Heart valve defects often necessitate the use of mechanical or tissue valve replacements. These substitutes have a limited durability which, especially in pediatric patients, requires frequent replacements.1–4 To overcome limitations of the currently used prostheses, tissue engineering of heart valves might be a future option for therapeutic approaches. The strategy of tissue engineering includes the use of autologous cells in order to finally fabricate tissue with the ability to integrate into the surrounding tissue to grow and to remodel.
The successful construction of living, autologous cardiac valve implants necessitates cells that resemble native cellular structures found in cardiac valves in phenotype and functionality. Human heart valve tissue is assembled mainly from two cell types: endothelial and interstitial cells. The interstitial cells are fibroblast as well as myofibroblast types of cells. They are thought to produce and maintain the complex extracellular matrix (ECM) that provides the valve with the necessary physical strength and flexibility.5 The cellular layer that covers matrix and interstitial cells is composed of endothelial cells that provide multiple functions including the prevention of thrombus formation and serve as a selective barrier. The recreation of these cellular structures is one of the aims in tissue engineering. Numerous cell sources were proven feasible for the fabrication of heart valves as the identification of the optimal cell sources is still under investigation in numerous groups. Current efforts to develop tissue engineered heart valves utilize autologous cells commonly isolated from the saphenous vein, umbilical cord or from umbilical cord blood.6,7 Also bone marrow-derived cells,8 as well as cryopreserved umbilical cord cells were successfully applied to create tissue engineered heart valves.9 Most of the authors focus on valvular prostheses for the right ventricular outflow tract (low pressure system) in pediatric surgery.
The aim of the present experiment was to evaluate and to compare different cell sources with cultured cells of native pulmonary valve. Vascular myofibroblasts isolated from saphenous veins (HVM) and myofibroblasts isolated from the umbilical artery (HUAM) were examined as comparable to pulmonary valve interstitial cells (HPVIC). Human pulmonary heart valve endothelial cells (HPVEC) were compared to human vascular endothelial cells derived from the vena saphena magna (HVEC) and to human umbilical vein endothelial cells (HUVEC), two cell sources previously used for graft endothelization.10,11 While ongoing in vitro and in vivo experiments seem promising,12–14 a thorough understanding of differences in morphology and gene expression patterns of these cells and of the cells of the pulmonary heart valve is lacking. Consequently, we inspected the expression levels of marker genes in the chosen cell sources in morphological, immunohistochemical and quantitative real-time polymerase chain reaction (QPCR) studies to determine their properties similar to cells of the pulmonary valve. Marker genes were selected based either on their described prominent expression and functional role in endothelia15–21 or on a fibroblast-specific expression and their participation in the ECM which is critical for structure and function of the valve.5,21,22
Materials and Methods
Anonymous use of human primary material for this experimental study was approved by the local ethics committee.
Isolation of HPVEC and HPVIC
After heart transplantation, the replaced heart was stored in M199 medium (500 ml) plus antibiotic cocktail (ciprofloxacin: 3 mg/100 ml, amikacin: 2 mg/100 ml, vankomycin: 2 mg/100 ml, metronidazol: 2 mg/100 ml, amphotericin B: 0.5 mg/100 ml, all from Sigma, St. Louis, MO) at 4°C. Up to 48 h later, the leaflet of the pulmonary valve was cut out and cells were isolated by enzymatic digestion according to the following procedure: after a brief wash in phosphate buffered saline (PBS), HPVEC were isolated by digestion in 0.1% collagenase II (Worthington Biochemical Corporation, Lakewood, NJ) for 5 minutes at 37°C. The enzymatic reaction was stopped with M199 with 20% FCS. A second 1 h digestion was performed to isolate HPVIC. Both cell types were centrifuged at 1,000 rpm for 10 minutes. HPVEC and HPVIC were resuspended in EC-M and SMC-M (Promocell, Heidelberg, Germany), respectively. The cell suspensions were plated out in T25 flasks.
