CHAPTER 5
SYNOVIUM AND CARTILAGE IN HEALTH AND DISEASE
- SECTION ONE: NORMAL ANATOMY OF SYNOVIUM
SECTION ONE
NORMAL ANATOMY OF SYNOVIUM
H. RALPH SCHUMACHER JR.
Many descriptions of normal synovial membrane have been published.(11) Most studies have been performed in a variety of small animals and show generally consistent findings among species. Detailed reports include guinea pig,(32) monkey,(23) rat,(30) rabbit,(8,16) pig,(22) cat,(18) chick,(13) frog,(15) and dog synovium.(2,26) This section of the chapter is a summary of the various reports; included are those instances in which rare, possibly unique, findings have been suggested for certain animals. The references at the end of the chapter usually identify in which animals the studies were done. There are definite differences in the appearance of synovium in different synovial joints and even in different parts of the same joints of individual animals.(5) Although these differences have been the subject of little study, they need to be considered in evaluating pathologic processes in the synovium.
GROSS APPEARANCE
The synovial surface (when viewed by arthroscopy or at surgery without a tourniquet) appears pale pink, smooth, and shiny with a moistening of synovial fluid. Folds are seen at some places, such as at the insertion of the periosteum into the subchondral bone, but villi are seen only microscopically.
Measurable volumes of synovial fluid cannot be aspirated from truly normal joints of small animals. The few drops obtainable are viscous and clear. Few cells, almost all of which are mononuclear, are present. (See Chapter 86.)
LIGHT MICROSCOPY
The innermost structures of synovium consist of one to three layers of cells generally identified as synovial lining cells (SLC). The number of these cells varies in different parts of the joint(3) and may also appear to vary when the cells are cut tangentially. Normal lining cells are oval or slightly flatter with their long axes parallel to the surface of the synovium. Lining cells are closely spaced but are not continuous. Small processes are present but are generally best appreciated on epon-embedded rather than paraffin sections. Mitotic figures are rare, although studies with Feulgen cytophotometry suggest that lining cells can divide.(12)
No specific feature separates the lining cells from the deeper structures, to which there are various modes of transition. In fibrous synovial membrane, lining cells tend to be few and to lie directly on dense fibrous tissue. In areolar or fatty synovial membrane, more small vessels are appreciated immediately under the SLC and in connective tissue septae extending between fat cells. The surface of areolar synovium tends to be more undulating, while on fibrous areas of synovium the surface is flatter. In fibrous synovium, fibrocytes with elliptic nuclei are seen along with the collagen and occasional elastic fibers.(3) Alcian blue also stains for profuse glycosaminoglycan in the matrix.(3)
Very superficial capillaries and small venules are often seen but may be inconspicuous unless dilated. Lumens can collapse and do so when flow is temporarily shunted away, as normally occurs. Arterioles and larger venules are seen deeper in the synovium. Even in some normal joints there can be thicker endothelium and vessel wall than might be expected merely by the size of the vessel lumen.(6)
Lymphatics are also present but are virtually never detected on routine studies of sections of normal tissue. Mast cells and a few macrophages are seen around vessels. The mast cells can be demonstrated only with special stains such as toluidine blue. They are infrequent in fibrous synovium. Nerves are scarce in the superficial synovium compared with the deep capsule. They can be identified with Gros-Bielschowsky stains(10) Their most commonly suggested role in superficial tissue is autonomic.
 |
FIG. 5-1 (A) Light microscopy shows areolar synovium with villous folds. Small vessels are seen immediately under the synovial lining cell. (SLC) (JS, joint space; V, venules) (original magnification x 200) (B) Fibrous synovium demonstrated by light microscopy. Note less villous surface, deep fibrous tissue instead of loose and fatty tissue, and otherwise similar structures. (original magnification x 200) |
ELECTRON MICROSCOPY
Scanning electron microscopy has aided in the characterization of the structure of the synovial surface. In the canine knee(33)areolar synovium can be confirmed to have an undulating or folded surface. Cytoplasmic processes of lining cells can be identified. In synovium overlying fat pads, a cobblestone appearance due to the fat cells can be seen on the surface. Fibrous synovium has a smoother surface with less evident cell processes.
Transmission electron microscopy(25) shows SLC arranged loosely in a granular and fibrillar matrix (Fig. 5- 2 and 5-3). Tight junctions are not present between lining cells (rare desmosomes have been described in normal mice); hence, the matrix is exposed to the joint space at many places. The most superficial cells often have many thin cytoplasmic processes, prominent vacuoles, dense bodies, and Golgi apparatus. This type of cell has generally been termed a type A lining cell.(1) It is active in phagocytosis of particles such as ferritin, carbon, or proteins(28) that have been experimentally injected either intra-articularly or intravenously.(21,23)
Generally most prominent deep to any type A lining cell are cells with fewer filopodia, abundant rough endoplasmic reticulum, mitochondria, and fewer of the structures typical of a type A cell. These cells have been termed type B. They are probably involved mainly in synthesis for secretion, although they also can phagocytose. Cells with features of both types exist in virtually all animals studied and have been termed type C. Cells typical of type B tend to have more homogeneous nuclear chromatin.(5) It is generally suspected but not yet established that transitions occur between the various types and that cells may have different features in nearby sections of the same cell. All SLC have organelles common to most cells, including mitochondria, microtubules, moderate and occasionally clustered microfilaments measuring 80 A to 100 A in diameter, pinocytic vesicles, glycogen particles, centrioles, occasional lipid droplets, and occasional cilia. Mitotic figures are very rare.
A variety of substances(4) have been identified as products of SLC, as determined mostly from histochemical and tissue culture studies. These include hyaluronic acid, collagenase, collagen, proteases, "connective tissue-activating peptide," nicotinamide adenine dinucleotide (NAD) diaphorase, lactic dehydrogenase (LDH), glucose 6-phosphate dehydrogenase, prostaglandins, plasminogen activator, and fibronectin.(27)Whether type A, B, or C cells are the origin of most of these products is not fully resolved. Hydrolytic enzymes appear to arise in large part from type A cells and are not seen in deeper synovial fibroblasts.(11) Definite type B lining cells are negative for peroxidase and for IgG or C3 receptors, which appears to separate them from monocytes,(l7,31) although there is recent evidence that monocytes are the origin of some of the type A lining cells.
 |
FIG. 5-2 Electron micrograph shows type B synovial lining cell surrounded by some finely granular and fibrillar material in the most superficial synovium. Collagen is seen deep to this cell. There is profuse rough endoplasmic reticulum (RER). Small processes of more superficial lining cells are seen toward the joint space (JS). (N, nucleus with a slightly irregular shape) (original magnification x 18,000) |
 |
FIG. 5-3 Electron micrograph shows type A synovial lining cell on the surface, type B cells deeper among collagen fibers, and superficial vessel (V) with fenestrated endothelium below. (original magnification x 9000) |
 |
FIG. 5-4 Electron micrograph shows fenestrated endothelium (E) of superficial vessel at higher magnification. (RBC, erythrocyte in the lumen) (original magnification x 30,000) |
 |
FIG. 5-5 Electron micrograph shows thicker-walled endothelium of synovial venule with mitochondria (M), scattered ribosomes, and a cluster of microfilaments (arrows). There is an erythrocyte (RBC) in the lumen. (P, pericyte) (original magnification x 42,000) |
 |
FIG. 5-6 Electron micrograph shows synovial mast cell with typical granules and long cytoplasmic processes (arrows). (original magnification x 17,000) |
Dense secretory vesicles associated with the Golgi apparatus of type B lining cells have been described as prominent in mouse and rat synovium but not in synovium of other animals. (19) Small amounts of basement membrane-like material can occasionally be seen adjacent to SLC, but there is no organized basement membrane separating the SLC from the deeper structures.
Matrix around the lining cells and deeper in the synovium consists of finely granular material, thin filaments of 90 A to 200 A diameter, and thicker mature collagen with typical 640-A banding.(20) Ruthenium red has been used to stain matrix proteoglycans and produces dark staining on the synovial surface, throughout the matrix, and on the surface of collagen fibers(14) It has been suggested that the matrix contributes some minor restriction to the transport of solutes,(29) but electrolytes, for example, are present in comparable levels in synovial fluid and serum. Larger particles such as carbon black also migrate readily from vessels through the matrix of rabbit synovium after intravenous injection.(24)
Some capillaries can be seen immediately beneath the SLC. Many of these most superficial vessels have fenestrated endothelial cells,(23,30) with the fenestrations tending, at least in the rat,(30) to be most prominent on the side of the vessel nearest the joint space. Fenestrations in capillaries and venules have a small diaphragm that appears to allow rapid passage of solutes (Fig. 5-4). Similar vessels are also seen in the renal glomerulus, choroid plexus, and intestinal villi, where there is prominent transport of fluid across the vessels. Electrolytes and low-molecular-weight proteins such as albumin seem to pass readily from vessels into the joint space. Thus it is suspected that these fenestrated vessels contribute most of the fluid for formation of the synovial fluid and that the lining cells add the hyaluronase protein.
Deeper capillaries and venules have dramatically thicker endothelium without fenestrations (Fig. 5-5). This endothelium is rich in organelles including mitochondria, dense bodies (lysosomes), vesicles, filaments,microtubules, rough endoplasmic reticulum, multivesicular bodies, and rodlike WeibelPalade bodies that are seen typically in endothelium. Intercellular junctions can open to allow escape of circulating particles, neutrophils during inflammation, or experimentally injected particles such as carbon black.(23) Even normal endothelium has prominent flaps and processes extending into the lumen and can phagocytize circulating particles. Both thick- and thin-walled vessels are surrounded by a single layer of basement membrane. Similar basement membrane also envelops pericytes of the larger vessels.
 |
FIG. 5-7 Electron micrograph shows dog synovial cell with unexplained crystalline array (arrow). Similar inclusions were seen in increased numbers in rheumatoid arthritis. (original magnification x 48,000) |
Arterioles are seen less commonly and are identified by their elastic and muscular walls.
Rare Iymphatics can be identified by the absence of intraluminal erythrocytes and by thin endothelium often closely approximated to the surrounding collagen.
Unmyelinated nerves can occasionally be seen in superficial synovium and are usually adjacent to small vessels. In the fibrous capsule, myelinated nerves and Ruffini corpuscles can be seen. (9)> Mast cells (Fig. 5-6) are also seen in the perivascular area and have characteristic laminated material in their granules. Fat cells and fibrocytes with surrounding collagen are seen in varying proportions in the various areas of synovium. Collagen in the synovial stroma is identified as predominantly types I and III. (7)
A few deep macrophages are seen even in normal synovium and may contain granular or other material in vacuoles. Unexplained crystalline arrays of 20 nm to 25 nm tubules have been seen in some normal dog synovial cells as well as in diseased tissue (Fig. 5-7).(26)
SECTION TWO
TRAUMA AND THE HEALING PROCESS OF SYNOVIUM
H. RALPH SCHUMACHER JR
RESPONSES TO EXPERIMENTAL BLUNT TRAUMA
A recent study (69)examined the sequential effects of blunt external trauma on the synovial membrane of dogs in an attempt to elucidate the process involved in the clear or hemorrhagic joint effusions that have been said to occur after moderate trauma not associated with fracture.(56,62) Anesthetized mongrel dogs weighing 15 kg to 26 kg were studied after metal weights of 245 g to 564 g were repetitively dropped from a height of 25 cm onto the lateral aspect of the right knee over a period of 1 to 4 days.
Carbon black was injected intravenously 10 minutes prior to the last trauma in order to better study extravasation. Joints were aspirated and dogs sacrificed 30 to 100 minutes after the last trauma. Synovial fluid volume was increased in the traumatized knees. Fluids of eight dogs were clear yellow and two were bloody. There were few nucleated cells, virtually all of which were synovial lining cells (SLC). Fat droplets stained with oil red O were found in all traumatized joints but not in controls.
