R

Hm

AKa

b

c

d

e

a

ARRA

KJAACMCR

1

Mbaoab

C1

0h

Annals of Anatomy 195 (2013) 488– 497

Contents lists available at ScienceDirect

Annals of Anatomy

journa l ho mepage: www.elsev ier .de /aanat

esearch article

eterotopic and orthotopic autologous chondrocyte implantation using ainipig chondral defect model

nke Lohana, Ulrike Marzahna, Karym El Sayeda, Christopher Bockb, Andreas Haischc, Benjamin Kohla,atharina Stoelzeld, Thilo Johna,e, Wolfgang Ertela, Gundula Schulze-Tanzil a,∗

Department for Orthopaedic, Trauma and Reconstructive Surgery, Charité-University of Medicine, Campus Benjamin Franklin, Berlin, GermanyDepartment of Nephrology and Internistic Intensive Care, Charité-University of Medicine, Campus Benjamin Franklin, Berlin, GermanyDepartment of Otorhinolaryngology, Head and Neck Surgery, Charité-University of Medicine, Campus Benjamin Franklin, Berlin, GermanyDepartment of Otorhinolaryngology, Head and Neck Surgery, Charité-University of Medicine, Campus Mitte, Berlin, GermanyDepartment for Orthopaedics and Trauma Surgery, DRK Kliniken Berlin Westend, Berlin, Germany

r t i c l e i n f o

rticle history:eceived 30 December 2012eceived in revised form 24 February 2013ccepted 2 April 2013

eywords:oint cartilagerticularuricularhondrocytesinipig

hondral defectepair

s u m m a r y

Implantation of non-articular (heterotopic) chondrocyte-based implants might be an alternativeapproach to articular cartilage repair. This strategy could be helpful in cases in which there are no ortoo few articular chondrocytes available. Therefore, this study was undertaken to compare joint cartilagedefect healing in the minipig model after implantation of heterotopic auricular and orthotopic articularchondrocytes.

Poly-glycolic acid (PGA) associated three-dimensional (3D) constructs were prepared culturing autol-ogous minipig-derived articular and auricular chondrocytes for 7 days in a dynamic culture system.Chondrocyte PGA constructs were implanted into 8 mm diameter and ∼1.1 mm deep chondral defectswithin the medial and lateral condyles of the minipig knee joints. Empty defects served as controls forassessment of the intrinsic healing response. Defect healing was monitored 6 months post implantationusing a macroscopic and microscopic score system and biomechanical analysis.

Neo-cartilage formation could be observed in the PGA constructs seeded with articular and auricularchondrocytes in vivo. The defect healing did not significantly differ at the macroscopic and histolog-ical level in response to implantation of either autologous articular or auricular chondrocytes seeded

constructs compared with the empty defects. Although the differences were not significant, the auricu-lar chondrocytes-based implants led to a slightly inferior repair quality at the macroscopic level, buta histologically superior healing response when compared with the empty defect group. However,biomechanical analysis revealed a higher stiffness in repair tissues produced by auricular chondro-cyte implantation compared with the other groups. Deduced from these results, articular chondrocytesrepresent the preferable cell source for implantation.

. Introduction

Injured cartilage possesses poor intrinsic healing capacity.atrix-assisted autologous chondrocytes implantation (MACI) has

een developed as a strategy for coverage of cartilage defects andchieving a hyaline-like repair tissue. For this technique, autol-

gous chondrocytes were isolated from a small cartilage biopsynd expanded in vitro. Then they were seeded on a supportingiodegradable scaffold and implanted into the cartilage defect.

∗ Corresponding author at: Department for Trauma and Reconstructive Surgery,harité-University of Medicine, Campus Benjamin Franklin, FEM, Garystrasse 5,4195 Berlin, Germany. Tel.: +49 30 450 552385; fax: +49 30 450 552985.

E-mail address: gundula.schulze-tanzil@charite.de (G. Schulze-Tanzil).

940-9602/$ – see front matter © 2013 Elsevier GmbH. All rights reserved.ttp://dx.doi.org/10.1016/j.aanat.2013.04.009

© 2013 Elsevier GmbH. All rights reserved.

However, the availability of autologous articular chondrocytes isvery limited restricting the application of this technique. Het-erotopic non-articular cartilage could serve as an alternative cellsource to cover joint cartilage defects as has been previously dis-cussed (El Sayed et al., 2010). Heterotopic cartilage is easier toharvest with lower biopsy morbidity and reveals a higher prolif-eration rate of chondrocytes. Heterotopic chondrocytes provide ahigh chondrogenic potential as has been revealed in many studies(El Sayed et al., 2010, 2013). They maintain their proliferative capac-ity for a longer life span and represent a healthy cartilage source,which is interesting in the light of the development of osteoarthri-

tis (OA) in response to articular cartilage injury. They might beable to transfer to the articular chondrocyte gene expression pro-file since non-articular nasoseptal chondrocytes which have beenexposed to mechanostimulation have been shown to express an

