Anti-Nuclear Antibody Production and Autoimmunity in Transgenic Mice that Over-Express the Transcription Factor Bright

Abstract

The B cell-restricted transcription factor, Bright, up-regulates immunoglobulin heavy chain transcription three- to seven-fold in activated B cells in vitro. Bright function is dependent upon both active Bruton’s tyrosine kinase and its substrate, the transcription factor, TFII-I. In mouse and human B lymphocytes, Bright transcription is down regulated in mature B cells, and its expression is tightly regulated during B cell differentiation. To determine how Bright expression affects B cell development, transgenic mice were generated that express Bright constitutively in all B lineage cells. These mice exhibited increases in total B220⁺ B lymphocyte lineage cells in the bone marrow, but the relative percentages of the individual subpopulations were not altered. Splenic immature transitional B cells were significantly expanded both in total cell numbers and as increased percentages of cells relative to other B cell subpopulations. Serum immunoglobulin levels, particularly IgG isotypes, were increased slightly in the Bright transgenic mice compared to littermate controls. However, immunization studies suggest that responses to all foreign antigens were not increased globally. Moreover, four week-old Bright transgenic mice produced anti-nuclear antibodies. Older animals developed antibody deposits in the kidney glomeruli, but did not succumb to further autoimmune sequelae. These data indicate that enhanced Bright expression results in failure to maintain B cell tolerance and suggest a previously unappreciated role for Bright regulation in immature B cells. Bright is the first B cell-restricted transcription factor demonstrated to induce autoimmunity. Therefore, the Bright transgenics provide a novel model system for future analyses of B cell autoreactivity.

Introduction

Appropriate regulation of immunoglobulin (Ig) synthesis by B lymphocytes is critical to maintain a balance between induction of strong humoral responses to pathogenic organisms and maintenance of tolerance to self antigens. Several autoimmune diseases, including systemic lupus erythematosus, are first characterized by high titers of anti-nuclear antibodies (ANA) in serum. In such cases, inappropriate production and regulation of these antibodies occur prior to presentation of other systemic disease responses and therefore, represent an early indication of tolerance failure. In most cases, ANA production and autoimmune disease progression is severely influenced by the genetic background of the animals used (reviewed in (1–3)), confirming the complex multigenic nature of most autoimmune diseases. The mechanisms that first lead to overproduction of ANAs and disease progression are unclear, but likely involve disruption of regulatory signals at multiple steps and may even originate from malfunctions in various cell types. Nonetheless, inappropriate antibody production is a defining and primary step in autoimmune disease.

Antigenic signals through the Ig receptor cause B lymphocytes to differentiate and secrete antibody in the case of foreign antigens, or to become tolerized if the antigen is a “self” antigen. B cell tolerance can occur by anergy with survival of anti-self cells that are nonfunctional; by receptor editing of light or heavy chains yielding Ig receptors that no longer bind self antigens; or by apoptosis and clonal deletion of the self–reactive B cells (4–8). Tolerant B cells are produced at two major checkpoints during B cell differentiation, the pre–B cell to immature B cell stage in the bone marrow, and in newly emigrating peripheral cells called transitional B cells (9,10). More recent studies also suggest that tolerance is influenced at later stages in the germinal centers (11). Fetal and neonatal B cells up to day seven after birth are also tolerant, and signals through the sIg receptor more closely resemble those seen in immature adult B cells (12). The complex nature of most autoimmune diseases suggests that not only is tolerance controlled at multiple checkpoints, but defects in multiple gene products can lead to breaks in tolerance (13–15). In some lupus patients, it is now clear that production of ANAs and early breaks in B cell tolerance can precede disease progression by many years (reviewed in (2)). Understanding the molecular events that can lead to production of autoantibodies, before complications due to secondary disease related events occur, is critical for early diagnosis and better disease treatment.

Bright (B cell regulator of Ig_H transcription) is a member of the ARID (A–T rich interaction domain) family of DNA binding proteins and interacts with the matrix association regions flanking both sides of the intronic heavy chain enhancer and 5' flanking sequences of several, but not all, variable heavy chain (V_H) genes (16–19). Several Bright binding sites were identified 5' of the basal promoter of the V1 S107 family gene, and these sites were required for IL–5 and antigen induced μ heavy chain transcription by Bright in vitro (16,20). More recent data indicate that Bright-enhanced transcription of a V1 reporter construct critically requires both the kinase activity of Bruton’s tyrosine kinase (Btk) and the Btk substrate, TFII-I (21,22). These data suggest that Bright and Btk share common pathways in some cases, but may also have independent functions.

Bright is expressed in multiple embryonic tissues, but becomes B cell–restricted in the adult (23). Bright transcription in vivo is tightly regulated during B cell differentiation, so that its expression is high in bone marrow pre-B cell subpopulations and in germinal center activated B cells, but most circulating and splenic mature B cells lack detectable Bright mRNA and protein (23,24). While Bright increases Ig transcription in vitro, it is not required for basal Ig production in mature B cells. Thus, the importance of Bright activity for immunoglobulin production and normal B cell development in vivo is unknown.

To determine whether appropriate regulation of Bright is important for B cell differentiation in vivo, transgenic mice constitutively expressing Bright in all B lineage cells were produced. Sera from non-immunized Bright transgenic mice contained slightly increased levels of total immunoglobulin compared to non-transgenic littermates. Expression of the V1 gene, previously shown to be regulated by Bright in vitro (21,22), was also enhanced in transgenic spleen cells relative to littermate controls. Because the V1 gene is used predominantly in the anti-self response against phosphorylcholine (PC) (25–27), antigen-specific responses reactive with this hapten were also examined. While anti-PC responses were significantly enhanced in the transgenic mice, responses to other foreign antigens did not differ from littermate controls. Strikingly, sera from even very young Bright transgenic mice contained ANAs. Moreover, these mice exhibit increases in B lymphopoeisis with significantly increased numbers of transitional type 1 immature B cells, a well-documented B cell tolerance checkpoint, in the spleen. These data suggest that inappropriate regulation of Bright alone during B cell development results in an early autoimmune phenotype.

