A zebrafish model of myelodysplastic syndrome produced through tet2 genomic editing

Abstract

The ten-eleven translocation 2 gene (TET2) encodes a member of the TET family of DNA methylcytosine oxidases that converts 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC) to initiate the demethylation of DNA within genomic CpG islands. Somatic loss-of-function mutations of TET2 are frequently observed in human myelodysplastic syndrome (MDS), which is a clonal malignancy characterized by dysplastic changes of developing blood cell progenitors, leading to ineffective hematopoiesis. We used genome-editing technology to disrupt the zebrafish Tet2 catalytic domain. tet2(m/m) (homozygous for the mutation) zebrafish exhibited normal embryonic and larval hematopoiesis but developed progressive clonal myelodysplasia as they aged, culminating in myelodysplastic syndromes (MDS) at 24 months of age, with dysplasia of myeloid progenitor cells and anemia with abnormal circulating erythrocytes. The resultant tet2(m/m) mutant zebrafish lines show decreased levels of 5hmC in hematopoietic cells of the kidney marrow but not in other cell types, most likely reflecting the ability of other Tet family members to provide this enzymatic activity in nonhematopoietic tissues but not in hematopoietic cells. tet2(m/m) zebrafish are viable and fertile, providing an ideal model to dissect altered pathways in hematopoietic cells and, for small-molecule screens in embryos, to identify compounds with specific activity against tet2 mutant cells.

FIG 1

Genome editing using targeted Fok1 cleavage to generate null alleles of the zebrafish tet2 gene. (A) Site-specific targeting for ZFN-directed Fok1 cleavage within exon 7 of the zebrafish tet2 gene. (B) Alignments of nucleotide sequences from wild-type and mutant tet2 alleles in the Tet2Δ4 and Tet2Δ5 zebrafish lines. ZFN binding sites appear in red. The dashes in the DNA sequence represent the nucleotides that are deleted during repair of Fok1-induced double-strand breaks. (C) Targeted Fok1-induced mutagenic lesions in tet2 produce frameshift mutations that lead to truncated protein products following short regions of novel amino acids, which are indicated in green. (D) The truncated protein products predicted by the mutant tet2 alleles encoding Tet2Δ4 and Tet2Δ5 lack most of the Tet2 catalytic domain. a.a., amino acids.

FIG 2

tet2 loss does not alter α-globin-expressing erythroid cells or cmyb-positive HPSCs during embryogenesis. (A1 to B3) In situ hybridization for α-globin (A1 toA3) at 32 h p.f. and cmyb (B1 to B3) at 36 h p.f. showed no differences in erythroid cells expressing α-globin or HSPCs expressing cmyb in tet2^wt/wt, tet2^wt/m, and tet2^m/m zebrafish embryos. (C1 to C3) o-Dianisidine staining for hemoglobin on 4-day-old tet2^wt/wt, tet2^wt/m, and tet2^m/m embryos showed no differences in circulating erythrocytes. (D1 to D3) Anti-GFP whole-mount immunostaining of the body region (CHT, magnified from the boxed area in the image at the top) of tet2^wt/wt, tet2^wt/m, and tet2^m/m embryos at 5 days p.f. in the Tg(cd41-EGFP) background. (D4) Scatter plot showing the number of cd41-GFP⁺ cells per embryo, counted in the CHTs of 5-day-p.f. tet2^wt/wt (n = 10), tet2^wt/m (n = 10), and tet2^m/m (n = 10) fish. Statistical analysis was performed using the unpaired Student t test. The long horizontal lines denote the mean values. The short horizontal lines above and below the means represent the error bars. (E1 to E6) Whole-mount myeloperoxidase staining of 7-day-old tet2^wt/wt, tet2^wt/m, and tet2^m/m embryos. (E2, E4, and E6) Mpx-expressing cells were counted in the CHT regions of the embryos. (E7) Scatter plot quantifying the mpx⁺ cells per embryo. The long horizontal lines denote the mean values. Statistical analysis was performed using an unpaired Student t test.

