Osteoarthritis and Cartilage 23 (2015) 300e307

Surgical induction, histological evaluation, and MRI identification ofcartilage necrosis in the distal femur in goats to model early lesions ofosteochondrosis

F. T�oth y *, M.J. Nissi z x, L. Wang z, J.M. Ellermann z, C.S. Carlson yy Veterinary Population Medicine Department, University of Minnesota, St. Paul, MN, USAz Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota, 2021 Sixth Street SE, Minneapolis, MN 55455, United Statesx Department of Orthopaedic Surgery, University of Minnesota, Minneapolis, MN, USA

a r t i c l e i n f o

Article history:Received 22 May 2014Accepted 6 November 2014

Keywords:Cartilage necrosisAdiabatic T1rMRIOsteochondrosisGoat

* Address correspondence and reprint requests to:erinary Population Medicine, College of Veterinarynesota; 435 AnSci/VetMed, 1988 Fitch Avenue, St. Pau624-7727.

E-mail addresses: ftoth@umn.edu (F. T�oth), nisslnwang1222@gmail.com (L. Wang), eller001@ucarls099@umn.edu (C.S. Carlson).

http://dx.doi.org/10.1016/j.joca.2014.11.0091063-4584/© 2014 Osteoarthritis Research Society In

s u m m a r y

Objective: Identify and interrupt the vascular supply to portions of the distal femoral articular-epiphysealcartilage complex (AECC) in goat kids to induce cartilage necrosis, characteristic of early lesions ofosteochondrosis (OC); then utilize magnetic resonance imaging (MRI) to identify necrotic areas ofcartilage.Design: Distal femora were perfused and cleared in goat kids of various ages to visualize the vascularsupply to the distal femoral AECC. Vessels located on the axial aspect of the medial femoral condyle(MFC) and on the abaxial side of the lateral trochlear ridge were transected in eight 4- to 5-day-old goatsto induce cartilage necrosis. Goats were euthanized 1, 2, 3, 4, 5, 6, 9, and 10 weeks post operatively andoperated stifles were harvested. Adiabatic T1r relaxation time maps of the harvested distal femora weregenerated using a 9.4 T MR scanner, after which samples were evaluated histologically.Results: Interruption of the vascular supply to the MFC caused lesions of cartilage necrosis in 6/8 goatkids that were demonstrated histologically. Adiabatic T1r relaxation time mapping identified these areasof cartilage necrosis in 5/6 cases. No significant findings were detected after transection of perichondrialvessels supplying the lateral trochlear ridge.Conclusions: Cartilage necrosis, characteristic of early OC, can be induced by interrupting the vascularsupply to the distal femoral AECC in goat kids. The ability of high field MRI to identify these areas ofcartilage necrosis in the AECC using the adiabatic T1r sequence suggests that this technique may beuseful in the future for the early diagnosis of OC.

© 2014 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.

Introduction

Osteochondrosis (OC), or osteochondritis dissecans as it is oftenreferred to in human medicine, is a developmental orthopedicdisease that affects juvenile humans1 and animals2,3. In the veter-inary medical field OC is defined as a focal failure of endochondralossification3, whereas in human medicine it is often considered an

F. T�oth, Department of Vet-Medicine, University of Min-l, MN 55108, USA. Tel: 1-612-

i@cmrr.umn.edu (M.J. Nissi),mn.edu (J.M. Ellermann),

ternational. Published by Elsevier L

acquired, idiopathic lesion of the subchondral bone resulting indelamination and sequestration with or without articular cartilageinvolvement and instability1. In spite of differences in diseasedefinition, a recent review focusing on the clinical aspects of thedisease presents compelling evidence that OC affecting animals andhumans is the same disease4. Predilection sites are often sharedamong species, e.g., human OC most frequently affects the femoralcondyles, which are also commonly affected sites in pigs5.

The etiology and pathogenesis of OC affecting humans are poorlyunderstood; genetic factors, ischemic events, (repetitive micro)trauma, and ossification abnormalities all are considerations1,6,7.These uncertainties likely arise from the fact that OC in humans israrely recognized before clinical signs become apparent. Thus, in-ferences regarding the etiology and pathogenesis of OC in humansare largely based on histological evaluation of osteochondral

td. All rights reserved.

F. T�oth et al. / Osteoarthritis and Cartilage 23 (2015) 300e307 301

fragments removed from symptomatic joints8. Although thesefragments accurately represent the end stage disease, they providelittle information regarding its pathogenesis. This limited under-standing of the pathogenesis of OC makes treatment selection anddisease prognostication in humans challenging6.

In children with open physes, stable OC lesions are known toheal in 50% of the cases by simply altering activity levels9. However,predicting which lesions will heal and which will progress andrequire surgical treatment is difficult, and the disease is rarelydetected before the onset of clinical signs. The ability of conserva-tive treatment to succeed, if applied early during the course of thedisease, provides a strong argument for development of diagnosticimaging methods capable of identifying these lesions during theearly stages of their development. Additionally, novel imagingtechniques capable of identifying and monitoring the developmentof early OC lesions will likely yield insight into the diseasepathogenesis.

