Contribution of αβ and γδ T cells to the generation of primary immunoglobulin G-driven autoimmune response in immunoglobulin-µ-deficient/lpr mice

Abstract

Class switch recombination (CSR) is a T-cell-dependent mechanism regulating isotype switching in activated mature B cells. Recently we showed that T-cell-independent CSRs occur spontaneously during B lymphopoiesis, but such cells are negatively selected by Fas signalling. In immunoglobulin µ-deficient mice, lack of Fas rescues isotype-switched B cells, resulting in generation of an autoimmune primary immunoglobulin G (IgG) repertoire in µMT/lpr mice. In the present study, we studied the role of αβ and γδ T cells in regulating this primary γH-driven repertoire. We found that a lack of αβ T cells significantly inhibited IgG production and autoimmunity in µMT/lpr mice, whereas a lack of γδ T cells resulted in augmented IgG production and autoimmunity. Also, a lack of T cells in µMT mice rescued isotype-switched B cells and serum IgG, probably owing to the lack of available FasL. We suggest that although CSRs in B-cell lymphopoiesis are T-cell independent, αβ T cells are important in the expansion of isotype-switched B-cell precursors and in promoting γH-driven autoimmunity, whereas γδ T cells regulate these cells.

Introduction

B-cell development in the bone marrow (BM) is guided by immunoglobulin gene rearrangements and the assembly of the B-cell receptor (BCR) genes.^1,2 The requirement of functional immunoglobulin H and L chain expression and signalling, in promoting B-cell development, has been demonstrated in several mutated mouse models (reviewed in refs ² and ³). In addition, B-cell development is controlled by negative selection, imposed by immune tolerance.^2,4,5 Thus, selection processes critically regulate B-cell development and are essential for preventing B-cell autoimmunity. In the periphery, B-cell tolerance mechanisms are controlled by T cells.¹ Upon BCR ligation by an antigen, appropriate T-cell help promotes mature immunoglobulin M (IgM)-expressing B cells to undergo proliferation, affinity maturation, class switch recombination (CSR) and differentiation to plasma or memory cells.⁶ In contrast, lack of T-cell help results in apoptosis, which is mediated by Fas signalling.^1,7–10 Thus, the Fas pathway is important in controlling activated mature B cells and in maintaining peripheral tolerance.

Mice deficient for functional Fas or Fas ligand (FasL) (lpr/lpr or gld/gld, respectively) are widely used to study the role of the Fas pathway in the immune system. These mice develop an autoimmune lupus-like disease, which is characterized by the presence of high titres of autoantibodies in serum, arthritis and glomerulonephritis.¹¹ Many studies have shown that both B and T cells are required for the development of autoantibodies and end-organ disease in lupus-prone mice.^12,13 However, αβ and γδ T cells contribute differently to murine lupus. While the αβ T cells play a central role in the pathogenesis of the disease and autoantibody production,^14–17 the γδ T cells are implicated in both propagation and regulation of the autoimmune response.^18–21

The role of Fas signalling at early stages of B lymphopoiesis is less known. Previous studies suggested that Fas has no role in the regulation of µH-driven B-cell development and central tolerance of IgM-expressing B cells.^22–24 However, we have recently shown that a small population of B-cell precursors undergoes spontaneous CSR, and these isotype-switched B cells are negatively regulated by Fas signalling.^25,26 If not deleted, these cells give rise to a primary γH-driven repertoire that may be subjected to further selection processes in the periphery. In mice where the µH transmembrane tail exon is targeted with a stop codon (µMT), B-cell development is blocked at the pro-B stage,²⁷ but lack of functional Fas rescues isotype-switched B cells and serum immunoglobulin G (IgG) production,^25,26 and leads to the generation of an oligo-monoclonal γH-driven autoimmune repertoire in µMT/lpr mice (Seagal et al., Int Immunol, in press). Lack of T cells in these mice suppressed IgG production significantly, but did not abolish it, suggesting that at least some CSR in B-cell development are T-cell independent, but further selection and expansion of autoimmune clones are regulated by T cells.²⁶ In this study we tested the role of αβ and γδ T cells in the development, selection and expansion of a γH-driven primary repertoire, and studied their contribution to the autoimmune response in µMT/lpr mice. Our results suggest that although CSR in B-cell lymphopoiesis is T-cell independent, αβ T cells play a critical role in the selection and expansion of isotype-switched B-cell precursors and in promoting γH-driven autoimmunity, whereas a lack of γδ T cells exacerbates the autoimmune response.