Gross examination revealed small hematomas in the subcutaneous tissue at the site of trauma. The synovium appeared pink and more hyperemic than that in the control knees but was not grossly hemorrhagic. Lines of black discoloration from the carbon black were seen only in the traumatized knees. Dissection of the adjacent structures showed no fractures or ligamentous tears.
Light microscopy identified significantly more vascular congestion in the traumatized knees but no differences in lining cell proliferation between acutely traumatized and control joints. Neither fibrin nor inflammatory cell infiltrates were found.
Electron microscopy showed no detectable differences between control and traumatized synovium as far as types of lining cells or connective tissue morphology. Gaps between endothelial cells were identified in only one synovium (a traumatized one), while extravascular carbon black particles were prominent in three traumatized sides but no controls. The extravasated carbon was identified in venule walls (Fig. 5-8), pericytes, macrophages, and interstitially even in synovial tissue on the side of the joint away from the administered trauma. Fat droplets were seen in vessel walls, interstitium, and lining cells only in the traumatized knees. No bone or cartilage fragments nor calcium apatite crystals were found.
 |
FIG. 5-8 Electron micrograph shows lipid droplet and dark carbon that had been injected intravenously in vessel wall of synovium distant from site of blunt trauma. (E, endothelium, L, lipid; P, pericyte) (original magnification x 16,000) |
Thus, this blunt trauma produced what appeared to be largely functional changes with diffuse increased vascular permeability. A demonstrable inflammatory cell response was not elicited. The release of lipid, which apparently had its origin in the adipose synovium,(44,52) has the potential for causing an inflammatory response as shown in studies in other dogs.69 No follow-up studies have been performed to see if the transudation and lipid release produce any later morphologic changes.
ACUTE AND CHRONIC HUMAN TRAUMATIC ARTHRITIS
Humans experiencing similar acute trauma have occasionally been studied by needle synovial biopsies and have shown only vasodilation, edema, and slight focal increases in SLC.(64) Long-standing traumatic arthritis complicated by damage to cartilage and menisci shows more striking morphologic changes. Light microscopic findings include mild proliferation of SLC,(58,66) some Iymphocytic infiltration, and vascular sclerosis. Electron microscopy shows marked increases and some dilation of rough endoplasmic reticulum of synovial cells, an increase in dense bodies, and an increase in lipid droplets. The contributions of on-going trauma due to the cartilage damage and the healing process are impossible to separate in these clinical studies.
SYNOVIAL REGENERATION AFTER SURGICAL OR CHEMICAL SYNOVECTOMY
Studies of the healing process that occurs after subtotal surgical synovectomy have been performed in normal rabbits,(34,45,48) rabbits with experimental immune synovitis,(38) and in humans with rheumatoid arthritis.(49,51,53) The effects of chemical synovectomy of rabbit knees have also been examined.(36)
In the studies on normal rabbits, surgical synovectomies were performed, wounds were closed and dressed, and the animals allowed activity as tolerated. Animals were sacrificed for study from 3 to 25 days in the various reports. Light microscopic studies have shown that synovial membrane with a typical lining layer regrows within approximately 30 to 60 days. During the first 3 days, fibrin and cell debris predominated. A few macrophages and polymorphonuclear leukocytes (PMNs) were also seen. By the fourth day fibrocytes had aligned roughly with the surface.(34) At 25 and 50 days it was believed that these same cells had become surface lining cells. Regeneration specifically did not seem to come from residual lining cells at the margin of the excised synovium.(34) There was hyperemia below, and tissue was still more fibrous than normal. At 85 days synovium looked normal except for some scarred areas.
By electron microscopy, elongated cells with profuse rough endoplasmic reticulum and uniform chromatin in their nuclei, appearing to be fibroblasts, had largely lined the joint cavity.(43) A few phagocytic cells were seen throughout the tissue and seemed to congregate deep to the fibrocytes, where they were believed to form the type A lining cells. Both types of cells populating the joint surface had many nucleoli. Throughout the first 5 weeks cells on the surface gradually became less flattened, and cell processes extended toward the joint space at 35 days when they completely lined the tissue surface. Synovial vessels were not discussed in these reports.(48)
In joint disease due to experimental arthritis or rheumatoid arthritis, synovium regenerated with features of the antecedent inflammatory disease. This was seen within 4 weeks in rabbit arthritis.(38)
Chemical synovectomy with nitrogen mustard(36) was followed by widespread necrosis not only of synovium but of superficial cartilage. In this experiment it was thought that repopulation of surface cells originated in osteoblasts at the junction of bone with synovium, although origin from deeper connective tissue did not appear to be excluded. Cellular covering with fibroblast-like cells, over and within the surface fibrin layer, antedated revascularization. Some areas that did not revascularize underwent cartilaginous metaplasia at about 6 weeks. Increased numbers of lining cells, up to three to four layers thick, were seen in some areas at 6 weeks. Occasional deep giant cells were seen.
By electron microscopy the cells invading through the surface fibrin to form the new lining were elongated and rich in rough endoplasmic reticulum and Golgi apparatus. They showed large mitochondria, as seen in wound healing elsewhere,(57)lipid deposits, and evidence of phagocytosis of debris.
Thin extracellular filaments were seen among the surface cells at 4 weeks, and mature collagen was evident at 6 weeks. Synovial fibroblastlike cells at this stage showed dilated rough endoplasmic reticulum and increasingly conspicuous cytoplasmic microfilaments and microtubules as seen in so-called myofibroblasts typical of wound healing.(40-42) Cilia were visible only late in synovial regeneration. In some joints regenerated capillaries seen at 3 months had thick walls, while others had thin walls and fenestrations typical of normal synovium.
Thus, what appears to be normal synovium regenerates in normal animals after subtotal surgical synovectomy or more complete chemical necrosis of the superficial synovium. Evidence to date suggests that regeneration of the typical lining surface comes from deep mesenchymal cells, not from residual lining cells or circulating blood cells. All properties examined so far seem to have returned to normal, at least in most parts of the tissue. That is, collagen and mucopolysaccharides appear,(36) phagocytosis occurs in the new lining, and adenosine triphosphatase (ATPase) is demonstrated as in normal synovium.(36)
The effect on the articular cartilage of temporary loss of synovium is not clear. In normal rabbits cartilage fibrillation seemed to be increased by synovectomy.(38) In animals in whom chemical synovectomy was performed, there was some direct injurious effect of the nitrogen mustard on the cartilage.
Interestingly, joints were not immobilized after any of these synovectomies in animals. What, if any, difference casting would make is not known. The administration of hydrocortisone acetate, 25 mg/week intramuscularly, was accompanied by some delay in regeneration of synovium in one study.(34)
SYNOVIAL HEALING AFTER INCISION
Wound and incision healing of the synovium has been studied only by Levene(41) in 1957 and in a recent project on eight dogs that my colleagues and I have done in our own laboratory.(46) In the latter project, sutured incisions as well as unsutured defects 1 cm' were studied in dogs who had been normal prior to surgery or in whom the cruciate ligaments had been sectioned 28 days before, leading to early osteoarthritis. Carbon black was injected intravenously to study vascular leakage.
In previously normal dogs, exudates containing neutrophils, erythrocytes, and fibrin filled the sutured incisions during the first 3 days (Fig. 5-9). Carbon leaked from vessels around the incisions. By 7 days fibroblastlike cells were prominent at the base and margin; these cells covered or nearly covered the surface by 10 days (Fig. 5- 10), at which time wounds also grossly appeared healed. Some necrotic cells and fibrin were walled off more deeply. There were scattered Iymphocytes and plasma cells.
At 10 to 14 days fibroblastlike cells with large pale nuclei increased on the surface and penetrated into the deeper tissue. These cells became rounder at 21 to 28 days. Proliferation of deep round cells was seen also (Fig. 5-10, C), with some suspected Iymphocytes or monocytes adjacent to the new surface cells. There was proliferation of remaining lining cells away from the wound with 1- 4 layers of cells.
In dogs with open wounds 1 cm,(2) virtually the same sequence was seen, and healing was no slower. In dogs with osteoarthritis due to previously sectioned cruciates, some SLC proliferation was present at the time of incision or wound. Following the incision, inflammation and edema were more striking than in the previous group, and wounds actually closed slightly more rapidly.
By electron microscopy, phagocytosis of cell debris, lipid, and granular material was prominent from day 1 through day 14 (Figs. 5-11 and 5-12). Some collagen was seen to be enfolded by predominantly fibroblastlike cells (Figs. 5-13). Many cells with predominantly fibroblastic characteristics had a few dense bodies.
 |
FIG. 5-9 Light micrograph shows mass of fibrin (F), neutrophils, and necrotic material filling an incision 1 day after surgery. (J joint space) (original magnification x 180) |
There were signs of activation of fibroblastlike cells containing prominent nuclei with sparse heterochromatin and nucleoli, and by the third day there were increasing amounts of dilated rough endoplastic reticulum, polyribosomes, and mitochondria (Fig. 5-13, B). At 14 days the fibroblastlike cells were clearly the predominant structure. Microtubules and microfilaments were seen but were not clearly increased. Some cells had cilia.
Many fibroblastlike cells were closely approximated to each other (Fig. 5-14) and to phagocytic cells with an occasional desmosome. From the first day on, coated pits and structures resembling hemidesmosomes were seen on a variety of mononuclear cells (Fig. 5-11). Nuclei were slightly irregular (see Fig. 5-13, A) but never extremely convoluted.
At 7 to 14 days, granular material and thin filaments (Fig. 5-14) of suspected immature collagen surrounded superficial cells. Increasing amounts of mature collagen were seen.
From day 3 on, vessels (see Fig. 5-12) adjacent to the wound had thick walls with prominent endothelial cells rich in organelles, including microfilaments and rough endoplasmic reticulum. Junctions between endothelial cells were unusually complex, endothelial processes extended into the lumen, and gaps were seen in some specimens with carbon Iying in the vessel wall. By day 7 vascular basement membranes were multilaminated. Some thrombosed vessels were seen. At day 14 there were some thin-walled suspected capillary sprouts.
 |
FIG.5-10 Lightmicrograph: fibroblastlike cells cover the surface in these specimens at 7 days (A) and 10 days (B) and have walled off some fibrin and cell debris below. Rounder mononuclear cells are mainly deep in the tissue and do not clearly make any major contribution to the surface closure. At 21 days (C) hyperplastic deep round mononuclears are seen and some rounder cells of unclear origin are near the surface. (original magnification x 400) |
From 21 to 56 days the superficial synovium showed increasingly mature collagen with fibrocytelike cells containing profuse rough endoplasmic reticulum that was no longer dilated (Fig. 5-15). There were small amounts of phagocytized debris in vacuoles. Vascular endothelium was still thick; inflammatory and purely phagocytic cells were not seen.
Few ultrastructural differences were seen between the animals with the sutured incisions and those with open wounds. The dogs with the sectioned cruciates had more thickened vessels with a greater number of gaps and more carbon leakage.
