Systemic Autoimmune Disease Caused by Autoreactive B Cells That Receive Chronic Help from Ig V Region-Specific T Cells¹

The highly diversified Ig V regions (1) contain clonally unique antigenic determinants called idiotypes (2). Ig is Ag processed and V region peptides are presented to CD4⁺ Id-specific T cells on MHC class II molecules (3, 4). Id-specific CD4⁺ T cells have been found in mice with lupus (5, 6, 7) and in humans (8, 9, 10) suffering from autoimmune diseases like systemic lupus erythematosus (SLE)³ (8, 9), arthritis (10), and multiple sclerosis (11). B lymphoma cells (4, 12), B cells from Ig L chain transgenic mice (13, 14, 15), and normal B cells (15) spontaneously display endogenous V region Id peptides on their MHC class II molecules and activate Id-specific T cells. It has been suggested by our group (14, 16) and others (6, 7) that Id-specific T cells may cause chronic autoimmune disease by helping autoreactive Id⁺ B cells. However, this possibility has been difficult to test, and is hitherto unproven in normal mice, due to low frequency sets of matching Id-specific T cells and Id⁺ B cells.

Taking a reductionistic approach to solve the low frequency problem, Id-driven T-B collaboration has recently been studied by us (13, 14, 17) and others (18) by use of two different models of paired Ig L chain and TCR transgenic mice with similar results. In particular, by use of adoptive transfer of lymphocytes, it was shown that transient Id-driven T-B collaboration resulted in switched IgG autoantibodies in reconstituted mice (17, 18). Ligation of the BCR was shown to be essential for IgM→IgG switch to occur (17). On this basis, a model was advanced in which autoreactive B cells that have their BCR ligated by autoantigen, and that simultaneously receive help from Id-specific T cells, differentiate into plasma cells producing switched IgG autoantibodies (17).

Spontaneous B cell presentation of Id peptides on class II molecules should be of no physiological consequence if T cells were tolerant. Indeed, T cells appear to be tolerant to abundant germline-encoded V region sequences (3, 19) in part due to deletion in the thymus (13). By contrast, T cells readily responded to rare Id sequences that express somatic mutations (3, 19, 20). Thus, although contracted in scope by T cell tolerance, Id-driven T-B collaboration could potentially occur in normal individuals, as well as in patients.

The recent demonstrations of transient Id-driven T-B collaboration were based on transfer of lymphocytes, and T cell tolerance mechanisms should not have come into play (17, 18). In this study, we have asked what happens when Id⁺ B cells and Id-specific T cells coexist in an individual from early ontogeny. This was accomplished by investigating doubly transgenic mice derived from paired Id-specific TCR transgenic and Ig L chain transgenic mice. More specifically, the Id⁺ mice are transgenic for a λ2 Ig L chain gene derived from the MOPC 315 myeloma (21); this λ2³¹⁵ L chain expresses a rare non-germline 91–101 CDR3 Id-peptide that is presented on an MHC class II molecule (I-E^d) to CD4⁺ cells (3). Importantly, because the Id⁺ mice are transgenic only for the λ2³¹⁵ L chain, they express a polyclonal H chain repertoire endowing their B cells with diverse BCR specificities, including autoantibody specificities (17). The transgenic αβ TCR is specific for the Id(λ2³¹⁵)-I-E^d complex (13).

The results show that although Id-specific T cell tolerance is profound in doubly transgenic mice, Id-reactive T cells increasingly escape negative selection with age. Due to this incomplete tolerance of Id-specific CD4⁺ T cells, chronic Id-driven T-B collaboration develops which results in production of IgG autoantibodies of diverse specificities and severe systemic autoimmune disease. These observations suggest that chronic Id-driven T-B collaboration may cause autoimmune disease, and may explain the significance of Id-specific T cells previously observed in autoantibody-associated diseases like SLE (5, 6, 7, 8, 9) and rheumatoid arthritis (10).