Anato

icTae2i(dapt(wnr2a2imcG

cmrtt

2

2

s(tlgTwq2FAwG3(BGpA

2

hpo(sdD

A. Lohan et al. / Annals of

ncreased level of lubricin (Candrian et al., 2008). Lubricin is a gly-oprotein found only in articular cartilage (Greene et al., 2011).he compatibility of non-articular heterotopic chondrocytes withrticular chondrocytes has been demonstrated in vitro using sev-ral co-culturing approaches (El Sayed et al., 2013; Kuhne et al.,010). Previously, nasoseptal chondrocytes have been successfully

mplanted into osteochondral joint cartilage defects in the rabbitVinatier et al., 2007, 2009). Hence, we analyzed articular cartilageefect healing after implantation of autologous heterotopic (non-rticular) chondrocytes which derived from elastic cartilage of theorcine auricle in the large animal minipig model and comparedhe outcome with that after implantation of autologous orthotopicarticular) chondrocytes. As a scaffold, poly-glycolic-acid (PGA)as chosen. PGA is well characterized for cartilage tissue engi-eering and has already been tested in several articular cartilageepair studies in the large animal model (Erggelet et al., 2007,009; Liu et al., 2002; Zhou et al., 2006). The minipig is a valu-ble model in cartilage repair studies (Ahern et al., 2009; Chu et al.,010; Gotterbarm et al., 2008) and suitable for conducting stud-

es using chondral defects to exclude the influence of ingrowingesenchymal stem cells deriving from the opened bone marrow

avities during osteochondral defect healing (Ahern et al., 2009;otterbarm et al., 2008).

However, it remains completely unclear whether auricularhondrocytes integrate into the defect and thus leads to develop-ent of a competent hyaline like and biomechanically competent

epair tissue when implanted in vivo. Likewise, it is unclear whetherhey might stimulate resident chondrocytes in a paracrine mannero re-enter the cell-cycle and to initiate proliferation.

. Materials and methods

.1. Isolation and culturing of minipig chondrocytes

Autologous cartilage was harvested from the auricle (two biop-ies each sized 6 mm in diameter were taken) and the knee jointcartilage was taken from the defect area, 8 mm in diameter, fromhe medial and lateral condyle of the ipsilateral leg to be implantedater into the cartilage defect of the contralateral leg) of Göttin-en minipigs (Ellegaard, Dalmose, Denmark) as described below.he connective tissue and perichondrium of auricular cartilageere carefully removed. Porcine cartilage samples were subse-

uently minced into 1 mm slices and enzymatically digested with% pronase (7 U/mg, Roche, Basel, Switzerland) diluted in Ham’s-12/Dulbecco’s modified Eagle’s (DMEM) medium 1:1 (BiochromG, Berlin, Germany) for 20 min at 37 ◦C and subsequently digestedith 0.1% (w/v) collagenase (≥0.1 U/mg, SERVA ElectrophoresismbH, Heidelberg, Germany) diluted in growth medium for 16 h at7 ◦C. Isolated chondrocytes were resuspended in growth mediumHam’s F-12/DMEM 1:1 containing 10% fetal calf serum [FCS,iochrom AG], 25 �g/mL ascorbic acid [Sigma–Aldrich, Munich,ermany], 50 IU/mL streptomycin, 50 IU/mL penicillin, 0.5 �g/mLartricin, essential amino acids, 2 mM l-glutamine [all: BiochromG]) and seeded at 2.8 × 104 cells/cm2 in culture flasks.

.2. Dynamic 3D PGA cultures

Biodegradable non-woven PGA meshes (36 mg/cm3, 1.1 mmeight, Alpha Research, Berlin, Germany) were sectioned intoieces of 1 cm diameter. One piece was soaked with 5 mLf a suspension of minipig chondrocytes in growth medium

36,000 chondrocytes calculated per mm3 PGA-scaffold wereuspended in 5 mL growth medium). Articular and auricular chon-rocytes of the second or third monolayer passage were used.ynamic cultures were performed for 1 week by stirring the cell

my 195 (2013) 488– 497 489

suspension in a bioreactor filter tube (TPP, Switzerland) underorbital shaking conditions (aptitude 10◦) with the floating PGA feltat 12 rpm in a rotatory shaker (digital tube roller, Stuart SRT9D,Bibby Scientific, USA) at 37 ◦C and 5% CO2. Growth medium waschanged every 48–72 h.

2.3. Assessment of cell vitality in PGA cultures

After washing with PBS, the 7-day-old scaffolds were incubatedin fluorescein diacetate (FDA, 3 �g/mL dissolved in acetone [stocksolution], Sigma–Aldrich and further diluted 1:1000 in PBS [work-ing solution]) for 15 min at 37 ◦C. Scaffolds were rinsed three timeswith PBS before being counterstained with propidium iodide solu-tion (PI, 1 mg/mL dissolved in PBS [stock solution], Sigma–Aldrichfurther diluted 1:100 in PBS [working solution]) for 1 min in thedark at room temperature (RT). The green or red fluorescence wasvisualized using fluorescence microscopy (Axioskop 40, Carl Zeiss,Jena) and an Olympus XC30 camera (Olympus Soft Imaging Solu-tions GmbH, Muenster, Germany).

2.4. Minipig chondral defect model: cartilage biopsies and PGAimplantation

The study was approved by the National Animal ProtectionCommission (LAGeSo, Berlin, Germany). Female, 1–3.5-year-oldGöttingen minipigs (Ellegaard, n = 4, body weights: between 39.5and 52.0 kg, group design: Table 1) were housed alone or in pairs.Animals were allowed unrestricted movement; they received somehay and standard food twice daily and water ad libitum.

Autologous cartilage was harvested from the auricle and theknee and all animals underwent parapatellar arthrotomy surgerieson both hind limbs under strictly aseptic conditions to implant thePGA scaffolds into the chondral defect in the medial und lateralcondyles of the right and left knee joints after a timespan of 4 weeks.

Auricular cartilage biopsies were taken under full inhalationnarcosis. After premedication (3 mg/kg azaperone [Janssen AnimalHealth, Neuss, Germany], 25 mg/kg ketamine [Serumwerk, Bern-berg; Germany], 3.5 mg/kg xylazin [Bayer Vital GmbH, Leverkusen;Germany] and 0.03 mg/kg atropine sulfate [Braun Melsungen AG,Melsungen, Germany], all intramuscular) and induction of anesthe-sia (10 mg/kg propofol [Fresenius Kabi, Bad Homburg, Germany],i.v.) animals were intubated and received isoflurane (Forene®,0.8–1.2%, Abbot GmbH & Co. KG, Wiesbaden, Germany) under con-trolled inspiration. Two 6 mm diameter cartilage/perichondriumtissue cylinders were punched out.

