Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Seeding density matters: extensive intercellular contact masks the surface dependence of endothelial cell–biomaterial interactions

Seeding density matters: extensive intercellular contact masks the surface dependence of... J Mater Sci: Mater Med (2011) 22:389–396 DOI 10.1007/s10856-010-4211-5 Seeding density matters: extensive intercellular contact masks the surface dependence of endothelial cell–biomaterial interactions • • • Yun Xia Melissa Prawirasatya Boon Chin Heng Freddy Boey Subbu S. Venkatraman Received: 23 August 2010 / Accepted: 8 December 2010 / Published online: 8 January 2011 Springer Science+Business Media, LLC 2010 Abstract The effects of seeding density have often been stents/grafts [2], complete endothelialization has become overlooked in evaluating endothelial cell-biomaterial even more crucial to prevent degradation debris from interactions. This study compared the cell attachment and entering the bloodstream and causing adverse complica- proliferation characteristics of endothelial cells on modi- tions. In order to promote such endothelialization, sub- fied poly (L-lactic acid) (PLLA) films conjugated to gelatin stantial effort has been focused on suitably functionalizing and chitosan at low and high seeding densities (5,000 and the surfaces of biomaterial. For example, extracellular 50,000 cells/cm ). During the early stage (2 h) of cell- matrix (ECM) proteins (e.g. collagen, fibronectin and biomaterial interaction, a low seeding density enabled us to laminin) and functional domains of ECM components (e.g. observe the intrinsic surface-dependent differences in cell RGD, YIGSR and REDV), have been immobilized on attachment capacity and morphogenesis, whereas extensive biomaterial surfaces [3, 4]. intercellular interactions at high seeding density masked A wide range of cell seeding densities from 4 9 10 to 5 2 differences between substrates and improved cell attach- 2 9 10 cells/cm have been used for in vitro studies on ment on low-affinity substrates. During the later stage of endothelial cell-biomaterial interaction [5–12]. To date, cell-biomaterial interaction over 7-days of culture, the there has been no established standard protocol to define proliferation rate was found to be surface-dependent at low the seeding density, which makes it difficult to compare seeding density, whereas this surface-dependent difference different studies done on the same substrate. The bio- was not apparent at high seeding density. It is recom- compatibility of certain materials and protocols evaluated mended that low seeding density should be utilized for using relatively high seeding densities, might not be evaluating biomaterial applications where EC density is appropriate for applications whereby the availability of likely to be low, such as in situ endothelialization of blood- cells is limited. For example, in the case of in situ endo- contacting devices. thelialization of blood-contacting devices, regrowth of the endothelial layer might be derived from the migration of ECs from adjacent tissue (‘‘trans-mural endothelializa- 1 Introduction tion’’) and/or attachment and proliferation or the circulat- ing endothelial cell precursors [13–15]. Available cell Rapid endothelialization is crucial for cardiovascular stents density in both situations is considered to be low [16, 17]. and grafts to prevent post-implantation thrombosis and It has long been recognized that cell-matrix adhesion is restenosis [1]. With increasing demand for biodegradable predominantly mediated by integrins, and adherent cells can sense their immediate environment through integrin- based adhesion complexes, namely focal adhesions (FA), Y. Xia  M. Prawirasatya  B. C. Heng  F. Boey  tightly associated with the actin cytoskeleton [18]. Cell– S. S. Venkatraman (&) cell adhesions are also sites of physical connection as well School of Materials Science and Engineering, as signaling transduction structures for regulating cell Nanyang Technological University, 50 Nanyang Avenue, behavior [19]. Cadherin is a key cell–cell adhesion mole- Nanyang 639798, Singapore cule localized at adherens junctions [20] which has a e-mail: assubbu@ntu.edu.sg 123 390 J Mater Sci: Mater Med (2011) 22:389–396 function similar to that of integrin, which may be consid- 2.2 Cell attachment and proliferation ered its counterpart in the FA complex. Integrins and cadherins are two distinct families of transmembrane cell HUVECs (Lonza) were cultured in endothelial growth adhesion receptors. While integrins allow cells to adhere to medium (EGM, Lonza) under 95% humidified atmosphere and 5% CO at 37C. Cells were dissociated with 0.025% the extracellular matrix, cadherins bind homotypically to cadherins on neighboring cells and are responsible for the trypsin–EDTA (Lonza) and washed in Dulbecco’s modified eagle medium (DMEM, Gibco) for 3 min by centrifugation development of adherens junctions in epithelial tissues. Arthur et al. [21] showed that the signaling cascades of to avoid any interference associated with adhesive proteins from serum, and then seeded on the various PLLA sub- both cell–matrix and cell–cell adhesion, transmitted through integrins and cadherins respectively, involve Rho strates in DMEM at either a low seeding density of 5,000 proteins, which are key regulators in reorganization of the cells/cm whereby cells are sparsely distributed on the actin cytoskeleton. Other studies have demonstrated cross- substrate, or a high seeding density of 50,000 cells/cm talk between cell-matrix and cell–cell junctions and both which is comparable to cell density at confluence. After 2 h types of junctions cooperatively regulate cell movement, incubation, unattached cells were gently rinsed off. The proliferation, adhesion and polarization [22–24]. number of attached cells was quantified by the WST-8 assay (Dojindo, Japan). Cell attachment percentage was quanti- Stimuli from neighboring cells via interaction of cell- surface receptors and secreted growth factors/cytokines are fied as N /N 9 100%, where N and N were 2h seeding 2h seeding the cell count at 2 h and the initial seeding respectively. The strongly dependent on the cell density. When the cell density is low, direct cell–cell contacts are limited and cell-bioma- cell count in each experimental condition was monitored on terial interaction is expected to be pre-dominantly influenced alternate days until confluence was reached. Cell doubling by cell-substrate contact. As cell density increases, cell–cell time of an exponential proliferation was calculated interaction becomes more extensive and is expected to according to the method as described previously [26]. profoundly influence cellular responses to biomaterials [25]. This study reports on the evaluation of the effects of initial 2.3 Mechanistic study on improved cell attachment seeding density on the EC-biomaterial interaction, by com- in high-density seeding paring a low seeding density of 5,000 cells/cm versus a high Two mechanisms are possible for the improvement of cell seeding density of 50,000 cells/cm using human umbilical vein endothelial cells (HUVECs). Three different PLLA attachment at high-density seeding. The secreted growth factors and cytokines by HUVECs at high-density seeding substrates, namely unmodified PLLA, PLLA–gAA–gelatin and PLLA–gAA–chitosan, were prepared as described pre- are more concentrated than that at low-density seeding. In order to study the influence of secreted growth factors and viously [26]. During the early phase of EC-biomaterial interaction, cell attachment and morphogenesis are the main cytokines on cell attachment, HUVECs of 5,000 cells/cm cellular responses, whereas cell proliferation is the main were seeded on pristine PLLA substrates in both fresh and cellular response during the later phase. conditioned medium for comparison. Conditioned medium here was prepared by incubating fresh EGM with confluent HUVECs in a tissue culture flask for 24 h, and then col- lected to centrifugation (3,600 rpm, 10 min) so as to 2 Materials and methods remove detached cells and debris. HUVECs were allowed to attach for 2 h. The cell attachment percentage was 2.1 Surface modification and characterization of PLLA studied using WST-8 assay as described earlier. It is also hypothesized that extensive cell–cell interaction PLLA (Purac Far East, Singapore) substrates were prepared between neighboring cells at higher density promotes cell as described previously [26]. Briefly, acrylic acids (AA, attachment. In order to evaluate this hypothesis, HUVECs Sigma-Aldrich) were graft polymerized on argon-plasma were seeded at a density of 5,000 cells/cm on both pristine treated PLLA surface. Gelatin (Type A, Sigma-Aldrich) and and pre-seeded PLLA. Pre-seeded PLLA here was prepared chitosan (Sigma-Aldrich) were then immobilized through by incubating HUVECs on PLLA substrates at a density of covalent bond formation between carboxylic groups found 25,000 cells/cm for 24 h. Cells were allowed to attach for on AA and amine groups found on gelatin or chitosan in 2 h, and the number of attached cells was quantified by the water soluble carbodiimide. PLLA–gAA–gelatin and WST-8 assay as described earlier. The cell attachment PLLA–gAA–chitosan refer to PLLA films modified with percentage on bare PLLA was calculated as N /5000 9 2h gelatin and chitosan respectively. Surface chemical com- 100%, where N was the cell density after 2 h incubation. 