вход по аккаунту


Mycobacteria and rheumatoid arthritis.

код для вставкиСкачать
43 1
Evidence for the involvement of a mycobacterial 65-kd heat-shock protein (hsp65) (or of crossreactive autoantigens) in the etiology of rodent models
of arthritis and of human rheumatoid arthritis (RA) has
become more provocative following the recent discovery that a change in the glycosylation of IgG, originally
described in RA, has also been observed in rodent
models of arthritis and in human mycobacterioses.
What is the relationship between these 2 sets of
findings, and do they point to a mycobacterial etiology
for RA? Experimental findings are analyzed below,
and 2 hypotheses are discussed.
Susceptible strains of rats develop arthritis 1014 days after injection of killed mycobacteria in a
water-in-oil emulsion (Freund’s complete adjuvant;
FCA). Histologically, adjuvant-induced arthritis does
not resemble the human disease because cartilage is
not affected, but this is attributable to the fact that rat
cartilage, in contrast to human cartilage, is much less
susceptible to cytokine-mediated degradation (Dingle
J: personal communication).
Recent studies have emphasized the relevance
of the rodent models. A T cell line isolated from the
lymph nodes of arthritic rats and cultured in the
presence of soluble antigens from Mycohacterium
From the Departments of Medical Microbiology and Immunology, University College and Middlesex School of Medicine,
London, United Kingdom.
Dr. Rook’s work is supported in part by G . D. Searle & Co.
Graham Rook, MD: Reader in Medical Microbiology; Peter
Lydyard, PhD: Senior Lecturer in Immunology; John Stanford,
MD: Reader in Medical Microbiology.
Address reprint requests to Graham Rook, MD, Department of Medical Microbiology, School of Pathology, University
College and Middlesex School of Medicine, Riding House Street,
London W I P 7PP. U K .
Arthritis and Rheumatism, Vol. 33, No. 3 (March 1990)
tuberculosis transferred the disease to irradiated, but
otherwise normal, recipients. An arthritogenic clone
derived from this cell line recognized cartilage components (1) and a 65-kd mycobacterial antigen that is a
homolog of a family of heat-shock proteins (2). When
administered in oil before FCA, mycobacterial hsp65
inhibited subsequent induction of arthritis (2). However, it is not yet clear whether these experiments
emphasize the arthritogenicity of mycobacteria or of
the 65-kd hsp, because this protein is a member of a
gene family that is highly conserved in all prokaryotes
and eukaryotes from which it has been sequenced.
This mycobacterial gene product shows 60% sequence
identity with the groEL gene product of Escherichiu
coli. It also shows a 47% sequence identity and an
additional 20% conservative change when compared
with the recently cloned human homolog (3). The
rodent sequence is probably very similar to the human one.
These findings indicate that mycobacterial heatshock protein would be an ideal candidate as an
autoimmunogen. However, since it is so highly conserved, many microorganisms other than mycobacteria could be capable of evoking autoimmunity to the
mammalian homologs. This is consistent with the
observation that pretreatment with mycobacterial
hsp65 inhibits induction of streptococcal cell wallinduced arthritis in the rat (4).However, this reasoning is incomplete because similar pretreatment with
the hsp65 of M tuberculosis inhibits induction of
pristane-induced arthritis in mice (3,and the Mycobacterium leprae homolog inhibits arthritis induction
by low doses of the synthetic lipoidal amine adjuvant
(CP-20961) in the rat (Billingham M: personal communication).
There are at least 3 possible explanations for
the results observed in the latter 2 models. First, the
host hsp6S homolog itself may function as the autoantigen. Second, some other cross-reactive host antigen
could be involved, as was suggested for cartilage
components in the adjuvant-induced arthritis model
described above (1). Third (although less probable), it
is conceivable that pretreatment with hsp6S exerts a
subtle regulatory effect on responses to unrelated
This dilemma is clinically important because
there is evidence for the involvement of the mycobacterial hsp6S or cross-reactive antigen in RA. Sera from
RA patients have been shown to contain elevated
levels of antibody that binds to the mycobacterial
hsp6S. IgG binding was found to be greater in RA than
in tuberculosis (TB) (6) or several other inflammatory
disorders (7). This has been observed in different
laboratories, using sera from different racial groups,
e.g., Kuwaitis (6) and British citizens (7). At least
some of this antibody seemed to be autoantibody,
since the human 65-kd protein, which is present in
large quantities in RA synovial fluid, was also present
in immune complexes precipitated from synovial fluid
(ref. 8 and Sharif M, Rook G, Lydyard P: unpublished
observations). However, antibody levels did not appear to be increased in RA patients when the E coli
homolog was used (7). Thus, more hsp65 homologs
must be tested before this finding constitutes evidence
that mycobacteria play a special role in RA induction.
