Mutant U2AF1 Expression Alters Hematopoiesis and Pre-mRNA Splicing In Vivo

Abstract

Heterozygous somatic mutations in the spliceosome gene U2AF1 occur in ∼ 11% of patients with myelodysplastic syndromes (MDS), the most common adult myeloid malignancy. It is unclear how these mutations contribute to disease. We examined in vivo hematopoietic consequences of the most common U2AF1 mutation using a doxycycline-inducible transgenic mouse model. Mice expressing mutant U2AF1(S34F) display altered hematopoiesis and changes in pre-mRNA splicing in hematopoietic progenitor cells by whole transcriptome analysis (RNA-seq). Integration with human RNA-seq datasets determined that common mutant U2AF1-induced splicing alterations are enriched in RNA processing genes, ribosomal genes, and recurrently mutated MDS and acute myeloid leukemia-associated genes. These findings support the hypothesis that mutant U2AF1 alters downstream gene isoform expression, thereby contributing to abnormal hematopoiesis in patients with MDS.

Figure 1. Global alterations in pre-mRNA splicing are distinct in human AML patients with U2AF1 mutations

Unsupervised principal component analysis of standardized splicing ratios of expressed 3′ splice sites of cassette and mutually exclusive exons (> 5 reads in ≥ half of samples) in human AML patients [n=6 mutant U2AF1(S34F/Y) (green dots), n=102 normal for spliceosome genes (gray dots), n=1 U2AF1(Q157P) (blue dot), n=1 SF3B1(K700E) (red dot)]. Putative spliceosome genes were used to denote “normal;” see also Table S1.

Figure 2. Generation of doxycycline-inducible U2AF1 transgenic mice

(A) Schematic of doxycycline-inducible U2AF1 transgene system. Human U2AF1(S34F) or U2AF1(WT) cDNA was integrated into a modified Col1a1 locus of mouse ES cells containing the reverse tetracycline transactivator (M2rtTA) expressed from the Rosa26 locus. Addition of doxycycline induces U2AF1 transgene expression. (B) Southern blot of ES cell or transgenic mouse tail genomic DNA using the 5′ Col1a1 probe to detect the integration of U2AF1(S34F) or U2AF1(WT) transgenes. Flip-in of the pBS31′ vector is detected by the appearance of a 4.1 kb band. Frt is the modified frt locus (6.7 kb), and WT is the wild-type locus (6.2 kb) (C) Doxycycline-dose response curve of U2AF1 transgene expression relative to endogenous mouse U2af1 in bone marrow cells measured by RT-PCR followed by pyrosequencing assay (n=2–7). Data is represented as mean (+/−SD). The 625 ppm doxycycline chow (*) was used in all subsequent experiments; ppm, parts per million. See also Figure S1.

Figure 3. Mature hematopoietic cell lineages are altered by in vivo expression of U2AF1(S34F)

(A) Peripheral blood white blood cell counts of mice transplanted with transgenic donor marrow following 1, 6, and 12 months of doxycycline induction of transgene expression (n=5–12). (B) Absolute count of peripheral blood monocytes (Gr-1⁺, CD115⁺), neutrophils (Gr-1⁺, CD115⁻), B cells (B220⁺), and T cells (CD3e⁺) based on WBC and flow cytometry following 1 month doxycycline (n=9–11). (C) Mature lineage distribution of bone marrow cells by flow cytometry for monocytes, neutrophils, B cells, and T cells following 1 month doxycycline (n=9–11). (D) Detection of apoptosis in bone marrow monocytes (Gr-1⁺, CD115⁺) following 5 days of doxycycline by flow cytometry for Annexin V⁺ (left panel, n=9–14) and intracellular phospho-H2AX (right panel, n=8–13). All data are represented as mean (+/−SD). (*p<0.05, **p<0.01, ***p<0.001) MFI, Mean Fluorescence Intensity. See also Figure S2.

