Naturally occurring osteoarthritis is one of the principal causes of lameness in horses and frequently contributes to the early retirement of equine athletes.1 Osteoarthritis is a degenerative disease characterized by proteolytic breakdown of the cartilage matrix, fibrillation and erosion of the cartilage surface, and release of breakdown products resulting in synovitis and associated subchondral bone and soft tissue changes. Although features associated with naturally occurring osteoarthritis in the distal portion of the limbs (distal limbs) of horses were described as early as 1973,2 descriptions of osteoarthritis in the DIPJs of horses are scarce in the veterinary literature.3–5
In horses, diagnosis of osteoarthritis in a DIPJ is made on the basis of clinical examination and perineural and intra-articular anesthesia that results in localization of lameness to the hoof as well as findings of various diagnostic imaging modalities such as radiography, ultrasonography, CT, and MRI. Because early lesions of hyaline cartilage may not cause obvious clinical signs,6,7 the use of imaging modalities to identify those lesions in a timely manner is essential to successful treatment and management of affected animals. Conventional MRI methods are based on the detection of differences in the water content of the tissues being scanned and translation of those differences into images. Conventional MRI methods have been used to identify morphological changes in cartilage, which probably represent the later stages of osteoarthritis.8 Those changes are likely preceded by biochemical changes in the extracellular cartilage matrix, which can be evaluated by cartilage-imaging techniques such as dGEMRIC.9 In human medicine, such MRI techniques have been developed and used to identify early stages of osteoarthritis.10–15
As cartilage degenerates, it loses essential negatively charged glycosaminoglycans.16 Delayed gadolinium-enhanced MRI of cartilage uses an anionic, paramagnetic contrast agent, Gd-DTPA2−, which is administered by either the IV or intra-articular route. With progressive loss of glycosaminoglycans, such as occurs in osteoarthritic joints, the negatively charged Gd-DTPA2− penetrates hyaline cartilage and adheres to the positively charged cartilage matrix in areas where glycosaminoglycans have been lost. Parametric mapping of cartilage can be performed by postprocessing of dGEMRIC images to create color maps that depict cartilage relaxation times.7,8
In the equine industry to date, dGEMRIC has been used primarily for research purposes, and only a few studies17,18,a describe its use in horses. In 1 study,17 a 1.5-T scanner was used to assess the feasibility of dGEMRIC in normal cadaveric metacarpophalangeal and metatarsophalangeal joints; however, the thin cartilage of those joints precluded optimal dGEMRIC evaluation. Relaxation times determined by dGEMRIC are similar for fresh, chilled, and frozen joints; and dGEMRIC cartilage maps can be used to obtain accurate measurements of normal cartilage thickness at the distal aspect of the third metacarpal and metatarsal bones when there was no contact with the hyaline cartilage of the first (proximal) phalanx.a In 5 ponies, the use of dGEMRIC and T2-weighted mapping was helpful for serial assessment of the healing of experimentally induced lesions of the femoral condyle that were treated with bone morphogenic protein.18 The premise that changes in the glycosaminoglycan concentration of hyaline cartilage can be accurately detected by dGEMRIC is promising, and the use of dGEMRIC to detect the early stages of osteoarthritis in the distal limbs of horses warrants investigation.
The objectives of the study reported here were to characterize the dGEMRIC features of healthy hyaline cartilage of the DIPJ, to determine whether dGEMRIC can be used to differentiate various stages of naturally occurring osteoarthritis of the DIPJ, and to correlate relaxation times determined by dGEMRIC with the glycosaminoglycan concentration, water content, and macroscopic and histologic findings of hyaline cartilage from DIPJs with and without osteoarthritis. We hypothesized that dGEMRIC would be feasible for imaging the DIPJ of horses and accurately differentiate the various stages of naturally occurring osteoarthritis and that relaxation times determined by dGEMRIC would be correlated with the glycosaminoglycan concentration and water content of cartilage.
Materials and Methods
Sample
The distal portion of the right or left forelimb was harvested from each of 12 adult (mean ± SD age, 15.2 ± 9.2 years; range, 6 to 32 years) warmblood horses that were euthanized for reasons unrelated to the musculoskeletal apparatus. The limb harvested from each horse was randomly selected by means of a coin toss, removed from the body at the level of the middle carpal joint, and stored at 4°C until scanned.
MRI
Each limb was scanned within 24 hours after harvest, and all 12 limbs were scanned on 12 different days over a 4-month period. For each limb, a vitamin E capsule was taped to the lateral aspect of the hoof, and the hoof was positioned with the dorsal hoof wall facing downward and the toe facing toward the gantry. The DIPJ of each limb was scanned at room temperature (approx 21°C) by use of a 16-channel knee coil and 3-T MRI scanner.b
A transverse localizer was used to identify the condylar and intercondylar sagittal slices for each DIPJ. For each lateral and medial midcondylar sagittal slice and central intercondylar sagittal slice positioned in the middle of the distal aspect of the P2 sagittal groove, T1preGd was measured with single-slice inversion recovery spin echo sequences (repetition time, 12 milliseconds; echo time, 5.6 milliseconds; field of view, 100 × 100 mm; matrix, 252 × 244; slice thickness, 3 mm; and receiver band width, 131.6 kHz/pixel). Synoviocentesis of the DIPJ was performed by insertion of a 21-gauge needle into the dorsal recess of the joint and aspiration of as much synovial fluid as possible. Then, 0.05 mL of Gd-DTPA2− (contrast)c was diluted in 5 mL of saline (0.9% NaCl) solution (Gd-DTPA2− dose, 0.025 mmol/joint) and injected into the DIPJ. The distal limb joints were manually flexed for 5 minutes after injection to distribute the Gd-DTPA2−. Two hours (120 minutes) after the Gd-DTPA2− injection, the limb was rescanned, and T1postGd was measured on midcondylar and central intercondylar sagittal slices in the same manner as the T1preGd.
