Determinants of Bioartificial Myocardial Graft Survival and Engraftment In Vivo

Determinants of Bioartificial Myocardial Graft Survival and Engraftment In Vivo Knut Mueller-Stahl, MD,a,b* Theo Kofidis, MD, PhD,a,c,d,e* Payam Akhyari, MD,a,c Darren H. L. Lee, MB, BCh,d Andre Lenz, MD,a,c Eliana C. Martinez, MD,e Felix Woitek, MS,f and Axel Haverich, MD, PhDa,c

Background: The specific interactions between tissue-engineered grafts and host tissue are frequently neglected. The aim of this study was to describe and quantify the fate of a tissue-engineered cardiac graft in vivo. Methods: Neonatal rat cardiomyocytes were cast into a collagen mesh, forming a bioartificial myocardial tissue (AMT). After 7 days in vitro, four groups were formed (Group A: sham; Group B: matrix; Group C: AMT [with additional host treatment with cyclosporine and prednisolone]; Group D: AMT; each n ⫽ 5) and the tissue grafts were implanted into the muscle pouch of adult rats at 14, 21 and 28 days. Implants were stained for troponin-T, BrdU, MF-20, desmin, vimentin, Flk-1, CD8, CD4, pentachrome, PSR and H&E. Results: AMT cell count, cell proportion, contractility, viability and metabolism proved stable in vitro. Grafted cells decreased over time and were detected in Group C until the end of the experiment (Day 28), and in Group D until Day 21. Angiogenesis began at the peripheries and slowly progressed toward the cores of the grafts. The thickness and collagen content of the matrix remained stable in Group C for 14 days, and decreased in all groups until Day 28 (thickness: Group B, ⫺66%; Group C, ⫺50%; Group D, ⫺100%). Grafts were predominately infiltrated by macrophages and stromal cells, and less so by lymphocytes (Group D ⬎ B ⬎ C). Conclusion: The differentiation of cardiac and non-cardiac grafted cells, infiltrating cells, scaffold kinetics and angiogenesis showed host immune responses and degree of angiogenesis to be the determinants for AMT graft survival. J Heart Lung Transplant 2008;27:1242–50. Copyright © 2008 by the International Society for Heart and Lung Transplantation.

The incidence of heart disease and myocardial infarction continues to increase. Current efforts to replace diseased heart tissue are based on either direct cell transfer or implantation of engineered tissue. The former has recently been entered into clinical trials, with heterogenous findings.1,2 Methodically, in the field

From the aLeibniz Institute of Biotechnology and Artificial Organs, Hannover Medical School, Hannover, Germany; bDepartment of Traumatology, Klinikum Bremen Mitte, Bremen, Germany; cDepartment of Thoracic and Cardiovascular Surgery, Hannover Medical School, Hannover, Germany; dDepartment of Cardiac, Thoracic and Vascular Surgery, National University Hospital, Singapore; eDepartment of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore; and fDepartment of Internal Medicine and Cardiology, Leipzig Heart Centre, University of Leipzig, Leipzig, Germany. *These authors contributed equally to the study. Submitted March 28, 2008; revised August 5, 2008; accepted August 6, 2008. Reprint requests: Theo Kofidis, MD, Department of Cardiac, Thoracic and Vascular Surgery, National University Hospital, 5 Lower Kent Ridge Road, Level 2, 119074 Singapore. Telephone: ⫹65-67722065. Fax: ⫹65-6776-6475. E-mail: [email protected] Copyright © 2008 by the International Society for Heart and Lung Transplantation. 1053-2498/08/$–see front matter. doi:10.1016/ j.healun.2008.08.003