For articular cartilage biopsies and PGA implantations intothe chondral defects, the animals were anesthetized as describedabove. A parapatellar skin incision of ∼5 cm was made. The fas-cia and the patellar ligament were split longitudinally within themedial third. The patella was luxated in a lateral direction. A punchwas used to create a reproducible 8 mm diameter and 1 mm deepfull thickness chondral defect in the medial and lateral condyles.The cartilage was harvested for chondrocyte isolation and chon-drocytes were seeded on PGA scaffolds before implantation intothe cartilage defects of the contralateral leg. Eight mm diameterpieces of the 1 cm diameter chondrocytes-seeded PGA constructswere punched out and implanted into the chondral defects. Fibringlue (Tissue Duo, Baxter, Deerfield, USA) was used to fix the scaf-folds and was allowed to solidify for 10 min. Additionally, a cartilagesuture (Vicryl 9.0, Ethicon, San Lorenzo, USA) was performed to fixthe scaffold. All animals were allowed full weight-bearing imme-diately after surgery. The minipigs received peri- and post-surgery

antibiosis and analgesia according to the clinical symptoms.

Four weeks after the articular biopsy, PGA implantation wasperformed into the contralateral hind limb and therefore, the ani-mals were anesthetized, operated and treated as described above.

490 A. Lohan et al. / Annals of Anatomy 195 (2013) 488– 497

Table 1Group design.

Empty defect Articular Auricular

Minipig 1a 2 defects (right knee)Minipig 2a 2 defects (right knee)Minipig 3 2 defects (left knee) 2 defects (right knee)Minipig 4 2 defects (left knee) 2 defects (right knee)

8reated

Cs

TndtaSDbiCwfiii

2

c

TS

mm diameter defects were created within the medial and lateral femur condyles.a Minipigs 1 and 2 received also two defects in the other (left) knee which were t

linical healing and gait alteration was monitored using a scoreheet (Table 2).

Animals were sacrificed 6 months (n = 4 defects in each group,able 1) after the last implantation of the construct using deeparcosis and 60 mL 1 M KCl (Fresenius Kabi) administration. Theefect site was macroscopically scored (Table 3) and then pho-ographically documented using a Canon EOS500D camera with

Canon EF-S60 mm macro objective (Canon, Krefeld, Germany).hore stiffness was evaluated using a durometer (type M, ASTM2240, Shore Instruments-Instron, Norwood, MA) as describedelow. Subsequently, explanted osteochondral blocks were fixed

n 4% paraformaldehyde (PFA) for 5 days and decalcified in 2 mMa2++−EDTA solution (Merck, Darmstadt, Germany) for 4 weekshich was changed daily before embedding in paraffin. Paraf-n sections (thickness: 7 �m) were prepared, histologically and

mmunohistologically stained and analyzed using a histological andmmunohistological scoring system (Tables 4 and 5).

.5. Macroscopic score to assess neo-cartilage quality

A macroscopic scoring system was developed to assess theartilage quality in vivo in the minipig (Table 3). It entails the

able 2core sheet clinical healing and gait alteration.

GaitPhysiological 5Unphysiological 4Lameness grade I (lameness after standing up) 3Lameness grade II (abbreviated step of one limb) 2Lameness grade III (abbreviated step of more limbs) 1Lameness grade III (one limb unloaded) 0

StandingAll limbs equally loaded 4One limb temporary under the body 3One limb always under the body 2One limb always unloaded 1Two limbs unloaded 0

LyingPhysiological 1Temporary induced 0Always induced 0

Standing upPhysiological 3Temporary delayed 2Always delayed 1Always lying 0

Signs of inflammation: right knee joint Yes NoRedness 0 1Swelling 0 1Induced articular filling 0 1Fluctuation 0 1Warmness 0 1

Signs of inflammation: left knee joint Yes NoRedness 0 1Swelling 0 1Induced articular filling 0 1Fluctuation 0 1Warmness 0 1

with another biomaterial (confidential data not shown).

examination of color, the depth of coverage, rough or smooth sur-face, integration, signs of inflammation after 6 months in vivo.

2.6. Histological analysis (hematoxylin eosin and alcian bluestaining)

For all histological staining procedures, paraffin sections (thick-ness: 7 �m) were prepared. Slices were deparaffinized in xylene(Carl Roth GmbH, Karlsruhe, Germany) and incubated in adescending alcohol series. For hematoxylin eosin (HE) staining,sections were stained for 4 min in Harris’ hematoxylin solution(Sigma–Aldrich), rinsed in water and counterstained for 4 min ineosin (Carl Roth).

For alcian blue (AB) staining, the sections were incubated for3 min in 1% acetic acid and then stained for 30 min in 1% AB (in 3%acetic acid) at pH 2.5 (Carl Roth). Subsequently, they were rinsed in3% acetic acid. Counterstaining of cell nuclei was performed usingnuclear fast red aluminum sulfate solution (Carl Roth) for 5 min.

Finally, all sections were rinsed with aqua dest. and subse-quently dehydrated in an ascending alcohol series. Then, thesections were embedded in Entellan (Merck, Darmstadt, Germany).All slices were analyzed using light microscopy (Axioskop 40 micro-scope: Zeiss Jena, Germany). Photos of the sections were takenusing an Olympus camera XC30. For HE and AB stained sections,

a histological scoring system was established to assess the qualityof neo-cartilage formation within the defect (Table 4).

