Cardiovascular Engineering and Technology, Vol. 9, No. 2, June 2018 ( 2016) pp. 181–192 https://doi.org/10.1007/s13239-016-0283-9 Intrinsic Cell Stress is Independent of Organization in Engineered Cell Sheets 1,2 1,2 3 1,2 INGE A.E.W. VAN LOOSDREGT, SYLVIA DEKKER, PATRICK W. ALFORD, CEES W.J. OOMENS, 1,2 1,2 SANDRA LOERAKKER, and CARLIJN V.C. BOUTEN Department of Biomedical Engineering, Eindhoven University of Technology, PO Box 513, 5600 MB Eindhoven, The Netherlands; Institute for Complex Molecular Systems, Eindhoven University of Technology, Eindhoven, The Netherlands; and Department of Biomedical Engineering, University of Minnesota, Minneapolis, MN, USA (Received 24 June 2016; accepted 11 October 2016; published online 24 October 2016) Associate Editor in Chief, Ajit P. Yoganathan and Guest Editors Hanjoong Jo and Craig Simmons oversaw the review of this article. Abstract—Understanding cell contractility is of fundamental INTRODUCTION importance for cardiovascular tissue engineering, due to its major impact on the tissue’s mechanical properties as well as Many cell types exert contractile stresses onto their 8,10,24,52 the development of permanent dimensional changes, e.g., by surroundings via stress ﬁbers. The main con- contraction or dilatation of the tissue. Previous attempts to tractile components of the stress ﬁbers are the actin quantify contractile cellular stresses mostly used strongly 8,39,52 ﬁbers and myosin motors. The stress ﬁbers can aligned monolayers of cells, which might not represent the be externally stimulated to exert stress (deﬁned here as actual organization in engineered cardiovascular tissues such as heart valves. In the present study, therefore, we investi- active cell stress) but also generate an intrinsic level of gated whether differences in organization affect the magni- stress without any external stimulation (deﬁned here tude of intrinsic stress generated by individual as intrinsic cell stress). Understanding and controlling myoﬁbroblasts, a frequently used cell source for in vitro the degree of cell stress is important for tissue engi- engineered heart valves. Four different monolayer organiza- neering in order to obtain mechanically functioning tions were created via micro-contact printing of ﬁbronectin lines on thin PDMS ﬁlms, ranging from strongly anisotropic tissues with a proper matrix organization. In cardio- to isotropic. Thin ﬁlm curvature, cell density, and actin stress vascular tissue engineering, for example, excessive ﬁber distribution were quantiﬁed, and subsequently, intrinsic (intrinsic) cellular stress can lead to tissue contraction, stress and contractility of the monolayers were determined by represented by leaﬂet shortening in case of tissue- incorporating these data into sample-speciﬁc ﬁnite element 11,30,44 engineered heart valves (TEHVs). Conversely, models. Our data indicate that the intrinsic stress exerted by the monolayers in each group correlates with cell density. insufﬁcient levels of (intrinsic) cellular stress can cause Additionally, after normalizing for cell density and tissue dilation, resulting in leaﬂet elongation in accounting for differences in alignment, no consistent differ- 42 TEHVs and aneurysm formation in vascular grafts. ences in intrinsic contractility were found between the The magnitude and direction of cellular stresses de- different monolayer organizations, suggesting that the intrin- pend on the (mechanical) environment, e.g., on stiff- sic stress exerted by individual myoﬁbroblasts is independent of the organization. Consequently, this study emphasizes the ness, applied strain, or architecture. For example, importance of choosing proper architectural properties for previous studies have demonstrated that stiff sub- scaffolds in cardiovascular tissue engineering, as these strates induce an increased development of stress ﬁ- directly affect the stresses in the tissue, which play a crucial bers and focal adhesions compared to soft substrates, role in both the functionality and remodeling of (engineered) enabling the cells to exert increased contractile stres- cardiovascular tissues. 8,10,19,29,51 ses onto their surroundings. Furthermore, cyclic uniaxial strain causes the cells and stress ﬁbers Keywords—Cardiovascular tissue engineering, Mechan- to orient perpendicular to the strain; a phenomenon otransduction, Cell alignment, Finite element modeling. 16,25,36 known as strain avoidance behavior. This cau- ses the cells to exert stresses in the direction perpen- dicular to the strain, changing the stress Address correspondence to Sandra Loerakker, Institute for directionality. Another phenomenon that affects cell Complex Molecular Systems, Eindhoven University of Technology, orientation is contact guidance, where the cells align Eindhoven, The Netherlands. Electronic mail: firstname.lastname@example.org 1869-408X/18/0600-0181/0 2016 The Author(s). This article is published with open access at Springerlink.com 182 VAN LOOSDREGT et al. in the direction of topographical environmental stress distribution inside these engineered tissues. In cues. fact, previous studies with cardiomyocytes have shown Numerous studies have been performed to improve that the stress developed by a complete monolayer 13,26,32 our understanding of stress ﬁber remodeling and cellular increases upon increasing cellular anisotropy. stress development, both in 2D and in 3D environments. However, these studies have not investigated whether In 3D, most studies have investigated the compaction of this was due to differences in intrinsic contractility of ﬁbroblast-loaded gels as a measure of cell contractil- the cells, or rather due to differences in alignment. 