Transcription Factor Gfi1 Restricts B Cell-Mediated Autoimmunity

Mice

All mice were bred and maintained under specific pathogen-free conditions in the central animal facility at Hannover Medical School (Hannover, Germany). Gfi1^−/− mice and Gfi1^GFP/+ mice were kindly provided by T. Möröy (University of Essen, Essen, Germany). Gfi1^−/− mice were generated by replacing exons 2, 3, 4, and 5 of Gfi1 by neomycin resistance gene, which results in the absence of Gfi1 protein in all organs (19). Gfi1^GFP/+ knockin mice were generated by replacing exons 3, 4, and 5 of Gfi1 with GFP cDNA. The levels of Gfi1 protein expression in Gfi1^+/+ and Gfi^GFP/+ mice are comparable. Gfi1^−/− and Gfi^GFP/+ mice were backcrossed with C57BL/6 for seven generations. NOD/SCID mice were obtained from Charles River Laboratories and were maintained in the pure NOD background. In all experiments, age- and sex-matched mice were used. All animal experiments were approved by the institutional review board of Hannover Medical School.

Flow cytometry

Single cell suspensions were analyzed by flow cytometry using FACScan or FACSCanto and CellQuest software, FACSDiva software (BD Biosciences), or Flow Jo software (Tree Star), respectively. Cell sorting of defined subpopulations was performed using a MoFlo cell sorter (DakoCytomation) or a FACSAria cell sorter (BD Biosciences), respectively.

The following mAbs (all from BD Pharmingen) were used: B220, IgM, CD19, and CD138. In all experiments, cells were also stained with corresponding isotype-matched mAbs. Cells reacted with biotinylated mAbs were incubated with fluorochrome-conjugated streptavidin-PerCP or streptavidin-allophycocyanin (BD Pharmingen). All fluorescence intensity plots are shown in log scales.

ELISA

Serum Ig levels were determined by enzyme-linked immunosorbent assay (ELISA), using Abs specific for each isotype (IgA, IgG2a, IgG2b, and IgM; BD Pharmingen). Anti-nuclear Abs were quantified using a commercially available ELISA kit (Alpha Diagnostics).

Western blotting

For Western blot analysis, 20 μg of protein was loaded on an 8 or 10% SDS gel, separated by electrophoresis, and blotted onto nylon membranes. The membranes were exposed to reacting with Erk1/2p, Erk1/2 (Cell Signaling Technologies), Lyn, p-Lyn Ccna2, Ccnb2, p21, and GAPDH (Santa Cruz Biotechnologies) followed by staining with secondary Abs conjugated to HRP. The enzymatic reaction was visualized using ECL reagents (ECL kit; Amersham Biosciences).

B cell proliferation assays and calcium flux

CD19⁺ cells were isolated from spleen using magnetic microbeads (Miltenyi Biotec). Purified B cells (2 × 10⁵) were cultured in 96-well plates containing complete RPMI 1640 supplemented with the stimuli recombinant murine IL-4 (Peprotech), anti-CD40 (BD-Pharmingen), anti-IgM (Southern Biotechnologies), or LPS (Sigma-Aldrich), each at a final concentration of 10 ng/ml for 2 days. On day 3 of culture, the cells were pulsed with 1 μCi of [³H]thymidine for 16 h. Incorporated [³H]thymidine was quantified by scintillation counting.

For calcium flux assays, purified total CD19⁺ or CD19⁺B220⁺CD138⁻CD43⁻ cells were prewarmed and analyzed on a FACSCanto flow cytometer for and initial 45 s and then stimulated with either goat anti-mouse IgM F(ab′)2 (Jackson Immunochemicals) or ionomycin (Sigma-Aldrich), and data collection was continued for 5 min. Data were analyzed with FlowJo.

In vivo B cell transfer

For B cell transfer experiments, 1 × 10⁶ CD19⁺B220⁺CD138⁻CD43⁻ splenic B cells obtained from 4-wk-old mice were injected i.v. into congenic NOD/SCID mice. Eight weeks after injection the mice were bled and anti-nuclear serum autoantibodies were measured by ELISA (Alpha Diagnostics). Twelve weeks after injection, mice were sacrificed and pathological studies were performed.

In vivo BrdU studies

In vivo incorporation of BrdU into CD19⁺CD138⁺ cells was assessed using an allophycocyanin BrdU flow kit (BD Pharmingen). After a single i.p. injection of BrdU (1 mg per 6 g of mouse weight; Sigma-Aldrich), an admixture of 1 mg ml⁻¹ BrdU was added to the drinking water for 5 days. Mice were sacrificed and bone marrow cells were prepared, stained with Abs recognizing CD19, B220, CD138 and BrdU, and analyzed by flow cytometry.

