Interleukin-1α induction of tensile weakening associated with collagen degradation in bovine articular cartilage

Abstract

Objective

To determine whether interleukin-1α (IL-1α) induces tensile weakening of articular cartilage that is concomitant with the loss of glycosaminoglycans (GAGs) or the subsequent degradation of the collagen network.

Methods

Explants of young adult bovine cartilage obtained from the superficial (including the articular surface), middle, and deep layers were cultured with or without IL-1α for 1 week or 3 weeks. Then, portions of the explants were analyzed for their tensile properties (ramp modulus, strength, and failure strain); other portions of explants and spent culture medium were analyzed for the amount of GAG and the amount of cleaved, denatured, and total collagen.

Results

The effect of IL-1α treatment on cartilage tensile properties and content was dependent on the duration of culture and the depth of the explant from the articular surface. The tensile strength and failure strain of IL-1α–treated samples from the superficial and middle layers were lower after 3 weeks of culture, but not after 1 week of culture. However, by 1 week of culture, IL-1α had already induced release of the majority of tissue GAGs into the medium, without detectable loss or degradation of collagen. In contrast, after 3 weeks of culture, IL-1α induced significant collagen degradation, as indicated by the amount of total, cleaved, or denatured collagen in the medium or in explants from the superficial and middle layers.

Conclusion

IL-1α–induced degradation of cartilage results in tensile weakening that occurs subsequent to the depletion of GAG and concomitant with the degradation of the collagen network.

The tensile integrity of articular cartilage decreases with age and in the presence of osteoarthritis (OA). In macroscopically normal adult human articular cartilage from the femoral condyles, both the tensile strength and stiffness of the superficial layer decrease ∼10% per decade of age after reaching peak values at the age of 24 years (1). In cartilage that is mildly fibrillated and osteoarthritic, tensile equilibrium moduli are reduced ∼70% and ∼85%, respectively, compared with that in young normal cartilage from the human knee joint (2). Such tensile deterioration of cartilage is detrimental to its load-bearing properties, with a resultant increase in transverse deformation in response to applied compressive load (3). Cartilage tensile properties are dependent primarily on the integrity of the collagen network (1), damage to which may result in an impairment of the normal counterbalance to proteoglycan-associated swelling (4) and lead to the cartilage swelling that occurs in OA (5, 6). Thus, the diminution of cartilage tensile integrity may facilitate progressive deterioration and development of end-stage OA.

The mechanism by which articular cartilage undergoes tensile weakening in arthritic degeneration remains to be established. In degenerate and osteoarthritic human articular cartilage from the femoral condyle, the percentage of degraded collagen (6) is higher than that in normal cartilage and shows specific alterations of the collagen network, including cleavage (7) and denaturation (8) of collagen molecules. Whether tensile weakening occurs concomitantly with collagen network degradation or with other preceding degradative processes remains to be established. Tensile softening and weakening of cartilage have been analyzed in samples subjected to extensive degradation of extracellular matrix components, including both proteoglycan and collagen (9, 10). In vitro degradation of the collagen network by the application of elastase or collagenase has been shown to result in an ∼45% (9) or ∼90% (10) decrease in the tensile strength of the superficial layer of macroscopically normal human articular knee cartilage. In naturally occurring arthritis, degradation of proteoglycan and collagen may be instigated by chemical stimuli and underlies the biomechanical weakening of the cartilage tissue.

Interleukin-1 (IL-1) is a mediator of inflammation and articular cartilage destruction in several arthritic diseases and is also a target for therapeutic intervention. IL-1 is a cytokine produced by activated synoviocytes, mononuclear cells, and chondrocytes (11–13), and levels of IL-1 are elevated in the synovial membrane, synovial fluid, and cartilage of patients with OA and in patients with rheumatoid arthritis (RA) (11, 14, 15). When injected into rabbit knee joints, IL-1 induces joint swelling, inflammation, and degradation of cartilage proteoglycan (16, 17), events that also occur in arthritis in humans. These effects were shown to be suppressed by intravenous administration of IL-1 receptor antagonist (18), with a decrease in inflammation and an inhibition of proteoglycan loss. Damage to articular cartilage may arise, in part, from direct activation of chondrocytes by IL-1.

