Lack of T cells in Act1-deficient mice results in elevated IgM-specific autoantibodies but reduced lupus-like disease

Authors

  • Angela C. Johnson,

  • Laura M. Davison,

  • Natalia V. Giltiay,

    Current affiliation:
    1. Department of Immunology, University of Washington, Seattle, WA
    Search for more papers by this author
  • Chairut Vareechon,

  • Xiaoxia Li,

  • Trine N. Jørgensen

    Corresponding author
    • Department of Immunology NE40, Lerner Research Institute, Cleveland Clinic Foundation, Cleveland, OH, USA
    Search for more papers by this author

Full correspondence: Dr. Trine N. Jorgensen, Department of Immunology NE40, Lerner Research Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA.

Fax: +1-216-444-9329

e-mail: jorgent@ccf.org

Abstract

Act1 is a negative regulator of B-cell activation factor of the TNF family (BAFF) and CD40L-induced signaling. BALB/C mice lacking Act1 develop systemic autoimmunity resembling systemic lupus erythematosus (SLE) and Sjögren's syndrome (SjS). SLE and SjS are characterized by anti-nuclear IgG autoantibody (ANA-IgG) production and inflammation of peripheral tissues. As autoantibody production can occur in a T-cell dependent or T-cell independent manner, we investigated the role of T-cell help during Act1-mediated autoimmunity. Act1-deficiency was bred onto C57Bl/6 (B6.Act1−/−) mice and B6.TCRβ−/−TCRδ−/−Act1−/− (TKO) mice were generated. While TCRβ/δ-sufficient B6.Act1−/− mice developed splenomegaly and lymphadenopathy, hypergammaglobulinemia, elevated levels of ANA-IgG, and kidney pathology, TKO mice failed to develop any such signs of disease. Neither B6.Act1−/− nor TKO mice developed SjS-like disease, suggesting that epigenetic interactions on the BALB/C background are responsible for this phenotype in BALB/C.Act1−/− mice. Interestingly, BAFF-driven transitional B-cell abnormalities, previously reported in BALB/C.Act1−/− mice, were intact in B6.Act1−/− mice and largely independent of T cells. In conclusion, T cells are necessary for the development of SLE-like disease in B6.Act1−/− mice, but not BAFF-driven transitional B-cell differentiation.

Introduction

Act1 (Traf3ip2, Ciks) is a negative regulator of CD40 and B-cell activation factor of the TNF family (BAFF)-receptor mediated signaling [1]. As such, B cells from Act1-deficient BALB/C mice and from B-cell-specific Act1-deficient mice (CD19-CRE+/−Act1−/fl) display increased survival in response to anti-CD40 antibody or BAFF (also known as Blys, THANK) treatment [1, 2]. BALB/C.Act1−/− mice develop signs of systemic autoimmunity including splenomegaly, lymphadenopathy and elevated serum anti-nuclear autoantibodies (ANA) starting as early as 3–4 weeks of age [1]. Likewise, both BAFF and CD40L transgenic mice have been shown to develop autoimmunity characterized by spontaneous B-cell activation and autoantibody production [3-5].

BAFF signaling is essential for B-cell maturation and survival as the immature T1 cells differentiate to transitional T2 and T3 B cells in the spleen (reviewed in [6]). In addition, it has been speculated that BAFF may function to maintain the mature pool of B cells [7]. The control of transitional B-cell differentiation is key to the elimination of potentially autoreactive B cells, and deficiency of Act1 in B cells lowers the threshold for B-cell elimination resulting in increased numbers of circulating mature autoreactive B cells [1, 2, 8]. Despite this, previous studies found that some autoantibody production is still measurable in Act1−/−BAFF−/− mice [1], suggesting that among the few B cells that effectively develop in the absence of BAFF signaling a population of autoreactive B cells remain.

Act1 is recruited to the CD40 receptor upon binding of CD40 ligand (CD40L, CD154) [1, 9]. CD40L-induced signaling results in B-cell differentiation, the initiation of germinal center (GC) formation, immunoglobulin (Ig) class switching and generation of memory B-cell responses [10-12], all of which are elevated in Act1-deficient mice. Thus Act1 is a negative regulator of CD40 intracellular signaling [1]. The main source of CD40L is activated T cells, however GC formation as well as autoantibody production have been found in T-cell-deficient mice [13, 14]. T-cell-independent GC formation and Ig class switching was also observed in mice overexpressing BAFF (BAFF-Tg) [15]. The exact mechanism for this phenomenon is not completely resolved, but several studies have pointed to a role for toll-like receptor (TLR)-signaling and/or BAFF itself [16-19]. Interestingly, autoantibody production in BAFF-Tg mice has been shown to rely on functional IL-1R/TLR signaling, but not T cells, as MyD88-deficient BM cells failed to support accelerated B-cell differentiation while TCR-deficient BAFF-Tg mice produced ANA equivalent to TCR-sufficient BAFF-Tg mice [17]. More recent data obtained from lupus-prone NZB mice support a role for both BAFF and T cells during B-cell development, separating the effect of B-cell survival (BAFF) from B-cell differentiation and antibody production (T cells) [20].

