C. Stoll and T. John contributed equally to this work.
Extracellular matrix expression of human tenocytes in three-dimensional air–liquid and PLGA cultures compared with tendon tissue: Implications for tendon tissue engineering†
Version of Record online: 24 FEB 2010
Copyright © 2010 Orthopaedic Research Society
Journal of Orthopaedic Research
Volume 28, Issue 9, pages 1170–1177, September 2010
How to Cite
Stoll, C., John, T., Endres, M., Rosen, C., Kaps, C., Kohl, B., Sittinger, M., Ertel, W. and Schulze-Tanzil, G. (2010), Extracellular matrix expression of human tenocytes in three-dimensional air–liquid and PLGA cultures compared with tendon tissue: Implications for tendon tissue engineering. J. Orthop. Res., 28: 1170–1177. doi: 10.1002/jor.21109
- Issue online: 19 JUL 2010
- Version of Record online: 24 FEB 2010
- Manuscript Accepted: 21 DEC 2009
- Manuscript Received: 16 AUG 2009
- air–liquid culture;
- extracellular tendon matrix;
- type I collagen
Tenocyte transplantation may prove to be an approach to support healing of tendon defects. Cell–cell and cell–matrix contacts within three-dimensional (3D) cultures may prevent tenocyte dedifferentiation observed in monolayer (2D) culture. The present study compares both neotissue formation and tenocyte extracellular matrix (ECM) expression in 2D and 3D cultures directly with that of native tendon, in order to determine optimal conditions for tendon tissue engineering. Primary human tenocytes were embedded in poly[lactic-co-glycolic-acid] (PLGA)-scaffolds and high-density cultures. Neotissue formation was examined by hematoxyline–eosine (H&E) and immunofluorescence staining. Gene expression of ECM proteins and vascular endothelial growth factor (VEGF) was compared at days 0 (2D), 14, and 28 in 3D cultures and tendon. Histomorphology of 3D culture showed tendon-like tissue as tenocyte cell nuclei became more elongated and ECM accumulated. Type I collagen gene expression was higher in 2D culture than in tendon and decreased in 4-week-old 3D cultures, whereas type III collagen was only elevated in high-density culture compared with tendon. Decorin and COMP were reduced in 2D and increased in 3D culture almost to ex vivo level. These results suggest that the 3D high-density or biodegradable scaffolds cultures encourage the differentiation of expanded monolayer tenocytes in vitro to tendon-like tissue. © 2010 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 28:1170–1177, 2010
Tendon healing is a time consuming process which leads to scar formation as a functionally inferior reparative tissue. Tendon injuries result in considerable morbidity, disabling the patients for several months or longer.1 Scars are characterized by a disorganized extracellular matrix (ECM), induce peripheral tendon adhesions and hence, result in reduced tendon stability and flexibility at the formerly injured site. The hypocellular histological structure, poor blood supply, and bradytrophic metabolism of tendons are major reasons for their limited self-healing properties.2, 3 Only 5% of the normal tendon tissue volume is represented by resident cells, the tenocytes, but 95% by tendon ECM. Tenocytes are highly specialized fibroblasts producing and regulating the abundant, but strictly organized tendon ECM.4 The tendon ECM mainly consists of parallel running collagen fibrils. Approximately 95% of collagen in normal tendon is type I, with types III and V also present in small amounts. Tendon also contains elastic fibers which are an important prerequisite for the elastic module of tendon.5 The main proteoglycan in tendon is decorin,6 whereby aggrecan is only present in small amounts.7 Furthermore, the non-collagenous protein COMP is a typical component of tendon ECM8 and scleraxis is a differentiation-associated transcription factor in tendon.9 Vascular endothelial growth factor (VEGF) is an important angiogenic factor in tendon which plays a role in tendon rupture, healing, and development.10 The expansion of differentiated tenocytes in vitro is a helpful approach to analyze tendon healing and to find tools to support tendon repair with the use of tissue engineering strategies. However, tenocytes expanded in monolayer culture proliferate slowly and display an unstable phenotype with increasing culture time and the tendency to dedifferentiate,11 recognizable for example, by morphological changes and a decrease in COMP expression. Three-dimensional (3D) cultures might be a strategy to maintain the differentiated tenocyte phenotype. The present study was accomplished to assess neotissue formation of 3D cultured tenocytes in direct comparison with 2D culture and native tendon.
