Characterization of self-T-cell response and antigenic determinants of U1A protein with bone marrow-derived dendritic cells in NZB × NZW F₁ mice

Abstract

Systemic lupus erythematosus (SLE) is characterized by the existence of a heterogeneous group of autoantibodies directed against nuclear intact structures, such as nucleosomes and small nuclear ribonucleoproteins (snRNPs). Autoantibodies against snRNPs are of special interest because they are detectable in the majority of SLE patients. Although the B-cell antigenic determinants have been well characterized, very limited data have been reported in regard to the T-cell epitopes of snRNPs. Furthermore, several studies have demonstrated that determination of the auto-T-cell epitopes recognized by freshly isolated T cells is difficult from unprimed lupus mice when self-antigen-pulsed B cells or macrophages are used as antigen-presenting cells (APCs) in vitro. In the present study, we showed a novel approach for determining the auto-T-cell epitopes, using bone marrow-derived dendritic cells (BMDCs) pulsed with the murine U1A protein – an immunodominant antigen of the U1 snRNPs – which is capable of activating freshly isolated T cells from unprimed (NZB × NZW) F₁ (BWF₁) mice in vitro. The T-cell epitope area was found to be located at the C-terminus of U1A, overlapping the T-cell epitope of human U1A that has been reported in human SLE. Identification of the autoreactive T-cell epitope(s) in snRNPs will help to elucidate how reciprocal T–B determinant spreading of snRNPs emerges in lupus. The results presented here also indicate that it is feasible to use this approach to further explore strategies to design immunotherapy for patients with lupus.

Introduction

Systemic lupus erythematosus (SLE) is characterized by a loss of tolerance to self-antigens and persistent production of autoantibodies.^1,2 Most of these autoantibodies are directed against intracellular macromolecules, such as nucleosomes and small nuclear ribonucleoproteins (snRNPs).¹ U1 snRNP is one of the U-type snRNP complexes (U snRNPs) that constitute a spliceosome. The U1 snRNP contains one U1 RNA molecule complexed with eight Sm proteins (B, B', D1, D2, D3, E, F, G) and three U1-specific proteins (A, C, 70K).³ Humans with SLE and MRL-lpr/lpr mice often produce immunoglobulin G (IgG) antibodies against components of the U1 snRNP particle.^3,4 Previous studies have shown that U1A is the immunodominant antigen of U1 snRNP,⁴ and that epitopes spread to other constituents of self-U snRNPs through intermolecular/intrastructural help.⁵ Therefore, identification of the auto-T-cell epitopes of U1A would enable further study of the diversification of the autoimmune responses to snRNP in lupus.

Although substantial efforts have been made to define the autoreactive T-cell response in lupus, little has been reported about the T-cell determinant of the self-antigen.^6–8 Datta and colleagues used T-cell hybridomas and T-cell lines derived from (SWR × NZB) F₁ mice and human lupus to define the critical autoepitopes for the nephritogenic autoantibody-inducing T-helper cells in the core histones of nucleosomes.^6,7,9 Singh and associates also showed that unprimed (NZB × NZW) F₁ (BWF₁) mice develop spontaneous T-cell autoimmunity to V_H-region peptides of syngeneic IgG anti-DNA monoclonal antibodies (mAbs).^8,10 Previous reports^7,8,11,12 have described difficulties in the determination of auto-T-cell epitopes recognized by freshly isolated T cells from unprimed mice cocultured with self-antigen-pulsed B cells or macrophages in vitro. Therefore, in the present study, bone marrow-derived dendritic cells (BMDCs) were used to assay the antigen-specific T-cell proliferative response to define the T-cell antigenic determinants of self-reactive T cells in BWF₁ mice.

It is well known that BMDCs play a central role in the induction of immunity in T cells, as well as B cells, in vivo.¹³ They not only express a high level of the major histocompatibility complex (MHC)–peptide complex, but also the other accessory molecules that interact with the receptors on the T cells to enhance adhesion and signalling. In addition, immature BMDCs can take up and present antigens so efficiently that picomolar and nanomolar concentrations of antigen suffice; much lower concentrations than the micromolar levels typically employed by B cells or macrophages.^13,14 Because BMDCs are the most potent antigen-presenting cells (APCs) known, their study may provide insight into the nature of the self-antigens in lupus.

The results of this study show that BMDCs are capable of processing and presenting U1A protein, resulting in the stimulation of autoreactive T cells of unprimed BWF₁ mice to proliferate and to secrete cytokines in an MHC class II-dependent manner in vitro. Furthermore, the major T-cell epitope was found to be located within the region containing amino acid residues 201–220 at the C-terminus of the U1A protein. The identification of an auto-T-cell epitope of U1A enabled us to further study the role of intramolecular and intermolecular epitope spreading in the diversification of the autoimmune response to the snRNP particle. The data presented here suggest that the approach used may also help in further analysis of the antigenic determinants of self-reactive T cells and in the exploration of strategies to design immunotherapy in SLE.