In the current study we investigated the role of T cells in Act1-deficient mice. In contrast to observations seen in BAFF-transgenic mice [17], we found that IgG-mediated systemic autoimmunity in B6.Act1−/− mice, despite showing BAFF-driven abnormalities among B-cell populations, is dependent on T cells.

Results

Act1-induced splenomegaly and lymphadenopathy develop only in the presence of T cells

Act1 is a negative regulator of B-cell activation and different-iation through its interaction with the intracellular signaling cascades triggered by CD40L and BAFF binding to their respective receptors (CD40, BAFF-R, TACI, or BCMA) [1, 2]. Deficiency of Act1 in BALB/C mice results in systemic autoimmunity characterized by the development of splenomegaly, lymphadenopathy, and elevated serum autoantibodies [1, 2, 8]. In order to define if T-cell help was required for the development of systemic autoimmunity, we generated αβ and γδ T-cell- and Act1-triple deficient mice (TCRβ/δ−/−Act1−/−; TKO) on the C57Bl/6 (B6) background. The development of splenomegaly and lymphadenopathy was intact in B6.Act1−/− mice, however T-cell deficiency completely abolished this phenotype, as TKO mice exhibited spleen and lymph node sizes and cellular levels equivalent to that of TCRβ/δ−/− and WT (B6) mice (Fig. 1A–B and E–F). As we had expected reduced spleen/LN size and cellularity in TCRβ/δ−/− mice, we further analyzed spleen cells for their relative levels of B- and T cells and found that levels of B cells were significantly elevated, making up the difference in total cellularity between WT and T-cell-deficient mice (Fig. 1C–D). In addition, B6.Act1−/− mice displayed elevated levels of non-B/T cells (manuscript in preparation).

Figure 1.

Lupus-like hypercellularity in B6.Act1−/− mice is dependent on T-cell help. (A–F) C57Bl/6 (WT), B6.Act1−/−, TCRβ/δ−/−, and TKO (TCRβ/δ−/−Act1−/−) mice were generated and analyzed for the development of splenomegaly and lymphadenopathy at 32–40 weeks of age. Data are presented as (A, E) mg tissue as well as (B, F) total cells isolated. (C, D) In addition, total numbers of B and T lymphocytes were calculated from spleen samples. Samples were collected over 12 months and evaluated at the time of harvest. Each symbol represents one individual mouse and the horizontal line signifies the mean. ** p < 0.01, *** p < 0.001; two-tailed Mann–Whitney test. # denotes data showing a trend toward statistical difference: p < 0.1.

Hypergammaglobulinemia and elevated levels of ANA in B6.Act1−/− mice is T-cell dependent

Elevated levels of serum total antibodies and anti-nuclear auto-antibodies (ANAs) are hallmarks of mouse lupus-like disease. We therefore assayed serum from aged (28–32-week old) WT, B6.Act1−/−, TCRβ/δ−/−, and TKO mice for levels of total serum immunoglobulins as well as antigen-specific anti-chromatin, anti-histone and anti-dsDNA IgG, and IgM antibodies. Similarly to BALB/C.Act1−/− mice, B6.Act1−/− mice developed hypergammaglobulinemia and elevated levels of serum ANA (Fig. 2B–G). We saw no difference in serum IgM levels between WT and B6.Act1−/− mice (Fig. 2A). In the absence of T cells, B6.Act1−/− mice developed significantly less total IgG antibodies (IgG, IgG1, and IgG2c, Fig. 2B–D) and anti-nuclear antigen specific IgG autoantibodies (anti-chromatin, anti-histone, and anti-dsDNA IgG autoantibodies) (Fig. 2E–G). In contrast, serum levels of anti-chromatin IgM, anti-histone IgM, and anti-dsDNA IgM were significantly elevated in TKO mice as compared with B6.Act1−/− mice (Fig. 2H–J), suggesting that BAFF-dependent survival and maintenance of (low affinity) self-reactive B cells was intact in these mice (see below). Thus, while T cells are required for the development of IgG-mediated lupus-like abnormalities in B6.Act1−/− mice, IgM-autoantibodies were elevated in a T-cell-independent manner.