MATERIALS AND METHODS
Tenocyte Isolation and 2D and 3D Cultivation
Human primary tenocytes were isolated from hamstring tendons of healthy middle-aged donors (averaged 31 years, male and female mixed) using explant culture as described previously,12 according to the Ethical Committee of the Charité University of Medicine (EA4/033/08, May 2008). Tenocytes cultures were grown at 37°C in a humidified atmosphere with 5% CO2 and the growth medium [DMEM/Ham's F-12 1:1 containing 10% fetal calf serum (both: Biochrom-Seromed, Munich, Germany), 25 µg/ml ascorbic acid (Sigma–Aldrich, Munich, Germany), 50 IU/ml streptomycin, 50 IU/ml penicillin, 2.5 µg/ml partricin, essential amino acids and L-glutamine (all: Biochrom-Seromed)] was changed every 3 days. For the experiments, tenocytes were cultured for 6–9 monolayer passages before being introduced in the 3D cultures. To prepare the high-density air–liquid cultures, 10 µl of a tenocyte cell pellet was pipetted onto a membrane filter with a pore diameter of 0.2 µm (Sartorius, Göttingen, Germany) on top of a stainless steel grid at the medium-air interface in a Petri dish.12 Tenocytes of passages 6–9 in adequate cell numbers (105 cells/mm3 PLGA) were mixed with fibrinogen (Tissuecol Duo S Immuno, Baxter, Unterschleißheim, Germany) to agglomerate onto the poly[lactic-co-glycolic-acid] (PLGA)-scaffold (Ethicon, Norderstedt, Germany) followed by polymerization with thrombin, diluted 1:10 (v/v PBS), for 10 min. Scaffolds and high-density cultures seeded with tenocytes derived from 4 or 7 different human donors were cultured for 14 or 28 days.
Life–Death Assessment of PLGA Cultures
After washing with PBS, the 14- and 28-day-old scaffolds were incubated in fluorescein diacetate (FDA; 3 mg/ml dissolved in acetone; Sigma–Aldrich) for 15 min at 37°C, rinsed three times with PBS before being counterstained with propidium iodide (PI) solution (0.1 mg/ml; Sigma–Aldrich) for 2 min in the dark. The green or red fluorescence was visualized using fluorescence microscopy (Olympus, Hamburg, Germany).
H&E-Staining and Indirect Immunofluorescence Labeling
Tendon tissue and 3D tenocyte cultures were embedded in OCT medium (TissueTek, Sakura Finetec, Zoeterwoude, The Netherlands). A 7-µm thick cryosections of cultures, native tendon or cover slips seeded with tenocytes were fixed for 15 min in 4% paraformaldehyde and stained for 4 min in Harris hematoxyline (Sigma–Aldrich), rinsed in water, and counterstained for 4 min in eosine (Carl ROTH GmbH, Karlsruhe, Germany). Slides were rinsed with Aqua dest. and covered with Entellan (Merck, Darmstadt, Germany). For immunofluorescence labeling slides were washed with Tris buffered saline (TBS: 0.05 M Tris, 0.015 M NaCl, pH 7.6), overlaid with protease-free donkey serum (5% diluted in TBS) for 10 min at RT before incubated with primary antibodies [polyclonal rabbit anti-(human)-type I collagen (1:50), Acris, Herford, Germany or monoclonal mouse-anti-(human)-aggrecan (1:20), RD systems, Minneapolis, MN] in a humid chamber 1 h at RT. Controls included absence of primary antibodies. Tenocytes were subsequently washed with TBS before and after incubation with donkey-anti-rabbit-Alexa-488 (Dianova, Hamburg, Germany) or donkey-anti-mouse-Cy-3 (Invitrogen, Carlsbad, CA) coupled secondary antibodies (diluted 1:200 in TBS), respectively, for 1 h at RT. Slides were covered with fluoromount mounting medium (Southern Biotech, Biozol Diagnostica, Eching, Germany) and examined under a fluorescence microscope (Axioskop 40, Carl Zeiss, Jena, Germany).
