Stem Cells to Viable Autologous Semilunar Heart Valve
http://www.100md.com
《循环学杂志》
the Department of Cardiovascular Surgery (F.W.H.S., T.E.P., Y.M., G.A.G., J.E.M.), Vascular Biology Program
Department of Surgery (Y.Y.), Department of Cardiology (M.C.S.), and Department of Surgery (D.L.M.), Children’s Hospital, Boston, Mass
Department of Pathology, Brigham and Women’s Hospital, Boston, Mass (E.B., F.J.S.)
Harvard Medical School, Boston, Mass (F.W.H.S., T.E.P., Y.Y., M.C.S., E.B., Y.M., G.A.G., D.L.M., F.J.S., J.E.M.)
University of Pittsburgh, Department of Bioengineering, Pittsburgh, Penn (G.C.E., M.S.S.).
Abstract
Background— An estimated 275 000 patients undergo heart valve replacement each year. However, existing solutions for valve replacement are complicated by the morbidity associated with lifelong anticoagulation of mechanical valves and the limited durability of bioprostheses. Recent advances in tissue engineering and our understanding of stem cell biology may provide a lifelong solution to these problems.
Methods and Results— Mesenchymal stem cells were isolated from ovine bone marrow and characterized by their morphology and antigen expression through immunocytochemistry, flow cytometry, and capacity to differentiate into multiple cell lineages. A biodegradable scaffold was developed and characterized by its tensile strength and stiffness as a function of time in cell-conditioned medium. Autologous semilunar heart valves were then created in vitro using mesenchymal stem cells and the biodegradable scaffold and were implanted into the pulmonary position of sheep on cardiopulmonary bypass. The valves were evaluated by echocardiography at implantation and after 4 months in vivo. Valves were explanted at 4 and 8 months and examined by histology and immunohistochemistry. Valves displayed a maximum instantaneous gradient of 17.2±1.33 mm Hg, a mean gradient of 9.7±1.3 mm Hg, an effective orifice area of 1.35±0.17 cm2, and trivial or mild regurgitation at implantation. Gradients changed little over 4 months of follow-up. Histology showed disposition of extracellular matrix and distribution of cell phenotypes in the engineered valves reminiscent of that in native pulmonary valves.
Conclusions— Stem-cell tissue-engineered heart valves can be created from mesenchymal stem cells in combination with a biodegradable scaffold and function satisfactorily in vivo for periods of >4 months. Furthermore, such valves undergo extensive remodeling in vivo to resemble native heart valves.
Key Words: stem cells ; tissue engineering ; valves
Introduction
Approximately 275 000 patients undergo heart valve replacement each year.1 Results of surgery are complicated by the morbidity associated with lifelong anticoagulation of mechanical valves and the limited durability of bioprostheses.2–4 Tissue engineering may provide a solution to these problems by providing living, autologous biological structures that do not require anticoagulation yet retain the capacity to remodel and repair, obviating the need for late valve replacement.
See p 2715
Stem cells are an attractive source of cells for tissue engineering because of their unique biological properties. There are major ongoing ethical concerns about the use of embryonic stem cells in conventional research and potential human therapies.5 However, stem cells have been isolated from a variety of somatic tissues.6 A subset of cells resides within human bone marrow that can be expanded in the laboratory and induced to differentiate into multiple mesenchymal tissue cell types in vitro or in vivo.7,8 These cells, mesenchymal stem cells (MSCs), thus display the important properties of self-renewal and plasticity.
Over the past 3 years, numerous studies have reported on the combination of cellular therapy and tissue engineering for heart valve replacement with various degrees of success. These studies have almost exclusively used an existing heart valve (either allograft or xenograft) as the substrate.9–11 In the present study, we aimed to engineer 3D "tissue" de novo using biodegradable polymer fibers and autologous MSCs and to prove its functionality in vivo. We first showed that MSCs could be isolated from ovine marrow in the same way that MSCs have been isolated from human marrow. A biodegradable scaffold was then developed to act as a temporary vehicle for these cells. The stem-cell tissue-engineered heart valve was thus created and evaluated in a sheep model, which extended to 8 months. Sheep were chosen because they are most widely favored for testing bioprosthetic valves as a result of their generally docile character and tendency to rapid degeneration of biological tissues.
Methods
Isolation and Culture of MSCs
Animals were sedated with ketamine at 25 mg/kg IM and xylazine at 0.05 mg/kg IM, intubated, and maintained on an inhaled 0.5% to 2% of isoflurane/oxygen mixture. A stab incision was made over the mid part of the sternum and approached with an 11-gauge bone marrow biopsy needle (MDTech). Then, 15 mL marrow was aspirated into a syringe containing 5 mL of 1000-IU/mL heparin sodium (Elkins-Sinn Inc). MSCs were obtained from ovine bone marrow using the method of Pittenger et al.12 Medium was changed at 24, 48, and 72 hours, by which time colonies of spindle-shaped cells were observed. Cells were passaged at low density (<104 cells/cm2) into successively larger tissue culture flasks and then into 1750-cm2 polystyrene roller bottles (Corning Inc). The culture medium comprised DMEM supplemented with 10% FBS and 1% antibiotic/antimycotic solution (GIBCO). Air in the roller bottles was displaced with a 5% CO2/air mixture, and bottles were rotated at 0.2 cycles per minute and 37°C. Cells were characterized before valve implantation.
Flow Cytometry
Cells were harvested with 0.05% trypsin/0.53 mmol/L EDTA solution (GIBCO) and resuspended in PBS containing 2 mmol/L EDTA and 0.5% BSA. We labeled 106 cells with mouse anti-sheep CD45 monoclonal antibody (Biosource) or mouse anti-sheep CD14 monoclonal antibody (Research Diagnostics) according to the manufacturer’s instructions. Antigens were detected by FITC-conjugated goat anti-mouse IgG. Cells were also labeled with isotype control antibodies. Samples were analyzed on a FACSVantage SE flow cytometer (BD PharMingen). Using CellQuest software (BD PharMinger, we acquired 10 000 to 50 000 events in list mode. The list mode files were analyzed with WinMDI software (Scripps Research Institute).
Adipogenic and Osteogenic Differentiation of Isolated Cells
Cells were incubated at 90% to 95% confluence in 35-mm tissue culture dishes in medium containing DMEM low glucose (GIBCO), 10% FBS (Hyclone), 100 nmol/L Dexamethasone (Sigma), 10 mmol/L ;-glycerolphosphate (Sigma), 50 μmol/L ascorbic acid (Sigma), and 1% penicillin-streptomycin-glutamine (GIBCO). The medium was replaced every 3 to 4 days. Cells containing lipid vacuoles and deposition of mineral were first observed after 3 weeks. After 4 to 5 weeks, adipocytes were stained with oil-red O (Sigma). Cells were fixed in 10% formalin and incubated twice with propylene glycol (Sigma) for 5 minutes. Oil-red O was added to the cells for 10 minutes at room temperature. After being washed in 85% propylene glycol and rinsed in distilled water, cells were counterstained with Gill’s hematoxylin No. 3 (Polysciences). To assess osteogenic differentiation, the mineral was stained with silver by the method of von Kossa. Cultures were fixed in 10% cold formalin and then stained with 2.5% silver nitrate (Sigma) for 30 minutes under a 60-W lamp. Cells were washed 4 times in distilled water and counterstained with hematoxylin.
Immunofluorescence
Cells were plated into 8-well chamber slides (Nalge Nunc International) for 24 hours before indirect immunofluorescence staining. Cells were washed with PBS and fixed with cold methanol. Slides were incubated for 1 hour at room temperature with mouse anti-human vimentin monoclonal antibody (Dako) at 1:300 dilution or anti–-smooth muscle actin (SMA) monoclonal antibody (Sigma) at 1:1000 dilution. Cells were then washed in PBS and incubated with FITC-conjugated goat anti-mouse IgG as a secondary antibody (Vector) at 1:200 dilution. Cells were examined by fluorescence microscopy with a Nikon microscope and IP Laboratory imaging software.
Fabrication of the Biodegradable Textile and Evaluation of the Heart Valve Scaffold
Processing of the Tissue-Engineered Heart Valves In Vitro
Biodegradable heart valve scaffolds were assembled using 1-mm PGA/PLLA nonwoven mesh and mounted on 50-mm sections of -in PFA tube (Cole Palmer Instrument Co). Each scaffold was inverted inside a 140-mL glass hybridization tube and sterilized by exposure to ethylene oxide. Approximately 1 billion cells were resuspended in 25 mL culture medium and transferred to the hybridization tube. Tubes were sealed and rotated at 4 cycles per minute and 37°C. At 6 hours, the medium was aspirated and centrifuged at 1000 rpm for 5 minutes, and the cells were resuspended in fresh culture medium.
After 12 hours, the cell-seeded scaffold was transferred to a 1750-cm2 roller bottle and incubated in culture medium as described above, along with 20 μg/mL basic fibroblast growth factor and 0.4 mg/L L-ascorbic acid 2-phosphate (Sigma-Aldrich). Medium and CO2 atmosphere were changed at 48-hour intervals, and culture was continued for 4 weeks. A preliminary assessment of tissue-engineered valves could be made by visual inspection of the surface of the graft through the roller bottle and by gently inverting the bottle so that fluid filled the outflow section of the conduit.
Surgical Technique
Autologous heart valves were implanted into 6 juvenile sheep (Dover/Suffolk cross). The sheep were housed and cared for according to US Department of Agriculture and institutional guidelines. Animals were sedated with ketamine 25 mg/kg IM and xylazine 0.05 mg/kg IM, intubated, and maintained on isoflurane. The heart was approached through a left fourth space thoracotomy incision, and cardiopulmonary bypass was established with an Optima XP oxygenator with integral 4 l hard shell reservoir (Cobe Cardiovascular Inc). A short segment of pulmonary artery was excised, native pulmonary valve leaflets were excised, and the autologous tissue-engineered valve was interposed between the 2 ends. Samples of each of the tissue-engineered valves were excised before implantation for recharacterization of cells and histology. The chest was closed, and the animal was allowed to recover.