Materials and Methods

Transgenic Mice

A full length cDNA for mouse Bright tagged at the carboxyl terminal end with His-Myc sequences (28) was ligated to the SV40 poly A site. The native Bright Kozak sequence was modified to (GCCACCATGC) (29), the resulting DNA was ligated to a 6.3 kb fragment containing the human CD19 promoter (30), and was cloned into pUC19. Excised DNA was injected into FVB/N blastocysts by the Oklahoma Medical Research Foundation Transgenic Core Facility. All animal care and procedures were performed with prior institutional approval and within the review board-specified guidelines. Toe DNA from 10–11 day old pups was assessed for the transgene with PCR primers from the Bright coding sequence (5’-GGAAGAGCAAGAGCTGGAAG-3’) and the His-Myc tag (5’-CAGATCCTCTTCTGAGATGAG-3’). Seven positive founders were obtained and were assessed for transgene expression by retroorbital bleeding and RT-PCR analyses of white blood cell RNA. Age-matched male mice8–15 weeks of age, unless otherwise indicated, were used for all assays.

Cell Preparation and Flow Cytometry

Mice were euthanized, thymus lobes and spleens were harvested, and single cell suspensions in RPMI with 7% FCS were produced using 70 µm strainers. Whole bone marrow cells were obtained from femurs by flushing with a 23 gauge needle containing PBS-3% FCS.

Cell surface phenotype analyses were performed on 1.5×10⁶ cells by flow cytometry using a FacsCalibur or LSRII (BD Biogenics, San Jose, CA). Cell sorting experiments were performed on a FACSARIA cell sorter (Becton Dickinson, Franklin Lakes, NJ). Antibodies purchased from BD were: fluorescein isothiocyanate (FITC)-conjugated CD19 (1D3), CD21 (7G6) and CD4 (RM4-4); phycoerythrin (PE)-conjugated CD8 (53−6.7), CD3 (145-2C11), CD43 (57), CD40 (1C10), CD69 (Hi.2F3), CD80 (1G10), CD86 (GL1) and CD23 (B3B4); allophycocyanin (APC)-conjugated CD45R/B220 (RA3-6B2), CD93/C1qRp (AA4.1); and peridinin chlorophyll-a protein (PerCP)-conjugated CD45R/B220(RA3-6B2). FITC-IgM, PE-IgD (11–26), goat anti-mouse IgM- and 1-A9,MHC II (KH116)-biotin, appropriate isotype controls and streptavidin conjugated-APC were from BD or Southern Biotech (Birmingham, AL). Cells were stained as previously described (23) and fixed in 0.2% paraformaldehyde overnight. Data were analysed using CellQuest Pro software (BD Biosciences).

Western Blots

Single cell suspensions were resuspended in SDS-sample buffer and electrophoresed through 7.5% SDS-polyacrylamide gels under standard denaturing conditions. Proteins were transferred to nitrocellulose membranes and developed with polyclonal rabbit anti-Bright as previously described (31). Blots were developed with alkaline phosphatase substrate (Bio-Rad, Hercules, CA).

ELISA and ANA Assays

Mice were anesthetized for retroorbital bleeding and sera were collected. Costar 96 well U bottom Polyvinyl plates (Corning, Corning, NY) were coated with 100 µl/well of 2 µg/ml of goat anti-mouse Ig in borate saline (pH 8.4) and incubated overnight at 4°C, washed 3× with PBS, blocked with 1% bovine serum albumin (Sigma-Aldrich, St. Lois, MO) in PBS for 1 hour at 20°C, and four dilutions of duplicate serum samples were added overnight at 4°C. Wells were washed 4×, and developed with isotype-specific alkaline phosphatase-labeled antibodies from the Clonotyping System-AP kit (Southern Biotechnologies, Birmingham, AL) and 4-nitrophenylphosphate disodium salt hexahydrate (Sigma). Reactions were stopped with 50 µl/well of 3N NaOH and read on a Dynex MRX microtiter reader (Dynatech Laboratories, Guernsey Channel Island, Great Britain). Standard curves were generated using serially diluted duplicates of known concentrations of each Ig isotype and Ig levels were determined using Excel software. For antigen-specific responses, plates were coated with NP-BSA or PC-BSA and duplicate sera samples were serially diluted. The last dilution to give a reading above background was taken as the final dilution. Detection of anti-double-stranded DNA was performed according to the manufacturer’s directions (kit #5100) using duplicate sera samples at multiple dilutions and the provided positive and negative controls (Alpha Diagnostics Intl. Inc., San Antonio, TX). Urine albumin concentrations were assessed with the Mouse Albumin ELISA Quantitation Kit (Bethyl Laboratories, Inc., Montgomery, TX) as per the manufacturer’s directions. A standard curve for albumin was calculated using a semi-logarithmic scale in Graph Pad Prism Version 4.0 (Graph Pad).

ANA assays were performed using the NOVA Lite HEp-2 kit (INOVA Diagnostics, Inc.) according to the manufacturer’s directions at 1:40, 1:120, 1:360, 1:1080 dilutions in PBS. Slides were incubated with anti-mouse IgG-FITC conjugate and visualized with a Zeiss Axioplan 2i microscope. Pictures were taken with an AxioCam HRm camera (Carl Zeiss International, Thornwood, NY). Data were analyzed using AxioVision LE.

Immunizations

Mice were immunized intraperitoneally with PC-KLH or NP-KLH (0.5 µg/ml) in Freund’s complete adjuvant (CFA) (Sigma, St. Louis, MO) and boosted on day 7 with the same dose of antigen. Serum was collected at days 0, 7, and 14 post-immunization.

RT-PCR Analyses

Total RNA was extracted from sorted B cell subpopulations using TriReagent (MRC, Cincinnati, OH) according to manufacturer’s instructions. cDNA was generated in reactions containing RNA from approximately 7.5×10³ cells as previously described (32). Levels of S107 family-V_H-specific IgM and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA were assessed as described (22). Samples were electrophoresed through 1.0% agarose gels, transferred to GeneScreen Plus hybridization membrane (PerkinElmer Life Science Products, Boston, MA) and hybridized with a ³²P-labeled V1 cDNA probe. Relative intensities of the PCR products were quantified using LumiAnalyst 3.0 software (Roche, Indianapolis, IN).

