In this study, we identified TCF7 as a regulator in the decision of EML multipotential hematopoietic precursor cell self-renewal and differentiation. We further identified RUNX1 as a partner and effector of TCF7 function. TCF7 and RUNX1 bind to a significantly overlapping set of target genes and likely function coordinately in regulating target genes. In particular, TCF7 and RUNX1 bind to and potentially regulate a network of transcription factors which characterize the gene expression pattern of CD34+ cells. We validated our hypothesis using functional tests.
is a member of the T-cell factor family of transcription factors that are the downstream effectors of the Wnt signal transduction pathways. Wnt signaling inhibits the degradation of beta-catenin protein by preventing its phosphorylation by GSK3 beta. In the absence of phsophorylation of N-terminal serine and threonine residues, beta-catenin accumulates and is translocated into the nucleus where it associates with members of the TCF family of transcription factors, and furnishes them with a transcriptional activation domain 
. Wnt signaling can act in a context dependent manner to either activate or repress transcription 
. The TCF family of transcription factors can also either activate or repress the transcription of genes responsive to Wnt signaling 
Wnt signaling has been implicated in self-renewal of hematopoietic stem cells and the growth of hematopoietic precursors 
although mice with defects in the Wnt signaling pathway continue to develop normal mature cells of the hematopoietic system. Although Tcf7
knockout mice have only been reported to show a defect in thymocyte development 
, there are multiple TCF family members in vivo
which can have redundant function and compensate for the loss of Tcf7
alone in the knock-out mouse model. In the EML system the expression level of other TCF family members are low and none of them displayed as remarkable differential gene expression as observed of Tcf7
. An analysis of the repopulating activity of the subpopulations of cells in Tcf7
−/− mouse bone marrow could be useful in the future to understand TCF7's roles in vivo; but such an analysis can be complicated by the stimuli from multiple signaling pathways in the bone marrow and the stem cell niche.
In addition to TCF7's role in thymocyte development, T-lineage specification and differentiation 
, the present study shows TCF7 plays a role in the EML model of hematopoietic precursor cell differentiation and function; and has a role in gene activation as well as repression, A number of components of the Wnt signaling pathway, such as Lrp5
, were expressed at higher levels in CD34+ than CD34− cells. However, the data presented here suggest that regulation by Wnt molecules is not a significant factor in EML cell growth and differentiation. RNA-Seq showed only low levels of the Wnt receptors frizzled 2, 5 and 7, and, at most, traces of Wnt9a and 10a mRNA in the CD34+ cell population, as well as a lack of mRNA for other known Wnt molecules in these cells. Even the trace amounts of Wnt9a
mRNA seen by RNA-Seq in CD34+ cells could not be detected by RT-PCR. Wnt9a
mRNAs were completely absent from the CD34− cells. EML cells are grown in conditioned medium and it is possible that this medium contributed some Wnt. However, EML cells also grow well in standard medium in the presence of only purified SCF so that Wnt in the conditioned media does not seem to be a necessary factor for the growth of the cells. Finally, when a Tcf/Lef
-GFP reporter was introduced into the EML cells, there was no GFP signal of Wnt induced activity in these cells, although the cells did respond to LiCl or external WNT ligands. We conclude that TCF7 may function in these cells by a pathway operative without autocrine Wnt signaling.
The present study suggests that TCF7 plays a dual role in global gene network by promoting the expression of large number of genes characteristic of self-renewing CD34+ cells (), while repressing genes activated in partially differentiated CD34− state (). The effects of Tcf7
knock-down is a result of a combination of both direct effects of losing TCF7 binding and secondary effects of removal of the short form of RUNX1 or other TCF7 targets. RUNX1 plays multiple roles in early hematopoietic development and, unlike TCF7 or Wnt, is necessary for emergence of hematopoietic stem cells 
. In sea-urchin embryos RUNX1 expression is linked to Wnt activity 
. In EML cells, TCF7 and RUNX1 bind to one another's promoter and thus may regulate each other. The Runx1
gene encodes both short and long isoforms, and these have antagonistic effects. The short isoforms promote maintenance and proliferation of progenitor cells, but the long isoforms promote differentiation and inhibit progenitor cell repopulation 
. Remarkably, knockdown of Tcf7
, caused the short isoforms of RUNX1 to disappear at the protein level, without changing the expression of the long isoforms. Therefore, TCF7, in addition to direct effects on the transcription of individual genes, may prevent differentiation by regulating the relative abundance of RUNX1 isoforms which have opposing effects on differentiation.
There may be additional unknown factors besides TCF7 and RUNX1 that are central for switching between the two cell types. Examination of the TCF7 binding targets whose expression is altered by Tcf7
inhibition showed that a STAT3 motif was one of the most frequently detectable transcription factor binding motifs (Figure S5
). Interestingly STAT3 was one of the transcription factors that are up-regulated in the CD34+ cells, suggesting that increased STAT3 levels might augment TCF7 mediated transcriptional changes in the CD34+ cells. Among other transcription factors whose mRNA levels are higher in CD34+ than in CD34− cells, Bcl9
are known modulators of the TCF7-beta-catenin transcriptional response. Interestingly, analysis of the binding targets of SCL/TAL1 in a stem/progenitor cell line HPC-7 indicates that they largely overlap with TCF7 binding genes (163 out of 243 SCL target genes are in common with TCF7 targets) 
is one of the TCF7 target genes; however, Tcf7
is not among the SCL/TAL1 target genes. Therefore, SCL/TAL1 may be a downstream mediator of TCF7 activation. In another context Bcl3
has been reported to be increased by SCF signaling 
, which may also contribute to the increase in Bcl3
mRNA levels seen in CD34+ as compared to CD34− cells. Overall, these results suggest that a major part of the transcriptional switch between CD34+ and CD34− cells is mediated by a small network of transcriptional mediators, with TCF7 central to the network. Finally, when Tcf7
level is knocked down by shRNA, although the transcription levels of many genes changed as mentioned earlier in this paper, CD34+ RNA is not reduced by the knockdown. Therefore, there may be additional transcription regulatory factors required for the switch and this is a topic of our future investigation.
Many models of regulated stem cell differentiation, such as sperm and egg production, skin regeneration, intestinal cell regeneration, and neural differentiation, involve control of the choice of differentiation versus precursor renewal that depends on contact or signaling to the stem cell from different types of cells in anatomically circumscribed niches. The EML system provides a clear example of a mammalian precursor cell that has the intrinsic ability to produce a quantitatively balanced ratio of renewing versus differentiated progenies. We have found that TCF7 is a regulator of the self-renewal and differentiation switch and further analysis of how it is controlled will be critical for understanding how this important process is regulated.