|
|
||||||||




* Division of Rheumatology, Toronto Western Hospital, University Health Network, and Toronto Western Research Institute, Toronto, Ontario, Canada;
Departments of Immunology and Medicine and
Institute of Medical Sciences, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada; and
Aventis Pasteur Limited, Connaught Campus, Toronto, Ontario, Canada
| Abstract |
|---|
|
|
|---|
1 domain of the MHC class I
molecule. The generation of CTL with specificity for B27 under these
conditions demonstrates that tolerance to B27 can be subverted by
Chlamydia. This indicates a dynamic interrelationship
between the pathogen and B27, which may have important implications for
B27-related spondyloarthropathies triggered by intracellular
bacteria. | Introduction |
|---|
|
|
|---|
It is known that Ag derived from intracellular bacteria may lead to activation of CD8+ lymphocytes via class I-mediated pathways (7), thereby invoking CD8+ CTL in a central role in resistance to these bacteria (8). CTL recognizing bacteria-derived peptides presented by self B27 are relevant in host defense (9, 10). Such a B27-dependent event might figure critically in the clearance of an arthritogenic agent. Alternatively, the CTL response might be directed to a self peptide, thereby initiating an autoimmune process (11). Presentation by HLA class I molecules of closely related peptides from B27 and bacteria could lead to T cell responses to self B27 or inactivation of antibacterial T lymphocytes (12). The peptide specificity of B27 also seems to figure critically in the spontaneous arthritis seen in certain lines of B27 transgenic (Tg) rats (13). Although the role of CTL in recognizing MHC class I-restricted Chlamydia peptides has been studied in both clinical and experimental systems (14, 15, 16, 17), these studies have generally not sought to address the interrelationship with concurrent B27-related immune responses. Kuon et al. (18) recently identified B27-restricted peptides from the Chlamydia proteome that were demonstrated specificity for CTL derived from B27-Tg mice. Addressing this issue in the current study, we used a Tg rat model involving both Chlamydia and B27 to study the influence that host genetic background may have on the process of generating anti-Chlamydia and anti-B27 CTL.
| Materials and Methods |
|---|
|
|
|---|
The Tg rat line 21-4L, bearing six copies each of the B*2705 and
human
2-microglobulin
(h
2m) on the inbred Lewis (LEW) background
(RT1l), was a kind gift of J. Taurog (University
of Texas Southwestern Medical Center, Dallas, TX)
(19). Non-Tg inbred LEW rats, 23 mo old, were purchased
(Harlan Sprague Dawley, Indianapolis, IN). Rats were maintained in
microisolators, and screened by FACS of their PBMC to confirm
appropriate MHC expression.
Antibodies
For flow cytometry, mAbs were derived from different hybridomas
(American Type Culture Collection (ATCC), Manassas, VA): HB24
(anti-H-2Dk), HB95 (W6/32, reacting to all
HLA-I in association with h
2m), HB119 (ME-1,
recognizing B27 as well as B7 and Bw22), and HB82 (BB7.2, anti-HLA
A2). The mAbs purified from hybridoma W3/25 (MCA55G, specific to rat
CD4) and MRC OX-8 (MCA48G, rat CD8 specific) (Serotec, Raleigh, NC)
were used for the in vitro blocking of the appropriate T subsets.
Isolation of genes
Full-length A2 and B27 cDNAs were generated by standard RT-PCR using mRNA isolated from splenocytes of Tg mice. To amplify the A2 and B27 genes, PCR was conducted, as previously described (20).
Plasmid DNA immunogens
A2 and B27 PCR products were cloned into the expression vector pcDNA3 (Invitrogen, San Diego, CA). Two eukaryotic expression vectors were constructed: pcDNA3-A2 and pcDNA3-B27 (18). The vector constructs, encoding class I HLA, were characterized by restriction enzyme analysis and confirmed by sequencing with T7 polymerase (Pharmacia, Piscataway, NJ). The completed plasmid DNA was amplified in the JM109 Escherichia coli and purified using EndoFree Megaprep Kit (Qiagen, Chatsworth, CA). In vitro cell surface expression of A2 and B27 proteins was confirmed by COS-7 cell transfections, according to previously described methods (21).
Immunization of rats
Groups of four B27-Tg rats were immunized with each of the following immunogens: B27-Tg splenocytes, non-Tg splenocytes, B27 DNA, A2 DNA, and vector DNA. The splenocytes were irradiated with 2000 rad, and 2 x 108 cells/animal, in 1.5 ml PBS, were i.p. injected. A total of 200 µg of each lyophilized closed circular plasmid construct DNA in 150 µl PBS was injected into the anterior tibialis muscle of a single hind leg of a rat anesthetized with Forane (Zenoca Pharma, Mississauga, Ontario, Canada). The rats were immunized at days 0 and 21, and subsequently boosted at day 42. Spleens were harvested for CTL activity measurements 34 wk postboost.
