Tissue Engineering Heart Valves: Valve Leaflet

Replacement Study in a Lamb Model

Toshiharu Shinoka, MD, Christpher K. Breuer, MD, Ronn E. Tanel, MD,
Gregor Zund, MD, Takuya Miura, MD, Peter X. Ma, PhD, Robert Langer, PhD,
Joseph I’. Vacanti, MD, and John E. Mayer, Jr, MD
Departments of Cardiovascular Surgery and Surgery, Children’s Hospital, Boston, and Department of Chemical Engineering.
Massachusetts Institute of Technology;, Cambridge, Massachusetts

Background. Valve replacements using either biopros- (n = 4) tissue engineered leaflets were implanted in 7
thetic or mechanical valves have the disadvantage that animals. In each animal the right posterior leaflet of the
these structures are unable to grow, repair, or remodel pulmonary valve was resected and replaced with an
and are both thrombogenic and susceptible to infection. engineered valve leaflet.
These characteristics have significantly limited their du- Results. All animals survived the procedure. Postoper-
rability and longevity. In an attempt to begin to over- ative echocardiography demonstrated no evidence of
come these shortcomings, we have tested the feasibility stenosis and trivial pulmonary regurgitation in the au-
of constructing heart valve leaflets in lambs by seeding a tografts and moderate regurgitation in the allogenic
synthetic polyglycolic acid fiber matrix in vitro with valves. Collagen analysis of the constructs showed devel-
fibroblasts and endothelial cells. opment of an extracellular matrix. Histologic evaluation
Methods. Mixed cell populations of endothelial cells of the constructs demonstrated appropriate cellular archi-
and fibroblasts were isolated from explanted ovine arter- tecture.
ies. Endothelial cells were selectively labeled with an Conclusions. This preliminary experiment showed that
acetylated low-density lipoprotein marker and separated a tissue engineered valve leaflet constructed from its
from the fibroblasts using a fluorescent activated cell cellular components can function in the pulmonary valve
sorter. A synthetic biodegradable scaffold constructed position. Tissue engineering of a heart valve leaflet is
from polyglycolic acid fibers was seeded with fibroblasts, feasible, and these preliminary studies suggest that au-
which grew to form a tissue-like sheet. This tissue was tograft tissue will probably be superior to allogenic
subsequently seeded with endothelial cells, which tissue.
formed a cellular monolayer coating around the leaflet.
Using these constructs, autologous (n = 3) and allogenic (Ann Thorac Surg 1995;6O:S513-6)

T issue engineering
neering principles
is a new science that applies engi-

effort to create viable structures

to the biological sciences in an
to replace diseased or
Material
Cell isolation,
and Methods
cell culture, cell sorting, polymer charac-
teristics, and cell seeding technique were described in
deficient nature structure. Our laboratory has focused on
detail previously [l].
the use of a biodegradable polymer scaffold as a cell
delivery system. Basic cell types to be used in engineered Cell Isolation and Culfwe
tissue can be grown in vitro using basic tissue culture Cells were collected from 20-day-old lambs. Using sterile
techniques. Once cells are attached to the three- surgical technique, 2- to 3-cm sections of femoral artery
dimensional biodegradable polymer, the resulting tissue were harvested. The tissue was minced into l- to 2-mm
construct can be implanted in vivu, where the cells have pieces, which were evenly distributed onto tissue culture
the potential for further growth and development. The plates. Dulbecco’s modified Eagle medium (Gibco GRL,
cellular structure and matrix develop as the polymer Grand Island, NY) supplemented with 10% fetal bovine
degrades, ultimately leaving only engineered tissue with- serum (Gibe0 GRL) and 1% GPS (L-glutamine, penicillin,
out foreign material. We have constructed tissue engi- and streptomycin; Irvine Scientific, Santa Clara, CA) was
neered valve leaflets and have implanted both allogenic used for tissue culture medium. The medium was
and autologous tissue in the pulmonarv valve position. changed every 7 days. The explanted tissue was placed in
We compared the function of these constructs in \rivo and a humidified incubator at 37°C with 5% CO, for 8 to 10
evaluated their morphologic and histologic changes. weeks. The cells migrated off from the explants and
formed mixed cell populations.
Cell Separation
.After the mixed cell population had grown to confluence
the cells were labeled with an acetylated low-density

