITC microparticles (bottom) and 100 μm thickness alginate layer containing Nile red microparticles (top). (j) Confocal image of
three-layer sheet prepared by subsequent deposition of 500 μm (bottom) fibrin-HA layer with blue micropartices, 200 μm (middle) alginate-
collagen layer with FITC-conjugated collagen and 150 μm (top) alginate layer with Nile red microparticles. Scale bars 2 mm (a left), 5 mm (a right),
2μm, 10 μm, 1 μm, 1 μm (f, from left to right), 100 μm (i and j).
Lab on a Chip Paper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab Chip This journal is © The Royal Society of Chemistry 2018
Fig. 2d shows the corresponding velocity profiles. In many
cases the viscosity of the cross-linker is much lower than the
one of the biopolymer solution, μ*∼0, leading to the linear
relationships . Note that the rela-
tionship is fulfilled even in the case of a very short
or absent overhanging roof structure at the cartridge exit
section.
The analytically predicted sheet thickness is in excellent
agreement with values measured for bioprinted fibrin-based
and alginate-based sheets. Sheet thicknesses between δ= 100
μm and 600 μm were reliably obtained using a microfluidic
cartridge with H= 1 mm. Thicker sheets, δ>600 μm, were
achieved by sequential deposition of multiple thinner sheets.
Sequential deposition is preferred over extruding thicker
sheets in a one-step process using a modified cartridge de-
sign with H>1 mm. The latter would increase the diffusion
time, ∼δ
2
/4, and consequently the time required for gelation,
and make the sheet thickness dependent on the inclination
angle of the deposition surface.
The handheld bioprinter is compatible with different bio-
polymers. In the presented work we consider the
polysaccharide-based biopolymer calcium alginate, and the
protein-based biopolymer fibrin. The compositions of the dif-
ferent considered bioinks and cross-linker compositions are
summarized in ESI†Table S1. Selected bioink choices are
highly biocompatible, biodegradable, provide a cellular
microenvironment conducive to cell proliferation and attach-
ment, and do not require secondary washing steps prior to
in vitro culture or direct in vivo application.
Understanding the kinetics of gelation within the depos-
ited bioink layers is crucial for the in situ deposition of tis-
sues. In the considered cases, gelation is a diffusion-limited
process that is initiated at the interface between the bioink
and cross-linker layers and propagates throughout the for-
mer. For alginate-based sheets, gelation is induced via rapid
ionic cross-linking by diffusion of Ca
2+
-ions from above.
39
Ge-
lation of fibrinogen is induced by an enzymatic reaction with
thrombin. Gelation is slower in this case because the process
is diffusion limited, and the diffusivity of thrombin is at a
comparable viscosity approximately ten times lower than the
one of Ca
2+
-ions. In order to retain the deposited layer thick-
ness and sheet architecture while gelation progresses, we in-
creased the viscosity of the fibrinogen-based bioink by
adding hyaluronic acid. We assessed the kinetics of gelation
by performing systematic turbidity measurements. Fig. 2e
shows data from time-resolved turbidity measurements
performed on in situ deposited fibrin-HA sheets with thick-
nesses of δ= 100 μm, 200 μm, 400 μm and 600 μm. Gelation
was induced by inter-diffusion of thrombin and calcium chlo-
ride contained in the cross-linker stream above, and in the
agarose-coated deposition surface below. The in situ turbidity
measurements reveal gelation times (estimated from the in-
flection time on the measured turbidity graph) of t
g
∼44 s,
64 s, 110 s and 160 s for sheet thicknesses of δ=100 μm, 200
μm, 400 μm and 600 μm. As we will discuss in more detail
below, rapid enough gelation is an important requirement for
in situ formation of sheets on inclined or compliant surfaces.
Fig. 2f shows representative scanning electron microscopic
(SEM) images to characterize the surface microstructure of
sheets that were obtained from the different bioinks. As the
sheets are small in thickness (100–600 μm), we don't expect
large non-homogeneities in pore size and microstructure
across the sheet thickness.
40
Fig. 2g and h show the Young's
moduli and elongations at break of in situ formed sheets, re-
spectively. Tensile properties were evaluated for the hydrated
sheets along the direction of deposition, x. At the selected
processing conditions, alginate-based sheets exhibited higher
Young's moduli, compared with fibrin-based ones. The latter
has a higher elasticity and a 2.6 times higher elongation at
break (at constant strain).