Mixed lymphocyte culture
Spleen cells harvested from individual immunized rats were cocultured in vitro for 5 days with stimulator splenocytes from naive B27-Tg rats at 5:1 responder/stimulator ratio. In some experiments, stimulators (12 x 107) were treated with 107 of Chlamydia trachomatis (Ct; LGV type II, strain 434) elementary bodies (EB) (Microbix Biosystems, Toronto, Ontario, Canada) in 1 ml of complete medium for 4 h at 37°C and under conditions of 5% CO2/95% air. Finally, stimulators were irradiated with 2000 rad.
CTL 51Cr release assay
A total of 106107 target cells was resuspended in 1 ml of complete medium containing 100 µCi Na2[51Cr]O4 (Amersham Pharmacia Biotech U.K. Limited, Little Chalfont, U.K.), and incubated under 5% CO2/95% air for 1.5 h at 37°C with occasional agitations (once in 5 min). Then the cells were washed out and added into 96-well round-bottom microtiter plates (Falcon 3077; BD Biosciences, Lincoln Park, NJ) (104 cells/well in 100 µl medium). A total of 100 µl/well of serial dilutions of in vitro restimulated effectors was also added. Spontaneous 51Cr release was measured in wells with targets and nonimmune spleen cells. Plates were centrifuged at 100 x g for 5 min, incubated for 6 h at 37°C under 5% CO2/95% air, and then recentrifuged at 1000 x g for 10 min. Supernatants were harvested, and radioactivity was measured using a gamma counter (Canberra Packard, Mississauga, Ontario, Canada). Percent specific lysis was calculated using the mean of triplicate samples, as follows: ((experimental 51Cr release cpm - spontaneous 51Cr release cpm)/(total 51Cr release cpm - spontaneous 51Cr release cpm)) x 100%. All experimental assessments were repeated from two to five times with different rats as donors and recipients of immune spleen cells. The data shown in the figures are representatives from three to five separate experiments. The spontaneous to total 51Cr release ratio was routinely <20% and averaged <14% at any E:T ratio. SD was always <5% specific lysis. The spontaneous release of 51Cr of targets incubated alone did not differ from wells in which targets were incubated with nonimmune spleen cells.
Targets
Synovial fibroblast lines were generated in non-Tg and B27-Tg
rats using induced synovitis, as we described (22). The
murine L cell line J26 (ATCC) transfected with
h
2m was previously cloned. Its clone J26.6 was
transfected with genomic DNA encoding B14, B27, B40, B44, or mutant
B27.C67Y (mutating Cys67 into Tyr), as we also
described (23, 24). Mouse L cells,
LTK-, transfected with
h
2m and B7, mutant B27.K70Q (the
Lys70 into Gln mutation), or mutant
B27.AKA6971TNT (the mutation of
Ala69Lys70Ala71
into ThrAsnThr), were a kind gift from J. Taurog (25).
B14, B40, and B44 were selected as controls having no known
relationship to SpA. The cell lines are checked periodically by FACS to
ensure stable expression of the transgenes. In some experiments,
targets were incubated with Ct (107 EB/ml) for
23 h at 37°C in humidified 95% air/5% CO2
and assaying, as described above.
Blocking of T lymphocyte subsets in vitro
Anti-CD4 or anti-CD8 mAbs at a concentration of 1 µg/ml were added to restimulated splenocytes at 107/ml, and maintained in complete medium for 1 h at +4°C.
| Results |
|---|
|
|
|---|
Irradiated naive B27+ Tg LEW spleen cells
were injected i.p. into syngeneic rats. The animals were boosted as
described, and the primed splenocytes were restimulated in vitro with
the Tg cells used for the in vivo immunization. In contrast to our
previous studies using non-Tg rats, B27+ Tg
animals were tolerant of immunization and restimulation with the self
B27+ cells, as reflected in the absence of any
CTL killing of B27+ targets. The T cell tolerance
was also observed when a plasmid DNA construct encoding full-length B27
was used, instead of the splenocytes, for in vivo immunization via an
i.m. route (Fig. 1
, A and
C). The figure is representative of three to five
experiments. We had previously demonstrated that both the plasmid DNA
and cellular strategies were effective in generating CTL against B27 in
non-Tg LEW rats (26).
|
|
We next addressed the specificity of CTL obtained from B27-Tg rats
primed in vivo with either syngeneic cells or B27 DNA and thereafter
restimulated in vitro with Chlamydia-treated B27-Tg
splenocytes. As expected, killing was observed, with recognition of
both Chlamydia-sensitized B27+ rat
fibroblasts (Fig. 1
A) and murine L cells (Fig. 1
C). When cytotoxicity against nonsensitized B27-Tg rat and
B27+ murine cell targets was examined, we
observed killing of all of the B27+ targets.