4,6,7,27 ity. In addition, studies have been performed to We hypothesize that the intrinsic contractility of investigate the inﬂuence of external cues, such as stiff- individual cells is independent of the monolayer 29 16 22 ness, cyclic strain, soluble factors or combinations organization, which would implicate that the intrinsic 17,27 of stimuli, on the actin organization in relation to stress generated by the complete monolayer will be compaction. Cellular stress development has also been dictated by the cellular organization only. To investi- examined in endogenously produced extracellular gate this hypothesis, we will focus on myoﬁbroblasts 44,46–48 matrices. Even though 3D studies are more rele- derived from the saphenous vein in this study as this vant for tissue engineering, extracting (individual) cell cell type is commonly used for creating tissue engi- 11,33,34,38,49 stress is challenging due to the complexity of the envi- neered cardiovascular tissues. We adapted ronment in which the cells reside. 2D studies with cells the thin ﬁlm method developed by Feinberg et al., 1 21 cultured on substrates may therefore be more suit- Alford et al. and Grosberg et al. to determine the able for providing the fundamental insights into stress intrinsic stress developed by a monolayer of myoﬁ- ﬁber remodeling and cellular stress development. broblasts with various degrees of alignment. The dif- Various experimental methods exist for measuring ferences in alignment were obtained by seeding the cell contractility in 2D, ranging from single cell cells onto micro-contact printed ﬁbronectin patterns 39,53 methods like traction force microscopy and set-ups with different orientations with respect to the long axis using micropost arrays, to monolayer methods such of the ﬁlms. The thin ﬁlm method was combined with as cyclic stretching of cell monolayers on ﬂexible sub- live imaging to determine the curvature and cell density 25 1,21 strates and the thin ﬁlm method. The latter of each individual ﬁlm. Separate samples were used to method is a suitable method for quantifying both stress stain the cell nuclei, F-actin and phosphorylated ﬁber remodeling and stress development with minimal myosin light chain, to determine nuclear and stress ﬁ- handling of the cells. ber organization and provide insight into intrinsic The thin ﬁlm consists of a thin layer of polydimethyl- stress ﬁber contraction potential. These experimental siloxane (PDMS) that is attached to a glass substrate via results were then combined with sample-speciﬁc ﬁnite the temperature sensitive polymer poly-N-isopropy- element modeling to determine the intrinsic stress ex- 14,20 lacrylamide (pIPAAm). The PDMS is subsequently erted in the ﬁlm direction by the complete monolayer micro-contact printed with lines of extracellular matrix and the normalized intrinsic cellular contractility. proteins toenhancecelladhesion and guide the cells intoa speciﬁc direction. Rectangular ﬁlmsare typicallycut from MATERIALS AND METHODS the PDMS, which partly release from the glass when the pIPAAM dissolves upon a decrease in cell culture med- Construct Fabrication 1,14,20 ium temperature. The curvature of the ﬁlms as a Thin ﬁlm constructs were fabricated as previously result of the contractile cell layer on top of the PDMS can 1,2,20 then be used to quantify the stress exerted by the mono- described. In brief, a layer of poly-N-isopropy- layer of cells. The thin ﬁlm method is therefore an elegant lacrylamide (pIPAAm; Sigma, Zwijndrecht, The method for determining the contractile properties of Netherlands) and a layer of polydimethylsiloxane 2,50,53 aligned contractile tissues, such as smooth, skele- (PDMS; Sylgard 184; Dow Corning, Auburn, MI) 40 21,31 tal or cardiac muscle tissue. were spin coated on a 25 mm diameter glass cover slip Tissue engineered heart valves and blood vessels can and cured overnight at 65 C. 2.5% of blue silicon dye 5,34,37,43 be created using (electrospun) ﬁbrous scaﬀolds (Silc-Pig; Smooth-On, Macungie, PA) was added to that allow for cell inﬁltration because of their high the PDMS in order to visualize the ﬁlms and different porosity. In this case the initial cell alignment is rotation speeds were used to create PDMS ﬁlms with determined by the scaffold ﬁber organization via the different thicknesses in order to account for the dif- mechanism of contact guidance. As the scaffold ﬁber ference in ﬁlm curvature between the different mono- organization is highly variable and never perfectly layer organizations. In addition, PDMS was spin aligned, the cell organization is also not perfectly coated on copper coated glass cover slips used for aligned, and it may be questioned whether the stress thickness measurements with an optical proﬁlometer exerted by perfectly aligned monolayers represents the (Pll 2300; Sensofar, Terrassa, Spain). In order to Intrinsic Cell Stress is Independent of Organization 183 determine the elastic modulus of the dyed PDMS, Netherlands) for 15 minutes, permeabilized with 0.5% rectangular bars were uniaxially strained with a tensile Triton-X 100 (Merck) for 5 minutes and subsequently tester (Z010; Zwick/Roell, Venlo, The Netherlands). incubated for 90 minutes with 2% BSA (Roche, Al- mere, The Netherlands)—1% horse serum (Sigma) in TBS supplemented with 0.05% Tween (Merck) to Micro-Contact Printing block non-speciﬁc binding. Mouse anti-phospho- PDMS stamps were fabricated using standard myosin light chain IIA (Cell Signaling, Danvers, MA) photolithography techniques. The stamps contain was used to label phosphorylated myosin IIA over- features of either 10 lm wide lines with 10 lm spacing night before addition of biotin labeled horse-anti- in between or a ﬁshnet pattern with 5 lm wide lines mouse secondary antibody (Vector, Burlingame, CA) with 10 lm spacing at an angle of ±15,30 or 45 for 90 minutes. Thereafter, the samples were incubated with respect to the 0 axis (Fig. 1). PDMS stamps were with streptavidin-Alexa 647 (Invitrogen) and phal- incubated with 50 mg/mL rhodamine ﬁbronectin (Cy- loidin-Atto 488 (Sigma) for 90 minutes. Before toskeleton, Denver, CO) in PBS for one hour, after mounting with mowiol (Sigma), the samples were which they were