in vitro analysis of an allogenic scaffold for tissue-engineered meniscus replacement

In Vitro Analysis of an Allogenic Scaffold forTissue-Engineered Meniscus Replacement

Dirk Maier,1 Klaus Braeun,2 Erwin Steinhauser,3 Peter Ueblacker,4 Michael Oberst,1 Peter C. Kreuz,1 Nadine Roos,6

Vladimir Martinek,5 Andreas B. Imhoff 6

1Department of Orthopaedic and Trauma Surgery, University of Freiburg, Hugstetter Strasse 55, Freiburg, 79106, Germany

2Department of Trauma and Reconstructive Surgery, Berufsgenossenschaftliche Unfallklinik Murnau, Murnau, Germany

3Department of Biomechanics, Technical University of Munich, Munich, Germany

4Department of Trauma, Hand and Reconstructive Surgery, University of Hamburg, Hamburg, Germany

5Department of Orthopaedics, University of Rostock, Rostock, Germany

6Department of Orthopaedic Sports Medicine, Technical University of Munich, Munich, Germany

Received 24 January 2006; accepted 14 February 2007

Published online 3 August 2007 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jor.20405

ABSTRACT: Scaffolds play a key role in the field of tissue engineering. Particularly for meniscusreplacement, optimal scaffold properties are critical. The aim of our study was to develop a novelscaffold for replacement of meniscal tissue by means of tissue engineering. Emphasis was put onbiomechanical properties comparable to nativemeniscus, nonimmunogenecity, and the possibility ofseeding cells into and cultivating them within the scaffold (nontoxicity). For this purpose, nativeovinemenisci were treated in vitro in a self-developed enzymatic process. Complete cell removalwasachieved and shownboth histologically and electronmicroscopically (n¼15). Immunohistochemicalreaction (MHC 1/MHC 2) was positive for native ovine meniscus and negative for the scaffold.Compared to native meniscus (25.8 N/mm) stiffness of the scaffold was significantly increased(30.2 N/mm, p< 0.05, n¼ 10). We determined the compression (%) of the native meniscus and thescaffold under a load of 7 N. The compression was 23% for native meniscus and 29% for the scaffold(p< 0.05, n¼ 10). Residual force of the scaffold was significantly lower (5.2 N vs. 4.9 N, p< 0.05,n¼10). Autologous fibrochondrocytes were needle injected and successfully cultivated within thescaffolds over a period of 4weeks (n¼ 10). To our knowledge, this study is thefirst to remove cells andimmunogenetic proteins (MHC1/MHC2) completely out of nativemeniscus and preserve importantbiomechanical properties. Also, injected cells could be successfully cultivated within the scaffold.Further in vitro and in vivo animal studies are necessary to establish optimal cell sources,sterilization, and seeding techniques. Cell differentiation, matrix production, in vivo remodeling ofthe construct, and possible immunological reactions after implantation are subject of furtherstudies. � 2007 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res

25:1598–1608, 2007

Keywords: meniscus transplantation; meniscus replacement; scaffold; tissueengineering; biomechanics

INTRODUCTION

The importance of meniscal function for theintegrity of hyaline cartilage of the knee jointis without controversy. The menisci contributeconsiderably to load bearing and shock absorp-tion.1–3 Various clinical4–6 and experimental7–9

studies could demonstrate the chondroprotectiveeffects of meniscal function. Degenerative ortraumatic loss of meniscal tissue may requirenumerous surgical procedures to treat secondarypathologies.10 Consequently, optimal treatmentof meniscus lesions should focus on preservationor restoration of meniscal functions.11 If repaircannot be achieved or fails secondarily, effortsshould be put on functional replacement. How-ever, adequate meniscal substitution remains tobe a great challenge.

In the past, various methods were studied in theexperimental and clinical context, e.g., allogenicmeniscus transplantation,12–14 autologous tissuetransplantation,15,16 and prosthetic meniscus

1598 JOURNAL OF ORTHOPAEDIC RESEARCH DECEMBER 2007

The study was performed at the Department of OrthopaedicSports Medicine, Technical University of Munich, Munich,Germany.

Correspondence to: Dirk Maier (Telephone:þþ49 (0)761 2702650; Fax: þþ49 (0)761 270 2887;E-mail: [email protected])

� 2007 Orthopaedic Research Society. Published by Wiley Periodicals,Inc.

replacement.17,18 In meniscus allograft transplan-tation, host reactions, e.g., towards major histo-compatibility complexes of the donor, lead toprogressive decellularization and consecutivefailure of the transplant.19 So far, none of the abovementioned replacement methods could prove along-term chondroprotective effect.

Recently, promising new strategies weredeveloped by means of tissue engineering.Meniscus replacement is performed with acellularmatrices or cell-seeded matrices. Cell-basedstrategies in particular seem to lead to superiorresults. Peretti et al.20 found better histologicalintegration in the repair of longitudinal meniscustears using autologous chondrocytes seeded ontodevitalized allogenic meniscus slices compared tothe unseeded control group. The ability of isolatedchondrocytes to adhere onto devitalized cartilagematrices has been examined previously by thesame author.21 However, little data exist onthe behavior of chondrocytes or meniscal cellsseeded into such matrices.