In contrast to humans, the pathogenesis of OC affecting animalshas been described in detail3. Histological studies using samplesobtained from predilection sites of OC in young, asymptomaticanimals have identified changes characteristic of specific stages ofthe disease. First, ischemic necrosis of the epiphyseal cartilagesubjacent to the articular cartilage (OC latens); followed by theassociated delay in the progression of the ossification front (OCmanifesta) after it reaches the area of necrotic cartilage. Finally,cleft formation through the area of necrosis (OC dissecans)10,11. Thepresence of necrotic vascular channels located within or adjacent tothe areas of necrosis provides strong evidence that vascular failureplays a central role in the development of cartilage necrosis, ahallmark of subclinical disease. However, the upstream eventstriggering ischemia in the epiphyseal cartilage in naturally occur-ring disease are yet to be identified. Recent work performed inhorses12 and pigs13 demonstrated that surgical interruption ofvascular supply to a portion of the articular-epiphyseal cartilagecomplex (AECC, immature joint cartilage at the ends of growinglong bones) results in development of lesions identical to thosepresent in naturally occurring OC. This ability to surgically induceOC-like lesions in animals suggests that development of a clinicallyrelevant animal model of the disease is feasible. However, the highincidence of naturally occurring disease, costs associated with theirhousing, and their large size are challenges in using horses and pigsas animal models. Goats have the potential to be an ideal model forsurgically-induced OC because of their low cost, long life span,absence of naturally occurring disease, and relatively small size thatallows evaluation of lesions by in vivomagnetic resonance imaging(MRI).

MRI has the potential to play a central role in the early diagnosisof OC by allowing noninvasive, in vivo investigations. Our recentstudies demonstrate that the vasculature of developing joints inpigs can be visualized using susceptibility weighted imaging14,15.Thus, this technique might be utilized to image failure of vascularsupply associated with OC. Relaxometry, a quantitative parametricMRI technique, is another method increasingly applied in theevaluation of cartilage16,17. T1r and adiabatic T1r, two of thesequantitative parameters, have received attention lately due to theirsensitivity to changes in cartilage proteoglycan content18e23. Re-sults from these studies suggest that cartilage necrosis and theassociated decrease in proteoglycan content, characteristic featuresof early OC, might be utilized for early diagnosis of this diseaseusing MRI.

In the present study, goat kids were used to study surgical in-duction of OC-like lesions. The objectives of the study reported herewere to: (1) assess the vascular supply to the AECC of the distalfemur in goat kids to identify vessels for surgical interruption, (2)induce cartilage necrosis in the AECC of the distal femur by

surgically interrupting the blood supply to a specific area, and (3)image early changes within the AECC after interruption of thevascular supply using high field MRI and validate these MRI find-ings by histological examination of the affected site.

Materials and methods

The study was conducted in two stages, which were approvedby the University of Minnesota Institutional Animal Care and UseCommittee (IACUC # 1207A18021).

First stage e vascular perfusion and clearing

Seven male dairy goat kids aged 4, 7, 14, 21, 28, 35, and 42 dayswere used to assess the vascular supply of the AECC of the distalfemur. The goats were sedated with an intramuscular injection of0.1 mg/kg xylazine, followed by 500 IU/kg heparin administeredintravenously. Immediately after heparin administration the goatswere euthanized and then eviscerated via a median laparotomy.The left external iliac artery was identified and catheterized with ared rubber catheter. The limb was then perfused with approxi-mately 150 mL radiographic contrast media (200 g BaSO4 [Sol-O-Pake] diluted to 1000 mL with 10% neutral buffered formalin). Af-ter completion of the perfusion, left distal femora were harvested,then cleared using the modified Spalteholz technique24. Briefly,samples were fixed in 10% neutral buffered formalin for 48 h thendehydrated by immersing them in sequentially increasing con-centrations of ethyl alcohol (70%, 80%, 90%, 95% then 100%).Dehydrated samples were cleared by placing them into concen-trated methyl-salicylate solution. Cleared specimens were visuallyinspected and then photographed to identify specific vessels as wellas the overall vascular architecture supplying the AECC.

Second stage e surgical interruption of vascular supply

During the second stage of the study we sought to surgicallyinterrupt the vascular supply to portions of the distal femur toinduce OC-like lesions. Initially, perichondrial vessels were trans-ected in six 14-day-old goats using the techniques described below;however, this did not result in OC-like lesion development identi-fiable histologically or byMRI over a period of 12weeks. A review ofthe results obtained in the first stage of the study revealed that, bythe age of 14 days, large numbers of vessels bridged the ossificationfront in the distal femoral epiphysis, providing an alternative bloodsupply to the perichondrial vessels that were surgically interrupted.Assuming that failure to induce necrosis in the 14-day-old goatswas due to this alternative blood supply, we elected to repeat theexperiment using 4- to 5-day-old goats, which had the mostextensive perichondrial blood supply with only a limited number ofvessels bridging the ossification front.