RNA extraction and real-time PCR

Total RNA was isolated using commercially available kit systems (Absolutely RNA Miniprep kit; Stratagene). cDNA was synthesized using an oligo(dT) primer and Expand reverse transcriptase (Roche). The PCR was performed in duplicates using the 7500 real-time PCR system and power SYBR Green PCR master mix (Applied Biosystems) according to the manufacturer’s instructions. Primer pairs used in the study are indicated in supplemental table 2.⁴

Histopathology and immunofluorescence

Animals were perfused with 4% paraformaldehyde. Tissues were embedded in paraffin. Immunocytochemistry was performed with a biotin-avidin technique using primary Abs against the following targets: CD3 (Serotex), B220 and Mac3 (BD Pharmingen), CD138 (BD Pharmingen), Ig (Amersham Biosciences), glial fibrillary acidic protein (NeoMarkers), and complement C9 (a gift of Dr. P. Morgan, Cardiff, University, Cardiff, U.K.).

To detect reactivity of autoantibodies to brain tissue, wild-type brain cryosections were blocked and exposed to blood serum obtained from Gfi1^+/+ or Gfi1^−/− mice, respectively. After a series of washing steps, the sections were incubated with FITC-conjugated goat-anti-mouse Ig (DakoCytomation) Abs. To detect Ig deposits in kidneys, renal cryosections were incubated with FITC-conjugated goat-anti-mouse Ig (DakoCytomation) Abs. Imaging analysis was done using a Zeiss Axiovert 200 microscope. Electron microscopy was performed following standard Epon embedding and ultrathin sectioning.

Microarray analysis and bioinformatics

Gene expression levels were determined by means of the Affymetrix microarray suite 5.0 (MAS 5.0). MAS 5.0 software algorithms allow quantitative estimation of gene expression and p values to establish a confidence level concerning the accuracy of measurement of an mRNA of interest (detection p value) and changes in gene expression (change p value). Concerning the measured p values, the criteria of present (P) or absent (A) define the quality of signal measurement, and increase (I) or decrease (D) define the signal change, respectively. For normalization, all array experiments were scaled to a target intensity of 150, otherwise using the default values of the microarray suite. Filtering of the results was done as follows. Genes are considered as regulated when their fold change is ≥1.5 or ≤−1.5, the change p value is not “NC” (no change), and at least one of the two compared signals was detected with high accuracy (absent call for both signals were not allowed). Differentially expressed transcription factors were manually determined based on the information available at the Gene Ontology database (www.geneontology.org/).

Statistical analysis

Data are presented as mean ± SEM. Statistical significance was assessed using a two-sided Student’s test. Values of p > 0.05 were considered to be non significant (NS) and values of p < 0.05 and p > 0.01 were represented as ∗. p < 0.01 and p > 0.001 were represented as ∗∗ and p < 0.001 were represented as ∗∗∗.

Autoimmunity in Gfi1-deficient mice

First, we analyzed paraffin-embedded brain sections from Gfi1^−/− and Gfi1^+/+ mice in greater detail. In line with previous reports (29), we were unable to document any gross structural abnormalities in the brains of Gfi1^−/− mice. However, we detected occasional meningeal inflammatory infiltrates (characterized by the presence of B cells, C9 deposition, and glial fibrillary acidic protein expression) and a striking deposition of Ig in Gfi1^−/− brains (Fig. 2⇓A). We also measured serum levels of anti-nuclear autoantibodies in 8-and 16-wk-old Gfi1^−/− and Gfi1^+/+ mice, respectively. Gfi1^−/− mice showed anti-nuclear Abs that increased with age (Fig. 2⇓B). Interestingly, serum of Gfi1^−/− mice gave rise to a nuclear staining pattern in native brain cryosections, whereas no autoantibodies could be detected in serum of wild-type control mice (Fig. 2⇓C). A pathophysiological role of anti-nuclear autoantibodies was further illustrated by analysis of kidneys revealing subendothelial electron-dense deposits in the glomerular basement membrane (Fig. 2⇓, D and E). These data suggest that Gfi1-deficient mice suffer from autoantibody-mediated immunopathology.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Autoimmunity in Gfi1^−/− mice. A, Brain sections from Gfi1^−/− and Gfi1^+/+ mice stained with Abs recognizing IgG immunoglobulins, B220, glial fibrillary acidic protein (GFAP), and C9. B, Anti-nuclear Ab titers in 8- and 16-wk-old mice as determined by ELISA. Results represent the mean values of six mice in each group. Error bars represent SEM. Data are representative of three independent experiments. C, Brain cryosections incubated first with serum of Gfi1^+/+ or Gfi1^−/− mice and subsequently with FITC-conjugated antimouse Ig secondary Abs. Data are representative of two independent experiments. D, Transmission electron micrographs of renal sections revealing electron-dense deposits in the subendothelial space of the glomerular basement membrane of Gfi1^−/− mice (PC, Podocyte; FP, foot processes of podocytes; EC, endothelial cell; arrows indicate subendothelial electron dense deposits). E, Renal cryosections stained with FITC-conjugated antimouse immunoglobulins. Data are representative of three independent experiments.