Materials and methods

Mice

Normal C57/BL6 (B6), C57/BL6-Fas^lpr/Fas^lpr (B6/lpr), C57/BL6-µMT/µMT (µMT) and C57/B6-TCR-β^–/–/TCR-δ^–/– (TCR-β^–/–/TCR-δ^–/–) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Mice carrying both lpr and µMT homozygous mutations (µMT/lpr) were generated by crossing B6/lpr and µMT mice. To generate µMT and µMT/lpr mice that are deficient in αβ or γδ T cells, or both, µMT and µMT/lpr mice were backcrossed with TCR-β^–/–/TCR-δ^–/– mice.^28,29 The progeny were screened by polymerase chain reaction (PCR) analysis to identify TCR-δ deficiency and by flow cytometry to identify TCR-β knockout animals by the lack of αβ T cells. All mice were on a C57/BL6 background. Mice were maintained under specific pathogen-free conditions and used at 3–5 months of age for the experiments.

Serum antibody detection

Detection of total IgG was performed by sandwich enzyme-linked immunosorbent assay (ELISA) using goat anti-mouse γ-chain specific reagents (Southern Biotechnology Associates, Birmingham, AL). Purified IgG provided the standard curve for calculation of antibody concentration, and results are expressed as µg/ml. Anti-chromatin IgG in serum was detected by ELISA, as previously described.²⁵ Affinity-purified mouse monoclonal IgG anti-double-stranded DNA (anti-dsDNA), 3H9, at a known concentration, served as a positive control and as a reference standard curve. Titres of anti-chromatin immunoglobulin for individual mice were calculated using the 3H9 antibody control standard curve and are expressed as µg/ml.

Enzyme-linked immunosorbent spot-forming cell assay (ELISPOT)

ELISPOT assays to detect antibody-forming cells (AFC) were performed as described previously.³⁰ Briefly, red blood cell (RBC)-depleted splenocytes were cultured in serial dilutions for 12–16 hr on nitrocellulose filters coated with mixture of goat anti-mouse kappa and goat anti-mouse lambda (2·5 µg/ml) (Southern Biotechnology Associates). Horseradish peroxidase (HRP)-labelled goat anti-mouse IgG (γ-chain specific) were used to detect bound antibodies. Spot signals were visualized by enhanced chemiluminescence (ECL). Spots were counted and AFC frequencies were calculated as the number of spots per 10⁶ spleen cells.

Western blot

Serum samples were analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) on a 10% gel. After transfer to nitrocellulose, membrane blots were blocked with 5% milk and reacted with biotinylated goat anti-mouse IgG (γ-chain specific) (Southern Biotechnology Associates). Streptavidin-conjugated HRP was used as a secondary probe, and specific bands were visualized by the ECL reaction.

Tissue array

Normal C57/BL6 mouse tissues were freshly frozen and homogenized in lysis buffer (1% Nonidet P-40, 0·1% sodium desoxycholate, 0·01% SDS in phosphate-buffered saline and protease inhibitors), followed by incubation for 30 min at 4°. After centrifugation, cleared tissue lysates were analysed by SDS–PAGE on a 12·5% gel. After transfer to nitrocellulose, membrane blots were blocked with 5% milk and then incubated with sera collected from normal mice or from TCR-δ^–/–µMT/lpr mice (sera were diluted 1 : 3000 and adjusted to contain ≈ 1 µg/ml of total IgG). HRP-conjugated goat anti-mouse IgG (Jackson Laboratories) was used for detection. Visualization of specific bands was performed by the ECL reaction.

Statistical analysis

Statistical differences between experimental groups were determined using the unpaired two-tailed Student's t-test.

Results

Serum IgG production in T-cell-deficient µMT/lpr mice

We have previously shown that at least some CSRs in B lymphopoiesis are T-cell independent, but T lymphocytes are required for the expansion and selection of isotype-switched B-cell precursors.²⁶ To directly test the role of αβ and γδ T cells in the rescue and selection of isotype-switched B-cell precursors, we generated µMT/lpr mice that were deficient in αβ, γδ, or both, T-cell subpopulations. Serum samples from mice at 3–5 months of age were assayed by ELISA to determine total IgG. It was found that a lack of Fas signalling results in the production of serum antibodies in µMT/lpr mice at levels that are significantly higher than those of normal mice (ref. ²⁵ and Fig. 1). However, the absence of γδ T cells results in significantly elevated levels (1·45-fold higher) of serum IgG relative to T-cell-sufficient µMT/lpr mice [1866 µg/ml relative to 1287 µg/ml (mean values), P<0·05], and 2·4-fold higher than in normal mice [1866 µg/ml relative to 778 µg/ml (mean values)]. In contrast, serum IgG levels were suppressed by 800-fold in µMT/lpr mice deficient in αβ T cells, relative to µMT/lpr mice [1·6 µg/ml relative to 1287 µg/ml (mean values)]. Interestingly, IgG levels in these mice were inversely correlated with the presence of γδ T cells. Thus, µMT/lpr mice, deficient in both αβ and γδ T-cell populations, developed significantly higher levels of serum IgG relative to µMT/lpr mice deficient in only the αβ T-cell population [29 µg/ml relative to 1·6 µg/ml (mean values), P < 0·05]. Further IgG isotypic analysis (Table 1) revealed that the elevation of serum IgG in µMT/lpr mice deficient in γδ T cells results primarily from the increased secretion of IgG1.

Figure 1.

Production of serum immunoglobulin G (IgG) in µMT/lpr mice lacking αβ, γδ, or both, T-cell subsets. Serum samples from 3–5-month-old mice, of the indicated genetic background, were collected and analysed by enzyme-linked immunosorbent assay (ELISA) to determine serum concentrations of total IgG. Concentrations were determined using an appropriate IgG standard curve, and results are expressed in µg/ml for individual mice and as group means. Each group contained six mice. *P-value of < 0·05 between µMT/lpr and µMT/lpr γδ^–/–, or between µMT/lpr αβ/γδ^–/– and µMT/lpr TCR-αβ^–/– mice.

Table 1.

Isotypic analysis of serum immunoglobulin G (IgG) from the indicated mice

Mouse strain	IgG1 (µg/ml)	IgG2a (µg/ml)
Normal	917·0±397·0	363·3±152·6
µMT	0·2±0·1	0·2±0·2
µMT/lpr	208·3±129·1	8166·6±3495·0
µMT/lpr γδ–/–	1632·9±1075·6	9174·0±4053·1
µMT/lpr αβ–/–	0·5±0·4	18·0±19·6
µMT/lpr αβ/γδ–/–	13·3±12·4	44·4±29·0

IgG1 and IgG2a serum concentrations were determined by enzyme-linked immunosorbent assay (ELISA) and calculated using an isotype-specific standard curve. Data are shown as mean ± standard error (SE), of at least five mice in each group.

Detection of AFCs in T-cell-deficient µMT/lpr mice

To confirm the serum IgG results, we determined the frequencies of IgG-producing AFCs in spleen tissue of these mice by ELISPOT, by using limiting dilution. As shown in Fig. 2, AFCs were detected in the spleens of all mice, except for µMT mice. Quantitative analysis revealed that a lack of γδ T cells significantly increased AFC frequencies by 1·3-fold relative to T-cell-sufficient counterparts (mean value of 464 relative to a mean value of 358, P < 0·1), whereas a lack of αβ T cells reduced it by 90-fold (mean value of 4·0 relative to a mean value of 358). As found for serum IgG, mice deficient in both T-cell populations produced 4·3-fold more AFCs relative to µMT/lpr mice deficient in only αβ T cells (mean value of 17 relative to a mean value of 4, P < 0·05). Our results suggest that the CSR in µMT/lpr mice is T-cell independent, but further selection and expansion of this primary γH-driven repertoire depends on αβ T cells and that γδ T cells may control this process.

Figure 2.

Frequencies of antigen-forming cells (AFCs) in µMT/lpr mice deficient in αβ, γδ, or both, T cell subsets. Spleen cells from the indicated mice, at 3–5 months of age, were analysed in an enzyme-linked immunosorbent spot-forming cell assay (ELISPOT) to determine AFC frequency. (a) Spleen cells, in serial dilutions, were placed on filters (10⁴−10⁷ cells/filter). Spots were counted on each filter (when possible, depending on the density of spots at each dilution), and the frequencies of IgG-producing AFCs were calculated and expressed as number of AFCs per 10⁶ spleen cells. Results are expressed as mean ± standard error of the mean (SEM) of at least five mice in each group. **P-value of < 0·1 relative to µMT/lpr mice, *P-value of < 0·05 relative to µMT/lpr αβ^–/–. (b) A representative ELISPOT membrane cultured with 10⁶ spleen cells from the indicated mice.

