Assessment of hydrogels for bioprinting of endothelial cells

Leo Benning,1 Ludwig Gutzweiler,2 Kevin Tro€ ndle,2 Julian Riba,2 Roland Zengerle,2,3,4
2 2
Peter Koltay, Stefan Zimmermann, G. Bjo € rn Stark,1 Gu € nter Finkenzeller1
1
Department of Plastic and Hand Surgery, Faculty of Medicine, Medical Center, University of Freiburg, Freiburg, Germany
2
Laboratory for MEMS Applications, IMTEK, Department of Microsystems Engineering, University of Freiburg,
Georges-Koehler-Allee 103, Freiburg 79110, Germany
3
Hahn-Schickard, Georges-Koehler-Allee 103, Freiburg 79110, Germany
4
FIT – Freiburg Centre for Interactive Materials and Bioinspired Technologies, University of Freiburg, Georges-Koehler-Allee
105, Freiburg 79110, Germany

Received 18 January 2017; revised 13 October 2017; accepted 2 November 2017

Published online 29 November 2017 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbm.a.36291

Abstract: In tissue engineering applications, vascularization
can be accomplished by coimplantation of tissue forming cells
and endothelial cells (ECs), whereby the latter are able to form
functional blood vessels. The use of three-dimensional (3D)
bioprinting technologies has the potential to improve the clas-
sical tissue engineering approach because these will allow the
generation of scaffolds with high spatial control of endothelial
cell allocation. This study focuses on a side by side compari-
son of popular commercially available bioprinting hydrogels
(Matrigel, fibrin, collagen, gelatin, agarose, Pluronic F-127,
alginate, and alginate/gelatin) in the context of their physico-
chemical parameters, their swelling/degradation characteris-
tics, their biological effects on vasculogenesis-related EC
parameters and their printability. The aim of this study was to
identify the most suitable hydrogel or hydrogel combination
for inkjet printing of ECs to build prevascularized tissue con-
structs. Most tested hydrogels displayed physicochemical
characteristics suitable for inkjet printing. However, Pluronic
F-127 and the alginate/gelatin blend were rapidly degraded
when incubated in cell culture medium. Agarose, Pluronic F-
127, alginate and alginate/gelatin hydrogels turned out to be
unsuitable for bioprinting of ECs because of their non-
adherent properties and/or their incapability to support EC
proliferation. Gelatin was able to support EC proliferation and
viability but was unable to support endothelial cell sprouting.
Our experiments revealed fibrin and collagen to be most suita-
ble for bioprinting of ECs, because these hydrogels showed
acceptable swelling/degradation characteristics, supported
vasculogenesis-related EC parameters and showed good print-
ability. Moreover, ECs in constructs of preformed spheroids
survived the printing process and formed capillary-like cords.
C 2017 Wiley Periodicals, Inc. J Biomed Mater Res Part A: 106A: 935–
V
947, 2018.
Key Words: 3D bioprinting, hydrogel, endothelial cell, vascu-
larization, tissue engineering

€ ndle K, Riba J, Zengerle R, Koltay P, Zimmermann S, Stark GB,

How to cite this article: Benning L, Gutzweiler L, Tro
Finkenzeller G. 2018. Assessment of hydrogels for bioprinting of endothelial cells. J Biomed Mater Res Part A
2018:106A:935–947.

INTRODUCTION in vivo upon implantation into immunocompromised ani-

In tissue engineering applications, a rapid neovasculariza-
tion of implanted grafts is essential to ensure the survival
of cells in the early postimplantational phase.1
The long-term survival of engineered grafts depends
strongly on oxygen delivery and nutrient exchange. These
factors are mainly provided by the vasculature. Therapeutic
approaches to induce vascularization comprise the use of
angiogenic growth factors such as VEGF or bFGF, delivered
as recombinant proteins or by gene therapeutic strategies2,3
and cell-based therapies using differentiated endothelial
cells such as human umbilical vein endothelial cells
(HUVECs) or endothelial progenitor cells. In those cell-based
therapies, it has been shown that human endothelial cells
were able to form functional long-lasting blood vessels
mals.4–7 In classical tissue engineering applications endothe-
lial cells are ex vivo seeded in biological matrices such as
fibrin or collagen and then implanted in experimental ani-
mals for the generation of a neovasculature. However, by
using this approach, endothelial cells (ECs) (either in form of
single cells or in form of preformed spheroids) are randomly
distributed within the matrices without control of the spatial
distribution of ECs and therefore without the opportunity to
develop predefined regions of prevascularization within the
biomatrix. More recently, three-dimensional (3D) bioprinting
technologies have been applied enabling the generation of
hydrogel scaffolds with defined architecture and high spatial
control of cell allocation within the 3D construct during the
printing process. Due to the possibility of tuning the spatial

Correspondence to: G. Finkenzeller; e-mail: guenter.finkenzeller@uniklinik-freiburg.de

Contract grant sponsor: Deutsche Forschungsgemeinschaft; contract grant numbers: FI 790/10-1, KO 3910/1-1

C 2017 WILEY PERIODICALS, INC.

