Basement membrane perforations guide anterior–posterior axis formation

Chen, Dong-Yuan; Claussen, Nikolas H.; Titus, Shiny; Hu, Wenqi; Weatherbee, Bailey A. T.; Mandelbaum, Rachel S.; Scott Jr, Richard T.; Seli, Emre; Streichan, Sebastian J.; Zernicka-Goetz, Magdalena

doi:10.1038/s41467-025-61441-6

Download PDF

Article
Open access
Published: 22 July 2025

Basement membrane perforations guide anterior–posterior axis formation

Nature Communications volume 16, Article number: 6763 (2025) Cite this article

4439 Accesses
12 Altmetric
Metrics details

Subjects

This article has been updated

Abstract

Establishment of the anterior-posterior (AP) axis is a critical symmetry-breaking event in mammalian development. In mice, this process involves the directed migration of the distal visceral endoderm (DVE). Here, we use targeted perturbations to demonstrate that asymmetric perforations in the basement membrane guide DVE migration. During implantation, matrix metalloproteinases in extra-embryonic tissues create uneven basement membrane perforations, establishing directional cues for cohesive DVE migration. Using light-sheet microscopy and tissue cartography, we show that migrating DVE deforms surrounding tissues. Physical modeling and live imaging of DVE protrusions indicate that basement membrane perforations orchestrate active force generation within the DVE. Extending these findings to human embryos and stem cell-derived models, we identify basement membranes with enriched perforations near the anterior hypoblast in embryos, suggesting a conserved mechanism for AP axis specification. These findings reveal an unrecognized role of basement membrane remodeling and mechanical heterogeneity in guiding directional tissue migration during mammalian development.

Tissue-intrinsic beta-catenin signals antagonize Nodal-driven anterior visceral endoderm differentiation

Article Open access 13 June 2024

AI-based approach to dissect the variability of mouse stem cell-derived embryo models

Article Open access 19 February 2025

Unraveling the regulative development and molecular mechanisms of identical sea urchin twins

Article Open access 05 September 2025

Introduction

The specification of the anterior–posterior (AP) axis has been extensively studied across various model organisms, revealing both conserved and unique mechanisms. In nematodes and frogs, the sperm entry point plays a pivotal role in polarizing the egg upon fertilization^1,2. In the fruit fly and zebrafish, asymmetric localization of maternal morphogens contributes to pre-pattern AP polarity already before fertilization^3,4,5. However, in mammals, AP axis specification occurs later in development.

In mice, the blastocyst implants at embryonic day 4.5 (E4.5), undergoing major morphogenetic changes across its three tissues⁶: the pluripotent epiblast (EPI), which will form the fetus; the extra-embryonic ectoderm (ExE), which will form the placenta; and the visceral endoderm (VE), which develops into the yolk sac. The critical symmetry breaking event that establishes the AP axis occurs at E5.5, when a specialized subset of the VE cells at the distal tip of the embryo, the distal visceral endoderm (DVE), migrates proximally to form the anterior visceral endoderm (AVE)^7,8,9,10 (Fig. 1a). The AVE cells secrete Nodal and Wnt antagonists, establishing signaling gradients that restricts primitive streak formation to the posterior epiblast, thereby establishing AP polarity at the onset of gastrulation^11,12. The AVE is conserved in human embryos^13,14, but how it becomes localized to one side of the epiblast remains unknown.

**Fig. 1: Modulation of the level of basement membrane perforations affects DVE migration.**

DVE migration involves both active cell movement and passive cell rearrangements. Live imaging studies revealed that migrating DVE cells extend filopodia-like, actin-rich protrusions, probing their environment¹⁵. Genetic analyses have demonstrated essential roles for actin regulators, such as Rac1 and the WAVE complex, in guiding the DVE^16,17. Importantly, DVE migration is a global morphogenetic event involving cell rearrangements within the VE, rather than an isolated movement^{18,19,20,21,22}.

While the cellular mechanisms underlying DVE migration have been extensively studied, how its proximal directionality and collective migration are regulated remains an open question²³. Understanding this directionality is critical as it dictates AP symmetry breaking post-implantation.

Accumulating evidence indicates that the extracellular matrix (ECM) plays a crucial role in regulating tissue architecture, cell fate specification, and morphogenesis²⁴. ECM mechanics, such as topography and stiffness, can provide patterning cues to influence cell behaviors^25,26. ECM can either guide or restrict cell migration, while cells, in turn remodel ECM through secretion and enzymatic modifications^27,28,29. Furthermore, ECM also modulates signaling gradients, reinforcing its role beyond structural scaffolding²⁴.

Before gastrulation, the basement membrane – composed of LAMININs, COLLAGEN IV, and NIDOGENs^30,31,32—lies between the VE and the EPI (Fig. 1b). We demonstrated previously that posterior basement membrane perforations facilitate the formation and extension of the primitive streak at gastrulation onset³³. Basement membrane formation is also essential for EPI polarization, indicating its importance in early post-implantation embryogenesis⁶.

Here, we combine basement membrane perturbation, in toto microscopy, and tissue cartography to investigate basement membrane remodeling during pre- to post-implantation transition. We identify asymmetric basement membrane perforations on the future anterior side of the embryo—before DVE specification—that are essential for unidirectional migration of the DVE. Mechanistically, we uncover spatially biased expression of matrix metalloproteinases corresponding to uneven basement membrane perforations and heterogeneous mechanical properties of the VE that prepattern DVE migration. Together, these findings reveal unrecognized role of basement membrane remodeling and mechanical heterogeneity in guiding directional tissue migration during the development of the mammalian embryo.

Results

Basement membrane is required for unidirectional, collective DVE migration

We recently discovered that perforations in the basement membrane of rapidly growing mouse embryos appear after implantation and become skewed towards the posterior to enable cell migration during gastrulation³³. However, whether these perforations are important for earlier stages of embryogenesis remains an open question. To investigate this, we collected E5.5 embryos from Cerl-GFP mice, which label the DVE^34,35,36, and treated them with a pulse of collagenase to globally deplete the basement membrane. Embryos were then cultured in vitro and GFP⁺ DVE cells were tracked using time-lapse confocal microscopy to assess DVE migration patterns. In control embryos, DVE cells migrated collectively from the distal tip toward the proximal end (Fig. 1c, d, and Supplementary Movie1). Strikingly, in collagenase-treated embryos, DVE cells migrated at twice the speed of controls and lost their cohesiveness (Fig. 1e–h, and Supplementary Movie2).

Since basement membrane loss led to a striking DVE migration phenotype, we next investigated the effects of excess basement membrane. Matrix-modifying enzymes, such as the matrix metalloproteinases (MMPs), are essential for patterning basement membrane perforations during gastrulation³³. To examine their roles during DVE migration, we inhibited MMP activities by treating Cerl-GFP E5.5 embryos with Batimastat, a broad-spectrum MMP inhibitor. Batimastat-treated embryos exhibited fewer basement membrane perforations than controls (Fig. 1i, j). Despite continued embryo growth and pro-amniotic cavity expansion, DVE cells migrated collectively as a group, extending basal protrusions visible during the time-lapse imaging. However, instead of migrating proximally towards the EPI-ExE boundary as in normal embryos^20,37,38 (Fig. 1k), DVE migration slowed by threefold, halted before it reached the EPI-ExE boundary, and redirected laterally along the circumference—behavior typically observed at the EPI-ExE boundary (Fig. 1k–n, and Supplementary Movie3). Together, these findings suggest that a perforated basement membrane is required for regulating the speed, cohesion, and directionality of DVE migration.

Aberrant DVE migration often leads to mispositioning of the primitive streak^11,16. To assess this, we cultured E5.5 embryos for 24 h and performed immunofluorescence for BRACHYURY (also known as TBXT), a marker of the primitive streak and the mesoderm. We observed multiple patches or aberrant distal localization of TBXT in the EPI of most collagenase-treated embryos (74%), but not in controls (10%) (Fig. 2a–f and Supplementary Fig. 1a, b). This phenotype resembles Laminin α5 knockout embryos, which lack key basement membrane component³⁹. Importantly, collagenase treatment at E5.75, after DVE migration is complete, did not affect TBXT localization (Supplementary Fig. 1c–f). These results indicate that the basement membrane integrity is specifically required for unidirectional, collective DVE migration and proper AP axis formation.

**Fig. 2: Heterogeneity of the basement membrane perforation cues DVE migration for proper localization of the primitive streak.**

DVE cells migrate toward locally depleted basement membrane

Our data suggest that the basement membrane might play an instructive role in directing DVE migration. To test this hypothesis, we aimed to locally deplete the basement membrane and assess its effect on DVE directionality. We first optimized a method for regional transgene overexpression in embryos⁴⁰. As a proof-of-principle, we electroporated one side of E5.5 embryos with a plasmid expressing Histone 2B-RFP (H2B-RFP). As expected, H2B-RFP was detected at one side of the embryo as early as 2.5 h post-electroporation (Supplementary Fig. 1g, h, and Supplementary Movie4).

To deplete the basement membrane at just one side of the embryo, we co-electroporated H2B-RFP with a membrane-tethered matrix metalloproteinase (hMT1-MMP) on one side of E5.25 Cerl-GFP embryos, six hours before DVE migration. MMPs degrade basement membrane components such as LAMININs, COLLAGENs, and FIBRONECTIN⁴¹. As a control, we electroporated embryos with H2B-RFP alone. After 24 h, we analyzed H2B-RFP and Cerl-GFP by immunofluorescence. In control embryos, H2B-RFP and Cerl-GFP co-localized in 50% of embryos (Fig. 2g–i), consistent with random distribution. Strikingly, in hMT1-MMP-expressing embryos, co-localization increased to 75% (15/20) (Fig. 2j–l), indicating that DVE cells preferentially migrated toward the side with reduced basement membrane levels and therefore that local degradation of the basement membrane is sufficient to direct DVE migration.

As an alternative approach, we co-cultured E5.0-E5.25 embryos with collagenase-soaked agarose beads that we placed on one side of the embryo to create an ectopic gradient of collagenase activity (Supplementary Fig. 1i, j). In control embryos with BSA-soaked beads, the DVE migration was equally split between moving toward or away from the bead. In contrast, 50% of embryos cultured with collagenase-soaked beads exhibited DVE migration toward the bead, while only 16% migrated away (Supplementary Fig. 1k), further supporting a bias toward regions of basement membrane depletion.