To obtain multilayered sheets of controllable thickness,
multiple sheets may be consecutively deposited using the
handheld Skin Printer. In addition to depositing multiple
sheets of the same composition, the stepwise approach en-
ables the biopolymer or the cellular composition to be altered
between layers. Fig. 2i and j show confocal micrographs of
multilayered sheets that were sequentially obtained. To
achieve the three-layered sheet shown in Fig. 2j the following
three layers were deposited from bottom to top: a 500 μm
layer of fibrin containing 0.1 μm polystyrene particles (blue
color); after 5 min a 200 μm layer of alginate containing FITC
conjugated collagen type 1; and after another 30 min a 150
μm layer of alginate containing 0.2 μm polystyrene particles
(conjugated with Nile red).
Architected biomaterial and tissue sheets
We now consider microfluidic bioprinter cartridges that aid
the formation of single-layered biomaterial or tissue sheets
with a composition that varies along either the lateral or the
extrusion direction. As schematically shown in Fig. 3a and
ESI†Video S1, architected sheets were obtained by indepen-
dently controlling the flow rates of a primary bioink, Q
M1
(in-
dicated by green color), and a secondary bioink, Q
M2
(indi-
cated by red color) as delivered from the two on board
syringe pumps. The cross-linker was supplied from an exter-
nal syringe pump in this case. The design of the microfluidic
cartridge was adapted from one that we had previously used
for the formation of stripe-patterned sheets without substrate
support.
23
We incorporated four equidistant stripes of a sec-
ondary biopolymer within the primary biopolymer. As shown
in Fig. 3b and c the relative stripe width, w
2
/w
0
, decreased
when we increased the relative rate at which the secondary
bioink was supplied compared with the primary one, Q
M2
/
Q
M1
, while keeping the total bioink flow rate, Q
M2
+Q
M1
,
unchanged. Stripes narrower than the width of the smallest
channel features of the microfabricated cartridge were
obtained by hydrodynamically focusing
41
the secondary bio-
ink within the biopolymer feature layer of the microfluidic
cartridge. In the examples presented so far, the primary and
secondary bioink solutions shared the same composition but
Lab on a ChipPaper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab ChipThis journal is © The Royal Society of Chemistry 2018
differed in their payload. However, primary and secondary bio-
ink solutions with different composition may also be used as
long as their cross-linking mechanisms are compatible. For ex-
ample, single-layered sheets with alternating composition were
obtained by co-flowing two different bioink solutions in-plane.
Fig. 3d shows a representative sheet with alternating alginate
and fibrin-HA stripes deposited on an agarose substrate coated
with calcium chloride and thrombin.
In addition, we formed sheets where spots of the second-
ary bioink were incorporated within the primary bioink. The
former was delivered from a well that was incorporated close
to the inlet of the 3D printed microfluidic cartridge, where a
time-dependent head pressure was applied (Fig. 3e). We then
delivered the primary bioink at a flow rate, Q
M1
, and applied
a square-wave pressure signal. The frequency and the duty cy-
cle of the pressure affected the spot volumes, as well as the
distance between spots in the direction of extrusion. For a
fixed Q
M1
, frequency and duty cycle, increasing the amplitude
of the head pressure decreased the distance between subse-
quent spots and increased the spot volume (Fig. 3f and g).
We analytically estimated the attainable resolution of the de-
position process (Fig. 3h, ESI†section S5) depending on the
inclination of the deposition surface, and the gelation kinetics.
As the sheet thickness increases, the permitted drainage time
scale for in situ formation of sheets with a desired spatial reso-
lution decreases, and the gelation time increases. The permit-
ted drainage time decreases with increasing inclination angle.
