KIMME L. HYRICH, TAE-HWAN KIM, URSULA PAYNE, XIANG ZHANG, BASIL CHIU, KAY MEHREN,
and ROBERT D. INMAN

ABSTRACT.

A 27-year-old female lowland gorilla developed an asymmetric oligoarthritis 3 months post-partum. There was no evidence of an antecedent gastrointestinal or genitourinary infection. Serum was negative for rheumatoid factor and antinuclear antibody. Synovial fluid revealed 2000 white blood cells with negative cultures and polarized microscopy. Studies on synoviocytes were the following: (1) FACS analysis revealed surface expression of a B27-like epitope of the cells. (2) Analysis of intracellular clearance kinetics of arthritogenic organisms showed peak intracellular colony-forming units at 48 hours after bacterial invasion, and clearance by 13 days post-invasion. (3) Interferon-g (0.1–10.0 ng/ml) accelerated intracellular microbicidal pathways in a dose-dependent fashion. These findings closely parallel those seen in human synoviocytes of patients with spondyloarthropathy. Primate and human seronegative arthritis share clinical and immunologic features, as well as aspects of host:pathogen defense mechanisms. The interplay of genetic and microbial factors underlying this arthritis appears to be conserved across these species boundaries. (J Rheumatol 2004;31:2300-4)

Key Indexing Terms:

ARTHRITIS
GORILLA
BACTERIA
PRIMATE

K.L. Hylich, MD, Clinical Fellow; T-H. Kim, MD, PhD, Research Fellow, Toronto Western Hospital, Hospital for Rheumatic Diseases, Hanyang University; U. Payne, BSc, Research Associate; X. Zhang, MD, PhD, Research Fellow; B. Chiu, PhD, Research Associate, Toronto Western Hospital; K. Mehren, BVSc, DVM, Staff, Toronto Zoo; R.D. Inman, MD, Professor of Medicine and Immunology, University of Toronto.

Address reprint requests to Dr. R.D. Inman, Arthritis Centre of Excellence, Toronto Western Hospital, 399 Bathurst Street, Toronto, Ontario M5T 2S8. E-mail: robert.inman@uhn.on.ca

Submitted December 22, 2003; revision accepted May 31, 2004.

While murine models of rheumatic diseases have been instructive in some instances, such as in systemic lupus erythematosus, spontaneous arthritis in the animal kingdom has received less attention in the rheumatology literature. We describe our recent experience with a gorilla mimicking human spondyloarthropathy (SpA).

Reactive arthritis (ReA), a subtype of SpA, refers to a nonseptic inflammatory joint disease accompanying an infection¹. The pathogenesis of ReA is unresolved, but studies of synovial tissues in ReA following Yersinia, Salmonella, and Chlamydia all point to local persistence of microbial antigens in the joint as a key component in maintaining the chronic synovitis. Since there is evidence of persistence of such antigens for prolonged periods after the initial infection^2,3, the concept has arisen that defective clearance of the arthritogenic pathogen may play an important role in the pathogenesis of ReA. There has been some evidence that HLA-B27 may modulate intracellular clearance of Salmonella⁴, but few studies have addressed the question using primary synoviocytes as host target cells. It has also been proposed that impairment in nitric oxide production may account for the altered kinetics of intracellular clearance⁵, although the HLA allele specificity of this event remains unresolved⁶. We utilized the opportunity to study host:pathogen interactions using the synoviocytes from our case of primate seronegative arthritis. Although we had no access to control synoviocytes to address the specificity of any MHC allele affect, we were able to observe that the kinetics of intracellular clearance of an arthritogenic pathogen paralleled exactly that seen in the human counterpart⁷.