Eight 4- to 5-day-old male dairy goat kids were selected forsurgical intervention. Briefly, goats were anesthetized with acombination of midazolam and ketamine administered intrave-nously and anesthesia was maintained by inhalation of sevofluranevaporized in oxygen. The right pelvic limb was prepared for asepticsurgery and the stifle joint was approached via a parapatellararthrotomy. The patella was luxated, the femorotibial joint wasbrought into complete flexion, and the infrapatellar fat pad wasretracted, exposing the intercondylar region of the distal femurcranial to the attachment of the cranial cruciate ligament. Aftertransecting the perichondrial vessel that was consistently presenton the axial aspect of the medial femoral condyle (MFC) (Fig. 1) anadditional ~3 � 4 mm (beaver blade, goats euthanized at 1, 2, 4, 6,and 10 weeks post operatively) or 5 � 5 mm (custom designedblade, goats euthanized 3, 5, and 9 weeks post operatively) incision

Fig. 1. Drawing demonstrating the surgical procedure used to interrupt the vascular supply to the distal femoral AECC. Red, dashed lines indicate location of incisions used tointerrupt perichondrial vessels on the axial aspect of the MFC (right) and lateral trochlear ridge (left). The incision extending into the epiphyseal cartilage is indicated by the beaverblade entering the MFC. Perichondrial vessels interrupted in the axial aspect of the MFC (right inset) and the abaxial aspect of the lateral trochlear ridge (left inset) are also shown inperfused, cleared specimens obtained from a 2-week-old goat.

F. T�oth et al. / Osteoarthritis and Cartilage 23 (2015) 300e307302

extending into the AECC at the cranio-axial aspect of the MFC wascreated (Fig. 1) parallel to the articular surface to interrupt anyblood supply to the AECC traversing the ossification front. We alsoattempted to induce OC-like lesions in the proximal portion of thefemoral trochlea by transecting perichondrial vessels that wereconsistently present on the proximo-abaxial aspect of the lateraltrochlear ridge (Fig. 1). These vessels were transected in two loca-tions 3e4 mm apart. Upon completing the procedure, the jointcapsule, subcutaneous tissues, and skin were closed in layers. Afterassisted recovery, goats were housed together in an approximately6 � 6 m pen until euthanasia. Goats were euthanized 1, 2, 3, 4, 5, 6,9, and 10 weeks post operatively and distal femora were harvestedand photographed.

MRI

MR imaging of the specimens was conducted on a 9.4 T scannerdriven with VnmrJ software (VnmrJ, version 3.2A, Agilent Tech-nologies, Santa Clara, CA, USA), using a quadrature volume trans-ceiver coil (Millipede, Varian NMR Systems, Palo Alto, CA) within6 h of euthanasia. The specimens were placed in flexible latexcontainers for optimal fitting in the coil and immersed in per-fluoropolyether for clean and susceptibility matched background.Three-dimensional gradient recalled echo images were acquired tolocalize the operated areas within the MFC and the lateral trochlearridge. Subsequently, adiabatic T1r relaxation time23,25 of thespecimens was measured in a single axial slice covering the oper-ated area as determined from 3-D gradient recalled echo images.The relaxation time was measured using a magnetization prepa-ration block consisting of 0, 4, 8, 12 and 16 adiabatic full passagepulses from the hyperbolic secant family18 of 6 ms duration at peakpower gB1,max ¼ 2.5 kHz coupled to a fast spin echo readoutsequence23,25. Slice thickness was 1 mm with imaging matrix of256 � 256; echo train length was 8; echo spacing was set to min-imum (~5 ms) and a repetition time of 5 s was used. The final im-aging resolution was dependent on the sample size and wasapproximately 150e165 mm in the image plane. The adiabatic T1rrelaxation time was calculated using monoexponential 2-parameter fit to the measured data in Matlab (MATLAB R2011b,The MathWorks, Natick, MA, USA).

Histology

At the conclusion of the MRI procedures, distal femora fromeach goat were placed in 10% neutral buffered formalin for 48 h

followed by decalcification using 10% EDTA. Decalcified sampleswere divided to separate the femoral trochlea from the condylesand then were trimmed into 2e3 mm thick slabs in the coronalplane, yielding five slabs from each condyle and trochlea. Slabswere routinely processed into paraffin blocks and five mm thicksections were obtained from the surface of each slab and stainedwith hematoxylin and eosin. Blocks containing the surgical incisionwere thoroughly explored. If necrotic cartilage canals and/or areasof cartilage necrosis were identified in any of the sections, addi-tional sections were obtained in 200e500 mm intervals adjacent tothe original section until the lesion was no longer present. Selectedsections containing necrotic cartilage or cartilage canals were alsostained with safranin-O and toluidine blue to demonstrate changesin the extracellular matrix (ECM). Histological sections corre-sponding with the MRI slice in which the adiabatic T1r relaxationtime was measured were identified and changes in the cartilagemaps corresponding with changes noted in histological sectionswere compared. For both histological and MRI evaluations, corre-sponding areas of the lateral femoral condyle (LFC) and the medialtrochlear ridge within the same stifle joint served as control sites.

Results

During the first stage of the study, perfusion and subsequentclearing of the left pelvic limb in seven goats enabled successfulvisualization of the vascular supply to the AECC. Perichondrialvessels that were consistently present at the axial aspect of theMFC(Fig. 1) and the abaxial aspect of the proximolateral femoraltrochlea (Fig. 1) were designated for transection in the next stage ofthe study. The most abundant perichondrial blood supply,including small branches extending into the distal femoral AECC,was noted in the 4-day-old goat. Although perichondrial bloodsupply remained apparent in older goats, vessels crossing theossification front became more prominent as the animals aged.