The in situ formation of substrate-adhesive undulated
sheets or arrays of parallel stripe filaments is schematically
shown in Fig. 3i and extends the attainable sheet morphol-
ogies. One cross-linker solution serves as the flow-confining
fluid and is referred to as the primary cross-linker solution
and delivered to the microfluidic cartridge from an external
syringe pump. The bioink is delivered by one of the on-board
syringe pumps. A secondary cross-linker solution is delivered
from the other on-board syringe pump, and distributed
within the biopolymer feature layer. As a result, stripes of the
secondary bioink and the cross-linker solutions are deposited
in alternating fashion along the lateral direction. For slow ge-
lation or a low viscosity of the cross-linker solution, the
stripe-patterned bioink spreads laterally until gelation is com-
pleted, producing undulated sheets (Fig. 3j and k). For more
rapid gelation, or an increased viscosity of the cross-linker so-
lution, the relaxation time due to the density difference
Fig. 3 In situ 3D bioprinting of architected sheets. (a) Schematic of biomaterials and cells organized into stripe patterns using microfluidic
cartridge. (b) Representative confocal image of stripe-patterned monolayer. (c) Relative stripe width w
stripe
/w
0
as function of flow rate ratio. (d)
Representative multi-material organization of sheets with alternating fibrin-HA (red color) and alginate (green color) stripes. (e) Schematic of bio-
materials and cells organized into spotted patterns using pressure-controlled reservoir. (f) Representative images for pressure-controlled spotting.
(g) Spot volume as function of reservoir head pressure during 0.2 s actuation. (h) Estimated nominal in-plane resolution for forming fibrin-HA sheet
onto flat but inclined surfaces, (inclination angle θ). Initial stripe resolution as obtained from 3D printed cartridge without (*) and with flow focusing
feature (**) (i) schematic of biomaterials and cells organized into undulated sheets or parallel fibers using microfluidic cartridge. (j) Representative
bright-field image of undulated sheet with 8 peaks. Image captured at 4 degrees. Insert shows enlarged image of two neighboring peaks at 2 de-
grees. (k) Confocal image of cross-section of an undulated sheet with 4 peeks. (l) Representative reconstructed confocal image of bi-layered sheet
cross-section. Bottom layer (green color) homogenous, top layer consists of 4 parallel stripes (red color). (m) Meshed pattern formed by successive
deposition of 8 parallel stripes perpendicular to one another. Scale bars 2 mm (b), 500 μm (d), 6 mm (f), 5 mm (j), 200 μm (k, l), 4 mm (m).
Lab on a Chip Paper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab Chip This journal is © The Royal Society of Chemistry 2018
between the biopolymer and cross-linker solutions is shorter
than the gelation time. An array of parallel substrate-attached
biopolymer filaments is obtained. Fig. 3l shows an array of
such stripe-filaments that was deposited in a single pass on
top of a homogeneous sheet. The filaments are disconnected
in the lateral direction. A mesh pattern as shown in Fig. 3m
was obtained in a case when another stripe array was depos-
ited in a second pass perpendicularly to a previously depos-
ited one.
In vitro application of handheld skin printer
The bioink used for in vitro experiments contained
hyaluronic acid, fibrinogen, and type-I collagen. Gelation of
the fibrinogen component was induced by the enzymatic ac-
tivity of thrombin at neutral pH and room temperature.
Hyaluronic acid was added to increase the viscosity and
“printability”of the bioink, without adversely impacting cell
viability. Selected deposition conditions are characterized by
low shear rates (on the order of 1 s
−1
) that are not expected
to affect cell viability or function. Fig. 4a and b show that hu-
man dermal fibroblasts (FB) embedded in the fibrin-based
sheets exhibited >90% viability based on a live/dead assay
performed after 10 days in culture. At day 0, we investigated
five different cell concentrations that ranged from 0.1 to 10
million/ml, and found the original cell seeding density in the
bioink to be consistent with the cell density assessed within
the planar tissue, based on Hoechst nuclear staining and
confocal microscopy (for more detail see ESI†section S1). No
cell or biomaterials clumps or aggregates were observed im-
mediately after sheet formation, indicating the delivered cells
to remain uniformly dispersed within the planar tissue (ESI†
Fig. S12). Cell numbers were quantified using a Hoechst nu-
clear stain and confocal microscopy and demonstrated the
increase in total cell numbers over a three-day culture period,
suggesting continued cell growth and proliferation (Fig. 4c).
To assess FB attachment and morphology, we selected a
bioink containing human dermal FBs in a fibrinogen–
collagen-HA solution. At time points 0, 3, 6 and 12 hour post
printing, sheets were fixed and stained for nuclei and cyto-
skeleton. The results suggest the cells are adapting to the 3D
scaffold without impacting morphology, as they exhibit elon-
gation and attachment within the first few hours after sheet
formation (Fig. 4d).