Evaluation of the Tissue-Engineered Heart Valves
A qualitative assessment of each valve was made in terms of tissue handling, ability to hold sutures, and hemostasis. Echocardiographic evaluation was carried out at the end of the procedure with a Sonos 5500 system (Philips). Two-dimensional imaging and color Doppler imaging on the short and long axes were performed. Regurgitation was graded as trivial, mild, moderate, or severe by the regurgitant jet width obtained with color Doppler. Proximal mean velocity and total maximum instantaneous and mean velocities were measured. The maximum instantaneous and mean gradients were calculated from the modified Bernoulli equation. Effective orifice area was calculated from the continuity equation. Hemodynamic function was reevaluated at 4 months.
Histology
Samples for histological examination were fixed in 4% paraformaldehyde, embedded in paraffin, cut in parallel 5- to 7-μm sections, mounted, and stained with Movat pentachrome to evaluate extracellular matrix. Immunohistochemistry was performed using the avidin-biotin-peroxidase method with monoclonal anti-vimentin, anti–-SMA, and anti–von Willebrand factor primary antibodies (DAKO Corp). Tissue sections were treated with hydrogen peroxide to inhibit endogenous peroxidase activity and then incubated with the primary antibody in PBS with 4% species-appropriate serum. Subsequent processing was carried out according to the manufacturer’s recommendations. The reaction was visualized with 3-amino-9-ethyl carbazole. Sections were counterstained with Meyer’s hematoxylin solution. Adjacent sections were treated with nonimmune IgG as controls for antibody specificity. Slides were viewed under a Nikon microscope. Normal ovine pulmonary valve leaflets were stained contemporaneously for comparison.
Statistical Analyses
All data are expressed as mean±SEM. Continuous variables were compared by use of a 2-tailed paired t test.
Results
Characterization of the MSCs
MSCs were isolated from the bone marrow harvested from juvenile sheep. Similar to human MSCs, these cells did not express the leukocyte common antigen CD45 or CD14, which were demonstrated by flow cytometric analysis (Figure 1a and 1c). To ensure immunoactivity of both anti-CD45 and anti-CD14 antibodies, Figure 1b and 1d shows that 98% of leukocytes isolated from ovine peripheral blood expressed CD45, and 6% were positive for CD14. The percentage of CD14+ cells is consistent with the proportion of monocytes present in peripheral blood. Similar to human MSCs, ovine MSCs expressed ;1 integrin CD29 but failed to express the thymocyte and lymphocyte marker CD4, endothelial antigen CD31, and von Willebrand factor (data not shown).
Multipotency of ovine MSCs was demonstrated by inducing cells to differentiate along adipogenic and osteogenic lineages using lineage-specific inducing media12 (Figure 1e and 1f). MSCs at low density displayed the spindle-shaped morphology characteristic of fibroblast-myofibroblast cell lineage and expressed the VA phenotype markers vimentin and -SMA14 (Figure 1g and 1h).
Cells harvested from the tissue-engineered valves at the time of implantation displayed an identical antigen profile. These cells also retained the capacity to differentiate into adipocytes and osteocytes.
Physical Properties of the Biodegradable Scaffold
A unique biodegradable scaffold was developed next that incorporated both PGA and PLLA fibers (Figure 2). Bending stiffness of fresh ovine pulmonary valve leaflets was 3.6±0.6 mN-mm2. Effective stiffness of the textile samples after 0 and 4 weeks in cell-conditioned medium is shown in Figure 2a. Initial stiffness of the scaffold was 99.3±10.5 mN-mm2. After 4 weeks in culture, the material became much more pliable; stiffness measured just 18.3±3.1 mN-mm2, which approaches that of the native ovine pulmonary valve leaflets (3.6±0.6 mN-mm2).
Tensile strength of samples of the scaffold as they degrade is displayed as a function of time in culture medium (Figure 2b). Tensile strength dropped very rapidly over the first 4 weeks but diminished much more slowly thereafter. At 24 weeks, tensile strength was 30% of the initial strength of the scaffold. Nevertheless, estimated burst strength was still 270 mm Hg for a 21-mm-diameter conduit, far exceeding the target strength for a pulmonary valve substitute.
Appearance and Geometry of the Tissue-Engineered Valves
The macroscopic appearance of the stem-cell tissue-engineered heart valves after removal from culture and before implantation is shown in Figure 3a. There was remarkable consistency in macroscopic appearance between each of the 6 grafts before implantation. The view from the ventricular side of one of the grafts shows the 3 valve leaflets in apposition, as they would appear during diastole (Figure 3b). In profile, bulges in the wall of the graft, marked with arrows, correspond to the recreated sinuses of Valsalva.
The tissue-engineered heart valves handled well surgically. There was no evidence of tearing or damage to the grafts during surgical manipulation. The tissue-engineered valves also held sutures well. The grafts were hemostatic, and bleeding through needle holes was easily controlled by reversal of heparin at the end of the operation.
Noninvasive echocardiographic imaging was used to assess the geometry of grafts after implantation. In profile, the valves displayed convex bulges on the surface of the graft analogous to the normal sinuses of Valsalva and smooth continuity between the valve leaflets and their respective sinuses. This geometry is also known to create vortexes in the flow of blood that initiate the prediastolic movement of leaflet closure. The short-axis view in mid systole shows the characteristically triangular orifice of the open valve and absence of "reverse bending" at the commissures (Figure 3c). The long-axis view shows the valve interposed between the right ventricle and the distal main pulmonary artery (Figure 3d). Measurement of root dimensions in vivo, obtained from the echocardiograms, showed a mean annulus diameter of 19.7±0.77 mm after implantation and a mean maximum sinus diameter of 23.8±0.87 mm.
Two animals died early, 1 at 24 hours and 1 after 7 days. The clinical history, antemortem examination, laboratory findings, and autopsy were consistent with respiratory failure and gastrointestinal hemorrhage, respectively, as the causes of death. The tissue-engineered valves from these animals were also explanted and examined grossly and histologically for signs of infection or thrombosis. There was no evidence of either of these pathologies.
After surviving the immediate postoperative period, the 4 animals remained in good general health; they continued to eat and drink normally. Figure 5 compares hemodynamic data obtained at 4 months in the anesthetized animals with measurements obtained under similar hemodynamic loading conditions at the time of implantation. There was an increase in maximum instantaneous gradient of 11±7.4 mm Hg (P=0.23), mean gradient increased by 5.2±4.9 mm Hg (P=0.36), and effective orifice area decreased by 0.05±0.19 cm2 (P=0.83). These differences failed to reach statistical significance.
Histology of the Tissue-Engineered Heart Valves
Similar to native valve leaflets (Figure 6a), extracellular matrix of the tissue-engineered valve leaflets was organized into layers (Figure 6b). An abundance of collagen fibers, staining yellow, was apparent on the outflow surface of the graft, analogous to the zona fibrosa of the native valve (Figure 6c). In contrast, the inflow surface of the valve was characterized by the presence of elastin fibers, staining black and characteristic of the native zona ventricularis (Figure 6d). Glycosaminoglycans, staining green, were disposed throughout much of the remainder of the valve leaflet, as seen in the native valve. The tissue-engineered leaflets, however, were much thicker than their native counterparts, and corrugations were absent from the outflow surface of the valve leaflets.
Phenotypic Expression in the Tissue-Engineered Heart Valve
Changes in structure were mirrored by changes in the distribution of cell phenotypes across the tissue-engineered valve leaflets. Tissue sections were stained separately for vimentin and -SMA. Comparisons were made between samples of valves harvested at the time of implantation and samples from valves explanted after 8 months in vivo. Sections of native ovine pulmonary valve leaflets were included as controls (Figure 7).
As in native valves, cells expressing vimentin were found to be evenly distributed throughout the full thickness of samples taken from tissue-engineered valves at the time of implantation. This pattern of distribution remained unchanged at 8 months (Figure 7a through 7c).
Samples of tissue-engineered valves taken at the time of implantation showed diffuse staining of -SMA throughout the sections (Figure 7d). In contrast, by 8 months, the -SMA–positive cells were confined to the subendothelial layer, in striking similarity to the pattern of staining observed in native valve leaflets (Figure 7e and 7f).
The endothelial cell marker von Willebrand factor was not present in the valves at the time of implantation (Figure 8a). However, explanted tissue-engineered grafts demonstrated a continuous and uninterrupted layer of von Willebrand factor–positive cells (Figure 8b).
Discussion
In several respects, the bone marrow is an ideal source of cells for tissue engineering. Marrow is easily accessible. The primary isolate is a suspension of cells that is much easier to process than solid tissue samples previously used in this field and presents much less opportunity for microbial contamination during harvest and processing. Finally, the unique properties of stem cells—multipotency and self-renewal—present clear advantages of tissue engineering over fully committed cells.
Although MSCs have been isolated from several different species,12,18 there is little or no published data in respect to MSCs in sheep marrow. Thus, our first step was to establish that MSCs could indeed be isolated from bone marrow in the ovine model. Flow cytometric analysis indicated that these cells did not express the leukocyte common antigen CD45 or CD14 expressed by monocytes, which distinguishes them from the much more numerous hematopoietic cell fraction of the marrow. Multipotency of these cells was also confirmed by their ability to differentiate into adipogenic and osteogenic lineages. Moreover, when cultured at low density, MSCs display the spindle-shaped morphology characteristic of fibroblast-myofibroblast cell lineage. They also express the VA phenotype markers vimentin and -SMA displayed by resting interstitial cell population of valves in culture.14 These findings lead us to hypothesize that MSCs isolated from ovine bone marrow could be used to provide cells of fibroblast-myofibroblast lineage for the interstitial cell fraction of a substitute heart valve.