Cell Culture Assays

Splenic B cells were T cell depleted using anti-Thy-1 and guinea pig complement and isolated by centrifugation through a ficoll gradient as previously described (31). Cells were plated at 1×10⁶ cells / ml in RPMI (supplemented with 10% heat-inactivated fetal calf serum, 100 U/ml penicillin, 100 µg/ml streptomycin, 5×10⁻⁵ M 2-mercaptoethanol, and 1 mM sodium pyruvate) alone or with 20 µg/ml LPS (E. coli 0111:1B4 (Sigma, St. Louis, MO)) or with CD40L-expressing SF9 or wild type SF9 control cells at a ratio of 1 SF9 cell to 10 B cells (31). After 72 hours, cells and supernatants were harvested for flow cytometry and ELISA. In some cases, wells were pulsed with 1µCi of ³H-Thymidine for 6 hours, harvested, and ³H incorporation was measured.

Histological Analyses

Spleens from euthanized animals were harvested and embedded in Tissue-Tek O.C.T. (Optimal cutting temperature, Electron Microscopy, Fort Washington, PA) and flash frozen for sectioning on a Leica CM3050 cryomicrotome (Leica, Vienna, Austria). Sections were rehydrated in PBS for 45 minutes and blocked with 2% BSA in PBS for 20 minutes at room temperature. Goat anti-mouse IgM-Alexa546 (Molecular Probes, Eugene, OR) and rat anti-mouse metallophilic macrophages (MOMA-1) (Serotec, Raleigh, NC) with goat anti-rat IgG-Alexa350 were used for staining. Sections were sealed with Prolong Antifade (Molecular Probes) and viewed with a Zeiss LSM150 confocal microscope and Ziess LSM Image Browser software (Carl Zeiss, Inc., Thornwood, NY).

Mice were anaesthetized intraperitoneally with 1ml of 20mg/ml Avertin (2, 2, 2-Tribromoethyl & Tertiary Amyl alcohol), prebled retroorbitally and kidneys were extracted after perfusion with Dulbeco’s PBS followed by chilled 2% paraformaldehyde (EMS Biosciences). Kidneys were cut across the juxtamedullary gap and transversely and incubated in 2% paraformaldehyde overnight prior to tissue processing & paraffin embedding or cryopreservation. Sections were prepared on positively-charged slides and blocked with 5% non-fat dairy milk for 45 minutes at 20°C. Sections were stained with an isotype control (donkey anti-goat IgG) or Cy3-conjugated AffiniPure Fab fragment (donkey anti-mouse IgG, Jackson Immunochemical) for 45 minutes at room temperature. Slides were rinsed with PBS + 0.05% Tween 20 for 5 minutes, twice with PBS, and DAPI (4', 6-diamidino-2-phenylindole, Sigma) was added for nuclear staining for 2 minutes. For complement C3 staining, Alexa Fluor 546® rabbit anti-mouse IgG (H+L) (Molecular Probes, Eugene, OR) and goat anti-mouse complement C3-FITC (MP Biomedicals,Inc-Cappel, Aurora, OH) were used. Slides were visualized with a Zeiss Axioplan 2i microscope and visualized with an AxioCam HRm camera (Carl Zeiss International) at 20× magnification. Glomerular and tubulointerstitial immunofluorescence staining was assessed by evaluating 10 average fields per animal and grading them on a scale from 1–4.

Urinalyses

Urine was collected from one year old transgenic and control mice and analyzed for protein using Multistix10 SG Reagent Strips (Bayer). Clinitek Reagent Strips (Bayer) were used to detect albumin and creatinine.

Results

Generation of transgenic mice over-expressing Bright

Transgenic mice were produced that expressed Bright from the human CD19 promoter previously used to express Btk in a B cell-specific fashion (30). A His-Myc tag was added to the carboxyl terminus of the protein causing the transgenic protein to be of slightly higher molecular weight thereby distinguishing it from the endogenous Bright protein (Figure 1a). The addition of the His-Myc tag did not alter DNA-binding or functional activity of Bright (21,28). Among seven founders produced, three did not breed well or maintain transgene expression in the progeny, and an additional line failed to exhibit robust protein expression. The remaining three lines were initially characterized. Two of those lines, the C line which expressed the highest level of transgenic Bright, and the R line that expressed lower transgenic protein levels compared to endogenous Bright, were chosen for more detailed evaluation (Figure 1b). Western blots of multiple tissues from both of these lines indicated that Bright expression did not occur broadly in tissues that do not contain B lymphocytes. Representative data are shown for line C (Figure 1c) and similar data were obtained for line R.

Figure 1.

Production of mice constitutively expressing Bright in all B lymphocytes. (a) Diagram of transgenic construct with CD19 promoter for B cell-specific expression. (b) A western blot was prepared using 10 µg of total spleen cell extract from two independent transgenic lines (R and C) and non-transgenic littermate controls (LM) and was developed with anti-Bright antibodies. The His-myc-tagged transgenic protein has a higher molecular weight than endogenous Bright. (c) Western blots of whole tissues from the transgenic C line and a littermate control were developed for the presence of transgenic (T) or endogenous (E) Bright. A positive control containing both proteins is shown in lane 1 (Control). (d) B cell subpopulations were isolated from littermate and C transgenic mice by flow cytometry using anti-B220, CD43 and IgM; 50,000 cells per lane were analyzed for Bright protein expression. Levels of transgenic and endogenous Bright protein in individual samples were quantified and are indicated as the ratio of transgenic to endogenous Bright (T/E). Similar results were found with the R line. (e) Splenic B cells were sorted into transitional 1, 2 and 3 (T1, T2, T3), marginal zone (MZ) and follicular (FO) subpopulations and were analyzed for Bright and actin protein expression. Signals were normalized to actin levels and relative Bright transgene expression is shown. All data are representative of at least three experiments.

