Competing regulatory modules control the transition between mammalian gastrulation modes

41 During mammalian gastrulation cells of the primary germ layers are generated in 42 anterior-to-posterior sequence employing different morphogenetic modes. Initial 43 gastrulation is characterized by cell ingression through the early primitive streak, 44 followed by posterior embryonic axis elongation via cell recruitment from progenitor 45 pools. Molecular details of different genetic programs controlling early and late 46 gastrulation remain ill described. Here, we employed stem cell -based mouse 47 gastruloids to reveal two consecutively acting regulatory modules that orchestrate 48 spatiotemporal progression of gastrulation. The early anterior module consists of the 49 Tbx transcription factor Eomes, and signalling molecules Nodal and Wnt3 that initiate 50 gastrulation and generate anterior mesoderm and definitive endoderm from the early 51 streak. The anterior module represses the second, Tbxt/Wnt3a posterior regulatory 52 module controlling axial extension at trunk levels. Both circuitries are self -reinforcing 53 while mutually repressing the counteracting module at multiple levels as the molecular 54 basis for the spatiotemporal progression of gastrulation along the AP axis. 55 56

57 The onset of gastrulation in mouse is marked by the emergence of the primitive streak 58 (PS), a transient structure formed by the ingression of epiblast cells that give rise to 59 the nascent mesoderm cell layer (Arnold & Robertson, 2009). Initially, the PS forms as 60 a confined hub of cells at the prospective posterior side of the cup -shaped embryo at 61 embryonic day 6.5 (E6.5). From here the PS progressively elongates distally until E7.5, 62 when the embryonic node forms at the most distal tip of the embryo (Downs & Davies, 63 1993; Williams et al., 2012) . Early ingressing cells predominantly generate nascent 64 mesoderm, including extraembryonic mesoderm (ExMes) and anterior, cranio-cardiac 65 mesoderm (AM), followed by definitive endoderm (DE) progenitors that subsequently 66 integrate into the outside endoderm layer (Kwon et al., 2008; Lawson et al., 1986; 67 Probst et al., 2021; Viotti et al., 2014) . After the first day of gastrulation, the 68 morphogenetic mode to generate the progenitors that give rise to cells for the posterior 69 extension of the body axis progressively changes (Aires et al., 2018; B énazéraf & 70 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 3 Pourquié, 2013; Duarte et al., 2023; Wymeersch et al., 2021). Different to epiblast cell 71 ingression through the early PS, in later phases of gastrulation the generation of 72 primordial tissues is fuelled from populations of proliferating progenitors. For example, 73 progenitors with neural and mesodermal potential (NMPs) are initially found in regions 74 of the node-streak border and the caudal lateral epiblast (Cambray & Wilson, 2007) . 75 NMPs persist as self-renewing progenitor pool in the chordo-neural hinge (CNH) of the 76 posterior tail bud regions of embryos from where they contribute to the generation of 77 mesoderm and spinal cord progenitors until gastrulation ceases (Cambray & Wilson, 78 2007; Wymeersch et al., 2019). 79 Gastrulation onset and the continuous patterning of the PS relies on instructive signals, 80 including Nodal/Smad2/3 and canonical Wnt signalling (Arnold & Robertson, 2009; 81 Bardot & Hadjantonakis, 2020; Morgani & Hadjantonakis, 2020; Robertson, 2014) . 82 Epiblast expression of Nodal initiates the formation of the PS and instructs cell 83 specification of early PS-derivatives, including Mesp1-expressing cranio-cardiac and 84 extraembryonic mesoderm (Costello et al., 2011; Lu & Robertson, 2004), and anterior 85 PS (APS) derived tissues, namely DE, node, notochord, and prechordal plate (Dunn 86 et al., 2004; Vincent et al., 2003) . PS formation additionally relies on canonical Wnt -87 signalling (Liu et al., 1999; Ten Berge et al., 2008). First, Wnt3 is expressed at the site 88 of the future PS and then along the PS until E7.5 (Liu et al., 1999; Rivera -Pérez & 89 Magnuson, 2005). Wnt3 expression is followed by the upregulation of Wnt3a in the late 90 PS and tail bud, that in concert with transcription factors (TFs) such as Tbxt (aka 91 Brachyury or T) and Cdx factors promotes mesoderm lineages and the elongation of 92 the posterior body axis (Amin et al., 2016; Aulehla et al., 2003; Herrmann et al., 1990; 93 Martin & Kimelman, 2008; Savory et al., 2009; Takada et al., 1994; Young et al., 2009). 94 This consecutive expression of different Wnt ligands during gastrulation suggests 95 individual regulatory functions in the course of germ layer formation (Wang et al., 2012; 96 Yamaguchi, 2008). 97 Cell type specification and germ layer morphogenesis additionally relies on sets of TFs, 98 including the Tbx factors Eomes and Tbxt. Eomes instructs the specification of AM 99 derivatives (Costello et al., 2011; Probst et al., 2021) and DE from the early PS (Arnold 100 et al., 2008; Teo et al., 2011), and Tbxt is required for axial mesoderm derivatives, and 101 paraxial mesoderm caudal to the first seven to ten pairs of somites (Koch et al., 2017; 102 Schüle et al., 2023). In the absence of both Tbx TFs all mesoderm and endoderm (ME) 103 gene programs entirely fail due to crucial requirements of Tbx factors for pluripotency 104 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 4 exit, the generation of chromatin-based competence for ME lineage specification, and 105 the concomitant repression of neuroectoderm lineage fates Previous studies suggest 106 that Eomes is regulated by and functionally associated with Nodal/SMAD2/3-signalling 107 (Brennan et al., 2001; Pfeiffer et al., 2018) . Similarly, Tbxt is induced and functionally 108 linked to canonical Wnt -signalling (Arnold et al., 2000; Martin & Kimelman, 2008; 109 Yamaguchi et al., 1999). However, to date, the molecular interdependencies of signals 110 and TF activities that orchestrate gastrulation along tight spatiotemporal patterns 111 remain poorly understood. 112 Here, we employed 3D mouse embryonic stem cell-based gastruloids (Beccari et al., 113 2018; McNamara et al., 2024; Rossi et al., 2021; van den Brink et al., 2014) to untangle 114 the dynamic interdependencies between signals and transcriptional regulators that 115 govern the progression between different gastrulation stages as found in E6.5 to ~E8.5 116 mouse embryos. To address individual functions of signalling components and TFs, 117 we used variations of the classical gastruloid protocol (Beccari et al., 2018; Turner et 118 al., 2017; van den Brink et al., 2014; Veenvliet et al., 2020) in combination with genetic 119 approaches (Wehmeyer et al., 2022). Hence, we could define two mutually competing 120 regulatory modules that contain the core molecular players essential for the 121 progression through gastrulation stages. A rapid switch between the se regulatory 122 circuitries accounts for proper timing and coordination of gastrulation modes and 123 coincides with the timepoint of formation of the node at E7.5 in the mouse embryo. 124 125

126 Consecutive morphogenetic modes of gastrulation correspond to a transition of 127 signalling signatures 128 Previous studies suggested that the cranial somites up to the level of somite pair 7-10 129 originate from early specified mesoderm and originate from a different progenitor pool 130 than later forming, more caudal somites (Guibentif et al., 2021; Kinder et al., 1999) . 131 These progenitor pools most likely also underlie different molecular regulation. This 132 notion is also supported by mouse mutants for Tbxt or Wnt3a that form most anterior 133 somites but lack somites posterior to the first 7 -10 pairs (Kispert & Herrmann, 1994; 134 Takada et al., 1994) . First, we aimed to inquire when different progenitor pools for 135 anterior (7 somites) mesoderm are generated. We 136 therefore re-analysed previous live embryo timelapse cell tracking movies (McDole et 137 al., 2018) to resolve when somitogenic mesoderm for somites at different levels is 138 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 5 generated. Analysed embryos comprised the time span from early PS formation to the 139 8-10 somite stage , corresponding to E 6.5 – E8.25 (McDole et al., 2018) . We 140 backtracked cells of the first 8 rows of cranial somites to the time of their origin (Fig. 141 1A). We found that the progenitors of most cranial somites are present in distal portions 142 of the PS and are already generated before the embryonic node is formed (Fig. 1A, 143 green cells, 11h40). Somitogenic mesoderm of more caudal somites is generated after 144 the node becomes morphologically discernible (Fig. 1A, purple cells, 42h00) , which 145 corresponds to the timepoint E7.5 (Downs & Davies, 1993). Notably, the timepoint of 146 E7.5 also indicates the onset of NMPs formation in vicinity to the embryonic node 147 (Wymeersch et al., 2021), further supporting that the origin and mode of somitogenic 148 mesoderm greatly differs between the most rostral and more caudal somites. To 149 correlate the onset of a morphogenetic transition of gastrulation at E7.5 to changes in 150 the signalling environment, we analysed published scRNA-seq data sets (Pijuan-Sala 151 et al., 2019). We specifically examined the expression dynamics of known regulators 152 of PS patterning, such as signalling molecules Nodal, Wnt3 and Wnt3a, and the Tbx 153 TFs Eomes and Tbxt (Fig. 1B). In accordance with previous expression analyses (Liu 154 et al., 1999; Lu & Robertson, 2004; Takada et al., 1994), the scRNA-seq data similarly 155 shows a rapid change of the signalling landscape around E7.5 (Fig. 1 B). Nodal and 156 Wnt3 are lost abruptly around E7.5 from the PS, and in reverse Wnt3a expression is 157 rapidly established and continues to be expressed in caudal mesoderm partially 158 overlapping with Tbxt (Fig. 1 B). Eomes, which is widely expressed throughout the 159 posterior epiblast, PS and nascent mesoderm during early gastrulation (E7.0) and 160 overlaps with Tbxt in the E7.0 and E7.5 PS, is similarly rapidly downregulated and 161 mostly absent at E8.0 (Schüle et al., 2023) . In sum, the highly dynamic expression 162 patterns of these key regulators of gastrulation tightly correspond to the transition of 163 gastrulation stages at the time of embryonic node formation at E7.5. 164 165 Gastruloids recapitulate the transition of gastrulation modes 166 To study the progression between gastrulation modes from early PS stages to 167 posterior axial elongation after E7.5, we employed embryonic stem cell (ESC) -based 168 gastruloids. We first applied the standard protocol of gastruloid formation using a pulse 169 of CHIR-treatment to induce canonical Wnt-signalling responses (van den Brink et al., 170 2014) (Fig. S1A) and analysed the protein dynamics of EOMES and TBXT as key 171 regulators of ME formation, by immunofluorescence (IF) staining (Fig. 1C). Similar to 172 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 6 the early embryo (Schüle et al., 2023) , EOMES and TBXT initially co -localise in 60h 173 gastruloids before TBXT becomes restricted to outer cells of gastruloids at 72 h. At 84 174 h obvious morphological asymmetries become apparent when TBXT is found at one 175 pole of the gastruloids, and EOMES is reduced and sparsely found at the opposite 176 pole. Elongation of gastruloids follows the loss of EOMES at 96 h, thereby mimicking 177 embryonic development after node formation (E7.5), when axial elongation is governed 178 by NMP-generated tissues. Hence, patterns of expression dynamics of EOMES and 179 TBXT in gastruloids largely reflect the observations at the embryonic streak. 180 Accordingly, gastruloids can be used as suitable model systems for studies of the 181 mechanisms that control the transition between gastrulation stages from E7.5 (Fig. 1D). 183 184 Combined activities of Tbx factors and signals instruct temporal progression 185 and patterning of gastruloids 186 To dissect the molecular regulation of gastrulation progression from early PS stages 187 to posterior axis elongation we first investigated functional requirements for Eomes and 188 Tbxt during gastruloid formation in context of different signalling environments. We 189 used either a standard CHIR pulse, or induced with the NODAL-analogue ACTIVINA 190 (ActA) (Fig. S1A, B). We compared gastruloids from wildtype (WT), Eomes- or Tbxt- 191 deficient (Eo - / - or Tbxt - / -) and double knockout (dKO) mESCs (Fig. S1C) (Schüle et 192 al., 2023; Tosic et al., 2019) . Expectedly, using the standard CHIR protocol, WT 193 gastruloids stereotypically develop a cell-dense region, representing the anterior pole, 194 and elongate posteriorly (Fig. 2A, S1D). The simultaneous deletion of both Eomes and 195 Tbxt (dKO) fully abrogates symmetry breaking, and dKO aggregates remain round 196 (Fig. S1D). Tbxt - / - gastruloids fail to elongate and remain oval -shaped at 120 h, as 197 also previously reported (Fig. 2B, S1D) (Wehmeyer et al., 2022) . Surprisingly, Eo - / - 198 gastruloids elongate and are morphologically similar to WT (Fig. 2C; Fig. S1D), despite 199 crucial morphogenetic roles of Eomes during early gastrulation (Arnold et al., 2008) . 200 We morphometrically quantified these observations by measuring the length and width 201 of gastruloids and calculated length/width ratios for each genotype and condition as an 202

