Download Depressed proliferative responses by peripheral blood mononuclear

Rheumatology 1999;38:631–635
Depressed proliferative responses by peripheral
blood mononuclear cells from early arthritis
patients to mycobacterial heat shock protein 60
J. M. Ramage and J. S. H. Gaston
Department of Medicine, University of Cambridge, Cambridge CB2 2QQ, UK
Abstract
Objectives. T-cell responses to mycobacterial heat shock protein 60 (M.hsp60) have been
implicated in the pathogenesis of adjuvant arthritis, but whether they play a role in
rheumatoid arthritis (RA) is undefined. We therefore examined T-cell responses to M.hsp60
and to other recall antigens in a cohort of patients with early RA and in healthy controls.
Methods. In vitro peripheral blood mononuclear cells’ (PBMC ) proliferative responses to
antigen were measured by [3H ]thymidine incorporation, and results correlated with clinical
and laboratory features of disease.
Results. Whereas responses to the recall antigens tetanus toxin and purified protein
derivative (PPD) were equivalent in the two groups, responses to both M.hsp60 and the
Escherichia coli hsp60 were lower in the RA patients. These results could not be explained by
either the higher prevalence of HLA-DR4 in the RA group, or the disease severity of the
patients.
Conclusion. In the light of results from the adjuvant arthritis model which suggest that
arthritis may be ameliorated by the actions of an hsp60-reactive T-cell population, the lack of
response to M.hsp60 in RA could contribute to disease persistence.
K : Rheumatoid arthritis, Heat shock protein 60, T cell.
Heat shock proteins have been implicated as possible
target antigens in rheumatoid arthritis (RA) primarily
because of work on experimental models of arthritis.
The most notable of these is adjuvant arthritis, in which
Lewis rats develop arthritis after injection with killed
Mycobacterium tuberculosis in oil [complete Freund’s
adjuvant (CFA)] [1]. From such rats, a mycobacteriaspecific T-cell line, A2, was generated [2, 3] which, on
transfer into naive irradiated Lewis rats, resulted in the
development of arthritis. A clone derived from this line,
A2b, was found to recognize mycobacterial heat shock
protein 60 (M.hsp60) and to cross-react with a component of articular cartilage proteoglycan [4, 5].
These experiments suggested that the M.hsp60-specific
T-cell clone was arthritogenic because of molecular
mimicry between M.hsp60 and proteoglycan. Therefore,
it would have been predicted that immunization with
M.hsp60 itself, instead of CFA, would also give rise to
arthritis. However, instead of developing arthritis, rats
immunized with M.hsp60 were protected from subsequent attempts to induce arthritis with CFA [4].
Similar effects of pre-immunization with M.hsp60 were
shown in other rodent models of arthritis, including
those induced by immunization with streptococcal cell
walls, type II collagen, the synthetic adjuvant CP20961,
mineral oil and pristane [6–8]. These results have
given rise to the idea of an hsp60-specific T-cell population which is able to downregulate or protect against
joint inflammation. In more recent experiments, protection could also be obtained by immunization with a
particular peptide within M.hsp60, a peptide differing
from that recognized by the arthritogenic clone A2.
Paradoxically, the protective peptide has an amino acid
sequence similar to that found in self (rat) hsp60, and
the protective T cells were also responsive to self-hsp60
[9]. However, protective T cells were only induced by
immunization with the peptide from M.hsp60 and not
with the corresponding peptide from self-hsp60. Thus,
from rodent studies, immune responses to M.hsp60
could be either arthritogenic, as suggested indirectly by
the work of van Eden and Holoshitz [2–5], or protective
as suggested by the more recent experiments.
The relevance of these observations in animal models
to human inflammatory arthritis, such as RA, remains
unclear, but there are two major possibilities. Firstly,
RA patients might make a response to M.hsp60 which
is arthritogenic by the same mechanism which operates
in adjuvant arthritis. Alternatively, RA patients might
lack the protective response to M.hsp60, which in
rodents is capable of downregulating arthritis due to
Submitted 15 September 1998; revised version accepted
25 February 1999.
Correspondence to: J. S. H. Gaston, University of Cambridge,
School of Clinical Medicine, Department of Medicine, Addenbrooke’s
Hospital, Level 5, Box 157, Hills Road, Cambridge CB2 2QQ, UK.
631
© 1999 British Society for Rheumatology
632
J. M. Ramage and J. S. H. Gaston
several kinds of antigenic or adjuvant challenge.