The hemizygous λ2³¹⁵ Ig L chain transgenic mice (Id⁺ mice) (21) are transgenic for the L chain genes of the MOPC 315 myeloma. The hemizygous αβ TCR transgenic mice (Vα1, Jα19; Vβ8.2, Jβ1.2) are specific for the Id(λ2³¹⁵)-peptide presented on I-E^d MHC class II molecules (13). Both transgenic strains are on a BALB/c background (>20 backcrosses). Homozygous Id⁺ mice and Id-specific TCR transgenic mice have been produced (B. Bogen, unpublished observation). TCR transgenic SCID mice were on a C.B-17 scid/scid background (14). BALB/c, C.B-17 SCID, and BALB/c Rag2^−/− mice were from Taconic M&B. OVA-specific DO.11.10 transgenic mice (22) were purchased (Jax Mice; The Jackson Laboratory). The following crosses were performed: 1) hemizygous Id-specific TCR transgenic (female) were crossed with hemizygous Id⁺ transgenic (male) to yield Id⁺ transgenic, TCR transgenic, wild-type BALB/c, and doubly transgenic offspring at Mendelian ratios; 2) TCR transgenic (homozygous (female)) were crossed with Id⁺ transgenic (homozygous (male)); and 3) Id⁺ transgenic (homozygous (female)) were crossed with TCR transgenic (homozygous (male)). In the second and third cross, all offspring were doubly transgenic; however, note that in the third but not the second, Id⁺ Ig was transferred to doubly transgenic offspring in utero. As a fourth option, offspring derived from a homozygous TCR trangenic (female) and a hemizygous Id⁺ transgenic (male) cross were also monitored and autopsied for pathology at Taconic M&B. In the latter experiment, serum samples were used to type mice as TCR transgenic or doubly transgenic in a blinded fashion after the experiment was terminated. The mice were either kept under specific pathogen-free conditions (M&B) or in a minimal disease unit located in Oslo, Norway. The Norwegian Animal Research Authority approved these experiments.

ELISAs have previously been described (21) for measurement of: Ig with either λ2 or λ3 (λ2/3) L chains; IgM with λ1, λ2/3, or κ; IgG1 with λ1, λ2/3, or κ; IgG2a with λ1, λ2/3, or κ; and IgG2b with λ1, λ2/3, or κ. Because expression of endogenous λ1, λ2, and λ3 L chain is negligible in λ2³¹⁵ transgenic mice (21), ELISAs based on mAbs with specificity for Vλ1/Vλ2 (9A8 mAb) and λ2/λ3 (2B6 mAb) measure almost exclusively transgene encoded λ2³¹⁵ Ig in these mice, i.e., Id⁺ Ig.

In ELISAs for detection of autoantibodies or immune complexes, wells were left uncoated (control) or were coated with recombinant human La, RO52 or RO60 (kindly provided by Dr. R. Jonsson, Broegelmann Research Laboratory for Microbiology, Bergen, Norway), purified human proteinase 3 (Wieslab), calf thymus histone IIA (Sigma-Aldrich), various purified mouse IgG isotypes (rheumatoid factor ELISA), or human C1q (Enzyme Research Laboratories). In some experiments, we used plates precoated with human extractable nuclear Ags (ENA; Sm, RNP, SS-A, SS-B, Jo-1, Scl-70; Shield Diagnostics). Diluted sera were added, and binding of autoantibodies was detected with secondary mAbs. The H chain-specific mAbs bio-mouse anti-mouse IgM^a (DS-1), bio-rat anti-mouse IgG1 (A85-1), bio-rat anti-mouse IgG2a (R19-15) were from BD Pharmingen. Negative controls for calculations of titers were sera from Id⁺ transgenic, TCR transgenic, and wild-type littermates. Individual doubly transgenic sera were compared with nine nondoubly transgenic littermate sera, and scored as positive if clearly above 95% confidence interval of these controls. The ELISAs were rerun several times and further tested against additional sets of nine control sera.

The following mAbs were affinity-purified and biotinylated or FITC-conjugated in our laboratory: transgenic TCR clonotype-specific GB113 (13), anti-CD4 (GK1.5), anti-CD8 (53.6-7.2), anti-Thy 1.2 (T24), anti-DEC205 (NLDC145), anti-CD11c (HB-224), 33D1 (TIB-227), anti-CD11b (TIB198), anti-F4/80 (HB198), anti-MHC class II (TIB120), 9A8 (anti-Vλ1/Vλ2), 2B6 (anti-Cλ2/Cλ3), anti-μ (Bet 2), and anti-κ (187.1). FITC anti-CD4 (RM4-5), PE anti-CD4, allophycocyanin anti-CD4, PE anti-CD69 (H1.2F3), PE anti-CD45R/B220 (RA3-6B2), allophycocyanin anti-CD45R/B220, PE anti-CD19 (1D3), bio anti-CD19, PE anti-CD11b (M1/70), allophycocyanin anti-CD11b, bio anti-IgM^a or PE anti-IgM^a (DS-1), bio anti-IgG1^a (10.9), bio-rat anti-mouse IgG1 (A85-1), bio anti-IgG2a^a (8.3), bio anti-IgG2a (R11-89), bio anti-IgG2b (R12-3), bio anti-IgG3 (R40-82), streptavidin-CyChrome, and streptavidin-PerCP were all from BD Pharmingen. Anti-BrdU FITC was from BD Biosciences. FITC-conjugated peanut agglutinin (PNA) was from Vector Laboratories. Quadruple stainings were performed with FITC-, PE-, PerCP-, and allophycocyanin-conjugated mAbs and acquired on FACSCalibur or FACSVantage (BD Biosciences) and files analyzed with WinMDI 2.7.

