Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces

journal homepage: www.elsevier.com/locate/colsurfb

Protocols

3D bioprinting and in vitro study of bilayered membranous construct with T

human cells-laden alginate/gelatin composite hydrogels
⁎ ⁎
Pengchao Liua,1, Hongzhou Shena,1, Yin Zhia, Jiawen Sia, , Jun Shia, , Lihe Guob,c,
⁎
Steve Guofang Shena,
a
Department of Oral and Craniomaxillofacial Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University College of Stomatology, Shanghai Jiao Tong
University School of Medicine, National Clinical Research Center for Oral Diseases, Shanghai Key Laboratory of Stomatology, Shanghai Research Institute of Stomatology,
Shanghai, 200011, People’s Republic of China
b
Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, 200031, People’s Republic of China
c
Sino-America United Stem Cell Research Center, Shanghai, 201203, People’s Republic of China

A R T I C LE I N FO A B S T R A C T

Keywords: Extrusion-based 3D bioprinting of cell-laden hydrogels is a potential technology for regenerative medicine,
Bioprinting which enables the fabrication of constructs with spatially defined cell distribution. However, the limited as-
Rheology sessment of rheological behaviors of hydrogel before printing is still a major issue for the advancement of 3D
Hydrogel bioprinting. In this work, we systematically investigated the rheological behaviors (i.e. viscosity, storage
Skin
modulus (G’), and loss modulus (G”)) of alginate/gelatin composite hydrogels first for 3D printing complex
Human cells
constructs. The rheological studies revealed that viscosity of alginate/gelatin hydrogels is temperature-depen-
dent and shear thinning. Sol-gel transition (intersection of G’ and G”) study provided indication for printing
temperature, which are in the range of 18.8 °C (H2/7.5) to 24.5 °C (H2/24.5). The alginate (2 wt%) /gelatin
(15 wt%) composite hydrogel sample was chosen to print the constructs and subsequent bioprinting. Complex
constructs (i.e. nose and ear) were obtained with high printing resolution (151 ± 13.04 μm) in a low tem-
perature (4 °C) chamber and crosslinking with 2 wt% CaCl2 subsequently without extra supports. Human am-
niotic epithelial cells (AECs) showed superior potential to differentiate into epithelial cells, while Wharton’s jelly
derived mesenchymal stem cells (WJMSCs) showed a superior angiogenic potential and fibroblastic phenotype.
For the in vitro study, AECs and WJMSCs as seed cells, encapsulated in alginate/gelatin composite hydrogels,
were bioprinted to form biomimetic bilayered membranous construct. High cell viability (> 95%) were ob-
served up to 6 days after printing. The presented 3D bioprinting of human AECs and WJMSCs-laden alginate/
gelatin composite hydrogels provides promising potentials for future skin tissue engineering.

1. Introduction
Skin, the largest organ of human body, plays a vital function in
sensation, protection and maintaining homeostasis. Skin is composed of
two primary layers: the epidermis (providing waterproofing and serving
as a barrier to infection) and the stromal layers (serving as a location for
the appendages of skin). Extensive and full-thickness skin wounds
caused by trauma, burns and chronic ulcers can be devastating to pa-
tients, even when treated [1,2]. Although various clinical treatments
have been shown to yield a reasonable outcome, up to 50% of the
wounds still fail to heal [3]. Three dimensional (3D) bioprinting, an
emerging additive manufacturing strategy, aims to produce re-
generative tissues and organs through layer-by-layer precise positioning
of living cells, biomaterials, and bioactive molecules [4,5]. In the past
decade, 3D bioprinting has made significant progress towards the re-
generation of transplantable tissues and even organs (e.g. skins, bones,
and ears) for restoring or repairing the damaged body [6–9]. Notably,
amniotic fluid-derived mesenchymal stem cells together with a fibrin/
collagen hydrogel were bioprinted into a customized membranous skin
substitute with high cell viability in vitro and promoted the skin
wounds healing [3]. More recently, a 3D bioprinted pigmented skin-like
membranous constructs were fabricated with very high resemblance
degree to native skin tissue in terms of the presence of well developed
stratified epidermal layers, pigment cells and the presence of a con-
tinuous layer of basement membrane proteins [10]. A novel 3D printing
platform was suggested for engineering a matured perfusable

⁎
Corresponding authors.
E-mail addresses: sjwlyl@163.com (J. Si), dr.shijun.oms@gmail.com (J. Shi), maxillofacsurg@163.com (S.G. Shen).
1
The first two authors contributed equally to this work.