Surgical procedures to interrupt blood supply to the distalfemoral AECC were completed in eight 4- to 5-day-old goatswithout any complications. After recovery, all animals exhibitednon-weightbearing lameness on the operated limb for 48e72 hwhich was treated with an NSAID (Flunixin meglumine 1.1 mg/kgsc) for 3 days and gradually resolved by the tenth post-operativeday. The goat that was euthanized 1 week post operatively devel-oped a seroma on the lateral side on the operated femorotibial jointwhich persisted until euthanasia. The goat that was euthanized 2weeks post operatively developed a hard, fibrous, pea-sizedswelling on the lateral side of the femorotibial joint involving the

Table

IHistologicalan

dMRIfindings

obtained

from

eigh

tgo

atkidsafterinterruption

oftheva

scularsu

pply

toportion

sof

thedistalfemur

Goa

tag

eat

thetimeof

euthan

asia

1wee

k2wee

k3wee

k4wee

k5wee

k6wee

k*9wee

k10

wee

k

Incision

Original

size

(mm)

~3�

4~3

�4

5�

5~3

�4

5�

5~3

�4

5�

5~3

�4

Hea

ling

No

Dee

pportion

isform

ing

gran

ulation

tissue

Dee

p~5

0%is

hea

led

Dee

p~7

5%is

hea

led

Dee

p~9

0%is

hea

led

Dee

pportion

isform

inggran

ulation

tissue

100%

hea

led

100%

hea

led

Cartilage

canals

Necrosisdistal

toincision

Necrosisin

axial~

50%

ofco

ndyle

Abs

entin

axial~

50%

ofco

ndyle

Abs

entin

axial~

20%

ofco

ndyle

Abs

entin

axial~

50%

ofco

ndyle

Abs

entin

axial~

50%

ofco

ndyle

Low

numbe

rsin

abax

ial50

%of

condyle

Abs

entor

chon

drified

through

out

Chon

drocy

teEx

tentof

necrosis

Adjacentto

incision

1.9mm

2area

inAEC

C1.8mm

2area

inAEC

C0.5mm

2area

inAEC

C1.2mm

2area

inAEC

CNot

apparen

t5.3mm

2area

atossification

fron

t0.6mm

2area

insu

bchon

dralbo

ne

Morpholog

ySh

rinka

ge,p

yknosis

Shrinka

ge,p

yknosis

Shrinka

ge,p

yknosis

Shrinka

ge,p

yknosis

Shrinka

ge,p

yknosis

Normal

Shrinka

ge,p

yknosis

Shrinka

ge,p

yknosis

ECMy

staining

Retained

Mild

lydecreased

Marke

dly

decreased

Mild

lydecreased

Marke

dly

decreased

Retained

Marke

dly

decreased

Nea

rlyab

sent

T1rh

orelaxa

tion

timein

necrotic

cartila

ge

N/A

zMild

lyincrea

sed

(~25

0ms)

Mod

eratelyincrea

sed

(~30

0ms)

Noch

ange

snoted

(~15

0ms)

Marke

dly

increa

sed

(~35

0ms)

Noch

ange

snoted

(~15

0ms)

Marke

dly

increa

sed

(~35

0ms)

Mild

lyincrea

sed

(~25

0ms)

OCclassification

N/A

OClatens

OClatens

OClatens

OClatens

N/A

OCman

ifesta

Hea

lingOC

man

ifesta

*In

the6-wee

k-oldgo

attheincision

into

thearticu

lar-ep

iphysea

lcartila

geco

mplexwas

crea

tedtoosu

perficially

relative

totherest

ofthean

imals.

yEC

M¼

extracellularmatrix.

zMRIstudymissedthearea

oflesion

.

F. T�oth et al. / Osteoarthritis and Cartilage 23 (2015) 300e307 303

periarticular soft tissues. The remaining six goats exhibited nocomplications prior to euthanasia at 3, 4, 5, 6, 9, and 10 weeks postoperatively.

Histological and MRI findings

Histological changes (Table I) in the goat euthanized 1week postoperatively were restricted to the AECC immediately adjacent anddistal to the surgical incision created in the axial aspect of the MFC.In the hematoxylin and eosin stained sections, the surgical incisionextending into the AECC was clearly visible and a thin zone ofnecrotic chondrocytes was present on either side of the incision.Cartilage canals located between the incision and the articularsurface appeared necrotic, characterized by loss of endothelium,absence of erythrocytes, and a lumen filled with amorphouseosinophilic material. Conversely, cartilage canals located outsidethis region lacked these features and appeared to be viable(Supplementary Fig. 1). Due to a mistake during setup, the MRIstudy missed the area of this lesion; thus, no MRI data wereavailable for this sample. No pathological changes were seen in thecorresponding control sections from the LFC.