dried using compressed air. The thin incubated with DAPI (Sigma) for 10 minutes. The ﬁlm constructs were treated with UV-ozone (PDS UV- mounted samples were visualized using both ﬂuores- ozone cleaner; Novascan, Ames, IA) for 8 minutes just cent (Axiovert 200 M; Zeiss, Sliedrecht, The Nether- before transfer of the ﬁbronectin onto the constructs. lands) and confocal microscopy (LSM 510 Meta; The stamps were positioned in such a way that the 0 Zeiss). Fluorescent microscopy images at 20 times axis of the stamp coincided with the length direction of magniﬁcation were analyzed using custom Matlab the to be cut ﬁlms. After 10 minutes of conformal (MathWorks, Natick, MA) scripts to determine the contact, the constructs were rinsed three times with actin ﬁber and nuclear orientation with respect to the 1,23 PBS and stored in PBS at 4 C until use. ﬁbronectin orientation. The actin ﬁber orientation was 16–18 determined as described previously. For each image a histogram containing the ﬁber percentage per Cell Seeding and Culture angle was obtained. The actin stress ﬁber distribution Human myoﬁbroblasts were harvested from the was subsequently quantiﬁed by ﬁtting the following vena saphena magna obtained from patients according curve to each histogram: to Dutch guidelines of secondary used material and ! ! were seeded at passage 7 onto the thin ﬁlm constructs ðÞ c l u ¼Ac þ exp ð1Þ sf at a seeding density of 8400 cells/cm . After seeding, 2r the HVSCs were cultured at 37 C and 5% CO for 2 days in growth medium consisting of Advanced with the main ﬁber direction (l) in the 0 direction, c DMEM (Invitrogen, Breda, The Netherlands) supple- the ﬁber angle and A a scaling factor. The offset (c) mented with 10% Fetal Bovine Serum (Greiner Bio- and dispersity (r) were ﬁt and used as parameters in the ﬁnite element model described below (Table 1). One), 1% GlutaMax (Invitrogen) and 1% penicillin/ The nuclear orientation was quantiﬁed by thresh- streptomycin (Lonza, Basel, Switzerland). olding the DAPI image and ﬁtting an ellipse through each nucleus after which the angle of the major axis of Cell Orientation Analysis the ellipse was determined. As for the actin ﬁbers, After culture, half of the samples were ﬁxated in histograms were constructed containing the nuclear 3.7% formaldehyde (Merck, Schiphol-Rijk, The orientation percentage per angle. The nuclear aspect FIGURE 1. Schematic overview of the micro-contact printing layout of the four different ﬁbronectin patterns. The ﬁbronectin lines are depicted in grey, the spacing in black and the angle a is depicted in the top left corner. The short arrows represent 5 lm and the long arrows represent 10 lm. 184 VAN LOOSDREGT et al. TABLE 1. Overview of speciﬁc and common parameters order to dissolve the pIPAAm, and enable the HVSCs used in the ﬁnite element model. to deform the PDMS layer. A picture of the initial curvature (0 h) was taken at room temperature using a Fibronectin angle t (lm) r () c (–) pdms stereomicroscope (Discovery.V8; Zeiss) after which the 0 8.2 13 0.065 thin ﬁlms were placed back at 37 C and 5% CO . 15 8.2 18 0.086 Another picture was taken after 1 h when the con- 30 7.1 23 0.138 tractile equilibrium was reached. 45 6.8 54 2.091 Common parameters Analysis of Intrinsic Cell Stress t (lm) 3.2 cell The curvature of the ﬁlms was determined using E (kPa) 0.7 cell Matlab by analyzing the projection length of the bent m (–) 0.3 cell ﬁlms. The length and width of each ﬁlm were obtained E (MPa) 1.52/1.91 pdms m (–) 0.49 pdms from images of the undeformed ﬁlms. Thereafter the intrinsic cellular stress was obtained via sample-speciﬁc ﬁnite element modeling in Abaqus (Dassault Syste` mes ratio was calculated by dividing the length of the major Simulia Corp., Providence, RI). The cell (t ) and PDMS cell axis by the length of the minor axis. Phosphorylated (t ) thickness (Table 1), and the length and width of pdms myosin light chain was visualized together with actin at each ﬁlm were used as geometrical input for creating a 63 times magniﬁcation, to investigate potential co-lo- double-layered ﬁnite element mesh. Both layers consisted calization of the two major stress ﬁber components. of 200 quadratic brick elements (C3D20), with the bot- The monolayer thickness was determined by analyzing tom layer representing PDMS and the top layer repre- the z-stacks obtained at 40 times magniﬁcation with 2 senting the cell monolayer. The PDMS layer was ﬁxed at confocal microscopy, and used as a parameter in the one of the short edges torepresent the experimentalsetup. ﬁnite element model (Table 1). ThePDMSlayer was assignedwithcompressibleNeo- Hookean material properties: Intrinsic Cell Stress Assay ln J G 2=3 r ¼ j I þ B J I ð2Þ After culture, the other half of the samples was J J stained with Hoechst (10 lg/ml; Invitrogen) for with shear modulus G = E/2(1 m), compression 15 minutes, subsequently rinsed 3 times with PBS and modulus j = 2G(1 + m)/3(1 2m), B = FÆF and growth medium was added to the samples until further J = det(F), where F represents the deformation gradi- use. The stained nuclei were used to determine the ent tensor. Parameter values are indicated in Table 1. nuclear aspect ratio and orientation with respect to the The cell layer of the model was a ﬁber-reinforced ﬁbronectin lines on each individual ﬁlm (as described layer with an active, ﬁbrous, component (r ) repre- ca above). In addition, the cell density (d) was determined senting the stress ﬁbers, and a passive, compressible by counting the nuclei in these images. After staining Neo-Hookean, component (r ) representing the other cp the nuclei, the samples were transferred to a Petri dish cellular components. r was calculated by assuming cp with preheated growth medium. The long edges of Neo-Hookean material behavior (Eq. (2); Table 1), eight rectangular ﬁlms were cut from the constructs while r is determined from the stress exerted by the ca and the excess PDMS was removed. The