Vascularized Lung Tissue Engineering

Full text

Vascularized lung tissue engineering
Yifan Yuan1
1Department of Anesthesiology, Yale University, New Haven, CT 06519;
Email: yifan.yuan@yale.edu

Abstract
Lung transplantation is an effective treatment for severe and end stage lung diseases. A
readily available functional bioengineered lung from an autologous source is critical for
lung transplantation. Previous studies have decellularized lungs and repopulated them
with functional cell population, however, it failed due in large part to impaired formation
of vascular compartment. Pulmonary vasculature not only exerts a physical barrier to
separate blood and tissue, but is a metabolically active barrier that controls exchange of
gas, macromolecules, and immune cells and as such is very critical for lung tissue
engineering. Many strategies including proper endothelial cell populations, seeding
techniques and culture conditions have been optimized to achieve endothelial coverage of
entire lung scaffold. In this article, strategies and challenges related to vascularized lung
tissue engineering will be reviewed.
Key words: Lung tissue engineering, pulmonary vasculature, endothelium, seeding,
angiogenesis, endothelial maturation, proteases, shear stress, substrate stiffness, pericytes,
bioreactors

1. Introduction
Chronic lung disease is the third leading killer for both men and women in the United
States. Lung transplantation is an effective treatment for severe and end stage lung
diseases such as chronic obstructive pulmonary disease, idiopathic pulmonary fibrosis,
cystic fibrosis, or idiopathic pulmonary arterial hypertension. Although current strategy
maintains 80% for one-year survival and below 50% for 5-year survival, many
complications such as malignancy are initiated due to lifelong immunosuppression.
Furthermore, lungs are susceptible to injury after brain death than other organs, resulting
in major mortality on the waiting list due to donor shortage. Thus, a readily available
functional bioengineered lung from an autologous source is critical for lung
transplantation.
The focus of lung tissue engineering has been the development of synthetic or natural
scaffold with stem or progenitor cells to regenerate lung tissue with functional pulmonary
vasculature. However, current technologies failed to recapitulate the complex 3-
dimensional architecture of the lung and further, clinical translation of this technique
remains unknown. In 2010, two separate groups made modest progress by decellularizing
rodent lungs and demonstrated the decellularized organ not only preserved the 3D
architecture of alveoli but also removed cellular components that minimized
immunogenicity, making ‘off the shelf’ bioengineered lung possible. They have further
repopulated the decellularized lung with endothelium and epithelium and managed to
obtain cell growth in vitro. However, after several hours of transplantation, bioengineered
lung failed due in large part to impaired formation of the vascular compartment, leading

to thrombosis and pulmonary edema. Therefore regeneration of functional pulmonary
vasculature is critical for lung tissue engineering.
2. Pulmonary Vasculature
Pulmonary circulation consists of three major parts: arteries, capillaries, and veins.
Morphometric approaches indicate that capillary walls dominate ~2% of lung anatomic
volume and are responsible for gas exchange and nutrient transport to the whole
parenchyma, which comprises 70 - 80% of lung anatomic volume. Extra-alveolar
structures, on the other hand, including extra-alveolar blood vessels such as pulmonary
arteries and veins take a relatively small fraction of lung volume, and supply nutrients to
airways that are ~7% of lung volume. Pulmonary capillaries are small vessels from 5 to 8
μm in diameter, having a wall of one endothelial cell thickness. The endothelial cells,
which form a monolayer lining the inner surface of blood vessels, synthesize many
factors such as nitric oxide, prostacyclin, tissue plasminogen activator, thrombomodulin,
heparan-sulfate, and the endothelial protein C receptor (a receptor associated with the
activation of protein C) to control thromboresistance, vascular barrier, and vessel wall
inflammation. Dysfunction of pulmonary microvascular endothelial cells increases
permeability, leading to extravascular leak of protein-rich edema, polymorphonuclear
leukocyte influx, microvascular thrombosis, and further lung dysfunction (e.g. acute lung
injury, sepsis). For example, endothelial activation by pro-inflammatory cytokines such
as tumor necrosis factor α (TNF-α), interferon γ (IFN-γ), or components of the bacterial
wall such as lipopolysaccharide can induce production of pro-inflammatory molecules

including intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion protein
1 (VCAM-1), tissue factor, and disruption of cell-cell contacts, leading to increased
leukocyte adhesion, vascular leakage, thrombosis and pulmonary edema. Abnormal
hemodynamics such as disturbed flow increase expression level of tissue factor in ECs
and add to progression of thrombosis. Pulmonary vascular endothelial cells not only
provide a barrier for exchange of macromolecules and immune cells, but also act as
supporting cells for epithelial function. Co-culture of human pulmonary microvascular
ECs with A549 cells (an epithelial cell line) improved their transelectrical resistance,
tight junction ZO-1 expression, and adherens junction E-cadherin expression. Thus,
pulmonary vascular endothelial cells, especially microvascular endothelial cells, play a
regulatory role in pulmonary vascular homeostasis, alveolar function, and further lung
regeneration.
3. Endothelial cell sources for lung tissue engineering
3.1. Pulmonary endothelial cells
It is well known that endothelium from different organs is phenotypically distinct. For
example, microvascular ECs from lung secrete 1.5 – 15 times more urokinase-type
plasminogen activator antigen, an enzyme involved in alveoli repair during injury and
surfactant protection, than human hepatic microvascular ECs, human umbilical vein ECs,
angioma ECs, and lung fibroblasts. Lung microvascular ECs uniquely expressed higher
levels of factor VIII, vWF and UEA binding, and have lower Dil-Ac-LDL uptake
compared to ECs from brain and liver. Under treatment of endothelial barrier disruptor