RNA Isolation and RTD-PCR Analysis
Human tendon was embedded in RNA stabilizing RNAlater (Qiagen, Hilden, Germany) for 24 h at 4°C immediately after surgery and then stored at −80°C. Tendon tissue was separated in liquid nitrogen and directly transferred in 2 ml Qiazol Reagent (Qiagen) for 10 min at RT. The total of 200 µl chloroform was added, mixed, and incubated for 5 min at RT. Samples were centrifuged at 12,000g for 15 min (4°C). The clear supernatant was separated and 1:1 70% ethanol was added. Then, with the use of the Qiagen RNeasy Mini Kit (Qiagen) RNA was isolated. This kit was directly used for RNA isolation of cultured tenocytes. RNA quantity and quality was evaluated with the RNA 6000 Nano assay (Agilent Technologies, Waldbronn, Germany). Equal amounts of total RNA were reverse transcribed with the use of the Qiagen QuantiTect® reverse transcription Kit (Qiagen) according to the manufacturer's instructions. One microlitre aliquot of the cDNA was amplified by RTD-PCR in a 20-µl reaction mixture using specific primer pairs for type I collagen (241462_Hs_Col1A1_FAM_1; Qiagen), type III collagen (Hs 00164103_m1), elastin (Hs00355783_m1), decorin (Hs00370384_m1), COMP (Hs 00164359_m1), scleraxis (Hs03054634_g1), aggrecan (Hs00202971_m1), sox9 (Hs 00165814_m1), β1-integrin (Hs00559595), and the house-keeping gene hypoxanthin–guanin–phosphoribosyl–transferase (HPRT, endogenous control; all: Applied Biosystems, Foster City, CA) as well as VEGF (VEGFA_gi_71051575; Qiagen). Evaluations were performed using the TaqMan Gene Expression Assay (Applied Biosystems) in an Opticon 1–real-time cycler (Opticon™ RTD-PCR; Biorad, Munich, Germany). Data analysis was performed according to Pfaffl.13
Normalized data were expressed as the mean and standard error of mean (mean ± SEM). Differences between experimental groups were considered significant at p < 0.05 as determined by ANOVA with the Bonferroni test.
Characteristics of Tenocyte 2D Culture
On a whole, during the monolayer passages 6–9 which were used for preparation of 3D cultures, no major differences in morphological characteristics could be recognized. Monitoring of gene expression for types I and III collagen, decorin, scleraxis, and sox9 via real-time RT-PCR showed no significant regulation within passages 6–9 (data not shown).
Viability of Tenocytes in PLGA Cultures
The vitality of tenocytes cultured on PLGA-scaffolds could be visualized by intracellular assembly of green fluorescent FDA (Fig. 1). There was no visible accumulation of dead cells at the investigated time points in 3D PLGA culture. Human tenocytes were viable after prolonged culture in PLGA. Furthermore, a slight increase in the tenocyte cell number by proliferation on the scaffold was discernible.
