OsteoArthritis and Cartilage (2007) 15, 48e58

ª 2006 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.joca.2006.06.009

InternationalCartilageRepairSociety

Differential cartilaginous tissue formation by human synovial membrane,fat pad, meniscus cells and articular chondrocytes1

A. Marsano M.Sc.y, S. J. Millward-Sadler Ph.D.z, D. M. Salter M.D.z, A. Adesida Ph.D.x,T. Hardingham Ph.D.x, E. Tognana Ph.D.k, E. Kon M.D.{, C. Chiari-Grisar M.D.#,S. Nehrer M.D.#, M. Jakob M.D.y and I. Martin Ph.D.y*yDepartments of Surgery and Research, University Hospital Basel, Basel, SwitzerlandzDepartment of Pathology, University of Edinburgh, UKxUK Center for Tissue Engineering, University of Manchester, UKkFidia Advanced Biopolymers, Biosurgery Division, Abano Terme, Italy{ Istitutes Orthopedics Rizzoli, Bologna, Italy# Department of Orthopaedics, Medical University of Vienna, Austria

Summary

Objective: To identify an appropriate cell source for the generation of meniscus substitutes, among those which would be available by arthros-copy of injured knee joints.

Methods: Human inner meniscus cells, fat pad cells (FPC), synovial membrane cells (SMC) and articular chondrocytes (AC) were expandedwith or without specific growth factors (Transforming growth factor-beta1, Fibroblast growth factor-2 and Platelet-derived growth factor bb,TFP) and then induced to form three-dimensional cartilaginous tissues in pellet cultures, or using a hyaluronan-based scaffold (Hyaff�-11),in culture or in nude mice. Human native menisci were assessed as reference.

Results: Cell expansion with TFP enhanced glycosaminoglycan (GAG) deposition by all cell types (up to 4.1-fold) and messenger RNAexpression of collagen type II by FPC and SMC (up to 472-fold) following pellet culture. In all models, tissues generated by AC containedthe highest fractions of GAG (up to 1.9% of wet weight) and were positively stained for collagen type II (specific of the inner avascular regionof meniscus), type IV (mainly present in the outer vascularized region of meniscus) and types I, III and VI (common to both meniscus regions).Instead, inner meniscus, FPC and SMC developed tissues containing negligible GAG and no detectable collagen type II protein. Tissuesgenerated by AC remained biochemically and phenotypically stable upon ectopic implantation.

Conclusions: Under our experimental conditions, only AC generated tissues containing relevant amounts of GAG and with cell phenotypescompatible with those of the inner and outer meniscus regions. Instead, the other investigated cell sources formed tissues resembling onlythe outer region of meniscus. It remains to be determined whether grafts based on AC will have the ability to reach the complex structuraland functional organization typical of meniscus tissue.ª 2006 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.

Key words: Chondrogenesis, Meniscus, Fibrocartilage, Tissue engineering.

Introduction

Meniscus is a complex fibrocartilaginous tissue, which isessential in the knee joint for shock absorption, load distri-bution, maintenance of stability and protection of articularcartilage1e3. Injuries to the meniscus are often treated bypartial or total meniscectomy, which is known to be associ-ated with detrimental changes in joint function, ultimatelyincreasing the risk of early degenerative joint diseases4e6.Surgical approaches currently in use to substitute the

1Supported by the Swiss Federal Office for Education and Sci-ence Bundesamt fur Bildung und Wissenschaft (B.B.W.) under theFifth European Framework Growth Program (MENISCUS REGEN-ERATION, contract G5RD-CT-2002-00703).

*Address correspondence and reprint requests to: Ivan Martin,Ph.D., Institute for Surgical Research and Hospital Management,University Hospital Basel, Hebelstrasse 20, ZLF, Room 405, 4031Basel, Switzerland. Tel: 41-61-265-2384; Fax: 41-61-265-3990;E-mail: imartin@uhbs.ch

Received 21 October 2005; revision accepted 17 June 2006.

48

damaged meniscus (e.g., the use of allografts or of a colla-gen-based material) can initially restore a stable and pain-free joint, but long-term clinical results, especially relatedto the protection of the articular surface, are stilluncertain7,8. Recently, tissue engineering strategies havebeen proposed for the generation of meniscus substitutes,based on the loading and culture of suitable cells intoappropriate biodegradable porous scaffolds9e12. The prom-ising approach has been pre-clinically validated by severalstudies11,13, indicating the potential of cell-based meniscussubstitutes to improve healing of meniscus tissue.

Several cell sources have been used for meniscus repair,including meniscus fibrochondrocytes11,12,14,15, chondro-cytes13,16,17 or bone marrow-derived mesenchymalprogenitor cells18e20, but to the best of our knowledge a directcomparison of the different cell types has not yet been estab-lished. With the ultimate goal to identify an appropriate cellsource for the generation of meniscus substitutes, the aimof our work was to compare the growth and post-expansionchondrogenic capacity of different cell types that would be

49Osteoarthritis and Cartilage Vol. 15, No. 1

readily available during meniscectomy or arthroscopic exam-ination of an injured knee joint. In particular, we investigatedthe following human cell types: (1) inner meniscus cells (IMC)(hereafter referred to as ‘‘meniscus cells’’), which constitutethe native target tissue and can be expanded even followingmeniscus injury15; (2) fat pad cells (FPC), known to includemultipotent mesenchymal progenitor cells21; (3) synovialmembrane cells (SMC), reported to have a progenitor na-ture22,23 and to be involved in healing of cartilage lesions24,25;and (4) articular chondrocytes (AC), currently in clinical usefor articular cartilage repair and capable of differentiationinto a variety of mesenchymal tissues26e28. All cells were ex-panded with or without growth factors previously reported toenhance chondrogenesis of different chondrocyte types29,30,and then induced to form three-dimensional (3D) cartilagi-nous tissues in a scaffold-free model system (i.e., pellet cul-ture), or using a hyaluronan-based scaffold, following in vitroculture or in vivo ectopic implantation. Each model includeda culture phase in the presence of chondrogenic supple-ments, typically used to induce the expression of glycosami-noglycan (GAG) and collagens31,32. Tissues were assessedfor the content of GAG, and the expression pattern of differentcollagen types was compared to that in the different regionsof native human menisci.