thrombin, lung microvascular ECs display the highest affinity and response, followed by
brain ECs and liver ECs. Therefore, lung ECs display a significantly different phenotype
from ECs of other organs, and as such the ideal way to repopulate lung vasculature would
be to use cells directly from lung vasculature.
The pulmonary vasculature composes of three anatomic compartments connected in
series: the arterial tree, an extensive capillary bed, and venular tree. Capillary endothelial
cells, the most abundant EC, make up 30% of the cells in human lungs. These cells are
adjacent to alveoli and directly responsible to form a metabolically active surface for
exchange of gas, macromolecules, and immune cells, and display a significantly different
phenotype from ECs in large vessels. Endothelium in vessels with internal diameters less
than 38 μm uniformly binds Griffonia simplicifolia, a lectin that specifically interacts
with α-galactose, and have no Weibel-Palade bodies. Emerging evidence indicates
pulmonary microvascular endothelial cells form significantly tighter barriers than do
pulmonary artery endothelial cells. Trans-electrical resistance and hydraulic conductivity
data revealed that microvascular ECs in basal conditions form tighter cell-cell barriers
compared to ECs from vein and artery. This may be due to the differential expression of
activated leukocyte cell adhesion molecule (ALCAM), a cell adhesion molecule of the
immunoglobulin super-family, which is significantly more abundant in pulmonary
microvascular ECs than in pulmonary artery ECs. This glycoprotein has been shown to
co-localize with adherens junctions such as vascular endothelial cadherin (VE-Cadherin).
Knockout of this gene results in reduced expression of barrier junction proteins such as
occludin, α-catenin, claudin-5, and ZO-1 and increased blood permeability.

Microvascular ECs also proliferate faster than do ECs from large vessels. In basal culture
conditions and under the same medium, the population doubling times of microvascular
ECs and artery ECs are 39 h and 58 h, respectively. This finding was correlated with the
higher proportion of endothelial colony forming cells in microvascular ECs compared to
those in macrovascular ECs. We have recently repopulated rat lung microvascular ECs
into a decellularized lung scaffold, and demonstrated ECs not only maintained their
expression of CD31 but also displayed tight junctions observed by TEM after 8 days
culture. Recent work has co-cultured pulmonary microvascular ECs with arterial and
venous ECs in decellularized lung scaffolds, and demonstrated that positive VE-cadherin
staining was only observed once three cell types are present, suggesting a necessity of
cell-cell cross-talk for barrier formation. However, it is well documented that lung
allograft endothelium can initiate immune rejection through presentation of alloantigens
to circulating T cells, natural killer cells and macrophages, which limits the use of
primary pulmonary ECs for bioengineered lung applications.
3.2. Blood outgrowth endothelial cells
A subpopulation of circulating, bone-marrow-derived cells, termed endothelial progenitor
cells, has been shown to contribute to neovascularization, and home to sites of vascular
trauma potentially participating in re-endothelialization of damaged or denuded surfaces.
These cells can be easily derived from a standard apheresis procedure from autologous
sources, and if maintained in culture they undergo a transition to cells with a phenotype
similar to mature endothelial cells. These blood outgrowth endothelial cells (BOECs) can
achieve at least 30 population doublings in serial passaging, re-plate into at least

secondary and tertiary colonies, and retain high levels of telomerase activity. They also
exhibit surface marker profiles (i.e. CD31, 133, 144, 15, 146, and KDR) and cytoplasmic
vWF production similar to mature ECs. Various groups have demonstrated that protein
and gene expression between BOECs and mature endothelial cells are highly similar. For
example, Medina and colleagues used 2D-PAGE to investigate the proteomes of BOECs
and human dermal microvascular endothelial cells (DMECs) and observed BOECs
demonstrated over 90% of the tested protein profile in common with DMECs. When
combined with synthetic or natural materials, BOECs can form de novo functional blood
vessel in vivo. Yoder et al have isolated and cultured BOECs in collagen/fibronectin gels,
and transplanted them into immunodeficient mice. At 14 to 30 days, BOECs formed
chimeric vessels, which were perfused with mouse red blood cells, suggesting that
BOECs resemble mature ECs in phenotype and can be directly involved in vessel
formation.
Multiple in vitro studies have seeded BOECs on surfaces of synthetic vascular grafts or
vascular scaffolds and demonstrated that human BOEC adherence, growth, and
phenotype are supported on various synthetic coatings (e.g. poly(1,8-(octanediol citrate),
polyurethane, polyglycolic-acid) under both static and flow culture conditions. When
seeded and cultured on three-dimensional biodegradable vascular scaffolds (polyurethane
foam or polyglycolic acid/poly-4-hydroxybutyrate mesh) under dynamic culture in vitro,
human umbilical cord blood OECs display suitable cell-to-polymer attachment and
growth on both polymers while maintaining endothelial phenotype (CD31, vWF, and
eNOS) after 12 days of culture. Further preclinical studies have demonstrated autologous

BOECs-seeding improved thromboprotective properties of vascular implants. Jantzen and
co-workers seeded autologous porcine BOECs at the point-of-care in the operating room
onto Titanium (Ti) tubes for 30 minutes followed by implanting into the inferior vena
cava of swine (n=8). After 3 days, a blinded analysis revealed that all 4 cell-seeded
implants are free of clot, whereas 4 controls without BOECs were either entirely
occluded or partially thrombosed. Quint and coworkers have cultured aortic smooth
muscle cells on a degradable polyglycolic acid mesh scaffold in bioreactors for 10 weeks
followed by decellularization. Porcine BOECs were seeded on the lumen of vessels for 3
days, and these tissue-engineered blood vessels were implanted as porcine carotid
interposition grafts. After 30 days of implantation, they found that all BOEC-seeded
TEBV remained patent, whereas only 3/8 control vein grafts were patent. They have also
determined that less neointimal hyperplasia were observed compared to control vein
grafts. Thus, BOECs potentially preserve endothelial properties on biomaterial scaffolds
and may be a proper candidate for vascularization of lung scaffolds. However, the
phenotype and proliferative potential of BOECs can be impaired in cardiopulmonary
disease patients who are the major cohort requiring lung transplantation. Furthermore,
reseeding with BOECs necessarily requires large volumes of apheresis product, and can
lead to risk of several complications related to leukapheresis procedures such as
hypocalcemia due to citrate anticoagulation.
3.3. iPSC-ECs
As a source for regenerative therapies, human induced pluripotent stem cells (iPSCs)
offer immense potential and was recognized by the 2012 Nobel Committee in Medicine.