Bovine cartilage explant cultures have been used to identify biochemical pathways by which chemical stimuli affect cartilage composition, structure, and biomechanical function and relationships between these properties. Incubation of adult bovine cartilage explants with growth factors that maintained proteoglycan content was shown to also maintain the compressive modulus (19). In contrast, incubation of adult or immature bovine cartilage explants with IL-1 induced a decrease in the compressive modulus as well as a correlated decrease in proteoglycan content (20). Treatment of adult human and bovine cartilage explants in vitro with IL-1 was shown to induce aggrecan cleavage and loss (21, 22), as well as subsequent collagen cleavage and denaturation (23, 24). This occurred through stimulation of chondrocyte synthesis and secretion of proteases (25–27) such as aggrecanase 1 (ADAMTS-4) and aggrecanase 2 (ADAMTS-5), which mediate IL-1–induced aggrecan degradation (28, 29) and the matrix metalloproteinases (MMPs) collagenase and stromelysin, which degrade proteoglycan subunits (30) and type II collagen (23, 24).

While cytokine-induced effects on the composition of the collagen network (23, 24, 27) have been studied extensively, the concomitant effect on collagen network function and cartilage tensile properties has not been studied previously. Thus, the objective of the present study was to determine whether IL-1 induces tensile weakening of articular cartilage and whether such weakening is associated temporally with the initial loss of proteoglycan or the subsequent degradation of the collagen network in the proteoglycan-depleted tissue.

MATERIALS AND METHODS

Materials and reagents.

Materials for cartilage explant isolation and culture and for extraction of degraded collagen were obtained as described previously (6, 19). Enzyme-linked immunosorbent assays (ELISAs) to detect cleaved (using the polyclonal antibody Col2-3/4C_short) and denatured (using the monoclonal antibody Col2-3/4m) collagen epitopes were obtained from Ibex (Montreal, Quebec, Canada), and recombinant human IL-1α was obtained from R&D Systems (Minneapolis, MN).

Cartilage isolation and culture.

Osteochondral fragments were obtained from the medial and lateral aspects of the patellofemoral groove of young adult (1–2 years old) bovine knee joints (n = 4). From each fragment, cartilage was cut sequentially, parallel to the articular surface, to a thickness of ∼0.3 mm to yield cartilage slices from the superficial, middle, and deep layers, with the superficial cartilage explants having an intact articular surface. From the slices, a total of 192 cartilage explants measuring 10 mm long and 5 mm wide were prepared. The cartilage explants were weighed wet and were either tested immediately for tensile properties (week 0) or were incubated in medium (Dulbecco's modified Eagle's medium with 10 mM HEPES, 0.1 mM nonessential amino acids, 0.4 mML-proline, 2 mML-glutamine, 100 units/ml of penicillin, 100 μg/ml of streptomycin, 0.25 μg/ml of amphotericin B, and 25 μg/ml of ascorbate) supplemented with 0.01% bovine serum albumin alone (basal medium) or with an additional 5 ng/ml of recombinant human IL-1α.

Cartilage explants were incubated at 37°C in an atmosphere of 5% CO₂, with 3 medium changes each week, for 7 or 21 days. The spent medium from each sample was pooled each week. After culture, each cartilage explant was weighed wet, measured for thickness, and cut into 2 portions, a tapered specimen for biomechanical testing and adjacent tissue for biochemical analysis. Biochemical measures of week 0 explants were not performed. However, based on a previous study (23), under basal conditions, tissue contents on day 0 and at day 7 would be expected to be similar for Col2-3/4m and Col2-3/4C_short and to be decreased by ∼25% for glycosaminoglycan (GAG).

Biomechanical analysis.

Tapered specimens were tested in tension, and data were analyzed to determine tensile ramp modulus, strength, and strain at failure, as described previously (31). Each tapered specimen had a gauge area of 4 mm × 0.8 mm (length by width), and specimens were elongated at a constant rate of extension (5 mm/minute) until failure, while monitoring load. Throughout mechanical testing, specimens were kept hydrated by immersion in a phosphate buffered saline bath or drip at room temperature (∼22°C) and an approximately physiologic pH (7.2). From the displacement and load data, tensile mechanical properties were determined. Stress (in MPa) was calculated as the load normalized to the initial width and thickness of the gauge region. Strain (unitless) was calculated by normalizing the displacement data to the initial sample length, which was taken as the initial grip-to-grip length. The ramp modulus was calculated as the slope of the stress–strain curve between 25% and 75% of the maximum strain, and the strength and failure strain were recorded as the stress and strain, respectively, at which the maximum stress was achieved.