read-out for elongation (Fig. S1E, F, Supplementary Table 1). 203 Next, we phenotypically analysed gastruloids, when induced with ActA (25 ng/ ml), that 204 also was applied as pulse between 48 and 72 h (Fig. S1B). As previously reported 205 (Turner et al., 2017; van den Brink et al., 2014) , ActA suppresses axis elongation in 206 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 7 WT aggregates, which acquire irregular round to triangular shapes at 120 h similar to 207 Tbxt - / - aggregates (Fig. 2A, S1D ). Unexpectedly, Eo - / - (ActA)-induced gastruloids 208 consistently elongate and are morphologically alike CHIR-treated gastruloids. These 209 findings suggest that Tbxt promotes axis -elongation in conjunction with CHIR/Wnt -210 stimulation, and that Eomes mediates elongation -inhibiting activities of the 211 Nodal/ActivinA-signalling pathway. 212 To molecularly characterize these morphological findings in gastruloids we 213 transcriptionally profiled CHIR - and ActA-induced gastruloids from WT, Tbxt - / - and 214 Eo - / - cells at three time point (72 h, 96 h and 120 h) using bulk RNA sequencing and 215 analysed differentially expressed genes (DEGs) when compared to WT gastruloids 216 (Fig. 2D, E, F, S2A-D). Gross numbers of DEGs indicate that the deletion of Tbxt 217

in larger transcriptional differences in CHIR treated gastruloids (Fig. S2A), and 218 the deletion of Eomes leads to largest effects in ActA induced gastruloids (Fig. S2B), 219 suggesting functional synergism between pairs of Tbxt/Wnt- (CHIR)-signalling and 220 Eomes/Nodal/ActA-signalling, respectively. GO Term analyses of DEGs confirm the 221 phenotypic observations of disturbed AP axis formation and patterning of Tbxt - / - 222 (CHIR)-induced gastruloids, that is remarkably reestablished in ActA -treated 223 aggregates when Eomes is deleted (Eo - / - (ActA); Fig. S2C, D). 224 To illustrate the transcriptional patterns characteristic for gastruloid development, we 225 generated heatmaps that show the sequential expression of marker genes along the 226 anterior-to-posterior axis (n=97 genes, Supplementary Tables 2, 3). The heatmap for 227 WT, CHIR-treated gastruloids shows the orderly pattern of early arising (72 h) anterior 228 marker genes (e.g., Snail, Evx1, Lhx1, Mesp1 ) and the transition to later expressed 229 marker genes of posterior identities (e.g., Cdx2, Wnt5a, Hoxa9, Hoxc10 ) (Fig. 2D, 230 CHIR-treatment). Heatmaps across different induction regimes and genotypes show 231 gross alterations of spatiotemporal patterns. Pulsing WT gastruloids with ActA 232 attenuates expression of posterior and increases expression of anterior markers (Fig. 233 2D). Additionally, ActA leads to the globally delayed expression of early anterior marker 234 genes from 72 to 96 h in WT and Tbxt-deficient gastruloids, supposedly reflecting the 235 pluripotency-maintaining effect of ActA/Nodal-signalling (Vallier et al., 2009) (Fig. 2D, 236 E). Tbxt - / - gastruloids show markedly increased expression of anterior and severely 237 reduced expression of posterior marker genes, irrespective of pulsing with either CHIR 238 or ActA (Fig. 2E). In reverse, Eomes-deficient ( Eo - / -) gastruloids show a gross 239 reduction of early anterior and increased expression of posterior markers above levels 240 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 8 found in WT gastruloids , irrespective of the initial pulse of CHIR or ActA (Fig. 2F, 241 compare to Fig. 2D). Importantly, despite ActA -treatment Eo - / - gastruloids strongly 242 express markers of posterior axis identity at 120 h fitting to the observed elongated 243 phenotype (Fig. 2C). The heatmap analyses of AP-axis markers from RNA-seq data 244 were confirmed by whole-mount in situ hybridisation (WISH) of early anterior ( Lhx1) 245 and later, posterior (Cdx2, Hoxb9) expressed marker genes (Fig. 2G-I). WISH marker 246 analysis validated that Nodal/ActA promote s anterior fate, while Wnt/CHIR-signals 247 instruct posterior identities (Fig. 2G-I). These patterning functions of signals critically 248 rely on the Tbx factors Eomes and Tbxt. Irrespective of the initial inductive signalling 249 pulse, gastruloids acquire posterior identities and elongate in the absence of Eomes 250 (Fig. 2F, I), while Tbxt-deficient gastruloids acquire mainly anterior fate identities (Fig. 251 2E, H). This suggests that activities of Nodal- and Wnt-signalling in gastruloids are 252 predominantly mediated by the presence or absence of Eomes or Tbxt, respectively 253 (Fig. 2J). 254 255 EOMES and BRACHYURY show mutual repressive regulation 256 Anterior-posterior marker expression in gastruloids largely relies on opposing activities 257 of Eomes and Tbxt (Fig. 2). Competing activities are also reflected in differential protein 258 distribution within gastruloids (Fig 1D). To capture the time period when the AP-259 asymmetry is established, we performed IF staining for EOMES and TBXT in CHIR -260 treated WT gastruloids collected from the start of CHIR treatment at 48 h in 6 h intervals 261 until 96 h (Suppinger et al., 2023) (Fig. 3A, partially also shown in Fig. 1 C). EOMES 262 and TBXT initially co -localise from 60 h until 72 h , before becoming restricted to 263 different domains at 84 h (Fig. 3A). To address if the expression of both Tbx factors 264 impact each other, w e compared IF pattern found in WT gastruloids to Eo - / - and 265 Tbxt - / -gastruloids (Fig. 3B, C). In Eo - / - gastruloids, polarization of TBXT to one side 266 of the gastruloid takes place , h owever, overall protein distribution is less confined 267 compared to WT gastruloids (Fig. 3B). In Tbxt - / - gastruloids, EOMES is first induced 268 similar to WT, but in contrast to the downregulation in WT, increased EOMES staining 269 persists at 96 h in Tbxt - / - (Fig. 3C). To relate IF staining to gene expression levels we 270 quantified expression from RNAseq data at 72, 96 and 120 h (Fig. 3D, E). mRNA levels 271 grossly match the observed protein staining intensitie s, such that we find increased 272 Tbxt levels in Eo - / - gastruloids (Fig. 3E), and maintained Eomes expression in Tbxt - / - 273 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 9 gastruloids (Fig. 3D). These findings suggest a mutual negative regulation after an 274 early phase of co-expression by Eomes and Tbxt as schematically depicted (Fig. 3F). 275 276 Tbx TFs and signal ling ligands act as reinforcing functional modules of 277 Eomes/Nodal/Wnt3 and Brachyury/Wnt3a 278 Next, we aimed to investigate more deeply potential modes of mutual repression 279 between Eomes and Tbxt. Based on the delayed repression dynamics (Fig. 3) we 280 suspected cell non -cell-autonomous reciprocal repression mechanisms between 281 EOMES and TBXT in addition to previously reported direct repression of Tbxt-activities 282 by Eomes (Schüle et al., 2023). Hence, we analysed RNA expression patterns of the 283 upstream signalling regulators of Eomes and Tbxt expression, namely Nodal (Brennan 284 et al., 2001; Simon et al., 2017), Wnt3 (Pfeiffer et al., 2018), and Wnt3a (Arnold et al., 285 2000; Yamaguchi et al., 1999), using WISH at 48 h, 72 h, 84 h, 96 h, and 120 h in WT, 286 Tbxt - / - and Eo - / - gastruloids (Fig. 4A-C). In WT gastruloids Nodal expression 287 precedes the CHIR pulse at 48 h, expression peaks at 72 h, and is followed by 288 downregulation in most parts of the gastruloids until 96 h (Fig. 4A). This pattern is 289 reminiscent of the embryonic expression where Nodal is present in the epiblast before 290 PS formation and becomes confined to the streak followed by the rapid downregulated 291 at E7.5 (Collignon et al., 1996). In WT gastruloids Wnt3 and Wnt3a are first expressed 292 after the CHIR pulse post 72 h. Wnt3 is downregulated between 72 h and 96 h, while 293 Wnt3a remains highly expressed and becomes confined to the posterior tail-like region, 294 that coexpresses Tbxt (Wehmeyer et al., 2022) (Fig. 4A). Hence, also Wnt3 and Wnt3a 295 expression reflect the kinetics found in embryos where Wnt3 expression is lost from 296 the PS at E7.5 (Liu et al., 1999) , and Wnt3a expression is maintained in the tail bud 297 (Takada et al., 1994). 298 In Tbxt - / - gastruloids Nodal and Wnt3 expression levels are markedly increased and 299 maintained beyond 96 h, while Wnt3a expression is almost absent (Fig. 4B). In 300 contrast, Eo - / - gastruloids show the opposite expression dynamics of decreased 301 Nodal and Wnt3, while Wnt3a levels appear grossly normal and locali se to a single, 302 enlarged expression domain at the posterior pole at 96 h (Fig. 4C). These expression 303 patterns by WISH are also recapitulated in mRNA levels of RNAseq data (Fig. 4D, G. 304 J). We next asked if expression of Nodal, Wnt3 and Wnt3a are directly regulated by 305 EOMES and/or TBXT. Hence, we analysed chromatin binding of TBXT and EOMES 306 by ChIPseq ( chromatin immunoprecipitation coupled with sequencing , data from 307 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 10 Schüle et al., 2023). Here, we find binding of EOMES, TBXT or both to the genomic 308 loci of Nodal, Wnt3 and Wnt3a (Fig. 4E, H, K). To study the functional relevance of Tbx 309 TF binding we performed ATACseq (Assay for Transposase -Accessible Chromatin 310 with sequencing) in gastruloids as readout for Tbx factor activity (Schröder et al., 311 2025). At the Nodal locus, EOMES, and to minor degree also TBXT, bind the well -312 described proximal enhancer element (PEE) ~12 kb upstream of the promoter that 313 controls Nodal expression in the epiblast and the PS (Fig. 4E) (Norris & Robertson, 314 1999; Vincent et al., 2003) as well as to asymmetric element (ASE) enhancer of intron 315 1 (Norris et al., 2002) . Chromatin accessibility at the PEE and the ASE is severely 316 reduced in Eo - / - gastruloids, suggesting direct regulation by EOMES (Fig. 4E), while 317 accessibility is unaltered in Tbxt - / -. EOMES also binds to the Wnt3 locus, where 318 Eomes-deletion leads to reduced chromatin accessibility at a putative intronic 319 enhancer (Fig. 4H). These findings suggest direct positive regulation of Nodal and 320 Wnt3 by Eomes (Fig. 4F, I), but not by Tbxt. In reverse, a t the Wnt3a locus we find 321 reduced chromatin accessibility in Tbxt - / - gastruloids at TBXT bound site (Fig. 4K) 322 which could account for the reduced Wnt3a expression in Tbxt - / - gastruloids (Fig. 4L), 323 as also previously suggested in zebrafish (Martin & Kimelman, 2008). 