Research Projects

Transcriptional and epigenetic regulation in normal and malignant stem cells.

The decision between self-renewal and differentiation of stem cells is masterminded by complex transcriptional networks, which are also frequently mutated in human leukaemia (Look, 1997).  In acute myeloid leukaemia (AML) where leukaemic stem cells (LSCs) have been functionally identified, the most prevalent chimeric leukaemia associated transcription factors (LATFs) arise from mutations of the retinoic acid receptor (RARa), the core-binding factors (either of its two subunits AML1 or CBFβ), the mixed lineage leukaemia protein (MLL) or its downstream effector, Hox proteins (So and Cleary, 2004), which are all critical for initiation and/or maintenance of the self-renewal in normal hematopoietic stem cells (HSCs) (Figure 2).  It is clear that transcription factors do not work as monomers but function as transcriptional complexes by recruiting various transcriptional co-regulators/cofactors. These factors, including DNA binding cofactors and epigenetic regulators, are critical for directing or enhancing the DNA binding and transcriptional activity of the complexes.

While chimeric LATFs exhibit very diverse properties, they can be functionally classified according to their dominant transcriptional activities into activating and repressive transcription factors (So and Cleary, 2004) (Figure 3). Examples of dominant repressive LATFs include chimeric RAR? and AML1 oncoproteins. They have compromised ability to recruit transcriptional co-activators but preferably associate with transcriptional co-repressors such as SMRT and HDACs, which result in suppression of downstream target genes such as CEBP and p14/Arf by RAR? and AML1 oncoproteins, respectively (Linggi et al., 2002; Park et al., 1999; Pitha-Rowe et al., 2003). In contrast, dominant activating LATFs such as MLL fusion can constitutively activate transcription of downstream target genes including Hox genes for maintaining self-renewal (Ayton and Cleary, 2003; So et al., 2004). While these two groups of LATFs apparently function in different ways, emerging evidence indicate that most of these LATFs share critical common features essential for transformation including aberrant self-association (Kwok et al., 2006; Kwok et al., 2009; Liu et al., 2006; So and Cleary, 2004; Sternsdorf et al., 2006), recruitment of DNA binding cofactors (Yokoyama et al., 2005; Zeisig et al., 2007; Zhu et al., 2007) and transcriptional effector complexes (Arteaga et al., 2013; Cheung et al., 2007; Krivtsov et al., 2008; Okada et al., 2005; So and Cleary, 2003; So et al., 2003b; Villa et al., 2007; Zeisig et al., 2008) (Figure 3).   Notably, the aberrantly recruited transcriptional effector complexes usually contain core members of epigenetic regulators, which have well-defined catalytic activity and are ideal targets for development of specific small molecule inhibitors. Among them are protein methyltransferases (PMTs) including both lysine methyltransferases (KMTs) and protein arginine methyltransferases (PRMTs) that confer methyl-mark, which can be removed by histone dememthylases (HDMs) for dynamic gene regulation (Cheung and So, 2011). Thus identification of these critical transcriptional components has given renewed optimism for targeting these classically non-druggable LATFs.

Primary Focus:

  • To define the roles of protein methyltransferases (PMTs) in normal and malignant haematopoiesis
  • To discover the roles of histone demethylases (HDMs) and other transcription accessory factors in leukaemogenesis



Figure 2. Deregulation of Hox genes in acute leukemia.  During development, Hox gene expression is initiated by pair-rule proteins (e.g., Runx) and subsequently maintained by TrxG (e.g, MLL) and PcG (e.g., Bmi-1) proteins.  Expression of Hox is also regulated by retinoic acid signalling  pathway (e.g, RARa).  Hox proteins function as trimeric complexes with Pbx and Meis1 or Prep1 to bind and regulate expression of downstream targets.  All of these proteins except Prep1 have been either directly involved in gene fusions or transcriptionally de-regulated in human leukemia. 



Figure 3. Activating and repressive oncogenic transcription factors. Leukemia associated chimeric transcription factors (LATFs) can be functionally classified into activating (top) and repressive (bottom) transcription factors. An example of activating LATFs is MLL fusion, which will recruit Prmt1 or Dot1L histone modifying enzymes for constitutive activation of downstream targets. In contrast, repressive chimeric LATFs such as RARa fusion protein recruiting HDAC/PRC complexes will have a dominant effect on silencing of downstream targets. Development of small molecule inhibitors targeting the epigenetic modifying enzymes may represent a promising avenue for development of specific and effective cancer therapeutics against these classically untraceable transcriptional targets.

Back to top