To determine whether Bright transgene expression was maintained throughout B cell differentiation, consistent with expression of the CD19 promoter, individual B cell subpopulations were sorted and evaluated for transgenic and endogenous Bright expression by western blotting. Western blots of sorted bone marrow B cell subpopulations indicated that the Bright transgene is expressed at the early pro-B cell stage when the CD19 promoter begins to be expressed and is maintained throughout late B cell differentiation in the recirculating B cells that have lost endogenous Bright expression (Figure 1d). Comparison of transgenic versus endogenous Bright within the C transgenics suggests that the transgenic protein is much more abundant than endogenous Bright, particularly in more mature B cells where endogenous Bright levels were previously shown to decrease (23). Similarly, fractionation of splenic B cell subsets (Figure 1e) shows that transgene expression is maintained in splenic B cell subpopulations. Endogenous Bright was not detectable in immature transitional T2, T3 and follicular cells, and expression of transgenic Bright was also particularly low in the follicular B cells. Expression of both endogenous Bright and transgenic Bright was highest in the transitional T1 and marginal zone B cell populations. Relative levels of transgene expression among the different B cell subpopulations was quantified after normalization to actin levels and showed more than a 35-fold difference in transgenic Bright expression (Figure 1e). Immature B cells were previously shown to express increased levels of CD19 relative to mature B cells (33), so some of the variation in transgene expression among the different subpopulations may be the result of increased use of the CD19 promoter in those cell types. These data demonstrate that transgenic Bright protein is consistently expressed at higher levels than the endogenous protein throughout B cell differentiation in the transgenic mice.

Expansion of early B lymphocyte subpopulations in transgenic Bright mice

Flow cytometric analyses of the bone marrow subpopulations (Figure 2a) revealed that total numbers of B220⁺ B lymphocyte lineage cells in the transgenic mice were significantly increased by approximately 1.8 fold in the C line (Table I). However, this increase was not the result of expansion of individual subpopulations, but rather reflected a general increase in B lineage cells. While the R line that expressed lower levels of Bright consistently exhibited an approximately 1.4-fold increase in B lineage cells compared to littermate control values, those increases were not statistically significant by student’s T test. These data suggest that over-expression of Bright in bone marrow B lineage cells leads to expansion of those cells.

Figure 2.

Bone marrow and splenic B cell subpopulations in Bright transgenic mice. (a) Bone marrow from representative Bright C transgenic (C TG) and littermate control (LM) male mice were stained with anti-CD43, anti-IgM and anti-B220. Pro-B (Pro), pre-B (Pre), recirculating (RC) and immature (Imm) B cells were identified by FACs analyses as shown by the boxed areas. (b) Spleen cells from a littermate control (LM) and Bright C transgenic (C TG) mice were stained with anti-CD93 and B220, and the B220⁺ cells were gated into CD93⁺ (R2) and CD93⁻ (R3) cells for analyses of IgM and CD23 levels. Transitional type 1 (T1), 2 (T2) and 3 (T3) subpopulations and marginal zone (MZ) and follicular (FO) B cells were identified. Only T1 cells (LM-2.2% vs. TG-4.6%) and MZ cells (LM-1.9% vs. TG-7.6%) were significantly increased in the transgenics. Data are representative of 14 littermates and 16 transgenic mice.

Table I.

B cell subpopulations are increased in C transgenic bone marrow.

	Littermate	C Transgenic	R Transgenic
Lymphocytes	5.23 × 106 ± 0.50a	9.40 × 106 ± 0.88	6.70 × 106 ± 0.72
(B220+)		(0.0008)b	(0.1114)
Pro-B Cells	0.86 × 106 ± 0.09	1.42 × 106 ± 0.14	1.17 × 106 ± 0.12
(CD43+IgM−B220lo)		(0.0034)	(0.0555)
Pre-B Cells	2.19 × 106 ± 0.23	3.72 × 106 ± 0.30	3.01 × 106 ± 0.53
(CD43−IgM−B220lo)		(0.0006)	(0.1714)
Immature B	0.70 × 106 ± 0.13	1.31 × 106 ± 0.13	0.96 × 106 ± 0.13
(CD43−IgM+B220lo)		(0.0027)	(0.1731)
Recirculating B	0.84 × 106 ± 0.18	1.57 × 106 ± 0.22	1.30 × 106 ± 0.24
(CD43−IgM+B220hi)		(0.0178)	(0.1370)

^aTotal cell numbers per femur ± standard error.

^bp value

LM (N=7), C (N=6), R (N=7)

Thymic T cell numbers did not differ significantly between littermate controls and transgenic lines, as expected due to the CD19-specific expression of the transgene. All mice examined appeared to be generally healthy and had normal sized thymi consistent with absence of stress and illness. In addition, total numbers of CD3 positive T lymphocytes in the thymus (not shown) of the transgenic R and C lines did not differ significantly from numbers observed in the littermate controls.

Examination of splenic lymphocytes by flow cytometry indicated slightly increased numbers of total lymphocytes, including T cells, in the C line relative to the R line and littermate controls (Table II). Individual B cell subsets in the spleen were skewed such that early immature B cells were consistently increased in the Bright transgenic mice (Figure 2b). Marginal zone B cells (identified as IgM^hi, CD93⁻, CD23^−/lo, B220⁺) (34) were significantly increased in the C line by as much as 4-fold, but the R line that expresses lower levels of the transgene did not exhibit increased marginal zone B cells. On the other hand, the transitional type 1 (T1) immature B cell subpopulation was increased significantly in both the R and C lines using several different staining protocols (34–36), both when assessed as total cell numbers and when analyzed as percentages relative to other B cell subpopulations. IgM^hi, IgD^−/lo, CD21^−/lo T1 cell numbers (36) were 3.3 × 10⁶ ± 0.7 in littermate controls, compared to 8.5 × 10⁶ ± 1.5 (p=0.011) for the C line and 5.5 × 10⁶ ± 0.6 (p=0.034) for the R line. Increased percentages of T1 cells relative to T2 cells were easily observed by flow cytometry (Figure 2b). Transitional type 2 and 3 (T2 and T3) immature cell numbers were not significantly different from littermate controls in either transgenic line. Furthermore, limited analyses of a third transgenic line (A) showed 2.4-fold increases in T1 cell numbers and 1.5-fold increases in marginal zone B cells, while other B cell subsets including T2 cells were not different than the littermate controls (not shown). These data indicate that Bright over-expression results in expanded numbers of splenic T1 immature B cells in three independent lines and suggest that marginal zone B cell numbers are also increased in some lines relative to littermate controls.

