Materials and Methods

Animals—Fifteen dogs were used in the study. The dogs comprised 13 client-owned dogs with unilateral CrCL rupture and 2 healthy dogs that underwent sham operations on the left stifle joint. All clients provided written consent for inclusion of their dogs in the study, and the sham operations for the 2 healthy control dogs were approved by the ethical commission of the Faculty of Veterinary Medicine, Ghent University.

Unilateral CrCL rupture in the 13 dogs was diagnosed on the basis of clinical findings of unilateral hind limb lameness, palpable cranial drawer movement (except in 1 dog, which had effusion of the stifle joint, radiographic signs of osteoarthritis, and arthroscopic diagnosis of rupture of the caudolateral band of the CrCL), and positive results for a tibial compression test on standard mediolateral radiographic views. The diagnosis of CrCL rupture was confirmed in all dogs at the time of corrective surgery. During surgery, dogs were further categorized with regard to the degree of CrCL rupture (partial or complete rupture). Partial rupture was identified when the craniomedial or caudolateral band of the CrCL was ruptured, whereas complete rupture was identified when the entire CrCL was ruptured.

During corrective surgery, remnants of the ruptured CrCL were completely removed, and the joint was stabilized by use of an extracapsular technique that involved sutures of polyglactin 910.^a Osteoarthritis was subjectively graded during surgery on the basis of the degree of osteophytes on the ridges of the femoral trochlea; the scale ranged from 0 to 3 (0, no osteophytes; 1, slight degree of osteophytes; 2, moderate degree of osteophytes; and 3, extreme degree of osteophytes). Synovitis was also subjectively graded during surgery on the basis of the magnitude of villous hyperplasia of the synovial membranes; the scale ranged from 0 (no hyperplasia) to 3 (extreme hyperplasia). After surgery, dogs were treated by administration of amoxicillin^b (10 mg/kg, PO, q 12 h for 5 days) and carprofen^c (2 mg/kg, PO, q 12 h for 14 days).

Two healthy dogs without pathologic changes to the stifle joints, as determined by physical examination and radiographic assessment, underwent a sham operation of the left stifle joint. The stifle joint was approached through a lateral arthrotomy. The patella was moved medially, and a Gelpi retractor and Wallace retractor were used to open the stifle joint to inspect intra-articular structures. There were no signs of osteoarthritis or damage to intra-articular joint tissues. The joint was routinely closed by use of single sutures of polyglactin 910.^a After surgery, both dogs were administered amoxicillin^b (10 mg/kg, PO, q 12 h for 5 days) and carprofen^c (2 mg/kg, PO, q 12 h for 14 days). Both dogs were allowed limited exercise for 6 weeks, after which they were allowed unlimited exercise. Both dogs were maintained at the Faculty of Veterinary Medicine of Ghent University and were housed, fed, and cared for in accordance with institutional and ethical guidelines.

Radiographic evaluation—Standard mediolateral radiographic views of the affected stifle joints of client-owned dogs were obtained as part of the diagnostic evaluation. Radiographs were evaluated in accordance with modified criteria described elsewhere³¹ to determine the degree of osteoarthritis. The grading (scale of 0 [no osteoarthritis] to 3 [severe osteoarthritis]) was performed by one of the authors (TdB), who was not aware of the dog from which radiographs were obtained.

Collection and processing of SF samples—All dogs (13 dogs with CrCL rupture and 2 control dogs) were examined at 6-month intervals after the initial surgery (for the affected stifle joint in the 13 dogs or for the sham-operated stifle joint of the 2 control dogs). Samples of SF were aspirated from both stifle joints and the left shoulder joint at the time of surgery (day 0) and at each subsequent examination (6, 12, and 18 months). When it was not possible to aspirate a sufficient amount of SF from a joint, a modified washing method that used vitamin B₁₂ as an internal standard was performed as described elsewhere.³² Briefly, a solution that contained 25% vitamin B₁₂^d and 75% saline (0.9% NaCl) solution^e (vol:vol) was injected into the joint; the joint was passively moved, and the solution was aspirated. Spectrophotometric measurements were used to determine the dilution factor of the sample. The SF was centrifuged (9,390 X g for 10 minutes at 8°C) to remove cells and debris. Supernatant was collected and stored at −20°C until testing.