Materials and methods

Mice

All mice were obtained from and maintained by the Animal Centre of the College of Medicine of National Taiwan University in a pathogen-free facility. Female (NZB × NZW) F₁ (BWF₁) mice were purchased from Jackson Laboratories (Bar Harbor, ME). Female BWF₁ and DBA-2 × NZW F₁ mice, 6–8 weeks of age, were used as the source of BMDCs. DBA-2 × NZW F₁ (H-2^d/u) mice are a non-autoimmune strain bearing MHC class II molecules identical to those of BWF₁ (H-2^d/u) mice. To confirm that DBA-2 × NZW F₁ mice do not spontaneously develop autoimmune disease, 14 female mice were monitored until 17 months of age. Survival was 100% at 17 months. These mice did not have proteinuria and did not develop detectable anti-DNA and anti-U1A antibodies.

Protein antigens and peptides

A full-length cDNA encoding murine U1A was subcloned into the vector pET28a (+) (Novagen, Madison, WI). Histidine-tagged U1A was expressed by isopropylthio-β-d-galactoside (IPTG) induction and purified on nickel-agarose (His-Bind resin), according to methods described by the manufacturer. The U1A was analysed for purity on a Coomassie Brilliant Blue-stained polyacrylamide gel. The molecular weight of histidine-tagged U1A is ≈ 35 × 10³ Da. Recombinant histidine-tagged hepatitis D virus (HDV) surface antigen (used as a negative control for U1A protein) was kindly provided by P.-J. Chen (National Taiwan University, Taipei, Taiwan). Both purified U1A protein and HDV surface antigen were passed through an endotoxin column (Pierce Chemical Co., Rockford IL) to remove potentially contaminating bacterial endotoxins acquired during preparation. No proliferative response was observed when endotoxin-free U1A or HDV surface antigen was cocultured with splenocytes from naive or antigen-immunized BALB/c mice (data not shown).

The overlapping peptides of U1A were designed to be 20 amino acid residues in length, overlapping by 10 amino acid residues. These series of peptides, ovalbumin (OVA)_323−339 and the histidine TAG control peptide (32 amino acids) were synthesized and high-performance liquid chromatography (HPLC) purified by the Genemed Synthesis Company (South San Francisco, CA).

Human gamma globulin (HGG) and OVA (Sigma Chemical Co., St. Louis, MO) were used as foreign antigens. They were dissolved in phosphate-buffered saline (PBS) and stored at −20°.

Immunization of the mice

Three-month-old DBA-2 × NZW F₁ mice were immunized intraperitoneally (i.p.) with Freund's complete adjuvant (FCA) (Difco Laboratories, Inc., Detroit, MI) alone or with an emulsion of 10 µg of OVA in FCA. Antigens in Freund's incomplete adjuvant (FIA) (Difco) were used for subsequent booster immunization 2 weeks later. After 5 days, T cells were purified to perform proliferation assays.

Enzyme-linked immunosorbent assay for determination of the level of autoantibody

Antibodies specific for U1A were measured in serum samples by using a standard enzyme-linked immunosorbent assay (ELISA), as previously described.¹⁵ Briefly, U1A protein was diluted with PBS to a concentration of 5 µg/ml. After overnight incubation at 4°, the plates were washed and blocked with gelatin postcoating solution. Diluted serum was then added to the appropriate wells, which were incubated at room temperature for 2 hr. Horseradish peroxidase-conjugated goat anti-mouse γ-chain-specific antibodies (Sigma) were added. After 2 hr of incubation, 2,2′-azino-bisC3-ethylbenzthiazoline-6-sulphonic acid (ABTS) solution (Sigma) was used as substrate and the absorbance was measured at 405 nm. The level of anti-U1A IgG was calculated as ELISA units/ml (EU/ml) by comparison with the antihistidine TAG mAb (AD1.1.1) (Serotec Ltd, Kidlington, Oxford, UK). The absorbance value generated by 250 ng/ml of antihistidine TAG mAb was defined as 1 EU/ml of anti-U1A IgG.

Generation of dendritic cells from bone marrow cultures

BMDCs were prepared as described previously.¹⁶ Briefly, bone marrow cells from femurs and tibias were depleted of red cells by using ACK lysis buffer. Approximately 1 × 10⁶ cells were placed in 24-well plates in 1 ml of medium supplemented with recombinant murine granulocyte–macrophage colony-stimulating factor (GM-CSF) (500 U/ml) and interleukin (IL)-4 (1000 U/ml) (Pepro Tech Inc., Rocky Hill, NJ). The culture medium was RPMI-1640 supplemented with 5% heat-inactivated fetal calf serum (FCS), 4 mm l-glutamine, 25 mm HEPES (pH 7·2), 50 µm 2-mercaptoethanol, 100 U/ml of penicillin, 100 µg/ml of streptomycin and 0·25 µg/ml of amphotericin. On alternate days, medium was aspirated off to remove lymphocytes and fresh medium containing GM-CSF and IL-4 was added. On day 8 of culture, non-adherent cells (BMDCs) were collected and analysed by flow cytometry for the expression of MHC class II, B7-1, B7-2, CD11c and 33D1. The purity of BMDCs was > 65%. BMDCs were also tested for their ability to activate naive allogeneic T cells in vitro using a mixed lymphocyte reaction assay (data not shown).