CASE REPORT

The Toronto Zoo in Toronto, Ontario, recently requested a consultation from our rheumatology service. Josephine, a 27-year-old female lowland gorilla (Gorilla gorilla) had recently developed arthritis. She presented with a one-week history of anorexia and lameness, particularly favoring her left hand and foot. She had had no previous medical problems and there was no observed trauma. There was no history of diarrhea or urinary tract infection. She was 3 months post-partum from a stillbirth, thought secondary to placental insufficiency. No cohabitating gorilla in the zoo had past or current symptoms to suggest a rheumatological problem.

Physical examination under general anesthesia revealed normal vital signs. Her weight was 100 kg. General physical examination by the veterinary staff was normal. There were no skin lesions. Joint examination revealed that the second and third metacarpophalangeal (MTP) joints on the left hand were warm and effused. Range of motion was preserved. The right hand was normal. There was swelling of the left first MTP joint with an extensor deformity of the first interphalangeal joint on the same toe. No other musculoskeletal abnormalities were noted.

Her white blood cell count was 6.1 ´ 10⁹/l, hemoglobin 116 g/l, and platelets 265 ´ 10⁹/l. Serum creatinine was 85 µmol/l and uric acid 160 µmol/l. Serum was negative for rheumatoid factor and antinuclear antibody. Synovial fluid from an aspirate of the second left metacarpophalangeal joint revealed 2000 leukocytes, with negative culture and crystal analysis by polarized microscopy. Radiographs of the hand and feet showed no evidence of erosive disease. Radiographs of the sacroiliac joints were not obtained.

Josephine was treated with naproxen sodium 375 mg bid for a presumed seronegative arthritis. She responded well with resolution of her effusions and resumption of normal activity. The drug was stopped after 2 months, with no recurrence of her arthritis at 6 months' followup.

Synoviocyte harvesting and FACS analysis. Synovial fluid was obtained using sterile technique from Josephine's affected joints. The fluid was added to a-modified Eagle's medium with 20% fetal calf serum (FCS). A primate adherent fibroblast cell line was subsequently obtained from cells that grew out of the fluid and maintained by splitting once weekly. Control synoviocytes were obtained from a B27/human ß₂-microglobulin double transgenic rat derived from a breeding pair generously provided by Dr. J. Taurog (Dallas, Texas, USA) and normal Lewis rats. They were grown out from synovial tissue explants as described⁸. These rodent fibroblasts were similar in appearance to Josephine's fibroblasts. FITC-conjugated mouse anti-human HLA-B27 antibody (clone m3) was obtained from Serotec (Oxford, England). A monolayer of the primate synoviocytes was removed from the flask with cell dissociation buffer (Gibco). Cell numbers were adjusted to 10⁶/ml, and 1.0 ml portions of cells were stained with the monoclonal antibody (1:5) at 4°C for 1 h. The stained cells were analyzed in a Becton-Dickinson FACScan.

Clearance of arthritogenic bacteria. Quantitative invasion assays were performed as described⁹. Briefly, bacteria were grown to mid-log phase, washed in phosphate buffered saline (PBS), and resuspended in Dulbecco's modified Eagle's medium (Gibco) with 5% FCS. Then 0.1 ml aliquots were added to the PBS-washed cell monolayer containing 0.3 ml media. To make immediate contact between cells and bacteria, the plate was spun for 10 min at 1800 rpm. The bacteria were incubated with the synoviocytes for 2 h at 37°C, at a bacteria:cell ratio of 2:1 to 10:1. After washing, gentamicin 100 µg/ml was added for 1 h to kill the extracellular bacteria. Cells were harvested at different time points and colony-forming units (CFU) were determined after cell lysis (with 0.2% Triton X-100) to calculate the total intracellular bacterial load. The organisms used were arthritogenic strains of Salmonella typhimurium and Yersinia enterocolitica 0:3 recovered, respectively, from patients with ReA during an epidemic of Salmonellosis and from a patient with post-Yersinia ReA. Interferon-g (IFN-g; Endogen, Cambridge, MA, USA) was used to observe its effect on clearance of bacteria. For comparison, synoviocytes harvested from a B27+ patient with SpA were studied in a similar manner.