Overall, these findings demonstrate that local basement membrane degradation influences DVE migration, reinforcing its role as a key regulator of early embryonic patterning.

Perforations in the basement membrane are enriched around migrating DVE

To systematically investigate the global patterning of the ECM, we used in toto imaging approaches to comprehensively map the basement membrane architecture. Using airyscan2 super-resolution microscopy, in toto imaging, and the software package ImSAnE^26,42, we generated two-dimensional projections of three-dimensional (3D) post-implantation embryos recovered from the mother at E5.5 when the DVE migration begins (Fig. 3a). We extracted and flattened a 3–6-micron-thick surface-of-interest, specifically highlighting the basement membrane. Re-projecting the surface back onto the 3D embryo confirmed that the flattening process preserved its 3D architecture (Supplementary Fig. 2a). A maximum Z projection of the basement membrane exhibited multiple perforations that were more pronounced in the basement membrane between the VE and EPI versus the VE and ExE, consistent with our previous findings³³ (Fig. 3b).

**Fig. 3: Basement membrane is heterogeneous before and during DVE migration.**

Our in toto super-resolution airyscan2 imaging combined with ImSAnE further demonstrated that the basement membrane perforations were not homogeneously distributed at E5.5 (Fig. 3c, d). Instead, perforations were significantly enriched beneath the region marked by the DVE reporter Cerl-GFP (Fig. 3e, f, and Supplementary Fig. 2b, c). COLLAGEN IV, another major component of the basement membrane, showed similar perforation patterns to LAMININ (Supplementary Fig. 2d). Interestingly, basement membrane perforations were enriched beneath the VE adjacent to the Cerl-GFP⁺ cells, towards which the cells would subsequently migrate, compared to the future posterior VE, where migration does not occur (Supplementary Fig. 2e–g). We did not observe a consistent difference in the level of the basement membrane perforations between the proximal and distal regions of the migrating DVE (Supplementary Fig. 2h, i). Since the DVE (Cerl-GFP⁺) cells migrate proximally, we hypothesized that the increased frequency of basement membrane perforations establishes the future migratory path, facilitating unidirectional DVE migration.

Basement membrane heterogeneity precedes DVE migration

To determine when basement membrane heterogeneity emerges, we analyzed E5.0-E5.5 embryos and mapped the appearance of basement membrane perforations relative to the expression of the DVE markers Cerl (also known as Cer1) and Lefty1. Consistent with a previous report³⁷, Lefty1 expression was sporadic across the VE at E5.0 (Supplementary Fig. 3a–c). We found LEFTY⁺ cells located asymmetrically at the proximal regions of the E5.0 embryo, often appearing in multiple LEFTY⁺ clusters around the embryo circumference (Supplementary Fig. 3d). In E5.25 and E5.5 embryos, Lefty1 mRNA and protein became consistently enriched at the distal tip (Supplementary Fig. 3a, b). Similarly, Cerl expression was sporadic in one or two VE cells at E5.0, but by E5.5 it was clearly localized to the distal tip (Fig. 3g and Supplementary Fig. 3e). These findings indicate that before DVE migrates, embryos do not show consistent asymmetric expressions of DVE markers.

Next, we examined basement membrane perforations across the circumference of E5.0 and E5.25 embryos. Interestingly, uneven basement membrane perforations were detected as early as E5.0, before the onset of DVE migration (Fig. 3h, i, and Supplementary Fig. 3d–g). These findings suggest that asymmetry in the basement membrane forms before the DVE is specified, potentially pre-patterning the migratory path.

Asymmetric MMP expression correlates with unidirectional collective DVE migration

To uncover the molecular mechanisms driving basement membrane heterogeneity, we analyzed single-cell RNA sequencing (scRNAseq) data on embryos from pre- to post-implantation stages (E3.5, E4.5, E4.75, E5.0, E5.25, and E5.5) (Fig. 4a). Our analyses focused on deferentially expressed genes in the embryos involved in basement membrane remodeling and identified two membrane-tethered metalloproteinases (MT-MMPs), Mmp15 and Mmp25, displaying dynamic expression patterns from E3.5 to E5.5 (Fig. 4b and Supplementary Fig. 4a). Other matrix-modifying enzymes such as ‘A-disintegrin and metalloproteinases’ (ADAMs), ‘A-disintegrin and metalloproteinase with thrombospondin motifs’ (ADAMTSs), and Tissue inhibitor of metalloproteinases (TIMPs) are either not upregulated during peri-implantation or are predominantly upregulated in the ExE or ExVE, and therefore are less likely to regulate basement membrane asymmetry for DVE migration (Supplementary Fig. 4a, b). Mmp25 displayed transient upregulation in 20% of the EPI at E4.5, increasing to 80% of the EPI at E4.75, before declining at E5.0 and E5.25 (Fig. 4b–d, Supplementary Fig. 5a–c). Mmp15 was transient upregulated in 10% of the primitive endoderm cells transforming into VE at E4.5, increasing to about 50% of VE cells at E4.75, and remained highly differentially expressed within a fraction of VE cells surrounding the EPI (EmVE) at E5.0, E5.25, and E5.5 (Fig. 4b–d, Supplementary Fig. 5a–c).

**Fig. 4: Asymmetric MMP expression is essential for DVE migration.**

To examine the spatial distribution of Mmp25 and Mmp15, we performed hybridization chain reaction (HCR) RNA fluorescence in situ hybridization (FISH)⁴³. Mmp25 mRNA expression was relatively homogeneous in the epiblast from E4.5 to E5.25 (Supplementary Fig. 5d–g). Mmp15 exhibited sporadic expression in the PE at E4.5 (Fig. 4e, Supplementary Fig. 5d, and Supplementary Movie5) and became enriched asymmetrically at E4.75, particularly at the tilted end of the VE and partially in the parietal endoderm (Fig. 4f and Supplementary Movie 6). At early E5.0-E5.25, Mmp15 expression remained higher on one side of the EmVE, coinciding with the region of increased basement membrane perforations (Fig. 4g–i, Supplementary Fig. 5h, i, and Supplementary Movie 7). Given that MT-MMPs degrade ECM proteins and activate other MMPs⁴⁴, their upregulation from E4.75 to E5.25 may contribute to the formation of uneven basement membrane perforations before DVE migration at E5.5.

MMP activity is required for unidirectional DVE migration

To test the role of MMPs in DVE migration, we inhibited MMP activity with Batimastat, a broad-spectrum inhibitor. We cultured Cerl-GFP embryos from E5.25 with DMSO control or a low dose (2 μM) of Batimastat (Fig. 4j, k and Supplementary Movie 8 and 9) for 15 h followed by time-lapsed live imaging. As expected, Batimastat treatment reduced MMP-dependent basement membrane perforations (Supplementary Fig. 5j, k). Strikingly, DVE failed to migrate unidirectionally as a cohesive group in the Batimastat-treated embryos, in contrast to the control embryos. Instead, individual DVE cells migrated toward the proximal end and, leading to patchy GFP⁺ cells distribution after 15 h (Fig. 4k, l, and Supplementary Movie 9). These findings suggest that MMP-dependent basement membrane remodeling is required before and during DVE migration to support the unidirectional and collective migration of the DVE. Continuous remodeling of basement membrane perforations appears essential for guiding proper DVE migration.

Uneven basement membrane perforations establish heterogeneous tissue mechanics

Our findings indicate that the basement membrane architecture is already heterogeneous before DVE migration, and this asymmetry modulates DVE migration. To determine if the basement membrane regulates the mechanical properties of the VE, we examined VE deformability using cell geometry as a proxy, as direct mechanical measurements in vivo are challenging. We optimized our ImSAnE-based 3D quantitative imaging platform^26,42 to perform morphometric analyses in toto and build a comprehensive atlas of VE cell behaviors.

To map VE cell behaviors, we flattened E5.0-E5.5 embryos expressing an E-cadherin-GFP knock-in (Cdh1-GFP)⁴⁵ (Fig. 5a, b). We then reconstructed embryos in silico with VE cell segmentation (Fig. 5c, d, and Supplementary Movie 10), and analyzed the morphometric parameters from E5.0 to E5.5, before and at the onset of DVE migration. Embryo volume increased about 2-fold from E5.0 to E5.5 (Fig. 5e). Aspect ratio along the proximal-distal axis increased from 1.6 to 2.1 (Fig. 5f). VE cell number doubled from about 100 to about 200 cells (Fig. 5g). These measurements provide a robust framework for staging embryos, in addition to the morphological features and the presence of DVE.

**Fig. 5: 3D morphometrics to identify cell behaviors in basement membrane-depleted embryos.**

To assess tissue-level fluidity, we examined hexagonal cell packing, which serves as a proxy for epithelial tissue fluidity^46,47,48. (Fig. 5h). E5.5 embryos were cultured for 3 h in the presence of DMSO (control) or collagenase (basement membrane-depleted). A decrease in the percentage of hexagonal cells reflects increase in cell rearrangements. Basement membrane depletion significantly reduced hexagonal packing, despite embryos having a similar number of VE cells (Fig. 5i, j). These data suggest that basement membrane perforations increase VE fluidity, potentially facilitating DVE migration.

To further investigate tissue fluidity at the cellular level, we mapped the cell shape index, defined by the ratio of perimeter divided by the square root of the cell’s surface area. Higher cell shape indices indicate more deformed, less round cells, which can correlate with lower cell shear modulus and increased tissue deformability (Supplementary Fig. 6a, b).

