Spontaneous Repair of Full- Thickness Defects of Articular Cartilage in a Goat Model A PRELIMINARY STUDY

Transcription

1 53 Spontaneous Repair of Full- Thickness Defects of Articular Cartilage in a Goat Model A PRELIMINARY STUDY BY DOUGLAS W. JACKSON, MD, PEGGY A. LALOR, PHD, HAROLD M. ABERMAN, DVM, AND TIMOTHY M. SIMON, PHD Investigation performed at the Orthopaedic Research Institute, Southern California Center for Sports Medicine, Long Beach, California Background: Full-thickness defects measuring 3 mm in diameter have been commonly used in studies of rabbits to evaluate new procedures designed to improve the quality of articular cartilage repair. These defects initially heal spontaneously. However, little information is available on the characteristics of repair of larger defects. The objective of the present study was to define the characteristics of repair of 6-mm full-thickness osteochondral defects in the adult Spanish goat. Methods: Full-thickness osteochondral defects measuring 6 6 mm were created in the medial femoral condyle of the knee joint of adult female Spanish goats. The untreated defects were allowed to heal spontaneously. The knee joints were removed, and the defects were examined at ten time-intervals, ranging from time zero (immediately after creation of the defect) to one year postoperatively. The defects were examined grossly, microradiographically, histologically, and with magnetic resonance imaging and computed tomography. Results: The 6-mm osteochondral defects did not heal. Moreover, heretofore undescribed progressive, deleterious changes occurred in the osseous walls of the defect and the articular cartilage surrounding the defect. These changes resulted in a progressive increase in the size of the defect, the formation of a large cavitary lesion, and the collapse of both the surrounding subchondral bone and the articular cartilage into the periphery of the defect. Resorption of the osseous walls of the defect was first noted by one week, and it was associated with extensive osteoclastic activity in the trabecular bone of the walls of the defect. Flattening and deformation of the articular cartilage at the edges of the defect was also observed at this time. By twelve weeks, bone resorption had transformed the surgically created defect into a larger cavitary lesion, and the articular cartilage and subchondral bone surrounding the defect had collapsed into the periphery of the defect. By twenty-six weeks, bone resorption had ceased and the osseous walls of the lesion had become sclerotic. The cavitary lesion did not become filled in with fibrocartilage. Instead, a cystic lesion was found in the center of most of the cavitary lesions. Only a thin layer of fibrocartilage was present on the sclerotic osseous walls of the defect. Specimens examined at one year postoperatively showed similar characteristics. Conclusions: Full-thickness osteochondral defects, measuring 6 mm in both diameter and depth, that are created in the medial femoral condyle of the knee joint of adult Spanish goats do not heal spontaneously. Instead, they undergo progressive changes resulting in resorption of the osseous walls of the defect, the formation of a large cavitary lesion, and the collapse of the surrounding articular cartilage and subchondral bone. Clinical Relevance: As surgeons apply new reparative procedures to larger areas of full-thickness articular cartilage loss, we believe that it is important to consider the potential deleterious effects of a zone of influence secondary to the creation of a large defect in the subchondral bone. When biologic and synthetic matrices with or without cells or bioactive factors are placed into surgically created osseous defects, the osseous walls serve as shoulders to protect and stabilize the preliminary repair process. It is important to protect the repair process until biologic incorporation occurs and the chondrogenic switch turns the cells on to synthesize an articular-cartilagelike matrix. It takes a varying period of time to fill a large, surgically created bone defect underlying a chondral surface. The repair of such a defect requires bone synthesis and the reestablishment of a subchondral plate with a tidemark transition to the new overlying articular surface. The prevention of secondary changes in the surrounding bone and articular cartilage and the durability of the new reparative tissue making up the articulating surface are issues that must be addressed in future studies. COPYRIGHT 2001 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED

2 54 Surgically created full-thickness articular cartilage defects have been studied in a variety of animal models to evaluate the effect of new procedures designed to elicit articular cartilage repair 1-8. Although these surgically created defects differ from the defects that occur in human patients following trauma and/or cartilage degeneration 9,10, they have been used to study techniques and methodologies that might eventually be used in humans. In large osteochondral defects in animal models, both the cartilage and the bone surrounding the defect undergo immediate changes and have different reparative responses depending on the size, shape, location, and depth of the lesion 11 as well as on the animal model being used. In the rabbit, 3-mm defects initially heal spontaneously with repair tissue composed of hyaline or fibrocartilage 8. The spontaneous repair of larger defects has not been systematically studied in a rabbit model, to our knowledge. In the goat, a 3-mm defect in the femoral groove (which is relatively small compared with the size of the joint) also undergoes repair initially 2. Moreover, lesions in the femoral groove tend to heal better than those in the rigorous loading environment of the femoral condyle. The repair of larger defects, especially those in the medial femoral condyle, has not been studied in a goat model, to our knowledge. New procedures designed to elicit articular cartilage repair must be analyzed, and the results must be compared with the spontaneous changes that occur in untreated defects of the same size. The untreated defects used for comparison must also be site-matched for the specific location under examination. It is essential to be certain that a new methodology enhances repair or prevents the deleterious effects of a similar untreated lesion, and this approach requires long-term controls. The goat model has been used to evaluate different techniques designed to elicit articular cartilage repair 2,5. The purpose of the present study was to characterize the sequential changes that occurred within and around a large full-thickness defect that was created in the weight-bearing region of the medial femoral condyle of the goat. Materials and Methods full-thickness defect, 6 mm in both diameter and depth, Awas created in the middle one-third of the medial femoral condyle of the right knee (stifle) joint of twenty-four adult female Spanish goats with use of a specially designed instrument. The center of the defect was in the middle portion of the medial femoral condyle. The articular cartilage in this region is approximately 2 mm thick, and the defect penetrated into and resulted in the removal of approximately 4 mm of the underlying subchondral bone. The defect was created unilaterally in eighteen goats and bilaterally in six goats. The untreated defects were allowed to heal spontaneously. The goats were killed humanely, the joints were removed, and the defects were examined at time zero (immediately postoperatively); at fifteen and thirty minutes; at two days; and at one, two, six, twelve, twenty-six, and fifty-two weeks. The six goats that were killed within two days postoperatively constituted the short-term group, and the eighteen goats that were killed one to fifty-two weeks postoperatively constituted the long-term group (Table I). The study protocol was reviewed and approved by the Institutional Animal Care and Use Committee, and all of the animals received humane care in compliance with the Guide for the Care and Use of Laboratory Animals, published by the National Institutes of Health. Changes in the adjacent surrounding cartilage, the characteristics of the repair tissue that formed in the defect, and changes in the subchondral bone underlying and adjacent to the defect were observed and recorded. The contralateral joint in the eighteen animals in the long-term group served as controls. Immediately after the animals were killed, both stifle joints (except those undergoing magnetic resonance imaging and computerized axial tomography) were opened, grossly evaluated, disarticulated, and fixed in neutral buffered formalin and then in ethanol. The specimens were cut into 2-mm slabs in the coronal plane, through the center of the defect. Each slab was placed onto autoradiographic film (X-OMAT AR; Eastman Kodak, Rochester, New York) and into an x-ray TABLE I Data on the Evaluation of Twenty-four Goats with Full-Thickness Cartilage Defects Time-Point* No. of Goats No. of Defects Histologic Sections Evaluation Microradiography MRI CT Short-term group 0 and 2 days and 30 mins Long-term group 1 wk wks wks wks wks wks *The defect was created bilaterally in the six goats in the short-term group. In three goats, the defect was created in the left knee at time zero and in the right knee at two days. In the remaining three goats, the defect was created in the right knee at fifteen minutes and in the left knee at thirty minutes. The defect was created unilaterally in the eighteen goats in the long-term group. The values are expressed as the number of defects. MRI = magnetic resonance imaging, and CT = computerized axial tomography.

3 55 Fig. 1 Gross appearance of the articular cartilage defect in the medial femoral condyle at selected time-intervals after creation of the defect. A: Time zero (immediately after creation of the lesion). B: One week. C: Two weeks. D: Six weeks. E: Twenty-six weeks. F: Fifty-two weeks. machine (Faxitron; Hewlett Packard, Wheeling, Illinois) set to a 20-kV peak. Following a fifteen-second exposure, the film was developed with an x-ray film processor (Mini Medical series; AFP Imaging, Elmsford, New York). The samples were decalcified and then were embedded in paraffin. Serial sections were cut and were stained with hematoxylin and eosin for routine histologic evaluation. Additional sections were stained with safranin O for evaluation of changes in the glycosaminoglycan content of the cartilage matrix. Magnetic Resonance Imaging Two knees that had a defect and two control knees were evaluated with magnetic resonance imaging in addition to histologic analysis at one year. The knees were removed 8 cm proximal and distal to the joint line immediately after the animals were killed. The knees were wrapped in a moistureproof plastic bag, extended, immersed in a saline bath, and placed in a standard knee-imaging coil. Magnetic resonance imaging scans were made on a high-resolution, 1.5-tesla scanner (Magnetom Vision; Siemens, Iselin, New Jersey). The standard examination consisted of contiguous 3-mm images in the coronal and sagittal planes. A turbo spin-echo sequence with a pulse-repetition time of 4100 to 4700 ms and an effective echo-delay time of 19 to 93 ms was used. An 18-cm field of view and a 256-by-256 matrix were used, resulting in a pixel size of mm. Computerized Axial Tomography Three specimens were evaluated with computerized axial tomography at one year. The specimens were prepared for analysis in a manner similar to that used for the knees that were evaluated with magnetic resonance imaging. One-millimeterthick images of each medial femoral condyle were then made in the coronal plane with use of a computerized axial tomography scanner (Somatom Plus 4; Siemens). Overlay Studies At time zero and at one year, the medial profile of the medial femoral condyle of knees with a defect and control knees was captured with use of image-analysis software (Cue-2; Olympus, La Palma, California). All images were made with the camera at the same calibration settings and magnification. The images were converted to binary black-and-white images that were then analyzed for curvature in order to determine whether the creation of the defect caused any change in the contour of the medial femoral condyle. For specimens that were evaluated at time zero, the profile of the normal condyle was captured, the condyle was drilled to create the defect, the profile of the same condyle was captured again, and the resulting image was overlaid onto the first image. For the specimens that were evaluated at one year, the profile of the contralateral condyle was captured, mirrored, and overlaid onto the profile of the condyle with the lesion.

4 56 Fig. 2 Microradiographic appearance of coronal sections of the lesion at selected time-intervals after creation of the defect. A: Time zero (immediately after creation of the lesion). B: Thirty minutes. C: Two days (note the blood clot filling the defect). D: Two weeks. E: Six weeks (note the widening of the defect wall). F: Fifty-two weeks (note the marked change in the geometry of the lesion compared with the geometry in A). Results he gross appearance of the lesions Tdemonstrated consistent and reproducible changes over time (Fig. 1, A through F). Immediately after the fullthickness defect was created (time zero), the edges of the cartilage defect were sharply defined. Within fifteen minutes the adjacent cartilage matrix began to flow over the rim of the defect, and it remained in that position at all of the subsequent time-periods. The adjacent articular cartilage surface demonstrated a gradual flattening surrounding the lesion in an associated zone of influence (Figs. 1 and 2). The zone of influence expanded over time. The defects were empty at time zero, but within fortyeight hours a blood clot had filled the defects to within 2 mm of the surface. The clot was attached to the rim of the cartilage defect, with a concavity toward the center. By twelve weeks, fibrillation of the articular cartilage adjacent to the defect and fraying of the inner rim of the medial meniscus were observed. By twenty-six weeks, repair tissue still had not totally filled the defect and the surface remained concave (Fig. 1, E). The articular cartilage surrounding the lesion had collapsed into the upper part of the defect (Fig. 1, E). By fifty-two weeks, the original defect opening had narrowed in three of the four animals, leaving a central 2 to 3-mm gap separating the edges of the cartilage surrounding the previously created defect (Figs. 1, F and 2, F). In one of the animals evaluated at fifty-two weeks, there was no central gap and the defect was occupied by tissue containing deep clefts that appeared to have originated from collapsed surrounding cartilage. Histologic sections also demonstrated progressive changes in the cartilage and bone surrounding the defect as well as within the repair tissue filling the defect (Figs. 3 through 6). A summary of the histologic changes is given in Table II. At time zero, histologic sections demonstrated a well-defined defect involving articular cartilage and subchondral bone. By thirty minutes, there was deformation of cartilage into the peripheral rim of the defect. However, no changes in the bone surrounding the defect were observed. Approximately 20% of the defect was filled with tissue composed of fibrin, red blood cells, and occasional neutrophils. Within forty-eight hours, chondrocyte clusters were present in the cartilage along the margins of the defect, particularly in areas where the cartilage matrix protruded into the defect. More neutrophils were seen in the tissue filling in the defect as well as in the surrounding bone. This tissue demonstrated a fibrinoid upper layer that was attached to the edges of the articular cartilage defect and had a concave surface (Fig. 3, A). By one week, extensive osteoclastic activity was observed in the trabecular bone and osteogenesis was noted at the base and in the walls of the defect. Chondrocyte cloning was more extensive in the articular cartilage edges of all of the defects, and it continued to be a prominent feature at later time-points. The concave surface layer of the tissue in the defect contained fibroblast-like cells and collagen fibers oriented parallel to the surface. Safranin-O staining showed a loss of proteoglycans in the

5 57 TABLE II Summary of Observations in the Subchondral Bone, Repair Tissue, and Surrounding Artcular Cartilage at Each Time-Point Time-Point* Short-term group No. of Defects Subchondral Bone 0 mins 3 Lesion creation, fresh cut bone edges Repair Tissue Observations Bleeding in defect Surrounding Articular Cartilage Lesion creation, welldefined sharp edges 15 mins 3 No visible change in bone Bleeding in defect Deformation into defect 30 mins 3 No visible change in bone Fibrin strands, red blood Deformation into defect cell clot, and sparse polymorphonuclear neutrophils 2 days 3 Increase in cellularity and Fibrin strands at surface Deformation into defect red blood cells in marrow attached to cartilage edges, central area containing fibrin-rich blood clot and more polymorphonuclear neutrophils Long-term group 1 wk 3 Bone resorption and formation in walls and base of defect 2 wks 3 Resorption of walls of defect, endochondral bone formation at base and walls of lesion 6 wks 3 Sparse osteoclastic activity except in subchondral area at lesion edges, sparse endochondral bone formation 12 wks 3 Structural collapse of lesion walls; endochondral bone formation under cartilage, walls, and base 26 wks 2 Defect walls extending farther outward due to structural collapse, sclerotic thickening, endochondral bone formation at periphery of lesion and subchondral area 52 wks 4 Sclerotic thickening, narrowing (3 goats) and closure (1 goat) of lesion aperture, vascular subchondral bone plate, osteophytes, sparse bone resorption, endochondral bone formation at periphery of lesion and subchondral area Concave upper layer containing fibroblast-like cells and collagen fibers, central area containing fibrin-rich material Concave upper layer similar to that at 1 wk, central area containing loose fibrous repair tissue Concave fibrous upper layer integrated with cartilage edge, collagen-fiber repair tissue oriented perpendicular to surface layer Concave fibrous upper layer, periphery of lesion demonstrating transition of fibrocartilage to vascular tissue with small central void Concave fibrous upper layer contiguous with adjacent surrounding cartilage, fibrocartilage at base, transition to fibrous tissue with central cyst, focal safranin-o staining at periphery and base of lesion Thin fibrous upper layer contiguous with cartilage edge, periphery demonstrating transition of fibrocartilage to disorganized fibrous tissue with central cyst, focal safranin-o staining at periphery and base of lesion Deformation into defect, chondrocyte clusters at cut edge Deformation of upper edges into defect, increased chondrocyte clusters at cut edge Deformation, flattening, chondrocyte clusters, cartilage edges being undermined by multinucleated giant cells in repair tissue Fibrillation, clefts, flattening, chondrocytes (single and clusters) present in edges Fibrillation, condylar flattening, chondrocytes (single and clusters), narrowing of aperture, safranin-o staining of cartilage to edge of lesion except superficial layer Fibrillation, flattening of newly formed repair cartilage and integration with old cartilage but with irregular surface, tidemark not reestablished, narrowing of aperture, safranin-o staining of cartilage to edge of lesion *The defect was created bilaterally in the six goats in the short-term group. In three goats, the defect was created in the left knee at time zero and in the right knee at two days. In the remaining three goats, the defect was created in the right knee at fifteen minutes and in the left knee at thirty minutes. The defect was created unilaterally in the eighteen goats in the long-term group.

6 58 articular cartilage adjacent to the periphery of the defect. By two weeks, osteoblastic bone formation was observed at the edges and base of the defect and was occasionally seen in the calcified cartilage zone at the edges of the defect. In addition, foci of endochondral bone formation could be seen at the base of the defect. Further collapse of the edges of the articular cartilage into the defect was noted (Fig. 3, B). By six weeks, the edges of the cartilage and subchondral bone adjacent to the defect were being undermined by multinucleated giant cells that were present in the tissue filling the defect (Fig. 3, C). Sparse osteoclasts were seen and little bone formation was observed in the surrounding bone. The fibrous tissue layer at the surface was still present. The cells and collagen fibers within the defect were oriented perpendicular to the fibrous tissue along the surface. By twelve weeks, structural collapse of the subchondral bone adjacent to the defect was noted (Fig. 3, D). Endochondral bone formation was noted beneath this cartilage-like material. Beneath the concave surface of the repair tissue, fibrocartilage was present at the periphery of the defect and vascular fibrous tissue was present within the center of the defect. By twenty-six weeks, the osseous walls of the defect had Fig. 3 Histologic sections at selected time-intervals after creation of the defect. A: Low-magnification view at two days, showing that a blood clot has nearly filled the defect (hematoxylin and eosin, 1). B: Lowmagnification view at two weeks, showing boneremodeling in the walls of the lesion and a dense cellular connective tissue lining the walls and penetrating into the marrow space (hematoxylin and eosin, 1). C: Low-magnification view at six weeks, showing a defect nearly filled with repair tissue and a small, central cyst-like structure. The repair tissue forms a concave surface that is integrated with the articular cartilage edges (hematoxylin and eosin, 1). D: Low-magnification view at twelve weeks, showing a defect containing repair tissue and a large, central cyst. The bone comprising the walls of the defect has been resorbed. In this specimen, fibrillation and clefts developed in the articular cartilage adjacent to the lesion (hematoxylin and eosin, 1). E: Lowmagnification view at twenty-six weeks, showing an increase in the size of the defect due to resorption of the bone in the sides and base of the defect. The walls of the defect have become sclerotic. A cyst containing loose fibrous tissue is present in the central portion of the lesion. This tissue appears to be integrated with the surrounding cartilage. The aperture of the lesion is narrowing as the surrounding articular cartilage and bone collapse into the lesion (hematoxylin and eosin, 1). F: View of the same area depicted in E, showing the extent of safranin-o staining. Note the focal areas of uptake around the periphery of the lesion (safranin O-fast green, 1). G: Low-magnification view of another defect at twenty-six weeks, showing extensive resorption of bone surrounding the defect. A central cyst contains loose fibrous repair tissue that appears to be integrated with the surrounding cartilage. The aperture of the lesion is narrowing as the surrounding articular cartilage extends into the lesion (hematoxylin and eosin, 1). H: View of the same area depicted in G, showing the extent of safranin-o staining. Note the lack of focal areas of uptake around the periphery of the lesion (safranin O-fast green, 1).

7 59 further extended outward because of bone resorption and continued structural collapse (Fig. 3, E through H). The walls were buttressed by a thickened shell of bone that was undergoing active bone formation (Fig. 3, E and G). Endochondral bone formation was seen in the base of the defect. The defect was filled with fibrocartilage, fibrous tissue, and a central cyst. The tissue filling the defect was still covered by a concave fibrous tissue layer that was contiguous with the surrounding cartilage. By fifty-two weeks, the initial cylindrical geometry of the defect had changed and was characterized by outwardly extending walls, collapse, and osteophyte formation (Fig. 4, A through L). The repair tissue in the lesion was fibrocartilaginous at the periphery. However, most of the defect contained disorganized fibrous tissue surrounding a central cyst. Safranin-O staining demonstrated small amounts of proteoglycan at the bone-fibrocartilage interface (Fig. 4, D, E, F, J, K, and L). The persistent layer of fibrous repair tissue at the surface was much thinner but was still contiguous Fig. 4 Histologic sections at fifty-two weeks after creation of the defect. A through F are from one goat, and G through L are from another goat. A: Low-magnification view showing the large increase in the size of the defect resulting from resorption and collapse of the bone in the walls and base of the defect. Sclerotic bone surrounds the lesion. Repair tissue in the cyst is mostly disorganized, with some fibrocartilage forming at the base and lower walls of the lesion. The aperture of the defect has narrowed greatly (hematoxylin and eosin, 1). B: Higher-magnification view showing the upper edge of the lesion with sclerotic subchondral bone and a thin band of repair tissue integrated with the cartilage edge (hematoxylin and eosin, 10). C: Highermagnification view showing the base of the lesion and the wall with fibrocartilaginous matrix repair tissue. The sclerotic bone has vascular elements adjacent to the repair tissue (hematoxylin and eosin, 10). D: View of the same area depicted in A, showing uptake of safranin O in the original articular cartilage, surrounding articular cartilage, and fibrocartilaginous matrix repair tissue at the periphery of the lesion (safranin O-fast green, 1). E: View of the same area depicted in B, showing uptake of safranin O (safranin O-fast green, 10). F: View of the same area depicted in C, showing uptake of safranin O (safranin O-fast green, 10). G: Low-magnification view of another defect at fifty-two weeks, showing extensive remodeling. The aperture of the lesion is quite narrow compared with the 6-mm aperture of the original lesion (hematoxylin and eosin, 1). H: Higher-magnification view showing the upper edge of the lesion, with sclerotic bone surrounding the defect and very little repair tissue attached to the cartilage surface (hematoxylin and eosin, 10). I: Higher-magnification view showing the interface of the base of the lesion with the fibrocartilaginous matrix repair tissue (hematoxylin and eosin, 10). J: View of the same area depicted in G, showing uptake of safranin O in the original articular cartilage and the fibrocartilaginous matrix repair tissue at the periphery of the lesion (safranin O-fast green, 1). K: View of the same area depicted in H, showing uptake of safranin O (safranin O-fast green, 10). L: View of the same area depicted in I, showing uptake of safranin O (safranin O-fast green, 10).

8 60 THE JOURNAL OF BONE & JOINT SURGERY JBJS.ORG VOLUME 83-A NUMBER 1 JANUARY 2001 Fig. 5 Histologic sections demonstrating common features observed at fifty-two weeks after creation of the defect. A: This section shows increased vascularity in the subchondral bone under the normal articular cartilage close to the lesion (hematoxylin and eosin, 50). B: In this section, the repair tissue filling the cyst consists of loose, disorganized fibrous tissue with sparse cellularity (hematoxylin and eosin, 50). C: In this section, a tartrate-resistant acid phosphatase staining shows osteoclastic activity (dark-staining cells [arrows]) in the bone at the base of the lesion ( 50). D: This section shows active osteoblastic bone formation on the wall and base of the lesion (hematoxylin and eosin, 50). E: This section shows the interface (arrows) between the native articular cartilage and the newly formed repair articular cartilage. The original tidemark is still evident but is absent from the area where the lesion was created. In this specimen, the newly formed articular cartilage is thinner than the original cartilage (hematoxylin and eosin, 50). F: This section of a different specimen also demonstrates the interface (arrow) between the original articular cartilage and the newly formed repair articular cartilage. The original tidemark is still evident but is absent from the area where the lesion was created (hematoxylin and eosin, 50). with the cartilage edges, and it followed the contour of the lesion walls more closely. The surrounding articular cartilage showed degenerative changes consisting of fibrillation, cleft formation, and condylar flattening. The original 6- mm-diameter aperture of the lesion had narrowed to 2-3 mm in three goats and was entirely closed over in one goat. Safranin-O staining was evident up to the edge of the articular cartilage but was absent in the most superficial layer (Fig. 4, E and K). Sclerotic bone surrounded the lesion and adjacent subchondral areas. Sparse bone resorption was observed at the base of the lesion. Features common to all of the specimens examined at one year are shown in Figure 5. These features included a vascular subchondral bone-cartilage interface near the edge of the lesion; disorganized fibrous repair tissue filling the lesion; sparse osteoclastic activity at the base of the lesion; marked endochondral bone formation; repair articular cartilage closely apposed to, but not always well integrated with, the native cartilage edge; and a

9 61 Fig. 6 Gross appearance, serial microradiographs, and histologic sections of three specimens at fifty-two weeks after creation of the defect. A, D, G, and J are from one goat; B, E, H, and K are from a second goat; and C, F, I, and L are from a third goat. D, E, and F are 1-mm sections taken 2 mm from the center of the lesion. G, H, and I are 1-mm sections taken from the center of the lesion. J, K, and L are safranin-o-stained histologic sections from a 1-mm wafer taken from the center of the lesion. A, B, and C show the gross appearance of the lesions at the time of evaluation. Note the variation among the animals in the size of the apertures of the lesions, which ranged from fully closed to 2 to 3 mm in diameter. D through I demonstrate the bone-remodeling changes and the development of the central cyst; these were consistent findings, although some variation among animals was observed. J, K, and L show consistent uptake of safranin O by the articular cartilage but with varying intensity in the repair tissue lining the lesion. The central cyst was consistently filled with repair tissue. poorly formed tidemark (Fig. 5, E and F). The surrounding articular cartilage had irregular surfaces, varied in thickness, and showed signs of degeneration, with fibrillation and clefts. The one-year evaluations showed variations and similarities in the reparative response among different animals (Fig. 6, A through L). On gross examination, the aperture of the lesion still appeared open, with a 1 to 3-mm diameter, or was completely closed by repair tissue. Serial microradiographs demonstrated large central cavitary lesions, but the size and morphology (geometry) of the cavitary

10 62 Fig. 7 Magnetic resonance images (A, B, and C) and a computerized axial tomographic image (D) of an articular cartilage lesion fifty-two weeks after creation of the defect. A: Coronal image demonstrating a clearly demarcated area of high signal intensity (arrow) at the site of the cavitary lesion in the medial femoral condyle. B: Sagittal image of the medial femoral condyle demonstrating an enlarged area of higher signal intensity at the site of the lesion. C: Sagittal image of the defect shown in B, made with fat suppression. D: Coronal computerized axial tomographic image through the center of the lesion, showing changes in the geometry of the lesion and the central cyst. The specimen is the same as that depicted in Fig. 2, F. Note the marked difference in comparison with Fig. 2, A. lesions varied (Figs. 6, D through I). Histologic sections stained with safranin O demonstrated proteoglycan in the native articular cartilage and repair cartilage adjacent to the lesion as well as focal proteoglycan synthesis in the fibrocartilaginous repair tissue at the periphery of the lesion (Fig. 6, J, K, and L). At one year, computerized axial tomography (three specimens) and magnetic resonance imaging (two specimens) consistently showed a large cavitary lesion in the region of the defect (Fig. 7, A through D). Resorption of the subchondral bone resulted in margins that were wider and deeper (an increase of up to 15%) than those of the original defect, with collapse of the articular surface relative to that of the contralateral control. The subchondral bone and its overlying cartilage surrounding the lesion had continued to collapse into the lesion. The collapse of the lesion walls appears to have contributed to the flattening of the condyle in the area of the defect. Discussion n the present study, the osteochondral defects that were cre- in the medial femoral condyle of the goat disrupted the Iated subchondral bone plate and established communication with the underlying marrow space. Over time, these large defects

11 63 demonstrated progressive changes in both the bone and the articular cartilage. In the first week, the initial cylindrical geometry of the osseous walls of the lesion showed evidence of resorption, which continued to progress, resulting in cavitation and a persistent central cyst. Simultaneous endochondral bone formation occurred at one year, apparently in response to the weakening bone structure. However, structural collapse still occurred by twelve weeks and was present at the completion of the study. Sclerotic bone developed around the cavitary lesion, and the opening of the lesion narrowed. The articular cartilage adjacent to the edge of the freshly cut defect quickly deformed, or flowed, into the lesion. This deformation became more pronounced at six months, with collapse of underlying bone and condylar flattening. At one year, repair tissue was seen adjacent to the surrounding cartilage, which varied in thickness, smoothness, integration, and quality. The 6-mm opening of the original defect gradually decreased in diameter. At one year, the large cavitary lesion was filled by fibrous tissue and a large central cyst. The articular cartilage adjacent to the defect was also affected, and it underwent changes in a region referred to as the zone of influence. There was deformation of the articular cartilage in this zone, characterized grossly by flattening and thinning. These changes in the zone of influence progressed throughout the duration of the study. The shape of the affected condyle at one year was characterized by a loss of contour caused by the creation of the defect followed by condylar flattening surrounding the defect. This encompassing zone of influence has been predicted in finite-element models of cartilage damage 12. As the size of a defect reaches a critical diameter beyond which healing is not possible, increased compressive stresses are predicted along its periphery, indicating that there is a mechanical overloading component to the creation of the zone. The bone compartment demonstrated a thickening at the periphery of the walls of these surgically created lesions, narrowing of the diameter of the upper bone margins, and a cystic vacuolization at the subchondral portion of the defect. Although bone has the capability of spontaneously regenerating itself 3, the bone compartment in these defects did not regenerate or repair itself completely. It appears that, following destruction of the subchondral plate in the weight-bearing surface of the medial femoral condyle in association with the size of these defects, there are structural and mechanical changes that substantially alter the bone-repair process. Major defects in the bone compartment remained present at one year, and no functional tissue developed within these surgically created defects. The size and location of the 6-mm-diameter and 6-mmdeep defects used in the present study presented a most challenging repair problem. The defects exceeded a critical size and did not spontaneously achieve complete repair. In contrast, Butnariu-Ephrat et al. reported that smaller, 3-mmdiameter defects underwent complete repair in a goat model 2. Convery et al., in a study of defects in the distal aspect of the femur of horses, reported that a large, 9-mm-diameter lesion did not heal but that a smaller, 3-mm-diameter lesion was fully repaired at approximately three months 11. In the present study, the creation of the lesions in the weight-bearing area of the medial femoral condyles exposed the defects to a changing mechanical environment with high loading conditions 13. The animals were allowed immediate weight-bearing as tolerated. A period of non-weight-bearing and passive motion might have altered the extent of the reparative process and the deleterious effects that were documented. By twelve months, the defects in our study were clearly visible and showed secondary changes that were suggestive of continued degeneration, making complete repair unlikely 9. From previous work