Tissue
Engineering:
The
University of Ghent in vitro experience.
Guido
Van Nooten
Hartcentrum, Ghent University Hospital
In
general terms, tissue engineering can be described as the development of viable
tissue by seeding living cells onto substrates that serve as scaffolds or
matrices in order to develop living constructs or tissues which are either
totally new or are representative of their naturally occurring counterparts. In
the Department of Cardiac Surgery at the University Hospital Ghent, tissue
engineering research is focused on the application of tissue engineering
techniques in the development of a tissue engineered aortic heart valve. The in
vitro research comprises four main areas:
I.
Selection and development of suitable scaffolds.
II.
Selection and characterization of appropriate cell types for tissue valve
engineering.
III.
Cell seeding.
IV.
Evaluation of tissue viability.
I.
Selection and development of suitable scaffolds.
At
present, scaffolds used in tissue valve engineering include biological matrices
created from porcine aortic valves, synthetic matrices manufactured from
biodegradable polymers such as polyglycolic acid and the less developed hybrid
matrices which use polymers embedded into biological matrices. Potential
scaffolds tissue from both are evaluated before and after conversion into
acellular matrices for specific properties required of the aortic valve and for
their biocompatibility (‘cell friendliness’). In our current work, we use
biological matrices derived from porcine and kangaroo aortic valves.
At
present biological valves are decellularized using a Modified
Detergent/Enzymatic treatment. Essentially tissues are subjected to alternating
hypotonic and hypertonic treatment to rupture cells before enzymatic digestion
of cellular debris and agitated extraction (gentle shaking during 24 hours).
Matrices prepared are evaluated for comparative purposes using the methodology
employed to assess the fresh tissue. Matrices are further tested for potential
toxicity to cells by evaluating cell survival using matrices prepared by our
protocol as substrates for cell expansion. For practical purposes, in order to
ensure a readily available supply of matrices of various sizes on short notice
for eventual clinical application, we have investigated the feasibility of
matrix preservation by cryopreservation.
Biological
valves as potential scaffolds are evaluated on:
(a)
Morphology, Histology and Ultrastructure
The normal aortic valve leaflet is comprised of structural (collagen
mainly collagen Type II, elastin and proteoglycans) and cellular (valvular
myofibroblasts and endothelial cells) elements. Myofibroblasts are mainly found
in the spongiosa while endothelial cells line the surface of the leaflet.
Classic methods such as light microscopy (LM), transmission electron microscopy
(TEM), scanning electron microscopy (SEM) and polarised light microscopy (PLM)
are used.
(b)
Tissue Biochemistry
To ensure adequate preservation of the extra-cellular matrix after
processing of tissue to produce a-cellular scaffolds, total collagen and
proteoglycan content are estimated by quantitative determination of
hydroxyproline and total uronic acid in fresh tissue.
(c)
Calcification Potential
Calcific deterioration is a major cause of failure of both native aortic
valves and currently available bioprosthetic heart valves. Biological tissues
that are potential sources of scaffolds are evaluated for their calcification
potential for comparative purposes with current bioprosthetic valves and tissue
engineered valves [1]. In keeping with the literature, tissues are evaluated
using the rat subcutaneous model to obtain as an initial indication of
calcification potential. Calcific deposits are identified histologically by Von
Kossa silver stain and determined quantitatively by Inductively Coupled Plasma
Spectrometry.
These
studies have shown that kangaroo and porcine valves differ in the proportional
thickness of the layers of the leaflet, despite having the same basic
histological components, namely collagen, elastin, proteoglycans and
fibroblastic cells.
II.
Selection and characterization of appropriate cell types.
Heart
valve fibroblasts are comprised of a heterogeneous population of mesenchymal
cells of which the major phenotype is that of myofibroblast.
Our initial experiments evaluated both adult bone marrow stem cells and
mesenchymal cell lines. Bone marrow stem cells as isolated by Prockop could not
be satisfactorily expanded under conditions available to us [2]. This, in
addition to the unpredictability of stem cell differentiation and recent reports
of unsuccessful use of stem cells in tissue valve engineering has resulted in a
temporary suspension of our efforts using Prockop cells. Our current work uses
mesenchymal cell lines and endothelial cells. Four mesenchymal cell types have
been isolated and characterised for application:
(a)
Human arterial mesenchymal cells (pulmonary artery)
(b)
Human Dermal Fibroblasts (neonatal foreskin)
(c)
Sheep Arterial Mesenchymal Cells (sheep carotid artery)
(d)
Sheep Dermal Fibroblasts.
Cells
are isolated from fresh tissue onto a plastic substrate and expanded in vitro
under appropriate conditions. Early passage cells were characterised and frozen
for future use. Mesenchymal cell used by us are classified as either
fibroblastic (< 50% myofibroblasts) after being scored morphologically,
immunohistochemically and metabolically. Morphologically we classify
myofibroblasts as spindle shaped cells with well-developed cell stroma and
prominent stress fibers. Immunohistochemically cell populations are
characterised by molecular markers using antibodies to vimentin (intermediate
filament present in mesenchymal cells), cytokeratin (intermediate filament
mostly found in epithelial cells), prolyl 4-hydroxylase (enzyme required for
hydroxyproline production in collagen synthesis) and smooth muscle alpha actin (a
SMA, the determinant characteristic of myofibroblasts, which is an isoform of
actin and found mainly in smooth muscle cells and which is responsible for
contraction and force generation in myfibroblasts). The metabolic and functional
properties of cells are evaluated by measuring the expression of extra-cellular
proteins (collagen type I, the adhesion molecule Tenacin-C and the pro-migratory
molecule N-Cadherin). Isolated human and sheep dermal and arterial mesenchymal
cell populations were found to comprise at best 50% myofibroblasts. To increase
myofibroblast count for cell seeding, we have investigated the
transdifferentiation of fibroblastic cell populations with the cytokine
transforming growth factor-b1
(TGF-b1).