Analysis of the peripheral B cell compartment in Gfi1-deficient mice

Because the onset of Ab-mediated autoimmunity has been linked to increased B cell proliferation and plasmacytosis in various transgenic mouse models (33), we characterized the peripheral B cell compartment in greater detail. Early investigations in Gfi1-deficient mice revealed a paucity of mature B cells in bone marrow (19, 20, 24, 26). In contrast, peripheral lymph nodes and spleens appeared enlarged in size when compared with those of control mice (Fig. 3⇓A). Overall, CD19⁺ B cells of Gfi1-deficient mice exhibited an activated phenotype as documented by surface expression of CD43 (supplemental figure 2A). Further, increased surface expression of MHC class II and costimulatory molecules such as CD40, CD80, and CD86 in CD19⁺CD43⁺ cells confirmed the hyperactivated phenotype of Gfi1-deficient B cells (supplemental figure 2B). Within the CD19⁺ splenic B cell pool, a relative increase of IgM⁺B220⁻ cells was seen, a subpopulation representing plasma blasts and terminally differentiated plasma cells as shown by expression of CD138 (Fig. 3⇓B). Increased numbers of CD138⁺ cells were also seen in peripheral and mesenteric lymph nodes (Fig. 3⇓C). Despite this marked increase in plasma cells in peripheral lymph nodes, the lymph node architecture remained preserved (Fig. 3⇓D). Furthermore, no destructive plasma cell infiltrates were seen in bones or other organs, suggesting that the plasma cell expansion represented benign hyperplasia. In view of our observation of increased plasma blasts/cells in Gfi1^−/− mice, we were interested to know whether Gfi1 is selectively expressed in this B cell subpopulation. We made use of a transgenic GFP reporter strain (34) that allows determination of the transcriptional activity of the Gfi1 locus by FACS analysis. As shown in Fig. 3⇓E, GFP expression was markedly enhanced in CD138⁺ B cells when compared with CD138⁻ B cells, suggesting that Gfi1 may indeed play a regulatory role in Ab-secreting cells. This increased expression level of the surrogate marker GFP in the GFP reporter cells was paralleled by increased transcriptional activity of the Gfi1 locus in CD138⁺ B cells when compared with CD138⁻ B cells by RT-PCR (supplemental figure 3), suggesting that this effect was not secondary to GFP accumulation in CD138⁺ B cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Plasmacytosis in Gfi1^−/− mice. A, Comparison of size of spleen and peripheral lymph nodes (pLN) in Gfi1^+/+ (a) and Gfi1^−/− (b) mice. B, FACS plots indicating a relative increase in B220⁻IgM⁺CD138⁺ plasma blasts/cells in the spleen of 12-wk-old Gfi1^−/− mice. Splenic B220⁻IgM⁺ cells (upper panel) of Gfi1 knockout mice were gated (G1) and analyzed for CD138 expression (lower panel); shaded histograms represent respective isotype control-treated cells. Data are representative of six independent experiments. C, Absolute numbers of plasma blasts/cells in peripheral lymphoid organs of 12-wk-old Gfi1^+/+ and Gfi1^−/− mice (n = 4 mice). Data are representative of eight independent experiments. pLN, Peripheral lymph node; mLN, mesenteric lymph node. D, H & E-stained sections of peripheral lymph nodes from Gfi1^+/+ and Gfi1^−/− mice. Complete lymph nodes and representative higher magnification windows show large clusters of plasma blasts/cells and preserved lymphoid architecture in Gfi1^−/− mice. Data are representative of three independent experiments. E, Transcriptional activity of Gfi1 locus determined by GFP expression in CD19⁺CD138⁻ and CD138⁺ B cell subsets in Gfi1^+/GFP mice. Black histograms represent GFP fluorescence in Gfi1^+/GFP cells, gray lines represent autofluorescence in Gfi1^+/+ cells. Data are representative of three independent experiments. F, Serum Ig isotype levels in Gfi1^+/+ and Gfi1^−/− mice as determined by ELISA. Results represent the mean values of Ig titers of four mice in each group. Error bars represent SEM. Data are representative of two independent experiments.