Keratinocytes (KCs) are the essential cell component of
the skin epidermal layer. A bioink consisting of 1.25 million
human KC ml
−1
in a mixture of fibrinogen and HA was used
to form a 200 μm thick sheet. Collagen-I was not added to
better mimic the epidermal layer undergoing wound repair,
to accelerate the degradation of the biomaterial matrix, and
thereby reduce cell–cell distances and aid KC cluster forma-
tion. On day 0, the cells were dispersed individually and
Fig. 4 In vitro characterization of in situ bioprinted tissues. (a) Homogenous printed sheet contains human dermal fibroblasts (FBs). Live cells
indicated by calcein stain (green). Dead cells indicated by fluorescent ethidium homodimer-1 (red). (b) Quantitative assessment of FB viability in
printed fibrin/HA/collagen-I bioink with >90% cell viability during 10 day culture. (c) Quantitative assessment of FB and KC cell numbers as an indi-
cation of cells proliferation over 3 days of in vitro culture. (d) FBs deposited within the bioink containing 1.25% fibrin, 0.25% collagen I, 0.25% HA
and stained with Hoechst (blue) and phalloidin (green) show excellent attachment and elongation during 12 hour. (e) Comparison between day
0 and day 3 of human keratinocytes (KCs) deposited in fibrin gel using immunofluorescent staining for cell nucleus (blue), actin (green), and
keratin-14 (red) indicating cell grouping and clustering by day 3. (f) Deposited monolayer sheet containing KCs in stripe patterns as visualized using
phalloidin immunostaining on day 0. (g) Bilayer construct printed in stepwise fashion keratinocytes (k14 & phalloidin co-stain, red) sequentially de-
posited on top of FBs (phalloidin, green) resembling bi-layered structure of skin. Scale bars 200 μm (a and g), 100 μm (d, e), 2 mm (f).
Lab on a ChipPaper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab ChipThis journal is © The Royal Society of Chemistry 2018
homogeneously distributed within the sheet (Fig. 4e, top row).
Within three days of 3D culture, KCs formed clusters as
shown in confocal micrographs, suggesting adhesion of KCs
and the formation of colonies, characteristics of a normal ep-
ithelial activity
42
(Fig. 4e, bottom row).
We obtained cell-laden stripe patterns in vitro,by
adding KCs to the secondary bioink as shown in Fig. 4f. The
width and distance of the stripes is governed by tuning the
volumetric flow rates of the secondary bioink, Q
M2
, and the
cross-linker, Q
M1
. The w
0
= 8 mm wide cartridge produced
equidistant stripes of width w
2
= 500 μm that were imaged by
using phalloidin staining.
To mimic the layered architecture of skin, a bi-layered
sheet containing both keratinocytes (KCs) and human dermal
FBs was deposited (Fig. 4g). First, a 500 μm thick layer was
formed from a bioink that contained 4 ×10
5
FBs ml
−1
in a
collagen/fibrinogen precursor solution. Second, a 100 μm
layer of KCs embedded within a fibrinogen/HA precursor so-
lution was deposited on top. Immunostaining of the bi-
layered construct specific for F-actin (Phalloidin green) and
keratin-14 (anti-K14, red) to visualize dividing basal
keratinocytes revealed cell compartmentalization in a strati-
fied structure with two distinct cell populations (phalloidin
green, K14 red).
In vivo application of handheld skin printer
We demonstrate the in vivo formation of architected sheets.
A murine wound model served as an example of a compliant
and curved surface that undergoes periodic respiratory mo-
tion. We characterized in situ sheet deposition by a combina-
tion of fluorescence and bright-field imaging during and im-
mediately after deposition. Mice were sacrificed 1 hour after
deposition. Fig. 5a and b show wide-field fluorescence images
of in situ formation of a biopolymer fiber array onto a 8 mm
diameter excisional wound, which is particularly challenging
because it requires the consistent sheets to be formed at a
short distance while deposition onto a curved and compliant
wound surface. The deposited stripe-patterned sheets or
stripe arrays remained firmly attached to the wound and peri-
odically followed respiratory cycles, or manual tissue defor-
mation (Fig. 5c). To improve contrast during imaging, we
added green fluorescent microparticles to the secondary bio-
ink. ESI†Fig. S11 shows images for the synchronous forma-
tion of stripe-arrays prepared from either alginate or fibrin-
HA onto an excisional wound located at the back of a mouse.
