Fisher John P. e.a. (ed.) Tissue Engineering

Подождите немного. Документ загружается.

mikos: “9026_c027” — 2007/4/9 — 15:53 — page 24 — #24

27-24 Tissue Engineering

[136] Rohr S. and Salzberg B. Characterization of impulse propagation at the microscopic level across

geometrically deﬁned expansions of excitable tissue: multiple site optical recording of transmem-

brane voltage (MSORTV) in patterned growth heart cell cultures. J. Gen. Physiol. 1994; 104:

287–309.

[137] Kucera J., Kleber A., and Rohr S. Slow conduction in cardiac tissue, II. Effects of branching tissue

geometry. Circ. Res. 1998; 83: 795–805.

[138] de Bakker J.M., van Capelle F.J., Janse M.J., Tasseron S., Vermeulen J.T., de Jonge N., and Lahpor J.R.

Slow conduction in the infarcted human heart. ‘Zigzag’ course of activation. Circulation 1993; 88:

915–926.

[139] Folch A. and Toner M. Microengineering of cellular interactions. Annu. Rev. Biomed. Eng. 2000;

2: 227–256.

[140] Khademhosseini A., Suh K.Y., Yang J.M., Eng G., Yeh J., Levenberg S., and Langer R. Layer-by-layer

deposition of hyaluronic acid and poly-L-lysine for patterned cell co-cultures. Biomaterials 2004;

25: 3583–3592.

[141] Fast V.G. and Ideker R.E. Simultaneous optical mapping of transmembrane potential and

intracellular calcium in myocyte cultures. J. Cardiovasc. Electrophysiol. 2000; 11: 547–556.

[142] Sperelakis N. and Haddad G. Developmental changes in membrane electrical properties of the

heart. In Sperelakis N., Ed. Physiology and the Pathophysiology of the Heart.3rded.Norwell,MA:

Kluwer Academic Publishers, 1995, pp. 669–700.

[143] Athias P., Frelin C., Groz B., Dumas J., Klepping J., and Padieu P. Myocardial electophysiology:

intracellular studies on heart cell cultures from newborn rats. Path. Biol. 1979; 27: 13–19.

[144] Joyner R. Interactions between spontaneously pacing and quiescent but excitable heart cells. Can.

J. Cardiol. 1997; 13: 1085–1092.

[145] Wagner M., Golod D., Wilders R., Verheijck E., Joyner R., Kumar R., Jonsma H., Van Ginneken A.,

and Goolsby W. Modulation of propagation from an ectopic focus by electrical load and by

extracellular potassium. Am. J. Physiol. (Heart Circ. Physiol.) 1997; 272: H1759–H1769.

[146] Antonutto G. and di Prampero P.E. Cardiovascular deconditioning in microgravity: some possible

countermeasures. Eur. J. Appl. Physiol. 2003; 90: 283–291.

[147] Decker M.L., Simpson D.G., Behnke M., Cook M.G., and Decker R.S. Morphological analysis of

contracting and quiescent adult rabbit cardiac myocytes in long-term culture. Anat. Rec. 1990;

227: 285–299.

[148] Clark W.A., Decker M.L., Behnke-Barclay M., Janes D.M., and Decker R.S. Cell contact as an

independent factor modulating cardiac myocyte hypertrophy and survival in long-term primary

culture. J. Mol. Cell. Cardiol. 1998; 30: 139–155.

[149] Simpson D.G., Decker M.L., Clark W.A., and Decker R.S. Contractile activity and cell–cell contact

regulate myoﬁbrillar organization in cultured cardiac myocytes. J. Cell Biol. 1993; 123: 323–336.

[150] Clark W.A., Rudnick S.J., LaPres J.J., Andersen L.C., and LaPointe M.C. Regulation of hypertrophy

and atrophy in cultured adult heart cells. Circ. Res. 1993; 73: 1163–1176.

