Collagen-induced arthritis in common marmosets: a new nonhuman primate model for chronic arthritis

Healthy male and female marmosets of adult age ( > 1.5 years) were supplied by the Institute's Animal Science Department. Individual data of the animals are presented in Table 1. Before inclusion into an experiment, each monkey received a complete health check by the veterinary staff, including assessment hematological, serological and microbiological abnormalities. During the experiments the monkeys were housed in pairs in spacious cages with cage enrichment and padded shelter provided. The daily diet during the study consisted of commercial food pellets for New World monkeys (Special Diet Services, Essex, UK) supplemented with rice, raisins, peanuts, marshmallows, biscuits, fresh fruit, grasshoppers, and maggots. Drinking water ad libitum, slurry enriched with vitamin D, and vegetables of the season were provided.

In accordance with the Dutch law on animal experimentation, all study protocols and experimental procedures were reviewed and approved by the Institute's Ethics Committee before the experiments started.

For the pilot study (Experiment I) a semi-purified batch collagen type II (CII) was prepared, extracted from bovine hyaline cartilage. For Experiments II and III, two industry-grade commercially available types of CII were used, respectively of bovine (bo-CII) or chicken (ch-CII) origin (MD Biosciences, Zürich, Switzerland). Collagen was dissolved in 0.1 M acetic acid to a final concentration of 5 mg/ml and mixed with an equal volume of complete Freund's adjuvant (CFA) (DIFCO, Detroit, MI, USA). A stable emulsion was prepared by gentle stirring of the protein/CFA emulsion for 60 minutes at room temperature. CIA was induced by injection of 0.4 ml emulsion into the dorsal skin distributed over four spots of 100 μl. The final amount of CII the animals received was 1 mg/animal. When no clinical signs of arthritis were visible at day 28, the animals were boosted with CII in incomplete Freund's adjuvant (IFA) at day 35 (Experiment II)/day 28 (Experiment III) subcutaneously in the flank.

The presence of clinical signs was recorded by daily cage-side monitoring of behavioral changes (apathy, loss of appetite) or pain (avoidance of limb usage). Twice per week all monkeys were sedated by intramuscular injection of 0.1 ml/kg ketamine (10 mg/ml) for determination of bodyweight - an accepted surrogate disease marker for the CIA model - determination of body temperature, blood collection and inspection of the limbs for redness and/or swelling of the joints. Observations were recorded using the discomfort management schedule (Table 2).

All hematological and clinical chemistry analyses were performed at the Biomedical Primate Research Centre on a Sysmex Sf-3000 (Goffin Meyvis, Etten-Leur, The Netherlands) and a COBAS INTEGRA-400+ (Roche, Almere, The Netherlands), respectively.

Urinary excretion of the collagen crosslinks hydroxylysylpyrridinoline (HP) and lysylpyrridinoline (LP) were determined twice weekly, starting from the day of CIA induction. For that purpose, each animal's urine was collected overnight in a metabolic cage. After centrifugation, the clear supernatant was isolated and stored at -20°C. Reverse-phase high-performance liquid chromatography was used to determine HP and LP levels in hydrolyzed urine samples as described previously [14]. The levels of HP and LP were normalized to creatinine levels (nmol levels/mmol creatinine) to compensate for a possible dilution by spilled drinking water.

Three independent experiments have been performed, involving a total of 22 marmosets. In the first experiment, four marmosets were immunized according to the standard protocol that was developed for CIA induction in genetically susceptible rhesus monkeys; that is, a single immunization with semi-purified bo-CII in CFA [14, 16]. All four monkeys displayed signs of CIA; that is, painful joints, expressed by relief of pressure from the hind legs and reluctance to manually take an offered reward (marshmallow), and a marked loss of body weight (Figure 1a). Sustained clinical arthritis, however, was observed in only two monkeys (Mi121 with 10 affected joints at day 29, and Mi125 with >20 affected joints at necropsy; Figure 1a). The two remaining animals displayed minimal clinical signs of arthritis transiently and received a subcutaneous booster immunization with bo-CII in IFA at day 114. This did not aggravate joint swelling but did boost the induction CII-specific IgM and IgG in Mi137. Figure 1b shows the prominent joint swelling in the extremities of the most severe arthritic case, Mi125. The X-ray scans depicted in the same figure, however, do not show prominent bone deformations in the affected hands. Histologically, varying degrees of synovitis were observed in the joints of two out of the four monkeys (except Mi121 - which was found dead in the cage on day 30 - and Mi137). Mi125 showed prominent signs of synovitis (Figure 1c). Furthermore, histologically Mi119 displayed signs of synovitis in clinically nonaffected joints. Erosion of the cartilage surface was observed to start to develop in both Mi125 and Mi119. It was concluded from this experiment that the CIA induction protocol developed for the rhesus monkey was suboptimal for the marmoset, probably because marmosets were less susceptible to the disease.

In a second experiment, industry-grade commercial batches of CII preparations of bovine and chicken origin were tested. The immunization protocol comprised inoculation with CII in CFA (intracutaneously) followed by booster immunization with CII in IFA (subcutaneously). From each immunization group of six animals, three animals were sacrificed at the peak of the disease for histopathological examination just after the onset of clinical disease. The remaining three animals per immunization group were allowed to develop full clinical arthritis. The animals were selected alternately to be sacrificed early or late in the disease course. Clinical signs of CIA were observed in all 12 cases (Integrated Discomfort Score ≥3), but no major differences between both collagen batches were observed (Figure 2a). A relapsing/remitting disease course with only moderate weight loss was observed in several cases, whereas in other monkeys a rather acute progressive course with serious weight loss was observed. This same clinical heterogeneity is observed in marmoset models of EAE [17]. This is probably a reflection of the genetically determined responsiveness to the collagen immunization, as can be seen in the twin pairs depicted in the top of Figure 2a.

To establish reproducibility of the induction protocol, ch-CII was selected for the third experiment (n = 6) - the most commonly used autoantigen for CIA induction in rodents. The images in Figure 2b confirm the high incidence of CIA in this species (five out of six animals), as well as the variable disease course. The one case in which clinical signs were not observed during the 100-day observation period did display weight loss, suggestive of subclinical disease.

In conclusion, sensitization of marmosets against CII by a first immunization with collagen in CFA followed by booster immunizations with collagen in IFA reproducibly induces clinical arthritis in marmoset monkeys.

The alternation of overt arthritis with episodes of complete remission that was observed in several cases could also be observed at the level of a single joint. As an example, data from M05003, which was immunized with bo-CII in Experiment II, are shown in Figure 3. Swollen joints at a given time point are indicated with red circles, while the arrows indicate joints that were found only transiently affected. The pictures at the bottom of Figure 3 show the severe swelling of interphalangeal joints and the severe deformation of the extremities at the end stage of the observation period.

Histological features of a representative animal with early onset arthritis (M03136) induced with ch-CII are depicted in Figure 4. A cross-section of a severely swollen joint (Figure 4a) was stained with H & E (Figure 4b). Figure 4c shows marked extraarticular inflammation (synovium and fibrous capsule) and loss of chondrocytes and collagen disruption. Figure 4d zooms in on an area where marked bone resorption takes place. Figure 4e shows the severe extraarticular inflammation of the periosteum and subcutis. Interestingly animals that showed marked inflammation at either the periosteum or the subcutis were mainly found in the early responders to induction (6/12 animals), which were also associated with a marked loss of body weight. Inversely, animals that showed a more chronic development of the disease (5/12 animals) displayed no such extraarticular phenomena.

The autoimmune attack on the joints in CIA models in mice, rats and rhesus monkeys is mediated by interplay of anti-CII T cells and humoral factors. In both rodent CIA models [18] as well as the CIA model in the rhesus monkey [16], a key pathogenic role of IgM autoantibodies was found. Rhesus monkeys that were genetically resistant to CIA were found incapable of producing adequate amounts of anti-collagen IgM [19]. Second, pre-sensitization of rhesus monkeys with heat-denatured CII not only abrogated the capacity to produce anti-CII IgM, but also induced resistance to CIA [20]. The relation between both autoimmune parameters and disease was therefore also tested in the novel CIA model in marmosets.