More extensive histological changes were noted in the goatseuthanized 2, 3, 4 and 5 weeks post operatively (Table I). A portionof the epiphyseal cartilage that was located between the incisionand the axial articular surface of the MFC was necrotic in theseanimals, characterized by shrunken chondrocytes with pyknoticnuclei. Pallor of the ECM was also apparent in multiple adjacentsections (both hematoxylin and eosin and special stains). Thesefindings are consistent with OC latens26. In the MRI images, a mild(~250 ms) to moderate (~300 ms) focal increase in adiabatic T1rrelaxation time was observed in the epiphyseal cartilage corre-sponding with the area of cartilage necrosis in goats euthanized 2(Fig. 2), 3, and 5 (Fig. 3) weeks post operatively. In the remainder ofthe MFC adiabatic T1r relaxation time was similar to that observedin the LFC. In the goat euthanized 4 weeks post operatively adia-batic T1r relaxation times of the lateral and MFCs appeared similar(~150 ms) (Supplementary Fig. 2).

Histological results of the MFC in the goat euthanized 6 weekspost operatively indicated that the surgical incision in this animalwas created too close to the articular surface, entering approxi-mately at the border between the articular and epiphyseal cartilage.Deeper portions of the surgical incisionwere filled with granulationtissue delaying the advancement of the surrounding ossificationfront. The superficial portion of the incision persisted with no orminimal signs of healing, and appeared similar to that present inthe goat euthanized 1 week post operatively. In the MRI images ofthe MFC an irregularity at the cartilageebone interface (ossificationfront) was clearly apparent, but adiabatic T1r relaxation times wereno different in the AECC between the MFC and LFC (~150 ms)(Supplementary Fig. 3).

Histological findings obtained from the goat euthanized 9weekspost operatively were consistent with an OC manifesta lesion.Progression of the ossification front in the axial portion of the MFCwas delayed by the presence of a necrotic island (5.3 mm2) ofepiphyseal cartilage. This island of necrotic epiphyseal cartilagewasclearly present in the MRI image as a focal area of markedly(~350 ms) increased adiabatic T1r relaxation time adjacent to theossification front (Fig. 4). No pathologic changes were noted in thecontrol LFC.

An island composed of necrotic cartilage was identified withinthe subchondral bone in multiple adjacent sections obtained fromthe MFC of the goat euthanized 10 weeks post operatively. Theappearance and location of this lesion, apparently bypassed by theossification front, was consistent with a healing OC manifestalesion. In the MRI images the epiphyseal cartilage appeared to have

Fig. 2. Adiabatic T1r relaxation time map (panel A) obtained from the distal femur of a goat 2 weeks post operatively with the affected area enlarged, demonstrating mildlyincreased relaxation time in the area corresponding with the area of decreased proteoglycan content in panel B. Arrowhead indicates minor alterations in the advancement of theossification front (inset, panel A). Photomicrograph of a safranin-O stained section (panel B) obtained from the axial aspect of the MFC of the same specimen (scale bar ¼ 500 mm).The surgical incision is partially healed (black arrows in panel B). Faint staining marked by a dotted line indicates decreased proteoglycan content in the area of cartilage necrosis.Panel C depicts high magnification image (scale bar ¼ 50 mm) of the area marked by the black box in panel B in a hematoxylin and eosin stained section. Chondrocytes to the right ofthe dotted diagonal line are necrotic, characterized by pyknotic nuclei and shrinkage.

F. T�oth et al. / Osteoarthritis and Cartilage 23 (2015) 300e307304

nearly identical adiabatic T1r relaxation times in the LFC and MFC.In the subchondral bone of the MFC, however, there was a well-defined area with mildly (~250 ms) increased adiabatic T1r relax-ation time corresponding with the island of necrotic cartilage seenhistologically (Fig. 5). No pathologic changes were noted in thecontrol LFC.

No significant lesions were noted in any of the sections obtainedfrom the trochlear ridges in any of the goats.

Discussion

In the study reported here, surgical interruption of the vascularsupply to a portion of the axial aspect of the MFC resulted in focalnecrosis of the epiphyseal cartilage in 6/7 goats operated at the ageof 4e5 days. In the eighth goat, euthanized 1 week after the sur-gery, cartilage canal necrosis, a precursor of cartilage necrosis, wasnoted. Necrotic islands of epiphyseal cartilage were differentiatedfrom surrounding normal tissues ex vivo using adiabatic T1rrelaxation time maps generated using a 9.4 T magnet in 5/6 cases.

Vascular perfusion and clearing allowed visualization of thevasculature supplying the distal femoral AECC in goats, enablingidentification of perichondrial vessels for surgical interruption.Interestingly, the vessels designated for transection on the abaxialaspect of the lateral trochlear ridge had a nearly identical appear-ance and location to those which were interrupted in horses inanother study12 in which OC-like lesions of the lateral trochlearridge were successfully induced. Conversely, interruption of these

Fig. 3. Adiabatic T1r relaxation time map (panel A) obtained from the distal femur of a goincreased relaxation time in the area corresponding with the area of decreased proteoglycanfrom the axial aspect of the MFC of the same specimen (scale bar ¼ 500 mm). Faint staining mnecrosis. Panel C depicts high magnification image (scale bar ¼ 50 mm) of the area marked bthe right of the dotted diagonal line are necrotic, characterized by pyknotic nuclei and shr

vessels in goats failed to produce OC-like lesions in any of theoperated animals. This difference between species is likelyexplained by variations in the blood supply to the distal femoralAECC. Indeed, re-evaluation of the perfused specimens obtainedfrom goats ranging from 1 to 6 weeks old demonstrated a largenumber of small vessels bridging the ossification front, with someof these vessels already present in the 4-day-old goat. Thesebridging vessels are particularly important, because they providean alternative blood supply to the AECC, thereby decreasing itsdependence on the perichondrial vessels.

