Tissue engineering for cardiac surgery

	INTRODUCTION

This chapter summarizes some of the progress that has been made in tissue engineering research as it relates to cardiac surgery and will seek to identify areas within the specialty where it is likely to have most impact. Even within this one specialty there are many different pathologies, and tissue engineering does not hold equal promise for all. Nevertheless, there are some areas where one might anticipate that tissue engineering solutions may have a profound impact on future surgical practice.

Tissue engineering is driven by the quest to restore function. It follows that in the creation of first-generation products, emphasis should be placed on their functional behavior in vivo rather than on esoteric parameters such as detailed tissue microstructure and elaborate histologic comparisons with normal tissues. At a very superficial level, it is clear that the heart has parts that function actively, such as the muscle and conduction pathways, and parts whose function is largely passive in response to intravascular flow and pressures created by the former elements. Historically, it has been possible to replace those parts of the heart that passively provide function, such as the cardiac valves and great arteries, baffles within the chambers of the heart, and patches to close defects between adjacent chambers with wholly synthetic elements. The results of these techniques have largely been very successful and hampered only by the interaction of body fluids with the surface of materials^1,2 or cavitation in the flow of blood through rigid moving parts.^3,4 With the possible exception of pacemakers, attempts to mimic actively functioning elements of the heart such as the myocardium have either failed or remain outside of wide clinical practice. There are still many problems to be overcome in respect of ventricular assist devices and artificial hearts.^5–7 So too it is anticipated that in the creation of tissue-engineered elements for the cardiovascular system, heart valves, conduits, and baffles/patches are likely to be somewhat easier to develop than a segment of graftable, coordinated, and contracting muscle. Tissue engineering holds one further promise that has peculiar significance to the pediatric population: the potential to create replacement materials that can grow or at least increase in size with the increase in size and attendant physiological demands on blood flow of the developing child.⁸ Although there remains no absolutely clear evidence to support the growth hypothesis, it is this anticipation that forms the cornerstone of research efforts in the field of tissue-engineered heart valves.

	THE CREATION OF HEART VALVES AND SIMPLE CONDUITS

 
It is natural that tissue engineering research should focus first on the pulmonary circulation where existing solutions are unsatisfactory,^9,10 and lower physiological pressures make this environment much more forgiving to the developing technology. The intention is then to translate these findings into valves and conduits for the systemic circulation that have the potential to benefit adult patients suffering from acquired valvular heart disease and ischemic heart disease. In a multidisciplinary collaborative effort that has come to characterize tissue engineering research in general, work at Children's Hospital, Boston, in conjunction with the Brigham and Women's Hospital and the Massachusetts Institute of Technology has focused on the optimizing the cell source, mechanisms for cell delivery, and scaffold design as they relate to the tissue engineering of heart valves. A considerable effort has been placed on selecting the optimal cell source for tissue engineering applications. In brief, the choice may be broadly categorized into stem cells or pleuripotent cells that exhibit the potential to differentiate into multiple cell types and fully committed or differentiated cells. The choice is further segmented between autologous and allogeneic cells, and between the anatomical source of those cells, i.e., from the proposed tissue that is to be fabricated or from anatomically distinct elements. The ultimate choice will likely reflect, however, a complex interplay between what is favored in terms of basic cell biology, what tissues are available for harvest in the individual patient, and what is clinically acceptable to both patient and physician. In the field of heart valves, considerable early success was achieved using immature myofibroblasts and endothelial cells derived from systemic arteries.^11–13 These cells were chosen by virtue of their anatomical location within the arterial limb of the cardiovascular system and their known ability to synthesize structural extracellular elements such as collagen. The prospect of harvesting segments of artery from an otherwise normal peripheral circulation presents a somewhat unattractive solution clinically, however, and led to the search for alternative cell sources. A comparison of myofibroblasts from the wall of the ascending aorta with those from discarded segments of saphenous vein revealed that the latter cells exhibit superior collagen formation and mechanical strength when cultured on biodegradable polyurethane scaffolds.¹⁴ Enhanced collagen formation may be a two-edged sword. The rapid formation of new tissue in the early culture period could give rise to uncontrolled proliferation and synthesis of matrix elements, similar to the development of intimal hyperplasia when venous grafts are introduced into the arterial circulation. Mature cells may, furthermore, present a problem of senescence in long-term culture, which typically spans many weeks. Techniques are currently being explored to increase the replicative lifespan of human vascular cells while maintaining their differentiated phenotype and capability of synthesizing extracellular matrix in vitro. The search for a more attractive source of cells has stimulated interest in the use of stem cells for tissue engineering. The latter may be isolated from a variety of tissues including skeletal muscle,¹⁵ adipose tissue,¹⁶ bone marrow,¹⁷ and peripheral blood.¹⁸ The latter study demonstrated that endothelial progenitor cells may be isolated from peripheral blood, expanded in vitro, and delivered onto decellularized allografts. These grafts remained functional for up to 130 days as carotid interposition grafts in sheep. The feasibility of using stem cells derived from the bone marrow of sheep to create semilunar valves has also been demonstrated by our own group both in vitro and in vivo (unpublished data). One final cell strategy is not to provide cells at all, but to allow grafts to "repopulate" in vivo from the host.¹⁹ In one study, porcine or sheep heart valves were decellularized and implanted as pulmonary or aortic valve replacements in sheep. Recellularization was evaluated histologically after intervals of between 3 and 11 months with cell phenotypes identified using specific antibodies. The heart valves showed progressive recellularization, with cell density and distribution at 11 months approaching that of native valve leaflets and localization of smooth muscle actin positive cells at the ventricularis/spongiosa interface. Of course this strategy is only applicable to grafts that are capable of functioning in the circulation as scaffolds alone, which is currently not the case with synthetic biodegradable scaffolds.