petri dish stress ﬁbers in a range of different directions: containing the sample was then transferred to a con- focal microscope (TCS SP5X; Leica, Son, The N i i * * Netherlands) to perform temperature- and CO -con- r ¼ u r e e ð3Þ ca max sf sf sf i¼1 trolled (37 C, 5% CO ) live imaging of the nuclei and ﬁbronectin lines on the ﬁlms. Hoechst was excited with where e is the stress ﬁber direction in the deformed sf a femtosecond pulsed laser (Chameleon; Coherent, conﬁguration, r is a measure for intrinsic cell con- max Santa Clara, CA) at 750 nm and a laser power of 10%. tractility, and u the actin stress ﬁber volume fraction sf Rhodamine ﬁbronectin was excited with a white light for each direction as obtained from the ﬂuorescent laser (Leica) at 535 nm and a laser power of 14%. images. r was iteratively increased until the curva- max 1024 9 1024 pixel images were taken with a scan speed ture of the ﬁnite element models matched the experi- of 400 Hz. We did not see any adverse effects of the mentally obtained curvature. The total intrinsic stress in imaging procedure on the cells (data not shown). Next, the cell layer of the model was obtained by adding the the ends of the rectangular ﬁlms were cut while the passive and active stress components: r = r + r . c cp ca medium was allowed to cool down below 32 Cin Intrinsic Cell Stress is Independent of Organization 185 The magnitude of the intrinsic stress component in the diﬀerent from each other (p < 0.002), except for the aspect ratio of the nuclei on the 30 and 45 ﬁbronectin lines. long axis direction of the deformed ﬁlm (e ) in the cell la layer (r ) was determined using: * * Curvature Increases with Increasing Cell Density r ¼ r e e : ð4Þ f c la la For all groups, 32 ﬁlms were manufactured to per- In order to compare intrinsic cell contractility between form curvature measurements on. In case of disconti- samples, we normalized r for the cell density max nuities in the ﬁbronectin pattern, the ﬁlm was not max included in the analysis. This resulted in analysis of r ¼ : ð5Þ norm respectively 32, 31, 29, and 26 ﬁlms for the 0,15,30, and 45 groups. Considerable diﬀerences in curvature between ﬁlms within the same group were present Statistical Analysis (Fig. 4a), due to local variations in cell density. Therefore, the number of nuclei on each ﬁlm was Quantitative data were analyzed with SPSS Statis- quantiﬁed via a Hoechst staining (Figs. 4b, 4c) in order tics 22 (IBM, Amsterdam, The Netherlands) and were to determine the correlation between cell density and considered signiﬁcant at p < 0.05. Differences in nu- ﬁlm curvature for each group at both time points. For clear aspect ratio were analyzed using a one-way all groups, positive correlations between cell density ANOVA with a Bonferroni post hoc test. Spearman’s and curvature were found, except for the 45 group at correlation coefﬁcient (q) was determined to investi- 0 h (Fig. 4d–4g). In addition, a minimum cell density gate correlations between cell density and either cur- was required for the cells to be able to bend the ﬁlm, vature, r or r . Differences in r between the f max norm which approximately equaled 150–200 cells/mm . four different alignment groups were analyzed using a non-parametric Kruskal–Wallis test with pairwise Wilcoxon rank sum tests with corrected levels as post Normalized Intrinsic Contractility Seems Independent of hoc analysis. Monolayer Alignment When ﬁlm curvature was absent, the intrinsic stress exerted by the monolayer was lower than the measure- RESULTS ment limit of this method. In that case, ﬁnite element simulations were omitted. Simulations were performed Monolayer Organization is Determined by the for the remaining 151 ﬁlms (both time points included). Fibronectin Pattern 26 simulations failed before reaching the experimentally The nuclear and actin orientation of the cells on observed curvature due to convergence issues, and were substrates with diﬀerent orientations of ﬁbronectin excluded from further analysis. Due to the lack of lines is shown in Fig. 2. For all groups, the orientation remaining samples with a cell density above 300 cells/ of the actin ﬁbers, the nuclei of the samples used for mm in the 0 group, no clear correlation was found staining, and the nuclei of the samples used for the between the cell density and the intrinsic stress in the stress measurements coincided. This indicates that direction of the ﬁlm (r ) for this group. For 15 samples both the actin orientation and the nuclear orientation there was a signiﬁcant correlation between the cell are a good measure for overall cellular orientation. The density and r . A similar correlation was observed for cells of the 0 group were primarily aligned in the the 30 samples, albeit with larger dispersity. For the direction of the ﬁbronectin lines, as demonstrated by samples with ±45 ﬁbronectin lines a signiﬁcant corre- the high peak in the orientation histograms at 0 for lation was observed at 0 h, however not as strong as for both the actin ﬁbers and nuclei (Figs. 2a and 2e). Upon the 15 and 30 samples, probably also due to the lack increasing the angle between the ﬁbronectin lines, the of samples with a high cell density. As the intrinsic stress peak at 0 ﬂattened out until a completely random in all directions is taken into account in r (Eq. (3)), max orientation was reached in the 45 group (Fig. 2dand its value was higher than r , which only includes the 2h). The mean actin ﬁber distribution of each group intrinsic stress in the direction of the ﬁlm (Eq. (4)). was successfully ﬁt using Eq. (2) (Fig. 3). The obtained Naturally, the difference between r and r increased max f parameters are shown in Table 1, and served as input with increasing ﬁbronectin angle. No signiﬁcant differ- for the computational model. ences in r were found between groups with the norm The nuclear aspect ratio was determined to be exception of the 30 group being signiﬁcantly higher 1.74 ± 0.09, 1.65 ± 0.05, 1.61 ± 0.05, and 1.60 ± 0.06 compared to the 0 group at 0 and 1 h. The median for the 0,15,30,and 45 group, respectively. This r ranged between 3.43 and 6.76 Pa at 0 h and norm indicates that the nuclei were all elliptical and signiﬁcantly between 4.80 and 8.76 Pa at 1 h (Fig. 6). 