At E5.5 embryos, we analyzed VE shape indices across four distinct regions: extra-embryonic VE (ExVE), posterior embryonic VE (pEmVE), Cerl⁺ DVE, and the Cerl^- anterior embryonic VE (aEmVE) (Fig. 6a, b). To test if the shape index serves as an indicator of the accessibility of tissue for migratory DVE cells, we compared the shape index between ExVE and EmVE. We found that ExVE had a lower shape index than EmVE, suggesting less deformability (Fig. 6c). This supports previous reports showing rare cell movements and rearrangements in the ExVE^18,20. Basement membrane perforation levels correlated with cell shape indices, linking ECM remodeling to VE tissue mechanics (Supplementary Fig. 6c). Cerl+ DVE exhibited the highest shape index, consistent with its motile nature (Fig. 6d). The anterior VE (aEmVE and DVE) had a higher shape index than pEmVE, indicating greater tissue deformations in regions surrounding the DVE (Fig. 6d). VE cells above the DVE exhibited higher shape indices than those at the opposite end of the embryo (Fig. 6d, e), suggesting regional mechanical differences that influence migration.

**Fig. 6: Heterogeneous properties of VE cells.**

To investigate whether basement membrane asymmetry influences VE tissue mechanics before DVE migration, we analyzed E5.0–E5.25 embryos, prior to CERL expression at the distal tip. VE segmentation and shape index mapping revealed early asymmetry in tissue mechanics (Fig. 6f, g). Dividing the embryo into two halves along the circumferential long axis (Supplementary Fig. 6d) showed a slightly higher shape index on one side (Fig. 6h, i). Collagenase treatment reduced this asymmetry, indicating that uneven basement membrane perforations influence early VE tissue mechanics (Fig. 6j).

Together, these results suggest that uneven basement membrane perforations pre-pattern DVE migratory path. This indicates a positive feedback loop, reinforcing VE deformability and promoting directed DVE migration once cells begin to move.

Physical modeling predicts basement membrane perforations control migratory forces in the DVE

The differences in cell shape indices within the VE suggest two possibilities: the tissue is deformed because of a cell-autonomous property (low shear modulus), or the tissue is responding to external forces in a non-cell autonomous manner⁴⁹. To test if a more deformable VE alone is sufficient to guide DVE migration, we applied a physical modeling approach (Fig. 6k-l). The DVE cells are thought to respond to a distal-to-proximal chemoattractant⁵⁰. While our model remains agnostic to the nature of this attractant, it assumes a proximal-distal morphogen gradient toward which the DVE cells migrate. In addition, the model incorporates: (1) The DVE cells adhere to one another and (2) migration and adhesion forces are balanced by friction with the VE and the basement membrane. Using these assumptions, we simulated a minimal model that treats DVE cells as active particles (detailed in Methods).

We found that modulating the resistance (friction index) experienced by DVE cells was insufficient to direct DVE migration (Fig. 6m). Basement membrane perforations were required to generate sufficient migratory forces, effectively guiding DVE cells in the correct direction. DVE located on the side with more basement membrane perforations established themselves as leader cells, pulling the remaining DVE cells toward their side through stronger proximal-directed forces (Fig. 6l and Supplementary Movie 11). If the basement membrane was depleted everywhere, the model predicted faster migrations, but with DVE cells splitting into multiple patches, instead of cohesive migration (Fig. 6m–o, and Supplementary Movie12). This prediction aligns with our experimental observations (Fig. 1e).

This model also predicted that proximal DVE migration would cease when the basement membrane lacks perforations due to a lack of migratory forces (Supplementary Fig. 6e). These predictions were also consistent with our results when remodeling was inhibited (Fig. 1m). Furthermore, DVE cells had shorter protrusions in batimastat-treated E5.5 embryos compared to the controls (Fig. 6p–r), further supporting the role of MMP-driven basement membrane remodeling to control the speed and the directionality of the migrating DVE.

Thus, our findings demonstrate that uneven basement membrane perforations, created by early biased MMP expression, serve as a guiding cue for unidirectional, collective DVE migration. Moreover, MMP-dependent basement membrane remodeling remains essential not only before migration to establish directional cues, but also during migration to sustain proper speed and directionality.

Basement membrane perforations and AP axis formation in human embryos

Unlike the cylinder-like structure of the mouse embryo at implantation, the human embryo develops into a disc-like bilaminar structure⁵¹. Whether these differences in morphology and tissue architecture influence distinct mechanisms of AP axis determination remains unknown. Before implantation, human blastocysts consist of three major cell types, the EPI, hypoblast (HYPO), and the trophectoderm (TE) (Fig. 7a). We previously discovered that the human HYPO contains a subset of cells that express CER1 and LEFTY1, homologous to mouse DVE markers¹³. These cells likely serve as an anterior signaling center (anterior hypoblast) to pattern the EPI. Recent data revealed that the anterior hypoblast specification requires Nodal signaling⁵², a conserved mechanism in mice. However, how the localization of CER1⁺ and LEFTY1⁺ cells at one end of the HYPO before gastrulation is determined remains undetermined.

**Fig. 7: Basement membrane heterogeneity and the AP axis specification in human embryos.**

To explore whether the basement membrane plays a role in human embryogenesis, we first analyzed human embryo scRNAseq datasets from our lab and others^{13,52,53,54,55,56}. We found that genes encoding basement membrane components, such as COL4A1, LAMA1, NID1, and LAMB1, were upregulated and differentially expressed in the EPI and HYPO from peri- to post-implantation development. This mirrors gene expression profiles observed in mice during the same developmental transition. (Fig. 7b).

Using our in vitro protocol⁵⁷, we cultured 5-6 day post-fertilization (dpf) human blastocysts to post-implantation stages (7-9 dpf). At 7 dpf, cultured embryos established a Carnegie stage 4 (CS4)-like structure, with POU5F1⁺ (also known as OCT4) EPI and GATA6⁺ HYPO (Fig. 7c–e). Airyscan2 super-resolution imaging revealed LAMININ expression forming multiple fibrils and punctate structures in the extra-embryonic region (Fig. 7d). A dense LAMININ layer was observed between the EPI and HYPO, resembling the basement membrane in the mouse embryo with multiple small perforations (Fig. 7d). By 9 dpf, cultured human embryos developed Carnegie stage 5a (CS5a)-like structure, with the formation of the amniotic cavity at the POU5F1⁺ region and the primary yolk sac at the GATA6⁺ region (Fig. 7f–h). At this stage, the LAMININ layer between the EPI and HYPO became denser and more continuous, forming a sheet-like structure (Fig. 7g). Importantly, multiple perforations were detected with biased localization within this basement membrane (Fig. 7i), reminiscent of the mouse embryo at E5.0-E5.5.

At 7 dpf, CER1⁺ cells were initially spread out, but localized to a subset of HYPO by 9 dpf (Fig. 7d and Supplementary Fig. 7a). At 9 dpf, CER1⁺ cells clustered at one end of the GATA6⁺ HYPO cells, where the basement membrane showed more perforations (Fig. 7j–l). These findings suggest that heterogeneous basement membrane perforations may serve as a cue for anterior hypoblast localization, potentially influencing AP axis determination in human embryos.

hESC-based embryo model recapitulates basement membrane perforations

To further investigate whether a basement membrane becomes perforated as the AP axis forms in human embryos, we turned to human embryonic stem cell (hESC)-based embryo model. We generated post-implantation bilaminar disc-like structures from human expanded potential stem cells (hEPSCs)⁵⁸ and examined whether a basement membrane-like structure forms in this model at the interface of EPI- and HYPO-like lineages. At Day5, embryo-like structures derived from RUES2-GLR cell line⁵⁹ formed two compartments, one marked with SOX2-Citrine, resembling the EPI and another marked with SOX17-tdTomato, resembling the HYPO (Supplementary Fig. 7b). At the SOX2⁺-SOX17⁺ cell interface, we detected a dense layer of LAMININ with perforations, similar to those in 9dpf in vitro cultured human embryos (Supplementary Fig. 7c).

These data demonstrate that the basement membrane become heterogeneous in early post-implantation human embryos and embryo-like models. The regional enrichment of basement membrane perforations correlates with anterior hypoblast localization, suggesting a potentially conserved mechanism of body axis establishment in mice and humans.

Discussion

The directional migration of the DVE is required for AP axis formation during early embryonic development. Here, we identify an asymmetric distribution of basement membrane perforations and tissue mechanics that precede DVE formation and migration. This suggests that basement membrane asymmetry may help establish the AP axis by directing DVE migration, which positions the AVE as it becomes specified (Fig. 8)^14,20,36,38.

**Fig. 8: Schematics of basement membrane remodeling and AP axis specification.**

Before DVE migration, basement membrane perforations create a bias in tissue deformability within the non-migratory VE, establishing directional cues for cohesive DVE migration. Once DVE migration begins, basement membrane perforations modulate both the non-migratory VE and the migratory DVE, guiding the direction of migration. The continued requirement for basement membrane remodeling suggests that an initially weak directional bias is amplified during migration, potentially by MMPs expressed in the DVE.

Interestingly, DVE can still migrate directionally despite the disruption of the basement membrane. This suggests that basement membrane remodeling does not impact a chemoattractant, which might originate from the ExE border^17,60. Overall, our data support a positive feedback loop between ongoing basement membrane remodeling and unidirectional migration of the DVE.

Our findings indicate that the basement membrane fine-tunes the strength of DVE migratory force. How does basement membrane heterogeneity modulate migration strength? One possibility is that ECM per se is a physical barrier. The basement membrane could provide resistance to amoeboid-like migration, with the DVE cells actively probing for a path of least resistance⁶¹. Another possibility is that the basement membrane modulates tissue mechanics via mechanotransduction. The mechanical properties of the basement membrane may influence cell migration via mechanosensitive pathways. A mechanosensitive mechanism regulating junctional remodeling during morphogenesis could play a role²⁶. Finally, the basement membrane heterogeneity may influence tissue deformability both before and during DVE migration⁶². Since directly measuring mechanical properties is challenging, we used cell geometry and physical modeling to infer tissue fluidity (deformability). Our morphometric analyses indicate that the VE exhibit heterogeneity in the shape index before DVE migration, aligning with a recent study report that spatial differences in cell shape increase from the onset of DVE migration⁶³. While the shape index can indicate tissue deformability, it cannot distinguish between passive deformation and active forces. Future studies should explore whether early asymmetry in the shape index predicts direction of DVE migration.