Cell Delivery Methodologies

A reliable method for delivering cells to polymer scaffolds in vitro is fundamental to the development of tissue-engineered structures. In early tissue engineering methodologies developed in our laboratory, valves were made on an individual basis by delivering cells in a concentrated paste onto scaffolds held in an open dish. These constructs were then left for an arbitrary period of time for cell attachment before being immersed in cell culture medium.^13,20 There were a number of theoretical problems intrinsic to this technique. First, it is very difficult to obtain an even distribution of cells over complex surfaces such as that of a semilunar heart valve scaffold; second, there is the potential for cell death through depletion of metabolites and accumulation of metabolic waste products in this extremely concentrated paste during the time allowed for cell attachment. An alternative technique is to deliver cells from suspension by immersing the entire scaffold within that suspension. However, in a static system, cells that do not immediately encounter the scaffold tend to deposit onto the base of the container and are effectively lost. It therefore becomes necessary to develop a mechanism for continual resuspension of unattached cells. Shaker and spinner flasks have been used in the past for this purpose.²¹ However, the physical forces created by such violent movements have the disadvantage of tending to disrupt nascent cell-polymer interactions. We compared the efficacy of two rotating systems as vehicles for cell delivery to our cardiovascular scaffolds, a rotating wall vessel (RWV, Synthecon Inc., TX), originally developed by NASA scientists to study the behavior of cells under conditions of microgravity, and a system of rotating individual sealed tubes (RIST). Scaffolds were fabricated from polyglycolic acid mesh dip-coated in poly-4-hydrobutyrate according to a method developed previously¹³ and placed in each of the cell culture systems. Cell-scaffold constructs were assayed for DNA content as a surrogate for cell number after 5 days. Total DNA per conduit was 226 ± 7 µg for conduits seeded in the RIST and 396 ± 18 µg for conduits in the RWV, suggesting that the RWV culture system provides a superior environment to the RIST for delivering cells onto polymer scaffolds. In this system, scaffolds are placed in the annular space between a cylindrical outer wall and a central core membrane; the entire assembly is placed in an incubator with atmosphere enriched with 5% CO₂ maintained at 37°C. This maintains a relatively constant physiological milieu protecting constructs from fluctuating temperature, pH, and mechanical forces disruptive to cell attachment. Indeed, the RWV has also seen wider application within tissue engineering, specifically to mature small pieces of PGA mesh seeded with chondrocytes.²² Such constructs exhibited far superior tissue formation to those grown in static culture or to the mixer flasks alluded to above.²³ It has further been shown that cell delivery systems that utilize rotation favor the uniform distribution of cells over polymer scaffolds and by corollary the uniform formation of tissue on valves engineered in vitro (FWH Sutherland and TE Perry, unpublished data). Work is currently underway to devise a system that combines the benefits of gas exchange afforded by the RWV with a scaffold-enclosing membrane that restricts the movement of cells to the vicinity of the scaffold yet allows free movement of media, nutrients and dissolved gases with the much larger volume of the RWV. At the time of writing we are seeking to overcome technical difficulties related to the selecting a membrane of suitable pore size for this purpose. In summary, the ability to deliver cells to cardiovascular scaffolds in a uniform and predictable manner will permit the generation of groups of scaffolds with identical numbers of cells. This is a prerequisite to the design of experiments to interrogate subsequent steps in the tissue engineering process. Ongoing research in this area is essential in order to facilitate scale-up and commercial application of tissue engineering for the benefit of patients.

Optimal Scaffold Design

A fundamental difference in the various tissue engineering strategies that have been adopted in the cardiovascular field centers on the choice of scaffold. The options fall broadly between the use of decellularized tissues or wholly synthetic matrices. Until recently, a significant advantage of decellularized grafts has been their established and appropriate geometry for valves and the density of extracellular material contained therein, which contributes significantly to the short-term integrity of small vessel conduits¹⁸ or heart valves.²⁴ Tissue-engineered grafts made in this manner and using a variety of cell types have proved functional following implantation in animal models.¹⁸ However, the disadvantages of decellularized grafts include the relative shortage of available homografts and immunogenecity problems that may arise from the use of decellularized xenogeneic tissues. Perhaps more importantly, the density of residual extracellular matrix that proves attractive from a structural point of view prevents the penetration of seeded cells into the interstices of the matrix. Furthermore, given the complex and ill-understood interaction between the cytoskeleton and extracellular matrix in normal tissues, it may be naïve to expect cells seeded onto decellularized tissues to assume those same relationships with foreign matrix elements.

One further scaffold that deserves mention is small intestinal submucosa (SIS). Grafts of SIS are prepared by removing the serosal layer, inverting the material, and removing the surface mucosa. The remaining submucosa and stratum compactum constitute the graft material.²⁵ SIS is an acellular structure that elicits relatively little cross-species immunogenicity.²⁶ In these respects, SIS represents the ideal naturally occurring material and preliminary studies have suggested that it may provide a very favorable substrate for tissue regeneration. Indeed, SIS has been implanted, as a substitute patch, into a variety of anatomical locations including the ureter,²⁷ dura,²⁸ and small bowel itself.²⁹ Each study has demonstrated regeneration of host tissue without histologic evidence of sensitization or graft rejection. SIS has also recently been sutured into a tubular form and implanted as an interposition graft into the carotid artery of dogs. Mechanical properties, over time, appear to approach those of the native carotid artery, angiograms show uniform flow without dilatation at two months, and the luminal surface of explanted grafts appears white, shiny, and glistening.³⁰ These encouraging results prompted interest in the use of this material for the fabrication of a substitute heart valve. Unfortunately, a major stumbling block remains in how to fabricate the highly complex, multiply curved, surface geometry of a heart valve and accompanying root from flat sheets of tissue. This is a problem that has faced heart valve manufacturers for many years. One could certainly draw upon traditional methods of heart valve assembly that, for example, utilize Elgiloy stents. However, this approach would seem to offer little or no advantage over the well-proven, stented, pericardial valves in use today.

Biodegradable polymers that have thus far been used for tissue engineering applications are mainly based on materials already in clinical use. In the group of macromolecules of natural origin collagen, alginate, agarose, hyaluronic acid derivatives, chitosan, and fibrin glue have been used as scaffolds. Man-made polymers such as polyglycolide (PGA), polylactides (PLLA, PDLA), poly(caprolactone) (PCL), and poly(dioxanone) (PDS) have also been studied. More recently, the fabrication of biodegradable elastomers has broadened the biomechanical possibilities of available scaffolds even further. Appropriate selection of a scaffold material in terms of its mechanical properties and degradation characteristics must be matched to each specific tissue engineering application.