However, there is some controversy over the extent of
the immune response to M.hsp60 in RA. Early studies
in RA patients suggested that T-cell responses to
M.hsp60 were readily observed in most patients and
that higher responses were present in synovial fluid
[10–12]. The latter observation may be explained by the
preferential recruitment of memory T cells to sites of
inflammation such as arthritic joints; similar responses
to M.hsp60 were seen in T cells from pleural effusions.
In addition, the specificity of the responses to M.hsp60
observed in RA has been questioned, since later studies
indicated that many may have been due to contamination of the recombinant M.hsp60 with Escherichia coliderived proteins [13]. Taking these factors into consideration, preliminary evidence suggested that RA patients
might in fact be less responsive to M.hsp60 than controls
or patients with other forms of arthritis (including the
human equivalent of adjuvant arthritis, associated with
intra-vesical treatment with BCG [14]).
The current study was designed to investigate T-cell
proliferative responses to M.hsp60 in a cohort of patients
with early RA. We show that peripheral blood mononuclear cells’ (PBMC ) proliferative responses to
M.hsp60 were significantly lower in RA patients than
in healthy controls, whereas responses to other recall
antigens such as purified protein derivative (PPD) and
tetanus toxoid were similar in the two groups. This
difference extended to responses to another bacterial
hsp60, GroEL. Despite the high prevalence of HLADR4 in early RA patients, the lower response to
M.hsp60 was not confined to DR4+ patients and did
not correlate with other clinical characteristics.
Methods
Patients
Blood samples were obtained from 26 healthy individuals and 58 patients attending an early arthritis clinic.
The patients each had a disease duration of <5 yr. The
average age of the patients was 58 yr (range 22–84,
median 60), 31 were rheumatoid factor seropositive
(>60 U/ml ) and 35 were receiving disease-modifiying
anti-rheumatic drugs. Thirty patients had laboratory
evidence of active disease [erythrocyte sedimentation
rate ( ESR) >20 mm/h and/or C-reactive protein (CRP)
>10 mg/ml ]. HLA typing was performed by Jeff Faint,
Department of Rheumatology, Birmingham, as previously described [15].
Antigens
The following antigens were used: recombinant
Mycobacterium leprae heat shock protein 60 (M.hsp60)
(a kind gift of Dr P. J. Jenner, N.I.M.R., Mill Hill,
UK ) used at 10 mg/ml; GroEL (a kind gift of Dr
P. Lund, University of Birmingham, Birmingham,
UK ) used at 10 mg/ml; PPD of tuberculin (Statens
Seruminstitut, Copenhagen, Denmark) at 10 mg/ml; tetanus toxoid (Statens Seruminstitut, Copenhagen,
Denmark) at a 1:200 dilution.
Proliferation assays
PBMC were obtained following centrifugation on FicollPaque (Pharmacia Biotechnology Ltd, Milton Keynes,
UK ) at 220 g for 30 min. A total of 1 × 105 cells were
cultured in 1640 RPMI (Sigma, Poole, UK ) with 5%
human pre-screened male AB+ or A+ serum and 5 m
HEPES (Sigma) in triplicate in 200 ml volume microtitre tissue culture plates (Falcon, Becton-Dickinson
Labware, Lincoln Park, CA, USA) with the appropriate
concentration of antigen for 6 days. Proliferation was
measured by [3H ]thymidine (0.15 mCi/well ) incorporation over the last 16 h of culture. Lymphocyte responses
were expressed as a stimulation index (SI ): the mean
counts per minute (c.p.m.) in the presence of antigen
divided by the mean c.p.m. without antigen.
Statistical analysis
Proliferative responses to antigens were compared
between patients and healthy controls using the Mann–
Whitney test, as a normal distribution for the T-cell
response to antigen was not assumed. Differences in the
proportions of responders in different groups were
assessed by calculating x2.
Results
Rheumatoid arthritis patients are poor responders to
M.hsp60
Figure 1a shows the stimulation index obtained for the
proliferative response to M.hsp60 by PBMC from each
healthy control and early arthritis patient. The median
response for healthy controls was 6.8, compared to 1.2
for RA patients (P < 0.001, Mann–Whitney). Hence,
there was a significantly lower proliferative response to
M.hsp60 from T cells obtained from this cohort of RA
patients compared with those from healthy controls.
Furthermore, if a SI 3 was taken as a criterion for
a positive response, most of the RA patients failed to
respond: only 15/58 were responders, as compared to
21/26 healthy controls ( Table 1), a significantly lower
proportion (P = 0.009, x2).