https://doi.org/10.1016/j.colsurfb.2019.06.069
Received 22 January 2019; Received in revised form 27 June 2019; Accepted 28 June 2019
Available online 29 June 2019
0927-7765/ © 2019 Elsevier B.V. All rights reserved.
P. Liu, et al.
vascularized 3D human skin equivalent composed of epidermis, dermis,
and hypodermis, reflecting the actual complexity of native human skin
[11].
3D bioprinting technologies can be classified into three categories:
extrusion-based, inkjet-based, and laser-assisted bioprinting on the
basis of working principle [12]. To date, extrusion-based bioprinting
has been accepted as the most promising approach for generating skin
or soft tissues’ constructs. In this technology, hydrogels are not only
used as the building materials but also as cell delivery vehicles for their
unique physical, chemical and biological properties, which can emulate
the hygroscopic nature of extracellular matrix and provide an ideal
hydrate environment for cell proliferation and differentiation [13].
The ideal hydrogels for 3D bioprinting should fulfill three major
requirements: (i) good biocompatibility to realize high cell viability, (ii)
proper rheological behaviors during printing, (iii) rapid crosslinking
ability to retain the 3D structures after printing [14]. To achieve su-
perior biocompatibility, natural hydrogels are commonly used for 3D
bioprinting, such as hyaluronic acid [15], chitosan [16], collagen [17],
and alginate [18]. However, these bioprinted constructs are usually
featured with unsatisfied shape fidelity and tend to collapse due to
limited viscosity or delayed crosslinking. To overcome the current
limitations, composite hydrogels are often adopted. Several studies
have reported 3D printed alginate/gelatin composite hydrogels for the
application in tissue engineering, including aortic value conduit [19],
bone [20,21], and eye cornea [22]. Although alginate hydrogels show
more viscous behaviors than elastic throughout different concentra-
tions, while gelatin hydrogels are highly elastic rather than viscous.
Mixtures of alginate and gelatin are proven to be ideal for 3D bio-
printing due to a synergetic effect between them [14].
The rheological behaviors should be taken into account when de-
veloping or evaluating hydrogels for 3D printing [23]. Investigation of
rheological properties of hydrogels plays an important role in 3D bio-
printing, which remains poorly understood. Many studies did not take
rheology into consideration during bioink evaluation or mainly focused
on viscosity [24]. Since viscosity alone can not describe the complex
behavior of hydrogel-based bioinks during the printing process [25],
storage modulus (G’), loss modulus (G”) and the loss tangent (tan
δ = G”/G’) should also be studied. However, only limited literatures are
reported on the items recently [14].
In the last decade, increasing evidences have suggested that pla-
centa and umbilical cord, which are routinely discarded after delivery
as biological wastes, represent a promising reservoir of stem cells.
Particular attention has been focused on AECs and WJMSCs. The
human AECs and WJMSCs can be easily harvested from term amniotic
membrane and umbilical cord Wharton’s jelly, expanded and banked at
clinical-scale cell numbers without invasive procedures and ethical
concerns [26,27]. Moreover, both AECs and WJMSCs possess the ability
to promote wound healing, regenerate damaged tissues and evade im-
munosurveillance after transplantation [28]. These advantages make
AECs and WJMSCs fascinating sources of stem cells for skin tissue en-
gineering. Notably, several recent studies indicate that multipotent
human AECs are more prone to differentiate into keratinocytes while
placenta and umbilical cord derived mesenchymal stem cells are more
prone to differentiate into dermal fibroblasts and endothelial cells
[29–31]. Thus, human AECs and WJMSCs could respectively be used
for the fabrication of epidermis and stromal layers of skin substitutes.
In this present study, bilayered human AECs and WJMSCs-laden
alginate/gelatin composite hydrogels’ membranous construct was fab-
ricated by 3D bioprinting, which presented one potential method for
the preparation of skin substitutes. We investigated the rheological
behaviors (i.e. viscosity, G’, and G”) of alginate/gelatin composite hy-
drogels to obtain their optimized parameters for 3D bioprinting. 3D
constructs with high printing precision and shape fidelity were obtained
in a low temperature (4 °C) chamber and crosslinking with 2 wt% CaCl2,
subsequently. The mechanical properties of hydrogels were also studied
using a compression test method. Afterwards, AECs and WJMSCs were
1027
Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034
successfully isolated, characterized and bioprinted into biomimetic two
primary layers of skin: the epidermis and the stromal layers in vitro,
respectively. Cells’ viability and proliferation were investigated. To our
knowledge, this is the first report related to the use of 3D bioprinted
alginate/gelatin composite hydrogels to form bilayered constructs for
potential application in skin tissue engineering.
2. Materials and methods
2.1. Hydrogel fabrication
Sodium alginate and gelatin (Type A, 240 bloom derived from
porcine skin, Aladdin biochemical technology co., LTD, Shanghai)
powder was sterilized using ethylene oxide before the preparation of
composite hydrogels. In a typical experiment, 15 g gelatin was dis-
solved in 100 mL Dulbecco’s Modified Eagle Medium (DMEM, Gibco)/
F12, followed by the addition of 2.0 g alginate. The mixture was then
vigorously stirred under 90 °C to form a homogeneous precursor solu-
tion (denoted as H2/15). For comparison, four other formulations of
hydrogel precursors were prepared with different proportions of algi-
nate/gelatin: 2/7.5, 2/10, 2/12.5, and 2/17.5 (denoted as H2/7.5, H2/
10, H2/12.5, H2/17.5, respectively). All hydrogel precursors were
stored in sterile conditions at 4 °C.
2.2. Rheological analysis of alginate/gelatin composite hydrogels
Rheological analysis was carried out by a rotational rheometer
(Physica MCR 301, Anton Paar, Germany) equipped with two con-
centric parallel plates (diameter: 25 mm). The gap was 1 mm for all
measurements. Steady shear experiments were carried out in the shear
rate range of 0.01-300 s−1. Storage modulus (G’) and loss modulus (G”)
were assessed by oscillatory tests over a temperature ramp from 40 °C to
0 °C with cooling rate of 2 °C/min, and the oscillation frequency was
1 Hz. Prior to all rheological tests at each temperature, the fresh pre-
pared and degassed samples were held at that specific temperature for
2 min without pre-shearing or oscillating. To prevent dehydration
during rheological measurements, a thin layer of low-viscosity paraffin