Biochemical analysis.

Previous in vitro studies (23) have developed methods, which were adopted for the present study, showing that IL-1α treatment of cartilage explants induces collagen degradation, altering the collagen in both the culture medium and tissue (Figure 1). Newly synthesized and intact collagen molecules may be deposited in the tissue or released into culture medium. Such collagen molecules may be subsequently catabolized into cleaved, denatured, or fragmented forms, the latter of which diffuse easily within cartilage and out into the culture medium. Treatment of cartilage with α-chymotrypsin selectively releases cleaved and denatured collagen, but not intact collagen, and subsequent treatment of the tissue with proteinase K completely solubilizes the residual collagen. All of these forms of collagen molecules can be measured as hydroxyproline (23) in the medium, in α-chymotrypsin extracts of tissue, and in proteinase K digests of tissue. In culture medium and α-chymotrypsin extract, cleaved and denatured collagens can be delineated by competitive ELISAs with the Col2-3/4C_short and Col2-3/4m antibodies, respectively (7, 8).

Figure 1. Measures of collagen network degradation after explant culture of young adult bovine cartilage. The components present in tissue or medium fractions were intact collagen, cleaved collagen, denatured collagen, and smaller fragments of collagen. Amounts of cleaved collagen, denatured collagen, and total collagen (encircled plus signs) were quantified by enzyme-linked immunosorbent assay (ELISA) for Col2-3/4C_short epitope, ELISA for Col2-3/4m epitope, and by colorimetric assay for hydroxyproline content (33), respectively. aCT = α-chymotrypsin; ProK = proteinase K.

Based on these findings, the surrounding residual tissue was analyzed for the quantity of matrix components, including sulfated GAG as an index of proteoglycan, and cleaved and denatured collagen. The tissue was weighed wet and then treated with 1 mg/ml of α-chymotrypsin for 18 hours at 37°C to extract denatured collagen (6). The residual tissue was solubilized with 0.5 mg/ml of proteinase K for 24 hours at 60°C. The spent medium, α-chymotrypsin extracts, and proteinase K digests were analyzed for GAG (32) and hydroxyproline (33) contents. Because the amount of hydroxyproline in medium was small, medium from multiple blocks from the same animal were pooled and analyzed. The spent medium and α-chymotrypsin extracts were analyzed for cleaved (Col2-3/4C_short) and denatured (Col2-3/4m) collagen by ELISA (7, 8). Controls for each assay were prepared in the appropriate buffer for the samples being analyzed.

Hydroxyproline content was converted to collagen content using a mass ratio of collagen to hydroxyproline equal to 7.25 (34). GAG content was calculated by comparison to known concentrations of shark chondroitin sulfate. The contents of GAG, collagen, Col2-3/4C_short, and Col2-3/4m were scaled up from the amount in the tissue used for this analysis to the amount in the whole explant (based on postculture measurements of wet weight), and these quantities were then normalized to the preculture wet weight. Therefore, both the medium and tissue contents of each particular component represent the values relative to the wet weight of the initial tissue explant. Presentation of the data in this form provides values that are comparable with those normalized to the wet weight at the end of culture, since the changes in wet weights were minimal (ranging on average from –8% to 2% in the different experimental groups). The amounts of these constituents in media were reported as the amount above that in basal (day 0) culture media. The percentage of collagen in α-chymotrypsin (degraded) was calculated as that in α-chymotrypsin compared with the sum in the α-chymotrypsin and proteinase K solutions.

Statistical analysis.

Data are expressed as the mean ± SEM. The effects of culture on tensile properties were assessed by one-way analysis of variance (ANOVA), followed by Dunnett's test, with data from week 0 as the control. The effects of IL-1α on mechanical and biochemical parameters were assessed using repeated-measures ANOVA, with tissue depth (superficial, middle, deep) as a repeated factor and with 1-week and 3-week experiments being performed and analyzed separately. For analysis of medium, the week in culture (first, second, or third week) was an additional repeated factor, and data from day 0 were not included, since the data were reported as values above amounts in day 0 culture medium. When IL-1α or depth from the articular surface had an effect (for P < 0.05), planned comparisons were made between treatment groups at each depth.