Table II.

B cell subpopulations are altered in transgenic spleens.

	Littermate	C Transgenic	R Transgenic
B Lymphocytes	4.56 × 107 ± 0.38a	7.34 × 107 ± 0.74	4.36 × 107 ± 0.48
(CD19+)		(0.0081)b	(0.7950)
T Lymphocytes	4.58 × 107 ± 0.40	6.96 × 107 ± 0.98	4.01 × 107 ± 0.64
(CD3+)		(0.0206)	(0.5260)
Follicular B Cells	24.89 × 106 ± 2.63	30.03 × 106 ± 4.34	21.08 × 106 ± 3.38
(IgMlo/−CD93−CD23hiB220+)		(0.3456)	(0.4185)
Marginal Zone B Cells	5.18 × 106 ± 1.00	19.34 × 106 ± 4.52	4.04 × 106 ± 0.50
(IgMhiCD93−CD23−/loB220+)		(0.0075)	(0.3925)
Transitional 1 B Cells	5.93 × 106 ± 0.79	13.56 × 106 ± 2.26	8.88 × 106 ± 0.77
(IgMhiCD93+CD23−/loB220+)		(0.0055)	(0.0269)
Transitional 2 B Cells	8.12 × 106 ± 1.32	8.36 × 106 ± 1.26	10.56 × 106 ± 1.65
(IgMhiCD93+CD23hiB220+)		(0.9051)	(0.2987)
Transitional 3 B Cells	3.75 × 106 ± 0.62	2.86 × 106 ± 0.42	5.01 × 106 ± 0.94
(IgMloCD93+CD23hiB220+)		(0.3103)	(0.3097)

^aTotal cell numbers per spleen ± standard error.

^bp value

LM (N=10), C (N=16), R (N=9)

Splenic architecture is not grossly perturbed in Bright transgenic mice

Spleen sections from unimmunized transgenic and littermate controls were analyzed for marginal zone B cells and follicle formation by staining with anti-IgM and anti-MOMA (37). Figure 3 shows typical splenic architecture in both transgenic and littermate controls. Consistent with the increased numbers of marginal zone B cells identified by FACs analyses in Table II, the C transgenic mice appeared to have slightly larger marginal zone areas (red) relative to normal control sections. However, transgenic mice did not exhibit any gross abnormalities in spleen histology.

Figure 3.

Enhanced marginal zone in Bright transgenic mice. Spleen sections from unimmunized male 12 week-old mice were sectioned and stained for IgM (red) and MOMA (green) to identify B cells and metallophillic macrophages, respectively. Two views of follicles are shown centered (left panels) and off-center (right panels) for a representative littermate control (LM) and a Bright C transgenic line (C TG). Magnification was 25×.

Bright trangenic mice exhibit slight increases in serum Ig levels

Because Bright upregulates immunoglobulin heavy chain transcription in in vitro model systems (20,21), we asked if serum antibody levels were increased in the transgenic mice. Sera were obtained from eight to twelve week-old mice and analyzed by ELISA for IgM and IgG isotypes (Figure 4). IgM levels were abnormally high only in the C line, consistent with the expansion of IgM-secreting marginal zone B cells observed in that line (Table II). Marginal zone B cells account for approximately half of serum IgM levels (37,38). IgG_2a and IgG_2b levels were statistically increased approximately two-fold relative to levels from the littermate controls in both the R and C lines. On the other hand, circulating IgG₃ levels were slightly decreased in the C transgenic line relative to the littermate controls, but not in the R line. The reasons for this apparent decrease are unclear, but xid mice also exhibit reduced levels of IgG₃ (39). IgG₁ levels could not be accurately assessed in the FVB/N mice because anti-isotype antibodies failed to react with the sera from this strain in a titratable fashion, albeit standards and sera from C57Bl/6 mice titered normally. Together, these data indicate that over-expression of Bright results in slight increases in serum Ig levels.

Figure 4.

Transgenic serum immunoglobulin levels are increased relative to control mice. Serum samples from 12 littermate, 12 C and 10 R male transgenic mice (8–12 weeks of age) were measured by ELISA for IgM and IgG antibodies and averaged from duplicate samples at three dilutions. Concentrations were determined using purified antibody controls of each specific isotype to produce standard curves. Error bars are indicated. Asterisks indicate values significantly different than those of the littermate controls. Student’s T test p values are as follows: IgM- C 0.00067, R 0.45; IgG_2a- C 0.000001, R 0.036; IgG_2b- C 0.00000004, R 0.0013; and IgG₃- C 0.02, R 0.75.

Transgenic B cells do not display a hyper-activated phenotype

To determine if the increased Ig production resulted from hyperactivation, or increased susceptibility to activation, of B cells in the transgenic mice, splenic B cells were stimulated for three days in culture with media alone, LPS, CD40 ligand-expressing SF9 insect cells, or SF9 cells alone. Quadruplicate samples from six transgenic mice and six non-transgenic littermates were evaluated for ³H-thymidine incorporation after six hours. None of the unstimulated transgenic cultures exhibited significantly increased ³H-thymidine incorporation relative to the control B cell cultures, suggesting that the transgenic B cells were not constitutively activated. No statistically relevant differences were observed between the transgenic and wild type B cells in their ability to upregulate activation and costimulatory markers (MHC II, CD69, CD40, CD80 or CD86) after treatment with LPS or CD40 ligand (data not shown). Furthermore, supernatants taken from the CD40 ligand- and LPS-activated cells contained similar levels of secreted Ig in both control and transgenic cultures suggesting that the increased levels of serum antibodies observed in the transgenic mice were not due to increased numbers of previously activated, or more easily activated B cells. Therefore, the transgenic B cells do not appear to be hyper-activated compared to wild type controls.

