Tbl1 promotes Wnt-ß-catenin signaling-induced degradation of the Tcf7l1 protein in mouse embryonic stem cells.

Activation of the Wnt-ß-catenin signaling pathway by CHIR99021, a specific inhibitor of GSK3ß, induces Tcf7l1 protein degradation, which facilitates the maintenance of an undifferentiated state in mouse embryonic stem cells (mESCs); however, the precise mechanism is still unclear. Here, we showed that the overexpression of transducin-ß-like protein 1 (Tbl1, also known as Tbl1x) or its family member Tblr1 (also known as Tbl1xr1) can decrease Tcf7l1 protein levels, whereas knockdown of each gene increases Tcf7l1 levels without affecting Tcf7l1 transcription. Interestingly, only Tbl1, and not Tblr1, interacts with Tcf7l1. Mechanistically, Tbl1 translocates from the cytoplasm into the nucleus in association with ß-catenin (CTNNB1) after the addition of CHIR99021 and functions as an adaptor to promote ubiquitylation of the Tcf7l1 protein. Functional assays further revealed that enforced expression of Tbl1 is capable of delaying mESC differentiation. In contrast, knockdown of Tbl1 attenuates the effect of CHIR99021 on Tcf7l1 protein stability and mESC self-renewal. Our results provide insight into the regulatory network of the Wnt-ß-catenin signaling pathway involved in promoting the maintenance of naïve pluripotency.

Inhibition of protein kinase C increases Prdm14 level to promote self-renewal of embryonic stem cells through reducing Suv39h-induced H3K9 methylation.

Inhibition of protein kinase C (PKC) efficiently promoted the self-renewal of embryonic stem cells (ESCs). However, information about the function of PKC inhibition remains lacking. Here, RNA-sequencing showed that the addition of Go6983 significantly inhibited the expression of de novo methyltransferases (Dnmt3a and Dnmt3b) and their regulator Dnmt3l, resulting in global hypomethylation of DNA in mouse ESCs. Mechanistically, PR domain-containing 14 (Prdm14), a site-specific transcriptional activator, partially contributed to Go6983-mediated repression of Dnmt3 genes. Administration of Go6983 increased Prdm14 expression mainly through the inhibition of PKCδ. High constitutive expression of Prdm14 phenocopied the ability of Go6983 to maintain` mouse ESC stemness in the absence of self-renewal-promoting cytokines. In contrast, the knockdown of Prdm14 eliminated the response to PKC inhibition and substantially impaired the Go6983-induced resistance of mouse ESCs to differentiation. Furthermore, liquid chromatography-mass spectrometry profiling and Western blotting revealed low levels of Suv39h1 and Suv39h2 in Go6983-treated mouse ESCs. Suv39h enzymes are histone methyltransferases that recognize dimethylated and trimethylated histone H3K9 specifically and usually function as transcriptional repressors. Consistently, the inhibition of Suv39h1 by RNA interference or the addition of the selective inhibitor chaetocin increased Prdm14 expression. Moreover, chromatin immunoprecipitation assay showed that Go6983 treatment led to decreased enrichment of dimethylation and trimethylation of H3K9 at the Prdm14 promoter but increased RNA polymerase Ⅱ binding affinity. Together, our results provide novel insights into the pivotal association between PKC inhibition-mediated self-renewal and epigenetic changes, which will help us better understand the regulatory network of stem cell pluripotency.

Rhox6 regulates the expression of distinct target genes to mediate mouse PGCLC formation and ESC self-renewal.

BACKGROUND: Mouse embryonic stem cells (mESCs) not only retain the property of self-renewal but also have the ability to develop into primordial germ cell-like cells (PGCLCs). However, knowledge about the mechanisms of transcriptional regulation is still limited. Rhox6, a member of the homeobox family that is located on the X chromosome, is highly expressed within PGCLCs in vivo and in vitro. However, the detailed effects of Rhox6 on PGCLC specification and mESC maintenance remain unclear. RESULTS: In this study, we found that overexpression of Rhox6 favors the formation of PGCLCs, while depletion of Rhox6 inhibits the generation of PGCLCs. Mechanistically, Rhox6 directly induces the expression of Nanos3 during the specification of PGCLCs. Subsequently, downregulation of Nanos3 expression is sufficient to decrease the ability of Rhox6 to induce PGCLC formation. Moreover, we found that depletion of Rhox6 expression facilitates the self-renewal of mESCs. High-throughput sequencing revealed that suppression of Rhox6 transcription significantly increases the expression of pluripotency genes. Functional studies further demonstrated that Rhox6 directly represses the transcription of Tbx3. Therefore, knockdown of the expression of the latter impairs the self-renewal of mESCs promoted by Rhox6 downregulation. CONCLUSIONS: Our study reveals that overexpression of Rhox6 is beneficial for PGCLC generation through induction of Nanos3, while downregulation of Rhox6 contributes to mESC self-renewal by increasing Tbx3. These findings help elucidate the early development of mouse embryos.

Bilateral Cross-Modality Graph Matching Attention for Feature Fusion in Visual Question Answering.

Answering semantically complicated questions according to an image is challenging in a visual question answering (VQA) task. Although the image can be well represented by deep learning, the question is always simply embedded and cannot well indicate its meaning. Besides, the visual and textual features have a gap for different modalities, it is difficult to align and utilize the cross-modality information. In this article, we focus on these two problems and propose a graph matching attention (GMA) network. First, it not only builds graph for the image but also constructs graph for the question in terms of both syntactic and embedding information. Next, we explore the intramodality relationships by a dual-stage graph encoder and then present a bilateral cross-modality GMA to infer the relationships between the image and the question. The updated cross-modality features are then sent into the answer prediction module for final answer prediction. Experiments demonstrate that our network achieves the state-of-the-art performance on the GQA dataset and the VQA 2.0 dataset. The ablation studies verify the effectiveness of each module in our GMA network.