Thus, wounds in synovium healed rapidly in all dogs studied; the synovial surface was restored or nearly restored in all animals by 10 days. Inflammation was surprisingly minimal. The dogs in whom the cruciates had been sectioned prior to creation of the new wound had a somewhat more proliferative and inflammatory reaction in what appeared to be more vascular and edematous tissue. Wounds actually closed slightly more rapidly than in the normal dogs. Some residual vessel abnormalities as a result of the previous surgery or the osteoarthritis in the abnormal dogs may have contributed to these differences.
 |
FIG. 5-11 Electron micrograph shows cell rich in rough endoplasmic reticulum at the margin of the incision after 1 day. Note lipid (L), a vacuole with unidentified material, coated pit (arrow), and hemi- desmosomelike structure (double arrows). (F,fibrin)(original magnification x 46,000) |
 |
FIG. 5-12 Electron micrograph shows venule adjacent to synovial incision at 3 days. Note large endothelial cells with many cell processes and rich in organelles, including rough endoplasmic reticulum and dense bodies. Intercellular junctions are long and tortuous. One endothelial nucleus (N) contains two nucleoli. (L Lumen) (original magnification x 8000) |
 |
FIG. 5-13 (A) Electron micrograph, 7 days after incision, shows fibroblast in connective tissue below wound with phagocytized (or enfolded) collagen fibers (arrow), only moderate rough endoplasmic reticulum in this section, many mitochondria, and convoluted nucleus (N) with a nucleolus. (original magnification x 10,000) (B) Electron micrograph shows fibroblastlike cells (F) extending into superficial fibrin and debris at 7 days. Note also one Iymphocyte (LY). (original magnification x 13,000) |
 |
FIG. 5-14 Electron micrograph shows interdigitation of fibroblast (F) and another less easily identifiable cell. Thin filaments surround the fibroblast. (C, collagen) (original magnification x 20,000) |
 |
FIG. 5-15. Electron micrograph shows possible capillary sprout (C) with incomplete layer of endothelium and basement membrane (BM) surrounding an empty lumen. Cells on the surface at 21 days have fibroblastlike features with occasional dense bodies. (original magnification x 8000) |
Compared with the healing of skin wounds,(40,55,57) synovial wound healing shows a similar presence of fibrin and neutrophils during the first days both in our dogs and in previous studies in rats.(47) In synovium PMNs were still visible at 3 days, whereas they had largely disappeared from skin wounds healing by primary intention.(55) In synovium, fibroblastlike cells covered the surface but left necrotic debris and fibrin deep to the surface even at 7, 10, and 14 days after suturing. This was not the case with primary skin healing, which was complete by 7 to 14 days. The synovial surface did appear healed by 10 to 14 days, which was only slightly slower than the skin healing. Levene, in his studies on synovial healing in rats, also suggested that healing was slower than in the skin but apparently found complete synovial surface integrity by 7 days.(47) In the dogs, the skin incisions, although not examined histologically, were fully healed by 7 days.
Studies to date have not shown dramatic evidence of contracting fibroblasts or myofibroblasts(41) with vast amounts of microfilaments, folded nuclei, or junctions between fibroblasts as described by Gabbiani and co-workers(40,42) in granulation tissue fibroblasts. Each of these changes was seen to a mild degree. Because some such findings are also seen in normal synovium, one may speculate that some chronic trauma occurs with joint motion, leading to some connective tissue response at all times. Normal type B SLC, although resembling fibro- blasts in many ways, seem to have more microfilaments of 80 A to 100 A (50,63) than do resting fibroblasts.(41) Occasional "attachment sites"(54) were seen among the micro-filaments in our synovial wounds; we have not observed these in normal synovium. They have been suggested to be involved in cell contraction, as in smooth muscle.(4l) The hemidesmosomes seen in our healing synovium may be important in some way in wound contraction, since they have also been found in granulation tissue.(41) Desmosomes seen in skin wounds(40) are rarely seen in synovium(63) and have not yet been seen in these dogs. Structures similar to the coated pits seen in our healing synovium have also been found as part of cell-to-cell interaction during development and inflammation(43); their role in these dogs remains to be determined. Other studies of small wounds in other tissues have not shown such contracting fibroblasts.(55) It may be that these do not occur to the same degree in relatively small wounds. We did not see any definite differences between open wounds and those that were sutured.
Wound closure is probably much more complex than can be appreciated by these plain morphologic studies. For example, the role of myosin and other intracellular proteins that are not part of the microfilaments in fibroblasts is just beginning to be studied.(71) Components of the extracellular finely granular and fibrillar material seen in our healing wounds have not been fully identified. Fibronectin, which is known to be produced by fibroblasts,(37) may be an important component along with immature collagen. Presumed serum proteins, platelets,(35) and cell debris, especially in the early stages in the interstitium, are probably factors in favoring the early migration of fibroblasts.(65) Increased contribution of these substances might explain the more rapid healing seen in the dogs with previously sectioned cruciates. It has been suggested that cell mitosis is not needed for fibroblast migration(68); in our wounds we have not identified mitoses. Levene, in his light microscopic study of rat synovial healing, did describe mitoses in fibroblasts but not in lining cells.(47)
Capillary proliferation has been well documented in wounds of other tissues. We clearly demonstrated thickening of endothelium with increased organelles, including prominent rough endoplasmic reticulum as described by others,(39) but did not identify definite pseudopods of vessels actually growing into new areas, as has been demonstrated with tumor implants. Occasional suspected capillary sprouts with incomplete endothelial linings were seen up to 21 days. Such small vessels did not have pericytes. Gaps and leakage of carbon persisted in our wounds up to 14 and probably 21 days. Proliferating vessels in other wounds have been shown to be similarly leaky.(39,59) Whether vessels remain leaky for a longer time in joints post trauma as a result of the difficulty in fully protecting the joints is not clear. Some intravenously injected carbon tracer appears to leak from normal synovial vessels(60,61) more often than from vessels in parenchymal organs. Whatever factors stimulate vascular proliferation must subside or be inhibited, since proliferation subsides as the wound heals, leaving mostly fibrous tissue at the synovial wound as elsewhere.
Our light microscopic studies and those of Levene(47) clearly suggest that the cells repopulating the synovial surface are fibroblastic in origin and seem to be migrating in from the margins. Electron microscopic findings are consistent with this, since most surface cells are rich in rough endoplasmic reticulum and seem to create a layer on the surface without prominent deep connections. Some surface cells, like those in the normal state, have varying amounts of vacuoles and evidence of phagocytosis; hence, we cannot totally exclude a monocyte origin. Study of these cells with electron microscopy, histochemistry, and monoclonal antibodies against monocytes will be needed to fully resolve the question of origin. Ross and Odland(57) have suggested that monocyte interaction with fibrocytes is important in skin wounds. In their study cells were suspected of being monocytes on the basis of routine ultrastructure. Some such close association of various cells was seen in our specimens but without desmosomes or other junctions.
Cilia were prominent in cells in healing wounds but were seen also in normal SLC.(64) Their possible roles have not been investigated. Cilia were also prominent in synovium regenerating after chemical synovectomy36 and in fibroblasts grown in tissue culture.(67,70)
SECTION THREE
THE ARTICULAR CARTILAGE
HENRY J. MANKIN
The hyaline articular cartilages constitute a unique and extraordinary body tissue (or perhaps more appropriately, "organ") the structural, biochemical, and metabolic characteristics of which endow the mammalian diarthrodial joint with a remarkable resiliency and almost frictionless movement. Historically, the articular cartilages drew the attention of the earliest dissectors, who described their gross characteristics long before the modern era of scientific investigation. With the advent of the microscope, the histologic characteristics of cartilage were studied, and before the turn of the 20th century, investigators had described the sparse cellularity, inert appearance, and peculiar staining characteristics of the cartilage matrix. The tissue was known to be avascular and aneural and to have limited capacity for repair prior to 1920, and, because of the homogeneous staining and the effete appearance of the cells, cartilage was considered to be metabolically inert.
Over the last 3 decades, technologic advances have allowed an increasing number of scientists to study the structure, chemistry, and metabolism of the articular cartilages, and a large body of information has accumulated that clearly defines not only the unique structural arrangement of the tissue and the complexity of the polydisperse matrix components but a surprisingly active set of metabolic processes.
STRUCTURE
Inspection of articular cartilage from joints of a young adult mammal shows the surface to be smooth, shiny, and dense white (Fig. 5-16). On palpation, the tissue is semisolid and elastic in the sense that it indents with pressure and is restored to its original contour upon release. The color of immature articular cartilage is somewhat bluish (presumably because of the underlying blood vessels); in aged animals the cartilage takes on a yellow cast, the nature of which is not well understood. The yellowish color is apparently not related to degenerative disease but appears to be a natural consequence of the aging process.(l50) A search for increased lipid or other pigmentary materials that could account for the yellowish coloration has led to inconclusive results.(208,287)
 |
FIG. 5-16 Gross specimen of a young adult human distal femur and patella obtained at autopsy. The cartilage is dense white, glistening, and the surface appears smooth. |
The thickness of cartilage varies from joint to joint, site to site within each joint, and from species to species.(75,207,266) As a general rule, the thickness correlates well with the size of the animal, although humans appear to have thicker cartilages than many species that are considerably larger in bulk. The thickness in humans ranges from approximately 0.5 mm to over 5 mm on the patella. Cartilages are firmly fixed to the underlying subchondral bone and are very resistant to shearing stress. Although the surface of articular cartilage looks smooth grossly, histologic sections, incident light microscopy, and, more recently, scanning microscopic studies show that the contours are irregular (Fig. 5-17) (96,102,132,133,180,292) Investigators have demonstrated shallow pits and raised mounds that appear to vary with the age and species of the animal, as well as irregular undulations and furrows in certain regions of the cartilage.(131,134,135,137,2l5)
 |
FIG. 5-17 Scanning electron microscopic study of the surface of adult articular cartilage shows the pits and mounds that generally conform in size (30um-50um) to the elongated cells of the gliding zone (original magnification x 300) (Courtesy of Ian Clarke, PhD) |
The system of lubrication of articular cartilage has been the subject of extensive study over the past 20 years. Important in the lubrication is the presence of a thin layer of proteinaceous material adherent to the surface;(239,277) this material serves as a boundary sub- stance, which when augmented by water (which is believed to "weep" from the cartilaginous surface(l82,183) contacts a similar layer of proteinaceous material on the adjacent cartilaginous surface. Thus, in circumstances of normal joint function, cartilage does not touch cartilage; rather, a boundary layer of hydrated proteinaceous molecules interfaces with a similar such material on the opposing surface.(183,278,291) The greater the axial load on the cartilage, the more water weeps from the depth of the tissue, thus theoretically increasing (but at least maintaining) the efficiency of the lubricating system, despite heavy loads. The coefficient of friction has been calculated for articular cartilage and is extremely low as compared with that for man-made machinery (at least an order of magnitude better than most ball bearing systems). (175,278,291) The presence of the boundary substance explains why there is limited attritional loss from the articular cartilage and also explains the necessity for a high concentration of unbound water (see below).
Examination of a histologic section of adult articular cartilage shows the cells to be quite sparse (Fig. 5-18),(178,271-273) and calculations of cell density show the value to be low (adult rabbits have fewer than 2 x 105 cells/mm3 of cartilage(194). The value varies from species to species, joint to joint, and site to site, but even in immature animals it rarely rises above 3 x 105 cells/ mm3.(194) The majority of the cartilage is in the form of a hyperhydrated proteinaceous matrix that provides the structural semirigidity necessary for its function as a weight-bearing and gliding surface.
Despite the appearance of random distribution, careful study of the cells of articular cartilage shows considerable heterogeneity of cell type and also of arrangement of the cells within the cartilage matrix (Fig. 5-19) (78,103,141,162,273) By custom, the cartilaginous surface is divided into four more or less distinct zones. The most superficial of these, Iying just subjacent to the surface, is called the gliding or tangential zone. Because this zone does not take a stain well, it was called the lamina splendens by early microscopists. Electron microscopic study of this region shows that it consists of a fine-fibered filamentous feltwork that presumably serves as the material to which the boundary protein adheres.(76,295) Immediately subjacent to this region is the "skin" of articular cartilage in which electron microscopic studies have shown bundles of collagen running parallel or tangential to the surface but in various planes (Fig. 5-20) (l65,181,208,227,265,269,295) The cells of the tangential zone are elongated, either spindle or oval in shape, and show small terminal processes and other characteristics suggestive of fibroblasts on electron microscopic studies (Fig. 5-21).(107,229,295) Studies of the chemical composition of the matrix in this region show this portion of the cartilage to be rich in collagen and relatively poor in the other major component, proteoglycan (see below). (204,205,295)
 |
FIG. 5-18 Low-power photomicrograph of the articular cartilage covering the distal femur of an adult rabbit Note the sparse cellularity and heterogeneity of cell population and distribution. The calcified zone is separated from the radial zone by a "tidemark," and the mature cortical bone of the bony end-plate can be seen contiguous with the calcified cartilage. (H&E, x 100) |
 |
FIG. 5-19 Graphic representation of the vanous zones of adult articular cartilage. Note the lamina splendens, a layer of fine filaments at the very surface of the gliding zone, in which the cells are elongated and tangentially arranged. ln the transitional zone the cells are rounded or ovoid and appear to have a random distribution, while cells on the radial zone are arranged in short columns. The "tidemark" separates the radial from the calcified zone, in which the cells are heavily encrusted with apatitic salts. The mature cortical bone of the end-plate shows haversian systems directed parallel to the joint surface (rather than along the long axis of the bone, as is seen in the shaft). |
 |
FIG. 5-20 Electron micrograph of the surface of adult articular cartilage shows the fine fibers and filaments of the lamina splendens and the parallel bundles of collagen of the tangential zone. Note that the bundles of collagen run parallel to the surface and form a "skin." (original magnification x 2l,000) (Courtesy of Charles Weiss, MD) |
Beneath the tangential layer is a broader area labeled the transitional zone in which the cells are oval or rounded, quite small, and inert-appearing on light microscopy (Fig. 5-21). On hematoxylin and eosin staining the cells are thought to lie in lacunae (an artifact of this particular staining technique), and with most light microscopic techniques they shrink away from the matrix margins. Light microscopy shows the matrix to be homogeneously stained, somewhat eosinophilic with hematoxylin and eosin and bright red with safranin-O stain; there is increased concentration of the latter immediately surrounding the cells (territorial zone) and lesser staining in the regions between the cells (interterritorial zone). Electron microscopic studies of the transitional zone show that the chondrocytes have all the characteristics of cells actively engaged in protein synthesis, with a well-developed rough-surfaced endoplasmic reticulum, a well-defined Golgi apparatus, mitochondria, secretory vacuoles, and intracellular filamentous structures (Fig. 5- 22).(109,251,295) Most of the cells have a single cilium and numerous footlike processes.(272,296) Careful examination of electron micrographs shows no evidence of a true "lacuna" but instead demonstrates a region immediately surrounding the cell (the moat), which on histochemical and immunochemical studies has been shown to be very rich in water and proteoglycan and poor in collagen (91,97,133,227,273) The collagen fibers immediately around the moat region are fine and may be of a different type than the collagen in the interterritorial zones (see below). (165) The collagen fibers in the transitional zone appear to be randomly arranged, although scanning studies have suggested the possibility that they align at angles of 45¡ and 135¡ to one another (Fig. 5-23).(269) The collagen fibers vary considerably in size in this region but are approximately 60 nm in diameter and show the characteristic 64-nm periodicity. (165,208 295) Fine fibers fill the interfibrillar space and often appear to be attached to the collagen fibers. Special studies with ruthenium red and other dyes disclose that proteoglycan attaches closely to the fibers in relatively large quantities at intervals along the collagen chain (l63,263,264)
 |
FIG. 5-21 High-power photomicrograph of gliding and transitional zones of articular cartilage shows the elongated fibroblast-like cells of the gliding zone and the rounded, plump cells of the transitional zone. Cells of the transitional zone appear to be lacunae. (H&E, x 400) |
 |
FIG. 5-22 Electron micrograph of a cell of the transitional zone. The cell has a well-defined nuclear membrane, well-developed endoplasmic reticulum, Golgi apparatus, and many vacuoles. original magnification x 22,000) (Courtesy of Charles Weiss, MD) |
 |
FIG. 5-23 Electron micrograph of the collagen fibers of the transitional zone of articular cartilage shows the random orientation of the fibers. Note the 64-nm periodic binding evident on many of the fibers. Most fibers are about 6 nm in width but may become larger with advancing age. (original magnification x 30,000) |
The zone immediately subjacent to the transitional zone is called the radial zone and is characterized by cells lined up in short irregular columns (approximately 5 to 7 cells per column; Fig. 5-24). The cells are smaller and appear less active than those of the transitional zone (although they have the same characteristic organellar structure, cilia, and footlike processes) and have been shown to have in association with them matrix vesicles containing calcific material. The collagen fibers between the cell columns lie in large dense bundles perpendicular to the surface of the cartilage and penetrate through to the calcified zone below.(77,139)
The calcified zone lies immediately subjacent to the radial zone and is characterized by heavy incrustation of the matrix and of the cells with hydroxyapatite salts.(139,166,150,274) Many of the cells of this region show organellar disintegration suggestive of cell death.(251,272,273,295) Studies of the viability of articular cartilage using tritiated cytidine (as a marker for ribonucleic acid [RNA] metabolism) show no incorporation in this cell layer, suggesting that the cells are in fact limited in their capacity to carry out synthetic activities.(189) The calcified zone varies in thickness from species to species and site to site, but the greatest variation is the result of age.(166) A recent study by Lane and co-workers(167) suggested that active remodeling is occurring in this region. The cartilage at the deepest portion of the calcified zone merges with the cortex of the underlying subchondral bony end-plate, which demonstrates the characteristic haversian systems of cortical bone. These systems differ from similar systems in the cortex in only one way: they are directed in the transverse plane parallel to the surface of the joint, rather than along the long axis of the bone.
 |
FIG. 5-24 The radial zone, tidemark, and calcified zone of adult articular cartilage The cells of the radial zone line up in short irregular columns. The tidemark may consist of several lines seen on H&E-stained sections. Cells of the calcified zone appear to be inert on light microscopy and show organellar degeneration on electron microscopy. (H&E, x 400) |
A region of considerable interest to cartilage microscopists is the "tidemark," a thin, wavy, bluish line on hematoxylin and eosin staining that separates the radial from the calcified zones (Fig. 5-24) (l03,112,l26,l39) As the animal ages, frequent reduplications are noted, with as many as seven or more tidemarks seen at some sites. (139,167) No local increase in proteoglycan or other materials that might account for the bluish color (not seen with some other stains) can be discerned, and the nature of the tidemark is obscure.(139) It has been postulated by Redler and co-workers(240) that the line represents a change in direction of the collagen fibers as they pass from the radial zone to become embedded in the calcified zone. Such an alteration in structure might be protective against shearing stress, and the tidemark region may serve as a tethering site for the perpendicularly oriented collagen fibers as they become anchored in the calcified zone.
NUTRITION
When considering the structure of articular cartilage, it should be noted that there are no blood vessels, lymphatic channels, or neural elements that enter or pass through adult articular cartilage. (78) Furthermore, the chondrocytes are separated from the blood vessels of the underlying bone by a zone of dense calcification and the mature cortex of the underlying subchondral bony end-plate. Numerous studies have been performed to assess whether diffusion from the underlying bone can provide nourishment to the cartilaginous surface, and it is now well established that in the adult animal transport of nutrients by this route does not occur.(92,l93,201,203,224,275,290) Conversely, studies in which nutrients have been injected into the joint cavity have demonstrated rapid transport through the "skin" of cartilage and the matrix to the cells, suggesting strongly that the major source of nutrients for the tissue is synovial fluid. Since the synovial fluid is in itself an ultafiltrate of plasma (certain large proteins are excluded),(243) it is apparent that chondrocytes receive their nutrition through a double diffusion system. Nutrients must first diffuse across the synovial barrier into the synovial fluid and then across the matrix of articular cartilage to reach the cell. Maroudas(197,198,200) has demonstrated that the matrix of articular cartilage is not freely permeable and that diffusion of nutrients is in large measure dependent on size and charge. The data from her studies suggest that most of articular cartilage will not admit molecules larger than a hemoglobin molecule, although some recent studies have suggested that there may be "passages" for larger molecules such as albumin.(200)
In light of the foregoing, it is apparent that the tissue lives in "isolation." The transport of humoral messages, such as are given to other organs by rapid transit of information peptides and proteins, may not be possible in cartilage because of the double diffusion barrier; or if transport does occur, it may be considerably slower than in other more vascular tissues. Furthermore, neural impulses that regulate many of the body processes cannot provide information to cartilage because there is no nerve supply; and all of the immune information that is transmitted either by immune globulins or Iymphocytes is likely to be excluded on the basis of the size of the messenger proteins or cell. In theory at least, the chondrocytes can receive only limited information regarding the rest of the body state by the standard neural, Iymphatic, or humoral pathways. On the other hand, if the cell is pressure sensitive (and there is some evidence to suggest that this is the case(222) it may in fact derive considerable information by alterations in the local matrix caused by loading, unloading, or movement. Since articular cartilage is the principal participant in a structure designed to move and bear load, it would seem logical that these two activities might be the means by which the cartilage receives messages to alter synthetic activity. Clearly there is considerable work to be done in this area, and a number of centers are carrying out research to establish the relationship between chondrocyte activities and alterations in the physical state of the joint.
CHEMISTRY
Considering the sparse cellularity of the articular cartilages, it is apparent that for the most part the chemistry of articular cartilages represents an analysis of the composition of the matrix. Cartilage is a hyperhydrated tissue, with values for water content ranging from 68% to almost 80% by weight (97,184,202,208,288) The remaining 20% to 30% of wet weight in tissue is mostly in the form of two macromolecular materials: collagen, which makes up approximately 60% of the dry weight; and a combination of a linear protein and polysaccharides, known as proteoglycan, which accounts for a large part of the remainder.(202,216) The ash content has been estimated at approximately 5% to 6%, and the residual is composed of trace amounts of lipid, phospholipid, Iysozyme, and an as yet uncharacterized component that is believed to be glycoprotein(216) or possibly a "matrix protein" with a molecular weight of 200,000 daltons to 300,000 daltons (Table 5-1).(231)
 |
TABLE 5-1 Adult Articular Cartilage: Approximate Biochemical Composition |
The water of articular cartilage is extraordinarily important in maintaining the resiliency of the tissue and, as discussed above, also makes an important contribution to boundary lubrication. The water content remains high throughout the life of the animal, with only a modest decrease noted with very advanced age.(74,209,288) The content increases with immobilization,(121) denervation,(155) and osteoarthritis. (185) The mechanism of water binding within the cartilage is poorly understood. Since most of the extracellular matrix consists of almost equal parts of collagen and proteoglycan, it would seem appropriate that these macromolecules are somehow involved in holding the water. Although there are a number of possible ways in which collagen or proteoglycan could bind water, the most logical one, particularly in view of the free exchange of water with the synovial fluid, is complex gel formation in which both the collagen and proteoglycan participate.(83,158,169,174) The water of the gel, although unable to "flow," is believed to be freely ex-changeable with the water of fluids on the other side of the membrane (the synovial fluid) and is subject to all of the physical and chemical laws that govern osmotic solutions.(197,198,226) Experiments have shown that tritiated water administered into the knee joint of a rabbit equilibrates rapidly with the serum (within 30 minutes), suggesting a free exchange of water with the synovial fluid (Fig. 5 25).(158) In further support of the gel theory is the finding that the water can be rapidly removed by vacuum dessication, and only a small portion (perhaps 3% to 4%) is bound so tightly as to resist drastic treatment such as heating to 90 C (158)
The proteoglycans of articular cartilage are complex macromolecules consisting of a linear protein core to which are linked long-chain polysaccharide moieties (glycosaminoglycans).(228) The glycosaminoglycans are polyanionic in charge, based on the regular occurrence of carboxyl groups and sulfates along the macromolecules. (It should be noted that proteoglycans were formerly called either protein polysaccharides or chondromuco-proteins, and the glycosaminoglycans were formerly termed mucopolysaccharides). The standard model for the steric structure and chemical configuration of proteoglycans as proposed by Mathews and Lozaityte(296) is that of a "test tube brush," in which the central core of the molecule consists of a linear protein and linked to it and radiating from it at right angles are numerous glycosaminoglycan chains, stiffly extended in space. This model has been partially confirmed by elegant studies by Rosenberg and co-workers,(246) who performed ultra high power electron microscopy of mixed proteoglycan-cytochrome C monolayers.
 |
FIG. 5-25 Study to define the rate of "clearance" of freshly administered tritiated water and urea 14C by adult rabbit articular cartilage. The isotopes were injected intra-articularly, and the rabbits were serially killed and the cartilage assayed for content of the two isotopes. Values, expressed in terms of the serum values, show that the water quickly equilibrates with the synovial fluid and serum, and by 15 minutes only 5% is retained. At 30 minutes the values are equivalent to that of the animal's serum. These data suggest free exchange of water with synovial fluid (and serum) and support the concept that the majority of the water of cartilage is held in a gel. |
In considering the proteoglycan molecule it would seem logical to describe each of the component parts and then to define the current concept of their molecular organization. The glycosaminoglycan molecule consists of long-chain, unbranched, repeating polydimeric sac- charides, only three of which can be found within the proteoglycan subunit: chrondroitin 4-sulfate; a stereoisomer, chondroitin 6-sulfate; and keratan sulfate (Fig. 5-26).(217,257,262) The chrondroitin sulfates are the most prevalent glycosaminoglycans in cartilage and account for 55% to 90% of the material depending principally on the age of the subject.(216) The repeating disaccharide units are N-acetyl galactosamine and glucuronic acid. The position of the sulfate on the N-acetyl galactosamine defines the isomer: when the sulfate is present on the fourth carbon the molecule is called chondroitin 4-sulfate, and when linked to the sixth, it is called chondroitin 6-sulfate. The average chain length for chondroitin sulfate in articular cartilage is 25 to 30 disaccharide repeating units(156,219,262)
 |
FIG. 5-26 Repeating dimeric sugar units of the four major glycosaminoglycans found in proteoglycan aggregate of articular cartilage Hyaluronate, present only in the aggregate, consists of a polymer of the dimer shown, which includes glucuronic acid, an N-acetyl glucosamine The chondroitin sulfates consist of repeating units of glucuronic acid and N-acetyl galactosamine. The position of the sulfate (on the fourth or sixth carbon) determines whether the polymer is chondroitin 4- or chondroitin 6-sulfate. Keratan sulfate consists of repeating units, each of which contains a galactose and an N-acetyl glucosamine. Chondroitin 4-sulfate, chondroitin 6-sulfate, and keratan sulfate are the principal glycosaminoglycans of the proteoglycan subunit. |
The formula for keratan sulfate is not as well defined as that for the chondroitin sulfates. There are a number of species of keratan sulfate that occur in various body sites, and the exact formula and position of the sulfate remain unclear. The repeating structure for the "ideal" molecule is a galactose linked to an N-acetyl glucosamine with a sulfate on the sixth carbon (Fig. 5-26). Keratan sulfate chains in human proteoglycan are much shorter than those of chondroitin sulfate, consisting of five to ten repeating dimeric units.(84,l30,2l7,245,26l,262) The combination of the glycosaminoglycans and a protein core constitutes the proteoglycan molecule (Fig. 5-27). The protein core measures approximately 180 nm to 210 nm in length(219,245) and is rich in serine, glutamic acid, proline, and glycine.(219,245) All three glycosaminoglycans are linked to the protein at specific sites. The linkage region for the chondroitin sulfates consists of a galactosylgalactosyl-xylosyl serine chain,(153,173) while that for keratan sulfate is a neuraminyl-galactosylgalactosamine-threonine unit.(90)
Extensive chemical studies of the proteoglycan sub-unit have demonstrated that the glycosaminoglycan chains are not distributed uniformly along the entire length of the core protein (Fig. 5-27).(151,152,245) At one end (the hyaluronate-binding region) the macromolecule has few or no polysaccharide chains. Adjacent to this section is a linear segment to which are attached principally keratan sulfate moieties and a few recently described oligosaccharide chains. The remainder of the core is of variable length and has linked to it chondroitin sulfate chains and a smaller number of keratan sulfate and oligosaccharide moieties. The asymmetric nature of the proteoglycan subunit accounts for the puzzling finding by early investigators of fairly marked variation in the concentrations of the glycosaminoglycans (chondroitin 4-sulfate, chondroitin 6-sulfate, and keratan sulfate) at different ages, in different sites, and indeed in different zones of the articular cartilage. Based on studies by Roughley(249) and Sweet and co-workers,(281) it is most likely that the synthetic product is the same but that the molecules are either synthesized with different core protein lengths or perhaps partially degraded from the free terminal end to produce a macromolecule with varying numbers of the various glycosaminoglycan chains linked to it. All, however, appear to have identical hyaluronate binding-and keratan sulfate-rich regions.(82,157,248)
 |
FIG. 5-27 The proteoglycan subunit consists of a linear protein core to which long chains of chondroitin sulfate (cs) shorter chains of keratan sulfate (KS), and even shorter oligosaccharide moieties are asymmetrically linked. The hyaluronate-binding region is protein-rich, with virtually no sugars present, while the remaining two sections contain principally keratan sulfate and oligosaccharides and chondroitin sulfate (with a few keratan sulfate) chains respectively. |
From data reported by Rosenberg(245,247) and others,(149,157,217,252) it seems unlikely that in normal cartilage any but a small fraction of the proteoglycan exists as the free subunit. The majority of the macromolecular material probably forms high-order aggregates containing many sub-units and having molecular weights of 60 x 106 daltons to 150 x 106 daltons.(217,234,245) These complex materials together with their gel-trapped water occupy an enormous domain, which, because of the polyanionic nature of the glycosaminoglycans, is strongly electroneg- active and evokes a resistance to compressive force that contributes significantly to the resiliency of the tissue.(217) Studies have established that hyaluronic acid (see Fig. 5-26), present in only a small quantity in articular cartilage, forms the long filamentous backbone for the aggregate (142,143,147,148) The polymer ranges from 40,000 nm to 420,000 nm in length and attached to it at regular intervals and at approximately right angles are monomeric subunits.(245) Proteoglycan subunits bind to the filament at their protein-rich terminal ends adjacent to the keratan sulfate-rich portion of the macromolecule (Fig. 5-28). Studies have demonstrated that the hyaluronic binding is a function of the protein core, and the binding is at a single site. Hyaluronate can bind as much as 250 times its weight of proteoglycan. Although hyaluronate is clearly the major factor in the aggregation of proteoglycan subunits, proteinaceous constituents known as link proteins have been found as components of the proteoglycan aggregates from every cartilage thus far examined.(219,223,255) The two proteins have molecular weights of approximately 44,000 daltons and 48,000 daltons and can now be isolated in sufficiently purified states to raise specific antibody.(236,285) How the link protein stabilizes the binding of proteoglycan subunits to the hyaluronic acid is not as yet known.
 |
FIG. 5-28 Scheme for the structure of the proteoglycan aggregate- gate. Many proteoglycan subunits are arranged at right angles at regular intervals along a long filament of hyaluronic acid. The subunits are linked at their hyaluronate-binding region associated with the linkage is a stabilizing material, glycoprotein link The macromolecule is extremely large (60-150 x 106 daltons) and exerts a strong electronegativity in aqueous solution. (Courtesy of Lawrence C. Rosenberg, MD) |
A property imparted to the cartilage by the proteoglycan is that of metachromatic staining. Dyes such as alcian blue, toluidine, azure A, crystal violet, and others serve as counterions to the polyanionic glycosaminoglycans.(237) Minute crystals of dye polymerize in a spatial relationship dictated by the loci of available sulfate and carboxyl groups, and when a critical concentration is reached, the arrangement will change the spectrum of transmitted light from the orthochromatic to the meta-chromatic color.(237,260) This phenomenon, known as meta-chromasia, is reasonably specific for the glycosaminoglycans and is of great value as a qualitative and semiquantitative indicator of the concentration and distribution of the macromolecule within the cartilage. Recently Rosenberg(244 ) reintroduced the use of safranin- O, an analine dye that interacts with chondroitin sulfate or keratan sulfate in solution. In its orthochromatic spectrum, safranin-O has an essentially one-to-one re- lationship with the negatively charged groups of the proteoglycan. Thus, this dye, which is quite stable on stained sections, is very useful in the study of cartilaginous tissue.
Well over 50% of the dry weight of adult articular cartilage consists of collagen, which for many years was thought to be similar and perhaps identical to that isolated from skin. However, compelling evidence now exists to indicate that cartilage collagen is a different genetic species. The fibers are smaller than those seen in skin and bone, are less soluble, and have an organization and structure important to the mechanical structure of the articular cartilage.
Chemically, in its simplest form (tropocollagen), collagen from cartilage measures approximately 1.5 nm in diameter and 300 nm in length. Each of these molecules consists of three polypeptide chains coiled in a rigid left-handed helical structure.(144,165) Each chain consists of about 1000 to 1050 amino acids and has a molecular weight of less than 100,000 daltons (123,212,232,233) molecular weight for the entire tropocollagen molecule is estimated to range between 270,000 daltons and 300,000 daltons.(l23,l44,232)
In terms of the molecular conformation of the tropocollagen molecule, each one of the three alpha chains consist of a major central region where the amino acid sequences are repetitions of a glycine-x-y triplet in which x is frequently proline and y is less frequently hydroxy- proline.(212,2l3) Tropocollagen molecules are ordered to form fibrils, and fibrils, by interfibrillar cross-links, are then ordered to form fibers. The fibers in articular cartilage as seen on electron microscopy may vary in width from 10 nm to 100 nm a value that may be exceeded in aging or osteoarthritis (see Fig. 5-8).(293)
The major difference between cartilage collagen (type 11 collagen) and that found in skin and bone (type I collagen) is that it consists of three identical alpha chains instead of two different chains (type I collagen consists of two alpha, and one alpha2 chains in a triple helix, while type II has three alpha, chains in the same arrangement.(94,210,214) Furthermore, the alpha l type II chains have a different chemical composition than the alpha, type I chains, with minor differences in the content of glutamic acid, alanine, and leucine(210,211,214) but major variations consisting of a fivefold increase in the concentration of hydroxylysine(210)and an almost ninefold increase in the content of hydroxylysine-linked carbohydrate.(85,270)
Considerably less information is available regarding other organic materials that are present in small amounts in cartilage. Recently, an "adhesive" protein, chondronectin, has been found in cartilage and presumably is partially responsible for maintenance of the collagen structure in relation to the chondrocyte. (154,169) Lipids form 1% or less of the wet weight of human adult articular cartilage and are found both in the cells and in the matrix. (104,136,238) Both neutral lipids and phospholipids have been described.(88,238) Their relationship to cell metabolism or matrix components is obscure.
Lysozyme, an enzyme that hydrolyzes the beta 1-4 linkage between N-acetyl muramic acid and N-acetylglucosamine, has been found to be present in many connective tissues and has been noted in high concentration in epiphyseal cartilage and, according to recent studies, in articular cartilage as well.(161,163,268) There are no known substrates present within the articular cartilage that could yield to the action of the enzyme, and its extracellular location has been somewhat of a puzzle. Recent studies by several investigators have indicated that additional proteins, as yet poorly defined, may be present in considerable concentration in cartilage.(215,217,231) Muir(216,218) has suggested that this material is a glycoprotein, but recent studies by Paulsson and co-workers(23l) support the concept that it is a "matrix protein" with a molecular weight of 200,000 daltons to 300,000 daltons.
METABOLISM
As noted above, early investigators thought articular cartilage to be metabolically inert. Over the past 25 years, however, there has been ample demonstration of a surprisingly active level of metabolism. Articular chondrocytes synthesize and assemble the matrix components and direct their distribution within the tissue. The synthetic apparatus is complex in that not only are proteins (protein core of the proteoglycan, collagen, glycoproteins, and enzymes) synthesized by the standard genetic pathway, but sugars are assembled into glycosaminoglycans, linked to the protein, and separately sulfated. All of these actions take place under avascular and at times hypoxic conditions, with considerable variation in local pressure and physical-chemical state.