MRI analysis
For each DIPJ, 11 ROIs at the distal aspect of P2 and proximal aspect of P3 were selected for analysis (Figure 1). The locations of those ROIs were standardized and determined on the basis of measurements made on each sagittal midcondylar and intercondylar MRI image (Figure 2). Briefly, a circle of best fit was placed over the distal aspect of P2, and a line was drawn from the center of the circle distally along the longitudinal axis of the P2 metaphysis to represent the center of rotation for the DIPJ. The dorsal ROIs (1 and 7) for the distal aspect of P2 extended 5° dorsal to palmar from a line drawn at a 50° angle dorsal to the proximal aspect of the line representing the center of rotation for DIPJ where there was no contact between the cartilage at the distal aspect of P2 and the cartilage at the proximal aspect of P3 when the joint was in a neutral position. The central ROIs (2 and 8) for the distal aspect of P2 extended 5° dorsal to palmar from the line representing the center of rotation for the DIPJ. The palmar ROIs (3 and 9) for the distal aspect of P2 extended 5° dorsal to palmar from a line drawn at a 55° angle palmar to the proximal aspect of the line representing the center of rotation for DIPJ where there was no contact between the cartilage at the distal aspect of P2 and the cartilage at the proximal aspect of P3 when the joint was in a neutral position. The dorsal ROI (13) for the proximal aspect of P3 was located at the most dorsal aspect of the hyaline cartilage and extended for 10° in a palmar direction. The central ROIs (11, 14, and 17) for the proximal aspect of P3 extended 5° dorsal to palmar from the line representing the center of rotation for the DIPJ. The palmar ROI (15) for the proximal aspect of P3 was located at the most palmar aspect of the hyaline cartilage and extended for 10° in a dorsal direction. A free-hand tool of a commercially available software programd was used to manually draw each circular ROI (mean ± SD number of pixels, 198.5 ± 13.8) over the designated areas of the hyaline cartilage of the distal aspect of P2 and proximal aspect of P3 on images obtained before and after contrast administration. For each ROI, the bulk T1 relaxation time was measured 3 times by each of 2 observers, who were supervised by a senior investigator (PK) and unaware of (blinded to) the osteoarthritis status of each DIPJ. The mean relaxation time for each ROI was then calculated for each observer.
Schematic illustrations of the hyaline cartilage surface of the distal aspect of P2 (A) and proximal aspect of P3 (B) in horses that depict the locations of 11 ROIs that were analyzed on dGEMRIC images and from which osteochondral core biopsy specimens were obtained for macroscopic and histologic assessment. DDFT = Deep digital flexor tendon. lat = Lateral. med = Medial. NB = Navicular bone.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.287
Sagittal intercondylar T2 turbo spin echo MRI image of the DIPJ of a cadaveric equine forelimb. A circle of best fit has been placed over the distal aspect of P2 to help determine the locations of the 11 ROIs depicted in Figure 1. The yellow line that extends from the center of the circle distally along the longitudinal axis of the P2 metaphysis represents the center of rotation for the DIPJ. The dorsal ROIs (1 and 7) for the distal aspect of P2 extended 5° dorsal to palmar (white dashes) from a line drawn at a 50° angle dorsal to the proximal aspect of the center of rotation for the DIPJ, where there was no contact between the cartilage at the distal aspect of P2 and the cartilage at the proximal aspect of P3 when the joint was in a neutral position. The central ROIs (2 and 8) for the distal aspect of P2 extended 5° dorsal to palmar (white dashes) from the line representing the center of rotation for the DIPJ. The palmar ROIs (3 and 9) for the distal aspect of P2 extended 5° dorsal to palmar (white dashes) from a line drawn at a 55° angle palmar to the proximal aspect of the center of rotation for the DIPJ, where there was no contact between the cartilage at the distal aspect of P2 and the cartilage at the proximal aspect of P3 when the joint was in a neutral position. The dorsal ROI (13) for the proximal aspect of P3 was located at the most dorsal aspect of the hyaline cartilage and extended for 10° (white dashes) in a palmar direction. The central ROIs (11, 14, and 17) for the proximal aspect of P3 extended 5° dorsal to palmar (white dashes) from the line representing the center of rotation for the DIPJ. The palmar ROI (15) for the proximal aspect of P3 was located at the most palmar aspect of the hyaline cartilage and extended for 10° (white dashes) in a dorsal direction. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.287
Macroscopic cartilage assessment
After MRI scanning was completed, each DIPJ was disarticulated, and the cartilage surface of the distal aspect of P2 and the proximal aspect of P3 were visually inspected for macroscopic degenerative changes at the ROIs evaluated on MRI images. For each ROI, the cartilage was scored on a scale of 0 to 3 by use of a modified Outerbridge grading system,19 where 0 = normal (no visible lesions); 1 = soft, discolored, or swollen cartilage; 2 = partial-thickness fissures, fragmentation, or erosion of the cartilage; and 3 = full-thickness fissures, fragmentation, or erosion of the cartilage. When multiple lesions were present within a joint, the highest (worst) cartilage score was used for statistical analysis.
Tissue specimen collection
An osteochondral autograft transfer system cutting tubee was used to obtain osteochondral core tissue specimens (diameter, 8 mm) from each ROI in the distal aspect of P2 and proximal aspect of P3. A saw was used to cut a 1,000-μm-thick slice from the center of each specimen, which was processed for histologic evaluation. A scalpel was used to separate the cartilage from the subchondral bone for the remainder of each tissue specimen, and the harvested cartilage specimens were stored frozen at −80°C until glycosaminoglycan and water content analyses were performed.
Histologic and immunohistochemical analyses
Osteochondral specimens designated for histologic evaluation were fixed in 4% paraformaldehyde for 48 hours, decalcified in 25% EDTA for 4 weeks, and embedded in paraffin. Specimens were then cut into 2- to 3-μm-thick sections and stained with H&E, safranin O fast green, and toluidine blue stains for histologic evaluation. Sections designated for immunohistochemical staining for detection of collagen type II were deparaffinized, incubated and digested with hyaluronidase, washed, and incubated with rabbit anti-collagen II antibodyf as the primary antibody and biotinylated goat anti-rabbit antibodyg as the secondary antibody.
