New ZealandNorwayPakistanPeruPolandPeurto RicoRomaniaRussiaSingaporeSouth AfricaSudanSwedenThailandTurkeyUnited KingdomUSAUzbekistan

Molecular Mimicry

The Geographical Distribution
of Immune Responses to Klebsiella in Ankylosing
Spondylitis and its Relevance to Therapy


(1)Immunologv Section. Division of Life Sciences. King's College, London. England.
(2)Department of Rheumatology. Hospital Clinic i Provincial de Barcelona, Barcelona. Spain.
(3)Department of Orthopaedic Surgery, Shiga University of Medical Science. Otsu. Japan.


Summary    The discovery that HLA-B27 is linked to ankylosing spondylitis (AS) and HLA-DR1/DR4 to rheumatoid arthritis (RA) has provided new approaches to the study of the possible causation of these diseases.

  Several theories have been proposed, to explain these associations but only one, namely "molecular mimicry", has provided a specific aetiological agent for each of these diseases.

  Molecular mimicry between HLA-B27 and two molecules in Klebsiella microbes: nitrogenase and pullulanase D has been reported whilst in Proteus microbes, the haemolysin molecule shows sterochemical similarity to HLA-DR1/DR4.

  Elevated immune responses to Klebsiella microbes have been demonstrated in AS patients from 10 different countries and this wide geographical distribution suggests that the same aetiological agent is probably acting in producing this condition.

  Furthermore RA patients show similar immune responses to Proteus microbes. Whether AS or RA are caused by these bacteria can only be resolved by tissue typing all rheumatological patients early, in the course of their disease and then assessing their response to antibiotic chemotherapy in longitudinal studies involving double-blind crossover trials.

  It is possible that in the future, the course of AS or even RA could be modified by adequate antibiotic chemotherapy or even diets which affect the substrates on which these bacteria grow.

Key words Ankylosing Spondylitis, Klebsiella, Treatment of AS.



  The discovery in the 1970's that 96% of ankylosing spondylitis (AS) patients are HLA-B27 positive, whilst only 8% of the general U.K. population carry this antigen, and the subsequent demonstration that over 90% of rheumatoid arthritis (RA) patients possess HLA-DR1/DR4 subtypes carrying the EQR(K)RAA susceptibility sequence which is present in 30 to 40% of controls has altered the way we look at inflammatory arthritis forever. Any theory proposing an aetiological hypothesis for either AS or RA, must at the same time provide a satisfactory explanation for these links with HLA molecules. Two main theories have so far been proposed to explain the association of HLA with disease.


  The first theory is based on the observations of Bjorkman et al., where crystallographic studies have shown the presence of grooves between alpha-1 and alpha-2 domains in either class I or class II MHC molecules, which could accommodate short pathogenic peptides (8-20 amino acids long) and thereby cause disease (1). No such pathogenic peptides occupying the HLA-B27 or HLA-DR1/DR4 cavities have so far been identified and therefore the validity of the theory as it applies to either AS or RA cannot be tested. In the absence of such specific peptides, no therapeutic test can be made to determine, whether withdrawal of antigens containing such peptides, cinically modifies the course of the disease.
  The second theory is based on the idea that the HLA molecule shares molecular mimicry or similarity with some antigens present in infectious organisms and following exposure, disease occurs through reactive inflammation and immune mechanisms. The prototype of a classical reactive arthritis is rheumatic fever, where molecular mimicry between streptococci and cardiac valve glycoproteins, basal ganglia and fibro-muscular tissue leads to disease, 4 to 6 weeks after a streptococcal upper respiratory tract infections;


  Molecular similarity has been demonstrated between HLA-B27 and Klebsielia nitrogenase reductase enzyme in that the sequence QTDRED is common to both molecules and AS patients appear to have antibodies to that sequence (2). Rat antisera raised against 16-mer peptides of Klebsiella nitrogenase reductase, containing in the middle the QTDRED sequence, were able to detect by immunoperoxidase, synovial biopsies obtained from HLA-B27 positive AS patients but not B27 negative biopsies obtained from RA patients (3). Rabbit anti-Klebsiella sera were able to distinguish HLA-B27 positive lymphocytes from either AS patients or healthy controls, when compared to lymphocytes obtained from HLA-B27 negative individuals (4). However it is not clear whether this enzyme is present in the bowel flora.


  The starch debranching enzyme pullulanase has 4 components (A,B,C and D) and is found mainly in Klebsiella microbes, although it is also present in some saprophytic bacteria. The search for other molecules which might crossreact with HLA-B27 was started when it was suggested that Klebsiella nitrogenase may not be expressed in bacteria of the bowel flora.
  Database analysis of published Klebsiella protein sequences showed that molecular mimicry is present between Klebsiella pneumoniae puld secretion protein (DRDE) and HLA-B27 (DRED) (5). IgG antibodies in AS patients were elevated against both a 16-mer synthetic peptides of HLA-B27 and pulD, containing the crossreacting sequences, when measured by enzyme linked immunosorbent assays (ELISA) and compared to controls (p< 0.001). The presence of antibodies in AS patients to a 16-mer peptide of HLA-B27 confirmed the data previously published by the Oldstone group (2).
A second type of crossreactivity was identified with pulA of the pullulanase enzyme which has repeating sequences of the trimer gly-X-pro and this shows molecular mimicry with types 1, III and IV collagen respectively. Increased collagen deposition is a feature of the disease and sera from AS patients were found to have increased levels of anti-collagen antibodies (5). The relevance of these antibodies in the pathogenesis of the disease remains to be determined.


Elevations in immune responses to Klebsiella in AS patients have been reported from 10 countries over the last 12 years:

1. England

  In 1983, serum IgA antibodies against Klebsiella were found to be elevated in 43 active (ESR>15 mm/hr) AS patients when compared to 39 inactive (normal ESR) AS patients, 13 psoriatic patients, 38 RA patients and 57 healthy controls, whilst no such elevations were found when the same sera were tested against E.coli and Canadian albicans (6).
  In 1984, the same group showed in a second study that active AS patients did not have antibodies against related Gram-negative bacteria such as Salmonella, Yersinia or Pseudomonas but persisted in showing elevated levels of anti-Klebsiella antibodies (7).

2. USA

  In 1987, Schwimmbeck and coworkers demonstrated that AS patients had elevated levels of antibodies to short Klebsiella peptides containing sequences crossreacting with HLA-B27 (2).

3. Finland

  In 1991 Maki-Ikola and coworkers, in a study of 99 AS patients showed that they had antibodies to Klebsiella and E.coli but not to a number of other microbes (8).

4. Scotland

  In 1992, O'Mahony and coworkers demonstrated that 14 active AS patients, as well as 14 patients with Crohn's disease had elevated levels of anti-Kiebsiella antibodies when compared to controls (9).

5. Canada

  In 1993, 33 AS patients from Toronto and Winnipeg were found to have elevated levels of antibodies to lipopolysaccharides from Klebsiella and Shigella but not against LPS from E.coli, Salmonella, Yersinia and Campylobacter (10).

6. Germany

  In 1994, in a study of 41 AS patients from Kiel specific antibodies were found to a restricted number of KJebsiella capsular polysaccharides (I 1).

7. Spain

  In 1994, in a study of 84 Catalan patients with AS, elevated levels of antibodies to Klebsiella were found by immunofluorescence (12).

8. Mexico

  In 1995, elevated levels of both humoural and cellular responses against Klebsiella were found in 13 AS patients when compared to controls (13).

9. Japan

  In 1995, in a study of 54 Japanese AS patients, 50 RA patients and 50 healthy controls, the AS patients were found to have antibodies to Klebsiella but not Proteus, whilst the RA patients had antibodies to Proteus but not against Klebsiella; thus each group was a specificity control for the other disease (14).

10. The Netherlands

  Antibodies to Klebsiella have been demonstrated in Dutch AS patients (15).


  Molecular mimicry has also been demonstrated between the disease susceptibility sequence EQR(K)RAA found in HLA-DR1/DR4 subtypes and the ESRRAL sequence of Proteus haemolysin (16,17). A second sequence IRRET was found in Proteus urease, which resembles the LRREI sequence of alpha-2 domain in type XI collagen, a component of hyaline cartilage (17). Antibodies to both Proteus molecules, haemolysin and urease are significantly elevated in RA patients when compared to either active AS patients or healthy controls (17). Furthermore elevations of IgG antibodies to Proteus mirabilis in RA patients has been reported from London (18), Winchester (19), Dublin (20), Newcastle (21), Oslo (22), Brest (23), Amsterdam (15) and Otsu in Japan (14).
  A molecular mimicry link between Klebsiella/Proteus and HLAB27/HLA-DRI/DR4 is undeniable and a serological link between AS/RA and these microbes has been confirmed by more than one group, but the pathological relevance remains unclear.


  However the observation of specific antibodies against these bacteria in both AS and RA provides a new clinical dimension to the understanding, early diagnosis and treatment of these diseases. One could rephrase the scientific problem of AS by asking "Why do AS patients have antibodies to Klebsiella?" The following hypotheses could be proposed as tentative solutions to such questions, to be resolved by longitudinal studies:

1 . The presence of anti-Klebsiella antibodies in AS patients in many countries, clearly suggests that AS is a "reactive arthritis following" Klebsiella infection in a HLA-B27 positive individual".
2. Similarly, the presence of anti-Proteus antibodies in RA patients in many countries again suggests that RA is a "reactive arthritis following Proteus infection in an HLA-DR1/DR4 individual".


  The empirical connections between AS,HLA-B27 and Klebsiella on the one hand and RA,HLA-DR1/DR4 and Proteus on the other, suggests an entirely new approach to the practice of rheumatology in the future. The specific immunogenetic and microbiological discoveries of the last two decades in AS and RA, provide a framework, whereby early stages of these diseases can be identified, before radiological changes have occurred and responses to early treatment evaluated in prospective longitudinal studies. It is not inconceivable that early therapeutic intervention may modulate or reduce disease expression in AS or RA, similar to the situation of prompt treatment of streptococcal infections which may have probably contributed to the marked decrease in the incidence of rheumatic fever.

  How could this be achieved? Any patient with a rheumatological symptom (backache, arthralgia or arthritis, fibro-muscular pain, morning muscle stiffness, effusion) should be tested on his/her first visit to a rheumatology department for:

a. Inflammatory activity (ESR and serum CRP)
b. HLA status (HLA-B27 and HLA-DR1/DR4)
c. Antibody titre (IgA anti-Klebsiella and IgG antiproteus levels).

  The results of one blood sample would assign the prospective patient to one of four taxonomic categories:

1. B27 group which may lead to AS
2. DR1/DR4 group which may lead to RA
3. Overlap group with both B27 and DR1/DR4 alleles
4. A "control" group with no predisposing B27 or DR1/DR4 molecules.

  The serial measurements of acute phase reactants and antibacterial antibodies would provide objective laboratory parameters to be evaluated in conjunction with clinical features and radiological changes, in longitudinal studies and responses to therapeutic interventions determined in double-blind cross-over trials.
  The most important question to be answered by such studies is whether reduction of anti-bacterial titres in either B27 or DR1/DR4 positive patients, delays or modifies the clinical course of AS or RA. Such studies cannot be carried out in patients with advanced radiological changes, when damage is irreversible and permanent. Only determination of HLA status would identify potential patients with genetic predispositions, who may have been exposed to environmental agents which appear to be associated with inflammatory disease activitv in AS and RA.
  It may be romantic to suggest that inflammatory arthritis could be associated with common microbes we all possess, but to propose that "complex cascades of CD numbers and interleukins" (24) specifically define RA is clearly wrong, since these features belong to "inflammation", a process common to both active AS and RA.
  In the immortal words of the Bard:
  "To be or not to be, 'the recipient of new ideas' that is the question." Hamlet (Shakespeare)


1. Bjorkman, P.J., Saper, M.A., Samraou, B., Benett, WS., Strtominger. J.L., Wiley, D.C. Structure of human class I histocompatibility antigen HLA-A2. Nature 1987,329,506-512.
2. Schwimmbeck. P.L., Yu. D.T.Y., Oldstone, M.B.A. Autoantibodies to HLA-B27 in the sera of HLA-B27 patients with anklylosing spondylitis and Reiter's syndrome: molecular mimicry with Klebsiella pheumoniae as potential mechanism of autoimmune disease. J exp Med 1987, 166, 173-181.
3. Husby, G., Tsuchiya, N., Schwimmbeck, PL., Keat, A., Pahle. J.A., Oldstone, M.B.A., Williams. R.C. Cross-reactive epitope with Klebsiella pneumoniae nitrogenase in articular tissue of HLAB27 positive patients with ankylosing spondylitis. Arthritis Rheum 1989, 32, 437-445.
4. Baines, M., Ebringer, A., Avakian, H., Samuel. D., James. D.C.O. The use of enzyme immunoassay (EIA) and radiobindifig assay to investigate the crossreactivity of Klebsiella antigens and HLAB27 in ankylosing spondylitis patients and healthy controls. Scand J Rheum 1990, 19. 341-349.
5. Fielder, M., Pirt, S.J., Tarpey. I., Wilson, C., Cunningham, R. Ettelaie, C., Binder. A., Bansal, S., Ebringer. A. Molecular mimicry and ankylosing spondylitis: possible role of a novel sequence in pullulanase of Klebsiella pneumoniae. FEBS Letters 1995, 369, 243-248.
6. Trull. A.K., Ebringer. R., Panayi, G.S., Colthorpe. D., James. D.C.O., Ebringer, A. IgA antibodies in ankylosing spondylitis. Scand J Rheum 12, 249-253, 1983.
7. Trull, A., Ebringer, A., Panayi, G., Ebringer, R., James, D.C.O. HLA-B27 and the immune response to enterobacterial antigens in ankylosing spondylitis. Clin Exp Immunol 1984, 55, 74-80.
8. Maki-lkola. O., Lehtinen. K., Granfors, K., Vainiopaa, R., Toivanen. P. Bacterial antibodies in ankylosing spondylitis. Clin exp Immunol 1991. 84.472-475.
9. OMahony. S., Anderson. N.. Nuki. G., Ferguson, A. Systemic and mucosal antibodies to Klebsiella in patients with ankylosing spondylitis and Crohn's disease. Ann Rheum Dis 1992,51, 1296-1300.
10. Thomson. G.TD., Alfa, M., Orr. K., Inman, R.D., Olson. N., Thomson, B.R.J. Humoral immune response to bacterial lipopolyvsaccharide in adult idiopathic spondyloarthropathies. American Colllege of Rheumatoltgy: 57th Annual Meeting. San Antonio 1993. Abstract Vol: S224. C218.
11. Sahly. H., Podschun. R., Sass. R., Broker, B., Kekow. J., Gross, W.L., Ullmann, U. Serum antibodies to klebsiella capsular Polysaccharides in anklyiosing spondylitis. Arthritis Rheum 1994.37. 754-759.
12. Colladta, A., Gratactos. J., Ebringer, A., Rashid, T., Marti, A., Sanmarti, R., Munoz-Gomez, J. Serum IgA anti-klebsiella antibodies in anklyiosing sptondylitis patients from Catalonia. Scand J Rheum 1994. 23. 119-123.
13. Gircia-Latorre, E.A., Dominguez-Lopez. M.L., JimenezZamudio, L.A., Burgos-Vargas. Granados-Arreola. J. Cellular immune response to different antigenic fractions of Klebsiella pneumonia in anklylosing spondylitis. 9th Int Cong Immunol. (San Francisco) Abst Vol. pg. 165, 1995.
14. Tani, Y., Tiwana, H., Hukuda, S., Nishioka. J., Fielder. M., Bansal, S., Ebringer, A. Antibodies to Klebsiellia and Proteus in Japanese patients with ankylosing spondylitis and rheumatoid arthritis. Arthr Rheum (Submitted for publication).
15. Sprenkels, S. (Personal communication).
16. Ebringer, A., Cunningham. P., Ahmadi, K., Wrigglesworth, J., Hosseini. R., Wilson. C. Sequence similarity between HLA-DRI and DR4 subtypes associated with rheumatoid arthritis and Proteus/Seiratia membrane haemolysins. Ann Rheum Dis 1992, 51. 12451246.
17. Wilson, C., Ebringer. A., Ahmadi. K., Wrigglesworth, J., Tiwana, H., Fielder. M., Binder, A., Ettelaie. C., Cunningham, P., Joannou. C., Bansal, S. Shared amino acid sequences between major histocompatibility complex class 11 glycoproteins, type XI collagen and Proteus mirabilis in rheumatoid arthritis. Ann Rheum Dis 1995. 54. 216-220.
18. Ebringer, A., Ptaszvnska. T. Corbett, M., Wilson, C., Macafee, Y., Avakian, H., Baron. P., James. D.C.O. Antibodies to Proteus in rheumatoid arthritis. Lancet 1985, ii, 305-307.
19. Ebringer, A., Cox, N.L., Abuljadayel. I., Ghuloom, M., Khalafpour, S., Ptaszynska. T., Shodjai-Moradi, F., Wilson, C. Klebsiella antibodies in ankylosing spondylitis and Proteus antibodies in rheumatoid arthritis. Brit J Rheum 1988, 27 (Suppl. 2), 72-85.
20. Rogers, P., Hassan, J., Bresnihan. B., Feighery, C., Whelan, A. Antibodies to Proteus in rheumatoid arthritis. Brit J Rheum 1988. 27, (suppl. 2). 90-94.
21. Deighton. C.M., Gray, J.W, Bint, A.J., Walker. D.J. AntiProteus antibodies in rheumatoid arthritis same-sexed sibships. Brit J Rheum 1992, 31, 241-245.
22. Kjeldsen-Kragh, J., Rashid, T., Dybwad, A.. Sioud, Haugen. M., Forre, O., Ebringer, A. Decrease in anti-Proteus mirabilis but not anti-Eschericihia coli antibody levels in rheumatoid arthritis patients treated with fasting and a one year vegetarian diet. Ann Rheum Dis 1995, 54, 221-224
23. Fielder, M., Tiwana, H., Youinou. R. Le Goff. P., Deonarain. R., Wilson. C., Ebringer, A. The specificity of the anti-Proteius antibody response in tissue typed rheumatoid arthritis (RA) patients from Brest. Rheumatol Int 1995,15. 79-82.
24, Gaston, H. Proteus - is it a likely aetiological factor in chronic polyarthritis? Ann Rheum Dis 1995, 54. 157-158.

Correspondence to: Dr. Alan EBRINGER,
Division of Life Sciences, King's College, Campdcn Hill Road. Kensington. LONDON W8 7AH. U.K.

Valid HTML 4.0 Transitional