Stripe arrays prepared in fibrin-HA are less distinct compared
with the alginate case, a difference that we attribute to the
slower gelation in the former case (Fig. 5d).
Fig. 5 In vivo compatibility of in situ bioprinting in small and large animal models. (a) Fluorescent image of stripe-patterned sheet formation di-
rectly onto murine excisional wound. (b) Representative image of 4 stripes in situ deposited onto 8 mm wound model. Dashed circle indicates
wound edge and arrow indicates initiation phase. (c) Striped sheets remain adherent to wound bed during respiration or stretching, and the printed
geometry retains its shape. 1 μm green fluorescent microparticles incorporated as payload (a–c). (d) Representative normalized fluorescence inten-
sity across striped alginate sheet (solid line) and fibrin sheet (dashed line) in situ formed on the back of a murine excisional wound. (e) Representa-
tive photograph showing in situ deposition of δ= 250 μm thick fibrin-HA/collagen sheet on top of a full thickness excisional porcine wound using
handheld skin printer (top); close-up view of sheet formation within wound bed with a w
0
= 2 cm microfluidic cartridge (bottom). (f) Control on
day 0 and printed 5 min after in situ formation of biomaterial sheet. (g) Trichrome staining indicating the extent of granulation tissue formation and
reepithelialization. Arrows indicate the border between newly formed granulation tissue and intact skin. Arrowheads marks epithelialized area. Ar-
rowhead at the center of treated wound shows complete re-epithelialization, while central arrowhead in control wound shows non re-
epithelialized zone at wound center. Scale bars 2 mm (a–c and g left), 10 mm (f), 1 mm (g right).
Lab on a Chip Paper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab Chip This journal is © The Royal Society of Chemistry 2018
Finally, we selected a porcine wound model in a proof-of-
principle demonstration of in situ depositing a homogenous
hemostatic barrier on an excisional wound under clinically
relevant conditions. Wounds were marked on the back of the
animal before the operation. For pain control, animals re-
ceived a basic dosage of tramadol (4–6mgkg
−1
) which may
escalate when needed on the day of the surgery until
euthanization. After the excision, 20 mm ×40 mm full-
thickness excisional wounds were covered by direct deposi-
tion of a homogenous fibrin-HA sheet (δ= 300 μm, w=20
mm, Fig. 5e). We compared to a contralateral wound without
deposited sheet (4 wounds as control on same animal, ESI†
Video S2). Bleeding stopped after approximately 5 min in the
wounds covered by the in situ deposited sheets while control
wounds achieved hemostasis after tens of minutes (Fig. 5f).
Microscopic analysis of H&E staining on cross-sectional sam-
ples of the harvested healed wounds sacrificed on day 20 re-
vealed that both treated and control wounds formed com-
plete granulation tissue as indicated by the continuous red-
stained region connecting the wound edges, and exhibited
comparable levels of collagen deposition and cellularity as
shown by the underlying purple region populated homo-
genously by dark blue cell nuclei (Fig. 5g). Only 1 out of the 4
control wounds showed complete re-epithelialization while 3
out of 4 treated wounds exhibited complete re-
epithelialization (non-significant parametric test) (ESI†Fig.
S13). Healed wounds were stained with Keratin 10, an epithe-
lial differentiation marker (ESI†Fig. S14a), and with alpha-
smooth muscle actin, a transient myofibroblast marker (ESI†
Fig. S14b), where no significant difference was observed be-
tween treated wounds and controls. The selected bioinks did
therefore not inhibit granulation tissue formation or re-
epithelization.
Summary and conclusions
We presented a handheld bioprinter for the formation of bio-
material and tissue sheets with local control over biomaterial
composition, and colloidal or cellular payload. The approach
requires minimal operator training for the in situ formation
of biomaterial and tissue sheets onto either flat surfaces
(in vitro) or wound beds (in vivo). The approach side steps
the washing and incubation steps, scanning of wound sur-
faces or multi-axis printhead translation that would otherwise
be required by many conventional bioprinters. Fragile sheets
that correctly mimic the soft microenvironment of many cells
can be in situ delivered for in vivo application.
We reported conditions for consistent sheet deposition
using bioinks that consisted of alginate, fibrin, collagen and
hyaluronic acid. The biopolymer solution was pre-mixed with
cells prior to deposition to deliver cells in a controlled and
cytocompatible manner and promote cell interaction with the
surrounding microenvironment.
