FRANÇOIS BERGERON, Ph.D.,1 DANIEL DUPUIS, Ph.D.,3 EVE LANGELIER, Ph.D.,3DENIS RANCOURT, Ph.D.,3 FRANÇOIS A. AUGER, M.D.,1 and LUCIE GERMAIN, Ph.D.1
ABSTRACT
In the rapidly growing field of tissue engineering, the functional properties of tissue substitutes arerecognized as being of the utmost importance. The present study was designed to evaluate the ef-fects of static mechanical forces on the functionality of the produced tissue constructs. Living tissuesheets reconstructed by the self-assembly approach from human cells, without the addition of syn-thetic material or extracellular matrix (ECM), were subjected to mechanical load to induce cell andECM alignment. In addition, the effects of alignment on the function of substitutes reconstructedfrom these living tissue sheets were evaluated. Our results show that tissue constructs made fromliving tissue sheets, in which fibroblasts and ECM were aligned, presented higher mechanical re-sistance. This was assessed by the modulus of elasticity and ultimate strength as compared with tis-sue constructs in which components were randomly oriented. Moreover, tissue-engineered vascularmedia made from a prealigned living tissue sheet, produced with smooth muscle cells, possessedgreater contractile capacity compared with those produced from living tissue sheets that were notprealigned. These results show that the mechanical force generated by cells during tissue organiza-tion is an asset for tissue component alignment. Therefore, this work demonstrates a means to im-prove the functionality (mechanical and vasocontractile properties) of tissues reconstructed by tis-sue engineering by taking advantage of the biomechanical forces generated by cells under staticstrain.
1Laboratoire d’Organogénèse Expérimentale, Hôpital du Saint-Sacrement du CHA, and Department of Surgery, 2Present ad-dress: Molecular Medicine Program and Centre for Stem Cell and Gene Therapy, Ottawa Health Research Institute, Ottawa, ON,Canada.
3Department of Mechanical Engineering, Laval University, Québec, PQ, Canada.
90
INTRODUCTION
THE OPTIMAL FUNCTIONALITY of a tissue depends on itsappropriate histological organization. Thus, the
alignment of cells, as well as of the extracellular matrix(ECM), is an essential indicator of tissue integrity. Forexample, longitudinal alignment of endothelial cells fa-vors their adherence to the basal lamina under shear stress
and circular orientation of the smooth muscle cells max-imizes their ability to regulate vascular tone.1,2 Further-more, abnormal tissue structures resulting from variouspathologies are responsible for functional defects. In ath-erosclerosis, the thickening of the subintimal spacecauses narrowing of the lumen, resulting in a reductionof blood flow to downstream tissues. In the wound-heal-ing process, aberrant tissue reorganization results in func-
tional problems.3 The formation of hypertrophic scarsover joints restricts the mobility of the articulation.4
Therefore, it is crucial to have a better understanding ofthe factors involved in tissue organization and remodel-ing.
The importance of mechanical stimulation in tissue re-modeling was deduced from empiric data obtained afterthe successful treatment of patients with hypertrophicscars by compression.5 The influence of physical andchemical stimuli on cell alignment and tissue reorgani-zation was further studied with three-dimensional tissueequivalents produced from fibrin and collagen gel-basedconstructs. It was shown that the addition of chemical ad-ditives to the medium, such as ribose or aprotinin, in-creases cross-linking of collagen and decreases matrixdegradation.6–8 The application of magnetic fields eitherbefore or after gel polymerization,9–12 as well as sta-tic13–18 or dynamic19–23 mechanical forces, increases thealignment of collagen fibers in vascular equivalents.However, the effect of these parameters on vascular func-tion such as vasoconstriction could not be evaluated inthese gel-based constructs because of their low mechan-ical resistance.
One method of tissue engineering has led to the pro-duction of human blood vessels (TEBVs) with high me-chanical resistance.24 This process, named the self-as-sembly approach, consists of coaxing cells to secrete theirown ECM by the addition of ascorbic acid in order to ob-tain a living tissue sheet. Once formed, the living tissuesheet can be either rolled or stacked to create three-di-mensional blood vessels or skin substitutes.24–26 The highresistance of these tissue constructs allows them to be im-planted24 and evaluated by measuring their vasocontrac-tile response within the tissue-engineered vascular media(TEVM).27
In this article we took advantage of this tissue-engi-neering approach leading to the production of a naturalhuman tissue from cells only, without the addition of syn-thetic or ECM components. We evaluated the effect ofmechanical load on the functionality of the ensuing tis-sue. Our results show that proper alignment of the com-ponents of the living tissue sheet construct increased itsmechanical strength. Accordingly, the contractile re-sponse of TEVM produced with prealigned living tissuesheets was greater than that of TEVM in which cells werenot prealigned. Thus, control of the alignment of livingtissue sheet components, using static force, improves thefunctional characteristics of these tissues.
