Regulation of B Cell Tolerance by the Lupus Susceptibility Gene Ly108

See allHide authors and affiliations

Science  16 Jun 2006:
Vol. 312, Issue 5780, pp. 1665-1669
DOI: 10.1126/science.1125893


The susceptibility locus for the autoimmune disease lupus on murine chromosome 1, Sle1z/Sle1bz, and the orthologous human locus are associated with production of autoantibody to chromatin. We report that the presence of Sle1z/Sle1bz impairs B cell anergy, receptor revision, and deletion. Members of the SLAM costimulatory molecule family constitute prime candidates for Sle1bz, among which the Ly108.1 isoform of the Ly108 gene was most highly expressed in immature B cells from lupus-prone B6.Sle1z mice. The normal Ly108.2 allele, but not the lupus-associated Ly108.1 allele, was found to sensitize immature B cells to deletion and RAG reexpression. As a potential regulator of tolerance checkpoints, Ly108 may censor self-reactive B cells, hence safeguarding against autoimmunity.

Loss of B cell tolerance is the hallmark of systemic lupus erythematosus (SLE), or lupus, an antibody-mediated chronic autoimmune disease affecting multiple organs. However, the precise means by which tolerance is breached in lupus, and the underlying genes and molecules responsible, remain obscure. The origin of lupus in both mice and humans appears to be polygenic, involving more than a dozen potentloci(13), although of these, at least one locus on chromosome 1 appears to be shared by both species. The z allele of Sle1 and its component sublocus Sle1b, derived from the lupus-prone NZM2410/NZW strain of mice, are linked to a variety of lupus-related disease phenotypes, including antinuclear antibodies (ANAs), splenomegaly, and glomerulonephritis (4, 5). A locus on human chromosome 1 orthologous to murine Sle1 has also been implicated in lupus susceptibility (6). Located within the Sle1bz sublocus, the SLAM gene family–encoded costimulatory molecules are among the first candidate genes to be identified as being linked to spontaneous lupus in mice (7). Extensive polymorphisms and expression differences were documented in several members of this gene cluster between the normal C57BL/6 (B6) strain and the lupus-prone B6.Sle1bz/z congenic mice bearing the z haplotype of SLAM (7).

To determine the mechanisms by which B cell tolerance might be infringed in lupus, the B cell repertoire of B6.Sle1z/z congenic mice (8) was modified to an essentially monoclonal specificity by breeding onto them a B cell receptor (BCR) transgene specific for lysozyme (HEL) (9). In these B6.HELIg mice, the HELIg BCR transgene leads to a near-uniform pool of HEL-specific B cells bearing a BCR heavy chain of immunoglobulin Ma (IgMa) allotype and high titers of serum antibodies to HEL of IgMa allotype (9). Developmental exposure of these B cells to a strong cross-linking surrogate self-antigen [in the form of cell-membrane–bound HEL (mHEL)], leads to clonal deletion in the bone marrow (10). However, when exposed to a weaker surrogate self-antigen [in the form of soluble HEL (sHEL)], the same HEL-specific B cells are censored by an alternate tolerogenic mechanism known as anergy (9). To ascertain how Sle1z might influence B cell tolerance, B6.Sle1z/z.HELIg mice were bred to mHEL or sHEL mice.

B6.Sle1z/z mice display several features of systemic autoimmunity (8), which are now established as B cell intrinsic (11). Whereas B6.Sle1z/z mice exhibit increased B cell and T cell activation, splenomegaly, and ANAs, these features were not apparent in B6.Sle1z/z.HELIg mice (fig. S1). However, B6.HELIg.mHEL and B6.Sle1z/z.HELIg.mHEL mice showed a comparable loss of transgenic B cells in the spleen and bone marrow, and an absence of IgMa antibodies to HEL in their sera (fig. S2). Thus, Sle1z does not impede deletion of B cells upon strong engagement of the surrogate self-antigen, mHEL. This is consistent with previous reports showing that central tolerance of high avidity anti-self B cells is intact in other models of murine lupus (12, 13).

To ascertain whether Sle1z influenced anergy induction upon weak engagement by self-antigen, B6.Sle1z/z.HELIg.sHEL mice were next examined. A measurable number of B6.Sle1z/z.HELIg.sHEL mice beyond 3 months of age, but not B6.HELIg.sHEL controls, exhibited elevated titers of IgMa antibodies to HEL with levels comparable to that seen in non–lysozyme-exposed B6.HELIg mice (Fig. 1A and fig. S3). However, neither strain exhibited significant ANA titers (Fig. 1B and fig. S3).

Fig. 1.

B6.Sle1z/z.HELIg.sHEL mice exhibit breach in B cell tolerance. (A and B) Levels of IgMa antibodies to HEL (A) and IgG antibodies to DNA/histone (B) in 8- to 14-month-old mice were measured by enzyme-linked immunosorbent assay. (C to E) Whole splenocytes from B6.HELIg and B6.HELIg.sHEL mice, with or without Sle1z, were surface-stained for IgMa, IgMb, B220, and Ig-λ and analyzed. (C) Representative contour plots of IgMa versus IgMb, pregated on B220+ cells. Boxed numbers at the top left corner indicate percentage IgMb+IgMa-lo cells. (D) Scatterplot relating the percentage of B cells expressing IgMb to the age of mice. (E) Percentage of IgMa+ cells expressing Ig-λ. (F to H) Whole splenocytes [(F) and (H)] or purified B cells (G) from various strains were stimulated as indicated for 18 hours (F) or 48 hours [(G) and (H)] and assayed for CD86 levels (F), proliferation (G), or apoptosis (H) (18). Each dot represents an individual mouse. The solid and dotted horizontal lines indicate group means and B6.HELIg mean values, respectively. Bars indicate mean + SEM of five to six mice each. The P value results of Student's t test are indicated. *, Mann-Whitney test (E) or paired t test [(F) and (H)]. NS, not significant.

In contrast to spleens from B6.HELIg.sHEL mice, which show reduced numbers of B cells (9), those from B6.Sle1z/z.HELIg.sHEL mice exhibited considerable B cell expansion (tables S1 to S3). Coexpression of soluble HEL in the B6 Tg background normally results in an age-dependent increase in the proportion of splenic B cells bearing the endogenous IgMb receptor versus the Tg IgMa (Fig. 1C and tables S1 to S3), possibly as a consequence of active receptor revision in response to HEL or selective expansion of B cells that are not HEL-reactive. However, this proportion of splenic B cells expressing endogenous IgMb heavy chain was considerably reduced in B6.Sle1z/z.HELIg.sHEL mice as compared with their B6.HELIg.sHEL counterparts (Fig. 1, C and D). Furthermore, IgMa+ B cells from B6.HELIg.sHEL spleens exhibited an age-dependent shift toward the use of nontransgenic Ig-lambda light chains, which was attenuated in the presence of Sle1z (Fig. 1E). In contrast to the spleen, there was negligible IgMb usage among bone marrow B cells in both strains (fig. S4). Thus, the revision away from transgenic BCR to endogenous BCR usage appears to be a peripheral rather than a central bone marrow B cell event that occurs in response to ongoing exposure to surrogate self-antigen, sHEL.

B6.HELIg.sHEL B cells have been documented to display reduced functional response to BCR ligation, indicative of efficient tolerization (14). Notably, however, B6.Sle1z/z.HELIg.sHEL B cells exhibited stronger up-regulation of CD86, and higher proliferation and reduced apoptosis in response to BCR cross-linking, as compared with B6.HELIg.sHEL B cells (Fig. 1, F to H). Thus, consistent with the in vivo findings, our in vitro studies also indicate that B6.HELIg.sHEL B cells are inadequately anergized on the B6.Sle1z/z background.

To examine the functional properties of Sle1z-bearing B cells before exposure to self-antigen, B cells from B6.HELIg and B6.Sle1z/z.HELIg mice were compared. Although total splenic B cells from both strains showed similar activation and apoptotic profiles upon BCR cross-linking, B6.Sle1z/z.HELIg spleens exhibited a significant expansion of B220+AA4.1+CD21CD23 (T1) transitional immature B cells, with a resultant increase in the T1:follicular B cell ratio compared with B6.HELIg controls (Fig. 2A).

Fig. 2.

Sle1/Sle1bz-bearing immature B cells have impaired responses to BCR cross-linking. (A) Numbers of splenic B cell subsets (left y axis) and T1:follicular B cell ratios (right y axis) in 6- to 10-week-old B6.HELIg, B6.Sle1z/z.HELIg, and B6.Sle1bz/z.HELIg mice are shown. Horizontal lines indicate group means. (B and C) Displayed are calcium flux responses in IL-7–driven, bone-marrow–derived immature B cells to 20 μg/ml F(ab)′2 antibody to IgM, measured as described (18). (B) Representative plots of digital signal processing (DSP) time versus free/bound calcium. (C) Area under calcium-flux curve in the three strains, compiled from two independent experiments. (D) Displayed are the mean + SEM (n = 3) stimulation indices of IL-7–driven, bone-marrow–derived immature B cells cultured in triplicate for 48 hours with 25 μg/ml F(ab)′2 antibody to IgM, 10 μg/ml biotin-labeled F(ab)′2 antibody to IgM + 10 μg/ml avidin, 1 μM ionomycin, 25 μg/ml antibody to Igκ, or 2 μg/ml sHEL. Representative data from three independent experiments are shown. (E) Real-time polymerase chain reaction analysis of SLAM molecules in B6.HELIg and B6.Sle1z/z.HELIg IL-7–driven bone-marrow–derived immature B cells. Bars represent mean + SEM expression levels in B6.Sle1z/z.HELIg, relative to B6.HELIg (n = 6 each), after normalizing against GAPDH. (A) to (E) P values pertain to Student's t test comparison of B6.HELIg to either B6.Sle1z/z.HELIg or B6.Sle1bz/z.HELIg.

Because the Sle1bz sublocus accounts for most of the autoimmune phenotypes attributed to the Sle1z locus (5), B6.Sle1bz/z.HELIg mice were next generated and were also found to exhibit a significant expansion in T1 B cells (Fig. 2A). Collectively, these findings suggested a relative block in the transition from T1 to mature B cells in the presence of Sle1bz. Potentially, such a block could result from impaired deletion of immature B cells, leading us to examine the function of Sle1bz-bearing immature B cells. Bone-marrow–derived immature B cells were cultured in interleukin-7 (IL-7) (fig. S5) (1518). After cross-linking of the BCR, immature B cells from B6.Sle1z/z.HELIg and B6.Sle1bz/z.HELIg mice were seen to flux considerably less calcium (Fig. 2, B and C), underwent cell death less readily, and showed an increase in proliferation, as compared with B6.HELIg controls (Fig. 2D). These results suggest that the Sle1bz allele may impede antigen-triggered death of immature B cells, possibly as a consequence of impaired BCR signaling.

The candidate gene(s) for Sle1bz have recently been elucidated to belong to the SLAM gene cluster of activation and costimulatory molecules (7, 19). Among these, polymorphic variants of the Ly108 gene–encoded protein were the most differentially expressed on immature B cells. Compared with B6, B6.Sle1z/z/Sle1bz/z immature B cells expressed four times as much Ly108.1 splice isoform but one-fourth as much Ly108.2 (Fig. 2E). To study the influence of these two splice isoforms on immature B cell function, the genes encoding the two Ly108 isoforms were transfected into the immature B cell line, WEHI-231 (WEHI), which expresses very low levels of endogenous Ly108 (fig. S6). In line with the findings observed in primary immature B cells (Fig. 2, B to D), W.Ly108.1 transfectants exhibited reduced calcium flux (Fig. 3A), reduced RAG reexpression (Fig. 3B), and decreased cell death (Fig. 3, C to E), compared with W.Ly108.2, upon triggering the BCR with a variety of stimuli. In particular, the W.Ly108.2 transfectants bearing the normal B6 allele of Ly108 were exquisitely sensitive to BCR cross-linking; even monovalent Fabs, low doses of F(ab)′2 antibody to IgM, and ionomycin resulted in a strong calcium flux, RAG reexpression, and near-maximal cell death (Fig. 3).

Fig. 3.

Ly108.2 but not Ly108.1 promotes strong calcium flux, RAG reexpression, and death in immature B cells following BCR ligation. (A) Displayed calcium flux responses to 1 μM ionomycin or 20 μg/ml antibody to IgM F(ab)′2 in W.Ly108.1 and W.Ly108.2 stable WEHI-231 transfectants. Data shown are representative of at least two independent experiments and clones. (B) WEHI transfectants (n = 4 independent clones each) were stimulated for 36 hours with 25 μg/ml F(ab)′2 antibody to IgM or 1 μM ionomycin. RAG2 message (mean + SEM) was measured in real time, normalized to GAPDH, and expressed relative to the level in unstimulated, untransfected WEHI cells (18). (C and D) Stable WEHI transfectants were stimulated as indicated for 48 hours and assayed for the percentage of apoptotic cell loss (C) and proliferation (D) (18). Each bar represents mean + SEM of n = 2 (mock) or n = 4 to 5 each (W.Ly108.1 and W.Ly108.2) independent transfectants, cultured in triplicate. Representative data are shown from three independent experiments. (E) Displayed are 200X images of WEHI transfectants grown with or without 10 μg/ml F(ab)′2 antibody to IgM for 72 hours. Images shown are representative of four to five independent stable transfectants of each type. (B) to (D) P values pertain to Student's t test comparison of W.Ly108.1 versus W.Ly108.2.

The above findings indicate that at least some of the genes that cause lupus may function by crippling multiple B cell tolerance mechanisms such as deletion, receptor revision, and anergy induction by tuning down BCR signaling at the immature stage of development. Among the SLAM molecules, Ly108 (and possibly others) may function as molecular rheostats, determining the stringency with which self-reactive B cells are censored during early development. Whether the dramatically different properties of the two Ly108 isoforms emanate from the dissimilar number of intracellular tyrosine switch motifs they possess warrants further investigation.

Given the present finding that the Ly108 isoforms influence B cell tolerance in a generalized, non–antigen-specific manner, it is intriguing that the serology in B6.Sle1/Sle1bz/z mice is chromatin-centric rather than polyclonal (7, 8, 20). We believe this outcome might relate to the fact that the early B cell repertoire at the pre-B and immature B cell stages may already be nuclear-antigen skewed to begin with, even in normal individuals, as illustrated by recent studies (21, 22). Hence, the emergence of ANAs and the pathognomonic hallmark of lupus, the LE (lupus erythematosus) cells (23), may be the consequence of a generalized tolerance defect in immature B cells that already encode nuclear-antigen reactivity following primary immunoglobulin gene rearrangement. Unraveling the molecular cascades through which Ly108 dictates how immature B cells toggle between life and death is the challenge that lies ahead.

Supporting Online Material

Materials and Methods

Figs. S1 to S6

Tables S1 to S3


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article