Measurement of nitric oxide (NO) production. NO production of the synoviocytes was measured as nitrate in the supernatants of the infected cells in the presence or absence of IFN-g. The accumulation of nitrate was determined with the diazotization reaction with the Griess reagent (0.1% naphtylenediamine dihydrochloride and 1% sulfanilamide in 5% phosphoric acid). Cell-free supernatants of the infected cells with or without IFN-g were collected at days 0, 1, 2, 3, 7, 10, and 13. The samples (100 µl) were mixed with equal volumes of Griess reagent and incubated at room temperature for 10–30 min. Absorbance was measured at 570 nm and nitrite concentrations were interpolated from standard curves prepared with NaNO₂ ^5-7.

Results. To address surface expression of HLA-B27 or cross-reactive determinants, the synoviocytes were stained with a monoclonal HLA-B27 antibody. Synoviocytes were stained at a level intermediate between the positive control (synoviocytes from a B27-transgenic rat) and the negative control (synoviocytes from a matched nontransgenic Lewis rat; Figure 1).

To analyze bacterial clearance by synoviocytes, the cells from Josephine were incubated with 2 arthritogenic pathogens, S. typhimurium and Y. enterocolitica, at different timepoints (Figure 2A). For both pathogens, intracellular CFU rose over day 1 and 2 and declined steadily to day 13. This is comparable to an analysis of clearance kinetics using human HLA-B27+ synoviocytes derived from a patient with ReA (Figure 2B). Clearance kinetics of the pathogens of both primate and human target cells reflected comparable time courses⁷.

Because of the implicated microbicidal effects of IFN-g and NO, we addressed the production of NO by the primate synoviocytes cocultivated with S. typhimurium for 7 and 10 days, respectively. Figure 3A indicates that at these 2 timepoints, increasing amounts of IFN-g had no consistent effect on NO production. This is in contrast to the dose-dependent microbicidal activity of IFN-g represented in Figures 3B and 3C. IFN-g thus appears to accelerate clearance of Salmonella, but not via NO pathways.

DISCUSSION

There is no perfect animal model for seronegative or seropositive arthritis. Case reports suggest primates are afflicted with several joint diseases comparable to humans. There are reports of chronic osteoarthritis in lowland gorillas¹⁰. Rheumatoid-like disease has been described in rhesus monkeys (Macaca mulatta) and a lion-tailed macaque (Macaca silenus)^11-13. There has also been evidence of calcium pyrophosphate dihydrate-like disease in these same monkeys and Barbary ape¹⁴.

In our case there were features mimicking human SpA, including a seronegative asymmetric oligoarthritis occurring after a potential infectious event (stillbirth). In this case, however, it may be more accurate to define this simply as seronegative arthritis since there was no documented antecedent infection or skin or gastrointestinal disease. Radiographs of the sacroiliac joints was not obtained. Most research in the area of primate arthropathies has focused on the occurrence of SpA-like disease. SpA has been documented as a pan-mammalian phenomenon, with high penetrance in many groups. As many as 25% of bears¹⁵ and 35% of rhinoceroses¹⁶ have been reported to manifest this clinical problem. The diagnosis of SpA was clearly identified as a population phenomenon in gorillas in 1989¹⁷. While it is quite reasonable to consider the possibility of selective disadvantage in the wild (for lame or slower-moving animals), that question likely pertains to zoo specimens as well. The frequency of 20% in skeletal samples of lowland gorillas is indistinguishable from that subsequently identified in zoologic parks^18,19. This contrasts with a SpA prevalence of 1% to 2% in humans. This study of gorillas described an asymmetrical subchondral erosive disease with perilesional, smooth, billowy new bone formation in the peripheral joints, and erosions, ankylosis, and syndesmophytes in the sacroiliac joints and spine. These findings are in keeping with the pattern of ankylosing spondylitis (AS) in humans. Resistance to nonsteroidal antiinflammatory drug (NSAID) therapy has proved to be common in gorilla SpA and sulfasalazine (SSZ) is now considered standard treatment for this disease¹⁹. In our case, Josephine responded well to NSAID, and SSZ therapy was not instituted.

