Abstract
Throughout development, numerous binary cell fate decisions are made where cells activate competing transcriptional networks and subsequently upregulate one network over others based on extrinsic cues. Cell fate decisions are regulated by an interplay of different factors such as epigenetics and chromatin context. The multimeric cohesin complex has crucial roles in cell division and regulation of gene expression via chromatin organisation. Cohesin complexes obligatorily carry RAD21, SMC1 and SMC3A subunits and either STAG1 or STAG2. Given this interchangeability, homozygous loss of function mutations in STAG1 or STAG2 are generally compatible with life.
Acute myeloid leukaemia (AML) is a haematopoietic malignancy characterised by increased proliferation of immature cells and decreased differentiation. STAG2 is the most commonly mutated cohesin gene in AML and also in other solid tumours, where mutations are acquired in somatic cells. Despite this, STAG2-mediated cancer pathogenesis remains underexplored. Besides, germline mutations in STAG1 and STAG2 have been recently reported in patients with cohesinopathy, a group of developmental disorders caused by mutations in the cohesin subunit genes. Unlike in solid tumours, STAG2 mutations in both AML and cohesinopathies are not frequently associated with chromosomal aneuploidy suggesting transcriptional dysregulation as the key pathogenic mechanism here. Growing evidence supports shared and unique gene regulatory functions for STAG2 and its paralogue STAG1 in normal homeostasis and tumour suppression but the associated mechanisms are not completely understood. In this project, I investigated the role of Stag1/2 in development using zebrafish.
Zebrafish have four copies of the Stag genes namely, stag1a, stag1b, stag2a and stag2b. I found all four paralogues to be expressed maternally and zygotically during zebrafish embryogenesis, suggesting crucial functions. Next, I generated loss of function germline zebrafish mutants for three of the four paralogues using CRISPR-Cas9 mutagenesis: stag1anz204, stag1bnz205 and stag2bnz207. No germline mutations could be recovered for stag2a paralogue. This paralogue showed high maternal expression, low but detectable zygotic expression and was also expressed in adult zebrafish ovary suggesting a possible role in germ cell development which might explain the inability to recover stable mutations. Loss of function for the remaining three Stag paralogue had different consequences on zebrafish embryonic development. stag1anz204 mutants were phenotypically normal while stag1bnz205 and stag2bnz207 mutants had misplaced pigment cells, defective body axis patterning and notochord malformations. With respect to embryonic haematopoiesis, expanded runx1+ haematopoietic progenitors and primitive myeloid skewing followed by reduced definitive HSPCs were seen with stag1a loss. In contrast, stag1b loss did not impact embryonic haematopoiesis but altered the spatiotemporal expression of runx1 and sox2. In stag2bnz207 mutants, runx1+ haematopoietic precursors were reduced specifically in the posterior lateral mesoderm in a manner previously reported rad21 zebrafish mutants. The above results suggest that the four zebrafish stag paralogues are sub-functionalised during embryogenesis and that stag1a and stag2b have non-redundant haematopoietic functions.
AML is a heterogeneous disease marked by the presence of combinatorial de novo mutations. I aimed to model combinatorial cohesin mutations with two frequent co-occurring mutations that are found in AML namely, tet2 mutation and RUNX1-RUNX1T1 translocation. For this, I generated a novel tet2nz207 mutant line carrying a disrupted functional oxygenase domain resulting in reduced primitive erythropoiesis and reduced definitive HSPC numbers. The haematopoietic defects in combinatorial tet2 and cohesin mutants remain to be evaluated. The RUNX1-RUNX1T1 fusion protein was expressed under the control of the Runx1+23 enhancer in the background of cohesin mutants. Abnormal expansion of posterior haematopoietic tissue and aberrant pu.1+ myeloid marker expression was seen in combinatorial mutants. While the above are preliminary observations, further characterisation of these novel mutant lines is expected to reveal common disease mechanisms. Existing AML animal models fail to capture the haematopoiesis-restricted and somatic nature of mutations. To address this problem, I generated macrophage-restricted tet2 mutant lines and draculin-restricted stag2b mutant lines that generate continual de novo mutations by tissue-specific production of Cas9 mRNA.
In summary, my research findings suggest that cohesin-STAG1 and cohesin-STAG2 have non-overlapping roles in transcriptional regulation during embryonic development. The novel combinatorial and tissue-specific mutant lines generated will be useful to further elucidate the transcriptional dysregulation associated with STAG2 mutations in acute myeloid leukaemia.