This apparent difference in blood supply prompted a minoralteration in the surgical technique employed in the 4- to 5-day-oldgoats whereby, in addition to interrupting the perichondrialvasculature (the only procedure performed in the 14-day-oldgoats), a horizontal incision extending into the AECC of the MFCwas also created to interrupt the vessels bridging the ossificationfront. It is likely that this additional step was critical in the suc-cessful induction of cartilage necrosis in six of the goats, as none ofthese animals developed a lesion in the trochlear ridges where onlythe perichondrial vessels were transected. These findings aredifferent from those obtained in 2-week-old horses12 and pigs up to6weeks old13, where cartilage necrosis was consistently induced byinterruption of the perichondrial blood supply. This species differ-ence in dependence on perichondrial blood supply of the AECC atolder ages might provide a partial explanation for the high inci-dence of OC in pigs27 and horses20 and the lack of reports in theliterature describing OC in goats.

at 5 weeks post operatively with the affected area enlarged, demonstrating markedlycontent in panel B. Photomicrograph of a safranin-O stained section (panel B) obtainedarked by a dotted line indicates decreased proteoglycan content in the area of cartilagey the black box in panel B in a hematoxylin and eosin stained section. Chondrocytes toinkage.

Fig. 4. Adiabatic T1r relaxation time map (panel A) obtained from the distal femur of a goat 9 weeks post operatively with the affected area enlarged, demonstrating markedlyincreased relaxation time in the area corresponding with the area of decreased proteoglycan content in panel B. Photomicrograph of a safranin-O stained section (panel B) obtainedfrom the axial aspect of the MFC of the same specimen (scale bar ¼ 500 mm). Faint staining marked by a dotted line indicates decreased proteoglycan content in the area of cartilagenecrosis. Delayed progression of the ossification front in the area of necrotic cartilage is also apparent, which is consistent with an OC manifesta lesion. Panel C depicts highmagnification image (scale bar ¼ 50 mm) of the area marked by the black box in panel B in a hematoxylin and eosin stained section. Chondrocytes to the right of the dotted diagonalline are necrotic, characterized by pyknotic nuclei and shrinkage. Chondrocyte clones are present in the viable cartilage to the left of the dotted diagonal line.

F. T�oth et al. / Osteoarthritis and Cartilage 23 (2015) 300e307 305

Our results, along with data obtained from studies using imag-ing and clearing techniques of the distal femora in pigs11 andhorses28,29, indicate that changes in the vascular architecture of theAECC occur at different times in various species. It is possible thatgoats have already passed a ‘threshold’ in joint development at thetime of birth, and the vascular architecture of their AECC bothmakes them resistant to naturally occurring OC and hampers sur-gical induction of OC-like lesions. This is in contrast to humans,where development of joints is still ongoing a decade after birth,and may explain in part why OC lesions are diagnosed at a moreadvanced age in humans than in animals6,30. It would be ideal totime the surgical intervention to induce OC-like lesions in anyspecies based on the degree of skeletal development as determinedby the degree of closure of the epiphyseal growth cartilage, how-ever this information is currently unavailable. Based on the volumeratio of the epiphyseal cartilage to the secondary ossification center(unpublished data) joints of goats appear to be more developed(increased volume of secondary ossification center compared toepiphyseal cartilage) than pigs or humans at the time of birth,which might also contribute to their apparent resistance to OC.

Another important finding of the present study is the remark-able ability of extensive lesions affecting the epiphyseal cartilage ingoats to heal, demonstrated by the fact that none of the goatsdeveloped clinically apparent disease (OC dissecans). Indeed, in thegoat euthanized 10 weeks post operatively, the ossification fronthad completely bypassed the necrotic cartilage, leaving it

Fig. 5. Adiabatic T1r relaxation time map (panel A) obtained from the distal femur of a goatincreased relaxation time in the area corresponding with the area of necrotic cartilage in panaspect of the MFC of the same specimen (scale bar ¼ 500 mm). An island composed of necrorelative to the epiphyseal cartilage, is retained in the subchondral bone. Necrotic cartilageincorporated in subchondral bone as the animal aged and the ossification front progresbar ¼ 50 mm) of the area marked by the black box in panel B in a hematoxylin and eosin stain

surrounded by bone. Rapid replacement/bypass of the necroticportion of the epiphyseal cartilage by bone appears to be aneffective strategy in preventing the progression of OC manifestalesions. Studies of spontaneous disease in swine suggest that areasof necrotic cartilage surrounded by bone gradually decrease in sizeand eventually disappear3,13.