RESULTS

Findings of the biomechanical analysis of bovine cartilage.

IL-1α treatment lowered the tensile integrity of cartilage samples in a manner that was dependent upon the culture duration and depth of the sample from the articular surface. In particular, the tensile integrity of IL-1α–treated samples, as compared with those incubated in basal medium as well as with the week 0 samples, was lower after 3 weeks of culture, but not after 1 week of culture. The tensile ramp modulus, strength, and failure strain of samples cultured in basal medium for 1 week (Figures 2B, E, and H) and 3 weeks (Figures 2C, F, and I) of culture were similar to those of week 0 samples (P = 0.1–0.8) (Figures 2A, D, and G).

Figure 2. Effect of interleukin-1α (IL-1α) on tensile properties of young adult bovine cartilage explants. Analyses of A–C, tensile ramp modulus, D–F, strength, and G–I, strain at failure were performed on explants of cartilage (n = 7–16) from the superficial (S), middle (M), and deep (D) layers. Samples were either tested immediately for tensile properties (week 0) (A, D, and G) or were incubated with or without 5 ng/ml of IL-1α for 1 week (B, E, and H) or for 3 weeks (C, F, and I). Values are the mean and SEM. ∗ = P < 0.05; ∗∗∗ = P < 0.005 by Dunnett's test for comparisons with week 0 samples and by planned comparisons between samples cultured with or without IL-1α. Asterisks within the bars indicate comparisons with week 0 samples; asterisks above the bars indicate comparisons between samples cultured with or without IL-1α.

After 1 week of culture, the tensile ramp modulus (Figure 2B), strength (Figure 2E), and failure strain (Figure 2H) of samples cultured in medium with IL-1α were similar to those in samples cultured in basal medium (P = 0.3, P = 0.5, and P = 0.3, respectively) as well as those in the week 0 samples (P = 0.1–0.4, P = 0.1–0.4, and P = 0.3–0.8, respectively) (Figures 2A, D, and G). The tensile strength and failure strain were dependent upon cartilage depth (P < 0.005 for each comparison), whereas the ramp modulus was not (P = 0.2), and there was no interaction effect between depth and IL-1α treatment (P = 0.5–0.8). Planned comparisons of treatment at each depth revealed no significant effect of IL-1α treatment after only 1 week of culture for ramp modulus (P = 0.2–0.5), strength (P = 0.2–0.8), or failure strain (P = 0.4–0.9) compared with samples cultured in basal medium.

In contrast, after 3 weeks of culture, treatment with IL-1α caused a decrease in tensile strength and failure strain of cartilage explants, with effects most pronounced in samples from the superficial and middle layers. Each of the tensile properties was dependent on depth (P < 0.05 for each comparison). After 3 weeks of culture, IL-1α–treated samples showed a tendency for an overall decrease in tensile strength, with a decrease (31%; P < 0.05) in superficial samples and no effect on middle (25%; P = 0.18) or deep (5%; P = 0.7) samples as compared with those cultured in basal medium (Figure 2F), and a decrease in superficial (30%; P < 0.05) and middle (33%; P < 0.05) samples, but not deep samples (23%; P = 0.6), as compared with week 0 samples (Figure 2D). Failure strain was also affected by IL-1α treatment, being lower in IL-1α–treated superficial (33%; P < 0.05) and middle (38%; P < 0.005) samples, but not deep samples (16%; P = 0.2), than in those cultured in basal medium (Figure 2I) and lower in IL-1α–treated middle (47%; P < 0.005) and deep (30%; P < 0.05) samples, but not superficial samples (30%; P = 0.1), than in week 0 samples (Figure 2G). Ramp modulus was not affected by IL-1α treatment in superficial, middle, or deep samples as compared with samples cultured in basal medium (P = 0.2–0.8) (Figure 2C) as well as with week 0 samples (P = 0.5–0.9) (Figure 2A).

Findings of the biochemical analysis of bovine cartilage.