Killing was comparable, whether (Fig. 1
, B and D)
or not (Fig. 1
, A and C) the targets had first
been sensitized with Chlamydia. No cytotoxicity was found on
non-Tg LEW synovial fibroblasts and murine L cells transfected with
third party B44. Furthermore, no such killing was seen with splenocytes
from control animals immunized with HLA-A2, vector DNA alone, or non-Tg
LEW splenocytes (data not shown). These observations point
toward the existence of autoreactive anti-B27 CTL.
Inhibition of CTL specific to Chlamydia and self HLA-B27 by anti-CD8 mAb
We evaluated whether CD4+ or
CD8+ cells were responsible for the lysis of
Chlamydia+ and B27+
targets. Splenocytes from B27-Tg animals challenged in vivo with B27
DNA, then stimulated in vitro with Chlamydia-treated
B27+ cells were incubated with anti-CD4 or
anti-CD8 mAb. Lysis of non-Tg fibroblasts sensitized with
Chlamydia was decreased by 61% with the anti-CD8
treatment of effector cells. For the corresponding experiment using
B27-Tg cells in vivo, the anti-CD8 inhibition was 40%. In
contrast, the anti-CD4 treatment of the same cells from these two
sets of animals demonstrated no decrease in CTL activity (Table I
). Thus, this cytotoxicity can be
predominantly attributed to CD8+ effectors. Some
small degree of lysis of targets may be due to nonspecific or
cross-reactive killing by other activated spleen cells. Recognition of
either the B27+ targets alone or the
Chlamydia-sensitized targets was also significantly
inhibited by the anti-CD8 mAb, but not by anti-CD4 mAb,
implicating CD8+ effectors as key in these
recognition events (Table I
).
|
To determine whether the autoreactive anti-B27 CTL
recognition was specific for self B27, we used a panel of L cell
targets transfected with h
2m and HLA-B allele
controls other than B44. We observed no killing by the autoreactive CTL
of targets expressing B7, B14, or B40 (Fig. 3
, lanes 57). The effect of
CTL recognition of mutated amino acid residues in the hypervariable
region of the
1 domain of B27 was then
examined. Killing by the autoreactive CTL was still observed for
targets expressing B27.C67Y, to a degree comparable as native B27 (Fig. 3
, lanes 12). Thus, the autoreactive CTL did
not discriminate the Cys
Tyr mutation at position 67. However, these
CTL did not recognize targets expressing B27.AKA6971TNT
mutant molecules (Fig. 3
, lane 4). Indeed, the single
residue mutant B27.K70Q was not recognized, indicating that mutation of
B27 Lys
Gln at position 70 alone was sufficient to abrogate the
autoreactive B27-specific CTL recognition (Fig. 3
, lane
3).
|
| Discussion |
|---|
|
|
|---|
Rats Tg for B27 and h
2m have proved useful for
investigating the basis of B27-associated SpA. Tg rat lines expressing
high levels of B*2705/
2m develop a spontaneous
disease with striking resemblance to human SpA, and this disease is
critically dependent on B27-bound peptides (18). Tg rats
from lines that express lower copy numbers of B27 develop no
spontaneous disease. In the present study, we have used the 21-4L, one
such low-copy line on LEW background.
As was seen in the xenogeneic B27 response observed with our non-Tg animals (24), we have also observed in this study that prior exposure to self B27 results in a reduced threshold for generating a primary in vitro CTL response to Chlamydia. Unexpectedly, this precondition of in vitro exposure to Chlamydia led to induction of the autoreactive CTL with specificity for B27, in the animals otherwise tolerant to self B27, by a single exposure in vitro to Chlamydia following the in vivo priming with self B27. These data implicate a dynamic interrelationship between immune recognition of Chlamydia and breakdown of tolerance to self B27. This phenomenon might be explained by cross-reactive recognition of self and microbial peptides. CTL with either of these two specificities may play a pathogenic role in B27-associated SpA. The arthritogenic peptide model (27) for the pathogenesis of SpA postulates that the B27-binding motif sets the stage for presentation of peptides to potentially pathogenic CTL. These peptides might be of endogenous origin or exogenous origin or both, and cross-recognition of such peptides might be favored by B27-restricted T cells. This notion of potentially cross-reacting peptides is supported by the observation that B27-restricted CTL with specificity for both self and bacterial peptides can be found in synovial fluid of SpA patients (28). The notion of such cross-reactivity has been proposed to be a central factor in the pathogenesis of B27-related SpA (29)
The interrelationships of infection and autoimmunity are complex, and may invoke a variety of pathways in addition to molecular mimicry (12). It has been argued that mimicry may arise as a secondary phenomenon caused by alteration of host antigenic determinants through tissue injury and the creation of neoepitopes. This process of tissue injury uncovering cryptic epitopes does not seem pertinent to our present observation of in vitro stimulation. Activation of otherwise anergic autoreactive T cells can result from up-regulation of costimulatory molecules on APCs, being a necessary precondition for T cell activation in conjunction with cognate binding of the MHC-peptide complex by TCRs. It is evident that autoreactive T cells do survive thymic deletion strategies to some extent, as must be the case for our B27-Tg rats. Peripheral tolerance mechanisms (30) then become the means of keeping such cells in check and preventing an ongoing autoimmune assault on host tissues.
Our studies on the mutant B27 targets implicate Lys at position 70 as being critical for an epitope recognized by the autoreactive CTL. In the studies of Lopez de Castro and colleagues (31, 32), the residue Lys70 plays a key role in the specificity of CTL. It was of interest that in our study the Cys67Tyr mutation did not affect autoreactive CTL recognition of B27. A Cys67Ser mutation of B*2705 has been felt to still predispose Tg animals to arthritis in vivo (33), and to weaken B pocket interactions affecting alloreactivity in vitro (34). Studies in progress in our laboratory are addressing whether the epitope of xenoreactive anti-B27 CTL differs from that of autoreactive anti-B27 CTL.
Activation of quiescent autoreactive T cells might implicate a sequence homology between the respective peptides from Chlamydia and B27, but this may not necessarily be the precondition for cross-reacting T cell responses. Studies by Wucherpfennig and Strominger (35) demonstrate that T cell clones may show significant stimulatory responses to peptides that are widely divergent in their amino acid sequences. Tertiary structure, hydrophobicity, and electrostatic charges are all potential contributors to interactions with cross-reacting TCRs. It is of interest that the residue Lys70 does not point upward, in a way that would reflect direct contact with the TCR, but forms a salt bridge with residue Asp74, which in turn figures critically in peptide binding (28, 29). The elution studies by Jardetzky et al. (36) indicate that a significant population of peptides normally resident in the B27 peptide-binding cleft is derived from B27 itself. It remains a distinct possibility that the B27-mediated sequelae of Chlamydia infection such as reactive arthritis are fundamentally autoimmune processes. Antibiotic therapy does not alter the natural course of this arthritis (37), nor has recent experience with anti-TNF therapies unmasked any occult or subclinical infectious arthritis (38). Thus, if the persistent inflammation is maintained by a sustained source of microbial peptides, the source of such exogenous peptides is obscure. Our results provide experimental evidence of a break in tolerance to B27 occasioned by exposure to Chlamydia. This may provide a mechanistic bridge between what begins as a septic event, but which evolves into an autoimmune process. It also raises the possibility that peptide therapy may be a feasible approach to treating this condition in the future.
| Acknowledgments |
|---|
| Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Robert D. Inman at the current address: The Arthritis Center of Excellence, Toronto Western Hospital, Fp 1-222, 399 Bathurst Street, Toronto, Ontario, Canada. M5T 2S8. E-mail address: robert.inman{at}uhn.on.ca ![]()
3 Abbreviations used in this paper: SpA, spondyloarthropathy; Ct, C. trachomatis; EB, elementary body; h
2m, human
2-microglobulin; LEW, Lewis; Tg, transgenic. ![]()
Received for publication March 6, 2002. Accepted for publication July 23, 2002.
| References |
|---|
|
|
|---|
2 microglobulin transgenic rats. J. Clin. Invest. 98:945.[Medline]
Interferon production by cytotoxic T lymphocytes is required for resolution of Chlamydia trachomatis infection. Infect. Immun. 66:5457.
2m: an animal model of HLA-B27-associated human disorders. Cell 63:1099.[Medline]
1-globin gene in a new SV40 host-vector system. Cell 27:279.[Medline]
This article has been cited by other articles:
![]() |
J. J. Cragnolini, N. Garcia-Medel, and J. A. Lopez de Castro Endogenous Processing and Presentation of T-cell Epitopes from Chlamydia trachomatis with Relevance in HLA-B27-associated Reactive Arthritis Mol. Cell. Proteomics, August 1, 2009; 8(8): 1850 - 1859. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. J. Cragnolini and J. A. L. de Castro Identification of Endogenously Presented Peptides from Chlamydia trachomatis with High Homology to Human Proteins and to a Natural Self-ligand of HLA-B27 Mol. Cell. Proteomics, January 1, 2008; 7(1): 170 - 180. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |