ZBTB33 is mutated in clonal hematopoiesis and myelodysplastic syndromes and impacts RNA splicing

Abstract

Clonal hematopoiesis results from somatic mutations in cancer driver genes in hematopoietic stem cells. We sought to identify novel drivers of clonal expansion using an unbiased analysis of sequencing data from 84,683 persons and identified common mutations in the 5-methylcytosine reader, ZBTB33, as well as in YLPM1, SRCAP, and ZNF318. We also identified these mutations at low frequency in myelodysplastic syndrome patients. Zbtb33 edited mouse hematopoietic stem and progenitor cells exhibited a competitive advantage in vivo and increased genome-wide intron retention. ZBTB33 mutations potentially link DNA methylation and RNA splicing, the two most commonly mutated pathways in clonal hematopoiesis and MDS.

Keywords: clonal hematopoiesis; hematologic malignancies; hematopoietic stem cells; myeloid neoplasia; somatic driver gene discovery.

Figure 1.

Detection of recurrent somatic mutations in large blood exome sequencing data sets. A, Lollipop plots showing the specific mutations identified in SRCAP, YLPM1, ZBTB33, and ZNF318 in 45,676 exomes from ExAC. Missense mutations are shown as green squares, truncating mutations (including nonsense mutations, frameshift insertions/deletions, and splice-site mutations) are shown as black circles, and in-frame mutations are shown as brown hexagons. BTB, broad-complex, tramtrack, and bric a brac protein–protein interaction domain; NLS, nuclear localization signal; SA, spindle-associated domain. B, Graph comparing the number of mutations identified in specific genes in 45,676 ExAC samples versus in 39,007 samples from the TOPMed cohort. Novel candidate CHIP genes are labeled in red. C–F, Graphs showing the prevalence of mutation in ZBTB33 (C), ZNF318 (D), YLPM1 (E), and SRCAP (F) among individuals from ExAC and TOPMed in different age groups. Error bars represent 95% confidence intervals.

Figure 2.

Identification of mutations in ZBTB33, YLPM1, SRCAP, and ZNF318 in a cohort of 1,206 patients with MDS. A, Graph depicting the number of mutations in potential new CHIP genes identified by targeted exome sequencing of 1,206 patients with MDS. B, Pie chart showing the sex for the 16 cases with ZBTB33 mutations. C, The VAF of each ZBTB33 mutation is plotted as an individual point, with bars representing mean and SEM. D, Lollipop plot showing the specific mutations identified in ZBTB33 relative to ZBTB33's functional domains in 1,206 MDS samples. Missense mutations are shown as green squares and truncating mutations are shown as black circles. BTB, broad-complex, tramtrack, and bric a brac protein–protein interaction domain; NLS, nuclear localization signal; SA, spindle-associated domain. E, Comutation plot showing mutations in MDS-associated genes identified in the 16 cases with ZBTB33 mutations. Genes encoding splicing factors and genes with three or more mutations are shown.

Figure 3.

Expansion of Zbtb33-edited HSPCs in mouse transplant models. A, Schematic of noncompetitive transplant setup. HSPCs from male mice expressing Cas9 were lentivirally transduced with an sgRNA targeting Zbtb33 or a negative control sgRNA targeting a noncoding region and transplanted into lethally irradiated mice (n = 7 per group). B, PB was drawn every 4 to 6 weeks; DNA was extracted, PCR amplified, and sequenced; and the percentage of reads with indels near the CRISPR cut site was measured. For each mouse, the indel percentage at each time point was normalized to that at week 4. Data, mean ± SEM. Prism was used to perform a linear regression for each group of mice and compute whether the slope was significantly nonzero. P = 0.0038 for mice transduced with Zbtb33 sgRNA and P = 0.80 for control sgRNA. C, Schematic of competitive transplant setup. n = 8 recipients. D and E, The percentage of cells expressing RFP or BFP in the CD45.2⁺ (D) or CD45.2⁺ CD11b⁺ (E) PB at each time point, as measured by flow cytometry. Data, mean ± SEM. The ratio of percentage of RFP⁺ to percentage of BFP⁺ cells was calculated for each mouse at each time point, and Prism was used to perform a linear regression and compute whether the slope was significantly nonzero. P = 0.00006 for the CD45.2⁺ PB and P = 0.0002 for the CD45.2⁺ CD11b⁺ PB. F, The percentage of RFP- or BFP-expressing cells in the c-kit⁺–enriched BM 44 weeks after transplant, as measured by flow cytometry. Data are plotted as individual mice (n = 7), with bars representing the mean and SEM. P = 0.019, computed using a two-tailed paired t test.

Figure 4.

Interaction of ZBTB33 with splicing-associated proteins and increased IR upon Zbtb33 loss. A and B, Volcano plots visualizing significant protein interacting partners enriched in the WT ZBTB33-V5 IP compared with control (A) and differentially enriched in the WT ZBTB33-V5 and ZBTB33 R26C-V5 IPs (B). Proteins involved in RNA splicing are colored red. C, Schematic depicting experimental setup for transplant to isolate mouse LSKs for RNA-seq experiment. Thirty-eight weeks after transplant, BM was harvested from recipient mice (n = 5). RFP⁺ and RFP⁻ recipient LSKs were isolated by FACS, followed by RNA extraction and RNA-seq. D, Scatterplot comparing constitutive IR in RFP⁻ LSKs (n = 5) versus RFP⁺ LSKs (n = 5). Axes measure the fraction of mRNAs with spliced introns. Red and blue dots represent introns that met the thresholds for significance (P ≤ 0.05) and effect size (absolute percentage change in isoform usage of ≥10% or log fold change ≥ 2) and were retained less or more frequently, respectively, in RFP⁺ compared with RFP⁻ cells. E, Bar graphs plotting the percentage of significant IR events that were increased (blue) or decreased (red) in SF3B1 single-mutation MDS samples (n = 4) versus healthy controls (n = 3; left) and in ZBTB33/SF3B1 comutation MDS samples (n = 3) versus SF3B1 single-mutation MDS samples (n = 4; right).