Table 3Macroscopical score.

Color of defect areaBright white 2Dull white, translucent 1Yellow, rose, brown, red 0

Level of the defectAt the niveau of surrounding healthy tissue 3Slight above the level 2Prominent above the level 1No defect filling 0

Defect surfaceIntact, smooth 3Smooth and harsh area 2Uneven, harsh 1No defect filling 0

Defect marginDetectable 2Hardly detectable 1Not detectable 0

Integration to the surrounding tissueConnected 2Slightly connected 1Not connected 0

InflammationNot existing 2Moderate (swelling, redness) 1Severe (osteophytes) 0

A. Lohan et al. / Annals of Anato

Table 4Histological score based on O’Driscoll (HE and AB stainings).

Quality of the predominant tissueCellular morphology

Hyaline articular cartilage 4Hyalin-like or heterogen 3Incomplete differentiated mesenchyme 2Fibrous tissue or bone 0

Presence of Alcian blue positive extracellular matrixNormal or nearly normal 3Moderate 2Slight 1None 0

Structural characteristicsSurface regularity

Smooth and intact 3Superficial horizontal lamination 2Fissures – 25–100% of the thickness 1Severe disruption, including fibrillation 0

Structural integrityNormal 2Slight disruption, including cysts 1Severe disintegration 0

Thickness of the repair tissue100% of normal adjacent cartilage 250–100% of normal cartilage 10–50% of normal cartilage 0

Bonding to the adjacent cartilageBonded at both ends of graft 2Bonded at one end, or partially at both ends 1Not bonded 0

Absence of cellular changes or degenerationHypocellularity

Normal cellularity 2Moderate hypocellularity 1Severe hypocellularity 0

Chondrocytes and matrix stainingNormal cell morphology, normal staining 3Cell cluster, normal staining 2Cell degeneration, reduced staining 1Severe cell degeneration, poor or no staining 0

Absence of degenerative changes in adjacent cartilageNormal cellularity, no clusters, normal staining 3Normal cellularity, mild clusters, moderate staining 2Mild to moderate hypocellularity, slight staining 1Severe hypocellularity, poor or no staining 0

Table 5Immunohistological score.

Collagen type IICollagen type II content

Normal, physiological 3Slightly reduced 2Moderately reduced 1Not detectable 0

Collagen type II distributionHomogeneous, physiological 1.5Heterogeneous 1Predominantly heterogeneous 0.5

Zonation of the surface compared to profound defectIncreased staining 1.5Equal staining 1Decreased staining 0.5

Quality of the ECM>90% smooth, homogeneous 350–90% smooth, homogeneous 250–90% fibrous 1>90% fibrous 0

Collagen type ICollagen type I content

Normal, physiological (not detectable) 2Slightly increased 1Severely increased 0

Collagen type I distributionHomogeneous, physiological 1.5Heterogeneous 1Predominantly heterogeneous 0.5

my 195 (2013) 488– 497 491

2.7. Type II, type I and type X collagen immunolabeling

Paraffin sections of the tissue engineered constructs were rinsedin Tris buffered saline (TBS: 0.05 M Tris, 0.015 M NaCl, pH 7.6)before incubation with 5 mg/mL pronase (7 U/mg, Roche, Basel,Switzerland) diluted in TBS for 5 min at 37 ◦C. Sections were sub-sequently rinsed with TBS, blocked with protease-free donkeyserum (5% diluted in TBS) for 30 min at RT, rinsed and incubatedwith polyclonal rabbit anti-type II or -type I collagen antibody(27.5 �g/mL, Acris Antibodies, cat. no. R1039X, R1038, Herford,Germany) and monoclonal mouse anti-type X collagen (Abcam,cat. no. ab499-45200) in a humidified chamber overnight at 4 ◦C.Sections were subsequently washed with TBS before incubationwith donkey-anti-rabbit-Alexa-Fluor®488 (10 mg/mL, Invitrogen,cat. no. A2106, Darmstadt, Germany) or donkey-anti-mouse-Alexa-Fluor®488 (Invitrogen, cat. no. A212002) secondary antibody for30 min at RT. Cell nuclei were counterstained using 4′,6-diamidino-2-phenylindole (DAPI) (0.1 �g/mL, Roche). Labeled sections wererinsed several times with TBS, embedded in Fluoromount G (South-ern Biotech, Biozol Diagnostica, Birmingham, USA) and examinedusing fluorescence microscopy (Axioskop 40, Carl Zeiss). As iso-type control rabbit IgG (Invitrogen, cat. no. O2-6102) or mouseIgG (Invitrogen, cat. no. MG100) were used. A decalcified femoralbone-articular cartilage tissue section derived from a minipig kneewithout defect served as a positive control. A scoring system wasused to compare the results (Table 5). Images were taken using aXC30 camera (Olympus, Europa Holding, Hamburg, Germany).

2.8. Biomechanical testing

Shore stiffness was evaluated using a durometer (type M,ASTM D2240, Shore Instruments-Instron, Norwood, MA) (Simonand Aberman, 2010). Three defined points of measurement weremarked within the defect area. In each case, two additional mea-surement points were indicated above this area and two belowthis area. All points were marked using a positioning device. At thesame, similar measurement points were marked in a control area(healthy cartilage) of the condyle. At each given point (defect areaand control area) 16 measurements were performed.

2.9. Statistical analysis

All values were expressed as mean and all values are shown.An outlier was detected using GRUBB’s test and data was analyzedusing unpaired Mann–Whitney-test (IBM SPSS Statistics 19). Forbiomechanical analysis a one-way ANOVA was used to determinesignificant differences between the groups. Statistical significancewas set at a p value of ≤0.05.