2h position and wettability were characterized by X-ray The cell attachment percentage on pre-seeded PLLA was Spectrometer and contact angle respectively [26]. calculated as (N -N /5000 9 100%, where N and N 2h pre) 2h pre 123 J Mater Sci: Mater Med (2011) 22:389–396 391 were the cell density after 2 h incubation and the cell density of pre-seeding HUVECs respectively. All experi- ments were carried out in 24-well tissue culture plate. 2.4 Immunofluorescence analysis Cells were fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 for 10 min, and incubated with 10% goat serum (DAKO) in PBS for 30 min at room temperature. Vinculin was labeled with mouse anti- human vinculin antibody (clone h Vin-1, Sigma-Aldrich) and visualized with goat anti-mouse IgG-FITC conjugate (Sigma-Aldrich). Actin filaments and nuclei were labeled with Alexa Fluor 568 (Molecular Probe) and DAPI (Molecular Probe) respectively. Cell images were captured with an Olympus IX71 inverted fluorescent microscope and Leica TCS SP5 laser-scanning spectral confocal micro- scope (CLSM). Cell morphometric parameters, spreading Fig. 1 a Extensive intercellular contact masks the surface depen- area and aspect ratio, were analyzed by the Image Pro Plus dence of cell adhesion. Surface-dependent cell adhesion was analyzed software using a modified method of Treiser et al.[27]. with ANOVA, *P \ 0.05. Density-dependent cell adhesion was analyzed with paired Student’s t-test, #P \ 0.05. b Growth factors and cytokines secretion showed no significant effect on cell attach- 2.5 Statistical analysis ment by comparing seeding in fresh medium (A) with that in conditioned medium (B). Pre-seeded cells on PLLA greatly promoted Three experimental replicates (n = 3) were utilized in the cell attachment (**P \ 0.05) by comparing seeding HUVEC on bare cell attachment and proliferation study. The results are PLLA (A) with that on 25,000cell/cm pre-seeded PLLA (C). After 2 h culture, unattached cells were removed by PBS rinsing. The presented as mean ± SD. The cell spreading area and number of remaining cells were measured by WST-8 reagent aspect ratio were plotted as box plots to show the data distribution and significant differences between measured biomolecules on the PLLA surface enhanced cell attach- groups. At least 50 cells were analyzed for each experi- mental group. Each box encompasses 25–75 percentiles, ment: 90% attachment was observed on PLLA–gAA–gel- atin and 50% on PLLA–gAA–chitosan. These observations with extending-lines covering the 95th and 5th percentiles, the thin line representing the median (50th percentile), agreed well with our previous study [26]. In contrast, the and the thick line representing the mean values. Values surface dependence of cell attachment was not observed at outside the 95th and 5th percentiles were treated as outliers, high seeding density of 50,000 cells/cm . Instead, all sub- and are represented by diamond dots. Density-dependent strates appeared to be equivalent: about 70% of seeded cellular responses were analyzed using Student’s t-test. cells were able to attach on all three PLLA substrates Surface-dependent cellular responses were analyzed using regardless of their differences in surface properties. High analysis of variance (ANOVA). A P value of less than 0.05 seeding density significantly improved cell attachment on was used to infer statistical significance of differences. PLLA and PLLA–gAA–chitosan ( P \ 0.05). Interest- ingly, cell attachment percentage on PLLA–gAA–gelatin decreased from 90 to 70%. However, the total cell count on 3 Results and discussions PLLA–gAA–gelatin in fact increased from 4,500 to 35,000. It is conceivable that high seeding density 3.1 Extensive intercellular contact masks the surface enhanced cell attachment during the early phase of cell- dependence of cell attachment biomaterial interaction and masked any differences in the substrate surface. Surface dependence of cell attachment Cell attachment percentage values were compared between was only observed at low seeding density by eliminating low and high seeding density experimental groups, with the intensive cell–cell interaction. results illustrated in Fig. 1a. In the case of 5,000 cells/cm The improvement of cell attachment at high seeding seeding density, the cell attachment percentage appeared density can possibly arise from two mechanisms: (a) In comparison to the low seeding density, the concentration of to be surface dependent (ANOVA, *P \ 0.05). After 2 h incubation, about 30% of seeded cells had adhered secreted growth factors and cytokines at high seeding on unmodified PLLA. The presence of immobilized density is higher, which could contribute to stronger cell 123 392 J Mater Sci: Mater Med (2011) 22:389–396 attachment signaling stimulation; (b) In comparison to the low seeding density, cell–cell interaction at high seeding density is more extensive, which could contribute to cell attachment signaling activation through crosstalk between cell-substrate and cell–cell adhesion. To investigate the first possible mechanism, conditioned medium which is the EGM incubated with HUVECs for 24 h was used to represent the medium at high density seeding. Cell attachment on PLLA at 5,000 cells/cm in conditioned medium was compared with that in fresh medium. As illustrated in Fig. 1b, cell attachment per- centage values were about 37% in both conditions. This implies that increased secretion of growth factors and cytokines at high seeding densities had negligible effect on cell attachment during 2 h incubation. To investigate the second possible mechanism, cell attachment on PLLA at 5,000 cells/cm was compared with that on pre-seeded PLLA. As shown in Fig. 1b, a signifi- cant increase of cell attachment percentage from 37 to 72% was observed by pre-seeding PLLA. This strongly sug- gested that the presence of pre-seeded HUVEC promoted the attachment of incoming cells. This simple test dem- onstrated a possible scenario occurring at high seeding density, whereby adherent cells hastened the attachment of cells from free suspension, through cell–cell interaction followed by triggered cell-substrate attachment. Another possible scenario is that high seeding density favors the formation of cell–cell adhesion in suspension even prior to cell-substrate attachment. Clumps of associated cells sed- iment and attach on the substrate simultaneously, such that cell attachment was significantly improved even on low cell affinity substrates. At present, we cannot distinguish between these two possible mechanisms. Zhu et al. [28, 29] showed that cell attachment on gelatin- and chitosan- modified PLLA (modified through aminolysis) was comparable at a seeding density of 12 9 4 2 10 cell/cm , which agrees with our results at high seeding density. However, our study demonstrated the superiority of cell attachment on PLLA–gAA–gelatin compared to Fig. 2 Surface chemistry, rather than seeding density, influences cell morphogenesis. HUVECs were seeded on PLLA (A and B), PLLA– PLLA–gAA–chitosan at a relatively low seeding density of gAA–gelatin (C and D), and PLLA–gAA–chitosan (E and F) at low 5,000 cell/cm . The data showed clearly that differences seeding density of 5000/cm (A, C and E) or high seeding density of among substrates are more pronounced at low seeding 50,000/cm (B, D and F). After 2 h culture, unattached cells were density, but was masked at high seeding density. This removed by PBS rinsing. The remaining cells were fluorescently stained to label nuclei (DAPI), vinculin (FITC) and F-actin (Alexa should be taken into consideration whenever biomaterials Fluor 568). The images were captured by Leica TCS SP5 CLSM. are to be evaluated for use in applications involving low Scale bar 20 lm cell densities. 