FIG 3

Analysis of 5hmC expression in zebrafish embryonic hematopoietic stem cells. (A1 to A7) Whole-mount immunohistochemistry for 5hmC expression on 36-h p.f. wild-type (A2 to A4) and tet2^m/m (A5 to A7) embryos showed no 5hmC expression (red signal) in hematopoietic stem cells expressing cmyb (green signal and white arrows). 5hmC is expressed in other cell types (red signal and yellow arrows) in the ICM. (A2 and A5) Single-scan imaging for GFP (green signal). (A3 and A6) 5hmC (red signal). (A4 and A7) merged GFP (green signal)-5hmC (red signal). (B1 to B7) Whole-mount immunohistochemistry for 5hmC expression on 48-h p.f. wild-type (B2 to B4) and tet2^m/m (B5 to B7) embryos showed 5hmC expression (red signal) in a subset of cells expressing cmyb in the caudal hematopoietic tissue (green signal and white arrows). (B2 to B7) Single-scan imaging for GFP (green signal) (B2 and B5), 5hmC (red signal) (B3 and B6), and merged GFP (green signal)-5hmC (red signal) (B4 and B7). White arrows, GFP-positive cells; yellow arrows, 5hmC-positive cells; orange arrows, GFP-5hmC-positive cells.

FIG 4

tet2 loss leads to reduced 5hmC expression in the kidney marrow in tet2^m/m adult fish. (A1 to A3) Immunohistochemistry for 5hmC expression on paraffin sections from adult fish shows loss of 5hmC expression in the blood cell progenitors of the kidney marrow in the tet2^m/m (A3) not the tet2^wt/wt fish. However, the renal glomerular cells (red outlines) in the same field of tet2^m/m adult fish express 5hmC. tet2 loss does not affect 5hmC expression levels in cell nuclei in brain (B1 to B3), intestine (C1 to C3), and muscle (D1 to D3).

FIG 5

Analysis of tet1, tet2, and tet3 expression in different blood cell types. qRT-PCR in cDNA from the erythrocytes, lymphocytes, progenitor cells, and myelomonocytes isolated from the kidney marrow was performed to examine the expression of each tet family member in tet2^wt/wt, tet2^wt/m, and tet2^m/m adult zebrafish. The expression levels are shown relative to β-actin. For this experiment, kidney marrow from 9 tet2^wt/wt, tet2^wt/m, and tet2^m/m adult zebrafish was drawn in 3 groups, each consisting of 3 animals. The bars represent the averages of triplicate runs, and the error bars represent standard errors of the mean (SEM).

FIG 6

tet2^m/m zebrafish develop premyelodysplasia at 11 months of age. (A) Forward versus side scatter analysis plots for kidney marrow cell populations in 11-month-old tet2^wt/wt, tet2^wt/m, and tet2^m/m fish. (B) Analysis of the kidney marrow cell populations of 11-month-old fish with loss of tet2 showing a significant decrease in erythrocytes and a significant increase in the myelomonocyte population in tet2^wt/m and tet2^m/m fish compared with tet2^wt/wt fish. A significant increase in the progenitor cell population is observed only in the tet2^m/m fish. Ey, erythrocytes; Ly, lymphocytes; Pr, progenitor cells; My, myelomonocytes.

FIG 7

Morphological analysis of blood cell types in the kidneys of tet2^wt/wt, tet2^wt/m, and tet2^m/m 11-month-old fish. Analysis of kidney smears revealed the presence of all hematopoietic lineages in tet2^wt/wt, tet2^wt/m, and tet2^m/m 11-month-old fish. (A1 to A3) May-Grünwald–Giemsa staining of kidney marrow smears for tet2^wt/wt fish shows no defects in maturation or the morphology of the blood cells. (B1 to B3) May-Grünwald–Giemsa staining of tet2^wt/m kidney marrow smears shows normal morphology of mature erythrocytes and mature myeloid and hematopoietic progenitor cells (6 of 10) (B1 and B2), as well as dysplastic myeloid and progenitor cells (4 of 10) (B2 and B3), in a subset of tet2^wt/m fish. The erythrocyte lineage in the tet2^wt/m mutants (4 of 10) (B2 and B3) also shows morphological differences compared with tet2^wt/wt erythrocytes. (C1 to C3) May-Grünwald–Giemsa staining of kidney marrow smears for tet2^m/m erythrocytes (10 of 10) shows dysplasia in the myeloid and progenitor cell lineages; also, the erythrocytes have a dark basophilic cytoplasm compared with tet2^wt/wt erythrocytes. Red arrows, mature erythrocytes; orange arrows, progenitor cells; black arrows, mature myeloid cells; green arrow, eosinophil; asterisks, dysplastic myeloid cells.

FIG 8

Analysis of total blood counts in tet2^wt/wt, tet2^wt/m, and tet2^m/m 11-month-old fish. (A) Representative schematic of the forward versus side scatter analysis plots for the peripheral blood populations in 11-month-old tet2^wt/wt, tet2^wt/m, and tet2^m/m fish. (B) Analysis of the total peripheral blood counts for 11-month-old tet2^wt/wt, tet2^wt/m, and tet2^m/m fish shows no significant differences among the three different genotypes. (C) Comparison of absolute cell numbers for erythrocytes, progenitor cells, myelomonocytes, and lymphocytes in the peripheral blood of 11-month-old tet2^wt/wt, tet2^wt/m, and tet2^m/m fish reveals no differences among the three genotypes. The short horizontal lines represent the error bars, and the long horizontal lines represent the mean values.

FIG 9

tet2^m/m zebrafish develop MDS at 24 months of age. (A) Forward versus side scatter analysis plots for kidney marrow cell populations in 24-month-old tet2^wt/wt, tet2^wt/m, and tet2^m/m fish. Analysis of the kidney marrow cell populations of 24-month-old fish with loss of tet2 shows a significant decrease in erythrocytes and a significant increase in the myelomonocyte population in tet2^wt/m and tet2^m/m fish compared with tet2^wt/wt fish. A significant increase in the progenitor cell population is observed only in the tet2^m/m fish. (B to D) May-Grünwald–Giemsa staining of kidney marrow smears at the 24-month-old stage for tet2^wt/wt (B), tet2^wt/m (C), and tet2^m/m (D) fish shows the presence of dysplastic myeloid and progenitor cells and the presence of the basophilic cytoplasm in a subset of tet2^wt/m fish and all tet2^m/m fish. Red arrows, mature erythrocytes; orange arrows, progenitor cells; black arrows, mature myeloid cells; asterisks, dysplastic myeloid cells.

FIG 10

Analysis of total blood count in tet2^wt/wt, tet2^wt/m, and tet2^m/m 24-month-old fish. (A to D) Comparison of absolute cell numbers for erythrocytes (A), progenitor cells (B), myelomonocytes (C), and lymphocytes (D) in the peripheral blood of 24-month-old tet2^wt/wt, tet2^wt/m, and tet2^m/m fish reveals a decrease in the total number of erythrocytes but not the other blood cell types in the tet2^m/m fish compared with tet2^wt/wt fish. Statistical analysis was performed using an unpaired Student t test. The long horizontal lines denote the mean values. (E) Scatter plot showing the cytoplasm/nucleus ratio of the circulating erythrocytes from 24-month-old tet2^wt/wt and tet2^m/m fish. Statistical analysis was performed using an unpaired Student t test. The long horizontal lines denote the mean values. (F and G) Analysis of the peripheral blood smears by MGG staining reveals a difference in the morphology of the erythrocytes between 24-month-old tet2^wt/wt (F) and tet2^m/m (G) fish.