To assess whether plasmacytosis in Gfi1-deficient mice might lead to hypergammaglobulinemia, serum Ig levels were quantified by ELISA. As shown in Fig. 3⇑F, Gfi1^−/− mice showed a selective increase in IgA, IgG2a, and IgE isotypes but normal levels of IgG1, IgG2b, IgG3, and IgM levels.

Functional characterization of peripheral B cells

Next, we asked the question of whether the observed plasmacytosis might ensue from unrestricted proliferation of B-lineage cells. Purified CD19⁺ cells were stimulated in vitro and their proliferation was measured by [³H]thymidine incorporation. In comparison to control B cells, Gfi1^−/− B cells showed a higher rate of proliferation in response to anti-CD40, anti-IgM, and LPS (Fig. 4⇓A), suggesting that Gfi1 restricts B cell hyperproliferation. We performed in vivo BrdU labeling studies to assess whether the proliferative capacity of CD138⁺ cells is increased in the absence of Gfi1. As expected, the proliferation of CD138⁺ cells was much higher in Gfi1-deficient mice compared with wild-type littermates (Fig. 4⇓B), suggesting that the splenic CD138⁺ cells of Gfi1-deficient mice represent mostly plasma blasts. Further, we investigated whether Gfi1-deficient B cells show enhanced differentiation into CD138⁺ cells in vitro in response to LPS. Interestingly, both absolute numbers and Ig secretion of in vitro differentiated CD138⁺ cells were increased in the absence of Gfi1 (Fig. 4⇓, C and D).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Functional analysis of peripheral B cells in Gfi1^−/− mice. A, Proliferation capacity of B cells in response to indicated stimuli. Purified CD19⁺ cells were cultured in the presence various stimuli for 48 h and pulsed with [³H]thymidine. Results represent the mean values of triplicates. Error bars represent SEM. Data are representative of two independent experiments. α- represents “anti-.” B, Proliferation of CD19⁺CD138⁺ B cells in vivo. Gfi1^+/+ and Gfi1^−/− mice were fed with BrdU through drinking water for 4 days. CD19⁺CD138⁺ cells were pregated and histograms were generated for quantifying the incorporated BrdU. C and D, In vitro differentiation and Ig secretion of Gfi1-deficient CD19⁺CD138⁺ cells. Sorted CD19⁺CD138⁻ CD43⁻ B cells (5 × 10⁶) were cultured in vitro in the presence of LPS for 5 days. On day 6, cells were counted and their absolute numbers were determined (C), supernatants were harvested, and their Ig levels were determined by ELISA (D). E, Western blot analysis of Erk1/2 and Lyn tyrosine phosphorylation (p-Erk and p-Lyn, respectively) in purified CD19⁺ splenic B cells upon stimulation with anti-IgM Abs. Data are representative of two independent experiments. F, Intracellular calcium release in of Fluo3-AM-loaded B cells upon stimulation with anti-mouse IgM F(ab′)₂. Ionomycin (Ino)-treated cells were used as positive controls. Data are representative of three independent experiments.

Because Gfi1 may be involved in the transcriptional regulation of multiple genes controlling growth and differentiation, we were interested in analyzing signals emanating from the BCR that are integrated by a cascade of phosphorylation and dephosphorylation events. We analyzed protein lysates of Gfi1-deficient and wild-type B cells for the specific levels of Lyn and Erk phosphorylation, two important elements in the BCR-mediated signal cascade. Although the overall expression levels of both Lyn and Erk were comparable in Gfi1-deficient and wild-type B cells, the phosphorylated forms were increased in Gfi1-deficient B cells (Fig. 4⇑E and supplemental figure 4). To measure immediate effects of BCR stimulation, we monitored the release of calcium upon BCR crosslinking. Compared with wild-type B cells, Gfi1-deficient B cells showed a more rapid release of calcium (Fig. 4⇑F and supplemental figure 5), suggesting that Gfi1 has a role in integrating the strength of BCR signaling.