186 VAN LOOSDREGT et al. FIGURE 2. Actin (green) and nuclei (blue) of myoﬁbroblasts cultured on ﬁbronectin lines (grey) with four different orientation angles (a–d; scale bar is 50 lm) and corresponding histograms of the actin and nuclear orientation (e–h; mean 6 standard error of mean). The green markers represent the actin ﬁbers (40 images), blue markers represent the nuclei of the stained samples (40 images) and the red markers represent the nuclei of the samples used for stress measurements (26–32 ﬁlms). Stress Fiber Organization is Similar in All Groups present and oriented along the longitudinal direction of the cells. Phosphorylated myosin light chain was Stainings for actin and phosphorylated myosin light observed to co-localize with the actin ﬁbers, conﬁrming chain, the major stress ﬁber components, are shown in the ability of the stress ﬁbers to contract. Fig. 7. In all groups, actin ﬁbers were abundantly Intrinsic Cell Stress is Independent of Organization 187 DISCUSSION of the tissue. In previous attempts to quantify the contractile cellular stresses by means of the thin ﬁlm 2,21,31,40,50,53 Understanding cell contractility is of fundamental method, mostly strongly aligned mono- importance for cardiovascular tissue engineering, due layers of cells were used, which might not represent the to its major impact on the tissue’s mechanical prop- actual cellular organization in engineered cardiovas- erties as well as the development of permanent cular tissues. In the present study, we investigated dimensional changes, e.g., by contraction or dilatation whether differences in alignment would affect the magnitude of the intrinsic stress generated by individ- ual myoﬁbroblasts. We hypothesized that the intrinsic contractile stress exerted by the myoﬁbroblasts is independent of the monolayer organization, as a result of which the total intrinsic stress exerted by the monolayer should be dictated by the actual cell align- ment. To test our hypothesis, patterns of ﬁbronectin lines were micro-contact printed on thin ﬁlm con- structs in order to create monolayers with varying degrees of cell alignment. The intrinsic stress exerted by each monolayer in the direction of the ﬁlm was determined from the curvature of the thin ﬁlms, and was found to correlate positively with the cell density. Importantly, after accounting for differences in cell alignment and normalizing for cell density, no consis- tent differences in intrinsic cellular contractility were found between the different monolayer organizations, suggesting that the intrinsic stress exerted by mono- FIGURE 3. Mean actin orientation (grey triangles) with cor- layers of myoﬁbroblasts can indeed be predicted from responding ﬁt (black lines) for 0 (upward triangles; dotted line), 15 (downward triangles; dash dot line), 30 (left point- the cellular organization. These ﬁndings are supported ing triangles; dashed line), and 45 (right pointing triangles; by the similarity in staining for stress ﬁber organization solid line). observed in the different groups. FIGURE 4. Representative top view image of bent ﬁlms with 30 ﬁbronectin lines at equilibrium (1 h), the black bars represent the projection length and the white bars the initial length; scale bar is 1 mm (a). Examples of confocal images of nuclei (blue) on ﬁbronectin lines (grey) at 0 (b) and 645 (c); scale bar is 50 lm. Density-curvature plots of the ﬁlms at 0 h (red triangles) and 1 h (black diamonds) for the 0 (d), 15 (e), 30 (f), and 45 (g) groups. Spearman’s correlation coefﬁcient is depicted in the top left corner for each density-curvature plot (n 5 26–32). *p < 0.05, **p< 0.01. 188 VAN LOOSDREGT et al. FIGURE 5. Density–stress plots for the stress in the length direction of the ﬁlm (r ; a–d) and for the maximum stress ﬁber stress (r ; e–h) for the 0 (a, e), 15 (b, f), 30 (c, g), and 45 (d, h) group at 0 h (red triangles) and 1 h (black diamonds). Spearman’s max correlation coefﬁcient is depicted in the top left corner of each density-stress plot (n 5 13-25). *p < 0.05, **p < 0.01. larger than 1.60 indicating the presence of elliptical ** nuclear shapes and thus elongated cells (Fig. 2a–d). ** The nuclear aspect ratio increased upon increasing 30 cellular alignment, suggesting that the cells adopted a more elongated shape for higher degrees of cellular alignment. The minimum cell density that was required to in- duce signiﬁcant curvature of the ﬁlms was 150–200 cells/mm , regardless of the cellular organization (Fig. 4). In a study that investigated collagen gel compaction by osteoblasts, a comparable threshold 2 15 value of 100 cells/mm was found. When the threshold density was exceeded, both the intrinsic 0 h 1 h 0 h 1 h 0 h 1 h 0 h 1 h stress in the direction of the ﬁlm (r ) and the measure 0° 15° 30° 45° for intrinsic cell contractility (r ) correlated with cell max density and increased over time (Fig. 5). These corre- FIGURE 6. r normalized for cell density (r )at0 h max norm lations were less strong in the 0 and 45 group, (red) and 1 h (black), n 5 13–25. **p< 0.01. probably due to the low number of samples with a high Using a simple and controlled method consisting of cell density (> 300 cells/mm ). Few studies have been micro-contact printing diﬀerent ﬁbronectin patterns, published on the effect of cell density on contraction cell sheets with organizations ranging from highly using gel compaction assays without a predeﬁned cel- 9,12,15,35 aligned to completely random were successfully cre- lular organization. Similar to the results ated. As the actin ﬁber orientation and the nuclear obtained in our study, they found that the initial orientation were