Biomatrix/Polymer Composite Material for Heart
Valve Tissue Engineering
CARDIOVASCULAR
Christof Stamm, MD, Amir Khosravi, MD, Niels Grabow, MS, Kathleen Schmohl, PhD,
Nadine Treckmann, BS, Anne Drechsel, BS, Ma Nan, PhD, Klaus-Peter Schmitz, PhD,
Axel Haubold, PhD, and Gustav Steinhoff, MD
Department of Cardiac Surgery, Institute for Biomedical Engineering, Research Center for Cardiac Tissue Replacement, University
of Rostock, Rostock, Germany
Background. Decellularized extracellular matrix has
been suggested as a scaffold for heart valve tissue engi-
neering or direct implantation. However, cell removal
impairs the physical properties of the valve structure and
exposes bare collagen fibers that are highly thrombo-
genic. Matrix/polymer hybrid valves with improved bio-
logical and mechanical characteristics may be
advantageous.
Methods. Porcine aortic valves were decellularized
enzymatically and impregnated with biodegradable
poly(hydroxybutyrate) by a stepwise solvent exchange
process. Biocompatibility was tested in vitro using cell
proliferation and coagulation assays. Proinflammatory
activity was assessed in vivo by implantation of matrix/
polymer patches in the rabbit aorta. Biomechanic valve
properties and fluid dynamics were tested in a pressure/
flow-controlled pulse duplicating system. Matrix/
polymer hybrid valves were implanted in pulmonary
and aortic position in sheep.
T he currently available tissue valve prostheses based
on aldehyde-fixed xenogenic tissue are inevitably sub-
ject to calcium phosphate deposition and degeneration. The
tanning process effectively devitalizes the native cell popu-
lation, denaturizes antigenic protein domains, and changes
the scaffold protein architecture rendering in vivo repopu-
lation with recipient cells impossible. Furthermore, none of
the currently available heart valve prostheses has the po-
tential for growth, limiting their use in infants and children.
Goal of the current heart valve tissue engineering efforts is
therefore the development of a valve prosthesis that com-
bines unlimited durability with physiologic blood flow
pattern and biologically inert surface properties [1–3].
Recently, extracellular heart valve matrix was suggested as
a scaffold for tissue engineering, providing the natural valve
architecture and ideal conditions for repopulation with
recipient cells [4]. However, there are at least two major
problems: first, the mechanical tissue properties deteriorate
Accepted for publication March 25, 2004.
Presented at the Fortieth Annual Meeting of The Society of Thoracic
Surgeons, San Antonio, TX, Jan 26 –28, 2004.
Address reprint requests to Dr Steinhoff, Department of Cardiac Surgery,
University of Rostock, Schillingallee 35, D-18057 Rostock, Germany;
e-mail: gustav.steinhoff@med.uni-rostock.de.
© 2004 by The Society of Thoracic Surgeons
Published by Elsevier Inc
Results. Biocompatibility assays indicated that human
blood vessel cells survive and proliferate on matrix/
polymer hybrid tissue. In vitro activation of cellular and
plasmatic coagulation cascades was lower than with
uncoated control tissue. After implantation in the rabbit
aorta, matrix/polymer hybrid patches healed well, with
complete endothelialization, mild leukocyte infiltration,
and less calcification than control tissue. Matrix/polymer
hybrid tissue had superior tensile strength and suture
retention strength, and hybrid valves showed good fluid
dynamic performance. The two valves in aortic position
performed well, with complete endothelialization and
limited inflammatory cell invasion after 12 weeks. Of the
two valves in pulmonary position, one failed.
Conclusions. Matrix/polymer hybrid tissue valves have
good biological and biomechanic characteristics and may
provide superior replacement valves.
(Ann Thorac Surg 2004;78:2084 –93)
© 2004 by The Society of Thoracic Surgeons
when cells are removed and the tertiary structure of fibrous
valve tissue constituents is altered during the decellulariza-
tion process; second, open collagen surfaces are highly
thrombogenic, because collagen directly induces platelet
activation as well as coagulation factor XII. To address these
issues, biomaterial/polymer composite materials based on
decellularized vascular matrix scaffolds that are coated with
biodegradable polymers were developed; the hypothesis
that such hybrid tissues exhibit improved biocompatibility
in vitro and in vivo was tested.
Material and Methods
Animals received humane care in compliance with the
Guide for the Care and Use of Laboratory Animals prepared
by the National Academy of Sciences and published by
the National Institutes of Health (National Institutes of
Health Publication No 86 to 23, revised 1996). The proto-
col was reviewed and approved by the local animal care
committees.
Decellularization
Enzymatic removal of cells without altering the biochem-
ical characteristics of the extracellular matrix by exposure
0003-4975/04/$30.00
doi:10.1016/j.athoracsur.2004.03.106
Ann Thorac Surg
2004;78:2084 –93 BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR
to chemical fixatives was performed as previously de-
scribed [1]. Briefly, hearts were harvested from porcine
cadavers (courtesy of Agricultural Research Center,
Dummerstorf, Germany). The aortic root was prepared,
washed in phosphate buffered saline (PBS), and incu-
bated in 0.05% trypsin solution for 48 hours at 37°C,
followed by 3 washing steps for 1 hour each. For longer
storage, the specimens were lyophilized (at ⫺40° C and