A matrix serves as a scaffold for the ingrowth ofcells and supports the remodeling process of thetissue.22 Both synthetic-resorbable and naturalmaterials were tested as scaffolds. In experimentalanimal studies, fibrocartilage-like tissue couldbe found about 3 months after implantationof acellular porous polymer scaffolds.17,23,24

Compared to control groups, degeneration ofhyaline cartilage progressed slower but could notbe stopped. Collagen scaffolds derived fromporcine small intestine submucosa (SIS) failed toshow consistent results in experimental animalstudies.25,26 Twelvemonths after implantation of atype I collagen scaffold developed from bovineAchilles tendons (CMI1) in dogs, fibrocartilage-like tissue was present histologically. A significantchondroprotective effect could not be demon-strated.27 Two clinical pilot studies showed animprovement of the knee function and the level ofactivity.28,29 Scaffolds are necessary for tissueengineered meniscus replacement, althoughpreviously tested scaffolds often failed. Majorreasons are biomechanical failure, cell toxicity,and immunological response of the host towardsthe scaffold or towards its degradation pro-ducts.30,31 Rare information can be found in theliterature on biomechanical properties of scaffoldsand engineered tissue in particular.32,33

The aim of our in vitro study was to developa novel cell-loaded scaffold for replacement ofmeniscal tissue (partial or total) by means of tissueengineering. The scaffold is designed for use as aviable construct, which is seeded with autologous

cells prior to host implantation. Emphasis was puton biomechanical properties matching those ofnative meniscus, nonimmunogenecity in vitro,nontoxicity, and the possibility of seeding cells intoand cultivating them within the scaffold.

MATERIALS AND METHODS

Meniscus Specimens

Fifty-six ovine knee menisci were excised after eutha-nasia of 37 Merino sheep. The animals were used in aprevious experiment that did not influence the structureof the menisci. Mean age was 8 years (range: 6.5–9.4 years) and mean weight was 80 kg (68–94 kg).The menisci were cleared of soft tissue and washed in40 ml physiologic saline solution (DeltaSelect, Munich,Germany) for 1 h at 378C. The menisci were divided intohalves. For cryopreservation, they were incubated for2 h in 40 ml medium with 4% dimethylsulfoxide (Merck,Darmstadt, Germany) and kept frozen at �858C.Complete medium was used consisting of Dulbecco’smodified Eagle’s medium (DMEM) with 10% fetalbovine serum (FBS) and 1% penicillin and streptomycin(PS) (Biochrom, Berlin, Germany). The mean period ofcryopreservation was 21 days.

Scaffold Processing

Forty-one menisci were used to establish the protocol ofscaffold processing. Before and after each step ofprocessing, histological analyses were performed. Opti-mal dosage of each substance was determined byexperimental rows with increasing concentrations anddurations. The menisci were put into a 50-ml FalconTM

tube (BD Biosciences, Heidelberg, Germany) with 40 mlof enzymatic solution and shaken at 120 rpm inhorizontal position at 378C in an incubator shaker.Enzymes were dissolved in complete medium. Thesamples consecutively were treated with 0.25% trypsinsolution with 0.02% EDTA (Biochrom, Berlin, Germany),collagenase solution with 3 mg of collagenase A with aspecific activity of >0.15 U/mg (Roche Diagnostics,Mannheim, Germany), and protease solution with15 mg of protease with a specific activity of 4.8 U/mg(Sigma-Aldrich, St. Louis, MO). After each enzymaticprocess, the samples were washed in separate tubescontaining 40 ml physiologic saline solution. They wereplaced horizontally and shaken at 120 rpm at 48C. Thewashing process consisted of three cycles of 1 minfollowed by a fourth cycle of 12 h. After each cycle, thewashing solution was changed. After trypsin treatment,the washing process was repeated with distilled water.After the last washing process, the scaffolds wereshaken horizontally at 120 rpm in 40 ml physiologicsaline solution at 48C for 3 days.

Histology

Histological analyses included hematoxylin and eosin(H&E), elastic van Gieson (EVG), and alcianblue (ALC).

TISSUE-ENGINEERED MENISCUS REPLACEMENT 1599

DOI 10.1002/jor JOURNAL OF ORTHOPAEDIC RESEARCH DECEMBER 2007

Cryocuts of 7 mm were created at �238C (Microm;Walldorf, Germany). Staining was performed accordingto standard protocols.34 Before and after each step ofprocessing, histological analyses were performed. Sam-ples for histology were taken as complete cross-sectionsof the middle portions of the specimens.

Immunohistochemistry

Immunohistochemistry for major histocompatibilitycomplex (MHC) class I and II was performed on cryocutsof 5 mm before and after scaffold processing. Twoprimary antibodies were used with an incubation periodof 1 h, a mouse anti-MHC I monoclonal antibody and arat anti-MHC II monoclonal antibody (DakoCytomation,Hamburg, Germany), diluted 1:20 with phosphatebuffered saline (PBS). A standard avidin-biotin complex(Kit K5001/5005; DakoCytomation, Hamburg, Ger-many) was used as secondary antibody for MHC I, anda rabbit anti-rat monoclonal antibody for MHC II.Incubation period for secondary antibodies was30 min. Ovine small intestine tissue was used fornegative control.