Isolation of HUVEC and HUAM
The umbilical cord was stored after delivery in M199 medium at 4°C for up to 48 h. To isolate HUVEC cells, the umbilical cord vein was cannulated, rinsed with buffered medium, and incubated with 0.1% collagenase II for 20 min at 37°C. After centrifugation, cells were resuspended in EC-M and plated out in a T25 culture flask.
To isolate HUAM cells, a small section of one umbilical cord artery was isolated and cut into approximately 1 × 1 mm sections. These were placed in a 35 mm Falcon dish (Becton Dickinson, Plymouth, UK) and covered with a glass cover slip, attached to the bottom of the culture dish with three silicon points. After 1 week, myofibroblast cells started to grow out of the sections.
Isolation of HVEC and HVM
Vein sections were retrieved from routine aortocoronary bypass operations and stored in M199 medium at 4°C for up to 48 h. HVEC cells were isolated in the same procedure as described for the umbilical cord. For the isolation of HVM, the vein section was filled with 0.1% collagenase II and incubated 2 times for 30 minutes each. Cells were centrifuged at 1,000 rpm for 10 min, resuspended in SMC-M, and plated out in a T25 flask.
All the cells used for the immunocytochemistry were at passage four. Cells were plated into eight-well chamber slides (Thermo Fisher, Waltham, MA) for 48 h. Cells were washed with PBS, fixed with 96% Ethanol and stored at −80°C until the staining procedure. Before staining, the chamber slides were fixed 2 min with acetone. Staining was performed by use of the avidin-biotin-peroxidase technique (DAKO, Glostrup, Denmark).
All endothelial cells were stained with antibodies against von Willebrand factor (vWF) (rabbit antihuman, DAKO), CD31 (mouse antihuman, DAKO), and CD90 (mouse antihuman, Dianova, Hamburg, Germany).
HPVIC, HUAM, and HVM were stained with antibodies against CD31 (mouse antihuman, DAKO), CD90 (mouse antihuman, Dianova), smooth muscle α-actin (αSMA) (mouse antihuman, DAKO), smooth muscle myosin (SMM) (mouse antihuman, DAKO), collagen I (mouse anticow, Sigma), collagen III, collagen IV (both mouse antihuman, Sigma), and fibronectin (rabbit antihuman, Sigma).
Qualitative Inspection of Immunostaining
Visual inspection was independently conducted by three researchers. If more than 90% of all observed cells were positively stained for an intracellular marker, the expression was considered strong and classified as +++. Observation of 50% to 80% stained cells was assigned as moderate expression ++, 10% to 40% as a weak expression level with +. Very weak expression strength with approximately 2% of the cells stained positively was termed ±. For extracellular proteins, we referred to the plane two-dimensional spread of the immunocytochemical stain within the cell-covered area to assess the level of expression.
Quantitative Real-time Polymerase Chain Reaction (PCR)
All the cells used for the RNA isolation were at passage four or five. Total RNA was isolated using Trizol (Invitrogen, Carlsbad, CA). RNA was treated with TURBO DNAse (Applied Biosystems, Foster City, CA) to degrade genomic DNA, and quantitated by UV absorbance at 260 nm. Complementary cDNA was synthesized using 2 μg of eluted RNA, random hexamer primers and the revertAid™ H Minus M-MuLV first strand cDNA Synthesis Kit (Fermentas, Burlington, Canada). Primers and probes for real-time quantitative PCR were obtained from Applied Biosystems TaqMan Gene Expression Assay catalogue. These assays come in a 20X reaction mix and are optimized to give approximately 100% efficiency. For the QPCR comparison of HPVEC, HVEC, and HUVEC, five different pairs of primers were used: eNOS (90 bp, Hs00167166_m1), vWF (79 bp, Hs00169795_m1), GPR3 (83 bp, Hs00270991_s1), KCNN4 (148 bp, Hs00158470_m1), and ICAM-1 (138 bp, Hs00277001_m1). For the QPCR analysis of HPVIC, HVM, and HUAM, the five different pairs of primers selected were: Collagen IV (78 bp, Hs00166712_m1), Elastin (71 bp, Hs00355783_m1), Laminin β2 (64 bp, Hs00158642_m1), and Vitronectin (77 bp, Hs00169863_m1). Housekeeping gene GAPDH (122 bp, Hs33679905_m1) was selected as control. QPCR was performed in duplicates using the Applied Biosystems PRISMa 7700 system and Goldstar TaqMan Universal Master Mix, (Eurogentec, Brussels, Belgium). Cycling conditions included an initial denaturation step of 10 min at 94°C, followed by 40 cycles of 15 s of denaturation at 94°C and 60 s of annealing and extension step at 60°C. Also negative controls without template cDNA were used.