3.2 Surface chemistry, rather than seeding density, differences upon comparing cells grown on the same sub- 2 2 influences cell morphogenesis strate at low (5,000 cells/cm ) and high (50,000 cells/cm ) seeding densities. On PLLA–gAA–gelatin (Fig. 2c,d), cells observed at both seeding densities exhibited the charac- After 2 h incubation, cell morphology was observed with immunostaining of F-actin, vinculin and nuclei. As shown teristic morphology of spreading HUVEC. Bundles of in Fig. 2, cell morphology did not exhibit distinctive F-actin were found in cell lamellipodia, which are the 123 J Mater Sci: Mater Med (2011) 22:389–396 393 broad, flat protrusions at the leading edge of a motile cell. Filopodia, the thin finger-like structures filled with tight parallel bundles of F-actin, were conspicuously protruding from the lamellipodia. Both of these are well-known structures involved in cell spreading and migration [30]. On PLLA–gAA–chitosan (Fig. 2e,f) seeded with either 5,000 or 50,000 cells/cm , the cells were observed to have significantly reduced spreading area in high-density seed- ing. This was further confirmed in the subsequent mor- phometric parameter analysis. Nevertheless, the typical spreading cell morphology was observed under both seeding density conditions, i.e. defined lamellipodia and filopodia with F-actin bundles localized mostly near the cell periphery. On PLLA (Fig. 2a,b), no obvious difference in cell morphology between low- and high-density seeding was found, other than the reduced cell spreading area. Compared with the cell morphology on gelatin- and chitosan-modified PLLA, much thinner F-actin bundles were observed beneath the cell membrane on bare PLLA and there was hardly any vinculin expression, thus indi- cating poor EC affinity for unmodified PLLA as demon- strated in our previous study [26]. Cell morphology was evaluated using two morphomet- ric parameters, cell spreading area and aspect ratio. Cell spreading area measures the extent of cell spreading on a substrate as projection area. As shown in Fig. 3a, the cell spreading area observed from a confluent EC monolayer grown on tissue culture plate (TCPS) was 1,400 lm on Fig. 3 Seeding density influenced (a) cell spreading area and (b) aspect ratio. Cell morphometric indicators observed on (A) TCPS, average. After 2 h incubation, cell spreading under any of (B and C) PLLA, (D and E) PLLA–gAA–gelatin, (F and G) PLLA– the experimental conditions was not able to reach that gAA–chitosan. Cells on TCPS were confluent and incubated for value. It is also noted that cell spreading areas from the 4 days. Low seeding density of 5000 cells/cm was used on B,D, high seeding density group were significantly smaller than F. High seeding density of 50,000 cells/cm was used on C, E, and G. High resolution immunofluorescent images of individual cells those observed in the low seeding density group. Spatial were taken by Leica TCS SP5 CLSM, and then analyzed by Image restriction is believed to be the main reason for the Pro Plus software. At least 50 cells were measured for each condition decrease in spreading area in the high seeding density group. Normal EC displays a typical ‘cobble-stone’ morphol- PLLA and PLLA–gAA–chitosan is believed to be the main ogy at confluence, with an epithelioid phenotype. By reason for the compromised cell morphology observed on contrast, when cells are sparse or when intercellular junc- both substrates. Unlike PLLA–gAA–gelatin, PLLA and PLLA–gAA–chitosan contained no cell attachment mole- tions are disrupted, a fibroblastoid/mesenchymal mor- phology predominates. Cell aspect ratio is defined as the cules/ligands on their surfaces. Cell attachment in serum- ratio of the major axis to the minor axis of an equivalent free medium relied largely on hydrophobic or electrostatic ellipse. In Fig. 3b, the cell aspect ratio of confluent EC interaction between substrate and attachment ligands on the monolayer grown on TCPS displaying cobble-stone mor- cell membrane. Amongst the high-density seeding experi- phology was about 2.38 on average. mental groups, cell aspect ratio on PLLA–gAA–gelatin Amongst the low-density seeding experimental group, was reduced to 1.90, mostly due to cell–cell contact inhi- the mean aspect ratio on PLLA–gAA–gelatin was about bition. The cell aspect ratio on PLLA and PLLA–gAA– 2.80, suggesting a mesenchymal phenotype. By contrast, chitosan was similar to that at low-density seeding. It aspect ratios on PLLA and PLLA–gAA–chitosan were suggests that the dominant factor for aspect ratio is surface about 1.5, which were significantly less than that of cobble- chemistry. The data presented in Fig. 3 suggest that cell morpho- stone morphology. This implied that although cells adhered on both surfaces, their cell morphology was somehow genesis during the early stage of cell-biomaterial interac- compromised. The absence of cell attachment ligands on tion was strongly dependent on substrate chemistry rather 123 394 J Mater Sci: Mater Med (2011) 22:389–396 than seeding density. However, seeding density did influ- ence some cell morphometric parameters. Reduced cell spreading area was observed on all three PLLA substrates and a less elongated cell phenotype was seen on PLLA– gAA–gelatin, at high seeding density. 3.3 Extensive intercellular contact masks surface dependence of cell proliferation When endothelial cells are sparsely-seeded in the sub- confluent state, they are actively proliferating and are sensitive to growth-factor stimulation. Once confluence has been achieved, the cells are contact-inhibited in their growth and protected from apoptosis [31]. Figure 4a shows the cell proliferation profiles on the various PLLA substrates as a function of culture duration, when HUVECs were sparsely seeded at a low density of 5,000 cell/cm . Cells grown on TCPS and PLLA–gAA– gelatin proliferated with a doubling time to about 15–16 h. In contrast, cell proliferation showed stagnation on PLLA– gAA–chitosan and even negative on PLLA during the first 72 h, with exponential growth taking place only after 72 h. Figure 4b illustrates the cell proliferation profiles of various PLLA substrates, at a high seeding density of 50,000 cells/cm . Cells plated on TCPS, PLLA–gAA– gelatin and PLLA–gAA–chitosan exhibited similar prolif- eration behavior. Cell proliferation started a few hours after seeding with a doubling time of about 7–9 h during the first 24 h. After 24–72 h, cell proliferation slowed down and eventually stopped. The number of HUVEC Fig. 4 Extensive intercellular contact masks surface dependence of on TCPS and PLLA–gAA–gelatin did not increase sig- cell proliferation. Cell proliferation was monitored over 7 days of nificantly after 72 h, suggesting that confluence was culture using (a) low seeding density of 5,000 cells/cm , and (b) high achieved and that cell–cell contact inhibited proliferation. seeding density of 50,000 cells/cm Cells grown on PLLA displayed an exponential prolif- eration pattern, which was different from the other three substrates. attachment and proliferation over PLLA. Clearly, PLLA– Comparing Fig. 4a to Fig. 4b, the profound impact of gAA–gelatin is a better candidate for such applications, a initial seeding density on cell proliferation is clearly evi- conclusion that could be reached only through studies at dent. First, low density seeding revealed the superiority of low seeding density. After confluence was reached, cell morphology was PLLA–gAA–gelatin in supporting cell proliferation com- pared to PLLA–gAA–chitosan, while extensive intercel- observed through cytoskeletal and focal adhesion immu- lular contact at high density seeding masked this difference nostaining (Fig. 5). It was found that cell morphology was in EC behavior. Second, by plating a confluent monolayer indistinguishable on the various PLLA surfaces at either at high seeding density, the proliferation stagnation on low (a, c and e) or high (b, d and f) seeding density. The PLLA–gAA–chitosan and PLLA was overcome. In appli- characteristic cobble-stone morphology was observed with cations involving the in situ endothelialization of cardio- bundles of stress fibers extending from the nucleus to vascular implants, such as coronary stents and patent lamellipodia across the entire cytoplasm and terminating at foramen ovale (PFO) occluders, the number of circulating FA and cell–cell junctions (Fig. 5g,h). During cell spreading and proliferation, cells produce their own ECs or EC-progenitor cells is relatively low [32]. At high