BMDCs were incubated with different doses of the protein antigens, as indicated in the legends to the figures, on day 6 and then harvested on day 8. For peptide pulsing, BMDCs were incubated with peptides (100 µg/ml) for 3 hr at 37° on day 8. Cells were washed extensively to remove free antigens and irradiated with 2500 rads, then resuspended in AIM-5 medium (Gibco/BRL, Gaithersburg, MD) containing 1 × TCM (mouse serum replacement; Celox, St Paul, MN) to avoid non-specific stimulation. In addition, no peptide inhibited the mitogen responses when purified T cells were cocultured with peptide-pulsed BMDCs (100 µg/ml) or with irradiated splenocytes in the presence of U1A peptides (25 µg/ml).

Proliferation assays

Responder T cells were purified by either nylon wool or magnetic-activated cell sorter (MACS) methods. The enriched non-B cells isolated by passing splenocytes over nylon wool columns were incubated at 37° for 1 hr to remove macrophages. The purity of these T cells was analysed by flow cytometry: there were < 5% B cells and > 87% T cells. In some experiments, Thy1.2⁺ T cells were positively selected from splenocytes by anti-Thy1.2-coated magnetic microbeads (Miltenyi Biotec, Auburn, CA) according to the manufacturer's instructions. The purity of Thy1.2⁺ T cells was > 95%, as determined by flow cytometry, and the ratio of CD4⁺ to CD8⁺ T cells was ≈ 2 : 1. Purified T cells (1–2 × 10⁵ cells per well) were cocultured with BMDCs (2500 cells per well) in the presence or absence of anti-IA^d/E^d (2G9; PharMingen, San Diego, CA), anti-IA^b (25-9-3; PharMingen), anti-CD4 (GK 1.5; Leinco Technologies, Inc., St. Louis, MO) or anti-CD8¹⁷ (3.155; Leinco), for 4 or 5 days. The T-cell proliferation assays were conducted 4–7 days after coculture of purified T cells and syngeneic BMDCs. The optimal proliferation appeared at 4 or 5 days of culture (data not shown). During the last 4–6 hr of culture, 1 µCi of [³H]thymidine was added to each well. The cells were collected onto glass fibre filters using an automated multisample harvester. [³H]Thymidine incorporation was then measured in a dry scintillation counter (Packard Instrument Co., Meridan, CT). The stimulaton index (SI) was calculated by dividing the mean counts per minute (c.p.m.) incorporated in cultures of T cells plus antigen-pulsed BMDCs (in the presence or absence of blocking mAb) by the mean c.p.m. in control cocultures of T cells plus non-antigen-pulsed BMDCs. A positive response was defined as an SI of > 3·0.

For the experiments shown in Fig. 3(b), total splenocytes (3 × 10⁵ cells per well) from 12-week-old BWF₁ mice were cocultured with different concentrations of U1A proteins for 5 days. In separate experiments, Thy1.2⁺ T cells (1·5 × 10⁵/well) were incubated with syngeneic irradiated non-T cells (B cells + macrophages) (1·5 × 10⁵/well), with or without U1A peptides (25 µg/ml), for 5 days. Sixteen to 18 hours before harvesting, [³H]thymidine was added to each well. The non-T-cells (B cells + macrophages) from 2-month-old BWF₁ mice were prepared by negative selection with MACS (Miltenyi Biotec) using anti-Thy1.2 mAb (Miltenyi Biotec). The purity of non-T-cells (B cells + macrophages) (> 95%) was determined by using flow cytometry.

Figure 3.