Using the relative quantitative reverse transcription (RT)-PCR method, the cycle thresholds of the genes of interest were compared to the housekeeping gene to determine relative changes in expression. We used the ΔCT target−ΔCTcontrol (ΔΔCT) method of detecting differences in the threshold cycles between the target and control genes (ΔCT target = CT target−CT housekeeping gene; ΔCTcontrol = CT control−CT housekeeping gene). The formula 2(ΔΔCT) was used to calculate the fold change. The Mann-Whitney test was performed, considering a p value < 0.05 as statistically significant.
Inspection of Morphology
Human umbilical vein endothelium proves a rich source for large amounts of cells. HVECs, as well as HVM cells are more aged and yield lesser numbers of passages. Endothelial cells in general look similar and robust. The endothelial cells isolated from the saphena magna as well as from the umbilical cord form a round firm cobblestone pattern that is more tightly packed than the elongated firm cobblestone pattern of cells isolated from the pulmonary valve (Table 1, Figure 1). Pulmonary valve interstitial cells, vein myofibroblasts, and myofibroblasts from umbilical cord all show elongated cell bodies typical for myofibroblast cells. In HUAM cultures, cells are filamentous and adjoined (Table 2, Figure 1).
Comparison of Immunocytochemical Markers
Endothelial cells were investigated with antibodies for vWF, CD31, and CD90. While staining for vWF is ubiquitously detected in the three types of cultures, its cellular localization differs: It is predominantly cytoplasmic in HPVEC and HVEC whereas it appears to be membrane bound in HUVEC (Figure 2, a, b, c and Table 1). The same difference with regard to its localization is found for CD31. HPVEC and HVEC show cytoplasmic localization of CD31, whereas HUVEC display membrane bound CD31 (Figure 2, d, e, and f). Endothelial cells of all sources stain negative for CD90 (Figure 2, g, h, i). Myofibroblast cells isolated of heart valve, vein and umbilical cord were compared morphologically (Table 2) and probed with antibodies to CD31, CD90, αSMA, SMM, collagen I, III, and IV, as well as fibronectin (Figure 3 and Table 1). Staining for the endothelial cell surface marker CD31 is, as expected, negative for all myofibroblast cells. The fibroblast specific marker CD90, however, is expressed in all three myofibroblast cell types in a uniform cytoplasmic pattern. Surprisingly, αSMA is barely expressed in HPVICs; it is expressed in a few isolated cells in HVMs, and uniformly in HUAMs. Smooth muscle myosin is present in clusters of HPVICs, in single enlarged HVMs and barely in HUAM cells. Collagen I appears (similar to the smooth muscle myosin expression) in groups of HPVIC, on very few HVM, and on approximately 45% of HUAM cells. Collagen III is detected on single cells in HPVIC cultures, not at all on HVM, but on about half of HUAM; however, collagen III in HPVICs decreases after passage 2 (data not shown). Collagen IV is located on about 80% of HUAM cells (u in Figure 3), albeit at lower levels on about 20% of HPVICs and on about 40% of HVMs (s, t, in Figure 3). Fibronectin is expressed at high levels by all cells of all types of myofibroblasts (v, w, x in Figure 3).
