The EphrinB2/EphB4 axis in regulating hESC pluripotency and differentiation

Human embryonic stem cells (hESC) have an inexhaustible ability to divide and renew, and under the appropriate conditions, differentiate and change into any cell type in the body. This balance between pluripotency and self-renewal is a complex and carefully choreographed response of the hESC to local microenvironmental cues. Understanding the molecular regulators of this balance, and the various signals that are integrated by hESC to maintain their pluripotency and self-renewal characteristics are critical for the expansion and differentiation of hESC to specific cell types. EphrinB2 and ephB4 are cell surface molecules that mediate and transduce signaling cascades upon interaction with each other. Cell-cell contacts between ephrinB2 and ephB4 expressing cells provide guidance cues for cell migration and boundary formation in many developmental systems such as the formation of neurons and blood vessels. Importantly, ephrinB2 has been determined to be a molecular marker of “stemness” and is expressed in human embryonic stem cells, neural stem cells and hematopoietic stem cells. Despite much evidence from model systems that ephrinB2/ephB4 axis may be intimately involved in ESC fate (survival, self-renewal, and pluripotency), this particular axis has not been carefully studied in human ESC due to the lack of highly specific reagents to block cognate ephrinB2-ephB4 interactions. Intriguingly, the envelope protein from an exotic and highly lethal virus called Nipah virus, binds ephrinB2 more “tightly” than the EphB4 receptor, and can therefore compete or interfere with normal ephrin-B2-EphB4 interactions. Using an arsenal of reagents based on engineered versions of this viral envelope protein, which retains the ephrinB2 binding properties without the virulence of the actual virus, we had proposed to interrogate the role of the ephrinB2-ephB4 axis in regulating hESC’s ability to proliferate, self-renew, and differentiate into any cell type that make up the human body. hESCs prefer to grow in clusters and propagate as a complex and dynamic ecosystem of cells where any given cell may have different capacity for pluripotency or self-renewal. We first asked if ephrinB2 was homogenously expressed on hESCs, and if not, does ephrinB2 mark for a subpopulation within hESC cultures with distinct properties? To do this, we infected hESCs with GFP-reporter lentiviruses bearing Nipah envelope proteins (NiVpp for NiV pseudotyped particles), which can only infect ephrinB2+ cells. We found that NiVpp consistently infected only 5-20% of hESCs through primary and secondary rounds of infection even though we can purify the initially infected subpopulation to near homogeneity (>85%) between rounds of infection. This suggests that ephrinB2 is not a stable cell surface marker for a distinct subpopulation of cells. However, ephrinB2+ hESCs do appear to represent a subpopulation of hESCs with decreased self-renewal capacity when subjected to the appropriate tests. Interestingly, these NiVpp infected hESCs still maintained the ability to form teratomas, albeit smaller ones, when injected in SCID mice. In toto, our results show that pluripotency and self-renewal are distinct and dissociable properties of hESCs and they do not necessarily reside within one particular cell in the hESC culture. Next we sought to determine whether the ephrin2-EphB4 axis plays a role in regulating the ability of hESCs to differentiate into the three major germ layers that make up the cells of the various organs and tissues in the human body. Differentiation is a carefully choreographed molecular and cellular response to local environmental determinants. In vitro formation of embryoid bodies, where expression of genetic markers for all three germ layers can be detected, is surrogate in vitro assay for pluripotency. Under standard conditions, embryoid bodies form extremely heterogenous spherical clusters that make it difficult to reproducibly quantify any differences in germ layer commitment that might result as a consequence of antagonizing EphrinB2-EphB4 interactions. Thus, we optimized a “spin embryoid body” assay where the number of hESC per embroid body formed could be carefully controlled. Under these conditions, ephrin-B2 expression increased dramatically between days 10-15, closely mirroring the upregulation of ectoderm markers (the germ layer that forms cells like neurons), and to a lesser extent, mesoderm markers (the germ layer that forms cells like endothelial cells). Enoderm markers (the germ layer that forms cells like those that line the gut) are dramatically downregulated during the first 10 days, and do not peak until days 15-20. These exciting results from our first year suggest ephrinB2-EphB4 interactions likely play a role in regulating ectoderm and mesoderm formation, and that antagonizing this axis using our Nipah envelope based reagents will illuminate these early differentiation processes.

