Intervertebral disc degeneration adjacent to a lumbar fusion. An experimental rabbit model.

Our study establishes a rabbit model of disc degeneration which requires neither a chemical nor physical injury to the disc. Disc degeneration similar to that seen in man was created at levels proximal (L4-L5) and caudal (L7-S1) to a simulated lumbar fusion and was studied for up to nine months after arthrodesis. Loss of the normal parallel arrangement of collagen bundles within the annular lamellae was observed in intervertebral discs adjacent to the fusion at three months. By six months there was further disorganisation as well as loss of distinction between the lamellae themselves. By nine months the structure of the disc had been replaced by disorganised fibrous tissue, and annular tears were seen. There was an initial cellular proliferative response followed by loss of chondrocytes and notochordal cells in the nucleus pulposus. Degeneration was accompanied by a decrease in the monomer size of proteoglycans. Narrowing of the disc space, endplate sclerosis and the formation of osteophytes at adjacent disc spaces were observed radiologically.

Temporal stability of the Minnesota Multiphasic Personality Inventory (MMPI) in patients undergoing lumbar fusion: a poor predictor of surgical outcome.

In this prospective study, the Minnesota Multiphasic Personality Inventory (MMPI) was administered to patients before and after lumbar spines fusion to investigate the stability of MMPI scores after surgical intervention and to attempt to correlate MMPI scale scores with outcome data. Sixty-eight patients were included. Testing was performed before surgery and at a mean of 1.5 years after surgery. Clinical outcome ratings were assigned by using criteria of pain relief and analgesic use. In addition, demographic variables known to affect outcome were analyzed. Sixty percent of the patients had a successful clinical outcome. Positive outcome correlated with the demographic factors of occupation (homemaker) and solid fusion. MMPI scales were stable across time, with no difference between groups. Independent t tests were used to study preoperative MMPI scores with respect to clinical outcome. Unsatisfactory outcomes were associated with higher scores on scales 1, 3, and 10 before surgery. Postoperative testing revealed significant outcome correlations--higher scores on scales 1, 2, 3, 5, 7, and 8 associated with an unsatisfactory outcome. However, discriminant function analysis of preoperative MMPI data was able to classify outcomes correctly in only 58.8% of the cases. The utility of the MMPI as a predictor of outcome after surgical intervention is quite limited. Use of group data and testing before and after surgery does not appear to influence this conclusion. Although the scores as a group were stable across time, the amount of variance in outcome that could not be accounted for by using MMPI scales as predictors was unacceptably large.

Results of surgery for spinal stenosis adjacent to previous lumbar fusion.

The literature provides little data to guide surgical management of spinal stenosis adjacent to previous lumbar fusion. Thirty-three consecutive patients who had surgical decompression for spinal stenosis at the lumbar segments adjacent to a previous lumbar fusion were studied. The mean interval between fusion and the adjacent segment surgery was 94 months. Of the 33 patients, 26 were followed for 3-14 years (mean: 5 years) after adjacent segment surgery and were clinically evaluated and independently completed an outcome questionnaire. Of the 26 patients, 15 rated their outcome as completely satisfactory, 6 were neutral toward the surgery, and 5 considered their surgery a failure. The surgery was generally effective at improving or relieving lower extremity neurogenic claudication. The strongest independent predictive factor of patient dissatisfaction was ongoing postoperative low back pain (r = 0.7, p = 0.001). A higher back pain score at follow-up was associated with continued narcotic use (p = 0.001) and decreased ability to perform activities of daily living (p = 0.05). Six patients required further lumbar surgery during the follow-up period. This study provides the longest published follow-up data of surgical results for symptomatic spinal stenosis adjacent to a previously asymptomatic lumbar fusion.

Management of metastatic disease of the spine.