How does post-implantation basement membrane heterogeneity relate to pre-implantation molecular asymmetry in AP axis specification? Nodal signaling is essential for DVE specification^38,64,65,66. Lefty1, which is induced by Nodal signaling, is a DVE marker expressed in a subset of PE cells in pre-implantation embryos. Its asymmetric expression has been proposed as a cue for the specification of the AP axis³⁸. However, whether the expression of Lefty1 peri- to post-implantation is stochastic or prespecified by another asymmetry remains unclear³⁷. Here we show that the expression of Lefty1 correlates with the tilting of the inner cell mass (ICM) of the implanting blastocysts^37,38,67. This suggests that ICM tilting could lead to asymmetric Nodal expression in some pre-implantation embryos, potentially influencing later DVE migration asymmetry. Despite these findings, the connection between pre-implantation asymmetry and DVE migration asymmetry requires future studies.

Our study proposes a model in which the basement membrane responds to molecular asymmetry during the pre- to post-implantation transition, creating physical heterogeneity that regulates DVE migration after implantation. It is intriguing to speculate that the asymmetry in MMP expression during peri-implantation maybe be Nodal signaling dependent, as previously reported for MMP expression before and during gastrulation³³. Our scRNAseq analysis indicates a low but persistent correlation between Nodal signaling effectors, such as Lefty1 and Eomes, from peri- to post-implantation stages (Supplementary Fig. 7d). Furthermore, disruption of Nodal signaling-related planar cell polarity leads to defects in the cohesiveness of DVE migration¹⁸. These observations support the idea that Nodal signaling contributes to basement membrane asymmetry directing DVE migration.

We observed asymmetric expression of the matrix metalloproteinase Mmp15 in the EmVE, yet embryos from Mmp14/Mmp15 double knockout become only exhibits defects around E9.5⁶⁸. Given that Batimastat is a broad-spectrum MMP inhibitor, it is likely that other ECM components or matrix modifying enzymes contribute to establishing and reinforcing the basement membrane perforation asymmetry. The asymmetry could also be amplified by other MMPs expressed in the DVE.

Our findings reveal that extra-embryonic tissues, including the PE and VE, display asymmetric Mmp15 expression, potentially contributing to basement membrane heterogeneity. The asymmetric perforations in the basement membrane provide directional cues for collective DVE migration, crucial for patterning in the EPI. Previously, we demonstrated that a different set of MMPs (MMP2 and MMP14) exhibit asymmetric expression in the EPI around E6.25, just before the primitive streak formation and extension³³. Together, these results underscore the key roles of basement membrane remodeling in patterning the embryo across different stages of embryogenesis. Specifically, basement membrane asymmetries inform cell behaviors during the earliest symmetry breaking events in the EPI, guide AP axis specification, and facilitate expansive embryo growth before and during gastrulation (Fig. 8).

Understanding the AP axis establishment in the human embryo remains challenging due to limited access to human embryos, particularly at the post-implantation stage when AP axis forms. We previously identified an anterior hypoblast in the human embryo with a molecular signature akin to mouse DVE¹³. A similar structure can also be identified in stem cell-based human embryo models^{58,69,70,71,72}. Our study demonstrates that the human basement membrane shares molecular components with mice and our preliminary data reveal a perforated architecture with higher perforation frequency near the anterior hypoblast. While it remains unclear whether cell migration is the mechanism that establishes the anterior hypoblast, our findings suggest that basement membrane remodeling may play a role in the patterning of the human embryo.

Several stem cell-based models mimic human post-implantation at the bilaminar disc stage^{58,69,70,71,72}. However, AP symmetry breaking generally occurs with low efficiency, as evidenced by the sporadic presence of CER1⁺ cells at one end of the HYPO-like compartment, with TBXT⁺ EPI cells at the opposite end. Our findings offer a potential framework to improve the robustness and reproducibility of these models, which could lead to deeper insights into human embryogenesis.

Methods

Animal husbandry

All animal procedures were approved by the Institute Animal Care and Use Committee (IACUC), protocol number 1772, at the California Institute of Technology (Caltech) and performed in accordance with NIH guidelines. Animals were housed in HEPA-filtered microisolator caging systems in a temperature and humidity-controlled environment (temperature 21.7–23.9 °C; humidity 30–70%) in an automated light-dark cycle room, with 13-h light and 11-h dark period.

Mouse embryo collection and culture

Timed pregnant females were obtained from natural mating by crossing 7-week-old CD-1 females (Charles River Strain Code:022) were mated with either B6CBAF1/J (The Jackson Lab Strain No:100011), Cdh1-GFP, or Cerl-GFP males. Embryos The presence of a vaginal plug in the female was denoted as embryonic day 0.5 day post coitum (E0.5). Embryos ranging from E4.75-E5.5 were dissected out from the uteri in warmed 37 °C M2 media (Sigma #M7167). After surgically removing Reichert’s membrane with tungsten needles, embryos were kept in M2 media at 37 °C within 45 min before subsequent experiments.

Embryos were cultured at 37 °C and 5% CO₂ using a modified DMEM-Rat serum-Human serum (DRH) culture media⁷³, which consists of 25% DMEM (DMEM/F-12), 50% of rat serum (made in-house), 25% of human serum (Sigma Catalog No.H4522), Penicillin-Streptomycin (Gibco #15140122). For live imaging, embryos were cultured in Fluorobrite DMEM (Gibco Catalog No. A1896701), supplemented with 50% of rat serum (made in-house), 25% of human serum (Sigma Catalog No.H4522), 1x GlutaMAX (Gibco #35050061), 1x MEM Non-Essential Amino Acids Solution (Gibco #11140050), and Penicillin-Streptomycin (Gibco #15140122). For E5.0 embryo culture, 8 nM β-estradiol (Sigma #E8875), 200 ng/mL Progesterone (Sigma #P8783), and 25 μM N-acetyl-L-cysteine (Sigma #A7250) were supplemented to the DRH media. Rat serum was collected either from male Sprague Dawley rat (Strain Code: 400) or Long-Evans rat (Strain Code: 006) aged 6 to 18 months and followed the protocol for collection⁷⁴.

For acute basement membrane depletion, embryos were incubated with 1 mg/mL of collagenase type IV (Gibco # 17104019) in M2 media at 37 °C for 25 min, washed with M2 3 times and incubated with 0.1 mg/mL collagenase in DRH media at 37 °C and 5% CO₂.

For co-culture experiments with collagenase-soaked beads, E5.0-E5.25 Cerl-GFP embryos were embedded in 0.15% low-melting point (LMT) agarose (Sigma #A4018) and cultured with agarose beads (Bio-Rad #1537302) soaked in either bovine serum albumin or 10 mg/ml Collagenase IV (Fisher Scientific # 17104019) in PBS either at 37 °C for 1 h or overnight at 4 °C. Beads were washed in PBS 3 times and 1 time with 0.15% LMT-agarose in DRH media before incubation with the embryos.

Embryo fixation and immunostaining

Embryos were fixed in 4% paraformaldehyde (Electron Microscopy Sciences #11586711) in PBS for 20 min at room temperature, washed twice in PBS, and followed by 30 min of permeabilization in PBS containing 0.3% Triton X-100 and 0.1 M glycine at room temperature. Embryos were blocked for 1 h in PBS containing 5% Normal donkey serum, 1% bovine serum albumin (BSA), and 0.1 % Tween20 and incubated with primary antibodies diluted in blocking buffer. After 3 10-minute washes in PBST (PBS with 0.1 % Tween20), embryos were incubated with fluorescently conjugated Alexa Fluor secondary antibodies (Fisher Scientific) for 2 h at room temperature in blocking buffer. Embryos were washed 3 times for 10 min in PBST and were mounted in mounting solution (1:1 PBS/Glycerol). Primary antibodies used are described in the reporting summary. Counterstain used: DAPI (1:5000; Thermo Fisher #D3571). Immunofluorescence images were processed and analyzed using Fiji. Cell number in the in vitro cultured human embryos were counted manually using Imaris. Hypoblast is defined by GATA6⁺ cells that line and are in direct contact with the basal distal side of the epiblast; these cells were counted, as opposed to the total GATA6⁺ cells, which includes parietal lineage. CER1⁺ cells in contact with the epiblast were counted.

Electroporation

We adopted a protocol that allows us to introduce transgenes predominantly at one side of the embryos⁴⁰. E5.25 Cerl-GFP embryos were recovered in M2 media and the Reichert’s membrane was manually removed with tungsten needles. Embryos were staged under the epifluorescent microscope to ensure no Cerl-GFP⁺ cell was present at the distal tip of the embryo. The injection pipette was made from capillaries (Harvard Apparatus) pulled on an automatic micropipette puller (Sutter instrument). Leica micromanipulator and micropipette glass needles back-filled with 1 μg plasmid carrying PCI-H2B-RFP in control or 1 μg PCI-hMT1MMP-EGFP-IRES2-H2BRFP were injected into the pro-amniotic cavity of the embryo with FemtoJet 5247 microinjector (Eppendorf) with an injection pressure of 300 hPa, injection time 0.10 sec, and compensation pressure 40 hPa. Injection needle was inserted from the ExE side of the embryo to reduce damaging the EmVE. Embryos were washed twice with HEPES-buffered saline (HBS) and were transferred to a microslide electrode chamber (BTX order #45-0104) filled with 40 μL of HBS. The proximal-distal axis of the embryos was aligned perpendicular to the electric current to allow targeted electroporation at one side of the embryos. Embryos were electroporated using a BTX ECM 2001 electroporation system, with a 30 Volt charging voltage, and 6 square-wave pulses for 1 second each with a 200 milliseconds interpulse gap. We constructed PCI-hMT1MMP-EGFP-IRES2-H2BRFP by cloning hMT1-MMP-eGFP into the multiple cloning site of PCI-H2B-RFP.