seeding density, PLLA–gAA–gelatin and PLLA–gAA– extracellular matrix with time. The self-synthesized matrix may in turn mask the intrinsic effects of the original sub- chitosan showed equivalent performance in cell attachment and proliferation. However, at low seeding density, PLLA– strate on cell proliferation. The results of this study cor- gAA–chitosan showed only marginal improvement in cell roborate with earlier reports—that differences in cell 123 J Mater Sci: Mater Med (2011) 22:389–396 395 cell attachment behavior. Extensive intercellular interac- tions at high seeding density masked differences amongst substrates, and enhanced cell attachment on all substrates examined. During the later stage of cell-biomaterial inter- action, cell proliferation profile was found to be surface-depen- dent at low seeding density, whereas surface dependence was masked at high seeding density. It is recommended that low seeding density should be utilized for in vitro evaluation the compatibility of biomedical materials. Acknowledgments Authors would like to thank National Research Foundation of Singapore for funding the work with their Competitive Research Grant and Dr. Wong Yee Shan for critical reading. References 1. Bhattacharya V, Cleanthis M, Stansby G. Preventing vascular graft failure: Endothelial cell seeding and tissue engineering. Vasc Dis Prev. 2005;2:21–7. 2. Venkatraman SS, Boey F, Lao LL. Implanted cardiovascular polymers: Natural, synthetic and bio-inspired. Prog Polym Sci. 2008;33:853–74. 3. de Mel A, Jell G, Stevens MM, Seifalian AM. Biofunctional- ization of biomaterials for accelerated in situ endothelialization: a review. Biomacromolecules. 2008;9:2969–79. 4. Eisenbarth E, Velten D, Breme J. Biomimetic implant coatings. Biomol Eng. 2007;24:27–32. 5. Lu A, Sipehia R. Antithrombotic and fibrinolytic system of human endothelial cells seeded on PTFE: The effects of surface modification of PTFE by ammonia plasma treatment and ECM protein coatings. Biomaterials. 2001;22:1439–46. 6. Yang J, Bei J, Wang S. Enhanced cell affinity of poly (D, L- lactide) by combining plasma treatment with collagen anchorage. Biomaterials. 2002;23:2607–14. 7. Gumpenberger T, Heitz J, Ba ¨uerle D, Kahr H, Graz I, Romanin C, Svorcik V, Leisch F. Adhesion and proliferation of human Fig. 5 Substrate dependence of cell morphology disappeared at the endothelial cells on photochemically modified polytetrafluoro- point of confluence. Fluorescent micrographs of HUVEC were taken ethylene. Biomaterials. 2003;24:5139–44. on PLLA (A and B), PLLA–gAA–gelatin (C and D), and PLLA– 8. Miller DC, Thapa A, Haberstroh KM, Webster TJ. Endothelial gAA–chitosan (E and F). Low seeding density of 5,000 cells/cm was and vascular smooth muscle cell function on poly(lactic-co-gly- used in A, C and E, and high seeding density of 50,000 cells/cm was colic acid) with nano-structured surface features. Biomaterials. used in B, D and F.(G) Matured FA clusters were observed in 2004;25:53–61. HUVEC. (H) F-actin bundles terminated at FA clusters by colocal- ´ ´ 9. Berard X, Remy-Zolghadri M, Bourget C, Turner N, Bareille R, ization of F-actin and vinculin. The remaining cells were fluorescently Daculsi R, Bordenave L. Capability of human umbilical cord stained to label nuclei (DAPI), vinculin (FITC) and F-actin (Alexa blood progenitor-derived endothelial cells to form an efficient Fluor 568). The images were captured by Olympus IX71 inverted lining on a polyester vascular graft in vitro. Acta Biomater. microscopy. Scale bar 50 lm 2009;5:1147–57. 10. Boura C, Kerdjoudj H, Moby V, Vautier D, Dumas D, Schaaf P, behavior rising from differences in seeding conditions Voegel JC, Stoltz JF, Menu P. Initial adhesion of endothelial cells (substrate and/or density) are more apparent during the first on polyelectrolyte multilayer films. Bio-Med Mater Eng. 2006;16:S115–21. few days of cell–biomaterial interaction [33]. 11. Chen ZG, Wang PW, Wei B, Mo XM, Cui FZ. Electrospun collagen-chitosan nanofiber: A biomimetic extracellular matrix for endothelial cell and smooth muscle cell. Acta Biomater. 4 Conclusion 2010;6:372–82. 12. Crombez M, Chevallier P, Gaudreault RC, Petitclerc E, Manto- vani D, Laroche G. Improving arterial prosthesis neo-endotheli- This study has demonstrated the important role of seeding alization: Application of a proactive VEGF construct onto PTFE density in cell–biomaterial interaction. During the early surfaces. Biomaterials. 2005;26:7402–9. stage of cell–biomaterial interaction whereby cell attach- 13. Brewster LP, Bufallino D, Ucuzian A, Greisler HP. Growing a ment is the pre-dominant cellular activity, low seeding living blood vessel: Insights for the second hundred years. Bio- density enabled us to observe surface chemistry-dependent materials. 2007;28:5028–32. 123 396 J Mater Sci: Mater Med (2011) 22:389–396 14. Avci-Adali M, Paul A, Ziemer G, Wendel HP. New strategies for 23. Schwartz MA, Ginsberg MH. Networks and crosstalk: Integrin in vivo tissue engineering by mimicry of homing factors for self- signalling spreads. Nat Cell Biol. 2002;4:E65–8. endothelialisation of blood contacting materials. Biomaterials. 24. Geiger B, Bershadsky A, Pankov R, Yamada KM. Transmem- 2008;29:3936–45. brane crosstalk between the extracellular matrix and the cyto- 15. Rotmans JI, Heyligers MMJ, Verhagen HJM, Velema E, skeleton. Nat Rev Mol Cell Biol. 2001;2:793–805. Nagtegaal MM, de Kleijn DPV, de Groot FG, Stroes ESG, 25. Nagahara S, Matsuda T. Cell-substrate and cell-cell interactions Pasterkamp G. In vivo cell seeding with anti-CD34 antibodies differently regulate cytoskeletal and extracellular matrix protein successfully accelerates endothelialization but stimulates intimal gene expression. J Biomed Mater Res. 1996;32:677–86. hyperplasia in porcine arteriovenous expanded polytetrafluoro- 26. Xia Y, Boey F, Venkatraman SS. Surface modification of poly(L- ethylene grafts. Circulation. 2005;112:12–8. lactic acid) with biomolecules to promote endothelialization. 16. Hill JM, Zalos G, Halcox JPJ, Schenke WH, Waclawiwm MA, Biointerphases. 2010;5:FA32–40. Quyyumi AA, Finkel T. Circulating endothelial progenitor cells, 27. Treiser MD, Liu E, Dubin RA, Sung H-J, Kohn J, Moghe P. vascular function, and cardiovascular risk. N Engl J Med. Profiling cell-biomaterial interactions via cell-based fluorore- 2003;348:593–600. porter imaging. BioTechniques. 2007;43:361–8. 17. Lin A, Ding X, Qiu F, Song X, Fu G, Ji J. In situ endothelial- 28. Zhu Y, Gao C, Liu X, He T, Shen J. Immobilization of biomac- ization of intravascular stents coated with an anti-CD34 antibody romolecules onto aminolyzed poly(L-lactic acid) toward acceler- functionalized heparin-collagen multilayer. Biomaterials. 2010; ation of endothelium regeneration. Tissue Eng. 2004;10:53–61. 31:4017–25. 29. Zhu Y, Gao C, Liu Y, Shen J. Endothelial cell functions in vitro 18. Geiger B, Spatz JP, Bershadsky AD. Environmental sensing cultured on poly(L-lactic acid) membranes modified with differ- through focal adhesions. Nat Rev Mol Cell Biol. 2009;10:21–33. ent methods. J Biomed Mater Res A. 2004;69A:436–43. 19. Dejana E. Endothelial cell-cell junctions: happy together. Nat 30. Mattila PK, Lappalainen P. Filopodia: molecular architecture and Rev Mol Cell Biol. 2004;5:261–70. cellular functions. Nat Rev Mol Cell Biol. 2008;9:446–54. 20. Nelson WJ. Regulation of cell–cell adhesion by the cadherin– 31. Liebner S, Cavallaro U, Dejana E. The multiple languages of catenin complex. Biochem Soc Trans. 2008;36:149–55. endothelial cell-to-cell communication. Arterioscl Throm Vas. 21. Arthur WT, Noren NK, Keith B. Regulation of rho family 2006;26:1431–8. GTPases by cell-cell and cell-matrix adhesion. Biol Res. 2002;35: 32. Huang Y, Venkatraman SS, Boey FYC, Umashankar PR, Moh- 239–46. anty M, Arumugam S. The short-term effect on restenosis and 22. Sakamoto Y, Ogita H, Hirota T, Kawakatsu T, Fukuyama T, thrombosis of a cobalt-chromium stent eluting two drugs in a Yasumi M, Kanzaki N, Ozaki M, Takai Y. Interaction of integrin porcine coronary artery model. J Interv Cardiol. 2009;22:466–78. a b with nectin: implication in cross-talk between cell-matrix 33. Reilly GC, Engler AJ. Intrinsic extracellular matrix properties v 3 and cell-cell junctions. J Biol Chem. 2006;281:19631–44. regulate stem cell differentiation. J Biomech. 2010;43:55–62. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Materials Science: Materials in Medicine Springer Journals

Seeding density matters: extensive intercellular contact masks the surface dependence of endothelial cell–biomaterial interactions

Loading next page...