Accurate assessment of spinal cord stability, pain, and neurologic function is essential to any rational approach to metastatic disease of the spine. The Harrington classification is useful in terms of selecting treatment options and planning appropriate interventional and adjuvant therapy. For milder cases, with bone involvement in the absence of structural deformity, adjuvant therapy usually is appropriate. In cases of bone collapse, with or without neurologic compromise, surgical consideration should be given to patients with appropriate life expectancy. Surgical principles revolve around anterior decompression and reconstruction, for anatomic and biomechanical reasons. More widespread use of anterior instrumentation may obviate the need for subsequent posterior stabilization, but additional long-term study is required in this area.

Peridiscal vascular anatomy of the lower lumbar spine. An endoscopic perspective.

STUDY DESIGN: An analysis of the vascular anatomy relative to the underlying discs in the lower lumbar spine performed by using radiologic studies. OBJECTIVE: To define better the vascular anatomy of this region from the perspective of the endoscopic spine surgeon, and to draw conclusions regarding surgical accessibility of the lower lumbar intervertebral discs. SUMMARY OF BACKGROUND DATA: Cadaveric studies have highlighted the high degree of anatomic variability among the vascular structures overlying the anterior elements of the lower lumbar spine. An endoscopic approach to the intervertebral disc at these levels often is limited by arterial or venous structures. The growth of laparoscopic techniques to approach the intervertebral disc has renewed interest in the vascular anatomy of this region. METHODS: A computer-generated series of abdominal vascular studies performed for unrelated indications were scrutinized to ensure clear demarcation of the overlying vascular structures and underlying bony anatomy of the lumbar spine in a true anteroposterior plane. For analysis, 21 arterial (abdominal angiograms) and 22 venous (venograms) studies were selected. For the arterial data, level of aortic bifurcation and course of common iliac vessels were determined. The venous data were analyzed in a similar manner using the confluence of the common iliac veins as the major landmark. To draw conclusions about anterior surgical accessibility of the disc, the location of vessels was measured with respect to intervertebral disc spaces. RESULTS: The bifurcation level of the abdominal aorta was quite variable, occurring anteriorly to the L3 vertebral body in 3 of 21 studies (14%), and anteriorly to L4 and L5 in 10 (48%) and 7 (38%) of the studies, respectively. The cranial half of L5 was the most common single area of bifurcation (in one third of the studies). Venous anatomy was more consistent, with 86% of the patients displaying an iliac vein confluence at L5, and 14% having a confluence overlying L4. From an arterial perspective the L3-L4, L4-L5, and L5-S1 discs were readily accessible in 10%, 38%, and 95% of the studies, respectively. From a venous perspective, these levels were accessible in 100%, 73%, and 60% of the studies, respectively. CONCLUSIONS: The results of this study suggest that a laparoscopic approach to the L3-L4 intervertebral disc will seldom be accomplished without significant retraction of the aorta. Access to the L4-L5 space will be accomplished readily in approximately one third of the patients. In the others, it will require significant vascular dissection. The L5-S1 space, conversely, will be readily accessible by the authors' definitions without significant vessel dissection in the majority of patients.

Hardware failure in an unconstrained lumbar pedicle screw system. A 2-year follow-up study.