IL-1α treatment had a significant effect on the release and retention of matrix components by cartilage explants. The effects were dependent on the culture duration and the depth of the explant. Release of matrix components into the medium was stimulated by IL-1α treatment, with the release of collagen network components being delayed relative to the release of GAG. Because the pattern of release of matrix components into the medium from samples cultured for 1 week was similar to that of matrix components released after 1 week of culture from samples that were cultured for a total of 3 weeks, data are shown only for the samples cultured for 3 weeks.

In particular, IL-1α stimulated the release of GAG from cartilage samples by 1 week of culture (P < 0.005), with an effect on the cumulative release after 2 and 3 weeks of culture (Figure 3A); this release was dependent on the culture duration (P < 0.005) and cartilage depth (P < 0.05), with an interactive effect of duration and depth (P < 0.001). After 1 week of culture, the amount of GAG released into the medium was significantly higher (91–116%; P < 0.005 for each comparison) in IL-1α–treated samples from the superficial, middle, and deep layers than in untreated controls (Figure 3A). Since GAG release from IL-1α–treated samples increased little during weeks 2 and 3, whereas untreated controls still released moderate amounts of GAG, the difference in the cumulative release of GAG between IL-1α–treated and control samples subsequently diminished during week 2 (45–70%; P < 0.05 for each comparison) and week 3 (20–45%; P = 0.082, P < 0.05, and P = 0.14, for the 3 cartilage layers, respectively).

Figure 3. Effect of interleukin-1α (IL-1α) on the cumulative release of matrix components from young adult bovine cartilage into the medium. Explants of cartilage (n = 14–16) from the superficial (squares), middle (triangles), and deep (circles) layers were cultured for 3 weeks in the presence (broken lines; open symbols) or absence (solid lines; solid symbols) of 5 ng/ml of IL-1α. Samples were then analyzed for the cumulative release of A, glycosaminoglycan (GAG), B, collagen (COL), C, cleaved collagen (Col2-3/4C_short), and D, denatured collagen (Col2-3/4m) into the medium. Values are the mean and SEM. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.005 by planned comparisons between samples cultured with or without IL-1α.

The effect of IL-1α on the release of collagen into the medium (determined from the hydroxyproline content in the medium) was delayed relative to the IL-1α–enhanced release of GAG into the medium. IL-1α induced the release of collagen from cartilage samples (P < 0.05) (Figure 3B), with the release being dependent on the culture duration (P < 0.005), but not the cartilage depth (P = 0.67), and without an interaction effect (P = 0.78). The amount of collagen released from IL-1α–treated samples during week 1 of culture was slightly higher (105–177%) than that released from samples cultured in basal medium. However, IL-1α increased the release of collagen dramatically during subsequent weeks of culture, being higher by 156–368% at week 2 and higher by 213–788% at week 3.

The effect of IL-1α on the release of the collagen degradation markers Col2-3/4C_short and Col2-3/4m epitopes into the medium was also delayed relative to the IL-1α–enhanced release of GAG into the medium. IL-1α enhanced the release of Col2-3/4C_short (Figure 3C) and Col2-3/4m (Figure 3D) epitopes (P < 0.01 for each comparison). The cumulative release of Col2-3/4C_short was dependent on the culture duration (P < 0.005) and cartilage depth (P < 0.005), with an interaction effect (P < 0.005). The cumulative release of Col2-3/4m was also dependent on the culture duration (P < 0.005), but not cartilage depth (P = 0.30), and there was no interaction effect (P = 0.07).

After 1 week of culture, the release of Col2-3/4C_short and Col2-3/4m epitopes was not affected by IL-1α treatment in samples from the superficial, middle, or deep layers (P = 0.11–0.47). In contrast, IL-1α enhanced the cumulative release of the Col2-3/4C_short and Col2-3/4m epitopes during week 2 of culture. The Col2-3/4C_short epitope release was higher in IL-1α–treated superficial layer samples (125%; P < 0.05) but not middle (P = 0.05) or deep (P = 0.3) layer samples, whereas the release of Col2-3/4m epitope was higher in IL-1α–treated superficial (596%; P < 0.005) and middle (517%; P < 0.05) layer samples, but not deep layer samples (P = 0.1). During week 3, the release of Col2-3/4C_short and Col2-3/4m epitopes was higher with IL-1α treatment for samples from the superficial (548% [P < 0.005] and 2,503% [P < 0.005], respectively) and middle (373% [P < 0.005] and 1,350% [P < 0.05], respectively) layers, but not those from the deep (P = 0.08–0.15) layer.