0003-4975/95/$9.50
SSDI 0003-4975(95)00733-4
s514 CASTANEDA FESTSCHRIFT SHINOKA ET AI Ann Thorac Surg
TISSUE ENGINEERING HEART VALVES 1995;6O:S513-6

Fig 1. Right posterior leaflet was resected and replaced with a tissue
engineered leaflef.

lipoprotein (LDL) marker, which was selectively taken

into endothelial cells via the scavenger passway. After a
24-hour incubation period the cells were separated into
two cell groups (LDL positive and LDL negative) using a
fluorescent activated cell sorter. The two cell populations
were separately passaged so as to obtain sufficient num-
bers of cells for cell seeding.

Bioabsorbable Polymer
The scaffold was composed of a polyglactin woven mesh Fig 3. Echocardiography through the window demonstrates fhick-
surrounded by two nonwoven polyglycolic acid mesh cnczd leaflet CT’EL) and a trivial pulmonary regurgitation PR).
sheets, which measured 3 cm by 3 cm by 3.2 mm. The PV ~rrlmonary uulve; RVOT = right ventricular ou$!ow tract.)
matrix was greater than 95% porous before seeding. The
scaffold was designed to degrade by hydrolysis over a 6-
to S-week period. Cell Seeding
One million LDL-negative cells (a combination of both
fibroblasts and smooth muscle cells) were seeded onto
the polymer each day for 2 weeks. Culture media was
changed on a daily basis. After 2 weeks the polymer-l
LDL-negative cell construct was seeded with three mil-
lion LDL-positive cells (endothelial cells), which formed
a monolayer around the tissue engineered construct.

Pulmona y Valve Leaflet Replacement

Allogenic tissue engineered leaflets were implanted into
20-day-old lambs (n = 4). Autologous tissue engineered
leaflets were implanted into the donor lambs (n = 3) from
which the cells had been harvested (average age, 82.6
days). Anesthesia was induced with 30 mglkg of ket-
amine and maintained with continuous infusion of 0.2
mg . kg ’ * min ’ of propofol. The chest was entered
through a left thoracotomy at the third intercostal space.
Using femoral arterial and right atria1 cannulation, nor-
mothermic total bypass was established. With the heart
Fig 2. Postmortem specimen showed that implanted allogenic tissue beating, a longitudinal pulmonary arteriotomy was
engineered leaflet was thickened. made. The right posterior pulmonary leaflet was excised
Ann Thorac Surg CASTAtiEDA FESTSCHRIFT SHINOKA ET AL S515
1995;60:S513-6 TISSLE ENGINEERING HEART VALVES

and replaced with the engineered construct, which had

been cut into a semicircular shape and sutured using
bioabsorbable material (5-O PDS; Ethicon, Somerville,
NJ) (Fig 1). The pulmonary arteriotomy was closed and
the lamb weaned off of bypass. Before the chest was
closed a section of the left third rib was removed to make
a window for postoperative echocardiography. Postoper-
atively lambs receiving allogenic tissue engineered leaf-
lets were immunosupressed using 2 mg . kg ’ . day ’ of
azathioprine, 1 mg . kg ’ . day ’ of prednisolone, and 20
mg . kg*+‘day ’ of cyclosporine. All animals received
humane care in compliance with the “Guide for the Care
and Use of Laboratory Animals” published by the Na-
tional Institutes of Health (NIH publication 85-23, revised
1985).