There has recently been interest in the MHC classifications among primates. Studies of MHC Class I genes have shown that homologs for HLA-A, B, E, and F are preserved in primates. Chimpanzees have shown preserved B locus alleles homologous to HLA-B15, HLA-B48, HLA-B57/58, and HLA B27/7²⁰. Similar homology has been reported between the rhesus macaque Mamu-B*03 molecule and human HLA-B27^21,22. A Gogo-B (Gorilla gorilla) allele (B*01) with limited nucleotide sequence similarity to HLA-B27 was reported in association with gorilla SpA^19,23. Subsequently, a National Zoo Committee project documented that a B27 epitope is extremely common in gorillas. In our study, the use of antisera for detection of HLA-B27 is complex, with issues of specificity still unresolved, especially in the gorilla. Using FACS analysis of the gorilla synovial fibroblasts we found evidence of an HLA-B27-like epitope on the surface of the cells.

The significance of this homology and association with clinical disease, however, remains unresolved in our case. Of the 3 gorillas described with the HLA-B27-like determinants, one developed a ReA after infection with Shigella flexneri²³ and one demonstrated clinical signs indistinguishable from AS²⁴. An unrelated gorilla in the same zoo as one of the animals in this report also tested HLA-B27 positive by standard human leukocyte tissue-typing methods and displayed no clinical disease. A recent report analyzed the MHC class I cDNA expressed in a cohort of rhesus macaques who developed ReA after an outbreak of shigellosis²². The investigators identified molecules that could bind peptides similar to HLA-B39. However, they could not find any statistically significant relationship between any particular MHC Class I molecule and development of ReA. They did not identify a molecule similar to HLA-B27, suggesting perhaps that ReA could occur in their animal model in the absence of HLA-B27.

It would appear that the clinical and biological features in our primate recapitulate the human counterpart of seronegative arthritis. The synovial fibroblasts expressed a structure on the surface similar to HLA-B27. We recently published our experience with bacterial clearance in synoviocytes from SpA⁷. Challenge of the gorilla synoviocytes with intracellular pathogens showed clearance kinetics similar to human SpA synoviocytes⁷. NO is thought to be important in the primary response against arthritogenic bacteria, and IFN-g has been known to be an efficient inducer of NO synthase in macrophages²⁵. Our recent analysis on human synoviocytes confirmed an interaction of NO and IFN-g in the clearance of intracellular pathogens, although the clearance profile was not clearly associated with B27 expression.

Our case illustrates that the cell biology of SpA may be common in diverse primate species, implying that the genetic susceptibility to this disease may be conserved across the species.

REFERENCES

Search PubMed for:

2. Inman RD, Whittum-Hudson JA, Schumacher HR, Hudson AP. Chlamydia and associated arthritis. Curr Opin Rheumatol 2000;12:254-62. [MEDLINE]

3. Granfors K, Jalkanen S, von Essen R, et al. Yersinia antigens in synovial fluid cells from patients with reactive arthritis. N Engl J Med 1989;320:216-21. [MEDLINE]

4. Laitio P, Virtala M, Salmi M, Pelliniemi LJ, Yu DT, Granfors K. HLA-B27 modulates intracellular survival of Salmonella enteritidis on human monocytic cells. Eur J Immunol 1997;27:1331-8. [MEDLINE]

5. Virtala M, Kirveskari J, Granfors K. HLA-B27 modulates the survival of Salmonella enteritidis in transfected L cells, possibly by impaired nitric oxide production. Infect Immun 1997;65:4236-42. [MEDLINE]

6. Ekman P, Saarinen M, He Q, Virtala M, Salmi M, Granfors K. Human monocytic U937 cells kill Salmonella in vitro by NO-independent mechanisms. Infect Immun 1999;67:3670-3. [MEDLINE]