324 In summary, these data suggest that Eomes acts by a self-reinforcing feed-forward 325 loop with the upstream regulators Nodal and Wnt3 that together build a functional 326 module that controls cell lineage specification and axial identity of early formed anterior 327 cell types. In contrast, Tbxt forms a n autoregulatory feed-forward loop with the 328 upstream signalling regulator Wnt3a, both being crucially required for posterior axis 329 elongation. T ogether the two functional modules of Eomes/Nodal/Wnt3 and 330 Tbxt/Wnt3a orchestrate patterning along the AP-axis of the PS. 331 332 Eomes induces anterior fates in gastruloids by repression of Tbxt functions 333 Our previous experiments demonstrated that ActA treatment of WT gastruloids 334 abrogates gastruloid elongation and expression of posterior markers in the presence 335 of Eomes (Fig. 2A, D, G ), but not in Eomes - / - gastruloids (Fig. 2C, F, I, illustrated in 336 Fig. 3J). We performed simultaneous IF for EOMES and TBXT in ActA- induced WT 337 and Eo - / - gastruloids (Fig. 5A) to test if the inhibition of axis-elongation by ActA results 338 from the absence of TBXT in WT gastruloids. However, in WT(ActA) gastruloids TBXT 339 is abundant and also shows the confined localization to one pole of the gastruloid at 340 120 h, albeit at reduced staining intensities when compared to CHIR induction (Fig. 341 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 11 5A, compare to Fig. 3A, E). Expectedly, ActA -pulsed WT gastruloids show high 342 EOMES levels (Fig. 5A) that remain elevated at timepoints past 84 h, when EOMES is 343 downregulated in WT(CHIR) gastruloids ( Fig. 5A, compare Fig. 3A). ActA induced 344 Eo - / - gastruloids undergo axis elongation similar to CHIR-treated WT gastruloids (Fig. 345 3A). In accordance, TBXT staining in Eo - / -(ActA) gastruloids remarkably resembles 346 patterns of WT(CHIR) gastruloids, where TBXT first shows circumferential staining that 347 later localizes to the posterior, extending tail bud-like region (Fig. 5A, compare to 3A). 348 Thus, differences of TBXT patterns or levels are unlikely to accounts for the gross 349 phenotypic differences between ActA treated WT and Eo - / - gastruloids. Therefore, we 350 further investigated the mechanisms by which EOMES suppresses gastruloid 351 elongation. We previously reported on competing activities of Eomes and Tbxt at the 352 level of newly forming chromatin accessibility in differentiating mESCs (Schüle et al., 353 2023). In the current study, we thus tested if the forced expression of Eomes can also 354 repress Tbxt functions required for axis elongation of gastruloids after the CHIR pulse. 355 Remarkably, the doxycycline-induced expression of Eomes (Wehmeyer et al., 2022) 356 simultaneous to the CHIR pulse (Fig. 5B, C) fully abrogates elongation of gastruloids 357 (Fig. 5D, E). Inhibition of elongation is not due to absence of TBXT, since we found 358 presence of TBXT after EOMES-induction by IF staining at 72 h with similar distribution 359 but at slightly reduced intensit ies compared to WT gastruloids (Fig. 5F). From 96 h , 360 TBXT is less confined and more dispersed after forced Eomes-expression, and at 120 361 h it is absent from the extending posterior pole . Mildly reduced levels of Tbxt mRNA 362 are also seen in RNAseq analysis of gastruloids after forced Eomes-expression (Fig. 363 5G). The heatmap analysis of anterior-posterior genes in Eomes-induced gastruloids 364 shows a transcriptional pattern that is strongly biased towards early expressed anterior 365 marker genes (Fig. 5H, Supplementary Table 4). This pattern is similarly found in ActA-366 pulsed WT gastruloids, and in Tbxt - / - CHIR-pulsed gastruloids (compare with Fig. 2D, 367 E). Thus, we next assessed if the lack of posterior marker expression after forced 368 Eomes expression is due to compromised TBXT transcriptional functions, and 369 performed WISH for Tbxt target genes, Rspo3 and Msgn1 (Evans et al., 2012; Schüle 370 et al., 2023). Remarkably, EOMES expression drastically reduces expression of both 371 Rspo1 and Msgn1 at 96 h and 120 h (Fig. 5I). This confirms previous observations of 372 repressed Tbxt-functions by Eomes in embryoid bodies by currently poorly understood 373 mechanisms (Schüle et al., 2023). 374 375 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 12 Eomes represses canonical Wnt3a signalling at multiple levels 376 Tbxt and canonical Wnt-signalling act synergistically during posterior axial elongation 377 (Arnold et al., 2000; Yamaguchi et al., 1999) . Thus, we reasoned that repress ion of 378 Tbxt functions might be additionally relayed through the suppression of the canonical 379 Wnt-cascade by Eomes. First, we analysed expression of Wnt3a, normally acting in a 380 positive feed-forward loop with Tbxt, in ActA-induced gastruloids in the presence (WT) 381 and absence of Eomes (Eo - / -) by fluorescent in situ hybridization (FISH) (Fig. 6A). 382 While Wnt3a expression is largely absent from ActA -induced WT gastruloids, we find 383 Wnt3a mRNA at the posterior pole of Eo - / -(ActA) gastruloids at 120 h (Fig. 6A), and 384 by analysis of RNAseq data at 96 and 120 h (Fig. 6B). The analysis of ATACseq data 385 of WT and Eo - / - gastruloids reveals a putative regulatory region 3’ adjacent to the 386 Wnt3a gene locus (Fig. 6C) that is bound by EOMES and TBXT (Fig. 4K). The dynamic 387 changes of chromatin accessibility at this region closely correlate to expression 388 dynamics of Wnt3a observed in different conditions, and at different timepoints (Fig. 389 6C, D), suggesting that EOMES and TBXT directly contribute to the control of Wnt3a 390 expression in an antagonistic manner . To assess more broadly canonical Wnt -391 signalling activities in gastruloids, we compared heatmaps of RNAseq data 392 representing Wnt pathway components and the signalling signature (Fig. 6 D). Here, 393 we find a signature of Wnt pathway activation (Axin2, Tcf7, Lef1 levels) in CHIR-treated 394 WT, and in ActA-treated Eo - / - gastruloids, but only largely reduced signature 395 expression in ActA-treated WT gastruloids. The positive signature of canonical Wnt 396 pathway activation in Eo - / - gastruloids is accompanied by the reduced expression of 397 inhibitory regulators of the Wnt pathway such as Tcf7l1 and Dkk1 (Fig. S3A, B). We 398 also performed heatmap analysis of CHIR-pulsed gastruloids, that show an active Wnt 399 signalling signature in WT and Eo - / - gastruloids, that is broadly reduced in Tbxt - / - 400 gastruloids in accordance with the observed phenotype (Fig. S3C). 401 To corroborate our finding that Eomes broadly impacts canonical Wnt signalling , we 402 engineered a mESC cell line that harbours a Tcf/Lef:H2B-GFP canonical Wnt reporter 403 cassette (Ferrer-Vaquer et al., 2010) and the DOX -inducible Eomes expression 404 cassette (Fig. 6E; TRE:Eo; Tcf/Lef:H2B-GFP) and generated gastruloids using CHIR 405 induction (Fig. 6F, G). We compared reporter expression in CHIR induced gastruloids 406 in the presence and absence of forced Eomes expression. While at 72 h there was no 407 major difference, at 96 h H2B-GFP reporter expression was significantly reduced after 408 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 13 forced Eomes expression (Fig. 6F, G), confirming the observation of globally reduced 409 canonical Wnt-signalling responses by transcriptional profiling (Fig. 6D). 410 In summary, we find that Nodal/ActivinA-signalling fully relies on Eomes to suppress 411 activities of the posterior functional module of Tbxt together with canonical Wnt3a 412 signalling for posterior axis elongation. Repression of the posterior module is achieved 413 at multiple different levels, such as reduced expression of Wnt3a and Tbxt (Fig. 5A, 414 6A), repression of signal transduction of the canonical Wnt -cascade (Fig. 6F) , and 415 repression of TBXT chromatin functions (Fig. 4K, 5C) (Schüle et al., 2023). 416 417

418 Previous studies uncovered the core signals and transcriptional regulators that 419 orchestrate the generation of cells that build the basic body plan in a correct spatial 420 and temporal order. Here, we studied the underlying molecular regulation and 421 interdependencies of the factors that pattern the gastrulation process of mammalian 422 embryos. We combined genetic and pharmacological approaches in highly versatile 423 mouse gastruloids to overcome the experimental limitations of mammalian 424 development. Our experiments reveal the cross-regulation of two competing functional 425 modules that control spatial patterning and the temporal progression during the first 426 two days of gastrulation as represented in a comprehensive model (Fig. 6H). The 427 consecutively acting regulatory modules guide the proper specification of different cell 428 types that is accompanied by a transition of different morphogenetic modes for cell 429 recruitment during early and late germ layer formation. Gastrulation is initiated when 430 combined activities of core signals of the anterior module, Nodal and Wnt3 induce 431 Eomes expression that is crucial for the specification of early AM (including cranial, 432 cardiovascular and anterior paraxial mesoderm) and DE from epiblast cells while they 433 ingress through the PS. Eomes reciprocally maintains Wnt3 and Nodal expression in 434 a mutually reinforcing feed forward loop that dominates the signalling environment 435 during the first day of gastrulation. The formation of the embryonic node at E7.5, marks 436 a rapid change in the signalling environment by the abrupt loss of expression of the 437 core components of the anterior functional module, Nodal, Wnt3 and Eomes, when 438 Tbxt and Wnt3a of the posterior functional module, gain functional competence. Both, 439 Tbxt and Wnt3a are expressed before E7.5, but activities are suppressed by at least 440 two mechanisms that largely rely on Eomes downstream of Nodal/Smad2/3 signalling: 441 First, by direct inhibition of Tbxt target genes (Schüle et al., 2023) (Fig. 5I), and, 442 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 14 second, by suppression of the canonical Wnt -signalling cascade through 443 transcriptional regulation of pathway members (Fig. 6). 444 Surprisingly, irrespective of the high levels of homology of amino acid sequence 445 between Wnt and Wnt3a (Roelink & Nusse, 1991) , both ligands show substantially 446 different functions during gastrulation. In the early PS Wnt3 promotes epiblast cells to 447 exit pluripotency, initiates Tbxt expression and contributes to the specification of early 448 PS derivatives in the context of the anterior regulatory module (Dias et al., 2025) 449 characterized by high Nodal signalling. In the late PS, Wnt3 is replaced by Wnt3a to 450 provide a dominant canonical Wnt signalling environment promoting NMP generation 451 and axial elongation. To date it is unclear where these substantial functional 452 differences originate from. These could result from the temporal context in which both 453 ligands act, or from differences in levels of expression and/or signalling functions of 454 both ligands. 