oil was spread on the exposed surface of hydrogels.
2.3. Mechanical test
Mechanical properties of the hydrogel cylinders were measured
through uniaxial compression testing approach in materials testing
system (HY-940FS, Shanghai Hengyu Co., Ltd). After removing any air
bubbles by centrifugation, alginate/gelatin composite precursor solu-
tions were transferred into a designed mold made from polytetra-
fluoroethylene. By which, hydrogel cylinders with 8 mm in diameter
and 10 mm in height were fabricated. After storing in 4 °C for 1 h and
subsequent crosslinking in 2 wt% CaCl2 bath at room temperature for
30 min, the cylindrical hydrogels were set on the lower plate and
compressed by the upper plate at a strain rate of 5 mm min−1. The
compressive elastic modulus (E) was taken from the slope of the stress-
strain curve at a strain between 0.1 and 0.3 while the toughness was
determined from the area underneath the stress-strain curve. The de-
tailed test procedure refers to previous report [32].
2.4. Bioprinting system
An open-source 3D bioprinter (Fig. 1) with two-syringe was de-
signed and set up by Medical 3D Printing Innovation Research Center in
Shanghai Jiao Tong University. The syringes (50 mL) were equipped
with heatable print-heads to regulate thermo-behavior of hydrogels.
Coordinate-based printing paths were designed with software (Cura
15.02. 1 version). Hydrogel was heated up to flow, loaded into syringes
equipped with an iron nozzle. The hydrogel flowed to the tip driven by
the compressed air. The printer was located in a low temperature (4 °C)
P. Liu, et al.
Fig. 1. Details of the 3D bioprinter with a low temperature chamber: (A) outer appearance, (B) inner appearance, (C) syringe heater and tightened with a fixation
device, and (D) printed hydrogel constructs on the substrate in the low temperature (4 °C) chamber.
chamber.
2.5. Isolation and culture of human AECs and WJMSCs
The appropriate use of human tissue and cells in this study was
approved by the institutional Patients and Ethics Committee. Human
placenta, umbilical cord, and peripheral blood were collected from
healthy donors. The human AECs, WJMSCs were isolated as previously
described [33]. Briefly, the amniotic membrane and Wharton’s jelly
were cut into small pieces and cultured in α-MEM medium (Hyclone,
USA) supplemented with 20% fetal bovine serum (FBS; Gibco, China) at
37 °C, 5% carbon dioxide, and 95% air humidity. The culture medium
was changed every three days until cells migrated onto the culture dish.
The human peripheral blood mononuclear cells (PBMCs) were isolated
from peripheral blood samples by density gradient centrifugation with
Ficoll-Paque plus (GE Healthcare, USA). The isolated AECs, WJMSCs
were cultured and expanded in DMEM medium (Invitrogen, China)
supplemented with 20% FBS and 1% penicillin-streptomycin (In-
vitrogen, China); the isolated PBMCs were cultured in RPMI 1640
medium (Invitrogen, China) supplemented with 2 mM L-Glutamine
(Invitrogen, China), 2.5% HEPES, 20% FBS and 1% penicillin-strepto-
mycin. Morphology of both human AECs and WJMSCs was photo-
graphed using a light microscope (Axio Scope A1, Zeiss, Germany).
2.6. Bioprinting of bilayered membranous construct and cell viability
assessment
Before bioprinting, human AECs and WJMSCs were respectively
washed twice with PBS and fluorescently labeled with 10 μmol Dil and
Dio (Beyotime, China) for 5 min. The labeled and unlabeled AECs and
WJMSCs were separately centrifuged and resuspended in H2/15 pre-
cursor solution at a density of 1 × 106 cells/ mL and placed on ice.
Then H2/15 precursor solutions containing AECs and WJMSCs were
placed into two 50 ml syringes, which were put into the heating tube of
3D printer. The heat temperature, air pressure, inner diameter of
nozzle, the moving velocity of stage were 30 °C, 0.2 MPa, 0.33 μm, and
7 mm/s, respectively. According to designed models and path, four
groups of cell-laden membranes: single-layered AECs laden group;
single-layered WJMSCs laden group; bilayered AECs and WJMSCs laden
group; bilayered Dil-labeled-AECs and Dio-labeled-WJMSCs laden
group were printed successfully in the low temperature (4 °C) chamber.
After bioprinting, the bilayered Dil-labeled-AECs and Dio-labeled-
WJMSCs laden membranous constructs were immediately scanned by a
1028
Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034
confocal microscope (Nikon, Japan). The images were acquired and
analysed using the NIS Viewer software (Nikon, Japan). The single-
layered AECs-laden membranes, single-layered WJMSCs-laden mem-
branes, bilayered-layered AECs and WJMSCs-laden membranes were
collected and cultured in 24-well plates for up to 6 days. As a control,
human AECs and BM-MSCs were seeded at 1 × 105 cells/ well into 24-
well plates and cultured for 5 days accordingly. Cell proliferation of
each group (Cell density: 1 × 105 cells/ well) at day 2, 4 and 6 was
examined using a Cell Counting Kit-8 (CCK-8, Dojindo, Japan) ac-
cording to the manufacturer’s protocol. Experiments were performed in
triplicate for each group. For further visualisation of live and dead cells,
the bilayered AECs and WJMSCs laden membranes (Cell density:
1 × 105 cells/ well) at day 2, 4 and 6 were washed with PBS and in-
cubated with 2 μmol calcein-AM and 4 μmol propidium-iodide (PI)
using a calcein-AM/PI double stain kit (Dojindo Laboratories, Japan)
for 20 min. Then the samples were thoroughly rinsed and visualized
using a fluorescence microscope (DP71, Olympus, Japan).
2.7. Statistical analysis
All measurements were collected and expressed as mean ±
standard deviation (SD). Numerical data were analyzed using two-way
ANOVA and Student’s t-test. A p value < 0.05 was considered statisti-
cally significant. SPSS 16.0 (IBM, USA) software and Graphpad prism
5.0 (GraphPad Software, USA) software were utilized to analyze and
demonstrate the statistical significance of the assays. For microarray
analysis, comparative genes expression analyses were performed with
the Genesrping (Agilent Technologies, USA) software. Differentially
expressed genes of human AECs and WJMSCs were identified through
fold change. The threshold set for up- and down-regulated genes in per
pair-wise comparison was absolute mean fold change ≥ 3 and
p < 0.05.
3. Results and discussion
3.1. Gelation mechanism of alginate/gelatin composite hydrogels during 3D
printing
High precision and shape fidelity are usually great challenges for 3D
printing of hydrogels, especially for natural polymers derived hydro-
gels. It is because that they can not usually rapidly gel during the 3D
printing process to support themselves and prevent the deformation or
collapse [34]. To address this, alginate/gelatin precursor solutions in
P. Liu, et al. Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034