[151] Ferrick K.J., Maher M., Roth J.A., Kim S.G., and Fisher J.D. Reproducibility of electrophysiological

testing during antiarrhythmic therapy for ventricular arrhythmias unrelated to coronary artery

disease. Pacing Clin. Electrophysiol. 1995; 18: 1395–1400.

mikos: “9026_c028” — 2007/4/9 — 15:53 — page1—#1

Tissue Engineering of

Heart Valves

K. Jane Grande-Allen

Rice University

28.1 The Native Heart Valve as a Design Goal............... 28-2

Anatomy and Terminology • Microstructure and Material

Behavior • Valvular Cells

28.2 Approaches to the Tissue Engineered Heart Valve .... 28-6

Biodegradable Polymeric Scaffolds • Decellularized Leaﬂet

Scaffolds • Cell Seeding • Natural Materials • Building

Block Approach • Cell Origin • Bioreactors for

Conditioning and Proof of Concept • In Vivo Testing in

Animal Models • Clinical Experience

28.3 Conclusions and Future Challenges .................... 28-13

Deﬁning Terms .................................................. 28-14

References ....................................................... 28-15

Heart valves are essential to the normal function of the heart and cardiovascular/cardiopulmonary systems.

When functioning properly, the heart valves allow unrestricted, unidirectional blood ﬂow through the

heart for subsequent distribution throughout the body. Consequently, valve disease or dysfunction can

result in signiﬁcant harm, as the reduction in the forward ﬂow of blood limits the oxygenation of the tissues

and can induce cardiac, cardiovascular, or cardiopulmonary compensation. Valve disease is prevalent in

our society, with valve replacement or repair in approximately 90,000 people in the United States in 2001

[1] (275,000 worldwide [2]). Moreover, valve disease can be either congenital or acquired. For example,

approximately 9 to 14 of every 10,000 children born are affected with the Tetralogy of Fallot [3,4],

a congenital heart disorder characterized by a narrowing of the pulmonary valve among other anomalies.

Acquired valve disease can affect people of all ages and may be due to an infectious agent (rheumatic

heart disease, endocarditis), systemic diseases (lupus, carcinoid syndrome), other cardiac disease, trauma,

pharmacologic agents, aging-related changes, or many other causes, some of which remain unknown [5].

The majority of current treatments for heart valve disease involve elective surgical replacement of

the valve with a mechanical, bioprosthetic, or cryopreserved allograft (homograft) valve. The allograft

is the treatment of choice for children, because bioprosthetic valves will calcify rapidly in children and

mechanical valves cannot growwith the child [6]. Aortic and pulmonary allografts havealso been used very

successfully in adults, with the pulmonary conduit having a 90% freedom from replacement at 20 years

[7], but the vascular remnant of these allografts eventually calciﬁes. Unfortunately, allografts, much like

other donated organs, are in scarce supply. Moreover, allografts needs to be matched to the recipient tissue

type to prevent immunological rejection [7], which narrows the diminishing pool of donated organs

28-1

mikos: “9026_c028” — 2007/4/9 — 15:53 — page2—#2

28-2 Tissue Engineering

even further. Alternative options such as mechanical or bioprosthetic heart valves can be used in many

situations, and are widely available, but have their own limitations [5]. Mechanical heart valves require

anticoagulation therapy, which some patients cannot tolerate. Bioprosthetic valves do not require any

anticoagulation, but do not contain any living tissues, and they undergo stiffening, calciﬁcation, and

structural deterioration in vivo as a result of their glutaraldehyde ﬁxation during manufacturing [8].

Bioprosthetic valves demonstrate a freedom from structural deterioration of 49% at 10 years and only

32% at 15 years, [9] and eventually require another surgical replacement. Overall, there is great need

for a living, unﬁxed tissue-engineered heart valve (TEHV) or valved conduit in adults who require valve

replacements. A TEHV with the potential for growth would also provide pediatric patients with a superior

alternative for the treatment of valve defects.

28.1 The Native Heart Valve as a Design Goal

28.1.1 Anatomy and Terminology

The aortic valve is one of four valves in the heart, but it is replaced most frequently, and therefore will be

discussed in greater detail. The aortic valve consists of three pieces of connective tissue (the right, left, and

noncoronary leaﬂets) that are attached to the aorta at one edge, and are free to move at the other edge.

