OUP user menu

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

Ching-Hsin Ku , Philip H. Johnson , Puspa Batten , Padmini Sarathchandra , Rachel C. Chambers , Patricia M. Taylor , Magdi H. Yacoub , Adrian H. Chester
DOI: http://dx.doi.org/10.1016/j.cardiores.2006.03.022 548-556 First published online: 1 August 2006


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.

  • Heart valve
  • Mechanotransduction
  • Extracellular matrix
  • Tissue engineering
  • Interstitial cells

1. Introduction

Replication of cellular function is a key goal in the quest to tissue engineer a heart valve. The maintenance of aortic valve cusp integrity and function is critically dependent on the responses of the extracellular matrix (ECM) and valve interstitial cells to the physical forces to which they are exposed. These forces can vary widely during the cardiac cycle and initiate specific movements of different regions of the aortic valve [1,2]. The responsiveness of any cell type, that may be considered for tissue engineering, to the mechanical environment experienced by heart valves will be a key requirement for their ability to mimic the function of the cells that reside within the cusp tissue.

In this respect, mesenchymal stem cells form an important source of cells that, under the correct mechanical and ECM environment, may be directed to differentiate into a phenotype that resembles and functions as a valve interstitial cell. Mesenchymal stem cells can be isolated from bone marrow and fulfil the criteria of stem cells in that they are able to undergo self-renewal, and differentiate into other cell types with a different tissue origin, phenotype and function [3].

Cells respond to changes in their mechanical environment in a number of different ways [4,5]. Responses may include an altered pattern of gene expression, changes in the release of growth factors and vasoactive agents, and remodelling of the ECM. The specific ECM secretory response or kinetics of collagen production and other ECM components of valve interstitial cells in response to varying degrees of stretch has not been addressed. Defining these properties of valve interstitial cells will be a significant step in understanding valve function and developing strategies for tissue engineering heart valves. In this study we have examined the effect of varying degree and duration of stretch, on native valve interstitial cells and mesenchymal stem cells, a potential source of cells for a construct, with regard to the pattern of gene expression, synthesis and deposition of collagen.

2 Methods

The study received ethical permission from the Royal Brompton and Harefield NHS Trust Ethical Committee, that gave consent for the use of bone marrow samples for research, that would otherwise be disposed of. The study complied with the Declaration of Helsinki [6].

2.1 Valve interstitial cell culture

Interstitial cells were isolated from porcine aortic valve leaflets by enzymatic digestion as described previously by our group [7]. Porcine hearts from juvenile pigs were obtained from a slaughter house (Cheal Meats, Chelmsford, UK) and were 85–92 kg is size. In brief, isolated valve leaflets were placed in a solution of Dulbecco's Modified Eagle Medium (DMEM) containing 1000 U/ml of collagenase type II, agitated initially for 5 min to remove the endothelial cells, and then after a change of media for a further 45 min at 37 °C. The reaction was stopped by the addition of heat-inactivated foetal calf serum (FCS). Cells were then washed and seeded on to 6-well plates, where they were left in DMEM supplemented with 100 μg ml− 1 penicillin, 100 U ml− 1 streptomycin, 4 mM l-alanyl-l-glutamine and 10% heat-inactivated FCS culture media to adhere and grow to confluence for 3–7 days. Prior to the experiment, cells were stained with a panel of antibodies that included mouse monoclonal antibodies against smooth muscle α-actin, smooth muscle myosin heavy chain and β-tubulin (Sigma, Poole, UK), CD31, prolyl 4-hydroxylase, vimentin and desmin (Dako, Cambridge, UK). We have previously compared the phenotype of cells in the native human valve leaflet and cultured valve interstitial cells [7,8]. In situ, almost all interstitial cells expressed vimentin, many expressed fibroblast surface antigen, smooth muscle α-actin and smooth muscle myosin, but only a few expressed prolyl 4-hydroxylase and desmin. When cultured and seeded on glass coverslips, fewer cells expressed smooth muscle myosin and more cells expressed prolyl 4-hydroxylase and the fibroblast surface antigen. Cells were used for experiments between passages 3 and 7.