Microscopy

We carried out selective plane illumination microscopy (SPIM) to acquire the data for the morphometric analysis. SPIM was carried out on a custom-built microscope⁷⁵. In brief, the SPIM microscope employs a pair of two orthogonal illumination and observation objectives, a rotation stage to acquire multiple views of the same sample, and electronic synchronization of the illumination laser and camera readout for scattering reduction (“confocal” SPIM⁷⁵). Samples were mounted in an agarose column and imaged, creating z-stacks with a resolution of 0.26*0.26*1.5 μm. We acquired 8 views separated by 45-degree rotations of each sample. Fluorescent beads (Fluoresbrite multifluorescent 0.5-μm beads 24054, Polysciences Inc.) were diluted into the agarose at concentration 1:1000. To combine the multiple views of each sample into a single volumetric image, we used the Fiji plugin Multi-view Reconstruction⁷⁶. This plugin detects the fluorescent beads which serve as fiduciary markers to register the 8 views of the sample to a common frame and combines the 8 views into a single volumetric image with resolution 0.26*0.26*0.261 μm using a Bayesian deconvolution algorithm. We used Zeiss LSM980 with airyscan2 and Leica SP8 for confocal imaging. Objectives used include Fluotar VISIR 25x/0.95 water immersion objective (Leica), HC PL APO CS2 40x/1.10 water immersion objective (Leica), Plan-Apochromat 20x/0.8 M27 (Zeiss), LD LCI Plan-Apochromat 25x/0.8 Imm Corr DIC M27 (Zeiss), and C-Apochromat 40x/1.2 W Corr water immersion objective (Zeiss).

HCR RNA-FISH

HCR probes were formulated by Molecular Technologies, Beckman Institute, California Institute of Technology. All equipment and tools were sterilized with 70% ethanol followed by RNAase Zap before the experiment. We followed in situ HCR v3.0 protocol summarized here⁴³ [https://files.molecularinstruments.com/MI-Protocol-RNAFISH-Mouse-Rev7.pdf], with 2 slight modifications. Briefly, embryos were dissected in M2 media at room temperature and were fixed in 4% paraformaldehyde overnight at 4 °C. Embryos were incubated with 10 µg/mL proteinase K at 37 °C for 3 min. Samples were incubated in 50% glycerol/PBS, and were cleared with 80% glycerol/PBS overnight at 4 °C.

Tissue Cartography and Morphometrics extraction

To analyze the 3D data both from SPIM and confocal microscopy, we employed the ImSAnE⁴² and blender-tissue-cartography (btc) software suites⁷⁷ to carry out tissue cartography. In brief, embryos were detected in volumetric imaging data using ilastik⁷⁸, and ImSAnE/btc was used in combination with the morphological snakes algorithm⁷⁹ to create a triangular mesh of the embryo surface. A 1 μm thick stack of VE layer was selected as the surface of interest for analysis and plotted back onto the 3D mesh. To analyze the basement membrane, a 9μm thick stack, including the basal surface of the VE, was selected. We used ImSAnE to create three different charts of the embryo (an equidistant projection of the distal pole, and two complementary cylindrical projections). For morphometric analyses, we employed custom MATLAB and python scripts to segment cell outlines and compute the best polygonal approximation of each cell. Cell area and perimeter were computed after mapping back from the ImSAnE-generated charts into 3D. We fused three projections described above for optimized analyses. For analyses of the basement membrane and DVE in E5.5 embryos, we employed btc. With improved tissue cartography, a single conformal chart centered around the distal pole was sufficient to capture the entirety of the embryo (EPI and VE), and a thinner 4μm thick stack was sufficient to capture the entirety of the basement membrane across the embryo. To analyze the perforations of the basement membrane, the basement membrane was segmented using ilastik. The size of the holes was measured in 2D, mathematically correcting for cartographic area distortion. For analyses of the basement membrane and DVE marker in E5.0 and E5.25 embryos, we first define the proximal-distal axis. The location of the distal tip and the EmVE-ExVE boundaries were defined by the geometry of the embryo. We focused on the EmVE and defined a polar coordinate system for the surface of the embryo with the distal tip as the origin. The radial coordinate of this coordinate system is the distance (on the embryo surface) from the distal tip. Using the azimuthal angle of the coordinate system, we divide the embryo into equally sized region of interest (ROIs) around its circumference. We computed the area fraction covered by basement membrane perforations and the number of Lefty⁺ cells of each ROIs.

A complete documentation of the btc software workflow, including embryo segmentation, meshing, cartographic projection, and the methods used to correct for cartographic distortion is available at⁷⁷.

Physical model of DVE migration and cell shape

For the simulations, we model the DVE cells as i = 1,...,N active particles with position ${{{\boldsymbol{x}}}}_{{{\boldsymbol{i}}}}$. We used a fixed value of N = 10. For the purpose of the model, we map the embryo onto a two-dimensional cartographic plane, whose origin is the distal pole of the embryo. We use $\phi,\,r$ for the polar angle and radius. The initial positions of the active particles are sampled from a Ginibre random point process. Bar plots in Fig. 6 indicate averages over 20 random initial conditions. The VE cells are not modeled directly, and only enter the model via the friction $\gamma$ the DVE cells experience. The equations of motion are

$$\gamma \left({\phi}_{i}\right)\frac{{{\rm{d}}}{{\boldsymbol{x}}}_{{\boldsymbol{i}}}}{{{\rm{d}}}t}=- \alpha {\nabla}_{{{\boldsymbol{x}}}_{{\boldsymbol{i}}}} {\sum}_{ij}V\left(\left|{{\mathbf{\ x}}}_{{\boldsymbol{i}}}-{{\mathbf{\ x}}}_{{\boldsymbol{j}}}\right|\right)+\beta \left({\phi}_{i}\right){\hat{{\mathbf{\ r}}}}$$

The pair-wise potential V models both short-range repulsion due to steric repulsion as well as medium-range cell–cell adhesion:

$$V\left(r\right)={{\mbox{relu}}}\left(2\left({r}_{0}-r\right)\right)+{{\mbox{relu}}}\left(r-{r}_{0}\right)-{{\mbox{relu}}}\left(r-2{r}_{0}\right)$$

where ${r}_{0}$ sets the cell size and relu (x) = x if x > 0, else 0. Both the friction index and the strength of the migratory force are potentially modulated azimuthally, due to basement membrane heterogeneity:

$$\gamma \left(\phi \right)={\left({\gamma }_{0}+{\gamma }_{1}\frac{1+\cos \left(\phi \right)}{2}\right)}^{-1},\beta \left(\phi \right)={\beta }_{0}+{\beta }_{1}\frac{1+\cos \left(\phi \right)}{2}$$

The two hypotheses considered above, modulation of VE mechanical properties only, or modulation of DVE migratory forces are represented by the following possibilities: ${\gamma }_{1} \,\ne\, 0,{\beta }_{1}=0$ versus ${\gamma }_{1}=0,{\beta }_{1} \,\ne\, 0$. The parameter values used were $\left({{{\rm{\beta }}}}_{0},{{{\rm{\beta }}}}_{1}\right)=\left(1,\,0\right),\,\left({{{\rm{\gamma }}}}_{0},{{{\rm{\gamma }}}}_{1}\right)=\left(1/2,\,1/2\right)$ for the friction modulation, $\left({\beta }_{0},{\beta }_{1}\right)=\left(1/2,\,1/2\right),\,\left({\gamma }_{0},{\gamma }_{1}\right)=\left(1,\,0\right)$ for the force modulation, and $\left({\beta }_{0},{\beta }_{1}\right)=\left(1,\,0\right),\,\left({\gamma }_{0},{\gamma }_{1}\right)=\left(1,\,0\right)$ for the simulations without any modulation. We always used α = 0.4. The direction of the migratory cue is always the same - it is always azimuthally symmetric and could be derived from a morphogen gradient. It is cell adhesion that translates the stronger forces experienced by cells on one side into motion of all cells in a coherent direction. To simulate the equations of motion, we used the SciPy ordinary differential equation solver⁸⁰.

To investigate the influence of mechanics on cell shape, we employed a physical model of tissue geometry to generate a physical model of cell shape⁸¹. In brief, cells i are modeled as 2 d polygons with tri-cellular vertices r_ijk, and the shape of the cells is determined by minimizing the cells’ elastic energy. The elastic energy combines strong interfacial tensions (matching the straight cell–cell interfaces observed in the data) and a cellular bulk modulus λ and shear modulus μ and has the form

$${{\rm{E}}}={\sum }_{{{\rm{ij}}}}{{{\rm{T}}}}_{{{\rm{ij}}}}{{{\rm{l}}}}_{{{\rm{ij}}}}-{{\rm{p}}}{\sum}_{{{\rm{i}}}}{{{\rm{A}}}}_{{{\rm{i}}}}+{\sum}_{{{\rm{i}}}}{{\rm{\lambda }}}{[{{\rm{Tr}}}({{{\rm{S}}}}_{{{\rm{i}}}}-{{{\rm{S}}}}_{0})]}^{2}+{{\rm{\mu }}}{{\rm{Tr}}}[{({{{\rm{S}}}}_{{{\rm{i}}}}-{{{\rm{S}}}}_{0})}^{2}]$$

Where l_ij are the cell interface lengths, T_ij the interface tensions, p is the pressure, and A_i are the cell areas. S_i is the cell shape tensor defined by S_i = Σ_j r_ij ∧r_ij / l_ij (the sum is over the set of neighbors of the cell i). To generate the images shown in Supplementary Fig. 6a and b, we computed cell shapes for λ = μ > 0 and λ > 0, μ = 0 (vanishing shear modulus) at identical values of the interfacial tensions. The interfacial tensions were obtained by sampling a random Voronoi cell tiling. Note that this model is different from the “vertex model” in which elastic energy is determined by cell area and perimeter, and where high shape indices are observed when interfacial tensions become zero. The numerical simulations were carried out using the code available at github.com/nikolas-claussen/CE_simulation_public.

Ethics statement

Human embryo work was approved under Caltech IRB (#19-0950), Caltech Human Embryo and Stem Cell (HESC) Committee (#22-0101), and IVIRMA New Jersey (WCG IRB # 20031397 and 20050731). Informed consent was obtained from all participants in the study. These were patients from IVIRMA and USC-fertility, which had surplus embryos after completing their IVF treatment. Patients were informed of the general conditions of the donation, objectives, and methodology of human embryo research. They were offered counseling and alternatives, including discarding embryos and continued cryopreservation. Patients were informed that they would not benefit directly from the donation of embryos to research. Patients did not receive any financial compensation. Embryos were not created for research.