STUDY DESIGN: A consecutive study of patients who underwent lumbar spinal arthrodesis with an unconstrained pedicle screw system. OBJECTIVES: To determine the rate of arthrodesis and of clinical success and to examine and characterize the cases of hardware failure with the AO/Dynamic Compression Plate system (Synthes, Paoli, PA). SUMMARY OF BACKGROUND DATA: Although the advantages and disadvantages of nonconstrained versus constrained systems have been studied extensively, instrumentation failure has not. Additionally, the association between pseudarthrosis and hardware failure per se is unclear. METHODS: Seventy-four consecutive cases of lumbar spinal fusion are reviewed. Standard outcome scores based on pain relief and medication usage were tabulated, along with pertinent demographic data. The patients were observed at five intervals after surgery for at least 2 years (range, 24 to 35 months; mean, 27 months). Standard statistical analyses were used to analyze data. Status of the arthrodesis was determined by standard radiographic criteria. RESULTS: The overall fusion rate was 61%. At final follow-up, 60% of patients believed that their back pain had improved, whereas 70% believed that their limb pain had improved. The presence of a solid fusion (r = 3.3, P = 0.010) was correlated positively with a successful clinical outcome; the presence of pseudarthrosis and preoperative narcotic use were negatively correlated with a successful clinical outcome. Twenty-two percent of patients (16) experienced hardware failure. Twelve of the 16 had pseudarthrosis; in the majority of these patients, hardware failure occurred at the level of the pseudarthrosis. CONCLUSIONS: The results of this study demonstrate an extremely high rate of hardware failure and pseudarthrosis using an unconstrained pedicle screw system. Interestingly, the initial rate of pain relief was higher and declined over time and was quite possibly associated with loosening of the hardware. Based on these data, it is difficult to recommend the use of an unconstrained fixation system in the lumbar spine.

Occipitocervical neutral position. Possible surgical implications.

STUDY DESIGN: The study defines the occipitocervical neutral position using cervical radiographs from 30 subjects. OBJECTIVE: To identify reproducible radiographic measures of the occipitocervical neutral position that can be used during surgery to optimize fusion position. SUMMARY OF BACKGROUND DATA: When performing rigid internal fixation of the occiput to the cervical spine, the ability to determine that the occiput is in a neutral position in relation to the cervical spine is important. Currently, no objective radiographic measures for the occipitocervical neutral position exist. METHODS: Thirty flexion, extension, and neutral lateral cervical spine radiographs radiographs interpreted as normal by an experienced radiologist were studied. The occipitocervical angle and occipitocervical distance were defined and calculated. Two investigators, an orthopedic resident and an experienced orthopedic spine surgeon, measured the occipitocervical angle and occipitocervical distance independently on all radiographs in a blinded manner. Correlation coefficients were obtained to determine interobserver reliability. RESULTS: The mean occipitocervical angles were 24.2 degrees, 44.0 degrees, and 57.2 degrees in flexion, neutral, and extension, respectively. The mean occipitocervical distances were 21.5 mm in neutral, 28.0 mm in flexion, and 14.8 mm in extension. The differences in the occipitocervical angle and occipitocervical distance in neutral, flexion, and extension were statistically significant (P < 0.05 and < 0.001, respectively). There were no significant interobserver differences in any of the measurements. CONCLUSIONS: The radiographic measures of the occipitocervical neutral position reported in this study are reliable, repeatable, and simple to determine on routine lateral radiographs. These measurements should be a valuable intra-operative tool for achieving occipitocervical fusion in appropriate alignment.

A critical appraisal of the pain associated with iatrogenic adult spinal deformity, and treatment by spinal cord stimulation. A case report and review of the literature.

Objective. Over the past two decades, with the increasing use of spinal instrumentation to treat deformity, surgical restoration has become more frequent. A complication of surgical reconstruction for adult scoliosis, the iatrogenic flat back syndrome, has been described. Results of surgical realignment have been disappointing in terms of pain relief. The objective of this study is to examine a case of so-called flat back deformity and describe treatment via spinal cord stimulation. Materials and Methods. This is a case report of an individual who presented a spine center with the flat back deformity and pain. Results. Following exclusion of reversible compressive pathology, a 46 year-old male with the iatrogenic flat back deformity and lumbosacral and thigh pain was treated with implantation of a thoracic spinal cord stimulator. Significant pain relief was obtained without resorting to additional major reconstructive surgery. Conclusions. Literature on the pain associated with flat back deformity is incomplete regarding description and characterization of pain. Given the unpredictability of spinal osteotomy to address this pain, a possible alternative treatment strategy is presented. This involves the use of selective pharmaceutical therapy and spinal cord stimulation. Based on the response of this patient to spinal cord stimulation, it is a possible that a component of this persistent pain is neuropathic, despite the fact that preoperative imaging studies failed to disclose a significant compressive lesion.