Stainings of organ sections were performed by embedding organs in OCT, and 5-μm frozen sections were mounted on l-polylysine-coated glass slides, air-dried overnight, blocked with 30% heat-aggregated rat serum, and stained with biotinylated or FITC-conjugated mAbs (see earlier discussion), Alexa Fluor 488 and 546 goat anti-mouse IgG (F(ab′)₂), Alexa Fluor 488 goat anti-mouse-γ2a, and Alexa Fluor 546 goat anti-mouse-γ1 (Molecular Probes), streptavidin-Cy2 and Cy3 (Amersham), 4′,6′-diamino-2-phenylindol-hydrochloride (DAPI; Molecular Probes). Kidney tubuli cells were in some experiments counterstained with rhodamine-conjugated PNA (Vector Laboratories). FITC-conjugated F(ab′)₂ goat anti-mouse C3 was from Cappel/ICN. Diluted sera were added to HEp2 cells (Immuno Concepts), to monkey kidneys and esophagus sections (The Binding Site), to rat stomach/kidney sections, or to SCID mouse kidney sections prepared in our laboratory. Abs were detected with FITC anti-IgM (II/41; BD Pharmingen), Alexa Fluor 488 goat anti-IgG2a and Alexa Fluor 546 goat anti-IgG1 (Molecular Probes). Photographs were made with a charge-coupled device camera (Hamamatsu).

Positive selection of B and T cells was done by flow cytometric sorting of B cells on a FACSVantage by sterile double stainings of spleen and lymph node suspensions and acquisition of B220⁺ or Thy1⁺ cells, respectively. Cells were acquired from within a lymphocyte forward and side light scatter gate. B and T cell purity and appropriateness of the gates were tested as described (14) and measured >97%.

Id-specific doubly transgenic T cell lines were derived from doubly transgenic or TCR transgenic mice by stimulating lymph node or spleen cells as discussed in the presence of lymphokines (murine rIL-2, 20 U/ml). Th2 TCR transgenic lines were generated in the presence of IL-4 (14). Proliferation of B cells in T-B collaboration experiments was measured by culturing B cells with irradiated T cells (1000 rad), with a [³H]TdR overnight pulse from day 3 to 4. Ig secretion was assayed by culturing splenic B cell (5 × 10⁵/ml) in the presence or absence of T cells (1 × 10⁵/ml) for 6 days. Supernatants were removed and Ig concentrations measured by ELISA. T cell proliferation assays were as described, in the presence or absence of IL-2 (4 U/well), as indicated (14). BrdU was provided as an i.p. injection (1 mg) and continuously in the drinking water (1 mg/ml) (14). Blood, lymph nodes, and spleens were harvested as described (14).

T or B cells, or bone marrow cells thrice washed and resuspended in protein-free RPMI 1640, were filtered through a sterile nylon sheet (40 μm mesh) and injected in a volume of 0.15 ml into TCR transgenic SCID, SCID, or BALB/c Rag2^−/− mice tail veins as indicated.

Six to 8-wk-old singly transgenic Id⁺ or nontransgenic Id⁻ mice were sacrificed and bone marrow was gathered by flushing RPMI 1640 (Sigma-Aldrich) through the femur shafts. Cellular suspensions were treated with an erythrocyte lysis buffer, cells were washed four times in RPMI 1640, and were filtered through a nylon mesh with pore diameter 40 μm as described (23). Four million bone marrow cells were injected i.v. into TCR transgenic SCID mice or into nontransgenic SCID littermates.

Colons were flushed to remove feces, minced into small pieces with scissors, placed in petri dishes in RPMI 1640 with 10% FCS containing collagenase (400 U/ml; Sigma-Aldrich) and DNase (0.33 mg/ml; Sigma-Aldrich) at 37°C for 15–20 min. Within this incubation period, cells were released by intermittent pipetting with a wide-bored pipette. Cell suspensions were thereafter washed before use.

We followed offspring from a hemizygous Id-specific TCR transgenic (female) crossed with a hemizygous Id⁺ transgenic (male) for disease development. Doubly transgenic progeny failed to thrive from 40 to 60 days of age, lost weight and died prematurely with a median survival of 90–110 days for males and 120–180 days for females. By contrast, singly transgenic and nontransgenic mice developed normally, had a normal longevity, and were healthy (Fig. 1, A and B). The microbiological status (specific pathogen-free or minimal disease unit) of the animal housing facilities had no major influence on disease development (compare Fig. 1, B and C). Transfer of maternal Id⁺ Ig had no impact on disease induction because results were the same regardless of whether the mother of the doubly transgenic mice was singly Id⁺ transgenic or singly TCR transgenic (Fig. 1 D).