455 Previous studies demonstrated key requirements of combined Tbxt and Wnt3a 456 activities for the elongation of the posterior body axis (Martin & Kimelman, 2008), that 457 is fuelled by cell recruitment from NMPs that initially form in vicinity of the node 458 (Cambray & Wilson, 2007; Wymeersch et al., 2019) . Presented lineage tracing (Fig. 459 1A) demonstrates that the mesoderm progenitors for the first 7-10 pairs of somites are 460 generated before node formation (<E7.5) and thus under the influence of the anterior 461 regulatory module from the PS. These findings align with phenotypes of Tbxt and 462 Wnt3a mutant embryos that suggested different molecular requirements, independent 463 of TbxtWnt3a for the generation of mesoderm of most anterior and of trunk and tail 464 somites (Guibentif et al., 2021). The molecular switch from the anterior to the posterior 465 regulatory module is rather abrupt and can be pinpointed to E7.5. It is currently not 466 clear if the morphogenetic transition of cell recruitment from PS ingression to NMP -467 derived mesoderm also occurs in a rapid step, or if this reflects a gradual change in 468 gastrulation modes (Mittnenzweig et al., 2021; Tzouanacou et al., 2009; Wilson et al., 469 2009; Wymeersch et al., 2021). 470 A major question that remains is how the abrupt breakdown of the anterior module is 471 initiated, when the feed -forward loop of Nodal, Wnt3 and Eomes suddenly breaks 472 down. Classical CHIR-pulsed gastruloids in which the anterior program is initiated but 473 rapidly overpowered by the posterior module suggests that Wnt3a signalling plays a 474 role in this transition. In agreement, we find extended Nodal/Wnt3/Eomes expression 475 in Tbxt - / - gastruloids (Fig. 4B). However, the rather rapid kinetics of the breakdown of 476 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 15 the Nodal/Wnt3/Eomes feed-forward loop, and the presence of Tbxt and Wnt3a 477 already before E7.5 strongly suggest additional mechanisms of regulation. These 478 could include suppression of Nodal-signalling, e.g. by increased expression of NODAL-479 antagonists Lefty1/2 or negative regulation of the critical NODAL co -receptor Cripto 480 (Tdgf1), by transcriptional mechanisms such as rapid nuclear degradation of Eomes, 481 or by rapid changes in the biophysical property of cells at the E7.5 PS that might impact 482 on signalling as described previously (Brennan et al., 2001; Gsell et al., 2025; Pfeiffer 483 et al., 2018). Based on the different levels of reciprocal regulation between the early 484 and late acting module that we discovered in this study and that were partially also 485 described previously (Ben-Haim et al., 2006; Martin & Kimelman, 2008; Pfeiffer et al., 486 2018; Yamaguchi et al., 1999) , we assume that it is the sum of synergistically acting 487 regulatory mechanisms that generate the highly robust patterns along the embryonic 488 AP-body axis. 489 In addition to providing a framework for the integration of transcription factors with 490 signals during mammalian gastrulation, this study provides an explanation for the 491 variable and often low representation of anterior fates, in particular AM and DE, in 492 classical gastruloids treated with a CHIR pulse. Suggested from the current study, and 493 also confirmed by Dias and colleagues (Dias et al., 2025), an experimental increase of 494 Nodal-signalling levels preceding the CHIR -pulse would allow to shift gastruloids 495 towards the anterior cell types of AM and DE to resemble earlier phases of PS 496 development and gastrulation in this model system. 497 Future studies will be required to integrate how additional cues, including biophysical 498 properties of 3D model systems, contribute to the remarkable robustness of stereotypic 499 embryonic patterning. This robustness is largely based on the formation of patterns as 500 intrinsic feature of self -organizing biological systems. Interestingly, the formation of 501 molecular patterns is independent of the underlying morphogenetic event since 502 gastruloids lack the equivalent of a PS or the epiblast. Existing and future embryoid 503 models will aim for adding back additional components that will allow to also 504 recapitulate morphogenetic aspects of embryogenesis in vitro. 505 506 507 508 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 16 Key messages 509 • The sequential switch between modes of gastrulation from the early primitive 510 streak to stem cell -generated axial elongation is accompanied by changes in 511 transcriptional and signalling signatures. 512 • Eomes/Nodal/Wnt3 and Brachyury/Wnt3a act as two competing regulatory 513 modules that coordinate the developmental progression from the primitive 514 streak stage to axial elongation. 515 • The transition from early gastrulation to the formation of posterior trunk 516 structures and axial elongation relies on the functional downregulation of the 517 early Eomes/Nodal/Wnt3 and activation of the later Brachyury/Wnt3a module. 518 • Functional modules show reciprocal regulation by feed -forward and feedback-519 regulation. 520 521

522 Cell lines and culture 523 E14-based A2lox mESCs were used throughout th is study (Iacovino et al., 2014) . 524 Tbxt/Bra - / - and Eo - / - (single knockout) and double knockout (dKO) mESCs and A2lox 525 mESCs harbouring dox -inducible expression cassette s for Eomes.GFP were 526 described previously (Tosic et al., 2019) . Tcf/Lef:H2B-GFP reporter cells were 527 generated by transgenic integration of the linearized Tcf/Lef:H2B -GFP reporter 528 plasmid (addgene, Plasmid #32610) (Ferrer-Vaquer et al., 2010) , followed by 529 screening for GFP-reporter expression in single clones. ESCs were transfected with 6 530 µg linearized vector by electroporation (Nucleofector ESC kit, Lonza), cultured on 0.1% 531 gelatine-coated dishes in Dulbecco’s modified Eagle’s medium (DMEM) containing 15 532 % fetal bovine serum (FBS, Gibco), 2 mM L-glutamine, 1X non-essential amino acids 533 (NEAA), 1 mM sodium-pyruvate, 1X penicillin/streptomycin (all from Gibco), 100 µM β-534 mercaptoethanol (Sigma), Leukemia inhibitory factor (ESGRO LIF, Merck Millipore, 535 1000 U/ml), and 2i: CHIR99021 (Axon Medchem, 1386, 3 µM) and PD0325901 (Axon 536 Medchem, 1408, 1 µM). The medium was changed daily and mESCs passaged every 537 other day. 538 539 Generation of gastruloids 540 Gastruloids were generated according to standard protocols (van den Brink et al., 541 2020) as Matrigel-free version. In brief, mESCs were dissociated and resuspended in 542 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 17 ESGRO Complete Basal Medium (Merck Millipore, SF002-500) and distributed as 300 543 cells per well into ultra-low attachment plates ( Greiner, 650970). 48 h after seeding, 544 the aggregates were pulsed with either 0 .3 mM CHIR (Axon Medchem, 1386) or 25 545 ng/µl Activin A (R&D Systems, 338 -AC) for 24 h . For induced gene expression from 546 the A2lox locus, gastruloids were treated with 1µg/ml doxycycline (Sigma -Aldrich, 547 D9891) simultaneous to the CHIR pulse. The medium was changed at 72 h and 96 h. 548 549 Whole mount in situ hybridization 550 Whole-mount in situ hybridization (WISH) was performed according to standard 551 protocols (Behringer, 2014). Probes and respective information can be requested from 552 the authors. In brief, gastruloids were fixed in 4% PFA / PBS o/n at 4°C, dehydrated 553 by a methanol series and stored in 100% methanol at -20°C. After rehydration 554 gastruloids were bleached in 6 % H2O2 for 5 min, digested by 1.6 μg/ml ProteinaseK / 555 PBT for 2 min, and postfixed in 4% PFA / 0.2 % glutaraldehyde for 20 min before 556 prehybridization for 4 h and hybridization o/n according to standard protocols. DIG -557 labelled RNA probes are detected using anti-Digoxigenin-AP Fab fragments (Roche) 558 in 1% sheep serum, 2% BBR in MAB (0.1 M Maleic acid, 0.3 M NaCl, NaOH, 1% 559 Tween-20 in H 2O, pH 7.5) and incubation at 4°C o/n. The antibody was washed by 560 repeated washes in MAB (>24 h, RT), and colour reaction s performed in BM purple 561 staining solution (Roche) for 2-6 h at RT. 562 563 Fluorescent in situ hybridization 564 For fluorescent in situ hybridization, we used HCR™ RNA-FISH probe sets, amplifiers 565 and buffers according to the manufacturer’s protocols (Molecular instruments). In brief, 566 gastruloids were collected, fixed in 4% PFA / PBS for 2 h, and stored in methanol as 567 described for WISH. After rehydration, gastruloids were digested by 1.6 μg/ml 568 Proteinase K / PBT for 2 min and postfixed in 4% PFA for 20 min before 569 prehybridization for 30 min at 37°C. Hybridization was performed at 37°C in probe 570 solution o/n. On the following day, gastruloids were repeatedly washed, incubated at 571 RT in hairpin solution o/n, and washed with SSCT (5x sodium chloride sodium citrate, 572 0,1% Tween 20) before image acquisition. 573 574 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 18 Immunofluorescence staining 575 Gastruloids were fixed in 4% PFA / PBS for 1 h at 4°C, permeabilized in 0.3% Triton 576 X-100/ PBT for 30 min, and blocked in 1% BSA / PBT for 1h at RT. Primary antibody 577 incubation was performed at 4°C o/n in 1% BSA / PBT, gastruloids washed 4x in PBT 578 before secondary fluorescence -conjugated antibody incubation for 3 h followed by 579 DAPI staining for 30 min at RT. Primary antibodies used were a-BRACHYURY (TBXT) 580 (R&D Systems; AF2085) and a-EOMES (abcam; ab23345) at suggested dilutions. 581 Secondary anti-goat and anti-rabbit Alexa Fluor antibodies conjugated with 488-, 546- 582 or 647-flurophores (Thermo Fisher) were used at 1:1000 dilution. 583 584 Imaging 585 Images were acquired on a Leica DMi8 Thunder Imager System or a Leica M165FC 586 Stereo microscope. Images were processed in the Leica LASX software and by Affinity 587 Photo. All brightfield and fluorescent images that were acquired with the Thunder 588 microscope are 3D-maximum projections of z-stacks. 589 590 RNAseq and analysis 591 For RNA sequencing, 12-48 gastruloids per replicate were used and total RNA isolated 592 using the RNeasy Mini kit (Qiagen). Concentration of the isolated RNA was quantified 593 using the NanoDrop (Thermo Fisher). Library preparation and sequencing was 594 performed by Novogene Services, UK , from a minimum of three biological replicates 595 for each condition. 596 FASTQ pre-processing was performed using the Galaxy biocomputing platform (Afgan 597 et al., 2018). Low-quality bases and adapter-containing reads were trimmed using Trim 598 Galore (Galaxy Version 0.4.3.1), and reads were aligned to the mouse reference 599 genome mm10 (mm10_UCSC_07_15, RNA STAR71 Galaxy Version 2.7.2b) using 600 default parameters. Duplicates were removed using RmDup (Galaxy version 2.0.1) 601 and genomic features were counted using htseq-count (Galaxy Version 0.9.1) with the 602 following settings: mode -union, stranded -no, minimum alignme nt quality -10 and 603 feature type -exon. Differentially expressed genes (DEGs) were analysed using 604 DESeq2 functions in R with filtering for adjusted p -value1.5 for 605 upregulated genes and log2(FC)<-1.5 for downregulated genes comparing WT versus 606 knockout gastruloids (Tbxt - / - or Eo - / -) individually at each time point. Venn diagrams 607 represent the overlap of up- and downregulated genes in Tbxt - / - and Eo - / - compared 608 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 19 to WT in the respective condition. For GO term analysis, all DEGs for a genotype and 609 signalling condition were analyzed using Enrichr (Xie et al., 2021) . Terms from GO 610 Biological Process 2023 are shown with their adjusted p -value that were computed 611 using the Benjamini-Hochberg method and the total number of overlapping genes are 612 indicated. Heatmaps were plotted with galaxy (version 23.1.2.dev0) using the 613 heatmap2 tool. Prior, z-scores were calculated from the normalized counts using table 614 compute. For heatmaps in Fig. 2 and Fig. 5, z-scores were computed row-wise from 615 each Supplementary Table. CHIR and ActA conditions were computed separately and 616 TRE.Eo gastruloids were computed in comparison to CHIR -treated WT gastruloids. 