BCL6 Controls the Expression of the B7-1/CD80 Costimulatory Receptor in Germinal Center B Cells

Cell Lines.

All B cell lines were cultured in IMDM supplemented with 10% FCS, 100 U/ml penicillin, 100 U/ml streptomycin, and 2 mM l-glutamine. The Phoenix viral packaging line was grown in DMEM, 10% FCS, 300 μg/ml Hygromycin, 1 μg/ml Diphtheria Toxin. The 293T fibroblast cell line expressing CD40L or CD8 was maintained in DMEM with 10% FCS and 300 μg/ml Hygromycin. NTera-2 cells were cultured in DMEM containing 10% FCS, penicillin/streptomycin and 2 mM l-glutamine.

Retroviral Infection and FACS^® Sorting.

PINCO-HA-BCL6 and PINCO-HA-BCL6_ΔPEST were constructed by blunt-end ligation of HA-BCL6 or HA-BCL6_ΔPEST sequences (18) into the BamHI site (filled-in) of the PINCO retroviral expression vector (34). The amphotropic packaging cell line Phoenix was transfected with PINCO, PINCO-HA-BCL6, and PINCO-HA-BCL6_ΔPEST by the calcium-phosphate/chloroquine method as described in http://www.stanford.edu/group/nolan/NL-phnxr.html, and viral supernatants were collected 48 h after transfection. For infection, Ramos B cells (5 × 10⁴ /ml) were resuspended in the viral supernatant supplemented with 0.5 μg/ml polybrene, centrifuged for 45 min at 1,800 rpm, and incubated at 37°C for 2 h. Four cycles of infection were performed before cells were placed back in growth medium. Green fluorescent protein (GFP) expression in Phoenix (∼100%) and Ramos cells (10–30%) was analyzed by fluocytometry using a FACSCalibur™ (Becton Dickinson); transduced GFP⁺ Ramos cells were selected by cell sorting using a FACStar™ (Becton Dickinson). Clones were isolated and characterized by Western blot analysis for HA-BCL6 and HA-BCL6_ΔPEST expression using anti-BCL6 (N3; Santa Cruz Biotechnology, Inc.) and anti-HA antibodies (clone 12CA5; Boehringer).

CD40 Stimulation.

Induction of CD40 signaling in the B cell lines was performed either by treatment with agonistic anti-CD40 antibody (G28–5 hybridoma from American Type Culture Collection; 0.5 μg/ml) or by cocultivation with 293T fibroblast monolayers engineered to express the CD40 ligand (CD40L) or CD8 (as negative control). For cocultivation, B cells and 293T cells were mixed at 2:1 ratio in 6 well plates, centrifuged at 1,000 rpm for 5 min, and incubated at 37°C for the indicated interval before harvesting.

Flow Cytometric Analysis.

Ramos cell clones (PINCO, PINCO-HA-BCL6, and PINCO-HA-BCL6_ΔPEST) stimulated with CD40 signaling by cocultivation with 293T (CD40L) fibroblasts were analyzed on a FACSCalibur™ (Becton Dickinson) for surface molecule expression using PE-conjugated anti-human B7.1 (CD80), B7.2 (CD86), FAS (CD95), CD23, and CD54 antibodies (BD Biosciences).

Northern and Western Blot Analysis.

Total RNA was isolated using the Trizol-reagent (GIBCO BRL) and Northern blot analysis was performed by standard methods with P³²-labeled probes for BCL6 and β-actin. The protocol for Western blot analysis of BCL6 expression has been described (18).

Quantitative RT-PCR Analysis.

2 μg of total RNA were reverse transcribed into cDNA by using the First Strand cDNA Synthesis Kit (Invitrogen). Quantitative RT-PCR was performed with the DNA Master SYBR Green I kit reagents in the LightCycler instrument (Roche), and the LightCycler Software 3 (Roche) was used for data analysis. The oligonucleotide sequences used are as follows. For CD80 (CD80): 5′-CAC CTG GCT GAA GTG AC-3′ and 5′-GTC AGG CAG CAT ATC AC-3′; for FAS: 5′-CAA ATG CAG AAG ATG TAG-3′ and 5′-GAT TCA TGA GAA CCT TGG-3′; for CD38: 5′-GCC CAG ACT GGA GAA AGG AC-3′ and 5′-CAA CCA CAG CGA CTG GCT CA-3′.

Transient Transfection and Reporter Gene Assays.