2.2 Mesenchymal stem cell culture

Mesenchymal stem cells were isolated from human bone marrow from healthy human donors aged 8–46 years that were no longer required for use in bone marrow transplants. Stem cells were cultured as previously described [9]. Briefly, heparinized bone marrow was mixed with an equal volume of phosphate buffered saline (PBS) and centrifuged at 900 g for 10 min at room temperature. Cells were washed and re-suspended in PBS to a final density of 4 × 107 cells/ml. A 5 ml aliquot of cells was then layered over a 1.073 g/ml Percoll solution (Amersham) and centrifuged at 1100 g for 30 min at 20 °C. Mononuclear cells collecting at the interface were recovered and washed in PBS at 900 g. The pellet was re-suspended in low glucose DMEM containing 10% FCS (pre-selected for MSC growth), 100 μg ml− 1 penicillin, 100 U ml− 1 streptomycin, 4 mM l-alanyl-l-glutamine and plated at a density of 2 × 105 cells/cm2.

When the cells had reached passage three, they were routinely characterised by immunocytochemistry as well as flow cytometry for a variety of intracellular and cell surface antigens respectively. Mesenchymal stem cell isolates that exhibited positive expression of CD105 (SH2), CD29, CD44, in the absence of CD45, MHC class II, CD31, CD14, CD34, B7 and vWF expression were chosen for further experiments, including their capacity to differentiate into osteogenic, adipogenic and chondrogenic lineages. Cells were used for experiments between passages 3 and 7.

2.3 Application of cyclic stretch to cultured cells

Cells were plated out at 2.5–3.0 × 105 cells/well into BioFlex culture plates (Dunn LabTech., Germany) pre-coated with collagen type I in culture media and allowed to reach confluence over a 3 day period. In preliminary experiments it was found that adhesion to the substratum was only obtained with type I collagen and fibronectin-coated membranes, but not with those coated with elastin, laminin or amino groups. Initial experiments also demonstrated that the collagen coated membranes did not interfere with the collagen assays (data not shown).

The cells were mechanically loaded on the Flexercell FX 4000T cell-straining device, in which negative pressure is applied beneath the culture plates via a vacuum pump and monitored by a pressure transducer, which permits precisely defined, multi-radial uniform stretch of up to 30%. The stretch can be applied with any waveform and frequency across the cell culture membrane. In the present experiments, cells were subjected to square waveforms at 0.6 Hz at forces that generated a 7%, 10%, 14% and 20% stretch. Cells were stretched cyclically for up to 4 days and each stretch experiment was carried out a minimum of three times. Cells grown on BioFlex plates in an identical manner to that described above, but not stretched, acted as the control (non-stretched) group. Upon completion of the stretching protocol, or at designated time points during it, samples of media were removed from the culture plates and stored at − 80 °C. Throughout the duration of the stretching protocols there was no measurable change in either the cell number or the incidence of cell death (data not shown), indicating that subsequent changes in levels of collagen were not related to changes in the numbers of cells in each well.

2.4 Measurement of collagen by quantification of total [3H]-proline

As the major biosynthetic destination of proline is collagen, the incorporation of the amino acid into newly synthesised proteins provides a reliable index of collagen synthesis. The amounts of radioactivity incorporated into soluble and into insoluble protein fractions provide an assessment of collagen released into the media (soluble protein), and that released and then subsequently incorporated in and around the cell layer (insoluble protein) [10].

For the [3H]-proline experiments, culture media containing 3.55 MBq/ml [3H]-proline was added to each well. After the appropriate experimental period, media were removed from the wells. Trichloroacetic acid (TCA) was then added to the media to give a final concentration of 10%, and left on ice for one hour. Precipitated protein was collected by centrifugation at 3000 g for 30 min, washed with 4 ml ice-cold 10% TCA to remove any unincorporated labelled proline and centrifuged again. The supernatant was carefully removed and the pellet suspended in 0.3 ml of 0.3 M NaOH/0.3% SDS. Preparations were then warmed to 37 °C until solubilised and added to 4 ml of liquid scintillant. The cell layer was washed twice with PBS and removed from Flexercell plates by scraping into 1 ml of ice-cold 10% TCA. Precipitated protein was collected by centrifugation at 14,000 g for 20 min. The cell layer precipitate was solubilised at 37 °C for one hour in 0.3 ml of 0.3 M NaOH/0.3% SDS and added to 4 ml of liquid scintillant. Radioactivity was counted in both the media and cell layer samples on a Packard Tricarb 1600 TR liquid scintillation analyser (Parkard, UK).