These free edges meet centrally to close the valve and keep the blood from reentering the left ventricle. The

valve is located in the bulbous base of the aorta, which is known as the aortic root (an anatomic recreation

for a TEHV is shown in Figure 28.1). There are several distinct anatomic regions of the valve leaﬂet itself

(Figure 28.2). The leaﬂet attachment edge inserts into the aortic root wall at the crown-shaped annulus

[10]. The common region where two leaﬂets insert into the root wall is their commissure. The leaﬂet

belly, or body, is the main portion of the leaﬂet (0.4 mm thick in humans [11]), and bears the majority

of the pressure load when the valve is closed [12]. The coaptation area is the 0.5 to 0.6 mm thick [11]

region of the leaﬂet that is in contact with the two other leaﬂets when the valve is closed. The free margin

is the unattached edge of the leaﬂet, and suspends the leaﬂet between the tops of the commissures, much

like cables of a suspension bridge [10,13]. Finally, the central portion of the edge of the valve leaﬂet is the

Nodule of Arantius, a thickened area (0.95 to 1.2 mm [11]) that helps maintain valve closure [14].

The pulmonary valve (the other “semilunar” valve) is located in the pulmonary root between the right

ventricle and pulmonary artery, and is thinner and more delicate than but otherwise almost identical to the

aortic valve. The semilunar valves are quite different structurally from the “atrioventricular” valves. The

mitral valve consists of two differently shaped leaﬂets attached at their outer border to the junction between

the left atrium and left ventricle. The free edges and ventricular surfaces of the leaﬂets are connected to

the papillary muscles of the left ventricle by numerous chordae tendineae. Likewise, the tricuspid valve is

located between the right atrium and right ventricle. The tricuspid valve also contains chordae, but has

three differently shaped leaﬂets as opposed to two. The tricuspid leaﬂets and chordae are thinner, shorter,

and more delicate than those in the mitral valve.

28.1.2 Microstructure and Material Behavior

The semilunar valve leaﬂets consist of three histologically deﬁned layers: the ventricularis forms the lower

surface, the ﬁbrosa forms the upper surface, and the spongiosa layer lies in between [15] (Figure 28.3 and

Figure 28.4). The ventricularis contains a meshwork of elastic ﬁbers along with loosely scattered collagen

ﬁbers [15]. The predominant elastic makeup allows this layer to expand in response to tension in the

closed state of the valve, and then retract when the valve opens in response to ventricular ejection [15].

The ﬁbrosa contains collagen ﬁbers (predominantly type I), which are aligned largely circumferentially,

although radially aligned ﬁbers are found near the root-valve annulus [15]. The collagen ﬁber bundles

in the ﬁbrosa serve as the main source of strength for the diastolic pressure. The ridged appearance

of the ﬁbrosa is attributed to a corrugation of that tissue layer, in addition to the collagen bundles

[16]. The spongiosa is a gelatinous layer containing loose connective tissue that is rich in proteoglycans

mikos: “9026_c028” — 2007/4/9 — 15:53 — page3—#3

Tissue Engineering of Heart Valves 28-3

FIGURE 28.1 PGA/P4HB scaffold after cell seeding and 2 weeks of bioreactor conditioning. (Reprinted from

Hoerstrup, S.P., Sodian, R., Daebritz, S., Wang, J., Bacha, E.A., Martin, D.P., Moran, A.M., Guleserian, K.J.,

permission from Lippincott Williams & Wilkins.)

(PGs), and serves as a mechanism for compressive resistance [17] and shear between the ﬁbrosa and

ventricularis [16].

The mitral and tricuspid valves have a similar laminated structure, except the respective outer layers are

“upside down” from the arrangement shown in Figure 28.3. In these valves, the thick, heavily collagenous

layer is located on the ventricular side, whereas the thin, predominantly elastic layer is found on the atrial

side; these layers are also separated by a spongiosa. These similarities, which may be beneﬁcial in future

tissue engineering of atrioventricular valves, end with the chordae tendineae, which are not found in

semilunar valves. The chordae are strong, thin, cable-like structures that contain a core of highly aligned

collagen inside a thin outer sheath of elastic ﬁbers and endothelial cells.