V 935
distribution of ECs during ex vivo printing, optimization of
vessel formation upon implantation should be achieved. Cur-
rent bioprinting techniques involve a rapid prototyping pro-
cess and can be divided in three main categories: inkjet,
microextrusion and laser assisted bioprinting.8 However, in
all 3D bioprinting processes hydrogels, often also referred to
as “bioinks”, are used for cell printing. These hydrogels have
to be compatible with the encapsulated cells, the printing
process and must also provide the desired mechanical stabil-
ity of the printed 3D construct. Hydrogels currently used in
the field of 3D bioprinting are predominantly based on either
naturally derived polymers (e.g., Matrigel, fibrin, collagen, gel-
atin, agarose, and alginate) or synthetic molecules (e.g., Plur-
onic F127).9 However, the optimal hydrogel has to be
determined for each cell type and each printing technology.
For the particular field of 3D bioprinting is at present unclear
which hydrogel is best suited to support cellular functions of
ECs.
In this in vitro study, we performed side by side compar-
isons of the above mentioned hydrogels in the context of
their physicochemical parameters (viscosity, surface tension,
storage, and loss modulus), their swelling or degradation
characteristics, their biological effects on vasculogenesis-
related EC parameters (cell viability, cell proliferation, and
endothelial cell sprouting) and their printability. The goal of
this study was to identify a hydrogel or a hydrogel combina-
tion with physicochemical characteristics suitable for inkjet
printing and concurrently with optimal long-term stability
and EC cytocompatibility.
MATERIALS AND METHODS
Preparation of hydrogels
Matrigel. Growth factor reduced Matrigel (354230; Corn-
ing) mainly contains Laminin-111, Collagen IV, Entactin, and
heparin sulfate from murine Engelbreth-Holm-Swarm sar-
coma.10 These components support structural and prolifera-
tive functions of various cell lines. However, the exact
curing mechanism of Matrigel still remains to be elucidated.
As irreversible thermal curing is induced at 108C, prepara-
tions were kept on ice before their final compositions were
achieved. Our experimental setups include samples with
100, 66 and 33% Matrigel. Dilutions were prepared with
phosphate buffered saline (PBS) (L182-10; Biochrom).
Fibrin. Fibrinogen (341576; Calbiochem) and thrombin
(605190; Calbiochem) are the well-known components of in
vivo coagulation processes. The serin protease thrombin
allows proteolytic processing of fibrinogen and the subse-
quent formation of an enzymatically cross-linked protein
complex named fibrin. Physiological plasma concentrations
of fibrinogen range from 2 to 4 mg/mL, whereas 0.1 m/mL
thrombin is sufficient for the initiation of the cross-linking
process. Due to the temperature-dependent kinetics of
thrombin, all preparations were kept on ice. We prepared
all samples with 10 mg/mL fibrinogen and evaluated the
effects of 10, 5 and 2 m/mL thrombin. Both components
were mixed 1:1. Dilutions from stock solution were pre-
pared with PBS.
936 BENNING ET AL.
Collagen. Collagen-I (345236; Corning) is collected from rat
tails and thoroughly purified by the manufacturer. The
highly conserved domain of Gly-X-Y-Gly is common to all
types of collagen and enables the protein to form stable tri-
ple helices in a pH neutral environment. During preparation,
all collagen samples were kept in an acidic environment
and could be handled in a highly viscous state. Samples
were prepared to final concentrations of 1, 2 and 3 mg/mL
according to the manufacturer’s instructions. Gelation was
induced by neutralization with equimolar 1 M NaOH.
Gelatin. Gelatin Type-A (G2500-100G; Sigma) consists of
collagen that has been processed using an acidic denatura-
tion method. The specific gelatin in use originates from por-
cine skin. All gelatin concentrations were dissolved in PBS.
Due to gelatin’s gel-sol transition at about 378C, we required
a covalent stabilization procedure to obtain a hydrogel that
remained in its semi-solid state during incubation. There-
fore, we kept each gelatin sample at room temperature and
induced a covalent cross-linking by adding 0.3% glutaralde-
hyde (G6257; Sigma). After 12 h, all samples were washed
thoroughly with PBS and possible aldehyde residues were
inactivated through the addition of 50 mM glycine (357002;
Calbiochem). After another washing step with PBS, samples
were ready to be seeded with cells with final concentrations
of 5, 10 and 15% gelatin. Dilutions were prepared with PBS.
Alginate. Alginic acid sodium salt (180947; Sigma) was
used for this study. Alginate is a polyanionic polysaccharide
sourced from brown algae. Due to its poor solubility, prepa-
rations were kept on a rocking table at 378C overnight.
Cross-linking of alginate requires bivalent cations (e.g., cal-
cium), which were added as 0.2 M CaCl2 (21114; Fluka).
After 5 min, remaining CaCl2 was taken out of each well
and each sample was rinsed thoroughly with PBS. Samples
were prepared to final concentrations of 1, 1.5 and 2% algi-
nate. Dilutions were prepared with PBS.
Alginate/gelatin. A common blend of hydrogels consists of
alginate and gelatin, which combines different suitable prop-
erties for tissue engineering while compensating some of
each other’s less well suited characteristics. We prepared
samples with 5% gelatin and 1% alginate, 10% gelatin and
1% alginate and 15% gelatin and 1% alginate. The gelatin
components in these samples were not covalently cross-
linked through the glutaraldehyde treatment, as outlined
above. Dilutions were prepared with PBS.
Pluronic F-127. Pluronic F-127 (P2443; Sigma) is the only
synthetic hydrogel used in this study and is an amphiphilic
blockcopolymer that consists of ethyleneoxide and propyle-
neoxide. Due to its poor solubility, all preparations were dis-
solved on a rocking table at 48C overnight since the
reversible thermal cross-linking of Pluronic F-127 is induced
at room temperature. Samples of 15, 20 and 25% were pre-
pared. Dilutions were prepared with PBS.
BIOPRINTING OF ENDOTHELIAL CELLS