2.5 Response of collagen gene expression by cells to stretch

Estimates were made of genes involved in collagen synthesis and processing. Expression of mRNA for collagen type I, II and III (COL1A1, COL2A1, COL3A1) and lysyl oxidase (LOX) genes was determined in aortic valve interstitial cells and human mesenchymal stem cells after 14% stretch for 2 days. Intron-spanning oligonucleotide primers were designed using Primer3 (http://www.basic.nwu.edu/biotools/Primer3.html) software from the nucleotide sequences encoding human COL1A1, COL2A1, COL3A1 and LOX genes available from Genbank (http://www.ncbi.nlm.nih.gov/entrez/). Primers were synthesised by TAGN ltd, Gateshead, UK (sequences are shown in Table 1). The highly expressed housekeeping 28S ribosomal gene was amplified in a separate PCR reaction to assess the quality of each cDNA template.

View this table:
Table 1

Primers sequences used for RT-PCR analyses

Name5←sequence→3Product sizeNCBI Reference Sequences (Ref Seq)

mRNA was isolated from stretched and non-stretched cells grown on BioFlex 6-well plates using an RNAeasy extraction kit (Qiagen) and according to manufacturer's instructions, except that the cells were lysed by the direct addition of 350 μl of lysis buffer to the cell layer covering the base of the well. cDNA was synthesised using SuperScript™ First-strand Synthesis system for cDNA synthesis (Invitrogen, UK) according to the manufacturer's protocol. PCR was carried out to amplify cDNA using 1 μl of appropriately diluted cDNA (up to ten-fold dilution) for 40 cycles, comprising 30 s denaturing at 94 °C, 30 s annealing at 54 °C for COL1A1, COL2A1, COL3A1 and LOX (30 s at 51 °C for 28S) and 30 s extension at 72 °C, followed by a 10 min extension step at 72 °C. The reaction mix (50 μl) contained 5 μl of 10 × PCR buffer with (NH4)2SO4, 2.5 μl of 4 mM each dNTP, 3 μl of 25 mM MgCl2, 1 μl of 50 μM each gene specific primer, and 3 μl of Taq DNA polymerase diluted 1/10 in 50% glycerol (all Fermentas Inc., UK). Negative controls for each tissue sample consisted of a cDNA reaction mixture minus reverse transcriptase enzyme. PCR products were mixed with PCR loading buffer and loaded onto a 2% agarose minigel in 1 × TBE and run for 20 min at 100 V to assess quantity and quality of synthesised product.

2.6 Phenotypic changes of mesenchymal stem cells

Following 3 days of 14% stretch, human MSCs on the central portion of the membranes were immediately fixed by immersion in 100% ice-cold acetone for 10 min, and washed at least three times with PBS for 3–5 min each and placed on glass slides. The seeded membranes were then blocked for non-specific labelling in 1% BSA in PBS for 30 min and the primary antibodies collagen I (Oncogene, 1 to 10 dilution), collagen III (Oncogene, 1 to 10 dilution), prolyl 4 hydroxylase α-unit (Chemicon, 1 to 10 dilution), prolyl 4 hydroxylase β-unit (DAKO, 1 to 50 dilution), osteocalcin (Abcam, 1 to 200 dilution) and alkaline phosphatase (Sigma, 1:4000 dilution) applied at the appropriate dilution in 1% BSA in PBS for 1 h at room temperature. Then the membranes were washed three times with PBS for 3–5 min each and incubated with secondary antibody (Alexa fluor 594 goat anti-mouse IgG for monoclonal primary antibodies and Alexa fluor 488 goat anti-rabbit IgG for polyclonals) diluted 1 to 250 in 1% BSA in PBS for 1 h at room temperature, washed five times with PBS for 3–5 min each and mounted on glass slides using Permafluor mounting medium.

2.7 Measurement of the alkaline phosphatase activity

Alkaline phosphatase (ALP) activity in mesenchymal stem cells following 14% stretch for up to 4 days was measured by means of a colorimetric assay. ALP activity has previously been used as a marker of osteoblast cell phenotypes [11,12]. Cells were washed twice with PBS and cell layers were extracted with 200 μl of lysis Buffer (0.2% NP-40, 1 mM MgCl2). The cell extract was transferred to a microtube and centrifuged for 5 min at 10,000 rpm. The enzyme activity in the supernatant was assayed with 10 mM (final concentration) p-nitrophenyl phosphate (Sigma, UK) as a substrate in 0.1 M glycine buffer, pH 10.4, containing 1 mM ZnCl2 and 1 mM MgCl2. The ALP activity (nmol p-nitrophenol (pNp) produced/min) was calculated at the absorbance of 405 nm. Mesenchymal stem cells cultured in osteogenic medium (DMEM containing 10− 8 M dexamethasone, 50 μg/ml ascorbate β-phosphate and 10 mM β-glycerophosphate (Sigma, UK)) together with 10% FCS was used to induce ALP activity and served as a positive control. The protein content in each sample was determined with the BCA protein assay (Sigma, UK) using bovine serum albumin as a standard sample. The specific activity of ALP was given as nmol pNp/min/mg protein.

