Search

Cell identity is the control of gene transcription by Transcription Factors (TF) and chromatin. The extent to which human cell identity can be recapitulated outside of the cell using TFs and nucleosomes in unknown. I investigate if hESC-like nucleosome positioning, MNase-sensitivity and OCT4 and SOX2 (OS) binding can be recapitulated in vitro. Nucleosomes reconstitute well into cell-like positions upon human DNA sequences in vitro as determined by MNase-seq. These in vitro nucleosomes display fragility, but these nucleosomes are not fragile in hESCs. A sub-population of 110bp nucleosomes are reconstituted, which demonstrate similar MNase-resistance as the 150bp population of nucleosomes. The addition of OS to the reconstituted nucleosomes reveals SOX2-lead binding to MNase-sensitive 110bp nucleosomes, recapitulating OS binding to MNase-sensitive nucleosomes in hESCs. However, binding locations and hESC motifs are not recapitulated in vitro. These results contribute to the understanding of human nucleosome properties, and pioneer factor targeting of chromatin.

In multicellular biology, an astounding array of cellular identities are specified from the same genome, by drawing on a finite pool of transcription factors and signalling pathways in different permutations. How are signal and transcription network interplay computed by the cell to instruct identity? Must a given identity always be established by the same mechanism? To address these questions, I created defined and tractable systems based on reprogramming from epiblast stem cells to naïve pluripotency. By independent modulation of genetic and signalling parameters followed by isolation of successfully transitioning cells, I showed that naïve pluripotency can be established via distinct routes. These differ in their transcriptional trajectories and in their mechanisms, each with different genetic and signal requirements. Relative to development, one route initially moves forward, with productive cells acquiring mesodermal signature prior to naïve pluripotency induction. In contrast, another route overshoots backwards, transiently resembling the earlier embryo and recapitulating key aspects of naïve epiblast establishment in vivo. These remarkably distinct routes ultimately converge on the same naïve pluripotent endpoint, revealing surprising flexibility for the establishment of a single identity from a single origin. This provides evidence for cellular identity as a multidimensional attractor state (Kauffman, 1993), and extends the paradigm by which transcription factors and signals are used in different combinations to generate different cell types: they can also be used in different ways to generate the same cell type. How can the naïve pluripotent identity be reached via such different routes? I reconciled route differences, finding precise Oct4 expression as a unifying, essential and sufficient feature. Different routes achieve this required Oct4 expression by diverse logics, highlighting that successful identity change does not simply require activation of the destination network but instead pivots on the mechanism that permits a transition to occur. Fixing Oct4 at this level is sufficient to reprogram EpiSCs and more advanced cell types under signal instruction alone, yet is also compatible with bonafide development when returned to the in vivo context. At this expression level, I therefore deem Oct4 to be a 'transition' factor, a novel concept that I define as permitting identity change in various directions depending on environmental instruction. In sum, this thesis furthers our knowledge of naïve pluripotency induction in vitro and in vivo, and provides conceptual advances in our understanding of cell identity transitions.

Armadillo family proteins are present throughout the animal kingdom. In fruit flies, frogs, fish and mammals these proteins are essential for correct embryonic patterning via the Wnt/Wingless signal transduction pathway, acting downstream of glycogen synthase kinase-3 (gsk3). In addition, the same proteins are essential for regulated cell to cell adhesion. These proteins must be tightly controlled to prevent the growth and metastasis of tumours in the adult organism. However, the relationship between their developmental and adhesive role is not well understood at present. The cellular slime mould Dictyostelium discoideum is a model system in which to study multicellular development and cell adhesion. The Dictyostelium homologue of glycogen synthase kinase-3, gskA, is essential for the correct proportioning of cell types during development. The downstream targets of gskA remained to be discovered at the start of this work. I have isolated an Armadillo family gene, aardvark (aar), in Dictyostelium and created a knock-out mutant. The mutant phenotype demonstrates that aar has an important role in the later stages of development. Aardvark is crucial for defining the final structure of the fruiting body, particularly the assembly of the stalk tube and its associated cell-cell adhesion junctions. Perturbation of this structure in the aar mutant leads to a duplication of the developmental axis. Aar also has a cell-autonomous role, promoting the terminal differentiation of stalk and spore cells. Furthermore, I present evidence that two other Armadillo homologues are also present in Dictyostelium. These proteins are developmentally regulated and their levels and distribution are altered in a gskA- mutant. This thesis suggests that the molecular control of pattern formation and cell to cell adhesion may be very highly conserved throughout evolution and that in Dictyostelium, as in animals, Armadillo homologues act to coordinate morphogenesis and cell identity.