Evaluation of Tissue Engineered LeafIets Fig 4. Histologic section of an allogenic tissue engineered leaflet im-
plmrterf for 7 day showed thnt fhe Iraflets had the appropriate cellu-
Valve function was evaluated using Doppler echocardi- lar ~zrchitecbrc. The cellular matrix uas underdeveloped, and there
ography on postoperative days 2, 7, 14, and 21. A Vitro- ZLWS soww evidence of in$wnmation. Arrows indicate polyglycolic
dyne V-1000 mechanical tester (Lifeco, Burlington, VT) Ilc-id mzsl~ fibers.
was used to evaluate maximal tensile strength of one
tissue engineered leaflet before implantation 121, A 4-h?-
droxyproline assay [3] was used to measure collagen con- 3.1 MPa; our constructs, 3.4 MPa) before implantation.
tent in one tissue engineered leaflet that had been The 4-hydroxyproline assay demonstrated that tissue
implanted for 2 weeks and compared with native pulmo- engineered allogenic leaflet that had been implanted for
nary valve in the same animal. After animals wertb 7 days contained 30% of the collagen content of the native
sacrificed, valve leaflet tissue was fixed in formalin and pulmonary valve leaflet.
set in paraffin, and slides were stained with hematoxvlin Histologic examination of the tissue engineered con-
and eosin in 1 lamb that received an allogenic leafle;. structs for the l-day specimens showed that the leaflets
had the appropriate cellular architecture and resembled
native valve tissue histologically (Fig 4). However, the
Results
cellular matrix was less well developed and there was
We have constructed and implanted four allogenic leaf- evidence of an inflammatory reaction.
lets and three autologous pulmonary valve leaflets. All
animals survived the operation and were followed up for
various time periods. Two of 3 lambs receiving autolo- Comment
gous valves had no complications and are doing well. The Tissue engineering is a multidisciplinary science that
other 1 was sacrificed on the first postoperative day for uses basic principles from engineering and biology to
the histologic examination. Three of 4 lambs receiving construct tissue from their cellular components [4, 51.
allogenic valves had serious infectious complications Our laboratory has previously investigated the use of
after the operation and were sacrificed at 7,8, and 19 davs synthetic biodegradable polymeric scaffolds for other
after operation. The other 1 was sacrificed on the first tissue engineering applications 161. Cells are seeded onto
postoperative day for histologic examination of tht, irn- biodegradable, synthetic polymer scaffolds, and these
planted leaflet. One allogenic leaflet, which had been cells attach to the polymer, multiply, and develop into
implanted for 7 days, disappeared completely. The other tissue. Over time the polymer degrades as the cells
two showed significant shrinkage and deterioration (Fig secrete their own matrix and organize into functional
2). The two autologous valve leaflets that were allowed to tissue that can be used for reconstructive or transplanta-
remain in place functioned appropriately as demon- tion surgery 171.
strated by echocardiography. The tissue engineered Valvular heart disease is a significant cause of morbid-
valves were thicker than the native leaflets and appeared ity and mortality. Although the currently used replace-
to be less flexible than the native tissue. However, echo- ment valves, including mechanical valves, bioprosthetic
cardiographic examination of the pulmonary valve dem- xenograft valves, and homografts, are efficacious, they
onstrated no evidence of stenosis and trivial pulmonarv are still far from ideal. Mechanical valves require long-
regurgitation in the lambs receiving autologous leaflets life anticoagulation therapy to reduce their rate of throm-
(Fig 3). Additionally, these studies have shown that the boembolic complications. Bioprosthetic valves have lim-
autologous leaflets have maintained these size and shape ited durability and may calcify prematurely in young
after implantation. patients. Both mechanical and bioprosthetic valves are
The maximal tensile strength was measured using an nonliving structures and do not have ability to grow,
lnstron tensometer and showed similar values as com- repair, or remodel. Currently, homograft valves are fro-
pared with native valve tissue (native pulmonarv valve. len in liquid nitrogen and stored until used. Cryopreser-
S516 CASTANEDA FESTSCHRIFT SHINOK. El .AL Ann Thorac Surg
TISSUE ENGINEERING HEART \‘AL\‘ES 1995;60:5513-6