2.8 Data analysis

Each individual experimental observation was done in triplicate and the mean value calculated. Due to the variability of the counts per minute in control wells between different cell isolates, results were calculated as the percentage of the control (non-stretched cells) and expressed as mean±standard error of the mean and ‘n’ values represent the number of different cell isolates used for each experiment. Statistical analysis was performed using an ANOVA with SPSS software (version 10.5). P values equal to or less than 0.05 were regarded as significant.

3 Results

3.1 [3H]-Proline incorporation in response to stretch

3.1.1 Valve interstitial cells

The incorporation of [3H]-proline into the cell layer and in the medium of porcine valve interstitial cells was dependent upon the degree of stretch and the amount of time for which it was applied. After 2 days of stretch there were significant increases above the control in [3H]-proline incorporation at 10%, 14% and 20% stretch in the cell layer. However, when measured in the medium, this effect only reached significance at 14% and 20% (Fig. 1a and 1b). There was no difference between the response at 7% stretch in either the cell layer or the medium. As a degree of stretch greater than 14% did not appear to be associated with any great increase in collagen synthesis, and because at above this level on occasions cells detached from the membrane, 14% stretch was used in subsequent experiments to determine the time course of the effect. Levels of [3H]-proline-labelled protein increased in both the cell layer and the medium over a 5-day period (Fig. 2a and b). This effect achieved significance at 2 days in both the cell layer and the medium, and continued to increase after 3 days in the cell layer and in the medium.

Fig. 2

Time course of the increase in the incorporation of [3H]-proline by porcine valve interstitial cells into (a) the medium and (b) the cell layer in response 14% stretch over a 1, 2, 3, 4 and 5 day period. Results are expressed as mean±S.E.M. of the percentage increase in the stretched compared to non-stretched controls. =P<0.05, n=4.

Fig. 1

Increase in the incorporation of [3H]-proline by porcine valve interstitial cells into (a) the medium and (b) the cell layer in response to 7%, 10%, 14% and 20% stretch over a 2 day period. Results are expressed as mean±S.E.M. of the percentage increase in the stretched compared to non-stretched controls. =P<0.05, n=4.

3.2 Mesenchymal stem cells

Similar to that observed with valve interstitial cells, mesenchymal stem cells were also able to increase the incorporation of [3H]-proline in response to stretch. However, mesenchymal stem cells were more sensitive to the temporal effect of 14% stretch, with significant increases of [3H]-proline evident in the cell layer after 1 day. Further increases were seen after 2 and 3 days of stretch (Fig. 3a and b). The magnitude of the increase measured in the cell layer and the supernatant was similar, and corresponded broadly with the effects seen in the valve interstitial cells.

Fig. 3

Time course of the increase in the incorporation of [3H]-proline by mesenchymal stem cells into (a) the cell layer and (b) the medium in response 14% stretch over a 1, 2 and 3 day period. Results are expressed as mean±S.E.M. of the percentage increase in the stretched compared to non-stretched controls. =P<0.05, n=4.

3.3 Influence of stretch on collagen gene expression

Expression of the COL1A1 gene was detectable at similar levels of intensity in stretched and non-stretched valve interstitial and mesenchymal stem cells (Fig. 4a and b). However, PCR analysis showed the band representing the gene for COL3A1 was markedly more intense in valve interstitial cells that had been stretched at 14% for 2 days compared with cells that had received no stretch (Fig. 4a). In contrast, the COL3A1 transcript was present in the stretched and non-stretched mesenchymal stem cells (Fig. 4b). There was no expression of the COL2A1 in mesenchymal stem cells. The expression of LOX was markedly more intense in mesenchymal stem cells after 2 days stretch, as compared with the non-stretched cells. (Fig. 4b).

Fig. 4

Expression of the COL3A1 LOX or 28S genes in (a) porcine valve interstitial cells exposed to 14% stretch (S) or non-stretch (N) for 2 days and (b) mesenchymal stem cells.

