TET1 is a tumor suppressor of hematopoietic malignancy

Abstract

The methylcytosine dioxygenase TET1 ('ten-eleven translocation 1') is an important regulator of 5-hydroxymethylcytosine (5hmC) in embryonic stem cells. The diminished expression of TET proteins and loss of 5hmC in many tumors suggests a critical role for the maintenance of this epigenetic modification. Here we found that deletion of Tet1 promoted the development of B cell lymphoma in mice. TET1 was required for maintenance of the normal abundance and distribution of 5hmC, which prevented hypermethylation of DNA, and for regulation of the B cell lineage and of genes encoding molecules involved in chromosome maintenance and DNA repair. Whole-exome sequencing of TET1-deficient tumors revealed mutations frequently found in non-Hodgkin B cell lymphoma (B-NHL), in which TET1 was hypermethylated and transcriptionally silenced. Our findings provide in vivo evidence of a function for TET1 as a tumor suppressor of hematopoietic malignancy.

Figure 1. Tet1-deficiency drives B cell malignancy upon advanced age

a) Kaplan-Meier survival curve of Tet1-deficient mice with heterozygous (Tet1^+/−, n = 44) and homozygous (Tet1^−/−, n = 78) deletion compared to wild-type mice (Tet1^+/+ n = 28). * P = <0.0005. b) Peripheral blood smears stained with Wright-Giemsa; arrows indicate normal lymphocyte (left panel) and aberrant lymphocyte (right panel). Representative of n = 8–10 mice per genotype. c) Lymph nodes from Tet1^+/+ and Tet1^−/− mice; representative of n = 8–10 mice per genotype. Scale bar = 1cm. d). Flow cytometric analysis of malignant cells in the lymph nodes of moribund Tet1^−/− compared to age-matched Tet1^+/+ mice. Upper panels; side-scatter (SSC) vs. B cells (B220⁺). Middle Panels; B220⁺ gated cells stained for progenitor/precursor B cells (IgM⁻IgD⁻), immature B cells (IgM⁺IgD⁻), transitional B cells (IgM⁺IgD^low) and mature B cells (IgM⁺IgD^high). Lower panels; B220⁺ gated cells stained for progenitor B cell marker CD43 vs. CD19 staining. Data are representative of 3 independent experiments, n = 8–10 mice per genotype. Histological analysis by H&E staining of e) liver, lung, kidney and f) lymph node sections from sick Tet1^+/− and Tet1^−/− mice compared to Tet1^+/+ controls. Arrows indicate infiltration of lymphocytes. C = cortex, M = Medulla; Scale bar = 100 μm in all panels g) Immunohistochemistry of lymph nodes from Tet1^+/+ and sick Tet1^−/− mice stained with anti-Bcl-6, anti-IRF4, anti-CD138 antibodies (brown) and hematoxylin. Scale bar = 100 μm in all panels. Data are representative of n = 4 mice per genotype.

Figure 2. Whole-exome sequencing of Tet1-deficient tumors reveals mutations of B-non Hodgkin’s lymphoma

a) Total number of non-synonymous single nucleotide variants (nsSNVs) and insertions and deletions (Indels) detected by whole exome sequencing in thirteen Tet1-deficient tumor cell populations (T1-13) in order of increasing mutation load. b) Total number of nsSNVs divided into A, T, C or G base substitutions, ordered from left to right in tumors according to increasing total exonic variations. c) The frequency of missense, indel, splicing, ncRNA, and nonsense mutations found in Tet1-deficient tumors and d) the average frequency of nsSNVs that are transition and transversion mutations (mean ± SEM, n = 13) and e) according to base substitution frequency per tumor. f) Average mutation frequency of base substitutions in Tet1-deficient tumors according to trinucleotide context (mean ± SEM, n = 13). Circos plots of g) the most frequently mutated genes and their co-occurrence in Tet1-deficient tumors; red = B-NHL recurrently mutated genes, pink = T11 nsSNVs, blue = indels, black = IgM⁺ tumors, orange = IgM⁻ tumors. Circos plots of mutations and their co-occurrence in h) histone-modifying enzymes and i) histone cluster 1 genes.

Figure 3. TET1 is hypermethylated and transcriptionally down-regulated in B-NHL

a) Methylation profiling by HELP-assay of the TET1 promoter in DLBCL and FL patients compared to normal naïve (NB) and centroblast (CB) B cells. b) mRNA expression analysis of TET1 in B-NHL patients. c) MassARRAY analysis of TET1 CpG methylation in 26 FL patients compared to normal Germinal Center B (GCB) cells and d) RNA-seq of TET1 and TET2 expression in the same 26 FL patients and normal GCB cells. Each circle indicates an individual patient in all panels; * P = <0.05, ** P = <0.005, *** P = <0.0005.

Figure 4. Tet1-deficient hematopoietic stem display increased self-renewal in vivo with a bias toward B cell differentiation

Competitive bone marrow (BM) reconstitution assays. Primary transplants were performed with CD45.2⁺ Tet1^+/+ and Tet1^−/− total BM cells (200,000 per mouse), mixed in equal ratio with CD45.1⁺ support BM cells and transplanted into lethally irradiated recipient mice. 20 weeks post transplant, CD45.2⁺ Tet1^+/+ and Tet1^−/− purified LT-HSCs (500 per mouse) were serially transplanted into lethally irradiated recipient mice with CD45.1⁺ support BM. a) Frequency of donor-derived CD45.2⁺ cells in the peripheral blood (PB) of primary and secondary transplanted mice. Data are the average of two independent experiments (mean ± SEM, n = 3 recipient mice per donor BM), n = 2 donor BM per genotype, per experiment. b) Average frequency of CD45.2⁺Lineage⁺ cells stained for CD11b/Gr1 (M), B220 (B) and CD3 (T) surface markers in peripheral blood 20-weeks post transplant in secondary recipient mice (mean ± SEM, n = 6 mice per genotype). Loss of Tet1 cooperates with Bcl2 overexpression to drive B lymphocytosis in mice. Purified LSK cells from Tet1^+/+ and Tet1^−/− mice were transduced with either pMIG-Bcl2 or an empty vector control retrovirus, and transplanted into lethally irradiated recipient mice (5000 LSKs were injected per recipient with 200,000 wild-type support bone marrow cells). c) Total numbers of GFP⁺ leukocytes, GFP⁺ B lymphocytes and GFP⁺ myeloid cells monitored 4, 8 and 10 weeks post-transplant in the peripheral blood of recipient mice. d) Representative flow cytometric analysis of peripheral blood for CD45.2⁺ GFP⁺ cells, with B220, IgM and IgD staining as indicated (n = 6–12 mice per genotype). Small horizontal lines indicate the mean. * P = <0.01 ** P = <0.001, *** P = <0.0001 in all panels.

Figure 5. Loss of Tet1 in hematopoietic stem cells promotes differentiation with a lymphoid bias

a) Tet1 mRNA expression abundance in long-term hematopoietic stem cells (LT-HSC), multi-potent progenitors (MPP), lymphoid-primed multipotent progenitors (LMPP), common lymphoid (CLP) and myeloid progenitors (CMP), megakaryocyte and erythroid progenitor (MEP), granulocyte and macrophage progenitor (GMP), progenitor B (ProB), precursor B (PreB), Immature B (ImmB), Mature splenic B (MatB), immature myeloid cells in the bone marrow (ImmGM) and mature myeloid cells in the spleen (MatGM) of wild-type mice. Tet1 mRNA expression was normalized to Hprt mRNA (mean ± SEM, n = 3 experiments). b) Loss of Tet1 mRNA expression upon deletion in hematopoietic stem and progenitor cells (Lin⁻cKit⁺ cells) measured by qPCR and normalized to Hprt (mean and SEM, n = 3 experiments). c) Intracellular flow cytometric analysis of 5-hydroxymehtylcytosine (5hmC) in total bone marrow (BM) and B cells from Tet1^+/+ and Tet1^−/− mice (data are representative of 3 experiments). d–f) Frequency of Lin⁻Sca1⁺cKit⁺ (LSK) subsets in old Tet1-deficient mice compared to wild-type mice (mean ± SEM, n = 6–8 mice per genotype). * P = <0.01, ** P = <0.001, *** P = <0.0001. g) Representative flow cytometric analysis of Tet1^+/+ and Tet1^−/− LSKs. Upper panels = LT-HSC (CD150⁺CD48⁻), ST-HSC (CD150⁻CD48⁻), MPP1 (CD150⁺CD48⁺) and MPP2 (CD150⁻CD48⁺); Lower panels = LMPP (Flt3⁺CD34⁺). Data are representative of 6–8 mice per genotype. h) Microarray analysis of LSK cells. Significantly differentially expressed genes (P = <0.05) with changes >2-fold are displayed. GSEA analysis of RNA-seq from purified Tet1^+/+ and Tet1^−/− MPP cells with corresponding heatmaps of leading edge gene expression changes in i) B lineage and j) histone cluster 1 genes.

Figure 6. Aberrant DNA hydroxymethylation of Tet1-deficient stem and progenitor cells

DNA sequencing data was obtained from purified bone marrow LSK cells of 6 month-old Tet1^+/+ and Tet1^−/− mice using HELP-GT (n = 2 per genotype) and RRBS (n = 3 per genotype) assays for quantitation of 5hmC and 5mC, respectively. Graphic representation of the differentially methylated genomic regions are displayed as circos plots for a) 5hmC and b) 5mC with respect to their chromosomal location. Bar graphs depict the overall frequency of c) differential hydroxymethylation (%DhMC) compared to d) differential methylation (%DMC) per chromosome and e) genomic location. Differential hyper- and hypo-methylated sites were calculated for CpG sites with methylation difference cutoff >25%, q-value < 0.01.

Figure 7. Enhanced colony formation and accumulation of DNA damage in Tet1- deficient progenitor B cells

Colony-formation assays in methylcellulose media were performed to measure a) pro-B and b) pre-B self-renewal capacity. Cells were passaged every 7 days for 3 successive weeks (P1-3), (mean ± SEM, n = 3 experiments). Quantitative mRNA expression analysis of pro-B cells from P1 for the relative fold-change in expression of c) histone cluster 1 gene variants and d) DNA repair genes in Tet1^−/− compared to Tet1^+/+ (mean + SEM, n = 3 experiments, P = <0.05). e) Immunofluorescence staining of γH2AX foci in purified ProB cells from Tet1^+/+ and Tet1^−/− mice, scale bar = 10μm. Quantitation of f) the frequency of γH2AX positive cells and the number of foci per positive cell (mean ± SEM, n = 3 per genotype). In all panels: * P = <0.05, ** P = <0.005.