MATERIALS AND METHODS
Cell isolation and culture
Human fibroblasts were enzymatically dissociated fromthe dermis, using collagenase H (Roche, Indianapolis, IN),
TISSUE FUNCTIONALITY IN RESPONSE TO LOADING
and plated as described previously.24 Human vascularsmooth muscle cells (VSMCs) were isolated from humanumbilical vein by the method of explantation.28 Cells andexplants were grown in the Dulbecco–Vogt modificationof Eagle’s medium–Ham’s F12 (DMEM:Ham; ratio, 3:1)(GIBCO/Invitrogen, Grand Island, NY), supplementedwith 10% fetal bovine serum (FBS; HyClone, Logan, UT)and antibiotics (penicillin [100 U/mL] and gentamicin [25�g/mL]). Cells were used at passage 4 to 6.
Preparation of aligned and randomly orientedliving tissue sheets
To favor the production of an abundant extracellularmatrix (ECM), human fibroblasts or VSMCs were cul-tured in DMEM:Ham supplemented with 10% FBS, L-ascorbic acid (50 �g/mL; Sigma-Aldrich Canada,Oakville, ON, Canada), and antibiotics. Under these cul-ture conditions, cells progressively self-assembled theneosynthesized ECM proteins into a deep and adherentliving tissue sheet. Once matured, the cohesive living tis-sue sheet was detached from the support and used for tis-sue construction. The time of detachment was numberedas day 0 for our experiment.
The method used to axially align cells and ECM pro-teins within the living tissue sheet is illustrated in Fig. 1.After detachment from the culture flask (Fig. 1A), oneof the extremities of the living tissue sheet was rapidly,but carefully, attached to a side of the frame by gentlyclipping it, using Ligaclips (Ethicon Endo-Surgery,Cincinnati, OH) (Fig. 1B). The other extremity wasclipped to the opposite side of the frame (Fig. 1C). Theliving tissue sheet on the frame was then placed in a bac-teriological petri dish and cultured for a further 21 days(Fig. 1D).
As control, all four sides of the living tissue sheet wereattached on the frame in order to conserve a random or-ganization of cells and ECM (Fig. 1D�).
Contribution of cells to tissue compaction
Contribution of cells to the compaction of tissue wasassessed as follows. On day 0, living tissue sheets weremechanically detached from the culture flask and an-chored on their opposite sides to the frame. Attached liv-ing tissue sheets were cultured in DMEM:Ham supple-mented with 10% FetalClone II (HyClone), L-ascorbicacid (50 �g/mL), and antibiotics for 21 days. For con-trol, cells present in living tissue sheets were disruptedby osmotic shock, using sterile distilled water, or treatedwith inhibitors (0.1% sodium azide [Sigma-Aldrich] orcytochalasin D [2 �g/mL; Calbiochem, La Jolla, CA])before their detachment from tissue culture flasks. Cy-tochalasin D was removed from living tissue sheets bywashing four times with fresh medium.
At each time point (i.e., 0, 1, 2, 3, 4, 5, 7, 14, and 21
91
days) digital images were taken with a Coolpix 4500(Nikon, Melville, NY). The surface area of living tissuesheets was calculated with freeware NIH Image. Resultswere represented as the relative compaction of the livingtissue sheet by comparing the surface area at each timepoint with that at time 0.
Tissue-engineered vascular media
TEVMs were prepared as previously described.24
Briefly, freshly detached living tissue sheets, or livingtissue sheets cultured for 7 days in a planar form to in-duce axial alignment, were wrapped around tubular man-drels.
To allow for further maturation of the tissue beforeprocessing for vasocontractile studies, living tissue sheetsand TEVMs were cultured for 21 days in medium con-taining ascorbic acid (50 �g/mL). Medium was changedthree times per week.
Mechanical features of a living tissue sheet
Mechanical properties of freshly detached living tis-sue sheets (n � 4), and of living tissue sheet containingrandomly oriented (n � 4) or prealigned (n � 4) compo-nents, were measured directly with a previously describedmechanical stretching apparatus29 performing simple ten-sile tests. Both extremities of each living tissue sheet werefixed on anchoring jaws, one mobile and the other con-nected to a load cell. The prealigned living tissue sheetswere stretched in a direction parallel to the orientation oftheir components (cells and ECM). Living tissue sheetsin which the components were randomly oriented, andfreshly detached living tissue sheets, were attached bytheir opposite sides. Once a living tissue sheet was an-chored, the apparatus stretched it by pulling on the mo-bile jaw at a constant rate of 1 mm/s. Data generated from
GRENIER ET AL.
the developing resistance as a function of the deforma-tion were recorded and processed with data acquisitionsoftware. The tensile stress was calculated by dividingthe measured force by the initial cross-sectional area ofthe living tissue sheet. The strain was calculated by di-viding the length increase of the living tissue sheet by itsoriginal length. To determine the relative stiffness andthe resistance of the tissue, stress–strain curves were plot-ted. The modulus of elasticity (Young’s modulus) wasdetermined by calculating the slope of the linear portionof these curves. The ultimate tensile strength was thehighest stress recorded before the failure of the stretchedtissue.