1242

of direct cell transfer, a cell suspension is administered by injection into a focal site of the organ, and organization of the grafted cells should take place in vivo to improve organ function. By tissue engineering, a functional construct is created in vitro, and then transferred in vivo. The currently available methods are limited by: (1) scar tissue preventing the interconnection of grafted and host cells; (2) injection of non-contractile cells, matrix or fluid improves heart function through wall thickening (law of Laplace) as well as post-infarction scarring and reactive angiogenesis (inflammation); and (3) the fact that most cells are lost to post-infarction ischemia and inflammation after injection.3–5 The fate of grafted tissue depends not only on graft composition, but also on location, immunologic barriers and time of implantation in vivo. These processes are mediated by multiple interactions of host and graft tissue, which remain poorly characterized. Therefore, qualitative analysis of bioartificial grafts in vivo is often limited by low specificity and susceptibility to random observations.6 We present a control-group– based comparative quantitative analysis to describe the fate of a bioartificial myocardial tissue construct, using specific histologic markers both in vitro and in vivo.

The Journal of Heart and Lung Transplantation Volume 27, Number 11

METHODS All procedures conformed to the Guide for the Care and Use of Laboratory Animals (published by the National Academy Press, Washington, DC, 1996) and the “Guidelines for the Use of Laboratory Animals for Research” (Hannover Medical School, Hannover, Germany). Cell Isolation and Characterization Neonatal Wistar rat (outbred strain) hearts were excised at the atrioventricular junction and digested with trypsin and DNAse as previously described.7 Viability, cell identity and cell location were evaluated by co-localization using immunohistochemistry (MF-20, troponin-T and 5-bromo-2-deoxyuridine [BrdU]) and labeling procedures (carboxyflourescein diacetate-succinimidyl ester [CFDA-SE] and 4=,6-diamidino-2-phenylindole [DAPI]) after isolation and on day of implantation.8 Briefly, 500 ␮g of CFDA-SE (Molecular Probes, Leiden, The Netherlands) powder was dissolved in 90 ␮l of dimethylsulfoxide (DMSO), resulting in a 10-mmol/liter stock solution. For labeling purposes, we used 10 ␮mol/ liter of the stock solution in 15 ml of phosphate-buffered saline (PBS), free of magnesium and calcium, containing 5 mmol/liter of ethylene-diamine tetraacetic acid (EDTA). After centrifugation (600 rpm for 10 minutes), a pellet of freshly isolated cells was resuspended in pre-warmed labeling dye and incubated for 15 minutes (5% CO2, 37°C). Cellular uptake of CFDA-SE was stopped by adding 250 ␮l fetal bovine serum (FBS), followed by an additional incubation for 30 minutes. Finally, the labeling procedure was terminated with two cycles of centrifugation (600 rpm, 15 minutes) and resuspension in 15 ml of pre-warmed PBS, supplemented with 0.1% bovine serum albumin (BSA). Finally, the cell pellet was resuspended in modified Eagle’s medium, supplemented with 1 ml 10% FBS, 5 ml penicillin/streptomycin and 110 ␮mol/liter BrdU, yielding a cell concentration of 3 ⫻ 106 cells/ml. For cell characterization the labeled cell solution was diluted to 5 ⫻ 105 cells/ml. Silane-coated slides were clamped in cytofuge devices (Cytospin 2; Shandon, Astmoore, UK). Each device was loaded with 200 ␮l of the cell suspension and spun for 5 minutes at 600 rpm. The cytospot slides were then removed, dried at room temperature in the dark, and frozen at ⫺80°C. Preparation and Cultivation of Tissue-engineered Myocardial Tissue Graft (AMT Construct) An equine collagen (Type I) matrix (Tissue Fleece; Baxter Deutschland, Heidelberg, Germany) was minced into pieces of 18 mm ⫻ 18 mm ⫻ 2 mm and placed into prepared cell culture dishes. The matrix was seeded

Mueller-Stahl et al.