Bone-marrow-derived dendritic cells (BMDCs) pulsed with U1A protein elicited an autoreactive T-cell response from unprimed BWF₁ mice in vitro. (a) Purified T cells isolated from unmanipulated 12-week-old BWF₁ mice or DBA-2 × NZW F₁ mice were cocultured with syngeneic BMDCs pulsed with increasing concentrations of U1A. After 5 days, the proliferation assay was performed. Results are expressed as stimulation index (SI) (mean ± SD, n = 3–6). The mean background (T cells plus non-antigen-pulsed BMDCs) counts per minute (c.p.m.) was 1375. Results are representative of two independent experiments. An asterisk indicates a statistically significant difference (P < 0·01) in [³H]thymidine incorporation. (b) BMDCs were incubated with U1A (10 µg/ml), human gamma globulin (HGG) (100 µg/ml), hepatitis D virus (HDV) surface antigen (20 µg/ml) or histidine TAG control peptide (1 µg/ml). Experimental conditions were similar to those described in Figure 3(a). Results are shown as SI (mean ± SD, n = 3–6). Background (T cells plus non-antigen-pulsed BMDCs) c.p.m. varied from 733 to 2016. Splenocytes (SC) from 12-week-old BWF₁ mice (n = 4) were cocultured with U1A proteins (0·1, 1 or 10 µg/ml) for 5 days. The value of baseline proliferation ranged from 253 to 550. Dotted horizontal lines indicate an SI of 3·0. These data are representative of four separate experiments.

Cytokine secretion

Culture supernatants were collected at 24, 48 and 72 hr following coculture of 2 × 10⁵ purified T cells and 2500 irradiated non-antigen-pulsed or U1A-pulsed BMDCs. The concentrations of IL-2, IL-4, IL-5 and interferon-γ (IFN-γ) were determined by using the Quantikine M ELISA Kit (R & D Systems, Minneapolis, MN). The level of cytokines from T cells stimulated with OVA or HGG-pulsed BMDCs was similar to that of T cells cultured with non-antigen-pulsed BMDCs (data not shown).

Generation of peptide-specific T-cell lines¹¹

Thy1.2⁺ T cells (1·5 × 10⁶/well) from 7-month-old BWF₁ mice were cultured with U1A peptide-pulsed syngeneic BMDCs (2·5 × 10⁴ per well) in serum-free medium (AIM-5 containing TCM) for 7 days in 24-well plates. On day 3, recombinant IL-2 was added to each well at a concentration of 50 U/ml. On day 7, the culture medium was replaced by RPMI-1640 containing 10% FCS and IL-2 (50 U/ml). On day 10, T-cell lines (5 × 10⁴ per well), harvested by Lympho-Paque (Accurate Chemical & Scientific Co., Westbury, NY), were restimulated with OVA or U1A protein-pulsed syngeneic BMDCs (1500 per well) for 5 days. During the last 4–6 hr of culture, 1 µCi of [³H]thymidine was added to each well.

Statistical analysis

The Wilcoxon test was applied to identify significant differences in the level of anti-U1A IgG in BWF₁ mice of different ages. The Mann–Whitney U-test was used to identify statistically significant differences in the proliferation assays between different groups. A P-value of < 0·01 was considered to be statistically significant.

Results

BWF₁ mice develop an antibody response specific to U1A protein

Initially, in order to study the role of snRNPs in BWF₁ mice, the level of the anti-U1A antibody response was monitored regularly. Mice were bled monthly, starting at 10 weeks of age. Anti-U1A IgG levels in sera were then determined by using an ELISA. Values that were greater than the mean value + 3 standard deviations (SDs) (horizontal line) from BALB/c mice were considered to be ELISA positive. As shown in Fig. 1, the earliest detectable anti-U1A IgG antibody response was seen at 12 weeks of age (in five of 15 animals), with 15 of 15 (100%) mice becoming antibody positive by 16 weeks of age. Therefore, the concentration of anti-U1A IgG dramatically increased from 10 to 20 weeks of age and slightly decreased from 20 to 28 weeks of age. This decrease may be because some of the mice, which had a high titre of anti-DNA IgG that was accompanied with a higher level of anti-U1A IgG, died. Furthermore, the titre of autoantibody showed a statistically significant increase (P < 0·01) in comparison to that of 10-week-old BWF₁ mice. In addition, the level of the anti-U1A antibody response in DBA-2 × NZW F₁ mice was similar to that of the age-matched BALB/c mice (data not shown). Therefore, the U1A protein in BWF₁ mice may gradually lose the B-cell tolerance present in early life. Based on the model of T–B diversification, which described the reciprocal T–B-determinant spreading in SLE, as proposed by several studies,^10,18–21 U1A-specific T cells might be activated when autoimmune responses expand and spread in the disease-developing BWF₁ mice.

Figure 1.

The level of anti-U1A immunoglobulin G (IgG) in BWF₁ mice over time with age. Sera obtained from 15 BWF₁ mice at each time-point shown were tested for anti-U1A IgG by enzyme-linked immunosorbent assay (ELISA), as described in the Materials and methods. Sera was diluted 1 : 100 for anti-U1A antibodies. Values that were greater than the mean ELISA units + 3 SD (horizontal line) from 4-month-old BALB/c mice (n = 6) were regarded as positive.