Contraction experiment
TEVM contraction experiments were done as de-scribed previously,27 using traditional TEVMs preparedwith freshly detached living tissue sheet (n � 8), TEVMsprepared with living tissue sheet in which cells and ECMwere prealigned (n � 8), or TEVMs in which longitudi-nal alignment of components was induced by attachingthe extremities of the freshly detached living tissue sheetafter wrapping on the mandrel. The contraction assayswere conducted with TEVMs cultured for 21 days on atubular support. Contractile functions were assessed bychallenging the rings with a cumulative dose (10�8–10�5
M) of histamine (Sigma-Aldrich), a well-known vasoac-tive agent as previously described.27 Briefly, rings weremounted in a myograph and the mechanical activity wasrecorded by an isometric force transducer.
Planar (non-tubular) living tissue sheets were alsoprocessed for contraction experiments as follows. Pre-aligned living tissue sheets (n � 4) were fixed to the forcetransducer in a direction parallel (n � 4) or perpendicu-lar (n � 4; control) to the alignment of the components.
92
FIG. 1. Preparation of a planar aligned living tissue sheet. The living tissue sheet is mechanically detached from the cultureflask (A), anchored on two sides to a frame side by side (B and C), using Ligaclips, and maintained in culture for 21 days (D).A living tissue sheet can be kept randomly oriented by anchoring it on its entire periphery (D�).
Histology
Tissue preparations were fixed overnight in Histo-choice (Amresco, Solon, OH) and embedded in paraffin.Five-micrometer-thick sections were stained with Mas-son’s trichrome solution.
Immunohistofluorescence
For immunofluorescent staining, 5-�m-thick frozentissue sections were fixed for 10 min in methanol andprocessed as previously described.24 For VSMC pheno-typic characterization and organization, sections were in-cubated with mouse monoclonal antibodies directedagainst �-smooth muscle actin (�-SM-actin) (DakoCy-tomation, Mississauga, ON, Canada), type I and type IVcollagens (Chemicon International, Temecula, CA), andrhodamine-labeled-goat anti-mouse IgG (Cedarlane Lab-oratories, Hornby, ON, Canada). For controls, the pri-mary antibody was omitted. Nuclei were stained bluewith Hoechst 33258.
For confocal observations, 5-mm-diameter punch biop-sies of tissue preparations were harvested from the livingtissue sheets on day 0 and day 7. Samples were then fixedovernight in 4% formaldehyde. Before labeling, the tissueswere rinsed in phosphate-buffered saline (PBS) and thecells were permeabilized with 100% methanol for 5 min.Tissues were incubated with a primary antibody against �-SM-actin or type I collagen, and with a secondary antibodycoupled to Alexa-594 fluorochrome (polyclonal rabbit anti-mouse; Molecular Probes, Eugene, OR).
Transmission electron microscopy
Samples were fixed in 2% glutaraldehyde and processedfor electron microscopy as previously described.30
TISSUE FUNCTIONALITY IN RESPONSE TO LOADING
Statistical analysis
The Student unpaired t test or analysis of variance(ANOVA) was used for statistical analysis.
RESULTS
Characterization of aligned and randomlyoriented living tissue sheets
Living tissue sheets were produced from either humanvascular smooth muscle cells (VSMCs) or fibroblasts cul-tured in the presence of ascorbic acid for 9 to 14 days.Under these conditions, cells became rapidly confluent,forming three or four superposed cell layers included inan extracellular matrix comprising collagen as shown bytransmission electron microscopy (Fig. 2A). Living tis-sue sheets made from smooth muscle cells may sponta-neously detach from the bottom of the plastic flask whenthe culture time is prolonged. This is indicative of mul-tiple internal tension points developing and accumulat-ing on the surface of the culture flask. In living tissuesheets, the cells were randomly oriented in the plane par-allel to the surface of the culture flask (Fig. 2B). Afterdetachment from the flask, the living tissue sheet re-tracted to 25 to 40% of its original size. This was char-acterized by the presence of cell clusters and loose col-lagen fibrils resulting from the loss of accumulatedinternal tension (Fig. 2C). In contrast, the natural de-tachment of living tissue sheets produced from fibroblastsis occasional and gradual over several days.
To induce the alignment of cells and ECM, two ex-tremities of retracted living tissue sheets, freshly detachedfrom the culture flask, were anchored on the frame. Fig-
93
FIG. 2. Microscopic view of a living tissue sheet produced with fibroblasts by the self-assembly approach. (A) Transmissionelectron micrograph of a transverse section of living tissue sheet attached to the bottom of a plastic flask. (B) Histological ap-pearance (Masson’s trichrome staining) of a living tissue sheet before detachment from plastic flask in comparison with (C) afreshly detached living tissue sheet. Under both conditions, cells and ECM fibers are randomly oriented. Scale bars: (A) 1 �m;(B and C) 50 �m.
ure 3 shows a microscopic view of an anchored livingtissue sheet as a function of maturation time. Before de-tachment, cells were randomly oriented within the livingtissue sheet (Fig. 3A). After detachment from the cultureflask, opaque zones consisting of clustered cells weresoon observed as the sheet spontaneously contracted (0h; Fig. 3B). During the next days, these clustered zonestended to dissociate as cells reorganized and began theirreorientation in a uniaxial direction (Fig. 3C). Cells con-tinued their reorganization along the strain with time,contracting the living tissue sheet while reorienting them-selves in a uniaxial direction parallel to the developingmechanical load (Fig. 3D). This continued until a paral-lel orientation of cells and ECM fibers was observed af-ter 7 days (Fig. 3E). The controls, in which all four sidesof the living tissue sheet were anchored to the frame (Fig.1D�), did not realign, as cells maintained a random ori-entation in the plane.