1243

either with labeled (CFDA-SE) or non-labeled freshly isolated neonatal cardiac cell suspension (3 ⫻ 106 cells in 1 ml of suspension). During a cultivation of 7 days (37°C, 5% CO2, saturated humidity, change of culture medium every second day), morphology, contractility and metabolic activity (glucose, lactate) were analyzed every second day. On Day 7, co-localization with immunohistochemistry, hematoxylin– eosin (H&E) and pentachrome staining and Live/Dead Assay (Molecular Probes, Leiden, The Netherlands) was repeated on constructs that had been seeded with labeled cell suspension. Cellular counts were obtained on Days 4, 7 and 14 by applying the MTS Assay (CellTiter 96 Aqueous One Solution Kit, Cat. No. #G5421; Promega, Mannheim, Germany). Implantation Procedures and Post-operative Treatment Adult Wistar rats (allogeneic, outbred strain) were anesthetized with a cocktail of 0.3 ml of 10% ketamine (Graeub/Albrecht, Germany) and 0.05 ml 2% xylazine (Bayer, Germany). An intermuscular pouch of 2 ⫻ 2 cm was created between the musculus pectineus and adductor longus. One control group (Group A, n ⫽ 15) underwent only sham surgical treatment. The collagen patches (Group B, n ⫽ 15) and the AMT (Groups C and D, n ⫽ 30, in vitro cultivation for 7 days each) were transferred into the pouch (Figure 2a), fixed with two or three sutures (Prolene 5-0; Ethicon, Hamburg, Germany), and the skin was closed. Half of the animals with AMT implants received daily intraperitoneal immunosuppression (Group C, cyclosporine 5 mg/kg body weight and prednisolone 5 mg/kg body weight). Animals were killed after 14, 21 and 28 days (n ⫽ 5 each per group) and specimens excised en bloc. Histology Tissue constructs were washed in PBS three times, cut into two pieces, and fixed in 4.7% formaldehyde. Specimens were dehydrated and embedded in paraffin blocks. Cross-sections of the constructs (6 to 7 ␮m) were stained with H&E. To assess development of the cardiac grafts in vivo, the specimens were analyzed for specific cardiac characteristics in morphology (rodshaped, perinuclear free space, striated myofilaments). For immunohistochemistry as well as pentachrome and Sirius red staining, sections were first liberated from paraffin by rehydration and incubation in boiling citrate buffer. Immunohistochemistry was performed using primary antibodies directed against troponin-T, MF-20, desmin, vimentin, BrdU, Flk-1, CD8 and CD4. We also used the following secondary antibodies: biotinylated horse–anti-mouse IgG; labeled donkey–anti-mouse–IgG–Cy3TM; labeled donkey–antimouse–IgG–Cy2TM, labeled goat–anti-mouse IgG and labeled goat–anti-rabbit IgG. Haerrishaemalaun was

1244

Mueller-Stahl et al.

used for counterstaining non-fluorescent slides and 4=,6-diamidino-2-phenylindole (DAPI) for fluorescent slides. Quantitative Data Analysis Quantitative analysis of the slides was divided into a global (whole implant) and zonal model to differentiate the core of each implant from the periphery. Both models were evaluated by taking random photomicrographs (n ⫽ 5 to 10, magnification ⫻400) of three different levels of the harvested cross-sectioned grafts, in all groups, and on Days 0, 14, 21 and 28 in vivo. “Periphery” was defined as a rectangular area of the aforementioned micrograph reaching the host– graft interface of the sectioned specimen. “Core” was defined as a rectangular area, centered in the bisector of the cross-sectioned specimen. Five independent observers were asked to assess the same digitalized photomicrographs for the quantity (0 —nothing, 1—few, 2—moderate, 3—many, 4 —very many, 5—almost collagen) and the architecture of the collagen fibers (0 —no architecture, 1—architecture non-woven, 2—architecture woven) within the AMT constructs. The results on the day of implantation served as reference values for the in vivo study. All data are expressed as mean and SD. Comparison of two groups was performed using Student’s t-test for independent variables. Multiplerange differences were analyzed by analysis of variance, using the post hoc Bonferroni/Dunnett test. Data from the blind observer study were calculated by Mann– Whitney U-test. SPSS, version 11.0/11.5 (Superior Performing Software Systems), was used for statistics. RESULTS In Vitro Phase Cell isolation revealed 73% viable cardiomyocytes (vCM), 16% viable non-cardiomyocytes (vnCM), 6% dead cardiomyocytes (dCM) and 5% dead non-cardiomyocytes (dnCM) (Figure 1a). Simultaneously, the AMT cell count decreased from 3 ⫻ 106 cells to 2.5 ⫻ 106 cells. Immunohistochemistry for co-localization on Day 7 showed a vCM:vnCM ratio of 9:1, indicating that cell loss was mainly due to non-cardiac cells. The AMT showed spontaneous contractions from Day 3. Crosssections on Day 7 showed a highly interconnected tissue network consisting of cell– cell and cell– collagen fiber contacts (Figure 1c, arrows). There were neither differences in cell count for cardiomyocytes using troponin-T or MF-20 staining nor in cell count for non-cardiomyocytes when comparing co-localization (vnCM), vimentin or BrdU staining. However, cardiac cell density was higher along the periphery than in the center of the AMT constructs (19.4 ⫾ 5.0 vs 14.3 ⫾ 4.5, p ⫽ 0.01), but this was not

The Journal of Heart and Lung Transplantation November 2008

the case for non-cardiac cells (Figures 1c and 5a and b). There were no dead cells within the AMTs. Production of extracellular matrix proteins (blue area) correlated positively with cell density within the AMT (Figure 1e). Cell-seeded matrix was thicker than unseeded (Figure 5a). Consumption of glucose and production of lactate decreased from Day 4 onward. In Vivo Studies Morphology in situ. The upper edge of the implants was covered by a thin layer of connective tissue (Figure 2b). In Group D, only the sutures remained on Day 28. In contrast to Group C, neo-vascularization into the upper edge of the grafts could be seen in Group D from Day 14, and to a lesser extent in Group B from Day 21 as well (Figure 2b, arrows). However, no contractions were visible at the edge of the grafts in Groups C or D. There were no signs of infection or inflammation in any group. Cell survival within the graft: protective effects of immunosuppression. Grafted cardiomyocytes and noncardiomyocytes were observed in Group C throughout the experiment. Without immunosuppression (Group D), cell count declined significantly and was detectable only until Day 21. The maximal duration of cardiac graft survival was 12 weeks with immunosuppression. Cell count was 2.4 ⫾ 1.4 (whole implant), whereas zonal evaluation could not be performed at a mean thickness of 310 ⫾ 90 ␮m (n ⫽ 3) due to zonal overlapping in the photomicrograph setting. In Group C, the decrease of cell count was higher in the center than along the periphery on Days 14 and 28. On Day 28, in Group C, 25.0% (peripheral) and 12.3% (central) of grafted cardiomyocytes remained, whereas cell count in Group D was negative (Figures 2e and f and 3a). The morphologic changes of maturation were seen along the periphery of Group C, but was not detected elsewhere. Cell count of non-cardiac grafted cells in the grafts was about 10% or less of cardiac grafted cells (Groups C and D) (Figures 3a and 3b). Global analysis also revealed a higher cell count in Group C than in Group D at all times. The density of non-cardiac cells decreased in Group D from Day 0 to Day 14 and from Day 0 to Day 28 in Group C (Figures 2g and h and 3b). However, there was no statistical significance between Groups C and D, nor between central and peripheral sites in each group. Ingrowth of vessels. Microscopic evaluation revealed greater angiogenesis in the centers of Groups B and D, but not Group C. Peripheral angiogenesis was higher in

The Journal of Heart and Lung Transplantation Volume 27, Number 11

Mueller-Stahl et al.

