DMEM, gentamicin, non-essential amino acids (NEAA), N-(2-hydroxyethyl)-piperazine-N'-2-ethanesulfonic acid (HEPES), trypsin-ethylenediaminetetraacetic acid, proteinase K, PBS and Alexafluor 488 anti-rabbit and Alexafluor 594 anti-mouse secondary antibodies were from Invitrogen (Carlsbad, CA, USA). Antibodies to aggrecan G1 and the NITEGE neoepitope were prepared as previously described 35. The antibody to type I collagen was from Abcam (Cambridge, MA, USA), and the antibody to bovine decorin (LF-94) was a gift from Dr. Larry Fischer (National Institute of Dental and Craniofacial Research, Bethesda, MD, USA). Fluorescein isothiocyanate (FITC)-conjugated secondary antibody was from Chemicon (Temecula, CA, USA). Guanidine hydrochloride, chondroitinase ABC, keratanase I, chloramine-T, para-(dimethylamino)-benzaldehyde (pDAB), non-immune rabbit IgG, alkaline phosphatase-conjugated secondary antibody, 4',6-diamidino-2-phenylindole dihydrochloride (DAPI), and other histologic reagents were from Sigma (St. Louis, MO, USA). Verhoeff's stain was from Poly Scientific (Bay Shore, NY, USA). Keratinase II was from Associates of Cape Cod (Falmouth, MA, USA). Dimethylmethylene blue (DMMB) was from Polysciences (Warrington, PA, USA). Recombinant human IL-1α (rhIL-1α) was from Peprotech (Rocky Hill, NJ, USA).

Protease inhibitor cocktail I was from Calbiochem (San Diego, CA, USA). The enhanced chemifluorescence (ECF) substrate was from Amersham (Piscataway, NJ, USA). Metalloproteinase inhibitors (RO3310769, aggrecanase-1,2-selective; RO1136222, MMP-selective; RO4002855, broad-spectrum metalloproteinase) with previously described selectivity profiles 27 were provided by Roche - Palo Alto (Palo Alto, CA, USA).

Immature (one to two weeks) bovine menisci (Research 87, Cambridge, MA, USA) were harvested aseptically within 24 hours of slaughter. Coronal 'slabs' (3 to 4 mm thick, Figure 1a) were either fixed in 10% neutral-buffered formalin (NBF) at 4°C for 48 hours and embedded in paraffin or fixed in 10% NBF for four hours at 4°C, incubated in 30% sucrose/PBS overnight at 4°C, embedded in optimal cutting temperature (OCT) embedding medium, and frozen in liquid-nitrogen cooled isopentane. Paraffin-embedded tissue was sectioned at 3 μm, stained for sGAG with safranin-O, and counterstained with fast green and hemotoxylin. Blood vessels were identified in sections stained with Verhoeff's elastin stain, differentiated in 2% ferric chloride, rinsed in 5% sodium thiosulfate, and counterstained with van Gieson's collagen stain. Frozen tissue was sectioned at 7 μm and co-immunostained for the aggrecan NITEGE neoepitope and collagen type I. Sections were digested with 0.5% trypsin for 10 minutes at 37°C and blocked for one hour at room temperature with blocking buffer containing 2% goat serum, 0.1% gelatin, 1% bovine serum albumin, 0.1% Triton, and 0.05% Tween-20 in PBS prior to incubation with antibodies. Primary antibodies were detected by indirect immunofluoresence with Alexafluor 488- and Alexafluor 594-conjugated secondary antibodies. Following tissue culture, 3 to 4 mm thick coronal slabs were fixed in 10% NBF for 48 hours at 4°C, embedded in paraffin, and sectioned at 3 μm. The NITEGE neoepitope was detected by indirect immunofluoresence with a goat anti-rabbit FITC-conjugated secondary antibody. All antibodies were used at 10 μg/mL with incubations for 60 to 90 minutes at room temperature. Cell nuclei were stained with DAPI.

Species-matched non-immune IgGs were used as negative controls. Images of histochemically stained sections were captured on a Nikon E600 microscope (Nikon USA, Melville, NY, USA). Immunofluorescent sections were imaged on a Zeiss AxioVert 200 M microscope using the Apotome optical sectioning module (Zeiss, Jena, Germany).