3.4 Phenotypic changes of mesenchymal stem cells

The immunohistochemical staining of mesenchymal stem cells demonstrated that collagen type III, prolyl 4 hydroxylase α-unit and prolyl 4 hydroxylase β-unit were generally enhanced after 3 days of 14% stretch, as compared with the non-stretched condition (Fig. 5). The influence of stretch on expression of collagenous proteins was more evident for collagen type III than collagen type I at 3 days. In addition, prolyl 4 hydroxylase β-unit seemed to be more intensely expressed than the α-unit. No significant change was observed in collagen I. Osteocalcin and ALP were not detected after 3 days of stretch (data not shown).

Fig. 5

Representive images of the expression of collagen type I (A and B) collagen type III (C and D), prolyl 4-hydroxylase α-unit (E and F) and prolyl 4-hydroxylase β-unit (G and H) in mesenchymal stem cells. Non-stretched cells are shown in panels A, C, E, and G, and cells stretched for 3 days in panels B, D, F and H.

3.5 ALP activity

Our results showed that ALP activity of mesenchymal stem cells did not significantly alter in response to mechanical stretch as compared with the non-stretched cells (Fig. 6). However, these cells do have the capacity to differentiate since there was a significant increase in ALP activity for the cells cultured in the osteogenic medium at day 4 (P<0.05).

Fig. 6

The ALP activity of mesenchymal stem cells exposed to 14% stretch (S) or non-stretch (N) for 4 days. Cells incubated in the osteogenic conditioned medium are used as a positive control (Osteo).

4 Discussion

This study has documented the kinetics of total collagen synthesis and has examined the expression of two major collagen genes, COL1A1 and COL3A1, at the level of the mRNA in response to varying degrees of stretch. This response is dependent upon both the duration of the stretch and the magnitude of the force applied. These properties of valve interstitial cells show how the cells may interact with their mechanical environment to ensure integrity and adaptation of valve tissue. In this respect, the ability to secrete collagen represents a fundamental mechanism whereby the cusp can continue to maintain the load-bearing structures that allow the valve to function adequately. In addition we have also shown that similar responses to stretch are also exhibited by mesenchymal stem cells. The ability of mesenchymal stem cells to reproduce the function of valve interstitial cells is important for the replication of native valve function in a tissue engineered heart valve.

Molecular analysis suggests that an increased expression of COL3A1 gene is a factor in the enhanced incorporation of [3H]-proline seen in the valve cells. However, both COL1A1 and COL3A1 gene expression could be detected in non-stretched mesenchymal stem cells. Thus, the collagen species responsible for the increase in mesenchymal stem cells remains unclear and additional analysis with accurate quantitative assessment of gene expression is required, together with measurement of the expression of other collagen isoforms.

Type I and III collagens are present in valve cusps in the ratio 3:1. In addition, about 2% of the collagen present is collagen V. These are fibrillar collagens that form banded fibrils and provide tissues with tensile strength, as well as influencing cell attachment and migration. Type I collagen is the most abundant and widely distributed collagen in the body and is synthesised in response to injury. Type III collagen is similar in structure to type I, but less abundant, and usually occurs in the same fibril with type I collagen. Interestingly, type III collagen is often encountered in areas of rapid new collagen synthesis [13]. Our results with the valve interstitial cells are in agreement with studies using cardiac fibroblasts that have reported an increased level of collagen type III mRNA, but not that of collagen type I in response to cyclic mechanical stretch [14]. This pattern was not seen in the mesenchymal stem cells. However, the rate of synthesis and breakdown of these different collagen species by valve cells in intact cusp tissue and how this may be influenced by the different types of force to which the valve is exposed (stretch, pressure, flow) is not known. It is likely that the expression of different types of collagen is differentially regulated by specific components of mechanical strain [15,16]. It has been shown that increased levels of static or cyclic pressure are capable of inducing collagen synthesis in valve cells, although these studies did not characterise the collagen isoforms involved [17,18]. The possibility exists that expression and release of collagens other than collagen type III may be linked to other periods or patterns of mechanical stimulation. The stretching protocols use in this study do not take into account the rate of change in application of the force. Differences in dp/dt in vivo may also influence the function of the cells. Additional studies are required to examine the influence of different types of force, and how it is applied, on the synthesis and secretion of each of the three most abundant collagen isoforms present in the valve.

