B-cell-specific conditional expression of Myd88p.L252P leads to the development of diffuse large B-cell lymphoma in mice

Abstract

The adaptor protein MYD88 is critical for relaying activation of Toll-like receptor signaling to NF-κB activation. MYD88 mutations, particularly the p.L265P mutation, have been described in numerous distinct B-cell malignancies, including diffuse large B-cell lymphoma (DLBCL). Twenty-nine percent of activated B-cell-type DLBCL (ABC-DLBCL), which is characterized by constitutive activation of the NF-κB pathway, carry the p.L265P mutation. In addition, ABC-DLBCL frequently displays focal copy number gains affecting BCL2 Here, we generated a novel mouse model in which Cre-mediated recombination, specifically in B cells, leads to the conditional expression of Myd88(p.L252P) (the orthologous position of the human MYD88(p.L265P) mutation) from the endogenous locus. These mice develop a lymphoproliferative disease and occasional transformation into clonal lymphomas. The clonal disease displays the morphologic and immunophenotypical characteristics of ABC-DLBCL. Lymphomagenesis can be accelerated by crossing in a further novel allele, which mediates conditional overexpression of BCL2 Cross-validation experiments in human DLBCL samples revealed that both MYD88 and CD79B mutations are substantially enriched in ABC-DLBCL compared with germinal center B-cell DLBCL. Furthermore, analyses of human DLBCL genome sequencing data confirmed that BCL2 amplifications frequently co-occurred with MYD88 mutations, further validating our approach. Finally, in silico experiments revealed that MYD88-mutant ABC-DLBCL cells in particular display an actionable addiction to BCL2. Altogether, we generated a novel autochthonous mouse model of ABC-DLBCL that could be used as a preclinical platform for the development and validation of novel therapeutic approaches for the treatment of ABC-DLBCL.

Figure 1

Construction of a conditional Myd88^p.L252P allele. (A) Targeting of the Myd88 locus in C57BL/6N Tac ES cells. The endogenous Myd88 locus was targeted with the linearized vector described in the supplemental Data. The targeted allele before (middle panel) and after Flp-mediated recombination of FRT and F3 sites (bottom panel) is schematically depicted. The Southern blots of BauI, Eco91I, and KpnI digested genomic DNA probed with a 5′, a 3′, and a Neo probe, respectively, are shown below the schematic illustration of the targeting strategy. Positions of restriction sites and probes are shown in the schematic drawing above. (B) Myd88^p.L252P mRNA is expressed upon Cre-mediated recombination in MEFs. Myd88^wt/wt and Myd88^{c-p.L252P/c-p.L252P} MEFs were isolated. RNA was isolated from both cell lines before LentiCre application (Sanger sequencing chromatograms, top and middle panels) and the Myd88 mRNA sequence was determined after reverse transcription. The wild-type sequence was recovered from both cell lines. After LentiCre application and puromycin selection, only the p.L252P sequence could be recovered from Myd88^{c-p.L252P/c-p.L252P} MEFs (Sanger sequencing chromatogram, bottom panel). (C) The Myd88^p.L252P isoform is expressed in Myd88^{c-p.L252P/c-p.L252P} MEFs after LentiCre-mediated recombination. Myd88^wt/wt and Myd88^{c-p.L252P/c-p.L252P} MEFs were LentiCre exposed and puromycin selected, as in (B). Whole-cell lysates were separated on sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and blotted onto polyvinylidene difluoride membranes before Myd88 and β-actin, which served as loading controls, and were visualized by immunoblotting. Both Myd88^wt and Myd88^p.L252P proteins were expressed at equal levels. (D) Conditional LentiCre-mediated Myd88^p.L252P expression leads to p65 Ser-536 phosphorylation. Myd88^wt/wt and Myd88^{c-p.L252P/c-p.L252P} MEFs were transduced with LentiCre and puromycin selected, as in (B). Upon selection, cells were lysed, proteins were separated on SDS-PAGE, and pSer-536 p65 was visualized by immunoblot. SAH, short arm of homology; LAH, long arm of homology; pA, polyadenylation signal sequence.