Scanning Electronmicroscopy

Scanning electronmicroscopy (SEM) was performedafter the process to evaluate ultrastructure of thescaffolds. The specimens were vacuum-dried for 24 hin a concentrator (Hetosicc CD 4; Jouan Nordic, Allerød,Denmark) and then fixed and gold sputtered beforeanalysis in the microscope (Jeol 6300; Tokyo, Japan).

Glycosaminoglycan Content

The sulphated glycosaminoglycan (SGAG) content wasmeasured with a modified dimethylmethylene bluemethod35 before and after the process (n¼ 10). Smalltissue chips were gained from a complete cross-section ofthe middle part of the specimen and vacuum-dried for24 h. A solution of 4 M guanidinhydrochloride, 0.01 M

disodium-EDTA, and 0.05 M sodiumacetate was usedfor glycosaminoglycan extraction. Two milliliters ofsolution were used for a tissue sample of 5 mg in a15-ml FalconTM tube. After an extraction period of 2 h at48C on a shaker (120 rpm), the tube was put in acentrifuge at 10,000 rpm for 5 min. A sample of 100 mlwas diluted with distilled water at a ratio of 1:3. Forfurther processing, the BlyscanTM sulphated glycosami-noglycan assay (Biocolor, Newtownabbey, NorthernIreland) was used. Glycosaminoglycan contents weremeasured with an automatic spectral photometer(Spectrometer DU1 640; Beckman Coulter, Krefeld,Germany) at a wavelength of 650 nm. The contents werecalculated as percentages of dry weight. The data werestatistically analyzed (nonparametric Student’s t-test).

Biomechanics

The medial menisci of 10 individuals were divided intotwo halves. One half was left untreated (nativemeniscus) and the other was processed according tothe protocol (scaffold). The treatment of the halves waschosen double-blinded and randomized. The menisciwere positioned on an even desk (horizontal plane) andthe undersurface of the menisci was orientated parallelto desk. Cylinders with a diameter of 6 mm were cut outof the middle portions perpendicular to the horizontalplane. The samples were put in a specially designeddevice and the superior part of the cylinder was cut togain approximate parallel planes in relation to theundersurface. Cylinders of about 3 mm height resultedfor biomechanical analysis. The samples were placedhorizontally on a specially designedmetallic plate with acircular sink (diameter 6.0 mm, depth 0.3 mm) toprevent the samples from dislocating during axialloading. The tip of the indenter consisted of a steel ballof 6.0 mm diameter (Fig. 1). Cyclic indention tests wereperformed as minimally constraint compression–relaxation tests. A universal testing machine (Zwicki1120; Zwick, Ulm, Germany) was used with a calibrated

Figure 1. (a) Indenter and metallic cylinder; (b) sample in minimally constraint compression.[Color figure can be viewed in the online issue, which is available at http://www.interscience.wiley.com.]

1600 MAIER ET AL.

JOURNAL OF ORTHOPAEDIC RESEARCH DECEMBER 2007 DOI 10.1002/jor

load sensor of a maximum of 20 N and an accuracy of 1%(KAP-S; A.S.T., Dresden, Germany). The samples werekept moist throughout the experiments. Physiologicsaline solution was applied as a water bath at roomtemperature before testing. The indenter position wascalibrated prior to each test. The indenter position wasset zero at the level of the base of the cavity. A preload of0.1 N was applied to the specimen. The sample heightwas determined under preload conditions. Testingwas carried out in displacement control. A test cycleconsisted of the following phases: 1) preloading ofthe sample with 0.1 N; 2) dynamic compression of thesample with a constant load velocity of 5 mm/min until7 N were reached; 3) static compression of the sample for60 s (indenter position was fixed when 7 N werereached); and 4) relaxation of the sample with a constantload velocity of 1 mm/min until a load of 0.15 N. After aninterval of 60 s, the next cycle started until a totalnumber of five repetitive cycles was obtained (Fig. 2). Asalready mentioned above, the specimens were placedminimally constrained to allow quasi-free expansionduring testing. Load, indenter position, and time werelogged and displayed online by the test softwareTestXpert (Version 8.1; Zwick, Ulm, Germany). Threematerial characteristic values were determined: 1)stiffness was determined from the linear-elastic slopeof the loading curve between 2 N and 5 N; 2) relativecompression (%) of the sample (indenter position inrelation to absolute sample height) was determinedunder a load of 7 N at the end of the dynamiccompression phase; and 3) the residual forcewas definedas the load measured at the end of the static compres-sion phase. Material characteristic test data wereexported from TestXpert to Excel (Excel X; Microsoft,

Redmond, WA) and nonparametric statistical analysesof cycles 1, 3, and 5 were performed (Student’s t-test).

Scaffold Seeding and Culturing

Ten processed scaffold halves of medial menisci of10 individuals were sterilized by high-pressure hydro-static sterilization36 (600 MPa). Autologous fibrochon-drocytes were gained from the lateral menisci of theipsilateral knee joints. The menisci were excised aftereuthanasia and cut into small pieces of approximately2–3 mm3 and washed three times in 40 ml phosphatebuffered saline (PBS). For cell isolation, a collagenasedigestion protocol was used according to Webber andcoworkers.37 For all cell culture procedures, completemedium was used. The meniscus cells were expanded inmonolayer culture in a cell cultivator (InnovaTM 4200;New Brunswick Scientific, Edison, NJ) at 378C at anatmosphere of 5% CO2. The culture was split twicebefore freezing the cells at portions of 105 at �858C. Thescaffolds were seeded manually under sterile conditionsby needle injection technique. Two milliliters of com-plete medium containing 105 cells/ml were injectedwith a 26-gauge cannula (Dispomed Witt, Gelnhausen,Germany) into each scaffold. Injections were adminis-tered homogenously 20� perpendicular and 10� paral-lel to the orientation of the a.–p. axis. The scaffold waspenetrated completely and the needle was slowlyremoved under continuous application of cell solution.The seeded scaffolds were cultivated in separate platescontaining 20 ml of complete medium. The medium waschanged every 2 days. Samples were taken 7, 14, and28 days after seeding for H&E stain.

RESULTS

Histology

The histological analyses after the process showedcomplete cell removal out of the tissue. Comparedto native meniscus (Fig. 3a), the scaffolds (Fig. 3b)were free of synovial, endothelial, and meniscalcells. Throughout the complete cross-section of thescaffolds, micropores were detected. Microporeswere distributed homogeneously. The pores werelocated within an intact collagen bundle network.Native meniscal tissue did not show micropores.Histomorphologically, main orientation of thecollagen bundles was found to be unchanged inEVG staining. Sulphated glycosaminoglycanswere markedly reduced after the process through-out the complete cross-section of the scaffolds inALC staining. Elimination of sulphated glyco-saminoglycans was distributed homogenously.

Immunohistochemistry

Specific positive reactions were found for synovialand endothelial cells (MHC II) and fibrochondro-cytes (MHC I) in native meniscus (Fig. 4a).

Figure 2. Load curve consisting of five repetitive cycles withpreload, dynamic and static compression, and relaxation. Alinear slopewas found during dynamic compression. The load attheendof thestatic compressionphase (residual force) ismarkedblue. [Color figure can be viewed in the online issue, which isavailable at http://www.interscience.wiley.com.]



Fibrochondrocytes were MHC II negative;synovial and endothelial cells were MHC Inegative. Scaffold tissue did not show any expres-sion of MHC I or MHC II antigen (Fig. 4b).

Glycosaminoglycan Content

Mean SGAG content in native meniscal tissue was3% (SD� 0.4). After processing, mean SGAGcontent was reduced to 1.1% (SD� 0.4, p< 0.01).

Scanning Electronmicroscopy

SEM of native meniscus showed a dense extra-cellular matrix permeated by channels (approxi-

mately 5–25 mm). Circumscribed pores couldnot be detected. The extracellular matrix ofthe scaffold appeared to be macerated. We foundgaps and micropores within the collagen net-work (approximately 5–150 mm). Circumscribedchannels could not be detected. The collagenbundles appeared to be intact but were packedlooser. No cells or remnants of cellular materialwere detected within the scaffold (Fig. 5).

Biomechanics

The mean sample height was 3.4 mm (� 0.2 mm).All samples could be loaded up to 7 Nwithout signsof plastic deformity. Slopes of stress–relaxation

Figure 3. (a) Native ovine meniscus (H&E, original magnification �30); (b) decellularizedmeniscus scaffold (H&E, originalmagnification�30). [Color figure can be viewed in the online issue,which is available at http://www.interscience.wiley.com.]

Figure 4. (a) Native ovine meniscus with specific positive reaction for MHC II (originalmagnification�60); (b) decellularizedmeniscus scaffold with negative reaction forMHC II (originalmagnification �60). [Color figure can be viewed in the online issue, which is available at http://www.interscience.wiley.com.]

1602 MAIER ET AL.


curves were always found to be linear between 2 Nand 5 N. The scaffold stiffness exceeded stiffness ofnative meniscus significantly throughout testingtissue by about 20% during cycle 3 and about 17%during cycle 5 (p< 0.05). During cycle 1, nosignificant difference was found between native(mean 18.5 N/mm� 2.2 N/mm) and processed(mean 18.1 N/mm� 2.2 N/mm) tissue (Fig. 6).Both stiffness of native meniscus and scaffoldmaterial increased significantly throughout test-ing by 39% and 69%, respectively (p< 0.05)

(Fig. 6a). Mean compression of sample heightwas 23% (�6%) for native meniscus after cycle 5.Mean scaffold compression was found to be 29%(�6%). Scaffold compression exceeded compres-sion of native meniscus significantly by 26%(p< 0.05) (Fig. 6b). The residual force of thescaffold material was significantly lower comparedto native meniscus throughout biomechanicaltesting (p< 0.05). Mean initial (after cycle 1)residual forces for native meniscus and scaffoldwere 3.5 N (�0.6 N) and 2.4 N (�0.8 N),

Figure 5. (a–d) SEM images of native meniscus [(original magnification) a: �120; b: �370; c:�800; d: �6,500]; (e–h) SEM of the scaffold [(original magnification) e: �120; f: �370; g: �800; h:�6,500].