Agarose. An especially low-temperature gelling agarose
(A2576; Sigma) was investigated since most general agarose
preparations require temperatures of far above 378C to
enter a homogeneous liquid state. Agarose was heated to
458C and plated. After all preparations were cooled down to
room temperature, the samples were ready for further proc-
essing. Samples were prepared with 2, 3 and 4% agarose.
Dilutions were prepared with PBS
Physicochemical characterization of hydrogels
Criteria for free-flying droplet generation using inkjet-based
dispensing systems are that delivered kinetic energy must
overcome surface energy and viscous losses of the liquid.
For this reason surface tension and viscosity are the domi-
nant physicochemical parameters in order to make a state-
ment about non-contact dispensing ability of hydrogel
components. Hydrogels are basically mixtures of liquids and
solid 3D matrices. Due to the elastic modulus dominating
the viscous modulus, hydrogels are usually printed prior to
gelation in their liquid state. Gelation of hydrogel compo-
nents used in this work is either induced chemically (fibrin,
collagen, and alginate) or by temperature (gelatin and aga-
rose <378C; Pluronic F127 and Matrigel >108C). Corre-
spondingly, thrombin, fibrinogen, collagen, and alginate (the
latter two in the non-cross-linked state) can be printed at
room temperature, while gelatin and agarose must be
printed above 378C and Pluronic and Matrigel below 108C.
Determination of surface tension
Surface tension of hydrogel components was measured accord-
ing to the pendant drop method using an OCA15 plus contact
angle measurement device from DataPhysics Instruments
GmbH, Germany. Five individual measurements of each hydro-
gel component were conducted, while the system was recali-
brated with water before measuring the following component.
As explained above, surface tension of the components was
measured at their respective printing temperatures.
Determination of viscosity
Like surface tension, viscosity of hydrogel components was
measured at dedicated printing temperatures using a MCR
52 rotational rheometer from Anton Paar GmbH, Austria.
Dynamic viscosity was measured from 1 to 2000 s21 since
most biopolymer solutions show a non-linear relation
between shear rate and shear stress as they are non-
Newtonian fluids. Hence, dependent on the specific printing
system different shear rates can act, resulting in different
viscosities of the same liquid component.
Determination of viscoelastic properties–storage and
loss modulus
Following in vitro cultivation procedure of bioprinted grafts,
implantation can be a challenging step due to handling
issues. Mechanical stability of hydrogels must be ensured
during implantation processes and can be quantitatively
described by the hydrogels dynamic shear modulus (G*),
composed of storage (G0 ) and loss modulus (G00 ). Characteris-
tic for most hydrogels is that the storage modulus exceeds
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | APR 2018 VOL 106A, ISSUE 4
ORIGINAL ARTICLE
the loss modulus and is constant over a broad frequency
range which allows the definition of a plateau modulus. To
compare results the plateau moduli were measured at 1 Hz
using a MCR 52 rotational rheometer from Anton Paar
GmbH, Austria (plate diameter: 25 mm; constant torque: 2.5
mN/m). Before each plateau measurement, frequency sweeps
from 1 to 10 Hz were performed to ensure that measure-
ments are indeed within the plateau regime. Measurements
were conducted at graft incubation temperature of 378C.
Swelling/degradation assays
Triplicates of all hydrogels named above were immersed in
Endothelial Cell Growth Medium (C22010; Promocell)
(ECGM) and incubated for 7 days to evaluate possible swel-
ling or degradation in an aqueous environment. After
obtaining blank measurements of each specific dish, 600 mL
of each hydrogel were plated in tissue culture dishes (35–
3001; Falcon) and the respective gelling procedures were
induced. 600 mL of ECGM was added on top of each hydro-
gel as soon as the ideal semi-solid state of each hydrogel
was reached. Weight measurements were conducted on day
0, 1, 4 and 7. Prior to each measurement, the entire super-
natant was removed from the dish to allow the weight mea-
surement of the remaining hydrogel. After weigh, fresh
media was pipetted onto each sample. Weight alterations
were monitored over the full course of the experiment.
HUVEC spheroid printing
HUVEC spheroids (each consisting of 250 HUVECs) were
printed between hydrogel layers using an inkjet-based dis-
penser (PipeJet P9; Biofluidix GmbH, Germany) to investi-
gate HUVEC viability. Spheroid concentration was adapted
to 3000 spheroids per mL and suspended in ECGM contain-
ing 1% penicilin/streptomycin and 10% FCS. Hydrogel com-
ponents were added to the spheroid suspensions, spheroid
suspension dispensed between fibrin layers contains a final
concentration of 5 mg/mL fibrinogen while 3 mg/mL colla-
gen was added to the spheroid suspension dispensed
between collagen layers. The solutions were loaded individ-
ually into the PipeJet reservoir connected to the protruded
polymeric nozzle (inner nozzle diameter: 500 mm). Droplets
of the HUVEC spheroid solutions (single droplet volume: 50
nL) were printed on the bottom of 24 well plates coated
with 150 mL fibrin (10 mg/mL fibrinogen crosslinked with
10 m/mL thrombin) or 150 mL collagen (3 mg/mL). Dis-
pensed spheroids were covered with 150 mL of the respec-
tive hydrogel and the wells were filled with 500 mL ECGM,
1% penicillin/streptomycin, 10% FCS. Spheroids were incu-
bated for 48 h and pictures were taken subsequent to print-
ing, after 24 and 48 h of incubation to observe the
sprouting process, indicating HUVEC viability.
Cell culture
Human umbilical vein endothelial cells (HUVECs) were
obtained from Promocell and cultured in ECGM. Media were
supplemented with 1% penicilin/streptomycin and 10%
fetal calf serum (FCS) (S0615; Biochrom). Media were
changed every 3 days and cells were trypsinized and split
937
when confluence reached >90%. Cells from passages three
to eight were used in all experiments.
Spheroid assay
To investigate the hydrogels’ ability of promoting cell
sprouting, a spheroid assay was conducted. First, a methocel
stock solution was prepared by dissolving 6 g methylcellu-
lose in 500 mL Endothelial Cell Basal Medium (C22210;
Promocell). The solution was supplemented with 50 mL
FCS. Second, cells were prepared as outlined above, counted
by a Casy1 cell counter (Sch€arfe Systems) and centrifuged
for 5 min at 1200 rpm. The resulting pellet was resolved in
an 80/20% preparation of ECGM and methocel stock solu-
tion and plated on standard cell culture dishes in 25 mL
drops. Culture dishes were carefully turned upside down to
allow spheroid formation in hanging drops. Cells were incu-
bated for 24 h at 378C before newly formed spheroids were
harvested. Spheroids were embedded into the various
hydrogels. For each hydrogel, 50 spheroids (500 HUVECs
per spheroid) were seeded into 500 mL hydrogel solution in
24 well plates. Freshly prepared gels were transferred into
a humidified incubator (378C, 5% CO2) and after polymer-
ization, 250 mL ECGM supplemented with 10% FCS was
added per well. After 24 h, gels were photographed using a
phase contrast microscope.
Viability assay
To determine the hydrogels’ biocompatibility, a Live/Dead
Viability/Cytotoxicity Kit (L3224; Life Technologies) was
used in accordance to the supplier’s instructions. 1 3 104
HUVECs each were seeded in all evaluated hydrogels in trip-
licates and incubated for a period of 24 h. Following the
extraction of ECGM and a washing step with PBS, the freshly
prepared Live/Dead solution was added to each sample.
Samples are kept at room temperature for 30 min before
pictures were obtained from an Axiovert 100 fluorescence
microscope (Carl Zeiss Jena GmbH, Jena). Fluorescent sig-
nals of the samples are based on the intracellular process-
ing of the cell-permeant calcein acetoxymethyl ester to
calcein, which emits fluorescence at 515 nm. Due to the
demand of an active processing, calcein AM is used to iden-
tify live cells. Ethidium homodimer 1 (EthD1) emits fluores-
cence at 635 nm when intercalating DNA. Since EthD1 can
only permeate damaged cell membranes, it visualizes dead
cells.
Cell proliferation assay
Cell activity on hydrogels was evaluated over the course of 7
days by using the cell activity assay CellTiter96 Aqueous One
Solution (G3580; Promega). The addition of the MTS derivate
allows a colorimetric determination of the cells’ activity
seeded onto the hydrogels. Samples of each hydrogel were
prepared in triplicates for measurements on day 1, 4, and 7.
Twelve well plates were preloaded with 215 mL hydrogel per
well. 1 3 104 HUVECs were plated onto each hydrogel.
Media were changed on day 3 and 6. Prior to the addition of
the reacting agent, all samples were washed thoroughly with
PBS. After 3 h of incubation at 378C, the evaluation was
938 BENNING ET AL.
carried out on a colorimetric Tecan SpectraFluor Plus reader
(Tecan Group AG, M€annedorf, CH). Increasing optical den-
sities reflect an increasing cell number.
RESULTS
The hydrogels investigated in our experiments have already
been used before by other groups in bioprinting approaches.
Therefore, we investigated the hydrogels in the concentra-
tion range already published. In our experiments, each
hydrogel was prepared and investigated in three different
concentrations.
Determination of surface tension
Respective surface tensions of the hydrogel components are
shown in Figure 1. Values for collagen and thrombin range
from 68.16 to 74.83 mN/m and do, with the exception of
alginate (1%), not significantly exceed the surface tension of
water. Therefore, they are in good compliance with water-
based liquids and can be employed for inkjet-based printing
platforms. The surface tensions of gelatin, alginate/gelatin
blends and fibrinogen are about 10 mN/m lower and cover
a range from 62.43 to 66.80 mN/m but still can be consid-
ered to be printable using inkjet-based dispensers as they
usually cover the higher surface tension range from 50 mN/
m on.11 Values for Pluronic and Matrigel are not stated
since gelation started within the measurement as a result of
the nonexistent cooling opportunity. Since cooling of those
hydrogels is important for inkjet-based printing, the respec-
tive dispensers must be cooled below 108C making the pro-
cess more difficult especially for Matrigel for which gelation
is irreversible.
Determination of viscosity
Viscosity of the hydrogel components can be seen in Figure
2. Pluronic F-127, gelatin and thrombin are the only hydro-
gel components showing a Newtonian behavior of shear
rate versus shear stress while all other hydrogels show a
shear thinning phenomenon in correspondence to their
pseudoplastic properties, which are typical for biopolymer
solutions. This means the viscosity decreases with increas-
ing shear rates. In contrast, the viscosity of water as a New-
tonian fluid stays constant within the measured shear rate
range. Highest viscosity was found for agarose whereas
thrombin solutions are most similar to water. The 4% aga-
rose solution could be most challenging for inkjet printers
requiring high kinetic energies for dispensing: even at the
highest shear rate applied, viscosity did not come below
257 mPa s. Moreover agarose and gelatin need to be heated
above 378C to keep them liquid, necessitating a dispenser
specific heating system. All other hydrogel compounds
exhibit viscosities ranging from 1 to 80 mPa s (for shear
rates between 1 and 2000 s21) which is within the range
covered by most inkjet printers.11
Determination of viscoelastic properties
Viscoelastic properties were measured to determine the
hydrogels ability for 3D printing. Hydrogels exhibiting high
storage moduli have a larger potential to resist
BIOPRINTING OF ENDOTHELIAL CELLS