BWF₁ mice, but not non-autoimmune mice, develop a spontaneous T-cell proliferative response to U1A protein presented by BMDCs in vitro

To study the potential role of BMDCs as the APCs, T cells from OVA-immunized DBA-2 × NZW F₁ mice were freshly isolated and cocultured with OVA-pulsed syngeneic BMDCs (Fig. 2). A positive response was defined as an SI of ≥ 3·0. BMDCs were found to be able to process and present OVA peptides to stimulate the T-cell proliferative response in a dose-dependent manner. The data demonstrated that antigen-specific T cells isolated from DBA-2 × NZW F₁ mice did respond to antigen-pulsed syngeneic BMDCs in vitro.

Figure 2.

Bone marrow-derived dendritic cells (BMDCs) could present ovalbumin (OVA) to stimulate T cells from OVA-immunized DBA-2 × NZW F₁ mice. Purified T cells from phosphate-buffered saline (PBS) or OVA-immunized DBA-2 × NZW F₁ mice were cocultured with non-antigen- or OVA-pulsed syngeneic BMDCs for 5 days. [³H]Thymidine was added during the last 4–6 hr of culture. The T-cell proliferation was calculated as the stimulation index (SI) (mean ± SD, n = 3). Background (T cells plus non-antigen-pulsed BMDCs) counts per minute (c.p.m.) varied from 323 to 804. The dotted horizontal line indicates an SI of 3·0.

To examine the existence of U1A-specific T cells in BWF₁ mice, syngeneic BMDCs pulsed with different concentrations of U1A were used to detect the proliferative response of freshly isolated splenic T cells from BWF₁ and DBA-2 × NZW F₁ mice. As shown in Fig. 3(a), T cells from 3-month-old unprimed BWF₁ mice responded to U1A-pulsed BMDCs in a dose-dependent manner. In contrast, purified T cells from DBA-2 × NZW F₁ mice did not respond to U1A-pulsed syngeneic BMDCs. Purified T cells from BWF₁ mice responded to U1A antigen, but not to titrated concentrations of histidine TAG-control peptide, foreign HGG or recombinant HDV surface antigen processed and presented by BMDCs. Only the data of one concentration of each antigen is summarized in Fig. 3(b). In addition, no proliferative response was observed when splenocytes from the BWF₁ mice were cocultured with different concentrations of U1A antigen (Fig. 3b).

BMDCs pulsed with U1A stimulate the auto T-cell response in an MHC class II-dependent manner in vitro

To determine which population of T cells responded to the U1A-pulsed BMDCs, purified T cells from BWF₁ mice were stimulated with U1A-pulsed BMDCs in the presence of either anti-CD4 or anti-CD8 mAbs. As shown in Fig. 4(a), U1A-induced proliferation was completely inhibited by anti-CD4, but not anti-CD8, mAb, suggesting that CD4⁺ T cells were responsible for proliferation in the in vitro culture. To ensure that the proliferative response stimulated by U1A-pulsed BMDCs is MHC class II dependent, blocking antibody (2G9, specific for I-A^d/I-E^d molecules) was used to block the cognate interaction between the T-cell receptor (TCR) and MHC class II molecules. However, in order to exclude the effect of non-T cells contaminating the purified T-cell population, Thy1.2⁺ T cells were positively selected on a MACS column and the purity was > 95%. As shown in Fig. 4(b), the response was inhibited by anti-I-A^d/I-E^d antibody. In contrast, a control blocking antibody, specific for I-A^b molecules, had no effect on the interaction between the self-antigen pulsed BMDCs and the autoreactive T cells.

Figure 4.

U1A-pulsed bone marrow-derived dendritic cells (BMDCs) stimulated the proliferative response of autoreactive CD4⁺ T cells in a major histocompatibility complex (MHC) class II-dependent manner in vitro. Purified T cells (a) or Thy1.2⁺ T cells (b) from 12-week-old BWF₁ mice were cocultured with U1A (10 µg/ml)-pulsed syngeneic BMDCs in the presence or absence of blocking monoclonal antibody (mAb), as indicated above. The concentrations of mAbs were as follows: anti-IA^d/E^d mAb at 25 µg/ml, anti-IA^b mAb at 10 µg/ml, and anti-CD4 and anti-CD8 mAbs at 5 µg/ml. Responses are presented as mean stimulation index (SI) (± SD) from one of three similar experiments. The mean background counts per minute (c.p.m.) was 807 (a) and 1007 (b). Dotted horizontal lines indicate an SI of 3·0.