 
/lp/springer-journals/seeding-density-matters-extensive-intercellular-contact-masks-the-vTEeJAgJQx

References (34)

Publisher
Springer Journals
Copyright
Copyright © 2010 by Springer Science+Business Media, LLC
Subject
Materials Science; Biomaterials; Biomedical Engineering; Regenerative Medicine/Tissue Engineering; Polymer Sciences; Ceramics, Glass, Composites, Natural Materials; Surfaces and Interfaces, Thin Films
ISSN
0957-4530
eISSN
1573-4838
DOI
10.1007/s10856-010-4211-5
pmid
21221736
Publisher site
See Article on Publisher Site

Abstract

J Mater Sci: Mater Med (2011) 22:389–396 DOI 10.1007/s10856-010-4211-5 Seeding density matters: extensive intercellular contact masks the surface dependence of endothelial cell–biomaterial interactions • • • Yun Xia Melissa Prawirasatya Boon Chin Heng Freddy Boey Subbu S. Venkatraman Received: 23 August 2010 / Accepted: 8 December 2010 / Published online: 8 January 2011 Springer Science+Business Media, LLC 2010 Abstract The effects of seeding density have often been stents/grafts [2], complete endothelialization has become overlooked in evaluating endothelial cell-biomaterial even more crucial to prevent degradation debris from interactions. This study compared the cell attachment and entering the bloodstream and causing adverse complica- proliferation characteristics of endothelial cells on modi- tions. In order to promote such endothelialization, sub- fied poly (L-lactic acid) (PLLA) films conjugated to gelatin stantial effort has been focused on suitably functionalizing and chitosan at low and high seeding densities (5,000 and the surfaces of biomaterial. For example, extracellular 50,000 cells/cm ). During the early stage (2 h) of cell- matrix (ECM) proteins (e.g. collagen, fibronectin and biomaterial interaction, a low seeding density enabled us to laminin) and functional domains of ECM components (e.g. observe the intrinsic surface-dependent differences in cell RGD, YIGSR and REDV), have been immobilized on attachment capacity and morphogenesis, whereas extensive biomaterial surfaces [3, 4]. intercellular interactions at high seeding density masked A wide range of cell seeding densities from 4 9 10 to 5 2 differences between substrates and improved cell attach- 2 9 10 cells/cm have been used for in vitro studies on ment on low-affinity substrates. During the later stage of endothelial cell-biomaterial interaction [5–12]. To date, cell-biomaterial interaction over 7-days of culture, the there has been no established standard protocol to define proliferation rate was found to be surface-dependent at low the seeding density, which makes it difficult to compare seeding density, whereas this surface-dependent difference different studies done on the same substrate. The bio- was not apparent at high seeding density. It is recom- compatibility of certain materials and protocols evaluated mended that low seeding density should be utilized for using relatively high seeding densities, might not be evaluating biomaterial applications where EC density is appropriate for applications whereby the availability of likely to be low, such as in situ endothelialization of blood- cells is limited. For example, in the case of in situ endo- contacting devices. thelialization of blood-contacting devices, regrowth of the endothelial layer might be derived from the migration of ECs from adjacent tissue (‘‘trans-mural endothelializa- 1 Introduction tion’’) and/or attachment and proliferation or the circulat- ing endothelial cell precursors [13–15]. Available cell Rapid endothelialization is crucial for cardiovascular stents density in both situations is considered to be low [16, 17]. and grafts to prevent post-implantation thrombosis and It has long been recognized that cell-matrix adhesion is restenosis [1]. With increasing demand for biodegradable predominantly mediated by integrins, and adherent cells can sense their immediate environment through integrin- based adhesion complexes, namely focal adhesions (FA), Y. Xia  M. Prawirasatya  B. C. Heng  F. Boey  tightly associated with the actin cytoskeleton [18]. Cell– S. S. Venkatraman (&) cell adhesions are also sites of physical connection as well School of Materials Science and Engineering, as signaling transduction structures for regulating cell Nanyang Technological University, 50 Nanyang Avenue, behavior [19]. Cadherin is a key cell–cell adhesion mole- Nanyang 639798, Singapore cule localized at adherens junctions [20] which has a e-mail: assubbu@ntu.edu.sg 123 390 J Mater Sci: Mater Med (2011) 22:389–396 function similar to that of integrin, which may be consid- 2.2 Cell attachment and proliferation ered its counterpart in the FA complex. Integrins and cadherins are two distinct families of transmembrane cell HUVECs (Lonza) were cultured in endothelial growth adhesion receptors. While integrins allow cells to adhere to medium (EGM, Lonza) under 95% humidified atmosphere and 5% CO at 37C. Cells were dissociated with 0.025% the extracellular matrix, cadherins bind homotypically to cadherins on neighboring cells and are responsible for the trypsin–EDTA (Lonza) and washed in Dulbecco’s modified eagle medium (DMEM, Gibco) for 3 min by centrifugation development of adherens junctions in epithelial tissues. Arthur et al. [21] showed that the signaling cascades of to avoid any interference associated with adhesive proteins from serum, and then seeded on the various PLLA sub- both cell–matrix and cell–cell adhesion, transmitted through integrins and cadherins respectively, involve Rho strates in DMEM at either a low seeding density of 5,000 proteins, which are key regulators in reorganization of the cells/cm whereby cells are sparsely distributed on the actin cytoskeleton. Other studies have demonstrated cross- substrate, or a high seeding density of 50,000 cells/cm talk between cell-matrix and cell–cell junctions and both which is comparable to cell density at confluence. After 2 h types of junctions cooperatively regulate cell movement, incubation, unattached cells were gently rinsed off. The proliferation, adhesion and polarization [22–24]. number of attached cells was quantified by the WST-8 assay (Dojindo, Japan). Cell attachment percentage was quanti- Stimuli from neighboring cells via interaction of cell- surface receptors and secreted growth factors/cytokines are fied as N /N 9 100%, where N and N were 2h seeding 2h seeding the cell count at 2 h and the initial seeding respectively. The strongly dependent on the cell density. When the cell density is low, direct cell–cell contacts are limited and cell-bioma- cell count in each experimental condition was monitored on terial interaction is expected to be pre-dominantly influenced alternate days until confluence was reached. Cell doubling by cell-substrate contact. As cell density increases, cell–cell time of an exponential proliferation was calculated interaction becomes more extensive and is expected to according to the method as described previously [26]. profoundly influence cellular responses to biomaterials [25]. This study reports on the evaluation of the effects of initial 2.3 Mechanistic study on improved cell attachment seeding density on the EC-biomaterial interaction, by com- in high-density seeding paring a low seeding density of 5,000 cells/cm versus a high Two mechanisms are possible for the improvement of cell seeding density of 50,000 cells/cm using human umbilical vein endothelial cells (HUVECs). Three different PLLA attachment at high-density seeding. The secreted growth factors and cytokines by HUVECs at high-density seeding substrates, namely unmodified PLLA, PLLA–gAA–gelatin and PLLA–gAA–chitosan, were prepared as described pre- are more concentrated than that at low-density seeding. In order to study the influence of secreted growth factors and viously [26]. During the early phase of EC-biomaterial interaction, cell attachment and morphogenesis are the main cytokines on cell attachment, HUVECs of 5,000 cells/cm cellular responses, whereas cell proliferation is the main were seeded on pristine PLLA substrates in both fresh and cellular response during the later phase. conditioned medium for comparison. Conditioned medium here was prepared by incubating fresh EGM with confluent HUVECs in a tissue culture flask for 24 h, and then col- lected to centrifugation (3,600 rpm, 10 min) so as to 2 Materials and methods remove detached cells and debris. HUVECs were allowed to attach for 2 h. The cell attachment percentage was 2.1 Surface modification and characterization of PLLA studied using WST-8 assay as described earlier. It is also hypothesized that extensive cell–cell interaction PLLA (Purac Far East, Singapore) substrates were prepared between neighboring cells at higher density promotes cell as described previously [26]. Briefly, acrylic acids (AA, attachment. In order to evaluate this hypothesis, HUVECs Sigma-Aldrich) were graft polymerized on argon-plasma were seeded at a density of 5,000 cells/cm on both pristine treated PLLA surface. Gelatin (Type A, Sigma-Aldrich) and and pre-seeded PLLA. Pre-seeded PLLA here was prepared chitosan (Sigma-Aldrich) were then immobilized through by incubating HUVECs on PLLA substrates at a density of covalent bond formation between carboxylic groups found 25,000 cells/cm for 24 h. Cells were allowed to attach for on AA and amine groups found on gelatin or chitosan in 2 h, and the number of attached cells was quantified by the water soluble carbodiimide. PLLA–gAA–gelatin and WST-8 assay as described earlier. The cell attachment PLLA–gAA–chitosan refer to PLLA films modified with percentage on bare PLLA was calculated as N /5000 9 2h gelatin and chitosan respectively. Surface chemical com- 100%, where N was the cell density after 2 h incubation. 