Fig. 2. Rheological properties as a function of

temperature for alginate/gelatin hydrogels
with different formulations: H2, H2/7.5, H2/
10, H2/12.5, H2/15 and H2/17.5. (A) Curves
of viscosity versus temperatures from 0 to
40 °C; (B) The magnified curves of alginate/
gelatin hydrogels in Stage II; (C) G’ and G”
versus temperatures from 0 to 40 °C; (D) Sol-
gel transition temperatures of alginate/gelatin
hydrogels.

Table 1
Results of uniaxial compression testing on hydrogels.
Samples Maximum stress Maximum Elastic modulus Toughness (kJ/
(kPa) strain (100%) (kPa) m3)

H2/7.5 438.7 ± 43.6 61.9 ± 2.0 280.0 ± 65.7 76.2 ± 8.8

H2/10 359.0 ± 48.1 64.4 ± 2.3 230.8 ± 41.4 69.1 ± 3.4
H2/12.5 372.0 ± 74.7 67.3 ± 4.2 199.3 ± 14.5 63.6 ± 13.6
H2/15 554.5 ± 76.1 73.1 ± 2.7 206.1 ± 11.5 106.4 ± 13.3
H2/17.5 517.5 ± 2.1 72.3 ± 1.5 192. 3 ± 3.9 97.8 ± 3.2

our study were printed in low temperature (4 °C) chamber first, fol-
lowed by Ca2+ crosslinking after printing. Low temperature leads to
instantaneous gelation and primary gel strand stability of composite
hydrogels, which is due to form triple-helices that are stabilized by
hydrogen bonds in gelatin. Afterwards, the printed whole construct is
immersed in a CaCl2 bath, which crosslinks the alginate present in
composite hydrogels. Long-term stability is ensured finally by the
crosslinked alginate (Scheme 1, Supporting Information). Printing in a
low temperature chamber is a key point for the preparation of complex
constructs with large size rather than only onto a cooled substrate.

3.2. Rheological behaviors of alginate/gelatin composite hydrogels

Understanding the rheological behaviors of hydrogel is a pre-

Fig. 3. (A) Viscosity variation as a function of shear rate for H2/15 at different
temperatures (Inset shows the magnified curves of viscosity versus shear rate);
(B) Log-log plot of G′, G′′ vs. shear rate for H2/15 at different temperatures. The
data are shifted along the vertical axis by 10a to avoid overlapping.
requisite for extrusion-based 3D bioprinting. For alginate/gelatin
system, effects of temperatures, concentration ratios and shear rates on
rheological behaviors (i.e. viscosity, G’, and G”) were investigated
systematically, respectively.
Temperature, determines the initial process of gelation, were stu-
died first. Viscosity of pure alginate solution (2 wt%) and alginate/ge-
latin composite hydrogels decreased with increasing temperatures
(Fig. 2A and B), which was common for polymer solutions. For pure
alginate solution, the dependence of viscosity on temperature is almost
negligible. So high printing resolution of alginate can not be achieved
by adjusting temperatures. For alginate/gelatin composite hydrogels
with different concentration ratios, the declining process of viscosity
can be divided into two stages. In stage I, the declining rate of viscosity

1029
P. Liu, et al. Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034

Fig. 4. Characterization of human AECs and WJMSCs. (A) AECs display homogeneous cobblestone-like morphology similar to epithelial cells and are positive for
cytokeratin-19 and E-Cadherin, while human WJMSCs display characteristic spindle-shaped fibroblast morphology and are negative for cytokeratin-19 and E-
Cadherin. (B) Comparing to the proliferation of PBMCs alone (Negative group) and PBMCs stimulated by equal amounts of allogenic PBMCs (Positive group), both
AECs and WJMSCs inhibit the MLR in a cell dose-dependent manner, ☆: p < 0.05. (C) Flow cytometric analysis demonstrates that AECs and WJMSCs are positive for
SSEA4, CD44, CD90, CD105, HLA-abc and negative for CD34, CD45 or HLA-dr.

was very fast in lower temperature range (< 25 °C). The apparent
viscosity was higher with more gelatin amount. The experimental data
were linearly fitted (Figure S1 and Table S1, Supporting Information).
For all the samples, R2 is greater than 0.999, which proved that the
linear model was very suitable to describe the rheological property of
alginate/gelatin composite hydrogels in stage I. In stage II, the viscosity
of all samples changed slightly. 3D printing should be conducted in
stage II.
Effects of temperatures on G’ and G” are further studied to illumi-
nate the complex behavior of hydrogels, which were reported in limited
literatures [14]. For pure 2% alginate solution, G′′ was always higher
than G′ (Fig. 2C), which indicated no sol-gel transition in the range of
0–40 °C. Poor precision and shape fidelity were usually found for
printing of pure alginate solutions, although some researchers made
attempts such as increasing the viscosity [35] or extruded with ion
crosslinker (such as Ca2+) simultaneously [36]. Sol-gel transition
temperature (Gelation point) is defined as the intersection of G’ and G”
curves during temperature decrease at an oscillatory measurement. For
the alginate/gelatin solutions, the sol-gel transition temperatures in-
creased from 18.8 to 24.5 °C with the increase of gelatin concentration
from 6.85 (H2/7.5) to 14.64 wt % (H2/17.5) (Fig. 2D). Above sol-gel
transition temperatures, alginate/gelatin solutions exhibit more viscous
than elastic, which is suitable for printing. So for the composite hy-
drogels, printing temperatures should be set in the near regions above
sol-gel transition temperatures, respectively.
To study the effect of shear rate on rheological behavior of alginate/
gelatin composite hydrogels, viscosity and dynamic modulus (i.e. G’
and G”) of H2/15 were studied. Fig. 3A shows that viscosity decreases
with increasing shear rate, indicating a shear-thinning behavior. Shear
thinning happens when the hydrogels are extruded from a syringe to a
narrow nozzle during the 3D printing process. As a result, a reducing
viscosity of the hydrogels is in favour of extruding easily through a fine
nozzle to protect the cells against the shear stress [34].
Fig. 3B shows the effect of shear rate on G’ and G”. We observed that
G’ was always higher than G” below 25 °C, indicating a dominant elastic
behavior of composite hydrogels. No intersection between the G’ and G”
further excluded the sol-gel transitions. In this temperature region
(< 25 °C), H2/15 with gel state was not suitable for extrusion. How-
ever, sol-gel transition happens above 30 °C, for the observation of cross
section between G′ and G′′ curves. And the sol-gel transition points shift
to high shear rates zone with increasing temperatures.
3.3. 3D printing of alginate/gelatin composite hydrogels
H2/15 sample was chosen by 3D printing for different constructs
(Figure S2, Supporting Information). By adjusting the printing para-
meters, the sample can be easily printed to form not only simple con-
structs (tubular, cylindrical, and cuboid), but also complex constructs

1030
P. Liu, et al. Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034

Fig. 5. Microarray analysis of global gene expression of human AECs and WJMSCs. (A) Hierarchical cluster analysis of the 911 differentially expressed genes resulted
in two main clusters, AECs and WJMSCs respectively (Red: up-regulated genes; green: down-regulated genes). (B) Gene ontology enrichment analysis of the dif-
ferentially expressed genes. (C) Real-time PCR verifies the microarray results and shows that FGF-7, CXCL-5, MMP-2, VEGF-A are highly enriched in WJMSCs
comparing to AECs (☆: p < 0.05) (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

1031
P. Liu, et al. Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034

Table 2 wound healing. According to previous studies, human AECs were

Microarray results of differentially expressed genes in WJMSCs associated with shown to express EGF, KGF, PDGF-B and HGF to improve wound
skin regeneration and angiogenesis comparing to AECs. healing and differentiate into keratinocytes-like cells and stratified
Gene Symbol Gene Accession Fold change (abs) Regulation epithelial cells, while MSCs may elicit the majority of their wound
healing properties by secreting growth factors and cytokines such as
Keratin-6 NM_005554 708.3 down VEGF, KGF, FGF and TGF-β [40,41]. However, their detailed me-
Keratin-8 NM_001256282 379.7 down
chanism and differences are still unclear. In this study, the comparison
Keratin-24 NM_019016 357.7 down
Keratin-5 NM_000424 275.6 down of global gene expression of AECs and WJMSCs revealed 911 differ-
Keratin-19 NM_002276 145.7 down entially expressed genes. Hierarchical cluster analysis of the differen-
Keratin-17 NM_000422 138.7 down tially expressed genes resulted in two main clusters: AECs and WJMSCs,
Keratin-18 NM_000224 130.5 down
respectively (Fig. 5A). Gene ontology enrichment analysis of the dif-
LAMA-4 NM_001105206 73.4 up
Nidogen-1 NM_002508 4.5 up
ferentially expressed genes revealed a more increased expression of
FGF-2 NM_002006 3.4 up genes involved in cell junction assembly, cell adhesion, epidermis de-
FGF-7 NM_002009 5.8 up velopment, wound healing and VEGF signaling pathway, etc (Fig. 5B).
TGF-b1 NM_000660 2.2 up While AECs expressed high level of keratins, the up-regulated genes in
CXCL-5 NM_002994 19.7 up
WJMSCs associated with skin regeneration and angiogenesis included
MMP-2 NM_001127891 27.8 up
VEGF-A NM_001025366 5.6 up vascular endothelial growth factor (VEGF)-A, VEGF-C, fibroblast
VEGF-C NM_005429 7.0 up growth factor (FGF)-2, FGF-7, MMP-2, chemokine C-X-C motif ligand
(CXCL)-5, Nidogen-1, LAMA-4 and TGF-β1 (Table 2). Notably, these
genes have been widely reported to play substantial roles in new vessel
such as ear and nose without any supports. These experiment results formation and cutaneous wound healing [42–44]. Of special interest,
confirmed that our materials’ system can be used for printing of refine VEGF, a family member of platelet-derived growth factors, promotes
structures with high fidelity. During the printing process, we find that the endothelial cells proliferation, collagen deposition and angiogenesis
the sol-gel transition process is needed to control precisely. By which, in vitro and in vivo [45]. FGF-7, also known as keratinocyte growth
the materials can be extruded with gel strand rather than droplets. H2/ factor, specifically promotes granulation tissue formation, collagen
15 sample was printed through a nozzle with inner diameter of 110 μm. deposition and skin re-epithelialization during wound healing [43,44].
The obtained gel strand was with diameter at 151 ± 13.04 μm (Figure Moreover, CXCL-5 and MMP-2, important stimulators of inflammatory
S3, Supporting Information). Our results showed that alginate/gelatin cell recruitment, extracellular matrix remodelling, angiogenesis and re-
composite hydrogels could be printed to form constructs with high re- epithelialization of the wounds, are both found to be significantly
solution without any supports. higher in cutaneous wound tissue treated with MSCs [42,46]. Thus,
real-time PCR was further used to verify the microarray results for FGF-
3.4. Mechanical properties 7, CXCL-5, MMP-2, VEGF-A and showed that these four crucial genes
for cutaneous wound healing are highly enriched in WJMSCs com-
Compression testing was carried out to study the mechanical paring to AECs, indicating a superior angiogenic potential of WJMSCs
properties of the prepared composite hydrogel cylinders directly. The (Fig. 5C).
compression testing results including the maximum stress and strain,
and toughness were showed in Table 1. As can be seen from Table 1, 3.6. Bioprinting and cell viability assessments of human AECs and
H2/15 showed the highest maximum stress, strain, and toughness, WJMSCs-laden bilayered membranous construct
which are 554.5 ± 76.1 kPa, 73.1 ± 2.7%, and 106.4 ± 13.3 kJ/m3,
respectively (Figure S4, Supporting Information). The wound-healing process is a complex alignment of various bio-
logical events and coordinated collaboration of several different cells
3.5. Phenotypes of human AECs and WJMSCs and tissues. Among them, keratinocyte differentiation, fibroblast pro-
liferation, epithelial-mesenchymal interaction and neovascularization
The selection of cell types and establishment of consistent cell banks are most important for wound healing. Notably, in vitro and in vivo
are crucial steps to develop a cell-based treatment strategy. In parti- studies have demonstrated that human AECs show high proliferation
cular, human AECs and WJMSCs with confirmed multi differentiation activity and superior potential to differentiate into epithelial cells,
capacity, immunomodulatory property, immunoprivileged character- while WJMSCs are more prone to exert paracrine functions and dif-
istic, non-tumorigenicity, high yielding expansion availability and low ferentiate into dermal fibroblasts and vessel forming endothelial cells
processing costs show most promising prospects for wound repair and [37,39,40,45,46]. Moreover, MSCs have been reported to exert para-
clinical cell therapy [37]. AECs displayed homogeneous cobblestone- crine functions which can result in accelerated proliferation of human
like morphology similar to epithelial cells while WJMSCs displayed AECs, indicating a crucial role of epithelial-mesenchymal interaction
characteristic fibroblast-like morphology. Moreover, AECs were posi- during cutaneous wound healing [47]. Our results together with these
tive for cytokeratin-19 and E-Cadherin while WJMSCs were negative reports supported the hypothesis that human AECs and WJMSCs could
for these markers (Fig. 4A). Notably, the expression of E-cadherin and respectively be used for the reconstruction of the epidermis and stromal
CK19 have been used as epithelial stem cell markers in the skin and layers of skin. Thus, the AECs and WJMSCs were separately en-
shows a critical function in embryo development, which indicates a capsulated in the H2/15 alginate/gelatin composite hydrogel and bio-
superior potential of AECs in keratinocyte differentiation and epidermal printed into the upper and lower layer of the membrane, respectively
regeneration [38,39]. In agreement with previous reports, both AECs (Figure S5, Supporting Information). Fig. 6A shows the confocal volume
and WJMSCs could inhibit the allogeneic PBMCs proliferation in a cell view and cross-sectional view of the bioprinted bilayered membranes.
dose-dependent manner and were positive for SSEA4, CD44, CD90, The uniform distribution of Dil-labeled-AECs (red fluorescence) and
CD105, HLA-abc, while both cell types are negative for CD34, CD45 or Dio-labeled-WJMSCs (green fluorescence) were observed in the upper
HLA-dr (Fig. 4B, C). The immunoprivileged characteristics of AECs and layer and lower layer, respectively. The CCK-8 assay showed that both
WJMSCs are therefore releasing the risks of rejection upon transplan- AECs and WJMSCs proliferated properly from day 2 to day 6. Bio-
tation. printing did not alter the proliferation activity of the two cell types at
Accumulating studies have shown that both human AECs and MSCs each predetermined time point (Fig. 6B). Cell viabilities within the
implantation can significantly promote skin re-epithelialization and bioprinted bilayered membranes were further confirmed using the live