In a preliminary study, coronal sections (Figure 1a) were dissected from immature bovine menisci and cultured in six-well plates with or without 20 ng/mL rhIL-1α in serum-free media consisting of DMEM, 10 mM NEAA, 50 μg/mL gentamicin, and 50 μg/mL ascorbic acid under standard conditions (5% CO₂, 95% humidity and 37°C). Media were exchanged after 48 hours. After four days, the tissue was fixed and processed for immunofluorescent detection of the NITEGE neoepitope.

Full-thickness cylindrical specimens (3 mm in diameter) were harvested from the middle region of menisci and cut to 2 mm in thickness with superficial tissue excluded (Figure 1b), generating samples with well-controlled geometries and low variability in initial wet mass (16.1 ± 1.7 mg). Explants were pre-cultured for 72 hours in 0.5 mL of 'basal' serum-free media consisting of DMEM, 10 mM NEAA, 50 μg/mL gentamicin, and 50 μg/mL ascorbic acid, followed by up to 12 days with or without 20 ng/mL rhIL-1α. Some explants were additionally treated with 20 μM of the aggrecanase-selective inhibitor, 5 μM of the MMP-selective inhibitor, or 5 μM of the broad-spectrum metalloproteinase inhibitor (doses based on previous studies 27). Media were exchanged every 48 hours. Explants (n = 6 per group) were collected at days 0, 4, 8, and 12 days of culture and stored frozen in PBS with protease inhibitors. Day 0 and 8 explants were lyophilized and extracted for 48 hours at 4°C in 10 volumes of extraction buffer (4 M guanidine hydrochloride, 10 mM 2-(N-morpholino)ethanesulfonic acid, 50 mM sodium acetate, 5 mM ethylenediaminetetraacetic acid, 5 mM iodoacetic acid at pH 6.5) with protease inhibitors.

Day 0 and day 12 explants were thawed, weighed, and measured using digital calipers. All tests were performed in PBS with protease inhibitors after compression by 10% of the explant thickness and 12 minutes of stress relaxation. Explants were tested in torsional shear on a CVO120 rheometer (Bohlin, East Brunskwick, NJ, USA) with a 0.5% nominal shear strain at 0.001 to 0.1 Hz to determine dynamic shear moduli G*. Following equilibration to the free-swelling state, the samples were recompressed and tested in oscillatory unconfined compression on an ELF3200 (Enduratec, Minnetonka, MN, USA) with a 1.5% strain amplitude at 0.001 to 0.1 Hz to determine the dynamic compressive moduli E*.

Explants from day 0, 4, and 12 were lyophilized, weighed, and digested in 0.0125 mg/mg tissue proteinase K. Explant digests, explant extracts, and conditioned media were assayed for sGAG content by the DMMB dye-binding method 36. Digests and conditioned media were assayed for hydroxyproline content by the chloramine-T/pDAB reaction 37, with collagen content calculated by assuming a collagen:hydroxyproline mass ratio of 8:1 38.

The inner (about 5 mm), middle (about 5 mm), and outer (about 8 mm; along lines drawn in Figure 1a) of immature bovine menisci were dissected, minced, and extracted in extraction buffer (80 mg wet tissue per ml) for 48 hours at 4°C. Conditioned media from the inhibitor study were pooled from days 2 and 4, 6 and 8, or 10 and 12 (from six explants per condition), and equal portions of day 0 and day 8 explant extracts were pooled (six explants per condition and time point). Proteoglycans were isolated from the media and extracts by ethanol precipitation, dried, and deglycosylated by overnight digestion in chondroitinase ABC and keratanases I and II. Equal volumes of conditioned media or native tissue extracts, or equal quantities of sGAG (5 μg for explant extracts), were separated by electrophoresis and transferred to nitrocellulose. Blotting equal volumes of media or tissue extract afforded direct comparison of the abundance per wet weight of the proteins of interest. Blotting equal amounts of sGAG (as with explant extracts) allowed comparison of the abundance of sGAG-bearing proteins per μg total sGAG. Following blocking, membranes were probed with primary antibodies at 1 μg/mL. Blots were developed using an alkaline phosphatase secondary antibody and exposure to ECF. Bands were visualized using a Fuji FLA3000 (Fuji, Tokyo, Japan).

The main effects of inhibitor treatments (at each time point) were evaluated using a general linear model in MINITAB Release 14 (MINITAB, States College, PA, USA). Pairwise comparisons between groups were determined by Dunnett's test with IL-1-only treated samples as controls and significance at P < 0.05.

Safranin-O staining revealed marked regional variations in sGAG content within the immature meniscus with the most intense staining within the inner region (Figure 2a). There were also sGAG-rich pockets in the middle and outer regions colocalized with cells and between collagen fibril bundles. These patterns are consistent with previously reported regional variations in canine meniscus aggrecan content 7, porcine aggrecan gene expression 12, and ovine proteoglycan synthesis rates 39. Verhoeff's stain demonstrated the presence of blood vessels in the middle and outer regions of the immature meniscus. Vessels were identified by elastin staining (black) and the presence of open lumens. The inner region was consistently devoid of blood vessels.

The abundance of the aggrecan NITEGE neoepitope, a product of aggrecanase activity, varied across coronal sections of the immature meniscus (Figure 2a, lower panels), with weak staining in the inner region but intense intra- and peri-cellular staining in the middle and outer regions. The middle and outer regions also exhibited NITEGE staining along and within collagen fibril bundles. NITEGE appeared in regions where safranin-O staining was weak and blood vessels were present. Immunoblots of tissue extracts (Figure 2b) supported the staining patterns, demonstrating regional variations in the abundance of NITEGE in both medial and lateral menisci.

When cultured coronal sections of meniscus were examined by immunohistochemistry, the unstimulated controls showed regional variations in NITEGE abundance similar to freshly isolated tissue and an increase in staining in the surface zone of the inner region (Figure 3). In contrast, IL-1-stimulated tissue exhibited marked increases in NITEGE staining throughout the inner and middle regions. The middle region of IL-1-stimulated meniscus demonstrated enhanced NITEGE staining around fibril bundles, seen as a brightly stained fine network that was less intense in the inner region and absent from the outer region. These data indicate that IL-1 stimulation of immature meniscal fibrocartilage triggers the exposure of the NITEGE neoepitope, presumably through aggrecanase-mediated degradation of aggrecan.

IL-1 stimulation triggered significant, time-dependent reductions in the sGAG content of meniscal explants (Figure 4a). Inhibition of aggrecanase-1 and -2 did not significantly enhance sGAG retention relative to IL-1-only controls at either day 4 or 12. Conversely, explants treated with an MMP-selective inhibitor exhibited sGAG depletion similar to IL-1-only controls over days 0 to 4 but had significantly higher sGAG contents than IL-1-only controls at day 12. Explants treated with a broad-spectrum inhibitor exhibited a blunted acute IL-1-induced release of sGAG and maintained significantly higher sGAG contents than IL-1-only controls at days 4 and 12. These data suggest that aggrecanases-1 and -2 participate in the early release of sGAG in response to IL-1, but MMPs and/or other aggrecanases contribute substantially to the release of sGAG later in the degradative process.

The size and structure of aggrecan in meniscal explants from day 0 and 8 were examined by western blot for the aggrecan G1 domain (Figure 4b). The sGAG contents of day 8 explant extracts (Figure 4b, lower panel) are provided for reference and are consistent with the temporal trajectories of explant digest sGAG content depicted in Figure 4a. A doublet migrating at 66 to 70 kDa was an abundant product in explant extracts and confirmed as the G1-NITEGE fragment in blots for the NITEGE neoepitope (not shown). The blot reveals a paucity of high molecular weight aggrecan at day 0 (and later culture times), although this may be attributable to relatively poor extraction of sGAG-bearing proteoglycans from the explants. We previously found that (with the guanidine extraction buffer used in the present study) proteoglycans are more difficult to extract from meniscal fibrocartilage than from age- and species-matched articular cartilage 40. Over the course of eight days in culture, the abundance of 66 to 70 kDa aggrecan increased slightly in untreated explants and was markedly reduced in IL-1-treated tissue. The aggrecanase-1,2-selective inhibitor had little effect on the IL-1-induced depletion of this species, whereas both the MMP-selective and broad-spectrum inhibitors abrogated this effect of IL-1 on the explants, suggesting that MMPs mediate the release of aggrecan-G1 from IL-1-stimulated meniscal fibrocartilage.

We also examined the abundance of the small leucine-rich proteoglycan decorin in the explant extracts (Figure 4b, middle panel). Decorin was readily detected in untreated controls but was almost absent in IL-1-stimulated explants. Interestingly, treatment with any of the metalloproteinase inhibitors abolished the IL-1-induced depletion of decorin. These results underscore the susceptibility of other proteoglycans to IL-1-induced ECM catabolism, and suggest that some sGAG released from the explants is associated with non-aggrecan proteoglycans.

Assays of conditioned media confirmed that IL-1 stimulated the release of sGAG and aggrecan from meniscal fibrocartilage (Figure 5). Release of sGAG peaked early in the experiment (days 0 to 4) and was substantially lower at later timepoints (Figure 5a). The amount of sGAG released from unstimulated controls was about 50% of that released from IL-1-treated explants. None of the inhibitor treatments significantly reduced IL-1-induced sGAG release over the first 96 hours, but treatment with the MMP-selective or broad-spectrum inhibitors significantly reduced sGAG in the media at later time points.

Using antibodies to the G1 domain and the NITEGE neoepitope, various aggrecan fragments were detected in the conditioned media (Figure 5b). Untreated controls released high molecular weight species at about 400 kDa (full-length aggrecan), about 280 kDa, and about 250 kDa as well as a low molecular weight species migrating at about 70 kDa and bearing the NITEGE neoepitope (Figure 5b, lower panels). Media from IL-1-stimulated samples contained the 250 kDa G1 species and the 66 to 70 kDa NITEGE doublet found in media from unstimulated controls, as well as additional G1 species migrating at about 100 kDa and about 45 kDa. Previous work with the anti-G1 antibody has shown that the 45 kDa band is due to reactivity with link protein 15. The bulk of aggrecan G1 was released from IL-1-stimulated fibrocartilage within the first 96 hours, when a large fraction of sGAG was released to the media. Treatment with the aggrecanase-1,2-selective inhibitor diminished the IL-1-induced release of the 250 kDa G1 species and NITEGE-bearing fragments to media over the first 96 hours, and abolished release of the 100 kDa and 45 kDa G1 species. The MMP-selective inhibitor also appeared to diminish aggrecan release to the media, but this inhibitor did not block release of the 100 kDa or 45 kDa species and release of NITEGE was similar to IL-1-only controls. Significantly, the broad-spectrum inhibitor restored the temporal pattern of aggrecan (G1 and NITEGE) release in IL-1-stimulated explants to that of unstimulated controls.

Assays of explant digests demonstrated that IL-1 triggered a loss of collagen (Figure 6a), and treatment with either the MMP or broad-spectrum metalloproteinase inhibitor abolished IL-1-induced collagen degradation. Unstimulated controls exhibited low rates of collagen release, and IL-1-stimulation elevated release from day 6 to 12 (Figure 6b). Addition of the aggrecanase-1,2-selective inhibitor did not affect collagen release, consistent with its specific effects on ADAMTS-4 and -5, which have no proteolytic activity against collagen. Interestingly, this pattern is in contrast to articular cartilage, where aggrecanase inhibition delayed the onset of IL-1-induced collagen destruction 2627 presumably by blocking access of the MMPs to the collagen rather than through direct MMP inhibition. Treatment with the MMP-selective or broad spectrum metalloproteinase inhibitors potently inhibited IL-1-induced collagen depletion in meniscal explants, confirming that MMPs were primary mediators of collagen destruction in this culture system.

In both unconfined compression and torsional shear tests, day 0 tissue exhibited trends of increasing modulus with frequency, and these trends were maintained in unstimulated controls following 12 days of culture (Figure 7). Significantly, the compression and shear properties of these explants were maintained over 12 days of culture, despite evidence of persistent aggrecanolysis (Figures 4 and 5), suggesting that generation of NITEGE in the immature bovine meniscus is compatible with maintenance of this tissue's mechanical properties (perhaps due to contributions by proteoglycans other than aggrecan). Stimulation with IL-1 led to reductions in both compression and shear moduli, and the aggrecanase- and MMP-selective inhibitors did not significantly ameliorate these losses. IL-1-stimulated explants cultured with the broad-spectrum inhibitor, however, had significantly higher compression and shear moduli than IL-1-only controls.