Comparison of Marker Gene Expression via QPCR
To gain deeper insights into the heterogeneity of the different cells, the expressions of some significant genes were analyzed via quantitative real-time PCR. The expression levels of the endothelial signaling factors eNOS and GPR3 are not significantly different in cells of different sources (Figure 4A). KCNN4 is expressed more than 100-fold more abundantly in HVEC than in HPVEC or HUVEC cultures and the difference is significant. The slight four-fold down-regulation of ICAM transcripts in HUVECs as detected via QPCR is not statistically significant. HVEC display a significant three-fold upregulation of vWF compared with HPVEC and HUVEC cells. Note that in immunocytochemical stainings, all three cell sources are characterized by a strong expression of vWF.
In the group of the HPVICs, relative expression levels of factors involved in the formation of the ECM was analyzed. There is a discreet raise of collagen IV expression by HUAM and HVM cells that is not statistically significant when compared with HPVICs. In the immunocytochemical analysis of HUAM cells this could be observed, too. Expression of elastin is slightly but not significantly reduced in HUAM cells. Laminin levels are higher in HVM cells than in HUAM cells. The difference is statistically significant albeit without distinction to the HPVIC cells aimed for. Though overall relative transcript levels of laminin are the highest detected. Vitronectin does not differ in myofibroblast cells from pulmonary valve, from umbilical cord, and from the vena saphena magna (Figure 4B).
Tissue engineering aims to overcome the limitations of currently used prostheses by reengineering heart valve constructs that resemble native tissue including the ability to grow and to remodel. This strategy worked well in certain in vitro and in vivo experiments. However, the optimal cell source is a crucial point and one of the major challenges to be addressed for the final success of tissue engineering of heart valves. In previous investigations, the producibility of tissue engineered heart valves using peripheral vascular cells,23,24 umbilical cord cells,25,26 and amniotic cells27 has been demonstrated, and with all these cell sources, promising tissue engineered valvular constructs were obtained. Based on our results with peripheral vascular cells and umbilical cord vascular cells, we speculate that all cell sources can potentially be used for tissue engineering of heart valves, and there is no actually tested cell source completely identical with the native human pulmonary valvular cells.
For valve tissue engineering applications, one critical requisite will be to isolate defined phenotypes of endothelial and myofibroblast cells so that differentiation can be controlled during the tissue development in vitro. However, the phenotypic characterization of cell sources for tissue engineering of heart valves has often been controversial.28
In previous investigations, the different phenotypes of endothelial cells from different cell sources were proven, e.g., peripheral vascular endothelial cells and vascular endothelial cells.4,28,29 In our study, endothelial cells isolated from saphenous vein and from umbilical vein were compared morphologically and based on marker gene expression cells to endothelial cells of the pulmonary valve. The relevant factors are expressed by all the inspected endothelial cells. We show that endothelial cells isolated from saphenous vein, umbilical cord, vein, and pulmonary valve have no significant difference in the expression of eNOS, GPR3, and ICAM-1. Furthermore our experiment demonstrates almost no difference of morphological characteristics of HVEC and HUVEC compared to HPVEC and all endothelial cell cultures stained positive for CD31 and vWF. Endothelial cells from saphenous vein showed significant higher expression of vWF and KCNN4. These findings suggest the difference of peripheral vascular endothelial cells compared to umbilical cord endothelial cells and pulmonary valve endothelial cells.
For tissue engineering of cardiovascular structures using biodegradable scaffold materials, it is of upmost interest to generate ECM that leads to the formation of functional cardiovascular tissue. Therefore, the presence of myofibroblasts and the formation of an intact ECM in tissue engineered heart valve constructs is an important factor and was described by previous investigators.4,29,30 In our study, we compared the expression of markers, such as CD90, αSMA, SMM, collagen I, collagen III, collagen IV, elastin, laminin and vitronectin in peripheral vein and umbilical cord vein myofibroblasts with human pulmonary valve interstitial myofibroblast cells. PCR analysis of all myofibroblast cell sources shows similar characteristics compared to human pulmonary valve cells. In immunostainings of the myofibroblast sources considered for tissue engineering of heart valves, HVM cells are characterized by a lack of collagen III expression. Regarding the αSMA expression, HVM cells more closely resemble HPVICs than HUAM cells. In this context, the more filamentous and elongated morphology of the HUAM cells is plausible. Moreover, results from immunocytochemistry indicate that production of collagen IV by HUAM cells is stronger than by HPVIC as well as HVM cells. Contrarily, earlier studies had shown more elevated collagen IV levels in native pulmonary valve interstitial cells, notwithstanding the interstitium’s layered structuring.22 Overall, expression levels of both HVM and HUAM cells are similar to HPVICs. The found differences regarding collagen III/IV, and αSMA truely give no clear indication that either vein myofibroblasts or umbilical arterial myofibroblasts are superior in replacing valve interstitial cells in tissue engineered heart valve constructs, but might, if applied, prove beneficial in the course of tissue engineering of heart valves.
