Search

The importance of HLA-B27 in the pathogenesis of ankylosing spondylitis is uncertain: current evidence favours a role for the B27 molecule itself. The possibility that quantitative differences in HLA-B27 expression may exist between patients with ankylosing spondylitis, family members, and control subjects positive for B27 was examined using appropriate monoclonal antibodies, flow cytometry, and a 'model lymphocyte' coated with a known number of mouse immunoglobulin binding sites. No differences were found between the groups. HLA-A2, examined for comparison, was expressed in greater amounts than HLA-B27, but each contributed only 10-20% of the total class I antigens. Homozygotes expressed twice the amount of antigen expressed by heterozygotes. Synovial lymphocytes expressed more class I antigens than peripheral lymphocytes.

Cytotoxic studies indicate cross-reactivity between some enteric organisms and cells obtained from the majority of patients with ankylosing spondylitis. Preliminary studies suggest that the factor responsible for cross-reactivity may be generated by a bacterial plasmid. However, the mechanism mediating the interaction between the HLA-B27 positive cell and the bacterial antigen is at present unknown.

In this paper, we report the presence on Epstein-Barr virus-transformed lymphoblastoid cell lines on platelets and on fibroblasts of an HLA-B27-associated cell surface complex (antigenically related to some antigens of Klebsiella K43 and K21) which is identical to or cross-reactive with the determinant present on the peripheral blood lymphocytes (PBL) of B27-positive patients with ankylosing spondylitis (AS). By contrast, no Klebsiella K43 markers could be demonstrated on the spermatozoa of B27+ AS+ individuals even though these cells expressed the HLA-B27 alloantigen. No B27-associated K43 antigen was detected on the erythrocytes of patients or of normal controls. The B27-associated membrane marker is still detectable on lymphoblastoid cell lines after 20 generations and on fibroblasts after about 10 generations. This finding implies that the continued expression of Klebsiella-modified B27 structure is generally determined and does not require the repeated exposure of the cell surface to Klebsiella antigen. These data suggest that certain non-lymphoid as well as lymphoid cells may be involved in the complex sequence of events leading to the clinical manifestation of AS.

Although sometimes used to refer to any sterile arthritis occurring in association with infection, the term 'reactive arthritis' is better reserved for arthritis following sexually acquired nonspecific urethritis or enteric infections with organisms such as Shigella, Salmonella, Yersinia and Campylobacter, because these arthropathies are unified by a number of shared clinical characteristics and an association with HLA B27. This review suggests that these arthropathies may also share a common pathogenic pathway, triggered by an ' arthritogenic factor' common to the diverse microbes which cause the disease and modified by genetic factors other than HLA B27. Although uncommon, reactive arthritis is important because it could provide the key to understanding the other seronegative arthropathies and mechanisms basic to chronic inflammatory synovitis.

Although the strong association of HLA-B27 with ankylosing spondylitis (AS) is well documented, the significance of this association is largely unknown. It has recently been reported that Klebsiella pneumoniae was present in the faeces of patients with AS more frequently than in healthy individuals, and that there appeared to be some cross-reactivity between HLA-B27 and Klebsiella. We have therefore attempted to gain an insight into the possible role of Klebsiella in the pathogenesis of AS. The results presented here indicate that HLA-B27 positive individuals with AS have a significantly lower in vitro lymphocyte responsiveness to Klebsiella antigens, as compared with B27 positive and B27 negative healthy controls. By contrast, B27 positive patients with AS as well as B27 positive and negative controls respond equally well to phytohaemagglutinin, Staphylococcus, Streptococcus, Yersinia and Shigella. To investigate a possible cross-reactivity between Klebsiella and HLA-B27, antisera were raised in rabbits against various isolates of Klebsiella. An antiserum to one isolate (427) of Klebsiella lysed the lymphocytes of B27 positive AS patients but not of B27 positive or B27 negative controls. These observations strongly suggest that some Klebsiella antigens cross-react with a gene product closely associated with HLA-B27 or possibly with a modified B27 antigen in patients with AS.

Culture filtrates of some Klebsiella isolates contain a factor(s) capable of specifically modifying the HLA-B27-positive lymphocytes of normal individuals, resulting in a phenotypic change similar to that seen on lymphocytes from patients with ankylosing spondylitis (AS). To further delineate the nature of the interaction between HLA-B27 and some Klebsiella products we have undertaken a chemical characterization of B27+AS+-cross-reactive Klebsiella antigens from the culture filtrate and the bacterial cell-membrane. Biogel P-100 chromatography of the Klebsiella K43-derived modifying factor from the culture filtrate and from an NP-40-solubilized membrane extract gave a molecular weight of 26,000-30,000. Isoelectric focusing revealed that the modifying factor had an isoelectric point of approximately 5.4. Membrane-associated modifying factor activity was found only in outer-membrane preparations indicating that the cross-reactivity between Klebsiella K43 and B27+AS+ cells is defined, at least in part, by outer-membrane antigens. These studies demonstrate that membrane components of Klebsiella K43 share antigenic determinants with a modifying factor, which is released into the culture medium, and that these components are capable of specifically altering the HLA-B27 antigen or an associated cell-surface structure. Such a modification occurring in vivo following exposure to Klebsiella, or to antigenically related organisms, could explain the triggering of the B27-associated arthropathies such as AS.

It has been suggested that some Klebsiella sp may cross-react with a cell surface determinant on the lymphocytes of B27+AS+ individuals. Studies were undertaken to identify culture filtrates capable of rendering the lymphocytes of B27-positive healthy controls susceptible to lysis by the anti-Klebsiella antiserum. Polyacrylamide gel electrophoresis of cell wall material of Klebsiella K43 prepared by sonication, high-speed centrifugation, and nonidet solubilization, demonstrated the presence of four major protein bands. When antisera raised in rabbits to each of these were tested for their cytotoxic effect on the lymphocytes of B27+AS+ individuals, an antiserum to one component only, of 40-52 K daltons molecular weight, reproduced the activity of the whole serum. Studies on the K43 filtrate indicated that the 'modifying' factor appeared to reside in a 25-50 K dalton component. Immunoelectrophoresis against anti-K43 serum demonstrated overlapping bands in the culture filtrate and the 40-52 K dalton cell wall fraction and these appeared to be identical on immunodiffusion. Antibody to the cell wall component removed both the 'modifying' activity and the appropriate protein band from the filtrate. The results suggest that a 40-52 K dalton component of the Klebsiella K43 cell wall is cross-reactive with a determinant on the lymphocytes of B27+AS+ individuals and is similar to or identical with a modifying factor in K43 culture filtrate which renders the lymphocytes of B27-positive healthy controls susceptible to lysis by anti-Klebsiella antiserum.

Cytotoxic T lymphocytes, induced by stimulating the PBMC of an HLA-B27+ normal individual (B27+, AS-) with the PBMC of an HLA-identical sibling suffering from ankylosing spondylitis (AS) (B27+, AS+), specifically lyse B27+, AS+ PBMC but not PBMC from HLA-27+ or B27-, AS- normal controls, or from HLA-B27- AS patients (B27-,AS+). CTL of similar specificity can also be raised by immunizing in vitro B27+,AS- cells with autologous cells modified by cross-reactive bacterial antigens. These results suggest that CTL can recognize certain bacterial antigens in association with HLA-B27 and that this interaction may lead to an inflammatory episode during the initial stages of the disease.

Cultured skin fibroblasts of HLA-B27-positive clinically normal individuals specifically bind, and are modified by, a factor in the culture filtrate of some Klebsiella isolates. These modified fibroblasts are serologically similar to the cells (lymphocytes and fibroblasts) of B27-positive patients with ankylosing spondylitis (B27+AS+). By contrast, B27+AS+ cells fail to bind the factor, presumably because a receptor present on B27+AS- cells has already been blocked or modified by a Klebsiella antigen in vivo.

We have previously shown that antibodies raised in rabbits to certain enteric bacteria will specifically lyse, in a 51Cr release assay, the peripheral blood lymphocytes (PBL) of 80% of HLA-B27 positive patients with ankylosing spondylitis (B27+ AS+) but not the PBL of HLA-B27 positive normal controls (B27+ AS-). Other laboratories have been unable to reproduce these findings. This study was designed to ascertain whether this lack of reproducibility was due to a peculiarity of our B27+ AS+ patients or to technical difficulties in the complement mediated 51Cr release assay. We have shown in this blind study that the PBL of 16 out of 18 B27+ AS+ patients from a New Zealand population were lysed by our antisera but none of the PBL of 20 B27+ AS- normal controls were lysed. The phenomenon of 'cross reactivity' between certain enteric bacteria and B27+ AS+ PBL is not confined to the Sydney AS population.

It has been shown that HLA-B27 lymphocytes from healthy individuals (B27+ ankylosing spondylitis [AS]-), which are not lysed by an antiserum against Klebsiella K43, can be rendered susceptible to lysis after incubation in the culture filtrate of Klebsiella K43. This finding is compatible with a specific modification by a Klebsiella K43-derived soluble factor of a B27-associated lymphoid cell component. Preliminary characterization of the factor has indicated that it is nondialyzable, but it is heat labile at 56 degrees C for 30 min and has a 35,000-50,000 mol wt. The modifying factor activity of the filtrate is destroyed by neuraminidase but not by trypsin and alpha-chymotrypsin. Furthermore, the ability of the factor to convert B27+AS- lymphocytes can be specifically absorbed by B27+AS- lymphocytes, but not by B27+AS+, B27-AS+, or by B27-AS- lymphocytes, which suggests that B27+AS- cells carry a hypothetical receptor which can specifically bind a Klebsiella K43 antigenic determinant. These results imply that the modification by environmental agents of specific major histocompatibility complex-associated gene products may be an important element in the pathogenesis of the HLA-B27-linked seronegative arthropathies.

Previous studies have shown that antisera raised in rabbits to certain enteric bacteria (cross-reactive bacteria) are capable of specifically lysing in a 51chromium-release lymphocytotoxicity test the lymphocytes of HLA-B27-positive (B27+) patients with ankylosing spondylitis (AS). This study investigated the clinical relevance of this finding by ascertaining whether Escherichia coli isolated from the rectal swabs of 20 B27+ AS patients (B27+ AS+) and 46 controls (35 B27- AS- and 11 B27+ AS-) were able to absorb the lymphocytotoxic activity of these antisera. All isolates from B27+ AS+ patients and one from a B27- AS- individual were capable of removing this activity. These organisms persisted in the bowel flora of five selected patients for at least 9 months. Cross-reactive bacteria were also found in a range of gram-positive organisms, including streptococcal, staphylococcal, and clostridial species. The relevance of these findings is discussed in terms of a pathogenetic concept relating the persistence of cross-reactive bacteria in the bowel flora of B27+ individuals to an early event in the development of AS.

A comparison was made between HLA-B27 positive patients with peripheral arthritis and patients with ankylosing spondylitis (AS). The distribution of peripheral joint involvement in HLA-B27 positive seronegative arthritis was not the same as that in AS; significant differences were also noted in the sex ratio and incidence of uveitis. Followup of the patients with peripheral arthritis revealed that 25% had developed a further criterion for Reiter's syndrome, 45% had persistent seronegative arthritis and in 55% the arthritis had resolved.