To explore whether BMDCs from DBA-2 × NZW F₁ mice can process and present U1A to stimulate U1A-specific T cells from BWF₁ mice, as both strains of mice are MHC matched, the T-cell proliferation assays were conducted using a combination of the BMDCs and T cells from each strain. As shown in Fig. 5, U1A-pulsed BMDCs from DBA-2 × NZW F₁ mice were able to stimulate the autoreactive T cells in BWF₁ mice (left). Conversely, we expected that BMDCs from BWF₁ mice would not elicit the proliferative response of T cells from MHC-matched, non-autoimmune DBA-2 × NZW F₁ mice. However, these T cells responded vigorously to either non-antigen- or antigen-pulsed BMDCs from BWF₁ mice, although the value of SI was < 3 (right). The background c.p.m. (non-antigen-pulsed BMDCs + T cells) of this combination for 3-, 5- or 7-day culture was 18 895, 62 310 or 55 895, respectively. Thus, the value of background c.p.m. in this experiment (Fig. 5, right) was much higher than that of syngeneic coculture. Notably, this phenomenon was only observed in BMDCs from BWF₁ mice cocultured with T cells from DBA-2 × NZW F₁ mice. Although more detailed studies are needed, this response might be mediated by minor histocompatibility antigens that were encoded by genes other than those of the H-2 locus.²²

Figure 5.

The T-cell proliferation against U1A stimulated by allogeneic bone marrow-derived dendritic cells (BMDCs). Thy1.2⁺ T cells from BWF₁ (left) or DBA-2 × NZW F₁ (right) mice at 6 months of age were cocultured with ovalbumin (OVA) (10 µg/ml) or U1A (10 µg/ml)-pulsed allogeneic BMDCs from DBA-2 × NZW F₁ (left) or BWF₁ mice (right) for 5 days, respectively. The mean background counts per minute (c.p.m.) was 3313 (left) or 62 310 (right). The data shown (mean ± SD, n = 3 each strain of mice) represent results obtained in two independent experiments. SI, stimulation index.

To further determine whether autoreactive T cells still exist in vivo in BWF₁ mice with the developed disease, purified T cells from 28-week-old mice were cocultured with U1A-pulsed BMDCs. As shown in Fig. 6, the proliferative response of these T cells could also be detected in autoimmune disease-developed BWF₁ mice.

Figure 6.

The T-cell response against U1A in young and old BWF₁ mice. Purified T cells from either 12-week-old or 28-week-old BWF₁ mice were cocultured with ovalbumin (OVA) (10 µg/ml) or U1A (10 µg/ml)-pulsed syngeneic bone marrow-derived dendritic cells (BMDCs). Data are expressed as the stimulation index (SI) (mean±SEM, n = 7 each group). The mean background counts per minute (c.p.m.) (T cells plus non-antigen-pulsed BMDCs) of young or old mice was 771 or 470, respectively. Dotted horizontal lines indicate an SI of 3·0. Results are representative of two separate experiments.

Autoreactive T cells stimulated with U1A-pulsed BMDCs secrete several kinds of cytokines in a dose-dependent manner in vitro

To further clarify the cytokine profile of U1A-specific T cells using BMDCs as APCs, the culture supernatants from the coculture of purified T cells and BMDCs were analysed for their cytokine content. The results are shown in Fig. 7. U1A-specific T cells from BWF₁ mice secreted IL-2, IL-4, IL-5 and IFN-γ in a dose-dependent manner. In contrast, the amount of cytokines secreted by purified T cells from unmanipulated DBA-2 × NZW F₁ mice was undetectable (data not shown). In addition, U1A-pulsed BMDCs elicited release of IL-2 and IFN-γ, as well as IL-4 and IL-5, which is a T helper 0 (Th0), or a mixed T helper 1 (Th1) and T helper 2 (Th2), pattern of cytokine production.

Figure 7.

Cytokine production of U1A-specific T cells from BWF₁ mice stimulated by bone marrow-derived dendritic cells (BMDCs). Purified T cells isolated from unmanipulated 12-week-old BWF₁ (n = 5, mean ± SD) mice were cocultured with irradiated syngeneic BMDCs pulsed with increasing concentrations of U1A, as indicated in the figure. The level of cytokines produced from ovalbumin (OVA)-pulsed BMDCs plus purified T cells was similar to that of non-antigen-pulsed BMDCs plus T cells. Culture supernatants were collected at 24 hr for interleukin (IL)-5, at 48 hr for IL-2 and interferon-γ (IFN-γ), and at 72 hr for IL-4. These data are representative of two independent experiments.

The auto-T-cell epitopes of the U1A protein are detected in BWF₁ mice, but not in DBA-2 × NZW F₁ mice, when BMDCs are used as APCs

We next used BMDCs to identify the T-cell epitope(s) of the U1A protein. A panel of 20-mer peptides were synthesized, each overlapping its neighbour by 10 amino acids and spanning the entire sequence of the U1A proteins. After peptide pulsing, the BMDCs were used to stimulate the Thy1.2⁺ T cells from unprimed BWF₁ or DBA-2 × NZW F₁ mice. As shown in Fig. 8(a), U1A_201−220 presented by BMDCs elicited a significant proliferative response with freshly isolated splenic T cells from all disease-developed BWF₁ mice. T cells from some BWF₁ mice, but not all individuals, were able to respond to U1A₃₁₋₅₀ or U1A_101−120. Furthermore, the magnitude of the proliferative response induced by U1A₃₁₋₅₀ and U1A_101−120 was lower than that induced by U1A_201−220. Thus, U1A_201−220 was the most dominant peptide. As expected, none of the peptides of U1A induced a proliferative response in T cells from DBA-2 × NZW F₁ mice (Fig. 8b). In addition, T cells from identical individuals did not respond to peptide-pulsed B cells isolated from young syngeneic mice (Fig. 8c). The control – OVA_323−339 – a known IA^d-binding peptide, also did not elicit any proliferative response in BWF₁ and DBA-2 × NZW F₁ mice (data not shown).