7. Inman RD, Payne U. Determinants of synoviocyte clearance of arthritogenic bacteria. J Rheumatol 2003;30:1291-7. [MEDLINE]

8. Inman RD, Chiu B. Synoviocyte-packaged Chlamydia trachomatis induces a chronic aseptic arthritis. J Clin Invest 1998;102:1776-82. [MEDLINE]

9. Kapasi K, Inman RD. The ME1 epitope of HLA-B27 confers Class I-mediated modulation of gram-negative bacterial invasion. J Immunol 1994;153:833-40. [MEDLINE]

10. Atkinson JP. A remembrance of Fred, the lowland gorilla. Arthritis Rheum 1996;39:891-3. [MEDLINE]

11. Lipman NS, Schelling SH, Otto G, Murphy JC. Juvenile rheumatoid arthritis in a rhesus monkey (Macaca mulatta). J Med Primatol 1991;20:82-8. [MEDLINE]

12. Bywaters EG. Observations on chronic polyarthritis in monkeys. J Roy Soc Med 1981;74:794-9. [MEDLINE]

13. Anderson ST, Schiller CA. Rheumatoid-like arthritis in a lion tailed macaque. J Rheumatol 1991;18:1247-50. [MEDLINE]

14. Kandel RA, Renlund RC, Cheng PT, Rapley WA, Mehren KG, Pritzker KP. Calcium pyrophosphate dihydrate crystal deposition disease with concurrent vertebral hyperostosis in a Barbary ape. Arthritis Rheum 1983;26:682-7. [MEDLINE]

15. Rothschild BM, Wang X-M, Cifelli R. Spondyloarthropathy in ursidae: A sexually transmitted disease? Natl Geographic Res Expl 1993;9:382-4.

16. Rothschild BM, Prothero DR, Rothschild C. Origins of spondyloarthropathy in Perissodactyla. Clin Exp Rheumatol 2001;19:628-32. [MEDLINE]

17. Rothschild BM, Woods RJ. Spondyloarthropathy in gorillas. Semin Arthritis Rheum 1989;18:267-76. [MEDLINE]

18. Rothschild BM, Martin LD. Paleopathology: Disease in the fossil record. London: CRC Press; 1993.

19. Neiffer DL, Rothschild BM, Marks SK, Urvater JA, Watkins DI. Management of reactive arthritis in a juvenile gorilla (Gorilla gorilla gorilla) with long-term sulfasalazine therapy. J Zoo Wildl Med 2000;31:539-51. [MEDLINE]

20. Watkins DI. The evolution of MHC Class I genes in primates. Crit Rev Immunol 1995;15:1-29. [MEDLINE]

21. Dzuris JL, Sidney J, Appella E, Chesnut RW, Watkins DI, Sette A. Conserved MHC Class I peptide binding motif between humans and rhesus macaques. J Immunol 2000;164:283-91. [MEDLINE]

22. Urvater JA, McAdam SN, Loehrke JH, et al. A high incidence of Shigella-induced arthritis in a primate species: major histocompatibility complex class I molecules associated with resistance and susceptibility, and their relationship to HLA-B27. Immunogenetics 2000;51:314-25. [MEDLINE]

23. Urvater JA, Hickman H, Dzuris JL, et al. Gorillas with spondyloarthropathies express an MHC class I molecule with only limited sequence similarity to HLA-B27 that binds peptides with arginine at P2. J Immunol 2001;166:3334-44. [MEDLINE]

24. Adams RF, Flinn GS, Douglas M. Ankylosing spondylitis in a nonhuman primate: a monkey tale. Arthritis Rheum 1987;30:956-70. [MEDLINE]

25. Xie QW, Cho HJ, Calaycay J, et al. Cloning and characterization of inducible nitric oxide synthase from mouse macrophages. Science 1992;256:225-8. [MEDLINE]