Nonwoven textiles manufactured from these materials have proved to be particularly well suited to tissue engineering applications because they can be made highly porous and exhibit extensive interporous connections that allow cells to penetrate and adhere to fibers through the full thickness of the scaffold. Furthermore, the fibers are not stretched taught as in a woven material, and therefore the material exhibits a degree of "give" that, it is hoped, should not restrict subsequent growth of the nascent tissue. From an engineering perspective, the Young's modulus is larger for nonwovens than for equivalent textiles made using a weave, and it is possible, at least in principle, to match the change in modulus of the nonwoven mesh as the polymer degrades over time with the requirement for growth imposed by the pediatric application.

Polyglycolic acid, PGA, was introduced into clinical practice as a biodegradable suture material and marketed under the trade name Dexon (U.S. Surgical, Norwalk, CT). It has since been adopted by many research groups as a scaffold for a variety of potential tissue engineering applications.³¹ The ability to extrude PGA into fibers permits it to be fabricated into nonwoven sheets with an open structure. Pore size has been shown to be important in the choice of scaffold for tissue engineering liver for a variety of cell types including hepatocytes.³² An open pore structure is thought to facilitate both cell delivery and subsequent cell proliferation by allowing free access to suspended cells, free diffusion of nutrients and dissolved gases, and removal of waste products of metabolism. These spatial properties combined with a consistent and rapid loss of polymer mass relative to other biodegradable materials make PGA an attractive choice for tissue engineering.

Biodegradable scaffolds for tissue engineering heart valves have been assembled from flat sheets of nonwoven PGA fibers, impregnated with the thermoplastic polymer poly-4-hydroxybutyrate (P4HB). These flat sheets were then assembled into a trileaflet structure by a series of two or more wraps around a cylindrical mandrel. Seams were heat welded at a temperature above the melting point of the P4HB to provide a tube-like, trileaflet-valved conduit.¹³

Despite promising early results using the PGA/P4HB composite, ongoing studies experienced a number of problems with the use of this particular scaffold material. Specifically, because of the rapid hydrolysis of PGA, there was a tendency for collapse of the cell-polymer construct during culture in vitro. A further, unfavorable, characteristic of PGA is that its tensile strength decays far more rapidly than the decay in its residual mass. At the time of implantation, the strength of the construct was highly dependent on the degree of new tissue growth and spontaneous rupture of the conduit wall occurred not infrequently at the time of implantation. One further but related problem was tearing of the grafts at the suture line during implantation. It appeared that most, if not all, of these problems were attributable to loss in tensile strength of the nonwoven mesh over time, itself a reflection of changes in the tensile properties of the individual fibers. Based on these problems we formally set out a number of criteria required of the scaffold at different time points in the tissue engineering process. It was clear that first and foremost we require a template in the geometry of a normal heart valve to which cells will attach at the time of cell delivery; second, we required a mechanical support for the cell-template construct that would maintain the structural integrity during culture, implantation, and into the short- to mid-term postoperative period; and finally, a skeleton was required onto which sutures could be anchored at the time of implantation and the resultant forces distributed through the whole body of the graft. We perceived that these requirements could be served by a composite nonwoven mesh incorporating the rapidly degrading PGA fibers to provide the geometric template with the much more slowly degrading PLLA fibers to provide the structural support and anchorage for the suture lines.