1032
P. Liu, et al. Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034

Fig. 6. (A) The confocal volume view and

cross-sectional view of the bioprinted bilayered
membrane show the uniformly distribution of
Dil-labeled-AECs (red fluorescence) and Dio-
labeled-WJMSCs (green fluorescence). (B) The
CCK-8 assay shows that human AECs, WJMSCs,
single-layered AECs-laden membrane (bio-
printed AECs), single-layered WJMSCs-laden
membrane (bioprinted WJMSCs), and the bi-
layered-layered AECs and WJMSCs-laden
membrane (bioprinted AECs and WJMSCs)
proliferate properly from day 2 to day 6. (C)
Live and dead cell staining displays a high cell
viability (＞95%) of the bioprinted bilayered
membranous construct at days 2, 4 and 6 (For
interpretation of the references to colour in this
figure legend, the reader is referred to the web
version of this article.).

and dead cell staining kit. High cell viabilities (> 95%) are maintained
at day 2, day 4 and day 6 (Fig. 6C).
4. Conclusions
In summary, we systematically studied the rheological behaviors of
alginate/gelatin composite hydrogels before printing and reported the
3D bioprinting of bilayered membranous constructs using human AECs
and WJMSCs as seed cells. Alginate/gelatin composite hydrogels went
through a two-step gelation mechanism: gelatin’s gelation induced by
low temperature and Ca2+ crosslinking of alginate. The rheological
studies revealed that viscosity of alginate/gelatin hydrogels varied in a
temperature-dependent manner. Proper printing temperatures were
obtained on the basis of sol-gel transition’s (intersection of G’ and G”)