Materials and methods

HUMAN MATERIAL

Meniscus, fat pad and synovial membrane tissues wereobtained from a total of 15 donors (26e68 years of age,mean 45), after informed consent. Biopsies were harvestedduring partial meniscectomy or anterior cruciate ligamentreconstruction, no longer than 2 months after trauma. Atthe time of surgery, knee joints were not overtly inflamedor osteoarthritic, although a few cases displayed minorinflammation (4/15 donors), International Cartilage RepairSociety score (ICRS) Grade 1 cartilage damage (2/15 do-nors), or moderate osteoarthrosis (3/15 donors). Meniscus(n¼ 15), fat pad (n¼ 7) and synovial membrane (n¼ 12) bi-opsies were all relatively small in size, with a weight rangeof, respectively, 10e500 mg, 10e100 mg, and 20e400 mg.Meniscus biopsies were only from the inner avascular re-gion, where injured tissue is typically resected. Synovialmembrane was from the suprapatellar recess and adiposetissue from the infrapatellar fat pad. For ethical reasons, ar-ticular cartilage could not be harvested from the same do-nors. Cartilage tissues were thus collected from the medialfemoral condyle of six cadavers (22e55 years of age,mean 42), with no known clinical history of joint disorders,within 24 h after death. Based on a preliminary study, thegrowth and differentiation ability of chondrocytes isolatedfrom macroscopically normal cartilage shortly after kneejoint trauma was within the range of chondrocytes isolatedfrom cadaver joints. The weight range for the articular carti-lage specimens was 500e1000 mg.

Complete human meniscus specimens were also obtainedwith ethical permission and consent from 32 individuals madeanonymous (18e86 years of age, mean 56), undergoingabove knee amputation or partial meniscectomy for traumaticor degenerative tears. Complete meniscus specimens wereprocessed for immunohistochemical analysis as describedbelow and used to define the pattern of collagen typesexpressed in the deep inner avascular and deep outer vascu-larized regions of the native meniscus tissue. Analysis did notinclude characterization of the superficial layer of themeniscus.

CELL ISOLATION AND CULTURE

Inner meniscus and articular cartilage biopsies were finelyminced and digested by incubation for 22 h at 37�C in0.15% type II collagenase (Worthington BiochemicalCorporation, Lakewood, NJ) (1 mL solution per 100 mgtissue)33. Meniscus cells and AC were resuspended inDulbecco’s Modified Eagle’s Medium (DMEM; 4.5 g/Lglucose with nonessential amino acids), containing 10%fetal bovine serum, 4.5 mg/mL glucose, 0.1 mM nonessentialamino acids, 1 mM sodium pyruvate, 100 mM 4-(2-hydrox-yethyl)-1-piperazineethanesulfonic acid (HEPES) buffer,100 U/mL penicillin, 100 mg/mL streptomycin and 0.29 mg/mL L-glutamine (control medium).

Fat pad and synovial membrane biopsies were grosslycut and digested by incubation, respectively, for 7 h andfor 22 h in 0.4% type II collagenase (1 mL solution per40 mg tissue). For the isolation of FPC, floating adipocyteswere removed by centrifugation at 300 g for 5 min34. Colla-genase concentration and incubation times were selectedby adapting a previously published protocol22, followingpreliminary tests.

Meniscus, FPC, SMC and AC were plated in tissue cul-ture flasks, at a density of 104 cells/cm2, and cultured ina humidified incubator at 37�C and 5% CO2. Throughoutthe phase of expansion in monolayer, all cell sourceswere cultured in control medium, without or with furthersupplementation of 1 ng/mL Transforming growth factor-beta1 (TGF-b1), 5 ng/mL Fibroblast growth factor-2 and10 ng/mL Platelet-derived growth factor-bb (TFP medium).The growth factors (all from R&D Systems, Minneapolis,MN) were selected based on their previously reported ca-pacity to enhance the growth and post-expansion chondro-genic ability of different human chondrocyte types29,30.

DIFFERENTIATION ASSAYS

Upon reaching subconfluence, meniscus, FPC, SMC andAC were detached using 0.05% trypsin/0.53 mM ethylene-diaminetetraacetic (EDTA) (GIBCO-BRL, CH) and replatedat a density of 5� 103 cells/cm2. Following an additionalpassage of expansion, cells were used in the three differentmodels described below.

Model I: chondrogenic differentiation in 3Dpellet culture

Meniscus, FPC, SMC and AC from all donors, expandedwith or without TFP, were suspended in a defined serumfree medium, consisting of DMEM supplemented withInsulineTransferrineSelenium (ITSþ1) (10 mg/mL insulin,5.5 mg/mL transferrin, 5 ng/mL selenium, 0.5 mg/mL bovineserum albumin, 4.7 ng/mL linoleic acid, from Sigma Chemi-cal, St Louis, MO), 1 mM sodium pyruvate, 100 mM HEPESbuffer, 100 U/mL penicillin, 100 mg/mL streptomycin,0.29 mg/mL L-glutamine, 0.1 mM ascorbic acid 2-phosphate,1.25 mg/mL human serum albumin, 10 ng/mL TGF-b1 and10�7 M dexamethasone. The serum free medium was initiallydeveloped for the chondrocytic differentiation of human mes-enchymal progenitor cells35 and is typically used for the redif-ferentiation of human AC29. Aliquots of 5� 105 cells in 0.5 mLof serum free medium were centrifuged in 1.5 mL conicalpolypropylene tubes (Sarstedt, Numbrecht, D) at 1300 rpmfor 4 min to form spherical pellets. These were placed ona 3D orbital shaker (Bioblock Scientific, Frenkendorf, CH)at 30 rpm in a humidified incubator at 37�C and 5% CO2.

Pellets were cultured for 2 weeks, with medium changestwice per week, and subsequently processed for histological,

50 A. Marsano et al.: Comparison of chondrogenic cells

immunohistochemical, biochemical or messenger RNA(mRNA) analysis as described below. Each analysis wasperformed independently in at least two entire pellets foreach primary culture and expansion condition.