Collagen crosslinks in human lumbar intervertebral disc aging.

STUDY DESIGN: Human lumbar intervertebral discs from individuals of varying ages were obtained at autopsy and analyzed for collagen crosslinks. OBJECTIVES: To analyze alterations in collagen crosslinks in human lumbar intervertebral discs with aging and disc degeneration. Crosslinks studied were pyridinoline, which is a collagen maturation crosslink, and pentosidine, a nonenzymatically initiated age-related crosslink. SUMMARY OF BACKGROUND DATA: Crosslinking of collagen fibers within the matrix affects intervertebral disc biomechanics. In various connective tissues, alterations in pyridinoline and pentosidine crosslinks have been shown to predispose the tissue to mechanical failure. Little is known about the fate of intervertebral disc collagen crosslinks with advancing age and disc degeneration. METHODS: Forty-two postmortem lumbar intervertebral discs were harvested from nine individuals whose ages were 24, 44, 47, 52, 67, 72, 75, 82, and 89 years. Degree of disc degeneration was graded macroscopically. Each lumbar disc was extracted with 4 mol/L guanidine hydrochloride, and the residual collagen was acid hydrolyzed and analyzed by reverse-phase high-performance liquid chromatography for pyridinoline and pentosidine crosslinks. RESULTS: The findings indicate a decrease in pyridinoline and an increase in pentosidine crosslink levels with disc aging. The decrease in pyridinoline crosslinks with disc aging is a novel finding and may have detrimental effects on matrix resilience. Increased pentosidine levels have been implicated in the age-related deterioration of connective tissue. With advancing degrees of macroscopic disc degeneration, pentosidine levels increase, and pyridinoline levels are diminished. CONCLUSIONS: Alterations in concentrations of pyridinoline and pentosidine collagen crosslinks occur with intervertebral disc aging and degeneration. These changes may contribute to the loss of disc integrity and play a role in the pathogenesis of the degenerative process.

Extradural sensory rhizotomy in the management of chronic lumbar radiculopathy: a minimum 2-year follow-up study.

STUDY DESIGN: Fifty-one consecutive patients who underwent extradural sensory rhizotomy for chronic radiculopathy after lumbar surgery were reviewed retrospectively. OBJECTIVES: To determine the effectiveness of sensory rhizotomy in the management of chronic radiculopathy in patients selected by extensive imaging techniques and selective nerve root sheath injections. SUMMARY OF BACKGROUND DATA: Results of more central ablative procedures for chronic benign pain problems have been disappointing, with variable reports of pain relief. METHODS: Fifty-one patients were reviewed. All patients underwent extensive evaluation to exclude reversible structural lesions, and all had the diagnosis of chronic radiculopathy confirmed by results of clinical and electrophysiologic examination. Selective nerve root sheath injections under fluoroscopic guidance confirmed the symptomatic nature of the segments. All blocks were repeated at least once. All patients underwent selective sensory rhizotomy or, in some cases, complete rhizotomy. After rhizotomy, 37 patients were available to be observed at selected time intervals for a minimum of 2 years. Clinical results were determined by the presence or absence of pain relief (visual analog scale), sensory and motor deficits, narcotic analgesic usage, and the patient's estimation of the effectiveness of the procedure. RESULTS: At 6 months after surgery, all 51 patients and the outcomes of their surgery were available for review. Fifty-five percent of patients rated were believed to have good or excellent outcomes, whereas the remainder had poor or failed outcomes. For the minimum 2-year follow-up period (range, 2-4.2 years), 37 patients were available for review. At final follow-up examination only 19% of the patients maintained good or excellent outcomes. CONCLUSIONS: The results of the rhizotomy procedures deteriorated over time. Possible reasons for the failure, other than temporal deterioration, were anatomic factors and lack of specificity of diagnostic techniques, specifically selective nerve root sheath injection. At this point rhizotomy cannot be recommended with any confidence whatsoever in the setting of chronic lumbar radiculopathy after lumbar surgery.