overlapping (Fig. 2e–2h), both compaction is higher in high-density gels compared to appeared to be good indicators of cell orientation and low-density gels. Moreover, the high-density gels also organization. Since previous research has shown that reached the maximum compaction at a faster rate the nuclear aspect ratio is correlated with the cellular compared to the low-density gels, although the actual 2,50,53 aspect ratio, we used this measure as an indicator maximum was the same for both types of gels. It is of cellular shape. In all conditions, the aspect ratio was however unclear if the maximum degree of compaction σ [Pa] norm Intrinsic Cell Stress is Independent of Organization 189 FIGURE 7. Representative ﬂuorescent microscopy images of actin (green), phosphorylated myosin light chain (red), ﬁbronectin (grey), and nuclei (blue). The angle of the ﬁbronectin lines is 0 in (a–c), 15 in (d–f), 30 in (g–i), and 45 in (j–l). Merged images are shown on the top row (a, d, g, j), actin and nuclei are shown on the middle row (b, e, h, k), and phosphorylated myosin light chain and nuclei are shown on the bottom row (c, f, i, l). Scale bar is 50 lm. in these gels is caused by direct cellular contractility when the cells are aligned compared to a random cell only. organization. In addition to that, similar to this study, When correcting the intrinsic stress exerted by the Knight et al. recently investigated multiple degrees of monolayer for diﬀerences in cell density and cellular anisotropy, demonstrating a doubling of the stress in organization, signiﬁcant diﬀerences in intrinsic cell con- the ﬁlm direction of anisotropic myocardial tissue tractility (r ; Fig. 6) were only found between the 0 compared to isotropic tissue, with a gradual decrease norm and 30 degree groups, which may be explained by the in global stress with decreasing anisotropy. However, combined effects of the high spread in r at the 30 as these stresses were not corrected for differences in norm group, the lack of samples with a high cell density in the 0 alignment and cell density, it remains unclear whether group, and the lack of low cell density samples in the 30 the stress generated by individual cardiomyocytes de- group. Taken together, no consistent signiﬁcant differ- pends on the local or global cell alignment. This ences in the normalized stress were observed between the uncertainty is even more emphasized by the fact that different groups (with the median r ranging from two studies that have normalized the globally found norm 3.43 to 8.76 Pa). Therefore, our data suggest that the cardiac tissue stress are contradictory to each other, intrinsic stress exerted by individual myoﬁbroblasts is where Feinberg et al. concluded that the force gen- independent of the monolayer organization. The actin erated by individual sarcomeres increases upon and phosphorylated myosin light chain staining support increasing alignment, and Van Spreeuwel et al. found this ﬁnding as no differences in the stress ﬁber organiza- that cardiomyocytes in both anisotropic and isotropic tion were found between groups (Fig. 7). tissues exert similar amounts of force. Previous studies that have investigated the rela- A limitation of the current study is the presence of a tionship between cell organization and stress develop- spatial variability in PDMS thickness between samples ment have mainly focused on myocardial that were manufactured with the same settings. As the 13,26,32,40,45 tissues. Most of these studies observed that stress that is necessary to bend the thin ﬁlm is strongly the stress developed by the complete tissue is higher dependent on the magnitude of the thickness, this 190 VAN LOOSDREGT et al. variability may have induced some uncertainty in the HUMAN STUDIES calculated stresses. Furthermore, it resulted in a de- No human studies were carried out by the authors crease in sample size of r , r , and r compared f max norm for this article. to the curvature data, due to the fact that in case of 0mm curvature the stresses could not be quantiﬁed as a result of low cell densities in combination with ANIMAL STUDIES relatively high local PDMS thickness. In addition, by normalizing the calculated intrinsic stress for cell No animal studies were carried out by the authors density, we assumed that all cells were exerting their for this article. intrinsic stress solely onto the PDMS layer, without pulling on their neighboring cells via cell–cell contacts. Future studies should point out whether this assump- OPEN ACCESS tion is completely valid. This article is distributed under the terms of the In summary, the results of our study suggest that the Creative Commons Attribution 4.0 International individual intrinsic contractility of myoﬁbroblasts is License (http://creativecommons.org/licenses/by/4.0/), independent of monolayer architecture, implying that which permits unrestricted use, distribution, and re- the architecture itself dictates the total intrinsic stress production in any medium, provided you give appro- distribution in the tissue. With regard to cardiovascular priate credit to the original author(s) and the source, tissue engineering, the initial organization of engineered provide a link to the Creative Commons license, and tissues is often imposed via the presence of ﬁbrous indicate if changes were made. 5,34,37,43 scaﬀolds. These scaffolds are essential in deliv- ering the correct material properties that will induce 3,30 physiological tissue deformations. As the scaffold architecture directly determines the direction in which REFERENCES individual cells exert stress and the magnitude of this stress is not affected by the architecture, the total stress Alford, P. W., A. W. Feinberg, S. P. Sheehy, and K. K. distribution in the tissue, generated by all cells in that Parker. Biohybrid thin ﬁlms for measuring contractility in tissue, is completely determined by the scaffold archi- engineered cardiovascular muscle. Biomaterials 31:3613– tecture. Hence, the fact that the intrinsic stress gener- 3621, 2010. Alford, P. W., A. P. Nesmith, J. N. Seywerd, A. Grosberg, ated by individual cells remains the same implies that and K. K. Parker. Vascular smooth muscle contractility the total intrinsic stress distribution can directly be al- depends on cell shape. Integr. Biol. 3:1063–1070, 2011. tered by changing the architecture of the scaffold. This Argento, G., M. Simonet, C. W. Oomens, and F. P. plays a crucial role in both the functionality and Baaijens. Multi-scale mechanical characterization of scaf- remodeling of (engineered) cardiovascular tissues. folds for heart valve tissue engineering. J. Biomech. 45:2893–2898, 2012. Balestrini, J. L., and K. L. Billiar. Magnitude and duration of stretch modulate ﬁbroblast remodeling. J. Biomech. Eng. 131:051005, 2009. Bouten, C. V., P. Y. Dankers, A. Driessen-Mol, S. Pedron, ACKNOWLEDGMENTS A. M. Brizard, and F. P. Baaijens. Substrates for cardio- vascular tissue engineering. Adv. Drug Deliv. Rev. 63:221– The authors thank Kerianne Steucke, Eric Hald and 241, 2011. Zaw Win for their assistance with the thin ﬁlm method. Brown, R. A. In the beginning there were soft collagen-cell gels: towards better 3D connective tissue models? Exp. Cell This work was supported by a grant from the Dutch Res. 319:2460–2469, 2013. Government to the Netherlands Institute for Regen- Brown, R. A., R. Prajapati, D. A. McGrouther, I. V. erative Medicine (NIRM Grant No. FES0908). Yannas, and M. Eastwood. Tensional homeostasis in dermal ﬁbroblasts: mechanical responses to mechanical loading in three-dimensional substrates. J. Cell. Physiol. 175:323–332, 1998. CONFLICT OF INTEREST Chen, C. S. Mechanotransduction—a ﬁeld pulling to- I. van Loosdregt, S. Dekker, C. Oomens, S. Loer- gether? J. Cell Sci. 121:3285–3292, 2008. Chieh, H. F., Y. Sun, J. D. Liao, F. C. Su, C. Zhao, P. C. akker and C. Bouten declare no conﬂict of interest. P. Amadio, and K. N. An. Effects of cell concentration and Alford has a patent ‘Devices comprising muscle thin collagen concentration on contraction kinetics and ﬁlms and uses thereof in high throughput assays for mechanical properties in a bone marrow stromal cell-colla- determining contractile function’, US 9 012 172 B2 gen construct. J. Biomed. Mater. Res. A 93:1132–1139, 2010. issued, and a patent ‘Muscle chips and methods of use Chiquet, M., L. Gelman, R. Lutz, and S. Maier. From mechanotransduction to extracellular matrix gene expres- thereof’, US 20140342394 A1 pending. Intrinsic Cell Stress is Independent of Organization 191 sion in ﬁbroblasts. Biochim. Biophys. Acta 1793:911–920, Knight, M. B., N. K. Drew, L. A. McCarthy, and A. 2009. Grosberg. Emergent global contractile force in cardiac Driessen-Mol, A., M. Y. Emmert, P. E. Dijkman, L. Frese, tissues. Biophys. J. 110:1615–1624, 2016. B. Sanders, B. Weber, N. Cesarovic, M. Sidler, J. Leenders, Kural, M. H., and K. L. Billiar. Regulating tension in R. Jenni, J. Grunenfelder, V. Falk, F. P. Baaijens, and S. P. three-dimensional culture environments. Exp. Cell Res. Hoerstrup. Transcatheter implantation of homologous 319:2447–2459, 2013. ‘‘off-the-shelf’’ tissue-engineered heart valves with self-re- le Digabel, J., M. Ghibaudo, L. Trichet, A. Richert, and B. pair capacity: long-term functionality and rapid in vivo Ladoux. Microfabricated substrates as a tool to study cell remodeling in sheep. J. Am. Coll. Cardiol. 63:1320–1329, mechanotransduction. Med. Biol. Eng. Comput. 48:965– 2014. 976, 2010. 12 29 Ehrlich, H. P., and T. Rittenberg. Differences in the Legant, W. R., A. Pathak, M. T. Yang, V. S. Deshpande, mechanism for high- versus moderate-density ﬁbroblast- R. M. McMeeking, and C. S. Chen. Microfabricated tissue populated collagen lattice contraction. J. Cell Physiol. gauges to measure and manipulate forces from 3D micro- 185:432–439, 2000. tissues. Proc. Natl. Acad. Sci. USA 106:10097–10102, 2009. 13 30 Feinberg, A. W., P. W. Alford, H. Jin, C. M. Ripplinger, A. Loerakker, S., G. Argento, C. W. Oomens, and F. P. A. Werdich, S. P. Sheehy, A. Grosberg, and K. K. Parker. Baaijens. Effects of valve geometry and tissue anisotropy Controlling the contractile strength of engineered cardiac on the radial stretch and coaptation area of tissue-engi- muscle by hierarchal tissue architecture. Biomaterials neered heart valves. J. Biomech. 46:1792–1800, 2013. 33:5732–5741, 2012. McCain, M. L., A. Agarwal, H. W. Nesmith, A. P. Ne- Feinberg, A. W., A. Feigel, S. S. Shevkoplyas, S. Sheehy, smith, and K. K. Parker. Micromolded gelatin hydrogels G. M. Whitesides, and K. K. Parker. Muscular thin ﬁlms for extended culture of engineered cardiac tissues. Bioma- for building actuators and powering devices. Science terials 35:5462–5471, 2014. 317:1366–1370, 2007. McCain, M. L., S. P. Sheehy, A. Grosberg, J. A. Goss, and Fernandez, P., and A. R. Bausch. The compaction of gels K. K. Parker. Recapitulating maladaptive, multiscale by cells: a case of collective mechanical activity. Integr. remodeling of failing myocardium on a chip. Proc. Natl. Biol. (Camb.) 1:252–259, 2009. Acad. Sci. USA 110:9770–9775, 2013. 16 33 Foolen, J., V. S. Deshpande, F. M. Kanters, and F. P. Mol, A., M. C. Rutten, N. J. Driessen, C. V. Bouten, G. Baaijens. The inﬂuence of matrix integrity on stress-ﬁber Zund, F. P. Baaijens, and S. P. Hoerstrup. Autologous remodeling in 3D. Biomaterials 33:7508–7518, 2012. human tissue-engineered heart valves: prospects for sys- Foolen, J., M. W. Janssen-van den Broek, and F. P. temic application. Circulation 114:I152–I158, 2006. Baaijens. Synergy between Rho signaling and matrix den- Mol, A., A. I. P. M. Smits, C. V. C. Bouten, and F. P. T. sity in cyclic stretch-induced stress ﬁber organization. Acta Baaijens. Tissue engineering of heart valves: advances and Biomater. 10:1876–1885, 2014. current challenges. Expert Rev. Med. Devices 6:259–275, Frangi, A., W. Niessen, K. Vincken, and M. Viergever. 