Consistent with the effect on the release of matrix components into the medium, IL-1α had a degradative effect on the residual collagen network in the cartilage samples that was delayed relative to the time course of the decrease in residual GAG. After 1 and 3 weeks of culture, IL-1α caused a marked decrease (P < 0.005 for each comparison) in the amount of GAG remaining in the cartilage samples (Figures 4A and B), with the amounts being dependent on the cartilage layer after the 3-week (P < 0.005), but not the 1-week (P = 0.2), culture duration. After 1 week of culture, the amount of residual GAG in IL-1α–treated samples (Figure 4A) was 53–75% lower in superficial, middle, and deep samples (P < 0.05 for each comparison) compared with samples cultured in basal medium. The amount of GAG remaining in the IL-1α–treated samples after 3 weeks of culture (Figure 4B) was considerably lower in superficial (79%; P < 0.005), middle (81%; P < 0.005), and deep (74%; P < 0.005) samples. Approximately 94% of the total GAG from samples cultured in basal medium and 85% from samples cultured in IL-1α were α-chymotrypsin–extractable.

Figure 4. Effect of interleukin-1α (IL-1α) on the amount of matrix components remaining in young adult bovine cartilage explants. Analyses of A and B, glycosaminoglycan (GAG), C and D, collagen (COL), E and F, collagen in α-chymotrypsin (COL in aCT), G and H, Col2-3/4C_short, and I and J, Col2-3/4m were performed on explants of cartilage (n = 7–16) from the superficial (S), middle (M), and deep (D) layers. Samples were cultured for 1 week (A, C, E, G, and I) or for 3 weeks (B, D, F, H, and J) in the presence or absence of 5 ng/ml of IL-1α. Degraded collagen was first extracted with α-chymotrypsin (striped bars), and residual tissue was digested with proteinase K (solid bars). Values are the mean ± SEM. ∗ = P < 0.05; ∗∗∗ = P < 0.005 by planned comparisons between samples cultured with or without IL-1α.

While it had no detectable effect on the overall amount of collagen remaining in the cartilage samples, IL-1α caused an increase in the percentage of collagen in the α-chymotrypsin fraction in a manner that was also delayed relative to the time course of the decrease in GAG, but that paralleled the time course of the decrease in tensile integrity. The amount of collagen remaining in the cartilage samples after 1 and 3 weeks of culture was similar between samples cultured in basal medium and samples cultured in medium with IL-1α (P = 0.9 and P = 0.3, respectively) (Figures 4C and D). However, the percentage of α-chymotrypsin–extractable collagen was significantly higher after 3 weeks (P < 0.005) (Figure 4F), but not after 1 week (P = 0.1) (Figure 4E), of IL-1α treatment. This was dependent on cartilage depth for the samples cultured for 1 week and 3 weeks (P < 0.005 and P < 0.01, respectively). The percentage of collagen in the α-chymotrypsin fraction was higher after the 3-week culture duration in samples from the superficial (102%; P < 0.005) and middle (39%; P < 0.005) layers, but not the deep layer (P = 0.3) samples (Figure 4F). The percentage of collagen in α-chymotrypsin was not higher after the 1-week culture duration in superficial, middle, or deep samples (P = 0.6–0.9) (Figure 4E).

IL-1α treatment had a differential effect on the amount of Col2-3/4C_short and Col2-3/4m epitopes found in the α-chymotrypsin extracts, with the delay of IL-1α–induced collagen network degradation being evident between 1-week and 3-week cultures. The presence of the collagen cleavage marker Col2-3/4C_short was not higher in samples cultured in IL-1α for 1 week (P = 0.2) (Figure 4G) or for 3 weeks (P = 0.06) (Figure 4H). After the 3-week culture duration, the amount of Col2-3/4C_short was higher in IL-1α–treated superficial layer samples (95%; P < 0.05), but not in middle or deep layer samples (P = 0.1–0.6). The presence of the Col2-3/4m epitope, though dependent on cartilage layer (P < 0.05 for each comparison), was not increased by IL-1α treatment after the 1-week (P = 0.9) (Figure 4I) or 3-week (P = 0.6) (Figure 4J) culture duration.