Human embryo thawing and in vitro culturing beyond implantation

Human embryo thawing was performed either at FEC Labs or in the Zernicka-Goetz Lab at Caltech. For thawing at FEC Labs, vitrified 5dpf or 6dpf human blastocysts were thawed using in-house thawing media from Reproductive Medicine Associates New Jersey (RMA NJ). The day before thawing, the TS Thawing solution (TS) and ORIGIO Sequential Blast without phenol red (Cooper Surgicals #83050010) were placed at 37 °C. For thawing, human embryos were immersed in 1 mL of the pre-warmed (TS: 1 M sucrose in HTF-HEPES BUFFER with 20% protein SPS) solution for 1 min. They were transferred to dilution solution (DS:0.5 M sucrose in HTF-HEPES buffer with 20% protein SPS) for 3 min, followed by incubation in washing solution WS1 solution (HTF-HEPES Buffer with 20% protein SPS) for 5 min, and WS2 solution (HTF-HEPES Buffer with 20% protein SPS) for 1 min. The embryos were incubated in the BLAST media at 37 °C with 5% CO₂. Thawed embryos were incubated in pre-equilibrated Sequential Blast™ culture medium under mineral oil (OVOIL vitrolife #10029) in a 37 °C, 5% CO₂ humidified incubator for 4 h to allow recovery. All these incubation steps were done in sterile Nunc 4-Well Dishes for IVF (Thermo Scientific #144444). For embryo thawing at Caltech, vitrified 5dpf or 6dpf human blastocysts were thawed with Vit Kit Thaw (FUJIFILM Irvine Scientific Catalog ID90137-SO), with 1-minute incubation in Thawing Solution, followed by 3-minute incubation in Dilution Solution, and 6 min in Washing Solution. The embryos were transferred to pre-equilibrated Continuous Single Culture-NX media (FUJIFILM Irvine Scientific Catalog ID90167) at 37 °C with 5% CO₂. The zona pellucida was removed by briefly treating the embryos with acidic Tyrode’s solution (Sigma # T1788) 4 h following thawing.

Embryos were cultured in pre-equilibrated in vitro culture1 (IVC1) medium with the addition of 50 ng/mL of Insulin Growth Factor 1 (IGF-1) (STEMCELL Technologies #78078). For 5dpf embryos, they were transferred to IVC1 the following day. For 6dpf embryos, they were transferred to IVC1 4 h following the thawing. Embryos were cultured in 8 well μ-slides tissue culture wells (Ibidi #80826) in 300 μL per embryo per well. Half-media was changed with IVC1 supplemented with IGF-1 every 24 h. Embryos were fixed with 4% paraformaldehyde/PBS for 30 min at room temperature at 7dpf and 9dpf.

Limitations of embryo study

There are limitations of the human 2D in vitro culture system and exclusion criteria for the analysis. We used surplus human embryos donated from patients underwent IVF treatment. Several embryos degenerated either upon thawing or during the in vitro culture. The embryos included in the analysis were attached to the dish and did not show signs of cell death. Nevertheless, there is higher variability in embryo quality and developmental potential among different embryos and patients.

While AP symmetry breaking in vivo occurs when the embryo begins to interact with maternal tissues at implantation, our in vitro culture system recapitulates key features of AP symmetry breaking - from DVE migration to the emergence of primitive streak. It would be interesting to explore the roles of the maternal environment in AP symmetry breaking in future studies.

Human stem cell-based embryo model

We used hEPSCs RUES2-GLR to generate peri-gastruloid with the version 2.0 protocol⁵⁸. On day0, hEPSCs were dissociated into single cells by incubation with TrypLE (Gibco# 12605036) at 37 °C for 5 min, followed by two hours of MEF depletion. After passing through a 40 μm filter, 45–70 cells per microwell were seeded into an AggreWell 800 plate (STEMCELL Technologies #34811) and were cultured with 2 mL of tHDM media per well supplemented with CEPT (50 nM Chroman 1, 5 μM Emricasan, 1x Polyamines, and 0.7 μM Trans-ISRIB). On day 1, half media change was performed by removing 1 mL of media from the well and adding 1 mL of fresh tHDM medium (N2B27 basal medium supplemented with 10 ng/mL FGF2, 10 ng/mL Activin-A, 1 μM CHIR99021 and 0.3 μM PD0325901) without CEPT. On day 2, the entire media was replaced with 2 mL of fresh N2B27 basal medium supplemented with 10 ng/mL FGF2, 10 ng/mL Activin-A, and 0.3 μM PD0325901 (tHDM (-Chir)). On day 3, half media change was performed with tHDM (-Chir). On day 4, media were replaced with IVC1 (Advanced DMEM/F12, 1x Glutamax, 1x Insulin-Transferrin-Selenium-Ethanolamine (ITS-X), 8 nM β-estradiol, 200 ng/mL progesterone, 25 μM N-acetyl-L-cysteine, 1x sodium pyruvate, and 20% heat-inactivated FBS) supplemented with 4% Matrigel (Corning #354277). Structures were fixed with 4% paraformaldehyde on Day5 for either 30 min at room temperature or overnight at 4 °C and underwent immunostaining as described above. hEPSCs were maintained in LCDM media (N2B27 medium with 10 ng/mL recombinant Human LIF, 1 μM CHIR 99021, 2 μM (S)-(+)-Dimethindene maleate, 2 μM Minocycline hydrochloride, 0.5 μM IWR-endo-1, and 2 μM ROCK inhibitor Y-27632) on irradiated MEF in a 37 °C incubator supplied with 5% CO₂ as described. N2B27 base medium consists of 50% DMEM/F12 medium (Gibco #11320033), 50% Neurobasal medium(Gibco #21103049), 50 U/mL penicillin/strep, 2 mM Glutamax, 1x N-2 supplement(Gibco #17502048), 1x B-27 supplement (Gibco #17504044), and 0.1 mM 2-Mercaptoethanol.

Human pluripotent stem cell work was approved under Caltech Human Embryo and Stem Cell (HESC) Committee (#22-0100). The embryo model does not contain trophoblast cells and is not considered to be an integrated human embryo model according to the International Society for Stem Cell Research (ISSCR). Our work complies with current ISSCR guidelines.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

For human embryo scRNAseq datasets, we used previously published and publicly datasets from Molè et al.¹³, ArrayExpress E-MTAB-8060; Blakely et al.⁵³, GSE66507 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE66507]; Petropoulos et al.⁵⁴, ArrayExpress E-MTAB-3929, [https://www.ebi.ac.uk/biostudies/ArrayExpress/studies/E-MTAB-3929?query=E-MTAB-3929]; Xiang et al.⁵⁵, GSE136447; Zhou et al.⁵⁶, GSE109555 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE109555]. Mouse sequencing data is deposited in the NCBI Gene Expression Omnibus (GEO) repository under the accession number GSE270803 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE270803]. All other data that support the findings of this study are available from the corresponding author upon reasonable request. Source data are provided with this paper.

Code availability

The MATLAB scripts, ImSAnE, used to carry out the tissue cartography are available here^26,42. The generation of the object for the human embryo combined scRNAseq dataset is available at https://github.com/bweatherbee/PeriImplantation and Seurat object is available at https://doi.org/10.5281/zenodo.7689579.

Change history

12 August 2025
In the version of the article initially published, nine of the supplementary figures were not correctly referenced in the Results and Discussion sections and this has been amended in the HTML and PDF versions of the article.

References

Goldstein, B. & Hird, S. N. Specification of the anteroposterior axis in Caenorhabditis elegans. Development 122, 1467–1474 (1996).
Article PubMed CAS Google Scholar
Munro, E., Nance, J. & Priess, J. R. Cortical flows powered by asymmetrical contraction transport PAR proteins to establish and maintain anterior-posterior polarity in the early C. elegans Embryo. Dev. Cell 7, 413–424 (2004).
Article PubMed CAS Google Scholar
Godt, D. & Tepass, U. Drosophila oocyte localization is mediated by differential cadherin-based adhesion. Nature 395, 387–391 (1998).
Article ADS PubMed CAS Google Scholar
González-Reyes, A. & Johnston, D. S. Patterning of the follicle cell epithelium along the anterior-posterior axis during Drosophila oogenesis. Development 125, 2837–2846 (1998).
Article PubMed Google Scholar
Schier, A. F. & Talbot, W. S. Molecular genetics of axis formation in zebrafish. Annu. Rev. Genet. 39, 561–613 (2005).
Article PubMed CAS Google Scholar
Bedzhov, I. & Zernicka-Goetz, M. Self-organizing properties of mouse pluripotent cells initiate morphogenesis upon implantation. Cell 156, 1032–1044 (2014).
Article PubMed PubMed Central CAS Google Scholar
Thomas, P. & Beddington, R. Anterior primitive endoderm may be responsible for patterning the anterior neural plate in the mouse embryo. Curr. Biol. 6, 1487–1496 (1996).
Article PubMed CAS Google Scholar
Thomas, P. Q., Brown, A. & Beddington, R. S. Hex: a homeobox gene revealing peri-implantation asymmetry in the mouse embryo and an early transient marker of endothelial cell precursors. Development 125, 85–94 (1998).
Article PubMed CAS Google Scholar
Rodriguez, T. A., Casey, E. S., Harland, R. M., Smith, J. C. & Beddington, R. S. Distinct enhancer elements control Hex expression during gastrulation and early organogenesis. Dev. Biol. 234, 304–316 (2001).
Article PubMed CAS Google Scholar
Perea-Gomez, A. et al. Initiation of gastrulation in the mouse embryo is preceded by an apparent shift in the orientation of the anterior-posterior axis. Curr. Biol. 14, 197–207 (2004).
Article PubMed CAS Google Scholar
Perea-Gomez, A. et al. Nodal antagonists in the anterior visceral endoderm prevent the formation of multiple primitive streaks. Dev. Cell 3, 745–756 (2002).
Article PubMed CAS Google Scholar
Yamamoto, M. et al. Nodal antagonists regulate formation of the anteroposterior axis of the mouse embryo. Nature 428, 387–392 (2004).
Article ADS PubMed CAS Google Scholar
Molè, M. A. et al. A single cell characterisation of human embryogenesis identifies pluripotency transitions and putative anterior hypoblast centre. Nat. Commun. 12, 3679 (2021).
Article ADS PubMed PubMed Central Google Scholar
Zhu, Q. et al. Decoding anterior-posterior axis emergence among mouse, monkey, and human embryos. Dev. Cell 58, 63–79.e4 (2023).
Article PubMed CAS Google Scholar
Srinivas, S., Rodriguez, T., Clements, M., Smith, J. C. & Beddington, R. S. P. Active cell migration drives the unilateral movements of the anterior visceral endoderm. Development 131, 1157–1164 (2004).
Article PubMed CAS Google Scholar
Rakeman, A. S. & Anderson, K. V. Axis specification and morphogenesis in the mouse embryo require Nap1, a regulator of WAVE-mediated actin branching. Development 133, 3075–3083 (2006).
Article PubMed CAS Google Scholar
Migeotte, I., Omelchenko, T., Hall, A. & Anderson, K. V. Rac1-dependent collective cell migration is required for specification of the anterior-posterior body axis of the mouse. PLoS Biol. 8, e1000442 (2010).
Article PubMed PubMed Central Google Scholar
Trichas, G. et al. Nodal dependent differential localisation of dishevelled-2 demarcates regions of differing cell behaviour in the visceral endoderm. PLoS Biol. 9, e1001019 (2011).
Article PubMed PubMed Central CAS Google Scholar
Trichas, G. et al. Multi-cellular rosettes in the mouse visceral endoderm facilitate the ordered migration of anterior visceral endoderm cells. PLoS Biol. 10, e1001256 (2012).
Article PubMed PubMed Central CAS Google Scholar
Shioi, G. et al. Apical constriction in distal visceral endoderm cells initiates global, collective cell rearrangement in embryonic visceral endoderm to form anterior visceral endoderm. Dev. Biol. 429, 20–30 (2017).
Article PubMed CAS Google Scholar
Stuckey, D. W. et al. Coordination of cell proliferation and anterior-posterior axis establishment in the mouse embryo. Development 138, 1521–1530 (2011).
Article PubMed PubMed Central CAS Google Scholar
Antonica, F., Orietti, L. C., Mort, R. L. & Zernicka-Goetz, M. Concerted cell divisions in embryonic visceral endoderm guide anterior visceral endoderm migration. Dev. Biol. 450, 132–140 (2019).
Article PubMed PubMed Central CAS Google Scholar
Stower, M. J. & Srinivas, S. The Head’s Tale: anterior-posterior axis formation in the mouse embryo. Curr. Top. Dev. Biol. 128, 365–390 (2018).
Article PubMed CAS Google Scholar
Walma, D. A. C. & Yamada, K. M. The extracellular matrix in development. Development 147 (2020).
Crest, J., Diz-Muñoz, A., Chen, D.-Y., Fletcher, D. A. & Bilder, D. Organ sculpting by patterned extracellular matrix stiffness. Elife 6, 773 (2017).
Article Google Scholar
Chen, D.-Y., Crest, J., Streichan, S. J. & Bilder, D. Extracellular matrix stiffness cues junctional remodeling for 3D tissue elongation. Nat. Commun. 10, 3339–15 (2019).
Article ADS PubMed PubMed Central Google Scholar
Haigo, S. L. & Bilder, D. Global tissue revolutions in a morphogenetic movement controlling elongation. Science 331, 1071–1074 (2011).
Article ADS PubMed PubMed Central CAS Google Scholar
Isabella, A. J. & Horne-Badovinac, S. Rab10-mediated secretion synergizes with tissue movement to build a polarized basement membrane architecture for organ morphogenesis. Dev. Cell 38, 47–60 (2016).
Article PubMed PubMed Central CAS Google Scholar
Sherwood, D. R. Basement membrane remodeling guides cell migration and cell morphogenesis during development. Curr. Opin. Cell Biol. 72, 19–27 (2021).
Article PubMed PubMed Central CAS Google Scholar
Pierce, G. B. The development of basement membranes of the mouse embryo. Dev. Biol. 13, 231–249 (1966).
Article PubMed CAS Google Scholar
Enders, A. C., Given, R. L. & Schlafke, S. Differentiation and migration of endoderm in the rat and mouse at implantation. Anat. Rec. 190, 65–77 (1978).
Article PubMed CAS Google Scholar
Leivo, I., Vaheri, A., Timpl, R. & Wartiovaara, J. Appearance and distribution of collagens and laminin in the early mouse embryo. Dev. Biol. 76, 100–114 (1980).
Article PubMed CAS Google Scholar
Kyprianou, C. et al. Basement membrane remodelling regulates mouse embryogenesis. Nature 582, 253–258 (2020).
Article ADS PubMed PubMed Central CAS Google Scholar
Mesnard, D., Filipe, M., Belo, J. A. & Zernicka-Goetz, M. The anterior-posterior axis emerges respecting the morphology of the mouse embryo that changes and aligns with the uterus before gastrulation. Curr. Biol. 14, 184–196 (2004).
Article PubMed CAS Google Scholar
Torres-Padilla, M.-E. et al. The anterior visceral endoderm of the mouse embryo is established from both preimplantation precursor cells and by de novo gene expression after implantation. Dev. Biol. 309, 97–112 (2007).
Article PubMed PubMed Central CAS Google Scholar
Morris, S. A. et al. Dynamics of anterior-posterior axis formation in the developing mouse embryo. Nat. Commun. 3, 673–10 (2012).
Article ADS PubMed Google Scholar
Hoshino, H., Shioi, G. & Aizawa, S. AVE protein expression and visceral endoderm cell behavior during anterior–posterior axis formation in mouse embryos: asymmetry in OTX2 and DKK1 expression. Dev. Biol. 402, 175–191 (2015).
Takaoka, K., Yamamoto, M. & Hamada, H. Origin and role of distal visceral endoderm, a group of cells that determines anterior–posterior polarity of the mouse embryo. Nat. cell Biol. 13, 743–752 (2011).
Article PubMed CAS Google Scholar
Miner, J. H., Li, C., Mudd, J. L., Go, G. & Sutherland, A. E. Compositional and structural requirements for laminin and basement membranes during mouse embryo implantation and gastrulation. Development 131, 2247–2256 (2004).
Article PubMed CAS Google Scholar
Soares, M. L., Torres-Padilla, M.-E. & Zernicka-Goetz, M. Bone morphogenetic protein 4 signaling regulates development of the anterior visceral endoderm in the mouse embryo. Dev. Growth Differ. 50, 615–621 (2008).
Article PubMed PubMed Central CAS Google Scholar
Almeida, L. G. N. et al. Matrix metalloproteinases: from molecular mechanisms to physiology, pathophysiology, and pharmacology. Pharmacol. Rev. 74, 712–768 (2022).
Article PubMed Google Scholar
Heemskerk, I. & Streichan, S. J. Tissue cartography: compressing bio-image data by dimensional reduction. Nat. Methods 12, 1139–1142 (2015).
Article PubMed PubMed Central CAS Google Scholar
Choi, H. M. T. et al. Third-generation in situhybridization chain reaction: multiplexed, quantitative, sensitive, versatile, robust. Development 145, dev165753–10 (2018).
Article PubMed PubMed Central Google Scholar
D’ortho, M. et al. Membrane-type matrix metalloproteinases 1 and 2 exhibit broad-spectrum proteolytic capacities comparable to many matrix metalloproteinases. Eur. J. Biochem. 250, 751–757 (1997).
Article PubMed Google Scholar
Christodoulou, N. et al. Morphogenesis of extra-embryonic tissues directs the remodelling of the mouse embryo at implantation. Nat. Commun. 1–12. https://doi.org/10.1038/s41467-019-11482-5 (2019).
Classen, A.-K., Anderson, K. I., Marois, E. & Eaton, S. Hexagonal packing of drosophila wing epithelial cells by the planar cell polarity pathway. Dev. Cell 9, 13–13 (2005).
Article Google Scholar
Curran, S. et al. Myosin II controls junction fluctuations to guide epithelial tissue ordering. Dev. Cell 1–20. https://doi.org/10.1016/j.devcel.2017.09.018 (2017).
Tetley, R. J. & Mao, Y. The same but different: cell intercalation as a driver of tissue deformation and fluidity. Philos. Trans. R. Soc. B 373, 20170328–10 (2018).
Article Google Scholar
Brauns, F., Claussen, N. H., Lefebvre, M. F., Wieschaus, E. F. & Shraiman, B. I. The geometric basis of epithelial convergent extension. eLife 13, RP95521 (2024).
Kimura-Yoshida, C. et al. Canonical wnt signaling and its antagonist regulate anterior-posterior axis polarization by guiding cell migration in mouse visceral endoderm. Dev. Cell 9, 639–650 (2005).
Article PubMed CAS Google Scholar
Rossant, J. & Tam, P. P. L. Early human embryonic development: blastocyst formation to gastrulation. Dev. Cell 57, 152–165 (2022).
Article PubMed CAS Google Scholar
Weatherbee, B. A. T. et al. Distinct pathways drive anterior hypoblast specification in the implanting human embryo. Nat. Cell Biol. 26, 353–365 (2024).
Article PubMed PubMed Central CAS Google Scholar
Blakeley, P. et al. Defining the three cell lineages of the human blastocyst by single-cell RNA-seq. Development 142, 3151–3165 (2015).
Article PubMed PubMed Central CAS Google Scholar
Petropoulos, S. et al. Single-cell RNA-Seq reveals lineage and x chromosome dynamics in human preimplantation embryos. Cell 165, 1012–1026 (2016).
Article PubMed PubMed Central CAS Google Scholar
Xiang, L. et al. A developmental landscape of 3D-cultured human pre-gastrulation embryos. Nature 577, 537–542 (2020).
Article PubMed CAS Google Scholar
Zhou, F. et al. Reconstituting the transcriptome and DNA methylome landscapes of human implantation. Nature 572, 660–664 (2019).
Article ADS PubMed CAS Google Scholar
Shahbazi, M. N. et al. Self-organization of the human embryo in the absence of maternal tissues. Nat. Cell Biol. 18, 700–708 (2016).
Article PubMed PubMed Central CAS Google Scholar
Liu, L. et al. Modeling post-implantation stages of human development into early organogenesis with stem-cell-derived peri-gastruloids. Cell 186, 3776–3792.e16 (2023).
Article PubMed CAS Google Scholar
Martyn, I., Kanno, T. Y., Ruzo, A., Siggia, E. D. & Brivanlou, A. H. Self-organization of a human organizer by combined Wnt and Nodal signalling. Nature 558, 132–135 (2018).
Article ADS PubMed PubMed Central CAS Google Scholar
Omelchenko, T. et al. β-Pix directs collective migration of anterior visceral endoderm cells in the early mouse embryo. Genes Dev. 28, 2764–2777 (2014).
Article PubMed PubMed Central Google Scholar
Yamada, K. M. & Sixt, M. Mechanisms of 3D cell migration. Nat. Rev. Mol. Cell Biol. 20, 738–752 (2019).
Article PubMed CAS Google Scholar
Mao, Y. & Wickström, S. A. Mechanical state transitions in the regulation of tissue form and function. Nat. Rev. Mol. Cell Biol. 1–17. https://doi.org/10.1038/s41580-024-00719-x (2024).
Stower, M. et al. Single-cell phenomics reveals behavioural and mechanical heterogeneities underpinning collective migration during mouse anterior patterning. bioRxiv https://doi.org/10.1101/2023.03.31.534937 (2023).
Brennan, J. et al. Nodal signalling in the epiblast patterns the early mouse embryo. Nature 411, 965–969 (2001).
Article ADS PubMed CAS Google Scholar
Norris, D. P., Brennan, J., Bikoff, E. K. & Robertson, E. J. The Foxh1-dependent autoregulatory enhancer controls the level of Nodal signals in the mouse embryo. Development 129, 3455–3468 (2002).
Article PubMed CAS Google Scholar
Takaoka, K. et al. The mouse embryo autonomously acquires anterior-posterior polarity at implantation. Dev. Cell 10, 451–459 (2006).
Article PubMed CAS Google Scholar
Takaoka, K., Nishimura, H. & Hamada, H. Both Nodal signalling and stochasticity select for prospective distal visceral endoderm in mouse embryos. Nat. Commun. 1–13. https://doi.org/10.1038/s41467-017-01625-x (2017).
Szabova, L. et al. Membrane-type MMPs are indispensable for placental labyrinth formation and development. Blood 116, 5752–5761 (2010).
Article PubMed PubMed Central CAS Google Scholar
Weatherbee, B. A. T. et al. Pluripotent stem cell-derived model of the post-implantation human embryo. Nature 622, 584–593 (2023).
Article ADS PubMed PubMed Central CAS Google Scholar
Oldak, B. et al. Complete human day 14 post-implantation embryo models from naive ES cells. Nature 622, 562–573 (2023).
ADS PubMed PubMed Central CAS Google Scholar
Pedroza, M. et al. Self-patterning of human stem cells into post-implantation lineages. Nature 622, 574–583 (2023).
Article ADS PubMed PubMed Central CAS Google Scholar
Okubo, T. et al. Hypoblast from human pluripotent stem cells regulates epiblast development. Nature 626, 357–366 (2024).
Article ADS PubMed CAS Google Scholar
Sturm, K. & Tam, P. P. L. Guide to Techniques in Mouse Development. [10] Isolation and culture of whole postimplantation embryos and germ layer derivatives, 164–190 (Elsevier, 1993).
Takahashi, M., Makino, S., Kikkawa, T. & Osumi, N. Preparation of rat serum suitable for mammalian whole embryo culture. J. Vis. Exp. e51969, https://doi.org/10.3791/51969 (2014).
Medeiros, G. et al. Confocal multiview light-sheet microscopy. Nat. Commun. 6, 8881 (2015).
Article ADS PubMed CAS Google Scholar
Preibisch, S. et al. Efficient Bayesian-based multiview deconvolution. Nat. Methods 11, 645–648 (2014).
Article PubMed PubMed Central CAS Google Scholar
Claussen, N., Regis, C., Wopat, S. & Streichan, S. Blender tissue cartography: an intuitive tool for the analysis of dynamic 3D microscopy data. bioRxiv https://doi.org/10.1101/2025.02.04.636523 (2025).
Article PubMed PubMed Central Google Scholar
Berg, S. et al. ilastik: interactive machine learning for (bio)image analysis. Nat. Methods 16, 1226–1232 (2019).
Article PubMed CAS Google Scholar
Álvarez, L., Baumela, L., Henríquez, P. & Márquez-Neila, P. Morphological Snakes. In: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition 2197–2202 https://doi.org/10.1109/cvpr.2010.5539900 (2010).
Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261–272 (2020).
Article PubMed PubMed Central CAS Google Scholar
Claussen, N. H., Brauns, F. & Shraiman, B. I. A geometric-tension-dynamics model of epithelial convergent extension. Proc. Natl Acad. Sci. 121, e2321928121 (2024).
Article PubMed PubMed Central CAS Google Scholar
Sun, J. et al. STIM1- and Orai1-mediated Ca2+ oscillation orchestrates invadopodium formation and melanoma invasion. J. Cell Biol. 207, 535–548 (2014).
Article PubMed PubMed Central CAS Google Scholar

Download references

Acknowledgements

The authors thank Ali H. Brivanlou and Jun Wu for providing RUES2-GLR hESCs; Meyer Barembaum for providing the plasmid PCI-H2B-RFP; Shengyu Yang for pLPCX2-hMT1-MMP-eGFP⁸². We thank Ali Ahmady for technical assistance with thawing human blastocysts; Lizhong Liu for the advice on the human embryo model; Jacqueline Scapa, Carol Readhead, and Ron Hadas for assistance in rat serum collection; Giada Spigolon for assistance in microscopy; Ana Sánchez for illustrations in Fig. 8; David Glover, Angela Anderson, Marianne Bronner, and our colleagues in the Zernicka-Goetz lab for their feedback. Some image analyses were performed in the Biological Imaging Facility, with the support of the Caltech Beckman Institute and the Arnold and Mabel Beckman Foundation. D.-Y.C. was supported by NIH/NICHD F32HD105442 and K99HD111676. N.H.C. was supported by NIGMS under award number R35-GM138203 and by NSF under award number PHY:2210612. S.J.S. was supported by NIGMS under award number 5R35GM138203. M.Z.-G. was supported by NICHD of the NIH under award number 5DP1HD104575. Human embryo work was supported by Open Philanthropy (M.Z.-G.), The Weston Havens Foundation, and NOMIS Foundation (M.Z.-G.).

Author information

Authors and Affiliations

Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, CA, 91125, USA
Dong-Yuan Chen, Wenqi Hu & Magdalena Zernicka-Goetz
Department of Physics, University of California Santa Barbara, Santa Barbara, CA, 93106, USA
Nikolas H. Claussen & Sebastian J. Streichan
Columbia University, Irving Cancer Medical Center, New York, NY, 10032, USA
Shiny Titus
Center for Stem Cell and Organoid Medicine, Perinatal Institute, Division of Developmental Biology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA
Bailey A. T. Weatherbee
Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, University of Southern California, Los Angeles, CA, 90033, USA
Rachel S. Mandelbaum
Foundation for Embryonic Competence, 140 Allen Road Basking Ridge, Basking Ridge, NJ, 07945, USA
Richard T. Scott Jr
Yale School of Medicine, 333 Cedar Street, New Haven, CT, 06510, USA
Richard T. Scott Jr
Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale School of Medicine, 333 Cedar St, New Haven, CT, 06520, USA
Emre Seli
IVIRMA Global Research Alliance, IVIRMA New Jersey, 140 Allen Basking Ridge, Princeton, NJ, 07920, USA
Emre Seli
Department of Physiology, Development and Neuroscience, University of Cambridge, Cambridge, CB2 3EG, UK
Magdalena Zernicka-Goetz

Authors

Dong-Yuan Chen
View author publications
Search author on:PubMed Google Scholar
Nikolas H. Claussen
View author publications
Search author on:PubMed Google Scholar
Shiny Titus
View author publications
Search author on:PubMed Google Scholar
Wenqi Hu
View author publications
Search author on:PubMed Google Scholar
Bailey A. T. Weatherbee
View author publications
Search author on:PubMed Google Scholar
Rachel S. Mandelbaum
View author publications
Search author on:PubMed Google Scholar
Richard T. Scott Jr
View author publications
Search author on:PubMed Google Scholar
Emre Seli
View author publications
Search author on:PubMed Google Scholar
Sebastian J. Streichan
View author publications
Search author on:PubMed Google Scholar
Magdalena Zernicka-Goetz
View author publications
Search author on:PubMed Google Scholar

Contributions

D.-Y.C.: Conceptualization, Methodology, Software, Formal analysis, Investigation, Visualization, Writing, Project administration, Funding acquisition. N.H.C.: Methodology, Formal analysis, Visualization, Writing. W.H.: Data Curation, Formal Analysis, Investigation. S.T.: Investigation. B.A.W.: Data Curation, Visualization. R.S.M.: Resources, Project administration. R.T.S.: Resources, Project administration. E.S.: Resources, Funding acquisition, Project administration. S.J.S.: Methodology, Visualization, Software. M.Z.-G.: Conceptualization, Supervision, Resources, Writing, Funding acquisition.

Corresponding author

Correspondence to Magdalena Zernicka-Goetz.

Ethics declarations

Competing interests

Emre Seli was a consultant for and received research funding from the Foundation for Embryonic Competence. Other authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Fan Zhou, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Information

Supplementary Movie 1

Supplementary Movie 2

Supplementary Movie 3

Supplementary Movie 4

Supplementary Movie 5

Supplementary Movie 6

Supplementary Movie 7

Supplementary Movie 8

Supplementary Movie 9

Supplementary Movie 10

Supplementary Movie 11

Supplementary Movie 12

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Chen, DY., Claussen, N.H., Titus, S. et al. Basement membrane perforations guide anterior–posterior axis formation. Nat Commun 16, 6763 (2025). https://doi.org/10.1038/s41467-025-61441-6

Download citation

Received: 10 April 2025
Accepted: 22 June 2025
Published: 22 July 2025
DOI: https://doi.org/10.1038/s41467-025-61441-6