2h position and wettability were characterized by X-ray The cell attachment percentage on pre-seeded PLLA was Spectrometer and contact angle respectively [26]. calculated as (N -N /5000 9 100%, where N and N 2h pre) 2h pre 123 J Mater Sci: Mater Med (2011) 22:389–396 391 were the cell density after 2 h incubation and the cell density of pre-seeding HUVECs respectively. All experi- ments were carried out in 24-well tissue culture plate. 2.4 Immunofluorescence analysis Cells were fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 for 10 min, and incubated with 10% goat serum (DAKO) in PBS for 30 min at room temperature. Vinculin was labeled with mouse anti- human vinculin antibody (clone h Vin-1, Sigma-Aldrich) and visualized with goat anti-mouse IgG-FITC conjugate (Sigma-Aldrich). Actin filaments and nuclei were labeled with Alexa Fluor 568 (Molecular Probe) and DAPI (Molecular Probe) respectively. Cell images were captured with an Olympus IX71 inverted fluorescent microscope and Leica TCS SP5 laser-scanning spectral confocal micro- scope (CLSM). Cell morphometric parameters, spreading Fig. 1 a Extensive intercellular contact masks the surface depen- area and aspect ratio, were analyzed by the Image Pro Plus dence of cell adhesion. Surface-dependent cell adhesion was analyzed software using a modified method of Treiser et al.[27]. with ANOVA, *P \ 0.05. Density-dependent cell adhesion was analyzed with paired Student’s t-test, #P \ 0.05. b Growth factors and cytokines secretion showed no significant effect on cell attach- 2.5 Statistical analysis ment by comparing seeding in fresh medium (A) with that in conditioned medium (B). Pre-seeded cells on PLLA greatly promoted Three experimental replicates (n = 3) were utilized in the cell attachment (**P \ 0.05) by comparing seeding HUVEC on bare cell attachment and proliferation study. The results are PLLA (A) with that on 25,000cell/cm pre-seeded PLLA (C). After 2 h culture, unattached cells were removed by PBS rinsing. The presented as mean ± SD. The cell spreading area and number of remaining cells were measured by WST-8 reagent aspect ratio were plotted as box plots to show the data distribution and significant differences between measured biomolecules on the PLLA surface enhanced cell attach- groups. At least 50 cells were analyzed for each experi- mental group. Each box encompasses 25–75 percentiles, ment: 90% attachment was observed on PLLA–gAA–gel- atin and 50% on PLLA–gAA–chitosan. These observations with extending-lines covering the 95th and 5th percentiles, the thin line representing the median (50th percentile), agreed well with our previous study [26]. In contrast, the and the thick line representing the mean values. Values surface dependence of cell attachment was not observed at outside the 95th and 5th percentiles were treated as outliers, high seeding density of 50,000 cells/cm . Instead, all sub- and are represented by diamond dots. Density-dependent strates appeared to be equivalent: about 70% of seeded cellular responses were analyzed using Student’s t-test. cells were able to attach on all three PLLA substrates Surface-dependent cellular responses were analyzed using regardless of their differences in surface properties. High analysis of variance (ANOVA). A P value of less than 0.05 seeding density significantly improved cell attachment on was used to infer statistical significance of differences. PLLA and PLLA–gAA–chitosan ( P \ 0.05). Interest- ingly, cell attachment percentage on PLLA–gAA–gelatin decreased from 90 to 70%. However, the total cell count on 3 Results and discussions PLLA–gAA–gelatin in fact increased from 4,500 to 35,000. It is conceivable that high seeding density 3.1 Extensive intercellular contact masks the surface enhanced cell attachment during the early phase of cell- dependence of cell attachment biomaterial interaction and masked any differences in the substrate surface. Surface dependence of cell attachment Cell attachment percentage values were compared between was only observed at low seeding density by eliminating low and high seeding density experimental groups, with the intensive cell–cell interaction. results illustrated in Fig. 1a. In the case of 5,000 cells/cm The improvement of cell attachment at high seeding seeding density, the cell attachment percentage appeared density can possibly arise from two mechanisms: (a) In comparison to the low seeding density, the concentration of to be surface dependent (ANOVA, *P \ 0.05). After 2 h incubation, about 30% of seeded cells had adhered secreted growth factors and cytokines at high seeding on unmodified PLLA. The presence of immobilized density is higher, which could contribute to stronger cell 123 392 J Mater Sci: Mater Med (2011) 22:389–396 attachment signaling stimulation; (b) In comparison to the low seeding density, cell–cell interaction at high seeding density is more extensive, which could contribute to cell attachment signaling activation through crosstalk between cell-substrate and cell–cell adhesion. To investigate the first possible mechanism, conditioned medium which is the EGM incubated with HUVECs for 24 h was used to represent the medium at high density seeding. Cell attachment on PLLA at 5,000 cells/cm in conditioned medium was compared with that in fresh medium. As illustrated in Fig. 1b, cell attachment per- centage values were about 37% in both conditions. This implies that increased secretion of growth factors and cytokines at high seeding densities had negligible effect on cell attachment during 2 h incubation. To investigate the second possible mechanism, cell attachment on PLLA at 5,000 cells/cm was compared with that on pre-seeded PLLA. As shown in Fig. 1b, a signifi- cant increase of cell attachment percentage from 37 to 72% was observed by pre-seeding PLLA. This strongly sug- gested that the presence of pre-seeded HUVEC promoted the attachment of incoming cells. This simple test dem- onstrated a possible scenario occurring at high seeding density, whereby adherent cells hastened the attachment of cells from free suspension, through cell–cell interaction followed by triggered cell-substrate attachment. Another possible scenario is that high seeding density favors the formation of cell–cell adhesion in suspension even prior to cell-substrate attachment. Clumps of associated cells sed- iment and attach on the substrate simultaneously, such that cell attachment was significantly improved even on low cell affinity substrates. At present, we cannot distinguish between these two possible mechanisms. Zhu et al. [28, 29] showed that cell attachment on gelatin- and chitosan- modified PLLA (modified through aminolysis) was comparable at a seeding density of 12 9 4 2 10 cell/cm , which agrees with our results at high seeding density. However, our study demonstrated the superiority of cell attachment on PLLA–gAA–gelatin compared to Fig. 2 Surface chemistry, rather than seeding density, influences cell morphogenesis. HUVECs were seeded on PLLA (A and B), PLLA– PLLA–gAA–chitosan at a relatively low seeding density of gAA–gelatin (C and D), and PLLA–gAA–chitosan (E and F) at low 5,000 cell/cm . The data showed clearly that differences seeding density of 5000/cm (A, C and E) or high seeding density of among substrates are more pronounced at low seeding 50,000/cm (B, D and F). After 2 h culture, unattached cells were density, but was masked at high seeding density. This removed by PBS rinsing. The remaining cells were fluorescently stained to label nuclei (DAPI), vinculin (FITC) and F-actin (Alexa should be taken into consideration whenever biomaterials Fluor 568). The images were captured by Leica TCS SP5 CLSM. are to be evaluated for use in applications involving low Scale bar 20 lm cell densities. 