Nociceptive tolerance is improved by bradykinin receptor B1 antagonism and joint morphology is protected by both endothelin type A and bradykinin receptor B1 antagonism in a surgical model of osteoarthritis

Over the course of two months post-operatively, animals were treated by weekly intra-articular injections of ETA and/or BKB1 specific peptide antagonists: BQ-123 (ETA antagonist; Sigma-Aldrich, Oakville, Ontario) [30, 31], R-954 (BKB1 antagonist; a kind gift from Pierre Sirois, IPS Thérapeutique, Sherbrooke, Quebec) [32, 33], both, or saline vehicle, was injected into the right knee at a dose of 30 nmol in a volume of 50 μL. Injections were performed under isoflurane anaesthesia, using a 28G needle; the procedure is described in detail in Additional file 1. Chemical structures of the antagonists are depicted in Additional file 2. Doses were based upon previously published reports [14, 19].

Over the course of the study, animal nociception was evaluated biweekly by the static weight bearing test. A static weight bearing apparatus was reverse-engineered from previously published reports [34‐36], designed, and machined by Usinage FB (Le Gardeur, Quebec). Design diagrams and photos are appended in Additional files 3 and 4.

After conditioning, animals were introduced to the apparatus and restrained in a plexiglass chamber with an angled base, such that each hind paw rested on a separate force plate connected to a load cell. The weight in grams distributed on each hind limb was recorded by a computer software interface (Futek USB software interface version 2.10). The static weight bearing distribution of each animal was recorded for 30 seconds; each data point was then taken as the mean of three 30-second readings. Data were transferred off-line to a personal computer, and the weight bearing on the right hind limb as a percentage of total weight bearing on both hind limbs was calculated by the following equation [37]:

All values are given as mean ± standard deviation (SD) per experimental group.

At four or eight weeks post-surgery, animals were sacrificed by cardiac puncture under deep isoflurane anaesthesia. The right knee was dissected, and 40-mm-long samples were cut and stored in phosphate-buffered saline until imaged by digital micro-X-ray (DX) and/or micro-magnetic resonance (MR). Samples were dissected the same day as the radiological scans.

All knee samples were X-rayed using a Faxitron MX-20 specimen X-ray system (Faxitron X-Ray Corporation, Lincolnshire, IL). Anteroposterior and lateral views were acquired at 5 × magnification (10 × 10 μm pixel size) using a dose of 26 kV for 6 seconds. Images were analyzed using OsiriX software (version 3.7.1) [41]. Radiological evidence of joint degradation was scored by two blinded examiners using an OA radiological score modified from Clark et al. [42] and Esser et al. [43]. Bone demineralization, subchondral bone erosion, and heterotopic ossification were all scored on a scale from zero (normal) to three (marked degenerative changes). Total scores were calculated by summing the individual scores for each index, with a maximum possible score of nine.

After radiological examination, knee samples were fixed in 10% neutral buffered formalin for two weeks, decalcified with RDO Rapid Decalcifier (Apex Engineering Products, Aurora, Illinois) for three days, circulated, and embedded in paraffin. Five-micron sagittal sections were acquired from the middle of the knee joint. Histomorphological staining was performed as previously described [45]: slides were deparaffinized, rehydrated, stained with Safranin O (which colors proteoglycans red), counterstained with Fast Green FCF (which colors proteins green) and with Weigert's hematoxylin (which colors nuclei black), dehydrated, cleared, and mounted in Permount. Representative digital photomicrographs were acquired with a Leica DM R microscope (Wetzlar, Germany) fitted with a QImaging Retiga 1300 B camera (Surrey, British Columbia), controlled by QCapture software (version 2.95.0). Images were captured at 50 × (low-power) or 200 × (high-power) magnification, and subsequently color-matched and balanced using Adobe Photoshop CS3.

Four slides from each condition were scored by two blinded examiners using the Osteoarthritis Research Society International (OARSI) histopathology assessment system [46], which assigns numeric values to grade, or depth progression into cartilage (0-6), and stage, or extent of joint involvement (0-4); multiplying grade and stage yields a total OA score with a maximum value of 24. Scores were averaged in between the two examiners; inter-examiner variation was within ± 5%.

Additional 5-micron sections were processed for immunohistochemical detection of type II collagen. Slides were deparaffinized, rehydrated, and washed in phosphate-buffered saline (PBS). Sections were incubated in 2 mg/mL hyaluronidase for 30 minutes at 37°C, followed by permeabilization with 0.3% Triton X-100 for 30 minutes at room temperature. Endogenous peroxidase activity was then quenched with 2% hydrogen peroxide in PBS for 15 minutes. Sections were blocked with normal mouse serum (Vector Laboratories, Burlingame, California) for 1 hour, after which they were blotted and then incubated with monoclonal mouse anti-rat type II collagen (clone SPM239; Spring Bioscience, Pleasanton, California) for 18 hours at 4°C. Sections were then washed in PBS, incubated with biotinylated anti-mouse IgG (Vector) for 1 hour at room temperature, and stained using the avidin-biotin complex method (Vectastain ABC kit; Vector). Color was developed using 3,3'-diaminobenzidine (Dako Diagnostics, Mississauga, Ontario) containing hydrogen peroxide. Slides were counterstained with Harris modified hematoxylin, dehydrated, cleared, mounted, and examined by light microscopy as described above.

To determine the effects of ETA and/or BKB1 local antagonist treatment on nociception in a surgical OA model, the static weight bearing asymmetry of the animals was measured repeatedly over the course of the study (Figure 1). Pre-operative baseline values for all groups indicated hind limb weight bearing symmetry (49.89 ± 0.42%). Unoperated vehicle-treated animals showed no important changes in hind limb weight bearing from baseline pre-operative values over the course of the study, staying roughly within ±4% of even weight distribution. Sham-operated vehicle-treated animals displayed an initial weight bearing imbalance 14 days post-operatively (36.47 ± 1.12%), but recovered weight bearing symmetry quickly thereafter (44.84 ± 0.33% by day 26 post-operatively). ACLT saline-treated animals showed significant weight bearing imbalance two weeks post-operatively, down to 33.66 ± 2.05% weight on the right leg, suggesting severe nociception. All animals had similar nociceptive tolerance at the last measured time-point (day 50 post-operatively), indicating nociceptive adaptation, but drug-treated animals were able to recover faster than saline-treated animals (up to 40.54 ± 3.36% weight on right leg by day 40 post-operatively, for BQ-123 and R-954 dual treatment).

Repeated measures analysis of variance of the static weight bearing data, followed by Tukey post-hoc hypothesis tests (Table 2), demonstrated that treatment with R-954, or both BQ-123 and R-954, significantly ameliorated nociceptive tolerance in ACLT animals over the study period, as compared to saline-treated positive controls (0.0001 ≤ P ≤ 0.0002). When administered alone, BQ-123 did not result in statistically significant increased nociceptive tolerance (P = 0.1847). Sham surgery was found to be slightly less nociceptive than ACL transection (P = 0.019), confirming that ACLT is necessary for a maximal nociceptive response. Furthermore, nociception in the sham-operated animals was comparable to unoperated animals, with no statistically significant difference calculated (P = 0.8746).

Right knee joints were dissected at the end of the study period and imaged by DX (Figure 2) and MR (Figure 3) to examine the radiological effects of antagonist treatments. ACLT rapidly induced radiological evidence of OA: knee joints showed signs of degradation such as subchondral bone remodeling, osteophyte formation (Figure 2c and Table 3), cartilage layer thinning (Figure 3c), and lengthened cartilage T₂ (Table 4). Neither sham surgery nor intra-articular injection affected joint radiomorphology (Figures 2a,b and 3a,b). DX analysis of antagonist-treated knee joints showed less subchondral bone remodeling and heterotopic ossification than saline-treated animals (Figure 2d,e,f and Table 3). Dual ETA/BKB1 antagonism appeared to be slightly more protective than single antagonism: less subchondral bone remodeling and greater trabecular integrity was observed in the dual-antagonist-treated animals than in the single-antagonist-treated animals. Radiological scoring of the DX views for a panel of OA joint degenerative changes (Table 3 and Additional file 5) demonstrated that treatment with BQ-123, R-954, or both, significantly ameliorated radiological indices of disease progression in ACLT animals, as compared to saline-treated positive controls (0.0020 ≤ P ≤ 0.0214, one-way ANOVA with Holm post-hoc). MR analysis of knee joints revealed that antagonist-treated animals had greater cartilage thickness and fewer cartilage lesions (Figure 3d,e,f), as well as shorter cartilage T₂ (Table 4, statistical significance not achieved) than saline-treated ACLT animals. These data suggest that antagonist treatment protected joint radiomorphology after ACLT.

To investigate the effects of ETA and/or BKB1 antagonist treatment on histological indices of disease, rat knee joints were processed to assess cartilage proteoglycan content and joint histomorphology (Figure 4 left and middle columns). ACLT saline-treated animals lost most proteoglycan staining when examined at eight weeks post-operatively, with severe articular surface disruptions and osteophyte formation (Figure 4g,h). In contrast, cartilage proteoglycans were detected in the knees of ETA and/or BKB1 antagonist-treated animals (Figures 4j,k,m,n and 4p,q), indicating that treatment protects cartilage structural components. As well, articular surface integrity was preserved to a greater extent, with dual antagonism appearing to be most protective (Figure 4p,q). Neither sham surgery nor intra-articular injection of saline vehicle negatively affected joint histomorphology (Figures 4a,b and 4d,e). Mean OARSI scores (Table 5 and Additional file 6) indicate that ETA and/or BKB1 antagonist treatment significantly reduced the amount of affected joint tissue and the degree of histopathology, as compared to saline-treated positive controls (P < 0.0001 for all comparisons, one-way ANOVA with Holm post-hoc).

Type II collagen, the major structural collagen of cartilage, was detected by immunohistochemistry (Figure 4 right column). ACLT saline-treated animals displayed significant losses of articular surface type II collagen (Figure 4i) with some localization in the deep zones of cartilage, reflecting cartilage remodeling processes. Animals treated with ETA and/or BKB1 antagonists (Figure 4l,o,r) displayed varying degrees of protection, retaining some type II collagen staining. Neither sham surgery nor intra-articular injection of saline vehicle negatively affected joint type II collagen expression (Figures 4c and 4f): protein was localized in the superficial zone of articular cartilage, indicating functional joint tissue.