vation is known to preserve fibroblast viability [8]. How- whether tissue leaflets will be able to withstand hemo-
ever, even though homograft valves are viable they are dynamic stress. In this feasibility study we replaced only
allografts and are therefore subject to rejection. In addi- one leaflet in the low-pressure pulmonary circulation. It
tion, homograft supply is limited due to donor organ is unknown whether our constructs could tolerate sys-
scarcity. temic pressures. It may be possible to “condition” leaf-
The option of creating valve tissue from autologous lets in vitro before implantation, but additional studies
cells offers many potential advantages. These include will be necessary in this area as well.
elimination of the problems of rejection and donor organ Clearly the applications for a single leaflet are limited,
scarcity. These leaflets would be viable structures with but we are planning to make a three-leaflet valve using a
the potential advantage of greater durability and longev- modified Love’s method [lo]. In Love and associates’
ity because the tissue could use naturally existing mech- original studies autologous pericardium was used to
anisms for repair and remodeling. There should be no construct a valve leaflet, but the tissue was briefly fixed in
foreign body response or need for long-term anticoagu- glutaraldehyde and was not endothelialized. These two
lation, and there may even be the potential for growth, variables may prove to limit the long-term durability of
which will be important in the pediatric age group. these constructs. Our tissue contained both autologous
In our first experiments we performed allograft im- fibroblasts and endothelial cells, which could potentially
plantations. More recently we have begun autologous improve the durability and longevity of our leaflets.
tissue implantation. Despite the use of triple-agent im- In conclusion, these preliminary results demonstrate
munosupression, the allogenic cells appeared to have that the construction of autologous valve leaflets is fea-
evoked an acute inflammatory response. The basis for the sible. It appears that autologous tissue leaflets will be
severity of this immune reaciion is unclear. Whether the superior to allogenic tissue leaflet with respect to longev-
engineered tissue has added immunogenicity or whether ity and function.
the foreign body reaction caused by the synthetic matrix
may augment the rejection phenomena is uncertain. In
addition these animals had several infectious complica- References
tions including wound infections and infectious pericar-
ditis associated with the immunosuppression regimen. 1. Breuer CK, Shinoka T, Tanel RE et al. Tissue engineering
lamb heart valve leaflets. Biotech Bioeng (in press).
The autologous tissue engineered leaflets were not 2. Sato M, Maeta H, Okamura K, Ohshima N. Relationship
subject to the same problems as the allogenic tissue between mechanical and hydrodynamic properties of bio-
engineered leaflets. There was no evidence of valve prosthesis produced from canine aortic valve. Artif Organs
shrinkage or the development of stenosis or regurgitation 1985;9:184-91.
3. Bergman 1, Loxley R. Two improved and simplified methods
during the 4-week period during which we evaluated for the spectrophotometric determination of hydroxyproline.
valve function using echocardiography. In the 2 lambs Anal Chem 1963;35:1961-5.
receiving autologous tissue implantation there were no 4. Langer R, Vacanti Jl’. Tissue engineering. Science 1993;260:
infectious complications, leading us to believe that the 920-6.
.5. Vacanti JP. Beyond transplantation. Third annual Samuel
infectious complications in the allograft implants were
Jason Mixter lecture. Arch Surg 1988;123:545-9.
secondary to immunosuppression and not to surgical 6. Vacanti JP, Morse MA, Saltzman WM, Domb AJ, Perez-
technique. Atayde A, Langer R. Selective cell transplantation using
These studies are quite preliminary and should be bioabsorbable artificial polymers as matrices. J Pediatr Surg
considered only as initial feasibility studies. Numerous 1988;23:3-9.
7. Vacanti CA, Langer R, Schloo B, Vacanti JP. Synthetic
issues remain to be addressed. One issue is the selection polymers seeded with chondrocytes provide a template for
of the appropriate origins of the cells for cell cultures. We new cartilage formation. Plast Reconstr Surg 1991;88:753-9.
used arterial or dermal fibroblasts as a scaffold cell, not 8. Watts LK, Dully I’, Field RB, Stafford EG, O’Brien MF.
valve myofibroblasts. The characteristics of cells are very Establishment of a viable homograft cardiac valve bank: a
rapid method of determining homograft viability. Ann Tho-
different depending on their site of origin. However we
rat Surg 1976;21:230-6.
also know that cells can differentiate and reorganize in 9. Noishiki Y, Tomizawa Y, Yamane Y, Okoshi T, Satoh S,
the physiologic environment. Noishiki and associates [9] Matsumoto A. Acceleration of neointima formation in vas-
have shown that under certain conditions dermal fibro- cular prostheses by transplantation of autologous venous
blasts were able to differentiate into smooth muscle-like tissue’ fragments. ‘Application to small-diameter grafts.
I Thorac Cardiovasc Surg 1993;105:796-804.
cells and endothelial-like cells. Therefore, it may not be 10. iove JW, Schoen FJ, Br&nock EM, Shermer SP, Love CS.
necessary that cells used to form tissue engineered leaf- Experimental evaluation of an autologous tissue heart valve.
lets originate from leaflet tissue. A second question is J Heart Valve Dis 1992;1:232-41.