TBX3 and HAND2 Controlled Gene Regulatory Networks in Establishment of Axis Polarity in Mouse Limb Buds

The limb bud is an outstanding model to study the precise control of the spatial and temporal developmental genes expression and its alterations resulting in phenotypic manifestations. The limb bud outgrowth and patterning is governed by a mesenchymal organizing center called the zone of polarizing activity (ZPA). The ZPA secrets the Sonic Hedgehog (SHH) morphogen that patterns the limb bud mesenchyme along its anterior-posterior (AP) axis. However, the limb bud mesenchymal AP polarity is already set before Shh expression is activated. The AP polarity in the nascent mesenchyme is established by a mutually antagonistic interaction which restricts the GLI3 repressor isoform to the anterior and the HAND2 transcription factor to the posterior mesenchyme (Galli et al., 2010; Osterwalder et al., 2014; te Welscher et al., 2002). However, HAND2 itself is not sufficient to establish a sharp boundary between the Hand2 and Gli3 expression domains but also requires Tbx3, which is itself a direct transcriptional target of HAND2. Genetic analysis has shown that TBX3 participates in positioning the posterior Gli3 expression boundary by inhibiting its expression in the posterior mesenchyme (Osterwalder et al., 2014). Prior to my study, the transcriptional target and networks governed by TBX3 during the establishment of AP limb bud polarity were mostly unknown. Thus, to identify the trans-acting interactions of TBX3 with the cis-regulatory modules (CRMs) located in the genomic landscapes of candidate target genes, I performed an unbiased genome-wide ChIP-seq analysis. As there is no ChIP-seq grade antibody, I first had to insert a 3xFLAG epitope tag into endogenous TBX3 protein using CRISPR/Cas9 genome editing in the mouse. The resulting Tbx33xF allele provided me with a highly sensitive tool to detect TBX3 in embryonic stem cells and embryonic tissues. Therefore, I was able to use this allele for TBX33xF ChIP-seq analysis during the critical stages of early mouse forelimb bud development. By intersecting different genome-wide TBX33xF and HAND23xF ChIP-seq with ATAC-seq and RNA-seq datasets, I was able to identify both TBX3-specific gene regulatory networks (GRNs), and GRNs shared with HAND2. Follow-up analysis identified the TBX3-specific and shared target genes that function in mesenchymal AP polarity establishment during the onset of limb bud development. By combining gene regulatory network analysis with a whole-mount in situ hybridization (WISH) screen, I gained insights into Tbx3-deficiency-related limb skeletal phenotypes that affect AP and proximal-distal (PD) axis patterning. Furthermore, my analysis showed that TBX3 transcriptional complexes interact with CRMs in the Gli3 cis-regulatory landscape, possibly contributing to repression to its expression from the posterior limb bud mesenchyme. In addition, my analysis reveals the existence of the TBX3 autoregulatory loop. Finally, the in-depth analysis showed that both TBX3 and HAND2 interact with CRMs in the Tbx2, Lmo1, and Gli3 cis-regulatory landscapes. I was also involved in a second project, which revealed an unexpected essential function of HAND2 in heart valve development. In particular, HAND2 is a key regulator of the endothelial-mesenchymal transition (EMT) of atrioventricular canal (AVC) cells that initiate cardiac cushion formation, which subsequently will give rise to the mitral and tricuspid valves.
Taken together, my PhD project sheds light onto the TBX3-specific and HAND2 shared GRNs that have essential morpho-regulatory functions during mouse embryonic development.