Chlorine-36 in fossil rat urine: an archive of cosmogenic nuclide deposition during the past 40,000 years.

Knowledge of the production history of cosmogenic nuclides, which is needed for geological and archaeological dating, has been uncertain. Measurements of chlorine-36/chlorine (36Cl/Cl) ratios in fossil packrat middens from Nevada that are radiocarbon-dated between about 38 thousand years ago (ka) and the present showed that 36Cl/Cl ratios were higher by a factor of about 2 before approximately 11 ka. This raises the possibility that cosmogenic production rates just before the close of the Pleistocene were up to 50% higher than is suggested by carbon-14 calibration data. The discrepancy could be explained by addition of low-carbon-14 carbon dioxide to the atmosphere during that period, which would have depressed atmospheric radiocarbon activity. Alternatively, climatic effects on 36Cl deposition may have enhanced the 36Cl/Cl ratios.

Anterior cervical pseudarthrosis. Natural history and treatment.

STUDY DESIGN: A retrospective study of long-term clinical outcomes in 48 patients with pseudarthroses after anterior cervical discectomy and fusion. OBJECTIVES: To determine the natural history, risk factors, and treatment outcomes in a large population with documented pseudarthrosis after anterior cervical discectomy and fusion. SUMMARY OF BACKGROUND DATA: Recent reports suggest that pseudarthrosis after anterior cervical discectomy and fusion adversely affects clinical outcome. Little data regarding cervical pseudarthroses have been published, and conclusions have been drawn from reports with small patient populations and short-term follow-up periods. METHODS: Forty-eight patients with radiographically documented pseudarthrosis after anterior cervical discectomy and fusion were studied. Patients were examined and radiographs made at regular intervals (mean follow-up, 66 months). Clinical results were based on patients' assessment of pain, prescription drug use, activity level and Odom's criteria. Clinical outcomes in patients who underwent surgical repair of the pseudarthrosis are reported. RESULTS: Of the 48 patients, 32 (67%) with pseudarthroses were symptomatic at latest follow-up or at the time of further surgery. Of the 32 patients, 9 had a symptom-free period of at least 2 years after the anterior cervical discectomy and fusion before redeveloping cervical symptoms after a traumatic episode. Of 48 patients with pseudarthroses, 16 (33%) remained asymptomatic at a mean of 5.1 years after anterior cervical discectomy and fusion. A younger age at the time of anterior cervical discectomy and fusion increased the likelihood of the pseudarthrosis becoming symptomatic. After multiple level anterior cervical discectomy and fusion, the caudal-most operated level accounted for 82% of the pseudarthroses. Sixteen patients had an anterior repair of the pseudarthrosis, and fusion was achieved in 14. Six patients underwent posterior pseudarthrosis repair, and all healed. In patients in whom fusion was achieved with a second cervical operation, the results were excellent in 19 and good in 1. CONCLUSION: A pseudarthrosis after anterior cervical discectomy and fusion is frequently associated with a poor clinical outcome. Surgical repair of the pseudarthrosis with an anterior or posterior approach seems to have a high likelihood of a successful clinical outcome.

Effects of basic fibroblast growth factor, transforming growth factor-beta 1, insulin-like growth factor-1, and insulin on human osteoarthritic articular cartilage explants.