One of the factors that led to the general impression that articular cartilage was inert was the early demonstration that although the tissue had a well-defined glycolytic system,(179) oxygen consumption was considerably lower than in other tissues. Later reappraisals of these data, however, pointed out that the rate of oxygen consumption per unit mass was low but because of the sparse cell population the rate per cell was considerably higher and probably approached that seen in other more vascular body tissues.(95,111,164) Despite this observation, there remains little doubt that articular cartilage uses principally anaerobic glycolysis for energy product(164) have shown that lactate concentration is high in articular cartilage slices and remains unchanged despite oxygen deprivation or the addition of sodium cyanide. On the other hand, the introduction of sodium monoidoacetate causes a marked diminution in the lactic acid production. The (LDH) isoenzyme pattern for articular cartilage from rabbits, cows, and humans has been shown to contain mostly LDH4 and LDHs, isoenzymes that are thought to be facilitators of anaerobic glycolysis.(285,259) Recent studies by Lane and co-workers(164) demonstrated that deoxyribo-nucleic acid (DNA) synthesis and proteoglycan synthesis were somewhat diminished at lower oxygen tensions in articular cartilage and that these activities became optimal at 21% oxygen. It is possible that the oxygen tension is one of the control mechanisms for the metabolism of articular cartilage.
In terms of the synthesis of matrix, approximately 30 years ago investigators first noted a surprisingly rapid rate of incorporation of radioactive sulfate (35-SO4), suggesting that at least one component of the matrix was being synthesized at a rapid rate.(89,114,115) Numerous investigations have now clearly demonstrated that the matrix of articular cartilage is locally synthesized by the chondrocytes and that the rate of synthesis for at least a portion of the proteoglycan is quite rapid.(191,194,294) That for collagen is probably slower but still occurs at a measurable rate.
The process by which the chondrocyte synthesizes and assembles the macromolecules in the matrix is complicated and has at least four components, all of which have been studied extensively. These include synthesis of the protein of proteoglycan; synthesis of glycosaminoglycans and assembly on the protein core; addition of the sulfates to the fourth or sixth carbons of chondroitin sulfate and the sixth carbon of keratan sulfate; and synthesis of the collagen. All of these activities occur intracellularly and are directly (or, in the case of sugar assembly and sulfation, indirectly) controlled by the genetic system of the cell.
Protein synthesis in the articular chondrocyte adheres to the standard genetic pathway in that DNA transcribes a message to messenger RNA, which then translates the message to the ribosome for the assembly of amino acids in appropriate sequence. The protein of proteoglycan and the collagen are synthesized in this fashion.(129) Collagen is synthesized as a procollagen pre-cursor, which then must undergo additional steps for hydroxylation of proline and glycosylation of hydroxylysine.(93,110) The chain is assembled intracellularly, and the triple helix is probably organized prior to extracellular extrusion. Once in the matrix, cross-linking occurs and the terminal ends of the larger procollagen molecule are "clipped" to produce the mature tropocollagen, which is then organized into a fibrillar arrangement (Fig. 5-29). (I0l,l29,159)
Sugar assembly is unique in the tissue-synthetic processes in the sense that it does not occur at the ribosome but probably is localized to the Golgi apparatus. Assembling disaccharide units is very costly in terms of energy unless appropriate intermediaries are used;(171)in the body tissues, the customary intermediary is a uridine diphosphate (UDP)-hexose that combines with another UDP-hexose or just hexose to form a disaccharide.(99,172,262) Chain assembly occurs through combinations of UDP intermediaries(Fig. 5-30) (242,257,262)
Sulfation of sugars also occurs at the Golgi apparatus and uses another nucleotide intermediary consisting of phosphoadenosine phosphosulfate (PAPS), which donates a sulfate to the fourth or sixth carbon of the sugar.(106) The active material is generated enzymatically from adenosine triphosphate (Fig. 5-31).(230,276)
Perhaps one of the areas of greatest interest and controversy in articular cartilage is the problem of "turn-over" of proteoglycan. In 1952 Bostrom(89) demonstrated a half-life of 17 days for 35-SO4 incorporated into costal cartilage of adult rabbits, and subsequent demonstrations of short half-lives for isotopically labeled proteoglycans by Schiller and co-workers(256) and Gross and co-workers(l40) clearly indicated that the proteoglycan macromolecule once synthesized is not inert. Numerous studies have since added values, including a study performed in our laboratory in 1969 that showed a half-life of approximately 8 days for newly synthesized proteoglycan in the cartilage from the knee joint of adult rabbits (Fig. 5-32).(194) The problem with the study (and all studies of this sort) is determination of the size of the rapidly turning over pool. The best estimates range from very small (less than one tenth of 1%) to approximately 8% to 10% of the proteoglycan of the cartilage. (192,199) The fast fraction is turning over so rapidly that it suggests that there are additional fractions that are also turning over at a slower rate. Since the isotope appears on autoradiographs to disappear uniformly throughout the articular cartilage mass, there is no evidence to support the concepts of a focal loss of the cartilage surface or that cartilage is in fact being "shed" in a manner analogous to that seen in the skin. These data strongly suggest the presence of an enzymatic internal remodeling system that has as its substrate the proteoglycan and that is presumably dictated by circumstances other than a compensation for attrition. The nature of this need for rapid turnover is not clear at the present time.
The second issue raised by the rapid turnover of a small "fast fraction" is the likelihood that the proteoglycan of articular cartilage is metabolically heterogeneous and that in fact there is a wide spectrum of fractions, some of which are quite inert and others that are turning over much more rapidly. The mean figure for all of the cartilage is, as Maroudas(199) has suggested, measurable in years, but individual fractions of the heterogeneous components may turn over at much more rapid intervals.
 |
FIG. 5-29 Steps in the synthesis of collagen in articular cartilage. A message from DNA via messenger RNA dictates the assembly of amino acids at the ribosome. In the presence of oxygen, ketoglutarate, Fe2+, and ascorbate some of the prolines and Iysines are hydroxylated. Glycosylation occurs at the Golgi apparatus, and the molecule is transferred to the membrane, where intramolecular cross-linking and triple-helix formation occurs prior to extrusion of the procollagen. The terminal ends are cleaved, and the material forms intermolecular cross-links to make fibrils and fibers within the matrix. |
 |
FIG. 5-30 Scheme for the assembly of chondroitin sulfate on the core protein of proteoglycan. The x-serine-y region probably serves as a primer. From glucose, through a series of nucleotide intermediaries, sugars are assembled in appropriate sequence as a result of the action of 15 enzymes (all locally synthesized). |
 |
FIG. 5-31 Scheme for sulfation of chondroitin using an adenosine phosphosulfate donor. Three enzymes are required for this step, which occurs at the Golgi apparatus. |
 |
FIG. 5-32 Rate of disappearance of glycine 3H and 35SO4 from articular cartilage of adult rabbit with time after intra-articular injection of the isotopic tracers. Values are expressed as percentages of the 0 hour concentrations (in cpm/mg dry weight). Animals were killed at intervals after injection and the cartilage assayed for incorporated radioactivity. As can be noted, the half-life for both isotopic substrates is about 8 days, suggesting a very rapid turnover for a portion of the proteoglycan of the tissue. |
ENZYMES
It has been known for years that certain degradative enzymes act on articular cartilage to destroy the matrix. Papain, a material extracted from the papaya plant, was noted to cause a loss of basophilia and metachromasia on histologic study, as well as profound depletion of the glycosaminoglycans on biochemical analysis.(l28,176)Fell and Thomas(128) noted that administration of vitamin A in large doses causes similar chondrolytic action. Subsequent studies suggested that vitamin A causes activation or release of potent autolytic enzymes contained within the chondrocyte but ordinarily stored in an inactive form in bodies called Iysosomes.(127,176) The enzymes were therefore known as Iysosomal enzymes.(108,109) Lysosomal bodies have been observed on electron microscopic studies of articular cartilage(251,295) (Fig. 5-33), and analyses of the tissues for Iysosomal marker enzymes such as acid phosphatase or B-glucuronidase have shown that these materials are present in very low concentrations in normal tissue but are increased by chemical maneuvers such as acidification or treatment with detergents, both of which destroy the envelope of the Iysosomal bodies.(120,284)
If one considers the classes of enzymes that could actively degrade the intact proteoglycan, only two are present in the body. The first of these is hyaluronidase, which could act on the hyaluronic backbone of the proteoglycan aggregate as well as on the chondroitin sulfate chains. Although studies by Bollet and co-workers(86) have failed to demonstrate evidence for the presence of a hyaluronidase in normal articular cartilage, it is possible that one of the cathepsins may work as a hyaluronidase. The second site on the proteoglycan molecule that is susceptible to attack is the core protein, and in 1964 Ali(73) first showed evidence for the presence of a Iysosomal acid protease that could cleave the protein of the proteoglycan. Subsequent studies by a number of workers have identified these materials as cathepsin D and B, both of which are known to occur in articular cartilage and appear to be more active in specimens from osteoarthritic joints.(8l,253,291) Studies by Poole,(235) using immunofluorescent antibodies for identification of localization of enzymes in articular cartilage, have shown both intracellular and extracellular cathepsin D in cartilage, particularly in tissues that are undergoing active resorption.
 |
FIG. 5-33 Electron micrograph of a cartilage cell specially stained to demonstrate Iysosomal bodies. The dense rounded body in the cytoplasm is presumed to be a Iysosome containing proteolytic enzymes (original magnification x 38,000) (Courtesy of Donald Chrisman, MD) |
Of considerable concern to investigators has been the pH at which the cathepsins act. It seems unlikely that a pH of 5. 5 or less (the optimal pH for the cathepsins) occurs in normal cartilage, suggesting the possibility that another enzyme, a neutral proteoglycanase, is operative in the tissue. Sapolsky and colleagues(254,255) and Ehrlich and co-workers(1l6-118) have independently described a Iysosomal enzyme, as yet only partially purified, that degrades the proteoglycan subunit at a neutral pH. This material may be harvested from the Iysosomal fraction of osteoarthritic articular cartilage, epiphyseal plate, or cultured chondrocytes.
Proteolysis is only one phase of the degradation of the proteoglycan, and it is likely that additional enzymes as yet unidentified act to further degrade the material. Cathepsin F has been found in cartilage, and its role is not clear.(79) Schwartz and co-workers(138,258) have demonstrated the presence of aryl sulfatase, which could serve to initiate degradation of the glycosaminoglycans.
Collagenase has been demonstrated in numerous body tissues, and the pattern of degradation has been well established.(122,145,146,179,298) Most collagenases act on collagen to cleave it into two fragments, one large and one small.(165,298) The material thus degraded is rendered more soluble and can then be further degraded by the action of the proteases.(298) To date, however, there is no direct evidence for the presence of collagenase in normal articular cartilage. Ehrlich and co-workers(119) were able to demonstrate an endogenous collagenase in osteoarthritic human articular cartilage by culture of the tissue for 7 days, followed by chromatography of the culture medium on a heparin-charged sepharose-4B affinity column activated by cyanogen bromide. It should be noted, however, that the turnover rates for collagen as have been recently described are sufficient to require some form of enzymatic degradation system for this otherwise insoluble material, and it is reasonable to suppose that the enzyme is present in normal articular cartilage but is either so tightly bound to an inhibitor(162) or present in such small quantities that separatory and assay systems are unable to detect it.(145)
IMMATURE ARTICULAR CARTILAGE AND THE EFFECTS OF AGING
Unlike many other body tissues, immature articular cartilage differs considerably from that of the adult. On gross inspection, the cartilage from an immature animal appears blue white (presumably because of the reflection of the vascular structures in the underlying immature bone) and is considerably thicker on cut section.(207) As will be discussed, however, the thickness is primarily a function of the dual nature of the cartilage mass; it serves not only as a cartilaginous articular surface for the joint but also as a microepiphyseal plate for endochondral ossification of the underlying bony nucleus of the epiphysis.