These cells, derived from differentiated adult cells through genetic reprogramming, have
provided an exciting alternative for bypassing ethical concerns related to embryonic stem
cells derivation and potential issues of allogeneic immune rejection. Differentiation
toward endothelial cells from iPSCs has been extensively studied and reviewed
elsewhere. The iPSCs are usually cultured into three-dimensional embryoid bodies in
suspension culture followed by differentiation into 3 germ layers: endoderm, ectoderm,
and mesoderm, where endothelial cells belong. Culture strategies including using growth
factors, co-culture with parenchymal cells, and 2 D culture on coatings enriched with
ECM proteins have been used to differentiate ECs from mesodermal cells. For example,
inhibition of TGF-β in pluripotent stem cell differentiation increased VE-Cadherin+ cells
by 10 fold. Subsequent inhibition of TGF-β in culture improved the proliferation of
purified endothelial cells up to 36 fold. Co-culture of mouse bone marrow stromal cell
line OP9 with iPSCs improved the emergence of CD31+ endothelial cells to as early as 8
days. Transduction of Oct4 and Klf4 into human neonatal fibroblasts in the presence of
soluble factors promotes the induction of endothelial reprogramming. Clusters of induced
endothelial cells resemble primary human ECs phenotype and protein profiles such as
CD31, VE-Cadherin, and vWF were obtained at 28 days.
Many studies have tried to improve the yield of iPSC-differentiated ECs for scale-up
applications. Sahara et al screened > 60 small molecules that would promote endothelial
differentiation and found that administration of BMP4 and a GSK-3β inhibitor in an early
phase and treatment with VEGF-A and inhibition of the Notch signaling pathway in a
later phase led to the emergence of endothelial cells within six days. These ECs were able

to form functional capillaries in vivo with anastomosis to the host vessels when
transplanted into immunocompromised mice. The iPSC:EC ratio of 1:20 was reported to
be a significant improvement over prior publications. A recent work by Prasain et al have
significantly improved the yield of iPSC:EC ratio to 1:1 × 108. They have identified a
Neuropilin-1 (NRP-1)+/CD31+/CD144+ subpopulation from iPSC-differentiated
endothelial cells displays functional properties similar to umbilical cord blood endothelial
colony forming cells, with high clonal proliferative potential and robust in vivo vessel-
forming ability. The ESC-NRP-1+CD31+ECFC and iPSC-NRP-1+CD31+ECFC
maintained a stable endothelial phenotype and function and did not undergo replicative
senescence for 18 passages in vitro. These iPSC-NRP-1+CD31+ECFC have shown
capacity to form human vessels in mice and to repair the ischemic mouse retina and limb
for > 6 months with lack of teratoma formation potential. Thus, these iPSC-NRP-
1+CD31+ECFC may provide clinically relevant number of functional cells and may be a
proper EC candidate for vascularization of lung tissue engineering.
However, many challenges associated with iPSC-differentiated ECs remain to be
overcome to achieve this goal. Firstly, iPSC-derived ECs may possess a different
phenotype compared to mature microvascular pulmonary ECs. Reed et al have directly
compared the phenotype among ESC-differentiated ECs with HUVEC, human lung
microvascular endothelial cells, BOECs, and human aortic endothelial cells and
demonstrated that only ESC-differentiated ECs express extremely low endothelin-1, the
cardioprotective hormone prostacyclin, and have completely different morphological
response to shear stress. Secondly, iPSC-derived cells have been shown to present poor

levels of maturation with high inter-line variability, and such immaturity may lead to
undesirable side effects, including cancer. Thirdly, the intrinsic qualities of self-renewal
and pluripotency of iPS cells are responsible for tumorigenic potential, which is a huge
safety issue for clinical applications. Finally, although it has been assumed that these
autologous cells should prevent immunorejection by an autologous host, their
immunogenicity has not been extensively investigated. Thus, iPS cell-based therapies are
still in their infancy, and many hurdles remain to be overcome before their clinical
applications become a reality.
4. EC seeding in lung scaffold
Early studies demonstrated that endothelial cell seeding into decellularized organs is
possible. Peterson et al have repopulated decellularized lung scaffolds through injection
into the pulmonary artery and found that the phenotype and survival of rat lung
microvascular ECs were maintained in the scaffold. However, after implantation, H/E
staining revealed that red blood cells invaded the alveolar septa, and some red blood cells
were present in airspaces due to the severely damaged alveoli-vascular barrier. Similarly,
Ott et al have managed to seed HUVECs into decellularized lung scaffolds and
maintained their survival. However, they observed severe interstitial edema after 6 h of
orthotopic transplantation of regenerated lungs, which maybe due to leakage from
immature vessels. Thus, there is a need for an optimized seeding and culture strategy to
increase EC coverage throughout the entire lung scaffold.

A recent study by Ren et al has moved the field forward by significantly improving the
EC seeding efficiency in lung scaffold. They have first evaluated the seeding method
through perfusing 0.2 μm microspheres into an accelular scaffold. Under physiological
pressure, the arterial perfusion of microspheres results in very limited seeding into the
capillary bed, which is likely due to the rapid loss of hydrostatic pressure as the aqueous
phase diffuses through the permeable basement membrane. Simultaneous gravity-driven
perfusion of red and green microspheres sequentially into pulmonary artery (PA) and
pulmonary vein (PV) led to significantly less leak of microspheres into the alveolar space
or airways. The seeding of fluorescently labeled HUVECs through both PA and PV
displayed similar distribution with that of the microspheres. Additionally, co-seeding the
lung with HUVECs and supporting human mesenchymal stem cells resulted in 75%
endothelial coverage of the entire lung. Vascular barrier function assayed by FITC-
conjugated 500-kDa dextran perfusion revealed that HUVEC-hMSC regenerated lungs
gradually increased intravascular dextran retention and reached 80.2% ± 5.4% on day 8
relative to freshly isolated cadaveric rat lungs. After 3 days of transplantation, fluorescent
microangiography revealed that pulmonary vessels remained perfusable.
Stabler et al have studied the posture of the lung during seeding and demonstrated that a
supine position rather than an upright position improved re-endothelization. Under
arterial perfusion at 1 mL/min, they seeded the lung with pooled rat lung microvascular
ECs in either the supine or upright position. The average cell number in right upper and
left lobes in the supine position is significantly higher than that in the upright position.
Additionally, supine position allows for greater cell retention within large diameter