Increased autoimmunity in µMT/lpr mice deficient in γδ T cells

We have previously shown that µMT/lpr mice develop anti-chromatin serum reactivity and lymphadenopathy, which is correlated with age.²⁵ To test the importance of αβ and γδ T-cell populations in generating the primary γH-driven autoimmune repertoire in µMT/lpr mice, we studied the reactivity of serum samples from 3–5-month-old µMT/lpr deficient in αβ, γδ, or both, T-cell populations, to chromatin by ELISA. As previously shown, µMT/lpr mice at 3–5 months old produce low levels of anti-chromatin immunoglobulin (ref. ²⁵ and Fig. 3), which were significantly detected in 65% of these mice (seven of 11, Fig. 3). We found that a lack of γδ T cells exacerbates autoimmunity, as µMT/lpr mice deficient in γδ T cells produced significantly higher titres (20-fold) of anti-chromatin immunoglobulin (mean of 7·2 µg/ml relative to a mean of 146 µg/ml, P<0·05), which were detected in 100% of the mice (five of five). These mice also developed lymphadenopathy. It should be noted that high titres of anti-chromatin IgG and lymphadenopathy develop in 100% of µMT/lpr mice only at 5–7 months.²⁵ No chromatin serum reactivity or lympahdenopathy was detected in µMT/lpr mice deficient in αβ T cells only, or both αβ and γδ T cells.

Figure 3.

Elevated levels of anti-chromatin immunoglobulin G (IgG) in γδ T-cell-deficient µMT/lpr mice. Serum samples from 3–5-month-old mice from the indicated genetic background were analysed by enzyme-linked immunosorbent assay (ELISA) for anti-chromatin IgG, as described in the Materials and methods. Antibody concentrations were determined using the anti-double-stranded DNA (dsDNA) monoclonal antibody (mAb), 3H9, at known concentrations, and are expressed in µg/ml for individual mice and group means. Results are expressed as mean ± standard error of the mean (SEM) for at least five mice in each group. *P-value of < 0·05 relative to µMT/lpr.

Tissue reactivity in µMT/lpr mice deficient in γδ T cells

The anti-chromatin reactivity developed in µMT/lpr mice deficient in γδ T cells prompted us to look for other tissue antigens that are targeted by this autoimmune response. To do so we performed a tissue array and tested serum samples from 3–5-month-old µMT/lpr TCR-δ^–/– and control mice for additional tissue reactivity. Tissue array analysis revealed that all serum samples from normal mice showed the same pattern of bands, probably corresponding to immunoglobulin H and L chains found in the assayed tissues and/or to cross-reactive proteins (Fig. 4, top panel). However, serum samples from µMT/lpr TCR-δ^–/– mice revealed additional bands that were tissue specific (Fig. 4, centre and bottom panels, specific bands are indicated using arrows). For each µMT/lpr TCR-δ^–/– mouse, unique tissue reactivity was detected either by a single band in a single tissue (Fig. 4, centre panel), or by multiple bands in several tissues (Fig. 4 bottom panel). Such tissue reactivity was obtained for 100% of the µMT/lpr TCR-δ^–/– mice tested. It should be noted that similar tissue reactivity was obtained only in 30% of 3–5-month-old µMT/lpr mice (Seagal et al., Int Immunol, in press). Thus, as shown in other autoimmune models,^17–19 a lack of γδ T cells exacerbates the onset of autoimmunity, suggesting a role for γδ T cells in controlling the γH-driven autoimmune response in µMT/lpr mice.

Figure 4.

Tissue reactivity in γδ T-cell-deficient µMT/lpr mice. Serum samples from 3–5-month-old normal and µΜT/lpr TCR-δ^–/– mice were tested in a tissue array for tissue reactivity, as described in the Materials and methods. Shown are representative blot arrays (out of five in each group) obtained for a normal (top) and two different (3 months old) µΜT/lpr/TCR-δ^–/– serum samples (middle and bottom). Serum samples were diluted 1 : 3000 and adjusted to contain ≈ 1 µg/ml total IgG. Tissue-specific bands, identified by µΜT/lpr sera, are indicated by arrows. Tissues used are as follows: lane 1, brain; lane 2, heart; lane 3, skin; lane 4, kidney; lane 5, pancreas; lane 6, spleen; lane 7, muscle; lane 8, lung; lane 9, liver.