Figure 2.

Lupus-like serological abnormalities in B6.Act1−/− mice are dependent on T-cell help. Serum samples were obtained from the indicated 28–32-week-old mice and analyzed for levels of total (A) IgM, (B) IgG, (C) IgG1, and (D) IgG2c as well as (E) ANA: anti-chromatin IgG and (H) IgM, (F) anti-histone IgG and (I) IgM, and (G) anti-dsDNA IgG and (J) IgM. Serum was collected and stored over a 12-month period. All samples were run at the same time to minimize assay-to-assay variation. Each symbol represents one individual mouse and the horizontal line signifies the mean. * p < 0.05, ** p < 0.01, *** p <0.001; two-tailed Mann–Whitney test. # denotes data showing a trend toward statistical difference: p < 0.1.

IgG immune complex deposition within the kidney glomeruli is T-cell dependent

Mouse lupus-like disease is most commonly associated with renal abnormalities such as mesangial cell hyperproliferation, glomerular IgG-immune complex (IgG-IC) deposition, and complement factor C3 fixation [21]. Aged BALB/C.Act1−/− and BAFF-Tg mice have abnormal kidney glomeruli with signs of mesangial proliferation and mononuclear cell infiltrates [8, 17, 22]. Analyses of B6.Act1−/− and TKO kidneys showed moderate hypercellularity of the glomerular mesangium and occasional obstruction of the capillary lumina, while WT mice displayed a largely normal glomerular morphology (Fig. 3A). We were unable to find areas of extensive mononuclear cell infiltrates and signs of tubulointerstitial disease in any of the mice (data not shown).

Figure 3.

T-cell deficient B6.Act1−/− mice are protected from glomerulonephritis and IgG-IC deposition in kidney glomeruli. (A–C) Kidneys were harvested from 32–40-week-old WT, TCRβ/δ−/−, B6.Act1−/−, and TKO mice and either fixed in 10% formalin or quick frozen in OCT as described in Materials and methods. (A) Formalin-fixed samples were embedded in paraffin and 5-μm sections were stained with H&E. Data shown represent the average of four individual mice per genotype. The mesangial cell index is given by number of mesangial cells/glomerulus and normalized to levels in WT mice. (B, C) Quick-frozen samples were sectioned (5 μm) and stained for the presence of IgG (red, (B)) or IgM (red, (C)) and complement factor C3 (green) by immunofluorescence staining. Pictures shown are representative of results from three individual mice per genotype. Two independent sections were analyzed per mouse representing more than 20 glomeruli/genotype. All immunofluorescence stainings were performed at the same time and pictures were taken using identical camera settings to allow for direct comparison. IgG and IgM staining per glomerulus was measured based on the density (red) per glomerulus per mouse and is displayed after normalization to the levels in WT mice (WT = 1.0).

We next tested kidneys from WT, TCRβ/δ−/−, B6.Act1−/−, and TKO mice for immunoglobulin deposition and C3 fixation. B6.Act1−/− mice exhibited significantly elevated IgG deposition within the kidney glomeruli (Fig. 3B, red stain, p < 0.001 as compared with WT), while we were unable to detect increased IgG deposition in kidneys of TCRβ/δ−/− and TKO mice. In contrast, analyses of IgM deposition showed elevated levels in TCRβ/δ−/− and TKO mice (Fig. 3C, both: p < 0.001 as compared with WT). Finally, as BAFF-Tg mice have been found to express elevated levels of deposited IgA, we tested kidneys for the deposition of IgA immune complexes. Neither B6.Act1−/−, DKO, nor TKO mice displayed any signs of elevated IgA staining (Supporting Information Fig. 1). Ig deposition during lupus-like disease is known to fixate complement involved in the development of renal disease. We detected no significant C3 fixation in any of the mouse strains, including B6.Act1−/− (Fig. 3B, C and Supporting Information Fig. 1, green stain), suggesting that the Ig-IC deposited in the kidneys of B6.Act1−/−, TCRβ/δ−/−, and TKO mice (IgG and IgM containing IC, respectively) were not sufficient or of the correct type to attract and fixate complement. It should be noted that this was not due to the relatively young age (20 weeks) of the mice, as staining of kidneys from 8–12 month-old B6.Act1−/− and TKO mice also failed to show glomerular C3 fixation (data not shown). Also, this observation correlates with the fact that none of the mice (up to 12 months of age) developed renal failure as determined by elevated proteinuria levels (data not shown).