3.2 Surface chemistry, rather than seeding density, differences upon comparing cells grown on the same sub- 2 2 influences cell morphogenesis strate at low (5,000 cells/cm ) and high (50,000 cells/cm ) seeding densities. On PLLA–gAA–gelatin (Fig. 2c,d), cells observed at both seeding densities exhibited the charac- After 2 h incubation, cell morphology was observed with immunostaining of F-actin, vinculin and nuclei. As shown teristic morphology of spreading HUVEC. Bundles of in Fig. 2, cell morphology did not exhibit distinctive F-actin were found in cell lamellipodia, which are the 123 J Mater Sci: Mater Med (2011) 22:389–396 393 broad, flat protrusions at the leading edge of a motile cell. Filopodia, the thin finger-like structures filled with tight parallel bundles of F-actin, were conspicuously protruding from the lamellipodia. Both of these are well-known structures involved in cell spreading and migration [30]. On PLLA–gAA–chitosan (Fig. 2e,f) seeded with either 5,000 or 50,000 cells/cm , the cells were observed to have significantly reduced spreading area in high-density seed- ing. This was further confirmed in the subsequent mor- phometric parameter analysis. Nevertheless, the typical spreading cell morphology was observed under both seeding density conditions, i.e. defined lamellipodia and filopodia with F-actin bundles localized mostly near the cell periphery. On PLLA (Fig. 2a,b), no obvious difference in cell morphology between low- and high-density seeding was found, other than the reduced cell spreading area. Compared with the cell morphology on gelatin- and chitosan-modified PLLA, much thinner F-actin bundles were observed beneath the cell membrane on bare PLLA and there was hardly any vinculin expression, thus indi- cating poor EC affinity for unmodified PLLA as demon- strated in our previous study [26]. Cell morphology was evaluated using two morphomet- ric parameters, cell spreading area and aspect ratio. Cell spreading area measures the extent of cell spreading on a substrate as projection area. As shown in Fig. 3a, the cell spreading area observed from a confluent EC monolayer grown on tissue culture plate (TCPS) was 1,400 lm on Fig. 3 Seeding density influenced (a) cell spreading area and (b) aspect ratio. Cell morphometric indicators observed on (A) TCPS, average. After 2 h incubation, cell spreading under any of (B and C) PLLA, (D and E) PLLA–gAA–gelatin, (F and G) PLLA– the experimental conditions was not able to reach that gAA–chitosan. Cells on TCPS were confluent and incubated for value. It is also noted that cell spreading areas from the 4 days. Low seeding density of 5000 cells/cm was used on B,D, high seeding density group were significantly smaller than F. High seeding density of 50,000 cells/cm was used on C, E, and G. High resolution immunofluorescent images of individual cells those observed in the low seeding density group. Spatial were taken by Leica TCS SP5 CLSM, and then analyzed by Image restriction is believed to be the main reason for the Pro Plus software. At least 50 cells were measured for each condition decrease in spreading area in the high seeding density group. Normal EC displays a typical ‘cobble-stone’ morphol- PLLA and PLLA–gAA–chitosan is believed to be the main ogy at confluence, with an epithelioid phenotype. By reason for the compromised cell morphology observed on contrast, when cells are sparse or when intercellular junc- both substrates. Unlike PLLA–gAA–gelatin, PLLA and PLLA–gAA–chitosan contained no cell attachment mole- tions are disrupted, a fibroblastoid/mesenchymal mor- phology predominates. Cell aspect ratio is defined as the cules/ligands on their surfaces. Cell attachment in serum- ratio of the major axis to the minor axis of an equivalent free medium relied largely on hydrophobic or electrostatic ellipse. In Fig. 3b, the cell aspect ratio of confluent EC interaction between substrate and attachment ligands on the monolayer grown on TCPS displaying cobble-stone mor- cell membrane. Amongst the high-density seeding experi- phology was about 2.38 on average. mental groups, cell aspect ratio on PLLA–gAA–gelatin Amongst the low-density seeding experimental group, was reduced to 1.90, mostly due to cell–cell contact inhi- the mean aspect ratio on PLLA–gAA–gelatin was about bition. The cell aspect ratio on PLLA and PLLA–gAA– 2.80, suggesting a mesenchymal phenotype. By contrast, chitosan was similar to that at low-density seeding. It aspect ratios on PLLA and PLLA–gAA–chitosan were suggests that the dominant factor for aspect ratio is surface about 1.5, which were significantly less than that of cobble- chemistry. The data presented in Fig. 3 suggest that cell morpho- stone morphology. This implied that although cells adhered on both surfaces, their cell morphology was somehow genesis during the early stage of cell-biomaterial interac- compromised. The absence of cell attachment ligands on tion was strongly dependent on substrate chemistry rather 123 394 J Mater Sci: Mater Med (2011) 22:389–396 than seeding density. However, seeding density did influ- ence some cell morphometric parameters. Reduced cell spreading area was observed on all three PLLA substrates and a less elongated cell phenotype was seen on PLLA– gAA–gelatin, at high seeding density. 3.3 Extensive intercellular contact masks surface dependence of cell proliferation When endothelial cells are sparsely-seeded in the sub- confluent state, they are actively proliferating and are sensitive to growth-factor stimulation. Once confluence has been achieved, the cells are contact-inhibited in their growth and protected from apoptosis [31]. Figure 4a shows the cell proliferation profiles on the various PLLA substrates as a function of culture duration, when HUVECs were sparsely seeded at a low density of 5,000 cell/cm . Cells grown on TCPS and PLLA–gAA– gelatin proliferated with a doubling time to about 15–16 h. In contrast, cell proliferation showed stagnation on PLLA– gAA–chitosan and even negative on PLLA during the first 72 h, with exponential growth taking place only after 72 h. Figure 4b illustrates the cell proliferation profiles of various PLLA substrates, at a high seeding density of 50,000 cells/cm . Cells plated on TCPS, PLLA–gAA– gelatin and PLLA–gAA–chitosan exhibited similar prolif- eration behavior. Cell proliferation started a few hours after seeding with a doubling time of about 7–9 h during the first 24 h. After 24–72 h, cell proliferation slowed down and eventually stopped. The number of HUVEC Fig. 4 Extensive intercellular contact masks surface dependence of on TCPS and PLLA–gAA–gelatin did not increase sig- cell proliferation. Cell proliferation was monitored over 7 days of nificantly after 72 h, suggesting that confluence was culture using (a) low seeding density of 5,000 cells/cm , and (b) high achieved and that cell–cell contact inhibited proliferation. seeding density of 50,000 cells/cm Cells grown on PLLA displayed an exponential prolif- eration pattern, which was different from the other three substrates. attachment and proliferation over PLLA. Clearly, PLLA– Comparing Fig. 4a to Fig. 4b, the profound impact of gAA–gelatin is a better candidate for such applications, a initial seeding density on cell proliferation is clearly evi- conclusion that could be reached only through studies at dent. First, low density seeding revealed the superiority of low seeding density. After confluence was reached, cell morphology was PLLA–gAA–gelatin in supporting cell proliferation com- pared to PLLA–gAA–chitosan, while extensive intercel- observed through cytoskeletal and focal adhesion immu- lular contact at high density seeding masked this difference nostaining (Fig. 5). It was found that cell morphology was in EC behavior. Second, by plating a confluent monolayer indistinguishable on the various PLLA surfaces at either at high seeding density, the proliferation stagnation on low (a, c and e) or high (b, d and f) seeding density. The PLLA–gAA–chitosan and PLLA was overcome. In appli- characteristic cobble-stone morphology was observed with cations involving the in situ endothelialization of cardio- bundles of stress fibers extending from the nucleus to vascular implants, such as coronary stents and patent lamellipodia across the entire cytoplasm and terminating at foramen ovale (PFO) occluders, the number of circulating FA and