We further questioned whether this cell phenotype would remain stable throughout the lengthy tissue engineering process, an important issue that has been raised previously in respect to the processing of cells destined for human therapies.19 Therefore, samples of "tissue" were excised from tissue-engineered heart valves immediately before surgical implantation for enzymatic digestion, plating, and recharacterization. These studies confirmed that MSCs retained the VA phenotype and the capacity to differentiate into multiple mesenchymal cell types. In summary, the results strongly supported the use of marrow-derived MSCs to create autologous heart valves through tissue engineering.
A unique biodegradable scaffold was developed in this study to act as a temporary vehicle for the stem cells. Numerous studies have demonstrated that the open pore structure and extensive interporous connections displayed by the nonwoven fiber architecture provide an effective substrate for cell attachment and tissue formation in vitro.20,21 However, previous studies have failed to recreate the highly complex surface geometry of the native heart valve using this type of material or to provide for mid-term durability of these grafts while living elements of the engineered tissue mature and develop strength in vivo.21,22 The combination textile used here incorporated both PGA fibers to provide a physical template for cells to adhere to and PLLA fibers to provide sustained mechanical strength to the nascent tissue. PGA and PLLA currently are used in medical implants; extensive studies have found them to be nontoxic. Implants fabricated from PLLA degrade much more slowly than those synthesized from the unsubstituted glycolic acid monomer PGA.23 Because these polymers both degrade by hydrolysis, the textile underwent preliminary evaluation by immersion in cell-conditioned culture medium and interval testing of strength and flexibility. The results showed that the scaffold material was strong enough and pliable enough to function in its proposed physiological role in the pulmonary circulation after 4 weeks, the proposed time for implantation, and for many months thereafter. In fact, the results suggested that the material was still a little stiffer than native valve leaflets. Nevertheless, these values are similar to values observed in the glutaraldehyde-fixed pericardial bioprostheses used clinically.24
As to the potential biological response to scaffold resorption, the degradation products of PGA and PLLA are glycolic acid and L-lactic acid, the naturally occurring stereoisomer of lactic acid, respectively. These products are metabolized normally to carbon dioxide and water. Buildup of local concentrations could potentially increase pH around the scaffold by overtaxing the transportation capacity of surrounding tissues. Studies in vitro have shown that PGA incubation solutions are toxic after 10 days of incubation.25 However, studies such as this in which solutions are not changed during the incubation intervals provide a worst-case model of the effects of accumulation of degradation products and do not reflect the washout expected in vivo generally or in the unique environment where the tissue-engineered valve leaflets find themselves, in the path of the entire circulating cardiac output. Furthermore, the accumulation of toxic metabolites is more of a concern in orthopedic applications in which the mass of bulk polymer is significant and a mild, delayed foreign body reaction to implanted polyesters is described.26 Interestingly, a typical foreign body reaction was not observed in any of the explanted tissue-engineered valves even at the implant–native tissue interface where one might arguably expect it to be most evident.
Although the scaffold was evaluated in vitro in this study for just 6 months, studies have shown that PLLA can last up to 2 years in vivo23; therefore, the long-term mechanical properties of the tissue-engineered heart valve will ultimately need to be tested over a much more extended time. To mimic the complex structure of native valves, needle punching was used to fit the flat textile over molds. Computer simulations and finite-element models have shown that the geometry of native heart valves and their prosthetic counterparts is important for long-term durability of valve leaflets in vivo.27,28 It is to be anticipated that recreating this structure will prolong the longevity of the valves in vivo.
It took 2 to 3 weeks to expand a sufficient number of cells and 4 weeks of culture on the heart valve scaffolds to produce the tissue-engineered valve. It is possible that efficiencies could reduce this time. Nevertheless, the macroscopic appearance and surface geometry of the valves at the conclusion of the culture phase were remarkably consistent. Echocardiography showed that this geometry was maintained in vivo. The absence of reverse bending at the commissures, known to have a negative impact on the long-term durability of artificial heart valves,29 was also noted on echocardiography. Systolic gradients observed across the engineered valves at the time of implantation and after 4 months in all the surviving animals compared favorably with published values for bileaflet mechanical and stented biological valves of similar dimensions.15–17 Little regurgitation was noted at implantation. However, it was not possible to obtain adequate views of the valve to assess regurgitation at the 4-month time point owing to scarring and interposition of the lung. Unfortunately, transesophageal echocardiography was not available.
The native heart valve is a layered structure.3 The outer layer or zona fibrosa is composed principally of collagen; the middle region, zona spongiosa, contains an abundance of glycosaminoglycans that are hydrated in the living state to form a gel; and the inner layer, the zona ventricularis, is characterized by the presence of radially arranged elastin fibers.
At low magnification, there was striking similarity between the tissue-engineered heart valve and the native valve leaflets in terms of the distribution of the extracellular matrix elements collagen, glycosaminoglycans, and elastin. Furthermore, an abundance of collagen fibers was observed on the outflow surface of the tissue-engineered heart valve, reminiscent of the zona fibrosa. In contrast, the inflow surface of the valve contained relatively little collagen but was remarkable for the presence of a thin layer of elastin fibers characteristic of the zona ventricularis. Glycosaminoglycans were distributed through the remainder of the valve structure. Clear differences remained, however. The tissue-engineered leaflets were much thicker than their native counterparts and surface corrugations were absent from the tissue-engineered valve. In the absence of further remodeling, these differences in structure would likely have a negative impact on the longevity of the valve leaflets. However, these problems may be improved by using a much thinner scaffold material in the valve leaflets and possibly by applying physical stimuli to the valve during culture.30
Changes in structure of the extracellular matrix were mirrored by changes in the distribution of cell phenotypes across the valve leaflets. Myofibroblasts (VA phenotype) were confined to a layer immediately below the surface endothelium. In contrast, fibroblastlike cells expressing vimentin but not -SMA (V phenotype) were distributed throughout the remainder of the graft, reminiscent of the native valve.31 This distribution of cell phenotypes is similar across native human aortic and pulmonary valves. It is not clear to what extent the change in phenotypes observed on the tissue-engineered valve at explantation represents a true change in phenotype of the implanted cells in vivo or in growth of cells from outside the graft. Certainly, in the absence of specific cell labeling studies, it is not possible to ascertain the precise origin of the cell phenotypes observed in the explanted tissue-engineered valves. One possible explanation is that stem cells undergo milieu-specific differentiation into tissue-specific cells, ie, cells resident in the native pulmonary valve. In support of this hypothesis, it is known that fibroblasts isolated from cornea can undergo transformation in the opposite direction, ie, from fibroblasts to myofibroblasts, when cultured in vitro.32 Furthermore, the differentiation of stem cells within the engineered heart valve leaflets would be in concordance with other studies in which stem cells have been tracked to a variety of different organs where they undergo tissue-specific differentiation into resident cell phenotypes.18 Although the results of some of these studies have been called into question by the demonstration of cell fusion in vitro,33–35 this event is rare and is unlikely to be responsible for the distribution of cell phenotypes observed in the explanted tissue-engineered heart valves.
Interestingly, the explanted grafts also demonstrated a continuous and uninterrupted endothelial layer on the blood-contacting surface of the grafts. The importance of the endothelium rests partly in its antithrombogenic properties but more importantly in its potential to modulate the underlying interstitial cells. Vascular endothelial dysfunction is thought to play a role in the development of atherosclerosis in blood vessels.36 Moreover, similarities between the morphology of early lesions of nonrheumatic aortic stenosis and atherosclerotic plaques have raised the possibility of a common underlying origin.37 However, existing heart valve prostheses do not become endothelialized or lose any residual endothelium present on the graft at the time of implantation after a short period in vivo.38 The appearance of a vascular endothelium on the tissue-engineered grafts is therefore a significant finding that is expected to have a positive impact on their long-term function. Once again, the precise origin of these cells is less clear. The endothelial lining could arise from deposition of circulating endothelial cells.39 Alternatively, transdifferentiation of the implanted MSCs may be responsible by a process analogous to the transdifferentiation of valvular endothelial cells into mesenchymal cells observed in vitro.40,41
An expanding number of alternative cell sources that hold potential for tissue engineering of heart valves have been examined. The list encompasses cells from all types of tissues and stages of development. During fetal life, amniotic fluid is available for harvest,42 and at birth, both the placenta43 and cord blood44 can provide multipotent cells. After birth, fully committed cells can be obtained from related tissues21,22 or even from the very tissue of interest.45 However, it is generally acknowledged that stem cells present significant advantages for tissue engineering over differentiated cells5 and that MSCs obtained from bone marrow remain the most promising source of cells for recreating heart valves.
We have shown that it is possible to create autologous living heart valves from bone marrow–derived MSCs in combination with a synthetic biodegradable scaffold and demonstrated that such tissue-engineered structures function within the circulation for prolonged periods. These valves undergo extensive remodeling in vivo and adopt cellular phenotypes and structural organization strongly reminiscent of native heart valves. Long-term durability of the valves depends on a variety of factors, including the geometry of the valves, degradation profile of the polymer components over time, and cellular responses to these changes and to the local physiological environment. With further advances, this technology may provide a superior alternative for patients in need of heart valve replacement.
Acknowledgments
Dr Sutherland was supported by an International Research Fellowship sponsored by the British Cardiac Society and Merck, Sharpe and Dohme. This work was supported by the National Institutes of Health.
Footnotes
The online-only Data Supplement can be found with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.104.498378/DC1.
References
Schoen FJ. Pathology of heart valve substitution with mechanical and tissue prostheses. In: Silver MD, Gotlieb AI, Schoen FJ, eds. Cardiovascular Pathology. New York, NY: WB Saunders; 2001: 629–677.
Dalrymple-Hay MJ, Pearce RK, Dawkins S, Alexiou C, Haw MP, Livesey SA, Monro JL. Mid-term results with 1,503 CarboMedics mechanical valve implants. J Heart Valve Dis. 2000; 9: 389–395.
