Murine Embryonic Stem Cell Plasticity Is Regulated through Klf5 and Maintained by Metalloproteinase MMP1 and Hypoxia.

Hammoud AA1,2,3, Kirstein N1,2, Mournetas V1,2, Darracq A1,2, Broc S1,2, Blanchard C1,2, Zeineddine D3, Mortada M3, Boeuf H1,2.
  • 1Univ. Bordeaux, CIRID, UMR5164, F-33 000 Bordeaux, France.
  • 2CNRS, CIRID, UMR 5164, F-33 000 Bordeaux, France.
  • 3Lebanese University, Beyrouth, Liban.



Mouse embryonic stem cells (mESCs) are expanded and maintained pluripotent in vitro in the presence of leukemia inhibitory factor (LIF), an IL6 cytokine family member which displays pleiotropic functions, depending on both cell maturity and cell type. LIF withdrawal leads to heterogeneous differentiation of mESCs with a proportion of the differentiated cells apoptosising. During LIF withdrawal, cells sequentially enter a reversible and irreversible phase of differentiation during which LIF addition induces different effects. However the regulators and effectors of LIF-mediated reprogramming are poorly understood. By employing a LIF-dependent ‘plasticity’ test, that we set up, we show that Klf5, but not JunB is a key LIF effector. Furthermore PI3K signaling, required for the maintenance of mESC pluripotency, has no effect on mESC plasticity while displaying a major role in committed cells by stimulating expression of the mesodermal marker Brachyury at the expense of endoderm and neuroectoderm lineage markers. We also show that the MMP1 metalloproteinase, which can replace LIF for maintenance of pluripotency, mimics LIF in the plasticity window, but less efficiently. Finally, we demonstrate that mESCs maintain plasticity and pluripotency potentials in vitro under hypoxic/physioxic growth conditions at 3% O2 despite lower levels of Pluri and Master gene expression in comparison to 20% O2.

PMID: 26731538


Supplement: What ‘s new, what ‘s next:

Embryonic Stem Cells (ESCs), derived from the inner cell mass of blastocysts, are pluripotent cells, meaning that they can self-renew and are able to generate all the cell types of the body. Pluripotency is associated with plasticity, which is more related to the property of stem cells to go back and force from immature to more mature states and to change their fates. However, since almost any type of somatic cells can be reprogrammed (even still at a low efficiency) with gene cocktails up to the ESCs states, (eg, the famous iPSCs), we could consider that almost any cell has a certain capability to be or become plastic in a particular environment. This is evidenced by the presence of cancer stem cells in tumors, cell types which are dedifferentiated and have recovered stemness properties. Cancer could be then considered as an escape of normal control of the plastic potential of cells.

ESCs hold great promise for future cell replacement therapies, but these developments need an in-depth knowledge to understand and control the mechanisms of maintenance and exit from the undifferentiated state. The pluripotent state of ESCs can be captured in vitro under specific cell growth conditions, which vary between species, because of the embryonic origin of cells (pre- or post- implantation stages defining respectively the “naïve” and “primed” states of the derived ESCs). In addition, it is now well established that ESCs, maintained in vitro, are in unstable equilibrium, between undifferentiated and committed cell states, a balance which is regulated by extrinsic factors [like LIF (Leukemia Inhibitory Factor) for “naïve” murine ESC (mESC) or FGF2 (Fibroblast growth factor 2) for “primed” human ES cells (hESC)]. Recently new actors involved in maintenance of mESC pluripotency came along like the metalloproteinase MMP1. Produced by feeder cells (murine embryonic fibroblasts), MMP1 degrades the extracellular matrix of mESC, releasing trapped cytokines as the Ciliary Neurotrophic Factor (CNTF), which belongs to the LIF family of cytokine and which displays similar effect on mESCs cell pluripotency.

In this manuscript we have investigated mechanisms involved in mESCs plasticity, that we defined as the capacity of committed cells to respond to LIF and to go back to a more immature state. This mimicks the transition of “primed” to “naïve” state, or at least part of this transition, if we assume that mESCs grown for 2 days without LIF are similar but not identical to “primed” cells (Figure 1).  We set up an in vitro assay which can now be adapted to study impact of any parameter on mESCs. Since we have shown that the expression of the carcinoembryonic antigen-related cell adhesion molecule 1 (Ceacam1) protein (encoded by a “Pluri gene”, Figure 2) is well regulated in the plasticity test, we could take advantage of the Ceacam1 promoter to drive a destabilised EGFP (d2EGFP) protein and establish a knockin Ceacam1-d2EGFP cell line. Such a LIF/ Ceacam1-dependent fluorescent cell line could be a powerful system to perform large screenings of genes (with siRNA librairies) or chemicals (which affect signalling pathways) by flow cytometry, to better characterize actors of cell plasticity. Such a tool will also allow to cell sort EGFPlow  expressing cells after d1 or d2 of LIF withdrawal and to determine the real proportion of reverted cells. It could also help to better identify the heterogeneity of the starting cell population along with “single cell” transcriptomic approaches.