3. Results

3.1. Results of dynamic 3D PGA in vitro cultures

Autologous porcine articular as well as auricular chondrocytesadhered to the fibers of the scaffold after 1 week of culturing.Cell-matrix sails were observed spanning between the PGA fibers(Fig. 1a1 and b1). They survived in the scaffolds for the 1 week ofculture before implantation and in other experiments for 3 weeksof culturing (own results). Mostly viable chondrocytes could be

detected adhering on the PGA fibers whereby only a few dead cellscould be demonstrated (Fig. 1a2 and b2). Histological staining (HE:Fig. 1a3 and b3) revealed the formation of an ECM which containedsulfated proteoglycans (AB: Fig. 1a4 and b4).

492 A. Lohan et al. / Annals of Anatomy 195 (2013) 488– 497

Fig. 1. (a and b) In vitro cultured PGA scaffolds before implantation. Light microscopy and vitality analyses of PGA scaffold seeded with articular (a1 and a2) and auricular (b1

a of PGAa

3

rttpTpb

nd b2) chondrocytes after 1 week in vitro. Scale bars 200 �m. HE and AB staining

fter 1 week in vitro. Arrows: fibers. Scale bars 25 �m.

.2. Clinical results

Wound healing proceeded without any complications. Loadelief of the operated legs of the pigs was only transient. Sincehe minipigs are rather inactive, each animal was mobilized threeimes a week for around 15 min. This procedure was started 4 weeks

ost surgery, when none of the animals need analgesia any longer.he gait, clinical healing response and the presence of any signs ofainful knees, indicative of OA, were controlled weekly post surgeryy veterinarians using a score sheet (Table 2). It revealed no major

scaffold seeded with articular (a3 and a4) and auricular (b3 and b4) chondrocytes

differences when compared with the state before surgery, exceptfor the two animals which received the empty defects. These pigsdeveloped slight alterations of the gait 4 month post surgery (datanot shown).

3.3. Results of macroscopic and histological scoring:

neo-cartilage quality

Features of joint inflammation such as an increase in synovialfluid content, color or amount could not be detected. The cartilage

A. Lohan et al. / Annals of Anato

Fig. 2. (a and b) Macroscopic results of defect healing 6 months post constructimplantation. Results of the macroscopic scoring are summarized (a). Healthy carti-lage reached a maximal macroscopic score of 14 points. Representative photographsof the defect repair in the empty defect (b1) and in response to implantation of PGAscaffolds seeded with articular (b2) and auricular (b3) chondrocytes into chondraljoint defects 6 month after implantation.

a(adftrdseiiihacfiwaatraTfiopago(c

minipig study we found that the cartilage thickness in the lateralfemoral condyles was generally higher than in the medial condyle

round the defects of both condyles still had a typical white colorFig. 2a and b). In some cases a faint injection of vessels at the medialnd lateral margin of the trochlea groove could be detected. Theefect healing was analyzed using HE and AB stainings to assess sul-ated cartilage proteoglycan content. It did not significantly differ athe macroscopic (Fig. 2a and b) and histological level (Fig. 3a–c) inesponse to implantation of autologous articular or auricular chon-rocytes seeded constructs compared with the empty defects. Thecoring results did not reveal significant differences between thexperimental groups either at the macroscopic or at the histolog-cal level. The auricular chondrocyte-based implants were slightlynferior at the macroscopic level, but slightly better at the histolog-cal level compared with the empty defect group. In comparison toealthy cartilage, a reduced overall height of the repair tissue andn impaired or inhomogeneous distribution of sulfated proteogly-ans could be detected using AB staining. More pronounced surfacebrillation could be detected in the empty defect group comparedith the other groups. Nevertheless, defects which received either

rticular chondrocyte seeded scaffolds as well as defects filled withuricular chondrocyte loaded scaffolds revealed a tidemark forma-ion. The tidemark was often doubled (not shown). Areas containingather fibrous repair tissue could be detected in the empty defectss well as in the defects which received auricular chondrocytes.he surface of the repair tissue was smoother in most of the defectslled with articular chondrocyte seeded scaffold compared with thether groups. No features of inflammation could be detected inde-endently of the implanted cells. Healing response was detectablelso by a proliferation of chondrocytes within isogenic chondrocyteroups. Particularly in the bounding area and the superficial zonef the repair tissue, larger clusters of chondrocytes became evident

Fig. 3b and c, arrows). An enlargement of bone marrow cavitiesould be detected in the empty defects (not shown).

my 195 (2013) 488– 497 493

3.4. Immunohistological results of defect repair

Type II and type I collagen expression and distribution was eval-uated by immunolabeling using a special scoring system (Table 5and Fig. 4a). Type II is the main cartilage ECM component. Type I col-lagen is indicative of inferior repair tissue or bone formation. TypeII collagen expression was slightly less in the empty defect com-pared with the other groups (Fig. 4b). Some islands with enhancedtype I collagen expression could be detected in the empty defectgroup (Fig. 4c). Especially chondrocytes located close to the surfaceof the cartilage defect expressed type X collagen in all experimen-tal groups (Fig. 4d). Only a few type X collagen positive cells couldbe detected near the border between the intact and repaired areas(Fig. 4e).

3.5. Biomechanical analysis

The auricular chondrocytes and the empty defect showed aslightly higher shore degree compared to native cartilage andthe articular chondrocytes, suggesting a more bone like character(Fig. 5). In comparison to each control area and, hence, healthy car-tilage the defect repair areas revealed a higher degree of stiffness.