Morphology and Type I Collagen Deposition in Tendon, 2D, and 3D Cultured Tenocytes
The tissue histology of native tendon was compared with tenocyte monolayers and both high-density and PLGA associated cultures at 2 or 4 weeks using H&E staining. In native tendon tissue, collagen fiber bundles, and tenocytes were organized in a wavy longitudinal pattern with tenocytes lying in rows between them. The cell nuclei of tenocytes were small, elongated, and heterochromatic (Fig. 2A). In contrast, on monolayer cultured tenocytes, ECM deposition was barely detectable. Their cell nuclei contained one or more nucleoli and were larger and more euchromatic than nuclei of tenocytes in tendon or cultured in 3D culture. Monolayer-derived tenocytes generally had a more tenoblastic appearance (Fig. 2B). They exhibited multiple spindle- or sheet-like cytoplasmic processes forming some cell–cell contacts with neighboring cells. In high-density cultures ECM deposition was clearly detectable. ECM assembly around the tenocytes was enhanced at 4 weeks culture, where multiple ECM fibers could be observed. The cell-to-matrix ratio decreased from 2 to 4 weeks in high-density culture, but did not become the low cellularity observed in native hamstring tendons (Fig. 2C,D). Hence, the cellularity in 2-week-old 3D high-density cultures was higher and cells appeared more rounded compared with the 4-week-old cultures suggesting their tenoblastic character. The cell nuclei at the bottom of the 3D high-density culture became smaller, more elongated, and heterochromatic at 4 weeks of culture. However, the cells at the top of the culture exposed to the air remained round or ovoid. In PLGA-associated 3D cultures an ECM deposition by tenocytes was also clearly detectable after 4 weeks. Cells were arranged in groups between the PLGA fibers. Degradation of the PLGA fibers was present at 4 weeks of culturing time (Fig. 2E,F). A lower cellularity was evident at both culture time points compared to the high-density cultures. Compared to the monolayer cultured cells, tenocytes had smaller and more heterochromatic round nuclei. In native tendon, the abundant type I collagen fiber bundles had a strong parallel orientation with the tenocytes arranged between them (Fig. 3A). Synthesis of cell-associated type I collagen was evident in tenocyte monolayer cultures by anti-type I collagen immunolabeling (Fig. 3B). The amount of type I collagen accumulated between the cells increased from 2 to 4 weeks in high-density and PLGA cultures (Fig. 3C–F). Immunocytochemical analysis indicated no detectable changes in the protein expression of the typical cartilage matrix proteins type II collagen and aggrecan (data not shown).
Gene Expression of Tendon ECM Proteins, Tendon, and Cartilage Markers
Type I collagen gene expression was higher in 2D and 3D cultures than in native tendon tissue and decreased in both 3D culture systems at 4 weeks compared with the monolayers, however, the effect was not significant. The 4-week-old PLGA cultures expressed type I collagen at nearly the same level observed in native tendon. Type III collagen gene expression was significantly increased in 2-week-old 3D high-density cultures compared with tendon and 2D culture (Fig. 4A1–2). In contrast, it remained stable in 3D PLGA cultures. The type III/I collagen ratio was slightly elevated in both 3D culture systems (not significant) compared with native tendon and 2D culture whereby a high-donor variability could be observed in the PLGA cultures. Elastin expression did not show major differences in cultured tenocytes compared with native tendon (data not shown). Gene expression of the small proteoglycan decorin (Fig. 4B1), which is a typical component of tendon ECM, was significantly reduced in 2D culture versus native tendon. It increased in 4-week-old 3D high-density cultures almost to the ex vivo level and was significantly higher compared to the gene expression of 2D cultured tenocytes. Gene expression of COMP, an intrinsic constituent of the tendon ECM revealed a similar trend in high-density culture but only a very slight increase in 4-week-old PLGA cultures (not significant; Fig. 4B2). Gene expression of the tendon marker scleraxis decreased significantly in both 2D and 3D cultured tenocytes compared with native tendon. However, embedding of monolayer cultured tenocytes in 3D high-density cultures stimulated scleraxis expression slightly compared to monolayer cells (Fig. 4C1). Gene expression of the chondrogenic transcription factor sox9 was marked by a general rise in cultured tenocytes (not significant, data not shown). Expression of aggrecan, which is a typical cartilage ECM component declined significantly in tenocyte monolayer cultures. The introduction of monolayer cultured tenocytes in high-density culture revealed a slight recovery of gene expression (not significant); however, it did not reach the level observed in native tendon at 4 weeks of culture. The significant suppression of aggrecan gene expression of monolayer cultured tenocytes compared to native tendon remained mainly unaltered, when the cells were introduced for 2 or 4 weeks in the PLGA cultures (Fig. 4C2).
