Transcriptional remodeling by OTX2 directs specification and patterning of mammalian definitive endoderm

Abstract

The molecular mechanisms that drive essential developmental patterning events in the mammalian embryo remain poorly understood. To generate a conceptual framework for gene regulatory processes during germ layer specification, we analyzed transcription factor (TF) expression kinetics around gastrulation and during in vitro differentiation. This approach identified Otx2 as a candidate regulator of definitive endoderm (DE), the precursor of all gut- derived tissues. Analysis of multipurpose degron alleles in gastruloid and directed differentiation models revealed that loss of OTX2 before or after DE specification alters the expression of core components and targets of specific cellular signaling pathways, perturbs adhesion and migration programs as well as de-represses regulators of other lineages, resulting in impaired foregut specification. Key targets of OTX2 are conserved in human DE. Mechanistically, OTX2 is required to establish chromatin accessibility at candidate enhancers, which regulate genes critical to establishing an anterior cell identity in the developing gut. Our results provide a working model for the progressive establishment of spatiotemporal cell identity by developmental TFs across germ layers and species, which may facilitate the generation of gut cell types for regenerative medicine applications.

Figure 1.. OTX2 depletion impairs DE development in murine models of germ layer specification

A. Otx2 expression levels in gastrulation-stage (E6.5–8.5) mouse embryos (modified after Pijuan-Sala et a., 2019). The red dotted line indicates the position of DE. B. Tissue annotation of gastrulation-stage mouse embryos with tissues expressing Otx2 listed. Arrows indicate different developmental trajectories along which Otx2 is expressed. C. Design of the murine OTX2-dTAG allele. Coordinates indicate the position of the Otx2 STOP codon (mm10), which was replaced with a multipurpose cassette. Blue box indicates a linker peptide. D. HA Western blotting (WB) after treatment of OTX2-dTAG EpiLC or EpiLCs derived from parental wildtype (WT) mESCs with dTAG-13 or DMSO for the indicated periods of time. Asterisk indicates a non-specific band. E. Representative IF images after staining anterior micropatterns formed in presence (top) or absence (bottom) of OTX2 for marker proteins of indicated lineages. F. Average marker intensity (+/− standard error of the mean) at different radial positions in anterior micropatterns. N = colonies analyzed. G. Protocol for directed differentiation of EpiSCs into DE, indicating compounds applied and timing of dTAG-13 administration for experiments shown in Figs. 1I,J. H. Quantification of Otx2-EGFP levels at indicated stages of directed differentiation. Dotted line indicates level of background fluorescence measured in WT cells. (****)p<0.0001 with one-way ANOVA with Tukey’s multiple comparison test. I. Quantification by flow cytometry of CXCR4⁺PDGFRA⁻ (DE) and PDGFRA⁺CXCR4⁻ (mesoderm) cells following directed differentiation in control (DMSO) or OTX2-depleted (dTAG-13) conditions. (***)p<0.001 and (****)p<0.0001 with unpaired t-test. N=4 separate cultures from each of two independent cell lines. J. Quantification of IF imaging of SOX17 (an endodermal marker) in DE derived from two independent OTX2-dTAG EpiSC lines cultured in absence or presence of dTAG-13. (****)p<0.0001 with unpaired t-test. N=4 separate cultures per cell line.

Figure 2.. Transcriptional dysregulation in DE upon OTX2 depletion

A. Force-directed layout graph of scRNA-seq analyses on WT EpiSCs, DE treated with DMSO and DE treated with dTAG-13 at PS (OTX2^depl_PS) DE cells. B. Volcano plot showing expression changes (logFC) and statistical significance (padj) in OTX2^depl_PS DE with DEGs (logFC>1;padj<0.05) highlighted in blue (DEG^DOWN) and red (DEG^UP), respectively. DEGs highlighted in the text or in other figure panels are annotated. C. Gene ontology (GO) analysis for DEG^DOWN and DEG^UP in OTX2^depl_PS DE using ENRICHR, listing top ranked GO terms for each gene category. D. Normalized expression profiles of select DEG^DOWN (blue names) and DEG^UP (red names) in OTX2^depl_PS DE that associated with indicated cellular programs. E. Comparison of the fold-change effect (dTAG-13:DMSO) on expression of select DEGs when OTX2 is depleted during (dTAG_PS; x-axis) or after (dTAG_DE; y-axis) DE specification. Genes DEGs only upon OTX2 depletion at PS are highlighted in red, all other genes are DEGs in both conditions. F. Normalized expression levels of a panel of anterior-posterior gut tube markers (Pijuan- Sala et al., 2019) OTX2^depl_PS DE, OTX2^depl_DE DE and in control DE treated with DMSO. G. Protocol to differentiate DE into SOX2⁺PDX1⁺ gastric foregut progenitors. H. Quantification of SOX2⁺PDX1⁺ foregut clusters formed after OTX2 depletion at either PS or DE compared to cells differentiated in DMSO. (*)p<0.05 and (**)p<0.01 with one-way ANOVA and Kruskal-Wallis multiple comparison test.