Figure 6. (a–c) Stiffness, compression, and residual force were determined before and afterenzymatic digestion (clear, native ovine meniscus; hatched, decellularized meniscus scaffold).



respectively. This initial difference was about 46%.Difference of the residual force significantlyreduced down to 6% throughout testing(p< 0.05). Mean residual force after cycle 5 was5.2 N (�0.3) for native meniscus and 4.9 N (�0.2)for scaffold, respectively (Fig. 6c).

Scaffold Seeding and Culturing

Seeding scaffolds with autologous fibrochondro-cytes and cultivation was successful in all casesover a period of 28 days. No case of contaminationoccurred during that period. Cell distribution wasclustered along injection canals. Histomorphologi-cally, viable cells were detected throughout allareas of complete scaffold cross-sections after 7,14, and 28 days after seeding (Fig. 7 a–d).Fibrochondrocytes showed typical configurationof elongated fibroblast-like cells after 4 weeks ofcultivation (Fig. 7c,d).

DISCUSSION

Optimal scaffold properties are critical for success-fully replacing meniscal tissue by means oftissue engineering. Rare information exist inparticular on biomechanical properties of scaffoldsused for meniscus replacement experimentallyand clinically.38 However, the importance ofscaffold biomechanics is well accepted throughoutthe literature.39–41 De Groot et al.42 found a higher

percentage of fibrocartilage formation in polymericscaffolds with compressive moduli similar tothose of native meniscus. To date, the influenceof biomechanical scaffold properties on cell differ-entiation is not finally elucidated. Biomechanicalinsufficiency and secondary failure of the scaffoldremains to be a major complication in large animalstudies in vivo.30

We managed to design a scaffold that approx-imates the biomechanical functions of nativemeniscus. Viscoelasticity in particular was thesubject of biomechanical investigation. We devel-oped an advanced experimental setup allowinghighly precise, reliable, and comparable measure-ments of important parameters of vicoelasticity(stiffness, compression, relaxation behavior). Thecompression–relaxation test (indentation test) hadto be minimally constraint because technically itwas not possible to cut the surfaces of the samplesexactly parallel. Consequently, sample dislocationwas observed during axial loading without use of aminimal cavity. The constraint portion of the testwas kept minimal compared to the sample height(approximately 8.8%)andwas constant throughoutthe whole setup. Its influence on the results isestimated to be very low. We do not have reason topresume that our results have been influenced inan unsystematic way. However, it represents amethodological limitation. We used a ball indenta-tion test to minimize the stress concentrations(shear stress) at the contact area of the indenter

Figure 7. (a,b) Autologous fibrochondrocytes show homogenous distribution within the scaffold2 weeks after cultivation (H&E, original magnification �10); (c,d) typically shaped autologousfibrochondrocytes line up along micropores 4 weeks after cultivation [(original magnification)c: H&E, �20; d: H&E, �100]. [Color figure can be viewed in the online issue, which is available athttp://www.interscience.wiley.com.]

1604 MAIER ET AL.


with the sample. A flat indenter would haveprovoked such shear stress and tensile reactionspossibly influencing the results substantially.Cyclic testing was performed to simulate physio-logical loading conditions and in order to gainmoreinformation about the viscoelastic behavior of themeniscus. Stiffness, compression, and the residualforce were measured as important parameters ofviscoelasticity. The dynamic compression phaseanalyzed the elastic properties of the meniscus.‘‘Stiffness’’ was measured during dynamic com-pression of the meniscus. For stiffness calculation,we used the linear-elastic part of the loadcurve (Fig. 2). The ‘‘compression’’ of the samplecharacterizes the ability of the sample to expandand is an indicator of viscosity. The ‘‘residual force’’is important for characterization of viscoelasticbehavior of material. It is influenced by the abilityof the tissue to expand under unconstraint com-pression (viscosity) and by the amount of resetforces present in the tissue (elasticity). Tissueswithhigh residual forceswould behavemore elasticthan viscous and vice versa. The digestion processsignificantly changed stiffness, compression, andthe residual force. After cyclic testing, stiffnesswasfound increased by 17%, compression increased by26%, and the residual force decreased by 6%. Thebiomechanical changes are well explained by theenzymatic digestion process. Enzymatic GAG-extraction decreases the ability of cartilage (menis-cus) to maintain water. Considering the function ofGAG’s as important load distributors, the observedchanges are likely to be mediated mainly byenzymatic GAG-extraction.43,44 Water exit out ofthe tissue proceeds faster and more intense indigested meniscus. Therefore, the extracellularmatrix contributes more to the measurement ofstiffness during dynamic compression resulting inan increase of stiffness. Increase of stiffnessthroughout cyclic loading is probably caused byprogressive compression of the tissue. Lowerresidual forces mean a decrease of the elasticproperties of the tissue, since the reset forceswithin the tissue diminish.We believe the decreaseof the residual force after the digestion can beexplained by the structural changes of the extra-cellular matrix (mainly GAG-extraction) includingthe collagen matrix (micropores, loosening). Theresidual force increased throughout cyclic testingin both native and processedmeniscus indicating acompaction of the extracellular matrix (Fig. 2).However, we did not find a significant differencebetween the groups.