Schoen FJ, Levy RJ. Founder’s Award, 25th Annual Meeting of the Society for Biomaterials, perspectives: Providence, RI, April 28–May 2, 1999: tissue heart valves: current challenges and future research perspectives. J Biomed Mater Res. 1999; 47: 439–465.
Perron J, Moran AM, Gauvreau K, del Nido PJ, Mayer JE Jr, Jonas RA. Valved homograft conduit repair of the right heart in early infancy. Ann Thorac Surg. 1999; 68: 542–548.
McLaren A. Ethical and social considerations of stem cell research. Nature. 2001; 414: 129–131.
Korbling M, Estrov Z. Adult stem cells for tissue repair: a new therapeutic concept; N Engl J Med. 2003; 349: 570–582.
Phinney DG. Building a consensus regarding the nature and origin of mesenchymal stem cells. J Cell Biochem Suppl. 2002; 38: 7–12.
Devine SM, Cobbs C, Jennings M, Bartholomew A, Hoffman R. Mesenchymal stem cells distribute to a wide range of tissues following systemic infusion into nonhuman primates. Blood. 2003; 101: 2999–3001.
Booth C, Korossis SA, Wilcox HE, Watterson KG, Kearney JN, Fisher J, Ingham E. Tissue engineering of cardiac valve prostheses, I: development and histological characterization of an acellular porcine scaffold. J Heart Valve Dis. 2002; 11: 457–462.
Teebken O, Mertsching H, Haverich A. Modification of heart valve allografts and xenografts by means of tissue engineering. Transplant Proc. 2002; 34: 2333.
Schenke-Layland K, Opitz F, Gross M, Doring C, Halbhuber KJ, Schirrmeister F, Wahlers T, Stock UA. Complete dynamic repopulation of decellularized heart valves by application of defined physical signals: an in vitro study. Cardiovasc Res. 2003; 60: 497–509.
Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, Mosca JD, Moorman MA, Simonetti DW, Craig S, Marshak DR. Multilineage potential of adult human mesenchymal stem cells. Science. 1999; 284: 143–147.
Gloeckner DC, Billiar KL, Sacks MS. Effects of mechanical fatigue on the bending properties of the porcine bioprosthetic heart valve. Asaio J. 1999; 45: 59–63.
Messier RH, Jr., Bass BL, Aly HM, Jones JL, Domkowski PW, Wallace RB, Hopkins RA. Dual structural and functional phenotypes of the porcine aortic valve interstitial population: characteristics of the leaflet myofibroblast. J Surg Res. 1994; 57: 1–21.
Chafizadeh ER, Zoghbi WA. Doppler echocardiographic assessment of the St. Jude Medical prosthetic valve in the aortic position using the continuity equation. Circulation. 1991; 83: 213–223.
Milano AD, De Carlo M, Mecozzi G, D’Alfonso A, Scioti G, Nardi C, Bortolotti U. Clinical outcome in patients with 19-mm and 21-mm St. Jude aortic prostheses: comparison at long-term follow-up. Ann Thorac Surg. 2002; 73: 37–43.
Cosgrove DM, Lytle BW, Williams GW. Hemodynamic performance of the Carpentier-Edwards pericardial valve in the aortic position in vivo. Circulation. 1985; 72 (suppl II): II-146–II-152.
Jiang Y, Jahagirdar BN, Reinhardt RL, Schwartz RE, Keene CD, Ortiz-Gonzalez XR, Reyes M, Lenvik T, Lund T, Blackstad M, Du J, Aldrich S, Lisberg A, Low WC, Largaespada DA, Verfaillie CM. Pluripotency of mesenchymal stem cells derived from adult marrow. Nature. 2002; 418: 41–49.
Sutherland FW, Mayer JE Jr. Ethical and regulatory issues concerning engineered tissues for congenital heart repair. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2003; 6: 152–163.
Langer R, Vacanti JP. Tissue engineering. Science. 1993; 260: 920–926.
Shinoka T, Breuer CK, Tanel RE, Zund G, Miura T, Ma PX, Langer R, Vacanti JP, Mayer JE Jr. Tissue engineering heart valves: valve leaflet replacement study in a lamb model. Ann Thorac Surg. 1995; 60: S513–S516.
Hoerstrup SP, Sodian R, Daebritz S, Wang J, Bacha EA, Martin DP, Moran AM, Guleserian KJ, Sperling JS, Kaushal S, Vacanti JP, Schoen FJ, Mayer JE Jr. Functional living trileaflet heart valves grown in vitro. Circulation. 2000; 102 (suppl III): III-44–III-9.
Peltoniemi HH, Hallikainen D, Toivonen T, Helevirta P, Waris T. SR-PLLA and SR-PGA miniscrews: biodegradation and tissue reactions in the calvarium and dura mater. J Craniomaxillofac Surg. 1999; 27: 42–50.
Vesely I, Boughner D. Analysis of the bending behaviour of porcine xenograft leaflets and of neutral aortic valve material: bending stiffness, neutral axis and shear measurements. J Biomech. 1989; 22: 655–671.
Taylor MS, Daniels AU, Andriano KP, Heller J. Six bioabsorbable polymers: in vitro acute toxicity of accumulated degradation products. J Appl Biomater. 1994; 5: 151–157.
Bostman OM, Paivarinta U, Partio E et al. The tissue-implant interface during degradation of absorbable polyglycolide fracture fixation screws in the rabbit femur. Clin Orthop. 1992; Dec: 263–272.
Beck A, Thubrikar MJ, Robicsek F. Stress analysis of the aortic valve with and without the sinuses of Valsalva. J Heart Valve Dis. 2001; 10: 1–11.
Grande-Allen KJ, Cochran RP, Reinhall PG, Kunzelman KS. Finite-element analysis of aortic valve-sparing: influence of graft shape and stiffness. IEEE Trans Biomed Eng. 2001; 48: 647–659.
Thubrikar MJ, Skinner JR, Eppink RT, Nolan SP. Stress analysis of porcine bioprosthetic heart valves in vivo. J Biomed Mater Res. 1982; 16: 811–826.
Engelmayr GC Jr, Hildebrand DK, Sutherland FW, Mayer JE Jr, Sacks MS. A novel bioreactor for the dynamic flexural stimulation of tissue engineered heart valve biomaterials. Biomaterials. 2003; 24: 2523–2532.
Della Rocca F, Sartore S, Guidolin D, Bertiplaglia B, Gerosa G, Casarotto D, Pauletto P. Cell composition of the human pulmonary valve: a comparative study with the aortic valve: the VESALIO Project: Vitalitate Exornatum Succedaneum Aorticum labore Ingegnoso Obtinebitur. Ann Thorac Surg. 2000; 70: 1594–1600.
Masur SK, Dewal HS, Dinh TT, Erenburg I, Petridou S. Myofibroblasts differentiate from fibroblasts when plated at low density. Proc Natl Acad Sci U S A. 1996; 93: 4219–4223.
Vogel G. Stem cell research: studies cast doubt on plasticity of adult cells. Science. 2002; 295: 1989–1991.
Terada N, Hamazaki T, Oka M, Hoki M, Mastalerz DM, Nakano Y, Meyer EM, Morel L, Petersen BE, Scott EW. Bone marrow cells adopt the phenotype of other cells by spontaneous cell fusion. Nature. 2002; 416: 542–545.
Ying QL, Nichols J, Evans EP, Smith AG. Changing potency by spontaneous fusion. Nature. 2002; 416: 545–548.
Wever RM, Luscher TF, Cosentino F, Rabelink TJ. Atherosclerosis and the two faces of endothelial nitric oxide synthase. Circulation. 1998; 97: 108–112.
O’Brien KD, Reichenbach DD, Marcovina SM, Kuusisto J, Alpers CE, Otto CM. Apolipoproteins B, (a), and E accumulate in the morphologically early lesion of "degenerative" valvular aortic stenosis. Arterioscler Thromb Vasc Biol. 1996; 16: 523–532.
Mitchell RN, Jonas RA, Schoen FJ. Pathology of explanted cryopreserved allograft heart valves: comparison with aortic valves from orthotopic heart transplants. J Thorac Cardiovasc Surg. 1998; 115: 118–127.
Asahara T, Murohara T, Sullivan A, Silver M, van der Zee R, Li T, Witzenbichler B, Schatteman G, Isner JM. Isolation of putative progenitor endothelial cells for angiogenesis. Science. 1997; 275: 964–967.
Rafii S, Lyden D. Therapeutic stem and progenitor cell transplantation for organ vascularization and regeneration. Nat Med. 2003; 9: 702–712.
Paranya G, Vineberg S, Dvorin E, Kaushal S, Roth SJ, Rabkin E, Schoen FJ, Bischoff J. Aortic valve endothelial cells undergo transforming growth factor-beta–mediated and non–transforming growth factor-beta–mediated transdifferentiation in vitro. Am J Pathol. 2001; 159: 1335–1343.
Kaviani A, Guleserian K, Perry TE, Jennings RW, Ziegler MM, Fauza DO. Fetal tissue engineering from amniotic fluid. J Am Coll Surg. 2003; 196: 592–597.
Kaviani A, Perry TE, Barnes CM, Oh JT, Ziegler MM, Fishman SJ, Fauza DO. The placenta as a cell source in fetal tissue engineering. J Pediatr Surg. 2002; 37: 995–999.
Wu X, Rabkin-Aikawa E, Guleserian KJ, Perry TE, Masuda Y, Sutherland FW, Schoen FJ, Mayer JE Jr, Bischoff J. Tissue-engineered microvessels on three-dimensional biodegradable scaffolds using human endothelial progenitor cells. Am J Physiol Heart Circ Physiol. 2004; 287: H480–H487.
