Author ORCID Identifier

https://orcid.org/0000-0002-7937-531X

Date of Graduation

5-2020

Document Type

Dissertation (PhD)

Program Affiliation

Immunology

Degree Name

Doctor of Philosophy (PhD)

Advisor/Committee Chair

Sean Post, Ph.D.

Committee Member

Scott Evans, M.D.

Committee Member

Marina Konopleva, M.D., Ph.D.

Committee Member

Dean Lee, M.D., Ph.D.

Committee Member

M. James You, M.D., Ph.D.

Abstract

Acute myeloid leukemia (AML) is an often devastating hematologic malignancy with 5-year overall survival lingering near 20%. Acquiring a deeper understanding of molecular underpinnings of leukemogenesis will provide a basis for developing more effective therapeutic strategies for patients with AML.

Here, we identified overexpression of hnRNP K as a recurrent abnormality in a subset (~20%) of AML patients. High levels of this RNA-binding protein associated with inferior clinical outcomes in de novo AML. Thus, to evaluate its putative oncogenic capacity in myeloid disease, we overexpressed hnRNP K in murine hematopoietic stem and progenitor cells isolated from fetal liver cells (FLCs). We revealed that hnRNP K-overexpression alters self-renewal capacity and differentiation potential of these cells in vitro. Such findings were recapitulated in vivo, as murine recipients of hnRNP K-overexpressing FLCs developed fatal myeloproliferative phenotypes.

To elucidate mechanisms by which hnRNP K overexpression causes myeloid neoplasia, we took an unbiased approach utilizing RNA-immunoprecipitation sequencing (fRIP-Seq). Among RNA transcripts interacting with hnRNP K was RUNX1—a pivotal transcriptional regulator of definitive hematopoiesis commonly mutated or translocated in AML. Consensus hnRNP K binding sites were identified in the 5’ UTR and near the 3’ splice site in intron 5-6 of RUNX1. Fluorescence anisotropy studies confirmed these interactions were direct, and abrogated when hnRNP K binding sites within RUNX1 were mutated. Manipulating hnRNP K expression in human cell lines and murine FLCs substantially altered RUNX1 splicing surrounding exon 6. RNA-sequencing of FLCs confirmed these data, exposing RUNX1 as a significantly differentially spliced entity in the context of hnRNP K overexpression. Importantly, the protein product of this spliced product (RUNX1ΔEx6) exhibited disparate transcriptional activity in reporter assays compared to full-length RUNX1. Furthermore, we identified KH3 as the hnRNP K domain most critical for these splicing alterations; deletion of KH3 markedly abrogated hnRNP K overexpression phenotypes in vitro.

In sum, we established hnRNP K as an oncogene in myeloid leukemia that binds RUNX1 RNA, altering its splicing and subsequent transcriptional activity. These findings shed light on a mechanism of myeloid leukemogenesis, paving the way for drug discovery efforts to improve outcomes for patients with this disease.

Keywords

hnRNP K, RNA binding proteins, RNA splicing, AML, myeloproliferative disease, mouse model

Share

COinS