0.05 mbar) and rehydrated before further use, but for
implantation in sheep the valves were processed imme-
diately. Biocompatibility tests in vitro and in rabbits were
performed using human aortic wall tissue, which was
obtained during routine coronary artery bypass grafting
(CABG) operations and was processed in identical
fashion.
Polymer Coating
Biodegradable polymers were supplied by Tepha Inc.
(Cambridge, MA). Poly(3-hydroxybutyrate) (P3HB),
poly(4-hydroxybutyrate) (P4HB), and poly(3- hydroxybu-
tyrate-co-4-hydroxybutyrate) (P3/4HB) in powder form
were dissolved in chloroform at 50°C according to the
chosen concentration. Initially, a dip-coating process was
used for fabrication of biomatrix/polymer hybrid tissue:
Decellularized and lyophilized tissue was repeatedly
immersed in 2% to 6% (w/v) polymer solution followed
by solvent evaporation for two weeks until the chloro-
form content was less than 0.2%. Before further use, the
tissue was rehydrated in cell culture medium (in vitro
experiments) or phosphate buffered saline (in vivo exper-
iments). When it became evident that this coating may
not withstand systemic hemodynamic forces, the proto-
col was switched to a modified polymer impregnation
process: freshly decellularized valve tissue was subjected
to a wet dehydration process, replacing water with eth-
anol. Then, the specimens were immersed in 1% (w/v)
polymer solution for 30 minutes, followed by rehydration
and wet solvent elimination in PBS.
In Vitro Cell Proliferation
Whether biomatrix/polymer hybrid valve tissue can be
repopulated with blood vessel cells, without exhibiting
cytotoxicity, was tested in series of in vitro experiments.
Tissue samples were prepared as described above,
seeded with L929 mouse fibroblasts and incubated under
standard cell culture conditions for at least 72 hours. Cell
viability was then assessed using the CellTiter96 fluores-
cent cell proliferation assay (MTS test). Because cellular
adhesion, proliferation, metabolic activity, and resistance
to toxin are highly dependent on species and cell type,
similar tests were performed using a mixed population of
human vascular myofibroblasts and endothelial cells,
prepared from saphenous vein samples of CABG pa-
tients. The lumen was filled with collagenase A contain-
ing medium. After incubation for 20 minutes, detached
endothelial cells were flushed out and cultivated under
standard conditions. The remaining tissue was minced
and placed in smooth muscle cell growth medium. Myo-
STAMM ET AL 2085
fibroblasts migrate onto the dish surface, attach, and
proliferate. The MTS-tests were performed in decellular-
ized matrix treated with various polymer preparations,
but also with pure polymer samples as well as with
hydrid tissues following several sterilization and storage
CARDIOVASCULAR
protocols such as lyophilization, plasmasterilization,
FAD-sterilization, ethylene oxide sterilization, and
gamma sterilization (data not shown).
In Vitro Hemocompatibility
Complement and coagulation system activation in re-
sponse to different hybrid tissue preparations were stud-
ied in several in vitro assays. Activation of complement
factor C3 was assessed by ELISA for C3a-des-Arg, the
stable metabolite of activated C3 (Progen, Heidelberg,
Germany), following incubation of human plasma with
hybrid tissue samples for 60 minutes. Representative for
activation of the plasmatic clotting system, the concen-
tration of the prothrombin fragments F1 and F2 was
measured by ELISA (Behring, Marburg, Germany), again
after incubation of human plasma with hybrid tissue.
Finally, the response of the cellular clotting system to
biomatrix/polymer hybrid tissue was studied by measur-
ing platelet factor 4 in human plasma using the Asserach-
romPF4 assay system.
In Vivo Screening Tests
Intravascular biocompatibility of various hybrid tissue
preparations was tested in a rabbit model. Adult New
Zealand White rabbits were anesthetized, heparinized,
intubated, and ventilated. The abdomen was opened, the
abdominal aorta was dissected distal to the renal arteries,
clamped, and incised longitudinally. A patch of bioma-
trix/polymer hybrid tissue measuring approximately 5⫻3
mm was sutured in place using nonresorbable suture.
After 1, 3, or 6 months, the animals were sacrificed and
the aortic segment containing the patch was explanted
and prepared for histology. Sections were stained with
H&E or antibodies for immunohistology, and examined
by light microscopy. A scoring system was designed to
facilitate comparison of histologic findings. The following
histologic characteristics were studied: endothelialization
of the luminal patch surface, intima proliferation, inflam-
matory infiltration, calcification, cellular migration into
the patch material, thrombus formation, and formation of
a neo-elastica interna (Table 1).
Mechanical Testing
The biomechanical characteristics of various biomatrix/
polymer hybrid tissue preparations were extensively
evaluated and are described in detail elsewhere [5]. In
that study, suture retention strength of the aortic conduit
wall, as well as tensile strength and elastic properties of
the leaflets were measured. The fluid dynamic properties
of intact hybrid valves were tested in a pulse-duplicating
system according to ISO 5940 standards. High-resolution
videomorphometric analysis of leaflet motion was
performed.
 2086 STAMM ET AL Ann Thorac Surg
BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR 2004;78:2084 –93
CARDIOVASCULAR