1245

Figure 1. a: Analysis of cell proportion in isolated cardiac cell suspension. Cytospot, Immunohistochemistry using Troponin-T, CFDA-SE- and DAPI-Labeling, 400x. *viable cardiomyocyte (vCM); **viable non-cardiomyocyte (vnCM); ⫹ dead cardiomyocyte (dCM); ⫹⫹ dead noncardiomyocyte (dnCM). Bar length is 20 ␮m. b: Collagen sponge TissueFleece® before seeding procedure, PSR staining, cross section, magnification 200x. Bar length is 50 ␮m. c: AMT on day 7 in vitro, e.g. day of implantation. Troponin-T, CFDA-SE- and DAPI-labeling, magnification 400x. Cell-matrix and cell-cell interactions are marked by arrows (¡) Cross section. Bar length is 20 ␮m. d: AMT on day 7 in vitro, e.g. day of implantation, cross section. (Troponin-T staining, 400x). Highly interconnected cardiac tissue forming compact muscle strains of 10-100 ␮m. Pores are visible to the naked eye. Bar length is 100 ␮m. e: AMT on day 7 in vitro, day of implantation, cross section. Pentachrome staining. red: cytoplasm of cells; brown-black: cell nucleus; yellow: collagen fibers of TissueFleece®; blue area: production of extracellular matrix by seeded cells. Bar length is 20 ␮m.

1246

Mueller-Stahl et al.

The Journal of Heart and Lung Transplantation November 2008

The Journal of Heart and Lung Transplantation Volume 27, Number 11

Mueller-Stahl et al.

1247

Figure 3. a-b: CD-8 staining for lymphocyte identification (group C, day 14, a: periphery, b: core). There are only few positive cells in both areas. Bar length is 20 ␮m. c-h: Comparison of group B, C and D on day 14 to determine the amount of infiltrating host cells, using double staining for desmin (red ¡ myofibroblasts/myoblasts) and vimentin (green ¡ fibroblasts). Left column displays group B, whereas “c” stands for core, and “p” for periphery. Same orientation for the column in the middle (group C), and the right side (group D). Group D is heavily infiltrated by fibroblasts, but also myofibroblasts, the core is more infiltrated by fibroblasts. Same observations were made for group B, but frequently weaker, whereas group C did not show an increase in fibroblasts. Desmin positive cells have myofibrillar phenotypes and are organized in clusters (see Troponin-T). However, desmin does not allow discrimination of grafted cardiomyocytes from infiltrating myoblasts and myofibroblasts, which are normally round shaped. In the periphery, large, red areas display adjacent skeletal muscle of the host. Bar length is 100 ␮m.

Groups B and D than in Group C (p ⬍ 0.05) (Figure 2d, black arrow, and Figure 3c). Relative to the vascular network of native heart tissue (vascular distance ⱕ40 ␮m),9 vascular density in all grafts was greatly reduced (Figure 3d). The diameter of vascular structures was smaller in Group C on Days 14 and 28, compared with Groups B or D. Diameter of capillaries was thicker in the periphery than in central

position in Groups C and D until Day 21 (p ⫽ 0.000 to 0.008). Scaffold kinetics. Graft thickness remained constant in Groups B and C until Day 14 in vivo, but decreased in Group D by 30%. Subsequently, all grafts became thinner (p ⬍ 0.05 each) (Figure 2c, double arrow, and Figure 5a). The collagen content remained unchanged

Figure 2. a: Implantation procedure of grafts in the muscular pouch of adult rats in the hind limb area. Pouch is opened up by forceps, so that AMT (i) can be introduced without being folded. b: View in situ, group D on day 21 in vivo. Graft is covered by a transparent layer of stromal tissue. Sutures (blue), which are ment to prevent dislodgement of the grafts during implantation are visible. Vascular structures are visible by the naked eye, as they infiltrate from the lateral borders of the graft (white arrows). c: Microscopic overview of the graft (group B, day 21). Graft (1) lies between compact skeletal muscles (3), parted by a thin layer of connective tissue (2). Double arrow displays thickness of graft. Bar length is 150 ␮m. d: AMT receiving immunosuppressive therapy (group C, day 21, paracentral area). Yellow structures stand for collagen fibers of the matrix, still organized in a porous architecture (⫹). Embedded blue areas indicate production of extracellular matrix and remodeling processes (*). The black arrow shows a sprouting capillary, which infiltrated the graft. Bar length is 20 ␮m. e-f: Troponin-T immunohistochemistry tracks grafted cardiomyocytes (group C, day 14; e: periphery, f: core). Histology reveals cluster formation of cardiac cells, with a higher cell count in the periphery. Bar length is 20 ␮m. g-h: BrdU immunohistochemistry tracks grafted non-cardiac cells (group C, day 14; g: periphery, h: core). There are more cells in the periphery, but analysis did not show differences in cell density. Bar length is 20 ␮m.