617 For the heatmap representing components of Wnt pathway activation in Fig. 6 D and 618 Fig. S3C, z-scores were computed row-wise from all depicted condition. The gene list 619 for heatmaps that depict the spatiotemporal progression of gastruloid development 620 were generated semi -manually by selecting genes from clustered DEGs (CHIR, 621 RNAseq data) that reflect the temporal progression in WT gastruloids. In addition, the 622 list contains selected genes from GO terms pattern specification process 623 (GO:0007389) and axis specification (GO:0009798). The same gene lists were used 624 for all conditions (see Supplementary Tables 2-4). Boxplots showing RNA expression 625 levels were generated using geom_bar of the ggplot2 package in R. RNA expression 626 levels are plotted as normalised counts of three or four replicates. Error bars indicate 627 the standard error of the mean (SEM). 628 629 ATACseq and analysis 630 ATACseq was performed according to published protocols (Buenrostro et al., 2013) 631 with modifications as outlined below. In brief, gastruloids at indicated time points were 632 dissociated and washed. 100,000 cells were lysed in 50 µl ATAC-seq lysis buffer and 633 tagmentation performed with 50 µl of transposition reaction mix containing Tagment 634 DNA Enzyme 1 (Nextera DNA Library Preparation Kit, Illumina) at 37°C and 600 rpm 635 for one hour. DNA was purified using the Qiagen MinElute Kit and amplified by PCR 636 (12 - 13 cycles) using Custom Nextera PCR primer and the NEBNext Ultra II Q5 Master 637 Mix (Nextera DNA Library Preparation Kit (New England BioLabs, M0544S). PCR 638 products were purified ( Qiagen MinElute Kit), DNA concentration of the libraries 639 measured with a Qubit Fluorometer (Thermo Fischer, Q32854) and fragment size 640 determined with a Bioanalyzer (Agilent). Libraries with large fragment sizes were size 641 selected for fragments <1000 bp using SPRI select beads. Sequencing was performed 642 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 20 by Novogene Services, UK. Low -quality bases and adapter -containing reads were 643 trimmed using Galaxy platform Trim Galore (Galaxy Version 0.4.3.1). Reads were 644 mapped to the mm10 genome using Galaxy platform Bowtie2 v2.3.4.3+galaxy0) 645 (Langmead & Salzberg, 2012) with the paired end option and the duplicates removed 646 with RmDup (Galaxy version 2.0.1). To generated IGV tracks, the coverage files were 647 created using bamCoverage54 (v3.3.2.0.0) with bin size 10 bases and normalization 648 to RPKM and paired-end extension of the fragment size. 649 650 Data acquisition and statistics 651 For the statistical analysis of gastruloid phenotype, lengths and widths were measured 652 using the LASX software (Leica) f rom a minimum of three replicate experiments and 653 n≥20 gastruloids for each condition and replicate (Supplementary Table 1) . The 654 fluorescent intensity of Tcf/Lef:H2B-GFP reporter expression was quantified using 655 ImageJ2 (version 2.14.0/154f). A ROI was defined in the brightfield image and the 656 mean gray value in the GFP channel measured. We measured the fluorescent 657 intensities for n≥20 gastruloids per condition and time point. The statistical significance 658 between control and CHIR +DOX treatment was calculated using a two-tailed Welch's 659 t-test (unequal variances t-test). Box plots and bar graphs were plotted using ggplot2 660 in R. 661 662 Visualization of scRNAseq data from the mouse gastrulation atlas data 663 We used published single-cell RNA-seq data of whole mouse embryos (Pijuan-Sala et 664 al., 2019). We plotted UMAPs of genes of interest for the time points E7.0, E7.5 and 665 E8.0. UMAPS and overviews of all cell types were downloaded from 666 https://marionilab.cruk.cam.ac.uk/MouseGastrulation2018/ (accessed 09/24). UMAPs 667 in svg format were edited in Affinity Designer as follow s: c ells originating from 668 extraembryonic tissues were removed, (visceral endoderm, ExE endoderm, ExE 669 ectoderm and parietal endoderm). All cells expressing the indicated genes of interest 670 were plotted in one colour (irrespective of levels of log2 normalized counts). 671 672 Data availability 673 Sequencing datasets of CHIR and ActA-treated gastruloids have been deposited in the 674 Gene Expression Omnibus (GEO) under accession code GSE283355 (ATACseq) and 675 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 21 GSE283356 (RNAseq). ChIPseq data w ere previously published and are accessible 676 under GSE194192. 677 678 Acknowledgments 679 We thank T. Bass for technical assistance and M. Kyba for the A2lox.Cre mESC line, 680 the Freiburg Galaxy Team, Bioinformatics, University of Freiburg (Germany) funded by 681 the Collaborative Research Centre 992 Medical Epigenetics (DFG grant SFB 992/1 682 2012) and the German Federal Ministry of Education and Research BMBF grant 031 683 A538A de.NBI-RBC. This study was supported by the German Research Foundation 684 (DFG) through the Heisenberg Program (AR 732/3 -1), project grant (AR 732/2 -1), 685 project A08 of CRC 992 (project ID 192904750) to SJA, and Germany's Excellence 686 Strategy (CIBSS – EXC-2189 – Project ID 390939984) to KMS and SJA . The project 687 was supported b y the Ministry of Science and Education of the State of Baden 688 Württemberg to SJA. KMS is supported by the EQUIP Program for Medical Scientist, 689 Faculty of Medicine, University Freiburg and by a CRC 992 MEDEP -Fellowship 690 (project ID 192904750). KM is supported by the Medical Research Council as part of 691 UK Research and Innovation (MCUP1201/23). AD is supported by an EMBO 692 Postdoctoral Fellowship (ALTF 948 -2022) and a Generalitat de Catalunya AGAUR 693 Grant (2021 SGR 00175). AMA is funded by an ERC Advanced Grant (834580) and 694 the Maria de Maeztu Program for Units of Excellence in R&D (CEX2018-000792-M). 695 Competing interests 696 AMA is an inventor in two patents on Human Polarised Three -dimensional Cellular 697 Aggregates PCT/GB2019/052670 and Polarised Three -dimensional Cellular 698 Aggregates PCT/GB2019/052668. 699 700 Author contributions 701 AEW, JKS, FE, CMS, L Z and KMS performed experiments. AEW generated 702 gastruloids, performed ISH, IF, imaging and RNA isolation. JKS generated gastruloids, 703 analysed phenotypes, performed IF, imaging and ATAC. MT, AEW and JKS performed 704 bioinformatical analyses. FE generated ESC lines and gastruloids. CMS, LZ and KMS 705 contributed to ESC line generation and interpretation of data. KM performed and KM 706 and SJA analysed time-lapse imaging. AEW, SP and SJA, analysed and interpreted 707 the data. AD and AMA shared unpublished data and discussed d ata interpretation. 708 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 22 AEW, SP, AMA and SJA, produced figures, wrote and edited the manuscript with input 709 from all authors. SJA conceived the study. 710 711

712 Afgan, E., Baker, D., Batut, B., van den Beek, M., Bouvier, D., Cech, M., Chilton, J., 713 Clements, D., Coraor, N., Grüning, B. A., Guerler, A., Hillman -Jackson, J., 714 Hiltemann, S., Jalili, V., Rasche, H., Soranzo, N., Goecks, J., Taylor, J., 715 Nekrutenko, A., & Blankenberg, D. (2018). The Galaxy platform for accessible, 716 reproducible and collaborative biomedical analyses: 2018 update. Nucleic Acids 717 Research, 46(W1), W537–W544. https://doi.org/10.1093/nar/gky379 718 Aires, R., Dias, A., & Mallo, M. (2018). Deconstructing the molecular mechanisms 719 shaping the vertebrate body plan. Current Opinion in Cell Biology , 55, 81–86. 720 https://doi.org/10.1016/j.ceb.2018.05.009 721 Amin, S., Neijts, R., Simmini, S., Van Rooijen, C., Tan, S. C., Kester, L., Van 722 Oudenaarden, A., Creyghton, M. P., & Deschamps, J. (2016). Cdx and T 723 Brachyury Co -activate Growth Signaling in the Embryonic Axial Progenitor 724 Niche. Cell Reports , 17(12), 3165 –3177. 725 https://doi.org/10.1016/j.celrep.2016.11.069 726 Arnold, S. J., Hofmann, U. K., Bikoff, E. K., & Robertson, E. J. (2008). Pivotal roles for 727 eomesodermin during axis formation,epithelium-to-mesenchyme transition and 728 endoderm specification in the mouse. Development, 135(3), 501 –511. 729 https://doi.org/10.1242/dev.014357 730 Arnold, S. J., & Robertson, E. J. (2009). Making a commitment: Cell lineage allocation 731 and axis patterning in the early mouse embryo. Nature Reviews Molecular Cell 732 Biology, 10(2), 91–103. https://doi.org/10.1038/nrm2618 733 Arnold, S. J., Stappert, J., Bauer, A., Kispert, A., Herrmann, B. G., & Kemler, R. (2000). 734 Brachyury is a target gene of the Wnt/beta -catenin signaling pathway. 735 Mechanisms of Development , 91(1–2), 249 –258. 736 https://doi.org/10.1016/s0925-4773(99)00309-3 737 Aulehla, A., Wehrle, C., Brand -Saberi, B., Kemler, R., Gossler, A., Kanzler, B., & 738 Herrmann, B. G. (2003). Wnt3a Plays a Major Role in the Segmentation Clock 739 Controlling Somitogenesis. Developmental Cell , 4(3), 395 –406. 740 https://doi.org/10.1016/S1534-5807(03)00055-8 741 Bardot, E. S., & Hadjantonakis, A. -K. (2020). Mouse gastrulation: Coordination of 742 tissue patterning, specification and diversification of cell fate. Mechanisms of 743 Development, 163, 103617. https://doi.org/10.1016/j.mod.2020.103617 744 Beccari, L., Moris, N., Girgin, M., Turner, D. A., Baillie -Johnson, P., Cossy, A. -C., 745 Lutolf, M. P., Duboule, D., & Arias, A. M. (2018). Multi -axial self-organization 746 properties of mouse embryonic stem cells into gastruloids. Nature, 562(7726), 747 272–276. https://doi.org/10.1038/s41586-018-0578-0 748 Behringer, R. (with Gertsenstein, M., Nagy, K. V., & Nagy, A.). (2014). Manipulating 749 the mouse embryo: A laboratory manual (Fourth edition.). Cold Spring Harbor 750 Laboratory Press. 751 Bénazéraf, B., & Pourquié, O. (2013). Formation and Segmentation of the Vertebrate 752 Body Axis. Annual Review of Cell and Developmental Biology , 29(1), 1 –26. 753 https://doi.org/10.1146/annurev-cellbio-101011-155703 754 Ben-Haim, N., Lu, C., Guzman-Ayala, M., Pescatore, L., Mesnard, D., Bischofberger, 755 M., Naef, F., Robertson, E. J., & Constam, D. B. (2006). The Nodal Precursor 756 Acting via Activin Receptors Induces Mesoderm by Maintaining a Source of Its 757 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 23 Convertases and BMP4. Developmental Cell , 11(3), 313 –323. 758 https://doi.org/10.1016/j.devcel.2006.07.005 759 Brennan, J., Lu, C. C., Norris, D. P., Rodriguez, T. A., Beddington, R. S. P., & 760 Robertson, E. J. (2001). Nodal signalling in the epiblast patterns the early 761 mouse embryo. Nature, 411(6840), 965–969. https://doi.org/10.1038/35082103 762 Buenrostro, J. D., Giresi, P. G., Zaba, L. C., Chang, H. Y., & Greenleaf, W. J. (2013). 763 Transposition of native chromatin for fast and sensitive epigenomic profiling of 764 open chromatin, DNA -binding proteins and nucleosome position . 765 https://doi.org/10.1038/nmeth.2688 766 Cambray, N., & Wilson, V. (2007). Two distinct sources for a population of maturing 767 axial progenitors. Development, 134(15), 2829 –2840. 