Rheumatoid arthritis patients can make normal
proliferative responses to recall antigens
To address the possibility that the lower response seen
to M.hsp60 was due to a general inability of RA T cells
to respond to any antigens, responses to the recall
antigens PPD and tetanus toxoid were measured.
Responses to PPD ( Fig. 1b) were not significantly
different (P = 0.1033, Mann–Whitney) between the two
groups. The median PBMC response for healthy controls was 59 compared with 35 for RA patients.
Furthermore, the vast majority of individuals could be
classified as responders—all the healthy controls and
55/58 of the RA patients (P = 0.832, x2) ( Table 1).
Similarly, proliferative responses to tetanus toxoid
by the two groups were not significantly different
(P = 0.1485, Mann–Whitney). The median response for
PBMC proliferative responses to M.hsp60
(a)
(b)
(c)
(d)
633
F. 1. Healthy controls’ and RA patients’ PBMC proliferative responses to (a) M.hsp60, (b) PPD, (c) tetanus toxoid and (d)
GroEL. Each point on these graphs represents a response by each individual to the antigen tested, expressed as a stimulation
index. The median responses for each group are also indicated on the figures.
T 1. Number of healthy controls and RA patients whose PBMC
proliferate to the antigens M.hsp60, GroEL, PPD and tetanus toxoid
Number of responders (SI 3)
Antigens
RA patients
Healthy controls
x2 P
15/58
14/56
51/58
31/46
21/26
16/26
26/26
11/11
0.009
0.06
0.832
0.568
M.hsp60
GroEL
PPD
Tetanus toxoid
SI, stimulation index.
healthy controls was 20 compared with a median
response of 14 for RA patients ( Fig. 1c).
Rheumatoid arthritis patients are poor responders to
GroEL
To investigate whether the poor PBMC proliferative
response to M.hsp60 applied only to this bacterial hsp
or whether PBMC from RA patients also had low
proliferative responses to other heat shock proteins,
responses to the E.coli hsp60 protein, GroEL, were
measured. As shown in Fig. 1d, the median response for
RA patients (1.75) was again significantly below that
of the healthy controls (3.5) (P = 0.0023). As observed
for responses to M.hsp60, there was also a lower proportion of responders to GroEL in the RA group than in
the healthy controls ( Table 1), although this difference
failed to reach significance (P = 0.06, x2)
Low responses to M.hsp60 are not associated with
expression of HLA-DR4
To establish whether the high percentage of HLA-DR4+
individuals (58%) in the RA patient group might have
accounted for the low responses to M.hsp60, RA
patients were divided into those who expressed DR4
and those who did not. No significant difference between
634
J. M. Ramage and J. S. H. Gaston
F. 2. HLA-DR4+ and HLA-DR4 − RA patients’ PBMC
proliferative responses to M.hsp60. This figure utilizes the data
for RA patients represented in Fig. 1. RA patients are divided
into those who express the MHC class II allele DR4 and those
who do not. Each point represents the SI of an individual
experiment. The median response for the two groups is indicated. The median response for all the RA patients was
1.2 (Fig. 1).
in vitro PBMC responses to M.hsp60 by these two
groups was observed ( Fig. 2). The median response for
each group was similar: 1.3 for DR4+ individuals
compared with 1.2 for DR4 − individuals.
Influence of clinical characteristics on low response to
M.hsp60
The possible influence of the clinical characteristics of
the RA patients was also analysed. Patients were subdivided into further groups: (I ) those with evidence of
active inflammation ( ESR > 20 and CRP > 10); (II )
those seropositive for rheumatoid factor (titre >
60 U/ml ); (III ) those patients who were thought by
their rheumatologists to require disease-modifying drugs
(sulphasalazine, gold, methotrexate or penicillamine).
Mann–Whitney tests were performed to determine
whether any difference in response could be detected
within these RA patient subgroups, but no significant
differences were obtained using any of these criteria.
Discussion
This study has demonstrated that a cohort of RA
patients, with relatively early disease, had significantly
lower in vitro PBMC proliferative responses to the
bacterial heat shock proteins, M.hsp60 and GroEL,
compared to healthy controls. Moreover, PBMC from
a large percentage of these RA patients failed to respond
significantly in vitro to either hsp60 (SI < 3) at day 6.