Model II: chondrogenic differentiation in 3Dscaffold based cultures

Meniscus, FPC, SMC and AC from two donors,expanded with TFP, were seeded into non-woven meshes(5 mm diameter, 2 mm thick disks), made of esterified hya-luronan (Hyaff�-11, Fab, Abano Terme, IT) at the density4� 106 cells/scaffold (7� 107 cells/cm3). The scaffold wasselected based on previous reports indicating that it sup-ports redifferentiation of human AC36 and since it is alreadyin clinical use for articular cartilage repair37. Cells were re-suspended in 28 mL of control medium and slowly dispersedover the top surface of the dry meshes with a micropipette.The seeded scaffolds were transferred to a 37�C incubatorto allow for initial cell attachment. After 45 min, 100 mL ofmedium was carefully added to the base of each well. Aftera further 1.5 h of incubation, 2 mL of medium was slowlyadded along the side of each well to cover the scaffold.Cell-scaffold constructs were cultured in control mediumsupplemented with 1 ng/mL of TGFb-3, 0.1 mM ascorbicacid 2-phosphate and 10 mg/mL of human insulin, withmedium changes twice a week. After 2, 4 and 6 weeks,the resulting tissues were processed for histological,immunohistochemical, biochemical or mRNA analysis asdescribed below. Each analysis was performed indepen-dently in at least three engineered tissues for each primaryculture and cell source.

Model III: ectopic implantation in nude mice

Meniscus, FPC, SMC and AC from one donor, expandedwith TFP, were loaded and cultured for 2 weeks intoHyaff�-11 meshes as described for Model II and implantedsubcutaneously in nude mice (CD-1 nude/nude, CharlesRiver, Germany). After 6 weeks, mice were euthanisedand explants processed for histological, immunohistochem-ical, biochemical or mRNA analysis as described below.Each analysis was performed independently in two engi-neered tissues for each cell source.

ANALYTICAL METHODS

Proliferation rate during expansion

The number of doublings of each cell type during the sec-ond passage of expansion was determined as the logarithmin base 2 of the fold increase in the number of cells duringexpansion. The proliferation rate was defined as the num-ber of doublings during the second passage of expansiondivided by the time required for expansion and was ex-pressed as doublings/day.

Real time Reverse Transcription-Polymerase ChainReaction (RT-PCR) (global amplification)

Total RNA was prepared from pellet cultures using Tri-Reagent (Sigma, UK). Pellet cultures were ground up inthe Tri-Reagent using Molecular Grinding Resin (GenoTechnology Inc, St Louis, USA). Complementary DNA(cDNA) was synthesised from 10 to 100 ng of total RNAby global amplification38. Globally amplified cDNA wasdiluted 1:1000 and 1 ml aliquots of the diluted cDNA were

amplified by real time RT-PCR in a 25 mL reaction volumeon a MJ Research Opticon using a SYBR Green Core Kit(Eurogentec, Seraing, Belgium), with gene specific primersdesigned using ABI Primer Express software. Relativeexpression levels were normalised using glyceraldehyde3-phosphate dehydrogenase (GAPDH) and calculatedusing the 2�DCt method39. All primers were from Invitrogen,Paisley, UK. Primer sequences were for (1) GAPDH: for-ward 50e30 CACTCAGACCCCCACCACAC, and reverse50e30 GATACATGACAAGGTGCGGCT; (2) collagen typeI: forward 50e30 TTGCCCAAAGTTGTCCTCTTCT, andreverse 50e30 AGCTTCTGTGGAACCATGGAA; (3) colla-gen type II: forward 50e30 CTGCAAAATAAAATCTCGGTGTTCT, and reverse 50e30 GGGCATTTGACTCACACCAGT;(4) SOX9: forward 50e30 CTTTGGTTTGTGTTCGTGTTTTG, and reverse 50e30 AGAGAAAGAAAAAGGGAAAGGTAAGTTT. The expression levels of collagen types I, IIand SOX-9 were investigated to determine the relativeextent of cell differentiation towards a fibrocartilaginous orhyaline-like phenotype.

Biochemistry

Tissues generated by meniscus, FPC, SMC and AC weredigested with protease K (1 mg/mL protease K in 50 mMTris with 1 mM EDTA, 1 mM iodoacetamide, and 10 mg/mL pepstatin-A; 0.5 mL solution for cell pellets and 1 mLfor Hyaff�-11 based constructs) for 15 h at 56�C34. TheGAG content was measured spectrophotometrically usingdimethylmethylene blue40, with chondroitin sulfate as a stan-dard, and normalized to the DNA amount, measured usingthe CyQUANT cell proliferation assay kit (Molecular Probes,Eugene, OR), with calf thymus DNA as a standard.

Histology

Tissues were rinsed in Phosphate Buffer Solution (PBS),fixed in 4% buffered formalin for 24 h at 4�C, embedded inparaffin, and sectioned to a thickness of 5 mm for cell pelletsand of 7 mm for Hyaff�-11 based constructs. Sections werestained with Safranin-O for sulfated GAG.

Immunohistochemistry

Cryostat sections (3e5 mm thick) were cut from tissuesand pellets stored frozen at �70�C. Sections, on Super-frost slides (BDH, Poole, UK), were fixed for 5 min in icecold acetone and then stored at �20�C until required forimmunostaining by standard ABC immunoperoxidasemethod with a panel of antibodies against collagen typeI (COL-1 Mouse, Sigma, Gillingham, UK), collagen typeII (CIIC1, Developmental Studies Hybridoma Bank, IowaCity, USA), collagen type III (FH-7A, Sigma, Gillingham,UK), collagen type IV (CIV22 Dako, Ely, UK) and collagentype VI (poly, Dako, Ely, UK). Sections immunostainedwith anti-type I and type II collagen antibodies were pre-incubated with 1 mg/mL hyaluronidase (Sigma type 4hyaluronidase from bovine testes) for 30 min at 37�C.Negative control sections with appropriate immunoglobulinG (IgG) or non-immune animal serum were run in parallelwith the experimental sections to confirm the specificity ofany positive staining.