The interaction between the extracellular matrix (ECM) constituents within the valve microstructure

allows distensibility, strength, elastic recovery, viscoelasticity, and an even distribution of deformation over

a wide range of loading [18]. Like many other biological soft tissues, the stress–strain and load elongation

curves of heart valve tissues are characterized by a low pretransition elastic modulus at initial strain (due to

elastic ﬁbers), followed by a transition zone to a higher posttransition elastic modulus at higher strains (due

to the uncrimped collagen) [18]. The unique collagen and elastic ﬁber arrangements in the different layers,

however, bestow the leaﬂet with anisotropic behavior (Figure 28.5). The greater circumferential stiffness

(due to collagen) contributes to normal aortic valve function by restricting downward leaﬂet motion,

while the lower radial stiffness permits the inward motion toward leaﬂet coaptation. This properly closing

aortic valve will allow blood ﬂow from the left ventricle into the ascending aorta, and prevent reverse

ﬂow. During this functional cycle, the leaﬂets interact with complex patterns of blood ﬂow [5], and are

mikos: “9026_c028” — 2007/4/9 — 15:53 — page4—#4

28-4 Tissue Engineering

Commissures (2 of 3)

Annulus

Free margin

Coaptation area

Belly

Attachment edge

Nodule of arantius

(a)

(b)

FIGURE 28.2 Semilunar valve leaﬂet anatomy. (a) Illustration of aortic valve leaﬂets within an opened aortic root.

(b) Single valve leaﬂet.

200 mm

FIGURE 28.3 The histological layers of the aortic valve leaﬂet. Movat’s Pentachrome stain. F, ﬁbrosa; S, spongiosa;

V, ventricularis.

subjected to transvalvular pressures as high as 120 mmHg [19] and shear stresses as high as 7.9 Pa [20].

These magnitudes of load are lower in the pulmonary circulation, where the transvalvular pressure across

the pulmonary valve is only 25 mmHg [19].

28.1.3 Valvular Cells

Heart valves contain both endothelial and interstitial cells [21–23]. The valvular endothelial cells (VECs)

populate the outer surfaces of the valves, whereas “interstitial cells” are all the cells that populate the inside

mikos: “9026_c028” — 2007/4/9 — 15:53 — page5—#5

Tissue Engineering of Heart Valves 28-5

Collagen

crimp

Elastin

Systole

Corrugations

Diastole

FIGURE 28.4 Illustration of the histological layers of the valve during systole and diastole. (Reprinted from

Fibrosa

fresh

cirumferential

Ventricularis

fresh

cirumferential

Fibrosa

fresh

radial

Vertricularis

fresh

radial

300

200

100

0 20406080100

Strain (%)

Stress (kPa)

SEM

FIGURE 28.5 Radial and circumferential stress–strain curves of the ventricularis and ﬁbrosa of aortic valve leaﬂets.

(Reprinted from Vesely, I. and Noseworthy, R., Micromechanics of the ﬁbrosa and the ventricularis in aortic valve

of the leaﬂets. Although it is presumed that the VECs provide the tissue with a nonthrombogenic sur-

face, their functions otherwise are only beginning to be explored [23,24]. The valvular interstitial cells

(VICs), which are only slightly better understood, are responsible for the synthesis of extracellular matrix

components, including collagen, elastin, proteoglycans, and hyaluronan (for reviews see References 25

and 26). A key characteristic of VICs is that this group of cells exhibits a mixed phenotype of both ﬁbro-

blastic and smooth muscle cell characteristics and are yet uniquely different from both these cell types

[21,22,27,28]. It remains unclear whether this dual phenotype is caused by a single population of cells

that express both features simultaneously [22,27], a single population of cells that can switch between

these two phenotypes [21], or a population of several types of cells [29,30]. The different phenotypes of

VICs are typically distinguished by their morphological appearance and immunohistochemical staining

