In vitro T lymphocyte responses to proteinase 3 (PR3) and linear peptides of PR3 in patients with Wegener's granulomatosis (WG)

Abstract

T cell-mediated immunity is thought to play an important role in the pathogenesis of WG. In previous studies a minority of WG patients as well as some healthy controls showed in vitro proliferation of their peripheral blood mononuclear cells (PBMC) to PR3, the main autoantigen in WG. The relevant peptides responsible for this in vitro proliferation have not been identified. In order to define immunogenic peptides, PBMC of 13 WG patients in remission and 10 healthy controls were tested for proliferation to linear peptides of PR3 and to whole PR3. Fifty overlapping peptides spanning the whole PR3 sequence were synthesized. Peptides were tested in pools of five peptides and as single peptide. PBMC of two WG patients and one healthy control proliferated to whole PR3 and to peptide pools. In addition, 10 WG patients and eight healthy controls that did not proliferate to whole PR3 did proliferate to pools of PR3 peptides. Although more WG patients tended to react to particular peptide pools, no significant difference was seen between lymphocyte proliferation to PR3 peptides of WG patients and that of healthy controls. The pools of peptides recognized were mainly located at the N- and C-terminus of PR3. No correlation was observed between HLA type and proliferation on particular peptide pools. No proliferation of PBMC was observed to single peptides. In conclusion, T cells of WG patients proliferate in vitro more frequently to PR3 peptides than to the whole PR3 protein. Peptides derived from the signal sequence, the propeptide or peptides located at the C-terminus of PR3 induce highest levels of proliferation. No specific PR3 sequence could be identified that was preferentially recognized by PBMC of WG patients compared with controls.

INTRODUCTION

WG is characterized by granulomatous inflammation in the upper and lower respiratory tract, systemic vasculitis affecting small blood vessels, and pauci-immune necrotizing crescentic glomerulonephritis (NCGN) [1]. Furthermore, in the majority of WG patients [2], anti-neutrophil cytoplasmic antibodies (ANCA) directed against PR3 are present [3–5]. PR3-ANCA has been established as a specific marker for primary, pauci-immune, small vessel vasculitis, of which especially WG is a subgroup. Detection of these autoantibodies is helpful in the diagnosis and follow up of WG patients [6,7].

T cell-mediated immunity is thought to play an important role in the pathogenesis of granulomatous inflammation and vasculitis as found in patients with WG. Several findings implicate T cells in the pathogenesis of WG. First, direct involvement of T cells in tissue injury at the site of inflammation is suggested by the occurrence of large mononuclear infiltrates containing monocytes and CD4⁺ T cells as frequently found in the respiratory tract [8]. Furthermore, CD4⁺ and CD8⁺ T cells are present in the glomeruli and in the renal interstitium of patients with ANCA-associated renal vasculitis [9,10]. Second, activated CD4⁺ and CD8⁺ T cells have been detected in the circulation of patients with WG [11,12]. In agreement with these findings, markers of T cell activation such as soluble IL-2 receptor [13,14], soluble CD30 [15] and soluble CD4 [16], are elevated in the circulation of WG patients. Furthermore, increases in the levels of all these three markers have been shown to correlate with disease activity in WG. Third, ANCA are predominantly of the IgG1 and IgG4 subclass, suggesting an antigen-driven T cell-dependent immune response [17]. Finally, alternative therapeutics that are mainly directed against T lymphocytes, like anti-thymocyte globulin [18], humanized anti-CD52 MoAb [18] or cyclosporin [19], were effective in patients with severe WG.

Several groups have tried to analyse the in vitro proliferative capacity to PR3 of peripheral blood mononuclear cells (PBMC) from WG patients. However, only a small proportion of WG patients as well as some healthy controls showed proliferation to PR3 or crude granular extracts of neutrophils [20–26]. T cell proliferation to PR3 was found in a range of 24% [23] to 63% [24] of WG patients and in a range of 0% [21,23,26] to 42% [25] in healthy controls. Nonetheless, most authors found a significant difference between WG patients and healthy controls, as WG patients proliferated more frequently to PR3 compared with healthy controls [22,24–26]. King et al. investigated the relevant PR3 peptides responsible for the in vitro proliferation of PBMC from WG patients, but only one out of 18 patients responded to PR3 peptides [25]. So, the relevant peptides responsible for the in vitro proliferation to PR3 have not yet been identified.

The objective of this study was to analyse the proliferative capacity of PBMC from WG patients to peptides of PR3 as well as to PR3. In order to identify further relevant peptides, overlapping synthetic peptides were used spanning the entire sequence of PR3 including the signal and prosequence. Each peptide was 15 amino acids in length with an overlap of 10 amino acids, and all stretches of 11 amino acids of the PR3 sequence were available, in contrast to the previous study where not all stretches of nine amino acids were available, so some relevant immunogenic peptides might have been missed [25]. In previous studies, lymphocyte proliferation of WG patients and healthy controls to PR3 was found at PR3 concentrations ranging from 0·1 μg/ml [22] to 20 μg/ml [24]. As shown in those studies, different proportions of WG patients and healthy controls proliferate when varying PR3 concentrations are used [22,25,26]. Therefore, we used PR3 peptides at three concentrations in order to find optimal stimulation. Here we provide evidence that a higher level of proliferation of autoreactive T cells of WG patients was induced by peptides of PR3 than by the whole PR3 protein. Peptides derived from the signal sequence, the propeptide or the C-terminus of PR3 induced highest levels of proliferation by PBMC of WG patients and healthy controls. Nevertheless, no PR3 sequence could be identified that was specifically recognized by PBMC of WG patients.

PATIENTS AND METHODS

Patients and controls

Patients who were newly diagnosed between January 1987 and January 1997 with PR3-ANCA-associated WG were considered for inclusion in the study. Diagnosis of WG was established according to clinical and histological criteria [1]. Patients fulfilled the American College of Rheumatology criteria for WG [27] and all patients had biopsy-proven NCGN at the time of diagnosis. In order to exclude in vitro non-responsiveness due to immunosuppressive drugs or severe disease activity, only patients in remission without significant immunosuppressive treatment (cyclophosphamide <25 mg daily) at the time of testing were included. Previously, all patients has been treated with cyclophosphamide and prednisolone. At the time of diagnosis, sera had been tested for ANCA by indirect immunofluorescence (IIF) and by ELISA for antibodies to PR3, myeloperoxidase (MPO) and human leucocyte elastase (HLE). All patients were positive for PR3-ANCA only (see below). Thirteen WG patients were tested (seven males and six females). Their median age was 60 years (range 28–73 years). Clinical and laboratory data of the patients are outlined in Table 1. Sera drawn at the time of this study were tested for ANCA by IIF and PR3-ANCA by antigen-specific direct ELISA. The control group consisted of 12 healthy volunteers (eight males and four females) and their median age was 48 years (range 27–62 years).

Table 1.

Patient characteristics at the time of testing

Patient no.	Sex	Age*, years	C-ANCA titre	PR3-ANCA titre (U)	Treatment	Years after diagnosis†	Years after last period of disease activity‡
1	M	55	80	79	–	12·4	4·3
2	M	70	320	29	–	7·1	1·8
3	F	28	640	1179	–	8·5	7
4	M	50	–	5	–	7·4	7·4
5	M	60	160	230	–	1·9	1·9
6	M	73	160	136	–	1·8	1·8
7	F	65	320	2590	CP 25 mg	1·3	1·3
8	F	32	640	452	–	4·3	4·3
9	F	60	160	63	–	3·3	3·3
10	M	58	–	4	–	3·2	3·2
11	F	65	40	7	–	5·7	5·7
12	M	52	160	16	–	22·6	3·3
13	F	62	> 640	135	CP 25 mg	10·8	0·9

C-ANCA titre, cytoplasmic ANCA titre measured by indirect immunofluorescence; PR3-ANCA titre, ANCA titre measured by direct PR3 ELISA; treatment, immunosuppressive treatment; CP, cyclophosphamide.

^*Median age is 60 years.

^†Median disease duration after initial diagnosis is 5·7 years.

^‡Median time interval after the last period of disease activity is 3·2 years.

ANCA detection

ANCA were detected by IIF on ethanol-fixed granulocytes according to a standard protocol [2,28] with minor modifications [29]. Patient and control sera were tested at serial two-fold dilutions starting at a dilution of 1:20. Slides were read by two independent observers and titres of ≥1:40 were considered positive for ANCA.

Specificity for PR3, MPO or HLE was defined by antigen-specific capture ELISA, as previously described [30]. Sera were considered positive when values exceeded the mean +2 s.d. of normal controls (n = 50).

Titrers of PR3-ANCA were also assessed by antigen-specific direct ELISA [12]. In this ELISA a standard PR3-ANCA serum is defined as 100 U and the sera tested are calculated accordingly. Titres of ≥6 U are considered positive for PR3-ANCA, based on results of normal controls.

MHC typing

All WG patients and healthy controls were typed using the standard NIH microlymphocytotoxicity technique for HLA class I. HLA class II typing was performed by serology, using two-colour fluorescence technique, and in case of homozygosity or unclear results, additionally by a DNA-based technique, using sequence-specific primers (SSP) [31].

Antigens

Purification of PR3 and MPO

PR3 was purified according to the isolation procedure as described previously [32], with minor modifications. The α granules were pooled and disrupted by sonification, six times for 20 s (22 A). The granule extract was only applied to a Biorex-70 column (BioRad Labs, Richmond, CA). The PR3 batch was not contaminated by HLE, MPO or lactoferrin as tested by capture ELISA. MPO was isolated as described previously [33]. To avoid possible mitogenic and toxic effects, PR3 and MPO were boiled for 15 min. Endotoxin in the PR3 and MPO batch as tested by the limulus amoebocyte assay (Biowhittaker, Vervier, Belgium) was not detected at PR3 and MPO concentrations used in the proliferation assay.

Peptide synthesis

Fifty overlapping peptides were synthesized from the entire sequence of PR3 [34]; the peptides are listed in Fig. 1. The peptides were 15 amino acids (aa) long, except for peptide nos 39 and 49 which contained 16 aa and peptide no. 38 being 17 aa long. The peptides had an overlap of 10 aa. At the N-terminus of each peptide a cysteine residue was present. This residue was either supplementary (c-) or part of the normal PR3 sequence (C). These peptides were synthesized at the Peptide Centre of the research school of Infection and Immunity, Veterinary Faculty, Utrecht, The Netherlands. Peptides were prepared by automated simultaneous multiple peptide synthesis (SMPS) [35]. The SMPS set up was developed using a standard autosampler (Gilson 221) as described previously [35]. Briefly, standard Fmoc chemistry with in situ PyBop/NMM activation of the aa in a five-fold molar excess with respect to 2 mol/peptide PAL-PEG-PS resin (PerSeptive Biosystems, Foster City, CA) was employed. Peptides were obtained as C-terminal amides after cleavage with 90–95% TFA/scavenger cocktails. The composition and purity of the peptides was checked by reverse-phase high performance liquid chromatography (RP-HPLC) and ion-trap mass spectrometry.

Fig. 1.

Diagram illustrating the 50 linear peptides spanning the sequence of PR3. Numbering of the amino acids according to the sequence of PR3 published by Campanelli et al. [44]. Residue numbers −27 to −3 constitute the signal sequence; −2 to −1, the prosequence; 1–222 the mature form of PR3; residue 222–229 the C-terminal extension. The signal sequence, the prosequence and the C-terminal extension are removed during the processing of PR3. Residues of the catalytic triad H⁴⁴, D⁹¹ and S¹⁷⁶ are marked by gray boxes. Peptide numbers are typed in front of each sequence. At the N-terminus of each peptide a cysteine residue is present; this residue was either supplementary (c-) or part of the normal PR3 sequence (C).

Peptides were either used in pools of five (peptide nos 1–5, 6–10, 11–15, 16–20, 21–25, 26–30, 31–35, 36–40, 41–45, 46–50) or as single peptides. These pools of five peptides did not reduce the proliferative response of PBMC from a healthy control to pokeweed mitogen (PWM; Gibco Labs, Breda, The Netherlands) or a combination of anti-CD3 and anti-CD28 MoAbs.

Polyclonal T cell activation

A combination of anti-CD3 (clone WT-32) and anti-CD28 MoAbs (CLB, Amsterdam, The Netherlands, clone 20-4996) served as a positive control for the proliferative capacity of the PBMC.

Proliferation assay

PBMC were isolated from freshly drawn heparinized blood by Lymphoprep gradient centrifugation (Nycomed, Oslo, Norway). PBMC (1 × 10⁵ cells/well) were stimulated with PR3 (final concentration 10 μg/ml), MPO (5 μg/ml), pools of five peptides (final concentrations 2, 10 or 25 μg/ml per peptide) or single peptides (final concentration 25 μg/ml per peptide) in RPMI containing 15% pooled normal human serum and gentamycin (60 μg/ml; BioWittaker). Stimulation with anti-CD3 and anti-CD28 MoAbs served as positive controls for the proliferative capacity of PBMC. As negative control PBMC were incubated with RPMI alone. The PBMC were cultured for 6 days. Proliferation was assessed by ³H-thymidine incorporation (final concentration 0·1 μCi/ml) which was added for the final 16 h of culture. All assays with antigen were carried out in triplicate and assays without antigen (blank value) in six-fold. The cultures were harvested on glassfibre paper and incorporation of ³H-thymidine was determined by β-scintillation counting (Tri-Carb 1500 Liquid Scintillation Analyzer; Packard A Canberra Co. USA, Downers Grove, IL). The results were assessed by calculating the stimulation index (SI) as the ratio of the mean of the disintegrations per second (dps) of antigen-stimulated to the mean of unstimulated cultures. The proliferative assay was considered positive when SI > 3.

Proliferation assay of rabbit PBMC

Rabbits were immunized subcutaneously with 100 μg human PR3 and were boosted twice with 100 μg human PR3 subcutaneously. These rabbits developed antibodies to PR3 as assessed by antigen-specific direct ELISA. As a control, PBMC were used from rabbits that had been immunized and boosted twice with human serum amyloid P (SAP) [36]. Tests were performed 2 weeks after the last booster. PBMC of the rabbits were isolated from freshly drawn heparinized blood by Nycoprep gradient centrifugation (Nycomed). The PBMC proliferation assay was performed in sterile round-bottomed 96-well plates (Costar Europe, Badhoevedorp, The Netherlands) in 100 μl RPMI containing 10% fetal calf serum (FCS) and gentamycin. All assays with antigen were carried out in triplicate and assays without antigen (blank value) in six-fold. PBMC (1 × 10⁵ cells/well) were stimulated with human PR3 (final concentration 1 or 10 μg/ml) or human SAP (final concentration 1 or 10 μg/ml). Stimulation with PWM (final concentration 10 μg/ml) and phytohaemagglutinin (PHA; final concentration 2·5 μg/ml; Murex Diagnostics S.A., Chatillon, France) served as positive controls for the proliferative capacity of the PBMC. As negative control PBMC were incubated with RPMI alone. The PBMC were cultured for 4, 6 or 8 days. Proliferation was assessed as described above.

Statistical analysis

Test results from patients or healthy controls with high background values (higher than the mean blank value +2 s.d.) were discarded. Using this criterion one healthy control was eliminated from the analysis. Differences in age, SI or dps between WG patients and healthy controls were tested by means of Mann–Whitney U-test. Differences in the male/female ratio were tested by means of Fisher's exact test. The level of significance used was 0·05. All reported P values are two-sided.

RESULTS

Patients and controls

Thirteen patients with WG and 12 healthy controls were studied. All patients were in remission but previously had had active disease with renal involvement. Eleven of the 13 WG patients were still ANCA⁺ at the time of testing (Table 1). One healthy control was excluded from analysis due to high background values. The mean number of PBMC from WG patients after isolation was 1·0 × 10⁶ per ml blood (range 0·3–1·8). Patient no. 7 had a very low PBMC count, probably explained by the continuous use of 25 mg cyclophosphamide. The mean number of PBMC from healthy controls after isolation was 1·3 × 10⁶ per ml blood (range 0·8–2·27).

Proliferation in response to anti-CD3/anti-CD28, PR3 and MPO

PBMC of all patients as well as of all healthy controls proliferated in response to stimulation with a combination of antibodies to CD3 and CD28 (Fig. 2a). No significant differences were observed between the SI values of patients and healthy controls on stimulation with CD3/CD28. Figure 2b shows that two patients (patient nos 12 and 13) proliferated to PR3. PBMC of one healthy control proliferated to PR3 as well as MPO. No significant differences were observed between the SI values of patients and healthy controls on stimulation with PR3 or MPO.

Fig. 2.

Proliferative responses of peripheral blood mononuclear cells (PBMC) of patients with WG (•) and healthy controls (○). PBMC of WG patients (n = 13) and healthy controls (n = 11) were stimulated with anti-CD3/anti-CD28 (a), and PR3 (10 μg/ml) and myeloperoxidase (MPO; 5 μg/ml) (b) for 6 days. Lymphocyte proliferation was assessed by ³H-thymidine incorporation. Proliferative responses are depicted as stimulation index (SI). The median is shown for WG patients and healthy controls.

Proliferation in response to pools of PR3 peptides

Proliferation of PBMC to peptides was assessed in pools of five peptides at three peptide concentrations. The antigen concentration optimal for proliferation to a certain antigen differed between individuals, so the results are given as the maximal proliferation seen on any peptide concentration. PBMC of WG patients as well as of healthy controls proliferated incidentally to pools of PR3 peptides (Fig. 3a,b). Although the reaction of patients seemed higher than controls, no significant difference between the SI values of patients compared with controls was seen. Also no significant differences between the SI values of patients and controls were seen for each concentration of peptides separately. PBMC of 12 of the 13 (92%) WG patients tested proliferated to any of the PR3 peptide pools. Of these 12 patients two also reacted to whole PR3 (Fig. 3c,e, patient nos 12 and 13, respectively). Consequently, PBMC of 10 WG patients proliferated to any of the pools of PR3 peptides, whereas they did not proliferate to whole PR3. PBMC of nine of the 11 (82%) healthy controls tested proliferated to any of the pools of PR3 peptides. Of these nine controls one also reacted to whole PR3 (Fig. 3d). PBMC of eight healthy controls proliferated to any pool of PR3 peptides, whereas they did not proliferate to PR3 (an example is shown in Fig. 3f). PBMC of WG patients and healthy controls did not proliferate on each peptide pool to the same extent (Table 2). Many of the patients and controls showed proliferation to the C-terminal pool of peptides (peptide nos 46–50). Only two WG patients and none of the healthy controls proliferated to the pool with peptide numbers 26–30. No significant difference was seen between the number of patients compared with healthy controls that proliferated to the various peptide pools.

Fig. 3.

Proliferative responses of peripheral blood mononuclear cells (PBMC) of patients with WG and healthy controls. PBMC of WG patients (n = 13) (a) and healthy controls (n = 11) (b) were stimulated with pools of five peptides for 6 days. Results of maximal proliferation to any peptide concentration are shown. The median is shown for WG patients and healthy controls (a,b). Proliferative responses of individual patients that proliferated to whole PR3 (c,e) and of one control who proliferated to whole PR3 (d) and one that did not (f) are shown as well. Patient and control numbers are according to numbers shown in Table 3. Lymphocyte proliferation was assessed by ³H-thymidine incorporation. Proliferative responses are depicted as stimulation index (SI). The mean of triple values for individuals is depicted in (c–f) together with the s.d. values.

Table 2.

Number of patients with proliferation to particular pools of five peptides

		Number of patients/controls that proliferated
Peptide no.	Amino acid residues of PR3	WG patient (n = 13)	Healthy control (n = 11)
1–5	−27 to 8	6	5
6–10	−2 to 33	2	3
11–15	24–56	5	3
16–20	45–79	5	4
21–25	70–104	4	5
26–30	95–129	2	0
31–35	120–154	5	1
36–40	145–181	5	6
41–45	172–206	4	2
46–50	197–229	9	6

Proliferation was defined as stimulation index (SI) > 3 for any concentration of peptides used.

Amino acid residues according to peptide numbers in Fig. 1.

P values determined by Fisher's exact test.

Interestingly, most of the proliferative responses were seen after stimulation with pools of peptide nos 1–5 and 46–50. These peptides are located in the signal peptide and at the C-terminus of PR3. Furthermore, pools with peptide numbers 16–20, 21–25 and 36–40 were recognized by a substantial number of patients and healthy control PBMC. In pools with peptide numbers 16–20 the active site histidine is located, whereas in the pools with peptide numbers 36–40 the active site serine is located.

Proliferation in response to single peptides of PR3

To test whether the proliferation seen to pools of five peptides could also be detected using single peptides, the proliferative response of PBMC of four WG patients and four controls to single peptides of the pools with peptide nos 1–5, 16–20, 21–25 and 46–50 was measured (Fig. 4 shown for two patients and two healthy controls). The pools of these peptides were also included in the test. In contrast to previous experiments, no proliferation was seen to pools 1–5 and 16–20, but some single peptides were recognized (data not shown). To pools with peptides numbered 21–25 and 46–50 proliferation was seen by patient no. 12 and healthy control no. 9, but only single peptide no. 48 induced proliferation of PBMC from healthy control no. 9 (Fig. 4). In this experiment patient no. 12 that previously proliferated to PR3, did not proliferate to PR3. This experiment on proliferation to single peptides (Fig. 4) was performed 1 year after the previous experiment on proliferation to pools of PR3 peptides (Figs 2 and 3). In this year patient no. 12 had had a relapse of WG and received immunosuppressive treatment, which might explain the difference in proliferation to whole PR3 and pools of PR3 peptides.

Fig. 4.

Proliferative responses of peripheral blood mononuclear cells (PBMC) of patients with WG and healthy controls. PBMC of two WG patients (a) and two healthy controls (b) were stimulated with 25 μg/ml of single peptides, pools of these peptides, a combination of anti-CD3 and anti-CD28 or PR3 for 6 days. Lymphocyte proliferation was assessed by ³H-thymidine incorporation. Proliferative responses are depicted as stimulation index (SI).

Influence of HLA type on proliferative responses to peptides of PR3

In a lymphocyte transformation test (LTT) on whole PR3, PR3 is processed and is then presented by MHC class II molecules. In a LTT performed with peptides the presentation of these peptides and possibly also the proliferation on these peptides may be dependent on the HLA type of the PBMC tested, since peptides are probably not processed but compete with endogenous antigens for presentation via MHC class II [37] or class I complex. The HLA type of all tested WG patients and healthy controls was determined (Table 3). Interestingly, the two patients that had a proliferative response on PR3 (nos 12 and 13) had the same HLA-DR and DQ type, but the healthy control that also proliferated on PR3 (no. 8) had a different HLA type. Patients with the same HLA-DR type (nos 1 and 3, and nos 12 and 13) did, however, not proliferate to the same peptide pools. Also, healthy controls with an almost identical HLA type (nos 2 and 4) did not proliferate to similar peptide pools. Furthermore, patients or controls that proliferated to the same peptide pools did not share the same HLA-A, B, DR, or DQ type. So, the HLA-A, B, DR, or DQ type did not seem to influence the proliferation on pools of PR3 peptides. Influences at the level of subtypes cannot be excluded, as subtyping at DNA level was not performed.

Table 3.

HLA type of WG patients and healthy controls

Patient no.	HLA-A	HLA-B	HLA-DR	HLA-DQ
1	A1 A3	B17(57) B37	DR4 DR7	DQ3
2	A2 A3	B15(62) B22	DR4 DR8	DQ3
3	A2 A3	B15(62) B41	DR4 DR7	DQ3
4	A1 A2	B12(44) B15(62)	DR4 DR5(11)	DQ3
5	A2 A3	B7 B27	DR2(15) DR5(11)	DQ1 DQ3
6	A1 A2	B7 B15(62)	DR4 DR6 DR(14)	DQ1(5) DQ3(7)
7	A3 A19(29)	B12(44) B12(45)	DR4 DR5(11)	DQ3
8	A1 A3	B8 B18	DR3 DR5(11)	DQ2 DQ3
9	A2 A3	B7 B12(44)	DR2(15)	DQ1(6)
10	A1 A2	B12(44) B17(57)	DR2(15) DR4	DQ1 DQ3
11	A1 A19	B8 B40(60)	DR3 DR4	DQ3
12	A2 A3	B8 B35	DR3	DQ2
13	A1	B8	DR3	DQ2
Control
1	A2 A19(32)	B40(60)	DR5(11) DR6(13)	DQ1 DQ3
2	A2 A3	B8 B15(62)	DR3 DR4	DQ2 DQ3
3	A2 A11	B18 B35	DR1 DR3	DQ1 DQ2
4	A1 A28	B8 B15(62)	DR3 DR4	DQ2 DQ3
5	A1Aw19(30)	B8 B7	DR6(14) DR2(15)	DQ1(5) DQ1(6)
6	A1A19(31)	B8B40(60)	DR1 DR4	DQ1 DQ2
7	A2 A3	B35 B7	DR2 DR4	ND
8	A2 A19(30)	B7 B35	DR2 DR5(12)	DQ1 DQ3
9	A1A9(24)	B8B16(38)	DR2(15) DR3	DQ1 DQ2
10	A2	B40(60) B16(39)	DR6(13)	DQ1(6)
11	A3 A19(29)	B5(51) B12(45)	DR4	DQ3 DQ3

Proliferation of immunized rabbits in response to PR3

As a positive control two rabbits were immunized and boosted with human PR3 and two rabbits were immunized and boosted with human SAP. The rabbits produced antibodies to human PR3 or human SAP, respectively. PBMC of both rabbits proliferated to the positive stimulus PHA and proliferation was maximal after 6 days (Fig. 5). PBMC of rabbits immunized with human PR3 did not show a proliferative response on PR3 (Fig. 5a) whereas PBMC of rabbits immunized with human SAP did show a high proliferative response on SAP (Fig. 5b). The proliferative response on human SAP of PBMC of rabbits immunized with human SAP was also maximal after 6 days.

Fig. 5.

Proliferative response of peripheral blood mononuclear cells (PBMC) of two rabbits immunized with PR3 or two rabbits immunized with serum amyloid P (SAP). PBMC of rabbits immunized with PR3 (a) or SAP (b) were stimulated with phytohaemagglutinin (PHA; 2·5 μg/ml), PR3 (1 and 10 μg/ml) or SAP (1 and 10 μg/ml) for 4 (□), 6 (hatched) or 8 days (▪). Lymphocyte proliferation was assessed by ³H-thymidine incorporation. Proliferative responses are depicted as stimulation index (SI). The mean of three values is depicted together with the s.d. values.

DISCUSSION

This study, using whole PR3 and overlapping linear peptides spanning the entire sequence of PR3, shows that only 15% of WG patients in remission and 9% of healthy controls proliferate to PR3, whereas 92% of WG patients in remission and 82% of healthy controls proliferate to particular pools of linear peptides. The proliferation to peptides did not seem to be dependent on HLA type and no PR3 sequence could be identified that was specifically recognized by PBMC of WG patients.

Several groups have investigated the proliferation of PBMC from patients with WG to PR3 [20–26]. In previous studies it was shown that disease activity did not influence the proliferative response to PR3 [21,23,25], whereas immunosuppressive treatment did [21,25]. So, in our study we selected WG patients in remission without significant immunosuppressive treatment. Furthermore, it has been shown that patients positive for PR3-ANCA at the time of testing proliferated better to PR3 than those negative for ANCA [21]. In our study most patients were positive for PR3-ANCA at the time of testing. The two patients that showed a proliferative response to PR3 were positive for C-ANCA. In previous studies proliferation to PR3 was already seen at day 5 and was maximal at day 7 or 10 [24,25]. To compare our data with previous data we assessed the proliferative response to PR3 after 6 days of culture.

The PR3 peptides as used in this study can be presented by antigen-presenting cells (APC) without processing. Previous studies showed that exogenously added peptides compete for the presentation of endogenous antigens by the MHC class II complex [37]. Peptides eluted from MHC class II have lengths that can vary from 10 to 24 residues. Nine core residues of these peptides are bound in the peptide cleft of the MHC class II molecule [38,39]. In our study at least all stretches of nine amino acids of the PR3 sequence were available. Furthermore, the cysteine residues attached to the N-terminus of the linear peptides probably did not interfere with the peptide binding, as the N- and C-termini of the peptides bound to MHC II are different in length and are not involved in peptide binding [38].

In accordance with previous studies [20,22,24,25] we showed that PBMC from healthy controls also respond to PR3. T cell responses from healthy controls have also been seen for other autoantigens such as myelin basic protein [40], glutamate decarboxylase [41], insulin [42] and acetylcholine receptor [43].

In this study highest level of proliferation of WG patients and healthy controls was seen to peptide pools located at the N-terminus, were the signal peptide propeptide is located, and at the C-terminus, were the C-terminal extension of seven amino acids is located. These three sequences are all removed during processing of PR3 and, thus, are not present in the mature protein [44,45]. This may in part explain why more proliferation is seen to PR3 peptides compared with whole PR3. PR3-specific T cell epitopes may lie in these sequences as the signal peptide, propeptide and C-terminal extension of PR3 differ in various amino acids from HLE [46]. In a previous study by King et al. [25] one of 18 WG patients proliferated to linear peptides of PR3. These peptides were mainly located in the signal peptide and C-terminal extension and overlapped with the peptides on which proliferation was seen in our study.

With the same set of peptides used in the present study epitope mapping was performed with PR3-ANCA sera of WG patients [47]. Epitopes recognized by PR3-ANCA sera were located near the active site, whereas most proliferation of PBMC was seen to peptides located at the N- and C-termini of PR3. One of the six most important epitopes recognized by PR3-ANCA sera as determined in another study [48] was located in peptides 28 and 29. Negligible proliferation of PBMC was seen on the pools containing peptides 26–30 in our study. So, epitopes bound by PR3-ANCA sera of WG patients do not correspond to T cell epitopes recognized by PBMC of WG patients and healthy controls.

Several HLA-A2-restricted PR3-derived peptides have been identified [23]. For one of these HLA-A2-restricted peptides it has been shown that cytotoxic T lymphocytes (CTL) raised against this peptide preferentially lyse myeloid leukaemia cells [49] and inhibit granulocyte-macrophage colony-forming unit from HLA-A2⁺ patients with chronic myeloid leukaemia [50]. The sequence of this HLA-A2-restricted peptide (VLQELNVTV) is located in peptide 35. Only five of the 13 WG patients proliferated on a pool containing peptides 31–35, and only three of these five WG patients were HLA-A2⁺. So, this PR3-derived HLA-A2-restricted peptide did not fully correspond to epitopes recognized by PBMC of WG patients and healthy controls.

Finally, we assessed T cell responses to human PR3 in rabbits that see the human PR3 molecule as an exogenous antigen. Also in rabbits a clear proliferative response of PBMC to PR3 failed to appear comparable to the restricted response of PBMC from WG patients to PR3. Possibly, the antigen inducing T cell help for PR3-specific B cells in WG patients is not PR3 itself but a protein complexed to PR3.

In conclusion, our results demonstrate that both WG patients and healthy controls proliferate to PR3 peptides, and a minority proliferate to whole PR3. These data suggest that autoreactive T cells to peptides of the PR3 sequence are present in both WG patients and healthy controls. Only a minority of WG patients and healthy controls proliferated to whole PR3. A similar lack of response was seen in animals immunized with PR3. Possibly, the antigen-inducing T cell help for PR3-specific B cells in WG patients is not PR3 itself but a protein complexed to PR3. One other possibility is that continuing PR3-ANCA production is either T cell-independent or mediated through non-specific B cell stimulation.