These results are in agreement with those obtained by
Lai et al. [16 ] who also showed depressed responses to
M.hsp60 by PBMC from Chinese RA patients, compared to those of healthy controls. The lack of response
could not be accounted for by the high percentage of
HLA-DR4+ individuals in the study, as no significant
difference in proliferation was observed between DR4+
and DR4− RA patients. Although most RA patients
failed to respond to M.hsp60, some responder patients
were seen, showing that RA is not inevitably associated
with an inability to recognize M.hsp60. Moreover, in
the study by Lai et al. [16 ], inhibition of proliferative
responses to M.hsp60 was achieved using a polyclonal
antiserum against DRB1*0405, an allele associated with
RA in Chinese patients. This result suggests that HLADR4-restricted recognition of M.hsp60 can occur in RA
patients. Likewise Mustafa et al. [17, 18] generated
M.hsp60-specific DR4-restricted T-cell clones from DR4
homozygous individuals, but unusually these individuals
had received vaccination against M. leprae, and some
of the clones developed were M. leprae specific rather
than cross-reactive with Mycobacterium tuberculosis or
Mycobacterium bovis BCG. Therefore, there is some
evidence that DR4-restricted recognition of M.hsp60
can occur, and our findings cannot simply be explained
as being due to a lack of DR4-binding peptide epitopes
in M.hsp60—indeed, application of the algorithms
devised to predict such epitopes suggests that there are
several candidates.
A direct comparison of PBMC proliferative responses
to M.hsp60 by RA patients and healthy controls has
previously been reported by Burmester et al. [19] who
showed low responses not only in RA, but also in
normal controls. This result reflects the very limited
number of individuals in either group in this study who
responded significantly to the antigen (5/63). This is
surprising in view of the results in the present study,
and in other previous work, where the majority of
healthy individuals have been shown to respond to
M.hsp60.
The limited proliferation to M.hsp60 by lymphocytes
from RA patients, reported by other investigators,
has been used as evidence against a role for M.hsp60specific T cells in RA, working on the hypothesis
that M.hsp60-specific T cells might be arthritogenic.
However, it is of interest that PBMC from patients with
another inflammatory condition, Crohn’s disease, were
also shown to have lower proliferative responses to
M.hsp60 compared with those obtained from controls
[20]. This allows consideration of the second hypothesis
which emerges from the rodent studies, i.e. that inflammatory conditions such as RA or Crohn’s disease might
be associated with defective regulatory responses to
hsp60. In the adjuvant arthritis model, the regulatory
T cells had a low affinity for self-hsp60 and were
generated by immunization with M.hsp60 or a
M.hsp60-derived peptide [21]. If the induction of such
a regulatory self-hsp60-specific T-cell population in
humans also requires a satisfactory response to M.hsp60
or other bacterial hsp60, then a limited response by RA
patients to bacterial hsp60 might result in a failure to
generate the regulatory population. We have recently
demonstrated that healthy individuals possess T cells
able to respond to self-hsp60 [22]. These cells are within
PBMC proliferative responses to M.hsp60
the CD45RA+RO− subset of CD4+ T cells, usually
regarded as naive cells. Their activation in vivo may
require challenge with bacterial hsp60 which bear epitopes conserved in self-hsp60. During inflammatory
responses such as those which are induced by bacterial
infection, such self-reactive cells may be employed to
control the response through the recognition of selfhsp60 whose expression is commonly increased at sites
of inflammation. One hypothesis which is worth considering would be that a defect in this regulatory mechanism could result in an inability to downregulate
immune responses in the joint appropriately, irrespective
of the immune response responsible for triggering the
arthritis (which might or might not involve hsp). This
defect would then be associated with a chronic inflammatory response, chronicity being a defining characteristic of RA.
Acknowledgements
We thank Professor Paul Emery and staff at the Selly
Oak Hospital, Birmingham, for their assistance in
obtaining patient blood samples, and Jeff Faint for
performing the tissue typing. This work was supported
by the Medical Research Council.
9.
10.
11.
12.
13.
14.
15.
16.
References
1. Pearson CM. Development of arthritis, periarthritis and
periostitis in rats given adjuvant. Proc Soc Exp Biol Med
1956;91:95–101.
2. Holoshitz J, Naparstek V, Ben-Nun A, Cohen IR. Lines
of T lymphocytes induce or vaccinate against autoimmune
arthritis. Science 1983;219:56–8.
3. van Eden W, Holoshitz J, Nevo Z, Frenkel A, Klajman
A, Cohen IR. Arthritis induced by a T-lymphocyte clone
that responds to Mycobacterium tuberculosis and to
cartilage proteoglycans. Proc Natl Acad Sci USA
1985;82:5117–20.