Maish MS, Hoffman-Kim D, Krueger PM, Souza JM, Harper JJ3rd, Hopkins RA. Tricuspid valve biopsy: a potential source of cardiac myofibroblast cells for tissue-engineered cardiac valves. J Heart Valve Dis. 2003; 12: 264–269.(Fraser W.H. Sutherland, F)
Department of Surgery (Y.Y.), Department of Cardiology (M.C.S.), and Department of Surgery (D.L.M.), Children’s Hospital, Boston, Mass
Department of Pathology, Brigham and Women’s Hospital, Boston, Mass (E.B., F.J.S.)
Harvard Medical School, Boston, Mass (F.W.H.S., T.E.P., Y.Y., M.C.S., E.B., Y.M., G.A.G., D.L.M., F.J.S., J.E.M.)
University of Pittsburgh, Department of Bioengineering, Pittsburgh, Penn (G.C.E., M.S.S.).
Abstract
Background— An estimated 275 000 patients undergo heart valve replacement each year. However, existing solutions for valve replacement are complicated by the morbidity associated with lifelong anticoagulation of mechanical valves and the limited durability of bioprostheses. Recent advances in tissue engineering and our understanding of stem cell biology may provide a lifelong solution to these problems.
Methods and Results— Mesenchymal stem cells were isolated from ovine bone marrow and characterized by their morphology and antigen expression through immunocytochemistry, flow cytometry, and capacity to differentiate into multiple cell lineages. A biodegradable scaffold was developed and characterized by its tensile strength and stiffness as a function of time in cell-conditioned medium. Autologous semilunar heart valves were then created in vitro using mesenchymal stem cells and the biodegradable scaffold and were implanted into the pulmonary position of sheep on cardiopulmonary bypass. The valves were evaluated by echocardiography at implantation and after 4 months in vivo. Valves were explanted at 4 and 8 months and examined by histology and immunohistochemistry. Valves displayed a maximum instantaneous gradient of 17.2±1.33 mm Hg, a mean gradient of 9.7±1.3 mm Hg, an effective orifice area of 1.35±0.17 cm2, and trivial or mild regurgitation at implantation. Gradients changed little over 4 months of follow-up. Histology showed disposition of extracellular matrix and distribution of cell phenotypes in the engineered valves reminiscent of that in native pulmonary valves.
Conclusions— Stem-cell tissue-engineered heart valves can be created from mesenchymal stem cells in combination with a biodegradable scaffold and function satisfactorily in vivo for periods of >4 months. Furthermore, such valves undergo extensive remodeling in vivo to resemble native heart valves.
Key Words: stem cells ; tissue engineering ; valves
Introduction
Approximately 275 000 patients undergo heart valve replacement each year.1 Results of surgery are complicated by the morbidity associated with lifelong anticoagulation of mechanical valves and the limited durability of bioprostheses.2–4 Tissue engineering may provide a solution to these problems by providing living, autologous biological structures that do not require anticoagulation yet retain the capacity to remodel and repair, obviating the need for late valve replacement.
See p 2715
Stem cells are an attractive source of cells for tissue engineering because of their unique biological properties. There are major ongoing ethical concerns about the use of embryonic stem cells in conventional research and potential human therapies.5 However, stem cells have been isolated from a variety of somatic tissues.6 A subset of cells resides within human bone marrow that can be expanded in the laboratory and induced to differentiate into multiple mesenchymal tissue cell types in vitro or in vivo.7,8 These cells, mesenchymal stem cells (MSCs), thus display the important properties of self-renewal and plasticity.
Over the past 3 years, numerous studies have reported on the combination of cellular therapy and tissue engineering for heart valve replacement with various degrees of success. These studies have almost exclusively used an existing heart valve (either allograft or xenograft) as the substrate.9–11 In the present study, we aimed to engineer 3D "tissue" de novo using biodegradable polymer fibers and autologous MSCs and to prove its functionality in vivo. We first showed that MSCs could be isolated from ovine marrow in the same way that MSCs have been isolated from human marrow. A biodegradable scaffold was then developed to act as a temporary vehicle for these cells. The stem-cell tissue-engineered heart valve was thus created and evaluated in a sheep model, which extended to 8 months. Sheep were chosen because they are most widely favored for testing bioprosthetic valves as a result of their generally docile character and tendency to rapid degeneration of biological tissues.
Methods
Isolation and Culture of MSCs
Animals were sedated with ketamine at 25 mg/kg IM and xylazine at 0.05 mg/kg IM, intubated, and maintained on an inhaled 0.5% to 2% of isoflurane/oxygen mixture. A stab incision was made over the mid part of the sternum and approached with an 11-gauge bone marrow biopsy needle (MDTech). Then, 15 mL marrow was aspirated into a syringe containing 5 mL of 1000-IU/mL heparin sodium (Elkins-Sinn Inc). MSCs were obtained from ovine bone marrow using the method of Pittenger et al.12 Medium was changed at 24, 48, and 72 hours, by which time colonies of spindle-shaped cells were observed. Cells were passaged at low density (<104 cells/cm2) into successively larger tissue culture flasks and then into 1750-cm2 polystyrene roller bottles (Corning Inc). The culture medium comprised DMEM supplemented with 10% FBS and 1% antibiotic/antimycotic solution (GIBCO). Air in the roller bottles was displaced with a 5% CO2/air mixture, and bottles were rotated at 0.2 cycles per minute and 37°C. Cells were characterized before valve implantation.
Flow Cytometry
Cells were harvested with 0.05% trypsin/0.53 mmol/L EDTA solution (GIBCO) and resuspended in PBS containing 2 mmol/L EDTA and 0.5% BSA. We labeled 106 cells with mouse anti-sheep CD45 monoclonal antibody (Biosource) or mouse anti-sheep CD14 monoclonal antibody (Research Diagnostics) according to the manufacturer’s instructions. Antigens were detected by FITC-conjugated goat anti-mouse IgG. Cells were also labeled with isotype control antibodies. Samples were analyzed on a FACSVantage SE flow cytometer (BD PharMingen). Using CellQuest software (BD PharMinger, we acquired 10 000 to 50 000 events in list mode. The list mode files were analyzed with WinMDI software (Scripps Research Institute).
Adipogenic and Osteogenic Differentiation of Isolated Cells
Cells were incubated at 90% to 95% confluence in 35-mm tissue culture dishes in medium containing DMEM low glucose (GIBCO), 10% FBS (Hyclone), 100 nmol/L Dexamethasone (Sigma), 10 mmol/L ;-glycerolphosphate (Sigma), 50 μmol/L ascorbic acid (Sigma), and 1% penicillin-streptomycin-glutamine (GIBCO). The medium was replaced every 3 to 4 days. Cells containing lipid vacuoles and deposition of mineral were first observed after 3 weeks. After 4 to 5 weeks, adipocytes were stained with oil-red O (Sigma). Cells were fixed in 10% formalin and incubated twice with propylene glycol (Sigma) for 5 minutes. Oil-red O was added to the cells for 10 minutes at room temperature. After being washed in 85% propylene glycol and rinsed in distilled water, cells were counterstained with Gill’s hematoxylin No. 3 (Polysciences). To assess osteogenic differentiation, the mineral was stained with silver by the method of von Kossa. Cultures were fixed in 10% cold formalin and then stained with 2.5% silver nitrate (Sigma) for 30 minutes under a 60-W lamp. Cells were washed 4 times in distilled water and counterstained with hematoxylin.
Immunofluorescence
Cells were plated into 8-well chamber slides (Nalge Nunc International) for 24 hours before indirect immunofluorescence staining. Cells were washed with PBS and fixed with cold methanol. Slides were incubated for 1 hour at room temperature with mouse anti-human vimentin monoclonal antibody (Dako) at 1:300 dilution or anti–-smooth muscle actin (SMA) monoclonal antibody (Sigma) at 1:1000 dilution. Cells were then washed in PBS and incubated with FITC-conjugated goat anti-mouse IgG as a secondary antibody (Vector) at 1:200 dilution. Cells were examined by fluorescence microscopy with a Nikon microscope and IP Laboratory imaging software.
Fabrication of the Biodegradable Textile and Evaluation of the Heart Valve Scaffold
Processing of the Tissue-Engineered Heart Valves In Vitro
Biodegradable heart valve scaffolds were assembled using 1-mm PGA/PLLA nonwoven mesh and mounted on 50-mm sections of -in PFA tube (Cole Palmer Instrument Co). Each scaffold was inverted inside a 140-mL glass hybridization tube and sterilized by exposure to ethylene oxide. Approximately 1 billion cells were resuspended in 25 mL culture medium and transferred to the hybridization tube. Tubes were sealed and rotated at 4 cycles per minute and 37°C. At 6 hours, the medium was aspirated and centrifuged at 1000 rpm for 5 minutes, and the cells were resuspended in fresh culture medium.
After 12 hours, the cell-seeded scaffold was transferred to a 1750-cm2 roller bottle and incubated in culture medium as described above, along with 20 μg/mL basic fibroblast growth factor and 0.4 mg/L L-ascorbic acid 2-phosphate (Sigma-Aldrich). Medium and CO2 atmosphere were changed at 48-hour intervals, and culture was continued for 4 weeks. A preliminary assessment of tissue-engineered valves could be made by visual inspection of the surface of the graft through the roller bottle and by gently inverting the bottle so that fluid filled the outflow section of the conduit.
Surgical Technique
Autologous heart valves were implanted into 6 juvenile sheep (Dover/Suffolk cross). The sheep were housed and cared for according to US Department of Agriculture and institutional guidelines. Animals were sedated with ketamine 25 mg/kg IM and xylazine 0.05 mg/kg IM, intubated, and maintained on isoflurane. The heart was approached through a left fourth space thoracotomy incision, and cardiopulmonary bypass was established with an Optima XP oxygenator with integral 4 l hard shell reservoir (Cobe Cardiovascular Inc). A short segment of pulmonary artery was excised, native pulmonary valve leaflets were excised, and the autologous tissue-engineered valve was interposed between the 2 ends. Samples of each of the tissue-engineered valves were excised before implantation for recharacterization of cells and histology. The chest was closed, and the animal was allowed to recover.