Gene Expression of Cell Matrix Receptor β1-integrin and VEGF
β1-integrin cell matrix receptor gene expression remained mainly unaltered by culture conditions (data not shown). Gene expression of the angiogenic factor VEGF was tested. A significant lower VEGF gene expression was detected in tenocytes cultured in monolayer or high-density culture compared with native tendon. However, the differences between native tendon and PLGA-cultured tenocytes were not significant (Fig. 4D1).
Tenocyte 2D and 3D Cultures
3D culture systems represent a possibility to stabilize tenocyte phenotype providing more intimate cell–cell and cell–matrix contacts compared with monolayer cultured tenocytes as reported previously.12 The high-density culture system as well as the PLGA-scaffolds may thus promote tenocyte redifferentiation. Tenocytes tend to lose their differentiated phenotype during expansion in monolayer culture as reported by Yao et al.11
Expression of decorin which is an important tendon proteoglycan and the main tendon matrix protein type I collagen was diminished with increasing monolayer passage numbers.11, 14 However, both groups did not compare the expression profile of tenocytes in monolayer culture directly with expression in native tendon. A significant decline of decorin, aggrecan, and the differentiation-associated transcription factor scleraxis gene expression as well as a decrease of COMP could be detected in the 2D cultures compared with native tendon in the present study, which strongly suggests these extracellular tendon ECM proteins as predictive indicators for tenocyte dedifferentiation. In contrast, type I collagen which is the main tendon ECM protein was increased in 2D culture in the present study when directly compared to native tendon tissue. The deprivation of ECM during passaging may be a stimulus inducing this enhanced collagen synthesis as already proposed by Almarza et al.15 The gene expression of elastin and β1-integrins remained mainly unaltered as also noted by Yao et al.11 To promote tenocyte redifferentiation two 3D culture systems were used in the present study: in comparison to a matrix-free air–liquid 3D culture system (high-density culture) the fibrin-associated PLGA culture was chosen. Fibrin was used to acquire a homogenous cell distribution in the PLGA-scaffold. PLGA possesses a lower rate of degradation compared with other akin biodegradable matrices such as PGA (poly[glycolic acid]) and provides higher grade of stability of tenocyte constructs for a longer time period.16, 17 Embedding autologous cells in polymer-based scaffolds of PLGA or PGA allows for formation of mesenchymal repair tissue as shown in vitro,18 in the equine joint defect model19 and clinical studies.20 In addition, from the surgical point of view, resorbable polymer-based scaffolds are initially stable and ensure easy handling during surgery and secure and mechanically stable fixation of grafts or implants by for instance pin fixation, suturing, or anchoring in the bone.21 Therefore, resorbable polymer-based scaffolds may be promising biomaterials also in regard to cell-based tendon repair approaches. The results of the life-death assay with PLGA cultures indicated a high content of viable tenocytes. Histomorphology of 3D cultures revealed a clear morphological shift from the tenoblastic cell shape with round nucleus towards a more elongated tenocytic phenotype and the matrix-to-cell ratio increased from 2- until 4-week-old 3D cultures. Tenoblasts display a lower differentiation state linked with less specific synthetic activity.22, 23 This morphological shift observed after the transfer of the tenocytes into the 3D cultures is also known for development and aging of tendon whereby cellularity in mature tendon declines.24, 25 The inhomogeneous culture histology observed in high-density culture reflects the fact that under air–liquid conditions a gradient of oxygen and nutrients exists. The more elongated tenocyte-like cells could be observed in the lower cell layers. In native tissue, tenocytes align according to the direction of mechanical stimuli, which were absent in the present in vitro study. The PLGA cultures revealed an increasing ECM deposition at 4 weeks. Accordingly, Cao et al.26 also observed substantial ECM production in a tenocytes PGA construct at 4-week culture time in vitro. After 4 weeks in vitro cultivation degradation of PLGA-fibers was clearly visible. Deng et al.27 described changes in cross-section and surface consistency for PLGA in dependence of culture duration in vitro, too. The generally low cellularity of the PLGA-scaffolds requires further investigation with higher tenocyte cell numbers. Increasing accumulation of the main tendon ECM protein type I collagen could be observed in both 3D cultures as shown by immunofluorescence labeling. However, the ECM content during the 4-week culture period did not reach the amount of ECM observed in the native tendon.
Comparison of Gene Expression in Cultured Tenocytes and Tendon
The gene expression analysis revealed multiple differences in the regulation of typical tendon and cartilage ECM components in the 2D and 3D cultured tenocytes compared with the native tissue. Type I collagen rose in both 3D cultures indicating the more tenoblastic gene expression profile of in vitro expanded tenocytes which were introduced in the 3D culture. After 4 weeks, the expression decreased in 3D culture suggesting assimilation towards the expression level in tendon. Type III collagen expression revealed a similar tendency in high-density culture, however, it remained mainly unaffected in PLGA culture. In PLGA culture effects of fibrin on ECM gene expression such as collagen type III cannot be excluded. Haasper et al.28 showed decreased cell proliferation of bone marrow-derived stem cells and further, an inhibition of chondrogenic differentiation by fibrin glue could be found.29 The gene expression of the ECM proteins decorin, COMP, and tendon marker scleraxis which was reduced in 2D cultures, possibly as a result of tenocyte dedifferentiation, rose again in 3D cultures. This trend indicates a tendogenic stimulus of 3D conditions. It is well known that tenocytes are able to transdifferentiate to fibrochondrocytes under particular culture conditions.30, 31 The chondrogenic transcription factor sox9 and several typical cartilage ECM components are expressed also in tendon such as type II collagen and aggrecan.32 Gene expression of sox9 did not increase significantly in 2D and 3D cultured tenocytes and there was no elevation of type II collagen protein deposition as well as gene and protein expression of the typical cartilage ECM protein aggrecan observed. Aggrecan gene expression differed between the two 3D culture systems: in the PLGA culture a severe decrease could be observed compared to tendon and also the 2D cultures. In contrast, in high-density cultures the aggrecan expression increased compared to 2D culture. One can assume that aggrecan expression in the liquid milieu of PLGA culture could be inhibited by absent mechanical load, since aggrecan is responsible for tissue viscoelasticity by reversible water binding in response to pressure. An inhibitory stimulus of fibrin on aggrecan gene expression in PLGA culture is further imaginable. Under air–liquid conditions of high-density cultures cells might require more water-binding proteoglycans for optimal nutrition and cell homeostasis. Additionally, the mechanical load of multiple cell layers present in high-density culture might induce aggrecan expression. Looking at the expression of these cartilage components a major chondrogenic transdifferentiative shift might be excluded in both tenocyte 3D culture systems. Additionally, VEGF expression was tested to assess proangiogenic stimuli of the culture conditions. A significantly lower gene expression of VEGF was observed in monolayer and high-density cultured tenocytes compared with native tendon whereby the expression level in PLGA cultures did not significantly differ from tendon. Hence, no proangiogenic shift induced by particular culture conditions could be supposed. The higher VEGF mRNA content in tendon might also be influenced by remnants of very small microvessels in native tendon tissue.10 Altogether, we found some assimilation of tenocyte gene expression (type I collagen, decorin, and COMP) towards native tendon as well as tendon-like neotissue formation in tenocyte 3D cultures. The analyzed 3D cultures might serve as a basis to study possibilities for cell-based and biomaterial guided tendon repair. However, some persistent differences in neotendon formation and aberrances compared with native tissue dependent on the particular 3D culture system still exist. Furthermore, it became obvious that a suitable strategy for tissue engineering requires extended culture periods and additional factors like appropriate growth factor or mechanostimulation, since Fong et al.33 found substantial changes in gene expression, for example, up- and downregulation of particular growth factors, induction of MMPs and decrease in matrix components.
The authors thank Hannah Gough for her support. This study was supported by grants from the Berlin Brandenburg Center for Regenerative Therapies, the Else Kröner-Fresenius-Stiftung the Sonnenfeld Foundation, Berlin, and the Rahel Hirsh program of the Charité Medical School, Berlin, Germany.
- 282008. Influence of fibrin glue on proliferation and differentiation of human bone marrow stromal cells seeded on a biologic 3-dimensional matrix. Technol Health Care 16: 93–101., , .