Mechanisms of cartilage growth: Modulation of balance between proteoglycan and collagen in vitro using chondroitinase ABC

Authors


Abstract

Objective

To examine the cartilage growth–associated effects of a disruption in the balance between the swelling pressure of glycosaminoglycans (GAGs) and the restraining function of the collagen network, by diminishing GAG content prior to culture using enzymatic treatment with chondroitinase ABC.

Methods

Immature bovine articular cartilage explants from the superficial and middle layers were analyzed immediately or after incubation in serum-supplemented medium for 13 days. Other explants were treated with chondroitinase ABC to deplete tissue GAG and also either analyzed immediately or after incubation in serum-supplemented medium for 13 days. Treatment- and incubation-associated variations in tissue volume, contents of proteoglycan and collagen network components, and tensile mechanical properties were assessed.

Results

Incubation in serum-supplemented medium resulted in expansive growth with a marked increase in tissue volume that was associated with a diminution of tensile integrity. In contrast, chondroitinase ABC treatment on day 0 led to a marked reduction of GAG content and enhancement of tensile integrity, and subsequent incubation led to maturational growth with minimal changes in tissue volume and maintenance of tensile integrity at the enhanced levels.

Conclusion

The data demonstrate that a manipulation of GAG content in articular cartilage explants can distinctly alter the growth phenotype of cartilage. This may have practical utility for tissue engineering and cartilage repair. For example, the expansive growth phenotype may be useful to fill cartilage defects, while the maturational growth phenotype may be useful to induce matrix stabilization after filling defect spaces.

Articular cartilage normally functions as a low-friction, wear-resistant, load-bearing material that facilitates joint motion (1). The functional mechanical properties of this connective tissue are attributed to 2 of the molecular components of its extracellular matrix, proteoglycan and collagen (2, 3). The proteoglycan constituent imparts a fixed negative charge to the tissue that increases the propensity of the tissue to swell and to resist compressive loading (4, 5). The crosslinked collagen network resists the swelling tendency of the proteoglycans and provides the tissue with tensile and shear stiffness and strength (6, 7). The balance between the swelling propensity of proteoglycan molecules and the restraining function of the collagen network governs cartilage hydration and its load-bearing biomechanical function (8). For example, in osteoarthritic cartilage, the increased hydration and loss of mechanical integrity are due to a weakened collagen network that can no longer resist normal levels of swelling pressure (9).

Growth and remodeling are biologic processes that together transform cartilage tissue in vivo from an immature to a mature state. Tissue growth is generally defined as an increase in tissue size due to accretion of one or more solid tissue components similar to those already present, while tissue remodeling is defined as a change in tissue composition and/or structure of tissue components (10, 11). Many tissues, including articular cartilage, can expand due to accretion of fluid; however, this process is not generally considered growth, but rather tissue swelling. Two distinct mechanisms of tissue growth have been recognized: appositional growth, or growth at a tissue surface, and interstitial growth, or growth within tissue volume (12). While it is possible that tissues can grow appositionally in the absence of remodeling, interstitial tissue growth must involve both growth and remodeling, since accretion of a single tissue component will change the overall tissue structure and mechanical properties.

Tissues can exhibit growth and remodeling in the form of hyperplasia (increase in the number of cells), hypertrophy (addition of cellular components), deposition of the extracellular matrix components, or any number of these processes occurring concomitantly. In articular cartilage, the incidence of cell division is low and matrix deposition is the major contributor to the increase in size (i.e., growth) and changes in biochemical composition (i.e., remodeling) of this tissue in vivo (13–15). Since articular cartilage tissue may undergo both appositional and interstitial growth, and the major contributor to cartilage growth is matrix deposition, the term “growth” is used subsequently in this article to refer collectively to both growth and remodeling, which can occur in the presence or absence of cellular proliferation.

The metabolic balance between proteoglycan molecules and the components of the collagen network may be responsible for the evolution of cartilage function, structure, and composition during growth in vivo and during growth stimulated by serum in vitro (11). Fetal and postnatal growth of the articular cartilage normally involves a net deposition of collagen that is greater than that of proteoglycan, as well as an increase in mechanical integrity. During maturation of articular cartilage from fetus to skeletal maturity, there is an increase in collagen and pyridinoline (Pyr) crosslink densities, but little or no change in the content of glycosaminoglycans (GAGs) (14, 16–18). These biochemical changes are accompanied by an increase in tensile modulus and strength (15). In contrast to this type of in vivo growth and maturation, growth of cartilage explants in vitro in serum-supplemented medium results in a net deposition of proteoglycan that is greater than that of collagen, along with a decrease in mechanical integrity. For explants of immature cartilage, incubation in serum-supplemented medium results in an increase of tissue volume, maintenance of proteoglycan concentration, and a decrease in the concentrations of collagen and Pyr crosslinks (19–21). These changes in composition are associated with a decrease in tensile modulus and strength (21).