FIGURE 1. Surface tension of hydrogels. Surface tension was measured at the respective printing temperatures of the hydrogels. Gelatin and
agarose were measured at 378C, all other hydrogels were measured at RT. Shown are mean values 6 SD. n 5 5.
deformations induced by handling steps which is especially
important for the stability of grafts. Storage and loss moduli
for the hydrogels are given in Figure 3. With the exception
of alginate/gelatin blends, all hydrogels show increasing
storage moduli with increasing polymer concentrations.
Pluronic F-127 (15%) is the only material not to be consid-
ered as a suitable hydrogel, since its loss modulus exceeds
the storage modulus. This finding supports the results of
the respective swelling/degradation assay which show a
rapid degradation of the Pluronic F-127 samples and sug-
gests an insufficient curing process in the Pluronic F-127
hydrogel with low polymer content. In contrast, the other
Pluronic F-127 solutions exhibit the highest storage moduli
of 7.8 (20%) and 13.8 kPa (25%) while lowest amongst all
hydrogels measured were found for collagen with values
between 2 and 33 Pa.
Validation of swelling/degradation properties
Since all hydrogels must be kept in an aqueous environment
due to the cells’ demand for a constant nutrient supply, all
hydrogels were immersed in ECGM and incubated for up to
7 days to evaluate possible swelling or degradation. In
order to measure swelling (weight increase) or degradation
(weight decrease), hydrogels were weighed upon gelation
(day 0) and after 1, 4 and 7 days (Fig. 4). Matrigel at 100
and 66% concentrations as well as agarose swelled to the
highest degree by an increase in weight of approximately 50
and 17%, respectively. In contrast, fibrin hydrogels prepared
with thrombin concentrations of 10 and 5 m/mL, collagen at
1 and 2 mg/mL and alginate at 1% showed a moderate
decrease in weight after 7 days, while being stable at all
other tested concentrations. Stability was also detected for
gelatin at all tested concentrations. Interestingly, although
pure gelatin and alginate were stable over the 7 days time
period, alginate/gelatin mixtures showed a relatively strong
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | APR 2018 VOL 106A, ISSUE 4
ORIGINAL ARTICLE
decrease in weight, indicative for hydrogel instability. More-
over, Matrigel at 33% also revealed instability as indicated
by a weight decrease of approximately 60%. Pluronic F-127
showed the highest instability and was almost completely
degraded after 1 day of incubation in growth medium.
Viability of HUVECs seeded on various hydrogels
Next, we investigated the biocompatibility of the hydrogels
by measuring viability, proliferation and sprout formation of
HUVECs. Viability was determined by employing a Live/
Dead assay in which viable cells are characterized by green
fluorescence and dead cells by red fluorescence (Fig. 5).
Matrigel, fibrin, collagen, and gelatin hydrogels showed
excellent effects on EC viability indicated by the fact that
virtually no dead cells could be detected. In contrast, dead
HUVECs indicated by read fluorescence, could be detected
in the alginate group. In this case, viability of HUVECs was
only about 71%. Although, no dead cells could be detected
in the agarose and Pluronic F-127 group, it became evident
that these hydrogels are not compatible with HUVECs,
because cells only weakly attach to these hydrogels and
most cells, presumably dead cells, were lost during the
washing step before performing the Live/Dead assay. Poor
cell attachment is also indicated by the rounded phenotype
of the ECs, which is also seen in the alginate group and the
alginate/gelatin (1/5%) group. Interestingly, EC attachment
can be improved by increasing the gelatin concentration in
the alginate/gelatin hydrogels. In contrast to all other tested
hydrogels, Matrigel is outstanding in its well-known ability
to support cord formation of ECs which is illustrated in Fig-
ure 5, where ECs form capillary-like networks on Matrigel
but not on the other tested hydrogels.
Determination of cell proliferation
Cell proliferation was measured by MTS assays performed
on day 1, 4, and 7 after seeding ECs onto the hydrogels
939
FIGURE 2. Viscosity measurements of biopolymer solutions at respective temperatures below or above gelation (Matrigel and Pluronic F-127
<48C, thrombin and fibrinogen, collagen, alginate, and alginate/gelatin 5 238C, gelatin and agarose >378C). Newtonian behavior was observed
for Pluronic, gelatin, and thrombin. Matrigel, collagen agarose, and alginate show pseudoplastic behavior.