The incorporation of collagen into the ECM of the valve cusp involves the expression and secretion of individual pro-collagen proteins into the extracellular space, the synthesis of processing enzymes such as prolyl 4-hydroxylase, lysyl hydroxylase, prolyl 3-hydroxylase, C-terminal peptidase and lysyl oxidase; the modification and aggregation of the individual subunits and the initiation of covalent cross-linking in collagens [19,20]. The measurement of collagen present in the medium represents the secretion stage of this process. In contrast, levels determined in the cell layer provide an index of the ability of the processing enzymes in the cells to assemble the collagen into insoluble protein attached to the cells, the initial steps in the laying down of a collagen matrix. In contrast to previous studies with other cell types [21] that suggested 90% of newly synthesised collagen was present in the medium, the valve interstitial cells and mesenchymal stem cells appear to be efficient at incorporating the soluble protein into the insoluble form that becomes attached to the cells [22]. The expression of LOX mRNA by mesenchymal stem cells might catalyse the efficient deposition of collagen (insoluble cell layer) from soluble collagen (supernatant) after stretch. Additional studies are required to investigate the expression and activity of collagen synthesis enzymes, as well as the activity that regulates collagen breakdown, in these cell types to assess if their function is enhanced compared with other collagen-producing cells [21,23].

We have previously demonstrated the ability of valve interstitial cells to secrete collagen in response to humoral agents such as serotonin and angiotensin II [24]. However, the relationship between the action of such agents and the influence of mechanical force has not been examined. The responses observed to mechanical force are, however, greater than those documented for either serotonin or angiotensin II, suggesting that mechanotransduction pathways are more efficacious than those involving receptor-operated mechanisms. The ability to synthesise and release collagen represents only one aspect of the mechanisms that regulate the synthesis and degradation of the ECM, which relies on a balance between synthesis and degradation. The net synthesis of extracellular matrix proteins is regulated by the expression and activity of matrix metalloproteinases (MMPs) and their tissue inhibitors (TIMPs). These enzymes and inhibitors have been shown to be expressed in cusp tissue and play an important role in the maintenance and remodelling of the extracellular matrix [25]. Further studies are required to look at the expression and release of other matrix proteins such as fibronectin, elastin and proteoglycans, as well as regulation by mechanical force of MMPs and TIMPs.

It has been shown that the pressures experienced by the valve cusp are related to the degree of stretch experienced by the tissue (and therefore the cells) [26]. From zero to aortic diastolic pressures, the valve cells may be expected to increase in length by approximately 19%. In reality, this degree of stretch may not be experienced directly by the cells due to the shielding effect of the collagen bundles within the cusp. Thus the degree of stretch at which we have seen significant collagen production represents a realistic stimulation of the mechanical environment experienced by valve interstitial cells in vivo. Our data support the notion that when mesenchymal stem cells are grown onto scaffold material and subjected to physiological pressures in a bioreactor, they will be capable of secreting collagen as part of the remodelling process, in a manner that is dependent both on the magnitude and duration of the stretch experienced by the cells.

We have no evidence that stretch induced any differentiation of mesenchymal stem cells into osteoblast cell phenotypes. This observation is important, since it has been shown that valve cells have the ability to differentiate into bone-like cells, a mechanism that may be involved in valve calcification [27]. A limitation of this study is the duration of our experiments. In order to tissue engineer a heart valve, one might expect that conditioning of a valve construct may occur over a period of weeks rather than days. An inability of mesenchymal stem cell to differentiation into osteoblasts (or any other undesired cell phenotypes) will be required over a much longer time frame. In this study we chose to compare human mesenchymal stem cells with porcine valve interstitial cells. In general the specimens of human valves that are available for cell culture are from older people who have significant degrees of calcification. Cells that are isolated from human valves typically grow very slowly, making it questionable as to their validity in representing normal valve cell function. In a small number of experiments using human valve cells that were amenable to experimentation, we saw a similar pattern and degree of collagen synthesis in response to stretch compared with the porcine cells (data not shown).

The function of cells under different levels of mechanical stimulation is an important consideration when choosing a suitable cell source with which to tissue engineer a heart valve. The findings of this study demonstrate the ability of valve interstitial cells to rapidly secrete collagen isoforms in response to physiological degrees of stretch and the potential for this property of valve cells to be mimicked by mesenchymal stem cells. This demonstrates the capacity of mesenchymal stem cells to participate in the laying down of new ECM in tissue engineered constructs, a property of valve interstitial cells that mesenchymal stem cell will be required to reproduce if they are to be used to successfully tissue engineer a heart valve.


These studies were supported by grants from the Biotechnology and Biological Sciences Research Council and the Magdi Yacoub Institute. The authors would like to thank Mr. Joseph Antoniw for his technical support.


  • Time for primary review 18 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
View Abstract