Based on the known phenotypes of in vivo growth and maturation and in vitro growth during serum-supplemented culture of immature cartilage, we hypothesized that cartilage growth can result from a dynamic imbalance between the swelling pressure of endogenous (that present at the time of explant) and newly synthesized GAG and the restraining function of the collagen network. One way to test this hypothesis is to compare the growth of cartilage explants, some of which are depleted of proteoglycan before subsequent culture and should therefore have a reduced propensity to grow. A number of studies have used chondroitinase ABC to deplete cartilage of GAG and to investigate the immediate effects on both mechanical properties and biologic responses to such perturbation of the tissue. Chondroitinase ABC specifically depolymerizes chondroitin and dermatan sulfate (22), while collagen, the collagen network arrangement, keratan sulfate, and link protein remain unaffected by the enzyme (23–25). In response to chondroitinase ABC treatment, altered mechanical properties include a reduction of the initial slope of the stress-strain curve (26, 27) and an increase in the overall tensile modulus (28, 29). The biologic response of chondrocytes to chondroitinase ABC treatment appears to be minimal; no changes were detected in cell density and volume (30) or in gene expression of certain collagens, aggrecan, and decorin (31).

To address the hypothesis of the present study, the objective was to examine the effect of depletion of GAG content in immature bovine articular cartilage (by enzymatic treatment using chondroitinase ABC) on culture-associated variations in 1) tissue size, 2) accumulation of GAG and collagen network components, and 3) tensile mechanical properties.

MATERIALS AND METHODS

Sample preparation and culture

Articular cartilage was harvested from the patellofemoral groove of 3 newborn (1–3-week-old) bovine calves as described previously (21). Blocks, 9 mm × 3 mm in area, were prepared using a sledge microtome either to include the intact articular surface (the superficial layer, ∼0.4 mm thick) or to include the middle zone, starting at a distance of ∼0.6 mm from the articular surface (the middle layer, 0.25 mm thick). The term layer rather than zone is used to avoid associations with classic zonal classification of articular cartilage. Blocks were weighed wet (initial wet weight [WWi]) under sterile conditions.

Some blocks were either analyzed immediately or incubated in medium (Dulbecco's modified Eagle's medium supplemented with 20% fetal bovine serum [FBS] and 100 μg/ml of ascorbate) for 13 days. Other blocks were treated in medium with 2 units/ml protease-free chondroitinase ABC (Seikagaku America, East Falmouth, MA) for 4 hours at 37°C to deplete GAG by >90%. After treatment, the samples were washed in medium and then either analyzed or incubated in medium for 13 days. During the first 12 days of culture, medium included 4.5 μCi/ml 3H-proline and 1.8 μCi/ml 35S-sulfate. To remove unincorporated isotopes, blocks were washed, transferred to a new culture plate, and incubated for an additional day in medium without radiolabel. Spent medium was collected throughout the culture duration. At termination, blocks were weighed wet (final wet weight [WWf]) and punched to form a tapered tensile test specimen and residual cartilage.

Biochemical analysis

The residual cartilage and failed portions of the corresponding tensile strip (see below) were analyzed together to quantify the biochemical composition of the explants. Samples were lyophilized, weighed dry, and solubilized with proteinase K (21). Portions of the tissue digest were analyzed to quantify the content of DNA (32), GAG (33), and hydroxyproline (34). Digest portions from each animal/layer/experimental condition were pooled for analysis of Pyr (35). DNA was converted to cell number by using a conversion constant of 7.7 pg of DNA per cell (36). Hydroxyproline was converted to collagen by assuming a mass ratio of collagen:hydroxyproline equal to 7.25:1 (17, 37). Portions of spent medium were analyzed for GAG (33). Biochemical parameters were normalized to WWi of the tissue to represent constituent content and to WWf of the tissue to represent constituent concentration.