940 BENNING ET AL. BIOPRINTING OF ENDOTHELIAL CELLS


FIGURE 3. Viscoelastic properties of hydrogels. The dynamic shear modulus (G*) of hydrogels, which is composed of storage (G0 ) and loss mod-
ulus (G00 ), was determined at 1 Hz. Measurements were conducted at 378C. Shown are means 6 SD. n 5 3.
(Fig. 6). No cell proliferation was detectable on agarose,
Pluronic F-127, alginate and alginate/gelatin hydrogels. In
contrast, ECs proliferated very well on Matrigel, fibrin, colla-
gen, and gelatin, where cell numbers increased 7- to 8-fold
between day 1 and 7. In case of Matrigel, collagen, and gela-
tin hydrogels, we made the interesting observation that EC
proliferation correlated inversely with the respective hydro-
gel concentrations investigated in this study. This means,
the lower the concentration of the hydrogels, the greater
the cell proliferation.
In vitro sprouting of HUVEC spheroids embedded in
various hydrogels
To investigate the ability of hydrogels to support EC sprout-
ing, HUVEC spheroids were fabricated, seeded into the
hydrogels and analyzed after 24 h of incubation, as
described before.12 HUVEC spheroid sprouts have previously
been shown to form lumenized capillary-like structures.13
As shown in Figure 7, spheroid sprouting, that is, radial out-
growth of cells from the spheroid, can be detected in Matri-
gel at concentrations of 100 and 66%, in collagen at
concentrations of 2 and 3 mg/mL and in fibrin at all tested
thrombin concentrations. In contrast, gelatin and alginate/
gelatin hydrogels were unable to support HUVEC sprouting.
Agarose, Pluronic F-127 and alginate hydrogels were not
tested in this assay because of their ineffectiveness to sup-
port HUVEC attachment and proliferation, as already shown
in Figures 5 and 6.
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | APR 2018 VOL 106A, ISSUE 4
ORIGINAL ARTICLE
Bioprinting of HUVEC spheroids in fibrin and collagen
We further focused our interest on fibrin and collagen,
because Matrigel is not clinically approved and is more diffi-
cult to print since it has to be cooled during the printing
process. In order to investigate whether fibrin and collagen
can be used as bioinks in combination with our inkjet-based
bioprinter, HUVEC spheroids were printed in a grid-like pat-
tern in both hydrogels. Moreover, we also conducted this
experiment to evaluate whether HUVEC spheroids remain
intact and viable during the printing process. As can be
seen in Figure 8, HUVEC spheroids remained intact after
printing, as evidenced by the rounded spheroidal morphol-
ogy, which was maintained subsequent to printing. More-
over, viability and functionality of HUVECs was proven by
the fact that HUVEC spheroids form capillary precursors
after 24 and 48 h of incubation. Spheroid sprouting was
detectable in both hydrogels to a similar extend.
DISCUSSION
In this study, we evaluated eight hydrogels in various final
concentrations in terms of their physicochemical character-
istics, their swelling/degradation properties and their bio-
compatibility toward human endothelial cells for
prospective bioprinting applications. These hydrogels are
commercially available and have already been used for bio-
printing of ECs14–16 or other cell types.17–21 In prospective
studies, we intend to print ECs in conjunction with other
tissue specific cell types to produce prevascularized artificial
941
FIGURE 4. Swelling/degradation properties of hydrogels. Hydrogels at different concentrations were plated in tissue culture dishes and after
gelation overlaid with endothelial cell growth medium. The weight of the gels after removal of the growth medium was determined at day 0, 1,
4, and 7. Swelling is indicated by an increase in weight, whereas degradation of the hydrogels is indicated by a decrease in weight. Shown are
means 6 SD. n 5 3.

tissue grafts for implantation studies. Since ECs are very
demanding concerning their cultivation and their environ-
mental requirements, especially in the context of their
potential capacity to generate endothelial sprouts in vitro
and stable blood vessels in vivo, we decided to perform a
direct side by side comparison of the hydrogels.
Surface tension and viscosity have been measured
regarding inkjet-based printing ability. It was found that
Matrigel and Pluronic F-127 are difficult to handle due to
cooling issues during the printing process. High concen-
trated agarose is most challenging exhibiting high viscos-
ities. Agarose and gelatin need to be heated during the