4. Discussion

The use of heterotopic non-articular cartilage might providea valuable approach for the recruitment of additional autologouschondrocyte sources for cartilage repair. Recently, Vinatier et al.reported that nasoseptal chondrocytes implanted into osteochon-dral defects of knee joints in the rabbit model supported articularcartilage repair (Vinatier et al., 2007, 2009). A former study of ourgroup performed in the nude mice xenograft model suggested thatporcine auricular chondrocytes cultured on PGA had a high chon-drogenic potential. The quality of the neocartilage produced in thenude mice by implanted auricular chondrocyte PGA constructs wasmore consistent after 6 weeks in vivo compared with nasoseptaland articular chondrocytes (Lohan et al., 2011). Subsequently, weanalyzed osteochondral defect healing in response to implanta-tion of auricular chondrocyte seeded PGA constructs in the rabbitmodel [own unpublished results] whereby in this model an inferiordefect healing became evident with auricular chondrocytes com-pared with articular chondrocyte implantation. Now, we wanted toconfirm and extend these previous results with auricular chondro-cyte implantation using a large animal chondral defect model. Inthis model, bone marrow derived stem cells do not immigrate intothe defect. The swine was chosen as a model since porcine cartilageand chondrocytes share some similarities with human derived car-tilage (Kaab et al., 1998; Schulze-Tanzil et al., 2009). Furthermore,the thickness of the knee joint cartilage is higher in the pig com-pared with other large animals such as the sheep and goat (Ahernet al., 2009). For this reason, the Göttingen minipig was selectedas an adequate large animal model. Since the cartilage layer in theporcine knee joint is thick enough (1–2 mm) (Ahern et al., 2009) achondral defect minipig model was used for this study. Most (95%)of the human cartilage defects in the knee are chondral (Ahern et al.,2009). Traumatic cartilage defects can more often be found at thefemoral condyles in humans; hence, the femoral condyles of theminipigs as a high weight bearing defect localization were pre-ferred compared to the trochlea groove. In each group, we used themedial and lateral femoral condyles for defect creation. In another

[own non-published results]. These differences might also influ-ence the results.

494 A. Lohan et al. / Annals of Anatomy 195 (2013) 488– 497

Fig. 3. (a–c) Microscopic results of chondral defect repair 6 months post construct implantation. The diagram depicts the results of the histological scoring (a). Healthycartilage reached a maximal microscopic score of 24 points. Representative pictures of the HE (b) and AB staining (c) of the empty defect (b2,6 and c2,6), articular (b3,7 and c3,7)a ig joini s 500 �A

cttdtbucmdcdm

nd auricular (b4,8 and c4,8), chondrocyte seeded PGA scaffolds implanted into minipntact and regenerated areas are shown at low and higher magnifications: scale barrrow head: tidemark.

In contrast to osteochondral defects, where mesenchymal stemells (MSC) can immigrate from the opened bone marrow cavi-ies and contribute to the repair outcome, in a chondral model,he reparative response is mainly restricted to the implanted chon-rocytes (Chu et al., 2010). The implanted chondrocytes were notracked in this study since cell tracking could influence chondrocyteiology and hence, also the results of cartilage repair. So it remainsnclear as to whether at least the implanted, autologous chondro-ytes are indeed integrated into the repair tissue and survived for 6onth. The only sources for chondral defect repair apart from chon-

rocyte intrinsic repair capacity are the synovial tissue layer andells floating in the synovial fluid. It may be that implanted chon-rocytes may stimulate the healing response simply in a paracrineanner. In the case of osteochondral defects it has been concluded

t defects compared to native minipig cartilage (b1,5 and c1,5). The borders betweenm and 100 �m. Asterisk: defect areas. Arrows: proliferating chondrocyte clusters.

that paracrine stimuli by implanted chondrocytes can contribute tothe chondrogenic differentiation of immigrating MSCs (Gelse et al.,2009). A higher grade of intrinsic defect healing could be excludedin the present study since a critical defect size of 8 mm diameterwas created. In adult pigs, 6 mm diameter defects are critical defects(Ahern et al., 2009; Gotterbarm et al., 2008).

PGA was chosen as a well established reference material forcartilage repair (Erggelet et al., 2007, 2009; Liu et al., 2002; Zhouet al., 2006) which allows cartilage formation in vitro and in vivo (ElSayed et al., 2013; Lohan et al., 2011). PGA has been used for car-

tilage repair in the pig (Liu et al., 2002; Zhou et al., 2006) and in agoat model (Niederauer et al., 2000). However, most of these stud-ies dealing with implantation of PGA in pig joint cartilage defectsused an osteochondral model (Liu et al., 2002; Zhou et al., 2006).

A. Lohan et al. / Annals of Anatomy 195 (2013) 488– 497 495

Fig. 4. (a–e) Results of immunohistological scoring based on type II versus type I collagen and type X collagen are shown 6 months post construct implantation. The diagramsummarizes the results of the immunohistological scores (a). Healthy cartilage reached a maximal immunohistological score of 12.5 points. An outlier was excluded fromthe articular chondrocyte implantation group (GRUBB’s test). Representative pictures of the type II (b), type I (c) and type X collagen ((d) defect surface, (e) border betweeni l bone( fects.

a

Cccscctcbi

sacia

ntact and repaired area) immunolabeling of the healthy cartilage and subchondrab4, c4, d4, e4) chondrocyte seeded PGA scaffolds implanted into chondral joint dereas. Scale bars 50 �m.

artilage and bone healing has to be distinguished in the osteo-hondral defect model. MSCs derived from the bone marrowavity, which undergo enchondral ossification and chondrogene-is contribute to the healing process. Cartilage repair studies usinghondral models are rather scarce (Chang et al., 2006). PGA wasompletely degraded at the 6th month of investigation wherebyhe healing process was not completed at the same time. In mostases, repair tissue did not reach the niveau of bonding cartilage. Aiomaterial with a slower degradation rate would be advantageous

n future to support the repair process for a longer period of time.Neo-cartilage formation could be observed in the PGA con-

tructs seeded with articular and auricular chondrocytes in vitro

nd in vivo (El Sayed et al., 2013; Lohan et al., 2011). The auricularhondrocyte-based implants showed slightly inferior repair qual-ty at the macroscopic level and, biomechanical analysis revealed

greater stiffness in the repair tissues produced by auricular

(b1, c1, d1, e1), empty defects (b2, c2, d2, e2), articular (b3, c3, d3, e3) and auricularIsotype control: rabbit IgG (b5 and c5) and mouse IgG (d5 and e5). Asterisk: defect

chondrocyte implantation compared with the other groups. His-tologically and immunohistologically, the auricular chondrocytesshowed superior healing response when compared with the emptydefects or articular chondrocyte-based implants. The cell distri-bution in the repair tissue did not show major differences inthe present study. Zonal architecture was generally only partiallyrestored. However, the tidemarks were detectable, some of themwere doubled as a result of the transformation process during car-tilage repair. Several cell aggregates which result from chondrocyteproliferation could be observed in the defect or rim areas. Addition-ally, singular cell clusters were immunopositive for type X collagensuggesting that these cells might be prone to hypertrophy. Fur-

ther, one can speculate that these cartilage areas could undergoossification in future.

Not only in vivo but also in culture, auricular chondrocytes pro-duce elastic fibers (Moskalewski, 1981) which were not analyzed

496 A. Lohan et al. / Annals of Anato

Fig. 5. Biomechanical analysis of stiffness of the neocartilage within the jointdtc

iitcat2

iati

5

sdtotfimPcootastSpa

A

tSB

efects. A durometer was used to analyze the biomechanical stiffness of the repairissue (empty defect, articular and auricular chondrocyte seeded PGA scaffolds) inomparison to the healthy neighboring areas.

n the present study. Moreover, a less defined zonal architectures known for auricular cartilage. In addition to the higher cell con-ent of native auricular cartilage tissue, in vitro cultured auricularhondrocytes possess a four times greater proliferation rate thanrticular chondrocytes combined with a high chondrogenic poten-ial (Henderson et al., 2007; Panossian et al., 2001; Van Osch et al.,004).

Taken together, auricular chondrocyte implantation lead to annferior quality of the repair tissue concerning macroscopic defectppearance and biomechanical stiffness, but the histological struc-ure was slightly superior compared to articular chondrocyte basedmplants.

. Limitation and outlook

A control group which received the implantation of a cell-freecaffold was lacking in this pilot study. This group is necessary toistinguish more clearly the influence of the biomaterial alone onhe entire healing response. We had this control group in a previ-us study in the rabbit model [own unpublished results]. Fixation ofhe scaffolds was performed using cartilage sutures combined withbrin glue – this procedure still remains a problem since the sutureaterial or fibrin glue might affect healing. A rapid degradation of

GA might lead to implant loss. In future, auricular chondrocytesould be adapted for implantation into joint cartilage defects byptimization of tissue engineering strategies using growth factorr cytokine supplementation, further, conditioned media from cul-ured articular chondrocytes could be added or co-culturing withrticular chondrocytes could be performed. The results of this pilottudy allow a first insight into the healing process in responseo auricular chondrocyte implantation in a large animal model.ince the cartilage defect healing was not completed after 6-montheriod, a longer observation time of e.g. 12 months would be advis-ble.

cknowledgments

The authors also thank Ms. Hannah Gough for proofreading of

he manuscript. We thank Jens Oetvoes, Birgit Vogt and Ahmedherif for their support. This study was supported by grants fromayer Innovation GmbH and the Sonnenfeld Foundation.

my 195 (2013) 488– 497

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.aanat.2013.04.009.

References

Ahern, B.J., Parvizi, J., Boston, R., Schaer, T.P., 2009. Preclinical animal models insingle site cartilage defect testing: a systematic review. Osteoarthritis Cartilage17, 705–713.

Candrian, C., Vonwil, D., Barbero, A., Bonacina, E., Miot, S., Farhadi, J., Wirz, D., Dick-inson, S., Hollander, A., Jakob, M., Li, Z., Alini, M., Heberer, M., Martin, I., 2008.Engineered cartilage generated by nasal chondrocytes is responsive to physicalforces resembling joint loading. Arthritis Rheum. 58, 197–208.

Chang, C.H., Kuo, T.F., Lin, C.C., Chou, C.H., Chen, K.H., Lin, F.H., Liu, H.C., 2006. Tissueengineering-based cartilage repair with allogenous chondrocytes and gelatin-chondroitin-hyaluronan tri-copolymer scaffold: a porcine model assessed at 18,24, and 36 weeks. Biomaterials 27, 1876–1888.

Chu, C.R., Szczodry, M., Bruno, S., 2010. Animal models for cartilage regenerationand repair. Tissue Eng. B 16, 105–115.

El Sayed, K., Haisch, A., John, T., Marzahn, U., Lohan, A., Muller, R.D., Kohl, B., Ertel,W., Stoelzel, K., Schulze-Tanzil, G., 2010. Heterotopic autologous chondrocytetransplantation—a realistic approach to support articular cartilage repair? TissueEng. B 16, 603–616.

El Sayed, K., Marzahn, U., John, T., Hoyer, M., Zreiqat, H., Witthuhn, A., Kohl, B.,Haisch, A., Schulze-Tanzil, G., 2013. PGA-associated heterotopic chondrocytecocultures: implications of nasoseptal and auricular chondrocytes in articularcartilage repair. J. Tissue Eng. Regen. Med. 7, 61–72.

