Recognition of the 60 kilodalton cysteine-rich outer membrane protein OMP2 by CD4⁺ T cells from humans infected with Chlamydia trachomatis

Abstract

T cell-mediated immunity is important in the control of chlamydia infection but chlamydia-specific T cells are also implicated in the inflammation and tissue damage which characterize chlamydia associated diseases. To investigate target antigens of the T cell-mediated immune response to chlamydia infection, Chlamydia trachomatis-specific CD4⁺ T cell clones were isolated from a patient with chlamydia-induced reactive arthritis. T cell immunoblotting indicated that an antigen of ∼60 kilodaltons molecular mass was recognized, and recombinant 60 kilodalton cysteine-rich outer membrane 2 (OMP2) proved to be stimulatory. By using deletion constructs and synthetic peptides an epitope presented by HLA-DRB1*0401 was defined and proved to contain the nonamer peptide within the OMP2 sequence predicted to have the greatest binding affinity for DRB1*0401 The sequence of the epitope is conserved in all C. trachomatis strains but not in C. pneumoniae. Investigation of patients with acute urethritis and additional patients with sexually acquired reactive arthritis showed that OMP2-reactive T cells were readily detectable in peripheral blood and synovial fluid. Thus OMP2 is a target antigen of the T cell-mediated immune response to CT infection.

Introduction

Effective clearance of infection by Chlamydia trachomatis requires a T cell-mediated immune response and protection from re-infection also depends on cellular immunity [1–4]. Animal models have suggested that CD4⁺ T cells which produce interferon-γ are particularly important effector cells [5–7]. However, T cell-mediated immunity can be both beneficial and pathological, since inflammatory states which complicate chlamydia infection, such as trachoma, salpingitis and reactive arthritis, are also thought of as delayed-type hypersensitivity responses to bacterial antigens [8–10]. We suggest that T cell responses in subjects with uncomplicated chlamydia infection may differ from those with inflammatory sequelae. To explore the mechanisms of chlamydia-induced inflammation and to identify potential vaccine candidates our aim has been to identify targets of the human T cell responses to chlamydia infection.

To date only a few T cell stimulatory antigens have been identified. These include the major outer membrane protein (MOMP), heat shock protein 60 (hsp60), and the histone-like protein, Hc1 [11–15]. CT Hsp60 and Hc1 were identified by performing T cell immunoblotting with human CD4⁺ CT-specific T cell clones, and then testing recombinant antigens of an appropriate molecular weight. In humans immune responses to MOMP were demonstrated in PBMC from infected individuals by priming in vitro with purified MOMP to produce T cell lines. These were subsequently used to identify a number of epitopes, most of which were in portions of the molecule whose amino acid sequence is conserved in different serovars.

In this report we have cloned T cells responsive to CT elementary bodies (EBs), and then applied a T cell immunoblotting approach which allowed us to identify human CD4⁺ T cell clones which recognize the 60 kilodalton cysteine-rich outer membrane protein, OMP2. This protein is the second most abundant constituent of CT EBs. To confirm that OMP2 is the stimulatory antigen we have identified a stimulatory OMP2 peptide and examined potential cross-reactivity of such clones. To determine if OMP2 is able to stimulate proliferative resposnes in the polyclonal populations we examined responses of PBMC from infected individuals to the purified OMP2 protein.

Patients and methods

CT-reactive T cells were cloned from synovial fluid mononuclear cells (SFMC) of a patient with sexually acquired reactive arthritis as previously described (see [12]). Diagnosis of reactive arthritis was based on typical clinical findings with evidence of preceding genitourinary infection with CT. SFMC were also obtained from three patients with classical Reiter's syndrome (urethritis, conjunctivitis and arthritis) and one with reactive arthritis, all of whom had marked SFMC responses to CT elementary bodies (EBs); these were tested for their ability to recognize recombinant CT antigens. As a control a patient with reactive arthritis secondary to gastrointestinal infection with Yersinia enterocolitica was also tested.