942 BENNING ET AL. BIOPRINTING OF ENDOTHELIAL CELLS


FIGURE 5. Cytotoxicity testing of hydrogels. HUVECs were plated on
hydrogels and viability of cells was evaluated after 24 h by using a
Live/Dead assay. Representative images were taken using a fluores-
cence microscope. Viable cells show green fluorescence, dead cells
show red fluorescence.
printing process, since respective gelling temperature is
below 378C. Those hydrogels become liquid when heated
above 378C which is not suitable for grafts as they may lose
form stability when incubated or implanted. All other
hydrogels exhibit physicochemical properties ranging within
the spectrum typically covered by most inkjet-based dis-
pensers. To give precise information about the printing abil-
ity of hydrogel components one has to consider the specific
printing system. Since most inkjet suppliers indicate suita-
ble viscosity and surface tension ranges, printing ability can
be determined theoretically if orifice diameter (respectively
droplet diameter) and droplet velocity is known for the spe-
cific inkjet dispenser.22 However, the droplet velocity itself
is dependent on viscosity and surface tension. Furthermore,
droplet generation is a highly dynamic process which occurs
on a short time scale. Therefore, it is advisable to determine
the suitability of a specific hydrogel for non-contact dispens-
ing by experimental means.
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | APR 2018 VOL 106A, ISSUE 4
ORIGINAL ARTICLE
In contrast to previous reports, where common surface
tension measurement methods for biopolymer solutions
appeared not well suited,23 here, the pendant drop method
has been applied as it exactly mimics the process when a
droplet is released out of an inkjet nozzle. When using the
pendant drop method, liquid is released out of a syringe
nozzle until gravity exceeds surface tension and the droplet
breaks-off. As inkjet-based systems show similar kinetics,
except for the induced kinetic energy of the printer is
responsible for droplet generation, the method can be
applied in this study. Nevertheless, it may also have led to
surface tension values for water-based alginate solutions
which are in conflict with the value of water (72.8 mN/m):
the increased values could have been caused by biopolymer
film formation at the surface acting as a skin-like structure.
The swelling/degradation properties of hydrogels repre-
sent an important point during the in vitro cultivation of the
bioprinted graft and especially in the context of the long-
term stability of the graft after implantation. The ideal
hydrogel for our prospective application should show only
minimal swelling or degradation. In a 1 week swelling/deg-
radation in vitro study, where the hydrogels have been sub-
merged in ECGM, we have seen that Pluronic F-127 is not
suitable for our bioprinting approach, because the Pluronic-
based hydrogel is completely degraded after already 24 h of
cultivation in ECGM at 378C. The same is true for the algi-
nate/gelatin hydrogel, which was also unstable over the
observation period, although the components by themselves
showed acceptable swelling/degradation properties. So far,
the reason for this interesting result is not clear, but a simi-
lar phenomenon for alginate/gelatin has been reported by
another group, where alginate alone showed moderate swel-
ling, whereas the alginate/gelatin hydrogel showed moder-
ate degradation.24 Collagen, gelatin, agarose, and alginate
hydrogels showed acceptable swelling/degradation proper-
ties at all tested concentrations. Degradation of fibrin was
observed for thrombin concentrations of 10 m/mL while it
was stable for fibrin gels prepared with 2 and 5 m/mL
thrombin.
This can be explained by the fact that the thrombin con-
centration has a strong impact on the gelation time of fibri-
nogen but also on the stiffness of the resulting fibrin gel.25
Matrigel, showed an acceptable swelling/degradation
property at a final concentration of 66%. At lower concen-
trations (33%) Matrigel degraded, whereas swelling
occurred at higher concentrations (100%).
Another very important aspect is the biocompatibility of
hydrogels. In our study, we analyzed the biocompatibility of
the hydrogels toward ECs in terms of cell viability, prolifera-
tion and in vitro sprout formation.
As expected, Matrigel, fibrin, collagen, and gelatin appeared
as excellent hydrogels concerning EC viability: at all tested
concentrations, we have been unable to detect dead cells by
using a classical Live/Dead assay. In contrast, in the alginate
group, EC viability was only about 71%. Moreover, we
observed only very poor attachment of ECs to alginate hydro-
gels, illustrated by low cell numbers and a rounded phenotype
of the cells. However, incorporation of gelatin into the alginate
943
FIGURE 6. Effect of hydrogels on HUVEC proliferation. HUVECs were seeded onto the hydrogels and proliferation was determined by using a
MTS assay at day 1, 4, and 7. Shown are means 6 SD. n 5 3.
hydrogels improved cell attachment and viability. Poor cell
attachment was also detected on agarose and Pluronic F-127
hydrogels at all tested concentrations. The poor attachment
of cells to agarose, Pluronic F-127 and alginate can be
explained by the fact that these hydrogels do not contain
arginine/glycin/aspartate motifs and therefore development
of focal adhesion contacts between the hydrogels and ECs is
strongly impaired.26–28 In case of Matrigel, we observed
944 BENNING ET AL.
aligning of HUVECs in capillary-like endothelial networks
during the 24 h incubation period, a process that could not
be seen in the other tested hydrogels. Cord formation of
endothelial cells on Matrigel is a well-known phenomenon
that is used in Matrigel assays to evaluate the angiogenic
properties of endothelial cells and/or to evaluate the pro-
or anti-angiogenic activity of growth factors or other
compounds.29–31
BIOPRINTING OF ENDOTHELIAL CELLS
 ORIGINAL ARTICLE

FIGURE 7. Effects of hydrogels on sprout formation of HUVEC spheroids. HUVEC spheroids were embedded manually into hydrogels and ana-
lyzed for sprouting by phase contrast microscopy after 24 h. Scale bars: 100 mm.

Proliferation of HUVECs on hydrogels was monitored by a
MTS assay over a time period of up to 7 days. Matrigel, fibrin,
collagen, and gelatin hydrogels were able to support prolifera-
tion of HUVECs. Interestingly, we observed an inverse correla-
tion between gel concentration and proliferation rate on
Matrigel, collagen, and gelatin hydrogels within the tested con-
centration range, demonstrating that lower hydrogel stiffness
is associated with increased proliferation. On agarose, Pluronic
F-127, alginate and alginate/gelatin gels, little or no prolifera-
tion was observed, which may be explained by the observa-
tion, that these hydrogels do not support cell attachment.
Finally, as another very important parameter in our pro-
spective effort to create bioprinted prevascularized grafts,
we used the 3D HUVEC spheroid sprouting assay in order
to investigate which hydrogels support 3D formation of
capillary-like cords emerging from HUVEC spheroids in
vitro. The original spheroid sprouting assay is based on pre-
formed EC spheroids that are encapsulated into collagen
type-I gels. This assay was developed to study endothelial
cell differentiation, maturation and capillary formation32,33
and also proved to be suitable for the identification of pro-
or anti-angiogenic molecules.34 In the present study, HUVEC
spheroids were encapsulated in the different hydrogels. As
expected, collagen at concentrations of 2 and 3 mg/mL was
able to support spheroid sprouting. The same was true for
Matrigel (at concentrations of 100% and 66%) and fibrin