Evaluation of the Tissue-Engineered Heart Valves
A qualitative assessment of each valve was made in terms of tissue handling, ability to hold sutures, and hemostasis. Echocardiographic evaluation was carried out at the end of the procedure with a Sonos 5500 system (Philips). Two-dimensional imaging and color Doppler imaging on the short and long axes were performed. Regurgitation was graded as trivial, mild, moderate, or severe by the regurgitant jet width obtained with color Doppler. Proximal mean velocity and total maximum instantaneous and mean velocities were measured. The maximum instantaneous and mean gradients were calculated from the modified Bernoulli equation. Effective orifice area was calculated from the continuity equation. Hemodynamic function was reevaluated at 4 months.
Histology
Samples for histological examination were fixed in 4% paraformaldehyde, embedded in paraffin, cut in parallel 5- to 7-μm sections, mounted, and stained with Movat pentachrome to evaluate extracellular matrix. Immunohistochemistry was performed using the avidin-biotin-peroxidase method with monoclonal anti-vimentin, anti–-SMA, and anti–von Willebrand factor primary antibodies (DAKO Corp). Tissue sections were treated with hydrogen peroxide to inhibit endogenous peroxidase activity and then incubated with the primary antibody in PBS with 4% species-appropriate serum. Subsequent processing was carried out according to the manufacturer’s recommendations. The reaction was visualized with 3-amino-9-ethyl carbazole. Sections were counterstained with Meyer’s hematoxylin solution. Adjacent sections were treated with nonimmune IgG as controls for antibody specificity. Slides were viewed under a Nikon microscope. Normal ovine pulmonary valve leaflets were stained contemporaneously for comparison.
Statistical Analyses
All data are expressed as mean±SEM. Continuous variables were compared by use of a 2-tailed paired t test.
Results
Characterization of the MSCs
MSCs were isolated from the bone marrow harvested from juvenile sheep. Similar to human MSCs, these cells did not express the leukocyte common antigen CD45 or CD14, which were demonstrated by flow cytometric analysis (Figure 1a and 1c). To ensure immunoactivity of both anti-CD45 and anti-CD14 antibodies, Figure 1b and 1d shows that 98% of leukocytes isolated from ovine peripheral blood expressed CD45, and 6% were positive for CD14. The percentage of CD14+ cells is consistent with the proportion of monocytes present in peripheral blood. Similar to human MSCs, ovine MSCs expressed ;1 integrin CD29 but failed to express the thymocyte and lymphocyte marker CD4, endothelial antigen CD31, and von Willebrand factor (data not shown).
Multipotency of ovine MSCs was demonstrated by inducing cells to differentiate along adipogenic and osteogenic lineages using lineage-specific inducing media12 (Figure 1e and 1f). MSCs at low density displayed the spindle-shaped morphology characteristic of fibroblast-myofibroblast cell lineage and expressed the VA phenotype markers vimentin and -SMA14 (Figure 1g and 1h).
Cells harvested from the tissue-engineered valves at the time of implantation displayed an identical antigen profile. These cells also retained the capacity to differentiate into adipocytes and osteocytes.
Physical Properties of the Biodegradable Scaffold
A unique biodegradable scaffold was developed next that incorporated both PGA and PLLA fibers (Figure 2). Bending stiffness of fresh ovine pulmonary valve leaflets was 3.6±0.6 mN-mm2. Effective stiffness of the textile samples after 0 and 4 weeks in cell-conditioned medium is shown in Figure 2a. Initial stiffness of the scaffold was 99.3±10.5 mN-mm2. After 4 weeks in culture, the material became much more pliable; stiffness measured just 18.3±3.1 mN-mm2, which approaches that of the native ovine pulmonary valve leaflets (3.6±0.6 mN-mm2).
Tensile strength of samples of the scaffold as they degrade is displayed as a function of time in culture medium (Figure 2b). Tensile strength dropped very rapidly over the first 4 weeks but diminished much more slowly thereafter. At 24 weeks, tensile strength was 30% of the initial strength of the scaffold. Nevertheless, estimated burst strength was still 270 mm Hg for a 21-mm-diameter conduit, far exceeding the target strength for a pulmonary valve substitute.
Appearance and Geometry of the Tissue-Engineered Valves
The macroscopic appearance of the stem-cell tissue-engineered heart valves after removal from culture and before implantation is shown in Figure 3a. There was remarkable consistency in macroscopic appearance between each of the 6 grafts before implantation. The view from the ventricular side of one of the grafts shows the 3 valve leaflets in apposition, as they would appear during diastole (Figure 3b). In profile, bulges in the wall of the graft, marked with arrows, correspond to the recreated sinuses of Valsalva.
The tissue-engineered heart valves handled well surgically. There was no evidence of tearing or damage to the grafts during surgical manipulation. The tissue-engineered valves also held sutures well. The grafts were hemostatic, and bleeding through needle holes was easily controlled by reversal of heparin at the end of the operation.
Noninvasive echocardiographic imaging was used to assess the geometry of grafts after implantation. In profile, the valves displayed convex bulges on the surface of the graft analogous to the normal sinuses of Valsalva and smooth continuity between the valve leaflets and their respective sinuses. This geometry is also known to create vortexes in the flow of blood that initiate the prediastolic movement of leaflet closure. The short-axis view in mid systole shows the characteristically triangular orifice of the open valve and absence of "reverse bending" at the commissures (Figure 3c). The long-axis view shows the valve interposed between the right ventricle and the distal main pulmonary artery (Figure 3d). Measurement of root dimensions in vivo, obtained from the echocardiograms, showed a mean annulus diameter of 19.7±0.77 mm after implantation and a mean maximum sinus diameter of 23.8±0.87 mm.
Two animals died early, 1 at 24 hours and 1 after 7 days. The clinical history, antemortem examination, laboratory findings, and autopsy were consistent with respiratory failure and gastrointestinal hemorrhage, respectively, as the causes of death. The tissue-engineered valves from these animals were also explanted and examined grossly and histologically for signs of infection or thrombosis. There was no evidence of either of these pathologies.
After surviving the immediate postoperative period, the 4 animals remained in good general health; they continued to eat and drink normally. Figure 5 compares hemodynamic data obtained at 4 months in the anesthetized animals with measurements obtained under similar hemodynamic loading conditions at the time of implantation. There was an increase in maximum instantaneous gradient of 11±7.4 mm Hg (P=0.23), mean gradient increased by 5.2±4.9 mm Hg (P=0.36), and effective orifice area decreased by 0.05±0.19 cm2 (P=0.83). These differences failed to reach statistical significance.
Histology of the Tissue-Engineered Heart Valves
Similar to native valve leaflets (Figure 6a), extracellular matrix of the tissue-engineered valve leaflets was organized into layers (Figure 6b). An abundance of collagen fibers, staining yellow, was apparent on the outflow surface of the graft, analogous to the zona fibrosa of the native valve (Figure 6c). In contrast, the inflow surface of the valve was characterized by the presence of elastin fibers, staining black and characteristic of the native zona ventricularis (Figure 6d). Glycosaminoglycans, staining green, were disposed throughout much of the remainder of the valve leaflet, as seen in the native valve. The tissue-engineered leaflets, however, were much thicker than their native counterparts, and corrugations were absent from the outflow surface of the valve leaflets.
Phenotypic Expression in the Tissue-Engineered Heart Valve
Changes in structure were mirrored by changes in the distribution of cell phenotypes across the tissue-engineered valve leaflets. Tissue sections were stained separately for vimentin and -SMA. Comparisons were made between samples of valves harvested at the time of implantation and samples from valves explanted after 8 months in vivo. Sections of native ovine pulmonary valve leaflets were included as controls (Figure 7).
As in native valves, cells expressing vimentin were found to be evenly distributed throughout the full thickness of samples taken from tissue-engineered valves at the time of implantation. This pattern of distribution remained unchanged at 8 months (Figure 7a through 7c).
Samples of tissue-engineered valves taken at the time of implantation showed diffuse staining of -SMA throughout the sections (Figure 7d). In contrast, by 8 months, the -SMA–positive cells were confined to the subendothelial layer, in striking similarity to the pattern of staining observed in native valve leaflets (Figure 7e and 7f).
The endothelial cell marker von Willebrand factor was not present in the valves at the time of implantation (Figure 8a). However, explanted tissue-engineered grafts demonstrated a continuous and uninterrupted layer of von Willebrand factor–positive cells (Figure 8b).
Discussion
In several respects, the bone marrow is an ideal source of cells for tissue engineering. Marrow is easily accessible. The primary isolate is a suspension of cells that is much easier to process than solid tissue samples previously used in this field and presents much less opportunity for microbial contamination during harvest and processing. Finally, the unique properties of stem cells—multipotency and self-renewal—present clear advantages of tissue engineering over fully committed cells.
Although MSCs have been isolated from several different species,12,18 there is little or no published data in respect to MSCs in sheep marrow. Thus, our first step was to establish that MSCs could indeed be isolated from bone marrow in the ovine model. Flow cytometric analysis indicated that these cells did not express the leukocyte common antigen CD45 or CD14 expressed by monocytes, which distinguishes them from the much more numerous hematopoietic cell fraction of the marrow. Multipotency of these cells was also confirmed by their ability to differentiate into adipogenic and osteogenic lineages. Moreover, when cultured at low density, MSCs display the spindle-shaped morphology characteristic of fibroblast-myofibroblast cell lineage. They also express the VA phenotype markers vimentin and -SMA displayed by resting interstitial cell population of valves in culture.14 These findings lead us to hypothesize that MSCs isolated from ovine bone marrow could be used to provide cells of fibroblast-myofibroblast lineage for the interstitial cell fraction of a substitute heart valve.