Contrary to tendon studies, tensile testing didnot appear to be of particular relevance for this

study. The biomechanical test setup was chosenbecause it is a feasible and reliable instrument tostudy the viscoelastic properties of meniscaltissue. In our opinion, it represents advancementcompared to more traditional test methods (flatindentation test, constraint compression test) thathave considerable methodical restrictions whenstudying cartilaginous tissue. We therefore usedthis approved experimental setup for the bio-mechanical evaluation of meniscus, in particularconsideration of its viscoelastic properties.45

We treated native ovine meniscus with a self-developed nontoxic enzymatic process to obtain acell-free scaffold. Trypsin is successfully used, e.g.,for decellularization of heart valves.46,47 Usingsimilar concentrations and incubation periods, wecould not achieve complete cell removal in ovinemeniscus. Only outer areas of the meniscusshowed cell extraction, whereas extraction ofGAG was found throughout the complete cross-section histologically. Therefore, amulti-step enzy-matic process was established. We found that low-concentrated collagenase solutions are able tocreate micropores within the tissue and promotefurther cell removal. Concentrations applied were20–30-fold lower compared to meniscus cell isola-tion protocols.37,48 Mueller et al.49 used protease(2.5 mg/ml) for digestion of cell-seeded type Iand type II collagen-GAG matrices. Proteaseconcentrations used in our setup were approxi-mately sevenfold lower (0.375mg/ml).We achievedcomplete cell removal after protease incubationand the final washing procedure. Presumably,destruction of cell adhesion proteins promotes themechanism of ‘‘shaking out’’ the remaining cellsthroughout the porotic scaffold. SEM of the nativemeniscus showed circumscribed channels. Weinterpret those channels to function as diffusionchannels. SEM and histological analyses of thescaffold showed a much looser structure of thecollagen network with gaps and micropores. Amicroporotic structure facilitates cellular andvascular penetration for both in vitro and in vivotissue engineering applications.40,50 However, thetwo study groups examined cellular ingrowth afterimplantation of acellular scaffolds. Optimal archi-tectural properties (e.g., pore size, pore geometry,and pore volume) of cell-seeded scaffolds used formeniscal substitution still have to be defined.

Cell-based therapies already delivered promis-ing in vivo results.20,30,51 Stem cell research andgene therapymaywell be integrated into cell-basedstrategies. Martinek et al.52 demonstrated thefeasibility of gene transfer in meniscal allograftsin rabbits.



We managed to decellularize native meniscusand preserve important biomechanical properties.The resulting scaffold showed no histochemicalsigns of immunogenecity (MHC 1/MHC 2). Anontoxic scaffold for use of cell seeding andcultivation was obtained. To our knowledge, thisis the first study to seed meniscus cells into anallogenic meniscus scaffold and cultivate themsuccessfully over a period of 4 weeks as a viableconstruct. In our opinion, the presented scaffold is afirst, but important, step in the complex process ofreplacing a meniscus by means of tissue engineer-ing. The main problems seen in experimental andclinical use of scaffolds for meniscal replacementwere successfully addressed by this study. Furtherin vitro and in vivo studies are necessary. Thesterilization process also needs detailed investiga-tion prior to implantation studies. However, theinfluence of the sterilization process was notthe subject of the present study. For articularcartilage, we found no significant influence ofhigh-pressure sterilization on the biomechanicalproperties using an identical biomechanical testsetup.45

Seeding techniques, in particular, need to beoptimized and standardized. We observed acluster-like distribution pattern of the cells withinthe scaffold. However, we did not aim to reach auniform distribution of the cells in the context ofthis study but wanted to demonstrate the generaltechnical feasibility of seeding and cultivating cellswithin the scaffold. We used a manual needleinjection technique for reasons of practicability forthis invitro study.More standardized techniques ofcell seeding are required for scaffolds used in thecontext of in vivo studies (e.g., mechanical needleinjection). Also actual chemotactic seeding techni-quesmight behelpful to achieveamoreuniformcelldistribution throughout the scaffold.53

Cell differentiation and extracellular matrixproduction should be investigated in vitro. Theoptimal cell source (e.g., chondrocytes, meniscalcells, stem cells) also needs to be determined.Subsequent large animal studies (sheep) will helpto analyze the remodeling process of the trans-planted cell-seeded scaffold. The potential immu-nogenecity of the allogenic scaffold needs to betested out in vivo. It is also important to reanalyzebiomechanical properties of the construct afterimplantation in vivo.

The aim of this study is not to favor a xenogenicapproach of meniscal replacement. In our experi-mental rows, we also managed to decellularizehuman menisci at even lower enzymatic concen-trations. We therefore believe it is possible to

recellularize such a cell-free human allogenicscaffold with autologous human cells or humanstem cells in the described manner.

The authors see long-term clinical applicationsof tissue-engineered meniscus scaffolds in mini-mally invasive partial or total meniscus replace-ment. Since we suppose that immunologicalmatching will not be necessary, there is moreflexibility to custom-size the construct according toindividual requirements. Partial or total meniscusreplacement can be performed arthroscopically inmost cases.

ACKNOWLEDGMENTS

The electron microscopic analyses were performed withthe aid of the Department of Dermatology (Professor H.Behrendt, Center of Allergy and Environment (ZAUM),Technical University of Munich). We want to thankDr. I. Weichenmeier and Mrs. C. Weil for their kindassistance. The authors state that no professional orfinancial affiliations have biased the results of thisstudy. This study contains data of the doctoral thesis ofMrs. Nadine Roos.