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | APR 2018 VOL 106A, ISSUE 4 945
FIGURE 8. Printing of HUVEC spheroids. Spheroids (each consisting of 250 HUVECs) were suspended in fibrinogen solution and printed onto a
fibrin layer or suspended in collagen solution and printed onto a collagen layer at a pitch of 500 mm. After printing and gelation of the spheroid
containing hydrogels, the arrays were covered by an additional layer of the respective hydrogels and incubated in ECGM, 10% FCS. An inverted
microscope was used for taking single images at day 0 to day 2 after incubation.
(at all thrombin concentrations tested). In contrast, gelatin
and alginate/gelatin hydrogels were completely unable to
support spheroid sprouting. Concerning gelatin, the lacking
ability to promote sprout formation is most likely due to
the glutaraldehyde treatment as outlined in the materials
and methods section. Whereas the gelatin hydrogels in all
experiments prior to the spheroid embedment were treated
with glutaraldehyde in a cell-free environment, the sphe-
roids embedded in the cross-linked gelatine are likely to
become severely damaged due to the cytotoxicity of
glutaraldehyde.
Taking into account all of the results generated so far,
Matrigel, fibrin and collagen turned out to be most suitable
for bioprinting of ECs to produce prevascularized tissue
equivalents. However, in terms of a putative prospective
clinical application, one has to consider that Matrigel is not
clinically approved, because it is produced from murine
Engelbreth-Holm-Swarm sarcoma. Moreover, cooling of
Matrigel is necessary for inkjet-based printing. Therefore,
the respective dispensers must be cooled below 108C mak-
ing the process more difficult. Therefore, we have concen-
trated our further interest on fibrin and collagen and
investigated the printability of these hydrogels as well as
viability and functionality of printed HUVEC spheroids.
As expected, we were able to demonstrate printability of
both hydrogels with our printing system, as previously
reported by other groups working with inkjet print-
ers.18,35,36 Moreover, we were also able to show that HUVEC
spheroids were not damaged during the printing process
and were able to form capillary-like structures after print-
ing, further demonstrating viability and functionality of
HUVEC spheroids, bioprinted with both hydrogels.
In summary, among the hydrogels tested, our selection
process finally identified fibrin and collagen as ideal candi-
dates for bioprinting of HUVEC spheroids. This study paves
the way for further exploration of 3D printing technology
for bioprinting of ECs to produce prevascularized tissue
equivalents for various medical applications.
946 BENNING ET AL.
ACKNOWLEDGMENT
We thank Brunhilde Baumer for excellent technical assistance.
This work was supported by funding through the Deutsche
Forschungsgemeinschaft (FI 790/10-1 and KO 3910/1-1).
CONFLICTS OF INTERESTS
There are no financial conflicts of interests for the authors
and the research.
REFERENCES
1. Shieh SJ, Vacanti JP. State-of-the-art tissue engineering: From tis-
sue engineering to organ building. Surgery 2005;137:1–7.
2. Henry TD, Rocha-Singh K, Isner JM, Kereiakes DJ, Giordano FJ,
Simons M, Losordo DW, Hendel RC, Bonow RO, Eppler SM,
Zioncheck TF, Holmgren EB, McCluskey ER. Intracoronary admin-
istration of recombinant human vascular endothelial growth fac-
tor to patients with coronary artery disease. Am Heart J 2001;142:
872–880.
3. Baumgartner I, Pieczek A, Manor O, Blair R, Kearney M, Walsh K,
Isner JM. Constitutive expression of phVEGF165 after intramuscu-
lar gene transfer promotes collateral vessel development in
patients with critical limb ischemia. Circulation 1998;97:1114–
1123.
4. Schechner JS, Nath AK, Zheng L, Kluger MS, Hughes CCW,
Sierra-Honigmann MR, Lorber MI, Tellides G, Kashgarian M,
Bothwell ALM, Pober JS. In vivo formation of complex microves-
sels lined by human endothelial cells in an immunodeficient
mouse. Proc Natl Acad Sci USA 2000;97:9191–9196.
5. Alajati A, Laib AM, Weber H, Boos AM, Bartol A, Ikenberg K, Korff
T, Zentgraf H, Obodozie C, Graeser R, Christian S, Finkenzeller G,
Stark GB, He roult M, Augustin HG. Spheroid-based engineering
of a human vasculature in mice. Nat Methods 2008;5:439–445.
6. Enis DR, Shepherd BR, Wang Y, Qasim A, Shanahan CM,
Weissberg PL, Kashgarian M, Pober JS, Schechner JS. Induction,
differentiation, and remodeling of blood vessels after transplanta-
tion of Bcl-2-transduced endothelial cells. Proc Natl Acad Sci USA
2005;102:425–430.
7. Steffens L, Wenger A, Stark GB, Finkenzeller G. In vivo engineer-
ing of a human vasculature for bone tissue engineering applica-
tions. J Cell Mol Med 2009;13:3380–3386.
8. Murphy SV, Atala A. 3D bioprinting of tissues and organs. Nat
Biotechnol 2014;32:773–785.
9. Malda J, Visser J, Melchels FP, Jungst T, Hennink WE, Dhert WJ,
Groll J, Hutmacher DW. 25th anniversary article: Engineering
hydrogels for biofabrication. Adv Mater 2013;25:5011–5028.
BIOPRINTING OF ENDOTHELIAL CELLS