Quantitativemapping of adiabatic T1r relaxation time identifiedareas of cartilage necrosis within the AECC or the subchondral bonein 5/6 cases. One plausible explanation for missing the lesion in thegoat examined 4 weeks post operatively in the MRI study is thesubstantial difference in average slice thickness between histology(5 mm) and the MRI (1 mm). Due to the thicker slices, MRI mayoccasionally miss (‘average out’) a lesion that is only a couplehundred mm in size. Alternatively, a small lesion could also bemissed due to operator error upon slice selection for adiabatic T1rimaging, as was demonstrated by our unfortunate failure to identifythe correct region for MRI in the goat euthanized 1 week postoperatively. Nevertheless, results obtained from the remaining fivegoats suggest that MRI allows identification of lesions withincartilage. Longitudinal monitoring of the developing OC lesionsmight also become possible after the above sequences are testedin vivo at a clinically applicable field strength.

An important limitation of our study is its inability to assessbiological variability due to our results being derived from singleobservation per time point. Also, it appears that goats receiving anincision measuring exactly 5 � 5 mm (created with the custom

10 weeks post operatively with the affected area enlarged, demonstrating moderatelyel B. Photomicrograph of a safranin-O stained section (panel B) obtained from the axialtic cartilage (marked by the dotted line), characterized by markedly decreased stainingthat was initially surrounded by epiphyseal cartilage (as shown in Fig. 3) has becomesed towards the articular surface. Panel C depicts high magnification image (scaleed slide. Tissue to the right of the dotted diagonal line is composed of necrotic cartilage.

F. T�oth et al. / Osteoarthritis and Cartilage 23 (2015) 300e307306

designed blade) developed more uniform lesions than thoseoperated using the beaver blade; therefore, future studies wouldideally use this larger, more standardized incision.

While initial results are promising, further refinement of thesurgical technique presented here, as well as evaluation of addi-tional joints, should be performed before recommending the use ofgoats as a model of surgically-induced OC. In its current stage themodel is best suited to evaluate repair mechanism of ischemiccartilage necrosis rather than studying clinically apparent disease,i.e., OC dissecans. Our findings do suggest an explanation as to whygoats appear to be free of naturally occurring OC, and may providefurther insight into general disease pathogenesis. The ability ofultra high field MRI using the adiabatic T1r sequence to identifyareas of necrotic cartilage within the AECC is an important finding,which is currently being evaluated at clinically relevant fieldstrengths.

Author contributions

All authors contributed significantly to the article formulation.FT drafted the primary manuscript with input from CSC. All authorswere involved in critical revision of the manuscript and approvedthe final version to be published. Surgical procedures and histo-logical evaluation were performed by FT and CSC. MRI was per-formed and interpreted byMJN, LW, and JME. Responsibility for theintegrity of this work as a whole is assumed by T�oth (ftoth@umn.edu), and Carlson (carls099@umn.edu).

Role of the funding sourceSupport from NIH grants T32OD010993, K18OD010468, P41EB015894, R21AR065385,W.M. KECK Foundation, and University ofMinnesota College of Veterinary Medicine Comparative MedicineFaculty Grant were used during the completion of the study.

Conflict of interestThere are no conflicting interests to report.

Acknowledgments

The authors are grateful to Lindsey Harper, Paula Overn,Drs Annette McCoy, Elizabeth Pluhar, and Lindsey Mathews fortheir assistance with the histologic, anesthetic, and surgicalprocedures.

Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.joca.2014.11.009.

References

1. Kocher MS, Tucker R, Ganley TJ, Flynn JM. Management ofosteochondritis dissecans of the knee: current concepts re-view. Am J Sports Med 2006;34:1181e91.

2. Foland JW, McIlwraith CW, Trotter GW. Arthroscopic surgeryfor osteochondritis dissecans of the femoropatellar joint of thehorse. Equine Veterinary J 1992;24:419e23.

3. Ytrehus B, Carlson CS, Ekman S. Etiology and pathogenesis ofosteochondrosis. Vet Pathol 2007;44:429e48.

4. McCoy AM, Toth F, Dolvik NI, Ekman S, Ellermann J, Olstad K,et al. Articular osteochondrosis: a comparison of naturally-occurring human and animal disease. Osteoarthritis Cartilage2013;21:1638e47.

5. Fisher DR, De Smet AA. Radiologic analysis of osteochondritisdissecans and related osteochondral lesions. Contemp DiagRad 1993;16:1e5.

6. Crawford DC, Safran MR. Osteochondritis dissecans of theknee. J Am Acad Orthop Surg 2006;14:90e100.

7. Kocher MS, Czarnecki JJ, Andersen JS, Micheli LJ. Internal fix-ation of juvenile osteochondritis dissecans lesions of the knee.Am J Sports Med 2007;35:712e8.

8. Yonetani Y, Nakamura N, Natsuume T, Shiozaki Y, Tanaka Y,Horibe S. Histological evaluation of juvenile osteochondritisdissecans of the knee: a case series. Knee Surg Sports Trau-matol Arthrosc: official journal of the ESSKA 2010;18:723e30.

9. Cahill BR, Phillips MR, Navarro R. The results of conservativemanagement of juvenile osteochondritis dissecans using jointscintigraphy. A prospective study. Am J Sports Med 1989;17:601e5. discussion 605e606.

10. Carlson CS, Meuten DJ, Richardson DC. Ischemic necrosis ofcartilage in spontaneous and experimental lesions of osteo-chondrosis. J Orthop Res: official publication of the Ortho-paedic Research Society 1991;9:317e29.

11. Ytrehus B, Carlson CS, Lundeheim N, Mathisen L, Reinholt FP,Teige J, et al. Vascularisation and osteochondrosis of theepiphyseal growth cartilage of the distal femur inpigsedevelopment with age, growth rate, weight and jointshape. Bone 2004;34:454e65.

12. Olstad K, Hendrickson EH, Carlson CS, Ekman S, Dolvik NI.Transection of vessels in epiphyseal cartilage canals leads toosteochondrosis and osteochondrosis dissecans in thefemoro-patellar joint of foals; a potential model of juvenileosteochondritis dissecans. Osteoarthritis Cartilage 2013;21:730e8.

13. Ytrehus B, Andreas Haga H, Mellum CN, Mathisen L,Carlson CS, Ekman S, et al. Experimental ischemia of porcinegrowth cartilage produces lesions of osteochondrosis. J OrthopRes: official publication of the Orthopaedic Research Society2004;22:1201e9.

14. Nissi MJ, Toth F, Zhang J, Schmitter S, Benson M, Carlson CS,et al. Susceptibility weighted imaging of cartilage canals inporcine epiphyseal growth cartilage ex vivo and in vivo. MagnReson Med 2014;71:2197e205.

15. Toth F, Nissi MJ, Zhang J, Benson M, Schmitter S, Ellermann JM,et al. Histological confirmation and biological significance ofcartilage canals demonstrated using high field MRI in swine atpredilection sites of osteochondrosis. J Orthop Res 2013;31:2006e12.

16. Gold GE, Burstein D, Dardzinski B, Lang P, Boada F, Mosher T.MRI of articular cartilage in OA: novel pulse sequences andcompositional/functional markers. Osteoarthritis Cartilage2006;14(Suppl A):A76e86.

17. Nieminen MT, Nissi MJ, Mattila L, Kiviranta I. Evaluation ofchondral repair using quantitative MRI. J Magn Reson Imaging2012;36:1287e99.

18. Garwood M, DelaBarre L. The return of the frequency sweep:designing adiabatic pulses for contemporary NMR. J MagnReson 2001;153:155e77.

19. Li X, Benjamin Ma C, Link TM, Castillo DD, Blumenkrantz G,Lozano J, et al. In vivo T(1rho) and T(2) mapping of articularcartilage in osteoarthritis of the knee using 3 T MRI. Osteoar-thritis Cartilage 2007;15:789e97.

20. Regatte RR, Akella SV, Borthakur A, Kneeland JB, Reddy R.Proteoglycan depletion-induced changes in transverse relax-ation maps of cartilage: comparison of T2 and T1rho. AcadRadiol 2002;9:1388e94.

21. Wheaton AJ, Dodge GR, Borthakur A, Kneeland JB,Schumacher HR, Reddy R. Detection of changes in articular

F. T�oth et al. / Osteoarthritis and Cartilage 23 (2015) 300e307 307

cartilage proteoglycan by T(1rho) magnetic resonance imag-ing. J Orthop Res 2005;23:102e8.

22. Salo E-N, Liimatainen T, Michaeli S, Mangia S, Ellermann J,Nieminen MT, et al. T1rho dispersion in constituent-specificdegradation models of articular cartilage with correlation tobiomechanical properties. In: Intl. Soc. Magn. Reson. Med2012;vol. 20:51. Melbourne.

23. Ellermann J, Ling W, Nissi MJ, Arendt E, Carlson CS,Garwood M, et al. MRI rotating frame relaxation measure-ments for articular cartilage assessment. Magn Reson Imaging2013;31:1537e43.

24. Guyer MF. Objects of general interest. In: Animal Micrology.Practical Exercises in Zoological Micro-technique. Chicago:University of Chicago Press; 1953:110e3.

25. Michaeli S, Sorce DJ, Springer Jr CS, Ugurbil K, Garwood M.T1rho MRI contrast in the human brain: modulation of thelongitudinal rotating frame relaxation shutter-speed during anadiabatic RF pulse. J Magn Reson 2006;181:135e47.

26. Ytrehus B, Ekman S, Carlson CS, Teige J, Reinholt FP. Focalchanges in blood supply during normal epiphyseal growth arecentral in the pathogenesis of osteochondrosis in pigs. Bone2004;35:1294e306.

27. Ytrehus B, Grindflek E, Teige J, Stubsjoen E, Grondalen T,Carlson CS, et al. The effect of parentage on the prevalence,severity and location of lesions of osteochondrosis inswine. J Vet Med A Physiol Pathol Clin Med 2004;51:188e95.

28. Olstad K, Ytrehus B, Ekman S, Carlson CS, Dolvik NI. Epiphysealcartilage canal blood supply to the distal femur of foals. EquineVet J 2008;40:433e9.

29. Olstad K, Ytrehus B, Ekman S, Carlson CS, Dolvik NI. Early le-sions of articular osteochondrosis in the distal femur of foals.Vet Pathol 2011;48:1165e75.

30. Hixon AL, Gibbs LM. Osteochondritis dissecans: a diagnosis notto miss. Am Fam Physician 2000;61:151e6.