Triplication of a 21q22 region contributes to B cell transformation through HMGN1 overexpression and loss of histone H3 Lys27 trimethylation

Abstract

Down syndrome confers a 20-fold increased risk of B cell acute lymphoblastic leukemia (B-ALL), and polysomy 21 is the most frequent somatic aneuploidy among all B-ALLs. Yet the mechanistic links between chromosome 21 triplication and B-ALL remain undefined. Here we show that germline triplication of only 31 genes orthologous to human chromosome 21q22 confers mouse progenitor B cell self renewal in vitro, maturation defects in vivo and B-ALL with either the BCR-ABL fusion protein or CRLF2 with activated JAK2. Chromosome 21q22 triplication suppresses histone H3 Lys27 trimethylation (H3K27me3) in progenitor B cells and B-ALLs, and 'bivalent' genes with both H3K27me3 and H3K4me3 at their promoters in wild-type progenitor B cells are preferentially overexpressed in triplicated cells. Human B-ALLs with polysomy 21 are distinguished by their overexpression of genes marked with H3K27me3 in multiple cell types. Overexpression of HMGN1, a nucleosome remodeling protein encoded on chromosome 21q22 (refs. 3,4,5), suppresses H3K27me3 and promotes both B cell proliferation in vitro and B-ALL in vivo.

Figure 1. Segmental trisomy orthologous to human chr.21q22 promotes progenitor B cell transformation

(a) Regions orthologous to human chromosome 21 that are triplicated in Ts1Rhr and Ts65Dn mice or amplified in iAMP21 B-ALL. (b) Progenitor B cells (B220+CD43+) and Hardy subfractions as percentages of bone marrow (BM) cells (n=5–7/group in 2 independent experiments). (c) Subfractions from mixed populations in recipient BM 16 weeks after competitive transplantation (n=5/group). (d) B cell colonies across 6 passages (n=3 biological replicates/genotype representative of 3 independent experiments, mean values shown, *P<0.05, **P<0.01), and bright field microscopy of 3 Ts1Rhr and 3 WT passage 2 cultures (scale bar = 4 mm). (e) Myeloid colonies across 4 passages (n=3 mice per genotype; NS, not significant). (f) Leukemia-free survival of recipient mice after transplantation of Eµ-CRLF2 (C2)/Eµ-JAK2 R683G (J2)/Pax5^+/− (P5), with or without Ts1Rhr (Ts1) BM transduced with vector or dominant negative Ikaros (Ik6) (n=8–10 mice/group). (g) Leukemia-free survival of recipient mice after transplantation of BM transduced with BCR-ABL (n=10 mice/group).

Figure 2. Polysomy 21 B-ALL is associated with the overexpression of PRC2 targets

(a) Heat map of human genes orthologous to the 150 most upregulated genes from Ts1Rhr B cells in primary human pediatric B-ALLs (“DS-ALL” represents Down syndrome-associated ALL). Unsupervised hierarchical clustering by gene revealed the “core Ts1Rhr” gene set (boxed). (b) GSEA plots for the full and core Ts1Rhr gene sets in the AIEOP data set. ES, enrichment score. (c) GSEA plot of the core Ts1Rhr gene set in an independent ICH validation cohort. (d) Network enrichment map of MSigDB gene sets enriched (FDR<0.05) in the Ts1Rhr expression signature. (e) Unsupervised hierarchical clustering of H3K27me3-marked genes from the MIKKELSEN_MEF_H3K27me3 gene set in the AIEOP pediatric B-ALL cohort (karyotype shown). (f) GSEA plots of the top 100 genes from three PRC2/H3K27me3 gene sets as defined in the AIEOP patient cohort in the ICH validation cohort. (g) Quantitative histone MS for H3K27-K36 peptides (*P<0.05, n=3 samples per group per genotype).

Figure 3. Ts1Rhr B cells have reduced H3K27me3 that results in overexpression of bivalently marked genes

(a) Gene tracks showing occupancy of histone marks at the Plod2 promoter (one of the 50 core Ts1Rhr genes) in reads per million per base pair (rpm/bp). (b) Levels of H3K27me3 in Ts1Rhr and wild-type B cells at regions enriched for H3K27me3 in wild-type cells (***P<1e-16). (c) Histone marks at the promoters of genes that are upregulated or downregulated in Ts1Rhr vs. wild-type cells (**P< 1e-5). (d) Chromatin marks in wild-type B cells present at promoters of all genes (left) or genes that are upregulated in Ts1Rhr B cells (right, ***P<0.0001 compared to all genes by Chi-square with Yates’ correction). (e) Colony counts in the presence of DMSO or GSK-J4 (n=3–6 biological replicates per genotype, *P<0.05 compared to DMSO for same genotype). (f) Colony counts in the presence of GSK-126 or after withdrawal at passage 5 (*P<0.05 compared to GSK-126 for same genotype, #P<0.05 compared to other genotype or no withdrawal). Arrow indicates GSK-126 withdrawal. (g) Western blotting of passage 2 colonies after 14 total days in culture with DMSO, 1 µM GSK-J4, or 1 µM GSK-126. (h) Western blotting of colonies one passage (7 days) after continuation (+) or removal (−) of GSK-126.

Figure 4. HMGN1 overexpression decreases H3K27me3 and promotes transformed B cell phenotypes

(a) Western blotting of Ba/F3 cells transduced with empty virus or murine HMGN1 (n=3 independent biological replicates). (b) Relative shRNA representation over passages 1–3. Each line represents an individual shRNA (n=155 total). The five shRNAs targeting Hmgn1 are indicated. (c) GSEA plots for the full and core Ts1Rhr gene sets in HMGN1_OE transgenic B cells. (d) B cell colonies during repassaging of WT and HMGN1_OE BM (n=4–5 biological replicates per genotype from two independent experiments, *P<0.05). (e) Leukemia-free survival of recipient mice after transplantation of wild-type or HMGN1_OE bone marrow transduced with BCR-ABL (aggregate of three independent experiments [Supplementary Fig. 8], n=20 [WT] or n=28 [HMGN1_OE] per group, curves compared by log-rank test).