B-cell development and serum IgG production in µMT mice deficient in T cells

In µMT/lpr TCR-βδ^–/– mice, the small amounts of serum IgG observed probably correspond to the T-cell-independent CSR during B lymphopoiesis and survival of these cells owing to a lack of Fas signalling. To confirm this in a Fas-sufficient environment, a different approach was taken. As T lymphocytes are the primary source of Fas-ligand expression, we studied the development of B cells and serum IgG production in µMT mice that are deficient in αβ, γδ, or both, αβ and γδ T-cell subsets. We postulated that a lack of T cells would significantly limit the availability of FasL. That this experimental manoeuvre was successful is demonstrated in Fig. 5. We found that 3–5-month-old µMT mice, deficient in both αβ and γδ T cells, produce a significant amount of serum IgG, similar to that in µMT/lpr TCR-βδ^–/– mice (compare Fig. 1 with Fig. 5a). In the presence of only αβ or γδ T cells, very low levels of IgG were detected in 50–60% of the mice (3 µg/ml versus 10 µg/ml, respectively), although at much lower levels relative to mice deficient in both T-cell populations (23 µg/ml, P < 0·2, Fig. 5a). It should be noted that a high variation in the level of serum IgG was detected in mice from these two groups. The low levels of serum IgG that were detected in these mice were independently confirmed by Western blot analysis (Fig. 5b) and by the detection of AFCs in spleens of these mice (P < 0·1, Fig. 5c). Thus, the lack of Fas, or the lack of T cells (which are a major source of FasL), allows survival, but not expansion, of class-switched B cells and the production of low levels of serum IgG.

Figure 5.

Detection of serum immunoglobulin G (IgG) and antigen-forming cells (AFCs) in µMT mice deficient for different T-cell subsets. Serum samples from 3–5-month-old T-cell-sufficient µMT mice, and from µMT mice deficient in αβ, γδ, or both, T-cell subsets, were analyzed for IgG production. (a) The concentrations of serum IgG were determined by enzyme-linked immunosorbent assay (ELISA), using an appropriate IgG standard curve. The results are expressed in µg/ml for individual mice, and as group means. Each group contained at least four mice. *P-value of < 0·2 relative to µMT/lpr αβ^–/–. (b) Detection of the γH chain, by Western blotting, in serum samples that were found to contain IgG by ELISA screening. A total of 8 µl of serum samples from the indicated mice were used for detection. A representative blot is shown. (c) AFC frequencies in spleens of the indicated mice were determined by enzyme-linked immunosorbent spot-forming cell assay (ELISPOT). The results represent the mean ± standard error of the mean (SEM) of at least four mice in each group. *P-value of < 0·1 relative to µMT/lpr αβ^–/–.

Discussion

In µMT mice, B-cell development can be rescued only by a process of CSR, but such cells are negatively selected by Fas-mediated signalling.^25,26 In the absence of Fas, high titres of non-IgM autoimmune serum antibodies are produced and selection and expansion of this γH-driven primary repertoire is T-cell dependent.^25,26 The present study attempts to dissect the contribution of αβ and γδ T-cell subsets in generating the γH-driven B-cell autoimmunity in µMT/lpr mice. Our findings demonstrate that αβ T cells are required for the expansion and activation of the autoimmune γH-driven repertoire in µMT/lpr mice, whereas a lack of γδ T cells exacerbates the autoimmune response.

The µMT/lpr mouse model is very efficient for using to probe the nature of γH-driven primary repertoire, as in a normal mouse it is difficult to distinguish class-switched B cells generated in the BM from those generated during the germinal centre reaction. We demonstrated CSR in normal B lymphopoiesis in vivo and in vitro²⁶ and found it to be T-cell independent (ref. ²⁶ and this study). A similar survival of isotype-switched B cells was demonstrated in µMT mice overexpressing bcl-2.³¹ Although the presence of T cells is not an obligatory condition for the induction of isotype-switching in mature peripheral B cells, some form of activation is still required.^1,32 In developing B cells, however, it is yet to be determined whether signals that trigger CSR reflect a spontaneous mechanism aiming to generate a non-IgM primary repertoire, or perhaps these CSRs are the result of leakage of an imperfect biological system. Nonetheless, the survival and further development of these isotype-switched B-cell precursors is strictly regulated by Fas-mediated signalling, as shown previously.²⁶

Our data suggest that the selection and expansion of an autoimmune primary repertoire, driven by γH receptors, is T-cell dependent. There are several studies demonstrating a differential role of αβ and γδ T cells in regulating B-cell activation and formation of a secondary antibody repertoire. Thus, it was demonstrated that αβ T cells are especially important for generation of an optimal humoral response, while γδ T cells are less efficient in activating B cells, probably owing to an infrequent communication between APCs and antigen-specific γδ T cells.³³ In agreement with these findings, we demonstrate here that αβ and γδ T cells contribute differentially also in controlling the primary γH-driven repertoire. In the absence of αβ T cells, µMT/lpr mice develop very low frequencies of AFC and small amounts of serum IgG, whereas a lack of γδ T cells results in increased AFC numbers and significantly elevated IgG titres (Figs 1 and 2). The differential function of αβ and γδ T cells is also reflected in the induction and propagation of the autoimmune response. Previous studies by our laboratory suggested that µMT/lpr mice have a restricted B-cell repertoire, as these mice failed to respond to external T-dependent or T-independent antigens.²⁵ As αβ and γδ T cells have a different contribution in regulating the antibody repertoire driven by IgM-expressing B cells, it is important to further analyse their effect on the primary γH-driven repertoire that is generated in the T-cell-deficient µMT/lpr mice. Such repertoire analysis should address the context of clonal selection and expansion relative to biased, V(D) J recombination and the production of a skewed repertoire in B lymphopoiesis.

Previous studies in MRL/lpr mice have demonstrated that αβ T cells are especially important in autoantibody production,^{14–16,19,34,35} whereas the importance of γδ T cells in this process is controversial in the literature.^{18,20,21,36,37} This autoimmune response in MRL/lpr mice is T helper 1 (Th1) regulated.^38,39 Analysis of splenocyte somatic hybrids from MRL/lpr mice strongly implies that anti-dsDNA immunoglobulins arise in these animals by antigen-driven clonal expansion and somatic mutations.⁴⁰ The central role of CD4⁺αβ T cells in murine lupus was demonstrated by showing that the elimination or functional down-regulation of these cells significantly ameliorates the disease.^{15,16,18,19,37} In addition, a recent study suggests that a diverse repertoire of autoimmune αβ CD4⁺ T cells is required to generate a sustained effector B-cell response capable of mediating systemic autoimmunity.³⁴ In agreement with these studies, we have previously shown that murine lupus symptoms in µMT/lpr mice are Th1 regulated.³⁰ We now show here that αβ T cells are required not only for the generation and selection of a γH-driven repertoire, but also in shaping its autoimmune reactivity, as antibody titres found in αβ-deficient µMT/lpr mice were not autoreactive.

We found that a lack of γδ T cells exacerbate the autoimmune response in µMT/lpr mice, suggesting that γδ T cells may be involved in controlling the γH-driven autoimmune repertoire. Several studies describe the role of γδ T cells in regulating the autoimmune response. An absence of γδ T cells exacerbates murine lupus in MRL/lpr mice, and the γδ T-cell line derived from lupus-prone mice can trigger apoptosis of stimulated B cells, and can regulate the production of αβ T-cell-dependent anti-dsDNA autoantibodies.^18,36 Other studies have shown that γδ T cells are sufficient to initiate production of pathogenic anti-nuclear antibodies in MRL/lpr mice and to suppress the autoimmune response in interferon-γ (IFN-γ) transgenic mice.^18,19 A similar involvement of γδ T cells has been shown in other autoimmune diseases, including experimental allergic encephalomyelitis (EAE), arthritis and Graves' disease.^41–44 Anti-nuclear antibodies are thought to arise by a two-stage process, namely T-cell-independent polyclonal B-cell activation, followed by an αβ T-cell-dependent process of class switching and expansion of high-affinity autoreactive clones. In the µMT/lpr model we could not find any evidence for polyclonal B-cell activation (Seagal et al., Int Immunol, in press), but we clearly showed an important role of T cells in the expansion of isotype-switched B cells. The µMT/lpr mice deficient in γδ T cells develop an autoimmune antibody response earlier then the µMT/lpr control mice, as revealed by the anti-chromatin and tissue reactivity. This suggests that a lack of γδ T cells exacerbates autoimmunity driven by isotype-switched B cells. As each mouse developed an individual pattern of tissue reactivity (Fig. 4), it is possible that, in this mouse model, self-reactivity is required for T-cell-regulated selection, but the target self-antigen may have some degree of randomness. Such a response suggests that specific, perhaps Ag-driven, rather than polyclonal B-cell activation, is involved in the generation of a primary γH-driven autoimmune repertoire. B-cell repertoire analysis of individual µMT/lpr mice deficient in γδ T cells, as well as identifying the target tissue proteins revealed in the tissue array by each serum sample, should clarify this hypothesis. In the µMT/lpr mouse, this process may be αβ T-cell dependent, but regulated and suppressed by γδ T cells. In the absence of all T cells, IgG production was not completely abolished, but autoimmunity was not detected. This could be a result of the low levels of IgG or the lack of autoimmune specificities in the T-cell-deficient µMT/lpr mice. Both possibilities are not mutually exclusive and result from the role of the T cell in selection and expansion of autoreactive B cells, as has been demonstrated in different autoimmune models.^12,13

Many studies have shown that αβ T cells regulate the induction of CSR in mature activated B cells, selection into the germinal center and initiation of somatic hypermutation. This regulation is mediated by CD40L and the secretion of cytokines such as interleukin-4 (IL-4) and IFN-γ (reviewed in ref. 1). There are several studies showing that γδ T cells may also regulate these processes through the direct interaction and secretion of cytokines.^45,46 However, as suggested previously, the two T-cell populations may have differential stimulatory requirements and differ in their consequential help that is provided to the B cells.⁴⁷ How αβ and γδ T cells regulate the selection of isotype-switched B-cell precursors and the resulting primary IgG repertoire, is unknown. Clearly, CSR in B lymphopoiesis is T-cell independent, as IgG-expressing cells are generated in T-cell-deficient BM cultures of normal and µMT mice, as well as in µMT mice deficient in αδ and γδ T cells. The differential phenotype obtained in αβ-deficient and γδ-deficient µMT/lpr mice suggests that both T-cell populations may be involved in the regulation of isotype-switched B cells and in shaping the primary γH-driven antibody repertoire. The type of regulation may be different between each cell population, as described for IgM-expressing cells,⁴⁷ and may also affect the generation of autoimmunity.

One possible complication of using lpr mice is the physiological relevance of the results, as many abnormalities are developed in these mice. This is in addition to other potential complications in using mice with a mutated genome, such as µMT, TCR-β and TCR-δ mice. Hence, it is important to demonstrate the physiological relevance of the γH-driven pathway and its autoimmune potential in the most ‘normal’ conditions. We have previously shown that spontaneous CSR normally occurs in B-cell development in normal mice, as well as in µMT mice, in vivo and in vitro.²⁶ In addition, we found that blocking the Fas/FasL in µMT mice in vivo, using slow-release microcapsules loaded with soluble Fas, rescue IgG-expressing B-cell development and serum IgG production.²⁶ In the present study we further support these observations. As T lymphocytes are a major source of FasL,⁹ elimination of T cells resulted in the rescue of some isotype-switched B cells in µMT mice, but not in the expansion of these cells. Such expansion, as well as autoimmunity, is evident in T-cell sufficient, FasL-deficient µMT (µMT/gld) mice.²⁵ Thus, the lack of Fas or FasL allows isotype-switched B cells to develop, but further selection and expansion in the periphery is T-cell dependent. This suggests that B-cell development may compartmentalize the process of generation and selection of the γH-driven repertoire by anatomical sites as a possible mechanism for preventing autoimmunity.

Abbreviations

Ab

antibody

AFC

antibody-forming cell

BCR

B-cell receptor

BM

bone marrow

CSR

class switch recombination

ECL

enhanced chemiluminescence

ELISA

enzyme-linked immunosorbent assay

ELISPOT

enzyme-linked immunosorbent spot-forming cell assay, HRP, horseradish peroxidase

PCR

polymerase chain reaction

SDS–PAGE

sodium dodecyl sulphate–polyacrylamide gel electrophoresis

TCR

T-cell receptor