We also reported on a new role of the pleiotropic LIF-induced Klf5 gene in cell plasticity. Klf5 is a transcriptional factor involved in many processes like embryo implantation1, maintenance of mESC pluripotency2,3 and survival and maintenance of intestinal stem cells4–6 to cite few features of this protein in normal contexts. Klf5 is also activated in different types of cancer and, recently, an effect in cell migration was demonstrated through the KLF5/FYN/p-FAK axis in bladder cancer7. It is not yet clear how Klf5 but not JunB regulates the reversion process which is probably not just proliferation-dependant. The mechanism of action of Klf5 in cell plasticity could also account for its role in cancers and particularly in cancer stem cells in which it is highly expressed (our unpublished results).

Unexpectedly, by blocking PI3K pathway with the specific Ly29400, we did not detect any effect on cell plasticity at d1. This shows that plasticity (as defined in our study) is not linked with cell viability and pluripotency (which involves active PI3K). However we depicted a new role of PI3K to control cell fate choices at very early time of differentiation.

Our results obtained with MMP1, a novel actor of cell pluripotency8, is also intriguing since we have observed that MMP1 mimics only partly the LIF effect in the plasticity window. LIF and MMP1-specific pathways should be compared to get a more complete view of  their role in ESC fates. Also the potential impact of MMPs on hESCs is not yet known and desserves future studies.

Our results also allowed to revisit the effect of hypoxia on stemness and cell plasticity. We showed that mESCs maintain their pluripotency at least up to two weeks (this commented work and our unpublished results) and keep their plasticity at 3%O2.

Importantly we demonstrated that maintenance of cell pluripotency at low [O2]:

  • 1) Is less dependent on LIF and Master gene signaling (since expression of Phospho Tyr705-STAT3, gp190, Oct4, Nanog or Sox2 was lower at 3% compared to 20% O2), or
  • 2) Requires much less of these above proteins, or
  • 3) Involves new, not yet identified pathways which synergize with LIF pathway.

We postulate that mESCs have adapted a genetic program when cultivated under 20% O2, and that this one is not necessarily identical to the physiologic one activated in the ICM of blastocysts, in which [O2] is much lower than 20%O2. Comparative transcriptomic and proteomic data, from ESC derived under 3 or 20% O2, should help resolving and consolidating this issue. New stemness genes could also be identified on the way.

Interestingly we have also observed that, under hypoxia, cell clusters are less compact than at 20% O2 in correlation with a decreased expression of the junction adhesion molecule Jam2 protein (our unpublished results, A. Abou Hammoud, PhD Thesis, University of Bordeaux, December 2015). It is tempting to suggest that adhesion and /or cell-cell interaction are modified under hypoxia and this will deserve future investigations. Our work emphasizes the importance to cultivate cells in specific hypoxia chambers without normoxic shocks along the experiment, and to develop new cell culture medium, for exemple rich in anti-oxydants and supplemented with particular cytokines, [as already developed for hematopoetic stem cells9–11], which mimics at best low O2 concentration environment.

To conclude, our results shed  new lights on stemness and cell plasticity helping to better understand and control the potential of ESCs.


1          Sun X, Zhang L, Xie H, et al. Kruppel-like factor 5 (KLF5) is critical for conferring uterine receptivity to implantation. Proc Natl Acad Sci U S A 2012; 109: 1145–50.

2          Aksoy I, Giudice V, Delahaye E, et al. Klf4 and Klf5 differentially inhibit mesoderm and endoderm differentiation in embryonic stem cells. Nat Commun 2014; 5: 3719.

3          Ema M, Mori D, Niwa H, et al. Kruppel-like factor 5 is essential for blastocyst development and the normal self-renewal of mouse ESCs. Cell Stem Cell 2008; 3: 555–67.

4          Ghaleb AM, Nandan MO, Chanchevalap S, Dalton WB, Hisamuddin IM, Yang VW. Kruppel-like factors 4 and 5: the yin and yang regulators of cellular proliferation. Cell Res 2005; 15: 92–6.

5          Kuruvilla JG, Ghaleb AM, Bialkowska AB, Nandan MO, Yang VW. Role of Krüppel-like factor 5 in the maintenance of the stem cell niche in the intestinal crypt. Stem Cell Transl Investig 2015; 2.

6          Nandan MO, Ghaleb AM, Bialkowska AB, Yang VW. Krüppel-like factor 5 is essential for proliferation and survival of mouse intestinal epithelial stem cells. Stem Cell Res 2015; 14: 10–9.

7          Du C, Gao Y, Xu S, et al. KLF5 promotes cell migration by up-regulating FYN in bladder cancer cells. FEBS Lett 2016; 590: 408–18.

8          Przybyla LM, Theunissen TW, Jaenisch R, Voldman J. Matrix remodeling maintains embryonic stem cell self-renewal by activating Stat3. Stem Cells Dayt Ohio 2013; 31: 1097–106.

9          Duchez P, Rodriguez L, Chevaleyre J, et al. Interleukin-6 enhances the activity of in vivo long-term reconstituting hematopoietic stem cells in ‘hypoxic-like’ expansion cultures ex vivo. Transfusion (Paris) 2015; 55: 2684–91.

10        Ivanovic Z. Respect the anaerobic nature of stem cells to exploit their potential in regenerative medicine. Regen Med 2013; 8: 677–80.

11        Anaerobiosis and Stemness, 1st Edition | Zoran Ivanovic, Marija Vlaski-Lafarge | ISBN 9780128005408. (accessed March 11, 2016).