Osteochondral specimens stained with safranin O fast green stain were assessed by each of 3 observers, who were supervised by a senior investigator (MH) and were blinded to the osteoarthritis status of the joint from which each specimen was obtained. Each observer screened an overview of each section and evaluated the ROI with 4 to 5 hpf, and the extent and nature of tissue staining, chondrocyte structure and density, and number of chondrocyte clusters within each field were assessed by use of a modified scoring system20 adapted from Mankin et al.21 For each observer, the mean score was calculated for each section, and the section was classified into 1 of 3 categories on the basis of the overall mean histologic score (normal cartilage [histologic score, 0 to 1.9], mild to moderate osteoarthritis [histologic score, 2.0 to 8.0], or severe osteoarthritis [histologic score, 8.1 to 16]). For immunohistochemically stained sections, the extent and location of stain uptake was assessed for each ROI by 1 observer (ASB) under the supervision of a senior investigator (MH).
Cartilage water and glycosaminoglycan content
Cartilage water content was determined by use of a modification of a previously described method.22 Briefly, cartilage specimens were weighed before (wet weight) and after (dry weight) lyophilization with a speed vacuum machine for 4 hours.22 Cartilage water content was expressed as a percentage of the wet weight and was calculated as ([wet weight – dry weight]/wet weight) × 100.
Cartilage specimens designated for determination of glycosaminoglycan concentration were sent to the Tetec Laboratory, Reutilingen, Germany, for analysis. Cartilage glycosaminoglycan concentration was determined by use of a modified method of the dimethylmethylene blue assay developed by Farndale et al23 with light absorbance measured at 520 nm. Results were expressed in μg/mg of cartilage normalized on the basis of wet weight.
Statistical analysis
Data were stored in a commercially available spreadsheet program.h The interobserver reliability for mean histologic scores and T1preGds and T1postGds was evaluated with an ICC, whereby interobserver reliability was considered good when the ICC was > 0.7, optimal when the ICC was > 0.8, and excellent when the ICC was > 0.9.
The distribution of data for continuous variables (T1preGd, T1postGd, cartilage water content, and cartilage glycosaminoglycan concentration) was assessed for normality by use of the Kolmogorov-Smirnoff test. Results were reported as the mean ± SD for variables with parametric distributions and median (range) for variables with nonparametric distributions. For each ROI, the mean histologic score, T1preGd, and T1postGd for each observer were combined to calculate the overall mean values for those variables, which were used in all analyses. The 11 ROIs were further categorized into 2 different sets of cartilage-zone subgroups (condyle [ROIs 1, 2, 3, 7, 8, 9, 11, and 17] or condylar groove [ROIs 13, 14, and 15] and dorsal [ROIs 1, 7, and 13], central [ROIs 2, 8, 11, 14, and 17], or palmar [ROIs 3, 9, and 15] zones) on the basis of their location in the DIPJ.
The Spearman rank correlation coefficient (rS) was used to assess the correlation between the macroscopic and histologic scores; the respective correlations between cartilage glycosaminoglycan concentration and T1preGd, T1postGd, and macroscopic and histologic scores; and the respective correlations between cartilage water content and T1preGd, T1postGd, and macroscopic and histologic scores. All correlation analyses were performed with a commercially available statistical software program,i and values of P < 0.05 were considered significant.
A mixed generalized linear model approach was used to analyze the relationship between each dependent or outcome variable of interest (T1preGd and T1postGd). Independent variables evaluated as fixed effects in the respective models included Gd-DTPA2− (before or after administration; included only in the T1preGd model), cartilage surface (P2 or P3), both cartilage-zone subgroups (condyle-condylar groove and dorsal-central-palmar zones), macroscopic score (0 to 3), and osteoarthritis status (none [normal cartilage], minimal to moderate, or severe). Random variables for joint and replicate were included in both models to account for repeated measures within joints. Results of exploratory analyses indicated that use of a γ distribution to model the dependent variable provided the best fit for the data. Each model was built by means of backward elimination, and only independent variables with values of P < 0.05 were retained in the final model unless otherwise specified. Both generalized linear models were created with statistical software.j
Results
Specimens
A total of 132 osteochondral core tissue specimens were collected from the 12 DIPJs (specimens collected from each of the 11 ROIs for each joint), but 10 specimens were lost during processing. Therefore, tissue specimens from 122 ROIs were evaluated.
Macroscopic cartilage assessment
Of the 122 osteochondral core tissue specimens (122 ROIs) evaluated, 45 were assigned a macroscopic score of 0 (normal cartilage), 46 were assigned a macroscopic score of 1 (soft, discolored, or swollen cartilage), 21 were assigned a macroscopic score of 2 (cartilage with partial-thickness fissures, fragmentation, or erosions), and 10 were assigned a macroscopic score of 3 (cartilage with full-thickness fissures, fragmentation, or erosions). The macroscopic score was significantly (P < 0.001) associated with the T1preGd. The median T1preGd for specimens with a macroscopic score of 3 (695 milliseconds; range, 585 to 853 milliseconds) was significantly greater than that for specimens with a macroscopic score of 0 (611 milliseconds; range, 481 to 747 milliseconds; P < 0.001) or 1 (628 milliseconds; range, 471 to 789 milliseconds; P < 0.025) but did not differ significantly (P = 0.181) from the median T1preGd for specimens with a macroscopic score of 2 (659 milliseconds; range, 556 to 779 milliseconds). There was a significant (P < 0.001 for both comparisons) positive correlation between cartilage macroscopic score and mean overall histologic score (rS = 0.71) and mean histologic staining score (rS = 0.68).
MRI analysis
Evaluation of dGEMRIC color-coded maps obtained before and after contrast administration for each joint clearly revealed that cartilage glycosaminoglycan concentration decreased as the severity of degenerative lesions progressed, which corresponded to an increase in Gd-DTPA2− binding with the positively charged cartilage matrix in the glycosaminoglycandeficient areas and a progressive increase in the magnitude of the difference between T1preGd and T1postGd (Figure 3). The median (range) T1preGds and T1postGds for osteochondral core tissue specimens with normal cartilage (ie, no degenerative lesions) and degenerative lesions consistent with mild to moderate and severe osteoarthritis were summarized (Table 1). Interobserver reliability was good for T1preGd (ICC, 0.78) and excellent for T1postGd (ICC, 0.94).
Representative color-coded maps created by postprocessing of dGEMRIC images of the distal sagittal medial (A through D) and lateral (E and F) midcondylar regions of the P2 of equine cadaveric forelimbs obtained before (A, C, and E) and after (B, D, and F) 0.05 mL of Gd-DTPA2− diluted in 5 mL of saline (0.9% NaCl) solution (Gd-DPTA2− dose, 0.025 mmol; contrast) was injected into the dorsal recess of the DIPJ. The maps depict various standardized ROIs described in Figure 1 in which the cartilage was evaluated by dGEMRIC and macroscopic and histologic examination. The colored bar on the right side of each color-coded map details the cartilage relaxation time associated with each color. Black corresponds to a cartilage relaxation time of 1 millisecond and is indicative of the absence of glycosaminoglycans, whereas dark red corresponds to a cartilage relaxation time of 1,000 milliseconds and is indicative of high concentrations of glycosaminoglycans. Notice that the articular cartilage is typically depicted by light green and blue pixels in postcontrast images (B, D, and F) versus red and yellow pixels in precontrast maps (A, C, and E), which indicated that the T1postGd was less than the T1preGd. On precontrast maps, the relaxation time for cartilage generally decreased from the cartilage surface to the surface of subchondral bone (red pixels transitioning to yellow pixels). The cartilage surface was more clearly delineated on postcontrast maps than on precontrast maps. Within each row of images, the images to the right of the color-coded maps are photomicrographs of safranin O fast green–stained sections of the articular cartilage from the distal portion of P2 obtained from the ROIs that correspond to those highlighted in the color-coded maps. A and B—Normal cartilage with no histologic lesions consistent with osteoarthritis evident at ROIs 7 and 9. C and D—Mild osteoarthritis. In the postcontrast map (D), notice the presence of light to dark blue pixels among the normal green pixels for the articular cartilage at ROI 7, which indicated that the cartilage at that ROI had a shorter relaxation time and lower glycosaminoglycan concentration than normal cartilage. In the corresponding photomicrograph for ROI 7, there is an area in which approximately two-thirds of the hyaline cartilage has been eroded and the adjacent cartilage has a lack of staining and fibrillations on the articular surface. E and F—Moderate osteoarthritis. In the postcontrast map (F), notice the presence of light to dark blue pixels consistent with an abnormally short relaxation time and low glycosaminoglycan concentration in the articular cartilage at ROIs 1 and 2. In the corresponding photomicrographs, notice that there is an area in which approximately a third of the hyaline cartilage has been eroded at ROI 1 and an area in which approximately two-thirds of the hyaline cartilage has been eroded at ROI 2. There is also lack of staining and superficial fibrillation in the cartilage adjacent to the erosions at both ROIs. For all photomicrographs, bar = 100 μm.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.287
Median (range) cartilage T1preGd and T1postGd as determined by dGEMRIC and cartilage water content and glycosaminoglycan concentration for 122 osteochondral core tissue specimens obtained from standardized ROIs in the DIPJ of cadaveric forelimbs harvested from 12 adult warmblood horses with and without naturally occurring osteoarthritis of that joint.
| Variable | Normal cartilage (n = 40) | Mild to moderate osteoarthritis (n = 68) | Severe osteoarthritis (n = 14) |
|---|---|---|---|
| T1preGd (ms) | 607 (473–732) | 642 (463–797) | 692 (454–853) |
| T1preGd (ms) | 234 (182–330) | 217 (116–334) | 210 (124–295) |
| Water (%) | 71 (59–83) | 73 (48–82) | 77 (64–83) |
| Glycosaminoglycan (μg/mg) | 60 (21–104) | 50 (18–106) | 30 (15–73) |
A histologic score was assigned to each tissue specimen on the basis of a modified scoring system20 adapted from Mankin et al21 by each of 3 observers; the scores for the 3 observers were used to calculate the mean overall histologic score, which could range from 0 (no evidence of osteoarthritis; normal) to 16 (severe osteoarthritis). Each specimen was classified into 1 of 3 categories on the basis of its mean overall histologic score (normal cartilage [histologic score, 0 to 1.9], mild-moderate osteoarthritis [histologic score, 2.0 to 8.0], or severe osteoarthritis [histologic score, 8.1 to 16]).
Results of the mixed generalized linear model indicated that cartilage relaxation time was decreased significantly (P < 0.001) following Gd-DTPA2− administration and that T1preGd was significantly associated with cartilage surface, the dorsal-central-palmar cartilage zone, and osteoarthritis status but was not significantly associated with the condyle-condylar groove cartilage zone. The median T1preGd for the articular cartilage of P3 (675 milliseconds; range, 476 to 853 milliseconds) was significantly (P < 0.001) greater than that for the articular cartilage of P2 (601 milliseconds; range, 454 to 767 milliseconds). The median T1preGd for cartilage located in the dorsal zone (ROIs 1, 7, and 13) of the DIPJ (606 milliseconds; range, 463 to 739 milliseconds) was significantly less than that for cartilage in both the palmar (ROIs 3, 9, and 15; 629 milliseconds; range, 434 to 797 milliseconds; P = 0.041) and middle (ROIs 2, 8, 11, 14, and 17; 622 milliseconds; range, 489 to 923 milliseconds; P = 0.046) zones of the DIPJ. The median T1preGd for specimens with severe osteoarthritis (692 milliseconds; range, 454 to 853 milliseconds) was significantly greater than that for specimens with mild to moderate osteoarthritis (642 milliseconds; range, 463 to 797 milliseconds; P = 0.005) and specimens with normal cartilage (607 milliseconds; range, 473 to 732 milliseconds; P = 0.001). The median T1preGd did not differ significantly between cartilage located at the condyle (ROIs 1, 2, 3, 7, 8, 9, 11, and 17; 621 milliseconds; range, 434 to 923 milliseconds) and condylar groove (ROIs 13, 14, and 15; 667 milliseconds; range, 476 to 853 milliseconds), and that variable was omitted from the final model.
The median T1postGd for cartilage specimens with severe osteoarthritis (210 milliseconds; range, 124 to 295 milliseconds) was significantly shorter than that for specimens with mild to moderate osteoarthritis (217 milliseconds; range, 116 to 334 milliseconds; P < 0.001) or normal cartilage (234 milliseconds; range, 182 to 330 milliseconds; P < 0.001). Although T1postGd was significantly (P = 0.012) associated with the fixed effect for macroscopic score, the median T1postGd did not differ significantly for any of the pairwise comparisons between individual macroscopic scores. Results of the final mixed generalized linear model indicated that T1postGd was not significantly associated with cartilage surface (P = 0.21), the condyle-condylar groove cartilage zone (P = 0.30), or the dorsal-central-palmar cartilage zone (P = 0.17), although that zone was retained in the final model.
Histologic and immunohistochemical analyses
Interobserver reliability was excellent for the histologic score (ICC, 0.98). Normal hyaline cartilage was characterized by a smooth surface and uniform histologic stain uptake throughout the extracellular matrix atop normal subchondral bone and minimal uptake of the immunohistochemical stain within the superficial cartilage zone (Figure 4). Mild osteoarthritis was characterized by marked fibrillations and fissures in the hyaline cartilage surface with a lack of histologic stain uptake in the superficial cartilage zone and evidence of chondrocyte cluster formation; however, there was diffuse immunohistochemical stain uptake in the superficial cartilage zone, which indicated the presence of collagen type II immediately below the damaged cartilage surface. Moderate osteoarthritis was characterized by partial-thickness erosion of and lack of histologic stain uptake in the superficial hyaline cartilage layer, and similar to mild osteoarthritis, there was diffuse immunohistochemical stain uptake in the hyaline cartilage immediately subjacent to the damaged cartilage surface. Severe osteoarthritis was characterized by full-thickness fissures or erosions of the hyaline cartilage layer with lack of histologic stain uptake in focal areas of the remaining hyaline cartilage and marked chondrocyte cluster formation as well as cracks in calcified cartilage and a decrease in the amount and size of bone lacunae underlying it. In some areas, the calcified cartilage and underlying subchondral bone collapsed, and cartilage islands were present within the subchondral bone. The intensity of immunohistochemical stain uptake (ie, amount of denatured collagen type II) in the extracellular matrix of the remaining hyaline cartilage increased as osteoarthritis severity increased.
Representative photomicrographs of sections of articular cartilage stained with safranin O fast green stain (A through D) or immunohistochemical stain for detection of collagen type II (E through H) that were obtained from the distal aspect of P2 or proximal aspect of P3 of equine cadaveric forelimbs and that depict healthy hyaline cartilage (A and E) and hyaline cartilage with lesions consistent with mild (B and F), moderate (C and G), and severe (D and H) osteoarthritis. A and E—Normal hyaline cartilage was characterized by a smooth surface and uniform stain uptake throughout the extracellular matrix atop normal subchondral bone. There was minimal uptake of the immunohistochemical stain within the superficial cartilage zone. B and F—Mild osteoarthritis was characterized by marked fibrillations and fissures in the hyaline cartilage surface with a lack of histologic stain uptake in the superficial cartilage zone and evidence of chondrocyte cluster formation; however, there was diffuse immunohistochemical stain uptake in the superficial cartilage zone, which indicated the presence of collagen type II immediately below the damaged cartilage surface. C and G—Moderate osteoarthritis was characterized by partial-thickness erosion of and lack of histologic stain uptake in the superficial hyaline cartilage layer, and similar to mild osteoarthritis, there was diffuse immunohistochemical stain uptake in the hyaline cartilage immediately subjacent to the damaged cartilage surface. D and H—Severe osteoarthritis was characterized by full-thickness fissures or erosions of the hyaline cartilage layer with lack of histologic stain uptake in focal areas of the remaining hyaline cartilage and marked chondrocyte cluster formation as well as cracks in calcified cartilage and a decrease in the amount and size of bone lacunae underlying it. In some areas, the calcified cartilage and underlying subchondral bone collapsed, and cartilage islands were present within the subchondral bone. The intensity of immunohistochemical stain uptake (ie, amount of collage type II) in the extracellular matrix of the remaining hyaline cartilage increased as osteoarthritis severity increased.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.287
Biochemical analysis
The median (range) glycosaminoglycan concentration and water content for normal cartilage and cartilage with mild to moderate and severe osteoarthritis were summarized (Table 1). There was a significant negative correlation between cartilage glycosaminoglycan concentration and T1preGd (rS, −0.34; P < 0.001), mean histologic score (rS, −0.39; P < 0.001), and macroscopic score (rS, −0.37; P < 0.001); however, cartilage glycosaminoglycan concentration was not significantly (P = 0.321) correlated with T1postGd (Figure 5). Conversely, there was a significant positive correlation between cartilage water content and T1preGd (rS, 0.39; P < 0.001), mean histologic score (rS, 0.32; P < 0.001), and macroscopic score (rS, 0.24; P < 0.001). Cartilage water content was not significantly (P = 0.31) correlated with T1postGd.
Scatterplots depicting the respective correlations between cartilage glycosaminoglycan concentration and T1preGd (A), T1postGd (B), mean overall histologic score (C), and macroscopic score (D) as well as between cartilage water content and T1preGd (E), T1postGd (F), overall mean histologic score (G), and macroscopic score (H) for 122 osteochondral core tissue specimens obtained from standardized ROIs in the distal aspect of P2 or proximal aspect of P3 of 12 equine cadaveric forelimbs. Prior to collection of the tissue specimens, the DIPJ was disarticulated and visually inspected for degenerative lesions. A macroscopic score was assigned to each ROI on a scale of 0 to 3, where 0 = normal (no visible lesions); 1 = soft, discolored, or swollen cartilage; 2 = partial-thickness fissures, fragmentation, or erosion of the cartilage; and 3 = full-thickness fissures, fragmentation, or erosion of the cartilage. A histologic score was assigned to each tissue specimen on the basis of a modified scoring system20 adapted from Mankin et al21 by each of 3 observers; the scores for the 3 observers were used to calculate the overall mean histologic score, which could range from 0 (no evidence of osteoarthritis; normal) to 16 (severe osteoarthritis). The solid line in each scatterplot represents the line of best fit for the data. The Spearman rank correlation coefficient (rS) and associated P value are provided for each scatterplot. GAG = Glycosaminoglycan.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.287
Discussion
In the present study, dGEMRIC was used to evaluate the DIPJ of equine cadaveric forelimbs. Results indicated that the T1preGd and T1postGd determined by dGEMRIC were correlated with the extent of degenerative cartilage lesions observed both macroscopically and histologically. Additionally, there was a significant negative correlation between T1preGd and cartilage glycosaminoglycan concentration and significant positive correlation between T1preGd and cartilage water content. Thus, it appeared that dGEMRIC can be used to detect the presence and severity of osteoarthritis in the DIPJ of horses.
For the DIPJs evaluated in the present study, the median relaxation time for normal cartilage decreased from 607 milliseconds (range, 473 to 732 milliseconds) prior to Gd-DTPA2− (contrast) administration to 234 milliseconds (range, 182 to 330 milliseconds) after contrast administration. In a study17 by Carstens et al, in which dGEMRIC was used to evaluate cartilage in the metacarpophalangeal and metatarsophalangeal (fetlock) joints of equine cadavers, the mean relaxation time for normal cartilage 120 minutes after intra-articular contrast administration was 650 milliseconds. The difference in the T1postGds determined for the articular cartilage of the DIPJs evaluated in the present study and the articular cartilage of the fetlock joints evaluated in the Carstens et al study17 may reflect inherent differences in the composition or thickness of the articular cartilage of those 2 joints. In horses, the distance from the surface of the articular cartilage to the bone cartilage interface of the third metacarpal or third metatarsal bone as determined by light microscopy ranges from 0.79 to 0.99 mm.24 Because the articular cartilage of the equine fetlock joint is so thin, the investigators of that study24 had to analyze a combination of the thin opposing cartilage surfaces and a thin layer of synovial fluid at some sites in the joints evaluated. In the DIPJ, the mean cartilage thickness ranges from 2.1 mm dorsally to 3.1 mm palmarly.25 The fairly thick cartilage of the DIPJs evaluated in the present study enabled us to evaluate only the hyaline cartilage at the specified ROIs, which may have contributed to the discrepancies in the T1postGds for the cartilage of the DIPJs of this study and the cartilage of the fetlock joints of the Carstens et al study.17 The Gd-DTPA2−dose (0.025 mmol/joint) administered in the DIPJs of this study was the same as that used in the fetlock joints of the Carstens et al study.17 Because the DIPJ is smaller than the fetlock joint, the Gd-DTPA2− dose relative to the surface area of the joint for this study may have actually been greater than that for the Carstens et al study,17 which may also have contributed to the substantially lower T1postGd. In the present study, postcontrast images were acquired 120 minutes after contrast administration on the basis of findings of the Carstens et al study17 in which relaxation time appeared to have a decreasing pattern for images acquired > 120 minutes after contrast administration.
Proteoglycans and their glycosaminoglycan side chains are lost during cartilage degeneration, and the glycosaminoglycans diffuse into the adjacent tissue. Disruption in the cartilage matrix also results in an increase in cartilage water content. That pathogenic mechanism was supported by the findings of the present study. Cartilage glycosaminoglycan concentration decreased as macroscopic and histologic scores increased, which were indicative of increasing severity of osteoarthritis. Delayed gadolinium-enhanced MRI of cartilage was developed for the indirect assessment of hyaline cartilage glycosaminoglycan concentration.15,26–28 Compared with normal cartilage, greater concentrations of Gd-DTPA2− accumulate in hyaline cartilage with low glycosaminoglycan concentrations, which results in shorter T1postGds, as was observed in the present study. That finding was further supported by the fact that T1postGd decreased as the extent of safranin O fast green (histologic) stain uptake in cartilage decreased (ie, stain uptake was less than expected) and the cartilage water content increased, both of which were indicative of proteoglycan loss and progression of osteoarthritis. Overall, the findings of the present study supported our hypotheses that dGEMRIC could be used to image the DIPJ of horses and accurately differentiate various stages of naturally occurring osteoarthritis. However, T1postGd was not significantly correlated with cartilage glycosaminoglycan concentration or water content, and we cannot definitively explain that finding. It is possible that the glycosaminoglycan concentration and water content in the osteochondral core tissue specimens evaluated in this study were adversely affected by storage conditions or the laboratory techniques used, despite those analyses being performed by technicians at an experienced and qualified laboratory. Alternatively, the sample size may have been too small to detect significant correlations between T1postGd and cartilage glycosaminoglycan concentration and T1postGd and cartilage water content.
Cartilage zone or location within the DIPJ was significantly associated with T1preGd, with cartilage specimens obtained from the dorsal zone of the joint having a significantly shorter median T1preGd than cartilage specimens obtained from the middle or palmar zones of the joint. Variation of MRI features on the basis of topographic location (topographic variation) has been described for the human knee29 and likely reflects the response of hyaline cartilage to physiologic loading and the relationship between the biomechanical and biochemical constitution of the tissue, resulting in varying relaxation times. Regional differences have also been identified for cartilage extracellular matrix components of equine fetlock joints, with cartilage at the dorsal aspect of the joint having the lowest glycosaminoglycan concentration.22 To our knowledge, normal topographic variation and reference intervals for glycosaminoglycan concentration and water content for normal cartilage of the DIPJ of horses have not been reported. Thus, it is likely some of the variability observed for the DIPJs evaluated in the present study simply reflected normal inherent characteristics of that joint. For example, the dorsal zone of the DIPJ may undergo intermittent loading such that glycosaminoglycan production is impaired (thereby decreasing cartilage relaxation times as determined by dGEMRIC), whereas the central and palmar zones of the joint are exposed to fairly consistent loads, which leads to greater glycosaminoglycan production and increases cartilage relaxation times.
For the DIPJs evaluated in the present study, the median T1preGd for articular cartilage of P3 (675 milliseconds) was significantly greater than that for the articular cartilage of P2 (601 milliseconds), which suggested that the glycosaminoglycan concentration for P3 cartilage was greater than that for P2 cartilage, likely owing to continuous loading of the articular surface of P3. The number of DIPJs evaluated in this study was too small to create a reliable and complete topographic cartilage relaxation time map of healthy joints, but that was not one of our objectives. Nevertheless, the topographic variation described for the joints of this study should be considered when dGEMRIC images of the DIPJs of horses are interpreted.
The histologic scoring system used to assess lesion severity in the present study has been used to assess degenerative lesions associated with naturally occurring osteoarthritis in carpal bones20 and the distal aspect of metacarpal bones30 of horses. It has proven to be repeatable among observers and provides a reliable assessment of the degree of cartilage degeneration in patients with naturally occurring osteoarthritis in those studies20,30 as well as in the present study. For the DIPJs evaluated in the present study, the focal nature of the cartilage lesions observed within the articular surface was striking, with areas of diseased cartilage in close proximity to areas of normal cartilage within the same histologic specimen, and was consistent with findings in horses with post-traumatic carpal osteoarthritis.31
The present study had a few unavoidable limitations. Although a template was used to match the location of the ROIs evaluated on the MRI images with the locations from which osteochondral core tissue specimens were obtained, small deviations may have occurred. To preclude that limitation, the actual osteochondral core tissue specimens would need to be scanned and then histologically evaluated. However, for this study, we chose to scan the entire DIPJ to more closely mimic a clinical scenario. The cadaveric DIPJs were obtained from horses of various ages. Thus, the lesions observed were likely caused by a combination of factors including overload injuries and age-related osteoarthritis. Also, we did not consider that cartilage relaxation times may vary on the basis of spatial orientation or location. For the DIPJs of the present study as well as in a study29 of human knees, the T1preGd for articular cartilage decreased from superficial to deep tissues, whereas the T1postGd for articular cartilage increased from superficial to deep tissues. The variation in articular cartilage relaxation times from superficial to deep tissues warrants further investigation.
In horses, dGEMRIC has historically been used primarily for research purposes. The acquisition of dGEMRIC images is time-consuming and requires the administration of a contrast agent. Horses must be anesthetized and repositioned following contrast injection, which takes time and can be challenging. Currently available standing MRI systems have insufficient magnet strength, and the resolution is not appropriate for accurate dGEMRIC of equine cartilage, especially in areas where the cartilage is thin. Areas that warrant additional research include the effect of IV versus intra-articular administration of Gd-DTPA2−, effect of exercise on the Gd-DTPA2− uptake by hyaline cartilage, and determination of topographic variations in T1preGd and T1postGd for normal hyaline cartilage.
To our knowledge, the present study was the first to describe the use of dGEMRIC to evaluate the DIPJs of equine cadaveric forelimbs with and without osteoarthritis of varying severity. Results indicated that naturally occurring osteoarthritis significantly affects cartilage relaxation times as determined by dGEMRIC. Further research is necessary to investigate the relationship between cartilage relaxation times and cartilage glycosaminoglycan concentration. Nevertheless, the findings of this study provided a foundation regarding specific MRI sequences for use with dGEMRIC in future studies and may help facilitate early recognition of osteoarthritis in the DIPJ of horses in both research and clinical settings.
Acknowledgments
This manuscript represents a portion of a thesis submitted by Dr. Bischofberger to the Graduate School for Cellular and Biomedical Sciences, University of Bern, Switzerland, as partial fulfilment of the requirements for a Doctor of Philosophy degree.
Supported by the Swiss Veterinary Association.
The authors declare that there were no conflicts of interest.
The authors thank Drs. Andrea Bachmann, Regula Fürst, and Angela Kirchmeier for assistance with image analysis.
ABBREVIATIONS
| dGEMRIC | Delayed gadolinium-enhanced magnetic resonance imaging of cartilage |
| DIPJ | Distal interphalangeal joint |
| Gd-DTPA2− | Gadolinium diethylene triamine pentaacetic acid |
| ICC | Intraclass correlation coefficient |
| P2 | Second (middle) phalanx |
| P3 | Third (distal) phalanx |
| ROI | Region of interest |
| T1preGd | Precontrast T1-weighted relaxation time |
| T1postGd | Postcontrast T1-weighted relaxation time |
Footnotes
Carstens A. Delayed gadolinium enhanced magnetic resonance imaging and T2 mapping of cartilage of the cadaver distal metacarpus 3–metatarsus 3 of the normal Thoroughbred horse. PhD thesis, University of Pretoria, Pretoria, South Africa, 2013.
Phillips Health Care Ingenia, Phillips AG, Zurich, Switzerland.
Magnevist gadopentate dimeglumine, Bayer Health Care Pharmaceuticals, Zurich, Switzerland.
Relaxation Maps Tool, version 2.1.2, Koninklijke Philips Electronics N V, Phillips AG, Zurich, Switzerland.
Arthrex Inc, Naples, Fla.
Ab34712, Abcam Inc, Cambridge, Mass.
Thermo Fisher Scientific, Schaumburg, Ill.
Excel, Microsoft Corp, Redmond, Wash.
SPSS, IBM Corp, Armonk, NY.
R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available at: www.R-project.org. Accessed May 19, 2017.
References
1. Jeffcott LB, Rossdale PD, Freestone J, et al. An assessment of wastage in Thoroughbred racing from conception to 4 years of age. Equine Vet J 1982;14:185–198.
2. Nilsson G, Olsson SE. Radiologic and patho-anatomic changes in the distal joints and the phalanges of the Standardbred horse. Acta Vet Scand Suppl 1973;44:1–57.
3. Dyson SJ. Lameness due to pain associated with the distal interphalangeal joint: 45 cases. Equine Vet J 1991;23:128–135.
4. McKnight AL, Posh J. Articular cartilage lesions seen with MRI at 0.25T in the distal equine limb. J Equine Vet Sci 2012;32:667–671.
5. Kristiansen KK, Kold SE. Multivariable analysis of factors influencing outcome of 2 treatment protocols in 128 cases of horses responding positively to intra-articular analgesia of the distal interphalangeal joint. Equine Vet J 2007;39:150–156.
6. Bekkers JE, Creemers LB, Dhert WJ, et al. Diagnostic modalities for diseased articular cartilage—from defect to degeneration: a review. Cartilage 2010;1:157–164.
7. Pease A. Biochemical evaluation of equine articular cartilage through imaging. Vet Clin North Am Equine Pract 2012;28:637–646.
8. Choi JA, Gold GE. MR imaging of articular cartilage physiology. Magn Reson Imaging Clin N Am 2011;19:249–282.
9. Domayer SE, Welsch GH, Dorotka R, et al. MRI monitoring of cartilage repair in the knee: a review. Semin Musculoskelet Radiol 2008;12:302–317.
10. Gold GE, Burstein D, Dardzinski B, et al. MRI of articular cartilage in OA: novel pulse sequences and compositional/functional markers. Osteoarthritis Cartilage 2006;14 (suppl A):A76–A86.
11. Potter HG, Black BR. Chong le R. New techniques in articular cartilage imaging. Clin Sports Med 2009;28:77–94.
12. Trattnig S, Mamisch TC, Pinker K, et al. Differentiating normal hyaline cartilage from post-surgical repair tissue using fast gradient echo imaging in delayed gadolinium-enhanced MRI (dGEMRIC) at 3 Tesla. Eur Radiol 2008;18:1251–1259.
13. Trattnig S, Marlovits S, Gebetsroither S, et al. Three-dimensional delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) for in vivo evaluation of reparative cartilage after matrix-associated autologous chondrocyte transplantation at 3.0T: preliminary results. J Magn Reson Imaging 2007;26:974–982.
14. Bekkers JE, Bartels LW, Benink RJ, et al. Delayed gadolinium enhanced MRI of cartilage (dGEMRIC) can be effectively applied for longitudinal cohort evaluation of articular cartilage regeneration. Osteoarthritis Cartilage 2013;21:943–949.
15. Trattnig S, Mlynárik V, Breitenseher M, et al. MRI visualization of proteoglycan depletion in articular cartilage via intravenous administration of Gd-DTPA. Magn Reson Imaging 1999;17:577–583.
16. Frisbie DD. Synovial joint biology and pathobiology. In: Auer J, Stick J, eds. Equine surgery. 4th ed. St Louis: Elsevier Saunders, 2012;1096–1114.
17. Carstens A, Kirberger RM, Velleman M, et al. Feasibility for mapping cartilage T1 relaxation times in the distal metacarpus3/metatarsus3 of Thoroughbred racehorses using delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC): normal cadaver study. Vet Radiol Ultrasound 2013;54:365–372.
18. Menendez MI, Clark DJ, Carlton M, et al. Direct delayed human adenoviral BMP-2 or BMP-6 gene therapy for bone and cartilage regeneration in a pony osteochondral model. Osteoarthritis Cartilage 2011;19:1066–1075.
19. Outerbridge RE. The etiology of chondromalacia of the patella. J Bone Joint Surg Br 1961;43:752–757.
20. Lacourt M, Gao C, Li A, et al. Relationship between cartilage and subchondral bone lesions in repetitive impact trauma–induced equine osteoarthritis. Osteoarthritis Cartilage 2012;20:572–583.
21. Mankin HJ, Dorfman H, Lippiell L, et al. Biochemical and metabolic abnormalities in articular cartilage from osteoarthritic human hips. II. Correlation of morphology with biochemical and metabolic data. J Bone Joint Surg Am 1971;53:523–537.
22. Brama PA, Tekoppele JM, Bank RA, et al. Topographical mapping of biochemical properties of articular cartilage in the equine fetlock joint. Equine Vet J 2000;32:19–26.
23. Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation and discrimination of sulfated glycosaminoglycans by use of dimethylmethylene blue. Biochim Biophys Acta 1986;883:173–177.
24. Carstens A, Kirberger RM, Dahlberg LE, et al. Validation of delayed gadolinium-enhanced magnetic resonance imaging of cartilage and T2 mapping for quantifying distal metacarpus/metatarsus cartilage thickness in Thoroughbred racehorses. Vet Radiol Ultrasound 2013;54:139–148.
25. Olive J. Distal interphalangeal articular cartilage assessment using low-field magnetic resonance imaging. Vet Radiol Ultrasound 2010;51:259–266.
26. Bashir A, Gray ML, Hartke J, et al. Nondestructive imaging of human cartilage glycosaminoglycan concentration by MRI. Magn Reson Med 1999;41:857–865.
27. Laurent D, Wasvary J, Rudin M, et al. In vivo assessment of macromolecular content in articular cartilage of the goat knee. Magn Reson Med 2003;49:1037–1046.
28. Mlynárik V, Trattnig S, Huber M, et al. The role of relaxation times in monitoring proteoglycan depletion in articular cartilage. J Magn Reson Imaging 1999;10:497–502.
29. Kurkijärvi JE, Nissi MJ, Kiviranta I, et al. Delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) and T2 characteristics of human knee articular cartilage: topographical variation and relationships to mechanical properties. Magn Reson Med 2004;52:41–46.
30. Vinardell T, Dejica V, Poole AR, et al. Evidence to suggest that cathepsin K degrades articular cartilage in naturally occurring equine osteoarthritis. Osteoarthritis Cartilage 2009;17:375–383.
31. Bertuglia A, Lacourt M, Girard C, et al. Osteoclasts are recruited to the subchondral bone in naturally occurring post-traumatic equine carpal osteoarthritis and may contribute to cartilage degradation. Osteoarthritis Cartilage 2016;24:555–566.