The remodeling of living tissue sheets by cells was vis-ible macroscopically. On day 0 (Fig. 4A), that is, at thetime of attachment to the frame, living tissue sheets wereheterogeneous. The previously described cell clusters ap-pear as opaque zones or “domains.” During maturationof the tissue the macroscopic aspect drastically changed,revealing homogeneous tissue in which cells and ECMcomponents were aligned after 7 days of maturation (Fig.4B). The reorientation of cells and ECM components be-tween day 0 and day 7 was confirmed by the immunola-beling of cellular �-SM-actin filaments (Fig. 4C and D)and type I collagen fibers (Fig. 4E and F), revealing theirrealignment within the living tissue sheets.
Furthermore, ultrastructural analysis showed thatcells and their nuclei were longitudinally oriented.Within the tissue structure, collagen fibers were abun-dant and oriented in a direction parallel to the cell lon-gitudinal axis (Fig. 4G). This alignment was in the axisof the strain. Moreover, cells formed multiple gap junc-tions as they do in vivo in native tissues (Fig. 4H). Asa control, living tissue sheets were detached from cul-ture flasks and kept in culture medium without any an-chorage. Under these conditions, numerous degenerat-ing fragmented nuclei typical of apoptotic cells wereobserved by electron microscopy (Fig. 4I), indicating
GRENIER ET AL.
that cells died when tension was not restored within theliving tissue sheets.
Contribution of cells to tissue compaction
To determine the contribution of cells to tissue com-paction, we treated living tissue sheets before their detach-ment from tissue culture flasks by disrupting the cells byosmotic shock, using water; by inhibiting ATP production,using sodium azide; or by treatment with a cytoskeletal in-hibitor, cytochalasin D (Fig. 5A). When compared withnontreated living tissue sheets, inhibiting conditions ham-pered tissue compaction completely, indicating that cellsare essential for the process of tissue compaction. Sodiumazide treatment inhibited tissue compaction with a lag time,indicating that cells cannot compact tissue if actin poly-merization does not occur.
By treating living tissue sheets with cytochalasin D atdifferent time points, it was possible to entirely inhibitfurther tissue compaction even after it had been initiated.Inversely, when cytochalasin D was washed from the tis-sue, after a 7-day treatment, compaction can be reiniti-ated, indicating that cells present in the living tissue sheetwere viable and able to reverse the inhibition of com-paction (Fig. 5B).
Taken together, these results show that cells drive tis-sue compaction. This phenomenon was possible only ifthe cells and their cytoskeleton were intact.
Mechanical resistance of aligned and randomlyoriented living tissue sheets
The dermal component of skin is responsible for itsmechanical properties. Thus, the resistance of alignedversus randomly oriented and freshly detached living tis-sue sheets made of dermal fibroblasts was evaluated bytensile tests, using a traction apparatus. The stiffness ofthe tissue construct was evaluated on the basis of Young’smodulus. The abrupt slope of the graphic indicated thataligned tissues were almost 20 times stiffer than freshlydetached and randomly oriented living tissue sheets (Fig.6). The evaluation of ultimate tensile strength revealedthat the rupturing point of aligned living tissue sheets was
94
FIG. 3. Microscopic aspect of an aligned living tissue sheet as a function of maturation time. (A) Attached to the bottom ofthe culture flask; (B) freshly detached and then anchored on two opposite edges (0 h). Later time points: (C) 48 h, (D) 5 days,and (E) 7 days of maturation. Scale bar: 300 �m.
approximately seven times higher than that of freshly de-tached or randomly oriented living tissue sheets.
Histological and functional characterization ofTEVM produced from aligned or randomlyoriented living tissue sheets
The effect of alignment on the functionality of tissue-engineered vascular media, produced with smooth mus-cle cells, was evaluated by its vasocontractile response.Histological analyses revealed that TEVMs prepared withaligned living tissue sheets were more compact and cellnuclei were more elongated. This indicated that the cellswere more oriented in comparison with TEVMs preparedwith freshly detached nonaligned living tissue sheets(Fig. 7A and B).
The immunolabeling of TEVM with a differentiationmarker of VSMCs revealed that the expression of �-SM-actin was increased in cells of the aligned living tissuesheets (Fig. 7C and D). The orientation of ECM was vi-
TISSUE FUNCTIONALITY IN RESPONSE TO LOADING
sualized by type IV collagen staining (Fig. 7E and F).The elongation of cell nuclei observed by Hoechst bluestaining also confirmed cell alignment.