1248

Mueller-Stahl et al.

The Journal of Heart and Lung Transplantation November 2008

Figure 4. a: Troponin-T cell count for seeded/grafted cardiac cells from day of implantation until end of study (day 28) within the grafts, parted in periphery and core. Cell count decreases in both groups with grafted cells (group C and D), but cell loss is higher without immunosuppression (group D). Mostly proportional cell loss is higher in the core than in the periphery marking out the lack of nutrition within the grafts after implantation. b: BrdU cell count for seeded/grafted non-cardiac cells in the AMT from day of implantation until day 28, parted in periphery and core within the grafts. These non-cardiac grafted cells comprise 1/10 of all grafted cells compared to cardiac grafted cells. The development of cell loss in the group of BrdU-positive grafted cells is similar to that seen in the group of cardiac grafted cells. c: Count of vascular structures infiltrating grafts in vivo (group B, C, D). On day of implantation all grafts were avascular. The amount of vascular sprouting into the grafts depended on graft composition (matrix ⫾ grafted cells) and immunological status of the host. The stagnation in the time course declares this parameter rather to be driven by the demands of the host for the degradation of the grafts than by grafted cells on their own to improve the situation of malnutrition in vivo. D: Counted vascular structures in comparison to vascular density of native heart tissue in adult rats. Reference distance in adult rat was given with 40 ␮m. Photomicrographs were obtained at magnification 400x (120 ␮m ⫻ 160 ␮m), giving a total area of 19200 ␮m2. 100% displays native situation of healthy myocardium, whereas “actual” stands for count reached by grafted tissue, difference is displayed as “deficit”. After avascular implantation no sufficient vascular status is found in any implantation group compared to healthy myocardium.

in both areas (periphery, core) of Group C after 14 days. In Groups B and D, collagen content decreased more in the periphery than in the core, respectively, whereas decline was greater in Group D than in Group B (Figure 5c). In comparison to Day 0, collagen content was reduced in Group C on Day 21, then remained constant.

In contrast, it decreased further in Group B, but especially in Group D. Collagen architecture was better preserved in Group C than in Groups B or D after implantation (Figure 5d). Thereafter, Group C retained an unchanged, homogeneous architecture of collagen fibers until Day 14. On

The Journal of Heart and Lung Transplantation Volume 27, Number 11

Mueller-Stahl et al.

1249

Figure 5. a: Thickness of in vitro cultured constructs, both seeded and non-seeded, and development after implantation. Measurements have been made using Olympus software HD-50 (Olympus, Heidelberg, Germany) adapted to inverse microscope (Olympus, Heidelberg, Germany), magnification 4-10x. Thickness decreases in all groups over time, immunosuppression reduced this process. b: Cell count of infiltrating lymphocytes (CD-8), in peripheral and central portions of all grafts projected over time. Peripheral portions of the graft are significantly more densely infiltrated. Lymphocytes infiltration depends on the presence of grafted cells and on the immunologic status of the host. However immunosuppression does not prevent lymphocytes infiltration. c: Collagen content in the matrix grafts (group B, C, D), using blind-observer evaluation. Levels at day of implantation were used as references for in vivo values. The periphery and core of the grafts were analyzed separately. Collagen content only slightly decreases under immunosuppressive treatment, more if implanted matrix is free of cells and the most in combination of grafted cells and matrix. Due to muscle pouch implantation there is a compression force, which might lead to higher scores in the parameter of collagen content. D: Evaluation of the collagen architecture in vitro and its in vivo development, using blind observer evaluation. Observations on day of implantation (day 0) were used as references for in vivo stages of graft development following implantation. Values are displayed as percentages of reference values. Disarrangement and dissolution of the matrix structure are triggered more by grafted cells than by the collagen matrix itself; daily immunosuppressive treatment decreases these processes significantly.

Days 21 and 28, the remodeling processes progressed further. On Day 14, in Groups B and D, the collagen mesh was inhomogeneous. This disarray became more obvious in Group D than in Group B on Day 21 (p ⫽ 0.016). However, collagen architecture was not significantly altered in Group C between Days 14 and 28, nor in Group B between Days 21 and 28. Implants in Group D dissolved entirely after 28 days in vivo (Figure 5d). Infiltration directed from adjacent host tissue. Ingrowth of host cells was present in all groups, but differed in cell density and proportion. Group D was heavily infiltrated by macrophages, occasionally form-

ing giant cells, myocytes and lymphocytes; a homogeneous cell infiltration was seen after 21 days. Infiltration was slower in Group B, reaching full thickness on Day 28. In contrast, Group C had minimal infiltration on Day 14. Cell content increased until Day 28, but not as much as in Groups B or D (Figure 4c– h). Group D was more heavily infiltrated with lymphocytes than Groups B or C (p ⬍ 0.050 each, whole implants). Although cell content increased in Groups B and C, it remained stable in Group D over time. Similar findings were seen in zonal analysis (periphery vs core) (Figures 4a and b and 5b). In all groups, central regions were less infiltrated by lymphocytes than the periphery (p ⬍ 0.050 each).

1250

Mueller-Stahl et al.

Despite higher cell content in Group B, time course of lymphocyte infiltration was similar to Group C (Figure 5b). In addition, production of extracellular matrix was most abundant in Group D, followed by Group B, and finally Group C. These findings are consistent with cell infiltration from the host tissue into the grafts of Groups B, C and D. DISCUSSION Stable, contractile bioartificial myocardial (AMT) constructs can be engineered and successfully cultured over weeks in vitro,4,7,10 but there have been few reports quantifying cell density and proportion within these constructs. We found a gradient of cell density decreasing toward the core of the graft. Differences in seeding techniques, malnutrition and insufficient oxygen supply were likely causes of this gradient. Even using a porous matrix (pores mimic convective structures for supplying nutrients and oxygen in vitro), myofibril length was limited to 30 to 80 ␮m. Despite the non-physiologic cell composition (90% cardiomyocytes, 10% non-cardiac cells), our AMTs were contractile and formed a dense tissue, remaining stable over weeks.7 In vitro proliferation of non-cardiac cells is difficult to control. The application of an anti-proliferative agent, such as BrdU, stabilized cell culture conditions for this study. Both co-localization studies showed comparable results for non-myogenic cells on the day of isolation and on the day of AMT implantation, identifying BrdU as a powerful tool both for cell cycle control in vitro and for the detection of non-cardiac grafted cells in vivo. In vivo there was a continuous reduction in cardiac and non-cardiac cells over time. Immunosuppressive therapy (Group C) slowed this down sufficiently so that grafted cells remained viable until the end of the study. Group C also exhibited higher collagen content and better preservation of the collagen fiber architecture, emphasizing the need for immunosuppressive agents. Observations for maturation of cardiac cells (rodshaped with perinuclear free space, positive for orientated sacromericactin, connexin43 [Cx43], N-cadherin, and others) are controversial.3,4,10 In this study maturation of cardiac cells was only detected along the periphery, in the vicinity of vascular structures, and with immunosuppressive treatment. There was no significant inflammatory response in Group C to explain the loss of grafted cells. In Groups B and D (but not in Group C) neo-vascularization was eventually visible to the naked eye in situ. This suggests that the supply situation of the grafts is only sufficient in isolated focal sites after implantation. However, other researchers have observed vigorous ingrowth of capillaries into the grafts despite immunosuppression and even in syngeneic models.4,10

The Journal of Heart and Lung Transplantation November 2008

Despite interventions to promote angiogenesis (micospheres, gene transfer, implantation of endothelial cells, etc.), capillary density in experiments did not approach that of native heart vasculature.11–14 In today’s in vivo setups, angiogenesis driven by ischemia remains difficult to distinguish from that based on immunologic response.15 The qualitative assessment of vascular density (macro- or microscopic) is therefore inadequate to imply a sufficient supply of nutrition and oxygen. Comparative analysis of vascular structures relative to the vasculature of native heart tissue should be the assessment of choice. In conclusion, with respect to AMT, collagen fibers are of low, but non-specific immunogenicity.16 REFERENCES 1. Menasché P. Myoblasts-based cell transplantation. Heart Fail Rev 2003;8:221–7. 2. Wollert K, Meyer GP, Lotz J, et al. Intracoronary autologous bone-marrow cell transfer after myocardial infarction: the BOOST randomised controlled clinical trial. Lancet 2004;364:141– 8. 3. Reffelmann T, Leor J, Müller-Ehmsen D, Kedes L, Kloner RA. Cardiomyocyte transplantation into failing heart—new therapeutic approach for heart failure. Heart Fail Rev 2003;8:201–11. 4. Zimmermann WH, Didie M, Wasmeier GH, et al. Cardiac grafting of engineered heart tissue in syngenic rats. Circulation 2002; 106(suppl I):151–7. 5. Itescu S, Schuster MD, Kocher AA. New directions in strategies using cell therapy for heart disease. J Mol Med 2003;81:288 –96. 6. Kofidis T, Balsam LB, Robbins RC. A few critical aspects—and Achilles heels— of tissue engineering approaches to restore injured myocardium. J Thorac Cardiovasc Surg 2003;126:2113– 4. 7. Kofidis T, Akhyari P, Boublik J, et al. In vitro engineering of heart muscle: artificial myocardial tissue. J Thorac Cardiovasc Surg 2002;124:63–9. 8. Mueller-Stahl K, Kofidis T, Akhyari P, et al. Carboxyfluorescein diacetate succinimidyl ester facilitates cell tracing and colocalization studies in bioartificial organ engineering. Int J Artif Organs 2003;26:235– 40. 9. Korecky B, Hai CM, Rakusan K. Functional capillary density in normal and transplanted rat hearts. Can J Physiol Pharmacol 1982; 60:23–32. 10. Li RK, Jia ZQ, Weisel RD, et al. Survival and function of bioengineered cardiac grafts. Circulation 1999;100(suppl II):63–9. 11. Yau TM, Fung K, Weisel RD, et al. Enhanced myocardial angiogenesis by gene transfer with transplanted cells. Circulation 2001;104(suppl I):218 –222. 12. Sakikabira Y, Nishimura K, Tambara K, et al. Prevascularisation with gelatin microsperes containing basic fibroblast growth factor enhances the benefits of cardiomyocyte transplantation. J Thorac Cardiovasc Surg 2002;124:50 – 6. 13. Liekens S, De Clercq E, Neyts J. Angiogenesis: regulators and clinical applications. Biochem Pharmacol 2001;61:253–70. 14. Pepper MS. Role of the matrix metalloproteinase and plasminogen activator–plasmin systems in angiogenesis. Arterioscler Thromb Vasc Biol 2001;21:1104 –17. 15. Jackson JR, Seed MP, Kircher CH, Willoughby DA, Winkler JD. The codependence of angiogenesis and chronic inflammation. FASEB J 1997;11:457– 65. 16. Bailey AJ. The fate of collagen implants in tissue defects. Wound Rep Reg 2000;8:5–12.