This study evaluated the effects of basic fibroblast growth factor, transforming growth factor-beta 1, insulin-like growth factor-1, and insulin on the incorporation of thymidine and sulfate in human osteoarthritic articular cartilage. Tissue explants were obtained from 11 patients undergoing total knee arthroplasty and were categorized as nonfibrillated or fibrillated cartilage. The explants were cultured for 22 days, with changes of medium and growth factor every 72 hours, and labeled with [3H]thymidine and [35S]sulfate. Growth factors were used in the following concentrations: basic fibroblast growth factor at 1, 10, and 100 ng/ml; transforming growth factor-beta 1 at 0.5, 5, and 50 ng/ml; insulin-like growth factor-1 at 0.15, 1.5, and 15 ng/ml; and insulin at 0.05, 0.5, and 5 micrograms/ml. Basic fibroblast growth factor decreased thymidine incorporation to 70% and sulfate incorporation to less than 20% that of the growth factor-free controls. Transforming growth factor-beta 1 had no significant effect on thymidine incorporation, whereas the concentrations studied inhibited sulfate incorporation to approximately 40% that of the controls. At the concentrations tested, insulin-like growth factor-1 had no significant effect on incorporation of either thymidine or sulfate. In contrast, insulin significantly stimulated the incorporation of both. Compared with growth factor-free controls, insulin maximally increased thymidine incorporation by a factor (+/- SEM) of 2.36 +/- 0.47 and 1.69 +/- 0.19 in nonfibrillated and fibrillated explants, respectively; sulfate incorporation was maximally increased 1.60 +/- 0.24 and 1.92 +/- 0.29-fold for nonfibrillated and fibrillated explants, respectively. Of the factors tested, insulin demonstrated the greatest promise for promoting a synthetic response that may contribute to the regeneration of osteoarthritic cartilage.

Cementless total hip arthroplasty in patients with steroid-induced avascular necrosis of the hip. A 62-month follow-up study.

Twenty cementless porous-coated primary total hip arthroplasties (THA) were performed on 15 patients, all of whom were diagnosed with steroid-induced avascular necrosis (AVN). The average age of the patients at the time of surgery was 45 years. Minimum follow-up period for all patients was 24 months (average follow-up period: 62 months). Patients were rated using the modified Harris hip score, as well as serial radiographs. The average hip score at follow-up examination was 88, with 17 of the 20 hips having good or excellent clinical results. No revisions of the prostheses were performed. Using radiographic criteria, 12 femoral components met the criteria for bone ingrowth, seven were considered stable with fibrous fixation, and one femoral component was loose. When good initial femoral component fit was achieved, bone ingrowth reliably followed. Three acetabular components showed migration on serial radiographs. A high rate of acetabular component wear and osteolysis was noted. Avascular necrosis has been shown to adversely affect the outcome of hip arthroplasty surgery. Previous studies of patients with advanced AVN undergoing cemented THA report a high incidence of component loosening. Literature on the results of cementless THA in this patient group is sparse. The study demonstrates encouraging clinical results for cementless THA in steroid-induced AVN. Reliable femoral component fixation occurred if a good initial component fit was achieved; however, long-term acetabular loosening and wear remain serious concerns.

Evidence of proteoglycan/proteoglycan interactions within aggregates.

Nonaggregated proteoglycan monomers, digested fragments of the monomers, as well as link proteins have been shown to self-associate. These associations have not been shown to occur on the aggregate. However, previous reports, using the Kleinschmidt technique of monolayer electron microscopy, have noted proteoglycan subunits on the aggregate that appear to interact, either as branched proteoglycans or as proteoglycan subunits that appear to share the same attachment site on the hyaluronic acid chain. Branching and shared attachments were noted in all aggregates analyzed in this study. Increasing the average space between proteoglycan subunits on the reconstituted aggregate resulted in a significant decrease in branched proteoglycans, indicating either a weak association occurring on the aggregate, or an artifact created by a three-dimensional structure being reduced to a two-dimensional monolayer image. The shared attachments were independent of both the presence of link proteins and changes in spacing between proteoglycans, suggesting a proteoglycan-proteoglycan interaction occurring before aggregation. The interactions were not influenced by proteoglycan concentration at the time of aggregation. Link proteins, however, did increase the number of proteoglycans on the aggregate that could be cross-linked with a bifunctional reagent, suggesting that link proteins facilitate proteoglycan-proteoglycan interactions.