vessels than the upright position, with little difference in the small diameter distal vessels.
After 1 week of culture, FITC-dextran (150 kDa) assay revealed a markedly increased
vascular barrier in vascularized lung compared to decellularized lung.
Scarritt et al have compared the gravity-driven and pump-driven perfusion during
seeding. Gravity-driven seeding was accomplished by passively delivering cell
suspension from 20 cm above the lung to the vasculature. As for pump-driven perfusion,
cells were delivered from the seeding reservoir to the lung vasculature via pump-driven
flow at rates of either 1 ml/min or 3 ml/min. Gravity-driven seeding evenly distributed
cells and supported cell survival as well as re-lining of the vascular walls while perfusion
pump-driven seeding led to increased cell fragmentation and death. Similarly, our recent
published report has used diluted rat lung microvascular ECs to sequentially seed PV and
PA with gravity-driven perfusion at ~60 cm H2O above the lung. We noticed that gravity-
driven pressure results in even distribution of ECs throughout the lung after 4 days of
culture. The EC coverage reached almost 80% of the entire lung relative to native control.
Interestingly, in the endothelialized lung explant both large and small vessels were
perfused with red blood cells, suggesting thromboresistant function of cultured
endothelial cells in lung scaffold. However, the airspaces still contained an abundance of
blood indicating the existence of vascular barrier leakage.
For advances to be made in this area, considerations for re-endothelializing an acellular
scaffold include, but are not limited to, the cell size in comparison to the capillary
diameter, cell density, perfusion methods, and pressure level. Firstly, based on direct

measurements accumulated from casts of the pulmonary vasculature or from transmission
electron microscopy, average capillary diameter ranges from ~ 5 – 8 μm, which is close
to or slightly smaller than the size of most suspended mammalian cells. For example, the
diameter of bovine aortic endothelial cells is around 10 – 18 μm. Thus, the seeding
strategy may vary with the difference in cell size, which is also dependent on species, cell
cycle, and cell types. Secondly, cell-seeding density should be maintained relatively low.
High cell density increases the chances of cell clumping in large vessels and subsequent
reduction of cell delivery into microvasculature, and eventually results in patchy seeding.
However, low cell density may increase the duration of cell delivery and decrease cell
survival due to anoikis. Thirdly, although previous studies provided valuable initial
comparisons of pump and gravity -driven perfusion, the flow rate and seeding pressure
have not maintained consistent between two comparison groups, making it hard to
determine the cause of the different cell distribution. Further study is needed to determine
the effect of pressure and flow rate driven by either gravity or pump on EC seeding.
Lastly, pressure level is critical for EC seeding, as hydraulic pressure decreases
substantially during delivery into microvasculature due to leaky barrier and high
resistance. However, high seeding pressure increases matrix stretching and damages
ECM structures, leading to barrier breakdown. Low seeding pressure may not sufficiently
perfuse cells into the capillary bed. Recent literatures are reaching a consensus that dual-
side seeding from both PA and PV may increase seeding efficiency while limiting high
pressure. However, dual-side seeding fails to recruit the entire lung vasculature. When
seeding with 0.2 μm red and green microspheres through PA and PV respectively,
microsphere-opacified vascular channels of the two colors were generally mutually

exclusive, suggesting existence of non-seeded capillaries. Thus, full EC coverage in lung
scaffold may require further cell migration and proliferation.
5. Angiogenesis in lung scaffold
Angiogenesis depicts the formation of new vessel from existing blood vessels. This
process usually involves proliferation, migration, and differentiation of endothelial cells.
Angiogenic growth factors initially activate receptors on endothelial cells present in pre-
existing blood vessels and induce their release of proteases to degrade basement
membrane and further their migration from original vessel walls. The endothelial cells
then proliferate into the surrounding matrix and form solid sprouts connecting vessels
while recruiting supporting cells. Many molecules can regulate pulmonary angiogenesis
including vascular endothelial growth factor (VEGF), endothelial nitric oxide synthase
(eNOS), fibroblast growth factor (FGF), transforming growth factor (TGF)-β, and
substance P.
VEGF binds to 2 tyrosine-kinase receptors present on endothelial cells: Flt or VEGF
receptor 1 (VEGFR1); and KDR or VEGFR2. VEGFR1 binds VEGF with approximately
10-fold higher affinity than KDR, undergoes receptor autophophorylation, and stimulates
calcium influx. VEGFR1-deficient mice die in utero due to an early defect in the
development of haematopoietic and endothelial cells. Organized blood vessels were not
observed in the embryo or yolk sac at any stage, suggesting the role of VEGFR1 in
endothelial development. VEGF binding to VEGFR2 results in cell ruffling, mitosis,