10. Benton G, Arnaoutova I, George J, Kleinman HK, Koblinski J.
Matrigel: From discovery and ECM mimicry to assays and models
for cancer research. Adv Drug Deliv Rev 2014;79–80: 3–18.
11. Calvert P. Inkjet Printing for materials and devices. Chem Mater
2001;13:3299–3305.
12. Wenger A, Stahl A, Weber H, Finkenzeller G, Augustin HG, Stark
GB, Kneser U. Modulation of in vitro angiogenesis in a three-
dimensional spheroidal coculture model for bone tissue engineer-
ing. Tissue Eng 2004;10:1536–1547.
13. Korff T, Augustin HG. Tensional forces in fibrillar extracellular
matrices control directional capillary sprouting. J Cell Sci 1999;
112: 3249–3258.
14. Norotte C, Marga FS, Niklason LE, Forgacs G. Scaffold-free vascu-
lar tissue engineering using bioprinting. Biomaterials 2009;30:
5910–5917.
15. Boland T, Mironov V, Gutowska A, Roth EA, Markwald RR. Cell
and organ printing 2: Fusion of cell aggregates in three-
dimensional gels. Anat Rec A Discov Mol Cell Evol Biol 2003;272:
497–502.
16. Cui X, Boland T. Human microvasculature fabrication using ther-
mal inkjet printing technology. Biomaterials 2009;30:6221–6227.
17. Sakai S, Ito S, Inagaki H, Hirose K, Matsuyama T, Taya M,
Kawakami K. Cell-enclosing gelatin-based microcapsule produc-
tion for tissue engineering using a microfluidic flow-focusing sys-
tem. Biomicrofluidics 2011;5:13402.
18. Skardal A, Mack D, Kapetanovic E, Atala A, Jackson JD, Yoo J,
Soker S. Bioprinted amniotic fluid-derived stem cells accelerate
healing of large skin wounds. Stem Cells Transl Med 2012;1:792–
802.
19. Snyder JE, Hamid Q, Wang C, Chang R, Emami K, Wu H, Sun W.
Bioprinting cell-laden Matrigel for radioprotection study of liver
by pro-drug conversion in a dual-tissue microfluidic chip. Biofab-
rication 2011;3:034112.
20. Duarte Campos DF, Blaeser A, Weber M, Jakel J, Neuss S,
Jahnen-Dechent W, Fischer H. Three-dimensional printing of stem
cell-laden hydrogels submerged in a hydrophobic high-density
fluid. Biofabrication 2012;5:015003.
21. Fedorovich NE, Schuurman W, Wijnberg HM, Prins HJ, van
Weeren PR, Malda J, Alblas J, Dhert WJ. Biofabrication of osteo-
chondral tissue equivalents by printing topologically defined, cell-
laden hydrogel scaffolds. Tissue Eng Part C Methods 2012;18:33–
44.
22. Koltay P, Zengerle R. Non-contact nanoliter and picoliter liquid
dispensing. 14th International Conference on Solid State Sensors,
Actuators and Microsystems. Lyon, France; 2007. p 165–170.
23. Lee B-B, Chan E-S, Ravindra P, Khan TA. Surface tension of vis-
cous biopolymer solutions measured using the Nouy ring method
and the drop weight methods. Polym Bull 2012; 69:471–489.
24. Murphy SV, Skardal A, Atala A. Evaluation of hydrogels for bio-
printing applications. J Biomed Mater Res A 2013;101:272–284.
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | APR 2018 VOL 106A, ISSUE 4
ORIGINAL ARTICLE
25. Duong H, Wu B, Tawil B. Modulation of 3D fibrin matrix stiffness
by intrinsic fibrinogen-thrombin compositions and by extrinsic
cellular activity. Tissue Eng Part A 2009;15:1865–1876.
26. Bidarra SJ, Barrias CC, Fonseca KB, Barbosa MA, Soares RA,
Granja PL. Injectable in situ crosslinkable RGD-modified alginate
matrix for endothelial cells delivery. Biomaterials 2011;32:7897–
7904.
27. Wang W, Liu Z, Sun P, Fang C, Fang H, Wang Y, Ji J, Chen J.
RGD peptides-conjugated Pluronic triblock copolymers encapsu-
lated with AP-2alpha expression plasmid for targeting gastric can-
cer therapy in vitro and in vivo. Int J Mol Sci 2015;16:16263–
16274.
28. Kopf M, Campos DF, Blaeser A, Sen KS, Fischer H. A tailored
three-dimensionally printable agarose-collagen blend allows
encapsulation, spreading, and attachment of human umbilical
artery smooth muscle cells. Biofabrication 2016;8:025011.
29. Kuzuya M, Kinsella JL. Reorganization of endothelial cord-like
structures on basement membrane complex (Matrigel): Involve-
ment of transforming growth factor beta 1. J Cell Physiol 1994;
161:267–276.
30. Tsopanoglou NE, Haralabopoulos GC, Maragoudakis ME. Oppos-
ing effects on modulation of angiogenesis by protein kinase C
and cAMP-mediated pathways. J Vasc Res 1994;31:195–204.
31. Muhlhauser J, Merrill MJ, Pili R, Maeda H, Bacic M, Bewig B,
Passaniti A, Edwards NA, Crystal RG, Capogrossi MC. VEGF165
expressed by a replication-deficient recombinant adenovirus vec-
tor induces angiogenesis in vivo. Circ Res 1995;77:1077–1086.
32. Korff T, Augustin HG. Integration of endothelial cells in multicellu-
lar spheroids prevents apoptosis and induces differentiation.
J Cell Biol 1998;143:1341–1352.
33. Korff T, Kimmina S, Martiny-Baron G, Augustin HG. Blood vessel
maturation in a 3-dimensional spheroidal coculture model: Direct
contact with smooth muscle cells regulates endothelial cell quies-
cence and abrogates VEGF responsiveness. Faseb J 2001;15:447–
457.
34. Haspel HC, Scicli GM, McMahon G, Scicli AG. Inhibition of vascu-
lar endothelial growth factor-associated tyrosine kinase activity
with SU5416 blocks sprouting in the microvascular endothelial
cell spheroid model of angiogenesis. Microvasc Res 2002;63:304–
315.
35. Moon SJun, Hasan SK, Song YS, Xu F, Keles HO, Manzur F,
Mikkilineni S, Hong JW, Nagatomi J, Haeggstrom E,
Khademhosseini A, Demirci U. Layer by layer three-dimensional
tissue epitaxy by cell-laden hydrogel droplets. Tissue Eng Part C
Methods 2010;16:157–166.
36. Xu T, Binder KW, Albanna MZ, Dice D, Zhao W, Yoo JJ, Atala A.
Hybrid printing of mechanically and biologically improved con-
structs for cartilage tissue engineering applications. Biofabrication
2012;5:015001.
947