We further questioned whether this cell phenotype would remain stable throughout the lengthy tissue engineering process, an important issue that has been raised previously in respect to the processing of cells destined for human therapies.19 Therefore, samples of "tissue" were excised from tissue-engineered heart valves immediately before surgical implantation for enzymatic digestion, plating, and recharacterization. These studies confirmed that MSCs retained the VA phenotype and the capacity to differentiate into multiple mesenchymal cell types. In summary, the results strongly supported the use of marrow-derived MSCs to create autologous heart valves through tissue engineering.
A unique biodegradable scaffold was developed in this study to act as a temporary vehicle for the stem cells. Numerous studies have demonstrated that the open pore structure and extensive interporous connections displayed by the nonwoven fiber architecture provide an effective substrate for cell attachment and tissue formation in vitro.20,21 However, previous studies have failed to recreate the highly complex surface geometry of the native heart valve using this type of material or to provide for mid-term durability of these grafts while living elements of the engineered tissue mature and develop strength in vivo.21,22 The combination textile used here incorporated both PGA fibers to provide a physical template for cells to adhere to and PLLA fibers to provide sustained mechanical strength to the nascent tissue. PGA and PLLA currently are used in medical implants; extensive studies have found them to be nontoxic. Implants fabricated from PLLA degrade much more slowly than those synthesized from the unsubstituted glycolic acid monomer PGA.23 Because these polymers both degrade by hydrolysis, the textile underwent preliminary evaluation by immersion in cell-conditioned culture medium and interval testing of strength and flexibility. The results showed that the scaffold material was strong enough and pliable enough to function in its proposed physiological role in the pulmonary circulation after 4 weeks, the proposed time for implantation, and for many months thereafter. In fact, the results suggested that the material was still a little stiffer than native valve leaflets. Nevertheless, these values are similar to values observed in the glutaraldehyde-fixed pericardial bioprostheses used clinically.24
As to the potential biological response to scaffold resorption, the degradation products of PGA and PLLA are glycolic acid and L-lactic acid, the naturally occurring stereoisomer of lactic acid, respectively. These products are metabolized normally to carbon dioxide and water. Buildup of local concentrations could potentially increase pH around the scaffold by overtaxing the transportation capacity of surrounding tissues. Studies in vitro have shown that PGA incubation solutions are toxic after 10 days of incubation.25 However, studies such as this in which solutions are not changed during the incubation intervals provide a worst-case model of the effects of accumulation of degradation products and do not reflect the washout expected in vivo generally or in the unique environment where the tissue-engineered valve leaflets find themselves, in the path of the entire circulating cardiac output. Furthermore, the accumulation of toxic metabolites is more of a concern in orthopedic applications in which the mass of bulk polymer is significant and a mild, delayed foreign body reaction to implanted polyesters is described.26 Interestingly, a typical foreign body reaction was not observed in any of the explanted tissue-engineered valves even at the implant–native tissue interface where one might arguably expect it to be most evident.
Although the scaffold was evaluated in vitro in this study for just 6 months, studies have shown that PLLA can last up to 2 years in vivo23; therefore, the long-term mechanical properties of the tissue-engineered heart valve will ultimately need to be tested over a much more extended time. To mimic the complex structure of native valves, needle punching was used to fit the flat textile over molds. Computer simulations and finite-element models have shown that the geometry of native heart valves and their prosthetic counterparts is important for long-term durability of valve leaflets in vivo.27,28 It is to be anticipated that recreating this structure will prolong the longevity of the valves in vivo.
It took 2 to 3 weeks to expand a sufficient number of cells and 4 weeks of culture on the heart valve scaffolds to produce the tissue-engineered valve. It is possible that efficiencies could reduce this time. Nevertheless, the macroscopic appearance and surface geometry of the valves at the conclusion of the culture phase were remarkably consistent. Echocardiography showed that this geometry was maintained in vivo. The absence of reverse bending at the commissures, known to have a negative impact on the long-term durability of artificial heart valves,29 was also noted on echocardiography. Systolic gradients observed across the engineered valves at the time of implantation and after 4 months in all the surviving animals compared favorably with published values for bileaflet mechanical and stented biological valves of similar dimensions.15–17 Little regurgitation was noted at implantation. However, it was not possible to obtain adequate views of the valve to assess regurgitation at the 4-month time point owing to scarring and interposition of the lung. Unfortunately, transesophageal echocardiography was not available.
The native heart valve is a layered structure.3 The outer layer or zona fibrosa is composed principally of collagen; the middle region, zona spongiosa, contains an abundance of glycosaminoglycans that are hydrated in the living state to form a gel; and the inner layer, the zona ventricularis, is characterized by the presence of radially arranged elastin fibers.
At low magnification, there was striking similarity between the tissue-engineered heart valve and the native valve leaflets in terms of the distribution of the extracellular matrix elements collagen, glycosaminoglycans, and elastin. Furthermore, an abundance of collagen fibers was observed on the outflow surface of the tissue-engineered heart valve, reminiscent of the zona fibrosa. In contrast, the inflow surface of the valve contained relatively little collagen but was remarkable for the presence of a thin layer of elastin fibers characteristic of the zona ventricularis. Glycosaminoglycans were distributed through the remainder of the valve structure. Clear differences remained, however. The tissue-engineered leaflets were much thicker than their native counterparts and surface corrugations were absent from the tissue-engineered valve. In the absence of further remodeling, these differences in structure would likely have a negative impact on the longevity of the valve leaflets. However, these problems may be improved by using a much thinner scaffold material in the valve leaflets and possibly by applying physical stimuli to the valve during culture.30
Changes in structure of the extracellular matrix were mirrored by changes in the distribution of cell phenotypes across the valve leaflets. Myofibroblasts (VA phenotype) were confined to a layer immediately below the surface endothelium. In contrast, fibroblastlike cells expressing vimentin but not -SMA (V phenotype) were distributed throughout the remainder of the graft, reminiscent of the native valve.31 This distribution of cell phenotypes is similar across native human aortic and pulmonary valves. It is not clear to what extent the change in phenotypes observed on the tissue-engineered valve at explantation represents a true change in phenotype of the implanted cells in vivo or in growth of cells from outside the graft. Certainly, in the absence of specific cell labeling studies, it is not possible to ascertain the precise origin of the cell phenotypes observed in the explanted tissue-engineered valves. One possible explanation is that stem cells undergo milieu-specific differentiation into tissue-specific cells, ie, cells resident in the native pulmonary valve. In support of this hypothesis, it is known that fibroblasts isolated from cornea can undergo transformation in the opposite direction, ie, from fibroblasts to myofibroblasts, when cultured in vitro.32 Furthermore, the differentiation of stem cells within the engineered heart valve leaflets would be in concordance with other studies in which stem cells have been tracked to a variety of different organs where they undergo tissue-specific differentiation into resident cell phenotypes.18 Although the results of some of these studies have been called into question by the demonstration of cell fusion in vitro,33–35 this event is rare and is unlikely to be responsible for the distribution of cell phenotypes observed in the explanted tissue-engineered heart valves.
Interestingly, the explanted grafts also demonstrated a continuous and uninterrupted endothelial layer on the blood-contacting surface of the grafts. The importance of the endothelium rests partly in its antithrombogenic properties but more importantly in its potential to modulate the underlying interstitial cells. Vascular endothelial dysfunction is thought to play a role in the development of atherosclerosis in blood vessels.36 Moreover, similarities between the morphology of early lesions of nonrheumatic aortic stenosis and atherosclerotic plaques have raised the possibility of a common underlying origin.37 However, existing heart valve prostheses do not become endothelialized or lose any residual endothelium present on the graft at the time of implantation after a short period in vivo.38 The appearance of a vascular endothelium on the tissue-engineered grafts is therefore a significant finding that is expected to have a positive impact on their long-term function. Once again, the precise origin of these cells is less clear. The endothelial lining could arise from deposition of circulating endothelial cells.39 Alternatively, transdifferentiation of the implanted MSCs may be responsible by a process analogous to the transdifferentiation of valvular endothelial cells into mesenchymal cells observed in vitro.40,41
An expanding number of alternative cell sources that hold potential for tissue engineering of heart valves have been examined. The list encompasses cells from all types of tissues and stages of development. During fetal life, amniotic fluid is available for harvest,42 and at birth, both the placenta43 and cord blood44 can provide multipotent cells. After birth, fully committed cells can be obtained from related tissues21,22 or even from the very tissue of interest.45 However, it is generally acknowledged that stem cells present significant advantages for tissue engineering over differentiated cells5 and that MSCs obtained from bone marrow remain the most promising source of cells for recreating heart valves.
We have shown that it is possible to create autologous living heart valves from bone marrow–derived MSCs in combination with a synthetic biodegradable scaffold and demonstrated that such tissue-engineered structures function within the circulation for prolonged periods. These valves undergo extensive remodeling in vivo and adopt cellular phenotypes and structural organization strongly reminiscent of native heart valves. Long-term durability of the valves depends on a variety of factors, including the geometry of the valves, degradation profile of the polymer components over time, and cellular responses to these changes and to the local physiological environment. With further advances, this technology may provide a superior alternative for patients in need of heart valve replacement.
Acknowledgments
Dr Sutherland was supported by an International Research Fellowship sponsored by the British Cardiac Society and Merck, Sharpe and Dohme. This work was supported by the National Institutes of Health.
Footnotes
The online-only Data Supplement can be found with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.104.498378/DC1.
References
Schoen FJ. Pathology of heart valve substitution with mechanical and tissue prostheses. In: Silver MD, Gotlieb AI, Schoen FJ, eds. Cardiovascular Pathology. New York, NY: WB Saunders; 2001: 629–677.
Dalrymple-Hay MJ, Pearce RK, Dawkins S, Alexiou C, Haw MP, Livesey SA, Monro JL. Mid-term results with 1,503 CarboMedics mechanical valve implants. J Heart Valve Dis. 2000; 9: 389–395.