chemotaxis, and actin rearrangement. Pharmacological blockage of VEGFR2 by SU5416
inhibited VEGF-driven mitogenesis in HUVECs, in vivo angiogenesis, and vascular
permeability. VEGFR2 is also phosphorylated by autocrine VEGF produced by
endothelium, which is essential for endothelial survival. Lung contains the highest level
of VEGF mRNA compared to brain, kidney, liver, and spleen of the healthy adult rat.
Decreased expression of VEGF, Flt-1, and Tie-2 due to Bronchopulmonary Dysplasia
(BPD) induces pulmonary vascular dysfunction. Postnatal, intratracheal adenovirus-
mediated VEGF gene therapy improves survival, promotes lung capillary formation, and
preserves alveolar development of animals with BPD, suggesting its role in pulmonary
vessel formation and alveolar maintenance.
Endothelial NOS was extensively investigated as a means to improve postnatal
angiogenesis. Namba and colleagues injected eNOS cDNA into rats with ischemic
hindlimbs, and demonstrated that both peripheral blood flow and capillary number were
significantly increased 4 weeks after transfection. This effect was completely abolished
by the treatment with NOS antagonist NW-nitro- -arginine methyl ester (L-NAME). In a˪
3D fibrin model, stimulated NO production in HUVECs by VEGF or FGF increased
capillary-like structure formation. This effect can be completely blocked by L-NAME.
Similar effects have been observed in vitro using substance P or transforming growth
factor β. Although these studies indicate the important role of eNOS in angiogenesis, the
precise mechanism has not been elucidated. Previous studies has demonstrated that eNOS
may promote angiogenesis through enhancing cell migration. Murohara et al., have found
that L-NAME, but not its inactive enantiomer D-NAME, inhibited endothelial cell

sprouting from the scratched edge of the cultured bovine aortic endothelial cell
monolayer. Boyden chamber assays have revealed that L-NAME inhibited endothelial
cell migration but not proliferation. Ward and colleagues have transfected human eNOS
into myeloid angiogenic cells, and demonstrated that eNOS transduction significantly
increased cell migration and network formation but not proliferation. Thus, enhanced
eNOS expression may improve in vitro angiogenesis through cell migration but not
proliferation.
FGF is the first pro-angiogeneic factor that was identified. In chicken chorioallantoic
membrane (CAM), FGF was shown to induce angiogenesis. This effect was attenuated by
either inhibition of mitogen-activated protein (MAP) kinase (ERK) or blockage of MEK,
while the pre-existing blood vessel was not affected on the CAM, indicating that FGF
induces new blood vessel formation through a MEK-MAPK dependent mechanism. FGF
can also induce angiogenesis through activating other pro-angiogenic factors such as
VEGF. Addition of recombinant FGF-2 to cultured endothelial cells or upregulation of
endogenous FGF-2 results in increased VEGF expression. Blockade of VEGF inhibits
FGF-2-induced endothelial proliferation. In mouse cornea, systemic administration of
neutralizing VEGF antibody dramatically reduces FGF-2-induced angiogenesis,
demonstrating that VEGF is an important autocrine mediator of FGF-2-induced
angiogenesis.
During angiogenesis in lung scaffolds, culture condition with enriched pro-angiogenic
factors are typically preferred. Ren et al have cultured HUVECs-seeded lung scaffolds in

“Angiogenic medium” that had high levels of serum and angiogenic growth factor for 14
days. However, they have not obtained increased endothelial coverage nor barrier
formation compared to 1 day after seeding, which may be due to matrix degradation
during angiogenesis and impaired cell stability in the scaffold.
6. EC maturation in lung scaffolds
During EC angiogenesis, cells are activated upon exposure to pro-angiogenic factors.
They are prone to matrix degradation, they secrete pro-inflammatory molecules, and this
leads to increased permeability, vascular barrier breakdown and ECM remodeling. Thus,
maintenance of EC quiescence after proliferation is critical for stabilizing cell integrity,
with enhanced paracellular barrier and protection of the ECM structure. Many
mechanical and chemical cues including shear stress, substrate stiffness, protease
inhibitors, and mural cells such as pericytes have been discussed as having the potential
to regulate the quiescent state of endothelial cells, and can be used for vascularization in
lung scaffolds.
6.1. Mechanical regulation
6.1.1. Shear stress
Shear stress is the force per unit area that is created when a tangential force (blood flow)
acts on a surface (endothelium). Endothelium lining the cardiovascular system is highly
sensitive to hemodynamic shear stresses that act at the vessel luminal surface in the

direction of blood flow. Physiological shear stress has been demonstrated to maintain
endothelial quiescence and integrity. Passerini and colleagues identified approximately
2,000 differentially expressed genes between ECs isolated from a region that is
susceptible to atherogenesis, and undisturbed laminar flow in adult porcine aorta.
Compared to cells under disturbed flow, there was a downregulation of several
inflammatory cytokines, pro-angiogeneic factors, and pro-atherogenic factors in ECs
under undisturbed laminar flow. Additionally, physiological shear stress was shown to
reduce endothelial metabolism by decreasing glucose uptake and mitochondrial content
due in part to KLF4/eNOS signaling. Endothelial NOS is a mechanosensing protein that
has been shown to maintain vascular quiescence in response to shear stress. Using a
microfluidic tissue analog of angiogenic sprouting, Song et al have shown that
physiological shear stress (3 dynes/cm2) attenuated VEGF-driven morphogenesis and
proliferation. This effect was blocked when cells were treated with NOS inhibitor NG-
monomethyl-L-arginine monoacetate (L-NMMA). This effect has been further supported
by the in vivo observations that the basal capillary-to-muscle fiber ratios and capillary
density were around 20% higher in the muscle of eNOS-/- mice compared to wild type.
Interestingly, with the increase of shear stress, increased capillary-to-muscle fiber ratio
and capillary density were observed in the muscles of rats and mice, suggesting that
mechanotransduction is dependent on the rate of shear stress. Baeyens et al have
proposed that endothelial cells have their preferred fluid shear stress, or ‘set point’. They
have shown that upon exposure to 16 hours of laminar shear stress ranging from 2 to 60
dynes/cm2, HUVECs aligned in the direction of the flow, between approximately 10 and
20 dynes/cm2, but were misaligned or oriented perpendicularly (against the flow