B6.Act1−/− mice fail to develop Sjögrens syndrome-like disease

In addition to developing lupus-like disease, BALB/C.Act1−/− mice develop early and severe SjS-like disease [8]. In contrast, B6.Act1−/− mice failed to develop gross signs of SjS-like disease including enlarged submaxillary glands and elevated serum anti-SSB/La IgG autoantibodies (Fig. 4A–B). We did find occasional IgG deposition within the glands of B6.Act1−/− mice which appeared to be diminished in the absence of T cells, however both WT and B6.Act1−/− mice displayed areas of mononuclear cell infiltration (Supporting Information Fig. 2). T-cell deficiency only had little or no effect on IC deposition (compare TCRβ/δ−/− with WT, Fig. 4C). Thus, Act1-deficiency results in variable disease symptoms in B6 and BALB/C mice, suggesting that epigenetic interactions within different strains play a role in disease specifications. Such phenomenon is well established and has previously been reported to differentially affect the susceptibility to autoimmunity [23].

Figure 4.

B6.Act1−/− mice fail to develop SjS-like disease. (A) Submaxillary glands were harvested from 16–18-week-old mice and the weight was measured. (B) Serum levels of anti-SSB/La autoantibodies were measured in WT, B6.Act1−/−, TCRβ/δ−/−, and TKO mice (n = 6–10 per strain). Sera were obtained as described in the legend of Fig. 2. Each symbol represents one individual mouse and the horizontal line signifies the mean. (C) Frozen submaxillary glands were stained for deposition of IgG (green) and evaluated by immunofluorescence staining. All immunofluorescence stainings were performed at the same time and pictures were taken using identical camera settings to allow for direct comparison. Pictures are representative of three independent sections per strain.

T2/T3/MZ B-cell accumulation is intact in T-cell deficient B6.Act1−/− mice

After leaving the BM, immature T1 B cells travel to the spleen where they differentiate into T2 or T3 B cells in a B-cell receptor/BAFF-dependent manner [24], [25]. In BALB/C.Act1−/− and BAFF-Tg mice, B-cell hyperplasia and accelerated B-cell differentiation occur due to the cells’ heightened response to BAFF [2], and results in a skewing in the repertoire of transitional B cells toward the T2 B-cell phenotype (B220+AA4.1+CD23+IgM+), as well as increased levels of T3 and marginal zone (MZ) B cells [2, 25]. As T cells may represent a possible source of BAFF [26, 27], we evaluated if BAFF-driven T2/T3/MZ B-cell accumulation was present in TKO mice. Sixteen- to eighteen-week-old B6.Act1−/− mice expressed significantly increased numbers of total immature AA4.1+B220+ B cells (p < 0.05 as compared with WT mice, Fig. 5A). Levels of immature B cells were also increased in TCRβ/δ−/− mice and trended toward an increase in TKO mice (Fig. 5A). Mature B cells, including both MZ and FM B-cell subset, were significantly elevated in T-cell-deficient mice regardless of Act1 expression (Fig. 5A–B) as previously described by others [28], while classical PC (B220lowIgDCD138+) were significantly reduced as a result of T-cell deficiency (Fig. 5C and Supporting Information Fig. 3A–C). B6.Act1−/− mice also displayed elevated levels of MZ B cells, but we found no increase in the number of FM B cells (Fig. 5B).

Figure 5.

B6.Act1−/− mice develop B-cell hyperplasia, elevated levels of T2 and T3 transitional B cells and increased numbers of plasma cells. Seventeen- to -twenty-week-old WT (n = 8), TCRβ/δ−/− (n = 12), B6.Act1−/− (n = 5), and TKO (n = 12) mice were sacrificed and splenic levels of B-cell subsets were determined by flow cytometry. Absolute numbers of (A) immature (AA4.1+B220+) and mature B cells (AA4.1B220+), (B) mature B-cell subsets (MZ: CD21+CD23lowB220+ and FM: CD21lowCD23highB220+), (C) CD138+B220low plasma cells, and (D) immature transitional B-cell subsets (T1: AA4.1+B220+CD23IgMhigh, T2: AA4.1+B220+CD23+IgMhigh and T3: AA4.1+B220+CD23+IgM+/low) were calculated per mouse. Samples were obtained over a 12-month period and analyzed at the time of harvest. Each symbol represents levels in one individual mouse and the horizontal line signifies the mean. (E) The ratio of T2:T1 and T3:T1 were calculated based on the total number of cells (D). Each bar represents the mean + SEM. n = 5–12 per strain as described above. * p < 0.05; ** p < 0.01; * p < 0.001; two-tailed Mann–Whitney test. # denotes data showing a trend toward statistical difference: p < 0.1.

