ETV6 germline mutations cause HDAC3/NCOR2 mislocalization and upregulation of interferon response genes

Abstract

ETV6 is an ETS family transcription factor that plays a key role in hematopoiesis and megakaryocyte development. Our group and others have identified germline mutations in ETV6 resulting in autosomal dominant thrombocytopenia and predisposition to malignancy; however, molecular mechanisms defining the role of ETV6 in megakaryocyte development have not been well established. Using a combination of molecular, biochemical, and sequencing approaches in patient-derived PBMCs, we demonstrate abnormal cytoplasmic localization of ETV6 and the HDAC3/NCOR2 repressor complex that led to overexpression of HDAC3-regulated interferon response genes. This transcriptional dysregulation was also reflected in patient-derived platelet transcripts and drove aberrant proplatelet formation in megakaryocytes. Our results suggest that aberrant transcription may predispose patients with ETV6 mutations to bone marrow inflammation, dysplasia, and megakaryocyte dysfunction.

Keywords: Hematology; Transcription.

Figure 1. Mislocalization of ETV6 P214L in all peripheral cell types.

(A) Comparison via immunofluorescence microscopy of ETV6 localization in PBMCs from ETV6 P214L and R369Q patients and healthy controls. Representative PBMCS stained for DNA (blue) and ETV6 (green). In cells from healthy controls, ETV6 is concentrated in the nucleus. In contrast, cells from patients carrying ETV6 P214L and ETV6 R369Q demonstrated ETV6 concentrated in the cytoplasm, with scarce staining in the nucleus. Scale bar: 20 μm. Original magnification, ×8 (inset). (B) Nuclear-to-cytoplasmic intensity ratio of ETV6 is significantly reduced in ETV6 P214L PBMCs (unpaired t test, ****P < 0.0001). Fluorescence intensity was calculated in the nucleus and cytoplasm. n = 5 random fields of view used to calculate N/C intensity ratio, representing over 100 cells analyzed in each condition.

Figure 2. HDAC3/NCOR2 complex binds to ETV6 and is mislocalized in ETV6 P214L and R369Q PBMC, resulting in increased acetylation.

(A) Comparison via immunofluorescence microscopy of HDAC3 localization in PBMCs from ETV6 P214L and R369Q patients and healthy controls. Representative PBMCs stained for DNA (blue) and HDAC3 (red). In cells from healthy controls, HDAC3 is concentrated in the nucleus. In contrast, cells from patients carrying ETV6 P214L or R369Q demonstrated HDAC3 concentrated in the cytoplasm, with scarce staining in the nucleus. Scale bar: 20 μm. Original magnification, ×8 (inset). (B) Nuclear-to-cytoplasmic intensity ratio of HDAC3 is significantly reduced in ETV6 P214L PBMCs (unpaired t test, **P = 0.0071). Fluorescence intensity was calculated in the nucleus and cytoplasm. n = 5 random fields of view were used to calculate N/C intensity ratio, representing a minimum of 100 cells analyzed in each condition. (C) Comparison via immunofluorescence microscopy of NCOR2 localization in PBMCs from ETV6 P214L patients and healthy controls. Representative PBMCs stained for DNA (blue) and cofactors listed above (red). In cells from healthy controls, NCOR2 is concentrated in the nucleus. In contrast, cells from patients carrying ETV6 P214L and R369Q demonstrated NCOR2 concentrated in the cytoplasm, with scarce staining in the nucleus. Scale bar: 20 μm. Original magnification, ×8 (inset). (D) Nuclear-to-cytoplasmic intensity ratio of NCOR2 is significantly reduced in ETV6 P214L PBMCs (unpaired t test, *P = 0.0169). Fluorescence intensity was calculated in the nucleus and cytoplasm. n = 5 random fields of view were used to calculate N/C intensity ratio, representing a minimum of 100 cells analyzed in each condition. (E) Immunoprecipitation of ETV6 in healthy control PBMCs demonstrates ETV6 interacting with HDAC3 and NCOR2. α-Tubulin was included to demonstrate equal loading of protein in each sample. Representative image of 2 independent experiments. (F) Transient knockdown of ETV6 by siRNA in healthy control PBMCs results in increased acetylated histone H3 compared with control conditions. Representative image of 2 independent experiments. (G) CRISPR-mediated knockout of ETV6 in cord blood–derived CD34⁺ HSCs drives increased acetylation of histone H3. Representative image of 2 independent experiments.

Figure 3. Cytoplasmic localization of ETV6 in P214L and R369Q patient-derived PBMCs drives upregulation of interferon response genes.

(A) UMAP plot of cells sequenced from healthy controls (blue) and ETV6 P214L patients (red). (B) UMAP plot of cells sequenced from healthy controls (circles) and ETV6 P214L patients (triangles), denoting cell types. (C) Complex Venn diagram of upregulated genes across all 4 cell subsets, identifying a highly specific 22 gene signature of proinflammatory interferon response genes (in red). (D) Dot plot of top differentially expressed interferon response genes, split by ETV6 P214L patients versus age and sex matched controls.

Figure 4. ETV6 P214L–derived platelets and MKs lacking nuclear ETV6 demonstrate transcriptional dysregulation and impaired proplatelet formation.

(A) Proinflammatory interferon response genes are dysregulated in platelet transcripts from ETV6 P214L patients (n = 2) as compared with healthy controls (n = 5) (adjusted P > 0.0001 for all genes displayed). (B) Western blot demonstrating complete knockout of ETV6 protein mediated by CRISPR sgRNA. Representative of 2 independent experiments. Histone H3 loading control is previously shown in Figure 2G, as these experiments were conducted on the same set of knockout samples. (C) Transcriptional dysregulation of ETV6-knockout MK cells, upregulated genes enriched in the interferon response pathway, downregulated genes enriched in cytoskeletal biology. Heatmap displays mean transcript counts of 1 experiment in a series of 3 independent experiments. (D) Proplatelet formation is significantly impaired (Mann-Whitney U test, P < 0.0001, 3 independent experiments) when ETV6 is deleted in cord blood–derived MKs. CD34⁺ cells were differentiated into MKs, and proplatelet-forming cells were scored on day 12–14. More than 300 proplatelet-forming cells were counted per condition.