The slides were blindly examined by light microscopy andthe results were scored based on the percentage of positivetissue as: no matrix positive (�); <50% of matrix positive(þ/�); and >50% of matrix positive (þ).

51Osteoarthritis and Cartilage Vol. 15, No. 1

Table IImmunophenotypic characterization of native and engineered tissues

A: Meniscus B: Model I C: Model II D: Model III

IMC FPC SMC AC IMC FCP SMC AC AC

IN OUT CTR TFP CTR TFP CTR TFP CTR TFP

Col I þ þ þ þ þ þ þ þ þ þ þ þ þ þ þCol II þ/� � � � � � � � þ/� þ/� � � � þ þCol III þ þ þ/� þ þ/� þ þ/� þ þ þ þ þ þ þ/� þCol IV þ/� þ þ þ þ/� þ þ þ þ/� þ þ þ/� þ þ þCol VI þ þ þ þ þ þ þ þ þ þ þ þ þ þ þ

The expression of collagen types I, II, III, IV and VI was assessed in the inner avascular (IN) and outer vascular (OUT) regions of native

human meniscus (section A), in pellets generated by IMC, FPC, SMC cells and AC expanded without (control, CTR) or with TFP, after 2 weeks

(section B), in tissue constructs generated by the four different cell sources after 6 weeks of culture into Hyaff�-11 (section C), and in tissue

constructs generated by AC cultured for 2 weeks into Hyaff�-11 and implanted ectopically for 6 weeks (section D). Scores are based on the

percentage of positive tissue as no matrix positive (�); <50% of matrix positive (þ/�); and >50% of matrix positive (þ).

STATISTICAL ANALYSIS

Values are presented as mean� standard deviation. Dif-ferences between the experimental groups were evaluatedby ManneWhitney U tests and considered statisticallysignificant with P< 0.05.

Results

CHARACTERIZATION OF NATIVE HUMAN MENISCUS

The inner and outer regions of human native meniscusdisplayed a uniformly positive immunohistochemical stain-ing for collagen types I, III and VI. Instead, staining forcollagen type II, the major fibril collagen in articular carti-lage, was weakly positive only in the inner meniscus region.Staining for collagen type IV, usually absent from articularcartilage, was positive, predominantly in a dendritic pattern,in the outer vascular zone of the meniscus and also, but toa lower extent, in the inner meniscus (Table I, section A), aspreviously described for ovine meniscus41.

CELL PROLIFERATION DURING MONOLAYER EXPANSION

Meniscus, FPC, SMC and AC cultured in controlmedium proliferated at similar rates (Fig. 1). However,probably due to the limited biopsy size (i.e., if lowerthan 20 mg), some cultures of meniscus (three primaries),FPC (four primaries) and SMC (one primary) could not beexpanded.

TFP supplementation during expansion significantlyincreased the proliferation rate of all four cell types, andallowed cell expansion even in cases of limited biopsysize. In the presence of TFP, AC proliferated significantlyfaster than all other cell sources.

MODEL I: 3D PELLET CULTURES

After 2 weeks of culture, meniscus, SMC and ACexpanded in control medium formed spherical pellets, incontrast to FPC, which did not generate regular tissue struc-tures (Fig. 2). Following expansion in control medium, onlyAC generated tissues positively stained for GAG. Mediumsupplementation with TFP during expansion of all cell sour-ces resulted in pellets more intensely stained for GAG. Achondrocytic cell morphology and overtly positive staining

for GAG were observed in pellets generated by AC and inrestricted regions of pellets generated by FPC and SMC.

Pellets generated by all cell sources, expanded with orwithout TFP, contained similar amounts of DNA (data notshown). Following expansion in control medium, AC gener-ated pellets with significantly higher GAG/DNA fractionsthan the other cell sources [Fig. 3(A)]. Medium supplemen-tation with TFP during expansion induced a significantincrease in the GAG/DNA content of pellets generated byall cell sources (1.4, 4.1, 1.9 and 3.1 fold, respectively, formeniscus, FPC, SMC and AC). Following TFP expansion,tissues formed by AC contained the largest fractions ofGAG/DNA.

Following expansion in control medium and differentiationin pellets, AC expressed the highest mRNA levels ofcollagen types I, II and SOX-9 [Fig. 3(BeD)]. Mediumsupplementation with TFP during expansion induced a sig-nificant increase in the mRNA expression of collagen type IIby FPC and SMC (respectively 66- and 472-fold). FollowingTFP expansion and differentiation in pellets, AC expressedmRNA levels of SOX-9 higher than all other cell sourcesand of collagen type II higher than meniscus and SMC.

Pellets generated by all cell sources displayed positiveimmunohistochemical staining for collagen types I, III, IVand VI, with generally increased levels of staining followingTFP expansion (Table I, section B). Only AC, expandedwith or without TFP, formed tissues positively stained forcollagen type II, although only in discrete areas.

0

0.5

1

doub

lings

/day

*

**

*

IMC FPC SMC AC

Proliferation Index

CTRTFP

°

Fig. 1. Proliferation index for IMC, FPC, SMC and AC expandedwithout control (CTR) or with growth factors (TFP). *¼ significantlydifferent from same cells expanded without TFP. � ¼ significantly

different from IMC, FPC and SMC.

52 A. Marsano et al.: Comparison of chondrogenic cells

Fig. 2. Representative sections of pellets generated by meniscus (A, B), FPC (C, D), SMC (E, F) and AC (G, H) expanded with (B, D, F, H) orwithout (A, C, E, G) the growth factor combination TFP and stained by Safranin-O. Bar¼ 100 mm.

MODEL II: CULTURE INTO HYAFF�-11 MESHES

All cell sources, expanded in medium supplemented withTFP and loaded into Hyaff�-11 scaffolds, generated tissueswhich developed with in vitro culture time. In particular, DNAfractions of wet weight decreased with time, reaching after 6weeks similar levels in all constructs (between 0.062% and0.071% of the wet weight). The GAG content of all con-structs, expressed as percentage of wet weight, increasedwith culture time, and after 6 weeks’ culture was the highestin tissues generated by AC (1.90� 0.09% of the wet weight,respectively, 2.6-, 2.9- and 3.0-fold higher than in tissuesformed by meniscus, FPC and SMC), approaching thelevels measured in the inner region of native menisci

(approximately 2%, calculated based on the study of Auf-derHeide and Athanasiou 42). Histological staining of tissuecross-sections indicated that only AC were able after 6weeks to engineer cartilaginous tissues positively stainedfor GAG (Fig. 4).

The pattern of collagen types deposited in cell-Hyaff�-11constructs was similar to that described in cell pellets(Table I, section C). In particular, tissues generated by ACwere positively stained for all assessed collagen types,including collagen type II, thus indicating a hybrid fibro-hyaline cartilaginous nature (Fig. 5). Instead, tissues formedby meniscus, FPC and SMC were negatively stained forcollagen type II.

53Osteoarthritis and Cartilage Vol. 15, No. 1

MODEL III: ECTOPIC IMPLANTATION FOLLOWING CULTURE

INTO HYAFF�-11 MESHES

In order to assess the intrinsic capacity of engineeredcartilage tissues to further develop upon implantation in anenvironment supportive, but not inductive, of chondrogene-sis43, constructs generated by all cell sources, following 2weeks of pre-culture in vitro44, were implanted in subcutane-ous pockets in nude mice for 6 weeks. Following in vivoimplantation, the DNA content decreased and reached sim-ilar levels in all constructs (between 0.030% and 0.047% ofthe wet weight). In parallel, the GAG content increased in allconstructs, and reached the highest levels in tissues

AGAG/DNA

0

5

10

IMC FPC SMC AC

ug

/ u

g

CTRTFP

*

* * *

°

IMC FPC SMC AC

CTRTFP

BCollagen type I mRNA levels

1.E-04

1.E-02

1.E+00

1.E+02

no

rm

alized

to

G

AP

DH

°

IMC FPC SMC AC

CTRTFP

1.E-04

1.E-02

1.E+00

1.E+02

no

rm

alized

to

G

AP

DH

SOX-9 mRNA levels

D

°

°

IMC FPC SMC AC

CTRTFP

1.E-04

1.E-02

1.E+00

1.E+02

no

rm

alized

to

G

AP

DH

Collagen type II mRNA levels

C

°*§

§

*

Fig. 3. Fractions of GAG (A) and mRNA expression levels of colla-gen types I (B), II (C) and Sox9 (D) in pellets generated by IMC,FPC, SMC and AC, expanded without (control, CTR) or with growthfactors (TFP). *¼ significantly different from same cell source ex-panded without TFP. � ¼ significantly different from all the othercell sources, expanded in the same condition. x ¼ significantly dif-

ferent from AC, expanded in the same condition.

generated by AC (1.4� 0.3% of the wet weight, respec-tively, 2.6-, 1.9- and 3.5-fold higher than in tissues formedby meniscus, FPC and SMC). Histologically, tissues formedby AC were qualitatively similar to those maintained in cul-ture for a total of 6 weeks (model II), with a mixture of fibro-blastic and chondrocytic cell phenotypes, large intercellularspaces positively stained for GAG, and no evidence of vas-cularization (Fig. 6). Instead, tissues formed by meniscus,FPC and SMC contained mostly fibroblastic cells and negli-gible extracellular matrix, with no detectable positive stainfor GAG. As compared to AC-based constructs, the loweramount of extracellular matrix in meniscus, FPC- and syno-vial membrane cells-based tissues appeared to be associ-ated with a more advanced degradation of the esterifiedhyaluronan fibers, which were markedly more swollen, likelydue to increased absorption of water (Fig. 6). Immunohisto-logical staining of explant cross-sections indicated thattissues generated by AC were positively stained forall assessed collagen types, including type II (Table I,section D). Instead, the amount of extracellular matrix inmeniscus, FPC- and SMC-based tissues was too limitedto allow for reliable immunohistochemical characterization.

Discussion

In this study, with the ultimate goal of identifying a suitablecell source for engineering autologous meniscus grafts, wecompared the growth and post-expansion chondrogenic ca-pacity of human meniscus, FPC, SMC and AC using threedifferent model systems. Our results indicate that (1) ACreached the highest proliferation rates and (2) only tissuesformed by expanded AC contained molecules typicallyfound in the inner avascular region of the meniscus (i.e.,GAG and collagen type II), in the outer vascularized regionof the meniscus (i.e., collagen type IV), and common to bothregions (i.e., collagen types I, III and VI). Instead, under ourspecific experimental setup, tissues formed by meniscus,FPC, and SMC contained negligible GAG amounts andno detectable collagen type II.

We initially studied the response of the different cellsources to a specific growth factor combination (i.e.,TFP) supplemented during cell expansion. Medium sup-plementation with TFP induced a significant increase inthe proliferation rate during expansion and in the deposi-tion of GAG in pellet cultures by all cell sources. In addi-tion, it induced an upregulation of the mRNA levels ofcollagen type II in pellet cultures by FPC and SMC. Wepreviously reported that in addition to AC29, also chondro-cytes from ear and nasal septum increased their growthand post-expansion ability to generate cartilaginoustissues in response to TFP30. The fact that different chon-droprogenitor cells and mature chondrocytes from differentcartilage sources similarly responded to the TFP growthfactor combination during expansion suggests that thesedifferent cell types share certain pathways controllingchondrogenic commitment; in this context, it would beinteresting to assess which common specific gene setsare up- or down-regulated in the different cell sourceswhen exposed to TFP.