cell–cell junctions (Fig. 5g,h). During cell spreading and proliferation, cells produce their own ECs or EC-progenitor cells is relatively low [32]. At high seeding density, PLLA–gAA–gelatin and PLLA–gAA– extracellular matrix with time. The self-synthesized matrix may in turn mask the intrinsic effects of the original sub- chitosan showed equivalent performance in cell attachment and proliferation. However, at low seeding density, PLLA– strate on cell proliferation. The results of this study cor- gAA–chitosan showed only marginal improvement in cell roborate with earlier reports—that differences in cell 123 J Mater Sci: Mater Med (2011) 22:389–396 395 cell attachment behavior. Extensive intercellular interac- tions at high seeding density masked differences amongst substrates, and enhanced cell attachment on all substrates examined. During the later stage of cell-biomaterial inter- action, cell proliferation profile was found to be surface-depen- dent at low seeding density, whereas surface dependence was masked at high seeding density. It is recommended that low seeding density should be utilized for in vitro evaluation the compatibility of biomedical materials. Acknowledgments Authors would like to thank National Research Foundation of Singapore for funding the work with their Competitive Research Grant and Dr. Wong Yee Shan for critical reading. References 1. Bhattacharya V, Cleanthis M, Stansby G. Preventing vascular graft failure: Endothelial cell seeding and tissue engineering. Vasc Dis Prev. 2005;2:21–7. 2. Venkatraman SS, Boey F, Lao LL. Implanted cardiovascular polymers: Natural, synthetic and bio-inspired. Prog Polym Sci. 2008;33:853–74. 3. de Mel A, Jell G, Stevens MM, Seifalian AM. Biofunctional- ization of biomaterials for accelerated in situ endothelialization: a review. Biomacromolecules. 2008;9:2969–79. 4. Eisenbarth E, Velten D, Breme J. Biomimetic implant coatings. Biomol Eng. 2007;24:27–32. 5. Lu A, Sipehia R. Antithrombotic and fibrinolytic system of human endothelial cells seeded on PTFE: The effects of surface modification of PTFE by ammonia plasma treatment and ECM protein coatings. Biomaterials. 2001;22:1439–46. 6. Yang J, Bei J, Wang S. Enhanced cell affinity of poly (D, L- lactide) by combining plasma treatment with collagen anchorage. Biomaterials. 2002;23:2607–14. 7. Gumpenberger T, Heitz J, Ba ¨uerle D, Kahr H, Graz I, Romanin C, Svorcik V, Leisch F. Adhesion and proliferation of human Fig. 5 Substrate dependence of cell morphology disappeared at the endothelial cells on photochemically modified polytetrafluoro- point of confluence. Fluorescent micrographs of HUVEC were taken ethylene. Biomaterials. 2003;24:5139–44. on PLLA (A and B), PLLA–gAA–gelatin (C and D), and PLLA– 8. Miller DC, Thapa A, Haberstroh KM, Webster TJ. Endothelial gAA–chitosan (E and F). Low seeding density of 5,000 cells/cm was and vascular smooth muscle cell function on poly(lactic-co-gly- used in A, C and E, and high seeding density of 50,000 cells/cm was colic acid) with nano-structured surface features. Biomaterials. used in B, D and F.(G) Matured FA clusters were observed in 2004;25:53–61. HUVEC. (H) F-actin bundles terminated at FA clusters by colocal- ´ ´ 9. Berard X, Remy-Zolghadri M, Bourget C, Turner N, Bareille R, ization of F-actin and vinculin. The remaining cells were fluorescently Daculsi R, Bordenave L. Capability of human umbilical cord stained to label nuclei (DAPI), vinculin (FITC) and F-actin (Alexa blood progenitor-derived endothelial cells to form an efficient Fluor 568). The images were captured by Olympus IX71 inverted lining on a polyester vascular graft in vitro. Acta Biomater. microscopy. Scale bar 50 lm 2009;5:1147–57. 10. Boura C, Kerdjoudj H, Moby V, Vautier D, Dumas D, Schaaf P, behavior rising from differences in seeding conditions Voegel JC, Stoltz JF, Menu P. Initial adhesion of endothelial cells (substrate and/or density) are more apparent during the first on polyelectrolyte multilayer films. Bio-Med Mater Eng. 2006;16:S115–21. few days of cell–biomaterial interaction [33]. 11. Chen ZG, Wang PW, Wei B, Mo XM, Cui FZ. Electrospun collagen-chitosan nanofiber: A biomimetic extracellular matrix for endothelial cell and smooth muscle cell. Acta Biomater. 4 Conclusion 2010;6:372–82. 12. Crombez M, Chevallier P, Gaudreault RC, Petitclerc E, Manto- vani D, Laroche G. Improving arterial prosthesis neo-endotheli- This study has demonstrated the important role of seeding alization: Application of a proactive VEGF construct onto PTFE density in cell–biomaterial interaction. During the early surfaces. Biomaterials. 2005;26:7402–9. stage of cell–biomaterial interaction whereby cell attach- 13. Brewster LP, Bufallino D, Ucuzian A, Greisler HP. Growing a ment is the pre-dominant cellular activity, low seeding living blood vessel: Insights for the second hundred years. Bio- density enabled us to observe surface chemistry-dependent materials. 2007;28:5028–32. 123 396 J Mater Sci: Mater Med (2011) 22:389–396 14. Avci-Adali M, Paul A, Ziemer G, Wendel HP. New strategies for 23. Schwartz MA, Ginsberg MH. Networks and crosstalk: Integrin in vivo tissue engineering by mimicry of homing factors for self- signalling spreads. Nat Cell Biol. 2002;4:E65–8. endothelialisation of blood contacting materials. Biomaterials. 24. Geiger B, Bershadsky A, Pankov R, Yamada KM. Transmem- 2008;29:3936–45. brane crosstalk between the extracellular matrix and the cyto- 15. Rotmans JI, Heyligers MMJ, Verhagen HJM, Velema E, skeleton. Nat Rev Mol Cell Biol. 2001;2:793–805. Nagtegaal MM, de Kleijn DPV, de Groot FG, Stroes ESG, 25. Nagahara S, Matsuda T. Cell-substrate and cell-cell interactions Pasterkamp G. In vivo cell seeding with anti-CD34 antibodies differently regulate cytoskeletal and extracellular matrix protein successfully accelerates endothelialization but stimulates intimal gene expression. J Biomed Mater Res. 1996;32:677–86. hyperplasia in porcine arteriovenous expanded polytetrafluoro- 26. Xia Y, Boey F, Venkatraman SS. Surface modification of poly(L- ethylene grafts. Circulation. 2005;112:12–8. lactic acid) with biomolecules to promote endothelialization. 16. Hill JM, Zalos G, Halcox JPJ, Schenke WH, Waclawiwm MA, Biointerphases. 2010;5:FA32–40. Quyyumi AA, Finkel T. Circulating endothelial progenitor cells, 27. Treiser MD, Liu E, Dubin RA, Sung H-J, Kohn J, Moghe P. vascular function, and cardiovascular risk. N Engl J Med. Profiling cell-biomaterial interactions via cell-based fluorore- 2003;348:593–600. porter imaging. BioTechniques. 2007;43:361–8. 17. Lin A, Ding X, Qiu F, Song X, Fu G, Ji J. In situ endothelial- 28. Zhu Y, Gao C, Liu X, He T, Shen J. Immobilization of biomac- ization of intravascular stents coated with an anti-CD34 antibody romolecules onto aminolyzed poly(L-lactic acid) toward acceler- functionalized heparin-collagen multilayer. Biomaterials. 2010; ation of endothelium regeneration. Tissue Eng. 2004;10:53–61. 31:4017–25. 29. Zhu Y, Gao C, Liu Y, Shen J. Endothelial cell functions in vitro 18. Geiger B, Spatz JP, Bershadsky AD. Environmental sensing cultured on poly(L-lactic acid) membranes modified with differ- through focal adhesions. Nat Rev Mol Cell Biol. 2009;10:21–33. ent methods. J Biomed Mater Res A. 2004;69A:436–43. 19. Dejana E. Endothelial cell-cell junctions: happy together. Nat 30. Mattila PK, Lappalainen P. Filopodia: molecular architecture and Rev Mol Cell Biol. 2004;5:261–70. cellular functions. Nat Rev Mol Cell Biol. 2008;9:446–54. 20. Nelson WJ. Regulation of cell–cell adhesion by the cadherin– 31. Liebner S, Cavallaro U, Dejana E. The multiple languages of catenin complex. Biochem Soc Trans. 2008;36:149–55. endothelial cell-to-cell communication. Arterioscl Throm Vas. 21. Arthur WT, Noren NK, Keith B. Regulation of rho family 2006;26:1431–8. GTPases by cell-cell and cell-matrix adhesion. Biol Res. 2002;35: 32. Huang Y, Venkatraman SS, Boey FYC, Umashankar PR, Moh- 239–46. anty M, Arumugam S. The short-term effect on restenosis and 22. Sakamoto Y, Ogita H, Hirota T, Kawakatsu T, Fukuyama T, thrombosis of a cobalt-chromium stent eluting two drugs in a Yasumi M, Kanzaki N, Ozaki M, Takai Y. Interaction of integrin porcine coronary artery model. J Interv Cardiol. 2009;22:466–78. a b with nectin: implication in cross-talk between cell-matrix 33. Reilly GC, Engler AJ. Intrinsic extracellular matrix properties v 3 and cell-cell junctions. J Biol Chem. 2006;281:19631–44. regulate stem cell differentiation. J Biomech. 2010;43:55–62.

Journal

Journal of Materials Science: Materials in MedicineSpringer Journals

Published: Jan 8, 2011

There are no references for this article.