Immune response to stem cells and strategies to induce tolerance.

Although recent progress in cardiovascular tissue engineering has generated great expectations for the exploitation of stem cells to restore cardiac form and function, the prospects of a common mass-produced cell resource for clinically viable engineered tissues and organs remain problematic. The refinement of stem cell culture protocols to increase induction of the cardiomyocyte phenotype and the assembly of transplantable vascularized tissue are areas of intense current research, but the problem of immune rejection of heterologous cell type poses perhaps the most significant hurdle to overcome. This article focuses on the potential advantages and problems encountered with various stem cell sources for reconstruction of the damaged or failing myocardium or heart valves and also discusses the need for integrating advances in developmental and stem cell biology, immunology and tissue engineering to achieve the full potential of cardiac tissue engineering. The ultimate goal is to produce 'off-the-shelf' cells and tissues capable of inducing specific immune tolerance.

Characterization of structural and signaling molecules by human valve interstitial cells and comparison to human mesenchymal stem cells.

BACKGROUND AND AIM OF THE STUDY: Human mesenchymal stem cells (MSCs) are a potential cell source for the tissue engineering of biological structures, including cardiac valves. A comprehensive, phenotypic analysis of MSCs and, for the latter, their comparison with valve interstitial cells (ICs) is therefore essential. METHODS: Isolates of bone marrow-derived human MSCs and human cardiac valve ICs were extensively phenotyped for their expression of membrane proteins involved in adhesion and cell-cell communication, cytoskeletal components, extracellular matrix (ECM) proteins and gene expression of WNT/FZD/SFRP/DKK/LRP family members. RESULTS: MSCs and valve ICs (>80%) expressed fibroblast surface antigen, smooth muscle alpha-actin, vimentin and CD44; expression of MHC class I and II and calponin was inconsistent, and a small proportion expressed desmin and smooth muscle myosin. CD105 was weakly expressed by a low percentage of valve ICs (<10%) compared to MSCs (>90%). ECM components made by both cell types demonstrated similar levels and patterns of staining, although expression of elastin was not detected by both cell types. Adhesion molecule expression was highly variable among the MSC isolates and between the two cell types, with the predominant integrins being alphal, alpha3, alpha5, and beta1 by both cell types. PCR analysis of WNT/FZD/SFRP/LRP family members revealed a greater range of the WNT family of genes being expressed in MSCs compared to ICs. CONCLUSION: The study results provided an extensive fingerprint of valve ICs and of MSCs for the tissue engineering of biological structures and for the manipulation of their desired phenotype. MSCs represent a promising cell type for valve tissue engineering, and will require extensive phenotyping after differentiation.

Human mesenchymal stem cells induce T cell anergy and downregulate T cell allo-responses via the TH2 pathway: relevance to tissue engineering human heart valves.

To generate an ''off the shelf'' tissue-engineered heart valve, the cells would need to be of allogeneic origin. Here, we report the possibility of using human bone marrow-derived mesenchymal stem cells (MSCs) as a suitable allogeneic cell source for tissue-engineered heart valves. Proliferative responses of primary and primed CD4+ T cells to allogeneic MSCs were examined. A protein microarray system was used to detect soluble factors from supernatants collected from the T cell assays. MSCs are poor stimulators of primary and primed CD4+ T cell proliferation, despite provision of B7-1 trans-co-stimulation. MSCs not only directly inhibited primary and primed T cell responses to allogeneic peripheral blood mononuclear cells (PBMCs), but 24-h pre-culture of T cells with MSCs suppressed subsequent T cell proliferative responses to allogeneic PBMCs in a contact-dependent manner. Analysis of supernatants revealed a distinctly different cytokine profile after co-culture of T cells with MSCs than with PBMCs or endothelial cells. Pro-inflammatory Th1 cytokines interleukin (IL)-1alpha and beta, interferon (IFN)gamma, and tumor necrosis factor (TNF)alpha were downregulated, whereas, anti-inflammatory Th2 cytokines IL-3, IL-5, IL-10, and IL-13 and the Th2 chemokine I-309, a chemoattractant for regulatory T cells, were upregulated. Further analysis revealed that after co-culture with MSCs, the T cells exhibited a regulatory phenotype (CD4+ CD25(lo) CD69(lo) FoxP3+). MSCs downregulate T cell responses through direct contact and secretion of anti-inflammatory and tolerogenic cytokines, which may involve the recruitment of regulatory T cells. This implies that allogeneic MSCs could be a suitable cell source for tissue engineering a heart valve.

Collagen synthesis by mesenchymal stem cells and aortic valve interstitial cells in response to mechanical stretch.

OBJECTIVE: The synthesis of appropriate extracellular matrix by cells in tissue engineered heart valve constructs will be important for the maintenance of valve cusp integrity and function. We have examined and compared the capacity of mesenchymal stem cells to synthesise collagen in response to stretch in comparison with native aortic valve interstitial cells. METHODS: Cells were stretched on a Flexercell FX4000 apparatus and total collagen synthesis was measured by the incorporation of [3H]-proline. The effect of stretch on gene expression of different collagen types was assessed by RT-PCR. RESULTS: There was a significant (p<0.01) increase in [3H]-proline incorporation into stretched valve cells at 10%, 14% and 20% stretch. The response of mesenchymal stem cells at 14% stretch was similar to that seen in the valve cells. Incorporation of [3H]-proline into soluble proteins in the cell media was significantly higher (p<0.01) only at 14% and 20% stretch in valve interstitial cells. These effects were shared with mesenchymal stem cells at 14% stretch. RT-PCR experiments demonstrated that 14% stretch up-regulated levels of mRNA for COL3A1 gene (type III collagen) but did not increase the expression of COL1A1 gene (type I collagen) in valve interstitial cells. However, both collagen genes could be detected in non-stretched and stretched mesenchymal stem cells. There was no evidence that the mesenchymal stem cells had started to adopt an osteoblastic cell phenotype in response to stretch. CONCLUSIONS: Collagen synthesis by valve interstitial cells is dependent upon the degree and duration of stretch. This response can be mimicked closely by exposure of mesenchymal stem cells to the same stretching profile. These properties could have important implications for the choice of cells and programme of conditioning with which to tissue engineer heart valves.