Figure 3.. Dynamics of OTX2 genome occupancy during DE specification

A. Classification and number of OTX2 binding sites identified by CUT&RUN in EpiSCs and DE. B. GO analysis of genes in proximity of EPI, primed, and de novo OTX2 binding sites. C. HOMER TF motif analysis at EPI, primed, and de novo OTX2 binding sites. The analysis was restricted to TFs expressed in DE and/or EpiSCs. D. IGV tracks of OTX2 genome occupancy at select DEG^DOWN loci with examples of genes exhibiting primed (left two panels) or de novo (right two panels) OTX2 binding. E. IGV tracks of OTX2 genome occupancy at select DEG^UP loci with examples of primed (top panel) and de novo (bottom panel) OTX2 binding. F. Percentage of DEG^DOWN, DEG^UP and gene loci unaffected by OTX2 depletion (“unchanged”) with OTX2 peaks of increasing (Q1 to Q4) intensity. G. Distribution of DEG^DOWN gene loci with evidence of primed (site bound both in EpiSCs and DE) and de novo (site bound only in DE) OTX2 binding with example genes highlighted. Note that all gene loci with primed peaks also harbor additional de novo peaks but not vice versa. H. Normalized expression levels of DEG^DOWN with either primed or de novo only OTX2 binding in EpiSCs, DE_DMSO an DE_dTAG. (****)p<0.0001 with two-sided unpaired Wilcoxon test.

Figure 4.. OTX2-dependent remodeling of chromatin accessibility during DE specification

A. Volcano plot showing ATAC-seq signal in DMSO-treated DE versus DE treated with dTAG-13 at PS. Significantly different DARs (logFC>1;padj<0.05) are colored in blue (DAR^DOWN) and red (DAR^UP). B. Four categories of DARs (C1-C4) identified by supervised clustering of ATAC-seq signal intensity in EpiSCs, control DE and OTX2-depleted DE. C. Normalized expression levels of genes in vicinity of C1 to C4 DARs in indicated samples. (***)p<0.001 and (****)p<0.0001 with two-sided unpaired Wilcoxon test. D. Example of chromatin accessibility changes at DEG^DOWN loci representing major cellular programs affected by OTX2. E. Frequency and type of OTX2 binding at C1 to C4 DARs. F. HOMER TF motif enrichment at C1-C4 DARs. Analysis was restricted to TFs expressed in EpiSCs, DE, or both. Dotted lines indicate TFs exhibiting enrichment at similar DAR categories.

Figure 5.. Impaired human DE formation upon OTX2 depletion

A. Design of the human OTX2-dTAG allele. Coordinates indicate the position of the Otx2 STOP codon (in hg38) replaced with the multipurpose cassette. The blue box indicates a linker peptide. B. Anti-HA WB with OTX2-dTAG hESCs (n=3 lines) and parental cells after 1h of culture in the presence of dTAG-v1 or DMSO. * = minor band indicating incomplete cleavage of P2A fusion protein. C. Representative flow cytometry plots showing CXCR4 and CD117 expression in human DE derived from MEL-1 cells after culture in the presence of DMSO (top) or dTAG-v1 (bottom) from the PS stage onwards. Cells with a canonical DE cell surface phenotype are highlighted. D. Abundance of cells with the canonical DE phenotype (CXCR4⁺CD117⁺) in cultures initiated with two independent OTX2-dTAG hESC lines and exposed to either DMSO or dTAG-v1 in two different culture conditions (n=3 independent cultures). (***)p<0.001 or (****)p<0.0001 with multiple T-tests and Bonferroni-Dunn correction. E. Measurement by qPCR of effect of dTAG-v1 treatment on select human genes whose mouse homologues are altered in their expression levels by OTX2 depletion in DE. (*)p<0.05, (***)p<0.001 or (****)p<0.0001 with multiple T-tests and Bonferroni-Dunn correction. F. EGFP fluorescence levels in DE cultures established in either presence of DMSO or dTAG-v1. Cultures established from parental non-transgenic (WT) human ESCs serve as control for background fluorescence. G. Representative low-magnification fluorescent live cell images of OTX2-EGFP⁺ human DE derived in presence of DMSO (top) or dTAG-v1 (bottom). Images are of MEL-1 derived cells using Protocol 1, but similar differences in cell clustering were observed with H9- derived cells and with DE generated using media 2. H. Schematic highlighting major functions of OTX2 during mouse and human DE specification.