Three main factors are well known to be of general importance in the discipline of tissue engineering. Firstly, the applied cell source has to be capable of a replication of cellular structures and functions of the targeted tissue. Currently we are investigating additional alternative cell sources from umbilical cord blood or cord tissue that could provide the comparable phenotypic criteria to pulmonary valve endothelial as well as myofibroblast cells. Currently, also the possibility of differentiating cells from a single cell source is a subject of investigation. Secondly, the used scaffolds must be biocompatible and at the same time provide well-defined shaping. Biodegradable scaffolds should be gradually resorbed to bring out the functional tissue. Thirdly, an in vitro environment that ultimately elicits a proper tissue formation has to be set up under sterile conditions. As cells are likely to adapt and to undergo morphological as well as expression profile changes once placed in the appropriate heart environment, future experiments will further characterize the examined cell sources in vitro after bioreactor conditioning runs as well as in vivo in an animal model.
In summary, our results support the potential of using human peripheral vascular cells and umbilical cord vascular cells to fabricate tissue engineered heart valves. For infants, human umbilical cord cells may be isolated and used for cardiovascular tissue engineering applications. Further efforts are required to establish cell culture techniques to differentiate umbilical cord blood progenitor cells into endothelial and myofibroblastic cells for tissue engineering of heart valves. Therefore, it might be desirable to establish autologous human cell banks for pediatric patients, who are likely to require heart valve replacements in the early years of life.
This work was supported by grant no. 01GN0544 from the German Federal Ministry for Education and Research (BMBF). The authors thank Crispin A, Laubender R, Schotten KJ, Unlig A, and Akra B for their support.
1.Flanagan TC, Pandit A: Living artificial heart valve alternatives: A review. Eur Cell Mater
6: 28–45, 2003.
2.Schmidt D, Hoerstrup SP: Tissue engineered heart valves based on human cells. Swiss Med Wkly
135: 618–623, 2005.
3.Mayer J: Uses of homograft conduits for right ventricle to pulmonary artery connections in the neonatal period. Sem Thorac Cardiovasc Surg
7: 130–132, 1995.
4.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
47: 439–465, 1999.
5.Taylor PM, Batten P, Brand NJ, et al
: The cardiac valve interstitial cell. Int J Biochem Cell Biol
35: 113–118, 2003.
6.Brody S, Pandit S: Approaches to heart valve tissue engineering scaffold design. J Biomed Mat Res
83: 16–43, 2007.
7.Schmidt D, Hoerstrup SP: Tissue engineered heart valves based on human cells. Swiss Med Wkly
135: 618–623, 2005.
8.Schmidt D, Achermann J, Odermatt B, et al
: Prenatally fabricated autologous human living heart valves based on amniotic fluid derived progenitor cells as single cell source. Circulation
116: I64–I70, 2007.
9.Sodian R, Lueders C, Kraemer L, et al
: Tissue engineering of autologous human heart valves using cryopreserved vascular umbilical cord cells. Ann Thorac Surg
81: 2207–2216, 2006.
10.Bader A, Schilling T, Teebken OE, et al
: Tissue engineering of heart valves–human endothelial cell seeding of detergent acellularized porcine valves. Eur J Cardiothorac Surg
14: 279–284, 1998.