Public Summary of Scientific Progress Introduction. EphrinB2 and ephB4 are cell surface molecules that mediate and transduce signaling cascades upon interaction with each other. Cell-cell contacts between ephrinB2 and ephB4 expressing cells provide guidance cues for cell migration and boundary formation in many developmental systems such as the formation of neurons and blood vessels. Importantly, ephrinB2 has been determined to be a molecular marker of “stemness” and is expressed in human embryonic stem cells, neural stem cells and hematopoietic stem cells. Despite much evidence from model systems that ephrinB2/ephB4 axis may be intimately involved in ESC fate (survival, self-renewal, and pluripotency), this particular axis has not been carefully studied in human ESC due to the lack of highly specific reagents to block cognate ephrinB2-ephB4 interactions. Intriguingly, the envelope protein from an exotic and highly lethal virus called Nipah virus (NiV), binds ephrinB2 more “tightly” than the EphB4 receptor, and can therefore compete or interfere with normal ephrin-B2-EphB4 interactions. NiV envelope proteins pseudotyped onto lentiviral particles can also specifically transduced EphrinB2 expressing cells. Thus, using an arsenal of reagents based on engineered versions of this viral envelope protein, which retains the ephrinB2 binding properties without the virulence of the actual virus, we had proposed to interrogate the role of the ephrinB2-ephB4 axis in regulating hESC’s ability to proliferate, self-renew, and differentiate into any cell type that make up the human body. In Year 1, using NiV envelope mediated lentiviral transduction to mark ephrinB2+ hESCs, we found that ephrinB2+ cells were homeostatically maintained at 5-20% of total SSEA4+ hESCs, even if ephrinB2+ cells were purified to near homogeneity (>85%) between passages. These results indicate that ephrinB2 does not mark for a stable distinct subpopulation of hESCs; instead, ephrinB2 expression might represent a marker for intrinsic stem cell heterogeneity that needs to be maintained at a certain percentage of hESCs in culture in order for the line to maintain all the cardinal properties of stem cells. However, the subpopulation of ephrinB2+ hESCs do appear to have decreased self-renewal capacity, although they maintained the ability to form teratomas, albeit smaller ones, when injected in SCID mice. Using a “spin embryoid body” (spin EB) assay as an in vitro surrogate assay for pluripotency and monitoring the time-course and expression levels of various germlayer differentiation markers after formation of spin EBs, we found that ephrinB2 expression closely mirrored the upregulation of ectoderm markers (the germ layer that forms cells like neurons), and to a lesser extent, mesoderm markers (the germ layer that forms cells like endothelial cells and hematopoietic stem cells). These results suggest that ephrinB2-EphB4 interactions likely play a role in regulating ectoderm and mesoderm formation, and that antagonizing this axis using our Nipah envelope based reagents will illuminate these early differentiation processes. In Year 2, we examined the effects of antagonizing the ephrinB2-ephB4 axis by generating stable hESCs (H9 and UCLA1 cell lines) expressing the soluble NiV attachment glycoprotein (sNiV-G) or a short hairpin RNA against ephrinB2 (shB2). sNiV-G binds to ephrinB2 and should prevent bi-directional signaling via the ephrinB2-ephB4 axis, while shB2 knocks down ephrinB2 mRNA expression by 50-80%. sNiV-G expressing hESCs gradually lose their pluripotency markers (SSEA4 and Oct-4) while upregulating ectoderm markers like Pax6 by 100-fold. On the other hand, hESCs expressing the shB2 exhibited marked defects in ectoderm differentiation (pax6 and NeuroD) when assayed using the spin EB method under spontaneous differentiation protocols. When the spin EB method was performed under directed-mesoderm differentiation, shB2 hESCs showed a 10-fold decrease in CD34 levels compared to control hESCs, indicating a defect in endothelial cell and/or hematopoietic cell differentiation. Collectively, our results show that antagonizing the ephrinB2-ephB4 axis can affect the pluripotency of hESCs, specifically with regards to ectoderm and mesoderm differentiation. Interestingly, physically antagonizing the ephrinB2-ephB4 signaling in trans (via secreted sNiV-G binding to ephrinB2) and knocking down ephrinB2 expression in cis (via shB2 mediated decrease in ephrinB2 mRNA) appears to reveal the different roles that ephrinB2-ephB4 axis can play in ectoderm and mesoderm differentiation. Future experiments will examine these putative differences in greater detail, and also confirm their phenotype on hESC pluripotency in vivo via the use of teratoma formation assays.

Introduction. EphrinB2 has been determined to be a molecular marker of “stemness” and is expressed in human embryonic stem cells, neural stem cells and hematopoietic stem cells. However, the ephrinB2 signaling axis has not been carefully studied in human ESC due to the lack of highly specific reagents to block cognate ephrinB2-ephB4 interactions. Intriguingly, the envelope protein from Nipah virus (NiV) binds ephrinB2 with very high affinity and specificity, and can therefore compete or interfere with normal ephrinB2 interactions with its cognate Eph receptors. NiV envelope proteins pseudotyped onto lentiviral particles can also specifically transduced ephrinB2+ cells. Thus, using an arsenal of reagents based on engineered versions of this viral envelope protein, we had proposed to interrogate the role of the ephrinB2 signaling axis in regulating hESC’s ability to proliferate, self-renew, and differentiate into any cell type that make up the human body. In Year 1, using NiV envelope mediated lentiviral transduction to mark ephrinB2+ hESCs, we found that ephrinB2+ cells were homeostatically maintained at ~20% of total SSEA4+ hESCs, even after repeated purification between passages. Thus, ephrinB2 does not mark for an independent, stable subpopulation of hESCs. Instead, ephrinB2 may be an intrinsic marker of stem cell heterogeneity; perhaps an emergent marker that arises from the statistical mechanics model of pluripotency as recently proposed by MacArthur and Lemischka (Cell, 2013). EphrinB2 expression closely mirrored the upregulation of ectoderm markers, and to a lesser extent, mesoderm markers in a “spin embryoid body” (spin EB) assay, which we used as an in vitro surrogate assay for assessing pluripotency. Our results suggest that ephrinB2 signaling axis likely plays a role in regulating ectoderm and mesoderm formation, and that antagonizing this axis using our Nipah envelope based reagents will illuminate these early differentiation processes. In Year 2, we examined the effects of antagonizing the ephrinB2 signaling axis by generating stable hESCs (H9 and UCLA1) expressing the soluble NiV env glycoprotein (sNiV-G) or a short hairpin RNA against ephrinB2 (shB2). sNiV-G should bind avidly to ephrinB2 and antagonize the complex forward and reverse Eph receptor-Ephrin ligand signaling axis, while shB2 knocks down ephrinB2 mRNA expression. sNiV-G expressing hESCs gradually lose their pluripotency markers (SSEA4 and Oct-4) while upregulating ectoderm markers like Pax6 by 100-fold. On the other hand, hESCs expressing the shB2 exhibited marked defects in ectoderm differentiation (pax6 and NeuroD) and to a lesser extent, mesoderm differentiation (CD34) when assayed using the spin EB method under spontaneous differentiation or directed-mesoderm differentiation conditions, respectively. Collectively, our results show that antagonizing the ephrinB2 signaling axis can affect the pluripotency of hESCs, specifically with regards to ectoderm and mesoderm differentiation. Interestingly, physically antagonizing the ephrinB2 signaling in trans (via secreted sNiV-G binding to ephrinB2) and knocking down ephrinB2 expression in cis (via shB2 mediated decrease in ephrinB2 mRNA) appears to reveal the different roles that ephrinB2 signaling axis can play in ectoderm and mesendoderm differentiation. In Year 3, we characterized the self-renewal, survival, and pluripotency of the sorted shB2 and shNT H9 hESCs (shNT is a non-targeted shRNA used as a comparison control for shB2). Parental H9, shNT and shB2 H9 cells showed no significant differences in self-renewal assays over 5 passages (data not shown). In vivo teratoma formation assays demonstrated no obvious qualitative differences in pluripotency as all three germ layers were observed in each H9 hESC line. However, in vivo teratoma formation assays are inherently variable and not amendable to easy quantification. To address the impact of ephrinB2 antagonism on germ layer specification in a more quantitative and holistic fashion, we performed microarrays on H9, shNT, and shB2 hESCs, and on their derived EBs (days 6 and 13) under spontaneous differentiation conditions and compared their global gene expression profiles. Our analysis revealed a progressively larger number of genes were specifically up- or down-regulated in shEFNB2 cells compared to H9 and shNT cells on days 0, 6, and 13, respectively. Further analysis indicated that ephrinB2 knockdown in H9 hESCs may enhance the formation of mesoendoerm progenitors while inhibiting the differentiation of neuro-ectoderm lineages. Finally, functional mesoderm-directed differentiation assays revealed that shEFNB2 hESCs have an increased propensity to differentiate into one specific sub-type of mesenchymal cells. In sum, the findings of this dissertation suggest that the heterogeneity of ephrinB2 expression and perturbation of ephrinB2 signaling may both be manipulated to enhance directed differentiation of hESCs in vitro.