Erggelet, C., Endres, M., Neumann, K., Morawietz, L., Ringe, J., Haberstroh, K.,Sittinger, M., Kaps, C., 2009. Formation of cartilage repair tissue in articu-lar cartilage defects pretreated with microfracture and covered with cell-freepolymer-based implants. J. Orthop. Res. 27, 1353–1360.

Erggelet, C., Neumann, K., Endres, M., Haberstroh, K., Sittinger, M., Kaps, C., 2007.Regeneration of ovine articular cartilage defects by cell-free polymer-based

implants. Biomaterials 28, 5570–5580.Gelse, K., Brem, M., Klinger, P., Hess, A., Swoboda, B., Hennig, F., Olk, A., 2009.

Paracrine effect of transplanted rib chondrocyte spheroids supports formationof secondary cartilage repair tissue. J. Orthop. Res. 27, 1216–1225.

Gotterbarm, T., Breusch, S.J., Schneider, U., Jung, M., 2008. The minipig model forexperimental chondral and osteochondral defect repair in tissue engineering:retrospective analysis of 180 defects. Lab. Anim. 42, 71–82.

Greene, G.W., Banquy, X., Lee, D.W., Lowrey, D.D., Yu, J., Israelachvili, J.N., 2011.Adaptive mechanically controlled lubrication mechanism found in articular

joints. Proc. Natl. Acad. Sci. U.S.A. 108, 5255–5259.Henderson, J.H., Welter, J.F., Mansour, J.M., Niyibizi, C., Caplan, A.I., Dennis, J.E., 2007.

Cartilage tissue engineering for laryngotracheal reconstruction: comparisonof chondrocytes from three anatomic locations in the rabbit. Tissue Eng. 13,843–853.

Kaab, M.J., Gwynn, I.A., Notzli, H.P., 1998. Collagen fibre arrangement in the tibialplateau articular cartilage of man and other mammalian species. J. Anat. 193 (Pt1), 23–34.

Kuhne, M., John, T., El-Sayed, K., Marzahn, U., Aue, A., Kohl, B., Stoelzel, K., Ertel,W., Blottner, D., Haisch, A., Schulze-Tanzil, G., 2010. Characterization of auricu-lar chondrocytes and auricular/articular chondrocyte co-cultures in terms of anapplication in articular cartilage repair. Int. J. Mol. Med. 25, 701–708.

Liu, Y., Chen, F., Liu, W., Cui, L., Shang, Q., Xia, W., Wang, J., Cui, Y., Yang, G., Liu, D., Wu,J., Xu, R., Buonocore, S.D., Cao, Y., 2002. Repairing large porcine full-thicknessdefects of articular cartilage using autologous chondrocyte-engineered carti-lage. Tissue Eng. B 8, 709–721.

Lohan, A., Marzahn, U., El Sayed, K., Haisch, A., Kohl, B., Muller, R.D., Ertel, W.,Schulze-Tanzil, G., John, T., 2011. In vitro and in vivo neo-cartilage formationby heterotopic chondrocytes seeded on PGA scaffolds. Histochem. Cell Biol. 136,57–69.

Moskalewski, S., 1981. The elastogenetic process in transplants and cultures ofisolated auricular chondrocytes. Connect. Tissue Res. 8, 171–174.

Niederauer, G.G., Slivka, M.A., Leatherbury, N.C., Korvick, D.L., Harroff, H.H.,Ehler, W.C., Dunn, C.J., Kieswetter, K., 2000. Evaluation of multiphaseimplants for repair of focal osteochondral defects in goats. Biomaterials 21,2561–2574.

Panossian, A., Ashiku, S., Kirchhoff, C.H., Randolph, M.A., Yaremchuk, M.J., 2001.Effects of cell concentration and growth period on articular and ear

chondrocyte transplants for tissue engineering. Plast. Reconstr. Surg. 108,392–402.

Schulze-Tanzil, G., Muller, R.D., Kohl, B., Schneider, N., Ertel, W., Ipaktchi, K., Hunigen,H., Gemeinhardt, O., Stark, R., John, T., 2009. Differing in vitro biology of equine,ovine, porcine and human articular chondrocytes derived from the knee joint:an immunomorphological study. Histochem. Cell Biol. 131, 219–229.

Simon, T.M., Aberman, H.M., 2010. Cartilage regeneration and repair testing in asurrogate large animal model. Tissue Eng. B 16, 65–79.

Van Osch, G.J., Mandl, E.W., Jahr, H., Koevoet, W., Nolst-Trenite, G., Verhaar, J.A.,2004. Considerations on the use of ear chondrocytes as donor chondrocytes forcartilage tissue engineering. Biorheology 41, 411–421.

Anato

V

V

nasal chondrocytes and a silanized hydroxypropyl methylcellulose hydrogel.

A. Lohan et al. / Annals of

inatier, C., Bouffi, C., Merceron, C., Gordeladze, J., Brondello, J.M., Jorgensen, C.,

Weiss, P., Guicheux, J., Noel, D., 2009. Cartilage tissue engineering: towards abiomaterial-assisted mesenchymal stem cell therapy. Curr. Stem Cell Res. Ther.4, 318–329.

inatier, C., Magne, D., Moreau, A., Gauthier, O., Malard, O., Vignes-Colombeix, C.,Daculsi, G., Weiss, P., Guicheux, J., 2007. Engineering cartilage with human

my 195 (2013) 488– 497 497

J. Biomed. Mater. Res. A 80, 66–74.Zhou, G., Liu, W., Cui, L., Wang, X., Liu, T., Cao, Y., 2006. Repair of porcine artic-

ular osteochondral defects in non-weightbearing areas with autologous bonemarrow stromal cells. Tissue Eng. B 12, 3209–3221.