The effect of marginal osteophytes on reduction of varus-valgus instability in osteoarthritic knees.

The varus-valgus stability of 20 knees with unicompartmental osteoarthritis was studied in vivo at the time of total knee replacement. Intact osteoarthritic knees had an average of 11.0 degrees of varus-valgus motion. Removal of osteophytes from the osteoarthritic compartment significantly increased the motion to 13.1 degrees (P less than 0.05), while subsequent removal of osteophytes from the nonosteoarthritic compartment further increased motion to 14.7 degrees (P less than 0.025). In primarily unicompartmental osteoarthritis, marginal osteophytes appear to stabilize osteoarthritic knees, but can cause fixed deformity.

Cosmogenic chlorine-36 chronology for glacial deposits at bloody canyon, eastern sierra nevada.

Deposits from mountain glaciers provide an important record of Quaternary climatic fluctuations but have proved difficult to date directly. A chronology has been obtained for glacial deposits at Bloody Canyon, California, by measurement ofthe accumulation of chlorine-36 produced by cosmic rays in boulders exposed on moraine crests. The accumulation ofchlorine-36 indicates that episodes of glaciation occurred at about 21, 24, 65, 115, 145, and 200 ka (thousand years ago). Although the timing of the glaciations correlates well with peaks of global ice volume inferred from the marine oxygen isotope record, the relative magnitudes differ markedly. The lengths of the moraines dating from 115 ka and 65 ka show that the early glacial episodes were more extensive than those during the later Wisconsin and indicate that the transition from interglacial to full glacial conditions was rapid.

In situ cross-linking of cartilage proteoglycans.

Although the in vitro interactions between purified cartilage matrix components have been studied extensively, little is known about these interactions in situ. In this study, cartilage was treated with a cross-linking reagent with a span of 1.2 nm between its reactive terminal groups in order to preserve the native relationships between closely associated matrix components throughout extraction, purification, and preparation for electron microscopy. After in situ cross-linking, electron microscopy and gel chromatography revealed that about one-half of the guanidine hydrochloride extractable proteoglycans were polymeric, usually with two to five proteoglycan subunits in each polymer. Cross-linking consistently involved the thin segments of the proteoglycan subunits. Some of the proteoglycan polymers were capable of binding hyaluronic acid and were parts of aggregates under associative conditions. SDS-polyacrylamide gel electrophoresis revealed that link proteins were present within the polymers, and studies in which purified proteoglycans were cross-linked in vitro confirmed that the link proteins increased the proportion of polymeric proteoglycans. These findings suggest that individual proteoglycans within cartilage have intimate associations with other proteoglycans that are mediated by link proteins.

In vitro reconstruction of a cartilage matrix granule network.

The cartilage extracellular matrix contains electron-dense granules and fine filaments when studied electron microscopically after staining with ruthenium red. The matrix granules contain proteoglycans, while the filaments are thought to represent hyaluronic acid. In the present study partial extraction of proteoglycans from the cartilage prior to staining reduced the density of matrix components to allow visualization of a well-developed network involving the matrix granules and hyaluronic acid filaments. The matrix granules frequently had multiple filamentous attachments and the network appeared to be formed by intersecting filaments with the matrix granules at points of intersection. A similar network was created in Sepharose CL-2B beads when proteoglycans, link proteins, and hyaluronic acid were concentrated in the beads. Elimination of any one of these components resulted in failure to form a complete network. Purified proteoglycan monomers alone were sufficient to create matrix granules in the beads. Filaments were seen only when hyaluronic acid was added to the beads. The nature of the network suggests that some type of association between separate aggregates is occurring both within cartilage and within the Sepharose CL-2B beads.