1033
P. Liu, et al.
research. Shear thinning behaviors of alginate/gelatin hydrogels were
beneficial to printing. The alginate (2 wt%) /gelatin (15 wt%) compo-
site hydrogel sample was chosen to print the constructs with high
printing precision and shape fidelity. For the in vitro study, human
AECs showed a superior epithelial cells phenotype, while WJMSCs
showed a superior angiogenic potential and fibroblastic phenotype.
High cell viability were also observed up to 6 days after printing
(> 95%). The presented fabrication system and strategy of human AECs
and WJMSCs-laden alginate/gelatin composite hydrogels provides
promising potentials for future skin engineering by 3D bioprinting.
Declaration of Competing Interest
There are no conflicts of interest for this work.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (No.81570947 and No.81600827). P. Liu ac-
knowledges the China Postdoctoral Science Foundation (Grant No.:
2017M621496). Also we would like to thank all the donors as well as
the hospital support staff for their kindly cooperation.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.colsurfb.2019.06.069.
References
[1] J.S. Chen, V.W. Wong, G.C. Gurtner, Front. Immunol. 3 (2012) 192.
[2] N. Hirt-Burri, A.A. Ramelet, W. Raffoul, A. de Buys Roessingh, C. Scaletta,
D. Pioletti, L.A. Applegate, ISRN Dermatol. 2011 (2011) 1–16.
[3] A. Skardal, D. Mack, E. Kapetanovic, A. Atala, J.D. Jackson, J. Yoo, S. Soker, Stem
Cells Transl. Med. 1 (2012) 792–802.
[4] J. Groll, T. Boland, T. Blunk, J.A. Burdick, D.-W. Cho, P.D. Dalton, B. Derby,
G. Forgacs, Q. Li, V.A. Mironov, L. Moroni, M. Nakamura, W. Shu, S. Takeuchi,
G. Vozzi, T.B.F. Woodfield, T. Xu, J.J. Yoo, J. Malda, Biofabrication 8 (2016)
013001.
[5] S.V. Murphy, A. Atala, Nat. Biotechnol. 32 (2014) 773–785.
[6] W.L. Ng, S. Wang, W.Y. Yeong, M.W. Naing, Trends Biotechnol. 34 (2016) 689–699.
[7] Stephanie Knowlton, Sevgi Onal, H.Chu Yu, J.Jean Zhao, Savas Tasoglu, Trends
Biotechnol. 33 (2015) 504–513.
[8] S. Jana, A. Lerman, Biotechnol. Adv. 33 (2015) 1503–1521.
[9] R. Tevlin, A. Mcardle, D. Atashroo, G.G. Walmsley, K. Senarathyapa, E.R. Zielins,
K.J. Paik, M.T. Longaker, D.C. Wan, Title, J. Dent. Res. 93 (2014) 1187–1195.
[10] W.L. Ng, J.T.Z. Qi, W.Y. Yeong, M.W. Naing, Biofabrication 10 (2018) 025005.
[11] B.S. Kim, G. Gao, J.Y. Kim, D.-W. Cho, Adv. Healthc. Mater. 8 (2019) 1801019.
[12] F. Pati, J. Gantelius, H.A. Svahn, Angew. Chem. 55 (2016) 4650–4665.
[13] F.P.W. Melchels, M.A.N. Domingos, T.J. Klein, J. Malda, P.J. Bartolo,
1034
Colloids and Surfaces B: Biointerfaces 181 (2019) 1026–1034
D.W. Hutmacher, Prog. Polym. Sci. 37 (2012) 1079–1104.
[14] T. Gao, G.J. Gillispie, J.S. Copus, A.P.R. Kumar, Y.-J. Seol, A. Atala, J.J. Yoo,
S.J. Lee, Biofabrication 10 (2018) 034106.
[15] I. Donderwinkel, J.C.M. van Hest, N.R. Cameron, Polym. Chem. 8 (2017)
4451–4471.
[16] Q. Gu, E. Tomaskovic-Crook, G.G. Wallace, J.M. Crook, Adv. Healthc. Mater. 6
(2017) 1700175.
[17] A. Bell, M. Kofron, V. Nistor, Biofabrication 7 (2015) 035007.
[18] R. Xiong, Z. Zhang, W. Chai, Y. Huang, D.B. Chrisey, Biofabrication 7 (2015)
045011.
[19] B. Duan, L.A. Hockaday, K.H. Kang, J.T. Butcher, J. Biomed. Mater. Res. Part A
101A (2013) 1255–1264.
[20] Y. Luo, Y. Li, X. Qin, Q. Wa, Mater. Des. (2018) 12–19.
[21] X. Wang, P. Lu, Y. Song, Y. Sun, Y. Wang, Y. Wang, RSC Adv. 6 (2016) 6832–6842.
[22] A. Isaacson, S. Swioklo, C.J. Connon, Exp. Eye Res. 173 (2018) 188–193.
[23] J. Malda, J. Visser, F.P. Melchels, T. Juengst, W.E. Hennink, W.J.A. Dhert, J. Groll,
D.W. Hutmacher, Adv. Mater. 25 (2013) 5011–5028.
[24] A.L. Rutz, K.E. Hyland, A.E. Jakus, W.R. Burghardt, R.N. Shah, Adv. Mater. 27
(2015) 1607–1614.
[25] J. Gohl, K. Markstedt, A. Mark, K. Hakansson, P. Gatenholm, F. Edelvik,
Biofabrication 10 (2018) 034105.
[26] B. Farhadihosseinabadi, M. Farahani, T. Tayebi, A. Jafari, F. Biniazan,
K. Modaresifar, H. Moravvej, S. Bahrami, H. Redl, L. Tayebi, H. Niknejad, Artif.
Cells Nanomed. Biotechnol. (2018) 1–10.
[27] J.E. Davies, J.T. Walker, A. Keating, Stem Cells Transl. Med. 6 (2017) 1620–1630.
[28] L. Pu, M. Meng, J. Wu, J. Zhang, Z. Hou, H. Gao, H. Xu, B. Liu, W. Tang, L. Jiang,
Y. Li, Stem Cell Res. Ther. 8 (2017) 72.
[29] L. Xu, J. Zhou, J. Liu, Y. Liu, L. Wang, R. Jiang, Z. Diao, G. Yan, B. Peault, H. Sun,
L. Ding, Stem Cells Int. 2017 (2017) 3175748.
[30] S.C. Yu, Y.Y. Xu, Y. Li, B. Xu, Q. Sun, F. Li, X.G. Zhang, Eur. Rev. Med. Pharmacol.
Sci. 19 (2015) 4627–4635.
[31] E. Jin, T.H. Kim, S. Han, S.W. Kim, J. Tissue Eng. Regener. Med. 10 (2016)
613–622.
[32] Z. Zeng, H. Wang, Y. Morsi, X. Mo, Colloids Surf. B 161 (2018) 94–102.
[33] S. Jiawen, Z. Jianjun, D. Jiewen, Y. Dedong, Y. Hongbo, S. Jun, W. Xudong,
S.G. Shen, G. Lihe, Stem Cells Transl. Med. 3 (2014) 1504–1513.
[34] H. Li, C. Tan, L. Li, Mater. Des. 159 (2018) 20–38.
[35] K. Markstedt, A. Mantas, I. Tournier, H.M. Avila, D. Hagg, P. Gatenholm,
Biomacromolecules 16 (2015) 1489–1496.
[36] Q. Gao, Y. He, J.-z. Fu, A. Liu, L. Ma, Biomaterials 61 (2015) 203–215.
[37] J.W. Si, X.D. Wang, S.G. Shen, World J. Stem Cells 7 (2015) 149–159.
[38] W. Kong, S. Li, M.T. Longaker, H.P. Lorenz, Exp. Cell Res. 314 (2008) 1529–1539.
[39] L.W. Jiang, H. Chen, H. Lu, J. Dermatol. Sci. 81 (2016) 26–34.
[40] R. Mahmood, M.S. Choudhery, A. Mehmood, S.N. Khan, S. Riazuddin, Stem Cells
Int. 2015 (2015) 841062.
[41] A.N. Smith, E. Willis, V.T. Chan, L.A. Muffley, F.F. Isik, N.S. Gibran, A.M. Hocking,
Exp. Cell Res. 316 (2010) 48–54.
[42] F. Jiang, J. Ma, Y. Liang, Y. Niu, N. Chen, M. Shen, Biomed. Res. Int. 2015 (2015)
324014.
[43] Y. Qu, C. Cao, Q. Wu, A. Huang, Y. Song, H. Li, Y. Zuo, C. Chu, J. Li, Y. Man, J.
Tissue Eng. Regener. Med. 12 (2018) 1508–1518.
[44] C.Y. Lee, C.Y. Yang, C.C. Lin, M.C. Yu, S.J. Sheu, Y.H. Kuan, Mol. Med. Rep. 17
(2018) 8047–8052.
[45] A. Zhou, X. Zheng, L. Yu, M. Quan, X. Shao, Z. Jiang, Exp. Ther. Med. 10 (2015)
2161–2168.
[46] S.S. Mehanni, N.F. Ibrahim, A.R. Hassan, L.A. Rashed, Int. J. Stem Cells 6 (2013)
45–54.
[47] Q. Wu, T. Fang, H. Lang, M. Chen, P. Shi, X. Pang, G. Qi, Int. J. Mol. Med. 39 (2017)
918–926.