Upon further analysis of immature B-cell subsets, we found that B6.Act1−/− mice displayed a similar skewing in the repertoire from T1 to T2/T3 B cells as previously described for BALB/C.Act1−/− mice (Fig. 5D and Supporting Information Fig. 4) [2]. Interestingly, also TCRβ/δ−/− mice showed elevated levels of T2 and to a lesser extend T3 B cells, suggesting that either (i) B cells accumulated at the immature stage due to lack of additional T-cell-driven differentiation factors or (ii) that TCRβ/δ−/− mice expressed increased BAFF production and thus enhanced T2/T3 B-cell survival. It should also be noted that despite variable numbers of total transitional T1, T2, and T3 B cells, the ratios of T2:T1 and T3:T1 B cells were consistently increased in all gene-deficient mice (TCRβ/δ−/−, B6.Act1−/−, and TKO) as compared with WT mice (Fig. 5E).

Based on these data, we evaluated if T-cell deficiency affected BAFF signaling. We first tested mice for expression levels of TACI and BAFF-R on spleen-derived transitional B cells. In correlation with our previous observation [2], T1 and T2/T3 B cells from all strains expressed comparable levels of BAFF-R and TACI (Fig. 6A). We then tested levels of serum BAFF and found that B6.Act1−/− mice expressed levels similar to WT mice, while T-cell-deficient mice (TCRβ/δ−/− as well as TKO) displayed increased levels of BAFF (p < 0.0001, as compared with WT and B6.Act1−/−, respectively) (Fig. 6B). These data suggest that the increased levels of T2/T3 B cells observed in T-cell-deficient mice could in fact be driven by excess BAFF.

Figure 6.

TCRβ/δ-deficient mice express increased levels of serum BAFF, but no difference in levels of BAFF-R and TACI expression. (A) WT, TCRβ/δ−/−, B6.Act1−/−, and TKO mice were sacrificed at 16–18 weeks of age and levels of BAFF-R and TACI was evaluated on T1 (B220+AA4.1+CD23low) and T2/T3 (B220+AA4.1+CD23high) immature B-cell subsets. Samples were obtained over a 12-month period and analyzed at the time of harvest. Data shown were pooled from 6–8 mice analyzed per strain. (B) Levels of serum BAFF was detected in 17–20-week-old mice by ELISA as described in the Materials and methods. Sera were obtained and stored over a 12-month period. All samples were run at the same time to minimize assay-to-assay variation. n = 10 (WT); n = 9 (B6.Act1−/−); n = 9 (TCRβ/δ−/−); n = 11 (TKO). * p < 0.05; ** p < 0.01; *** p < 0.001; two-tailed Mann–Whitney test.

The number of MZ B cells increases in mice deficient in Act1 or T cells

Finally, accumulation of MZ B cells is a common readout in autoimmune mouse models and has been attributed a significant role in driving autoantibody production [29-31]. We tested spleen samples for numbers of MZ B cells (B220+AA4.1CD21+CD23low) by flow cytometry. Deficiency in either T cells (TCRβ/δ−/−) or Act1 (B6.Act1−/−) resulted in significantly increased levels of MZ B cells (p < 0.05 versus WT, Fig 7). Combined deficiency in TKO mice did not result in further increases.

Figure 7.

B6.Act1−/− as well as TCRβ/δ-deficient mice develop significantly increased levels of MZ B cells. Splenic MZ B cells (B220+CD21highCD23lowIgMhigh) were identified in 16–18-week-old WT (n = 7), TCRβ/δ−/− (n = 11), B6.Act1−/− (n = 5), and TKO mice (n = 11) by flow cytometry and total number of cells were enumerated. The mean number of cells (×106) is added above the graph for easier assessment. Samples were obtained as described in the legend of Fig. 5. * p < 0.05; ** p < 0.01; two-tailed Mann–Whitney test.