Anti-PC antibodies are over-produced in transgenic mice over-expressing Bright

Although many mouse heavy chain promoters do not contain obvious Bright consensus sites (18,40), the S107 family V1 gene that is used predominantly in responses to PC (25–27) is regulated by Bright and contains Bright binding sites in its 5’ flanking sequence (16). Therefore, we predicted that responses to PC-KLH (keyhole limpet hemocyanin) might be increased in the transgenic mice. Responses to antigen were examined by immunizing six transgenic and six littermate controls with the T-dependent antigens NP-KLH and PC-KLH. Mice were pre-bled and immunized on day 0, boosted and bled at day 7 and bled again at day 14 after the initial immunization. Hapten-specific anti-IgM and anti-IgG responses were measured using either NP-BSA, or PC-BSA coated ELISA plates and results were reported as the last positive dilution where both of the duplicate samples were above background (Figure 5). No significant differences were observed in either preimmune or primary 7 day immune IgG or IgM responses between the littermate and transgenic mice for either hapten. However, the transgenic mice consistently produced more anti-PC IgM antibody (average end dilution of 1: 18,300) at day 14 than did the control littermates (endpoint dilution of only 1:1070) (Figure 5a). Anti-PC IgG antibodies were low in both transgenic and littermate mice. Others have shown that the anti-PC response is almost entirely germline and of the IgM isotype (25). Anti-NP IgM antibodies were slightly lower in the transgenics (1:2200) than in the littermates (1:13,325), but NP-specific IgG at day 14 averaged 1:17,000 in the Bright C line transgenics and 1:13,000 in the littermates, a statistically insignificant difference. These results are consistent with previous data indicating that Bright enhances expression of PC-specific antibodies in vitro and suggest that the increase in humoral antibodies in the Bright transgenic mice does not reflect a global increase in all Ig production.

Figure 5.

Antigen-specific immune responses to phosphorylcholine are increased in Bright transgenic C mice. Sera were collected from preimmune and either PC-KLH (a and c) or NP-KLH (b and d) immunized 8–12 week old control and transgenic mice 7 and 14 days post-immunization (first and second bleeds) and were serially diluted in duplicate and assessed by ELISA for NP-specific and PC-specific IgM (a and b) and IgG (c and d) antibodies using hapten-BSA coated plates. The last dilution in both duplicates above background was taken as the endpoint. Data show average values from seven C transgenic mice and six littermates. Standard error bars are shown. The asterisk indicates significance. (e) S017 V1 mRNA was amplified by RT-PCR from FACs purified T1 and follicular (FO) littermate control and WT Bright transgenic B cells from unimmunized mice. Data for one littermate (LM) and two C transgenic (Tg1 and Tg2) mice are shown. Relative RNA levels per sample were standardized by amplification of GAPDH and are listed numerically below each sample.

To further analyze the repertoire of the transgenic B cells for increases in anti-PC responses, relative levels of the S107 family V1 heavy chain mRNA were examined in FACs purified B cells from both littermate controls and Bright transgenics. The S107 family V1 gene is the predominant heavy chain used against PC (25,41). T1 and follicular (FO) B cells were isolated as shown in Figure 2 and RNA was subjected to semiquantitative RT-PCR analyses for the presence of V1 hybridizing cDNA (Figure 5e). T1 cells from the transgenic mice contained approximately 50-fold more V1 heavy chain mRNA than did the littermate controls. Follicular B cells in the Bright transgenic mice also expressed significantly more (20-fold) V1 mRNA than the littermate controls. It is not clear whether the increase in V1 mRNA reflects an increase in the numbers of cells expressing V1 or in the levels of Ig produced per V1 positive cell. However, these data suggest that expression of Ig from the S107 family is more abundant in the transgenic mice relative to littermate controls even at immature stages of B cell differentiation.

Bright transgenic mice produce ANAs

PC-specific antibodies have previously been associated with autoimmune syndromes and atherosclerotic plaques (42). We therefore sought to determine if the Bright transgenic mice produced other antibodies that were self-reactive. Sera from 8 C and 12 R transgenic mice and 17 littermate controls were analyzed for reactivity with nuclear antigens by staining to Hep-2 cells. Representative positive staining is shown (Figure 6a). One hundred percent of the C transgenic mice examined showed ANA staining of HEp-2 cells at serum dilutions of 1:360. Only samples that resulted in obvious staining at this dilution were considered positive. None of the littermates analyzed were positive for ANA production. ANAs were already evident in the C mice analyzed at four weeks of age. Furthermore, the C transgenic line has now been backcrossed for six generations onto the non-autoimmune C57BL/6 background, and ANA production was retained in three of three young C transgenics on that background. Production of ANAs in the R Bright transgenic line appeared only in older male mice (4–6 months of age) and was present in only three of the 13 R mice examined. Thus, the R Bright transgenic mice that expressed lower levels of transgenic Bright exhibited an intermediate phenotype with respect to ANA production compared to the C transgenic mice that were all positive for ANA production from early ages.

Figure 6.

Bright transgenic mice produce anti-nuclear antigen antibodies. (a) HEp-2 cells were incubated with varying dilutions of sera from littermate or transgenic R and C Bright mice and were developed with anti-mouse IgG FITC. Data shown are representative of typical positive sera samples for each transgenic line at a 1:360 serum dilution. (b) A summary of the data for all mice analyzed is shown graphically. The highest antibody dilution that yielded positive staining as shown in the middle and right panels of (a) is indicated for male littermate controls (LM) and R and C transgenic (TG) mice. Each point represents sera from an individual animal.

Sera from mice that produced ANAs were analyzed by ELISA for activity with double-stranded DNA. Compared to four littermate controls that produced neither ANA nor anti-dsDNA antibody, three of three sera from the R transgenic line reacted with dsDNA and only two of the eleven C sera tested were reactive with dsDNA (not shown). Thus, the nature of the strong ANA response in the C line is likely to be polyclonal.

Over-expression of Bright in old transgenic mice results in kidney sequelae

Autoimmunity often leads to immune complex nephritis, with deposition of IgG antibodies in the kidney with age. Therefore, one year-old male mice were perfused with saline to remove circulating antibodies, and kidney sections were prepared and stained with anti-mouse IgG. Figure 7a shows representative kidney sections from a C transgenic and an age- and gender-matched littermate control stained with the nuclear stain DAPI (blue). Anti-mouse IgG is detected as red. Although the eight littermates tested showed low reactivity with the anti-mouse Ig reagent, twelve of the seventeen Bright transgenic mice (two-thirds) exhibited higher levels of mouse IgG deposition on the kidney basement membranes and/or in the glomeruli (Figure 7b). On the other hand, complement C3 staining was not remarkable in any of the mice examined (not shown). These data suggest that over-expression of the Bright transgene does not result in pathologic immune complexes and kidney malfunction.

Figure 7.

Kidney glomeruli exhibit immunoglobulin deposits in one year old male C transgenic mice. (a) Cryosections of kidney from perfused C Bright transgenics and littermates (LM) were co-stained with DAPI (blue, left panels) to show nuclei and anti-mouse IgG Fab’ antibodies (red, middle panels). Arrows indicate glomeruli with Ig deposits in the transgenic sections. The panel on the far right is an overlay of the two stains. Sections shown are representative of a very positive stain (4+). (b) Graphic representation of the IgG deposits observed in 8 littermate, 17 C transgenic and 6 R transgenic mice ranging in age from 7 to 9 months. Scoring was as follows: 1+ background staining; 2+ low staining; 3+ moderate staining; and 4+ bright staining. Each point indicates an individual animal assessed.

Nevertheless, to determine if IgG deposition was associated with renal impairment in these mice, we assessed urine protein excretion in the transgenic mice. While the R Bright transgenic mice exhibited no differences in urine protein levels compared to the littermate controls (Figure 8), eight of the twelve C Bright transgenic mice (67%) showed total protein levels in the high range. Only one littermate control of sixteen examined, exhibited high range proteinuria. A more in depth analysis of albumin and creatinine levels in the urine by HPLC (not shown) failed to show substantial increases in the majority of the mice. These data are consistent with the failure to find appreciable complement deposition in the kidney and suggest that renal function is not impaired in the Bright transgenic mice. Indeed, at least out to one year of age, mortality of the mice was not affected.

Figure 8.

One year old C Bright transgenic mice exhibited proteinuria. Urine protein levels were measured using Multistix 10 SG reagent strips. Samples were read as indicated. Each point represents an individual animal.

Discussion

Over-expression of the transcription factor, Bright/ARID3a in B lineage cells resulted in increased numbers of B lineage cells in both the bone marrow and periphery, increased levels of serum immunoglobulin and production of ANAs. These effects appeared to occur in a dose dependent fashion in three independent transgenic lines, but we cannot exclude the possibility that some of the variation in phenotype observed among the three lines were due to integration of the transgenes into other loci. Bright was originally identified as a protein that upregulated immunoglobulin heavy chain transcription, so the effects on serum antibody production were not unexpected. However, selective over-production of a subset of antibodies associated with autoimmune syndromes was observed. Intriguingly, immature transitional T1 cells in both lines, and marginal zone B cells in the line with highest transgene expression, were statistically increased relative to other splenic B lymphocytes in the Bright transgenic mice. These data suggest that inappropriate Bright regulation at the immature T1 stage may contribute to the breaches in B cell tolerance observed in the transgenic mice. Marginal zone B cells have also been implicated as a subset containing enriched numbers of autoreactive B cells (43) and the increased numbers observed in Bright transgenic mice may contribute to the autoimmune phenotype observed. These data suggest a previously unappreciated role for Bright in immature B cells and are the first to suggest the importance of a B cell–restricted transcription factor in tolerance maintenance.

Endogenous Bright expression in the B cell lineage is tightly regulated such that transcription in the bone marrow begins at the pro-B to pre-B cell stages, peaks in immature B cells and then is turned off in recirculating IgM⁺ mature B cells (23). Previous experiments indicated that bulk splenocytes also show very low levels of Bright expression, with the highest transcription observed in the PNA⁺ germinal center subpopulation (23). The Bright transgenic mice express the transgene throughout B cell differentiation (Figure 1) and at much higher levels than the endogenous proteins. Individual B cell subpopulations vary in the levels of transgene expression. This may be partially explained by the fact that expression is controlled by the CD19 promoter. CD19 expression varies during B differentiation with particularly high expression in immature B cells (33). Alternatively, protein stability may vary among the different B cell subsets. It may not be surprising that transgenic Bright had no apparent effect on follicular, T2 or T3 cells because endogenous Bright is not expressed in those subpopulations and additional co-factors and gene targets for Bright activity may be inactive in those cells. On the other hand, the splenic B cell populations with the highest Bright transgene expression exhibited the greatest increase in cell numbers suggesting a potential correlation between Bright over-expression and expansion of these cells. Increased numbers of T1 and marginal zone B cells have been correlated with autoimmune disease in other mouse and human systems where the expansion is proposed to be the result of survival of self-reactive cells that would normally undergo apoptosis (44–48). Further experiments will be required to determine the role Bright may play in proliferation and survival of T1 and marginal zone B cells.

Finding increased numbers of T1 cells in the transgenic mice was unexpected, but correlates with others’ studies suggesting an important role for Btk in these cells. Xid mice with a point mutation in Btk exhibit blocks in B cell differentiation at the immature B cell stage, although the block is incomplete. Several groups have demonstrated that B cells from mice deficient in Btk do not proliferate normally at the T1 immature stage and that T2 transitional stage xid B cells fail to respond to signals through the surface immunoglobulin receptor (9,49). Our data, using an in vitro model system, suggested that Btk is critically required for Bright-mediated upregulation of transcription of at least some heavy chains (21,22). We hypothesize that Bright and Btk function coordinately in transitional B cells to mediate further differentiation through this important tolerance checkpoint. Whether Bright activity is mediated through direct effects on Ig heavy chain expression or through other gene targets remains to be seen.