Anticollagen type I ELISA—Samples of SF obtained from the dogs were tested for antibodies against collagen type I. The ELISA used has been described elsewhere¹⁴; slight modifications were used for coating concentrations. Briefly, 96-microwell plates^f were coated with highly purified human type I collagen.^g Collagen was diluted in 50mM carbonate-bicarbonate coating buffer (pH, 9.4) to achieve a concentration of 10 Mg/mL. One hundred microliters of coating solution was added to each well and incubated for 24 hours at 4°C. Collagen-coated wells were then washed 4 times with PBS solution supplemented with 0.05% (vol: vol) Tween 20.^h Subsequently, 300 μL of blocking solution that consisted of PBS solution supplemented with 0.2% (vol:vol) Tween 80ⁱ was added to each well. Plates were incubated for 2 hours at 37°C, and blocked wells were then washed 4 times with PBS solution supplemented with 0.05% (vol:vol) Tween 20. Samples of SF were inactivated by incubation at 56°C for 30 minutes, and 20 or 50 μL of inactivated SF was added to the dilution solution (PBS solution supplemented with 0.05% Tween 20) in the wells in serial 2-fold dilutions. For aspirates obtained from a joint by use of injection of vitamin B₁₂, 200 μL was added to wells in serial 2-fold dilutions. Collagen-coated wells were incubated with 100 μL of the diluent alone to serve as a blank. Also, negative and positive control samples were included in each assay. Samples were incubated at approximately 25°C for 2 hours.

Unbound material was removed by washing the wells 4 times with washing solution. Bound antibodies were detected by the addition of 100 μL of an appropriate dilution of affinity-isolated rabbit anti-dog IgG (whole molecule) conjugated to horseradish peroxidase.^j The conjugate was incubated for 1 hour at 37°C. Wells were again washed 4 times with washing solution, and binding of conjugate was developed by the addition of 50 μL of 2a2-azino-di-(3-ethylbenzthiazoline sulfonate[6]) diammonium salt crystals substrate.^k After incubation for 1 hour at 37°C, absorbance was measured at 405 nm by use of an ELISA reader.^l

Nonspecific binding was controlled in parallel for each sample of SF by determining the absorbance obtained without collagen. To minimize the effects of variation among days and among plates, SF samples obtained from the same dog were always tested in a single batch and on the same plate.

Antibody titers—Anticollagen antibody titer of a sample was the highest dilution of that sample with an OD above the cutoff value. Samples of SF obtained from the sham-operated dogs were used to calculate the cutoff values used to determine antibody titers of the affected dogs. The cutoff value was determined by calculating the sum of the mean + 2 SD of the OD of each joint (left shoulder joint, sham-operated left stifle joint, and normal right stifle joint) at the corresponding time point (day 0 and 6, 12, and 18 months after surgery). Thus, the cutoff value of the left shoulder joint was used to determine the anticollagen type I antibody titer of the left shoulder joint of the affected dogs, the value of the sham-operated stifle joint was used to determine the anticollagen antibody titer for the CrCL-deficient stifle joint of the affected dogs, and the value of the contralateral stifle joint of the sham-operated dogs was used to determine the anticollagen antibody titer in the contralateral stifle joints of the affected dogs.

Statistical analysis—Mann-Whitney rank sum tests were used to compare data, such as age, body weight, duration of lameness, degree of osteoarthritis determined during surgery, degree of synovitis, and meniscus lesions, between dogs with partial and complete CrCL rupture. The Mann-Whitney rank sum test was also used to compare anticollagen type I antibody titers for affected stifle joints between dogs with partial and complete CrCL rupture, among the various joints, among the assessment time points, and between dogs that ruptured the contralateral CrCL during the study and dogs that did not. Correlation between the anticollagen type I antibody titers and radiographic osteoarthritis grade was investigated by use of the Spearman rank correlation test. Results were considered significant at values of P ≤ 0.05.

Results

Clinical variables—The 13 dogs with CrCL rupture comprised 5 Golden Retrievers, 2 American Staffordshire Terriers, 2 Labrador Retrievers, 2 Boxers, 1 Bernese Mountain Dog, and 1 mixed-breed (Labrador Retriever X White Shepherd) dog. Affected dogs consisted of 10 bitches (2 of which were spayed) and 3 males. The left and right stifle joints were affected almost equally. Six of 13 affected dogs sustained a CrCL rupture in the contralateral stifle joint during the 12 to 18 months of the study.

At initiation of the study, 8 dogs had partial CrCL rupture, 3 of which also had a medial meniscus tear (Table 1). The remaining 5 dogs had complete CrCL rupture, and 4 of them also had a concomitant medial meniscus tear. The number of dogs with medial meniscus damage did not differ significantly (P = 0.24) between those with partial and complete CrCL rupture.

Mean age of affected dogs at initiation of the study was 4.4 years (range, 1 to 8.2 years). Mean body weight of affected dogs was 33 kg (range, 23 to 58 kg). Dogs with complete CrCL rupture typically were younger than dogs with partial rupture, although the groups did not differ significantly (P = 0.42). No significant difference was found in mean body weight (P = 0.09), degree of osteoarthritis (P = 0.051), or synovitis (P = 0.66) between dogs with complete or partial CrCL rupture. Mean duration of lameness typically was longer in dogs with partial CrCL rupture (3.9 months; range, 1 to 12 months), compared with the duration for dogs with complete CrCL rupture (2.4 months; range, 1 to 5 months); however, these values did not differ signifi-cantly (P = 0.42).

The 2 sham-operated dogs were both Foxhounds. Mean age was 2.9 years (2.2 and 3.5 years old, respectively). Mean body weight was 23.3 kg (17.5 and 29 kg, respectively). One was a neutered female, and the other was a neutered male.

Anticollagen antibody titers—Cutoff values (mean OD + 2 SD) calculated for the left shoulder joint of the sham-operated dogs for the 4 assessment time points (day 0 and 6, 12, and 18 months after surgery) were 0.111, 0.182, 0.185, and 0.103, respectively. Cutoff values for the sham-operated stifle joint of the control dogs for the 4 assessment time points were 0.125, 0.131, 0.136, and 0.112, respectively. Cutoff values for the contralateral stifle joint of the control dogs for the 4 assessment time points were 0.094, 0.116, 0.119, and 0.085, respectively.

Comparisons were made to investigate whether there was a significant difference in antibody titer between dogs that sustained a CrCL rupture of the contralateral stifle joint during the study and dogs that did not. Twelve months after initiation of the study, the antibody titer for the contralateral stifle joint of dogs that sustained a subsequent CrCL rupture typically was higher, but not significantly so (P = 0.073), than for dogs in which the contralateral CrCL remained intact. The titer at this assessment time point differed signifi-cantly (P = 0.016) between the left shoulder joint and the contralateral stifle joint in dogs that ruptured the contralateral CrCL. However, the titer did not differ significantly between the left shoulder joint and the stifle joint with the initially ruptured CrCL (P = 0.413) or between both stifle joints (P = 0.556).

The anticollagen type I antibody titers of all 3 joints of each dog were summarized (Table 2). Most dogs had antibodies against collagen type I in the SF of all 3 joints. Three patterns could be identified. High antibody titers were detected for all 3 joints in 3 dogs (dogs 9, 11, and 13). A higher antibody titer was detected for 1 or both stifle joints than for the shoulder joint of 7 dogs (dogs 1, 2, 3, 4, 8, 10, and 12). Extremely low antibody titers or no detection of antibodies was evident for all 3 joints for 3 dogs (dogs 5, 6, and 7).

Eight of 13 dogs had partial CrCL rupture of the initially affected stifle joint, and 7 of these 8 dogs had anticollagen type I antibody titers for both stifle joints or for all 3 joints. The anticollagen type I antibody titers of these 7 dogs were high at the time of surgical intervention, after which 3 patterns of antibody response were evident for the surgically corrected stifle joints (Table 2). The antibody titer decreased in 2 dogs over time (dogs 3 and 12), remained constant in 3 dogs (dogs 8, 9, and 11), and increased in 2 dogs (dogs 1 and 2).

Five of 13 dogs had complete CrCL rupture of the initially affected stifle joint. Antibody titers in 4 of these 5 dogs with complete CrCL rupture were low to moderate for the affected stifle joint at the time of surgical correction but became undetectable during the 12 to 18 months of the study, whereas 1 dog had extremely high antibody titers initially that remained high throughout the study. The antibody titers for the contralateral stifle joint and left shoulder joint of these 5 dogs were similar to the titers for the stifle joint with the ruptured CrCL. Consequently, the mean antibody titers for dogs with a partially ruptured CrCL typically were higher, but not significantly so, than titers for dogs with a complete CrCL rupture at day 0 (P = 0.380) and 6 (P = 0.079), 12 (P = 0.186), and 18 (P = 0.480) months after surgery.

Six of 13 dogs sustained a CrCL rupture of the contralateral stifle joint during the study. One of the 6 dogs (dog 11) had similar antibody titers for all 3 joints, whereas the remaining 5 dogs (dogs 1, 2, 4, 8, and 10) had higher antibody titers for 1 or both stifle joints than for the left shoulder joint. In 2 dogs (dogs 2 and 8), the antibody titer increased in SF of the contralateral stifle joint before clinical signs of the subsequent rupture. In 2 other dogs (dogs 1 and 11), antibody titers in SF of the contralateral stifle joint were extremely high before the subsequent rupture and remained relatively high. In the final 2 dogs (dogs 4 and 10), the antibody titers were low or undetectable for the contralateral stifle joint before the subsequent rupture (Table 2).

We did not detect a correlation between the severity of osteoarthritis determined by radiographic evaluation and the anticollagen type I antibody titer for the stifle joint with the initially ruptured CrCL at any assessment time point (day 0, r = −0.291; 6 months, r = 0.128; 12 months, r = 0.386; and 18 months, r = −0.456). No correlation was found between the severity of osteoarthritis and the antibody titer for the contralateral stifle joint on day 0 (r = 0.322) or at 6 (r = 0.265) or 12 (r = 0.354) months after surgery; however, there was a positive correlation (r = 0.820) between severity of osteoarthritis and the antibody titer for the contralateral stifle joint at 18 months after surgery. Correlative analysis was not conducted for the shoulder joints because standard radiographs were not taken at every examination.

Discussion

To our knowledge, this is the first prospective study in which anticollagen type I antibodies have been measured sequentially over a period of 12 to 18 months in the SF of dogs with unilateral CrCL rupture. Anticollagen type I antibodies were investigated over time after surgery in the affected stifle joints and contralateral stifle joints to study the progression of antibodies before and after rupture of the CrCL. A remote joint (left shoulder joint was chosen arbitrarily) was also investigated to determine whether antibodies against collagen are primed for joints with pathologic changes or whether they are the result of a generalized immune response. Measurement of anticollagen antibodies in the sera of the dogs was not performed because other studies^13,17 have revealed that there is no correlation between the anticollagen antibody titers in sera and SF, with higher titers consistently detected in the SF.

Dogs with high antibody titers against collagen type I for both stifle joints or all 3 joints had partial CrCL rupture of the initially affected stifle joint (except for 1 dog). There is a need for continued antigenic stimulation to maintain active production of gammaglobulin.³³ Partially injured CrCLs stretch and tear more gradually over time, which constantly releases small quantities of collagen type I into the intra-articular space. Prolonged liberation and persistence of antigen in the joint could result in chronic joint inflammation and high amounts of antibodies in these dogs. Anticollagen type I antibodies form immune complexes with collagen type I.^15,34–36 Therefore, it is also possible that the low or undetectable anticollagen type I antibody titers found in the SF of dogs with complete rupture were attributable to the formation of immune complexes.

Six of 13 dogs sustained CrCL rupture of the contralateral stifle joint during the 18 months of the study, which is in accordance with other studies^2,6,7 of dogs with naturally developing cruciate disease. In 2 of 6 dogs that sustained a CrCL rupture in the contralateral stifle joint, an increase in anticollagen type I antibody titer was detected in SF obtained from this joint before rupture, which would suggest active antibody production before clinical signs of cruciate disease became apparent. Indeed, radiographic signs of osteoarthritis also increased in these joints before CrCL rupture, which indicated a chronic pathologic process in the joint.

One of the 6 dogs had extremely high antibody titers for the contralateral stifle joint that remained high throughout the study. The contralateral stifle joint of that dog had mild radiographic signs of osteoarthritis at the time of surgery to repair the initial CrCL rupture; those radiographic signs were suggestive of pathologic changes in the joint. In that dog, clinical signs of CrCL rupture were evident 12 months after surgical repair of the initially affected stifle joint.

Antibodies against collagen type I are not likely to be the sole causative factor for sustaining a CrCL rupture in dogs. Two other dogs with extremely high antibody titers in SF obtained from all 3 joints and at all assessment time points did not sustain a CrCL rupture in the contralateral stifle joint, nor did they develop osteoarthritis during the study. It is possible that our study period was too short and that the dogs would have sustained a CrCL rupture of the contralateral stifle joint in the future. Another explanation could be that functionality of the anticollagen antibodies in those 2 dogs differed from that of the dogs that sustained a CrCL rupture in the contralateral stifle joint. For example, antibodies from dogs that did not sustain CrCL rupture of the contralateral stifle joint may not have been able to maintain or perpetuate an inflammatory reaction, whereas antibodies from dogs that did sustain CrCL rupture of the contralateral stifle joint are able to participate in the inflammatory cascade.

When analyzing anticollagen type I antibody titers for all 3 joints, higher titers were detected for 1 or both stifle joints than for the left shoulder joint of 5 of 6 dogs that sustained a CrCL rupture of the contralateral stifle joint, with a distinct gradient toward the stifle joints. The remaining dog had high antibody titers for all 3 joints. A significant difference in anticollagen type I antibody titer was found between the left shoulder joint and contralateral stifle joint 12 months after initial surgery in dogs that sustained a CrCL rupture of the contralateral stifle joint. This could suggest that there is a local inflammatory process in these stifle joints that results in accumulation of collagen type I activated B lymphocytes and production of anticollagen type I antibodies. It is possible that these antibodies could maintain inflammation in the stifle joint and initiate CrCL degradation in these dogs.

We did not detect a correlation between anticollagen type I antibodies and the severity of osteoarthritis in the surgically repaired stifle joints at any examination. For the contralateral stifle joints, a positive correlation was only found 18 months after surgery on the initially affected stifle joint. Radiographic evaluation is not an accurate way to determine the severity of osteoarthritis because this disease is characterized by new bone formation as well as by synovial inflammation.^37,m To assess the true degree of osteoarthritis, histologic evaluation of the synovial membrane should be performed to determine correlations between the severity of osteoarthritis and anticollagen type I antibody titers. In another study,¹⁷ investigators measured anticollagen type I and II antibodies in the SF and determined the histologic grade of synovitis in dogs with naturally occurring CrCL rupture, and no correlation was detected. However, samples of synovial membrane were acquired from the lateral side of the joint, which has less synovial inflammation than is evident in the synovial membrane on the medial side.²⁹ Therefore, multiple samples from several sites of the same stifle joint should be acquired and examined histologically in future studies.

Cruciate disease is becoming more prevalent in largebreed dogs at younger ages, even before the dogs reach maturity.^6,38 Radiographic changes in the stifle joint of these dogs include osteophytes that are more advanced than can be related to the duration of lameness, suggesting an initial nonclinical pathologic change in the ligament that initiates a secondary osteoarthritic process.^3,6,38 Five dogs in the study reported here were between 1 and 3 years of age, and all dogs were large-breed dogs, which provides evidence that the disease is prevalent at an extremely young age. However, support for an active role of anticollagen type I antibodies in these young dogs could not be achieved from our results because 1 of the 3 young dogs that sustained a CrCL rupture in the contralateral stifle joint had low anticollagen antibody titers until the time of the subsequent rupture.

Although the number of dogs in the study reported here is relatively small, the population is representative for dogs with naturally developing CrCL disease because none of the dogs had a history of vigorous trauma and nearly half the dogs sustained a rupture of the contralateral CrCL during the study. However, the patient group was not homogeneous. Dogs represented various breeds, ages, and sexes. The etiopathogenesis for naturally developing CrCL rupture most likely differed among these dogs, and therefore it is not surprising that the acquired results varied considerably.

Finding a distinct anticollagen type I antibody gradient toward the stifle joint for 5 of 6 dogs that sustained a CrCL rupture of the contralateral stifle joint suggests that there is an inflammatory process in these joints before clinical signs of rupture. However, because the cutoff values for determining the anticollagen type I antibody titers were calculated on the basis of samples obtained from only 2 sham-operated dogs and because the mean and SD cutoff values differed widely, additional studies with more affected dogs and with a larger number of sham-operated dogs are needed to confirm this hypothesis. The study reported here did not provide evidence that anticollagen antibodies alone initiated CrCL damage in dogs because not all dogs with high antibody titers subsequently ruptured the contralateral CrCL.