To determine the functionality of TEVM as a functionof cell and ECM orientations, we challenged our differ-ent TEVM preparations with vasoactive agents in organbaths (Fig. 7G). No significant difference was observedfor the lowest doses of histamine (i.e., 10�8 and 10�7
M). However, at higher doses of histamine (more than10�6 M), TEVM produced with aligned living tissuesheets showed significantly better contractile responsesthan did TEVM prepared with randomly oriented livingtissue sheets (traditional). At 10�5 and 10�4 M doses,TEVM prepared with aligned living tissue sheet con-tracted 1.8 times more compared with traditional TEVM.As a control, TEVMs were reattached at both extremi-ties after rolling to induce longitudinal alignment. Whenthe cells and ECM were oriented in a direction parallelto the axis of the tubular mandrel, the contractile responsewas significantly lower than with the two other constructs
95
FIG. 4. Macroscopic aspect of a living tissue sheet with fibroblasts attached on a frame by two opposite edges on day 0 (A)and after 7 days of maturation (B). Note multiple contraction domains on the living tissue sheet as demonstrated by the presenceof opaque spots on day 0 (A). After 7 days of maturation, compaction of the living tissue sheet occurs. Confocal microscopy re-veals that on day 0 both cytoskeletal protein [�-SM-actin (C)] and ECM protein [collagen I (E)] are disorganized. On day 7,both of these proteins are aligned along the axis of the tension [�-SM-actin (D); collagen I (F)]. (G) Transmission electron mi-croscopy showing a group of cells parallel to one another and to the collagen present in ECM. (H) Cell–cell junctions are pres-ent as indicated by arrowheads. (I) Abundant apoptotic cells are observed in a living tissue sheet that was matured in the absenceof tension. Scale bars: (A and B) 8 mm; (C–F) 25 �m; (G and I) 2 �m; (H) 100 nm.
(Fig. 7G). These results indicate that cell orientation iscrucial for an optimal tissue function.
To confirm the effect of SMC orientation on the con-tractile response, we used planar living tissue sheets, inwhich cells and ECM were aligned. They were mounteddirectly on the myograph in a direction parallel or per-pendicular to the axis of the isometric force transducer(Fig. 7H). When 10�4 M of histamine was added, the liv-ing tissue sheets in which cells were aligned in a direc-tion parallel to the axis had a contractile response almostsix times greater than that of the other living tissue sheetplaced in a perpendicular orientation.
DISCUSSION
In this report, we took advantage of a new tissue-en-gineering approach to determine the response of cells to
GRENIER ET AL.
mechanical loading. We have established for the first timethat the functional characteristics of a tissue construct canbe improved by aligning cells and extracellular matrixcomponents, using a static mechanical load developedduring tissue compaction. This method could be usefulfor the study of cellular mechanisms involved in the re-modeling process that takes place during mechanicalloading.
The importance of the mechanical environment to thecells in tissues is reflected by cell death when mechani-cal loading is released. For example, apoptotic cell deathis induced when normally adherent cells are cultured insuspension because of an inappropriate mechanical en-vironment.31 In our experiments, apoptotic cells were ob-served in living tissue sheets that were not reanchored af-ter their detachment. This is consistent with previousobservations made in a collagen gel system, in which
96
FIG. 5. Contribution of cells to tissue compaction. (A) Graph showing the effect of sodium azide (�), cytochalasin D (�), andwater (�) on the compaction of living tissue sheet produced with fibroblasts (see Materials and Methods). (�) Control (non-treated). (B) Graph showing the effect of cytochalasin D and its reversibility at various time points. Treated from: day 0 to day21 (�), day 0 to day 7 (�), day 2 to day 21 (�) and control (nontreated) (�). Results presented are from a representative ex-periment (with five replicas) of two independent experiments performed. Tmt, treatment.
TISSUE FUNCTIONALITY IN RESPONSE TO LOADING 97
FIG. 6. Representative curves of stress (MPa) as a function of the strain (%) for prealigned (solid line), freshly detached (dot-ted line), and randomly oriented (dashed line) living tissue sheets produced with fibroblasts. Results presented are from one rep-resentative experiment of four independent experiments performed.