Magnetic resonance imaging of ferumoxide-labeled mesenchymal stem cells seeded on collagen scaffolds-relevance to tissue engineering.

Mesenchymal stem cells (MSCs) are a promising candidate cell for tissue engineering. Magnetic resonance imaging (MRI) has been proven effective in visualizing iron-labeled stem cells; however, the efficiency of this approach for visualization of cells seeded on scaffolds intended for use as tissue-engineered heart valves has not been assessed. MSCs were labeled by incubating for 48 h with ferumoxide and poly-L-lysine as transfecting agent. Any detrimental effect of iron labeling on cell viability, proliferation, and differentiation was examined using appropriate functional assays. Change in the nuclear magnetic relaxation properties of labeled cells was determined using in vitro relaxometry of cells seeded in 3-dimensional collagen gels. Images of labeled and non-labeled cells seeded onto 1% type I bovine collagen scaffolds were obtained using MRI. The presence of intracellular iron in labeled cells was demonstrated using Prussian blue staining, confocal microscopy, and electron microscopy. Cell viability, proliferation, and differentiation were comparable in labeled and non-labeled cells. The T2 relaxation time was 40% to 50% shorter in ferumoxide-labeled cells. Labeled cells seeded on scaffolds appeared as areas of reduced signal intensity in T2 weighted images. Ferumoxide labeling persisted and remained effective even on scans performed 4 weeks after the labeling procedure. Ferumoxide labeling of human MSCs seeded on collagen scaffolds is an effective, non-toxic technique for visualization of these cells using MRI. This technique appears promising for cell tracking in future tissue-engineering applications.

Loss of ecto-5'nucleotidase from porcine endothelial cells after exposure to human blood: Implications for xenotransplantation.

The endothelial cell surface expression of ecto-5'-nucleotidase (E5'N, CD73) is thought to be essential for the extracellular formation of cytoprotective, anti-thrombotic and immunosuppressive adenosine. Decreased E5'N activity may play a role in xenograft acute vascular rejection, preventing accommodation and tolerance mechanisms. We investigated the extent of changes in E5'N activity and other enzymes of purine metabolism in porcine hearts or endothelial cells when exposed to human blood or plasma and studied the role of humoral immunity in this context. Pig hearts, wild type (WT, n = 6) and transgenic (T, n = 5) for human decay accelerating factor (hDAF), were perfused ex vivo with fresh human blood for 4 h. Pig aortic endothelial cells (PAEC) were exposed for 3 h to autologous porcine plasma (PP), normal (NHP) or heat inactivated human plasma (HHP), with and without C1-inhibitor. Enzyme activities were measured in heart or endothelial cell homogenates with an HPLC based procedure. The baseline activity of E5'N in WT and T porcine hearts were 6.60 +/- 0.33 nmol/min/mg protein and 8.54 +/- 2.10 nmol/min/mg protein respectively (P < 0.01). Ex vivo perfusion of pig hearts with fresh human blood for 4 h resulted in a decrease in E5'N activity to 4.01 +/- 0.32 and 4.52 +/- 0.52 nmol/min/mg protein (P < 0.001) in WT and T hearts respectively, despite attenuation of hyperacute rejection in transgenic pigs. The initial PAEC activity of E5'N was 9.10 +/- 1.40 nmol/min/mg protein. Activity decreased to 6.76 +/- 0.57 and 4.58 +/- 0.47 nmol/min/mg protein (P < 0.01) after 3 h exposure of HHP and NHP respectively (P < 0.05), whereas it remained unchanged at 9.62 +/- 0.88 nmol/min/mg protein when incubated with PP controls. C1-inhibitor partially preserved E5'N activity, similar to the effect of HHP. Adenosine deaminase, adenosine kinase and AMP deaminase (other enzymes of purine metabolism) showed a downward trend in activity, but none were statistically significant. We demonstrate a specific decrease in E5'N activity in pig hearts following exposure to human blood which impairs adenosine production resulting in a loss of a cytoprotective phenotype, contributing to xenograft rejection. This effect is triggered by human humoral immune responses, and complement contributes but does not fully mediate E5'N depletion.

The cardiac valve interstitial cell.

Cardiac valve interstitial cells (ICs) are a heterogeneous and dynamic population of specific cell types that have many unique characteristics. They are responsible for maintaining the extracellular scaffold that provides the mechanical characteristics vital for sustaining the unique dynamic behaviour of the valve. A number of cellular phenotypes can be distinguished: some are sparsely arranged throughout the valve leaflets, whilst others are arranged in thin bundles. These cells express molecular markers similar to those of skeletal, cardiac and smooth muscle cells (SMCs) and in particular, many ICs express smooth muscle (SM) alpha-actin, a marker of myofibroblasts. In this respect, these cells exhibit a profile unlike skin fibroblasts, which may allude to their role in valve function.