direction) outside this range. NF-κB activation was decreased while Smad1 was
maximum between 10 to 25 dynes/cm2, indicating that HUVECs have a biphasic
response to shear stress such that anti-inflammatory, stabilization pathways lie in the
range of 10 to 20 dynes/cm2. Shear stress has also been discussed as a way to regulate
vascular barrier function. For example, Colgan and colleagues exposed bovine brain
microvascular endothelial cells to either steady or pulsatile shear stress (10 and 14
dynes/cm2, respectively) for 24 hours. Substantial increases of occludin and ZO-1
expression were observed in both steady and pulsatile flow, compared to static control.
Pulsatile but not steady shear stress markedly reduced transendothelial permeability to
14C sucrose. Thus, both physiological rate and pattern of shear stress are critical to
maintaining an endothelial functional and quiescent state.
6.1.2. Substrate stiffness
Matrix stiffness, a mechanical property inherits to the ECM, has an influential role in
numerous endothelial cells functions such as inflammatory response, cell proliferation,
angiogenesis, and cell integrity. The blood vessel wall has a stiffness of 1.2 – 2 kPa
without pressurization, and stiffening of blood vessel ECM during aging or
atherosclerosis progression induced upregulation of endothelial permeability and
leukocyte extravasation. This phenomenon has been supported by the in vitro
observations that increasing substrate stiffness from 5 kPa to 280 kPa dramatically
enhanced neutrophil transmigration on HUVEC-formed monolayers. The effect of
substrate stiffness was attenuated when cells were treated with ML-7, a myosin light
chain kinase inhibitor, or blebbistatin, an inhibitor of myosin II. Further study has

indicated that septin 9, a negative upstream effector of Rho-asscoiated kinase, attenuated
the up-regulation of DNA synthesis and cell growth by stiff substrates in ECs. These
results indicate a critical role of myosin in mechanotransduction by substrate stiffness in
angiogenesis, leukocyte adhesion, vascular permeability, and cell proliferation. Similarly,
Bordeleau and colleagues demonstrated that the angiogenic outgrowth, invasion, and
neovessel branching increased, but barrier function was impaired and vascular
endothelial cadherin was altered, with increased matrix stiffness. Interestingly, this effect
was not only related to myosin inhibition, but to the regulation of matrix
metalloproteinase (MMP) expression such as MT1-MMP, indicating the role of substrate
stiffness on protease production.
6.2. Chemical Regulation
6.2.1. Protease inhibitors
During angiogenesis, ECs are activated in response to various environmental cues and
produce proteases to improve cell proliferation, migration, and invasion. Numerous
proteases including matrix metalloproteinases (MMPs), a disintegrin and metalloprotease
domain(s) (ADAM), a disintegrin and metalloproteinase with thrombospondin motifs
(ADAMTS), and cysteine and serine proteases are involved in angiogenesis. For
example, adenoviral-mediated MMP9 downregulation inhibited endothelial cell
migration and tube formation. When co-cultured with gliobastoma spheroids, MMP9
inhibition resulted in reduced invasion. MT1-MMP, a membrane-type MMP, is involved
in EC lumen formation and generation of vascular guidance tunnels in 3D collagen
matrices. ADAMTS-13 can also enhance HUVEC migration, proliferation, and in vitro

tube formation, possibly through a VEGF/Akt dependent mechanism. Various molecules
such as tissue inhibitors of metalloproteinases (TIMP), and synthetic MMP inhibitors
have been described to inhibit matrix proteases, angiogenesis and improve ECs
maturation.
There are four members in the TIMP family including TIMP1, TIMP2, TIMP3, and
TIMP4, which exert different impacts on MMPs, ADAMs, and ADAMTSs. TIMPs have
various biological activities including inhibition of cell proliferation, migration, invasion,
and angiogenesis. Co-culture of pericytes with endothelial cells induced capillary tube
stabilization by blocking EC tube morphogenesis and regression in a 3-D collagen
matrix. This effect was regulated by production of TIMP-3 in pericytes and TIMP-2 in
ECs. Suppression of TIMP-2 in ECs and TIMP-3 in pericytes results in capillary tube
regression in a MMP-1, MMP-10, MT1-MMP, and ADAM-15 dependent manner.
Treatment with recombinant human TIMP-1 and synthetic MMP inhibitors, GM6001 and
MMP-2-MMP-9 inhibitor III, suppressed MMP-2 activity and migration of human
dermal microvascular endothelial cells in a dose-dependent manner. Interestingly, the
MMP-dependent inhibition of migration was associated with increased expression of the
junctional adhesion proteins VE-Cadherin, PECAM-1, and VE-cadherin at cell-cell
junctions. Thus, TIMPs play an important role not only on inhibition of matrix
degradation, but also on anti-angiogenesis and improving endothelial integrity. However,
TIMPs non-selectively target a broad range of proteases, making it hard to study the
specific regulation on a single protease. Indeed, small interfering RNA specifically
knocks down the level of a particular protease, however, low transfection efficiency and

detrimental effect associated electroporation and cationic transfection may limit their use
for scale-up studies. Many small molecules have been designed, especially in the cancer
field, to specifically inhibit proteases’ function to block tumor angiogenesis. For example,
succinyl hydroxamic acid 4a-o is a potent and selective inhibitor of MMP3
(IC50=5.9nM), and was >140-fold less potent against MMP-1 (IC50=51,000nM), MMP-
2 (IC50=1790nM), MMP-9 (IC50=840nM), and MMP-14 (IC50=1900nM). The
sulfonamide hydroxamate exhibits >10-fold selectivity for MMP-2 over MMP-8, 9, 14,
and >1000-fold selectivity over MMP-1, and 3. However, many of these small molecules
were designed for cancer research, and may not only target MMP signaling but also
induce endothelial apoptosis. Therefore, investigation of these inhibitors on ECs
behaviors may help our understanding of specific protease regulation on ECsstabilization
and maturation.
6.2.2. Sphingosine 1-phosphate
Sphingosine 1-phosphate (S1P), a lipid mediator produced by sphingolipid metabolism,
promotes endothelial cell spreading, vascular maturation/stabilization, barrier function,
and angiogenesis. In a confluent layer of HUVECs, S1P significantly increases the
abundance of VE-cadherin and β-catenin at the cell-cell contact regions and enhances
adherens junction assembly. This effect was attenuated by microinjection of S1P1 and
S1P3 inhibitors. Overexpression of S1P1 in HEK293 cells markedly increases the
expression level of P-cadherin and E-cadherin, but not α-catenin and β-catenin, and
induces formation of well-developed adherens junctions in a manner dependent on S1P
and the small guanine nucleotide binding protein Rho. Furthermore, S1P1 silencing leads