FIG. 7. Histological aspect of TEVM prepared traditionally (A, C, and E) or with an aligned living tissue sheet (B, D, and F).Masson’s trichrome staining (A and B); immunolabeling of �-SM-actin (C and D) and type IV collagen (E and F). Graphs il-lustrate dose–response curves for the contraction of tubular (G) and planar (H) constructs in response to cumulative doses of his-tamine (10�8–10�5 M). (G) TEVM were prepared with axially prealigned living tissue sheets in which cells and ECM are alignedcircumferentially to the axis of the tubular support (n � 8). TEVM were prepared traditionally (nonaligned) with freshly detachedliving tissue sheets (n � 8). In some cases they were reattached at both extremities after rolling to induce longitudinal alignment(n � 8). (H) Planar living tissue sheets were also processed for contraction experiments as described below. Prealigned livingtissue sheets (n � 4) were fixed in the apparatus parallel to the orientation of alignment. As for controls, aligned living tissuesheets (n � 4) were fixed perpendicularly to the orientation of alignment. Results presented are from a representative experimentof three independent experiments performed. Error bars indicate the SD. *p � 0.05, **p � 0.01, significantly different from con-trol, that is, traditional TEVM in (G) and perpendicularly fixed in (H). Scale bars: 50 �m.
cells rapidly die when the tension of a mechanicallyloaded collagen gel is released.32
In our self-assembly model, the thickness and resis-tance of our living tissue sheets allowed us to counteractthe effect of detachment and cell death by reanchoringthe freshly detached living tissue sheets on a frame.Therefore, it was possible to study the reorganization ofthe living tissue sheets after a temporary removal of ten-sion, and to examine the predominant role of cells ver-sus that of the ECM during tissue remodeling.
This remodeling phenomenon is known as compactionand has been previously described for the collagen gelsystem.33–35 However, the role of cells in this system isnot evident because the compaction could take place evenin the absence of cells by the use of magnetic field andcross-linking molecules.7,9–12,17,35–37 However, cellsmight play an active role in the alignment of collagenfibers after gelation. They reorganize their surroundingECM by pulling and aligning collagen fibers, in a direc-tion parallel to their bodies. For example, cells present inan anchored collagen gel will align in the plane of theanchorage because the strain will be generated in thisplane.13,14,38,39
In the present study, a living tissue sheet in which cellsand collagen fibers were randomly oriented in the planewas produced. Its reorganization was monitored after de-tachment from the bottom of the flask and subsequentreattachment by two of its opposite extremities, withoutinduced initial tension, a condition that does not induceapoptosis. Under these conditions, cells pull on the ECM,and align along the axis of the cellular self-generatedstrain, leading to compaction of the tissue. Thus, it ap-pears that cells play a central role in the reorganizationof the living tissue sheet, because collagen fibers werenot prealigned by external stimuli such as magnetic stim-ulation. In fact, our results suggest that tissue compactionis intimately driven by the cytoskeleton. This hypothesisis strengthened by the finding that the cytoskeleton isconnected to protein complexes that have extracellularlinks with the ECM.40 Our system will be useful in pro-viding a better understanding of the process of tissue re-modeling that is necessary to design better treatments inpathologies such as hypertrophic scars.
This work also revealed that the use of aligned livingtissue sheets for the production of two different tissuesimproves their respective functional properties. Indeed,the mechanical resistance increased in planar living tis-sue sheets when cells and ECM were aligned as a resultof the anchorage of their extremities. Moreover, TEVMproduced by wrapping aligned living tissue sheets arounda tubular mandrel presented a better contractile responseto vasoactive agents than did nonaligned ones. TEVMsmade with aligned living tissue sheets also presented amore compact ECM and an increase in cell differentia-tion exemplified by �-SM-actin content. The higher �-
GRENIER ET AL.
SM-actin expression may have contributed to the in-creased vasocontraction. Our observations are consistentwith those made in collagen gels, in which ultrastructuralfeatures characteristic of a higher differentiation statuswere induced by mechanical loading.20
In this article, we show for the first time that the align-ment of cells and ECM induced by static mechanicalforces can improve the functionality of tissue constructssuch as their mechanical and contractile properties. Tis-sue engineering is now evolving toward new goals, inwhich not only the appearance but also the functionalityof tissue substitutes is a priority.41 A better understand-ing of static mechanical loading, as described here, or ofcyclic strain, as presented by our team42 and Niklason etal.,43 is necessary to improve the quality and functional-ity of tissue substitutes such that they resemble, as closelyas possible, their natural counterparts. In addition, thesetissue substitutes provide methodologies to better under-stand the role of mechanical stimuli in the developmentof pathologies such as hypertrophic scars.
ACKNOWLEDGMENTS
We thank Dr. Charles J. Roberge and Dr. AnthonyScime for critical review of the manuscript. We alsothank Rina Guignard for technical assistance. This workwas supported by a grant from the Canadian Institute ofHealth Research (CIHR) to F.A.A. and L.G. L.G. was therecipient of scholarships from the CIHR and a holder ofa Canadian Research Chair on Stem Cells and Tissue En-gineering from the CIHR. G.G. was the recipient of aStudentship from the Fonds de la Recherche en Santé duQuébec, as was D.L. from the CIHR; and M.R.-Z. wasthe recipient of a fellowship from the Heart and StrokeFoundation of Canada.
REFERENCES
1. Dunkern, T.R., Paulitschke, M., Meyer, R., Buttemeyer, R.,Hetzer, R., Burmester, G., and Sittinger, M. A novel per-fusion system for the endothelialisation of PTFE grafts un-der defined flow. Eur. J. Vasc. Endovasc. Surg. 18, 105,1999.