43
Depending on the selected
cartridge width and deposition velocity we rapidly covered de-
position surfaces at rates of 0.3–1.6 cm
2
s
−1
, exceeding char-
acteristics of most extrusion-based bioprinters by at least one
order of magnitude. Additionally, we demonstrated different
sheet morphologies including stripes, spots and meshes. The
capacity of bioprinting undulate sheets may in the future al-
low mimicking the periodic protrusions (wavelength 50–400
μm, amplitude 50–200 μm) that separates the epidermal layer
from the dermal layer of intact skin in the bioprinting pro-
cess. These protrusions have been shown to increase the con-
tact area, improve adhesion between layers, influence epider-
mal cell proliferation and migration, and decrease the
chance of scar formation.
44,45
A first proof-of-concept experiment was conducted in a
murine wound model with the goal of demonstrating the in
situ deposition of an architected sheet in the form of a fiber
array onto a small and compliant wound surface. In a second
a proof-of-concept study we evaluated in a porcine full thick-
ness wound model the feasibility of using the handheld Skin
Printer for in situ biopolymer sheet deposition in a clinically
relevant setting. A hemostatic biopolymer sheet (without cell
load) was in situ fabricated. Wounds where monitored for 20
days and histological end point analysis was performed. The
porcine model demonstrated successful in situ bioprinting to
cover full thickness wounds with a homogeneous layer that
provided a non-detrimental hemostatic barrier immediately
after application where it did not impede normal re-
epithelization or wound contraction. Four porcine full thick-
ness wounds with sizes of 2 cm ×4 cm (1.24 in
2
) were cov-
ered with a δ= 0.3 mm thick and 20 mm wide sheet. The cur-
rent instrument accommodates up 3 ml of bioink solution,
which allows at the selected conditions the continuous cover-
age of ∼100 cm
2
(15.5 in
2
) during 0.8–2.1 min. Covering sig-
nificantly larger areas with homogeneous sheets will require
the bioink volume per filling to be increased. Comprehensive
in vivo experiments will be required to assess wound healing
of in situ deposited skin tissue sheets.
We expect the presented approach to be widely applicable
for the delivery of different natural and synthetic biopolymers
solutions as well as differentiated and non-differentiated
cells.
Author contributions
N. H., L. L., S. A. N., M. G. J and A. G. designed the research.
N. H. conducted the experiments and analyzed data. M. S.
and N. H. designed, machined, assembled and tested the
handheld Skin Printer and control box. N. H. designed the
microfluidic cartridge. N. H. and P. Q. B developed analytical
models and performed sheet thickness measurements. R. C.
performed SEM. S. A. N. and M. G. J supervised in vivo experi-
ments. R. C. and N. B. assisted with in vitro and in vivo exper-
iments. N. B. and S. A. N. analyzed in vivo results. N. H. and
A. G. wrote the manuscript. R. C., L. L, S. A. N., and M. G. J,
reviewed the manuscript.
Conflicts of interest
There are no conflicts of interest.
Lab on a ChipPaper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab ChipThis journal is © The Royal Society of Chemistry 2018
Acknowledgements
Helen Han Ming and Caiden Chih produced 3D rendered im-
ages. Shashi Malladi assisted with the porcine study and
wound dressing changes. Kavin Kowsary took bright field
photographs of the microfluidic cartridge and the in vivo exper-
iments and recorded supplementary videos. We acknowledge
an NSERC postgraduate fellowship (NH), a Barbara and Frank
Milligan graduate fellowship (NH), a Weber and Mariano Grad-
uate Scholarship (NH), a fellowship of the NSERC Training
Program Organ-on-a-Chip Engineering & Entrepreneurship
(RC), and the endowed Wallace G. Chalmers Chair in Engineer-
ing Design (AG). We are grateful for grant support from NSERC
DC and DAS (AG), Grand Challenges Canada (AG, MGJ), Medi-
cine by Design funded by the Canada First Research Excellence
Fund (Transition Award, AG; Seed Grant, SAN, MGJ), CIHR
(123336, MGJ), NIH RO1 (2R01GM087285-05A1, MGJ), CFI
Leader's Opportunity Fund (25407, MGJ) and a generous dona-
tion from Toronto Hydro (MGJ). Device fabrication was
performed at the Centre for Microfluidic Systems in Chemistry
and Biology. In vitro and animal studies were performed at the
Sunnybrook Research Institute.