Figure 2

B-cell–specific Myd88^p.L252P expression drives lymphoproliferation and lymphomagenesis in vivo. (A) B-cell–specific expression of Myd88^p.L252P significantly reduces overall survival in vivo. Kaplan-Meier curves illustrate the overall survival of M-Cd19 mice. M-Cd19 mice display a significantly reduced survival compared with the respective controls (log-rank test). (B) Serial MRI scans in 16-, 34-, and 70-week-old M-Cd19 mice revealed the occurrence of splenomegaly in 70-week-old M-Cd19 mice. (C) M-Cd19 mice display splenomegaly at the time of death. Preterminal M-Cd19 and Cd19 mice were sacrificed and spleen weight was recorded. Bars represent the average (n = 3); error bars represent standard deviations. (D) M-Cd19 mice develop lymphoproliferative disease and occasional lymphoma. The top panels show H&E staining of spleens and livers isolated from C57BL/6 and M-Cd19 mice at the time of death. Although the organ architecture appeared normal in C57BL/6 wild-type mice, the architecture of spleens isolated from M-Cd19 mice was largely disrupted by infiltration of small mature lymphocytes (LPD columns) or large blastoid cells (DLBCL columns). The bottom panel shows the partial and complete disruption of the spleen by infiltrates with high proliferative indices. (E) Immunohistochemical characterization of the liver infiltrates of M-Cd19 mice. Areas of infiltrates morphologically resembling DLBCL (marked with solid triangles) showed a homogeneous staining pattern of B220 and Irf4 positivity, whereas they were negative for Bcl6 and Cd138 (bottom panel). Infiltrated areas of small, mature lymphocytes (marked with open triangles) displayed a more heterogeneous staining pattern of positive and negative cells for B220, Irf4, and Cd138, whereas staining was largely negative for Bcl6 (middle panel). (F) The lymphoma cells infiltrating spleens and livers of M-Cd19 mice displayed a largely nuclear localization of p65, indicating NF-κB activation. Black arrows indicate hepatocytes, black arrowheads indicate cytoplasmic p65 staining in splenic lymphocytes in C57BL/6 wild-type mice, and white arrowheads indicate nuclear p65 staining in lymphoma cells in M-Cd19 mice.

Figure 3

MYD88 mutations and high expression of BCL2 are enriched in ABC-DLBCL. (A) MYD88 mutations are substantially enriched in human ABC-DLBCL. Human DLBCLs were stratified as ABC- or GCB-DLBCL following the Hans algorithm, as depicted in (B). After immunohistochemistry-based stratification, DNA was isolated from tissue sections and subjected to targeted deep sequencing by a multiplex PCR, which covered the ATM, BTK, CD79B, DDX3X, FBXW7, MAPK1, MYD88, NOTCH1, PIK3CA, PIK3CD, PTEN, PTPN6, SF3B1, TP53, and XPO1 genes. MYD88 mutations per se, and particularly the MYD88^p.L265P mutation were substantially enriched in ABC-DLBCL. Similarly, CD79B and PTEN mutations were enriched in ABC-DLBCL. (C) Distribution of MYD88 mutations detected in human DLBCL samples is shown in pie charts. The samples were classified into high and low or negative expression of BCL2, as shown in (D). (E) High protein expression levels of BCL2 are significantly enriched in ABC-DLBCL (Fisher’s exact test).

Figure 4

Combination of B-cell–specific BCL2 overexpression and Myd88^p.L252P expression drives ABC-DLBCL development in vivo. (A) Kaplan-Meier curve illustrates overall survival of M-B-Cd19 mice. M-B-Cd19 mice display a significantly reduced survival compared with the respective controls (log-rank test). (B) M-B-Cd19 mice develop splenomegaly at 25 to 35 weeks. MRI scans and quantification of spleen weights at autopsy are shown (two-tailed Student t test). (C) The architecture of spleens and lymph nodes isolated from M-B-Cd19 mice was completely disrupted by a homogeneous infiltrating population of large lymphoblastoid cells (DLBCL column). Small areas of more heterogeneous lymphoid cell populations consisting of small lymphoid cells were occasionally detectable (LPD column). The livers of M-B-Cd19 mice were diffusely infiltrated by large blastoid cells with DLBCL-like morphology. (D) Variable-diversity-joining–recombination analysis by Southern blot analysis revealed the presence of clonal populations in tumors of M-B-Cd19 mice. Tumors from leukemic Eµ:TCL1 mice were used as oligoclonal controls. The germ line configuration is also present in the WT controls and is depicted as “G.” Asterisks indicate clonal rearrangements. Samples of mice with lymphoma detected by histologic analyses are marked with white triangles (top). (E) Genealogic trees of murine DLBCL isolated from M-B-Cd19 mice. Igh rearrangements were cloned and individual clones were sequenced. Analysis focused on the Ighv segments starting with FR1. Genealogic tree was derived from 32 sequences. The tree-like structures demonstrate ongoing somatic hypermutation. Numbers in circles show number of sequence reads; numbering of mutations is in regard to the germ line sequence of Ighv1-26. (F) The Ki-67 index of DLBCL lesions from M-B-Cd19 mice was significantly higher than that in M-Cd19 mice (P = 2.27 × 10⁻¹², two-tailed Student t test; n = 3 mice per genotype; 20 fields of view per lesion). (G) Lymphoma infiltrates of M-B-Cd19 mice were immunohistochemically analyzed and stained positive for Irf4 and Cd138 and negative for B220 and Bcl6. The Ki-67 stainings are quantified in (F).