Several groups have demonstrated the presence of T
lymphocytes in RA synovial fluid that recognize mycobacterial hsp6S (9,lO). In addition, the overall pattern of skin test response to mycobacterial antigens in
RA patients was found to closely resemble that in
patients with mycobacterial infections, and there were
significant correlations with the presence of DR4 and
DR7 ( I 1).
A parallel line of investigation has also suggested a role for mycobacteria in the induction of RA.
There is a biantennary oligosaccharide on a conserved
glycosylation site on the C,2 domain of all IgG heavy
chains. Patients with RA have an increased percentage
of IgG molecules that lack galactose on both arms of
this oligosaccharide (12). This probably accounts for
the decrease in total galactose content that has been
reported (13). The resulting “glycoform” of IgG is
known as agalactosyl IgG, or GO, and is associated
with decreased activity of a galactosyl transferase in B
lymphocytes (14). GO levels are also elevated in TB
and Crohn’s disease (CD) (IS), and in leprosy, during
episodes of erythema nodosum leprosum (ENL) (16).
Since CD has been considered by some investigators
to be of mycobacterial origin (17,18), could an increased GO level be a marker of mycobacterial infection? Preliminary studies of families have suggested
that GO may be elevated in a few spouses of RA
patients (Roitt IM: unpublished observations). If confirmed, this observation could imply a shared environmental factor, such as an infection. The extraordinarily slow-growing, pleomorphic, acid-fast organisms
found in the lymph nodes of patients with CD (17) have
been characterized to the point where their genomic
DNA can be shown to be related to that of the
mycobacteria (Visuvanathan S: personal communication). We cannot rule out the possibility that RA is due
to infection with similar organisms.
However, more recent studies of GO suggest
that a raised level of this glycoform of IgG does not
always indicate infection with mycobacteria. GO levels
have been shown to be elevated in rats with adjuvantinduced arthritis (Rademacher T: personal communication) and in mice with arthritis induced by type I1
collagen in FCA (ref. 19 and Rook G, Whyte A:
unpublished observations). In these models, the only
mycobacteria known to be present are the killed ones
in the adjuvant. However, killed mycobacteria are not
essential for the induction of elevated GO levels in
rodents, since GO is also elevated in mice with arthritis
induced by intraperitoneal injection of pristane (Rook
G, Thompson S, Elson C: unpublished observations).
Elevated GO levels can also be dissociated from
arthritis. This is obvious in humans, where most cases
of TB or CD are not complicated by arthritis. Similarly, GO is strikingly elevated during the necrotizing T
cell-dependent inflammation provoked by schistosome ova in mice 42 days after infection with Schistosoma mansoni (Rook G, Dunne D: unpublished
observations). This model is of particular importance
because no mycobacteria are involved. Conversely,
heat-shock proteins may still be relevant because IgG
binding to the mycobacterial hsp6S increases in parallel with GO. This may be due to the anticipated
cross-reactivity with the schistosomal hsp65 homolog.
An elevated GO level is not confined to situations involving mycobacteria, or to arthritides. It
appears to be a correlate of a type of T cell-dependent
tissue-damaging pathology, accompanied by cytokine
release and an acute-phase response (19), similar to
that in TB, CD, RA, ENL, and murine schistosomiasis. GO is usually not elevated when there are T
cell-dependent granulomata without tissue destruc-
tion, as observed in sarcoidosis or leprosy without
E N L (15), or when there is an acute-phase response
without chronic T cell activity. It is not clear whether
the frequent correlation between GO and IgG binding
to heat-shock protein is a circumstantial result of a
tendency for both to occur during tissue-damaging
inflammation or is a consequence of a novel regulatory
role of heat-shock protein. Of interest is the finding
that in acute rheumatic fever, no increase in either GO
o r antibody to heat-shock protein is observed (20).
