FBXO11 inactivation leads to abnormal germinal-center formation and lymphoproliferative disease

Abstract

The BCL6 proto-oncogene encodes a transcriptional repressor that is required for the germinal center (GC) reaction and is implicated in lymphomagenesis. BCL6 protein stability is regulated by F-box protein 11 (FBXO11)-mediated ubiquitination and degradation, which is impaired in ∼6% of diffuse large B-cell lymphomas that carry inactivating genetic alterations targeting the FBXO11 gene. In order to investigate the role of FBXO11 in vivo, we analyzed GC-specific FBXO11 knockout mice. FBXO11 reduction or loss led to an increased number of GC B cells, to an altered ratio of GC dark zone to light zone cells, and to higher levels of BCL6 protein in GC B cells. B-cell receptor-mediated degradation of BCL6 was reduced in the absence of FBXO11, suggesting that FBXO11 contributes to the physiologic downregulation of BCL6 at the end of the GC reaction. Finally, FBXO11 inactivation was associated with the development of lymphoproliferative disorders in mice.

Figure 1

Conditional targeting of FBXO11 in mice. (A) Schematic representation of the FBXO11 targeting strategy. Indicated are restriction sites as well as expected fragment sizes detectable by Southern blot in the WT allele as well as in the targeted allele before and after Flp-mediated deletion of the neomycin-resistance cassette. CL, conditional allele; TV, targeting vector. (B) Southern blot analysis of BamHI digested genomic DNA displays the 15-kb WT band along with the 10-kb band after correct targeting in ES cells using a 5′ probe (left). Southern blot analysis of StuI-digested DNA displays the 20-kb WT band along with the 8.5-kb band after correct targeting in ES cells using a 3′ probe (right). (C) Double immunofluorescence staining of FBXO11 (green) and BCL6 (red) of a GC in FBXO11^+/+-Cγ1cre^tg/+ (WT) and FBXO11^fl/fl-Cγ1cre^tg/+ (KO) mice.

Figure 2

FBXO11 deletion leads to an increase in GC B cells and affects the DZ/LZ ratio. (A) Flow cytometry analysis of splenic mononuclear cells from age-matched FBXO11^+/+-Cγ1cre^tg/+ (WT), FBXO11^fl/+-Cγ1cre^tg/+ (HET), and FBXO11^fl/fl-Cγ1cre^tg/+ (KO) for lectin PNA and the B-cell marker B220. Mice were immunized with SRBCs 10 days prior to the analysis. (B) Percentages of B220⁺/PNA⁺ splenic GC B cells. n = number of analyzed mice, average ± standard deviation [SD]; *P < .05 (1-way ANOVA). (C) Average GC size (left), numbers (middle) and GC/spleen area ratios obtained by staining with anti-PNA at least 3 independent spleen sections for each mouse. Pix, pixels. n = number of analyzed mice, average ± SD; *P < .05 (1-way ANOVA). (D) Representative flow cytometry analysis of splenic GC DZ and LZ cells as defined by CXCR4 and CD86 expression. (E) DZ/LZ cell ratios determined as in panel D in FBXO11 WT, HET, and KO mice. n = number of analyzed mice, average ± SD; *P < .05 (1-way ANOVA). (F) DZ/LZ cell ratios in Iμ-HA-BCL6 mice and WT littermates. n = number of analyzed mice, average ± SD; *P < .05 (1-way ANOVA).

Figure 3

FBXO11 deletion increases BCL6 protein levels in vivo. (A) Immunoblot analysis of FBXO11, BCL6, and actin in GC B cells isolated from spleens of FBXO11^+/+-Cγ1cre^tg/+ (WT), FBXO11^fl/+-Cγ1cre^tg/+ (HET), and FBXO11^fl/fl-Cγ1cre^tg/+ (KO) mice. Protein expression levels were quantitated by densitometric analysis, normalized to actin, and fold changes are displayed relatively to WT. (B) Average BCL6 mean fluorescence intensity measured by flow cytometry analysis and displayed as fold change relative to WT mice. n = number of analyzed mice, average ± SD; *P < .05 (1-way ANOVA).

Figure 4

FBXO11 silencing impairs downregulation of BCL6 upon IgM treatment. Immunoblot analysis of FBXO11, BCL6, and actin protein expression in a Burkitt lymphoma cell line (Daudi) treated with IgM in the presence or absence of FBXO11 expression. Three different shRNAs (#2, top; #5, middle; #7, bottom) were used to silence FBXO11 expression. Data were quantitated by densitometric analysis, normalized to actin levels, and graphed relative to untreated cells (right). A representative experiment is displayed out of 2 to 3 independent assays that were performed for each shFBXO11.

Figure 5

FBXO11 deletion in mice leads to lymphoproliferative disease. (A) Incidence of florid follicular hyperplasia (FFH), lymphoproliferative disease (LPD), and DLBCL in 17- to 18-month-old FBXO11^+/+-Cγ1cre^tg/+ (WT), FBXO11^fl/+-Cγ1cre^tg/+ (HET), and FBXO11^fl/fl-Cγ1cre^tg/+ (KO) mice. N = number of analyzed mice; *P < .05 (1-way ANOVA). (B) Lymph node sections representative of no pathology (normal), FFH, LPD, and DLBCL were stained with hematoxilin/eosin (H&E), B220 (blue) and CD3 (brown), or BCL6 as indicated. Images were acquired using 10× and (for the insets) 20× objectives leading to an overall magnification of ×100 and ×200, respectively.