Schoen FJ, Levy RJ. Founder’s Award, 25th Annual Meeting of the Society for Biomaterials, perspectives: Providence, RI, April 28–May 2, 1999: tissue heart valves: current challenges and future research perspectives. J Biomed Mater Res. 1999; 47: 439–465.
Perron J, Moran AM, Gauvreau K, del Nido PJ, Mayer JE Jr, Jonas RA. Valved homograft conduit repair of the right heart in early infancy. Ann Thorac Surg. 1999; 68: 542–548.
McLaren A. Ethical and social considerations of stem cell research. Nature. 2001; 414: 129–131.
Korbling M, Estrov Z. Adult stem cells for tissue repair: a new therapeutic concept; N Engl J Med. 2003; 349: 570–582.
Phinney DG. Building a consensus regarding the nature and origin of mesenchymal stem cells. J Cell Biochem Suppl. 2002; 38: 7–12.
Devine SM, Cobbs C, Jennings M, Bartholomew A, Hoffman R. Mesenchymal stem cells distribute to a wide range of tissues following systemic infusion into nonhuman primates. Blood. 2003; 101: 2999–3001.
Booth C, Korossis SA, Wilcox HE, Watterson KG, Kearney JN, Fisher J, Ingham E. Tissue engineering of cardiac valve prostheses, I: development and histological characterization of an acellular porcine scaffold. J Heart Valve Dis. 2002; 11: 457–462.
Teebken O, Mertsching H, Haverich A. Modification of heart valve allografts and xenografts by means of tissue engineering. Transplant Proc. 2002; 34: 2333.
Schenke-Layland K, Opitz F, Gross M, Doring C, Halbhuber KJ, Schirrmeister F, Wahlers T, Stock UA. Complete dynamic repopulation of decellularized heart valves by application of defined physical signals: an in vitro study. Cardiovasc Res. 2003; 60: 497–509.
Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, Mosca JD, Moorman MA, Simonetti DW, Craig S, Marshak DR. Multilineage potential of adult human mesenchymal stem cells. Science. 1999; 284: 143–147.
Gloeckner DC, Billiar KL, Sacks MS. Effects of mechanical fatigue on the bending properties of the porcine bioprosthetic heart valve. Asaio J. 1999; 45: 59–63.
Messier RH, Jr., Bass BL, Aly HM, Jones JL, Domkowski PW, Wallace RB, Hopkins RA. Dual structural and functional phenotypes of the porcine aortic valve interstitial population: characteristics of the leaflet myofibroblast. J Surg Res. 1994; 57: 1–21.
Chafizadeh ER, Zoghbi WA. Doppler echocardiographic assessment of the St. Jude Medical prosthetic valve in the aortic position using the continuity equation. Circulation. 1991; 83: 213–223.
Milano AD, De Carlo M, Mecozzi G, D’Alfonso A, Scioti G, Nardi C, Bortolotti U. Clinical outcome in patients with 19-mm and 21-mm St. Jude aortic prostheses: comparison at long-term follow-up. Ann Thorac Surg. 2002; 73: 37–43.
Cosgrove DM, Lytle BW, Williams GW. Hemodynamic performance of the Carpentier-Edwards pericardial valve in the aortic position in vivo. Circulation. 1985; 72 (suppl II): II-146–II-152.
Jiang Y, Jahagirdar BN, Reinhardt RL, Schwartz RE, Keene CD, Ortiz-Gonzalez XR, Reyes M, Lenvik T, Lund T, Blackstad M, Du J, Aldrich S, Lisberg A, Low WC, Largaespada DA, Verfaillie CM. Pluripotency of mesenchymal stem cells derived from adult marrow. Nature. 2002; 418: 41–49.
Sutherland FW, Mayer JE Jr. Ethical and regulatory issues concerning engineered tissues for congenital heart repair. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2003; 6: 152–163.
Langer R, Vacanti JP. Tissue engineering. Science. 1993; 260: 920–926.
Shinoka T, Breuer CK, Tanel RE, Zund G, Miura T, Ma PX, Langer R, Vacanti JP, Mayer JE Jr. Tissue engineering heart valves: valve leaflet replacement study in a lamb model. Ann Thorac Surg. 1995; 60: S513–S516.
Hoerstrup SP, Sodian R, Daebritz S, Wang J, Bacha EA, Martin DP, Moran AM, Guleserian KJ, Sperling JS, Kaushal S, Vacanti JP, Schoen FJ, Mayer JE Jr. Functional living trileaflet heart valves grown in vitro. Circulation. 2000; 102 (suppl III): III-44–III-9.
Peltoniemi HH, Hallikainen D, Toivonen T, Helevirta P, Waris T. SR-PLLA and SR-PGA miniscrews: biodegradation and tissue reactions in the calvarium and dura mater. J Craniomaxillofac Surg. 1999; 27: 42–50.
Vesely I, Boughner D. Analysis of the bending behaviour of porcine xenograft leaflets and of neutral aortic valve material: bending stiffness, neutral axis and shear measurements. J Biomech. 1989; 22: 655–671.
Taylor MS, Daniels AU, Andriano KP, Heller J. Six bioabsorbable polymers: in vitro acute toxicity of accumulated degradation products. J Appl Biomater. 1994; 5: 151–157.
Bostman OM, Paivarinta U, Partio E et al. The tissue-implant interface during degradation of absorbable polyglycolide fracture fixation screws in the rabbit femur. Clin Orthop. 1992; Dec: 263–272.
Beck A, Thubrikar MJ, Robicsek F. Stress analysis of the aortic valve with and without the sinuses of Valsalva. J Heart Valve Dis. 2001; 10: 1–11.
Grande-Allen KJ, Cochran RP, Reinhall PG, Kunzelman KS. Finite-element analysis of aortic valve-sparing: influence of graft shape and stiffness. IEEE Trans Biomed Eng. 2001; 48: 647–659.
Thubrikar MJ, Skinner JR, Eppink RT, Nolan SP. Stress analysis of porcine bioprosthetic heart valves in vivo. J Biomed Mater Res. 1982; 16: 811–826.
Engelmayr GC Jr, Hildebrand DK, Sutherland FW, Mayer JE Jr, Sacks MS. A novel bioreactor for the dynamic flexural stimulation of tissue engineered heart valve biomaterials. Biomaterials. 2003; 24: 2523–2532.
Della Rocca F, Sartore S, Guidolin D, Bertiplaglia B, Gerosa G, Casarotto D, Pauletto P. Cell composition of the human pulmonary valve: a comparative study with the aortic valve: the VESALIO Project: Vitalitate Exornatum Succedaneum Aorticum labore Ingegnoso Obtinebitur. Ann Thorac Surg. 2000; 70: 1594–1600.
Masur SK, Dewal HS, Dinh TT, Erenburg I, Petridou S. Myofibroblasts differentiate from fibroblasts when plated at low density. Proc Natl Acad Sci U S A. 1996; 93: 4219–4223.
Vogel G. Stem cell research: studies cast doubt on plasticity of adult cells. Science. 2002; 295: 1989–1991.
Terada N, Hamazaki T, Oka M, Hoki M, Mastalerz DM, Nakano Y, Meyer EM, Morel L, Petersen BE, Scott EW. Bone marrow cells adopt the phenotype of other cells by spontaneous cell fusion. Nature. 2002; 416: 542–545.
Ying QL, Nichols J, Evans EP, Smith AG. Changing potency by spontaneous fusion. Nature. 2002; 416: 545–548.
Wever RM, Luscher TF, Cosentino F, Rabelink TJ. Atherosclerosis and the two faces of endothelial nitric oxide synthase. Circulation. 1998; 97: 108–112.
O’Brien KD, Reichenbach DD, Marcovina SM, Kuusisto J, Alpers CE, Otto CM. Apolipoproteins B, (a), and E accumulate in the morphologically early lesion of "degenerative" valvular aortic stenosis. Arterioscler Thromb Vasc Biol. 1996; 16: 523–532.
Mitchell RN, Jonas RA, Schoen FJ. Pathology of explanted cryopreserved allograft heart valves: comparison with aortic valves from orthotopic heart transplants. J Thorac Cardiovasc Surg. 1998; 115: 118–127.
Asahara T, Murohara T, Sullivan A, Silver M, van der Zee R, Li T, Witzenbichler B, Schatteman G, Isner JM. Isolation of putative progenitor endothelial cells for angiogenesis. Science. 1997; 275: 964–967.
Rafii S, Lyden D. Therapeutic stem and progenitor cell transplantation for organ vascularization and regeneration. Nat Med. 2003; 9: 702–712.
Paranya G, Vineberg S, Dvorin E, Kaushal S, Roth SJ, Rabkin E, Schoen FJ, Bischoff J. Aortic valve endothelial cells undergo transforming growth factor-beta–mediated and non–transforming growth factor-beta–mediated transdifferentiation in vitro. Am J Pathol. 2001; 159: 1335–1343.
Kaviani A, Guleserian K, Perry TE, Jennings RW, Ziegler MM, Fauza DO. Fetal tissue engineering from amniotic fluid. J Am Coll Surg. 2003; 196: 592–597.
Kaviani A, Perry TE, Barnes CM, Oh JT, Ziegler MM, Fishman SJ, Fauza DO. The placenta as a cell source in fetal tissue engineering. J Pediatr Surg. 2002; 37: 995–999.
Wu X, Rabkin-Aikawa E, Guleserian KJ, Perry TE, Masuda Y, Sutherland FW, Schoen FJ, Mayer JE Jr, Bischoff J. Tissue-engineered microvessels on three-dimensional biodegradable scaffolds using human endothelial progenitor cells. Am J Physiol Heart Circ Physiol. 2004; 287: H480–H487.
Maish MS, Hoffman-Kim D, Krueger PM, Souza JM, Harper JJ3rd, Hopkins RA. Tricuspid valve biopsy: a potential source of cardiac myofibroblast cells for tissue-engineered cardiac valves. J Heart Valve Dis. 2003; 12: 264–269.(Fraser W.H. Sutherland, F)