2. Thoumine, O., Ziegler, T., Girard, P.R., and Nerem, R.M.Elongation of confluent endothelial cells in culture: Theimportance of fields of force in the associated alterationsof their cytoskeletal structure. Exp. Cell Res. 219, 427,1995.
3. Gabbiani, G. The myofibroblast in wound healing and fi-brocontractive diseases. J. Pathol. 200, 500, 2003.
4. Scott, P.G., Ghahary, A., Chambers, M.M., and Tredget,E.E. Biological basis of hypertrophic scarring. Adv. Struct.Biol. 3, 157, 1994.
5. Rudolph, R., Vande Berg, J., and Ehrlich, H.P. Wound con-traction and scar contracture. In: Cohen, I.K., Diegelmann,
98
R.F., and Lindblad, W.J., eds. Wound Healing: Biochem-ical and Clinical Aspects. Philapelphia: W.B. Saunders,1992, p. 96.
6. Girton, T.S., Oegema, T.R., Grassl, E.D., Isenberg, B.C.,and Tranquillo, R.T. Mechanisms of stiffening andstrengthening in media-equivalents fabricated using glyca-tion. J. Biomech. Eng. 122, 216, 2000.
7. Girton, T.S., Oegema, T.R., and Tranquillo, R.T. Exploit-ing glycation to stiffen and strengthen tissue equivalentsfor tissue engineering. J. Biomed. Mater. Res. 46, 87, 1999.
8. Jockenhoevel, S., Zund, G., Hoerstrup, S.P., Chalabi, K.,Sachweh, J.S., Demircan, L., Messmer, B.J., and Turina,M. Fibrin gel: Advantages of a new scaffold in cardiovas-cular tissue engineering. Eur. J. Cardiothorac. Surg. 19,424, 2001.
9. Barocas, V.H., Girton, T.S., and Tranquillo, R.T. Engi-neered alignment in media equivalents: Magnetic pre-alignment and mandrel compaction. J. Biomech. Eng. 120,660, 1998.
10. Torbet, J., Freyssinet, J.M., and Hudry-Clergeon, G. Ori-ented fibrin gels formed by polymerization in strong mag-netic fields. Nature 289, 91, 1981.
11. Torbet, J., and Ronziere, M.C. Magnetic alignment of col-lagen during self-assembly. Biochem. J. 219, 1057, 1984.
12. Tranquillo, R.T., Girton, T.S., Bromberek, B.A., Triebes,T.G., and Mooradian, D.L. Magnetically orientated tissue-equivalent tubes: Application to a circumferentially orien-tated media-equivalent. Biomaterials 17, 349, 1996.
13. Auger, F.A., Berthod, F., Goulet, F., and Germain, L. Whatis new in mechanical properties of tissue-engineered or-gans. Curr. Top. Pathol. 93, 87, 1999.
14. Germain, L., Goulet, F., Moulin, V., Berthod, F., andAuger, F.A. Engineering human tissues for in vivo appli-cations. Ann. N.Y. Acad. Sci. 961, 268, 2002.
15. Hirai, J., Kanda, K., Oka, T., and Matsuda, T. Highly ori-ented, tubular hybrid vascular tissue for a low pressure cir-culatory system. ASAIO J. 40, M383, 1994.
16. Hirai, J., and Matsuda, T. Self-organized, tubular hybridvascular tissue composed of vascular cells and collagen forlow-pressure-loaded venous system. Cell Transplant. 4,597, 1995.
17. Huang, D., Chang, T.R., Aggarwal, A., Lee, R.C., andEhrlich, H.P. Mechanisms and dynamics of mechanicalstrengthening in ligament-equivalent fibroblast-populatedcollagen matrices. Ann. Biomed. Eng. 21, 289, 1993.
18. L’Heureux, N., Germain, L., Labbe, R., and Auger, F.A.In vitro construction of a human blood vessel from culturedvascular cells: A morphologic study. J. Vasc. Surg. 17, 499,1993.
19. Kanda, K., and Matsuda, T. Behavior of arterial wall cellscultured on periodically stretched substrates. Cell Trans-plant. 2, 475, 1993.
20. Kanda, K., and Matsuda, T. Mechanical stress-induced ori-entation and ultrastructural change of smooth muscle cellscultured in three-dimensional collagen lattices. Cell Trans-plant. 3, 481, 1994.
21. Kanda, K., Matsuda, T., and Oka, T. Mechanical stress in-duced cellular orientation and phenotypic modulation of 3-D cultured smooth muscle cells. ASAIO J. 39, M686, 1993.
22. Kanda, K., Matsuda, T., and Oka, T. In vitro reconstruc-
TISSUE FUNCTIONALITY IN RESPONSE TO LOADING
tion of hybrid vascular tissue: Hierarchic and oriented celllayers. ASAIO J. 39, M561, 1993.
23. Seliktar, D., Black, R.A., Vito, R.P., and Nerem, R.M. Dy-namic mechanical conditioning of collagen-gel blood ves-sel constructs induces remodeling in vitro. Ann. Biomed.Eng. 28, 351, 2000.