[21,27,28]. The ﬁbroblastic phenotype is marked by elongated cells that contain numerous organelles for

mikos: “9026_c028” — 2007/4/9 — 15:53 — page6—#6

28-6 Tissue Engineering

matrix synthesis, and stain for prolyl-4-hydroxlyase. The smooth muscle cell phenotype is denoted by

cobblestone cells that stain for smooth muscle α-actin and stress ﬁbers. VICs display this dual phenotype

consistently throughout passaging [22,27]. Although in native valves these different cell phenotypes are

slightly segregated [29,30], cells harvested from different regions of the valve had a consistent dual phen-

otypic appearance and growth characteristics [22]. It has also been difﬁcult to separate these phenotypes

in culture [28].

28.2 Approaches to the Tissue Engineered Heart Valve

The basic approach to constructing a TEHV, as with many other engineered tissues, is ﬁrst to seed an

appropriate cell type onto or within a suitable scaffold, and then to have a period of incubation during

which the cells remodel or otherwise become integrated with the scaffold, and form a neotissue. This deﬁn-

ition is intentionally ambiguous because a wide range of cells, scaffolds, and incubation environments

have been used. All of the proposed TEHVs, however, have been designed to have as many of the ideal

features of a heart valve substitute as possible [31,32] (i) maintain normal structural and biological func-

tion over the patient’s lifetime, (ii) not elicit any inﬂammatory, foreign body, or immunologic responses,

(iii) have antithrombotic surfaces and the potential for growth and self-repair, (iv) manufactured for each

individual, (v) easy to implant with little technical variability, and (vi) available in an unlimited supply

[31,32]. The majority of research of TEHVs has focused upon the structural, antithrombotic, immuno-

logic, and availability aspects of this lofty goal. Although the studies described here do not represent an

exhaustive discussion of TEHVs, several reviews provide more thorough detail [2,26,33].

28.2.1 Biodegradable Polymeric Scaffolds

Almost half of the proposed designs for TEHVs involve seeding cells on or within a polymeric biode-

gradable scaffold. The purpose in using a biodegradable scaffold is to anchor the seeded cells within an

environment that is originally strong enough to withstand the in vivo mechanical forces, yet will sub-

sequently degrade slowly, thereby transferring the function of load-bearing to the nascent ECM produced

by the cells. Ideally, scaffold degradation rate and cellular synthesis rate should be balanced so that the

scaffold has been completely degraded when the seeded cells have generated an amount of ECM compar-

able to native heart valves. The scaffold should also have initial material properties that are comparable to

native valves.

The ﬁrst such scaffold used to generate a TEHV was a woven mesh of 90% poly(glycolic acid)

(PGA)/10% poly(lactic acid) (PLA) sandwiched between nonwoven PGA mesh, which was seeded with

ovine vascular myoﬁbroblasts and endothelial cells and used to replace the right leaﬂet of the pulmonary

valve in a lamb model [34]. Although the resulting pulmonary valve was functional in the short term

(3 weeks), this scaffold was found to be too stiff and thick for long-term use [35,36]. Conversely, seeding

cells in PGA mesh alone produced a neotissue that was too delicate to handle [37], although the high

porosity of this scaffold (95%) encouraged high seeding efﬁciencies and subsequent ECM production [35].

To avoid the mechanical limitations of the previous scaffolds, Sodian et al. [38] developed a new TEHV

scaffold using poly(hydroxyalkanoate) (PHA), a thermoplastic, easily moldable polymer. The polymer was

cast into a valved-conduit-shaped mold, made porous through a salt leaching process, and seeded with

autologous ovine vascular myoﬁbroblasts and endothelial cells. Although this valved conduit functioned

normally in a sheep model for up to 17 weeks, the PHA scaffold material did not degrade fully by that time,

and the developing neotissue did not contain any histologically detectable elastin or have an endothelial

cell coating [38]. Moreover, PHA has a high echocardiographic density, which prevented the TEHV

performance from being evaluated by Doppler echocardiography.