Most doubly transgenic mice developed skin and gut disease, whereas fewer developed arthritis (Fig. 1,E). The skin lesions first appeared around day 40, but were highly variable between mice. Skin disease was first seen in the periorbital area, thereafter spreading to the snout and face, and in severe cases to the back or abdomen. Skin was inflamed, with hair loss and scarring (Fig. 2,A). Gut disease manifested itself as anal inflammation with bloody mucoid diarrhea, eventual rectum prolapses, and grossly enlarged colons upon autopsy (Fig. 1,E). Arthritis was characterized by marked swelling, erythema, and extensive bone erosion and pannus formation (Figs. 1,E, and 2, B and C).

Doubly transgenic mice were hypergammaglobulinemic with greatly elevated levels of Id⁺IgG1 and Id⁺IgG2a (p < 0.03, Mann-Whitney U test) whereas Id⁺IgM level was only moderately increased (Fig. 3,A, Id⁺ denoting transgenic λ2³¹⁵ L chain). Autoantibodies were increasingly detected with age; by 90 days almost all doubly transgenic mice had high titers of IgG1 and IgG2a autoantibodies binding HEp-2 cells, but far less IgM autoantibodies, suggestive of a mixed Th1 and Th2 involvement (Fig. 3, C and D). Autoantibodies bound nuclear and cytosolic Ags with a variety of staining patterns (L. A. Munthe and B. Bogen, manuscript in preparation). Antinuclear Abs were directed against a mix of six ENA that included Sm, RNP, SS-A, SS-B, Jo-1, Scl-70 Ags, as well as H2A and Ro52 (Fig. 3,D), and dsDNA (Crithidia luciliae stain) (Fig. 3,E). Moreover, IgG autoantibodies bound tissue sections with diverse staining patterns including anti-mitochondrial and anti-basal lamina (L. A. Munthe and B. Bogen, manuscript in preparation). More than one-third of doubly transgenic mice had IgG immune complexes in sera (Fig. 3,B) and deposition of IgG and C3 in kidney glomeruli (Fig. 2, D and E).

The development of autoimmune disease led us to reexamine the Id-specific T cell repertoire in doubly transgenic mice. Using two-fluorochrome flow cytometry, we have previously found that 3- to 5-wk-old doubly transgenic mice expressed high levels of both Id⁺ B cells and Id⁺ Ig, and had a pronounced thymic deletion of Id-specific T cells and extinguished peripheral Id responsiveness (13). With four-color flow analysis, we now confirm the previous findings of thymic negative selection because there was a striking deficiency of Id-specific CD4⁺CD8⁺ cells and Id-specific CD4 single positive T cells (Fig. 4, A and B). Nevertheless, Id-specific CD4⁺CD8⁻ (CD4 single positive) thymocytes were present in small amounts, ≈10⁵ cells/thymus (Fig. 4, A and B). This indicates that Id-specific negative selection is efficient in doubly transgenic mice, but not complete.

To investigate whether any mature Id-specific CD4 single positive thymocytes actually escaped to the periphery in doubly transgenic mice, we stained lymph nodes and spleens. Doubly transgenic mice had relatively normal frequencies and numbers of CD4 cells, but nearly all of these expressed endogenous TCR α- or β-chain. In fact, one-half of CD4⁺ cells lacked completely the transgenic TCRβ, which had been replaced with endogenous β-chains like Vβ6 (Fig. 4, C and D). The other half had reduced expression levels (Fig. 4,C, arrow). This observation is strikingly different from singly TCR transgenic mice in which almost all T cells express transgenic TCRβ and no endogenous β-chains (Fig. 4,C). Expression of endogenously rearranged TCR α-chains on peripheral CD4⁺ cells was similarly increased (data not shown). Nevertheless, despite the extensive expression of endogenous α- and β-chains in doubly transgenic mice, it was possible to find low numbers of Id-specific CD4⁺ T cells that stained with the transgenic TCR αβ (α_Tβ_T)-specific GB113 mAb and the transgenic TCRβ-specific F23.1 mAb (Fig. 4,E). Thus by day 20, ∼1.5% of CD4⁺ T cells expressed α_Tβ_T compared with 70% in the TCR transgenic controls. This low frequency of Id-specific CD4⁺ T cells increased to 4% in 120-day-old mice (Fig. 4,E). Similar findings were made for the spleen (Fig. 4 F). Thus, although reduction of peripheral Id-specific CD4⁺ T cells in doubly transgenic mice was pronounced, some few cells, whose Id-specific TCR were diluted with endogenous TCR α- or β-chains, could be increasingly found with age.