References
1 R. D. Mosteller, N. Engl. J. Med., 1987, 317, 1098.
2 F. M. Watt, Science, 2014, 346, 937–940.
3 J. van Kuilenburg, M. A. Masen and E. van der Heide, Proc.
Inst. Mech. Eng., Part J, 2013, 227, 349–361.
4 S. MacNeil, Nature, 2007, 445, 874.
5 R. V. Shevchenko, S. L. James and S. E. James, J. R. Soc.,
Interface, 2010, 7, 229–258.
6 D. M. Supp and S. T. Boyce, Clin. Dermatol., 2005, 23,
403–412.
7 I. Jones, L. Currie and R. Martin, Br. J. Plast. Surg., 2002, 55,
185–193.
8 I. Yannas, J. Burke, D. Orgill and E. Skrabut, Science,
1982, 215, 174–176.
9 J. C. Gerlach, C. Johnen, C. Ottomann, K. Bräutigam, J.
Plettig, C. Belfekroun, S. Münch and B. Hartmann, Int. J.
Artif. Organs, 2011, 34, 271–279.
10 F. Navarro, M. Stoner, C. Park, J. Huertas, H. Lee, F. Wood
and D. Orgill, J. Burn Care Res., 2000, 21, 513.
11 V. Falanga, S. Iwamoto, M. Chartier, T. Yufit, J. Butmarc, N.
Kouttab, D. Shrayer and P. Carson, Tissue Eng., 2007, 13,
1299–1312.
12 D. R. Griffin, W. M. Weaver, P. O. Scumpia, D. Di Carlo and
T. Segura, Nat. Mater., 2015, 14, 737–744.
13 R. L. Truby and J. A. Lewis, Nature, 2016, 540, 371–378.
14 S. V. Murphy and A. Atala, Nat. Biotechnol., 2014, 32, 773–785.
15 W. J. Liu, Y. S. Zhang, M. A. Heinrich, F. De Ferrari, H. L.
Jang, S. M. Bakht, M. M. Alvarez, J. Z. Yang, Y. C. Li, G.
Trujillo-de Santiago, A. K. Miri, K. Zhu, P. Khoshakhlagh, G.
Prakash, H. Cheng, X. F. Guan, Z. Zhong, J. Ju, G. H. Zhu,
X. Y. Jin, S. R. Shin, M. R. Dokmeci and A. Khademhosseini,
Adv. Mater., 2017, 29.
16 C. Colosi, S. R. Shin, V. Manoharan, S. Massa, M. Costantini,
A. Barbetta, M. R. Dokmeci, M. Dentini and A.
Khademhosseini, Adv. Mater., 2016, 28, 677–684.
17 C. Norotte, F. S. Marga, L. E. Niklason and G. Forgacs,
Biomaterials, 2009, 30, 5910–5917.
18 D. B. Kolesky, K. A. Homan, M. A. Skylar-Scott and J. A.
Lewis, Proc. Natl. Acad. Sci. U. S. A., 2016, 113,
3179–3184.
19 D. B. Kolesky, R. L. Truby, A. S. Gladman, T. A. Busbee,
K. A. Homan and J. A. Lewis, Adv. Mater., 2014, 26,
3124–3130.
20 C. Di Bella, S. Duchi, C. D. O'Connell, R. Blanchard, C.
Augustine, Z. Yue, F. Thompson, C. Richards, S. Beirne, C.
Onofrillo, S. H. Bauquier, S. D. Ryan, P. Pivonka, G. G.
Wallace and P. F. Choong, J. Tissue Eng. Regener. Med.,
2018, 12(3), 611–621.
21 Y. L. Han, J. Hu, G. M. Genin, T. J. Lu and F. Xu, Sci. Rep.,
2014, 4, 4872.
22 H. Onoe, T. Okitsu, A. Itou, M. Kato-negishi, R. Gojo, D.
Kiriya, K. Sato, S. Miura, S. Iwanaga, K. Kuribayashi-
Shigetomi, Y. T. Matsunaga, Y. Shimoyama and S. Takeuchi,
Nat. Mater., 2013, 12, 584–590.
23 L. Leng, A. McAllister, B. Zhang, M. Radisic and A. Günther,
Adv. Mater., 2012, 24, 3650–3658.
