Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Mar 21.
Published in final edited form as: Immunity. 2013 Mar 14;38(3):528–540. doi: 10.1016/j.immuni.2012.11.017

Signals via the Adaptor MyD88 in B cells and DCs Make Distinct and Synergistic Contributions to Immune Activation and Tissue Damage in Lupus

Lino L Teichmann 1,4, Dominik Schenten 2, Ruslan Medzhitov 2, Michael Kashgarian 3, Mark J Shlomchik 1,2
PMCID: PMC3638041  NIHMSID: NIHMS456340  PMID: 23499488

Summary

Detection of self nucleic acids by Toll-like receptors (TLR) preciptates autoimmune diseases, including systemic lupus erythematosus (SLE). It remains unknown how TLR signals in specific cell types contribute to distinct manifestations of SLE. Here, we demonstrate that formation of anti-nuclear antibodies in MRL.Faslpr mice entirely depends on the TLR signaling adaptor MyD88 in B cells. Further, MyD88 deficiency in B cells ameliorated nephritis, including antibody-independent interstitial T cell infiltrates, suggesting that nucleic acid-specific B cells activate nephrotoxic T cells. Surprisingly, MyD88 deletion in dendritic cells (DCs) did not affect nephritis, despite the importance of DCs in renal inflammation. In contrast, MyD88 in DCs was critical for dermatitis, revealing a separate pathogenetic mechanism. DC-expressed MyD88 promoted interferon-α production by plasmacytoid DCs, which was associated with Death domain-associated protein 6 upregulation and B lymphopenia. Our findings thus reveal unique immunopathological consequences of MyD88 signaling in B cells and DCs in lupus.

Introduction

Activation of the immune system by aberrant self nucleic acid detection has emerged as a fundamental mechanism in the pathogenesis of various autoimmune diseases, in particular systemic lupus erythematosus (SLE). In murine models of SLE, self nucleic acids stimulate Toll-like receptor 7 (TLR7) and TLR9, endosomal receptors that normally protect from infection by detecting foreign nucleic acids (Christensen et al., 2006). Both TLR7 and TLR9 signal via the adaptor MyD88. Deficiency for MyD88 abrogates most attributes of lupus in several lupus-prone mouse strains including MRL.Faslpr mice with polygenic susceptibility (Nickerson et al., 2010). Myd88−/− MRL.Faslpr mice have minimal organ disease and lack auto-Ab specificities that are dependent on TLR7 (anti-RNA, Anti-Sm) as well as TLR9 (anti-dsDNA, anti-nucleosome). MRL.Faslpr mice doubly-deficient for TLR7 and TLR9 largely mirror the Myd88−/− MRL.Faslpr phenotype (Nickerson et al., 2010). The similarity in phenotypes of these two mutant strains indicates that combined disruption of TLR7 and TLR9 signaling accounts for the loss of characteristic lupus features in Myd88−/− MRL.Faslpr mice, whereas interference with other MyD88-dependent pathways, such as IL-1 and IL-18 receptor signaling, does not essentially contribute to the Myd88−/− MRL.Faslpr phenotype.

These studies of global gene-deficiency in lupus-prone mice leave several fundamental questions unanswered. First, although B cells, DCs and other myeloid cells express TLR7 and TLR9, it is uncertain which of these cell types are directly activated by nucleic acid-sensing TLRs in lupus and what the requirements for this activation are. In lupus, B cells and DCs become spontaneously activated and can promote disease by several mechanisms. B cells and conventional DCs (cDCs), are major antigen (Ag)-presenting cells that regulate T cell-mediated autoimmunity (Chan and Shlomchik, 1998; Teichmann et al., 2010). B cells further contribute to lupus pathogenesis by auto-Ab secretion. Plasmacytoid DCs (pDCs) likely drive disease by type I interferon (IFN-I) production (Rönnblom et al., 2011).

In-vitro experiments suggest that lupus auto-Ags can indeed directly activate B cells and DCs via nucleic acid-specific TLRs. In B cells, targeting of mammalian DNA, RNA or immune complexes (ICs) that contain nucleic acids by the B cell receptor to endosomal TLRs leads to robust proliferation (Marshak-Rothstein, 2006). In DCs, uptake of ICs and delivery to endosomes mediated by FcγRs induces cytokine secretion (Lövgren et al., 2006). Of note, if DCs require nucleic acids to be complexed with auto-Abs for effective shuttling to endosomal TLRs, then TLR-mediated DC activation would depend on prior B cell activation. However, entry of nucleic acids to endosomes in DCs might also be facilitated by high-mobility group box-1 protein (HMGB1) (Tian et al., 2007) and the antimicrobial peptide LL37 (Lande et al., 2007).

Second, assuming that B cells and DCs are directly activated by endogenous TLR ligands in lupus, it is not clear whether direct TLR-driven activation is an indispensable requirement for these cell types to promote lupus pathogenesis. Conceivably other modes of activation could compensate for a loss of TLR stimulation. B cells for example can effectively be activated by synergistic engagement of the B cell receptor and CD40 (Bishop and Hostager, 2003), leading to class switch recombination and Ab forming cell (AFC) differentiation. In immunization studies, cDCs required direct activation by pattern recognition receptors, such as TLRs, to induce T cell differentiation (Joffre et al., 2009). Yet, in systemic autoimmunity this prerequisite might be weakened and cDCs might have disease-relevant functions that extend beyond T cell priming. Further, whether TLR-activation is necessary for pDCs to secrete IFN-I in lupus has not been established.

Finally, the functional consequences of TLR-mediated activation of distinct cell types by self nucleic acids for lupus development are unknown. Human SLE and murine lupus both display diverse clinical manifestations such as auto-Ab formation, excessive production of IFN-I and interferon γ (IFN-γ), hematological changes, organ infiltration, lymphadenopathy and splenomegaly. It is important to understand which aspects of disease pathogenesis specifically require B cells, DCs or both cell types to be directly activated by nucleic acid-sensing TLRs.

Dissecting how TLR signals in two major types of Ag-presenting cells regulate lupus requires the application of genetic tools for tissue-specific gene inactivation to a murine polygenic model of lupus. We combined two approaches—Cre-loxP recombination and mixed bone marrow chimeras—to inactivate Myd88, and hence TLR7 and TLR9 signaling, selectively in B cells or DCs in lupus prone MRL.Faslpr mice. This strategy allowed us to define the specific contributions of B cell- and DC-intrinsic MyD88 signaling to self-reactive B and T cell stimulation and how MyD88 signals in B cells and DCs cooperate with and complement each other in establishing the full clinical picture of lupus.

Results

Selective deletion of Myd88 in DCs and B cells

To study the role of MyD88 in DCs and B cells in lupus we generated CD11c-Cre (Caton et al., 2007) and CD19-Cre (Rickert et al., 1997) Myd88fl/fl MRL.Faslpr mice (called DC-Myd88Δ and B-Myd88Δ mice, respectively, hereafter). In DC-Myd88Δ mice deletion of the Myd88fl allele, as measured by quantitative real-time (qRT) PCR on genomic DNA, was 95% in cDCs (Table S1)—virtually complete given a sorting purity of ~97% for cDCs—and 78% in pDCs. Because CD11c expression is not restricted to DCs we also determined Myd88fl deletion in lymphoid cells, which was limited (<25%, Table S1). Of note, we have previously shown that in CD11c-Cre MRL.Faslpr mice Cre is not substantially expressed in marginal zone or metallophilic macrophages (Teichmann et al., 2010). In B-Myd88Δ mice we assessed deletion efficacy by qRT-PCR and by in-vitro CFSE proliferation assay using sorted splenic B cells (Table S1). Both assays showed that CD19-Cre mediated deletion of the Myd88fl allele was quite effective in B cells. Around 90% of Myd88fl alleles were deleted as measured by qRT-PCR and 87% of B cells did not divide in response to TLR9 stimulation. However, we reasoned that if there were selective pressure in favor of MyD88-sufficient self-reactive B cells to produce AFCs in lupus prone B-Myd88Δ mice, the extensive self-Ag-driven clonal expansion of residual MyD88-responsive B cells (~10%) could permit such cells to generate a substantial anti-self humoral immune response. In support of this notion, we found that in contrast to 90% deletion of the Myd88fl allele in B cells, deletion by qRT-PCR in splenic AFCs was only 48% (Table S1), revealing that MyD88-sufficient B cells did have an advantage over their MyD88-deficient counterparts to become AFCs. Thus, the utility of the CD19-Cre MyD88fl/fl approach to evaluate the dependence of lupus phenotypes on MyD88 in B cells is limited to instances in which partial MyD88-deficiency in B cells does impair an aspect of disease.

In view of this limitation, to generate lupus-prone mice with complete deletion of Myd88 in B cells, we lethally irradiated B cell-deficient Jh MRL.Faslpr mice and reconstituted them with a mixture of BM from Jh MRL.Faslpr (80%) and Myd88−/− MRL.Faslpr (20%) mice. In such chimeras (called B-Myd88−/− mice hereafter) all B cells originate from Myd88−/− stem cells whereas other BM-derived cell types are mostly (80%) of wild-type phenotype (Table S1). Irradiated lupus-prone mice develop a robust spontaneous anti-self B and T cell immune response but a caveat is that end-organ disease remains mild. Thus, except for data on organ manifestations, we present results from B-Myd88−/− chimeric mice in the main figures and results from B-Myd88Δ mice in supplemental figures.

B cell-intrinsic MyD88 signaling is absolutely required for ANA and essential for rheumatoid factor formation

In murine vaccination models the magnitude of the IgG response against the Ag is controlled by MyD88 in B cells or DCs depending on the type of the Ag (Hou et al., 2011). We asked whether MyD88 in B cells or DCs regulates the anti-self humoral immune response in lupus. Total serum IgG was decreased by 47% in B-Myd88−/− mice (Figure 1A) (35% in B-Myd88Δ mice (Figure S1A)), while deletion of Myd88 in DCs did not affect the serum IgG concentration (Figure 1B). However, the isotype subclass distribution of IgG was altered in DC-Myd88Δ mice, with an increase in IgG1 and a reduction in IgG2a. IgG anti-nuclear Abs (ANA) were completely absent in B-Myd88−/− mice whereas sera from littermate controls invariably produced homogenous staining patterns, indicating the presence of anti-DNA and anti-chromatin Abs (Figure 1C). All sera from DC-Myd88Δ mice contained IgG ANAs (data not shown). The distribution of dominant ANA staining patterns was similar in DC-Myd88Δ and control mice (Figure 1D). This suggested that ablation of Myd88 in DCs does not lead to a major loss or gross alteration in the composition of auto-Abs. Similar to the total IgG isotype subclass distribution, there was a relative increase of IgG1 compared to IgG2a ANAs in DC-Myd88Δ (Figure 1E).

Figure 1. B-cell intrinsic MyD88 signaling is absolutely required for ANA and critical for rheumatoid factor formation.

Figure 1

Open in a new tab

(A and B) Serum Ig isotype concentrations of (A) control (n = 12) and B-Myd88−/− (n = 12), or (B) control (n = 34) and DC-Myd88Δ (n = 22) mice.

(C) Representative Hep-2 ANA staining patterns of a control and a B-Myd88−/− mouse using serum dilutions of 1:100. ANAs are routinely detectable at serum dilutions of 1:16,000 to 1:32,000 in wild type MRL.Faslpr mice (not shown). Scale bars = 50 μm.

(D) ANA staining patterns produced by sera from control and DC-Myd88Δ mice. The number in the circles indicate the number of mice analyzed.

(E) Ratio of mean nuclear fluorescence intensity IgG1 divided by mean nuclear fluorescence intensity IgG2a. n = 11 for control and n = 23 for DC-Myd88Δ mice.

(F and G) Serum concentrations of anti-nucleosome IgG, IgG1 and IgG2a of (F) control (n = 12) and B-Myd88−/− (n = 12), or (G) control (n = 34) and DC-Myd88Δ (n = 22) animals.

(H and I) Number of anti-IgG1 and IgG2a rheumatoid factor secreting cells per spleen determined by ELISpot assay. (H) n = 6—12 for control and B-Myd88−/− mice, (I) n = 11—12 for control and n = 16 for DC-Myd88Δ mice.

Error bars show SEM. See also Fig. S1.

To quantitate serum auto-Abs we measured anti-nucleosome IgG, IgG1 and IgG2a by ELISA. As expected, anti-nucleosome IgG was virtually absent in B-Myd88−/− mice and 83% reduced in B-Myd88Δ mice (Figure 1F and Figure S1B). In DC-Myd88Δ mice serum anti-nucleosome IgG was 47% decreased, compared to littermate controls (Figure 1G). The reduction in anti-nucleosome IgG in DC-Myd88Δ mice was mainly a result of less IgG2a anti-nucleosome (Figure 1G). Thus, MyD88-dependent DC activation does contribute albeit weakly to auto-Ab production. To assess the rheumatoid factor (RF, anti-Fcγ auto-Abs) response, we enumerated RF-secreting cells by ELISpot assay. This approach is preferable to measuring serum concentrations because alterations in IgG serum concentrations in B-Myd88−/− and DC-Myd88Δ mice could lead to apparent changes in RF concentrations. Strikingly, spleens of B-Myd88−/− mice harbored 98% fewer anti-IgG1 and 99% fewer anti-IgG2a forming cells than controls (Figure 1H). Given the dominant role of TLR7 and TLR9 upstream of MyD88 in lupus-prone mice (Nickerson et al., 2010), this implies that the RF response is chiefly driven by direct B cell activation by ICs that contain self nucleic acids. In contrast, DC-Myd88Δ mice had only a 60% reduction in anti-IgG2a and no significant decrement (p = 0.29) in anti-IgG1 secreting cells (Figure 1I).

Taken together, ANA formation was entirely dependent and RF formation largely dependent on B cell-intrinsic MyD88 signaling. DCs activated through MyD88 contributed only modestly to auto-Ab generation, either directly, for example by secretion of B cell-trophic factors such as BAFF or APRIL (Rickert et al., 2011), or indirectly, for example by priming of T helper cells that augment TLR-driven autoreactive plasmablast responses (Odegard et al., 2008; Sweet et al., 2011).

AFC generation and class switch recombination in the spleen critically depend on MyD88 in B cells

Anti-DNA and RF AFCs in MRL.Faslpr mice are mainly located at extrafollicular sites in the spleen (Shlomchik, 2008). We determined the number of splenic AFCs by ELISpot and flow cytometry (intracellular-κhiCD138+CD19intCD44+TCRβ−). Corresponding to the profound impact of B cell-specific Myd88 deletion on serum auto-Abs concentrations, spleens from B-Myd88−/− mice contained around 90% fewer κ light chain Ab producing AFCs (Figure 2A, 2B and 2F). Most splenic AFCs in MRL.Faslpr mice are short-lived plasmablasts. Notably, residual AFCs in B-Myd88−/− mice displayed a higher κ light chain content and lower forward scatter as assessed by flow cytometry (Figure 2F), suggesting that such cells may represent more fully differentiated plasma cells rather than short-lived plasmablasts. Staining for active caspases confirmed that residual AFCs in B-Myd88−/− mice underwent apoptosis at a lower rate than their counterparts in controls (Figure 2C). In contrast, AFCs numbers in B-Myd88Δ mice, with only partial deletion of Myd88 in B cells, were unaffected (Figure S2A), indicative of very strong positive selection for the few B cells (~10%) that did not fully inactivate Myd88. AFC numbers were also not reduced in DC-Myd88Δ mice (Figure 2D and 2E) consistent with the modest contribution of DC-intrinsic MyD88 signaling to anti-nucleosome IgG and anti-IgG2a RF generation.

Figure 2. The spontaneous plasmablast response largely depends on B cell-expressed MyD88.

Figure 2

Open in a new tab

(A and B) Number of κ light chain AFCs per spleen determined by (A) ELISpot assay and (B) flow cytometry (intracellular-κhiCD138+CD19intCD44+TCRβ−). n = 12 for control and B-Myd88−/− mice.

(C) Staining of splenic κ light chain AFCs for active caspases to identify apoptotic cells. n = 4 for control and n = 5 for B-Myd88−/− mice.

(D and E) As in (A and B) but for control (n = 28) and DC-Myd88Δ mice (n = 16).

(F) Representative staining profiles of splenic AFCs of control (n = 12) and B-Myd88−/− mice (n = 12). Cells were first gated on CD44+TCRβ−. Values indicate percentage of live splenocytes (mean).

(G) ELISpot assays for IgG1, IgG2a and IgM AFCs in the spleen were performed. Numbers of IgG1 and IgG2a spots were normalized to numbers of IgM spots to demonstrate the effect of B cell-specific Myd88 deletion on class switch recombination. n = 12 for control and B-Myd88−/− mice.

(H) As in (G) but for control (n = 28) and DC-DC-Myd88Δ mice (n = 16).

Error bars show SEM. See also Fig. S2.

Class switching—measured by the ratios IgG1:IgM and IgG2a:IgM ELISpots—was markedly impaired in B-Myd88−/− mice (Figure 2G). In DC-Myd88Δ mice we found a small decrease in IgG2a and a more pronounced increase in IgG1 AFCs relative to IgM AFCs (Figure 2H), arguing against a general defect in class switch recombination. The shift from IgG2a to IgG1 in DC-Myd88Δ mice, which was also observed in IgG serum concentrations, is reminiscent of the phenotype of IFN-γ deficiency in MRL.Faslpr mice (Balomenos et al., 1998). IFN-γ is known to promote isotype switch to IgG2a.

The mechanisms that lead to extrafollicular B cell responses in MRL.Faslpr and many other lupus prone mouse strains—including BXSB/Yaa, NZB/W F1, B6.Sle123 and BAFF Tg—are poorly understood (Shlomchik, 2008). Evidence for the involvement of DCs in this process comes from C57BL/6 mice with DC-specific deficiency for the ubiquitin-editing protein A20, which display a strong extrafollicular plasmablast response (Kool et al., 2011). We investigated localization of the B cell response in control and DC-Myd88Δ mice by immunofluorescence microscopy. Spontaneous GC formation is minimal in older MRL.Faslpr mice, but the reasons for this are unknown. Remarkably, in spleens of DC-Myd88Δ mice, clusters of PNA+ cells that contain zones with and without follicular dendritic cells, indicative of GCs, were present (Figure S2B and S2C). In contrast, GCs were undetectable in control mice (Figure S2B), as expected. GC-like structures were also undetectable in B-Myd88−/− mice (data not shown), which was unsurprising given that MyD88 in B cells was essential for spontaneous B cell activation. These data suggest that persistent activation of DCs through the TLR-MyD88 pathway interferes with GC formation, shedding light on the mechanisms leading to a paucity of GCs in these lupus-prone mice.

In summary, the great majority of splenic AFCs in MRL.Faslpr mice were generated as a result of B cell-intrinsic MyD88 signaling. MyD88 in B cells also promoted class switch recombination whereas DC-expressed MyD88 influenced the direction of isotype switching. Continuous MyD88 signaling in DCs appeared to hinder GC formation.

DC- and B cell-intrinsic MyD88 signals regulate the size of the B cell compartment

B lymphocytopenia is a common feature of lupus patients with active disease and is also found in aged MRL.Faslpr mice. Notably, both, B-Myd88−/− and DC-Myd88Δ mice had increased B cell numbers compared to their respective littermate controls (Figure 3A and 3B). In aged MRL.Faslpr mice most B cells are located in the marginal zone rather than in follicles. As anticipated, immunofluorescence histology of spleens from littermate controls of B-Myd88−/− and DC-Myd88Δ mice demonstrated a lack of well-developed B cell follicles (Figure S3A and S3B). In contrast, B cell follicles were readily detectable in DC-Myd88Δ and B-Myd88−/− mice (Figure S3A and S3B). Thus, MyD88-dependent signal transduction in both B cells and DCs led to reduced overall numbers of B cells along with suppressed follicular B cell localization.

Figure 3. MyD88 signaling in DCs impaires B cell lymphopoiesis.

Figure 3

Open in a new tab

(A and B) Numbers of B cells (CD19hiCD138−) per spleen in (A) control (n = 12) and B-Myd88−/− (n = 12), or (B) control (n = 29) and DC-Myd88Δ (n = 16) mice.

(C) Gene expression of BAFF and APRIL in sorted splenic cDCs from 15 wk old control (gray, bars, n = 5) and DC-Myd88Δ (white bars, n = 6) mice as determined by qRT-PCR. Data represent normalized expression values relative to 6 wk old pre-autoimmune wildtype MRL.Faslpr mice.

(D) B cell lymphopoiesis in the bone marrow of control (n = 6) and DC-Myd88Δ (n = 7) mice. Shown are representative staining profiles of Hardy fractions A—E. Cells were first gated on CD11b−TCRβ−CD49b−. Values indicate percentage (mean) of parent gate.

(E and F) Number of (E) cells in Hardy fraction A—E and (F) total bone marrow cells after ACK lysis in the femur and tibia of both legs of control (n = 6) and DC-Myd88Δ (n = 7) mice.

(G) Cell death in Hardy fraction B+C cells. Dying cells were identified by staining with EMA. n = 6 for control and n = 7 for DC-Myd88Δ mice.

(H) Gene expression of DAXX in sorted bone marrow Hardy fraction A and B+C cells from 13—16 wk old control (gray bars, n = 5) and DC-Myd88Δ (white bars, n = 5) mice as determined by qRT-PCR. Data represent normalized expression values relative to total bone marrow cells from control mice.

(I) Expression of IFN-α genes, using pan-Ifna primers, in sorted bone marrow macrophages, pDCs and neutrophils from 16 wk old control (gray bars, n = 6) and DC-Myd88Δ (white bars, n = 6) mice as determined by qRT-PCR. Data represent normalized expression values relative to total bone marrow cells from control mice.

Error bars show SEM. See also Fig. S3.

We sought to further define how MyD88 signals in B cells and DCs promote alterations in the B cell compartment of MRL.Faslpr mice. B cells in spleens from B-Myd88−/− mice showed less proliferation (Figure S3C) and apoptosis (Figure S3D) suggesting less activation-induced cell death. Additionally, the larger follicular B cell population in B-Myd88−/− mice could partly be the result of impaired plasmablast differentiation, which normally might deplete follicular B cells (Figure 2A and 2B). However, such a process would not contribute to increased follicular B cell numbers in DC-Myd88Δ mice, as impaired AFC differentiation was not observed in these mice (Figure 2D and 2E). We hypothesized that in DC-Myd88Δ mice DCs produce greater amounts of BAFF, a B cell survival factor. Yet, BAFF mRNA levels in DCs from DC-Myd88Δ mice were substantially lower than those in DCs from control mice (Figure 3C), and similar results were obtained for the related cytokine APRIL. Further, Myd88 deletion in DCs did not affect proliferation or survival of B cells in the spleen (data not shown).

Because DC-specific Myd88 deletion had no detectable effect on peripheral B cell homeostasis, we asked whether B lymphopoiesis is regulated by DC-intrinsic MyD88 signaling. Increases in B cell progenitors in DC-Myd88Δ mice were evident starting at Hardy fraction B+C (CD45R+CD43+CD19+) (Figure 3D and 3E) corresponding to late pro-B and early pre-B cells. This effect was specific, as total bone marrow cellularity was unaffected by Myd88 deletion in DCs (Figure 3F). Hardy fraction B+C cells had a lower frequency of cell death in the absence of MyD88 in DCs (Figure 3G). Because survival and growth of fraction B+C cells is mediated by IL-7 (Hardy et al., 2007) those findings suggested stronger IL-7R signaling in DC-Myd88Δ mice. Importantly, IL-7R signaling is inhibited by IFN-I. This inhibition is dependent on the transcriptional repressor DAXX (Death domain-associated protein 6) (Gongora et al., 2001). Indeed, Hardy fraction B+C cells from control mice had a higher DAXX mRNA content than those from DC-Myd88Δ mice (Figure 3H). To identify the main source of IFN-I in the bone marrow we performed qRT-PCR for Ifna transcripts on various myeloid cell types. pDCs showed the highest Ifna mRNA levels and Ifna transcription in pDCs was strongly dependent on DC-MyD88 (Figure 3I). These data are consistent with a model in which MyD88 signaling in bone marrow pDCs leads to the production of IFN-α, which mediates DAXX-dependent inhibition of IL-7-induced growth and survival of B cell progenitor cells.

Full activation of cDCs in systemic autoimmunity requires DC- and B cell-intrinsic MyD88 signaling

We assessed whether Myd88 deletion in DCs affected their number and activation state. The frequency (Figure 4A) but not number (Figure 4B) of cDCs in the spleens of DC-Myd88Δ mice was slightly reduced. Frequency and number of pDCs were not altered (Figure 4A and 4B). cDCs from DC-Myd88Δ and control mice expressed comparable amounts of the co-stimulatory molecules CD40, CD80 and CD86 (Figure 4C). To assess cytokine production we performed qRT-PCR on freshly harvested, unmanipulated splenic cDCs from 16 wk old diseased DC-Myd88Δ and control mice, as well as pre-autoimmune 6 wk old wild-type MRL.Faslpr mice. mRNA amounts of p35 and p40—which encode the two subunits of IL-12—as well as Il6 were lower in cDCs from aged DC-Myd88Δ mice than in cDCs from aged controls (Figure 4D). Transcripts of p40 and Il6 in cDCs from old DC-Myd88Δ mice were even fewer than in those from young wild-type MRL.Faslpr mice (both p = 0.024). In contrast there was a slight increase in Tgfb1 transcripts in cDCs from old DC-Myd88Δ mice compared to those from old controls. Thus, in spontaneous age-dependent lupus, competence to synthesize IL-12 or IL-6, which are required for priming of Th1 or Th17 and Th22 cells, respectively, is critically dependent on DC-intrinsic MyD88 signaling.

Figure 4. Inflammatory cytokine production by cDCs is regulated by DC- and B cell-intrinsic MyD88 signaling.

Figure 4

Open in a new tab

(A) Representative staining profiles of splenic cDCs (CD11chiMHCII+CD19−) and pDCs (Siglec-H+BST2+) of control (n = 29) and DC-Myd88Δ (n = 16) mice. Values indicate percentage of live splenocytes (mean).

(B) Numbers of cDCs and pDCs in the spleen of control and DC-Myd88Δ mice gated as in (A).

(C) Representative staining histograms of gated cDCs from control (n = 29) and DC-Myd88Δ (n = 16) mice. Values indicate the mean of the MFI (mean fluorescence intensity).

(D) Gene expression of Il1b, Il6, p35, p40, Tgfb1 and Il10 in sorted splenic cDCs from 16 wk old control (n = 5) and DC-Myd88Δ mice (n = 6) as determined by qRT-PCR. Data represent normalized expression values relative to 6 wk old pre-autoimmune wildtype MRL.Faslpr mice.

(E) As in (C) but for control (n = 8) and B-Myd88−/− (n = 10) mice.

(F) As in (D) but for control (n = 4) and B-Myd88−/− (n = 4) mice. cDCs were isolated from BM chimeras 21 wks after transplantation.

Error bars show SEM.

Interestingly, we found that MyD88 signaling in B cells also controls DC activation in lupus. cDCs from B-Myd88−/− mice had lower surface expression of CD40 and CD86 than cDCs from controls (Figure 4E). Like cDCs from old DC-Myd88Δ mice, cDCs from aged B-Myd88−/− mice also contained fewer p35, p40 and Il6 transcripts and more Tgfb1 transcripts compared to cDCs from aged controls (Figure 4F). This result implies a forward loop in which B cells activated through MyD88 contribute to the environment in which cDCs become activated by DC-intrinsic MyD88 signaling, consistent with the notion of cDCs being activated by nucleic acid-containing ICs (Marshak-Rothstein, 2006). Together, these results indicate that both DC- and B cell-intrinsic MyD88 signaling are necessary to endow cDCs with optimal ability to prime a T helper cell response.

T cell activation depends on MyD88 in B cells but not DCs whereas T cell accumulation relies on MyD88 in both cell types

We next sought to clarify the contributions of MyD88 signals in B cells or DCs to the spontaneous T cell response. The characteristic splenomegaly and lymphadenopathy of MRL.Faslpr mice, owing largely to the accumulation of activated T cells, was substantially reduced in both B-Myd88−/− and DC-Myd88Δ mice (Figure 5A and 5B). Accordingly, splenic T cell numbers were lower in mice without MyD88 in B cells or DCs (Figure 5C). In B-Myd88−/− mice the percentage of CD4+ T cells with an activated-memory phenotype (CD44+) was appreciably lower (73.1%) than in controls (90.0%) (Figure 5D and S4A). Ablation of Myd88 in DCs did not impair spontaneous CD4+ T cell activation (Figure 5D and S4A), consistent with the minor role of DCs in CD4+ T cell activation in MRL.Faslpr mice (Teichmann et al., 2010). For comparison, we confirmed the previously reported finding that CD4+ T cell activation is strongly dependent on B cells in the MRL.Faslpr strain (Chan and Shlomchik, 1998). In B cell-deficient Jh MRL.Faslpr mice 38.8% of CD4+ T cells displayed an activated-memory phenotype compared to 86.6% in controls (Figure 5D and S4A). The greater decrease in CD4+ T cells with an activated-memory phenotype in B cell-deficient MRL.Faslpr compared to B-Myd88−/− mice indicates that there are MyD88-dependent and independent functions of B cells in activating CD4+ T cells. Overall, MyD88 signaling in B cells but not DCs promoted CD4+ T cell activation in lupus.

Figure 5. MyD88 signals in B cells and DCs promote Th1 and Tefh differentiation.

Figure 5

Open in a new tab

(A–C) Weight of (A) spleens and (B) the two largest axillary lymph nodes, and (C) T cell numbers (TCRβ+CD19−) per spleen of (upper graphs) control (n = 12) and B-Myd88−/− (n = 12) mice, or (lower graphs) control (n = 34) and DC-Myd88Δ mice (n = 22).

(D) CD44 and CD62L staining profiles of gated CD4+ T cells to identify naïve (CD44loCD62Lhi) and activated-memory (CD44hi) subpopulations. Contour plots show representative examples of B-Myd88−/− (upper panels), Jh MRL.Faslpr (middle panels) and DC-Myd88Δ (lower panels) mice with their respective controls. Values indicate percentage of CD4+ T cells (mean).

(E) FoxP3 and CD25 staining profiles of gated CD4+ T cells to identify Tregs. Contour plots show representative examples of B-Myd88−/− (upper panels) and DC-Myd88Δ (lower panels) mice with their respective controls. Values indicate percentage of CD4+ T cells (mean).

(F) Representative intracellular IFN-γ staining histograms of gated CD4+ (left side) and CD8+ (right side) T cells of PMA plus ionomycin stimulated splenocytes from B-Myd88−/− (upper panels), Jh MRL.Faslpr (middle panels) and DC-Myd88Δ (lower panels) mice with their respective controls. Values indicate percentage of CD4+ or CD8+ T cells (mean).

(G) CD62L and PSGL-1 staining profiles of gated CD4+ T cells after exclusion of CD45R+ cells to identify Tefhs. Contour plots show representative examples of B-Myd88−/− (upper panels) and DC-Myd88Δ (lower panels) mice with their respective controls. Values indicate percentage of CD4+ T cells (mean).

(D—G) n = 12 for B-Myd88−/− mice and n = 10—12 for controls; n = 5—8 for Jh MRL.Faslpr mice and n = 5—6 for controls; n ≥ 18 for DC-Myd88Δ mice and n ≥ 25 for controls.

Error bars show SEM. See also Fig. S4.

MyD88 signals in B cells and DCs promote Th1 and T extrafollicular helper (Tefh) cell differentiation

To test whether MyD88 in B cells or DCs is important in controlling CD4+ T cell differentiation fate we determined frequencies of relevant CD4+ T cell subsets. In B-Myd88−/− mice a smaller fraction of CD4+ T cells had a Treg phenotype than in controls (Figure 5E and S4B). Interestingly, in B-Myd88Δ mice, in which Myd88 deletion in B cell is incomplete, we found the opposite, a slight increase in Tregs among CD4+ T cells (Figure S4B). A notable difference between B-Myd88Δ and B-Myd88−/− mice is the lower disease activity in the latter. Tregs increase in number and frequency among CD4+ T cells over the disease course in MRL.Faslpr mice (Divekar et al., 2011), suggesting that MyD88 signaling in B cells, by driving inflammation, promotes peripherally induced Treg formation. Yet, other effects of MyD88 in B cells evidently inhibit Treg differentiation. In lupus prone and non-autoimmune mice loss of DCs leads to a loss of Tregs (Darrasse-Jèze et al., 2009; Teichmann et al., 2010). In contrast, in DC-Myd88Δ mice the frequency of Tregs among CD4+ T cells was higher than in controls (Figure 5E and S4B). This is consistent with our finding that DCs from these mice produce lower amounts of inflammatory cytokines and more TGFβ1, which induces FoxP3 in CD4+ T cells (Figure 4D).

Recently, a new CD4+ T cell population, called Tefhs, has been described that promotes B cell responses at extrafollicular sites, and which is prominent in MRL.Faslpr mice (Odegard et al., 2008). Like follicular helper T cells that reside in GCs, Tefhs depend on Bcl6, secrete IL-21 and downregulate P-selectin glycoprotein ligand 1 (PSGL-1). Unlike follicular B helper T cells, which express CXCR5, Tefhs express CXCR4, corresponding to their extrafollicular localization. The frequency of CD4+ T cells with a Tefh phenotype was decreased in both B-Myd88−/− and DC-Myd88Δ mice compared to their respective controls (Figure 5G and S4C). Thus, in addition to initiating AFC differentiation directly in B cells, B cell-intrinsic MyD88 signals also indirectly enhances the AFC response by supporting Tefhs. Further, the finding that DC-intrinsic MyD88 signaling leads to more Tefhs provides a mechanistic explanation for how DC-MyD88 promotes auto-Ab production.

Finally, we evaluated effects of MyD88 expression in B cells and DCs on T cell IFN-γ production. In human SLE, peripheral blood T cells produce excessive amounts of IFN-γ (Harigai et al., 2008). In MRL.Faslpr mice deletion of Ifng or Ifngr1 protects from disease (Balomenos et al., 1998; Schwarting et al., 1998). In B-Myd88−/− and DC-Myd88Δ mice the fraction of CD4+ T cells that produced IFN-γ upon phorbol myristate acetate (PMA) and ionomycin stimulation was lower than in controls (Figure 5F and S4D). The reduction in IFN-γ+ cells among CD4+ T cells in mice lacking MyD88 in B cells was small (31.4% in B-Myd88−/− mice vs. 38.1% in controls). B cell deficient Jh MRL.Faslpr mice, however, showed a marked decrease in CD4+ T cells that produce IFN-γ (42.2% in Jh MRL.Faslpr mice vs. 13.7% in controls), likely the consequence of substantially reduced CD4+ T cell activation in these mice. Differentiation of CD8+ T cells into IFN-γ+ effectors was marginally dependent on MyD88 in DCs but not B cells (Figure 5F and S4D), supporting the notion that B cells are not involved in priming CD8+ T cell responses in lupus. Notably, IFN-γ production in CD4+ and CD8+ T cells is driven by IL-12, whose expression was decreased in DCs isolated from both B-Myd88−/− and DC-Myd88Δ mice (Figure 4D). Taken together, MyD88 signals in B cells and DCs facilitated differentiation of T cell types that promote auto-Ab formation and inflammation.

Lupus nephritis is controlled by MyD88 in B cells but not DCs

To establish the impact of tissue-specific Myd88 deletion on clinical disease, we scored dermatitis in DC-Myd88Δ and B-Myd88Δ mice and their littermate controls. Because MRL.Faslpr females develop more severe dermatitis than males, we restricted our analysis to females. Dermatitis severity in B-Myd88Δ mice was indistinguishable from controls (Figure 6A). However, because of the incomplete deletion of Myd88 in B-Myd88Δ mice we cannot rule out that B cell-intrinsic MyD88 signaling promotes skin manifestations of lupus. Irradiation blocks dermatitis and hence we could not study B-Myd88−/− mice for this phenotype. Unlike in B-Myd88Δ mice, dermatitis in DC-Myd88Δ mice was greatly ameliorated compared to controls (Figure 6B). To evaluate renal disease we used semi-quantitative pathological scoring of severity of glomerular (GN) and interstitial nephritis (IN). Strikingly, B-Myd88Δ mice had markedly reduced GN (Figure 6C and 6E). For IN we observed a downtrend (Figure 6C), which did not quite reach significance (p = 0.08). For a quantitative assessment of IN we measured the area of all interstitial infiltrates relative to the total kidney section area. This demonstrated clearly that IN was reduced in B-Myd88Δ mice compared to controls (Figure 6D, 6F). Hence, T cell infiltrates in kidneys depend on B cell-expressed MyD88. In contrast, deletion of Myd88 in DCs had no effect on GN or IN (Figure 6G). Proteinuria, a functional measure, confirmed the histopathological findings: B-Myd88Δ mice had less proteinuria than control mice (Figure 6H), whereas DC-Myd88Δ mice showed no reduction (Figure 6I).

Figure 6. Deletion of Myd88 in B cells but not DCs protects MRL Faslpr mice from lupus nephritis.

Figure 6

Open in a new tab

(A and B) Dermatitis scores of (A) control and B-Myd88Δ mice, or (B) control and DC-Myd88Δ mice.

(C) Glomerular nephritis was scored from 0 to 6 and interstitial nephritis from 0 to 4. Shown are scores of control and B-Myd88Δ mice.

(D) The area of interstitial infiltrates expressed as a percentage of the total kidney section area is plotted for control (n = 24) and B-Myd88Δ (n = 17) mice.

(E) Representative images of H&E stained kidneys sections (scale bars = 100 μm) illustrating glomerular nephritis (arrows) in kidneys from control and B-Myd88Δ mice.

(F) Representative low magnification images of H&E stained kidneys sections (scale bar = 2 mm) illustrating perivascular infiltrates in kidneys from control and B-Myd88Δ mice.

(G) As in (C) but for control and DC-Myd88Δ mice.

(H and I) Proteinuria scores of (H) control and B-Myd88Δ mice, or (I) control and DC-Myd88Δ mice.

Error bars show SEM.

In conclusion, activation of B cells through MyD88 is essential for the development of lupus nephritis. Although deletion of DCs ameliorates kidney inflammation, deletion of Myd88 in DCs does not. This implies that signals that are not transduced by MyD88 are sufficient to induce DCs functions critical for nephritis. Interestingly, skin pathogenesis has different requirements, being dependent on DCs (Teichmann et al., 2010) and in particular MyD88 signaling in DCs.

Discussion

Our study provides detailed insight into how MyD88 signaling in B cells and DCs contribute to lupus pathogenesis and suggests a refined model of functional interactions (Figure S5—see legend for details). We formally established that ANA formation is entirely dependent on MyD88 in B cells. Deletion of Myd88 selectively in B cells ameliorated nephritis whereas DC-specific ablation of Myd88 did not. These findings emphasize the importance of B cells with specificity for nucleic acids or associated proteins in activating pathogenic T cells. That nephritis was independent of DC-expressed MyD88 was unexpected, because we previously showed that DCs per se are critical for nephritis development (Teichmann et al., 2010). Notably, dermatitis severity was impaired in the absence of MyD88 in DCs, revealing a pathogenetic mechanism distinct from that of kidney disease.

The complete loss of ANAs in MRL.Faslpr mice globally deficient for MyD88 (Nickerson et al., 2010) was fully recapitulated in MRL.Faslpr mice lacking MyD88 specifically in B cells. DC-expressed MyD88 had only minor effects on auto-Ab formation. In view of a recent study (Hou et al., 2011), which investigated the cell type specific role of MyD88 for different vaccination protocols, these results were surprising. In that work, the cell-specific requirements for MyD88 expression varied depending on the type of Ag. Responses to OVA-CpG were enhanced by MyD88 signaling in DCs but not in B cells. Other responses depended on B cell-expressed MyD88, but in all cases a role for MyD88 in other cell types was implied as decreases in serum concentrations of Ag-specific Abs in globally MyD88-deficient mice were more pronounced than those in mice selectively lacking Myd88 in B cells. Our findings thus contrast with those obtained from immunization of non-autoimmune mice and underscore the extraordinary dependence of ANAs on B-MyD88 in lupus. Notably, other possible modes of B cell activation, such as BCR and CD40 co-ligation, did not compensate for the loss of MyD88 in B cells. Consequently, activated auto-reactive T cells on their own lack the ability to break tolerance in B cells and induce AFC differentiation. Interestingly, B-Myd88−/− mice also had a >98% reduction in RF AFC numbers, indicating that RF B cells are almost exclusively driven by ICs containing nucleic acids.

Remarkably, ~99% of total splenic AFCs were generated as a result of B cell-intrinsic MyD88 signaling. Because AFCs are hardly detectable in the spleens of non-autoimmune mice, infectious agents are unlikely major stimuli for splenic AFC differentiation. Thus, the great majority of splenic AFCs in MRL.Faslpr mice is probably self-reactive.

Despite incomplete Myd88 deletion in B-Myd88Δ mice, GN and IN were both considerably ameliorated compared to controls. The development of GN is thought to depend on IC deposition, although some GN is observed in MRL.Faslpr mice lacking secreted Ig and thus ICs (Chan et al., 1999). Nonetheless, the reduction in auto-Abs in B-Myd88Δ mice might result in less IC deposition and GN. However, establishment of IN is not contingent on Ig (Chan et al., 1999) arguing that lower auto-Abs amounts do not explain attenuated renal interstitial infiltrates in B-Myd88Δ mice. Because IN is strongly dependent on CD4+ T cells (Jabs et al., 1992; Jevnikar et al., 1994), our results imply that activation of those CD4+ T cells that are essential for renal pathology is controlled by MyD88 in B cells. The simplest interpretation is that B cells activated via MyD88 present Ag to and activate nephrotoxic CD4+ T cells. That antigen presentation by B cells can lead to activation of naïve CD4+ T cells in-vivo has been directly documented (Rodríguez-Pinto and Moreno, 2005). Spontaneous CD4+ T cell activation was indeed impaired in B-Myd88−/− mice, although to a lesser extent than in B cell deficient Jh MRL.Faslpr mice suggesting MyD88-dependent and -independent Ag-presenting functions of B cells. Presumably, CD4+ T cells that are activated in a B-MyD88-dependent manner recognize peptides from nucleic acid-associated proteins. We therefore hypothesize that CD4+ T cells with these specificities can cause or at least indirectly promote tissue inflammation and damage. Defining the precise specificities of disease-relevant T cells will require the generation of T cell receptor transgenic mice for genuine lupus auto-Ags. Interestingly, B cell-specific deficiency for IL-6 attenuated experimental autoimmune encephalomyelitis, which was associated with an impaired Th17 cell response (Barr et al., 2012). A further explanation for reduced IN in B-Myd88−/− mice could thus be a lack of B cell-derived inflammatory cytokines.

We tested whether the ability of splenic cDCs to produce cytokines that are important for priming of T helper cell responses depends on MyD88 in DCs, B cells, or both. As expected from immunization studies (Joffre et al., 2009) MyD88 deficiency in cDCs led to substantial reductions in p35, p40 as well as Il6 mRNA amounts. Stimulation of the TLR-MyD88 axis in cDCs might be facilitated by endocytosis of nucleic acid containing ICs mediated by Fcγ receptors, or alternatively by uptake of nucleic acids bound to HMGB1 or LL37. Remarkably, decreases in inflammatory cytokine transcripts in cDCs in B-Myd88Δ mice were comparable to those in DC-Myd88Δ mice. The strong reliance of inflammatory cytokine production by cDCs on MyD88 in B cells suggests that it is mainly ICs that drive full cDC activation in lupus. Thus, our experiments provide evidence for the concept of DC activation by nucleic acid containing ICs in vivo, a mechanism that has only been shown in-vitro so far (Marshak-Rothstein, 2006).

DC-expressed MyD88 regulated several aspects of B cell biology: B cell lymphopoiesis, B cell positioning in the spleen, IgG isotype subclass distribution and to a lesser degree scale of IgG auto-Ab production. MyD88 signaling in DCs promoted B cell lymphopenia by impeding B cell lymphopoiesis. In B progenitors IFN-I induce DAXX, a transcription factor that represses IL-7-dependent growth and survival (Gongora et al., 2001). Our findings link IFN-I production by bone marrow pDCs, resulting from cell-intrinsic MyD88 signaling, to DAXX-dependent suppression of B lymphopoiesis in lupus. These data are consistent with the observation that a strong IFN-I signature in SLE patients is associated with more severe lymphopenia (Kirou et al., 2005) considered a risk factor for major infections (Ng et al., 2006). Numerous in-vitro studies have demonstrated the ability of human pDCs to secrete IFN-I in response to ICs containing TLR ligands (Rönnblom et al., 2011). However, in-vivo data on TLR-dependent IFN-I production by pDCs and its consequences in lupus have been missing so far.

Interestingly, deletion of MyD88 in DCs led to the preservation of B cell follicles and the development of GC-like structures in the spleens of MRL.Faslpr mice, both of which are typically effaced as these lupus-prone mice age. Infection with lymphocytic choriomeningitis virus disrupts normal splenic architecture subsequent to transient suppression of homeostatic chemokines by IFN-γ (Mueller et al., 2007). IFN-γ hyperproduction is found in both MRL.Faslpr mice and SLE patients (Harigai et al., 2008; Teichmann et al., 2010). Notably, DC-Myd88Δ mice harbor fewer IFN-γ secreting CD4+ and CD8+ T cells, suggesting that reduced IFN-γ induction by MyD88-deficient DCs could cause the preserved splenic architecture. The relationship between DC-intrinsic MyD88 signaling, IFN-γ and spleen architecture and function warrants further investigation in both mice and humans.

We have previously demonstrated that DCs (Teichmann et al., 2010) have nonredundant roles in dermatitis and nephritis pathogenesis. Unexpectedly, ablation of Myd88 in DCs attenuated dermatitis but not nephritis, revealing separate pathogenetic mechanisms. Unlike the kidneys, which are essentially sterile organs, the skin is physiologically colonized by a microflora and frequently exposed to pathogens. It is therefore possible that skin DC subsets partly differ in function compared to those in the kidneys. Alternatively, MyD88-dependent microbial sensing might contribute to dermatitis in MRL.Faslpr mice, although antibiotic treatment with co-trimoxazole or enrofloxacin does not prevent its development (data not shown). Generally, how DCs cause tissue damage is poorly understood and this question merits more study.

In summary, we have delineated the unique immunopathological effects of B cells and DCs that are directly activated by nucleic acid recognition through the TLR-MyD88 pathway. Based on these, we propose a hierarchy of events (Figure S5). Understanding the means by which discrete cell populations in complex multicellular networks become activated and the functional consequences of different modes of activation will likely enable the design of novel drugs for autoimmune disorders, for example by targeting DCs or DC subsets.

Material and methods

Mice

Myd88fl/+ C57BL/6 (Kleinridders et al., 2009) and CD11c-Cre BAC transgenic C57BL/6 (Caton et al., 2007) were backcrossed to MRL-MpJ-Faslpr/J mice (Jackson Laboratory) 10 and 17 times, respectively. CD19-Cre C57BL/6 (Rickert et al., 1997) were crossed 10 times to MRL-MpJ-Faslpr/2J and then 5 times to MRL-MpJ-Faslpr/J mice. Experimental mice were generated by interbreeding CD11c-CreMyd88fl/fl or CD19-Cre Myd88fl/fl with Myd88fl/fl MRL.Faslpr mice. Littermates without a Cre allele were used as controls. For mixed BM chimeras 6–7 wk old B cell-deficient Jh (Igh-Jtm1Dhu) MRL.Faslpr mice (Chan and Shlomchik, 1998) were irradiated with 800 cGy and injected intravenously with 8 × 106 Jh MRL.Faslpr together with 2 × 106 Myd88−/− MRL.Faslpr (Nickerson et al., 2010) bone marrow cells. Controls were generated in the same experiments using 2 × 106 Myd88+/+ instead of Myd88−/− MRL.Faslpr bone marrow cells. Cre-loxP mice were analyzed at 16 wks of age and BM chimeras 21 wks after transplantation, unless otherwise stated. All animals were maintained under specific pathogen-free conditions and handled according to protocols approved by the Yale Institutional Animal Care and Use Committee.

Flow cytometry

Surface staining was performed in ice-cold PBS with 3% calf serum in the presence of FcR blocking Ab 2.4G2. Ab clones used for FACS can be found in the Supplemental Information. Intracellular staining was performed using the BD Cytofix/Cytoperm and Perm/Wash buffers or, for intracellular FoxP3 staining, the eBioscience FoxP3 staining buffer set. For intracellular IFN-γ staining, 4 × 106 splenocytes were cultured for 4 hr at 37°C in 24-well plates in 2 ml culture medium containing ionomycin (750 ng/ml) and PMA (20 ng/ml). For the last 2 hrs brefeldin A (10 μg/ml) was added to the cultures. Live and dead cells were distinguished with ethidium monoazide (EMA), and gating strategies were used for doublet discrimination. Apoptotic cells were identified by staining for active caspases with FITC-VAD-FMK (SM Biochemicals). Cells were analyzed on a LSRII instrument (BD). For purification of populations, cells were sorted on a FACSAria II (BD).

qRT-PCR

To quantify deletion efficacy of Myd88, genomic DNA was extracted from FACS purified cells and used as template for qRT-PCR. The amount of Myd88 in each sample was normalized to the unaffected gene Il10. To measure gene transcript abundance of APRIL, BAFF, DAXX, pan-Ifna, Il1b, Il6, Il10, p35, p40 and Tgfb1 mRNA was isolated from FACS purified cells with the Qiagen RNeasy Plus Mini Kit and reverse transcribed with the Bio-Rad iScript cDNA Synthesis Kit. Gapdh was used as a housekeeping gene. Primer sequences are listed in the Supplemental Information. qRT-PCR was performed with the Agilent Brilliant II SYBR Green QPCR kit on a Stratagene Mx3000P instrument.

ELISpots

To detect AFCs that produce κ light chain Abs, IgG1, IgG2a or IgM 96-well Immulon 4 HBX plates were coated overnight at 4 °C with 5 μg/ml polyclonal goat-anti mouse κ (Southern Biotech; 1050-01). Nonspecific binding was blocked with 1% bovine serum albumin in PBS and samples were incubated at 37 °C. Alkaline phosphatase–conjugated secondary Abs (Southern Biotech; to κ (1050-04), IgG1 (1070-04), IgG2a (1080-04) or IgM (1020-04)) were detected with bromo-4-chloro-3-indolyl phosphate substrate (Southern Biotech).

For anti-IgG1 and anti-IgG2a RF producing AFCs, plates were coated with 5 μg/ml C18 (anti-(4-hydroxy-3-nitrophenyl)acetyl (NP) IgG1 λ) and 23.3 (anti-NP IgG2a λ), respectively. Biotinylated anti-κ (187.1) was used as secondary Ab, followed by alkaline phosphatase–conjugated streptavidin.

ELISA and Luminex

Serum anti-nucleosome IgG, IgG1 and IgG2a concentrations were determined by ELISA as previously described (Nickerson et al., 2010). The nuclesome-specific Ab clone PL2-3 was used as standard. Serum Ig isotype concentrations were measured by Luminex assay (Millipore) according to the manufacturer's protocol.

ANAs

HEp-2 immunofluorescence assays (Antibodies Inc.) were performed as previously described (Christensen et al., 2005) with serum diluted at 1:100. Images were captured on a Nikon Eclipse Ti-U microscope.

Evaluation of clinical disease

Dermatitis, nephritis and proteinuria were scored as described previously (Teichmann et al., 2010). Area of interstitial infiltrates relative to the total kidney section area was calculated with ImageJ software.

Statistical analysis and data display

Statistics were calculated by two-tailed Mann-Whitney U test with p values indicated throughout as *p < 0.05, **p < 0.01, ***p < 0.001. In scatter plots each dot represents an individual mouse and horizontal lines represent the median.

Supplementary Material

01

Highlights

  • B cell-intrinsic MyD88 signaling is absolutely required for ANA formation

  • Nephrotoxic T cells are controlled by B cell- but not DC-expressed MyD88

  • Dermatitis development is strongly dependent on MyD88 in DCs

  • DC-intrinsic MyD88 signals impair B cell lymphoiesis leading to B cell lymphopenia

Acknowledgements

We thank Yale Animal Resources Center for outstanding animal husbandry. Supported by National Institutes of Health grant R01-AR044077 and P01-AR050256 (M.J.S.), Deutsche Forschungsgemeinschaft (fellowship to L.L.T.), Howard Hughes Medical Institute (R.M.) and Cancer Research Institute (fellowship to D.S.). We thank Kevin Nickerson, Allison Campbell, and Ann Marshak-Rothstein for critical reading of the manuscript.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Balomenos D, Rumold R, Theofilopoulos AN. Interferon-gamma is required for lupus-like disease and lymphoaccumulation in MRL-lpr mice. J Clin Invest. 1998;101:364–371. doi: 10.1172/JCI750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Barr TA, Shen P, Brown S, Lampropoulou V, Roch T, Lawrie S, Fan B, O'Connor RA, Anderton SM, et al. B cell depletion therapy ameliorates autoimmune disease through ablation of IL-6-producing B cells. J Exp Med. 2012;209:1001–1010. doi: 10.1084/jem.20111675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bishop GA, Hostager BS. The CD40-CD154 interaction in B cell-T cell liaisons. Cytokine Growth Factor Rev. 2003;14:297–309. doi: 10.1016/s1359-6101(03)00024-8. [DOI] [PubMed] [Google Scholar]
  4. Caton ML, Smith-Raska MR, Reizis B. Notch-RBP-J signaling controls the homeostasis of CD8- dendritic cells in the spleen. J Exp Med. 2007;204:1653–1664. doi: 10.1084/jem.20062648. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chan O, Shlomchik MJ. A new role for B cells in systemic autoimmunity: B cells promote spontaneous T cell activation in MRL-lpr/lpr mice. J Immunol. 1998;160:51–59. [PubMed] [Google Scholar]
  6. Chan OT, Hannum LG, Haberman AM, Madaio MP, Shlomchik MJ. A novel mouse with B cells but lacking serum antibody reveals an antibody-independent role for B cells in murine lupus. J Exp Med. 1999;189:1639–1648. doi: 10.1084/jem.189.10.1639. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Christensen SR, Kashgarian M, Alexopoulou L, Flavell RA, Akira S, Shlomchik MJ. Toll-like receptor 9 controls anti-DNA autoantibody production in murine lupus. J Exp Med. 2005;202:321–331. doi: 10.1084/jem.20050338. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Christensen SR, Shupe J, Nickerson K, Kashgarian M, Flavell RA, Shlomchik MJ. Toll-like receptor 7 and TLR9 dictate autoantibody specificity and have opposing inflammatory and regulatory roles in a murine model of lupus. Immunity. 2006;25:417–428. doi: 10.1016/j.immuni.2006.07.013. [DOI] [PubMed] [Google Scholar]
  9. Darrasse-Jèze G, Deroubaix S, Mouquet H, Victora GD, Eisenreich T, Yao KH, Masilamani RF, Dustin ML, Rudensky A, et al. Feedback control of regulatory T cell homeostasis by dendritic cells in vivo. J Exp Med. 2009;206:1853–1862. doi: 10.1084/jem.20090746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Divekar AA, Dubey S, Gangalum PR, Singh RR. Dicer insufficiency and microRNA-155 overexpression in lupus regulatory T cells: an apparent paradox in the setting of an inflammatory milieu. J Immunol. 2011;186:924–930. doi: 10.4049/jimmunol.1002218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gongora R, Stephan RP, Zhang Z, Cooper MD. An essential role for Daxx in the inhibition of B lymphopoiesis by type I interferons. Immunity. 2001;14:727–737. doi: 10.1016/s1074-7613(01)00152-2. [DOI] [PubMed] [Google Scholar]
  12. Hardy RR, Kincade PW, Dorshkind K. The protean nature of cells in the B lymphocyte lineage. Immunity. 2007;26:703–714. doi: 10.1016/j.immuni.2007.05.013. [DOI] [PubMed] [Google Scholar]
  13. Harigai M, Kawamoto M, Hara M, Kubota T, Kamatani N, Miyasaka N. Excessive production of IFN-gamma in patients with systemic lupus erythematosus and its contribution to induction of B lymphocyte stimulator/B cell-activating factor/TNF ligand superfamily-13B. J Immunol. 2008;181:2211–2219. doi: 10.4049/jimmunol.181.3.2211. [DOI] [PubMed] [Google Scholar]
  14. Hou B, Saudan P, Ott G, Wheeler ML, Ji M, Kuzmich L, Lee LM, Coffman RL, Bachmann MF, DeFranco AL. Selective utilization of Toll-like receptor and MyD88 signaling in B cells for enhancement of the antiviral germinal center response. Immunity. 2011;34:375–384. doi: 10.1016/j.immuni.2011.01.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Jabs DA, Burek CL, Hu Q, Kuppers RC, Lee B, Prendergast RA. Anti-CD4 monoclonal antibody therapy suppresses autoimmune disease in MRL/Mp-lpr/lpr mice. Cell Immunol. 1992;141:496–507. doi: 10.1016/0008-8749(92)90166-m. [DOI] [PubMed] [Google Scholar]
  16. Jevnikar AM, Grusby MJ, Glimcher LH. Prevention of nephritis in major histocompatibility complex class II-deficient MRL-lpr mice. Journal of Experimental Medicine. 1994;179:1137. doi: 10.1084/jem.179.4.1137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Joffre O, Nolte MA, Spörri R, Reis e Sousa C. Inflammatory signals in dendritic cell activation and the induction of adaptive immunity. Immunol Rev. 2009;227:234–247. doi: 10.1111/j.1600-065X.2008.00718.x. [DOI] [PubMed] [Google Scholar]
  18. Kirou KA, Lee C, George S, Louca K, Peterson MG, Crow MK. Activation of the interferon-alpha pathway identifies a subgroup of systemic lupus erythematosus patients with distinct serologic features and active disease. Arthritis Rheum. 2005;52:1491–1503. doi: 10.1002/art.21031. [DOI] [PubMed] [Google Scholar]
  19. Kleinridders A, Schenten D, Könner AC, Belgardt BF, Mauer J, Okamura T, Wunderlich FT, Medzhitov R, Brüning JC. MyD88 signaling in the CNS is required for development of fatty acid-induced leptin resistance and diet-induced obesity. Cell Metab. 2009;10:249–259. doi: 10.1016/j.cmet.2009.08.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kool M, van Loo G, Waelput W, De Prijck S, Muskens F, Sze M, van Praet J, Branco-Madeira F, Janssens S, et al. The ubiquitin-editing protein A20 prevents dendritic cell activation, recognition of apoptotic cells, and systemic autoimmunity. Immunity. 2011;35:82–96. doi: 10.1016/j.immuni.2011.05.013. [DOI] [PubMed] [Google Scholar]
  21. Lande R, Gregorio J, Facchinetti V, Chatterjee B, Wang YH, Homey B, Cao W, Wang YH, Su B, et al. Plasmacytoid dendritic cells sense self-DNA coupled with antimicrobial peptide. Nature. 2007;449:564–569. doi: 10.1038/nature06116. [DOI] [PubMed] [Google Scholar]
  22. Lövgren T, Eloranta ML, Kastner B, Wahren-Herlenius M, Alm GV, Rönnblom L. Induction of interferon-alpha by immune complexes or liposomes containing systemic lupus erythematosus autoantigen- and Sjögren's syndrome autoantigen-associated RNA. Arthritis Rheum. 2006;54:1917–1927. doi: 10.1002/art.21893. [DOI] [PubMed] [Google Scholar]
  23. Marshak-Rothstein A. Toll-like receptors in systemic autoimmune disease. Nat Rev Immunol. 2006;6:823–835. doi: 10.1038/nri1957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Mueller SN, Hosiawa-Meagher KA, Konieczny BT, Sullivan BM, Bachmann MF, Locksley RM, Ahmed R, Matloubian M. Regulation of Homeostatic Chemokine Expression and Cell Trafficking During Immune Responses. Science. 2007;317:670–674. doi: 10.1126/science.1144830. [DOI] [PubMed] [Google Scholar]
  25. Ng WL, Chu CM, Wu AK, Cheng VC, Yuen KY. Lymphopenia at presentation is associated with increased risk of infections in patients with systemic lupus erythematosus. QJM. 2006;99:37–47. doi: 10.1093/qjmed/hci155. [DOI] [PubMed] [Google Scholar]
  26. Nickerson KM, Christensen SR, Shupe J, Kashgarian M, Kim D, Elkon K, Shlomchik MJ. TLR9 Regulates TLR7- and MyD88-Dependent Autoantibody Production and Disease in a Murine Model of Lupus. J Immunol. 2010;184:1840–1848. doi: 10.4049/jimmunol.0902592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Odegard JM, Marks BR, DiPlacido LD, Poholek AC, Kono DH, Dong C, Flavell RA, Craft J. ICOS-dependent extrafollicular helper T cells elicit IgG production via IL-21 in systemic autoimmunity. J Exp Med. 2008;205:2873–2886. doi: 10.1084/jem.20080840. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Rickert RC, Jellusova J, Miletic AV. Signaling by the tumor necrosis factor receptor superfamily in B-cell biology and disease. Immunol Rev. 2011;244:115–133. doi: 10.1111/j.1600-065X.2011.01067.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Rickert RC, Roes J, Rajewsky K. B lymphocyte-specific, Cre-mediated mutagenesis in mice. Nucleic Acids Res. 1997;25:1317–1318. doi: 10.1093/nar/25.6.1317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Rodríguez-Pinto D, Moreno J. B cells can prime naive CD4+ T cells in vivo in the absence of other professional antigen-presenting cells in a CD154-CD40-dependent manner. Eur J Immunol. 2005;35:1097–1105. doi: 10.1002/eji.200425732. [DOI] [PubMed] [Google Scholar]
  31. Rönnblom L, Alm GV, Eloranta ML. The type I interferon system in the development of lupus. Semin Immunol. 2011;23:113–121. doi: 10.1016/j.smim.2011.01.009. [DOI] [PubMed] [Google Scholar]
  32. Schwarting A, Wada T, Kinoshita K, Tesch G, Kelley VR. IFN-gamma receptor signaling is essential for the initiation, acceleration, and destruction of autoimmune kidney disease in MRL-Fas(lpr) mice. J Immunol. 1998;161:494–503. [PubMed] [Google Scholar]
  33. Shlomchik MJ. Sites and stages of autoreactive B cell activation and regulation. Immunity. 2008;28:18–28. doi: 10.1016/j.immuni.2007.12.004. [DOI] [PubMed] [Google Scholar]
  34. Sweet RA, Ols ML, Cullen JL, Milam AV, Yagita H, Shlomchik MJ. Facultative role for T cells in extrafollicular Toll-like receptor-dependent autoreactive B-cell responses in vivo. Proc Natl Acad Sci U S A. 2011;108:7932–7937. doi: 10.1073/pnas.1018571108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Teichmann LL, Ols ML, Kashgarian M, Reizis B, Kaplan DH, Shlomchik MJ. Dendritic cells in lupus are not required for activation of T and B cells but promote their expansion, resulting in tissue damage. Immunity. 2010;33:967–978. doi: 10.1016/j.immuni.2010.11.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Tian J, Avalos AM, Mao SY, Chen B, Senthil K, Wu H, Parroche P, Drabic S, Golenbock D, et al. Toll-like receptor 9-dependent activation by DNA-containing immune complexes is mediated by HMGB1 and RAGE. Nat Immunol. 2007;8:487–496. doi: 10.1038/ni1457. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS456340-supplement-01.doc (11.9MB, doc)

RESOURCES