to a reduction in expression of both VE-cadherin and PECAM-1, and the degree of S1P1
knockdown was correlated with the extent of suppression of VE-cadherin and PECAM-1.
Although S1P improves vascular maturation and regeneration, it has very short half-life
(~30 min) which necessarily limits its use in lung tissue engineering. Furthermore, as
oppose to endothelial maturation, S1P is a potent angiogeneic factor that improves
endothelial migration and proliferation and has been shown to be involved in matrix
degradation by activating MT1-MMP.
6.2.3. Cyclic AMP
Cyclic AMP (cAMP), a second messenger downstream of G-coupled receptor, has been
extensively studied to improve endothelial integrity. Elevation of cAMP promotes EC
barrier and further protects the lungs from edema development. Forskolin, a cAMP
activator, and isobutyl-1-methylxanthine treatment inhibited thrombin-induced ECs
barrier breakdown. Transfection with protein kinase A (PKA) inhibitor attenuated, while
mutant PKA inhibitor had no effect on, cAMP-induced barrier formation, suggesting a
key role of PKA in cAMP-regulated barrier protection. Pharmacological inhibitors of
PKA do not consistently reverse cAMP-enhanced endothelial cell function, suggesting
the existence of PKA-independent pathways. Treatment with 8-(4-chlorophenylthio)-2’-
O’methyladenosine-3’, 5’-cyclic monophosphate, a specific activator for Epac/Rap1
decreased permeability and increased VE-cadherin-mediated adhesion similar to
prostacyclin and forskolin. Similarly, activation of Rap1 resulted in a decrease in
permeability, enhancement of VE-cadherin-dependent cell adhesion and cortical actin,
whereas inactivation of Rap1 had the counter effect, suggesting a role for cAMP-Epac-

Rap1 signaling in barrier protection. Cyclic AMP has not only been shown to regulate
endothelial barrier, but also have impacts on angiogenesis and apoptosis. Inhibition of
endothelial PKA in mice led to perturbed vascular development, hemorrhage and
embryonic lethality. During perinatal retinal angiogenesis, inhibition of PKA resulted in
hypersprouting as a result of increased numbers of tip cells. Direct activation of PKA by
cAMP or by expression of the PKA catalytic subunit also induces endothelial cell
apoptosis, resulting in angiogenesis inhibition in vivo.
6.3. Pericytes
The stabilization and maturation of endothelial cells have been extensively studied in
pericytes. Although pericytes are in direct contact with endothelium, their regulation on
vascular maturation is primarily through paracrine signals, including platelet-derived
growth factor (PDGF) B receptor (PDGFR)-β; S1P1-endothelial differentiation
sphingolipid G-protein-coupled receptor-1 (EDG1); and Angiopoietin 1-Tie 2. Firstly,
ECs express PDGF-B to recruit pericytes by binding to PDGFRβ. The pericytes in turn
secrete vitronectin that initiates activation of integrin αvβ3 on ECs, leading to expression
of VEGF-A. Autocrine VEGF-A signaling in the endothelial cells then causes expression
of anti-apoptotic proteins to maintain endothelial cell survival. Secondly, the bioactive
lipid S1P is indispensable for vascular maturation. S1P1 mutants develop severed edema
and hemorrhages with no embryos surviving, due to lack of mural cell recruitment and
maintenance of vascular barrier. In addition, pericyte-derived S1P aids in maintenance of
microvascular stability by up-regulating the expression of N-cadherin and VE-cadherin,
and by down-regulating the expression of angiopoietin 2. Thirdly, Ang-1 is known to

stabilize nascent vessels and make them leak-resistant by facilitating communication
between ECs and mural cells. The blockage of pericyte resulted in poorly remodeled and
leaky vessel development and this can be restored by addition of recombinant ang-1.
Pericytes have been used to stabilize newly formed vasculature in decellularized lung
scaffolds. Ren et al have co-cultured iPSC-differentiated ECs with perivascular cells in
decellularized lung scaffolds for 6 days and obtained better endothelial integrity and
stability compared to ECs alone. Co-culture of adipose-derived stromal cells, which
contain a pericyte sub-population, with rat lung microvascular endothelial cells in a lung
scaffold improved EC long-term survival and vascular barrier function. Thus, pericytes
may represent a proper cell population to promote vascular maturation for long-term
culture in a decellularized lung scaffold. However, the plastic nature of pericytes and
their transition to myofibroblast phenotypes may damage vascular function and induce
lung fibrosis. Additionally, current co-culture studies have used similar seeding technique
for both pericytes and ECs. In native lung, pericytes are located on the opposite side of
endothelium. Further research on pericyte phenotype in decellularized lung scaffolds will
help to optimize culture protocol for maintaining their function.
7. Conclusion and Perspectives
Pulmonary vasculature not only exerts a physical barrier to separate blood and tissue, but
is a metabolically active barrier that controls exchange of gas, macromolecules, and
immune cells and as such is very critical for lung tissue engineering. However, many