Consistent with previous observations (13), ex vivo doubly transgenic lymph node and spleen cells did not proliferate in response to synthetic Id-peptide whereas singly TCR transgenic T cells responded intensely (data not shown). However, a very minor Id peptide-specific response of doubly transgenic lymph node cells could be obtained when IL-2 (4 U/well) was added to the cultures (Fig. 5,A), and the high background proliferation in absence of Id-peptide may be explained by T cells being activated in vivo (see below). However, because only very few Id-specific T cells are present in doubly transgenic lymph nodes, the proliferation of lymph node cells from doubly transgenic mice is not directly comparable to lymph nodes from singly TCR transgenic mice that have a high frequency of Id-specific T cells. We therefore generated short-term T cell lines cultured in the presence of irradiated APC, Id-peptide and IL-2; in such lines, more than three-fourths of T cells expressed the transgenic α_Tβ_T (data not shown). Nevertheless, when tested in the absence of IL-2, these doubly transgenic T cell lines proliferated less vigorously and required a much higher concentration of Id peptide than did T cell lines similarly obtained from TCR transgenic mice (Fig. 5,B). Expression of endogenous TCR chains that diluted Id-specific TCR expression in doubly transgenic T cells (Fig. 4, C–E) could have contributed to the reduced Id reactivity (23). Surprisingly, despite being only weakly reactive in terms of proliferation, doubly transgenic T cell lines efficiently helped Id⁺ B cells to proliferate and secrete Id⁺ Ig in an Id-dependent manner (Fig. 5 C).

When examining lymph node and spleen cells in diseased 120-days-old doubly transgenic mice, we found that Id-specific T cells were activated (CD69⁺) blasts, proliferated in vivo (incorporated BrdU), and had a memory (CD62L⁻CD44^high) phenotype (Fig. 6,A, and data not shown). This result is consistent with the high spontaneous background proliferation ex vivo of doubly transgenic lymph node cells in the absence of added Id peptide (Fig. 5 A). The detection of activated Id-specific T cells was age-dependent because far fewer BrdU⁺CD69⁺ Id-specific T cell blasts were found in healthy 20-days-old doubly transgenic mice (data not shown).

Id⁺ B cells in diseased 120-days-old doubly transgenic mice were, like Id-specific T cells, also activated (CD69⁺) and proliferated (BrdU⁺) (Fig. 6,B). A similar but much less frequent activation was found in healthy 20-days-old doubly transgenic mice (data not shown). Moreover, germinal center B cells and plasma cells were more frequent in doubly transgenic mice as compared with controls (Fig. 6 C). Taken together, both Id-specific T cells and Id⁺ B cells were activated in vivo in doubly transgenic mice, consistent with ongoing chronic Id-driven T-B collaboration.

We proceeded to investigate whether Id-driven T-B collaboration occurred in the diseased organs. In 6-wk-old doubly transgenic mice, colons were no different from singly transgenic or nontransgenic littermates (Fig. 2,F, inset, and data not shown). By contrast, in diseased mice (>120 days), colons were inflamed and contained cellular infiltrates that increased the size and cellularity of colons over 10 times (Fig. 2,F, note that the healthy colon (inset) is in the same scale). The inflammation was accompanied by grossly modified colon mucosal architecture, mucosal hyperplasia, infiltration of lymphocytes and macrophages, loss of goblet cells, increased vascularization, and areas of epithelial erosion (Fig. 2, F–H). Age-matched singly TCR transgenic, singly Id⁺ transgenic, or wild-type BALB/c littermate colons, showed no histological pathology.

Localized lesions involved discrete epithelial areas and could be seen as major erosions with effusion of mononuclear cells (Fig. 2,G) or as small point lesions found immediately above lymphoid follicles (Fig. 2,H). We reasoned that these isolated lesions could be caused by local immune responses and autoantibody production. The finding of focal C3 and Id⁺ Ig deposition near lymphocyte aggregates supported this hypothesis. Such deposits could be linear (Fig. 2,I, top) or diffuse surrounding lymphocyte aggregates (Fig. 2 I, bottom).

CD4⁺ T cells were present in great numbers in inflamed colons of 120-day-old doubly transgenic mice. Although most of these CD4⁺ T cells expressed endogenous α- and β-chains, bona fida Id-specific T cells were readily detectable and were activated (Figs. 7,A, and 2, J and L). In fact, doubly transgenic mice had greater numbers of Id-specific CD4⁺ T cells in the colon than did singly transgenic mice, whereas the reverse was true for lymph node and spleen (Fig. 4). Thus, T cell tolerance was not only increasingly broken with age, but Id-specific T cells localized to a large extent to diseased organs like the colon (Fig. 7,C). Moreover, the Id-specific CD4⁺ T cells in inflamed colons of doubly transgenic mice expressed the activation marker CD69, suggesting an ongoing activation (Fig. 7,A). Furthermore, clusters of blasts of Id-specific CD4⁺ T cells could be seen, suggesting that they proliferate in situ (Fig. 2,J). Consistent with this finding, ex vivo colon T cells responded to Id stimulation with proliferation, without addition of exogenous IL-2 (data not shown), which is in contrast to the unresponsiveness of lymph node or spleen T cells (see previous discussion and Fig. 5 A).

Id⁺ B cells in the lamina propria and submucosa of doubly transgenic mice (>120-day-old) were greatly increased in numbers and were activated CD69⁺B220⁺ MHC class II⁺ blasts (Fig. 7,B). As with Id-specific CD4⁺ T cells, Id⁺ B cells preferentially localized to the colon (Fig. 7,C). Moreover, the number of Id⁺IgG1⁺ and Id⁺IgG1⁺ cells was greatly increased, and Id⁺ B cells could be found in clusters of varying sizes, or as isolated cells, throughout the colon tissue (Fig. 2 K).

The highly frequent Id⁺ B cells in the colons were often surrounded and in tight contact with activated Id-specific CD4⁺ T cells (Fig. 2,L). Moreover Id⁺ Ig and C3 deposition as well as epithelial erosions were often found at such sites of Id-driven T-B collaboration (Fig. 2 I).

Whether autoimmune disease in doubly transgenic mice is caused by Id-driven T-B collaboration, it should be transferable by T and B cells. Indeed, disease could be transferred with sorted T and B cells from lymph nodes and spleens, but not B cells alone, when injected i.v. into BALB/c Rag2^−/− mice (Fig. 8). Within weeks, all recipients of T and B cells developed similar disease as that of the doubly transgenic donors and eventually died (Fig. 8, A and C). Disease phenotype was transferred to recipients, e.g., as seen for skin disease in Fig. 8,C. Disease in recipients was accompanied by high levels of Id⁺ IgG, high titers of autoantibodies, intense T and B cell activation and proliferation, germinal center development, and generation of plasma cells in both lymphoid organs and inflamed colons and activation of T and B cells in inflamed colons (Fig. 8, B–H). A preferential accumulation of Id⁺ B cells and Id-specific T cells was seen in the colons of diseased reconstituted mice, exactly as observed in doubly transgenic donors of lymphocytes (Fig. 7 C).

One may ask whether endogenous TCR α- and β-chains, so extensively expressed in doubly transgenic mice (Fig. 4, C–E) were necessary for disease development. To test this question, we transferred Id⁺ bone marrow cells into Id-specific TCR transgenic SCID mice that cannot express endogenous α- and β-chains due to recombination deficiency. This procedure resulted in enlarged lymph nodes and spleen, IgG autoantibodies, and gut disease in four of six mice, although less severe than that seen in doubly transgenic mice (data not shown). In such bone marrow chimeras, both Id⁺ B cells and Id-specific T cells were activated, proliferating blasts (Fig. 9, A and B), as proliferating CD4⁺ T cells in Id⁺ bone marrow→TCR transgenic SCID mice were GB113⁺. Disease development was Id-dependent because Id⁻ bone marrow transferred into TCR transgenic SCID mice or SCID mice, or Id⁺ bone marrow cells transferred into SCID mice resulted in neither autoantibodies nor disease. Consistent results were obtained in an alternative approach, in which cotransfer of high avidity Id-specific Th2 cells from TCR transgenic SCID mice and purified Id⁺ B cells from healthy Id⁺ mice induced disease in all three Rag2^−/− recipients, whereas transfer of OVA-specific TCR transgenic Th2 cells and Id⁺ B cells caused no disease (Fig. 9 C). Collectively, these experiments suggest that 1) Id-specific T cells do not need to express endogenously rearranged TCRs for disease to occur; 2) disease may develop in the absence of non-Id specific T cells; and 3) in these experiments, B and T cells each have only one complementary transgene, i.e., disease may occur without the presence of two transgenes in individual cells (as is the case in doubly trangenic mice).

Id-specific T cells have been described in SLE (8, 9), rheumatoid arthritis (10), and multiple sclerosis (11), but their role in the pathogenesis of these diseases is unclear. By using paired TCR transgenic and Ig L chain transgenic mice and adoptive transfer, we (17) and others (18) have shown that Id-specific T cells may transiently help Id⁺ B cells to secrete autoantibodies. We describe herein that chronic Id-driven T-B collaboration in doubly transgenic mice causes lethal systemic autoimmune disease. Thus, mice doubly transgenic for an Id⁺ Ig L chain and an Id-specific MHC class II-restricted αβ TCR developed, as young adults, hypergammaglobulinemia, IgG autoantibodies including antinuclear Abs, complement activation, glomerulonephritis, arthritis, skin disease, or inflammatory bowel disease (IBD). These findings were accompanied by lymphocyte infiltrations and ongoing Id-driven T-B collaboration in diseased organs. The disease appeared on a normally nonautoimmune background (BALB/c).

These results were surprising because previous evidence has suggested that T cell tolerance to abundant Id-peptides is rather complete (3, 13, 18, 19, 20, 24, 25). In particular, it was reported by us, using the same doubly transgenic mice as used in this study, that Id-specific T cells in very young (3- to 5-wk-old) mice were tolerant to abundant Id⁺ Ig L chains and Id⁺ B cells (13). Corroborating evidence was recently reported in another doubly transgenic Id-model (18). The present study confirms our previous findings (13) that there is a profound deletional T cell tolerance in very young doubly transgenic mice, and that the surviving T cells had greatly diluted their Id-specific TCR by expression of endogenous TCR α- and β-chains. However, as the mice aged, a low frequency of Id-specific T cells increasingly emerged. Thus, T cell tolerance to abundant Id was pronounced, but Id-specific T cells, with reduced Id reactivity due to expression of endogenous TCR (23), increasingly escaped with age. Consistent with our findings, leaky thymic negative selection has been described for other Ags (26, 27, 28, 29).

Id-specific T cells that escaped to periphery lymphoid organs in doubly transgenic mice were activated and incorporated BrdU in vivo. Although they often expressed endogenous TCR α- and β-chains and proliferated poorly ex vivo, they efficiently helped Id⁺ B cells to proliferate and secrete Id⁺ Ig. Consistent with this, Id⁺ B cells in peripheral lymphoid organs were activated, proliferated, and formed germinal centers. These findings strongly suggest ongoing Id-driven T-B collaboration in peripheral lymphoid organs of doubly transgenic mice. The pathogenicity of such Id-driven T-B collaboration was strongly supported in adoptive transfer experiments in which a mixture of Id-specific T cells and Id⁺ B cells from peripheral organs of doubly transgenic mice resulted in disease in recipients that closely mimicked the systemic autoimmunity and IBD-like disease of the donor.

Although Id-specific CD4⁺ T cells were scarce in lymphoid organs of doubly transgenic mice, they heavily infiltrated sites of disease like the lamina propria of colons, where Id-specific T cells were frequent and formed clusters of dividing blasts. Such activated Id-specific T cell blasts were accompanied by abundant Id⁺ B cells. Furthermore, Id-specific T and Id⁺ B cells were found in frequent synapses with each other, demonstrating Id-driven T-B collaboration in situ. As a result, Id⁺ B cells were activated and frequently had isotype switched to IgG1 or IgG2a. Whether these B cells produced IgG autoantibodies remains to be formally demonstrated, but suggestively, deposits of Id⁺ Ig and C3 were found in proximity to foci of Id-driven T-B collaboration.

One might ask why Id-driven T-B collaboration in doubly transgenic mice resulted in production of a range of different autoantibody specificities. Because Id⁺ B cells express a transgenic Id⁺ L chain, this fixed L chain is expressed in each B cell with different H chains resulting from clonally unique VDJ rearrangements. Thus, the total BCR repertoire in doubly transgenic mice could be substantial and should include autoreactive BCR, like in normal individuals (30). Transgenic B cells spontaneously present Id-peptide on their MHC class II molecules (14), therefore Id⁺ B cells with an autoreactive BCR should constitutively present Id-peptide on their class II molecules and receive help from Id-specific T cells (17, 18), thus short-circuiting the need for conventional self-reactive T cells. Additionally, formation of germinal centers (17) and somatic hypermutation driven by Id-specific T cells is likely to further expand the repertoire of autoantibodies.

In distinction to conventional T-B collaboration (31, 32) Id-driven T-B collaboration is nonlinked because TCR and BCR recognize different Ags (17). Nonlinked T-B collaboration makes it possible to distinguish the effects of BCR ligation and T cell help on B cells (17). Thus, in the presence of Id-specific T cell help, BCR ligation was additionally required for isotype switch and secretion of IgG autoantibodies to occur (17). BCR ligation could increase APC function by a number of mechanisms including up-regulation of Ag processing and costimulation (15, 33, 34, 35, 36, 37, 38, 39). Therefore, because autoantigens represent ubiquitous BCR ligands in vivo, self-reactive Id⁺ B cells could be selectively recruited into Id-driven T-B collaboration because ligation of their BCR could render them particularly efficient at presenting Id to T cells (17). In this respect, because Id-specific T cells may often be of low sensitivity due to tolerance mechanisms, BCR ligation by self-Ag may in fact be necessary for initiation of pathogeneic Id-driven T-B collaboration. In addition to BCR-ligation, Id-driven autoimmunity may be further facilitated by TLRs on the surface of B cells, as described in rheumatoid factor transgenic crossed with MyD88^−/− mice (40).

Although T and B cell specific for a single self-antigenic determinant dominate early in autoimmune disease, reactivity spreads to new epitopes with disease progression (41, 42). In this respect, inflamed colons of doubly transgenic mice had a pronounced infiltration of T cells that expressed endogenous TCR α- and β-chains to the exclusion of the transgenic TCR. It might well be that in the wake of initial Id-driven T-B collaboration, these infiltrating T cells of unknown specificity could be activated against food allergens, gut bacteria, or even autoantigens. In contrast, T cells with non-Id specificity were not necessary for disease induction as demonstrated in this study (Fig. 9) in transfer experiments involving recombination deficient transgenic T cells. Rather, these experiments, in addition to previous transfer experiments with negative control OVA-specific T cells (17), clearly demonstrate the requirement for Id-Ag specificity of T cells for development of autoimmunity.

Are these results obtained in a reductionistic doubly transgenic model relevant to development of autoimmune disease in normal individuals? Obviously, as most individuals do not suffer from autoimmune diseases, pathogenic Id-driven T-B collaboration must nearly always be counteracted by a number of mechanisms: 1) under physiological situations in which far fewer Id-specific T cells need to be tolerized, it is likely that T cell tolerance for abundant germline Id is rather complete (3, 13, 19, 20, 24, 25). Therefore, nonlinked Id-driven T-B collaboration is expected to be restricted to rare Id peptides that depend on N region diversity or somatic mutations (3, 16, 19, 20); 2) the L and H chain BCR must meet two conditions: specificity for a self-Ag and upon Ag processing, yield Id-peptides that fit the peptide binding motifs of the MHC class II molecules of the B cell (24) (few B cells may actually express a BCR, which meets these two requirements); 3) due to the low frequencies of each, Id⁺ B cells and Id-specific T cells may seldom meet during lymphocyte recirculation; 4) naive Id-specific T cells may be poorly activated by B cells, in fact they may require initial stimulation by dendritic cells Id-primed by secreted Ig (our unpublished observations); 5) nonlinked Id-driven T-B collaboration tapers off within weeks (17, 18), compatible with induction of deletion or anergy of interacting cells (43), or inhibition by immune regulatory mechanisms (44, 45). Thus, chronic Id-driven T-B collaboration may possibly only be achieved if apoptosis-resistant T and B cells are selected in the process or if suppression is dysfunctional. In summary, all these restraints are likely to make pathogenic Id-dependent T-B collaboration a very rare event. However, conditions that meet all these requirements may haphazardly be established with time, consistent with an age-dependent development of autoimmune disease.

Despite the extensive limitations just outlined, there is evidence that nonlinked Id-driven T-B collaboration may contribute to autoimmune disease in nontransgenic individuals. Ig-specific T cells have been found in rheumatoid arthritis patients although their fine specificity was not characterized (10). Moreover, Id-specific T cells have been suggested to exist in SLE patients (8, 9), and have been convincingly described in autoimmune BWF₁ mice (6) as well as in CD95 mutant lpr mice (7). Furthermore, it has very recently shown the coexistence of Id⁺ B cells and Id-specific T cells in the cerebral spinal fluid of patients with multiple sclerosis (11). Further studies on patients with autoimmune diseases are clearly warranted, to establish what contribution, if any, Id-driven T-B collaboration has in the pathogenesis of disease in humans.

In summary, establishment of chronic Id-driven T-B collaboration and autoimmune disease is likely to depend of a number of haphazardous events, as outlined, that would be difficult to predict by genetic analyses alone (46, 47). The likelihood of all these stochastic events to concur would increase with time, consistent with increased incidence of autoimmunity with age.

The expert technical help of H. Omholt and P. Hofgaard is appreciated. We thank Drs. K. Thompson and Z. Dembic for discussions and valuable comments.

The authors have no financial conflict of interest.