Formation of mesoderm from the pluripotent epiblast depends upon canonical Wnt/␤-catenin signaling, although a precise molecular basis for this requirement has not been established. To develop a robust model of this developmental transition, we examined the role of Wnt signaling during the analogous stage of embryonic stem cell differentiation. We show that the kinetics of Wnt ligand expression and pathway activity in vitro mirror those found in vivo. Furthermore, inhibition of this endogenous Wnt signaling abrogates the functional competence of differentiating ES cells, reflected by their failure to generate Flk1 + mesodermal precursors and subsequent mature mesodermal lineages. Microarray analysis at various times during early differentiation reveal that mesoderm-and endoderm-associated genes fail to be induced in the absence of Wnt signaling, indicating a lack of germ layer induction that normally occurs during gastrulation in vivo. The earliest genes displaying Wnt-dependent expression, however, were those expressed in vivo in the primitive streak. Using an inducible form of stabilized ␤-catenin, we find that Wnt activity, although required, does not autonomously promote primitive streak-associated gene expression in vitro. Our results suggest that Wnt signaling functions in this model system to regulate the thresholds or stability of responses to other effector pathways and demonstrate that differentiating ES cells represent a useful model system for defining complex regulatory interactions underlying primary germ layer induction.

SUMMARY Wnt signaling is required for development of mesoderm-derived lineages and expression of transcription factors associated with the primitive streak. In a functional screen, we examined the mesoderm-inducing capacity of transcription factors whose expression was Wnt-dependent in differentiating ESCs. In contrast to many inactive factors, we found that mesoderm posterior 1 (Mesp1) promoted mesoderm development independently of Wnt signaling. Transient Mesp1 expression in ESCs promotes changes associated with epithelial-mesenchymal transition (EMT) and induction of Snai1, consistent with a role in gastrulation. Mesp1 expression also restricted the potential fates derived from ESCs, generating mesoderm progenitors with cardiovascular, but not hematopoietic, potential. Thus, in addition to its effects on EMT, Mesp1 may be capable of generating the recently identified multipotent cardiovascular progenitor from ESCs in vitro.

Reactive oxygen species (ROS) are highly reactive molecules that arise from a number of cellular sources, including oxidative metabolism in mitochondria. At low levels they can be advantageous to cells, activating signaling pathways that promote proliferation or survival. At higher levels, ROS can damage or kill cells by oxidizing proteins, lipids, and nucleic acids. It was hypothesized that antioxidants might benefit high-risk patients by reducing the rate of ROS-induced mutations and delaying cancer initiation. However, dietary supplementation with antioxidants has generally proven ineffective or detrimental in clinical trials. High ROS levels limit cancer cell survival during certain windows of cancer initiation and progression. During these periods, dietary supplementation with antioxidants may promote cancer cell survival and cancer progression. This raises the possibility that rather than treating cancer patients with antioxidants, they should be treated with pro-oxidants that exacerbate oxidative stress or block metabolic adaptations that confer oxidative stress resistance.

The reprogramming of somatic cells to inducible pluripotent stem cells requires a mesenchymal-to-epithelial transition. While differentiating ESCs can undergo the reverse process or epithelial-to-mesenchymal transition (EMT), little is known about the role of EMT in ESC differentiation and fate commitment. Here, we show that Snail homolog 1 (Snail) is expressed during ESC differentiation and is capable of inducing EMT on day 2 of ESC differentiation. Induction of EMT by Snail promotes mesoderm commitment while repressing markers of the primitive ectoderm and epiblast. Snail's impact on differentiation can be partly explained through its regulation of a number of ESC-associated microRNAs, including the microRNA-200 (miR-200) family. The miR-200 family is normally expressed in ESCs but is downregulated in a Wnt-dependent manner during EMT. Maintenance of miR-200 expression stalls differentiating ESCs at the epiblast-like stem cell (EpiSC) stage. Consistent with a role for activin in maintaining the EpiSC state, we find that inhibition of activin signaling decreases miR-200 expression and allows EMT to proceed with a bias toward neuroectoderm commitment. Furthermore, miR-200 requires activin to efficiently maintain cells at the epiblast stage. Together, these findings demonstrate that Snail and miR-200 act in opposition to regulate EMT and exit from the EpiSC stage toward induction of germ layer fates. By modulating expression levels of Snail, activin, and miR-200, we are able to control the order in which cells undergo EMT and transition out of the EpiSC state. Stem Cells 2011;29:764–776

New therapies are required for melanoma. Here, we report that multiple cardiac glycosides, including digitoxin and digoxin, are significantly more toxic to human melanoma cells than normal human cells. This reflects on-target inhibition of the ATP1A1 Na+/K+ pump, which is highly expressed by melanoma. MEK inhibitor and/or BRAF inhibitor additively or synergistically combined with digitoxin to induce cell death, inhibiting growth of patient-derived melanomas in NSG mice and synergistically extending survival. MEK inhibitor and digitoxin do not induce cell death in human melanocytes or haematopoietic cells in NSG mice. In melanoma, MEK inhibitor reduces ERK phosphorylation, while digitoxin disrupts ion gradients, altering plasma membrane and mitochondrial membrane potentials. MEK inhibitor and digitoxin together cause intracellular acidification, mitochondrial calcium dysregulation and ATP depletion in melanoma cells but not in normal cells. The disruption of ion homoeostasis in cancer cells can thus synergize with targeted agents to promote tumour regression in vivo.