Articular cartilage was obtained from the knees of 2-year-old steer provided by a local slaughterhouse. Vardenafil was from Bayer AG (Leverkusen, Germany), tadalafil was from Elli Lilly (Indianapolis, IA, USA), hyaluronan binding protein (HABP) was from Calbiochem (Schwalbach, Germany), and hyaluronan (Healon^®) was a gift from Genzyme (Cambridge, MA, USA). Polyclonal antibodies to matrix metalloprotease (MMP)9 were from Biomol (Hamburg, Germany). Additional chemicals were from Sigma-Aldrich Chemical Corporation (Taufkirchen, Germany).

The hyaluronan synthase activity was determined by incorporation of radioactive [¹⁴C]glucuronic acid from UDP- [¹⁴C]GlcA and UDP-GlcNac 7. The cytotoxicity of the drugs was measured as described previously 34. For all experiments, the weight of the explants was determined immediately after cutting to minimize evaporation and the data were related to wet weight.

Cartilage explants were incubated in the absence or presence of IL-1 (2 ng/ml) and the inhibitors at various concentrations in serum-free Dulbecco's medium for 3 days. The amount of hyaluronan released into the culture medium was determined using biotinylated HABP, as described previously 30.

Cartilage explants were weighed (average wet weight 20 mg) and incubated in the absence and presence of IL-1 (2 ng/ml) and the inhibitors at various concentrations for 5 days. The tissues were extracted with 1.5 ml of a solution of 4 mol/l guanidinium hydrochloride, 0.1 mol/l ε-aminohexanoid acid, 5 mmol/l benzamidine, 10 mmol/l N-ethylmaleinimide and 0.5 mmol/l phenalmethylsulfonyl fluoride for 3 days at 4°C. The solution was centrifuged for 5 minutes at 10.000 g and the proteoglycans were determined in the supernatant using the alcian blue method, as described previously 35.

Chondrocytes were cultured in alginate beads, as described above, and supplemented with 25 μl [³⁵S]sulphate (0.5 mCi/ml) for 24 hours. The beads were washed three times with 102 mmol/l CaCl₂ to remove un-incorporated radioactivity and dissolved in 55 mmol/l sodium citrate. Proteoglycans were isolated by the alcian blue precipitation method 36 and aliquots were used in the determination of radioactivity.

The procedure for measurement of degraded collagen is described in detail in the report by Kosaki and coworkers 37. Cartilage explants were cultured in Dulbecco's medium for 28 days in the presence or absence of IL-1 (2 ng/ml), IL-17 (25 ng/ml), 2 μmol retinoic acid and hyaluronan export inhibitors, and media were changed every 2 days. The cartilage was weighed and extracted with 4 mol/l guanidinium hydrochloride in 0.1 mol/l Tris HCl (pH 7.3), 1 mmol/l Iodoacetamide, 1 mmol/l EDTA, and 10 μg/ml pepstatin A for 72 hours. The extracted explants were washed with 1 mmol/l iodoacetamide and 1 mmol/l EDTA, and 10 μg/ml pepstatin in 0.1 mol/l Tris-HCl (pH 7.3) three times for 2 hours. The denatured collagen was digested overnight at 37°C with a solution of 0.5 ml of α-chymotrypsin (0.5 mg/ml) in 1 mmol/l iodoacetamide and 1 mmol/l EDTA, and 10 μg/ml pepstatin in 0.1 mol/l Tris-HCl (pH 7.3). The mixture was centrifuged for 8 minutes at 10,000 g, and the supernatant containing the digested collagen was separated from the remaining insoluble matrix containing the intact collagen. The insoluble material was hydrolyzed with 500 μl of 6 mol/l HCl at 110°C for 20 hours. The hydrolysate was neutralized with 500 μl of 6 mol/l NaOH and undissolved material was removed by centrifugation. The amount of the collagen-specific amino acid hydroxyproline was determined. An aliquot (25 μl) was mixed with 975 μl citrate buffer (57 g sodium acetate, 37.5 g sodium citrate, 5.5 g citric acid and 385 ml 2-propanol in 1 l water). An aliquot (200 μl) of this mixture was added to 100 μl of a solution of 100 mg chloramine T in 1 ml water, 2 ml 2-propanol and 3 ml citrate buffer. After 20 minutes at room temperature, 100 μl of 6.2 mol/l perchloric acid was added and reacted for 12 minutes at room temperature. A solution of 100 μl Ehrlichs reagent (500 mg in 1.25 ml ethanol and 1.25 ml diethyleneglycol-monoethylether) was added and incubated at 60°C for 20 minutes. The adsorption was read at 565 nm and the content of hydroxyproline was calculated using 1 to 30 μg/ml calibration samples.

Chondrocytes were cultured in alginate beads for 1 week with 10% foetal calf serum in Dulbecco's medium. The medium was changed and supplemented with 1 mmol/l cysteine, 1 mmol/l pyruvate, 60 μg/ml β-aminopropionitril and 25 μg/ml ascorbic acid, and the beads were incubated for an additional 24 hours. The medium was replaced with serum-free medium containing the above supplements, IL-1, the inhibitors and [¹⁴C]proline (2 μCi/ml), and the cells were incubated for 24 hours. The beads were washed three times with 0.9% NaCl and CaCl₂ (116 mg/l) for 30 minutes to remove unincorporated radioactivity, dissolved in 500 μl of 55 mmol/l sodium citrate, and the radioactivity was determined.

Bovine cartilage explants were cultured in serum-free Dulbecco's medium for 5 days in the presence or absence of IL-1α (2 ng/ml) and 10 or 30 μmol/l zaprinast, vardenafil, or tadalafil. The protein concentrations of the culture media were determined and equal amounts of proteins were directly applied to a 7.5% SDS-polyacrylamide gel that contained 0.1% gelatin. After electrophoresis, the gel was washed twice with 2.5% Triton X-100 for 30 minutes, three times with water for 10 minutes, and incubated in a solution of 50 mmol/l Tris-HCl, 5 mmol/l CaCl, 1 μmol/l ZnSO₄ (pH 8.0) for 5 days at 37°C. The gel was stained with Coomassie blue.

Bovine cartilage explants were cultured in Dulbecco's medium and 10% foetal calf serum in the presence or absence of IL-1α (2 ng/ml) and the drugs zaprinast, tadalafil and vardenafil for 14 days. Pieces of 2 mm diameter and a volume of 3.15 mm³ were punched out. They were incubated in a solution of 10 units/ml of horseradish peroxidase (HRPO) in phosphate-buffered saline for 1 hour at 37°C. The pieces were rinsed with water and shaken in 250 μl phosphate-buffered saline overnight at 4°C to release the infiltrated enzyme. The concentrations of HRPO were determined in 50 μl of the supernatants. A solution (150 μl) of ABTS (2,2'-azino-bis [3-ethylbenzthiazoline-6-sulfonic acid]; 1 mg/ml) and 0.03% H₂O₂ was added, and after incubation for 30 minutes at 37°C the adsorbance at 405 nm was read.

Data are presented in the figures as mean ± standard deviation. The t-test was used, and a P value below 0.05 was considered statistically significant.

The drugs tadalafil, zaprinast and vardenafil were analyzed for their effects on hyaluronan export from bovine cartilage explants in tissue culture. Cartilage explants were incubated for 3 days in the presence and absence of IL-1α and increasing concentrations of the drugs. Figure 1a shows that IL-1α stimulated an increase in hyaluronan export by about sixfold, and the inhibitors partially reversed it. In control experiments, the inhibitors were analyzed for their effect on the hyaluronan synthase activity of chondrocytes cultured in alginate beads. Activity was reduced by less than 20% up to concentrations of 400 μmol/l (Figure 1b). The toxicity of the drugs was less than 10% at a concentration of 100 μmol/l for the three inhibitors.

The drugs tadalafil, zaprinast and vardenafil were analyzed for their effects on proteoglycan loss from IL-1α activated bovine cartilage explants. Proteoglycans were extracted from the tissues with guanidinium hydrochloride and determined colourimetrically. Figure 2a shows that IL-1α reduced the proteoglycan content in cartilage to less than 40% of that in the untreated control. The inhibitors protected the cartilage from proteoglycan loss. In a control experiment, the effect of the inhibitors on the proteoglycan synthesis rate was determined. Bovine chondrocytes were cultured in alginate beads and incubated with [³⁵S]sulphate in the presence of drugs, and inhibition of proteoglycan synthesis was found to be reduced by less than 25% (Figure 2b). These findings confirm earlier observations obtained with other drugs 30 and suggest that zaprinast, vardenafil and tadalafil prevented proteoglycan loss from osteoarthritic cartilage primarily by inhibition of hyaluronan over-production.

The drugs were analyzed for their effects on collagen degradation in IL-1α activated cartilage explants. Preliminary experiments revealed that induction of osteoarthritic reactions by IL-1α was not sufficient to detect measurable amounts of collagen degradation products. Degradation can be enhanced substantially by addition of IL-17 and retinoic acid. Therefore, these activators were added. Activated cartilage explants were incubated with tadalafil, zaprinast, or vardenafil for 28 days, extracted with guanidinium hydrochloride, and digested with chymotrypsin. Degraded collagen was measured as the amount of hydroxyproline that was susceptible to chymotrypsin. Figure 3a shows that cartilage activation reduced the amount of chymotrypsin-resistant collagen to 65%. Inhibition of hyaluronan export restored the content of intact collagen. In a control experiment, the effect of zaprinast on collagen synthesis was measured (Figure 3b). Bovine chondrocytes were cultured in alginate beads and incubated in culture medium containing [¹⁴C]proline in the absence and presence of zaprinast and incorporation of radioactivity into pepsin-resistant collagen was determined. The total amount of collagen was not altered significantly at concentrations up to 100 μmol/l zaprinast. These findings suggest that the drugs did not affect collagen synthesis and that the protection from collagen degradation could involve other mechanisms.

A possible explanation for the protective effect of hyaluronan export inhibition on collagen degradation could be that the altered composition and permeability of osteoarthritic cartilage allowed the diffusion of metalloproteases. It is known that chondrocytes produce gelatinases, particularly if they are activated by IL-1 38. We tested this possibility by measuring the release of gelatinases from IL-1α activated cartilage. Cartilage explants were incubated in the absence and presence of IL-1α and vardenafil, and enzymes released from the cartilage explants were analyzed by gel zymography. We also included dibutyryl-cGMP in the analysis, because cGMP has been shown to mediate IL-1 signalling in chondrocytes 39. Figure 4a shows three bands with molecular weights of 86 kDa, 66 kDa and 62 kDa. The upper band was probably pro-MMP9, because it reacted with monoclonal antibodies in Western blots (data not shown). The lower two bands comigrated with an authentic sample of pro-MMP2 and MMP2 gelatinases (from Dr R Dreier; data not shown). IL-1α enhanced the release of the gelatinases, and this release was not significantly altered by addition of dibutyryl-cGMP. Vardenafil reduced the gelatinase release in a concentration-dependent manner. Similar results were obtained with zaprinast and tadalafil (data not shown).

In a control experiment, we analyzed whether inhibition of hyaluronan export altered the syntheses of gelatinases by chondrocytes in unstimulated cartilage explants. The explants were incubated in the absence and presence of 100 μmol/l zaprinast, vardenafil and tadalafil, respectively, and the gelatinase activities were again analyzed by gel zymography. Figure 4b shows no differences in enzyme activities between the samples. These results showed that inhibition of hyaluronan export in IL-1 activated cartilage explants inhibited the release of gelatinases into the culture medium.

If the altered composition of arthritic cartilage with increased hyaluronan and decreased proteoglycan content was responsible for facilitated diffusion of degrading enzyme through the matrix, exogenous enzymes should infiltrate better. This hypothesis was tested using HRPO as an indicator protein. Cartilage explants were incubated in the absence or presence of IL-1α and zaprinast, vardenafil, or tadalafil in increasing concentrations. The explants were incubated with the indicator protein HRPO to allow diffusion into the cartilage. After extensive washing, the explants were further incubated in phosphate-buffered saline to liberate the infiltrated enzyme. The amount of liberated enzyme was determined by a colour reaction. Control experiments indicated that the drugs did not have any direct effect on the peroxidase activity at micromolar concentrations. Figure 5 shows that IL-1α treatment led to an increase of enzyme infiltration of about 350% over the unstimulated control (100%). Inhibition of hyaluronan export reduced the IL-1α induced enzyme infiltration almost to control values.

There are two possible mechanisms for the inhibitory action of the drugs zaprinast, vardenafil and tadalafil. Because of their PDE5 inhibitory activity, with Ki values of 300 nmol/l for zaprinast, 1.5 nmol/l for vardenafil and 2.9 nmol/l for tadalafil 4041, they will certainly raise the concentration of intracellular cGMP that could inhibit hyaluronan export by MRP5 8. It is also possible that they additionally act as MRP5 inhibitors, because they are structural analogues of cGMP; also, it is known that zaprinast inhibits transport at concentrations between 20 nmol/l and 250 μmol/l, depending on the transported substrate 334243. If the drugs acted only through inhibition of PDE5, then also other unrelated PDE5 inhibitors or addition of other cGMP analogs should have similar effects. We therefore analyzed hyaluronan export, proteoglycan loss and collagen degradation of IL-1 activated cartilage explants in the presence of dibutyryl-cGMP and ODQ (1H-124-oxadiazolo [4,3a]quinoxaline-1-one), which is a selective inhibitor of the soluble nitric oxide inducible guanylate cyclase. Figure 6 shows that none of the parameters was significantly altered by these treatments. Similar results were obtained with bromo-cGMP (data not shown). These findings suggest that alterations of the intracellular cGMP concentration did not account for the inhibitory effects of the drugs zaprinast, vardenafil and tadalafil.