RUNX1 Isoforms Regulate RUNX1 and Target-Genes Differentially in Platelets/Megakaryocytes: Association with Clinical Cardiovascular Events

RUNX1 is an essential transcription factor for hematopoiesis. Germline RUNX1 haplodeficiency is associated with thrombocytopenia, platelet dysfunction and predisposition to myeloid malignancies. Three major RUNX1 isoforms (A, B and C) are recognized and share a DNA binding RUNT domain; RUNX1A lacks the transactivation domain. Two distinct promoters regulate RUNX1 expression: distal P1 regulates RUNX1C; proximal P2 regulates RUNX1B and RUNX1A. Little is known regarding the differential effects of isoforms in RUNX1 autoregulation, target-gene regulation and association with clinical events. We performed studies in megakaryocytic HEL cells, HeLa cells (which do not express RUNX1) and in platelets of healthy subjects (n=74) and provide evidence for isoform-specific differential autoregulation and target-gene regulation by RUNX1. We show an association between expression of specific RUNX1-regulated genes in whole blood with cardiovascular events in patients.

There are 5 RUNX1 consensus binding sites in P1 promoter and 1 in P2 promoter within ~1000 bp from ATG. Wild type RUNX1 P1 and P2 promoter regions were individually cloned into PGL4 luciferase promoter vector. ChIP studies using PMA-treated megakaryocytic HEL cells showed RUNX1 binding to chromatin regions encompassing the RUNX1 binding sites in both promoters. Mutations of 2 RUNX1 binding sites in intron region of P1 promoter reduced promoter activity; mutations of first 2 RUNX1 binding sites at the exon region of P1 promoter increased activity. Mutation of single RUNX1 binding site in P2 promoter increased promoter activity. Thus, RUNX1 binds to P1 and P2 promoters to regulate activities.

To examine the RUNX1 autoregulation by individual isoforms, we co-transfected each isoform with P1 or P2 promoter vector in HeLa cells, which do not express endogenous RUNX1. In response to RUNX1B over-expression, both P1 and P2 promoters showed a dose-dependent decrease in promoter activity. RUNX1C over-expression increased P1 and P2 promoter activities. Thus, RUNX1B and RUNX1C regulate P1 and P2 promoters differentially. In HEL cells, which have endogenous RUNX1, RUNX1B overexpression decreased RUNX1C and RUNX1A mRNA/protein expression by ~50%. RUNX1C overexpression increased RUNX1B and RUNX1A mRNA/protein expression. Thus, RUNX1B decreases RUNX1C and RUNX1A expression; RUNX1C increases RUNX1B and RUNX1A.

We studied the regulation by RUNX1 isoforms of target genes PCTP (phosphatidylcholine transfer protein), MYL9 (myosin light chain), PDE5A (phosphodiesterase 5A) and F13A (factor XIIIA). In HeLa cells. RUNX1B overexpression increased PCTP, MYL9 and PDE5A (protein and mRNA); RUNX1C overexpression reduced PCTP and MYL9. In HEL cells RUNX1B increased PCTP, MYL9, PDE5A and F13A expression; RUNX1C reduced PCTP and MYL9 expression.

To obtain evidence in vivo, we studied the relationships between RUNX1 isoforms and their relationships to RUNX1-target genes by RNAseq in leukocyte-poor platelets (74 healthy donors; ages 30-75 yrs). RUNX1B correlated negatively with RUNX1C (R= -0.413 and p=1.5e-06) and RUNX1A (R= -0.493 and p=0.00e+oo); RUNX1C correlated positively with RUNX1A (R=0.45, p<0.0001). RUNX1B correlated positively with PCTP, F131A, PDE5A and RAB1B, and negatively with MYL9 (Figure 1); no significant relationships were observed with RUNX1C or RUNX1A.

To understand the relationships to cardiovascular events (MI or death), we explored whole blood gene expression using microarrays in 587 patients (2 cohorts, 190 and 397 patients) presenting to cardiac catherization for chest pain and followed for 3.8 years. In our previous studies RUNX1B transcripts were higher and RUNX1C lower in those with MI or death. We now found that higher expressions of F13A1(Figure 2) and RAB31, which have been shown by us to be RUNX1-targets, were associated with cardiovascular events.

Conclusions: RUNX1 isoforms autoregulate RUNX1 expression differentially and regulate target genes in a differential manner, a pattern consistent with in vivo platelet gene expression. The relative expression of RUNX1B and RUNX1C isoforms and of their target genes associate with clinical cardiovascular events. RUNX1-isoform specific effects need to be considered in clinical scenarios where RUNX1 is involved and in its pharmacologic modulation.