Figure 8.

Identification of auto-T-cell epitopes in the U1A protein using bone marrow-derived dendritic cells (BMDCs) as antigen-presenting cells (APCs). Thy1.2⁺ T cells from unprimed BWF₁ (a) or DBA-2 × NZW F₁ mice (b) at 10 months of age were cocultured with overlapping peptides of U1A-pulsed syngeneic BMDCs. T cells from identical BWF₁ mice were cocultured with purified non-T cells (B cells + macrophages) (c) in the presence or absence of U1A peptides. Results are expressed as the mean stimulation index (SI) ± SD calculated from six mice tested individually in two separate experiments. The mean background counts per minute (c.p.m.) was 657 (a), 1401 (b) or 242 (c). The c.p.m. elicited by OVA_323−339 was similar to that induced by non-antigen-pulsed BMDCs. Dotted horizontal lines indicate an SI of 3·0.

To determine whether the predominant peptides U1A_201−220 are naturally processed by BMDCs, T-cell lines were generated by the coculture of Thy1.2⁺ T cells from 7-month-old BWF₁ mice and U1A peptide-pulsed syngeneic BMDCs for 10 days. T-cell lines were then restimulated with U1A protein-pulsed syngeneic BMDCs. T cells from these disease-developed mice responded to U1A₃₁₋₅₀ and U1A_201−220, but not to U1A_261−280, in the primary proliferation assay (data not shown). As shown in Fig. 9, U1A₃₁₋₅₀- and U1A_201−220-specific T-cell lines were able to respond to U1A protein, but not to OVA processed and presented by BMDCs, although the magnitude of proliferative responses was not as high as that seen in the primary proliferation assay. In addition, T-cell lines stimulated with U1A_261−280-pulsed BMDCs did not respond to either OVA or U1A protein.

Figure 9.

Restimulation of peptide-specific T-cell lines with U1A protein-pulsed bone marrow-derived dendritic cells (BMDCs). T-cell lines were generated by the coculture of Thy1.2⁺ T cells from 7-month-old BWF₁ mice and U1A peptide-pulsed syngeneic BMDCs, as indicated on the x-axis. After 10 days, the T-cell lines were harvested and restimulated with ovalbumin (OVA) (10 µg/ml) or U1A protein (10 µg/ml)-pulsed syngeneic BMDCs. The mean background counts per minute (c.p.m.) ranged from 984 to 2053. Results are representative of two similar experiments and the values shown represent the mean stimulation index (SI) from triplicate wells.

Discussion

One of the several principal targets of autoimmune responses in human SLE and in murine models of the disease is the U1 snRNP particle.²³ U1A may be an initial target when an anti-snRNP response develops in MRL-lpr/lpr mice.⁴ Two major T-cell epitopes on U1A RNP have been identified in human lupus.²⁴ However, only limited studies have been carried out on the antigenic determinant of U1A in BWF₁ mice. This is, to our knowledge, the first report to demonstrate the existence of an anti-U1A IgG and U1A-specific CD4⁺ T cell in BWF₁ mice. In addition, this study shows that the use of BMDCs as APCs enables us to detect the existence of autoreactive T cells and furthermore it allows the identification of the auto-T-cell epitopes without the need for in vitro culture of T-cell lines or T-cell hybridomas, which is often employed in assays using other, less potent, APCs.^6,7,9 Importantly, once U1A epitopes have been identified in lupus mice, a disease model against snRNP particles can be generated in normal mice, permitting manipulations (that are not possible in humans) to mechanistically elucidate the development of autoimmunity.

Several studies have proposed a model of T–B diversification,^11,18–21 where autoreactive B cells, which take up an immunogenic particle such as intact U1 snRNP, drive the recruitment of a diverse population of CD4⁺ T cells. Thus, these T-helper cells provide help for B cells to produce antibodies of different specificities. The autoimmune response spreads in this way by reciprocal T–B stimulation until large cohorts of T and B cells have formed. Because antibodies to U1A arise first during the autoimmune response against snRNP particles,^4,25 U1A-specific T cells might play an important role in this autoreactive T–B interaction. According to a previous study, the region 201–241 of human U1A was reported to contain major T epitopes that were recognized by T cells from patients with mixed connective tissue disease.²⁴ The murine and human U1A proteins are 96% identical at the amino acid level, with the main difference occurring at the N-terminus.²⁶ Therefore, the sequence of murine U1A_201−220, which was identified in the present study, is exactly the same as the human U1A_196−215 sequence, which overlaps the 201–241 region at the C-terminus of human U1A.²⁴ It suggests that U1A-specific T cells from either human lupus or BWF₁ mice may recognize the same auto-T-cell epitope. It might also explain why normal mice immunized with native chimeric snRNPs (i.e. mouse snRNPs mixed with human U1A) can break tolerance and develop antibodies against mouse snRNPs.²⁵