4. van Eden W, Thole JER, van der Zee R et al. Cloning of
mycobacterial epitope recognised by T lymphocytes in
adjuvant arthritis. Nature 1988;331:171–3.
5. Holoshitz J, Matitau A, Cohen IR. Arthritis induced in
rats by cloned T lymphocytes responsive to mycobacteria
but not to collagen type II. J Clin Invest 1984;73:211–5.
6. van den Broek M, Hogervorst EJM, van Bruggen MCJ,
van Eden W, van Der Zee R, van Den Berg WB. Protection
against streptococcal cell wall-induced arthritis by pretreatment with the 65-kD mycobacterial heat shock protein. J Exp Med 1989;170:449–66.
7. Thompson SJ, Rook GAW, van der Zee R, Elson
CJ. Autoimmune reactions to heat-shock proteins in pristane-induced arthritis. Eur J Immunol 1990;20:2479–84.
8. Billingham MEJ, Carney S, Butler R, Colston MJ. A
mycobacterial 65-kD heat shock protein induces antigen-
17.
18.
19.
20.
21.
22.
635
specific suppression of adjuvant arthritis, but is not itself
arthritogenic. J Exp Med 1990;171:339–44.
Anderton SM, van der Zee R, Prakken B, Noordzij A,
van Eden W. Activation of T cells recognising self 60-KD
heat shock protein can protect against experimental arthritis. J Exp Med 1995;181:943–52.
Res PC, Schaar CG, Breedveld FC et al. Synovial fluid
T cell reactivity against 65kD heat shock protein of mycobacteria in early chronic arthritis. Lancet 1988;ii:478–80.
Gaston JSH, Life PF, Bailey LC, Bacon PA. In vitro
responses to a 65-kD mycobacterial protein by synovial
T cells from inflammatory arthritis patients. J Immunol
1989;143:2494–500.
Pope RM, Wallis RS, Sailer D, Buchanan TM, Pahlavani
MA. T cell activation by mycobacterial antigens in
inflammatory synovitis. Cell Immunol 1991;13:95–108.
Res PC, Telgt D, van Laar JM, Pool MO, Breedveld FC,
de Vries RRP. High antigen reactivity in mononuclear
cells from sites of chronic inflammation. Lancet
1990;336:1406–8.
Gaston J, Young J, Ramsey S, Scott D. Specificity of
synovial fluid T cells in iatrogenic human adjuvant arthritis. Arthritis Rheum 1996;39:S149.
Gough A, Faint J, Salmon M et al. Genetic typing of
patients with inflammatory arthritis at presentation can
be used to predict outcome. Arthritis Rheum
1994;37:1166–70.
Lai NS, Lan JL, Yu CL, Lin RH. A dominant human
leucocyte antigen DR4- Dw15 restricted mycobacterial
65kDa antigen-specific T-cell immunity in Chinese
patients with rheumatoid arthritis. Br J Rheumatol
1995;34:598–606.
Mustafa AS, Lundin KEA, Meloen RH, Shnnick TM,
Coulson AFW, Oftung F. HLA-DR4-restricted T-cell
epitopes from mycobacterial 60 000MW heat shock protein 60 (hsp60) do not map to the sequence homology
regions with the human hsp60. Immunology
1996;87:421–7.
Mustafa AS, Lundin KEA, Oftung F. Human T cells
recognize mycobacterial heat shock proteins in the context
of multiple HLA-DR molecules: studies with healthy
subjects vaccinated with Mycobacterium bovis BCG and
Mycobacterium leprae. Infect Immun 1993;62:5294–301.
Burmester GR, Altstidl U, Kalden JR, Emmrich F.
Stimulatory response towards the 65kDa heat shock protein and other mycobacteral antigens with rheumatoid
arthritis. J Rheumatol 1991;18:171–6.
Peetermans WE. Expression of and immune response to
heat shock protein 65 in Crohn’s disease. In: Stress proteins
in medicine. New York: Marcel Dekker Inc.,
1996:197–211.
van Eden W, Anderton SM, van der Zee R, Prakken BJ,
Broeren CPM, Wauben MHM. (Altered) self peptides
and the regulation of self reactivity in the peripheral T cell
pool. Immunol Rev 1996;149:55–73.
Ramage JM, Young JL, Goodall JC, Gaston JSH. T cell
responses to heat shock protein 60: differential responses
by CD4+ T cell subsets according to their expression of
CD45 isotypes. J Immunol 1999;162:704–10.