Poly L-lactic acid, PLLA, the -methyl–substituted form of PGA, was initially developed for the biodegradable suture market but failed to become popular in that capacity. The polymer hydrolyzes at a much slower rate than PGA, which is made from the unsubstituted glycolide monomer. The initial tensile strength of PLLA is, however, less than that of PGA or many of the other biodegradable sutures such as PDS, vicryl, or silk, making it less suitable as an absorbable suture material. Studies in vivo conducted on rats have shown the breaking force of PLLA to diminish by 20% during the first 2 weeks but to remain relatively constant for up to 4 weeks thereafter when used for fascial closure. Suture material in this study appeared intact at 28 weeks but had largely disappeared by 52 weeks.³³ In a separate study, 8-0 PLLA fibers were found to retain 70% to 100% of their strength over a 12-week period of in vitro degradation. The more sustained tensile properties of PLLA have already led to its use in orthopedic applications where fixation with biodegradable miniplates and screws avoids the need for subsequent removal. Comparison of PGA and PLLA miniscrews implanted into the calvarium of sheep showed fragmentation and hydrolysis of PGA at 4 to 6 weeks and total absorption by 12 weeks. In contrast, PLLA miniscrews retained their integrity and holding power at 26 weeks but had mostly resorbed by 2 years.³⁴ PLLA materials are now being investigated for use in a range of biodegradable applications including urethral and cardiovascular stents³⁵ where the rapid loss of tensile properties exhibited by PGA might permit early restenosis or closure. Stress-strain data have been collected for combinations of PGA/PLLA in nonwoven meshes using uniaxial tests of tensile strength and found to be highly suited to a role in the wall of tissue-engineered heart valves on the pulmonary side of the circulation. Further mechanical studies are underway, in conjunction with Dr. Michael Sacks' laboratory for soft tissue biomechanics at the University of Pittsburgh, to optimize the flexural behavior of this nonwoven textile for the valve leaflets.

The next problem has been to develop a means of assembling scaffolds from sheets of nonwoven textiles that recapitulate the surface geometry of the native pulmonary valve and trunk. On the surface of it this would appear to be a perfectly straightforward task. However, the necessity of fabricating such a structure from a nonwoven material required the development of some novel textile technologies. We have developed a means of fitting flat sheets of mesh over molds with multiple axes of curvature. This is achieved by needle punching around the periphery of the mold using custom-cut felting needles. The needle-punching process permits adjoining layers of nonwoven textiles to be united without evidence of layering in the final product.³⁶ While these technologies have been developed for use with the PGA/PLLA mesh described above, the methods remain broadly applicable to a wide range of nonwoven textiles in the fabrication of a variety of anatomical elements. In essence, these techniques permit the fabrication of a structure comprising the multiply curved surfaces of three symmetrical valve leaflets and their respective sinuses of Valsalva. This mode of assembly creates a wholly symmetrical and unitary structure with no sutures or residual evidence of lamination. The role of computational models with respect to heart valves presents a much less complex problem than presented below for the development of soft tissues. Nevertheless, models have proved useful in predicting and optimizing the geometry of valves. The role of computational fluid dynamics, important as it is to the design and characterization of flow through prosthetic mechanical valves, assumes much less importance to the tissue-engineered valve in which one seeks to mimic nature in terms of geometry and viscoelastic properties of the valve leaflets over time. Even with very stiff leaflets it is unlikely that the normal function of flexible tissue-engineered valve leaflets will generate microcavitation in the flow of blood between them.

As to the future, there are a number of emerging technologies that have the potential to provide scaffolds with the desired geometry, porosity, and connectivity between pores for tissue engineering applications. Specifically, advanced manufacturing technologies known as rapid prototyping or solid freeform fabrication are under investigation. In essence, rapid prototyping (RP) is the process of creating a three-dimensional (3-D) object through repetitive deposition of material in 2-D layers from a computer-aided design (CAD) virtual model of the object using a slicing algorithm.³⁷ Although RP has been available for a decade now and data on the geometry of anatomical elements has been widely available from MRI or CT images of normal human subjects for even longer, conventional RP techniques do not have sufficient resolution to fabricate scaffolds with the porosity and extensive interporous connections considered necessary for tissue engineering and which are currently available from textile technologies. One further perspective on the future of scaffolds for tissue engineering in general will be the ability to deliver cytokines locally to developing tissues through advances in the field of drug delivery in conjunction with existing human recombinant technologies. This has already proved possible in the engineering of regenerative elements for orthopedic surgery by the provision of bone morphogenic proteins (BMPs) from poly-D,L-lactic acid-p-dioxanone-polyethylene glycol block coplymer (PLA-DX-PEG), a synthetic biodegradable polymer used to tissue engineer bone.³⁸ As yet these techniques have not been applied widely to the engineering of soft tissues, however. It is clear that the success of tissue-engineered grafts in cardiac surgery will be intimately related to the properties of the underlying scaffold. Despite the promise of emerging technologies, it is likely that a plethora of first-generation products utilizing existing polymers and established technologies will emerge over the next decade.

Evaluation of Tissue-Engineered Heart Valves in Vivo

Tissue-engineered valve leaflets and valves proper have been evaluated in a series of experiments spanning several years and many design and tissue culture iterations.^11,13,39,40 Most recently, we have undertaken a detailed evaluation of the hemodynamic performance of tissue-engineered valves acutely in the sheep model. Valves grown over many weeks in culture (Fig. 65-1) exhibit the histology shown in Figure 65-2 with dense tissue formation on inner and outer surfaces and relatively less tissue in the depth of the structure. Following implantation (Fig. 65-3), valve gradients are typically found to be less than 5 mm Hg with only trivial regurgitation in the resting state. Through the use of echocardiography, the valve leaflets are seen to open symmetrically into their respective sinuses, creating an effective orifice area that is typically in the region of 2.0 to 2.5 cm² for a 21-mm valve at 5 L/min cardiac output. Echocardiography has furthermore enabled a detailed analysis of the shape and morphology of various components of the valve throughout the cardiac cycle. Figure 65-4 demonstrates the curvaceous outline of the valve sinus and the smooth continuity that exists between the sinus and the valve leaflets. Close inspection of the leaflets in the commissural regions in late systole (Fig. 65-5) shows an absence of the reverse bending that has been the cause of stress-related tears in explanted bioprostheses. Stress sharing between the valve leaflet and its respective sinus together with the absence of reverse bending are expected to have a significant impact on the longevity of tissue-engineered valves just as they have for classical biologic prostheses.

View larger version (132K): [in this window] [in a new window]   FIGURE 65-1 The tissue-engineered substitute pulmonary valve viewed from below prior to implantation.

View larger version (179K): [in this window] [in a new window]   FIGURE 65-2 Histology of the tissue-engineered substitute pulmonary valve showing dense surface tissue formation (a) and relatively fewer cells deep in the wall of the construct (b). H.& E. x100.

View larger version (117K): [in this window] [in a new window]   FIGURE 65-3 The tissue-engineered substitute pulmonary valve interposed within the main pulmonary artery as seen through a right fourth space thoracotomy. RV = right ventricle; PA = pulmonary artery.

View larger version (123K): [in this window] [in a new window]   FIGURE 65-4 Long axis echocardiogram of the substitute pulmonary valve in diastole showing proximal (p) and distal (d) suture lines and smooth continuity between the sinus (s) and the valve leaflet. RV = right ventricle; PA = pulmonary artery.

View larger version (133K): [in this window] [in a new window]   FIGURE 65-5 Short axis echocardiogram of substitute pulmonary valve in late systole showing the leaflet free margins (m) and absence of reverse bending at the commissures (c).

	THE CREATION OF HEART MUSCLE SUBSTITUTES

 
In regard to heart muscle, the central problem facing physicians is that adult myocardium lacks the ability to regenerate. Tissue engineering of heart muscle is principally aimed at improving the outlook for patients who have lost the function of substantial portions of their left ventricular wall because of ischemic heart disease. Conceivably a variety of other smaller patient groups may also potentially benefit from these advances. Approaches to the tissue engineering of myocardium have many similarities to those that address the fabrication of substitute heart valves, as outlined above. One principal difference, however, is the notion of using the scar itself as a scaffold. The idea that a scar may be repopulated with functional cells has emerged over the last decade and stimulated many research groups to look for a suitable cell type. A variety of cell types have indeed been injected into myocardium under different experimental conditions. While transformed cell lines appear intuitively very attractive, the risk of uncontrolled cellular proliferation is real. For example, primary myoblasts engineered to contain vascular endothelial growth factor (VEGF) and injected into myocardium in one study resulted in the formation of hemangiomas and the appearance of VEGF in the circulation.⁴¹ Until recently, however, most studies have concentrated on the use of fetal cardiomyocytes or satellite cells from skeletal muscle.

Cultured fetal cardiomyocytes, in particular, have been extensively investigated.^42–48 Fetal cardiomyocytes have been shown to survive after implantation into host myocardium.^49,50 Furthermore, the functional benefits of cardiomyocyte injection have been demonstrated in two well-established models of myocardial infarction created by cryonecrosis⁴³ and coronary artery ligation.⁴⁸ These cells have then been shown to form gap junctions with those of the recipient myocardium.⁴² The hope is that connections such as these should permit the transplanted cells to beat in synchrony with the recipient myocardium. In contrast to fetal cardiomyocytes, satellite cells within skeletal muscle are morphologically undifferentiated cells that lie dormant between the plasma membrane and basal lamina. With appropriate stimuli, these cells will proliferate and differentiate into new skeletal muscle fibers.⁵¹ The hypothesis that satellite cells might be able to undergo "milieu-dependent" differentiation into cardiac myocytes was first put forward by Chiu et al in 1995.⁵² They began by implanting autologous satellite cells isolated from skeletal muscle into canine myocardium in a cryoinjury model and were able to demonstrate the presence of striated muscle fibers at the implant sites. These cells showed morphologic characteristics of cardiac myofibrils such as the centrally located nuclei and the presence of intercalated disks. They later confirmed the feasibility of satellite cell differentiation into muscle fibers after implantation into the myocardium of isogenic rats using the cell marker 4',6-diamidino-2-phenylindone, which binds to DNA and to the protein tubulin to form a fluorescent complex.⁵¹ There are obvious advantages to using autologous donor myoblasts in preference to fetal tissues. Ethical and regulatory issues abound in the area of human fetal cells. Furthermore, cells transplanted into potential adult patients would, by necessity, be allogeneic, raising the possibility of rejection. Scorsin et al compared the use of cellular transplantation of fetal cardiomyocytes with that of neonatal skeletal myoblasts and uninjected controls in rats. Their model used coronary artery ligation to create a zone of myocardial infarction. Functional assessment was made by 2-D echocardiography. In this study, they were able to demonstrate a decrease in LV function following coronary artery ligation, as anticipated. However, a significant improvement was seen in the calculated ejection fraction of injured hearts that had been injected with both fetal cardiomyocyte and skeletal myoblast groups 1 month earlier. Although the sample size was quite small, there was no difference demonstrable between the two experimental groups. This and other studies have opened the door to the future possibility of "cellular cardiomyoplasty" by removing the requirement for fetal tissues.