REFERENCES

1. Ahmed AM, Burke DL, Yu A. 1983. In-vitro measurementof static pressure distribution in synovial joints—part II:retropatellar surface. J Biomech Eng 105:226–236.

2. Arnoczky SP, Warren RF, Spivak JM. 1988. Meniscalrepair using an exogenous fibrin clot. An experimentalstudy in dogs. J Bone Joint Surg [Am] 70:1209–1217.

3. Voloshin AS, Wosk J. 1983. Shock absorption of meniscec-tomized and painful knees: a comparative in vivo study.J Biomed Eng 5:157–161.

4. Burks RT, Metcalf MH, Metcalf RW. 1997. Fifteen-yearfollow-up of arthroscopic partial meniscectomy. Arthro-scopy 13:673–679.

5. Jackson RW, Rouse DW. 1982. The results of partialarthroscopic meniscectomy in patients over 40 years of age.J Bone Joint Surg [Br] 64:481–485.

6. Maletius W, Messner K. 1996. Chondral damage and agedepress the long-term prognosis after partial meniscec-tomy. A 12- to 15-year follow-up study. Knee Surg SportsTraumatol Arthrosc 3:211–214.

7. Kobayashi K, Amiel M, Harwood FL, et al. 2000. The long-term effects of hyaluronan during development of osteoar-thritis following partial meniscectomy in a rabbit model.Osteoarthritis Cartilage 8:359–365.

8. Ghosh P, Sutherland J, Bellenger C, et al. 1990. Theinfluence of weight-bearing exercise on articular cartilageof meniscectomized joints. An experimental study in sheep.Clin Orthop Relat Res 252:101–113.

9. Hede A, Svalastoga E, Reimann I. 1991. Articularcartilage changes following meniscal lesions. Repair and

1606 MAIER ET AL.


meniscectomy studied in the rabbit knee. Acta OrthopScand 62:319–322.

10. Agneskirchner JD, Burkart A, Imhoff AB. 2002. [Axisdeviation, cartilage damage and cruciate ligamentrupture—concomitant interventions in replacement ofthe anterior cruciate ligament]. Unfallchirurg 105:237–245.

11. Sudkamp NP, Kaab MJ. 2002. Biodegradable implants insoft tissue refixation: experimental evaluation, clinicalexperience, and future needs. Injury 33 (Suppl 2):B17–B24.

12. Wirth CJ, Peters G, Milachowski KA, et al. 2002. Long-term results of meniscal allograft transplantation. AmJ Sports Med 30:174–181.

13. Verdonk R. 2002. Meniscal transplantation. Acta OrthopBelg 68:118–127.

14. van Arkel ER, de Boer HH. 2002. Survival analysis ofhuman meniscal transplantations. J Bone Joint Surg [Br]84:227–231.

15. Kohn D, Rudert M, Wirth CJ, et al. 1997. Medial meniscusreplacement by a fat pad autograft. An experimental studyin sheep. Int Orthop 21:232–238.

16. Bruns J, Kahrs J, Kampen J, et al. 1998. Autologousperichondral tissue for meniscal replacement. J Bone JointSurg [Br] 80:918–923.

17. Klompmaker J, Veth RP, Jansen HW, et al. 1996. Meniscalreplacement using a porous polymer prosthesis: apreliminary study in the dog. Biomaterials 17:1169–1175.

18. Messner K, Gillquist J. 1993. Prosthetic replacement of therabbit medial meniscus. J Biomed Mater Res 27:1165–1173.

19. Rodeo SA, Seneviratne A, Suzuki K, et al. 2000.Histological analysis of human meniscal allografts. Apreliminary report. J Bone Joint Surg [Am] 82-A:1071–1082.

20. Peretti GM, Gill TJ, Xu JW, et al. 2004. Cell-based therapyfor meniscal repair: a large animal study. Am J Sports Med32:146–158.

21. Peretti GM, Bonassar LJ, Caruso EM, et al. 1999.Biomechanical analysis of a chondrocyte-based repairmodel of articular cartilage. Tissue Eng 5:317–326.

22. Coutts RD, Healey RM, Ostrander R, et al. 2001. Matricesfor cartilage repair. Clin Orthop Relat Res 391:S271–S279.

23. Tienen TG, Heijkants RG, Buma P, et al. 2003. A porouspolymer scaffold for meniscal lesion repair—a study indogs. Biomaterials 24:2541–2548.

24. Kobayashi M. 2004. A study of polyvinyl alcohol-hydrogel(PVA-H) artificial meniscus in vivo. Biomed Mater Eng 14:505–515.

25. Welch JA, Montgomery RD, Lenz SD, et al. 2002.Evaluation of small-intestinal submucosa implants forrepair of meniscal defects in dogs. Am J Vet Res 63:427–431.

26. Gastel JA, Muirhead WR, Lifrak JT, et al. 2001. Meniscaltissue regeneration using a collagenous biomaterialderived from porcine small intestine submucosa. Arthro-scopy 17:151–159.

27. Stone KR, Rodkey WG, Webber R, et al. 1992. Meniscalregeneration with copolymeric collagen scaffolds. In vitroand in vivo studies evaluated clinically, histologically, andbiochemically. Am J Sports Med 20:104–111.