Fig 1. Decellularization and polymer coating of porcine aortic valves. (A and B) Hematoxylin & eosin stain of the aortic media (A) before and
(B) after enzymatic decellularization demonstrates complete removal of all cellular material. (C–E) Electron microscopy of the luminal valve
surface (C) before decellularization, (D) after decellularization, and (E) after polymer penetration (here: P3/4HB). Note that the modified poly-
mer penetration process leaves the surface porous, facilitating recipient cell adhesion and penetration.

Large Animal Implantation underwent echocardiography and cardiac catheterization

Once the optimal biomatrix/polymer tissue composition
had been chosen, four porcine hybrid valves were im-
planted in sheep by Experimental Surgical Services
(ESS), University of Minnesota. Following approval by
the local regulatory bodies, two valves were implanted to
replace the native pulmonary root, and three valves were
implanted as freestanding aortic root replacement with
coronary reimplantation. One animal died while under-
going aortic valve surgery and was excluded from the
analysis. Three months after implantation, the animals
and were sacrificed. Morphologic analysis of the hybrid
valve was carried out by the local pathologist before the
valve tissue was divided in three parts and prepared for
further analysis by light microscopy, immunohistology,
and electron microscopy.
Histology
Hematoxylin & eosin, Mason’s trichrome, and Van Gie-
son staining of formalin-fixed paraffin-embedded tissue
was performed in standard fashion. For immunohistol-
Ann Thorac Surg STAMM ET AL 2087
2004;78:2084 –93 BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR

CARDIOVASCULAR
Fig 2. Cell proliferation on biomatrix/polymer hybrid tissue in vitro
(MTS test). (A) Cultivated L929 mouse fibroblasts; (B) mixed popu-
lation of human myofibroblasts and endothelial cells. Although
mouse fibroblasts appear to proliferate on P4HB and P3/4HB only,
with virtually no cell growth on P3HB, there is good proliferation of
human cells on all tested polymers. (OD ⫽ optical density; P3HB ⫽
poly[3-hydroxybutyrate]; P4HB ⫽ poly[4-hydroxybutyrate]; P3/4HB
⫽ poly[3-hydroxybutyrate-co-4-hydroxybutyrate].)

ogy, frozen sections were prepared from cryopreserved

tissue and incubated with monoclonal mouse antihuman
CD31 antibody (Dako, clone Nr. JC70A) or monoclonal
mouse antihuman smooth muscle actin antibody (Dako,
clone Nr. 1A4), followed by detection with peroxidase-
conjugated goat antimouse IgG secondary antibody
(Dako). For analysis of surface morphology and ultra-
structure, specimens were prepared for scanning elec-
tron microscopy and viewed using a Philips XL30ESEM
(FEI, Hillsboro, OR) electron microscope.
Results
The decellularization process removed all cellular com-
ponents of the porcine aortic valve leaflets and the aortic
wall (Figs 1A and 1B). After enzymatic removal of the
endothelial cell layer and cellular components of the
media, the fibrous network of collagen and elastin, and
Fig 3. Activation of the cellular clotting system, plasmatic clotting sys-
tem, and complement system in human plasma incubated with decellu-
larized matrix and biomatrix/polymer hybrid tissue assessed by ELISA
for (A) platelet factor 4, (B) prothrombin fragments F1 and F2 (pro-
thrombin), and (C) C3a-des-Arg. Shown are representative sets of ex-
periments using the same plasma sample. In the xenogenic setting (por-
cine matrix incubated with human plasma) polymer coating of
decellularized matrix attenuated the activation of both cellular and plas-
matic clotting system, while complement C3 activation remained un-
changed. (P3HB ⫽ poly[3-hydroxybutyrate]; P4HB ⫽ poly[4-hydroxy-
butyrate]; P3/4HB ⫽ poly[3-hydroxybutyrate-co-4-hydroxybutyrate].)
proteoglycans forms the luminal surface of the valve
scaffold (Figs 1C and 1D). As has been previously re-
ported, the enzymatic digestion process inevitably weak-
 2088 STAMM ET AL Ann Thorac Surg
BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR 2004;78:2084 –93
CARDIOVASCULAR

