Proteogenomic analysis reveals cytoplasmic sequestration of RUNX1 by the acute myeloid leukemia–initiating CBFB::MYH11 oncofusion protein
Ryan B. Day, Julia A. Hickman, Ziheng Xu, Casey D.S. Katerndahl, Francesca Ferraro, Sai Mukund Ramakrishnan, Petra Erdmann-Gilmore, Robert W. Sprung, Yiling Mi, R. Reid Townsend, Christopher A. Miller, Timothy J. Ley
Ryan B. Day, Julia A. Hickman, Ziheng Xu, Casey D.S. Katerndahl, Francesca Ferraro, Sai Mukund Ramakrishnan, Petra Erdmann-Gilmore, Robert W. Sprung, Yiling Mi, R. Reid Townsend, Christopher A. Miller, Timothy J. Ley
View: Text | PDF
Research Article Oncology

Proteogenomic analysis reveals cytoplasmic sequestration of RUNX1 by the acute myeloid leukemia–initiating CBFB::MYH11 oncofusion protein

Abstract

Several canonical translocations produce oncofusion genes that can initiate acute myeloid leukemia (AML). Although each translocation is associated with unique features, the mechanisms responsible remain unclear. While proteins interacting with each oncofusion are known to be relevant for how they act, these interactions have not yet been systematically defined. To address this issue in an unbiased fashion, we fused a promiscuous biotin ligase (TurboID) in-frame with 3 favorable-risk AML oncofusion cDNAs (PML::RARA, RUNX1::RUNX1T1, and CBFB::MYH11) and identified their interacting proteins in primary murine hematopoietic cells. The PML::RARA- and RUNX1::RUNX1T1-TurboID fusion proteins labeled common and unique nuclear repressor complexes, implying their nuclear localization. However, CBFB::MYH11-TurboID–interacting proteins were largely cytoplasmic, probably because of an interaction of the MYH11 domain with several cytoplasmic myosin-related proteins. Using a variety of methods, we showed that the CBFB domain of CBFB::MYH11 sequesters RUNX1 in cytoplasmic aggregates; these findings were confirmed in primary human AML cells. Paradoxically, CBFB::MYH11 expression was associated with increased RUNX1/2 expression, suggesting the presence of a sensor for reduced functional RUNX1 protein, and a feedback loop that may attempt to compensate by increasing RUNX1/2 transcription. These findings may have broad implications for AML pathogenesis.

Authors

Ryan B. Day, Julia A. Hickman, Ziheng Xu, Casey D.S. Katerndahl, Francesca Ferraro, Sai Mukund Ramakrishnan, Petra Erdmann-Gilmore, Robert W. Sprung, Yiling Mi, R. Reid Townsend, Christopher A. Miller, Timothy J. Ley

×

Figure 7

RUNX1 is dysregulated in CBFB::MYH11 AML.

Options: View larger image (or click on image) Download as PowerPoint
RUNX1 is dysregulated in CBFB::MYH11 AML. (A) Empty vector (EV; n = 8), ...
(A) Empty vector (EV; n = 8), CBFB::MYH11 (n = 8), or CBFBN104A::MYH11 (n = 4) was retrovirally expressed in murine hematopoietic cells, and RNA-Seq was performed at days 4 and 7. A heatmap of the top 100 DEGs shows that the CBFB::MYH11 transcriptional signature is abrogated by the CBFBN104A mutation. (B) Runx1 and Runx2 are upregulated in CBFB::MYH11 (red) but not CBFBN104A::MYH11 (blue) cells relative to EV cells (gray), suggesting that upregulation may be due to a CBF-sensing feedback loop. One-way ANOVA between all samples, *P < 0.05, **P < 0.01, ****P < 0.0001, nonsignificant comparisons unlabeled. Each point represents an individual sample, bar indicates mean, box indicates 95% confidence interval, whiskers indicate value range. (C) Human TCGA AML RUNX1/2 RNA-Seq data for healthy donor CD34+ (black, n = 3) and CBFB::MYH11 (red, n = 11), RUNX1::RUNX1T1 (green, n = 7), PML::RARA (yellow, n = 16), or NPM1c-mutant AMLs (purple, n = 21). RUNX1 was overexpressed in CBFB::MYH11 samples relative to all other subgroups; RUNX2 was overexpressed relative to CD34+ and NPM1c subgroups. One-way ANOVA between all subgroups, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, nonsignificant/non-CBFB::MYH11 comparisons not shown. (D) Human TCGA AML RUNX1 RNA-Seq data for healthy donor (CD34+; Pro, promyelocytes; Neu, neutrophils; Mono, monocytes; CD3, T cells; CD19, B cells) or AML samples with the indicated oncofusion/mutation. Red squares, biallelic RUNX1 mutations; black dashed line, mean healthy donor cell RUNX1 expression; red dashed line, mean AML RUNX1 expression. All AML samples have upregulated RUNX1 relative to healthy donor cells; AML samples with CBFB::MYH11, RUNX1, or CBFB mutations have mean RUNX1 expression above the AML average. (E) RNA-Seq data for RUNX1/2 from an independent cohort of healthy donor CD34+ (black, n = 3) and CBFB::MYH11 (n = 12, red), RUNX1::RUNX1T1 (n = 9, green), and NPM1c AMLs (n = 13, purple), confirming higher levels of RUNX1/2 expression in CBFB::MYH11 AMLs. One-way ANOVA between all subgroups, *P < 0.05, ***P < 0.001, nonsignificant/non-CBFB::MYH11 comparisons not shown.