768 https://doi.org/10.1242/dev.02877 769 Collignon, J., Varlet, I., & Robertson, E. J. (1996). Relationship between asymmetric 770 nodal expression and the direction of embryonic turning. Nature, 381(6578), 771 155–158. https://doi.org/10.1038/381155a0 772 Costello, I., Pimeisl, I. -M., Dräger, S., Bikoff, E. K., Robertson, E. J., & Arnold, S. J. 773 (2011). The T-box transcription factor Eomesodermin acts upstream of Mesp1 774 to specify cardiac mesoderm during mouse gastrulation. Nature Cell Biology , 775 13(9), 1084–1091. https://doi.org/10.1038/ncb2304 776 Dias, A., Pascual -Mas, P., Torregrosa -Cortés, G., McNamara, H. M., Wehmeyer, A. 777 E., Arnold, S. J., & Arias, A. M. (2025). A temporal coordination between Nodal 778 and Wnt signalling governs the emergence of the mammalian body plan (p. 779 2025.01.11.632562). bioRxiv. https://doi.org/10.1101/2025.01.11.632562 780 Downs, K. M., & Davies, T. (1993). Staging of gastrulating mouse embryos by 781 morphological landmarks in the dissecting microscope. Development, 118(4), 782 1255–1266. https://doi.org/10.1242/dev.118.4.1255 783 Duarte, P., Brattig Correia, R., Nóvoa, A., & Mallo, M. (2023). Regulatory changes 784 associated with the head to trunk developmental transition. BMC Biology, 21(1), 785 170. https://doi.org/10.1186/s12915-023-01675-2 786 Dunn, N. R., Vincent, S. D., Oxburgh, L., Robertson, E. J., & Bikoff, E. K. (2004). 787 Combinatorial activities of Smad2 and Smad3 regulate mesoderm formation 788 and patterning in the mouse embryo. Development, 131(8), 1717 –1728. 789 https://doi.org/10.1242/dev.01072 790 Evans, A. L., Faial, T., Gilchrist, M. J., Down, T., Vallier, L., Pedersen, R. A., Wardle, 791 F. C., & Smith, J. C. (2012). Genomic Targets of Brachyury (T) in Differentiating 792 Mouse Embryonic Stem Cells. PLoS ONE , 7(3). 793 https://doi.org/10.1371/journal.pone.0033346 794 Ferrer-Vaquer, A., Piliszek, A., Tian, G., Aho, R. J., Dufort, D., & Hadjantonakis, A.-K. 795 (2010). A sensitive and bright single -cell resolution live imaging reporter of 796 Wnt/ß-catenin signaling in the mouse . https://doi.org/10.1186/1471 -213X-10-797 121 798 Gsell, S., Tlili, S., Merkel, M., & Lenne, P. -F. (2025). Marangoni-like tissue flows 799 enhance symmetry breaking of embryonic organoids. Nature Physics, 21(4), 800 644–653. https://doi.org/10.1038/s41567-025-02802-2 801 Guibentif, C., Griffiths, J. A., Imaz-Rosshandler, I., Ghazanfar, S., Nichols, J., Wilson, 802 V., Göttgens, B., & Marioni, J. C. (2021). Diverse Routes toward Early Somites 803 in the Mouse Embryo. Developmental Cell , 56(1), 141 -153.e6. 804 https://doi.org/10.1016/j.devcel.2020.11.013 805 Herrmann, B. G., Labeit, S., Poustka, A., King, T. R., & Lehrach, H. (1990). Cloning of 806 the T gene required in mesoderm formation in the mouse. Nature, 343(6259), 807 617–622. https://doi.org/10.1038/343617a0 808 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 24 Iacovino, M., Roth, M. E., & Kyba, M. (2014). Rapid Genetic Modification of Mouse 809 Embryonic Stem Cells by Inducible Cassette Exchange Recombination. In M. 810 F. Ochs (Ed.), Gene Function Analysis (Vol. 1101, pp. 339 –351). Humana 811 Press. https://doi.org/10.1007/978-1-62703-721-1_16 812 Kinder, S. J., Tsang, T. E., Quinlan, G. A., Hadjantonakis, A. -K., Nagy, A., & Tam, P. 813 P. L. (1999). The orderly allocation of mesodermal cells to the extraembryonic 814 structures and the anteroposterior axis during gastrulation of the mouse 815 embryo. Development, 126(21), 4691–4701. 816 https://doi.org/10.1242/dev.126.21.4691 817 Kispert, A., & Herrmann, B. G. (1994). Immunohistochemical Analysis of the Brachyury 818 Protein in Wild -Type and Mutant Mouse Embryos. Developmental Biology , 819 161(1), 179–193. https://doi.org/10.1006/dbio.1994.1019 820 Koch, F., Scholze, M., Wittler, L., Schifferl, D., Sudheer, S. , Grote, P., Timmermann, 821 B., Macura, K., & Herrmann, B. G. (2017). Antagonistic Activities of Sox2 and 822 Brachyury Control the Fate Choice of Neuro -Mesodermal Progenitors. 823 Developmental Cell, 42(5), 514-526.e7. 824 https://doi.org/10.1016/j.devcel.2017.07.021 825 Kwon, G. S., Viotti, M., & Hadjantonakis, A. -K. (2008). The Endoderm of the Mouse 826 Embryo Arises by Dynamic Widespread Intercalation of Embryonic and 827 Extraembryonic Lineages. Developmental Cell , 15(4), 509 –520. 828 https://doi.org/10.1016/j.devcel.2008.07.017 829 Langmead, B., & Salzberg, S. L. (2012). Fast gapped -read alignment with Bowtie 2. 830 Nature Methods, 9(4), 357–359. https://doi.org/10.1038/nmeth.1923 831 Lawson, K. A., Meneses, J. J., & Pedersen, R. A. (1986). Cell fate and cell lineage in 832 the endoderm of the presomite mouse embryo, studied with an intracellular 833 tracer. Developmental Biology, 115(2), 325–339. https://doi.org/10.1016/0012-834 1606(86)90253-8 835 Liu, P., Wakamiya, M., Shea, M. J., Albrecht, U., Behringer, R. R., & Bradley, A. (1999). 836 Requirement for Wnt3 in vertebrate axis formation. Nature Genetics , 22(4), 837 361–365. https://doi.org/10.1038/11932 838 Lu, C. C., & Robertson, E. J. (2004). Multiple roles for Nodal in the epiblast of the 839 mouse embryo in the establishment of anterior -posterior patterning. 840 Developmental Biology, 273(1), 149–159. 841 https://doi.org/10.1016/j.ydbio.2004.06.004 842 Martin, B. L., & Kimelman, D. (2008). Regulation of Canonical Wnt Signaling by 843 Brachyury Is Essential for Posterior Mesoderm Formation. Developmental Cell, 844 15(1), 121–133. https://doi.org/10.1016/j.devcel.2008.04.013 845 McDole, K., Guignard, L., Amat, F., Berger, A., Malandain, G., Royer, L. A., Turaga, S. 846 C., Branson, K., & Keller, P. J. (2018). In Toto Imaging and Reconstruction of 847 Post-Implantation Mouse Development at the Single -Cell Level. Cell, 175(3), 848 859-876.e33. https://doi.org/10.1016/j.cell.2018.09.031 849 McNamara, H. M., Solley, S. C., Adamson, B., Chan, M. M., & Toettcher, J. E. (2024). 850 Recording morphogen signals reveals mechanisms underlying gastruloid 851 symmetry breaking. Nature Cell Biology , 26(11), 1832 –1844. 852 https://doi.org/10.1038/s41556-024-01521-9 853 Mittnenzweig, M., Mayshar, Y., Cheng, S., Ben-Yair, R., Hadas, R., Rais, Y., Chomsky, 854 E., Reines, N., Uzonyi, A., Lumerman, L., Lifshitz, A., Mukamel, Z., Orenbuch, 855 A.-H., Tanay, A., & Stelzer, Y. (2021). A single-embryo, single-cell time-resolved 856 model for mouse gastrulation. Cell, 184(11), 2825 -2842.e22. 857 https://doi.org/10.1016/j.cell.2021.04.004 858 859 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 25 Morgani, S. M., & Hadjantonakis, A.-K. (2020). Signaling regulation during gastrulation: 860 Insights from mouse embryos and in vitro systems. In Current Topics in 861 Developmental Biology (Vol. 137, pp. 391 –431). Elsevier. 862 https://doi.org/10.1016/bs.ctdb.2019.11.011 863 Norris, D. P., Brennan, J., Bikoff, E. K., & Robertson, E. J. (2002). The Foxh1 -864 dependent autoregulatory enhancer controls the level of Nodal signals in the 865 mouse embryo. Development, 129(14), 3455 –3468. 866 https://doi.org/10.1242/dev.129.14.3455 867 Norris, D. P., & Robertson, E. J. (1999). Asymmetric and node -specific nodal 868 expression patterns are controlled by two distinct cis -acting regulatory 869 elements. Genes & Development , 13(12), 1575 –1588. 870 https://doi.org/10.1101/gad.13.12.1575 871 Pfeiffer, M. J., Quaranta, R., Piccini, I., Fell, J., Rao, J., Röpke, A., Seebohm, G., & 872 Greber, B. (2018). Cardiogenic programming of human pluripotent stem cells 873 by dose-controlled activation of EOMES. Nature Communications, 9(1), 440. 874 https://doi.org/10.1038/s41467-017-02812-6 875 Pijuan-Sala, B., Griffiths, J. A., Guibentif, C., Hiscock, T. W., Jawaid, W., Calero-Nieto, 876 F. J., Mulas, C., Ibarra-Soria, X., Tyser, R. C. V., Ho, D. L. L., Reik, W., Srinivas, 877 S., Simons, B. D., Nichols, J., Marioni, J. C., & Göttgens, B. (2019). A single -878 cell molecular map of mouse gastrulation and early organogenesis. Nature, 879 566(7745), 490–495. https://doi.org/10.1038/s41586-019-0933-9 880 Probst, S., Sagar, Tosic, J., Schwan, C., Grün, D., & Arnold, S. J. (2021). 881 Spatiotemporal sequence of mesoderm and endoderm lineage segregation 882 during mouse gastrulation. Development, 148(1), dev193789. 883 https://doi.org/10.1242/dev.193789 884 Rivera-Pérez, J. A., & Magnuson, T. (2005). Primitive streak formation in mice is 885 preceded by localized activation of Brachyury and Wnt3. Developmental 886 Biology, 288(2), 363–371. https://doi.org/10.1016/j.ydbio.2005.09.012 887 Robertson, E. J. (2014). Dose-dependent Nodal/Smad signals pattern the early mouse 888 embryo. Seminars in Cell & Developmental Biology , 32, 73 –79. 889 https://doi.org/10.1016/j.semcdb.2014.03.028 890 Roelink, H., & Nusse, R. (1991). Expression of two members of the Wnt family during 891 mouse development —Restricted temporal and spatial patterns in the 892 developing neural tube. Genes & Development , 5(3), 381 –388. 893 https://doi.org/10.1101/gad.5.3.381 894 Rossi, G., Broguiere, N., Miyamoto, M., Boni, A., Guiet, R., Girgin, M., Kelly, R. G., 895 Kwon, C., & Lutolf, M. P. (2021). Capturing Cardiogenesis in Gastruloids. Cell 896 Stem Cell, 28(2), 230-240.e6. https://doi.org/10.1016/j.stem.2020.10.013 897 Savory, J. G. A., Bouchard, N., Pierre, V., Rijli, F. M., De Repentigny, Y., Kothary, R., 898 & Lohnes, D. (2009). Cdx2 regulation of posterior development through non -899 Hox targets. Development, 136(24), 4099 –4110. 900 https://doi.org/10.1242/dev.041582 901 Schröder, C. M., Zissel, L., Mersiowsky, S.-L., Tekman, M., Probst, S., Schüle, K. M., 902 Preissl, S., Schilling, O., Timmers, H. Th. M., & Arnold, S. J. (2025). EOMES 903 establishes mesoderm and endoderm differentiation potential through 904 SWI/SNF-mediated global enhancer remodeling. Developmental Cell , 60(5), 905 735-748.e5. https://doi.org/10.1016/j.devcel.2024.11.014 906 Schüle, K. M., Weckerle, J., Probst, S., Wehmeyer, A. E., Zissel, L., Schröder, C. M., 907 Tekman, M., Kim, G. -J., Schlägl, I. -M., Sagar, & Arnold, S. J. (2023). Eomes 908 restricts Brachyury functions at the onset of mouse gastrulation. Developmental 909 Cell, 58(18), 1627-1642.e7. https://doi.org/10.1016/j.devcel.2023.07.023 910 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 26 Simon, C. S., Downes, D. J., Gosden, M. E., Telenius, J., Higgs, D. R., Hughes, J. R., 911 Costello, I., Bikoff, E. K., & Robertson, E. J. (2017). Functional characterisation 912 of cis-regulatory elements governing dynamic Eomes expression in the early 913 mouse embryo. Development, 144(7), 1249 –1260. 914 https://doi.org/10.1242/dev.147322 915 Suppinger, S., Zinner, M., Aizarani, N., Lukonin, I., Ortiz, R., Azzi, C., Stadler, M. B., 916 Vianello, S., Palla, G., Kohler, H., Mayran, A., Lutolf, M. P., & Liberali, P. (2023). 917 Multimodal characterization of murine gastruloid development. Cell Stem Cell, 918 30(6), 867-884.e11. https://doi.org/10.1016/j.stem.2023.04.018 919 Takada, S., Stark, K. L., Shea, M. J., Vassileva, G., McMahon, J. A., & McMahon, A. 920 P. (1994). Wnt-3a regulates somite and tailbud formation in the mouse embryo. 921 Genes & Development, 8(2), 174–189. https://doi.org/10.1101/gad.8.2.174 922 Ten Berge, D., Koole, W., Fuerer, C., Fish, M., Eroglu, E., & Nusse, R. (2008). Wnt 923 Signaling Mediates Self-Organization and Axis Formation in Embryoid Bodies. 924 Cell Stem Cell, 3(5), 508–518. https://doi.org/10.1016/j.stem.2008.09.013 925 Teo, A. K. K., Arnold, S. J., Trotter, M. W. B., Brown, S., Ang, L. T., Chng, Z., 926 Robertson, E. J., Dunn, N. R., & Vallier, L. (2011). Pluripotency factors regulate 927 definitive endoderm specification through eomesodermin. Genes & 928 Development, 25(3), 238–250. https://doi.org/10.1101/gad.607311 929 Tosic, J., Kim, G. -J., Pavlovic, M., Schröder, C. M., Mersiowsky, S. -L., Barg, M., 930 Hofherr, A., Probst, S., Köttgen, M., Hein, L., & Arnold, S. J. (2019). Eomes and 931 Brachyury control pluripotency exit and germ-layer segregation by changing the 932 chromatin state. Nature Cell Biology , 21(12), 1518 –1531. 933 https://doi.org/10.1038/s41556-019-0423-1 934 Turner, D. A., Girgin, M., Alonso -Crisostomo, L., Trivedi, V., Baillie -Johnson, P., 935 Glodowski, C. R., Hayward, P. C., Collignon, J., Gustavsen, C., Serup, P., 936 Steventon, B., Lutolf, M., & Martinez, A. A. (2017). Anteroposterior polarity and 937 elongation in the absence of extraembryonic tissues and spatially localised 938 signalling in Gastruloids , mammalian embryonic organoids. Development, 939 dev.150391. https://doi.org/10.1242/dev.150391 940 Tzouanacou, E., Wegener, A., Wymeersch, F. J., Wilson, V., & Nicolas, J. -F. (2009). 