Neuroblastomas (NB) are embryonal childhood tumors that derive from the multipotent neural crest cells (NCCs) of the peripheral nervous system. NB accounts for more than 15% of all childhood cancer-related deaths. Despite the most intensive multimodal therapy, more than 50% of patients with high-risk NB relapse with often fatal, resistant disease. Novel therapies are desperately needed to improve cure rates. Previous studies proposed that the deregulation of normal neural crest developmental programs contributes to NB oncogenesis by retaining the highly migratory and proliferative traits of NCCs. Thus, activation or repression of neural crest developmental pathways have been implicated in NB pathogenesis. Recent data reported two identities in neuroblastoma cell lines: one establishing a more proliferative adrenergic (ADRN) cell state and a second establishing a more invasive, therapy-resistant mesenchymal (MES) cell state. Super-enhancer-associated transcription factor (TF) networks define cell identities in neuroblastoma (NB). Dysregulation of these TFs contributes to the initiation and maintenance of NB by enforcing early developmental identity states. We report the bHLH transcription factor TCF4 (E2-2) is a critical NB dependency gene that significantly contributes to these identity states through its heterodimerization with cell identity specific bHLH TFs. Mechanistically, we show that TCF4 promotes cell proliferation through direct transcriptional regulation of a MYC/MYCN oncogenic program. To identify potential therapeutic vulnerabilities, we characterized the TCF4 regulatory interactome and identified multiple epigenetic factors including HDACs and KDM1A. We determined that inhibitors to both HDACs and KDM1A, which often form complexes together, reduce TCF4 protein stability. Our work suggests that loss of TCF4 protein expression is an important biological readout for determining the efficacy of these epigenetic inhibitors in treating patients and could lead to improved patient outcomes.

Cell identity is shaped by complex gene expression programs, at the core of which lies the set of messenger RNAs (mRNAs) expressed in a given cell. While there is considerable knowledge about the involvement of mRNA transcription in cell fate control, the role of post-transcriptional mechanisms – which are governed by various RNA-binding proteins (RBPs) – is less well understood. By focusing on specific RBPs, the stem cell factor LIN28 and the terminal uridylyltransferases (TUTases) ZCCHC6/11, this dissertation explores novel post-transcriptional regulatory mechanisms that contribute to the control of cell identity. Chapter 2 addresses the question of how signaling and post-transcriptional regulation are integrated to influence cell fate. In particular, we investigated the role of LIN28 phosphorylation in pluripotent stem cells (PSCs). We found that MAPK/ERK, a central signaling regulator of pluripotency, phosphorylates LIN28 to increase its protein stability and thereby enhance LIN28’s translational regulation of its mRNA targets, which contributes to the control of pluripotency transitions. These findings establish a novel link between extracellular cues, mRNA translation, and cell fate regulation. Chapter 3 examines the role of ZCCHC6/11-mediated mRNA uridylation, a recently appreciated mechanism for promoting global mRNA decay, in the control of cell identity. In particular, we explored the functions of these TUTases in cancer cells, PSCs and muscle progenitors. We found that ZCCHC6/11 contribute to oncogenic transformation and support the growth and tumorigenicity of cancer cell lines. Mechanistically, these effects were associated with altered mRNA uridylation and turnover, including dysregulation of cell cycle factors and concomitant cell cycle impairment. Interestingly, we also report that ZCCHC6/11 promote a less differentiated cell state in both PSCs and lineage-restricted muscle progenitors. Our results reveal novel functions for ZCCHC6/11 and implicate uridylation-mediated mRNA turnover as a mechanism of oncogenesis. Collectively, our work uncovers new post-transcriptional regulatory mechanisms in the control of cell identity, with implications for stem cell and cancer biology.