In addition peripheral blood mononuclear cells (PBMC) were obtained from 26 patients with acute urethritis attending a genitourinary medicine clinic. PBMC were obtained at first attendence and before treatment with antibiotics, and screened for their ability to respond to CT EBs, recombinant CT antigens and control recall antigens. No details on whether patients were previously infected by CT were available, but all presented with new episodes of symptomatic urethritis.

Bacterial strains

The following Chlamydia trachomatis strains were used: E/DK20/ON (trachoma biovar: strain DK20) and L2/434/BU (lymphogranuloma venereum biovar: strain 434) were used. Growth and purification of EBs was performed as described elsewhere [16]. Organisms were not inactivated prior to use.

Generation of CT–specific T cell clones

SFMC were isolated obtained from heparinized, hyaluronidase-treated SF by centrifugation on Ficoll-Paque, washed in PBS and incubated for 6 days with 2 × 10⁸/ml purified DK20 CT EBs in complete medium (RPMI 1640 (gibco BRL, Paisley, Scotland, UK), 5% human serum, 25 mm HEPES, 2 mm glutamine 100 U/ml penicillin and 100 µg/ml streptomycin sulphate). The responding T cells were cloned by limiting dilution in 20 µl Teraski wells using 10⁶/ml irradiated autologous PBMC as antigen presenting cells, 100 U/ml recombinant IL-2 (Chiron, Harefield, Middlesex, UK) and CT EBs. Resulting clones were expanded at 21-day intervals by stimulation with irradiated (3 Gy) allogeneic PBMC, IL-2 (10 U/ml) and PHA (1 µg/ml), as described previously [12]. Specificity of the clones, tested at least 10 days after stimulation with PHA, was determined by measuring proliferation ([³H]-thymidine incorporation) by 2 × 10⁴ cloned T cells co-cultured for 3 days with 5 × 10⁴ autologous irradiated PBMC and EBs.

T cell immunoblotting

This procedure has been described in detail previously [11,17,18]. Briefly, 5 × 10⁸ chlamydia EBs were solubilized in 300 µl SDS-PAGE sample buffer, loaded to a gel and separated by SDS-PAGE. Proteins were electroblotted onto a 0·45 µ nitrocellulose sheet (100 V for 90 min) which was then sliced horizontally into 2-mm strips so that each strip contained chlamydia proteins of a particular molecular weight, estimated by using a standard curve run on the same gel. These strips were solubilized in 250 µl DMSO and fine particles of nitrocellulose bearing CT antigens precipitated by slow addition of 0·1 m sodium bicarbonate. Following extensive washing in sterile PBS, the preparations were resuspended in 1 ml PBS per cm of nitrocellulose and stored at −20°C. A 1 : 10 dilution of these preparations was added to T cell clone proliferation assays.

PBMC proliferation assays

PBMC or SFMC were cultured for 6 days in RPMI medium at a final concentration of 1 × 10⁵ cells per well in 200 µl (Falcon; Becton Dickinson, Franklin Lakes, NJ, USA) flat wells, with recall antigens (PPD and tetanus toxin, both obtained from Statens Seruminstitut, Denmark) at previously determined optimal concentrations (10 µg/ml). Proliferation was measured by the incorporation of (methyl-[³H]) thymidine ([³H]-thymidine) (Amersham UK) over the final 14 h of culture. For all proliferation assessments 1 µCi/well of [³H]-thymidine was used and samples were harvested onto printed glass fibre filter mats (Wallac, Turku, Finland) using a cell harvester (Skatron, Oslo, Norway). The filter mats were treated with scintillation fluid (OptiPhase™ HiSafe2, Wallac) and [³H]-thymidine incorporation measured using a β plate counter (1205 Betaplate™, Wallac).

Expression and purification of recombinant CT OMP2 and fusion proteins

DNA encoding the full length OMP-2 sequence (a kind gift from Dr I. Clarke, University of Southampton, UK [19]) was inserted into the appropriate sites of the expression vector, pQe60 (Qiagen, Crawley, UK), to enable expression of CT OMP-2 with a C-terminal polyhistidine motif and subsequent purification using nickel affinity chromatography. Briefly, overnight cultures were diluted 1 : 10 into 250 ml of l-broth medium containing 100 µg/ml ampicillin and grown at 37°C until the O.D. was between 0·3 and 0·6. Protein expression was induced with 2 mm IPTG for 3 h. Following centrifugation of the bacteria, the pellet was resuspended in PBS and the bacteria lysed by sonication. The inclusion bodies were obtained by centrifugation at 1000 g and resuspended in PBS containing 8 m urea and 10 mm imidazole. The protein was purified by passage of the Escherichia coli lysate through a nickel affinity column (Hi trap, Pharmacia, St Albans, Herts, UK). Unbound proteins were removed by washing with PBS containing 40 mm imidazole and the proteins bound specifically to the nickel sepharose were eluted by the addition of 300 mm imidazole. The eluate was dialysed in PBS and concentrated by centrifugation through a 30-kDa cut-off filter (Vivaspin, Binbrook, Lincoln, NE, USA). Purity of the protein was assessed using SDS-PAGE.

OMP-2 deletion mutants were prepared as follows: 1 ng of DNA encoding full length OMP2 serovar B was used as the PCR template using the following oligonucleotides; P1 5′gtccatggtggcgagtttatttgctagc3′, P2 5′gtccatggttcctaaatatgctacg3′, P3 5′gtccatggcttgtttgcgttgccca3′, P4 5′tggatccgatgtgtgtattctctgtatc3′, P5 5′gtccatggcaggagcagattggtcttat3′ P5a 5′gtccatggatacttgtgaccctgtttgt3′ with the OMP2 reverse primer 5′atggatccgatgtgtgtattctctgtatc3′. Amplification was performed in a total volume of 50 µl containing 10 mm Tris-Hcl pH 9·0, 1·5 mm MgCl₂, 200 µm of each dNTP and 2 units of Expand DNA polymerase mix (Roche, Lewes, East Sussex, UK). PCR amplification parameters were as follows; 94°C for 1 min, 60°C for 1 min, 72°C for 1 min 30 s for a total of 20 cycles. The resultant PCR products were digested with Nco I and Bgl II, purified and ligated in frame into the E. coli expression vector pQE60 predigested with Nco I and Bgl II. E. coli M15 pRep4 cells were transformed with the ligation mix and E. coli recombinants expressing OMP2 deletion mutants were selected. DNA sequencing was performed on all OMP2 expression DNA constructs to ensure no mutations had occurred during the PCR amplification steps. The recombinant polypeptides were obtained as described above, and lysates checked to confirm expression of the appropriate fragments by SDS-PAGE. An identical protocol was used to produce purified recombinant chlamydia hsp60.

Results

OMP2 is a target of the CD4⁺ T cell responses to CT

CT-specific CD4⁺ T cell clones were obtained from synovial fluid of a patient (AH) with CT-induced reactive arthritis. To identify the antigen recognized, we performed T cell immunoblotting experiments with T cell clone AH1·37. TCR analysis confirmed that AH1·37 was clonal and the receptor was cloned and sequenced (TCR AV1S3/AJ16; BV8S2/BJ2S3, full sequence available from the authors). Despite being clonal, proliferative responses were detected to several different molecular weight fractions (Fig. 1). We assumed that the T cell clone recognized an antigen within the stimulatory fraction with the highest molecular weight (∼60 kDa) and also responded to fragments of this antigen which preserved the T cell epitope. Accordingly we tested the clone's ability to recognize a recombinant major CT antigen of 60 kilodalton, the cysteine-rich outer membrane protein; the clone responded strongly to this recombinant antigen and not to controls (Fig. 2). The response to 10 µg/ml is shown, but titration of the response showed a plateau in response at concentrations between 1 and 10 µg/ml. Subsequent experiments identified two additional OMP2-specific clones from among the population of CT-specific clones derived from this patient. The patient from whom the clones were derived was typed for HLA-DR and shown to be express DRB*0401 and DRB*0101. Therefore an experiment using DRB*0401⁺ or DRB*0101⁺ antigen presenting cells (APC) was carried out; a response to CT EBs was only seen with DRB*0401⁺ cells. Subsequent work used APC from several DRB*0401⁺ donors, all of which presented OMP2 (or specific peptide, see below) to AH1·37 (e.g. response using allogeneic APC matched only for DRB*0401: T +APC 270 c.p.m.; T +APC and 1·6 µg/ml OMP2 12 081 c.p.m., T +APC + CT EBs (10⁷/ml) 9905 c.p.m.)

Fig. 1.

Proliferative responses ([³H]-thymidine incorporation) of T cell clone AH1·37 to chlamydia antigens fractionated by SDS-PAGE (final dilution of 1 : 10 for each solubilized fraction). Molecular masses of stimulatory fractions are shown. Responses to whole CT EBs (2·5 × 10⁷/ml) are also shown (LGV) and the response to antigen presenting cells in the absence of antigen (media).

Fig. 2.

Proliferative responses ([³H]-thymidine incorporation) of T cell clone AH1·37 to recombinant CT OMP2 and to a lysate of E. coli which does not express a recombinant protein (control). Both were tested at 10 µg/ml. Responses are also shown to whole CT EBs (10⁷/ml) from two biovars, LGV: strain 434 (LGV) and trachoma: strain DK20 (DK20), and to antigen-presenting cells in the absence of antigen (T + APC).

Identification of a T cell epitope in CT OMP2

To map the epitope recognized by AH1·37, a series of recombinant proteins was generated which had successively larger deletions of the OMP2 sequence. As shown in Fig. 3a, AH1·37 recognized full length OMP2 (P1), amino acids 109–547 (P2) and 322–547 (P4), but failed to respond to 439–547 (P5), although this latter fragment was clearly expressed at high levels when recombinant bacterial lysates were examined by SDS-PAGE (data not shown). These experiments localized the epitope to amino acids 332–439. A further recombinant P5a, amino acids 385–547, was generated and proved stimulatory (Fig. 3b), localizing the epitope to 385–439. To identify the T cell epitope within this 55 amino acid sequence, we synthesized three 14/15-mer peptides predicted on the basis of a published algorithm to bind strongly to HLA-DRB*0401 [20]. One of these, amino acids 400–414, elicited responses while the other two were not stimulatory (Fig. 4). Interestingly, 400–413 is predicted to be the strongest binding peptide of the three candidates. Examining the sequence of the epitope showed that it was completely conserved in all CT serovars, but the corresponding region of OMP2 from C. pneumoniae shows six amino acid substitutions, several of which are non-conservative. Accordingly, a peptide containing the C. pneumoniae OMP2 sequence was not stimulatory at any concentration up to 10 µg/ml (data not shown).

Fig. 3.

(a) Proliferative responses ([³H]-thymidine incorporation) of T cell clone AH1·37 to recombinant OMP2 polypeptides of differing length as detailed in Fig. 3, and to a lysate of E. coli which does not express a recombinant protein (control). All antigens were tested at 0·8 µg/ml final concentration. Background responses of T cell clone alone, antigen presenting cells and T cell clone + antigen-presenting cells in the absence of antigen are also shown. (b) Proliferative responses to an additional recombinant OMP2 polypeptide (P5a, amino acids 384–547) and to whole CT EBs (LGV). Other controls as in (a).

Fig. 4.

Proliferative responses ([³H]-thymidine incorporation) of T cell clone AH1·37 to three synthetic peptides tested at 0·8 µg/ml from the stimulatory OMP2 sequence (amino acids 385–439) each containing a putative DRB1*0401 binding motif. Background responses of T cell clone alone, antigen presenting cells and T cell clone + antigen-presenting cells in the absence of antigen are also shown.

T cell-mediated responses to OMP2 in CT-infected patients

To determine the extent to which cell-mediated responses to OMP2 occur amongst individuals likely to have been infected by CT, we surveyed a population of patients attending a genitourinary medicine clinic with symptoms of acute urethritis. PBMC were tested for their ability to repond to CT elementary bodies, recombinant CT OMP2, recombinant CT hsp60 and a control recall antigen (PPD or tetanus toxin). Of the 26 patients tested, 15 patients showed a significant response to CT EBs (SI > 5). In these the mean stimulation index for the response to OMP2 was 4·8 ± 1·2, compared to 17·9 ± 3·2 for the response to whole organisms; five patients had SIs > or equal to 5. By contrast responses to CT hsp60 were somewhat higher [4,7,± 1,1] with 11 patients having SIs > or equal to 5. Importantly, in the 11 patients who failed to respond to CT EBs (mean SI 2·5 ± 1·1) low responses to both recombinant CT antigens were recorded (mean responses to OMP2 and hsp60 were 2·3 ± 2·0 and 2·4 ± 1·0, respectively), while responses to PPD or tetanus toxin were maintained (mean 20·9 ± 12·5).

SFMC were available from four patients with sexually acquired reactive arthritis, in whom prominent proliferative responses to CT EBs had aleady been demonstrated; their clinical features are shown in Table 1. Although CT was not isolated by culture in any of these cases, the organism was demonstrated by PCR in synovial fluid in one case. All four showed significant responses to OMP2 (Table 1) but the corresponding responses to CT EBs were substantially higher. This suggests that although OMP2 represents a major component of the protein composition of EBs, other less abundant proteins may be more important as targets of the immune response to CT. As an example responses by these patients to recombinant CT hsp60 were similar in magnitude to those to OMP2 despite the fact that hsp60 is a much less abundant protein in EBs. SFMC from a control patient with ReA induced by Yersinia infection did not show significant responses to CT EBs, or to OMP2 and hsp60. This excludes the possibility of OMP2 acting as a non-specific mitogen or superantigen.

Table 1.

Responses of SFMC to CT antigens

	Clinical features	Background (c.p.m.)	Tetanus toxin (SI)	CT-EBs (SI)	OMP2 (SI)	HSP60 (SI)
PH	Urethritis, conjunctivitis, acute arthritis	1081	38	112	58	47
GA	Urethritis, conjunctivitis, acute arthritis; previous reactive arthritis	377	33	247	144	238
DH	Acute arthritis	780	< 1	224	11	9
	Antibodies to CT
TB	Urethritis, conjunctivitis, acute arthritis	4617	12	13	5	12
	Balanitis; CT detected in SF by PCR
Control	Acute arthritis	1768	23	2	1	3
	Antibodies to Y. enterocolitica

Discussion

In this paper we have identified OMP2 as a target of the T cell-mediated response to infection by CT, and mapped an epitope which is stimulatory when presented in the context of HLA-DRB*0401. Initial studies utilized CD4⁺ T cell clones obtained from the joint of a patient with CT-induced reactive arthritis, but subsequent studies using polyclonal population of T cells from peripheral blood or synovial fluid indicated that recognition of OMP2 is found commonly as a component of the cell mediated response to CT in both uncomplicated genitourinary infection with CT as well as CT-induced reactive arthritis.

OMP2 is the second most abundant protein in CT EBs and homologues occur in all chlamydia species. Studies on C. pneumoniae and C. psittaci show that during the chlamydia replication cycle OMP2 is expressed when reticulate bodies differentiate into elementary bodies [19]. Thus it is not surprising that this antigen elicits immune responses in vivo, and that OMP2-specific memory cells can be demonstrated by challenge with EBs in vitro. In mice infected with CT (mouse pneumonitis) by the intrauterine route, antibodies to OMP2 developed at 3 weeks post-infection [21]. Antibodies to OMP2 have already been reported to be present in the majority of chlamydia-infected patients, and this applies to infections with C. pneumoniae and C. psittaci in addition to CT [22–25]. Antibodies to OMP2 have also been noted in a 2-D immunoblotting study of cohort of reactive arthritis patients [26]. However, the antibody response to OMP2 may be of limited utility since in EBs OMP2 is confined to the inner surface of the EB outer membrane [19,27].

The presence of antibodies to a protein antigen such as OMP2 implies the existence of T helper cells. T cell lines reactive with OMP2 have been described in CT-infected mice [28], but these have not previously been identified in humans. It has also been suggested that OMP2-specific T cells might be able to provide help for MOMP-specific B cells, since uptake of EBs by MOMP-specific B cells would allow presentation of OMP2 to T cells and the receipt of help [29]. However, in a different experimental system such help was not apparent [30].

OMP2 shows a high degree of conservation between chlamydia species for most of its sequence but has a highly variable segment in the N-terminal 100 amino acids [31], a fact which has been used in developing diagnostic PCR methods for distinguishing between different chlamydia species [32]. Interestingly the epitope which we mapped (amino acids 400–413) was not within the highly variable segment, but nevertheless the clone was able to distinguish between CT and C. pneumoniae. Since only eight amino acid differences were noted in a comparison between serovars B and L1 of CT [33], it is not surprising that the clone responded to both DK20 and L2. The use of an algorithm [20] to predict the peptides most likely to be recognized on the basis of their ability to bind to HLA-DRB1*0401 was remarkably successful in this case, with the stimulatory peptide having by far the highest predicted affinity amongst the candidate peptides whose amino acid sequences fulfilled the minimal requirements for binding to DRB1*0401.

What is the significance of T cell-mediated immunity to OMP2? It is not yet clear which CT antigens are targeted by protective immune responses, or which are associated with immunopathology. Data on protective responses are being provided by animal vaccination studies, particularly using DNA vaccines [34,35], and it is notable that vaccination with plasmids encoding MOMP elicited immune responses but no protection. Such data are not available for OMP2, but the conservation of the sequence of OMP2 among different CT serovars may make it an attractive candidate compared to MOMP where T cell responses can be directed at highly variable portions of the MOMP sequence which would produce serovar-specific protection. Further data on the epitopes commonly recognized in the response to MOMP and the ability of OMP2-specific T cells to provide help for MOMP-specific B cells in humans (as suggested by a murine study [29]) are required to determine whether OMP2 or MOMP is the more attractive vaccine candidate.

In relation to immunopathology, it is interesting that immune responses, both antibody and T cells, to a less abundant protein hsp60, have been associated with disease [36–38]. In the patients with arthritis which we tested, responses to hsp60 were as great as, or greater than, those to OMP2, although this conclusion requires caution since full dose–response curves were not established for the two antigens which were simply compared at concentrations found previously to be optimal. However, in support of the conclusion, in a set of clones from one reactive arthritis patient derived by stimulation with EBs, the frequency of hsp60-specific clones was approximately twice that of OMP2 reactive clones. It is likely that hsp60 is more strongly immunogenic than the more abundant membrane proteins, since it is a common target of immune responses to all intracellular bacteria and has properties which enhance its immunogenicity [39–42]. However, despite the apparent association between immunopathology and responses to hsp60, responses to bacterial hsp60 can also be protective, as shown by immunization studies in mycobacterial infection [43,44].

Thus a full assessment of the role of T cell recognition of OMP2 in CT-infection and CT-associated diseases will require further assessment of the frequencies of OMP2-specific T cells, the cytokines they secrete, and whether they are generally found at sites of inflammation such as the reactive arthritis joint.