with goats, it has become apparent that defects of this size that are created in the medial femoral condyle behave differently from those that are created in the femoral groove 5. Lesions of similar size in the trochlear groove in the same goat and dog models generally heal, with bone filling the base and fibrocartilage filling the area above 5,14. Strategies for articular cartilage repair must consider the size, shape, depth, and location of the lesion. Investigators designing prospective studies involving large animals with large osteochondral lesions that disrupt the subchondral plate may want to consider providing initial protection from the potential detrimental effects of loading in weight-bearing areas during the initial repair process. Comparisons of defect location, shape, size, and rehabilitation are warranted to assess the spontaneous repair and potential deleterious responses following the surgical creation of a lesion. The progressive changes documented in these surgically created defects in the weight-bearing portion of the medial femoral condyle present problems with regard to both the articular cartilage and the bone compartment. If an enhancement of the cartilage reparative process or any regeneration methodology is developed 3,7,15, it should address the zone of influence with its thinning and flattening of the articular cartilage and surrounding condyle, deformation of the articular cartilage into the margins of the defect, regeneration or replacement of the subchondral bone plate, and regeneration of bone within the defect. It is unlikely that a successful regeneration of the overlying articular cartilage will be possible until these secondary changes are addressed in a new treatment technique. Douglas W. Jackson, MD Harold M. Aberman, DVM Timothy M. Simon, PhD Orthopaedic Research Institute, Southern California Center for Sports Medicine, 2760 Atlantic Avenue, Long Beach, CA Peggy A. Lalor, PhD SkeleTech, th Avenue S.E., Room 104, Bothell, WA Although none of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be directed to a research fund, foundation, educational institution, or other nonprofit organization with which one or more of the authors is associated. Funds were received in total or partial support of the research or clinical study presented in this article. The funding sources were the Douglas W. Jackson Orthopaedic Research Trust and an unrestricted research grant from Howmedica.

12 64 1. Breinan HA, Minas T, Hsu H-P, Nehrer S, Sledge CB, Spector M. Effect of cultured autologous chondrocytes on repair of chondral defects in a canine model. J Bone Joint Surg Am. 1997;79: Butnariu-Ephrat M, Robinson D, Mendes DG, Halperin N, Nevo Z. Resurfacing of goat articular cartilage by chondrocytes derived from bone marrow. Clin Orthop. 1996;330: Caplan AI, Elyaderani M, Mochizuki Y, Wakitani S, Goldberg VM. Principles of cartilage repair and regeneration. Clin Orthop. 1997;342: Hangody L, Kárpáti Z, Tóth J, Diószegi Z, Kendik Z, Bély M. Transplantation of osteochondral autografts of the weight-bearing surface of the knee joint and the patellofemoral joint in dogs. Hungarian Rev Sports Med. 1994;35: Huibregste BA, Samuels JA, O Callaghan MW. Development of a cartilage defect model of the knee in the goat for autologous chondrocyte implantation research. Trans Orthop Res Soc. 1999;24: Hurtig MB, Novak K, McPherson R, McFadden S, McGann LE, Muldrew K, Schachar NS. Osteochondral dowel transplantation for repair of focal defects in the knee: an outcome study using an ovine model. Vet Surg. 1998;27: Oka M, Chang Y, Hyon S, Ikada Y, Cha W, Nakamura T. Development of artificial osteo-chondral composite material. In: Transactions of the Fifth World Biomaterials Congress. Toronto: Society for Biomaterials; p 922. References 8. Shapiro F, Koide S, Glimcher MJ. Cell origin and differentiation in the repair of full-thickness defects of articular cartilage. J Bone Joint Surg Am. 1993;75: Buckwalter JA, Mankin HJ. Instructional Course Lecture, American Academy of Orthopaedic Surgeons. Articular cartilage. Part II: Degeneration and osteoarthrosis, repair, regeneration, and transplantation. J Bone Joint Surg Am. 1997;79: Pridie KH. A method of resurfacing osteoarthritic knee joints. In: Proceedings of the British Orthopaedic Association. J Bone Joint Surg Br. 1959;41: Convery FR, Akeson WH, Keown GH. The repair of large osteochondral defects. An experimental study in horses. Clin Orthop. 1972;82: Rosenberg TD, Weiss JA, Moulis PM, Deffner KT, Cooley VJ. Finite element simulation of stresses in chondral defects. Int Cart Repair Soc Newsletter. 1999;2: Braman JP, Crary JL, Norman AG, Clark JM. Acute deformation of experimental drill holes in loaded articular cartilage. Trans Orthop Res Soc. 1999;24: Breinan HA, Minas T, Barone L, Tubo R, Hsu H-P, Shortkroff S, Nehrer S, Sledge CB, Spector M. Histological evaluation of the course of healing of canine articular cartilage defects treated with cultured autologous chondrocytes. Tissue Eng. 1998;4: Lee CR, Spector M. Status of articular cartilage tissue engineering. Curr Opin Orthop. 1998;9:88-93.