Figure 4. Hematopoietic stem-enriched and progenitor cells are increased by in vivo expression of U2AF1(S34F), but have a competitive disadvantage in repopulation assays

(A, B) Myeloid progenitor colony forming CFU-C assays of bone marrow (A, n=9–11) and spleen cells (B, n=5–10) from mice transplanted with transgenic donor marrow following 1 month of doxycycline. (C, D) Flow cytometry for donor-derived common myeloid progenitors (CMP: CD45.1⁻, CD45.2⁺, lin⁻, c-Kit⁺, Sca-1⁻, CD34⁺, Fcγ⁻) in bone marrow (C, n=9–11) and spleen (D, n=5–6) of mice transplanted with transgenic donor marrow following 1 month of doxycycline. (E) Flow cytometry for donor-derived, stem-cell enriched fractions of bone marrow [KLS: CD45.1⁻, CD45.2⁺, lin⁻, c-Kit⁺, Sca-1⁺ cells (n=9–11)] following 1 month of doxycycline. Representative flow cytometry plots of U2AF1(WT)/rtTA and U2AF1(S34F)/rtTA-recipient mice showing c-Kit and Sca-1 antigen expression on CD45.2⁺, lin⁻ gated cells (left panel); quantification of multiple mice (right panel). (F) Intracellular flow cytometry of Ki67⁺ levels in bone marrow KLS cells following 5 days of doxycycline (n=9). (G) Peripheral blood chimerism of primary recipient mice transplanted with equal portions of transgenic (CD45.2⁺) and wild-type (CD45.1⁺/CD45.2⁺) bone marrow cells in competitive repopulation assays. Data is shown as fold change of transgenic (CD45.2⁺ chimerism) from the start of doxycycline (n=5–12). Differences (two-way ANOVA) are indicated at each time point as different from the U2AF1(S34F) genotype by: *wild-type nontransgenic mice, ^ΦrtTA only, and ^εU2AF1(WT)/rtTA. (H) Bone marrow cells from primary recipient mice of the same genotype were pooled equally and transplanted into secondary recipient mice (left panel, n=3–5); likewise, bone marrow from secondary recipients was transplanted into tertiary recipient mice (right panel, n=4–5). Fold change of transgenic cells (CD45.2⁺ chimerism) measured in peripheral blood from 1 month post-secondary and tertiary transplants. All data are represented as mean (+/−SD), except panel C as mean (+/−SEM). (*p≤0.05, **p≤0.01, ***p≤0.001) See also Figure S3.

Figure 5. U2AF1(S34F) expression alters splicing in mouse progenitors cells in vivo

(A) Unsupervised principal component analysis of standardized splicing ratios of expressed 3′ splice sites of cassette and mutually exclusive exons (> 5 reads in ≥ half of samples) in donor-derived bone marrow common myeloid progenitor cells (CMP) sorted from mice transplanted with U2AF1(S34F)/rtTA or U2AF1(WT)/rtTA bone marrow (n=3 pools of 5–7 mice each) following 5 days doxycycline. (B) Pathways enriched (GOseq FDR<0.1) in differentially spliced genes in U2AF1(S34F)/rtTA CMPs. (C) The consensus sequence surrounding the AG dinucleotide of the 3′ splice site of 544 randomly selected junctions not altered by U2AF1(S34F). (D) The 3′ splice site consensus sequence of U2AF1(S34F)-skipped exons (n=156) and (E) of skipped canonical splice junctions and their associated alternative 3′ junctions (n=42) identified by DEXSeq (FDR<0.1). Results are depicted in Logos plots. ss, splice site. See also Figure S4 and Table S2–S4.

Figure 6. Mutant U2AF1 alters splicing in common targets across species

(A) Venn diagram of the overlap of independently discovered mutant U2AF1-induced splice junction changes by DEXSeq analysis (FDR<0.1) in 3 RNA-seq datasets: transgenic mouse CMPs, primary human hematopoietic CD34⁺ cells, and human AML patient samples. *Indicates multiple significant junctions detected within the gene. (B) Correlation plots and r values for junctions identified by overlap of DEXSeq datasets shown in the Venn diagram. One junction per gene is shown. (C) Log2 fold change in junction expression [U2AF1(WT) versus U2AF1(S34F)] for overlapping genes (A) and recurrently mutated genes in MDS and AML identified by Fisher’s meta-analysis (Table 1), plotted in descending order based on CMP results. U2AF1 was excluded due to an inability to differentiate endogenous vs. exogenous U2AF1(WT) transcript in CD34⁺ cells. See also Figure S5 and Tables S5, S6.

Figure 7. Mutant U2AF1 alters splicing in MDS bone marrow cells

(A) Mutually exclusive exons in H2AFY detected by RT-PCR and gel electrophoresis (quantified in right panel); the diagram of the event measured and the gel image (left panel). (B, C) Quantification of alternative splice site utilization (B) and altered cassette exon events (C). All data are represented as mean (+/−SD). (*p≤0.05, **p≤0.01, ***p≤0.001; n=5–6)