Development of autoimmunity in Gfi1-deficient mice is B cell mediated

Because autoimmunity could result from multiple effectors in an abnormal hematopoietic system, we next sought to analyze the role of B cells in the development of autoimmunity. We transferred 10⁶ purified CD19⁺ B cells from Gfi1^−/− and Gfi1^+/+ mice into NOD/SCID recipient mice. Similarly as Gfi1^−/− mice, NOD/SCID mice injected with Gfi1^−/− B cells developed autoantibodies (Fig. 5⇓A), eczema, ataxia (data not shown), and Ig deposition in the brain, whereas adoptive transfer of control B cells had no effects (Fig. 5⇓B). Although we cannot rule out that immune effector cells such as dendritic cells (21) and/or T cells (24) may contribute to the autoimmunity in Gfi1-deficient mice, our findings support the notion that unrestricted activation and proliferation of B cells leads to CNS and renal autoimmunity in Gfi1^−/− mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

B cell-mediated autoimmunity in Gfi1-deficient mice. A, Anti-nuclear Ab titers in NOD/SCID mice injected with 1 × 10⁶ purified B220⁺CD19⁺ CD43⁻CD138⁻ splenic B cells from either Gfi1^+/+ or Gfi1^−/− mice. The purity of the sorted cells was >98%. Animals were sacrificed 8 wk after adoptive transfer and anti-nuclear Ab titers were determined by ELISA. B, Brain sections from NOD/SCID mice 12 wk after adoptive transfer of B220⁺CD19⁺CD43⁻CD138⁻ B cells of spleen from either Gfi1^−/− or Gfi1^+/+ mice. Sections were stained for IgG Ig.

Identification of downstream target genes of Gfi1 in B cells

Recent evidence has shown that autoreactive B cells can be activated by TLRs, in particular TLR7 and TLR9 (5). We therefore hypothesized that Gfi1 may control transcription of these TLRs or their downstream effectors. However, we could not detect any differences in the mRNA expression levels of TLR7, TLR9, MyD88, and NF-κB when comparing Gfi1^−/− and Gfi1^+/+ B cells (supplemental figure 6). To further elucidate the mechanism of Gfi1 in the control of B cell differentiation, we next hypothesized that Gfi1 may control a number of downstream target genes specifying B cell differentiation and cell cycle progression. Using Affymetrix microarray analysis, a comprehensive study of the transcriptome was performed in purified splenic CD19⁺CD138⁻ cells from Gfi1^−/− and Gfi1^+/+ mice, respectively. Six hundred and forty-eight genes were found to be differentially expressed (supplemental table 1). In particular, we identified increased expression of 16 transcription factors and decreased expression of 14 transcription factors in Gfi1^−/− B cells (Fig. 6⇓, A and B). These data indicate a complex dysregulation of transcriptional control in Gfi1^−/− B cells. Of interest, expression of Blimp-1 (Prdm1) and XBP-1, two master regulators of plasma cell differentiation, was increased in Gfi1^−/− B cells. In view of the hyperproliferative B cell phenotype, we also analyzed the protein expression profile of selected cyclin-dependent kinases. A markedly increased expression of cyclin A2 and B2 as well as decreased expression of the cell cycle inhibitor p21cip1 in Gfi1^−/− B cells was seen (Fig. 6⇓C). These data highlight that multiple pathways are in dysbalance in Gfi1-deficient B cells, yet further studies are needed to document whether these effects are directly or indirectly influenced by Gfi1.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Dysregulation of transcription factors and cell cycle control factors in Gfi1^−/− B cells. A and B, Graphic and numeric representation of expression profile analysis of transcription factors up-regulated (A) and down-regulated (B) in splenic CD19⁺CD138⁻ B cells of Gfi1^−/− mice compared with Gfi1^+/+ littermates. Relative expression (normalized to the median) is displayed as color (green = normalized expression level below the median; black = near to the median; and red = above the median). Fold change (Gfi1^−/− vs Gfi1^+/+) is calculated by Affymetrix MAS 5.0 software. A comprehensive data set is available in supplemental table 1 and in the Gene Expression Omnibus (National Center for Biotechnology Information) database under accession number GSE12545 (www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc = GSE12545). C, Western blot analysis of cyclin A2 (Ccna2), cyclin B2 (Ccnb2), and p21cip1 (p21) in purified CD19⁺ splenic B cells of Gfi1^−/− and Gfi1^+/+ mice.