These studies also led some groups to take a closer look at the myocardium itself. Although the lack of regenerative capacity of adult myocardium would appear to preclude its use as a potential cell source, some researchers have questioned this traditional dogma. Indeed, atrial cardiomyocytes may be cultured in the presence of fetal calf serum and resemble immature cardiomyocytes, more so even than ventricular cardiomyocytes, which may also be cultured in vitro.⁵³ Atrial cardiomyocytes are of course clinically quite accessible by endomyocardial biopsy. The Toronto group has transplanted cardiomyocytes cultured from the left atrial appendage of adult rats and expanded in vitro into a transmural scar created 3 weeks earlier at the time of cell harvest. Five weeks later ventricular function, evaluated after explantation in a Langendorff preparation, showed systolic and developed pressures as well as maximal rates of myocardial contraction and relaxation significantly better in the transplanted group than controls. Histologically, the cultured cells were shown to have survived within the scar tissue, even in the absence of an apparent blood supply. The area occupied by the LV scar was smaller and its thickness greater than that of controls and LV chamber volume was furthermore noted to be smaller in the transplanted group. Although the future prospects of cardiomyocyte transplantation or "cellular cardiomyoplasty" are intriguing, especially for patients with established scar formation who are not expected to benefit from coronary revascularization, the mechanism of action remains elusive. A criticism of early experimental models that used cryoinjury alone was a failure of the model to wholly reflect the occlusive pathology that characterizes ischemic heart disease. Nevertheless, the demonstration of these same functional improvements in hearts in which the model for injury has been created by coronary artery ligation has perplexed cardiac surgeons. One would anticipate that the survival of cells implanted deep within myocardial scar tissue would necessitate rapid and thorough revascularization. However, there appears to be little evidence of neovascularization in these studies. Indeed, the functional benefits may not be related to any ongoing cellular functionality but follow simply from changes to the physical characteristics of the scar itself. These and many other questions need to be addressed before this technique finds its way to the clinic. More generally, the survival of cells that are deeply embedded within cell-polymer constructs raises the question of how to provide for the rapid and thorough revascularization of tissue-engineered elements. One approach to this problem has been through the delivery of angiogenic molecules from the underlying polymer scaffolds. The delivery of specific growth factors, or plasmid DNA encoding the factors, with defined doses and rates at the site where blood vessel growth is desired may now be possible. Such systems will likely have significant utility in the engineering of large tissue volumes and may indeed provide novel, model systems for studying human angiogenesis in vivo.⁵⁴ The more traditional approach to tissue engineering, trying to create of a segment of cardiac muscle from its constituent elements in vitro, would seem to present a much more complex problem than that of creating a functional heart valve. While a heart valve functions largely passively in response to pressure changes in adjacent cardiac chambers and great arteries, cardiac muscle must assume an active role in its function. A noncontractile patch is destined to suffer from paradoxical motion if interposed into the ventricular free wall, a point well illustrated by transmural scars arising from earlier myocardial infarction. In essence, individual cells are required to contract and to do so in a coordinated manner in synchrony with the recipient myocardium. The creation of a contractile element is thus the core of the problem. Furthermore, it would appear that, at least intuitively, the final product will require a blood supply and that cells must be assembled in such a way that the integrity of the structure is maintained in the face of significant intraventricular pressures and shortening of tissue-engineered elements that allows them to contribute meaningfully to contraction of the structure as a whole. This equates to the building of a very complex system and one for which armchair (qualitative) thinking is wholly inadequate. For this reason, it is thought that modeling, through the use of computers, may provide a useful tool. Indeed, the results of modeling very complex systems are frequently counterintuitive.⁵⁵ The amount of biological data currently available that relates to the heart and generated by new technologies is truly overwhelming. There is, already, rich data available in the literature on the three-dimensional geometry of the heart as a whole. Anatomically detailed models of the heart include fiber orientations and sheet structure that permit one to reconstruct the electrical and mechanical behavior of the organ. Furthermore, accurate cellular models have also become highly sophisticated. Models exist for the electromechanical function of all of the main types of cardiac myocyte; in many cases there are multiple models for a given cell type. Connecting these two levels, an accurate reconstruction of the depolarization waveform has been beautifully demonstrated across a virtual framework of the heart. Extensions of this work by incorporation of the ventricular model into a virtual torso have enabled computers to simulate the conducting properties of different tissues and reconstruct the ECG chest and limb leads.⁵⁶ Despite, the complex anatomy and lack of symmetry in most soft tissues, a special computational challenge, material properties of the native heart and its ventricular mechanics have also been modeled during the early filling phase of diastole by coupling finite elasticity theory with finite element analysis.⁵⁷ These examples serve to illustrate the potential power that in silico biology has to offer tissue engineering. By way of example, in one potential iteration, anatomically detailed models of the coronary circulation, developed to study the effects of transmural pressure on blood flow, could be rapid prototyped and serve as the scaffold onto which various myocardial elements could be assembled. This model could be connected to existing models of cellular function or new models of specific cell types, such as skeletal myoblasts that are envisaged as potential cell sources, to predict the electrical behavior of the construct. Further connections with models that describe the material properties of those cells and the underlying scaffold over time might be able to predict the ventricular mechanics of potential tissue substitutes even before they have been formed. In contrast to models of individual cellular functions, which typically run on PCs, this kind of work that uses integrative elements from several models requires massive computing power and program codes, such as CMISS developed at the University of Auckland as part of the Cardiome project, to enable such models to communicate freely with one another. The true prospects for regenerative medicine in the future will likely be closely tied to the development of new supercomputers and to the continued sophistication of mathematical models that describe nature. We must anticipate that there will be many experimental iterations between now and the clinical reality of tissue engineering. Nevertheless, the mere prospect of being able to fashion substitute elements for the cardiovascular system in the laboratory provides a healthy stimulus for ongoing research in this area.