The CD80-TK-CAT reporter construct and its linker-scanning mutants (M1-M18) were kindly provided by L. Glimcher, Harvard University, Boston, MA (35). Transient transfections in NTera-2 cells were performed by calcium-phosphate precipitation using 2 μg of the reporter plasmid, various amounts of the BCL6 expression vector pMT2T-HA-BCL6, and the nuclear factor (NF)-κB c-Rel expression vector (p-MT2T-c-Rel), as indicated in the figure legend. The Renilla luciferase pRLTK vector (0.3 μg) was used as internal control for transfection efficiency. CAT activities were measured after 48 h using the CAT Enzyme Assay System from Promega, while luciferase activity was measured using the Dual-Luciferase Reporter Assay kit (Promega). The expression levels of BCL6 and its mutant were controlled by Western blot analysis using anti-HA antibodies.

Chromatin Immunoprecipitation Assay.

For chromatin immunoprecipitation (ChIP), Ramos cells (∼10⁸ total), or transiently transfected NTera-2 cells (one 10 cm plate per assay, harvested 48 h after transfection) were incubated with formaldehyde (1% final volume) for 10 min at 37°C. Cells were washed in PBS containing protease inhibitors, resuspended in 200 μl SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-Cl, pH 8.0), and incubated on ice for 10 min. After sonication to shear the chromatin (10–20 s for three times at 30% maximum power on ice), cellular debris was pelleted by 10 min centrifugation at 14,000 rpm. Supernatants were recovered and diluted 10-fold in ChIP dilution buffer (0.01% SDS, 1% Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM Tris-HCl, pH 8.0, protease inhibitors [Boehringer Mannheim]); 50 μl of this solution were removed to serve as DNA input control. The diluted supernatants were precleared with 50 μl Protein G-Sepharose beads (50% Slurry) and 20 μg sonicated salmon sperm DNA for 1–2 h at 4°C with rotation. After centrifugation to remove the beads, 2 μg of the respective antibody (anti-BCL6 [N3], anti-c-Rel [C], anti-p50 [NLS]; Santa Cruz Biotechnology, Inc.) were added for 2 h, followed by overnight incubation with 30 μl protein G-Sepharose at 4°C with rotation. In each experiment, one sample with no Ab was included as negative control for nonspecific binding. Pelleted beads were washed sequentially with 1 ml each of the following buffers for 5 min: (a) 0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8.0, with 150 mM NaCl; (b) 0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8.0, and 500 mM NaCl; (c) 0.25M LiCl, 1% NP-40, 1% NaDOC, 1 mM EDTA, 10 mM Tris-HCl, pH 8.0; (d and e) TE pH 8.0. Immune complexes were eluted off the beads by adding 250 μl 1% SDS in 0.1 M NaHCO₃ elution buffer and rotating the sample for 15 min at room temperature twice. To reverse the cross-links, 5 M NaCl was added to the eluates, followed by overnight incubation at 65°C and Proteinase K treatment (20 μg for 1 h at 45°C). DNA was purified by phenol-chloroform extraction, ethanol precipitated and resuspended in 30–50 μl of TE (200 μl for the total input sample). PCR reactions were performed using 2 μl of the above DNA preparations as template and the following conditions: 95°C for 2' and 25–30 cycles at 95°C for 30”, 58°C for 30”, 72°C for 1'. The oligonucleotide sequences were as follows: for the transfected TK-CAT and CD80-TK-CAT reporter genes: T3 primer (Promega) and 5′-GTC ACC TTA ATA TGC GAA GT-3′. For the CD80 promoter region (GenBank/EMBL/DDBJ accession no: XM_002948), three different primer pairs were used: (a) 5′-CCT TAT AGA CAT GAA CAG-3′ and 5′-CAC GGT GCT AGC CTG ACT-3′ for generation of fragment A, containing the M7 sequence (−3128 to –2946 from the transcription initiation start; reference 35); (b) 5′-CAA GAT TGC ACC ACT GCA CT-3′ and 5′-GGG AAC ATA GAT TGG AGG TA-3′ for fragment B (−1281 to −1054; reference 35); (c) 5′-CAA AGG AAC CCT TAG GGT CT-3′ and 5′-GAG CTC AAG CGA TTT TCC CA-3′ for fragment C (+296 to +485). PCR products were resolved on 2% agarose gels and visualized by ethidium bromide staining and/or by Southern blot hybridization using the P³²-labeled PCR product as a probe.

Analysis of BCL6-deficient Mice.

Three BCL6-deficient (Δ/Δ) mice and three heterozygous littermates controls (Δ/+) were immunized intraperitoneally with sheep red blood cells and killed 6 d later. Cell suspensions from spleen were prepared as previously described, stained with appropriate combinations of fluorochrome and/or biotin-labeled monoclonal antibodies and analyzed on a FACStar^Plus™ flow cytometer (Becton Dickinson). The antibodies used were APC-anti-B220, FITC-anti-IgD, PE-anti-CD80, PE-anti-CD23, and PE-anti-CD95 (BD Biosciences).