Expanded meniscus, FPC and SMC displayedevidences of differentiation towards the chondrogeniclineage, as assessed by the detection of type II collagenand Sox-9 mRNA expression. However, these cellsdeveloped merely fibrocartilaginous tissues, as indicatedby the low amounts of deposited GAG and by the immuno-histochemical detection of all the investigated collagen

54 A. Marsano et al.: Comparison of chondrogenic cells

Fig. 4. Representative Safranin-O stained sections of tissues generated by culture of meniscus (A), FPC (B), SMC (C) and AC (D) intoHyaff�-11 for 6 weeks. Bar¼ 300 mm.

types, with the exception of type II collagen. Previous stud-ies reported the production of type II collagen bymeniscus45, FPC21 and SMC22. The apparent discrepancymight be explained by the use of different detectionmethods and/or by the fact that biopsies taken with differ-ent methods or from healthy vs injured knees may result ina different quality of the isolated cells. Indeed, one mainlimit of our study is related to the use of likely heteroge-neous cells, without definition of typical phenotypes by as-sessment of surface markers. Our in vivo results areanyway consistent with the work by De Bari et al.46, re-porting that no stable cartilage could be formed ectopicallyby SMC, and with that by Park et al.47, indicating thatchondrogenesis by synovium-derived cells was not com-plete, and anyway inferior e with respect to type II colla-gen and Sox 9 mRNA expression e to that by AC. Interms of mRNA expression of chondrogenic genes andGAG deposition, our findings are also in line with other re-cent studies, reporting the chondrogenic differentiation ofhuman SMC in the presence of high concentrations(500 ng/mL) of bone morphogenetic protein (BMP)-223,48.However, since these studies did not include assessmentof type II collagen at the protein level, a direct comparisonwith our data, potentially addressing the role of BMP-2 inthe system, cannot be established.

In all investigated models, cartilaginous tissues gener-ated by AC contained all the investigated collagen typesin addition to abundant GAG and thus displayed a mixedhyaline/fibrous phenotype, consistently with several previ-ous studies33,49e51. The quality of engineered cartilagebased on AC seeding into Hyaff�-11 meshes improvedduring time in culture or during the time of subcutaneousimplantation, as assessed by the increased content ofGAG and decreased cellularity. Moreover, cells remainedphenotypically stable and tissues implanted in vivo werenot vascularized, also consistent with a previous report44.The fact that the GAG content of AC-based constructs

was in the range of that measured in the inner region ofthe native meniscus would be important to guarantee initialresistance of the graft upon implantation42,52 and biologicalresponsiveness to compressive loads53. Moreover, theexpression and deposition by AC of all collagen typesidentified in the native meniscus indicate a compatibilityof their phenotype with meniscus cells. Last but not least,the relatively higher proliferation rate of AC, especially inthe presence of the growth factor combination TFP, wouldallow to generate a sufficient number of cells in a shortertime.

Although we cannot exclude that different experimentalconditions (e.g., use of different sets of growth factors)might result in a different trend, data generated usingthe described model systems indicate that AC representa more appropriate cell source than meniscus, FPC orSMC for the generation of a meniscus substitute. Recentstudies have already explored the possibility to heallesions in the avascular region of the meniscus usinglamb16 or swine AC17, with promising results. Clearly,tissues engineered by AC only remotely mimic the com-plex structure of native meniscus; moreover, in our studywe have not investigated the spatial organization ofaggrecan or of collagen fibers, which is known to behighly specific in meniscus cartilage54,55. In this regard,it should be pointed out that differential localization ofspecific matrix molecules in AC-based tissues may beinduced by appropriate physical conditioning in vitro,e.g., by application of hydrodynamic forces56. Moreover,it is likely that the composition and mechanical propertiesof the graft will correctly develop by adapting to localmechanical forces upon implantation, as it is describedduring the process of embryonic joint development57e59.In this context, studies are ongoing in a sheep modelto assess the efficacy of autologous AC loaded intoappropriately shaped scaffolds as engineered substitutesfor total meniscectomies.

55Osteoarthritis and Cartilage Vol. 15, No. 1

Fig. 5. Representative immunohistochemical stain for collagen types I (A, B), II (C, D), III (E, F), IV (G, H) and VI (I, J) of cartilaginous tissuesgenerated by culture of AC (A, C, E, G, I) and synovial membrane cells (B, D, F, H, J) into Hyaff�-11 for 4 weeks. Immunohistochemical stains

were similar in tissues generated by meniscus, FPC and SMC. Bar¼ 100 mm.

56 A. Marsano et al.: Comparison of chondrogenic cells

Fig. 6. Representative Safranin-O stained sections of tissues generated by culture of meniscus (A), FPC (B), SMC (C) and AC (D) intoHyaff�-11 for 2 weeks, followed by 6 weeks of ectopic implantation. Bar¼ 300 mm.

Acknowledgements

We wish to thank A. Barbero and F. Wolf for scientific andtechnical assistance, L. Fairbairn for the immunohistochem-ical stainings, L. Grady for mRNA extraction, and C. De Bariand F. Dell’Accio for helpful discussions. Hyaff�-11 waskindly provided by Fidia Advanced Biopolimer, AbanoTerme, Italy. This work was supported by the Swiss FederalOffice for Education and Science (B.B.W.) under the FifthEuropean Framework Growth (MENISCUS REGENERA-TION, contract G5RD-CT-2002-00703). This study is dedi-cated to the late Dirk Schaefer, who inspired and criticallyguided the work.

References

1. Ahmed AM. The load-bearing role of the knee menis-cus. In: Mow VC, Arnoczky SP, Jackson DW, Eds.Knee Meniscus: Basic and Clinical Foundations.New York: Raven Press 1992:59e73.

2. Levy IM, Torzilli PA, Fisch ID. The contribution of theknee menisci to the stability of the knee. In:Mow VC, Arnoczky SP, Jackson D, Eds. Knee Menis-cus: Basic and Clinical Foundations. New York: RavenPress 1992:107e115.

3. Sweigart MA, AufderHeide AC, Athanasiou KA. Fibro-chondrocytes and their use in tissue engineering ofthe meniscus. In: Ashammakhi N, Ferretti P, Eds.Topics in Tissue Engineering, University of Oulu2003;1:1e18.

4. Burr DB, Radin EL. Meniscal function and the impor-tance of meniscal regeneration in preventing latemedical compartment osteoarthrosis. Clin OrthopRelat Res 1982;171:121e6.

5. Cox JS, Nye CE, Schaefer WW, Woodstein IJ. Thedegenerative effects of partial and total resection of

the medial meniscus in dogs’ knees. Clin Orthop RelatRes 1975;109:178e83.

6. Higuchi H, Kimura M, Shirakura K, Terauchi M,Takagishi K. Factors affecting long-term results afterarthroscopic partial meniscectomy. Clin Orthop RelatRes 2000;377:161e8.