Fig 4. (A) Explanted abdominal aorta of a rabbit 12 weeks after implantation of a patch consisting of decellularized human aortic wall seen
from the adventitial aspect. The arrow indicates the patch. (B–D) Photomicrographs of explanted specimens 12 weeks after implantation (he-
matoxylin & eosin staining; original magnification 100⫻). The solid arrows indicate patch material; interrupted arrows indicate neoelastica
interna. (B) Decellularized uncoated matrix is clearly distinct from the native aortic tissue with no signs of integration or resorption. There are
no regions of calcification, moderate inflammatory infiltration, and formation of a thick neoelastica interna. (C) Matrix coated with P4HB. The
patch is partially resorbed; the remaining material is heavily calcified. There is formation of a neoelastica interna as well as significant adven-
titial thickening. (D) P3HB-coated hybrid tissue is partially reabsorbed and well integrated in the native aortic tissue. There is little inflamma-
tory infiltration, thin neoelastica interna, and near-normal adventitial tissue.

ens the mechanical characteristics of the valve, so that it
may not be able to withstand hemodynamic forces in the
systemic circulation [5]. The polymer impregnation pro-
cess, however, led to improved biomechanic properties
in terms of suture retention strength and tensile tissue
Table 1. Morphologic Analysis of Biomatrix/Polymer Hybrid
Material in Vivo (Rabbit Model)
strength. Furthermore, the fluid dynamic and morpho-
metric characteristics of polymer-impregnated valves re-
sembled those of native heart valves. As opposed to the
initially used dip-coating procedure, the modified poly-
mer penetration protocol did not alter the surface mor-
phology of the decellularized matrix, preserving the
tissue native texture and microporosity, thus facilitating
adhesion and migration of recipient cells (Fig 1E).

Uncoated P3HB P4HB In Vitro Experiments

Endothelialization 2.5 (2–3) 1.5 (1–2) 2.5 (1–3) In a first series of screening tests, the proliferation
Intima proliferation 2 (1–3) 1.75 (1–2) 1.67 (1–2) capacity of L929 mouse fibroblasts on different biomatrix/
Inflammatory infiltration 2 (2) 1.5 (1–2) 1.67 (1–2) polymer preparations, with and without lyophilization
Calcification 2.75 (2–3) 2.5 (2–3) 1.67 (1–2) and sterilization was assessed. A summary of the cell
Cellular migration 2 (2) 1.75 (1–2) 1.5 (1–2) proliferation data by MTS test in biomatrix coated with
Thrombus formation 3 (3) 3 (3) 2.67 (1,3) P3HB, P4HB, and P3/4HB is shown in Figure 2A. It
Neo-elastica interna 1 (1) 1 (1) 1 (1) appears that P3HB almost completely inhibits cell growth
on the matrix. These experiments were then repeated
Histologic scores of uncoated decellularized valve tissue (uncoated) and
biomaterial/polymer composite tissue (P3HB, P4HB) 3 months after using a mixed population of human endothelial cells and
implantation in the rabbit aorta. Note that the experiments were per- myofibroblasts. It was found that they proliferate very
formed using the polymer dip-coating process. Shown is the arithmetic
mean of the scores in each group and the range of scores assigned in a
well on all matrix/polymer combinations (Fig 2B). Hence,
group (in parenthesis). A scoring system of 1–3 was devised and a specific decellularized biomatrix/polymer hybrid tissue has the
definition assigned to every for each parameter. In general, a score of 3 potential for repopulation even in the xenogenic setting.
represents a favorable result (ie, no calcification, complete endotheliali-
zation), whereas a score of 1 describes an unfavorable result (ie, complete Representative data on hemocompatibility of hybrid tis-
thrombotic occlusion, absence of a neo-elastica interna). sue are shown in Figure 3. Matrix impregnation with
P3HB ⫽ poly(3-hydroxybutyrate); P4HB ⫽ poly(4-hydroxybutyrate). P3HB attenuated the activation of platelet factor 4 in
Ann Thorac Surg
2004;78:2084 –93 BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR
human plasma, indicating less activation of the cellular
clotting system (Fig 3A). With P3/4HB copolymer there
was still some attenuation of PF4 release, while P4HB
alone appeared not to change the matrix-induced PF4
activation. Activation of the plasmatic clotting cascade as
assessed by the concentration of prothrombin fragments
F1 and F2 was partially suppressed when the matrix was
coated with biodegradable polymer, irrespective of the
PHB-type used (Fig 3B). Finally, activation of the comple-
ment system in response to hybrid tissue was estimated.
Again an attenuation of C3a-des-Arg production in the
presence of polymer-coated matrix compared to un-
treated decellularized matrix was observed (Fig 3C).
In Vivo Screening Tests
In the rabbit abdominal aorta patch implantation model,
various combinations of polymer-matrix composites, but
also noncoated xenogenic matrix patches were evaluated
(n ⫽ 3, each). Representative photomicrographs are
shown in Figure 4, and part of the qualitative histologic
evaluation data are summarized in Table 1. At follow-up,
all rabbit aortas were patent and free from blood clot
formation, except in one animal with a P4HB-treated
patch. There were no patch aneurysms despite implan-
tation in the high-pressure system. There was some early
inflammatory cell infiltration (4 weeks) that later resolved
in all patches (12 weeks). Recipient blood vessel cells had
migrated into to patch material in both P3HB and P4HB
coated matrix. Importantly, very little or no calcification
occurred in the decellularized and P3HB-coated matrices,
in contrast to autologous blood vessel control patches
and some P4HB-coated matrices. There was some degree
of intimal thickening with formation of a neo-elastica
interna in all hybrid tissue patches, but the vessel lumen
was not narrowed in any of the P3HB-coated patches. As
confirmed by immunohistology staining for CD31, there
was complete endothelial cell lining of the luminal sur-
face of matrix/polymer hybrid patches (data not shown).
Large Animal Implantation
The preliminary biocompatibility and mechanical tests of
various hybrid material preparations indicated that
P3HB-treated xenogenic matrix tends to exhibit better
STAMM ET AL 2089
Fig 5. Gross morphology of the two por-
cine biomatrix/polymer hybrid valves 12
weeks after implantation in pulmonary
position in sheep. (A) The first valve was
completely degenerated probably due to
CARDIOVASCULAR
bacterial endocarditis. (B) The second
valve was in excellent condition, with
near-normal gross morphology.
biocompatibility but is rather stiff and brittle, while
P4HB-coated matrix is soft and pliable, but probably at
the cost of earlier calcification in vivo. Based on these
results a copolymer consisting of 82% P3HB and 18%
P4HB processed in 1% solution with freshly decellular-
ized matrix was chosen for further testing in a large
animal model.
The first valve implanted in pulmonary position was
severely obliterated, with severe diffuse pio-granulo-
matous valvular endocarditis, chronic lympho-histocytic
and necrotizing bioprosthetic periarteritis, and intimal
fibrous-hyperplasia. Although microbiology at the time
of sacrifice was negative, bacterial endocarditis was prob-
ably the cause (Fig 5A). The second valve was in excellent
condition, with near-normal gross morphology (Fig 5B).
On histology, there was complete endothelial cell lining
and moderate multifocal granulomatous inflammation
limited to the leaflet hinge point (data not shown).
Particular attention was paid to the two valves im-
planted in aortic position. At the time of sacrifice, both
animals were in good condition without clinical signs of
valve dysfunction. Gross morphology and representative
photomicrographs are shown in Figure 6. The pathologist
described that there were focal fibrinous deposits on the
leaflets. The leaflets contained homogeneous eosino-
philic bundle of collagens, expanded by moderate mul-
tifocal granulomatous inflammation (epitheloid macro-
phages and giant multinucleated cells). As seen in Figure
6B, there was some intimal thickening particularly of the
luminal surface of the otherwise delicate leaflets. By
immunohistology, complete endothelial cell lining of the
leaflets and the conduit wall was found. There was
smooth muscle cell migration into the media of the
leaflets (Fig 7), but very little cellular infiltration of the
conduit wall (Fig 6E). By scanning electron microscopy,
the luminal surface of the leaflets in aortic position
showed a smooth texture without apparent interruptions
of the intimal integrity (Fig 8).
Comment
The approach described herein to improve tissue engi-
neering heart valve design consisted of the following
 2090 STAMM ET AL Ann Thorac Surg
BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR 2004;78:2084 –93
CARDIOVASCULAR

Fig 6. Xenogenic biomatrix/polymer hybrid valve 12 weeks after implantation in aortic position in sheep. (A) Gross morphology reveals the
valve is in good condition, and the distal and proximal suture as well as both coronary orifices are clearly visible. The arterial wall lining is
smooth, the leaflets are delicate and freely mobile; (B) cross section through a leaflet and its hinge point at the conduit wall (Mason’s
trichrome staining). The collagenous valve scaffold is intact. There is some intimal thickening with inflammatory infiltration on the luminal
aspect of the leaflet. Panels C–F reveal the arterial wall of the conduit (hematoxylin & eosin stain): (C) hyperplasia of the tunica intima (ar-
rows), for comparison; (D) shows the native aorta of the same animal; (E) tunica media of the arterial conduit wall in which, at this point,
only few cells appear to have migrated into the media of the conduit wall; (F) for comparison, the native aortic media of the same animal is
illustrated.

steps: (1) extraction of a porcine heart valve and removal
of all xenogenic cells by enzymatic digestion without
altering the biological properties of the valve matrix
components (no tanning process); (2) penetration of the
decellularized matrix with biodegradable polymer to
enhance the mechanical characteristics of the porous
valve scaffold and to cover thrombogenic matrix compo-
nents. It was shown that coating with poly(hydroxybu-
tyrate) does indeed improve biocompatibility and me-
chanical properties in vitro, and that such hybrid tissue
heals in well when inserted as a patch in the rabbit aorta.
Finally, P3/4HB-impregnated xenogenic hybrid valves
were implanted in sheep. The two valves implanted in
aortic position functioned well for up to 3 months and
partially developed the morphologic characteristics of a
native aortic valve.
Due to the multiple shortcomings of conventional
prostheses, heart valve tissue engineering strategies have
attracted considerable attention. The initial approach was
based on the fabrication of the entire valve scaffold from
biodegradable polymers, followed by in vitro seeding
with autologous cells and conditioning under simulated
in vivo conditions before implantation [1–3]. Early results
were promising, but it soon became clear that the com-
plex three-dimensional structure of the native valve can
hardly be achieved with current techniques, and the
structural and mechanical properties of the various poly-
mers are not ideal. Furthermore, in vitro seeding and
conditioning with cells of the future recipient is a time-
consuming process that would require sophisticated lab-
oratory equipment in the clinical setting. In addition, it
remains unclear whether the cells actually adhere to the
scaffold after implantation, and there is evidence that the
cells found on in vitro-seeded valves months after im-
Ann Thorac Surg STAMM ET AL 2091
2004;78:2084 –93 BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR

Fig 7. Immunohistology staining of a xe-

nogenic biomatrix/polymer hybrid valve
leaflet in aortic position 12 weeks after
implantation in sheep. (A) CD31 staining
demonstrating a confluent layer of endo-

CARDIOVASCULAR
thelial cells covering the leaflet predomi-
nantly on the luminal aspect but also on
the mural surface (original magnification
⫻100). (B) Higher magnification (⫻1000)
illustrating a monolayer of endothelial
cells with multiple CD31⫹ cells migrating
into inner layers of the leaflet. (C) Leaflet
stained for smooth muscle actin (⫻400);
multiple SMA⫹ cells are integrated in the
fibrous scaffold of the hybrid valve tissue.
(D) Stain of the corresponding negative
control.

plantation are in fact cells that have colonized the valve trix scaffold with those of an artificial polymer and
postimplantation, independent from the preseeding pro- demonstrated that polymer-treated valve scaffolds ex-
cess. More recently, natural xenogenic or allogenic heart hibit improved resistance to hemodynamic forces [5].
valve tissue has been propagated as a scaffold for heart These composite heart valves can now be implanted in
valve tissue engineering. In contrast to aldehyde-fixed vivo, and will be repopulated by recipient cells. The
tissue, enzymatically decellularized extracellular matrix polymer coating will then be degraded, while cells mi-
without tanning-induced crosslinks possesses epitopes grate into the extracellular matrix scaffold, restore the
for cellular adhesion receptors, facilitating repopulation natural tissue structure of the native heart valve, and
with tissue-specific celltypes but also inflammatory cells initiate the physiologic turnover of extracellular matrix
[6, 7]. Nonautologous matrix constituents such as colla- components. Once these processes have been completed,
gen, elastin, and proteoglycans have little antigenicity, the neo-valve should have biological and mechanical
given that cellular components are entirely removed [8, characteristics identical to those of the native valve.
9] It has recently been shown that mismatch of HLA-DR
and ABO antigens on endothelial cells in unmodified
valve allografts is associated with accelerated valve fail-
ure [10, 11]. To avoid the immunologic response to graft
cells, we have developed a decellularization process that
removes cells and cellular debris from the heart valve
scaffold by enzymatic digestion, without the use of ma-
trix-altering tanning procedures. Thus, the biological
properties of the extracellular matrix components are
preserved, and the decellularized valve can be repopu-
lated by cells derived from the recipient organism. In an
earlier series of experiments, we reseeded such decellu-
larized valve scaffolds with autologous blood vessel cells
in vitro and implanted those in pulmonary position in
sheep [4]. Three months after implantation, good valve
function and complete histologic restitution of valve
tissue, as well as a confluent endothelial surface were
found. However, aggressive enzymatic decellularization
inevitably weakens the valve tissue, so that the mechan- Fig 8. Surface morphology by scanning electron microscopy (origi-
ical properties do not allow for implantation in the high nal magnification ⫻200) of a xenogenic biomatrix/polymer hybrid
pressure system. Therefore, we sought to combine the valve 12 weeks in aortic position after implantation. Note that the
advantageous properties of the native extracellular ma- luminal surface of the leaflet is perfectly smooth.
 2092 STAMM ET AL
BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR
The coating process also serves to attenuate the pro-
coagulatory activity of bare matrix components. It is well
established that platelet activation occurs when platelets
come in contact with collagen. Platelets directly adhere
by binding of platelet collagen receptor to integrins on
CARDIOVASCULAR
collagen, and the intrinsic clotting cascade is initiated
when prekallikrein, kininogen, factor XI and factor XII
are exposed to collagen. In in vitro tests screening tests
we found that activation of both the cellular and the
plasmatic coagulation system is indeed attenuated by
polymer coating of decellularized matrix, while there was
no apparent change in the extent of complement
activation.
The use of uncoated decellularized matrix with or
without in vitro autologous preseeding for heart valve
replacement remains controversial. Recently, Leyh and
colleagues [12] reported that decellularized porcine pul-
monary valves implanted in pulmonary position in grow-
ing sheep functioned well for up to 24 weeks and exhib-
ited reconstitution of viable valve tissue and only mild
