A Novel Interaction between RUNX1 and the Splicing Factor PTBP1 Is Enriched in Leukemia Cells

RUNX1 and CBFβ form a critical hematopoietic transcription factor complex. RUNX1 is important not only for normal blood development, but also in leukemia. In the context of Inversion 16 acute myeloid leukemia (AML), RUNX1 is required for the activity of CBFβ-SMMHC (CM), the fusion oncoprotein generated by a chromosomal rearrangement and driver of this subtype of AML. The respective roles RUNX1 and CBFβ play in transcriptional regulation are well characterized but the role of other interacting partners is less well understood.

To identify novel RUNX1 binding partners, we immunoprecipitated the transcription factor in mouse leukemia cells expressing CBFβ-SMMHC (CM+) and analyzed precipitated proteins by mass spectrometry. One of the proteins identified was the splicing factor Polypyrimidine Tract Binding Protein 1 (PTBP1). PTBP1 is an RNA binding protein that regulates alternative splicing of mRNAs. PTBP1 is required for normal hematopoiesis, including roles in stem cell maintenance, erythroid differentiation, and B-cell selection. However, a role for PTBP1 in leukemia has yet to be described.

We confirmed the association of PTBP1 with RUNX1 via immunoprecipitation/western blot in mouse CM+ leukemia cells and the human Inv(16) cell line, ME-1. Next, we asked if this interaction is unique to inv(16) AML. We found that antibodies against RUNX1, but not isotype control antibodies, immunoprecipitated PTBP1 in a variety of myeloid (Molm-13, MV411, Kasumi-1) and lymphoid (Jurkat, MOLT-4, REH, SEM) human leukemia cell lines, indicating that this interaction occurs in multiple leukemia subtypes.

To identify the region of RUNX1 required for the interaction with PTBP1, we used two deletion mutants found in patients with RUNX1-Familial Platelet Disorder (RUNX1-FPD), RUNX1-R232 and RUNX1-Y287. We observed strong binding of PTBP1 with full length RUNX1, but not with either truncation mutant indicating that the RUNX1 C-terminus is required for interaction with PTBP1.

To determine location of this novel interaction, we used proximity ligation assay (PLA). We found that the RUNX1:PTBP1 association occurred exclusively in the nucleus in both mouse and human leukemia cells. Next, we asked if this association is observed in healthy hematopoietic cells. By PLA, we found that the RUNX1:PTBP1 interaction is significantly increased in mouse leukemic cells compared to wild type lineage depleted (lin-) mouse bone marrow cells. To determine if the decrease of PLA signal in healthy bone marrow cells was due to spatial separation of PTBP1 and RUNX1, we performed immunofluorescence staining. We observed diffuse nuclear staining of both proteins in the majority of both healthy and leukemic cells. These results indicate that the interaction between RUNX1 and PTBP1 is significantly increased in leukemia cells, despite both proteins being expressed in similar patterns in both healthy and leukemic cells.

During our experiments, we noticed large variations in PTBP1 expression levels in different mouse CM+ leukemia samples. To test whether PTBP1 levels correlate with a difference in leukemia stem cell activity, we performed colony assays using equal numbers of leukemia cells from samples with differing PTBP1 expression. We found that samples expressing high levels of PTBP1 resulted in a statistically significant increase in colony number. In addition, the colonies that formed from PTBP1 high samples were notably larger than those formed from PTBP1 low cells, implying a relationship between PTBP1 and leukemia stem cell self-renewal and proliferation. To determine if human AML samples show the same variability in PTBP1 levels, we performed western blot analysis using a panel of nine de novo and relapsed AML samples. Like our findings in mouse leukemia, human samples showed high variability in PTBP1 levels. We are currently addressing whether increased PTBP1 directly promotes leukemia progression.

Collectively, our results demonstrate that PTBP1, a splicing regulator, is a novel RUNX1 interacting partner whose association with RUNX1 is significantly increased in leukemia cells. In addition, our data imply a potential role in leukemia stem cell proliferation, self-renewal and a role for PTBP1 in the splicing of RUNX1 target genes, perhaps generating leukemia specific splicing variants. This work has the potential to identify a new component of the RUNX1 transcriptional complex important for leukemogenesis.