7. Rijk PC. Meniscal allograft transplantation e part I:background, results, graft selection and preservation,and surgical considerations. Arthroscopy 2004;20:728e43.

8. Rijk PC. Meniscal allograft transplantation e part II:alternative treatments, effects on articular cartilage,and future directions. Arthroscopy 2004;20:851e9.

9. Ibarra C, Koski JA, Warren RF. Tissue engineeringmeniscus: cells and matrix. Orthop Clin North Am2000;31:411e8.

10. Adams SBJ, Randolph MA, Gill TJ. Tissue engineeringfor meniscus repair. J Knee Surg 2005;18:25e30.

11. Ibarra C, Jannetta C, Vacanti CA, Cao Y, Kim TH,Upton J, et al. Tissue engineered meniscus: a potentialnew alternative to allogeneic meniscus transplanta-tion. Transplant Proc 1997;29:986e8.

12. Martinek V, Ueblacker P, Braun K, Nitschke S,Mannhardt R, Specht K, et al. Second generation ofmeniscus transplantation: in-vivo study with tissueengineered meniscus replacement. Arch OrthopTrauma Surg 2005;8:1e7.

13. Weinand C, Randolph MA, Peretti GM, Adams SB,Gill TJ. Cellular repair of meniscal tears in avascularregion. Trans Orthop Res Soc 2005 (Abstract).

14. Hidaka C, Ibarra C, Hannafin JA, Torzilli PA,Quitoriano M, Jen SS, et al. Formation of vascularizedmeniscal tissue by combining gene therapy with tissueengineering. Tissue Eng 2002;8:93e105.

15. Nakata K, Shino K, Hamada M, Mae T, Miyama T,Shinjo H, et al. Human meniscus cell: character-ization of the primary culture and use for tissueengineering. Clin Orthop Relat Res 2001;391:S208e18.

57Osteoarthritis and Cartilage Vol. 15, No. 1

16. Peretti GM, Caruso EM, Randolph MA, Zaleske DJ.Meniscal repair using engineered tissue. J OrthopRes 2001;19:278e85.

17. Peretti GM, Gill TJ, Xu JW, Randolph MA, Morse KR,Zaleske DJ. Cell-based therapy for meniscal repair:a large animal study. Am J Sports Med 2004;32:146e58.

18. Port J, Jackson DW, Lee TQ, Simon TM. Meniscalrepair supplemented with exogenous fibrin clot andautogenous cultured marrow cells in the goat model.Am J Sports Med 1996;24:547e55.

19. Walsh CJ, Goodman D, Caplan AI, Goldberg VM.Meniscus regeneration in a rabbit partial meniscec-tomy model. Tissue Eng 1999;5:327e37.

20. Zellner J, Kujat R, Nerlich M, Angele P. Cell-basedtissue engineering approach for repair of defects inthe avascular zone of meniscus. Trans Orthop ResSoc 2005 (Abstract).

21. Wickham MQ, Erickson GR, Gimble JM, Vail TP,Guilak F. Multipotent stromal cells derived from theinfrapatellar fat pad of the knee. Clin Orthop RelatRes 2003;412:196e212.

22. De Bari C, Dell’Accio F, Tylzanowski P, Luyten FP.Multipotent mesenchymal stem cells from adult humansynovial membrane. Arthritis Rheum 2001;44:1928e42.

23. Sakaguchi Y, Sekiya I, Yagishita K, Muneta T. Compar-ison of human stem cells derived from various mesen-chymal tissues: superiority of synovium as a cellsource. Arthritis Rheum 2005;52:2521e9.

24. Hunziker EB, Rosenberg LC. Repair of partial-thick-ness defects in articular cartilage: cell recruitmentfrom the synovial membrane. J Bone Joint Surg Am1996;78:721e33.

25. Miyamoto A, Deie M, Yamasaki T, Nakamae A,Shinomiya R, Ochi M. The synovium has important ef-fects to repair the cartilage defects. Trans Orthop ResSoc 2005 (Abstract).

26. Peterson L, Minas T, Brittberg M, Nilsson A, Sjogren-Jansson E, Lindahl A. Two- to 9-years outcome afterautologous chondrocyte transplantation of the knee.Clin Orthop Relat Res 2000;374:212e34.

27. Barbero A, Ploegert S, Heberer M, Martin I. Plasticity ofclonal populations of dedifferentiated adult human ar-ticular chondrocytes. Arthritis Rheum 2003;48:1315e25.

28. Dell’Accio F, De Bari C, Luyten FP. Microenvironmentand phenotypic stability specify tissue formation by hu-man articular cartilage-derived cells in vivo. Exp CellRes 2003;287:16e27.

29. Barbero A, Grogan S, Schafer D, Heberer M, Mainil-Varlet P, Martin I. Age related changes in humanarticular chondrocyte yield, proliferation and post-ex-pansion chondrogenic capacity. OsteoarthritisCartilage 2004;12:476e84.

30. Tay AG, Farhadi J, Suetterlin R, Pierer G, Heberer M,Martin I. Cell yield, proliferation, and postexpansiondifferentiation capacity of human ear, nasal, and ribchondrocytes. Tissue Eng 2004;10:762e70.

31. Martin I, Shastri VP, Padera RF, Yang J, Mackay AJ,Langer R, et al. Selective differentiation of mammalianbone marrow stromal cells cultured on three-dimen-sional polymer foams. J Biomed Mater Res 2001;55(2):229e35.

32. Pangborn CA, Athanasiou KA. Growth factors andfibrochondrocytes in scaffolds. J Orthop Res 2005;23:1184e90.

33. Jakob M, Demarteau O, Schafer D, Stumm M,Heberer M, Martin I. Enzymatic digestion of adulthuman articular cartilage yields a small fraction ofthe total available cells. Connect Tissue Res 2003;44:173e80.

34. Gronthos S, Franklin DM, Leddy HA, Robey PG,Storms RW, Gimble JM. Surface protein characteriza-tion of human adipose tissue-derived stromal cells.J Cell Physiol 2001;189:54e63.