DISCUSSION

IL-1α induced tensile weakening of, and the release of GAG and collagen components from, adult bovine cartilage explants, with characteristic time courses. The IL-1α–induced loss of GAG was not associated with a loss of tensile integrity, whereas alterations of the collagen network were. Initially (within the first week of culture), IL-1α reduced the amount of GAG remaining in the matrix (Figure 4A), without an associated effect on the collagen network (Figures 4C, E, G, and I) or a detectable decrease in tensile strength or failure strain compared with either week 0 samples (Figures 2D and G) or samples cultured in basal medium (Figures 2E and H). In contrast, after 3 weeks of culture, IL-1α treatment caused a decrease in tensile strength (Figures 2D and F) and failure strain (Figures 2G and I), as well as degradation of the collagen network. In particular, IL-1α induced an increase in the percentage of degraded collagen (Figure 4F) and cleaved collagen (Figure 4H), while the amount of GAG in the cartilage samples remained low (Figure 4B) and the total amount of collagen was not affected (Figure 4D). These results indicate that IL-1α induced degradation of collagen molecules, but not IL-1α–triggered depletion of proteoglycan, results in tensile weakening of articular cartilage.

Certain factors may limit the interpretation of the results of this study and the assessment of the relationships between alterations of the extracellular matrix and the mechanical integrity of articular cartilage. We used articular cartilage from adult animals, rather than animals undergoing rapid growth (35), because the former is more similar in tissue organization and composition to the human adult cartilage that is affected by OA (36), and the adult bovine knee joint offers a broad, relatively flat surface, which in contrast to cartilage from older animals or cartilage from humans, is generally free from degeneration.

However, certain compositional, structural, and functional differences between human and bovine articular cartilage do exist. Adult bovine articular cartilage has lower tensile stiffness and strength (31, 35, 37) than adult human cartilage (1, 2). The depth-dependent variation in tensile strength and stiffness of young adult bovine articular cartilage (31) is also somewhat different from that of adult human cartilage (1). In contrast, the failure strain (31, 37) of adult bovine cartilage is consistent in magnitude and depth-variation with that of young adult human cartilage (38). Generally, the collagen content is slightly lower and the GAG content slightly higher in bovine (35) than in human (38, 39) articular cartilage, and the crosslink composition of the collagen network can vary, with hydroxypyridinoline crosslinks being less abundant in adult bovine cartilage (35) than in adult human cartilage (3). Despite these differences, bovine cartilage has been used as a model to study age-related changes in normal human adult articular cartilage (31).

In addition, the properties of superficial, middle, and deep layers in this study do not necessarily represent properties of the superficial, middle, and deep “zones” as defined classically (36). The thickness of specimens was chosen to be similar to that used in previous studies (1, 10, 31, 35) to allow for direct comparisons, and samples including the articular surface (superficial region) were used because of its importance to the tensile properties of articular cartilage (38) and its sensitivity to aging (1), degeneration (2, 40), and the degradative effects of IL-1 (41).

Treatment of cartilage with IL-1α induced a number of the changes that are also observed in various types of arthritis. IL-1 is abnormally elevated in diseases such as OA and RA, and it is often used in culture systems to identify cascades of enzymatic activity associated with OA, such as that of the aggrecanases (42, 43) and MMPs (7, 24, 44). The addition of IL-1α to cartilage explant cultures resulted in a loss of matrix components similar to that which occurs in arthritis, although in the absence of mechanical loading and the complicating effects of surrounding joint fluids and tissues. The relative amounts of cleaved and denatured collagen as well as GAG released into the media (Figure 3) and remaining in the residual cartilage samples (Figure 4) following culture with IL-1α were consistent with those determined previously, as was the time course of release (23, 24). Indeed, the delay of collagen degradation compared with the loss of GAG may be due to the GAG shielding of collagen epitopes to cleavage (24, 27), as opposed to the delay in up-regulation of MMPs as compared with the up-regulation of aggrecanases (45). Using IL-1α in the culture of cartilage tissue explants in the present study allowed the biologic study of the relationship between cytokine-induced matrix degradation and alteration of mechanical integrity.

The major result of the present study was that, in such a biologically triggered system, collagen degradation in a GAG-depleted tissue, as opposed to GAG loss alone, is associated with tensile weakening of cartilage. The decrease in tensile strength seen in this study seems likely to be due to the alterations of the collagen network and highlights the importance of studies of collagen degradation in arthritis (7, 8). IL-1α treatment for 3 weeks resulted in a 40–100% higher percentage of α-chymotrypsin–extractable collagen in superficial and middle layer samples (Figure 4F). This increase is similar to the difference in the percentage of α-chymotrypsin–extractable collagen between normal and fibrillated cartilage (6).

The IL-1α–induced degradation of collagen was associated with specific alterations to the collagen network. In particular, the amount of Col2-3/4C_short was 95% higher in samples from the superficial layer and mildly higher (26–30%) in samples from the middle and deep layers (Figures 4H and J), differences similar to that (∼100%) between normal and OA cartilage (7). There was, however, no difference in the amount of denatured collagen (Col2-3/4m) between IL-1α–treated and control samples, which is in contrast to the higher amount (∼100%) present in OA cartilage compared with normal tissue (7, 8). Indeed, this may be why the tensile strength was only 28–31% lower in IL-1α–treated samples from the superficial and middle layers after 3 weeks of culture, whereas fibrillated cartilage has a tensile strength that is ∼50% lower than that of normal cartilage (40). While additional or other forms of enzymatic or mechanical cartilage degradation may contribute to the tensile weakening observed in OA, IL-1α–induced matrix degradation highlighted the role of collagen degradation in cartilage tensile weakening.

The association between alterations in cartilage tensile properties and collagen network structure induced by IL-1α in this study was consistent with the findings of previous studies of the tensile properties of native and degraded cartilage. The values for week 0 samples and samples cultured for 1 or 3 weeks in basal medium were consistent with the properties of mature adult bovine articular cartilage from the patellofemoral groove (31, 35). The dramatic decrease (45–90%) in the tensile strength of the superficial layer of human articular cartilage following treatment with collagen-degrading enzymes is consistent with the role of the collagen network (9, 10), although the extent of collagen cleavage and degradation was not determined in those studies. In contrast, the tensile strength of articular cartilage was maintained with enzymatic treatment targeting proteoglycan aggregate constituents (10, 46, 47), although such treatments and ionic alterations can affect viscoelastic tensile behavior (2, 10, 47).

Sufficiently long and intense exposure of articular cartilage explants to enzymatic degradation appears to be needed for tensile weakening. Indeed, the lack of effect of IL-1α on the deep layer tensile properties and residual collagen network may reflect a protective effect of proteoglycan on the collagen network, since collagen network degradation by collagenase is enhanced in an aggrecan-depleted matrix, and inhibition of aggrecanase blocks IL-1–stimulated collagen cleavage (27). The zone-specific effect may also be related to the more potent effect of IL-1 on chondrocytes from the superficial zone of cartilage than those of the deep zone (41).

Further analysis of the pathway linking biologic stimuli to biomechanical deterioration of articular cartilage may elucidate the role of specific factors, and their inhibitors, in the advancement of arthritic disease. Together with other cytokines and mechanical factors, IL-1 may contribute to functional deterioration of the collagen network of cartilage in OA. For example, it has been found that activation of latent MMPs in immature cartilage results in compressive softening as well as hypotonic swelling of the tissue that can be markedly inhibited by application of tissue inhibitor of metalloproteinases 1 (48). Other interesting targets may be receptors for advanced glycation end products and their ligands, which are present in the synovium of OA patients and increase the expression of MMPs (49). Cytokine-induced degradation may be the cause of the alteration in matrix metabolism and content of the cartilage extracellular matrix, both near and far from cartilage lesions (50, 51) and, thus, trigger weakening of cartilage throughout the joint. Natural or synthetic mediators or inhibitors of matrix-degrading enzymes may be tested to elucidate molecular pathways that lead to tensile biomechanical dysfunction of cartilage and to identify potential therapies that might prevent functional deterioration of cartilage in arthritis.