11.Kasimir MT, Weigel G, Sharma J, et al
: The decellularized porcine heart valve matrix in tissue engineering: Platelet adhesion and activation. Thromb Haemost
94: 562–567, 2005.
12.Hoerstrup SP, Kadner A, Breymann C, et al
: Living, autologous pulmonary artery conduits tissue engineered from human umbilical cord cells. Ann Thorac Surg
74: 46–52, 2002.
13.Hoerstrup SP, Zünd G, Cheng S, et al
: A new approach to completely autologous cardiovascular tissue in humans. ASAIO J
48: 234–238, 2002.
14.van Wachem PB, Reinders JH, van Buul-Wortelboer MF, et al
: Von Willebrand factor in cultured human vascular endothelial cells from adult and umbilical cord arteries and veins. Thromb Haemost
56: 189–192, 1986.
15.Desjardins F, Balligand JL: Nitric oxide-dependent endothelial function and cardiovascular disease. Acta Clin Belg
61: 326–334, 2006.
16.Muro S, Muzykantov VR: Targeting of antioxidant and anti-thrombotic drugs to endothelial cell adhesion molecules. Curr Pharm Des
11: 2383–2401, 2005.
17.Ruggeri ZM: The role of von Willebrand factor in thrombus formation. Thromb Res
120 (Suppl 1): S5–S9, 2007.
18.Uhlenbrock K, Huber J, Ardati A, et al
: Fluid shear stress differentially regulates gpr3, gpr6, and gpr12 expression in human umbilical vein endothelial cells. Cell Physiol Biochem
13: 75–84, 2003.
19.Si H, Heyken WT, Wölfle SE, et al
: Impaired endothelium-derived hyperpolarizing factor-mediated dilations and increased blood pressure in mice deficient of the intermediate-conductance Ca2+ activated K+ channel. Circ Res
99: 537–544, 2006.
20.Grgic I, Eichler I, Heinau P, et al
: Selective blockade of the intermediate-conductance Ca2+-activated K+ channel suppresses proliferation of microvascular and macrovascular endothelial cells and angiogenesis in vivo. Arterioscler Thromb Vasc Biol
25: 704–709, 2005.
21.Latif N, Sarathchandra P, Taylor PM, et al
: Localization and pattern of expression of extracellular matrix components in human heart valves. J Heart Valve Dis
14: 218–227, 2005.
22.Butcher JT, Simmons CA, Warnock JN: Mechanobiology of the aortic heart valve. J Heart Valve Dis
17: 62–73, 2008.
23.Shinoka T, Shum-Tim D, Ma PX, et al
: Tissue-engineered heart valve leaflets: Does cell origin affect outcome? Circulation
96: II102–II107, 1997.
24.Sodian R, Hoerstrup SP, Sperling JS, et al
: Evaluation of biodegradable, three-dimensional matrices for tissue engineering of heart valves. ASAIO
46: 107–110, 2000.
25.Schmidt D, Mol A, Breymann C, et al
: Living autologous heart valves engineered from human prenatally harvested progenitors. Circulation
114: I125–II31, 2006.
26.Allen J, Khan S, Serrano MC, Ameer G: Characterization of Porcine Circulating Progenitor Cells: Toward a Functional Endothelium. Tissue Eng
14: 183–194, 2008.
27.Flanagan TC, Black A, O’Brien M, et al
: Reference models for mitral valve tissue engineering based on valve cell phenotype and extracellular matrix analysis. Cells Tissues Organs
183: 12–23, 2006.
28.Kirkpatrick CJ, Otto M, Van Kooten T, et al
: Endothelial cell cultures as a tool in biomaterial research. J Mater Sci Mater Med
10: 589–594, 1999.
29.Rabkin E, Hoerstrup SP, Aikawa M, et al
: Evolution of cell phenotype and extracellular matrix in tissue-engineered heart valves during in-vitro maturation and in-vivo remodeling. J Heart Valve Dis
11: 308–314, 2002.
30.Neidert MR, Tranquillo RT: Tissue-engineered valves with commissural alignment. Tissue Eng
12: 891–903, 2006.