challenges such as proper EC population, seeding strategy, and culture strategy, remain to
be overcome to achieve a vascularized lung scaffold. Firstly, endothelial cells display
significant phenotypic heterogeneity and as such the ideal cell population for lung tissue
engineering should be from lung itself. However, the allogeneity of endothelium
necessarily initiates immune rejection through presentation of alloantigens to circulating
T cells, natural killer cells, and macrophages. Endothelial cells isolated from autologous
leukapheresis products are readily available, functional, and proliferative, however their
functions are largely impaired if harvested from cardiovascular patients, which are the
major cohort of lung transplantation recipients. iPSC-derived ECs are highly
proliferative, purified and functional and as such may be a proper cell candidate for
vascularization of lung tissue engineering. However, iPSC differentiation toward lung-
specific endothelium has been sparsely investigated, and therefore phenotypic
comparison between iPSC-ECs and pulmonary ECs may help to understand their
differences. Secondly, current seeding strategies have allowed evenly distributed EC
coverage of entire lung scaffolds, and 75-80% coverage has been reached after several
days of culture. However, when seeding with 0.2 μm red and green microspheres through
PA and PV respectively, microsphere-opacified vascular channels of the two colors were
generally mutually exclusive, suggesting the existence of non-seeded capillaries. Thirdly,
endothelial cell culture in lung scaffold requires two phases, an angiogenesis phase and
maturation phase. Many molecules such as VEGF and FGF have been shown to improve
angiogenesis however, their effects necessarily damage ECM due to proteases
production. Protease inhibitors can be used to reduce the basement membrane
breakdown, however, their target specificity and short half-life may require combination

of multiple drugs and repeated dosing. Pericytes produce many molecules including Ang-
1, S1P, and protease inhibitors to stabilize endothelium and as such may represent a better
way for vascular maturation. However, their plastic nature may allow myofibroblast
transition during culture.

8. Further Reading References
1. Petersen TH, Calle EA, Zhao L, Lee EJ, Gui L, Raredon MB, et al. Tissue-
engineered lungs for in vivo implantation. Science. 2010;329(5991):538-41.
2. Ott HC, Clippinger B, Conrad C, Schuetz C, Pomerantseva I, Ikonomou L, et al.
Regeneration and orthotopic transplantation of a bioartificial lung. Nat Med.
2010;16(8):927-33.
3. Gebb S, Stevens T. On lung endothelial cell heterogeneity. Microvasc Res.
2004;68(1):1-12.
4. Yuan Y, Altalhi WA, Ng JJ, Courtman DW. Derivation of human peripheral blood
derived endothelial progenitor cells and the role of osteopontin surface modification and
eNOS transfection. Biomaterials. 2013;34(30):7292-301.
5. Prasain N, Lee MR, Vemula S, Meador JL, Yoshimoto M, Ferkowicz MJ, et al.
Differentiation of human pluripotent stem cells to cells similar to cord-blood endothelial
colony-forming cells. Nat Biotechnol. 2014;32(11):1151-7.
6. Ren X, Moser PT, Gilpin SE, Okamoto T, Wu T, Tapias LF, et al. Engineering
pulmonary vasculature in decellularized rat and human lungs. Nat Biotechnol.
2015;33(10):1097-102.
7. Le AV, Hatachi G, Beloiartsev A, Ghaedi M, Engler AJ, Baevova P, et al. Efficient
and Functional Endothelial Repopulation of Whole Lung Organ Scaffolds. ACS
Biomaterials Science & Engineering. 2017;3(9):2000-10.
8. Baeyens N, Nicoli S, Coon BG, Ross TD, Van den Dries K, Han J, et al. Vascular
remodeling is governed by a VEGFR3-dependent fluid shear stress set point. Elife.
2015;4.
9. Fukuhara S, Sakurai A, Sano H, Yamagishi A, Somekawa S, Takakura N, et al.
Cyclic AMP potentiates vascular endothelial cadherin-mediated cell-cell contact to
enhance endothelial barrier function through an Epac-Rap1 signaling pathway. Mol Cell
Biol. 2005;25(1):136-46.
9. Glossary of Critical Terms
Lung tissue engineering
Development of synthetic or natural scaffolds with stem or progenitor cells to regenerate
lung tissue with functional pulmonary vasculature
Pulmonary vasculature
pulmonary arteries, pulmonary capillaries, and pulmonary veins
Endothelium
The cell monolayer lining the inner surface of blood vessels, synthesizes many factors
such as nitric oxide, prostacyclin, tissue plasminogen activator, thrombomodulin,
heparan-sulfate, and the endothelial protein C receptor (a receptor associated with the
activation of protein C) to control thromboresistance, vascular barrier, and vessel wall
inflammation

Angiogenesis
The formation of new vessel from existing blood vessels
Proteases
An enzyme that helps proteolysis
iPSC
These cells, derived from differentiated adult cells through genetic reprogramming, have
provided an exciting alternative for bypassing ethical concerns related to embryonic stem
cells derivation and potential issues of allogeneic immune rejection
Shear stress
Force per unit area that is created when a tangential force acts on a surface
Substrate stiffness
A mechanical property inherits to the ECM, has an influential role in numerous
endothelial cells functions such as inflammatory response, cell proliferation,
angiogenesis, and cell integrity
Pericytes
Contractile cells adjacent to the endothelial cells that line the capillaries
Bioreactors
An engineered system that support biological active environment
10. Abbreviations
TNF-α tumor necrosis factor α
IFN-γ interferon γ
ICAM-1 intercellular adhesion molecule 1
VCAM-1 vascular cell adhesion protein 1
ALCAM activated leukocyte cell adhesion molecule
VE-Cadherin vascular endothelial cadherin
BOECs blood outgrowth endothelial cells
DMECs dermal microvascular endothelial cells
Ti Titanium
iPSCs induced pluripotent stem cells
NRP-1 Neuropilin-1
PA pulmonary artery
PV pulmonary vein
VEGF vascular endothelial growth factor
eNOS endothelial nitric oxide synthase
FGF fibroblast growth factor

TGF-β transforming growth factor-β
BPD Bronchopulmonary Dysplasia
L-NAME NW-nitro- -arginine methyl ester˪
CAM chorioallantoic membrane
MAP mitogen-activated protein
L-NMMA NG-monomethyl-L-arginine monoacetate
MMP matrix metalloproteinase
ADAM a disintegrin and metalloprotease
ADAMTS a disintegrin and metalloproteinase with thrombospondin motifs
TIMP tissue inhibitors of metalloproteinases
S1P Sphingosine 1-phosphate
cAMP Cyclic AMP
PKA protein kinase A
PDGF platelet-derived growth factor
EDG1 endothelial differentiation sphingolipid G-protein-coupled receptor-1