Although serum Igs were increased in the Bright transgenic compared to control mice, immunization with NP-KLH resulted in normal levels of anti-hapten antibodies in vivo. This indicated that increased Bright expression did not uniformly increase all antibody production. Furthermore, stimulation of splenic B cells failed to show increased secretion of antibody in response to LPS or CD40 ligand stimulation, suggesting that transgenic B cells were not simply hyperactive relative to littermate controls. However, ANAs and immune responses to PC were significantly enhanced by over-expression of Bright (Figure 5 and Figure 6). The increased anti-PC responses observed in these mice were not unexpected because the S107 family V1 heavy chain gene predominantly responsible for immune responses to PC (25–27) is regulated by Bright in vitro (21). Consistent with our data, xid mice that lacked functional Btk and were defective in Bright DNA-binding activity (31) exhibited the opposite phenotype showing preferential loss of anti-PC responses relative to other types of antigens (50). Together, our data are consistent with a model whereby Bright is more critically required for expression of the V1 S107 heavy chain gene that predominates the anti–PC response than for other Ig_H genes (51). Anti-PC antibodies can also react with self antigens in atherosclerotic plaques, and over-expression of these antibodies in knock-in mice leads to down-regulation of anti-self cells (42,52–54). Thus, we speculate that Bright may preferentially affect a subset of Ig_H genes. Some of these may normally be self–reactive resulting in the ANA production observed in the Bright transgenic mice. Experiments to formally examine the V_H genes used in these mice for repertoire skewing are in progress.

The mechanisms by which ANAs are produced in Bright transgenic mice are unclear; however, the B cell-restricted expression of the transgene suggests that the effect is intrinsic to B cells. The immature B cell subpopulations expanded in the bone marrow and spleens of wild type Bright transgenic mice are the checkpoints in normal mice where autoreactive B cells are either deleted, become anergic or undergo receptor editing effectively removing self–reactive Ig (9,10). Because the ANA specificity is variable in these mice (only some produce anti-dsDNA antibodies) it has not been possible to directly correlate the ANA production with either the T1 or marginal zone B cell populations. However, we hypothesize that ANA production in the Bright transgenic mice is the direct result of breaks in tolerance at the T1 checkpoint. Accumulation of immature B cells in Bright transgenic mice could be the result of failure of immature B cells to undergo apoptosis, increased survival and migration from the bone marrow, increased proliferation of these cells and/or inappropriate responses to signaling through the sIg receptor. None of these are mutually exclusive. The underlying mechanism by which overexpression of Bright leads to this phenotype may be the direct result of upregulation of Ig_H transcription and surface Ig density at a specific stage of B cell differentiation. Several studies indicate that B cell receptor density is critical for signaling cell survival versus apoptosis (55,56). However, flow cytometry data do not show appreciable increases in IgM levels in transgenic B cells relative to control B cell populations, nor were surface IgG levels appreciably different (not shown). Alternatively, Bright could act through the regulation of other non–Ig gene targets. Although no additional targets of Bright have been experimentally identified, microarray data indicate several anti–apoptotic genes may be regulated by Bright (our unpublished results).

Transgenic mice expressing wild type Bright were originally produced as controls for other transgenic mice that express a dominant negative (DN) form of Bright (28). Because Bright associates with Btk, over-expression of Bright might have sequestered Btk such that it could not function properly leading to an additional Btk-deficient phenotype. However, the phenotype of the wild type Bright transgenic mice differs from Btk-deficient mice and from DN Bright transgenic mice using the identical CD19 promoter for transgene expression. Intriguingly, the DN Bright transgenic mice down-regulate transgene expression at the T1 to T2 transition, consistent with an important function for Bright in immature B cells (our unpublished results). On the other hand, wild type Bright transgenic mice, with B cells that maintained Bright expression continuously throughout B cell development, developed ANAs early in life and exhibited an autoimmune phenotype similar to that observed in Sle1 congenic mice (13). Sle1 congenic mice produce ANAs, but do not develop disease without expression of additional lupus-associated genes (15,57). Together, these data strengthen the hypothesis that Bright plays an important role in maintaining B cell tolerance at early stages before production of autoreactive antibodies.

Bright transgenic mice represent a new category of autoimmune mice. Although over-expression of cytokines such as BAFF can result in autoimmunity (58,59), in most autoimmune models that result in autoantibody production, cell surface receptor or intracellular signaling molecules, such as CD22, CD19, SHP-1 and Lyn, have been shown to play important roles (6,60–62). The Bright transgenic mice over-express a downstream target of these signaling processes, and represent the first transgenic model where an intrinsic B cell transcription factor capable of regulating endogenous Ig_H genes directly results in ANA production. Therefore, determining which checkpoints are overcome in these mice may lead to a better understanding of the initial events that are disrupted in autoimmune disease.

More recently, new data has suggested that B cells do not merely act as the end producer of antibodies in autoimmune disease, but they may also be important regulators of earlier events involving T cells. Using a diabetes mouse model, Tian et al., (63) presented data that suggested B cells, rather than other antigen-presenting cell types, are responsible for the spreading of T cell responses during type I diabetes. In another study, activated B cells from NZB mice that expressed high levels of CD86 induced hyper-responsiveness in CD4⁺ T cells from normal Balb/c mice (64). In this same study, B cells from tolerized mice induced T cell anergy in vitro (64). Others showed that B cells controlled T cell help and autoimmunity in lupus prone NZB/W mice (65). We showed previously that Bright is expressed in activated mature B cells in response to a number of stimuli (31). Therefore, it is interesting to hypothesize that over-expression of Bright in the transgenic mice may result in B cells preferentially capable of stimulating T cellmediated autoimmune responses. Further experiments to characterize T cell-mediated functions in the Bright transgenics are in progress.

While the data presented here expand the potential roles for Bright in the mouse, very little is known regarding the function of Bright in human B lymphocytes. Notably, while Bright activity is not detectable in peripheral blood B cells from normal individuals, peripheral blood from lupus patients is an abundant source of Bright protein and was used as a source of RNA for cloning human Bright (24). Experiments using peripheral blood cells from lupus patients also show an expansion of cells that resemble T1 transitional B cells (46,66). Furthermore, preliminary data suggest that Epstein Barr virus, sometimes causally linked with lupus, can induce Bright expression in human cells (67). ANA production in lupus and rheumatoid arthritis patients has been shown to occur years before disease onset (47,68,69). Understanding more about how Bright regulation is related to ANA production may lead to new insights into human autoimmune disease.

Abbreviations

V_H

variable heavy chain

Btk

Bruton’s tyrosine kinase

ANAs

anti-nuclear antibodies

PC

phosphorylcholine

KLH

keyhole limpet hemocyanin

DN

dominant negative