Transforming growth factor successfully transdifferentiates both human and sheep
fibroblasts to metabolically active and functional myofibroblasts and allows us
to produce large quantities of myofibroblasts for cell seeding [3].
Heart
valve leaflets are covered by a mono-layer of endothelial cells. These cells
express platelet endothelial cell adhesion molecules (PCAM-1), vascular adhesion
molecules (VCAM-1) and human lymfocyte antigen (HLA11). Each of these molecules
can interact with T-lymfocytes and are responsible for the immunogenicity of
heart valves. In addition, endothelial cells have been suggested as a possible
source of valvular myofibroblasts. Cells are expanded and sewn onto biological
matrices in an initial step to valve seeding.
It
has been well established that fibroblasts respond to stress by developing
stress fibers and thus transdifferentiating to myofibroblasts. Such stresses are
present in heart valve leaflets and it has been suggested that they may play a
role in both the formation of myofibroblasts and the orientation of these cells
in heart valve leaflets. Such orientation is important in the orientation of
newly produced collagen fibers in valvular leaflets. Matrices are seeded (4x106
cells /cm²) under, (a) static conditions; in supplemented or un-supplemented
culture medium and under (b) dynamic conditions; in a custom developed
bioreactor that allows seeding under specified flow and pressure and temperature
conditions representative of in vivo dynamic conditions. Successful cell seeding
requires proliferation, differentiation and invasion of cells into the
scaffolds. These processes are dependent on both cell-cell and cell-molecule
interactions. Growth factors either produced by the cells themselves or
exogenously added affect proliferation, differentiation and invasion by binding
as ligands onto enzyme linked receptors (serine/threonine and thyrosine
receptors) with consequent downstream effects. It is our aim to establish a
functional populated heart valve matrix by seeding human dermal mesenchymal
cells (hDMC) on a decellularized porcine heart valve matrix.
Under in vitro conditions, we treat the fibroblasts with combinations of
growth factors to stimulate them to proliferate, to invade and to
transdifferentiate into contractile myofibroblasts. We have shown previously
that TGF-b
stimulates transdifferentiation of fibroblasts into a-SMA
positive myofibroblasts. However TGF-b, as such, is unsuitable for effective population of a heart valve matrix
since it dose-dependently inhibits growth of dermal fibroblasts as evidenced by
nuclear Ki67 (protein involved in cell proliferation) staining, mitochondrial
dehydrogenase assay (MTT-test) assay and cell counting. We have focused our
efforts on finding a growth factor combination that maintains TGF-b
induced transdifferentiation but overcomes TGF-b induced growth inhibition. We
combined TGF-b with several growth factors such as IGF-I, IGF-II, EGF, bFGF, PDGF-A,
PDGF-B, PDGF-AB, since reversed transscriptive polyclonal chain reaction
(RT-PCR) revealed that the respective growth factor receptors were expressed in
hDMC cells. Nuclear Ki67 staining,
MTT assay and cell counting showed us that only EGF and bFGF were capable to
overcome TGF-b induced growth inhibition. However
bFGF but not EGF inhibited TGF-b induced a-SMA expression as evidenced by immunocytochemistry and Western blotting.
Tissue
viability is tested by:
(i)
characterisation of cells by morphology and
histochemistry as described above,
(ii)
evaluation cell proliferative activity using the
MTT assay and Ki67 staining
(iii)
evaluation of metabolism by collagen production.
Cell
invasion is evaluated microscopically (LM, TEM and Confocal EM). Tissue surface
is evaluated by scanning electron microscopy (SEM).
References
1. Van Nooten G. Artificiële hartkleppen: state
of the art. Verh
K Acad Geneeskd Belg. 2003;65(3):135-80.
2. Colter DC, Sekiya I, Prockop DJ.
Identification of a subpopulation of rapidly self-renewing and multipotential
adult stem cells in colonies of human marrow stromal cells. Proc Natl Acad Sci U
S A. 2001 Jul 3;98(14):7841-5.
3. Narine K, DeWever O, Cathenis K, Mareel M,
Van Belleghem Y, Van Nooten G. Transforming growth factor-beta-induced
transition of fibroblasts: a model formyofibroblast procurement in tissue valve
engineering. J Heart Valve Dis. 2004 Mar;13(2):281-9; discussion 289.
Legend
Figure1.
TGF-b
transdifferentation of human arterial mesenchymel cells (HAMC).
a.
Vimentin content in untreated cells
b.
Vimentin content in TGF-b
treated cells
c.
a-SMA
content in untreated cells
d.
a-SMA
content in TGF-b
treated cells