28. Stone KR, Steadman JR, Rodkey WG, et al. 1997.Regeneration of meniscal cartilage with use of a collagen

scaffold. Analysis of preliminary data. J Bone Joint Surg[Am] 79:1770–1777.

29. Rodkey WG, Steadman JR, Li ST. 1999. A clinical study ofcollagen meniscus implants to restore the injured menis-cus. Clin Orthop 367:S281–S292.

30. Martinek V, Ueblacker P, Braun K, et al. 2006. Secondgeneration of meniscus transplantation: in-vivo study withtissue engineered meniscus replacement. Arch OrthopTrauma Surg 126:228–234.

31. Mikos AG, McIntire LV, Anderson JM, et al. 1998. Hostresponse to tissue engineered devices. Adv Drug Deliv Rev33:111–139.

32. Peretti GM, Randolph MA, Zaporojan V, et al. 2001. Abiomechanical analysis of an engineered cell-scaffoldimplant for cartilage repair. Ann Plast Surg 46:533–537.

33. Duda GN, Haisch A, Endres M, et al. 2000. Mechanicalquality of tissue engineered cartilage: results after 6 and12 weeks in vivo. J Biomed Mater Res 53:673–677.

34. Bancroft JD, Gamble M. 2002. Theory and practice ofhistological techniques, 5th ed. London/NewYork: Church-ill Livingstone.

35. Farndale RW, Buttle DJ, Barrett AJ. 1986. Improvedquantitation and discrimination of sulphated glycosami-noglycans by use of dimethylmethylene blue. BiochimBiophys Acta 883:173–177.

36. Diehl P, Reuning U, Gollwitzer H, et al. 2004. Effect ofextracorporal high hydrostatic pressure on tumor celladherence and viability. Oncol Rep 12:369–373.

37. Webber RJ. 1990. In vitro culture of meniscal tissue. ClinOrthop Relat Res 252:114–120.

38. Buma P, Ramrattan NN, van Tienen TG, et al. 2004.Tissue engineering of the meniscus. Biomaterials 25:1523–1532.

39. van Tienen TG, Heijkants RG, de Groot JH, et al. 2003.Presence and mechanism of knee articular cartilagedegeneration after meniscal reconstruction in dogs.Osteoarthritis Cartilage 11:78–84.

40. van Tienen TG, Heijkants RG, Buma P, et al. 2002. Tissueingrowth and degradation of two biodegradable porouspolymers with different porosities and pore sizes. Bioma-terials 23:1731–1738.

41. Setton LA, Guilak F, Hsu EW, et al. 1999. Biomechanicalfactors in tissue engineered meniscal repair. Clin Orthop367:S254–S272.

42. de Groot JH, Zijlstra FM, Kuipers HW, et al. 1997.Meniscal tissue regeneration in porous 50/50 copoly(L-lactide/epsilon-caprolactone) implants. Biomaterials 18:613–622.

43. Zhu W, Mow VC, Koob TJ, et al. 1993. Viscoelastic shearproperties of articular cartilage and the effects of glycosi-dase treatments. J Orthop Res 11:771–781.

44. Schmidt MB, Mow VC, Chun LE, et al. 1990. Effects ofproteoglycan extraction on the tensile behavior of articularcartilage. J Orthop Res 8:353–363.

45. Diehl P, Naal FD, Schauwecker J, et al. 2006.[Biomechanical properties of articular cartilage after highhydrostatic pressure treatment]. Biomed Tech (Berl) 51:8–14.

46. Rieder E, Kasimir MT, Silberhumer G, et al. 2004.Decellularization protocols of porcine heart valves differimportantly in efficiency of cell removal and susceptibilityof the matrix to recellularization with human vascularcells. J Thorac Cardiovasc Surg 127:399–405.

47. Schenke-Layland K, Vasilevski O, Opitz F, et al. 2003.Impact of decellularization of xenogeneic tissue on



extracellular matrix integrity for tissue engineering ofheart valves. J Struct Biol 143:201–208.

48. Peretti GM, Caruso EM, Randolph MA, et al. 2001.Meniscal repair using engineered tissue. J Orthop Res19:278–285.

49. Mueller SM, Shortkroff S, Schneider TO, et al. 1999.Meniscus cells seeded in type I and type II collagen-GAGmatrices in vitro. Biomaterials 20:701–709.

50. Klompmaker J, Jansen HW, Veth RP, et al. 1993. Porousimplants for knee joint meniscus reconstruction: apreliminary study on the role of pore sizes in ingrowth

and differentiation of fibrocartilage. Clin Mater 14:1–11.

51. Ghadially FN, Wedge JH, Lalonde JM. 1986. Experimen-tal methods of repairing injuredmenisci. J Bone Joint Surg[Br] 68:106–110.

52. Martinek V, Usas A, Pelinkovic D, et al. 2002. Geneticengineering of meniscal allografts. Tissue Eng 8:107–117.

53. DeLong SA, Moon JJ, West JL. 2005. Covalently immobi-lized gradients of bFGF on hydrogel scaffolds for directedcell migration. Biomaterials 26:3227–3234.

1608 MAIER ET AL.


in vitro analysis of an allogenic scaffold for tissue-engineered meniscus replacement

Documents