Because the PHA did not degrade rapidly enough, valved conduits were next assembled from nonwoven

PGA mesh coated with a thin layer of poly-4-hydroxybutyrate (P4HB), a biodegradable thermoplastic

moldable polymer that provided the nonwoven PGA with additional strength [39]. After seeding with

mikos: “9026_c028” — 2007/4/9 — 15:53 — page7—#7

Tissue Engineering of Heart Valves 28-7

(a)

(a) (b)

(b)

(c)

FIGURE 28.6 After6weeksin vivo (a), the TEHV from Figure 28.5 demonstrates preliminary organization (50×).

After 16 weeks (b) and 20 weeks (c), the TEHV leaﬂet demonstrates organized, dense collagen on the outﬂow

surface, elastic ﬁbers on the inﬂow surface (arrow), and spongy organization within (both 100×). (Reprinted from

Hoerstrup, S. P., Sodian, R., Daebritz, S., Wang, J., Bacha, E.A., Martin, D.P., Moran, A.M., Guleserian, K.J.,

permission from Lippincott Williams & Wilkins.)

autologous ovine vascular cells and 2 weeks of dynamic conditioning in a bioreactor (Figure 28.1), these

TEHVs were implanted in the pulmonary position of sheep for 20 weeks. These TEHVs functioned well,

with only mild to moderate pulmonary regurgitation at 16 and 20 weeks. Upon explant, the TEHV leaﬂets

were found to have the normal three-layered structure of native heart valves (Figure 28.6), although their

biochemically measured concentrations of collagen, elastin, and GAGs, as well as their ultimate tensile

strength, were signiﬁcantly higher than normal native pulmonary leaﬂets. Overall, the PGA/P4HB has

been considered a very successful scaffold for TEHV development and is still being investigated [40,41].

Other biodegradable scaffolds that have been explored include biodegradable polyurethane, which was

found to have not degraded completely at 6 weeks [42]. Finally, new classes of biodegradable scaffolds

are being designed, such as a combination of poly(vinyl alcohol) (PVA) with brush groups of modiﬁed

PLA [43]. This novel scaffold should combine the advantages of PVA, a high water content hydrogel with

high elasticity and the ability to incorporate biologically active molecules on its hydroxyl groups, with the

features of PLA, which is biodegradable, can be crosslinked, and is hydrophobically attractive to cells.

28.2.2 Decellularized Leaﬂet Scaffolds

Although polymeric biodegradable scaffolds have a long history in TEHV designs, another early approach

that is still in active development today is the use of decellularized semilunar valve leaﬂets as a scaffold, with

the rationale that they would provide the requisite strength [44], and already contain ECM in the correct

microstructural arrangement [45]. Unlike polymeric designs, there is no need to fabricate or mold these

scaffolds. Moreover, removing the cells would presumably eliminate the most antigenic elements, thereby

avoiding any immunological response in the recipient. This reasoning has enabled the development

of decellularized scaffolds from not only human heart valves (from donated allograft organs [45–47] or

cadavers [48]), but also porcine heart valves. The predominant matrix element in these scaffolds, collagen,

is highly conserved between species and is thus considered minimally antigenic [49,50], although there

are concerns about the potential for transmission of xenogenic diseases [51].

A major area of research in the development of this scaffold is determining the best method to remove

the cells from the original valve leaﬂet. Several different methods, involving ionic and nonionic detergents,

mikos: “9026_c028” — 2007/4/9 — 15:53 — page8—#8

28-8 Tissue Engineering

solutions that are hypotonic or hypertonic, and enzyme treatments, have been attempted, abandoned,

debated, and revisited, with only few direct comparisons [52]. In early studies, a treatment involving

hypotonic saline to lyse the cells followed by two washes with the nonionic detergent Triton X-100 and

one enzymatic soak in DNAse and RNAse was effective in the removal of all cells and cell debris [44,53].