941 Redefining the Progression of Lineage Segregations during Mammalian 942 Embryogenesis by Clonal Analysis. Developmental Cell , 17(3), 365 –376. 943 https://doi.org/10.1016/j.devcel.2009.08.002 944 Vallier, L., Mendjan, S., Brown, S., Chng, Z., Teo, A., Smithers, L. E., Trotter, M. W. 945 B., Cho, C. H. -H., Martinez, A., Rugg -Gunn, P., Brons, G., & Pedersen, R. A. 946 (2009). Activin/Nodal signalling maintains pluripotency by controlling Nanog 947 expression. Development, 136(8), 1339 –1349. 948 https://doi.org/10.1242/dev.033951 949 van den Brink, S. C., Baillie-Johnson, P., Balayo, T., Hadjantonakis, A.-K., Nowotschin, 950 S., Turner, D. A., & Martinez Arias, A. (2014). Symmetry breaking, germ layer 951 specification and axial organisation in aggregates of mouse embryonic stem 952 cells. Development, 141(22), 4231–4242. https://doi.org/10.1242/dev.113001 953 Veenvliet, J. V., Bolondi, A., Kretzmer, H., Haut, L., Scholze -Wittler, M., Schifferl, D., 954 Koch, F., Guignard, L., Kumar, A. S., Pustet, M., Heimann, S., Buschow, R., 955 Wittler, L., Timmermann, B., Meissner, A., & Herrmann, B. G. (2020). Mouse 956 embryonic stem cells self -organize into trunk -like structures with neural tube 957 and somites. Science, 370(6522), eaba4937. 958 https://doi.org/10.1126/science.aba4937 959 960 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 27 Vincent, S. D., Dunn, N. R., Hayashi, S., Norris, D. P., & Robertson, E. J. (2003). Cell 961 fate decisions within the mouse organizer are governed by graded Nodal 962 signals. Genes & Development , 17(13), 1646 –1662. 963 https://doi.org/10.1101/gad.1100503 964 Viotti, M., Foley, A. C., & Hadjantonakis, A. -K. (2014). Gutsy moves in mice: Cellular 965 and molecular dynamics of endoderm morphogenesis. Philosophical 966 Transactions of the Royal Society B: Biological Sciences , 369(1657), 967 20130547. https://doi.org/10.1098/rstb.2013.0547 968 Wang, J., Sinha, T., & Wynshaw -Boris, A. (2012). Wnt Signaling in Mammalian 969 Development: Lessons from Mouse Genetics. Cold Spring Harbor Perspectives 970 in Biology, 4(5), a007963–a007963. 971 https://doi.org/10.1101/cshperspect.a007963 972 Wehmeyer, A. E., Schüle, K. M., Conrad, A., Schröder, C. M., Probst, S., & Arnold, S. 973 J. (2022). Chimeric 3D gastruloids – a versatile tool for studies of mammalian 974 peri-gastrulation development. Development, 149(22), dev200812. 975 https://doi.org/10.1242/dev.200812 976 Williams, M., Burdsal, C., Periasamy, A., Lewandoski, M., & Sutherland, A. (2012). 977 Mouse primitive streak forms in situ by initiation of epithelial to mesenchymal 978 transition without migration of a cell population. Developmental Dynamics , 979 241(2), 270–283. https://doi.org/10.1002/dvdy.23711 980 Wilson, V., Olivera -Martinez, I., & Storey, K. G. (2009). Stem cells, signals and 981 vertebrate body axis extension. Development, 136(10), 1591 –1604. 982 https://doi.org/10.1242/dev.021246 983 Wymeersch, F. J., Skylaki, S., Huang, Y., Watson, J. A., Economou, C., Marek -984 Johnston, C., Tomlinson, S. R., & Wilson, V. (2019). Transcriptionally dynamic 985 progenitor populations organised around a stable niche drive axial patterning. 986 Development (Cambridge, England) , 146(1), dev168161. 987 https://doi.org/10.1242/dev.168161 988 Wymeersch, F. J., Wilson, V., & Tsakiridis, A. (2021). Understanding axial progenitor 989 biology in vivo and in vitro . Development, 148(4), dev180612. 990 https://doi.org/10.1242/dev.180612 991 Xie, Z., Bailey, A., Kuleshov, M. V., Clarke, D. J. B., Evangelista, J. E., Jenkins, S. L., 992 Lachmann, A., Wojciechowicz, M. L., Kropiwnicki, E., Jagodnik, K. M., Jeon, M., 993 & Ma’ayan, A. (2021). Gene Set Knowledge Discovery with Enrichr. Current 994 Protocols, 1(3), e90. https://doi.org/10.1002/cpz1.90 995 Yamaguchi, T. P. (2008). Genetics of Wnt Signaling During Early Mammalian 996 Development. In E. Vincan (Ed.), Wnt Signaling (Vol. 468, pp. 287 –305). 997 Humana Press. https://doi.org/10.1007/978-1-59745-249-6_23 998 Yamaguchi, T. P., Takada, S., Yoshikawa, Y., Wu, N., & McMahon, A. P. (1999). T 999 (Brachyury) is a direct target of Wnt3a during paraxial mesoderm specification. 1000 Genes & Development , 13(24), 3185 –3190. 1001 https://doi.org/10.1101/gad.13.24.3185 1002 Young, T., Rowland, J. E., Van De Ven, C., Bialecka, M., Novoa, A., Carapuco, M., 1003 Van Nes, J., De Graaff, W., Duluc, I., Freund, J. -N., Beck, F., Mallo, M., & 1004 Deschamps, J. (2009). Cdx and Hox Genes Differentially Regulate Posterior 1005 Axial Growth in Mammalian Embryos. Developmental Cell , 17(4), 516 –526. 1006 https://doi.org/10.1016/j.devcel.2009.08.010 1007 1008 1009 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 28 Figure legends 1010 Fig. 1. Morphogenetic gastrulation modes and signalling environments switch 1011 at E7.5 1012 (A) Snapshot images showing cell tracking from beginning of gastrulation (01h30) to 1013 the time of node formation ( 11h40) until 9-somite stage (42h00) (Ventral views, full 1014 movie as Supplemental movie 1). Cells of newly forming somitogenic paraxial 1015 mesoderm are marked in green and purple depending to the time of origin, either prior 1016 to node formation (E.7.5, purple cells). 1017 Blue cells mark node and notochord. At 42h00 somites are indicated by arrows. The 1018 first two somite -pairs are not visible as epithelial structures (boxed area left), and 1019 unsegmented paraxial mesoderm is indicated as boxed area (boxed area right). The 1020 rostral 8 somites are almost exclusively formed from presomitic mesoderm, that was 1021 already formed prior to E7.5, while later emerging mesoderm (purple) contribute s to 1022 more caudal somites. Scale bars 100 µm. 1023 (B) Gene expression signatures around E7.5 shown as UMAPs of published single cell 1024 mouse embryo RNAseq data sets (Pijuan-Sala et al., 2019). Eomes and Tbxt are co-1025 expressed in subpopulations of the PS and mesoderm prior to the appearance of the 1026 node at E7.5 , overlapping with Nodal and Wnt3 expression. At E7.5, Eomes, Nodal 1027 and Wnt3 are downregulated. After E7.5 , emerging mesoderm is established in the 1028 presence of Tbxt and Wnt3a. Extraembryonic tissues are excluded from this UMAP 1029 representation. 1030 (C) Immunofluorescent staining (IF) of CHIR-treated wildtype (WT) mouse gastruloids. 1031 Dynamic changes in TBXT and EOMES protein distribution are shown in 12 h intervals. 1032 Similar to embryonic expression, TBXT and EOMES are initially co -staining (60 h) 1033 before their domains separate. At 84 h EOMES is downregulated and TBXT polarizes 1034 to one pole, from where gastruloid elongation occurs. n≥5. Scale bars 100 µm. 1035 (D) Schematic of gastrulation stages in embryos and the corresponding timepoints in 1036 mESC-derived gastruloids. Before E7.5, nascent mesoderm is generated by cell 1037 ingression through the PS which is re capitulated in gastruloids <84 h, when EOMES 1038 and TBXT are co-localizing. Around E7.5 the node forms at the distal tip of the embryo. 1039 Eomes, Nodal and Wnt3 are rapidly downregulated and neuro -mesodermal 1040 progenitors (NMPs) form in vicinity of the node. Similar regulation is found at 84 h in 1041 gastruloids. >E7.5 and >84 h, posterior elongation of the embryo and of gastruloids is 1042 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 29 accomplished in an environment of high WNT3a and TBXT levels . Primitive streak 1043 (PS), Node streak border (NSB), Neuro-mesodermal progenitor (NMP). 1044 1045 Fig.2 Anterior -posterior axis elongation in gastruloids relies on Tbx TFs t o 1046 transduce instructive signals. 1047 (A-C) Brightfield images of representative gastruloids at 120 h from (A) wildtype (WT), 1048 (B) Tbxt - / - or (C) Eomes-deficient (Eo - / -) mouse embryonic stem cells (mESCs) that 1049 were generated by pulsed induction (48 h -72 h) with CHIR, or ActivinA (ActA). Arrows 1050 indicate posterior elongation. n≥60. Scale bars 100 µm. Compare Fig. S1 for additional 1051 phenotypic analysis. 1052 (D-F) Molecular profiling by RNAseq. Heatmap depicts spatial and temporal 1053 expression changes of characteristic anterior and posterior marker genes (n=97, 1054 Supplementary Table 2,3). (D) CHIR-treated WT gastruloids indicate the sequence of 1055 early, anterior marker gene expression and the transition to later expressed posterior 1056 marker genes. ActA -treated WT gastruloids shows a delayed emergence of anterior 1057 markers and strongly reduced posterior marker genes. (E) Tbxt - / - gastruloids show an 1058 upregulation of anterior , and downregulation of posterior marker genes. (F) Eo - / - 1059 gastruloids exhibit the opposite pattern: an upregulation of posterior and decreased 1060 expression of anterior marker genes. Remarkably, th e gross anterior -posterior 1061 patterning of Tbxt - / - and Eo - / - gastruloids is largely independent of the upstream pulse 1062 being either CHIR or ActA. 1063 (G-I) Whole-mount in situ hybridization (WISH) of an anterior marker ( Lhx1), and 1064 posterior marker genes ( Cdx2, Hoxb9) illustrates anterior-posterior pattering at 24 h 1065 intervals of CHIR or ActA pulsed gastruloids of indicted genotypes, (G) WT, (H) Tbxt - / - 1066 and (I) Eo - / -. Marker genes reflect the patterning phenotypes deduced from heatmap 1067 analysis. n≥6. Scale bars 100 µm. 1068 (J) Schematic summarizing anterior -posterior patterning of WT, Tbxt - / - and Eo - / - 1069 gastruloids in different signalling environments by stimulation with either CHIR, or 1070 ActA. A - anterior. P - posterior. 1071 1072 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 30 Fig. 3. EOMES and TBXT are dynamically distributed and mutually repress each 1073 other. 1074 (A-C) Immunofluorescent staining of CHIR-treated WT, Eo - / - and Tbxt - / - gastruloids 1075 in 6 h intervals to demonstrate dynamic changes of TBXT and EOMES during the early 1076 phase (48 - 96 h) of asymmetry breaking along the anterior-posterior axis. (A) TBXT 1077 and EOMES initially co-localise (60 h) in WT gastruloids before domains separate (72 1078 h). At 96 h EOMES is largely absent and TBXT polarizes to one pole. (B) Eo - / - 1079 gastruloids exhibit premature TBXT at slightly increased levels. TBXT is less clearly 1080 polarized at 96 h compared to WT gastruloids. (C) Tbxt - / - gastruloids initiate EOMES 1081 similar to WT and maintain high levels at later timepoints (96 h), in contrast to WT that 1082 loose EOMES from 90 – 96 h. n≥5. Scale bars 100 µm. 1083 (D, E) RNA levels of Tbxt and Eomes in CHIR -treated WT , and Eo - / - or Tbxt - / - 1084 gastruloids reflect the protein dynamics. (D) At 96 h, Tbxt mRNA levels are increased 1085 in Eo - / -, and (E) Eomes levels are maintained longer in Tbxt - / - gastruloids until 96 h. 1086 Protein and mRNA expression dynamics suggest mutual negative regulation between 1087 Eomes and Tbxt as (F) summarized in a schematic. RNA expression levels are plotted 1088 as normalised counts of three replicates. Error bars indicate SEM. 1089 1090 Fig. 4. Tbx factors and signals act as self -reinforcing functional modules of 1091 Eomes/Nodal/Wnt3 and Tbxt/Wnt3a. 1092 (A-C) Expression analysis of main upstream signalling regulators of Eomes (Nodal, 1093 Wnt3) and Tbxt (Wnt3a) by WISH in WT , Tbxt - / - and Eo - / - CHIR-pulsed gastruloids 1094 (A) In WT, Nodal expression precedes the exogenous CHIR-pulse at 48 h, expression 1095 peaks at 72 h , and is subsequently downregulated. Wnt3 and Wnt3a are expressed 1096 after the CHIR pulse from 72 h. Wnt3 is downregulated between 72 and 96 h, while 1097 Wnt3a remains expressed in the posterior tail-bud like region. (B) Tbxt - / - gastruloids, 1098 show increased expression of Nodal and Wnt3 that are maintained beyond 96 h, while 1099 they are devoid of Wnt3a expression. (C) Eo - / - gastruloids show opposing expression 1100 patterns and lack Nodal and Wnt3 expression, while strongly expressing polarized 1101 Wnt3a at 96 h. n≥6. Scale bars 100 µm. 1102 (D-L) Analysis of the regulation and interdependency of signals and Tbx TFs. mRNA 1103 expression of (D) Nodal, (G) Wnt3 and (J) Wnt3a in WT, Tbxt - / - and Eo - / - CHIR-1104 pulsed gastruloids recapitulate observations of WISH analysis. 