Complement C4 maintains peripheral B-cell tolerance in a myeloid cell-dependent manner

Abstract

The factors that allow self-reactive B cells to escape negative selection and become activated remain poorly defined. Using a B-cell receptor-knock-in mouse strain, we identify a pathway by which B-cell selection to nucleolar self-antigens is complement-dependent. Deficiency in complement component C4 led to a breakdown in the elimination of autoreactive B-cell clones at the transitional stage, characterized by a relative increase in their response to a range of stimuli, entrance into follicles and a greater propensity to form self-reactive germinal centers. Using mixed bone marrow chimeras we found that the myeloid compartment was sufficient to restore negative selection in the auto-reactive mice. A model is proposed in which in the absence of complement C4, inappropriate clearance of apoptotic debris promotes chronic activation of myeloid cells, allowing the maturation and activation of self-reactive B-cell clones leading to increased spontaneous formation of germinal centers.

Introduction

Much of our understanding of B-cell tolerance is based on murine models. Characterization of immunoglobulin (Ig) transgenic (Tg) mice have revealed that self-reactive B cells are silenced in multiple ways including clonal deletion [1-3], receptor editing [4, 5], allelic inclusion [6] and anergy [7-10]. To persist and reach maturity, autoreactive B cells must bypass receptor editing at the immature stages in the bone marrow and must evade deletion during negative selection stages in the periphery. Additional B-cell-extrinsic mutations or dysfunctions are generally required for the activation and differentiation of autoreactive B-cell clones and precipitation of autoimmunity.

One important factor in permitting self-reactive B-cell activation is alterations in components of the complement system. Myeloid cells like macrophages are a major producer of complement components C1q, C4 and C3 [11]. Liver also produces most complement components except C1q, contributing to circulating serum levels of complement proteins [12]. Inherited deficiency in complement C1, C2 and C4 are highly associated with susceptibility to lupus [13, 14] and while in the general population, lupus occurs in women at a higher frequency than in men, it occurs with a similar frequency in individuals deficient in complement [13, 15]. In contrast, increased copies of C4 genes are protective against lupus [16]. Deficiency in C1q and C4 also predisposes mice to a lupus-like disease on certain backgrounds [17-19]. For example, C4^-/- and Cr2^-/- B6.lpr mice develop elevated autoantibody titers relative to complement sufficient controls and show evidence of glomerulonephritis [20]. Using the anti-hen egg lysozyme (Hel) B-cell Tg model, Prodeus et al. [20] reported that deficiency in C4 leads to a relative increase in mature self-reactive B cells that appear to partially escape anergy suggesting that complement might regulate B-cell tolerance and that the defect may be B cell intrinsic.

Another important class of factors in determining the fate of self-reactive B cells is Toll-like receptors (TLRs). Many of the classic lupus antigens derived from apoptotic cells such as ribo-nuclear proteins (RNPs) and DNA are ligands for TLR and internalization via the B-cell receptor (BCR) may enhance activation of anergic B cells through a two signal pathway [21]. Moreover, defects in clearance of apoptotic debris could result in triggering of TLR7 and TLR9 leading to elevated secretion of type I interferon and enhanced differentiation of autoreactive B cells [22,23]. For example, in the 564 Igi BCR knock-in mouse strain in which B cells are specific for a nucleolar antigen, self-reactive B cells are activated and secrete IgG autoantibody through a TLR7-dependent pathway despite apparent normal negative selection [24, 25]. To investigate a role for complement in B-cell tolerance to nucleolar antigen, C4-deficient mice were crossed with 564 Igi knock-in-line on a B6 background. Characterization of the mice identified a loss of tolerance of the autoreactive B cells at the transitional stage. In addition, deficiency of C4 resulted in a loss of B-cell anergy and an increased propensity to form self-reactive germinal centers (GCs). Using mixed bone marrow chimeras, we found that efficient B-cell selection and anergy was restored in the presence of a C4-sufficient myeloid compartment.

Results

564 autoantibodies recognize ribonucleoproteins

The 564Igi mouse model, originally described by Imanishi-Kari and colleagues [24], was found to produce autoantibodies. To identify the 564 antigen, 564Igi was mixed with nuclear and cytoplasmic extract of P3Ag cells and immune complexes separated on SDS-gels. A number of antigens were precipitated, suggesting that the epitope recognized by the 564 idiotype (Id) is a domain that may be common to multiple self-antigens (Figure 1A). Many of these proteins bore characteristics of ribonucleoprotein (RNP) complexes (Supporting Information Table 1). Pretreatment of extracts with RNAse abolished immune precipitation with the 564 antibody suggesting that the epitopes contained RNA (Figure 1B). One notable antigen identified by mass spec analysis was the Sjögren's Syndrome antigen B (SSB/La), a recognized lupus antigen which was further confirmed by probing immune precipitates with anti-SSB/La antibody (Figure 1C). The 564 antibody as well as sera derived from both 564Igi-C4^+/+ and 564Igi-C4^-/- mice were found to bind in a nucleolar pattern to nuclei immobilized on Hep 2 slides (Figure 1D). Microscopic study of kidney sections from 564Igi-C4^+/+ and 564Igi-C4^-/- mice stained with 564mAb revealed more 564 immune complexes deposited in the kidney of C4-deficient 564Igi animals (Figure 1E).

Fig 1. 564Igi mice produce anti-nucleolar autoantibodies recognizing an RNP.

(A) Immunoprecipitation (I.P.) using the 564Igi antibody or a control IgG of cytoplasmic (CE) or nuclear extracts (NE) from P3Ag myeloma cells. The band indicated by a bracket was determined by mass spectrometry to contain the SSB/La ribonucleoprotein. (B) RNase I treatment of CE and NE extracts leads to loss of binding by 564Igi mAb. (C) I.P. using either 564Igi mAb or control IgG of nucleolar preparations was probed later for SSB/La by western blot. Arrow indicates the expected SSB/La band size at around 47KDa. (D) Sera from B6, 564Igi-C4^+/+ and 564Igi- C4^-/- mice (1/100) were incubated on Hep2 slides. 564Igi mAb was used as control. Images were acquired at 20X magnification. (E) Kidney sections from 564Igi-C4^+/+ and 564Igi- C4^-/- mice were stained with anti-564 Ab and observed microscopically. Images were acquired at 60× magnification. Data shown are from one experiment representative of 3 experiments performed.

Relaxed negative selection in C4 deficient mice

Examination of primary and secondary lymphoid organs in 564Igi mice, indicated that re-circulating (mature; AA4.1^lo) Id⁺ B cells were more prevalent in the bone marrow (BM), lymph node and spleen of 564Igi-C4^-/- mice compared to 564Igi-C4^+/+ mice (Supporting Information Figure 1). This suggested that negative selection of Id⁺ B cells might be impaired in absence of C4. Self-reactive B cells progress through checkpoints during differentiation in both BM and spleen. Development of 564Igi Id⁺ cells in BM was similar between the two 564 strains. Splenic B cells can be divided into transitional (AA4.1^hi or AA4.1^int) and mature (AA4.1^lo) based on expression of the AA4.1 cell surface marker. Assessment of B-cell development in the spleen identified similar frequencies of self-reactive B cells in the AA4.1^hi population, which represents the most recent emigrant B-cell population in the spleen (Figure 2). However, comparison of mature Id⁺ cells (AA4.1^lo) revealed a significant reduction in the frequencies and cell numbers in 564Igi- C4^+/+ mice relative to C4^-/- mice (Figure 2). Thus, as reported earlier by Berland et al. [24], the Id⁺ B cells undergo negative selection in the 564 Igi- C4^+/+ mice at the transitional stage; however, in the absence of C4 we observed a less stringent selection. Using a different combination of markers (CD23, CD21 and IgM), we found a similar pattern of differential negative selection at the transitional stage in 564Igi-C4^-/- relative to C4-sufficient mice (results not shown). However, since cell surface IgM is down regulated on the Id⁺ B cells in both strains as described below, the latter approach is less reliable in phenotyping differentiation of transitional B cells. The overall results suggest that negative selection of B cells in spleen was impaired in the absence of C4 leading to an accumulation of mature, self-reactive B cells in different lymphoid organs.

Fig 2. C4 is required for efficient B-cell selection in spleen.

Splenocytes from 564Igi-C4^+/+ and C4^-/- mice were stained with 7AAD and antibodies against B220, AA4.1 and Id. Different B-cell developmental stages in spleen were identified as being AA4.1^hi (top), AA4.1^int (middle) and AA4.1^lo (bottom) and self-reactive (Id⁺) B-cell proportion; the total number in each of these stages is shown for the two 564 Igi strains. Each symbol represents an individual mouse and data shown are pooled from 5 independent experiments. Bars represent means. ***p<0.0001, Student's T-test.

B-cell anergy is impaired when C4 is absent

To assess whether mature Id⁺ self-reactive B cells were functional or were maintained in an anergic state, their surface phenotype and ability to become activated were characterized. In accordance with the results of Berland et al. [24], we observed down-modulation of surface markers associated with B-cell responsiveness in mature, self-reactive B cells (AA4.1^lo Id⁺) from C4-sufficient 564Igi mice (Figure 3A). These included IgM, IgD and CD21. In C4-deficient 564Igi mice, the mature Id⁺ B cells had down-regulated IgM but expressed near normal levels of IgD and CD21 (Figure 3A), indicating a mature, non-anergic phenotype.

Fig 3. C4 is required for maintaining anergic phenotype of self-reactive cells.

(A) Expression of IgM, IgD and CD21 was assessed on mature (AA4.1^lo) self-reactive (Id⁺) B cells from 564Igi-C4^+/+ and C4^-/- mice compared with non-transgenic, endogenous B cells. Data shown are representative of 3 independent experiments performed with 3 mice total. (B) Spleen sections from B6, 564Igi-C4^+/+ and C4^-/- mice were stained with anti-CD3 (T-cell area (T)), anti-B220 (B-cell follicles (B)) and anti-564 antibody (self-reactive B cells (Id)). ‘EF’ indicates extrafollicular foci of Id-producing cells. Images were acquired at 20× magnification. Data shown are representative of 4 independent experiments performed with 5 mice total. (C) Magnetically sorted CD43^- splenocytes from 564Igi-C4^+/+ and C4^-/- mice were cultured for 16 h with a range of stimuli. Expression of CD86 was analyzed on AA4.1^lo (mature) B cells by flow-cytometry to measure activation. Data shown are representative of 4 independent experiments performed with a total of 4 mice.

In the presence of a normal repertoire of non-self-reactive competitor B cells, anergic B cells are excluded from the B-cell follicles and accumulate at the boundary between the T-cell zone and the B-cell follicles [26, 27]. To determine whether Id⁺ B cells in the absence of complement C4 undergo follicular exclusion, cryosections of spleens were stained with antibodies against B220, CD3 and Id and analyzed by confocal microscopy. As expected, Id⁺ B cells in 564Igi mice underwent exclusion from B-cell follicles and marginal zones [24]. In contrast, Id⁺ B cells in spleens from C4-deficient 564 mice were not restricted at the T cell-B cell boundary but were instead distributed throughout the B-cell follicles (Figure 3B).

To examine the functional responsiveness of Id⁺ B cells from 564Igi-C4^-/- mice, naïve mature splenic B cells (AA4.1^lo, as gated in Figure 2) were analyzed by flow cytometry for the up-regulation of CD86 upon activation with various stimuli. Naïve B cells from both strains did not express CD86 when freshly isolated (data not shown), although Id⁺ B cells isolated from 564Igi-C4^-/- mice were spontaneously activated after 16 h in culture based on a shift in CD86 expression (Figure 3C). Self-reactive mature B cells derived from a 564Igi-C4^-/- mice also expressed CD86 in response to BCR and TLR ligation, unlike autoreactive cells isolated from C4-sufficient animals, which for the most part did not up-regulate CD86 (Figure 3C). As expected, we found CD86 up-regulation on Id^- B cells from both 564Igi-C4^-/- and – C4^+/+ mice in these stimulation cultures (Figure 3C). While the majority of self-reactive B cells remained anergic when derived from C4-sufficient animals, a minor population of Id⁺ B cells from 564Igi-C4^+/+ mice responded to BCR stimulation in this assay (Figure 3C). This might be due to allelic inclusion, where developing B cells rearrange and express two immunoglobulin surface receptors [2]. For self-reactive B cells engaging self-antigen, presence of a second receptor could dilute a strong BCR signal and allow them to escape negative selection and anergy. To investigate whether Id⁺ B cells exhibit allelic inclusion in the 564Igi model, we performed single cell sorting of CD138⁺ Id⁺ plasmablasts from the spleens of 564Igi-C4^-/- and –C4^+/+ mice. The single cells were analyzed for expression of the transgenic 564 light chain Jκ5 as well as the endogenous Jκ1, Jκ2, and Jκ4 light chains. While allelic inclusion is a rare occurrence, it was found to be a feature of Id⁺ B cells from both 564Igi- C4^-/- and –C4^+/+ mice (Supporting Information Figure 2A).

A classic feature of anergy in self-reactive B cells is also an impaired ability to flux Ca²⁺ in response to BCR cross-linking. Endogenous, non-self-reactive B cells in both strains mobilized calcium to the cytoplasm in a manner similar to that of B cells from a wildtype, non-564Igi mouse upon cross-linking of the BCR with anti-IgM F(ab’)₂ (Supporting Information Figure 2B). 564Igi B cells from both groups of mice were barely responsive to BCR ligation, consistent with results from an earlier study showing that Id⁺ B cells from 564Igi-C4^+/+Rag2^-/- mice did not flux calcium with anti-IgM stimulation [24]. Together, these findings indicate that while self-reactive B cells in a C4-deficient environment displayed some signs of anergy such as down-regulation of BCR and impaired calcium flux, they are more prone to mature and accumulate in the periphery and respond to BCR and TLR stimulation.

GC formation in 564Igi mice

Many autoimmune mouse models show spontaneous GC formation, allowing mutated, self-reactive autoantibodies to be generated [28, 29]. To test if the increased frequency of mature Id⁺ B cells in C4-deficeint mice resulted in elevated GCs [Peanut Agglutinin (PNA⁺) cells], spleens were harvested from B6 controls and 564 Igi mice. As expected, GCs were absent in majority of follicles in naïve B6 spleens. In 564Igi mice, however, GC formed spontaneously; and the frequency of GC was significantly elevated when C4 was absent (Figure 4A and B). Both Id⁺ and Id^-, PNA⁺Fas⁺ B cells were found to be elevated in 564Igi-C4^-/- mice relative to 564Igi-C4^+/+ controls by flow cytometry (Figure 4C and D). In support of increased GC formation, a significantly higher proportion of CD4⁺ follicular T helper cells (Tfh; CXCR5⁺ICOS⁺) were found in C4-deficient animals (Figure 4E). These findings demonstrate that when C4 is absent 564Igi B cells have a greater propensity to form GC.

Fig 4. Absence of complement C4 enhances self-reactive GC formation in 564 Igi mice.

(A) Spleen sections from B6, 564Igi-C4^+/+ and -C4^-/- mice were stained with PNA and antibodies against Id and B220. T, B and GC indicate T-cell area, B-cell follicles and germinal centers respectively. Images were acquired at 20× magnification. Data shown are representative of 3 independent experiments performed with a total of 4 mice. (B) To estimate the number of GCs present per follicle, spleen sections from B6, 564Igi-C4^+/+ and -C4^-/- mice were stained with PNA and anti-CD35 (8C12) antibody. Ratio of PNA⁺ centers to CD35⁺ areas (follicles) per spleen are depicted as a ratio. Images were acquired at 60X magnification. Symbols represent individual mice pooled from 5 independent experiments, bars represent means. (C) 564Igi-C4^+/+ and -C4^-/- splenocytes were stained with PNA, 7AAD and anti-B220 and anti-Fas antibodies. GC B cells were identified as B220⁺ cells that are PNA⁺Fas⁺. Symbols represent individual mice pooled from 4 independent experiments, bars represent means. (D) Splenocytes from 564Igi- C4^+/+ and -C4^-/- mice were stained with PNA, 7AAD and antibodies against B220 and Id. PNA⁺ Id B-cell frequencies are plotted for the two 564 Igi mice strains. Symbols represent individual mice pooled from 4 independent experiments, bars represent means. (E) Follicular helper T cells (Tfh) in spleens from 564 Igi mice were identified as live (7AAD-) CD4⁺ cells that are CXCR5⁺ ICOS⁺. Symbols represent individual mice pooled from 4 independent experiments, bars represent means. *p<0.05, **p<0.001, ***p<0.0001, Student's T-test.

Although self-reactive GC formation was greater when C4 was absent, anti-nucleolar IgG titer in serum was similar between both 564Igi strains (Supporting Information Figure 3A). Thus, the GC reaction in 564Igi-C4^-/- mice was less productive. Moreover, no difference in affinity maturation or isotype-switching of anti-nucleolar antibody was observed between the two 564Igi strains (Supporting Information Figure 3B and 3C). We found that the proportion of Id⁺ B cells undergoing apoptosis in the GC (PNA⁺ AnnV⁺) was significantly higher in 564Igi-C4^-/- mice compared to 564Igi-C4^+/+ control mice (Supporting Information Figure 3D), which could account for the comparative reduction in GC output observed in 564Igi-C4^-/- mice relative to the rate of GC B cell differentiation.

Myeloid C4 restores negative selection in C4-deficient mice

The major source of serum C4 is the liver; however, myeloid cells are sufficient in restoring humoral immunity to foreign antigens in C4-deficient mice. To determine if myeloid cells were sufficient to provide a protective effect to self-antigen in the 564 Igi mice a series of mixed BM chimeras were prepared (Figure 5). Control chimeras were created in which C4^+/+ and 564Igi-C4^+/+ (CD45.2⁺) BM was engrafted at a ratio of 9:1, respectively, into lethally irradiated C4^+/+ (CD45.1⁺) recipients. At this ratio approximately 90% of the myeloid compartment is populated by the C4^+/+ B6 BM and 10 % by the 564 Igi BM. Negative selection at the transitional stage in the spleens of these 9 × (C4^+/+) : 1 × (564Igi-C4^+/+) BM → C4^+/+ control chimeras was intact, as it was in 9 × (C4^+/+) : 1 × (564Igi-C4^-/-) → C4^-/- chimeric mice, indicating that a C4-sufficient myeloid compartment is adequate for the establishment of tolerance in B cells. In contrast, when no C4 was present in this mixed chimeric situation, in 9 × (C4^-/-) : 1 × (564Igi-C4^-/-) → C4^-/- chimeric mice, B-cell selection was significantly diminished. In BM chimeras containing 9 × (C4^-/-) : 1 × (564Igi-C4^+/+) → C4^-/- in which 10% of the BM compartment is able to produce C4, negative selection of autoreactive B-cell specificities occurred but with less efficiency than in an environment containing adequate C4. Thus, the overall results indicate that myeloid cells are a sufficient source of C4 independent of serum C4 to maintain tolerance of the self-reactive B cells.

Fig 5. Myeloid production of C4 is critical for negative selection of self-reactive B cells.

BM chimeric mice were created in which donor BM was mixed 9 parts (of either C4^+/+ or C4^-/-) to 1 part of 564Igi BM (either 564 Igi-C4^+/+ or 564Igi-C4^-/-) before transfer into C4^-/- recipient mice (except the first control panel where the recipients are C4^+/+). Mice were analyzed 6-8 weeks following transfer. Within the donor (CD45.2⁺) population, the proportion of 564Igi Id⁺ cells of a transitional (T; B220⁺ AA4.1^hi) or mature (M; B220⁺ AA4.1^lo) phenotype was analyzed. Symbols represent individual mice pooled from 2 independent experiments. *p<0.05, Student's T-test.

Discussion

Deficiency in complement C4 is a major susceptibility factor in human lupus; and on certain backgrounds predisposes mice to lupus-like disease. To dissect the role of C4 in regulation of auto-reactive B cells specific for a nucleolar antigen, we crossed the C4^-/- line with 564Igi mice [24]. Interestingly, the 564Igi specificity was found to recognize nucleolar autoantigens, many being RNP complexes including SSB/La, a common RNP antigen in both murine and human SLE [30]. Comparison of Id⁺ autoreactive B cells in the C4^-/- and C4^+/+ 564 Igi mice identified a loss of tolerance at the transitional stage in the spleen. This stage of differentiation of immature B cells represents a major step in negative selection of autoreactive B cells in the periphery [31-34].

In C4-sufficient 564 Igi mice, most anti-self B cells are deleted before they reach maturity and those that enter the mature population are, for the most part, maintained in a tolerant state [24]. We found that in the absence of C4, similar frequencies of immature, self-reactive B cells enter the spleen, but many more survive through the maturation stages. The self-reactive B cells were activated in response to BCR and TLR ligation, gained access to follicles and maintained near normal levels of surface IgD and CD21, unlike their counterparts in C4-sufficient mice. However, self-reactive Id⁺ cells from 564Igi-C4^-/- mice showed down-modulation of surface IgM, possibly resulting from internalization of the BCR following self-antigen ligation. Autoreactive B cells from C4-deficient animals spontaneously upregulated CD86 after overnight incubation in culture medium, possibly due to their binding of debris from dying cells in culture, thereby receiving both BCR and TLR signals. In contrast, endogenous non-self-reactive B cells in the same cultures were not activated.

Localization of anergic, auto-reactive B cells to the outer PALS is dependent on the presence of competitor B cells and on constant antigen receptor signaling [27, 35]. Autoreactive B cells have a greater dependence on B-cell activating factor (BAFF) for persistence and are therefore outcompeted by their non self-reactive counterparts in terms of entry and survival in the follicles [36]. In the 564Igi model [24], a diverse B cell pool is present which includes a sizable population of competitor non-transgenic B cells. Therefore, follicular localization of self-reactive B cells in C4-deficient 564Igi mice is not likely due to the lack of competition with endogenous non-self-reactive B cells.

A consequence of continued differentiation of the self-reactive B cells in the C4-deficient 564Igi mice is that they formed spontaneous GCs in higher proportions relative to those in 564 Igi- C4^+/+ mice. Whether this simply reflects a larger precursor frequency in these mice cannot be excluded. GC formation in autoimmune mice is not uncommon; they have been reported for multiple strains of lupus-prone mice such as MRL/lpr, BXSB, NZB/W and ABIN1 [28,29]. Although GC sizes and frequencies vary among the lupus-prone strains, the general size of GC in the 564 Igi mice for both C4^+/+ and C4^-/- are within an average size range at 8 weeks. Moreover, the relative frequencies of PNA⁺Id⁺ GC B cells and CD4⁺ CXCR5⁺ TfH in the 564 Igi-C4^-/- mice was similar to that observed in the mutant ABIN1 (D485N) strain which develop severe lupus-like disease [28]. Despite the relative increase in spontaneous GC in the C4-deficient mice, a pronounced increase in autoantibody titers of the switched isotypes was not observed. One explanation is that in the absence of C4, survival of GC B cells is reduced resulting in the output of fewer effector cells and this was supported by our finding of a relatively higher turnover and increased cell death of Id⁺ GC B cells in the 564 Igi-C4^-/- mice. Thus, C4 protein dependent signaling is required for efficient survival of self-reactive GC B cells as observed for non-self antigens [37].

The current studies build on our previous finding with the hen egg lysozyme Ig double Tg mice (anti-Hel × sHel) mice where we identified a less anergic phenotype of the auto-reactive B cells in mice crossed onto the Cr2^-/- or C4^-/- background relative to complement-sufficient controls [20]. In the current study, we identify the stage of B-cell differentiation (transitional) in which complement C4 is critical in negative selection against RNP antigens. In the mixed bone marrow chimeric mice, use of C4 deficient recipients negated any role for liver-derived complement C4, which is otherwise the major source of systemic C4 and instead indicated that myeloid cell derived C4 is responsible for the observed effects on B cell selection. Thus our finding that myeloid compartment dictates the fate of the auto-reactive B cells indicates that local C4 secretion is sufficient for regulation of self-reactive B cells and suggests several testable hypotheses. C4, like C1q, opsonizes apoptotic cells and enhances their clearance by phagocytic cells [38]. In its absence, excess apoptotic debri may accumulate and form immune complexes that are known to stimulate plasmacytoid DC to secrete Type I interferon in a TLR- dependent pathway [39] (Supporting Information Figure 4). For example, Elkon and colleagues [40] found that human DC cultured with lupus serum and apoptotic cells were activated and secreted type I interferon. Strikingly, adding C1q to the immune complexes significantly reduced the response [40]. Whether C4 protein has a similar effect on dampening the response to RNP-IC in the 564 Igi mice is not clear but is a testable model. Elevated levels of type I interferon could promote escape of self-reactive B cells from negative selection by inducing expression of TLR7 at the immature B-cell stage. Thus, in a two signal pathway as discussed above for activation of mature B cells, the combined signaling could induce B-cell escape of tolerance. Alternatively, IFN-α is known to induce DC to secrete B-cell activating factor (BAFF) [41]. BAFF is known to lower the threshold for escape of negative selection of self-reactive B cells at the transitional stage leading to escape of tolerance by the C4-deficient B cells [42].

In summary, the current study suggests a novel paradigm to explain a long-standing observation that deficiency in complement C4 almost always leads to lupus in humans. Thus, the finding that C4 participates in tolerance of lupus-antigen specific B cells at the transitional stage and that the effect is myeloid cell dependent suggests that complement regulates activation of myeloid cells which in turn influence differentiation of immature B cells.

Material and Methods

Mice

Generation of 564 Ig heavy and light chain knock-in (564Igi) mice has been described [24]. C4-deficient mice [43] were bred with 564Igi mice to generate 546Igi-C4^-/- mice. C57BL/6 mice were purchased from Jackson Laboratory. PCR analysis on tail DNA for the presence of transgenic heavy and light chains and FACS analysis of peripheral blood B cells for expression of the 564 receptor were performed. Mice expressing the transgenic receptor and heterozygous for transgenic [44] and endogenous IgH alleles (IgM^b) were used in experiments. Animals used were housed in specific pathogen free facility of Immune Disease Institute and Harvard Medical School and all procedures were conducted according to the institutional guidelines.

Detection of 564 antigens

Hep-2 substrate slides (Scimedx Corp.) were incubated with mouse sera, probed with Alexa488-conjugated goat anti-mouse IgG (Molecular Probes) and analyzed microscopically. Nuclear and cytoplasmic extracts from P3Ag myeloma cells were subjected to immunoprecipitation using either 564 (IgG2b) monoclonal antibody (clone C11) or an isotype control antibody under non-reducing conditions (IP wash buffer: 50 mM Tris-HCl, 300mM NaCl, 5mM EDTA and 0.1% Triton X-100 pH 7.4). Precipitated antigens were separated by gel electrophoresis, bands excised and sent to BIDMC/NRB Mass Spectrometry Facility, Beth Deaconess Medical Center (Boston MA 02115) for analysis.

Immunohistology

Spleens were frozen at –80°C in Tissue-Tek O.C.T embedding media [46]. Sections (6-8μm) were cut, fixed in cold acetone, blocked with BSA and 10% Fc-block, stained in PBS containing 1% BSA and 0.05% Tween-20.

Microscopy

Stained slides were mounted using Gel/Mount media (Biomeda Corp). Images were acquired with confocal microscope (BioRad) using SCAN software (BioRad).

Western Blot

Nucleolar preparations were lysed using RIPA buffer (Cell signaling technologies), cleared using donkey serum (Jackson Immunoresearch) and then immunoprecipitated with Protein G Agarose (Santacruz) and anti-Id antibody (clone 9D11). Mouse IgG was used as a control antibody. The precipitated proteins were run on an SDS-PAGE, probed with anti-SSB/La antibody and developed using a chemiluminescence kit (Pierce).

Antibodies

For FACS and microscopy, anti-564 (anti-idiotype; IgG1) monoclonal antibody (clone 9D11) was prepared and conjugated in-house. Monoclonal antibodies to CD3, CD4, CD5, CD19 (1D3), CD21/35 (7G6) CD86, CXCR5 were purchased from BD Biosciences and that to CD23, CD24, CD43, CD93, ICOS, B220 (RA3-6B2) from BioLegend. FITC-labeled anti-mouse IgG and IgM antibodies were from Southern Biotech. For microscopy studies, anti-CD3e (145-2C11), anti-B220 (RA3-6B2), anti-CD35 (8C12) antibodies were made in-house. Other antibodies used for microscopy were FITC-labeled PNA (EY laboratories, Inc), anti-Armenian hamster IgG-Alexa 488 (Jackson ImmunoResearch) and streptavidin-Alexa 633 (Molecular Probes).

Activation of B lymphocytes in vitro

Naïve (CD43^-) B cells from spleens were magnetically (Miltenyi Biotech) enriched to more than 90% purity. Cells were cultured in IMDM medium (Gibco) supplemented with 10% heat-inactivated fetal calf serum, 100 IU/ml penicillin, 100 mg/ml streptomycin (Gibco), 2 mM L-glutamine (Gibco) and 50 mM of 2-mercaptoethanol (Merck) in 96 well plates (Costar) at 1×10⁶ cells per well. Stimuli were added at the following final concentrations: 2 μg/ml goat anti-mouse IgM F(ab’)₂ (Jackson ImmunoResearch, Inc.), 20 ng/ml recombinant murine IL-4 (e-Bioscience), 2 μg/ml rat anti-mouse CD40 (e-Bioscience), 20 μg/ml LPS (Sigma), 10nM CpG oligonucleotide (ODN1826, 5’-tccatgacgttcctgacgtt-3’; Invivogen) and 100 nM gardiquimod (Invivogen).

FACS analysis

Generation of single-cell suspension, FACS staining and analysis were done as described earlier [45]. All analyses were performed on live-gated cells (7AAD-negative; BioLegend).

Generation of BM chimeric mice

Recipient mice were lethally irradiated with two doses of 550 Rad three hours apart then injected i.v. with 1×10⁷ donor-derived BM cells. Recipient mice were fed with neomycin-containing water for two weeks following reconstitution and analyzed 6-8 weeks after reconstitution.

Statistical Analysis

Results were expressed as the mean ± standard error of the mean (SEM). Differences between groups were analyzed using the Student's t tests. In the figures, p value < 0.05 has been depicted as ‘*’, p value < 0.001 as ‘**’ and p value < 0.0001 as ‘***’ for convenience.

Abbreviations

RNP

ribonucleoprotein

SSB/La

Sjögren's Syndrome antigen B

Id

Idiotype

PNA

Peanut Agglutinin

Tfh

follicular T helper cells

BAFF

B cell activating factor