Several studies^6–8,11,12 have shown that it is difficult to define auto-T-cell epitopes recognized by freshly isolated T cells from unprimed autoimmune mice when cocultured with self-protein-pulsed B cells or macrophages in vitro. There are two probable scenarios that may contribute to this difficulty. First, self-proteins may not be internalized and processed with sufficient efficacy to have adequate quantities of the relevant peptides for class II binding on B cells or macrophages in vitro.²⁷ To overcome this difficulty, large doses of overlapping peptides can be used to competitively replace non-irrelevant peptides that are predominantly displayed on the class II molecules of resting APCs.^7,27 Thus, APCs are able to express sufficient quantities of relevant peptide–MHC class II complexes to activate autoreactive T cells in vitro. However, it is possible that they may sometimes amplify some cryptic T-cell epitopes that are not major antigenic determinants in vivo. The second possible scenario is that the frequency of autoreactive T cells is too low to be detected when B cells or macrophages are used as APCs in vitro. A common strategy employed by a number of investigators for increasing auto-T-cell frequency is to generate autoantibody-inducing T-helper clones or hybridomas from mice developing lupus.^6,7,9,11,12 This enables detection of autoreactive T cells when irradiated splenocytes are used as APCs. However, some of the autoreactive T-cell populations may gradually disappear during in vitro culture. In addition, this method is time-consuming. Owing to these technical obstacles, it is difficult to clarify the autoepitope(s) of autoreactive T cells using this technique. In contrast, in the present study it was shown that BMDCs pulsed with a self-antigen, such as U1A protein, can be used to define antigenic determinants of freshly isolated self-reactive T cells in a more sensitive and definite manner in vitro.

U1A-pulsed BMDCs are able to stimulate T cells to secrete cytokines, as well as to proliferate. However, background levels of cytokine production have been observed, especially of IL-2 and IL-5 (Fig. 7). In theory, BMDCs can also present endogenous self-peptides by MHC molecules through uptaking bystander apoptotic cells,^28,29 even though there are no other antigens added during the in vitro culture. Therefore, this may indicate that the interaction of autoreactive T cells and non-antigen-pulsed syngeneic BMDCs may be sufficient to activate a small population of T cells to proliferate and to secrete a background level of cytokines. This may also be the reason why T-cell lines could be generated by non-antigen-pulsed or non-stimulatory peptide-pulsed BMDCs (such as U1A_261−280, as shown in Fig. 9). The population of T-cell lines enriched by U1A_201−220 or U1A₃₁₋₅₀ might also contain other specificities of autoreactive T cells. Therefore, the magnitude of proliferative responses after restimulation was not as high as that seen in the primary [³H]thymindine incorporation assay (as shown in Fig. 3). In addition, a previous study³⁰ has demonstrated that mechanical manipulation, such as pipetting and transferring BMDCs to fresh plates, induces maturation of BMDCs. It suggests that both U1A-pulsed and non-antigen-pulsed BMDCs used here had reached a similar level of maturation before they were cocultured with purified T cells. This point was also examined by phenotype analysis of BMDCs. U1A-, OVA- or non-antigen-pulsed BMDCs expressed similar levels of MHC class II (IA^d/IE^d), B7.1, B7.2 and CD40 molecules (data not shown).

In conclusion, our study has demonstrated that U1A-specific CD4⁺ T cells are present in disease-developing or disease-developed BWF₁ mice when BMDCs are used as APCs. Further characterization of these T cells would provide clues for the study of reciprocal T–B determinant spreading of snRNPs in lupus. In addition, the data presented here demonstrate the promise of the approach used in this study for further analysis of the antigenic determinants of self-reactive T cells and in the exploration of strategies for the design of lupus immunotherapy.^27,31 The findings of this study also suggest that this approach may more accurately reflect the real autoimmune response existing in vivo than that presented by the culture systems previously used.

Abbreviations

APC

antigen-presenting cell

BMDC

bone marrow-derived dendritic cell

BWF₁

(NZB × NZW) F₁

c.p.m.

counts per minute

ELISA

enzyme-linked immunosorbent assay

EU

ELISA unit

HGG

human gamma globulin

OVA

ovalbumin

SD

standard deviation

SI

stimulation index

SLE

systemic lupus erythematosus

snRNP

small nuclear ribonucleoprotein.