On histologic examination it is apparent that the immature articular cartilage is considerably more cellular than the adult tissue, and numerous studies have corroborated the increased number of cells per unit volume or mass (Fig. 5-34),(95,190,271,273) The hypercellularity appears fairly uniform throughout the cartilage, with little variation noted in cell density. The structural organization of the tissue also differs from that described for adults in that the zonal characteristics show major variations, particularly in the lower zones. (188,225) The gliding or tangential layer remains evident in immature cartilage, although the surface cells are somewhat larger and less discoid than those seen in adult cartilage. The midzone is wider and contains a large number of randomly arranged cells. In the lower zones, however, the orientation differs markedly in that at about the half-way mark in the distance from the surface to the underlying bone, the chondrocytes are arranged in irregular columns, and at further depth the columnar arrangement becomes more evident. The cells in these columns show characteristics consistent with those of chondrocytes of the epiphyseal plate. (141) In the uppermost portion of the column the chondrocytes are round or flat or, occasionally, triangular, indicating their participation in the process of diagonal division. With further distance from the surface the cells are increased in size and demonstrate shrunken pyknotic nuclei and large intracytoplasmic vacuoles that have been shown to contain glycogen. Vascular buds from the underlying bone invade the cartilage columns, destroying the cells, and calcification is noted in the matrix between the cartilage columns in a pattern resembling the zone of provisional calcification of the epiphyseal plate.
 |
FIG. 5-34 Low-power photomicrograph of articular cartilage from an immature rabbit. Note the increased cellularity absence of a tidemark, calcified zone, and cortical bony end-plate. The lower half of the cartilage mass is a microepiphyseal plate for growth of the underlying bony epiphyseal center. The zones where mitotic figures and thymidine 3H localization are noted are shown by the dotted lines. The more superficial one (l) is presumably for growth of the cartilage surface, and the deeper one (2) the proliferative zone of the microepiphyseal plate. (Safranin-O, fast green and iron hematoxylin, x l00) |
When immature articular cartilage is examined by light microscopy, mitotic figures are readily noted and all stages of mitosis can be seen. (124,136,190) Cell replication is not present uniformly throughout the tissue, however, in that in the very young animal the mitotic activity occurs in two distinct zones, one Iying subjacent to the surface and presumably accounting for the growth of the cellular complement of the articular portion of the cartilage mass, and a second Iying deeper to this region and comprising a narrow band of cells that morphologically resemble the proliferative zone of the microepiphyseal plate of the subjacent bony nucleus (Fig. 5-34).(186-188) As the animal ages and approaches maturity, the pattern of cell replication changes. The mitotic index is diminished and mitotic activity is confined to only one zone just above the zone of vascular invasion in the lowermost portion of the cartilage, an area that now demonstrates a more diffuse calcification. No evidence for cell replication can be found in the more superficial regions.
In the adult animal mitotic activity ceases with the development of a well-defined calcified zone, a tidemark, and in some species with closure of the epiphyseal plate (1l3,187,188,190,282,283) A careful search of normal articular cartilage from adult animals of all species has failed to demonstrate mitotic figures, and studies using tritiated thymidine have not demonstrated grains over the nucleus indicating DNA replication.(113,187) Although it has been suggested that the chondrocyte divides by amitotic division, there is probably no evidence for such an activity, and cytophotometry and cytofluorometry have failed to demonstrate evidence for the nuclear polyploidy in the adult tissue.(186)
In recent years investigations have provided data regarding variation of chemistry of articular cartilage with advancing age. Water content appears to be increased in immature animals and slowly diminishes to a standard figure that remains constant throughout most of adulthood.(87,121,209,280) As the animal becomes considerably far advanced in age, there may be some diminution. The collagen content of fetal articular cartilage is considerably diminished compared with that of mature animals. (121,267) Once the level climbs to the adult values (shortly after birth), the concentration is maintained throughout the life of the animal even through senescence.
The principal variation in the chemistry of matrix of articular cartilage with advancing age appears to be in the proteoglycan molecule. Proteoglycan content in articular cartilage is highest at birth and diminishes slowly through the period of immaturity.(250) The core protein is diminished in length in immature animals, but polymerization of the glycosaminoglycan is greater. (155,156,250,251,280) As the animal approaches adolescence and maturity, the proteoglycan core becomes longer, and the chain lengths, particularly for chondroitin sulfate, diminish. While chondroitin 4-sulfate has been noted to be extraordinarily high in concentration in immature animals, a fairly rapid diminution in this value occurs with aging. Furthermore, with advancing age the total chondroitin sulfate concentration falls, and keratan sulfate increases until, at approximately age 30 in humans, the value for keratan sulfate approximates 50% of the total glycosaminoglycan and remains constant throughout old age.(50,l25,229,22l,250,280) Inerot and co-workers(157) have shown a decrease in the size of the core protein of the proteoglycan subunit with advancing age and have suggested that the loss of the chondroitin sulfate-rich terminal end of the macromolecule explains the altered glycosaminoglycan distribution seen with advancing years.
When considering the alterations of metabolic activity in articular cartilage with advancing age, it is important to remember that only the more superficial part of the immature cartilage mass is "articular" and the remainder is "epiphyseal." Despite the difficulties that this causes in assessing the roles of syntheses of the matrix, the generally accepted view is that metabolic activity and, specifically, the synthesis of proteoglycan and collagen are increased in immature animals and gradually diminish to the adult level as the animal ages. (192,193,196,213,219) The adult rate of synthesis, once established, appears to remain constant throughout the life of the animal. Surprisingly few abnormalities are noted when considering the overall effect of aging on articular cartilage. As has been indicated, the cartilage may take on a yellowish discoloration with advanced age, the nature of which is obscure. Cartilage appears to be thinner, but accurate measurements have not been carried out to support this concept. Histologically, no alterations are evident on light microscopy, but electron microscopic studies have suggested an increased number of cells showing organellar degeneration and intracytoplasmic fine filaments. Electron microscopic study of the collagen fibers shows increased fiber size in aging similar to that seen in osteoarthritis. Occasionally giant fibers are seen, particularly in the deeper layers.
The proteoglycan of aged cartilage shows some variation, however, which is probably very significant both in the development of osteoarthritis (which has a predilection for the aged) and possibly in an alteration in the transfer of water. Both Roughley and White(250) and Inerot and co-workers(157) have shown a decrease in the proteoglycan content of the cartilage with advancing age and a moderate to marked decrease in the size of the proteoglycan subunit. Keratan sulfate is increased relative to chondroitin sulfate, suggesting that the terminal portion of the proteoglycan subunit has been cleaved. Of considerable interest has been the finding that aging articular cartilage shows diminished aggregation and that there is an increase in the quantity of proteoglycan subunit as compared with aggregate. Furthermore, additional hyaluronate does not appear to enhance the aggregation, suggesting that there is an abnormality of the hyaluronate-binding region of the core protein, synthesis of hyaluronate, or a change in the character or quantity of link protein.
SECTION FOUR
THE RESPONSE OF ARTICULAR CARTILAGE TO MECHANICAL INJURY
HENRY J. MANKIN
Injuries to articular cartilage are common events in mammals, and historically probably no experiment has been repeated more frequently than surgically injuring the joint and observing the manner in which the tissues heal over time. Despite the large number of studies done over the years, some confusion remains as to the nature of the response. Much of the controversy is based on the fact that cartilage responds to different forms of trauma in different ways (see below).
Before describing the response of articular cartilage to mechanical trauma, it is important to consider two issues: first, the question of whether adult articular cartilage can respond to appropriate stimuli by an increase in its synthetic activities for DNA and matrix components; and second, the "normal" course of events in the healing of body tissues after trauma and the aberrations seen in this pattern in the avascular articular cartilages.
THE POTENTIAL FOR HEALING
Central to any discussion of reparative responses on the part of the tissue is the question regarding the ability of a tissue to increase its rate of synthesis of DNA and protein. Some mammalian tissues repair with scar (e.g., skin, liver, kidney, lung, brain) while others (principally connective tissues) repair with a tissue resembling the parent tissue (e.g., bone, tendon, synovium). Regardless of the type of tissue involved in the response, the process is a cellular one in the sense that either fibroblasts (which produce the scar) or specific cells (e.g., osteoblasts, chondrocytes) must be present in large numbers at the site of injury in order to synthesize the repair tissue. For the most part fibroblasts or chondrocytes are "new" cells that must evolve by cell replication and modulation from existing cells or from cells that have migrated from the margins of the wound or from the blood vessels that are entering the tissue. It is important to recognize that DNA replication and cell division are essential characteristics of any repair process and that these processes plus those of modulation or differentiation are essential steps in the repair phenomenon.
Despite the years of study of articular cartilage, there remains some controversy as to whether or not chondrocytes divide and the method by which they do so. There is little doubt that chondrocytes from immature cartilage are capable of cell division,(316,333,337) and studies have demonstrated that mitotic figures can be found in the immature animal in two zones: one subjacent to the surface, presumably for the growth of the articular cartilage mass; and the second deeper layer that forms a proliferative zone for a microepiphyseal plate for growth of the underlying bony epiphyseal nucleus.(333,336) These findings have been corroborated by tritiated thymidine autoradiographs, which show cells with grains in the two zones.(333) As the animal ages, however, evidence of DNA replication becomes more difficult to find, and in older but still immature animals, mitotic figures are rare and are located principally in the basilar areas adjacent to the developing calcified zone and tidemark.(337) The animal at this point has had a sufficient growth of the articular cartilage mass to achieve adult proportions, and one presumes that DNA replication is necessary only for remodeling of the boundary area adjacent to the underlying bony end-plate.(368) Numerous investigators have sought evidence for cell replication in the adult animal, but to date, mitotic figures have not been found in "normal" cartilage.(315,335,367) It appears that once the calcified zone is established and the tidemark well defined in the basilar areas of articular cartilage, no mitotic activity occurs. Early investigators identified chondrocytes with peculiar pyknotic nuclear formations that were thought to be evidence for the somewhat mysterious process of "amitotic division."(3ll,3l6,327,339,) Such a view is hard to dispute on histologic grounds, but tritiated thymidine autoradiographs of mature articular cartilage fail to show grains over adult cells,(333) and cytophotometric study of DNA content of nuclei has shown only limited evidence for polyploidy. These data would indicate that the chondrocyte has little potential for cell replication and raise a major question: does the adult cartilage cell "turn off the switch" for DNA synthesis or does it, like brain cells, "break the switch"? The significance of the answer is obvious. If the cell at maturity disassembles or permanently injures the DNA replicative apparatus, it would theoretically be unable to respond to any stimulus by active mitosis (including lacerative or mechanical injury). If, on the other hand, the chondrocyte only "suppresses" the replicative apparatus (presumably as a result of some cytochemical alteration), it is possible that the inhibitory system can be overcome by appropriate stimuli. The answer to the question appears to lie in the study of osteoarthritis. Analysis of cartilage from joints with osteoarthritis has demonstrated an increased number of cells in clones, and evidence for DNA synthesis has come from tritiated thymidine metabolic studies and autoradiography; and there has even been histologic demonstration of mitotic figures (324,325,343,360,369) These data suggest that under circumstances of chronic injury as is seen in osteoarthritis, chondrocytes are capable of mounting a significant reparative response and can replicate their DNA to form new cells. Whether or not the process is effective in producing sufficient new tissue to repair the cartilaginous surface is not germane to the discussion; the fact remains that the chondrocyte can divide and does so in the adult animal afflicted by osteoarthritis. The second part of the same question relates to the capacity of the synthetic apparatus for protein and glycosaminoglycan in articular chondrocytes to vary the rate at which matrix components are synthesized. The issue here is whether such activities occur at a fixed rate or whether the cell can respond to various stimuli to produce increased quantities of matrix. Ample evidence now exists that articular chondrocytes from immature and adult animals can show altered rates of proteoglycan synthesis in response to such diverse physical and pathologic states as osteoarthritis(338,339,343); altered hydrostatic pressure(359); varied oxygen tension;(303,304,330,363) alterations in pH (364) calcium concentration(352), and substrate concentrations;(3l0,33l, 362), and the presence of growth hormone(344,353,366), ascorbate(332, 363), vitamin E30s cortisol(34l,342,345) and so forth Therefore, it is reasonable to suppose that even adult articular cartilage chondrocytes, if injured, have the capacity to regain their biologically suppressed DNA synthetic activities and are also capable of substantially increasing the rate of matrix synthesis. The possibility of repair in articular cartilage exists.