Histologic analysis

Histologic analysis of GAG was performed on tissue cryosections using Alcian blue dye. Tissue was embedded in OCT compound and snap-frozen by immersion in isopentane cooled with liquid nitrogen. Tissue was then sectioned normal to the articular surface at 5 μm, stained with Alcian blue (0.002% Alcian blue in 0.4M MgCl2, 0.025M sodium acetate, 2.5% glutaraldehyde [pH 5.6]), and imaged at 20× magnification with an inverted Eclipse TE 300 microscope (Nikon, Melville, NY) equipped with a SPOT RT camera (Diagnostic Instruments, Burlingame, CA).

Analysis of matrix metabolism

To assess matrix metabolism, other portions of the tissue digest and portions of the medium were analyzed for incorporated radioactivity. The contents of 35S and 3H radioactivity were determined in the tissue digest and the medium and used to estimate the absolute rates of sulfate incorporation and hydroxyproline formation as indices of sulfated GAG and collagen synthesis, respectively. A portion of spent medium was analyzed for content of 35S-GAG using the Alcian blue precipitation method (38). For 3H-hydroxyproline analysis using Dowex 50WX8 columns (Sigma, St. Louis, MO) (19, 39), portions of tissue digests and medium were pooled for each animal/layer/experimental condition.

Biomechanical analysis

Tapered tensile specimens were analyzed to determine mechanical properties as described previously (40). The thickness of the 0.8-mm–wide gage region of each tensile strip was measured using a contact-sensing micrometer. Tensile specimens were secured in clamps (4.0 mm apart) of a mechanical tester and elongated at a constant rate (5 mm/minute) until failure. Structural tensile parameters were obtained from the load-displacement curves. Structural tensile strength was determined as the maximum load sustained at failure. The displacement at failure was the displacement at which the maximum load was attained. Ramp stiffness was calculated as the slope of the linear regression of the load-displacement curve from 25% to 75% of the maximum load. Load and displacement were converted to stress (load normalized to the cross-sectional area of the gage region) and strain (elongation distance normalized to the initial gage length) to obtain material tensile parameters of tensile strength, strain at failure, and ramp modulus.

Statistical analysis

For each layer (superficial and middle), the effects of chondroitinase ABC treatment (untreated, treated with chondroitinase ABC) and culture duration (0 and 13 days) were assessed by two-way analysis of variance (ANOVA) with the donor animal as a random factor. For dependent variables where data for certain grouping factors were not applicable (e.g., sulfate incorporation), one-way ANOVA was used to examine the effect of experimental conditions. Tukey's post hoc testing was performed to compare groups. To analyze the effect of culture duration on wet weight, repeated-measures ANOVA was performed for each layer and experimental condition with wet weight (WWi, WWf) as a repeated factor. Data are expressed as the mean ± SEM. P values less than 0.05 were considered significant.

RESULTS

The extent of in vitro growth of articular cartilage blocks was markedly affected by chondroitinase ABC treatment as assessed by changes in thickness and wet weight (Figures 1A and B). In each layer, thickness and wet weight increased during culture (P < 0.001). While chondroitinase ABC treatment did not have an independent effect on thickness in the superficial layer (P = 0.16) and had only minor effects in the middle layer (P < 0.05), and there was interactive effect between chondroitinase ABC treatment and culture duration in each layer (P < 0.01). Treatment of cartilage with chondroitinase ABC on day 0 did not affect tissue thickness (P > 0.63), and, consistent with this, the change in wet weight of these explants was near 0% (P > 0.19). Untreated explants grew during culture, both in thickness (96% in the superficial layer, 51% in the middle layer; P < 0.001) and in wet weight (109% in the superficial layer, 57% in the middle layer; P < 0.001). In contrast, during culture after chondroitinase ABC treatment, explants from the middle layer did not grow (P = 0.70 for thickness and P = 0.11 for wet weight) and explants from the superficial layer grew (30% in thickness [P = 0.09] and 44% in wet weight [P < 0.001]), but the extent of this growth was less than that of the untreated explants (23% in thickness [P < 0.05] and 65% in wet weight [P < 0.001]).

Figure 1.

Effect of chondroitinase ABC (C-ABC) pretreatment on general indices of in vitro growth of calf articular cartilage explants from the superficial (S) and middle (M) layers. On the day of explantation, some blocks were treated with chondroitinase ABC. Blocks were either analyzed on day 0 (d0) or incubated for 13 days in medium (d13). Shown are A, thickness, B, change in wet weight, and C, percent water of cartilage blocks. Values are the mean and SEM (n = 6–9 blocks from 3 animals). ∗ = P < 0.05 versus d0 without chondroitinase ABC pretreatment; # = P < 0.05 versus d0 chondroitinase ABC; + = P < 0.05 versus d13 without chondroitinase ABC pretreatment. n/a = not applicable.

The content of water (Figure 1C) did not vary with chondroitinase ABC treatment and culture duration in the middle layer, but an interactive effect was observed in the superficial layer (P < 0.01). Treatment with chondroitinase ABC did not affect water content on day 0 (P > 0.73), but during subsequent incubation of these explants it decreased slightly (2%; P < 0.05). The content of water of the untreated explants of the superficial layer did not change during culture (P = 0.42), and on day 13 was slightly higher (by 2%; P < 0.05) than that in the superficial layer of the chondroitinase ABC–treated explants.

The content of cells (cells/WWi) (Figure 2A) did not vary with chondroitinase ABC treatment or culture duration (P > 0.14), and when normalized to WWf to give a measure of the concentration of cells (cells/WWf) (Figure 2A), it reflected the culture-associated changes in wet weight. Cells/WWf decreased during culture (P < 0.01). While chondroitinase ABC treatment had only a minor independent effect (P < 0.05), an interactive effect between chondroitinase ABC treatment and culture duration was found (P = 0.09 in the superficial layer, P < 0.01 in the middle layer). On day 0, treatment with chondroitinase ABC did not affect cells/WWf (P > 0.86). During culture of explants that exhibited volumetric growth, cells/WWf decreased (by 30–45% in the untreated explants [P < 0.001] and by 26% in the superficial layer of chondroitinase ABC–treated explants [P < 0.01]); cells/WWf did not change in explants that did not grow (P = 0.37 in the middle layer of chondroitinase ABC–treated explants).

Figure 2.

Effect of chondroitinase ABC pretreatment on changes in the biochemical composition of calf articular cartilage explants from the superficial and middle layers. Shown are A, the number of cells, B, glycosaminoglycan (GAG) content, C, collagen (COL) content, and D, pyridinoline crosslink (PYR) levels. To the left of the dotted line, data are expressed per initial wet weight (WWi), since for these explants final wet weight (WWf) equals or is similar to WWi. To the right of the dotted line, data are expressed per WWi (bar with an upward error bar) to represent constituent content and per WWf (bar with a downward error bar) to represent constituent concentration. Values are the mean and SEM (n = 6–9 blocks from 3 animals). ∗ = P < 0.05 versus d0 without chondroitinase ABC pretreatment; # = P ≤ 0.05 versus d0 chondroitinase ABC; + = P < 0.05 versus d13 without chondroitinase ABC pretreatment. See Figure 1 for other definitions.

The extent of volumetric growth was generally paralleled by variations in the tissue contents of GAG and collagen, but not in the tissue content of Pyr. In each layer, GAG content (GAG/WWi) (Figure 2B) was reduced by chondroitinase ABC treatment and increased during culture, with an interactive effect (P < 0.001). On day 0, treatment with chondroitinase ABC reduced GAG/WWi by 93–95% (P < 0.001). During culture of the untreated explants, GAG/WWi increased markedly (36–59%; P < 0.01). Similarly, during culture of chondroitinase ABC–treated explants, GAG/WWi increased from very low levels to levels higher than (64% in the superficial layer; P < 0.001) or similar to (P = 1.0 in the middle layer) those at explant, and similar to (P > 0.48 in the superficial layer) or lower than (26% in the middle layer; P < 0.001) those in the untreated explants on day 13. The content of collagen (collagen/WWi) (Figure 2C) increased during culture (P < 0.001 in each layer) and decreased with chondroitinase ABC treatment in the superficial layer (P < 0.01), without interactive effects (P > 0.25). Treatment with chondroitinase ABC did not affect collagen/WWi on day 0 (P > 0.28). During culture of both the untreated and chondroitinase ABC–treated explants, collagen/WWi increased (42–43% and 25–40%, respectively; P < 0.01). On day 13, collagen/WWi of the chondroitinase ABC–treated and subsequently cultured explants was similar to that of the untreated and cultured explants (P > 0.11). The cartilage Pyr content (Pyr/WWi) (Figure 2D) did not vary with chondroitinase ABC treatment or culture duration (P > 0.26).

When the content of extracellular matrix components was normalized to WWf to give an index of concentration in the tissue, the concentrations of GAG, collagen, and Pyr reflected the changes in wet weight during culture. In each layer, the concentration of GAG (GAG/WWf) (Figure 2B) was reduced by chondroitinase ABC treatment, and increased during culture, with an interactive effect (P < 0.001). During culture of the untreated explants, GAG/WWf decreased slightly (13–25%; P < 0.05). During culture of chondroitinase ABC–treated explants, GAG/WWf increased (P < 0.001) to levels similar to those in freshly explanted cartilage (P > 0.61) and in untreated explants in the middle layer (P = 0.39), or higher than those in the untreated explants in the superficial layer (by 43%; P < 0.001). Histochemical staining confirmed the quantitative changes in GAG/WWf, with uniform initial depletion by chondroitinase ABC and subsequent restoration seen on day 13 (Figure 3).

Figure 3.

Effect of chondroitinase ABC pretreatment on localization of glycosaminoglycan in calf articular cartilage explants from the superficial and middle layers. Tissue sections were stained with Alcian blue for analysis of untreated (A, B, E, and F) and chondroitinase ABC–treated (C, D, G, and H) explants on day 0 (AD) and day 13 (EH). Bar = 100 μm. See Figure 1 for definitions.

The concentration of collagen (collagen/WWf) (Figure 2C) decreased during culture in the superficial layer (P < 0.001) and increased with chondroitinase ABC treatment in the middle layer (P < 0.001), with an interactive effect in each layer (P < 0.05). During culture of the untreated explants, collagen/WWf either decreased (32% in the superficial layer; P < 0.001) or did not change (P = 0.75 in the middle layer). In contrast, during culture of chondroitinase ABC–treated explants, collagen/WWf remained unchanged (P > 0.86 in the superficial layer) or increased (17% in the middle layer; P = 0.05). As a result, on day 13, chondroitinase ABC–treated explants had a higher collagen/WWf (19% in the superficial layer [P = 0.12] and 50% in the middle layer [P < 0.01]) than the untreated explants. The cartilage Pyr concentration (Pyr/WWf) (Figure 2D) did not vary with chondroitinase ABC treatment (P > 0.41) or culture duration in the middle layer (P = 0.24), but there was a slight effect of culture duration in the superficial layer (P = 0.06). During culture of both the untreated and chondroitinase ABC–treated explants, Pyr/WWf decreased (39% and 12%, respectively).

Variations in the content of GAG in tissue and medium and collagen content in the tissue were generally paralleled by similar variations in the total incorporation of sulfate (sum of sulfate incorporation in tissue and medium) and hydroxyproline formation, respectively. Chondroitinase ABC treatment stimulated the total sulfate incorporation, as well as individual incorporation in tissue and medium (Figure 4A) during subsequent culture in the superficial layer (25–36%; P ≤ 0.08) but not in the middle layer (P > 0.22). Also, chondroitinase ABC treatment led to a decrease in GAG release to the culture medium (23% in the superficial layer [P = 0.12] and 71% in the middle layer [P < 0.001]). The formation of hydroxyproline (Figure 4B) did not vary with chondroitinase ABC treatment (P > 0.34).

Figure 4.

Effect of chondroitinase ABC pretreatment on sulfate incorporation and hydroxyproline (HYPRO) formation in calf articular cartilage explants from the superficial and middle layers. Dotted (upper) regions of bars represent sulfate (A) and hydroxyproline (B) released from tissue into the culture medium; solid (lower) regions of bars represent sulfate incorporation (A) and hydroxyproline formation (B) in the tissue. The overall bar height represents the total incorporated sulfate (A) and the total formed hydroxyproline (B). Values are the mean and SEM (n = 6–9 blocks from 3 animals). + = P ≤ 0.08 versus superficial zone on day 13 without chondroitinase ABC pretreatment. WWi = initial wet weight (see Figure 1 for other definitions).

Chondroitinase ABC treatment also affected the tensile biomechanical behavior of cartilage explants, as demonstrated by the average load-displacement and stress-strain curves and the mechanical properties derived from these curves for individual samples (Figures 5 and 6). Ramp stiffness and strength increased with treatment (P < 0.01) and decreased with culture duration in the middle layer (P < 0.05) but not in the superficial layer (P > 0.97), with an interactive effect (P < 0.01) in each layer. On day 0, chondroitinase ABC treatment led to a marked increase in ramp stiffness (110–190%; P < 0.001) in each layer and an increase in strength in the superficial layer (78%; P < 0.001) but not in the middle layer (P > 0.48), while displacement at failure decreased (32–38%; P < 0.001) in each layer. During subsequent culture of these explants, the mechanical properties were maintained (P > 0.24), so that on day 13 ramp stiffness and strength were higher (100–190% [P < 0.01] and 29–91% [P < 0.05], respectively) and displacement at failure was lower (29–35%; P < 0.05) than those at the time of explant. During culture of the untreated explants, mechanical properties were maintained (P > 0.62) in the superficial layer, but in the middle layer ramp stiffness and strength decreased (50–59%; P < 0.05). As a result, on day 13, ramp stiffness and strength were higher (216–390% and 124–155%, respectively; P < 0.01) and displacement at failure was lower (18–37%; P < 0.05) in chondroitinase ABC–treated explants than in untreated explants.

Figure 5.

Effect of chondroitinase ABC pretreatment on changes in structural tensile mechanical behavior and properties of calf articular cartilage explants. A and B, Load and displacement at failure of cartilage blocks in the superficial layer (A) and middle layer (B). Average load-displacement profiles preceding the failure strength of the weakest sample within a group are displayed for chondroitinase ABC–treated and untreated explants, which were analyzed on day 0 or on day 13. The load-displacement profiles are then extended (dotted lines) until the average maximum strength and failure displacement for each group (larger symbols). C, Ramp stiffness of cartilage blocks. Values are the mean and SEM (n = 6–9 blocks from 3 animals). ∗ = P < 0.05 versus d0 without chondroitinase ABC pretreatment; + = P < 0.05 versus d13 without chondroitinase ABC pretreatment for load (above label) and displacement at failure (below label). See Figure 1 for definitions.

Figure 6.

Effect of chondroitinase ABC pretreatment on changes in material tensile mechanical behavior and properties of calf articular cartilage explants. A and B, Stress and strain at failure of cartilage blocks in the superficial layer (A) and middle layer (B). Average stress-strain profiles preceding the failure stress of the weakest sample within a group are displayed for chondroitinase ABC–treated and untreated explants, which were analyzed on day 0 or on day 13. The stress-strain profiles are then extended (dotted lines) until the average maximum strength and failure strain for each group (larger symbols). C, Ramp modulus of cartilage blocks. Values are the mean and SEM (n = 6–9 blocks from 3 animals). ∗ = P < 0.05 versus d0 without chondroitinase ABC pretreatment; + = P < 0.05 versus d13 without chondroitinase ABC pretreatment for stress (above label) and strain at failure (below label). See Figure 1 for definitions.

The average stress-strain curves and the calculated material mechanical properties (Figure 6) accentuated the above-described trends in the structural mechanical properties. In addition, in the untreated explants of the superficial layer, both ramp modulus and strength decreased during culture (58–65%; P < 0.05), similar to variations in ramp modulus and strength in the middle layer.

DISCUSSION

The data presented here demonstrate 2 distinct in vitro growth phenotypes in explants of immature bovine articular cartilage, as indicated by changes in tissue size, content of matrix components, and the integrity of the collagen network. In the case of the first phenotype, incubation with 20% FBS resulted in expansive cartilage growth (Figure 7B) characterized by a marked increase in tissue thickness and wet weight (Figures 1A and B) and in the content of GAG and collagen (Figures 2B and C). The result was a reduction in the concentrations of GAG, collagen, and Pyr (Figures 2B–D) and a concomitant diminution in tensile mechanical integrity (Figures 5 and 6), similar to previous findings on growth of immature cartilage in serum-supplemented medium (19–21). These findings suggest that a free-swelling, serum-supplemented culture environment induces remodeling and reorganization of the tissue matrix that facilitates a relatively loose and weak collagen network and allows tissue expansion.

Figure 7.

Schematic of the manipulation of glycosaminoglycan (GAG) content and metabolism of GAG and collagen network (CN) components leading to different types of cartilage growth. Shown is the initial state (A) with the major components of the solid extracellular matrix. The expansive growth phenotype (B) is characterized by deposition of proteoglycan and collagen and an increase in tissue size. Enzymatic treatment to remove GAG (C) and subsequent growth lead to the maturational growth phenotype (D), with deposition of proteoglycan and collagen but with maintenance of tissue size.

In contrast, in the case of the second phenotype, treatment with chondroitinase ABC and subsequent incubation with 20% FBS induced maturational cartilage growth (Figure 7D). Chondroitinase ABC treatment on day 0 induced a depletion of matrix GAG, leaving the collagen network intact, and resulted in a marked increase in tensile mechanical integrity (Figures 5 and 6). During subsequent culture, the volumetric cartilage growth was slight (Figures 1A and B), the content of GAG was restored (Figure 2B), and the collagen content increased slightly (Figure 2C). Consequently, the concentrations of GAG, collagen, and Pyr remained relatively unchanged (Figures 2B–D). Notably, these changes were accompanied by maintenance of the tensile mechanical integrity at enhanced levels attained after treatment with chondroitinase ABC (Figures 5 and 6). These findings suggest that chondroitinase ABC treatment and subsequent free-swelling, serum-supplemented culture induce conditions that restrict volumetric tissue growth while enhancing the integrity of the collagen network. Thus, during in vitro growth, the expansive growth phenotype shows many hallmarks of a distinct immature and growing tissue, while the maturational growth phenotype exhibits signs of progression to a more mature tissue state.

The use of immature cartilage explants for studying mechanisms of articular cartilage growth required consideration of a number of issues. The tissue was harvested in layers using the top ∼1 mm of articular cartilage. The superficial layer included the articular surface and was prepared to include cells that are situated to become those present in mature articular cartilage. Because the various zones of normal cartilage exhibit differences in biochemical composition and mechanical properties, the middle layer was also analyzed, and it displayed an initial state different from that of the superficial layer. The patellofemoral groove was used as the source of tissue, similar to tissue used in previous studies (15, 19, 21). Consequently, the biochemical and biomechanical properties of tissue samples at the time of explant were similar to those reported previously (15, 21).

The individual direct contributions of GAG and the collagen network to the tensile properties of cartilage may explain some, but not all, of the increase in the overall tensile integrity after chondroitinase ABC treatment demonstrated in the present study and observed by others (28, 29). In normal cartilage, proteoglycans act to inflate the collagen network and establish a level of prestress (8), while in GAG-depleted cartilage, in the absence of prestress, the collagen network would be expected to relax and straighten. Although chondroitinase ABC treatment had little effect on thickness and wet weight (Figures 1A and B), the small decrease in wet weight due to treatment may indicate a contraction of the collagen network. The absence of prestress in the collagen network of treated cartilage may contribute to higher strength, while collagen network straightening can account for some of the decrease in failure strain and, consequently, an increase in the ramp stiffness/modulus.

The residual variation in tensile properties between fresh explants and chondroitinase ABC–treated explants may involve different interactions between collagen fibers within the collagen network. Although the nature of such interactions in normal cartilage is not well established, collagen-binding molecules (such as type IX collagen and decorin) can regulate collagen fibril length and diameter during development (41, 42) and may also mediate the adherence and sliding between collagen fibrils and act as functional noncovalent linkages within the extracellular matrix (43, 44). While aggrecan contains most of the chondroitinase ABC–susceptible material of the tissue, depletion of GAG with chondroitinase ABC may alter or displace some of these other chondroitin sulfate– or dermatan sulfate–containing molecules and allow new interactions between collagen fibrils that enhance the tensile integrity of the collagen network. Thus, although the tensile behavior of cartilage in the high-strain region is generally attributed to the tensile response of the collagen network (15, 26), these results demonstrate an indirect, but dramatic, contribution of the GAG component.

It appears that free-swelling growth of immature cartilage explants in vitro in the presence of serum involves a shift in the balance between the swelling pressure of the proteoglycan molecules and the restraining ability of the collagen network, in favor of the swelling pressure. Factors that contribute to this imbalance, and the consequent expansive cartilage growth, involve both the additional swelling pressure associated with the newly synthesized GAG (Figure 2B) and the remodeling of the collagen network that renders it soft and weak (Figures 5B and C and 6B and C). Pilot studies (data not shown) indicate that such growth is mediated by the metabolic activities of chondrocytes and is not a passive swelling process, since inhibition of biosynthesis by addition of cycloheximide to block protein synthesis or incubation at 4°C resulted in maintenance of cartilage explant geometry within 6% of the initial wet weight up to 2 weeks in serum-supplemented culture, much less than the observed increase in tissue volume during expansive growth (Figure 1B). Thus, the swelling pressure of intrinsic and newly synthesized GAG may be sufficient not only to expand the weakened endogenous collagen network, but also to prevent its functional reinforcement with the newly deposited collagen. Studies to investigate the spatial distribution of newly deposited GAG molecules and other matrix components may provide additional insights into the mechanisms of cartilage growth.

While remodeling of the collagen network during growth is an important determinant of volumetric growth of cartilage tissue, the structural organization of the collagen network and its resultant restraining function at the initial state also play a role. Although the superficial and middle layers were not compared directly, differences in statistical variations of certain parameters in the 2 layers were evident. During growth of the superficial layer, the collagen network remained functional, as indicated by the maintenance of structural mechanical integrity (Figures 5A and C), despite large extents of volumetric growth (Figures 1A and B). In contrast, during growth of the middle layer, even lower extents of volumetric growth induced a weakened collagen network (Figures 5B and C). These trends suggest that cartilage from the superficial layer is more resistant to free-swelling in vitro growth-associated collagen network weakening than cartilage from the middle layer, and that it may be better suited to accommodate expansion associated with growth in vivo. Depth-associated variations in matrix components and their structural organization, such as the orientation of the collagen fibrils, may be responsible for these differences.

Remarkably, chondroitinase ABC–treated and subsequently cultured explants were able to restore GAG content to native levels (Figures 2B and 3), a finding similar to that in previous studies that demonstrated the ability of cartilage to restore proteoglycan content and organization within 4–10 days after enzymatic digestion by increasing the rate of proteoglycan synthesis (45–48). Although chondroitinase ABC treatment stimulated sulfate incorporation in the superficial layer, the level of sulfate incorporation in the middle layer was similar to that of the untreated explants (Figure 4A), despite marked GAG deposition (Figure 2B). This could be largely explained by a higher rate of GAG turnover in the untreated explants, as indicated by the elevated GAG release to the culture medium, combined with preferential turnover of GAG in the matrix further removed from the cells, as indicated by a lower specific activity of GAG in the medium. Thus, while high rates of synthesis are needed to restore GAG content in chondroitinase ABC–treated explants, in untreated explants similar rates of GAG synthesis are needed to replenish nonradiolabeled GAG that is being turned over.

Remodeling and reorganization of cartilage tissue that occurs during chondroitinase ABC treatment and subsequent incubation with serum allows enhancement of tissue function and results in maturational cartilage growth. While the overall content of the newly synthesized and deposited GAG on day 13 was high, the majority of newly synthesized collagen was probably deposited during the time when tissue GAG content and the associated swelling pressure were low. This temporal mismatch in the contents of GAG and collagen may have permitted the development of a stiff collagen network that was able to resist the swelling pressure and tendency toward expansion when GAG content increased. Factors other than the dynamic imbalance between the swelling pressure of GAG and the restraining function of the collagen network may also contribute to the observed cartilage growth phenotypes. Alterations in matrix content induced by chondroitinase ABC may change the way certain chemical (medium components) or mechanical (e.g., osmotic) signals are perceived by the cells. This in turn may perturb the cell-mediated matrix metabolism in a way that can lead to the observed changes in the composition and the mechanical properties of the tissue.

While the relationship between manipulations in vitro and growth and maturation control mechanisms in vivo remains to be defined, the findings of the current study may have practical utility for tissue engineering and cartilage repair. The expansive growth phenotype may be beneficial for volumetric tissue growth and used to create cartilage tissue constructs in vitro or to fill cartilage defects directly in vivo. The maturational growth phenotype may be beneficial for promoting tissue integrity and used to induce matrix stabilization after generating a cartilage tissue construct of a desired volume or after filling defect spaces.

AUTHOR CONTRIBUTIONS

Dr. Sah had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Ms Asanbaeva and Dr. Sah.

Acquisition of data. Ms Asanbaeva.

Analysis and interpretation of data. Ms Asanbaeva and Drs. Masuda, Thonar, Klisch, and Sah.

Manuscript preparation. Ms Asanbaeva and Drs. Masuda, Thonar, Klisch, and Sah.

Statistical analysis. Ms Asanbaeva and Dr. Sah.

Acknowledgements

We give special thanks to Mary Ellen Lenz for careful reading of the manuscript.

Ancillary