Constitutive Expression of BCL6 Inhibits CD40-induced Activation of CD80 Gene Expression.

The observation that CD40 down-regulates BCL6 expression (14; unpublished data) raised the possibility that BCL6 may influence the expression of genes induced by CD40. To test this hypothesis, we infected Ramos cells, which represent transformed GC B cells, with a retroviral vector (PINCO; reference 34) that encodes HA-tagged BCL6 to obtain cells that constitutively express exogenous BCL6 since the retroviral promoter is resistant to CD40 induced down-regulation. The PINCO vector also encodes a GFP for efficient selection of transduced cells by FACS^®. Ramos cells were infected with vectors (Fig. 1 a) expressing either the wild-type BCL6 protein (PINCO -HA-BCL6) or a deletion mutant (PINCO-HA-BCL6_ΔPEST) that is still capable of transcriptional repression, but has a longer half life due to deletion of the PEST domains recognized by the ubiquitin-proteasome complex (18). Northern blot analysis of representative clones confirmed that endogenous BCL6 RNAs were down-regulated by anti-CD40 treatment in all transduced clones, while exogenous HA-BCL6 or HA-BCL6_ΔPEST RNAs were constitutively expressed (Fig. 1 b). Western blot analysis of the same cells showed that the BCL6 proteins were constitutively expressed in the BCL-transduced cells.

We then examined the effect of constitutive BCL6 expression on the up-regulation of several known CD40 target genes (CD80, CD86, CD95, CD11a/18, CD54, and CD23; references 36 and 37) by FACS^® analysis of transduced Ramos cells cocultivated with 293T cells expressing CD40L. Fig. 2 a shows that CD40-induced up-regulation of CD80 (mean fluorescence intensity), but not of CD95, was significantly (>50%) inhibited in multiple clones of PINCO-HA-BCL6- and PINCO-HA-BCL6_ΔPEST-transduced Ramos cells. Other CD40 target genes (CD86, CD54, CD11a/18, and CD23) were not significantly affected by BCL6 expression (data not depicted). A quantitative RT-PCR analysis of CD80 and CD95 transcripts also revealed that CD40-induced up-regulation of CD80, but not CD95, was significantly reduced cells constitutively expressing wild-type BCL6 and abrogated in those expressing BCL6_ΔPEST (Fig. 2 b). These results suggest that BCL6 inhibits CD40-induced expression of the CD80 gene.

BCL6 Represses NF-κB–induced CD80 Gene Transcription.

As CD40 signaling up-regulates expression of its target genes primarily by activation of the NF-κB transcription factor complex (38, 39), we examined whether the BCL6 transcriptional repressor can influence NF-κB–induced transcription. For this purpose, we used transient transfection/reporter gene assays in Ntera-2 cells (which lack endogenous NF-kB activity) to test whether BCL6 expression vector could modulate NF-κB–dependent (c-Rel) transcriptional activation of a CAT reporter gene linked to the 5′ enhancer region of the CD80 gene. This region was previously shown to support NF-κB-dependent CD80 gene expression in vivo (35). Fig. 3 a shows that BCL6 strongly inhibits c-Rel-induced transcriptional activation of CD80. This activity represents genuine transcriptional repression because it is significantly reduced with BCL6 deletion mutants lacking either the DNA binding (BCL6ΔZF) or the transcriptional repression (BCL6ZF) domain (8). In an analogous assay, BCL6 was also able to repress NF-κB p65–mediated transcriptional activation of the same reporter gene, as well as a reporter gene containing the B6BS BCL6-binding site (8) linked to a κB site (unpublished data). Together, these results indicate that BCL6 can modulate transcription driven by NF-κB response elements within experimental as well as native (CD80) regulatory sequences.

To localize the BCL6 response element(s) within the CD80 enhancer region, we screened a series of linker-scanning mutants of the CD80-TK-CAT reporter construct (M1-M18; reference 35) for mutants that retained NF-κB–induced activation, but displayed reduced BCL6-mediated suppression. The M9, M10, M16, and M17 mutants were insensitive to activation (Fig. 3 b) and no longer bound NF-κB in vitro (reference 35, and unpublished results). However, the M7 mutant, and in part the m15 mutant, retained NF-κB inducibility, but were significantly resistant to BCL6-mediated suppression. Notably, the M7 mutation affected a DNA site (ggtcaTCTTAGAA) closely resembling the consensus sequence previously identified for BCL6 binding in vitro (TTCYTNGAA; references 8 and 9), while the M15 region contains a slightly degenerated BCL6 binding site (see below). These results indicate that BCL6 can repress NF-κB–induced transcriptional activation of the CD80 enhancer, and that this repression is mediated by sequences distinct from those involved in NF-κB–mediated activation.

BCL6 Binds the CD80 Enhancer Region In Vivo.

To examine whether BCL6 binds the CD80 enhancer in vivo, we performed ChIP assays in Ntera-2 cells cotransfected with vectors expressing c-Rel, p50, BCL6, a control reporter lacking CD80 sequences (TK-CAT) and a reporter gene containing either the wild-type (CD80-TK-CAT) or the M7 mutant (CD80[M7]-TK-CAT) of the CD80 enhancer. Specific binding of BCL6 to the test (region B in the CD80 containing plasmids) versus the control region (region A in TK-CAT) was assessed in the same transfected cells by using the same PCR primers (see Fig. 4 legend). Fig. 4 a shows that c-Rel and p50 can bind specifically to the wild-type CD80 enhancer and that their binding is conserved in the M7 mutant. Conversely, BCL6 can also bind to the wild-type CD80 enhancer, but not to the M7 mutant, consistent with the functional data shown in Fig. 3 b. These results indicate that BCL6 can bind the CD80 enhancer region in vivo and identify the M7 motif as the region of BCL6 binding, which is clearly distinct from that bound by NF-κB factors.

To conclusively demonstrate the physiologic nature of BCL6 binding to CD80 sequences, we examined whether endogenous BCL6 binds the native CD80 enhancer region in vivo in nontransfected Ramos cells. We analyzed by ChIP the binding of BCL6 to the enhancer region containing the sequence identified by the M7 mutant (A: −3128 to −2946) or, as controls, to two other regions of the CD80 gene (B: −1281 to −1054; C: +296 to +485; Fig. 4 b). The results show that BCL6 can bind specifically to region A, but not to the control regions B and C. Analogous experiments showed a weak binding to the m15 region that contains a degenerated BCL6 binding site (not shown). These results demonstrate that the CD80 gene is a direct and physiologic target for BCL6 binding and transcriptional repression in vivo.

Relationship between BCL6 Expression and CD80 Expression in Lymphoid Cells.

The observation that BCL6 controls CD80 expression implies that the expression of these two molecules may be inversely related in B cells. Consistent with this notion, it has already been established that normal and neoplastic GC B cells and GC NHL cell lines, which all express BCL6 (13), express extremely low levels of CD80 (reference 32; and see Discussion). To examine this relationship more precisely, we examined BCL6 and CD80 expression in three pairs of Burkitt's lymphoma lines (KEM, MUTU, ODHI), each derived from the same tumor case, but selected in vitro to represent either a GC phenotype (type I) expressing BCL6, or a post-GC immunoblastic phenotype (type III) in which BCL6 expression is down-regulated (40, 41). RT-PCR analysis shows that BCL6 and CD80 RNA expression are mutually exclusive, and that CD80 RNA and protein are detectable only in type III cells that lack BCL6 (Fig. 5) .

Next, we used flow cytometric analysis to compare CD80 expression in the spleens of mice lacking BCL6 (BCL6Δ/Δ) and their phenotypically normal heterozygous controls (BCL6+/Δ) (17). In BCL6Δ/Δ mice, a direct comparison between purified similar cells differentially expressing BCL6 was not possible since these mice lack GC. Therefore, CD80 expression had to be compared between whole B220⁺ B cell populations. In wild-type mice, most B220⁺ B cells do not express or express very low levels of BCL6, but include 5–10% GC B cells expressing high levels of BCL6 protein; conversely, in BCL6Δ/Δ BCL6 protein whose levels are increased by its longer half-life and inability to negatively autoregulate BCL6 gene transcription (17, 42, 43). Flow cytometric analysis revealed that, even within whole B cell populations, a significantly higher expression of CD80 expression was detectable in the splenic B cells of BCL6Δ/Δ mice both in terms of number of CD80 positive cells (Fig. 6 a: representative analysis of one mouse) and mean levels of expression per cell (Fig. 6 b: summary of results for three mice pairs). Although BCL6 Δ/Δ mice display an inflammatory disease of the spleen (17) that may also contribute to increased CD80 expression, the results are consistent with a role for BCL6 in modulating CD80 expression in vivo. Taken together, these observations confirm that BCL6 and CD80 expression are inversely related in a variety of B cell types, consistent with the notion that BCL6 modulates CD80 expression in vivo.