Abstract
DNA methylation is an epigenetic modification prevalent in differentiated cells, playing a crucial role in defining their specialised functions. Equally, the removal of DNA methylation to a globally hypomethylated state enhances the pluripotency of ESCs by facilitating the erasure of cellular identity. DNA demethylation of ESCs can be initiated in culture with the use of two small inhibitory molecules, one of which inhibits GSK3b, thereby stabilising β-catenin and activating the canonical Wnt signalling pathway.
DNA hypomethylation is also a hallmark of cancer, where it may equally contribute to cellular dedifferentiation. Intriguingly, Wilms tumours with enhanced developmental potential and exaggerated global hypomethylation have stabilising mutations of β-catenin, functionally similar to small molecule inhibition of naïve mESCs. This raises the question of whether stabilised β-catenin is a common thread binding DNA demethylation in carcinogenesis and naïve pluripotency.
In Chapters 3 to 5, I investigate the association between Wnt signalling and methylation loss. Firstly, I quantified global methylation of 88 breast tumour samples and, by several independent approaches, identified tumours with hyperactive Wnt signalling. By these measures, I did not detect a correlation between hyperactive Wnt signalling and substantial global methylation erasure in breast cancer. Thus, such an association is not universally observed across different cancer types.
Additionally, I assessed the nature of the demethylation of Wilms tumours by comparing methylation levels at CpG islands to those of publicly available naïve human ESCs methylomes. Distinct patterns emerge, with CpG islands primarily hypermethylated in Wilms tumour and hypomethylated in ESCs, suggesting diverse mechanisms driving DNA demethylation in these two systems.
To determine the demethylating potential of stabilised β-catenin, I overexpressed mutated β-catenin, as observed in Wilms tumour, in mESCs. Subsequent low-coverage bisulfite-sequencing finds that β-catenin is insufficient to drive widespread methylation loss. This finding aligns with observations that GSK3b inhibition is not the primary driver of global methylation erasure, as indicated by treating mESCs individually with each of the two small molecule inhibitors that maintain naïve pluripotency. Collectively, these findings negate the role of stabilised β-catenin as a pivotal driver of demethylation in both naïve pluripotency and carcinogenesis.
Next, in a continuation of exploring the mechanisms governing DNA demethylation, I turn my attention to the TET demethylases. TET enzymes catalyse the oxidation of DNA methylation, prompting locus-specific active and passive DNA demethylation and, in turn, enhance pluripotency. However, the mechanism guiding TET targeting to specific sequences remains elusive. To address this, I rescued TET catalytic activity in mESCs by forced overexpression and calculated the rate of epigenetic memory erasure, finding that preferred TET targets included recognition sites for iconic methylation-sensitive and rapid-responding transcription factors, such as MYC and JUN/FOS. On the other hand, the least favourable TET motifs include DNA sites bound by the methylation-sensitive transcription factor OCT4 (Octamer-binding transcription factor 4). This observation, consistent with in vitro findings from collaborators, provides as explanation as to where methylation is erased in reprogramming events, both in vivo and in culture, underscoring TETs role in supporting pluripotency.
Emerging sequencing-based methods utilise enzymes, particularly TET, as alternatives to harsh chemical sodium bisulfite treatment for detecting and quantifying DNA methylation at base-pair resolution. Chapter 7 addresses whether inefficient TET targeting at non-preferred sequences introduces a sequence context biases, impacting which methylation marks evade TET oxidation and subsequently evade detection. This technical limitation is evaluated by calculating the difference in methylation levels of various sequence contexts quantified by WGBS and TET-dependent sequencing methods. When this difference in quantified methylation was compared to established TET targeting preferences, I found that indeed, sites least efficiently targeted by TET exhibit more under-quantification of methylation in TET-dependent sequencing methods in comparison to bisulfite-dependent methods.
This thesis acts to unravel the mechanisms underlying DNA demethylation concerning both pluripotency and cancer. My research collectively illuminates the potential significance of β-catenin being associated with methylation loss in these two systems, while also unveiling the targeting mechanism employed by TET enzymes. This understanding offers insights into how TET enzymes contribute to enhancing pluripotency and highlights the essential technical considerations that arise with their practical application.