This method preserved the majority of the thermal, physical, and material properties with the exception

of a slight swelling and a slight increase in the stress relaxation of the tissue. This successful treatment was

in contrast to their previous experimentation with the ionic detergent sodium dodecyl sulfate (SDS), in

which the leaﬂet matrix swelled up to three times and had signiﬁcant thermal denaturation [53]. On a

single wash basis, however, SDS appears to remove more cells than Triton X-100 [54]. A combination of

0.5% trypsin and 0.2% EDTA was also successful in removing cells from human and porcine valves [48,55],

as was a solution of 1% deoxycholic acid [56]. A solution of 0.1% N -cetylpyridinium chloride was shown

to remove cells effectively and to preserve the tissue’s microstructure and mechanics but this treatment

induced calciﬁcation when the decellularized leaﬂet was tested in a rat subcutaneous dermal model [57].

Booth et al. [52] found that solutions of 0.03 to 0.1% SDS and 0.5% Na deoxycholate in hypotonic solutions

worked best (with SDS causing a slight increase in tissue extensibility [58]), which they attribute to better

protease inhibition than in the previous studies that implicated SDS in ﬁber damage. Although a few

studies reported using protease inhibitors [48,53,55], such as phenylmethyl sulfonyl ﬂuoride (PMSF) and

EDTA, to block the endogenous lysosomal proteases released during cell lysis and to prevent degradation

of the matrix scaffold, certain protease inhibitors (including PMSF) are short-lived in aqueous solutions

and a more stable compound such as aprotinin may be preferable [52]. Many studies did not report any

use of protease inhibitors, which could result in partial degradation of the collagen and elastic matrix

components of the scaffold. The partial degradation of elastin in these scaffolds is considered particularly

risky given that decellularized aortic wall, found to contain an abundance of partially degraded elastin,

was prone to calciﬁcation in a rat subcutaneous dermal model [59].

Once prepared, the decellularized leaﬂet scaffold is almost always reseeded with autologous cells derived

from the same host animal to be used in the TEHV study. Several of these scaffolds have been reseeded

with endothelial cells only [56,60]. The main intent of this seeding is to form an antithrombotic coating

around the bare collagen [60], and is an especially important consideration in planning for human TEHV

use, because humans may have a more difﬁcult time endothelizing structures than do the sheep models

used in most of these studies [61]. Many other decellularized scaffolds, however, have also received a

preliminary reseeding with vascular myoﬁbroblasts in attempts to accelerate the eventual remodeling of

the matrix [51,55,62–64]. Despite the preliminary seeding of the decellularized scaffolds, dispersing the

cells within the existing matrix has proven difﬁcult, with some tendency for the cells to remain on the

surface or to merely line the largest pores [65]. In addition, Steinhoff et al. [62] found that the seeded cells

tended to make new matrix on top of as opposed to within the existing scaffold matrix.

The Synergraft™ valve (Cryolife, Inc., Kennesaw, Georgia) consists of a decellularized porcine pul-

monary root or composite aortic root (constructed from three noncoronary root-valve segments);

decellularized human allografts are also available [46,66]. In contrast to the other approaches, the Syner-

graft valved conduits were not reseeded with any cells before implantation in sheep models, but became

entirely repopulated with host cells and were completely functional for one year [45]. Although these scaf-

folds were developed with many techniques similar to those used for other TEHVs, they are not universally

considered tissue-engineered structures because they are not reseeded with cells before implantation.

28.2.3 Cell Seeding

The methods used to seed the cells on and within the polymeric biodegradable scaffolds and the decellu-

larized valve leaﬂets tend to be very straightforward: a concentrated solution containing the cells is dripped

onto the scaffold surface and the cells disperse by gravity [67,68]. This seeding dispersion was encouraged

by gentle agitation in some studies [30], but there was no report of any improved seeding efﬁciency due

to this method. Many TEHVs have been developed by seeding ﬁrst with myoﬁbroblasts (several million

cells), incubating 10 to 14 days, and then seeding with endothelial cells [34,37,38,62,69–72]. In preparation

mikos: “9026_c028” — 2007/4/9 — 15:53 — page9—#9

Tissue Engineering of Heart Valves 28-9

for this staged seeding process, a mixed population of vascular cells can separated into endothelial and

nonendothelial cells using ﬂuorescence-activated cell sorting (FACS)-based binding to acetylated low-

density lipoprotein (positive binding in endothelial cells [73]). Staged seeding is not necessarily required