1105 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 31 (E, H, K) ChIPseq and ATACseq coverage tracks at the gene loci of Nodal, Wnt3 and 1106 Wnt3a show ChIP-binding of TBXT and EOMES to putative enhancer sites (indicated 1107 in grey) that undergo dynamic changes of chromatin accessibility by ATACseq. (E, H) 1108 Accessibility at the Nodal and Wnt3 locus is reduced in the absence of Eomes. (K) The 1109 Wnt3a locus shows reduced accessibility in Tbxt - / - gastruloids. Counts normalized to 1110 RPKM are indicated in E, H, K. (F, I, L) Schematics summarizing the suggested cross-1111 regulation of signals and Tbx TFs. 1112 1113 Fig. 5. Repressive functions of ActivinA on axis elongation are mediated by 1114 EOMES 1115 (A) IF staining o f ActA-induced WT and Eo - / - gastruloids shows highly abundant 1116 EOMES in WT that is maintained until 120 h. TBXT is present and localized to one 1117 pole in WT and Eo - / - gastruloids, but posterior elongation is only promoted in absence 1118 of EOMES (Eo - / -). n≥5. Scale bars 100 µm. 1119 (B) Schematic of the dox-inducible cell line (TRE.Eo) for the forced expression of 1120 Eomes in an otherwise wildtype Eomes background, and (C) schematic of the protocol 1121 used to generate CHIR-pulsed gastruloids with forced Eomes.GFP expression 1122 (CHIR+DOX). Control gastruloids are treated with CHIR only. 1123 (D) Brightfield images at 120 h demonstrate that forced Eomes expression impairs axis 1124 elongation, even in CHIR-induced gastruloids as schematically illustrated in (E). 1125 (F) IF staining for EOMES, GFP and TBXT after forced Eomes.GFP expression in 1126 CHIR-pulsed gastruloids at indicated timepoints shows abundant TBXT at 72 and 96 1127 h in uninduced and Eomes-induced cells, but absence of TBXT at 120 h after forced 1128 Eomes expression. n≥10. Scale bars 100 µm. 1129 (G) mRNA expression levels of Eomes and Tbxt at 72 and 96 h grossly recapitulate IF 1130 staining in TRE.Eo gastruloids (+DOX vs. -DOX) and show reduced and prematurely 1131 downregulated levels of Tbxt. 1132 (H) Heatmaps of RNAseq analysis of CHIR -pulsed gastruloids at 72 and 96 h shows 1133 the absence of posterior marker gene expression and increased anterior signature of 1134 gastruloids after forced Eomes expression (+Dox). Control gastruloids ( -DOX) show 1135 the same expression dynamics as WT gastruloids (Fig. 2D). 1136 (I) WISH of two putative Tbxt-dependent target genes ( Rspo3 and Msgn1) with 1137 posterior expression in WT gastruloids confirms the absence of the posterior 1138 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 32 expression signature after forced Eomes-expression in gastruloids at 72, 96 and 120 1139 h. n≥5. Scale bars 100 µm. 1140 1141 Fig. 6. The canonical Wnt cascade of the posterior module is repressed at 1142 multiple levels by Eomes 1143 (A) Fluorescent in situ hybridisation (FISH) and (B) mRNA expression levels of Wnt3a 1144 in ActA-induced WT and Eo - / - gastruloids, suggests suppression of Wnt3a by Eomes. 1145 ActA-induced Eomes-deficient ( Eo - / -) gastruloids show similar Wnt3a patterns 1146 compared to CHIR-induced WT (see Fig. 4A). n≥8. Scale bar 100 µm. 1147 (C) ATACseq coverage tracks at the gene locus of Wnt3a show accessible chromatin 1148 at a putative enhancer of CHIR -treated WT gastruloids, and reduced chromatin 1149 accessibility when WT gastruloids are induced with ActA. ActA -pulsed Eo - / - 1150 gastruloids, restore levels of chromatin accessibility. 1151 (D) Heatmap representation of expression of components of the canonical Wnt -1152 signalling cascade. Heatmaps show normalized RNAseq expression data of indicated 1153 genes in CHIR - and ActA -treated WT gastruloids, and in ActA -treated Eo - / - 1154 gastruloids. CHIR-treated WT gastruloids exhibit an active Wnt signature (Axin2, Tcf7, 1155 Lef1), that is reduced in ActA -induced gastruloids. ActA -pulsed Eo - / - gastruloids 1156 establish an active Wnt- and tailbud signature (Cdx1, Cdx4, Tbxt, Rspo3). Gastruloid 1157 schematics above the heatmaps illustrate observed phenotypes and Wnt signature 1158 activities in different conditions. 1159 (E) Schematic of the Eomes-inducible (TRE.Eo) cell line that harbours the Tcf/Lef:H2B-1160 GFP transcriptional Wnt-reporter. 1161 (F) Bar chart representation of Wnt - reporter activity in CHIR-pulsed gastruloids after 1162 forced expression of Eomes. Fluorescent intensities of the H2B-GFP reporter are 1163 compared between uninduced ( -DOX) and Eomes-expressing (+DOX) gastruloids at 1164 72 h and 96 h, demonstrating reduced Wnt -reporter activation in Eomes-expressing 1165 cells at 96 h (***, p<0,001). 1166 (G) Examples of brightfield and GFP -fluorescent images of uninduced and Eomes-1167 expressing CHIR-pulsed gastruloids as used for reporter quantification in (F). n≥20. 1168 Scale bars 100 µm. 1169 (H) A comprehensive modal describing the interdependencies of Nodal and Wnt 1170 signals with Tbx TFs Eomes and Tbxt, that guide the progression of gastrulation from 1171 early primitive streak stages to axial elongation from progenitor pools. The anterior 1172 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 33 regulatory module (Eomes/Nodal/Wnt3) dominates the signalling landscape of the PS 1173 before E7.5 and represses activities of the posterior module ( Brachyury/ Wnt3a) that 1174 promotes caudal axial elongation from NMPs. The timeline and schematics of 1175 embryonic stages indicate the corresponding timepoint of regulation in gastruloids. 1176 1177 SUPPLEMENTARY FIGURES 1178 1179 Fig. S1. Extended phenotype analysis of anterior -posterior axis elongation in 1180 gastruloids 1181 (A, B) Schematics of induction protocols to generate gastruloids as models of primitive 1182 streak patterning and development by (A) pulsing with CHIR to mimic Wnt signalling 1183 or (B) induction with Activin A (ActA) to induce Nodal-signalling. 1184 (C) Schematics of cell lines used for the generation of gastruloids. Single knockouts 1185 were generated by the insertion of fluorescent reporters into one allele (EomesGFP and 1186 TbxtTomato) and frameshift deletions on the second allele. dKO cells are deficient for 1187 both Eomes and Tbxt. 1188 (D) Examples of brightfiel d images showing the phenotypic variation of WT, Tbxt - / -, 1189 Eo - / -, and dKO gastruloids at 120 h following induction with either CHIR or ActA. 1190 (E) Morphometric measurement of gastruloids of different genotypes and induction 1191 protocols. Gastruloids were measured along the maximal length and width as 1192 illustrated. Boxplots show the mean length (red) and mean width (blue) for each 1193 genotype after a CHIR- or ActA-pulse. For every condition three independent replicates 1194 with each n ≥20 gastruloids were measured (values provided in Supplementary table 1195 1). 1196 (F, G) Bar graphs show the calculated length/width ratios of gastruloids of different 1197 phenotype and induction protocol as read-out for phenotypic elongation. 1198 1199 Fig. S2. RNAseq analysis featuring the molecular differences of WT, Tbxt - / - and 1200 Eo - / - gastruloids 1201 (A-B) Venn diagrams showing gene numbers of up - and downregulated differentially 1202 expressed genes (DEGs) at 72, 96 and 120 h in Tbxt/- and Eo - / - gastruloids compared 1203 to WT and in (A) CHIR, or (B) ActA condition. Tbxt - / - gastruloids exhibit markedly 1204 increased numbers of DEGs after the CHIR -pulse, while Eo - / - gastruloids exhibit 1205 higher numbers of DEGs following an ActA-pulse. 1206 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 34 (C-D) GO term analysis for biological processes of all DEGs in Tbxt - / - and Eo - / - 1207 gastruloids in (C) CHIR, or (D) ActA conditions indicates gene function in axis 1208 specification and patterning. The -log(10) adjusted p -value for each term are plotted 1209 and total numbers of overlapping genes are indicated. 1210 1211 Fig. S3. Analysis of effectors and regulators of the Wnt pathway in CHIR and 1212 ActA treated WT, Tbxt - / - and Eo - / - gastruloids 1213 (A) ATACseq coverage tracks at the gene loci of Tcf7l1, and Dkk1 in CHIR pulsed WT 1214 and ActA-pulsed WT, and Eo - / - gastruloids at 72, 96 and 120 h. Putative repressor 1215 (Tcf7l1) or enhancer ( Dkk1) sites are indicated in grey. Dynamic accessibility at a 1216 putative repressor site at the Tcf7l1 locus is induced by CHIR in WT gastruloids, and 1217 by ActA in Eo - / - gastruloids, but not in ActA -pulsed WT gastruloids. The negative 1218 regulator of Wnt-signalling Dkk1 shows the opposite regulation at a putative enhancer 1219 element. Counts normalized to RPKM are indicated. 1220 (B) Bar graphs showing RNA expression levels of Tcf7l1 and Dkk1 that are generally 1221 reduced in Eo - / - gastruloids. 1222 (C) Heatmaps of RNA expression levels of genes indicative for active Wnt signalling 1223 (e.g. Rspo3, Axin2, Tcf7, Lef1) and posterior marker genes (Cdx1, Cdx4, Tbxt, Wnt3a) 1224 in CHIR, or ActA-pulsed WT, Eo - / -, and Tbxt - / - gastruloids. Tbxt - / - gastruloids show 1225 reduced Wnt - and posterior signatures. ActA -pulsed WT and Tbxt - / - gastruloids 1226 globally lack posterior and Wnt -signatures, that are partially restored in Eo - / -(ActA) 1227 gastruloids. 1228 1229 Supplementary Material 1230 1231 Supplementary Table 1 1232 Measured length and width of gastruloids of different genotypes (WT, Tbxt - / -, Eo - / -, 1233 dKO) and signalling pulses (CHIR or ActA ). The table shows the individual 1234 measurements for length, width and for length/width ratios. For each genotype and 1235 signalling pulse (CHIR or ActA) n≥20 gastruloids were measured in three independent 1236 experiments. The median of measured values is indicated for each replicate. Relates 1237 to Figure S1. 1238 1239 Supplementary Table 2 1240 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint 35 Temporal progression of anterior-to-posterior marker gene expression in CHIR-treated 1241 gastruloids. RNAseq data show averaged normalised counts of marker genes (n=97) 1242 for the three genotypes (WT, Tbxt - / -, Eo - / -) and three time points (72, 96 and 120 h). 1243 Relates to Figure 2. 1244 1245 Supplementary Table 3 1246 Temporal progression of anterior-to-posterior marker gene expression in ActA-treated 1247 gastruloids. RNAseq data shows averaged normalised counts of marker genes (n=97) 1248 for the three genotypes (WT, Tbxt - / -, Eo - / -) and three time points (72, 96 and 120 h). 1249 The same marker genes are listed as for CHIR-treated gastruloids in Table S2. Relates 1250 to Figure 2. 1251 1252 Supplementary Table 4 1253 Comparison of the spatiotemporal progression of anterior to posterior marker gene 1254 expression between gastruloids with induced EOMES -GFP expression (TRE.Eo, 1255 +DOX), uninduced controls ( -DOX) and CHIR -treated WT gastruloids. The same 1256 marker genes are listed as in heatmaps in Fig. 2. Table shows average normalised 1257 counts of RNAseq data. Relates to Figure 5. 1258 1259 1260 .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint .CC-BY-ND 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted May 8, 2025. ; https://doi.org/10.1101/2025.05.07.652670doi: bioRxiv preprint