35. Yoo JU, Barthel TS, Nishimura K, Solchaga L,Caplan AI, Goldberg VM, et al. The chondrogenicpotential of human bone-marrow-derived mesenchy-mal progenitor cells. J Bone Joint Surg Am 1998;80:1745e57.

36. Grigolo B, Lisignoli G, Piacentini A, Fiorini M, Gobbi P,Mazzotti G, et al. Evidence for redifferentiation ofhuman chondrocytes grown on a hyaluronan-basedbiomaterial (HYAff 11): molecular, immunohistochem-ical and ultrastructural analysis. Biomaterials 2002;23:1187e95.

37. Dickinson SC, Sims TJ, Pittarello L, Soranzo C,Pavesio A, Hollander AP. Quantitative outcome mea-sures of cartilage repair in patients treated by tissueengineering. Tissue Eng 2005;11:277e87.

38. Al-Taher A, Bashein A, Nolan T, Hollingsworth M,Brady G. Global cDNA amplification combined withreal-time RT-PCR: accurate quantification of multiplehuman potassium channel genes at the single celllevel. Yeast 2000;17:201e10.

39. Livak KJ, Schmittgen TD. Analysis of relative gene ex-pression data using real-time quantitative PCR andthe 2(�Delta Delta C(T)) method. Methods 2001;25:402e8.

40. Hollander AP, Heathfield TF, Webber C, Iwata Y,Bourne R, Rorabeck C, et al. Increased damage totype II collagen in osteoarthritic articular cartilagedetected by a new immunoassay. J Clin Invest 1994;93:1722e32.

41. Melrose J, Smith S, Cake M, Read R, Whitelock J.Comparative spatial and temporal localisation of perle-can, aggrecan and type I, II and IV collagen in theovine meniscus: an ageing study. Histochem CellBiol 2005 Sep;124(3e4):225e35. Epub 2005 Oct 28.

42. AufderHeide AC, Athanasiou KA. Mechanical stimula-tion toward tissue engineering of the knee meniscus.Ann Biomed Eng 2004;32:1161e74.

43. Reinholz GG, Lu L, Saris DB, Yaszemski MJ,O’Driscoll SW. Animal models for cartilage reconstruc-tion. Biomaterials 2004;25:1511e21.

44. Moretti M, Wendt D, Dickinson SC, Sims TJ,Hollander AP, Kelly DJ, et al. Effects of in vitro precul-ture on in vivo development of human engineeredcartilage in an ectopic model. Tissue Eng 2005;11:1421e8.

45. Verdonk PC, Forsyth RG, Wang J, Almqvist KF,Verdonk R, Veys EM, et al. Characterisation of humanknee meniscus cell phenotype. Osteoarthritis Carti-lage 2005;13:548e60.

46. De Bari C, Dell’Accio F, Luyten FP. Failure of in vitro-differentiated mesenchymal stem cells from thesynovial membrane to form ectopic stable cartilagein vivo. Arthritis Rheum 2004;50:142e50.

47. Park Y, Sugimoto M, Watrin A, Chiquet M, Hunziker EB.BMP-2 induces the expression of chondrocyte-specificgenes in bovine synovium-derived progenitor cellscultured in three-dimensional alginate hydrogel. Oste-oarthritis Cartilage 2005;13:527e36.

58 A. Marsano et al.: Comparison of chondrogenic cells

48. Mochizuki T, Muneta T, Sakaguchi Y, Nimura A,Yokoyama A, Koga H, et al. Higher chondrogenicpotential of fibrous synovium- and adiposesynovium-derived cells compared with subcutaneousfat-derived cells: distinguishing properties of mesen-chymal stem cells in humans. Arthritis Rheum 2006;54:843e53.

49. Jakob M, Demarteau O, Schafer D, Hintermann B,Dick W, Heberer M, et al. Specific growth factors dur-ing the expansion and redifferentiation of adult humanarticular chondrocytes enhance chondrogenesis andcartilaginous tissue formation in vitro. J Cell Biochem2001;81:368e77.

50. Roberts S, Hollander AP, Caterson B, Menage J,Richardson JB. Matrix turnover in human cartilagerepair tissue in autologous chondrocyte implantation.Arthritis Rheum 2001;44:2586e98.

51. Yaeger PC, Masi TL, de Ortiz JL, Binette F, Tubo R,McPherson JM. Synergistic action of transforminggrowth factor-beta and insulin-like growth factor-Iinduces expression of type II collagen and aggrecangenes in adult human articular chondrocytes. ExpCell Res 1997;237:318e25.

52. Fithian DC, Kelly MA, Mow VC. Material properties andstructureefunction relationships in the menisci. ClinOrthop Relat Res 1990;19e31.

53. Demarteau O, Wendt D, Braccini A, Jakob M, Schafer D,Heberer M, et al. Dynamic compression of cartilage

constructs engineered from expanded human articularchondrocytes. Biochem Biophys Res Commun 2003;310:580e8.

54. Kambic HE, McDevitt CA. Spatial organization of typesI and II collagen in the canine meniscus. J Orthop Res2005;23:142e9.

55. Valiyaveettil M, Mort JS, McDevitt CA. The concentra-tion, gene expression, and spatial distribution ofaggrecan in canine articular cartilage, meniscus, andanterior and posterior cruciate ligaments: a newmolecular distinction between hyaline cartilage andfibrocartilage in the knee joint. Connect Tissue Res2005;46:83e91.

56. Marsano A, Wendt D, Quinn TM, Sims TJ, Jakob M,Heberer M, et al. Bi-zonal cartilaginous tissues engi-neered in a rotary cell culture system. Biorheology,in press.

57. Chen AC, Wong VW, Sah RL. Maturation dependentbiomechanical properties of bovine meniscus. TransOrthop Res Soc 2002 (Abstract).

58. Mikic B, Johnson TL, Chhabra AB, Schalet BJ, Wong M,Hunziker EB. Differential effects of embryonic immobili-zation on the development of fibrocartilaginous skeletalelements. J Rehabil Res Dev 2000;37:127e33.

59. Ochi M, Kanda T, Sumen Y, Ikuta Y. Changes in thepermeability and histologic findings of rabbit menisciafter immobilization. Clin Orthop Relat Res 1997;334:305e15.