[74]; mixed vascular cells seeded onto PGA/PLA–PGA sandwich scaffolds tended to segregate during

incubation and in vivo implantation, forming neotissue with endothelial cells (staining for Factor XIII) on

the outside and myoﬁbroblasts (negative for Factor XIII) within. Other factors shown to improve seeding

efﬁciency include 24 or 36 h seeding intervals (as opposed to 2 or 12 h intervals [68]), using polymeric

scaffolds with high porosity such in PGA [67,68], mixing soluble collagen into the cell seeding solution

[70], and coating the scaffold with Matrigel before seeding [75].

28.2.4 Natural Materials

Although the majority of TEHVs to date have been constructed from either biodegradable polymeric or

decellularized leaﬂet scaffolds, there are a number of alternative scaffolds that have been developed using

natural polymers such as collagen. The rationale behind using these natural materials is that the synthetic

biodegradable polymers may have cytotoxic degradation products such as lactic acid, which can lower

the pH of the culture medium [76]. A very early approach, reported by Carpentier et al. [77], was to

inject solubilized collagen into a leaﬂet shaped mold. The resulting valve was implanted in sheep in the

tricuspid or mitral position and functioned for up to 10 months without incompetence. Upon explant,

ﬁbroblasts were found within the collagen leaﬂets, but the leaﬂets were determined to have thickened

slightly in vivo. More recently, valvular interstitial cells seeded within collagen sponge scaffolds were

found to demonstrate phenotypic characteristics very similar to cells within intact valve leaﬂets [78].

Rothenburger et al. [71,75] grew human and porcine vascular and valvular cells within a freeze-dried

porous type I collagen matrix in vitro and found that the cells produced the large hydrating proteoglycan

(PG) versican, the small collagen-binding PG decorin, ﬁbronectin, thrombospondin, and a medium-

sized heparan sulfate PG (possibly perlecan or syndecan). The production of these PGs and glycoproteins

indicate that the cells were interacting with and organizing the nascent ECM. Another natural material,

ﬁbrin, is being explored as a natural scaffold because ﬁbrin gels can be made autologously from a patient’s

own blood and thereby prevent an immunologic reaction [67,68,76]. Fibrin gel components can also be

coupled with exogenous biologically functional groups, such as growth factors, for improved neotissue

formation. In addition, the use of injection molding to cast a cell-seeded ﬁbrin gel ensures that the cells

will be evenly distributed throughout the TEHV. This initial even distribution of cells is an advantage

over the seeding of ﬁbrous or sponge structures, where seeding dispersion is due to gravity and barely

50% of cells attach [67]. The disadvantage with ﬁbrin gels is their initial weakness, which needs to be

improved before in vivo studies can be performed. Yet another natural polymer that has been proposed

is chitosan, a polysaccharide derivative that has been used for other tissue engineering applications [23].

Although three-dimensional chitosan scaffolds have yet to be tested in a TEHV, monolayer cultures of

VECs adhered better to chitosan surfaces — and chitosan/collagen IV combinations in particular — than

to PHA surfaces.

28.2.5 Building Block Approach

Almost without exception, the approach to TEHV development has been to use scaffolds that are destined

for degradation or remodeling by the cells that will populate these constructs. The exception to this

paradigm is the approach by Vesely et al. [79], in which the different ECM structural components are

derived and then assembled into an approximation of the native aortic valve microstructure in vitro.The

collagen bundles that provide the leaﬂet strength are replicated by neonatal rat aortic smooth muscle cells

(NRASMCs) seeded within a type I collagen gel and anchored to promote uniaxial or branched contrac-

tion. The lubricating glycosaminoglycan component will be represented by cross-linked hyaluronan (also

seeded with NRASMCs), and the network and sheets of elastic ﬁbers are synthesized by the NRASMCs atop

the crosslinked hyaluronan and around the collagen bundles. The ﬁnal valve structure will be assembled