calcification. On the other hand, allogenic sheep valves
prepared in the same manner calcified rapidly, associ-
ated with structural and functional deterioration. In a
similar model, the same group observed that in vitro
repopulation with autologous cells before implantation
led to severe leaflet degeneration in vivo, while valves
that were implanted unseeded had clearly superior mor-
phologic and functional characteristics [13]. In another
study, however, they observed progressive degeneration
of xenogenic decellularized aortic tissue when implanted
subcutaneously in rats, and concluded that some recel-
lularization process may be necessary before implanta-
tion [14].
Decellularized valves are already commercially avail-
able. Preclinical large animal testing in the xenogenic and
allogenic setting was very promising, with good hemo-
dynamic function, morphologic reconstitution, and little
calcification for almost 1 year [15–17]. The initial enthu-
siasm has been muted, however, by reports of early
xenograft failure in humans [18, 19]. Decellularized ho-
mografts appear to cause less problems in humans, but
the actual long-term benefit remains to be determined
[20, 21]
In conclusion, we believe that decellularized heart
valve matrix holds great potential for creation of viable,
long-lasting replacement valves, and that many of the
problems in the xenogenic setting, including impaired
biomechanics and residual antigenicity can be overcome
by pretreatment with biodegradable polymers. However,
more extensive long-term studies in large animals are
clearly needed before clinical use can be considered.
We thank the staff of ESS Experimental Surgical Services,
University of Minnesota, for their excellence in carrying out the
valve implantations in sheep. The contribution of Dr Heiko
Klinge to the early rabbit experiments is acknowledged. We also
appreciate the help of the Department of Cardiac Surgery at
University of Luebeck with performing the fluid dynamic
analysis.
Ann Thorac Surg
2004;78:2084 –93
References
1. Shinoka T, Breuer CK, Tanel RE, et al. Tissue Engineering
Heart Valves. Valve Leaflet Replacement Study in a Lamb
Model. Ann Thorac Surg 1995;60:513–16.
2. Sodian R, Hoerstrup SP, Sperling JS, et al. Early In Vivo
Experience with Tissue-Engineered Trileaflet Heart Valves.
Circulation 2000;102:III22–9.
3. Hoerstrup SP, Sodian R, Daebritz S, et al. Functional Trileaf-
let Heart Valves Grown In Vitro. Circulation 2000;102:III44 –9.
4. Steinhoff G, Stock U, Karim N, et al. Tissue engineering of
pulmonary heart valves on allogenic acellular matrix con-
duits-in vivo restoration of valve tissue. Circulation 2000;102:
III50 –5.
5. Grabow N, Schmohl K, Khosravi A, et al. Mechanical and
structural properties of a novel hybrid heart valve scaffold
for tissue engineering. Artificial Organs 2004 (in press).
6. Courtman DW, Pereira CA, Kashef V, McComb D, Lee JM,
Wilson GJ. Development of a pericardial acellular matrix
biomaterial: biochemical and mechanical effects of cell ex-
traction. J Biomed Mater Res 1994;28:655–66.
7. Bader A, Schilling T, Teebken O, et al. Tissue engineering of
heart valves – human endothelial cell seeding of detergent
acellularized porcine valves. Eur J Cardiotharac Surg 1998;
14:279 –84.
8. Yannas IV. Natural materials. In: Ratner BD, Hoffmann AS,
Schoen FJ, Lemons JE, eds. Biomaterial science. San Diego:
Academic Press, 1996:84 –94.
9. Wilson GJ, Courtman DW, Klement P, Lee JM, Yeger H.
Acellular matrix: a biomaterial approach for coronary artery
bypass and heart valve replacement. Ann Thorac Surg
1995;60(Suppl 2):S353.
10. Baskett RJF, Nanton MA, Warren AE, Ross DB. Human
leukocyte antigen-DR and ABO mismatch are associated
with accelerated homograft valve failure in children: Impli-
cations for therapeutic interventions. J Thorac Cardiovasc
Surg 2003;126:232–9.
11. Hogan PG, O’Brien MF. Improving the allograft valve: Does
the immune response matter? J Thorac Cardiovasc Surg
2003;126:1251–3.
12. Leyh RG, Wilhelmi M, Rebe P, et al. In vivo repopulation of
xenogenic and allogenic acellular valve matrix conduits in the
pulmonary circulation. Ann Thorac Surg 2003;75:1457–63.
13. Leyh R, Wilhelmi M, Haverich A, Mertsching H. A xenoge-
neic acellularized matrix for heart valve tissue engineering:
in vivo study in a sheep model. Z Kardiol 2003;92:938 –46.
14. Walles T, Puschmann C, Haverich A, Mertsching H. Acellu-
lar scaffold implantation–no alternative to tissue engineer-
ing. Int J Artif Organs 2003;26:225–34.
15. Goldstein S, Clarke DR, Walsh SP, Black KS, O’Brien MF.
Transpecies Heart Valve Transplant: Advanced Studies of
Bioengineered Xeno-Autograft. Ann Thorac Surg 2000;70:
1962–9.
16. O’Brien MF, Goldstein S, Walsh S, Black KS, Elkins R, Clarke
D. The SynerGraft valve: a new acellular (nonglutaralde-
hyde-fixed) tissue heart valve for autologous recellulariza-
tion. Semin Thorac Cardiovasc Surg 1999;11(suppl):194 –200.
17. Elkins RC, Dawson PE, Goldstein S, Walsh SP, Black KS.
Decellularized human valve allografts. Ann Thorac Surg
2001;71(suppl):S428 –32.
18. Simon P, Kasimir MT, Seebacher G, et al. Early failure of the
tissue engineered porcine heart valve SYNERGRAFT in
pediatric patients. Eur J Cardiothorac Surg 2003;23:1002–6.
19. Sharp MA, Phillips D, Roberts I, Hands L. A cautionary case:
the SynerGraft vascular prosthesis. Eur J Vasc Endovasc
Surg 2004;27:42–4.
20. Bechtel JF, Muller-Steinhardt M, Schmidtke C, et al. Evalu-
ation of the decellularized pulmonary valve homograft (Syn-
erGraft). J Heart Valve Dis 2003;12:734 –9.
21. Sievers HH, Stierle U, Schmidtke C, Bechtel M. Decellular-
ized pulmonary homograft (SynerGraft) for reconstruction
of the right ventricular outflow tract: first clinical experience.
Z Kardiol 2003;92:53–9.
Ann Thorac Surg
2004;78:2084 –93 BIOMATRIX/POLYMER COMPOSITE MATERIALS USED IN HEART VALVE REPAIR
DISCUSSION
DR SCOTT M. BRADLEY (Charleston, SC): In terms of the final
biopolymer you arrived at—I think it was a combination of two
polymers— can you give us an idea of how long that takes to
degrade in vivo.
DR STAMM: The lifetime in vivo of P3 and P4HB is different, the
half-lives are different. For P3HB, I believe they are in the
magnitude of several months, if not years; P4HB degrades much
faster. About the overall total degradation time or half-life of the
polymer combinations, nothing is known.
DR ANTONIO CORNO (Lausanne, Switzerland): The compos-
ite valves that you prepared are exposed to different regimen in
pulmonary artery position regarding the pressure and the oxy-
gen saturation. Did you detect any difference to the response of
this valve?
DR STAMM: The valve in pulmonary position that survived the
procedure more or less intact had a complete endothelial lining.
We also found some smooth muscle cells in the leaflets, but we
were not able to detect a significant difference in terms of
histologic appearance between both implantation sites, aortic or
pulmonary.
The Society of Thoracic Surgeons: Forty-first Annual
Meeting
Please mark your calendars for the Forty-first Annual
Meeting of The Society of Thoracic Surgeons, to be held
in Tampa, Florida, from Jan 24 –26, 2005. The program
will provide in-depth coverage of thoracic surgical topics
selected to enhance and broaden the knowledge of car-
diothoracic surgeons. Attendees will benefit from tradi-
tional Abstract Presentations, as well as Surgical Forums,
Breakfast Sessions, Surgical Motion Pictures, and Town
Hall Meetings on specific topics.
Advance registration forms, hotel reservation forms,
and details regarding transportation arrangements, as
well as the complete meeting program, will be mailed to
Society members this fall. Also, complete meeting infor-
mation will be available on the Society’s Web site at
© 2004 by The Society of Thoracic Surgeons
Published by Elsevier Inc
STAMM ET AL 2093
DR HENRY L. WALTERS III (Detroit, MI): Did you do any
scanning electron micrography on the autopsy specimens of the
CARDIOVASCULAR
valves that you put in circulation to determine the uniformity of
endothelialization of the leaflets?
DR STAMM: Do you mean like the image that I showed of the
leaflet in aortic position?
DR WALTERS: That looked like a cross-section light micro-
graph to me. But you had shown some scanning electron
micrograph specimens of your combination matrices, and I only
saw light micrographs of a cross-section of the autopsy speci-
mens after they had been in circulation. I thought the endothe-
lialization was one of the remarkable findings, and I just won-
dered how uniform it is and if you have done any studies on the
cells to assess their function.
DR STAMM: As yet we only have cross-sectional images of both
conventional staining and immunohistology, but the endothelial
lining is complete in all the sections that we’ve looked at. We
have also looked at the entire surface by scanning electron
microscopy and again observed a complete cell lining.
www.sts.org. Nonmembers who wish to receive informa-
tion on the Annual Meeting may contact the Society’s
secretary, Gordon F. Murray.
Gordon F. Murray, MD
Secretary
The Society of Thoracic Surgeons
633 N. Saint Clair St, Suite 2320
Chicago, IL 60611-3658
Telephone: (312) 202-5800
Fax: (312) 202-5801
e-mail: sts@sts.org
website: www.sts.org
Ann Thorac Surg 2004;78:2093 • 0003-4975/04/$30.00