24 M. Hospodiuk, M. Dey, D. Sosnoski and I. T. Ozbolat,
Biotechnol. Adv., 2017, 35, 217–239.
25 J. Malda, J. Visser, F. P. Melchels, T. Jungst, W. E. Hennink,
W. J. A. Dhert, J. Groll and D. W. Hutmacher, Adv. Mater.,
2013, 25, 5011–5028.
26 M. D. Balceniuk, N. A. Wingate, H. Krein, J. Curry, D.
Cognetti, R. Heffelfiner and A. Luginbuhl, Otolaryngol.–Head
Neck Surg., 2016, 155,76–80.
27 H. Yoshida, K. Hirozane and A. Kamiya, Biol. Pharm. Bull.,
2000, 23, 313–317.
28 Y. L. Jin, C. Liu, W. Chai, A. Compaan and Y. Huang, ACS
Appl. Mater. Interfaces, 2017, 9, 17456–17465.
29 N. Cubo, M. Garcia, J. F. Del Canizo, D. Velasco and J. L.
Jorcano, Biofabrication, 2016, 9, 015006.
30 V. Lee, G. Singh, J. P. Trasatti, C. Bjornsson, X. Xu, T. N.
Tran, S. S. Yoo, G. Dai and P. Karande, Tissue Eng., Part C,
2014, 20, 473–484.
31 L. J. Pourchet, A. Thepot, M. Albouy, E. J. Courtial, A. Boher,
L. J. Blum and C. A. Marquette, Adv. Healthcare Mater.,
2017, 6(4), 1601101.
32 L. Koch, A. Deiwick, S. Schlie, S. Michael, M. Gruene, V.
Coger, D. Zychlinski, A. Schambach, K. Reimers, P. M.
Vogt and B. Chichkov, Biotechnol. Bioeng., 2012, 109,
1855–1863.
33 S. Michael, H. Sorg, C. T. Peck, L. Koch, A. Deiwick, B.
Chichkov, P. M. Vogt and K. Reimers, PLoS One, 2013, 8,
e57741.
34 H. W. Kang, S. J. Lee, I. K. Ko, C. Kengla, J. J. Yoo and A.
Atala, Nat. Biotechnol., 2016, 34(3), 312–319.
35 S. Derler and L. C. Gerhardt, Tribol. Lett., 2012, 45,1–27.
36 B. Wijnen, E. J. Hunt, G. C. Anzalone and J. M. Pearce, PLoS
One, 2014, 9, e107216.
Lab on a Chip Paper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
Lab Chip This journal is © The Royal Society of Chemistry 2018
37 A. K. Au, W. Huynh, L. F. Horowitz and A. Folch, Angew.
Chem., Int. Ed., 2016, 55(12), 3862–3881.
38 G. K. Batchelor, An introduction to fluid dynamics, Cambridge
university press, 2000.
39 T. Braschler, A. Valero, L. Colella, K. Pataky, J. R. Brugger
and P. Renaud, Anal. Chem., 2011, 83, 2234–2242.
40 B. Thu, O. Gaserod, D. Paus, A. Mikkelsen, G. Skjak-Braek,
R. Toffanin, F. Vittur and R. Rizzo, Biopolymers, 2000, 53,
60–71.
41 J. B. Knight, A. Vishwanath, J. P. Brody and R. H. Austin,
Phys. Rev. Lett., 1998, 80, 3863.
42 A. Roshan, K. Murai, J. Fowler, B. D. Simons, V. Nikolaidou-
Neokosmidou and P. H. Jones, Nat. Cell Biol., 2016, 18,
145–156.
43 B. V. Slaughter, S. S. Khurshid, O. Z. Fisher, A. Khademhosseini
andN.A.Peppas,Adv. Mater., 2009, 21, 3307–3329.
44 G. F. Odland, Anat. Rec., 1950, 108, 399–413.
45 B. R. Downing, K. Cornwell, M. Toner and G. D. Pins,
J. Biomed. Mater. Res., Part A, 2005, 72,47–56.
46 A. Skardal, D. Mack, E. Kapetanovic, A. Atala, J. D. Jackson,
J. Yoo and S. Soker, Stem Cells Transl. Med., 2012, 1,
792–802.
Lab on a ChipPaper
Published on 11 April 2018. Downloaded by University of Toronto on 17/04/2018 08:58:31.
View Article Online