Discussion

BAFF-Tg mice are known to develop a SLE-like disease independently of T cells [17]. Act1 is well established as a negative regulator of BAFF signaling, and thus we expected the auto-immune phenotype of B6.Act1−/− mice to be T-cell independent as well. Upon analyzing T-cell-deficient B6.Act1−/− mice, it became clear that while all IgG-related abnormalities were absent in TKO mice, IgM-related autoimmune characteristics, including IgM anti-nuclear autoantibodies and IgM-IC deposition in kidney glomeruli, were retained or even elevated in these mice. Both TCRβ/δ−/− and TKO mice experienced similarly elevated IgM levels within the kidney glomeruli, that is, the deposition was not dependent on Act1-deficiency and did not correlate with specific levels of anti-nuclear IgM autoantibodies. Also, neither TCRβ/δ−/− nor TKO mice appeared to fixate complement and none of the mice developed overt renal disease, suggesting that the specificity of the deposited IgM antibodies is different from the specificity of IgM antibodies in B6.Act1−/− mice and has no or minor influence on disease development. Thus, not surprisingly we found that T cells are necessary for IgG, but not IgM, autoantibody production and IgG antibody-related symptoms in lupus-like disease in B6.Act1−/− mice.

Although the absolute number of T3 B cells was less in TKO mice than in B6.Act1−/− mice, the ratio of T3:T1 was similarly elevated in both strains as compared with WT mice, suggesting that this step in B-cell differentiation is T-cell independent. In fact, the absence of T cells alone (in TCRβ/δ−/− mice) led to elevated levels of T2 and T3 B cells and elevated ratios of T2:T1 and T3:T1. Serum BAFF levels were significantly higher in T-cell-deficient mice (13 ng/mL versus 10 ng/mL in WT and B6.Act1−/− mice) and could possibly be the mechanism driving this differentiation, however levels did not reach those seen in BAFF-Tg mice (>35 ng/mL, [21]), making further studies needed to firmly make such conclusion. T3 B cells have been shown to consist of primarily anergic B cells highly enriched for autoreactivity and may represent a population of cells specifically enriched during autoimmunity [32]. It has been suggested that the strength of BCR signaling during T1 B-cell stimulation decides whether the cells will differentiate along the T2-FM/MZ pathway (strong signal) or become anergic T3 B cells (attenuated signal). As increased BAFF signaling has been associated with increased survival of immature B cells with lower antigen-binding affinity (including potentially autoreactive B cells) [33], it is not surprising that many T1 B cells in Act1-deficient mice differentiate into anergic T3 B cells. Interestingly, our data imply that in TKO mice, when BAFF levels are increased at the same time as the response to BAFF is elevated, T3 cells are partially rescued shifting the balance toward the T2 and eventually MZ/FM B-cell subsets. This is consistent with data from BAFF-Tg mice, where the very high levels of BAFF (>35 ng/mL) favors accumulation of T2 B cells rather than T3 B cells [33]. Thus, the absolute level of serum BAFF and/or responsiveness to BAFF may be instrumental in driving immature B-cell differentiation, resulting in (i) controlled T2/T3 differentiation at normal BAFF levels, (ii) increased T3 B-cell differentiation at intermediate BAFF levels hereby preventing autoimmunity by anergizing potentially autoreactive B cells, and (iii) complete T2/FM/MZ differentiation at very high BAFF levels resulting in T-cell-independent autoimmunity as seen in BAFF-Tg mice.

MZ B cells are known to differentiate from T2 B cells in an NF-κB-dependent (p65 and c-Rel) manner [34], although the initiating signals inducing differentiation remain to be identified. It is interesting to note that BAFF-mediated signaling activates the NF-κB pathway [1, 2], that treatment of mice with recomb-inant BAFF significantly and specifically increases T2 and MZ B-cell compartments [35], and that BAFF-deficient mice lack MZ B cells [36]. Thus, it is possible that MZ B-cell differentiation is specifically driven by BAFF. In support hereof, we observed a positive correlation between BAFF levels in WT and TCRβ/δ−/− mice, although due to the small differences in BAFF levels the analysis failed to reach statistical significance (Pearson test: R2 = 0.29, p = 0.22, n = 7, data not shown). Due to the function of Act1 on BAFF responsiveness rather than BAFF production, we were unable to extend this analysis to Act1-deficient mice.

Given the many known roles of Act1, Act1-deficient mice develop a complex phenotype involving many cell subsets. Even in B cells, Act1 appears to play multiple roles (i.e. control of CD40 and BAFF-R-signaling and responsiveness to IL-17A). Interestingly, it has been shown that IL-17A functions to increase B-cell survival, proliferation, and differentiation and hence supports the generation and persistence of autoreactive B cells [37]. As Act1 is a positive regulator of IL-17A signaling and a negative regulator of BAFF, it follows that the balance of Act1 binding to either IL-17R or BAFF-R is crucial for maintaining B-cell tolerance (Fig. 8). T-cell-deficient Act1-sufficient mice express very little IL-17A (data not shown), increased BAFF, and accelerated B cell maturation (increased T2/T3, MZ, and FM), slightly elevated levels of anti-nuclear IgM antibodies and elevated deposition of IgM-IC in the kidney glomeruli (Fig. 8, bottom left panel). As expected all IgG and IgA production is abolished in the absence of T-cell help, that is, CD40 ligation (Fig. 8, bottom panels). Act1-deficiency on the other hand results in increased BAFF-mediated signaling driving T1 to T2/T3 B-cell maturation and elevated levels of MZ and FM B cells (Fig. 8, top right panel). We suggest that more self-reactive B cells (low BCR-antigen-binding affinity), which would normally have been deleted due to negative selection, survive, and differentiate as a result of BAFF hyperresponsiveness. In addition, Act1-deficiency increases CD40L-mediated Ig class switching and the differentiation of IgG-secreting plasma cells hence elevated levels of IgG autoantibodies (Fig. 8, top right panel). Whether lack of IL-17-mediated signaling in the absence of Act1 is counteracting this effect by diminishing B-cell survival is currently unknown. Finally, when combining TCR deficiency with Act1 deficiency (TKO mice) it follows that BAFF-mediated signaling is increased leading to increased levels of T2/T3 immature B cells, MZ and FM B cells including cells with self-reactivity. CD40L-dependent class switching is eliminated by the lack of T cells resulting in elevated levels of IgM-secreting anti-nuclear-specific plasma cells (Fig. 8, bottom right panel).

Figure 8.

Model of the involvement of BAFF, IL-17A, and T cells during immature B-cell differentiation and plasma cell generation. WT mice (upper left quadrant): Immature B-cell (IMB) survival and different-iation is under the control of Act1 (A) negatively affecting BAFF and CD40L and positively affecting IL-17A induced signaling (reviewed in [9]). TCRβ/δ−/− mice (lower left panel) express increased levels of BAFF by a yet unknown mechanism, resulting in elevated IMB survival and differentiation into marginal zone (MZ) and follicular mature (FM) B cells. Due to the lack of T cells (CD40L), however, no Ig class switching takes place and antibodies present are IgM only. B6.Act1−/− mice (upper right quadrant) have normal levels of BAFF, but increased responsiveness due to the lack of Act1 [1]. This drives IMB survival and differentiation into MZ and FM B cells [2]. Furthermore, as Act1-mediated negative regulation of CD40 signaling is lacking, the formation of IgG-secreting autoreactive plasma cells (PC) is increased. Combined deficiency of T cells and Act1 (lower right quadrant) results in both high levels of BAFF and BAFF hyperresponsiveness [1] driving the accumulation of MZ and FM B cells, but due to the lack of CD40L-expressing T cells, still inhibited Ig class switching. IL-17 is induced during inflammatory conditions and further enhances B-cell survival, proliferation, and differentiation in WT mice [37]. Lack of T cells or Act1 will each separately abrogate this pathway.

In conclusion, T-cell-deficient B6.Act1−/− mice represent a model in which BAFF-mediated signaling during transitional B-cell differentiation and maintenance can be further explored without the interference of CD40L-CD40-mediated B-cell activation and IL-17-dependent amplification of the response.

Materials and methods

Mice

All mice studied were on the C57BL/6 background. BALB/C.Act1−/− mice [1] were backcrossed more than 12 generations to C57Bl/6J. B6.129P-Tcrbtm1MomTcrdtm1Mom/J mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Triple knockout (TKO) animals (TCRβ−/−TCRδ−/−Act1−/−) and age-matched controls (WT, Act1 deficient (B6.Act1−/−), and TCRβ−/−TCRδ−/− double-deficient mice (TCRβ/δ−/−)) were generated in our Biological Research Unit at Lerner Research Institute. Both male and female mice were analyzed. Animals were maintained in accordance with guidelines provided by the Cleveland Clinic Foundation Animal Research Committee.

Measurement of serum immunoglobulins

Serum was collected from WT, B6.Act1−/−, TCRβ/δ−/−, and TKO mice and levels of serum immunoglobulins were measured by enzyme-linked immunosorbent assay (ELISA). Serum immunoglobulins (IgM, IgG, IgG1, and IgG2c) and anti-nuclear autoantibodies (ANA; anti-chromatin IgG, anti-chromatin IgM, anti-histone IgG, and anti-histone IgM) were detected as previously described [38] using HRP-conjugated anti-mouse IgG and anti-mouse IgM secondary antibodies (Southern Biotech). Anti-dsDNA IgG and anti-SSB/La IgG levels were determined using the manufacturer's protocol (Alpha Diagnostic International Inc., TX). Anti-dsDNA IgM levels were determined using the anti-dsDNA IgG kit, but replacing the secondary anti-IgG antibody with HRP-conjugated anti-mouse IgM. Levels of serum Igs were determined based on a colorimetric assay measured on a Victor 3 plate reader (Perkin Elmer, MA) at 450 nm.

Immunohistochemistry and immunofluorescent stainings

Kidneys were collected from four mice per strain (WT, TCRβ/δ−/−, B6.Act1−/−, and TKO). One half was fixed in 10% formalin and embedded in paraffin. Five-micrometer sections were generated and kidney morphology was detected by hematoxylin and eosin (H&E) staining of formalin-fixed samples. A total of three sections more than 30 μm apart were analyzed per mouse. Mesangial cellularity was determined in a blinded fashion by counting of nuclei within 2–3 glomeruli per section per mouse. Another half kidney was quick frozen in Tissue Tek® (Sakura, CA) on dry ice. Immunofluorescence staining were performed as previously described [38]. Briefly, 5 μm sections were obtained and at least two sections per mouse were analyzed. Frozen samples were fixed with cold (−20°C) acetone, washed with 1× phosphate buffered saline (1 × PBS, pH 7.4), and blocked with 10% normal goat serum (Invitrogen, CA) for 30 min. Antibodies specific to IgG, IgM, or IgA (Texas-red-conjugated goat anti-mouse IgG, Invitrogen) or complement factor 3 (FITC-conjugated goat anti-mouse C3, ICL Lab, OR) were diluted 1:750 and 1:500, respectively, in 1 × PBS and applied over night at room temperature in a humid chamber. After incubation, slides were washed extensively and mounted in 20% glycerol/PBS. All images were collected using a Leica DMR fluorescence microscope equipped with a cooled CCD camera (Q-Imaging Retiga EXi) and ImagePro Plus software (MediaCybernetics, MD). The density of IgG, IgM, and IgA staining was determined using ImagePro Plus and is given by the level of density (red)/glomulus area/mouse. Twenty-four- to twenty-six glomeruli representing 3–4 individual mice/strain were measured. The actual staining level (density/glomerulus) is displayed as fold of WT levels.

Flow cytometric analysis

Single-cell preparations of spleens and BM were generated according to standard procedures. Red blood cells were lysed in ACK-buffer (0.15 M NH4Cl, 0.01 M KHCO3, 0.1 mM EDTA) for 5 min on ice. Remaining cells were washed and resuspended in 1 × PBS. Cells were stained with fluorescently conjugated antibodies against CD3, B220, CD23, CD21, CD24, AA4.1 (CD93), CD138, IgM, IgD, GL-7, BAFFR, and TACI (all from eBioscience Inc., CA) in 1 × PBS for 20–40 min. All samples were fixed in 1% parafomaldehyde before analysis. Samples were run on a FACS Calibur (BD Biosciences, CA) and data analysis was performed using FlowJoTM (Tree Star Inc., OR). B cells and B-cell subsets were gated as previously described [2].

Serum BAFF analysis

Serum was obtained from 16–18–week-old mice (n = 7 per strain: WT, TCRβ/δ−/−, B6.Act1−/−, and TKO) and tested for levels of BAFF/BLyS/TNFSF13B by ELISA following the manufacturer's protocol (R&D systems, MN). Prior to application, serum samples were diluted 1:4 in assay diluent. Levels of serum BAFF were determined based on a colorimetric assay measured on a Victor 3 plate reader (Perkin Elmer) at 450 nm and concentrations were determined based on the supplied standard.

Statistics

Statistical analyses of flow cytometry data were performed using nonparametric Mann–Whitney t-tests (GraphPad Prism, version 4.03). Statistical p-values are given as *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.

Acknowledgments

We wish to thank Ami Saraiya, Ayesha Khan, and Abhishek Trigunaite for excellent technical help throughout this study. This study was supported by an NIH grant 5R01AI065470 (X.L.) and seed funding from the Cleveland Clinic Foundation (T.N.J.).

Conflicts of interest

The authors declare no financial or commercial conflict of interest.

Abbreviations
ANA

Anti-nuclear autoantibody

BAFF

B-cell activation factor of the TNF family

cLN

cervical lymph node

FM

follicular mature

IC

immune complex

MZ

marginal zone

SjS

Sjögren's syndrome

SLE

Systemic lupus erythematosus

Ancillary