The intriguing point is that all of these clinical
conditions (TB, E N L , CD, and schistosomiasis),
which can be associated with elevated GO levels, can
also be associated with RA-like syndromes. In 1897,
Poncet described a syndrome occurring in tuberculosis
patients that was quite distinct from tuberculous arthritis. Sporadic reports of Poncet’s disease still occur
and are reviewed elsewhere (21). Similar syndromes
that accompany multibacillary leprosy, especially in
patients with ENL, have been observed in patients in
India, Papua New Guinea, and Egypt (22,23), but not
in Thailand o r Malaysia. Similarly, an arthritis that
resembles RA has been seen in some patients with
CD (24). In addition, an arthritis resembling RA,
which responds to treatment of the underlying infection, has been seen in some schistosomiasis patients in
Egypt (25).
The simplest unifying hypothesis is that the
highly conserved and cross-reactive hsp65 is an important target antigen (not necessarily the only one) in the
joints of RA patients, and that a specific type of T
cell-mediated effector response to this antigen leads to
chronic disease. There are 2 obvious categories of
explanation for the simultaneously increased GO level.
GO may contribute in some unknown way to the
imrnunopathologic condition, or it may merely be a
correlate of a specific type of tissue-damaging cellmediated response involved in RA, TB, CD, and ENL.
How might the fine specificity and effector
functions of the human response to hsp65 homologs
(or other cross-reactive antigen) be regulated? In 2
rodent models where hsp65 has been shown to be
relevant and can be used to block induction of disease,
the incidence of arthritis is strongly influenced by the
gut flora. This suggests regulation by the nature and
degree of exposure to cross-reactive hsp65 homologs.
Thus, the susceptibility of germ free rats to adjuvant
arthritis can be increased or decreased by appropriate
reconstitution of the bowel flora (26). Similarly, the
bowel flora can induce resistance to the streptococcal
cell wall-induced arthritis in F344 rats, probably via a
mechanism involving immunologic tolerance (van den
Broek M: personal communication).
It is tempting to speculate that the gut plays a
similarly important role in humans. There are reports
of abnormalities of the gut epithelium in RA, although
it is difficult to distinguish between primary abnormalities and changes that are secondary to the effects of
antiinflammatory drugs (27). Patients with reactive
arthritis, like those with RA, have T lymphocytes that
recognize the 65-kd heat-shock protein in their joints
(9,lO). The gut epithelium contains T cells bearing ylS
receptors, and recent evidence indicates that these
cells have a tendency to recognize bacterial, possibly
mycobacterial, antigens (for review, see ref. 28). They
may also be involved in the regulation of oral tolerance
(ref. 29 and, for discussion, see ref. 10). Much attention will be focused on the presence of increased
numbers of ylS T cells in RA joints (lo), particularly if,
as reported for a single clone, these T cells recognize
the mycobacterial hsp65 (10) or other cross-reactive
bacterial antigens.
In conclusion, neither the arthritis models in
which hsp65 plays a role, nor the human or animal
conditions in which GO levels are elevated, have been
shown to be exclusively mycobacterial in origin. It is
conceivable that many hsp65-containing organisms are
equally important. The antibody and skin test studies
with mycobacterial antigens in RA are suggestive, but
do not yet provide genus-specific proof that mycobacteria play a role in RA.
Nevertheless, it is possible to make a case for a
special role for mycobacteria in RA in humans, and we
propose 2 hypotheses. First, mycobacteria possess
relevant adjuvant properties and readily induce raised
GO levels, and some patients with classic mycobacterial infections develop inflammatory arthritides.
Therefore, since there are mycobacterium-like organisms that are extraordinarily difficult to detect or grow
in culture (17,18), infection cannot, at present, be
ruled out.
An alternative speculation, suggested by the
role of the gut flora as described above, is a changing
degree of exposure to environmental mycobacteria.
The mycobacteria are a largely saprophytic genus.
They are so abundant in some environments that
direct Ziehl-Neelsen staining of soil or water samples
can reveal their presence. The immune response must
have evolved in the presence of massive quantities of
mycobacteria, which may be needed for the correct
“setting” of the immunologic network or T cell repertoire. However, as a result of changed lifestyles and
modern water supplies, their presence in t h e human
gut now varies considerably , being, for example, more
common in Africa than in E u r o p e (Portaels F: personal
communication), and it is not clear whether these
organisms are truly commensal.
T h u s , exposure t o t h e mycobacterial heatshock protein, unlike exposure to that of commensal
genera, is a variable that depends on how and where
one lives. Could this factor explain the changing
incidence of RA?
We are grateful to Drs. M. Billingham, M. van den
Broek, J. Dingle, D. Dunne, C. Elson, H. Gaston, F.
Portaels, T. Rademacher, I. M. Roitt, S. Thompson, and S.
Visuvanathan for permission to quote unpublished findings.
1. Van Eden W, Holoshitz J , Cohen I: Antigenic mimicry
between mycobacteria and cartilage proteoglycans.
Concepts Immunopathol 4: 144-170, 1987
Van Eden W, Thole JER, van der Zee R, Noordzij A,
van Embden JDA, Hensen EJ, Cohen 1R: Cloning of the
mycobacterial epitope recognised by T lymphocytes in
adjuvant arthritis. Nature 331: 171-173, 1988
Jindal S, Dudani AK, Singh B, Harley CB, Gupta RS:
Primary structure of a human mitochondria1 protein
homologous to the bacterial and plant chaperonins and
to the 65 kilodalton mycobacterial antigen. Mol Cell Biol
9:2279-2283, 1989
Van den Broek MF, van Bruggen MCJ, Hogervorst
EJM, van Eden W, van der Zee R, van den Berg WB:
Protection against streptococcal cell wall-induced arthritis by pretreatment with the mycobacterial 6SkDa heat
shock protein. J Exp Med 170:449-466, 1989
Thompson SJ, Bedwell A, Hooper DC, Burtles SS,
Rook CAW, van Embden JDA, Elson CJ: Pristaneinduced arthritis: a possible role for the mycobacterial
65kD heat shock protein, Stress Proteins in Inflammation. Edited by C Rice-Evans, R Burdon, V Winrow, D
Blake. London, Richelieu Press (in press)
Bahr GM, Rook CAW, A1 Saffar M, van Embden J,
Stanford JL, Behbehani K: Antibody levels to mycobacteria in relation to HLA type: evidence for nonHLA-linked high levels of antibody to the 6SkDa heat
shock protein of M. tuberculosis in rheumatoid arthritis.
Clin Exp Immunol 74:211-215, 1988
Tsoulfa G , Rook GAW, Bahr G, Sattar MA, Young DB,
Mehlert A, van Embden JDA, Hay F, Isenberg D,
Lydyard PML: Elevated IgG antibody levels to the
mycobacterial 65kDa heat shock protein are a charac-
teristic of patients with rheumatoid arthritis. Scand J
lmmunol 30519-527, 1989
8. Lydyard PM, Tsoulfa G , Sharif M, Smith M, Young DB,
Bahr GM, van Embden JDA, Hay FC, Isenberg DA,
Gupta RS, Lamb J, Mayanil CSK, Venner T, Rook
CAW: Antibodies to heat shock proteins in rheumatoid
arthritis, Stress Proteins in Inflammation. Edited by C
Rice-Evans, R Burdon, V Winrow, D Blake. London,
Kichelieu Press (in press)
9. Gaston JSH, Life PF, Bailey LC, Bacon PA: In vitro
responses to a 65kDa mycobacterial protein by synovial
‘I’ cells from inflammatory arthritis. J Immunol 143:
2494-2500, 1989
10. Holoshitz J, Koning F, Coligan JE, de Bruyn J, Strober
S: Isolation of CD4-CD8- mycobacteria-reactive T
lymphocyte clones from rheumatoid arthritis synovial
fluid. Nature 339:226-229, 1989
11. Bahr GM, Sattar MA, Stanford JL, Shaaban MA, Al
Shimali B, Siddiqui Z, Gabriel M, Al SaEar M, Shahin
A, Chugh TD, Rook GAW, Behbehani K: HLA-DR and
tuberculin tests in rheumatoid arthritis and tuberculosis.