The second issue that must be discussed in relation to the possible response of articular cartilage to mechanical trauma involves the pathophysiology of trauma. Since cartilage is an avascular tissue, significant modifications must occur in the normal response of the body's tissues to trauma, modifications that are likely to affect the healing process. The general response to injury in vascularized mammalian tissues is a phasic one, so similar for most organs and structures as to be almost stereotypic. The response is generally divided into three more or less distinct phases: necrosis, inflammation, and repair.
The phase of necrosis begins immediately after injury and is characterized by tissue death, which varies in extent depending on the type and degree of trauma, the dependence of local tissue on its blood supply, and the richness (and collateral circulation) of the vascular bed. Injuries of limited extent in a richly vascular tissue in which the cells are not remarkably susceptible to hypoxia may result in little or no necrosis. In circumstances in which the injury is extreme, the blood supply marginal, and the cells very susceptible to a decline in oxygen tension, an extraordinary amount of necrosis may evolve (typified by osteonecrosis of the femoral head that follows some fractures of the neck of the femur). The second phase, that of inflammation, follows almost immediately after the first and is similar to that seen in infectious or immune challenges of the tissues in that it is mediated almost entirely by the vascular system. An increase in blood flow, dilatation of the vascular channels, and increased permeability of the vessel walls produce the rapidly developing reactive engorgement and hyperemia. Transudation and cellular exudation lead eventually to the development of a large mass of cellular and proteinaceous material that fills the extracellular spaces in the traumatized area and forms a dense fibrin network, rich in inflammatory cells that have the potential for cell division and modulation to repair cells. With time, the fibrinous and cellular structure becomes organized into a primitive "glue," holding the wound edges together and sealing the wound.
The third phase of the response of tissues to trauma is that of repair, which supervenes when the fibrinous mass is invaded by newly formed blood vessels and the inflammatory cells modulate into fibroblasts to produce at first a loose vascular granulation tissue, subsequently a fibrous repair matrix, and finally scar that firmly welds the wound edges together and, by later contracting, closes the wound gap. As indicated above, in certain of the body's tissues this final phase is associated with the replacement of the damaged tissue by the same tissue as was originally injured (bone healing with bony callus; tendon healing with tendon) rather than the fibrous scar such as occurs in organs such as the skin, liver, and kidney. In these circumstances, the reparative material undergoes a sometimes prolonged remodeling process to restore or approximate normal anatomy.
In considering the application of this scheme to injuries to the hyaline articular cartilages, it is apparent that the response is likely to vary in an important way in view of the avascular state of the cartilage. Cartilage undergoes the same phase of necrosis as any other body tissues. The cells at the site of injury die, and the matrix is disrupted to varying degrees depending on the extent and type of trauma. Since chondrocytes are relatively insensitive to hypoxia, there is probably less cell death than one might see in other body tissues, such as bone. The second phase, inflammation, which is mediated almost entirely by the vascular system, is absent. No blood escapes from ruptured vessels, and no clot can be produced. There are no local blood vessels to undergo vascular dilatation, and the processes of transudation, exudation, and hematoma formation are absent. No fibrin is produced; hence, the fibrin clot that serves as the scaffolding for the ingrowth of repair tissue is absent. If one considers the third phase, repair, the absence of an inflammatory or vascular phase (necessary to bring in blood vessels and undifferentiated cells that could modulate or differentiate into either fibroblasts or chondroblasts) limits considerably the cells that are available to respond to the trauma. The burden for repair falls on the existing chondrocytes. As discussed above, they are capable of active DNA synthesis and increased matrix synthetic activity, but the task is too great for a small number of cells with limited potential for metabolic activity; as will be discussed below, experimental and clinical studies have borne out the sharp limitation of the repair potential in injuries to mature articular cartilage in which no vascular system is involved.
It should be noted, however, that if the basal layers of the cartilage are involved in the injury and, specifically, the damage extends to or through the vascular bony endplate of the underlying subchondral cortex, the previous discussion is inappropriate, since for such an injury all three phases of repair are possible. Necrosis is evidently present (both cartilage and bone), and since the bone is richly vascular, the inflammatory response can be anticipated to be very extensive. Furthermore, the underlying bone is a superb source of new blood vessels and primitive cells for differentiation and modulation to fibroblasts or chondroblasts, all of which are required for the phase of repair. Thus in deeper cartilaginous injuries (those that injure the underlying bone as well), one would anticipate a more stereotypic response in which all three phases are demonstrated and the classic form of repair occurs.
RESPONSE TO SUPERFICIAL LACERATIVE INJURY
The experiment in which a superficial (not extending to the underlying bone) laceration is produced in the articular cartilage of an animal's joint and the tissue then evaluated at regular intervals for evidence of healing has been repeated often. The earliest recorded observation was that of Hunter in 1743,(326), who stated, "from Hippocrates to the present age it is universally allowed that ulcerated cartilage is a troublesome thing and that, once destroyed, it is not repaired." Since then, numerous studies have defined the relatively meager and, indeed, ineffectual response of articular cartilage to lacerative injury (299,301,302,306,308,314,315,320,322,323,328,336,339,351.354,365) With the exception of Redfem,(357) who in 1851 made the observation that ". . . I no longer entertain the slightest doubt that wounds in articular cartilage are capable of perfect union by the formation of fibrous tissue out of the texture of the cut surfaces," all observers have demonstrated by gross, microscopic, and ultrastructural evaluation the apparent inability of the articular cartilage to produce sufficient new tissue to heal a lacerative injury or repair an ulcer on its surface. In injuries confined to the substance of the cartilage of adult animals (i.e., not violating the junction of the calcified zone and the underlying bony end-plate), the response has been shown to lack both the inflammatory and repair components of the standard process as observed in more vascular tissues. Furthermore, the reaction (or lack of it) appears to be independent of extent, depth, or orientation of the lesion and is characterized by abortive and disappointing at- tempts on the part of the cartilage to add cellular and matrix elements, which in the mature animal is almost never effective in healing the defect (Fig. 5-35).(307,314,320,340)
In several animal studies the biochemical characteristics of this response have been defined. Immediately after superficial lacerative injury, ghost cells are noted in the lacunae of chondrocytes adjacent to the margin of the slice or injury (phase of necrosis; Fig. 5-36). After one day, however, a relatively intense burst of mitotic activity is noted in the adjacent cartilage (Fig. 5-37).(334) The cell replicative activity is associated with a more sustained increase, though of perhaps a lesser degree, in the rates of synthesis of matrix components as measured by 35-S04 (an indicator of glycosaminoglycan synthesis) and tritiated glycine (an indicator of protein synthesis).(314,340) The levels of isotopically labeled substrate incorporation accelerate through the second day post injury, suggesting increments in the rates of matrix synthesis and, even more important, cell replication (as measured by tritiated thymidine incorporation).(340) This process, however, is unfortunately very short lived, and by one week after trauma, the metabolic activity is reduced to levels equivalent to those of cartilage samples from sham-operated or normal joints (Fig. 5-38).(340)
 |
FIG. 5-35 High-power photomicrographs of superficial slice injuries to articular cartilage of adult rabbit (confined to the cartilage with no injury to the underlying bone). (A) The immediate response ( I week following injury). (B) The lack of healing or significant alterations in the structure of the cartilage at 6 months. (H&E, x 400) |
 |
FIG. 5-36 Photomicrograph of the margin adjacent to a slice injury in adult rabbit articular cartilage one day after injury. Note the presence of ghost cells adjacent to the slice. (The dotted line indicates the margins of the slice.) (H&E, x 850) |
 |
FIG. 5-37 High-power photomicrograph of a thymidine 3H autoradiograph of articular cartilage of adult rabbit 2 days after a superficial lacerative injury . (The slice is not seen in the section. ) Note grains over several cells indicating DNA replication, a phenomenon not observed in normal adult cartilage. (H&E, x 1000) |
Thompson(310) examined the degradative side of the metabolic response and showed increased concentrations of cathepsins, B-glucuronidase, hexuronidase, and aryl sulfatase in the first week after creation of the superficial lacerative lesions. Continued study of the tissue, however, showed a decline to control values for all enzymes. For unknown reasons aryl sulfatase became elevated again at about 16 weeks post injury. Fuller and Ghadially (370) produced tangential partial-thickness defects in the articular cartilage of rabbits and studied them both histologically and by electron microscopy. Their results did not differ from those seen in studies of perpendicular slices in that they found no evidence for repair even in young animals. Their ultrastructural studies showed cell death at the margin of the wound but also signs of an increased metabolic activity in the surviving chondrocytes, presumably directed toward the synthesis of matrix proteins. Nucleolar hypertrophy, increased quantities of rough endoplasmic reticulum, and occasional increased numbers of Golgi complexes were noted after injury, but by 6 months the defect remained and, except for some minor remodeling of the surface, all healing processes had ceased.(320) Using the same model Ghadially and co-workers(372) studied the repair tissue over a 2-year period with the scanning electron microscope. Initially a new layer of homogeneous matrixlike fibers formed over the injured surface, but as the new material slowly remodeled, the pits and depressions characteristic of the surface of normal tissue appeared, and soon thereafter all of the newly formed material was lost. At 2 years the cartilage surface was found to be almost identical in structure to the appearance immediately after injury.
 |
FIG. 5-38 Metabolic activity in distal femoral articular cartilage of adult rabbit in which slices have been performed and the animals serially studied for rates of incorporation of thymidine 3H (an indicator of DNA synthesis) and 35-S04 and glycine 3H (indicators of matrix synthesis). Values are expressed as fractures or multiples of control (sham-operated) opposite femora Note the sharp burst of DNA synthesis and the lesser peaks for matrix synthesis in the first and second days post injury. By one week the values are the same as the sham-operated side. |
Longer term follow-up of superficial lacerative injuries (slices) has demonstrated no further healing, but oddly, neither has there been evidence of progression to osteoarthritis. In 1963, Meachim(346) described a "scarification" model in the rabbit in which multiple superficial slices were made in the articular cartilage in order to produce a disorder resembling chondromalacia. He and other investigators, however, have shown that the lesions remain stable and that only occasionally does one see evidence of early osteoarthritic change. (347,370) Rosenberg performed such a study and used routine histology and safranin-O staining (as a histochemical indicator of proteoglycan concentration) to observe the cartilage over time. At the end of one year he found the lacerative defects in the cartilage essentially unchanged.
One of the more provocative theories regarding the failure of superficial lacerations to repair is related to the effect of proteoglycan on clot formation. Although the cartilage itself is avascular, it is difficult to visualize a circumstance in which a lacerative injury can occur to the cartilage without injury to the highly vascular synovium. In most injuries to a joint, the synovium is torn and blood fills the cavity. Although there is a limited Rosenberg L: Unpublished data, 1979 quantity of fibrinogen available in synovial fluid, an ample amount should enter the clot from the ruptured blood vessels, and it would seem reasonable that the clot should adhere to the irregular surfaces of the cartilaginous defect. The failure of such a phenomenon to occur is probably related to the presence of hyaluronic acid (from synovial fluid) and more directly to a leakage of proteoglycans from the cartilage matrix. This leakage coats the collagen fibers at the surface of the defect and prevents adhesion of platelets. To assess this possibility, an experiment was performed in our laboratory in which superficial lacerative defects were made in the distal femur of adult rabbits.(317) One group of animals received three intra-articular injections of very low concentrations of crude papain (a proteoglycanolytic enzyme) at 48-hour intervals. When the animals were killed at 6 weeks, there was evidence of healing of the surgically created lacerative injuries in the papain-treated joints but not in the contro