24. L’Heureux, N., Paquet, S., Labbe, R., Germain, L., andAuger, F.A. A completely biological tissue-engineered hu-man blood vessel. FASEB J. 12, 47, 1998.
25. Auger, F.A., Rémy-Zolghadry, M., Grenier, G., and Ger-main, L. The self-assembly approach for organ recon-struction by tissue-engineering. E-biomed J. Reger. Med.1, 75, 2000.
26. Michel, M., L’Heureux, N., Pouliot, R., Xu, W., Auger,F.A., and Germain, L. Characterization of a new tissue-en-gineered human skin equivalent with hair. In Vitro Cell.Dev. Biol. Anim. 35, 318, 1999.
27. L’Heureux, N., Stoclet, J.C., Auger, F.A., Lagaud, G.J.,Germain, L., and Andriantsitohaina, R. A human tissue-en-gineered vascular media: A new model for pharmacologi-cal studies of contractile responses. FASEB J. 15, 515,2001.
28. Grenier, G., Remy-Zolghadri, M., Guignard, R., Bergeron,F., Labbe, R., Auger, F.A., and Germain, L. Isolation andculture of the three vascular cell types from a small veinbiopsy sample. In Vitro Cell. Dev. Biol. Anim. 39, 131,2003.
29. Langelier, E., Rancourt, D., Bouchard, S., Lord, C.,Stevens, P.P., Germain, L., and Auger, F.A. Cyclic trac-tion machine for long-term culture of fibroblast-populatedcollagen gels. Ann. Biomed. Eng. 27, 67, 1999.
30. Black, A.F., Berthod, F., L’Heureux, N., Germain, L., andAuger, F.A. In vitro reconstruction of a human capillary-like network in a tissue-engineered skin equivalent. FASEBJ. 12, 1331, 1998.
31. McGill, G., Shimamura, A., Bates, R.C., Savage, R.E., andFisher, D.E. Loss of matrix adhesion triggers rapid trans-formation-selective apoptosis in fibroblasts. J. Cell Biol.138, 901, 1997.
32. Grinnell, F., Zhu, M., Carlson, M.A., and Abrams, J.M. Re-lease of mechanical tension triggers apoptosis of human fi-broblasts in a model of regressing granulation tissue. Exp.Cell Res. 248, 608, 1999.
33. Grinnell, F. Fibroblast–collagen–matrix contraction:Growth-factor signalling and mechanical loading. TrendsCell Biol. 10, 362, 2000.
34. Harris, A.K., Stopak, D., and Wild, P. Fibroblast tractionas a mechanism for collagen morphogenesis. Nature 290,249, 1981.
35. Tranquillo, R.T. Self-organization of tissue-equivalents:The nature and role of contact guidance. Biochem. Soc.Symp. 65, 27, 1999.
36. Dickinson, R.B., Guido, S., and Tranquillo, R.T. Biasedcell migration of fibroblasts exhibiting contact guidance inoriented collagen gels. Ann. Biomed. Eng. 22, 342, 1994.
37. Guido, S., and Tranquillo, R.T. A methodology for the sys-tematic and quantitative study of cell contact guidance inoriented collagen gels: Correlation of fibroblast orientationand gel birefringence. J. Cell Sci. 105, 317, 1993.
38. Eastwood, M., Mudera, V.C., McGrouther, D.A., and
99
Brown, R.A. Effect of precise mechanical loading on fi-broblast populated collagen lattices: Morphologicalchanges. Cell Motil. Cytoskeleton 40, 13, 1998.
39. Lopez Valle, C.A., Auger, F.A., Rompre, P., Bouvard, V.,and Germain, L. Peripheral anchorage of dermal equiva-lents. Br. J. Dermatol. 127, 365, 1992.
40. Hinz, B., and Gabbiani, G. Mechanisms of force genera-tion and transmission by myofibroblasts. Curr. Opin.Biotechnol. 14, 538, 2003.
41. Guilak, F., Kapur, R., Sefton, M.V., Vandenburgh, H.H.,Koretsky, A.P., Kriete, A., and O’Keefe, R.J. Functionalassessment of engineered tissues and elements of tissue de-sign: Breakout session summary. Ann. N.Y. Acad. Sci. 961,207, 2002.
42. Goulet, F., Rancourt, D., Cloutier, R., Germain, L., Poole,
GRENIER ET AL.
A.R., and Auger, F.A. Tendons and ligaments. In: Lanza,R., Langer, R., and Vacanti, J., eds. Principles of TissueEngineering. San Diego, CA: Academic Press, 2000, p.711.
43. Niklason, L.E., Gao, J., Abbott, W.M., Hirschi, K.K.,Houser, S., Marini, R., and Langer, R. Functional arteriesgrown in vitro. Science 284, 489, 1999.
Address reprint requests to:Lucie Germain, Ph.D.
LOEX, Hôpital du Saint-Sacrement du CHA1050, chemin Sainte-Foy