Ann Rheum Dis 48:6348, 1989
12. Parekh RB, Dwek RA, Sutton BJ, Fernandes DL,
Leung A, Stanworth DR, Rademacher TW, Mizuochi T,
Taniguchi T, Matsuta K , Takeuchi F, Nagano Y,
Miyamoto T, Kobata A: Association of rheumatoid
arthritis and primary osteoarthritis with changes in the
glycosylation pattern of total serum IgG. Nature 316:
452-457, 1985
13. Tomana M, Schrohenloher RE, Koopman WJ, Alarcon
GS, Paul WA: Abnormal glycosylation of serum IgG
from patients with chronic inflammatory diseases. Arthritis Rheum 31:333-338, 1988
14. Axford JS, Mackenzie L, Lydyard PM, Hay FC, Isenberg DA, Roitt IM: Decreased B cell galactosyl transferase in rheumatoid arthritis. Lancet 11: 1486-1488, 1988
15. Radernacher T, Parekh RB, Dwek RA, Isenberg D,
Rook GAW, Axford JS, Roitt I: The role of IgG glycoforms in the pathogenesis of rheumatoid arthritis.
Springer Semin Immunopathol 10:231-249, 1988
16. Filley E, Andreoli A, Steele J, Waters M, Wagner D,
Nelson D, Tung K, Rademacher T, Dwek R, Rook
GAW: A transient rise in agalactosyl IgG correlating
with free IL-2 receptors during episodes of erythema
nodosum leprosum. Clin Exp lmmunol76:343-347,1989
17. Burnham WR, Lennard-Jones JE, Stanford JL, Bird
RG: Mycobacteria as a possible cause of inflammatory
bowel disease. Lancet 11:693-696, 1978
18. Stanford JL: Much’s granules revisited. Tubercle 68:
241-242, 1987
19. Rook GAW: Mycobacteria, the glycosylation of IgG,
and rheumatoid arthritis. Scand J Immunol 2R487-493,
20. Bahr GM, Yousof AM, Majeed HAM, Behbehani K,
Lubani M, Parekh RB, Dwek RA, Rademacher TW,
Young DB, Mehlert A, Steele J, Rook CAW: Neither
agalactosyl IgG nor antibodies to heat shock proteins are
raised in acute rheumatic fever. Ann Rheum Dis (in
Isaacs AJ, Sturrock RD: Poncet’s disease: fact or fiction? A reappraisal of tuberculous rheumatism. Tubercle
55: 135-142, 1974
Ramu G, Balakrishnan S: Arthritis in lepromatous leprosy. Lepr India 40: 1-8, 1968
Atkin SL, Welbury RR, Stanfield E, Beavis D, lwais B ,
Dick WC: Clinical and laboratory studies of inflammatory polyarthritis in patients with leprosy in Papua, New
Guinea. Ann Kheum Dis 46:688-690, 1987
Gravallese EM, Kantrowitz FG: Arthritic manifestations of inflammatory bowel disease. Am J Gastroenterol 83:703-709, 1988
Atkin SL, Kame1 M, El Hady AMA, El Badawy SA, El
Ghobary A, Dick WC: Schistosomiasis and inflammatory polyarthritis: a clinical, radiological and laboratory
study of 96 patients infected by S. mansoni with particular reference to the diarthrodial joint. Q J Med 59:
479-487, 1986
Kohashi 0, Kohashi Y, Takahashi T, Ozawa A, Shigematsu N: Reverse effect of gram-positive bacteria vs.
gram-negative bacteria on adjuvant-induced arthritis in
germfree rats. Microbiol Immunol 29:487497, 1985
Doube A, Collins AJ: Is the gut intrinsically abnormal in
rheumatoid arthritis? Ann Rheum Dis 47:617419, 1988
Raulet DH: Antigens for $6 T cells. Nature 339:
342-343, 1989
Kitamura K , Kiyono H, Fujihashi K, Eldridge JH,
Green DR, McGhee JR: Contrasuppressor cells that
break oral tolerance are antigen-specific T cells distinct
from T helper (L3T4+), T suppressor (Lyt-2+) and B
cells. J Immunol 139:3251-3259, 1988
Без категории
Размер файла
444 Кб
mycobacterium, arthritis, rheumatoid
Пожаловаться на содержимое документа