2009. Multiscale vessel enhancement ﬁltering. Med. Image Com- Nirmalanandhan, V. S., M. S. Levy, A. J. Huth, and D. L. put. Comput. Assist. Intervent. 1496:130–137, 1998. Butler. Effects of cell seeding density and collagen con- Fu, J., Y. K. Wang, M. T. Yang, R. A. Desai, X. Yu, Z. centration on contraction kinetics of mesenchymal stem Liu, and C. S. Chen. Mechanical regulation of cell function cell-seeded collagen constructs. Tissue Eng. 12:1865–1872, with geometrically modulated elastomeric substrates. Nat. 2006. Methods 7:733–736, 2010. Obbink-Huizer, C., C. W. Oomens, S. Loerakker, J. Foo- Grosberg, A., P. W. Alford, M. L. McCain, and K. K. len, C. V. Bouten, and F. P. Baaijens. Computational Parker. Ensembles of engineered cardiac tissues for physi- model predicts cell orientation in response to a range of ological and pharmacological study: heart on a chip. Lab. mechanical stimuli. Biomech. Model. Mechanobiol. 13:227– Chip 11:4165–4173, 2011. 236, 2014. 21 37 Grosberg, A., A. P. Nesmith, J. A. Goss, M. D. Brigham, Rocco, K. A., M. W. Maxﬁeld, C. A. Best, E. W. Dean, M. L. McCain, and K. K. Parker. Muscle on a chip: and C. K. Breuer. In vivo applications of electrospun tis- in vitro contractility assays for smooth and striated muscle. sue-engineered vascular grafts: a review. Tissue Eng. Part J. Pharmacol. Toxicol. Methods 65:126–135, 2012. B. Rev. 20:628–640, 2014. 22 38 Grouf, J. L., A. M. Throm, J. L. Balestrini, K. A. Bush, Sanders, B., S. Loerakker, E. S. Fioretta, D. J. Bax, A. and K. L. Billiar. Differential effects of EGF and TGF- Driessen-Mol, S. P. Hoerstrup, and F. P. Baaijens. Im- beta1 on ﬁbroblast activity in ﬁbrin-based tissue equiva- proved geometry of decellularized tissue engineered heart lents. Tissue Eng. 13:799–807, 2007. valves to prevent leaﬂet retraction. Ann. Biomed. Eng. Hald, E. S., K. E. Steucke, J. A. Reeves, Z. Win, and P. W. 44:1061–1071, 2016. Alford. Long-term vascular contractility assay using geni- Schwarz, U. S., and M. L. Gardel. United we stand: inte- pin-modiﬁed muscular thin ﬁlms. Biofabrication 6:045005, grating the actin cytoskeleton and cell-matrix adhesions in 2014. cellular mechanotransduction. J. Cell Sci. 125:3051–3060, Hinz, B., S. H. Phan, V. J. Thannickal, M. Prunotto, A. 2012. Desmouliere, J. Varga, W. O. De, M. Mareel, and G. Sun, Y., R. Duffy, A. Lee, and A. W. Feinberg. Optimizing Gabbiani. Recent developments in myoﬁbroblast biology: the structure and contractility of engineered skeletal muscle paradigms for connective tissue remodeling. Am. J. Pathol. thin ﬁlms. Acta Biomater. 9:7885–7894, 2013. 180:1340–1355, 2012. Tamiello, C., A. B. Buskermolen, F. P. Baaijens, J. L. Kaunas, R., S. Usami, and S. Chien. Regulation of stretch- Broers, and C. V. Bouten. Heading in the right direction: induced JNK activation by stress ﬁber orientation. Cell understanding cellular orientation responses to complex Signal. 18:1924–1931, 2006. biophysical environments. Cell. Mol. Bioeng. 9:12–37, 2016. 192 VAN LOOSDREGT et al. Tara, S., H. Kurobe, M. W. Maxﬁeld, K. A. Rocco, T. Yi, ture to reduce stress generation and retraction in engi- Y. Naito, C. K. Breuer, and T. Shinoka. Evaluation of neered heart valve tissues. Tissue Eng. Part C Methods remodeling process in small-diameter cell-free tissue-engi- 19:205–215, 2013. neered arterial graft. J. Vasc. Surg. 62:734–743, 2015. Weber, B., P. E. Dijkman, J. Scherman, B. Sanders, M. Y. Tara, S., K. A. Rocco, N. Hibino, T. Sugiura, H. Kurobe, Emmert, J. Grunenfelder, R. Verbeek, M. Bracher, M. C. K. Breuer, and T. Shinoka. Vessel bioengineering. Circ. Black, T. Franz, J. Kortsmit, P. Modregger, S. Peter, M. J. 78:12–19, 2014. Stampanoni, J. Robert, D. Kehl, M. van Doeselaar, M. van Loosdregt, I. A., G. Argento, A. Driessen-Mol, C. W. Schweiger, C. E. Brokopp, T. Walchli, V. Falk, P. Zilla, A. Oomens, and F. P. Baaijens. Cell-mediated retraction ver- Driessen-Mol, P. Baaijens, and S. P. Hoerstrup. Off-the- sus hemodynamic loading—A delicate balance in tissue- shelf human decellularized tissue-engineered heart valves in engineered heart valves. J. Biomech. 47:2064–2069, 2014. a non-human primate model. Biomaterials 34:7269–7280, van Spreeuwel, A. C., N. A. Bax, A. J. Bastiaens, J. Foolen, 2013. S. Loerakker, M. Borochin, D. W. van der Schaft, C. S. Win, Z., G. D. Vrla, K. E. Steucke, E. N. Sevcik, E. S. Chen, F. P. Baaijens, and C. V. Bouten. The inﬂuence of Hald, and P. W. Alford. Smooth muscle architecture within matrix (an)isotropy on cardiomyocyte contraction in cell-dense vascular tissues inﬂuences functional contractil- engineered cardiac microtissues. Integr. Biol. (Camb.) ity. Integr. Biol. (Camb.) 6:1201–1210, 2014. 6:422–429, 2014. Wyss, K., C. Y. Yip, Z. Mirzaei, X. Jin, J. H. Chen, and C. van Vlimmeren, M. A., A. Driessen-Mol, C. W. Oomens, A. Simmons. The elastic properties of valve interstitial cells and F. P. Baaijens. An in vitro model system to quantify undergoing pathological differentiation. J. Biomech. stress generation, compaction, and retraction in engineered 45:882–887, 2012. heart valve tissue. Tissue Eng. Part C Methods 17:983–991, Yamin, R., and K. G. Morgan. Deciphering actin 2011. cytoskeletal function in the contractile vascular smooth van Vlimmeren, M. A., A. Driessen-Mol, C. W. Oomens, muscle cell. J. Physiol. 590:4145–4154, 2012. and F. P. Baaijens. Passive and active contributions to Ye, G. J., Y. Aratyn-Schaus, A. P. Nesmith, F. S. Pas- generated force and retraction in heart valve tissue engi- qualini, P. W. Alford, and K. K. Parker. The contractile neering. Biomech. Model. Mechanobiol. 11:1015–1027, 2012. strength of vascular smooth muscle myocytes is shape van Vlimmeren, M. A., A. Driessen-Mol, C. W. Oomens, dependent. Integr. Biol. (Camb.) 6:152–163, 2014. and F. P. Baaijens. The potential of prolonged tissue cul-
Cardiovascular Engineering and Technology – Springer Journals
Published: Oct 24, 2016
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera