Methods for producing amniotic membrane in vitro

Abstract

Methods for producing extra-embryonic structures Described herein is a novel cell culture model of the post-gastrulation stage human amnion. Also described herein are methods for producing extra-embryonic structures in culture, in particular for producing an in vitro three-dimensional amnion model. The amnion models described herein, and amniotic membranes isolated from such amnion models, can be used in a variety of applications, such as in methods of drug screening, tissue engineering, regenerative medicine, wound healing and treating burns.

Description

[0001] Methods for producing extra-embryonic structures

[0002] FIELD OF THE INVENTION

[0003] The present invention provides an in vitro amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity; and methods for producing the same.

[0004] BACKGROUND

[0005] Central to the survival and development of a healthy embryo are several extra-embryonic tissues that are often overshadowed by the prominence of the embryo itself. The amnion is a thin, transparent, but robust membrane that forms a fluid-filled sac around the embryo. Thought to initially help pattern the embryo through providing signalling cues, the amnion later provides mechanical protection and nutritional support [1 , 2]. The amnion is formed by two cell layers: the inner epithelial layer of amniotic ectoderm, which forms direct contact with and regulates the homeostasis of amniotic fluid, and an outer layer of mesenchymal extra-embryonic mesoderm (Figure 1A) [3]. The extra-embryonic mesoderm also covers the yolk sac, which initially develops from the hypoblast cells of the inner cell mass in the form of the primary yolk sac and is later replaced by the secondary yolk sac [4]. The critical function of the yolk sac is to provide nutritional supply and gas exchange for the early embryo and initiate the process of hematopoiesis [4, 5]. The study of human extra-embryonic tissues has been limited due to technical and ethical restrictions associated with obtaining human embryo samples at the post-implantation stages of development [6-9]. Most our understanding of extra-embryonic development comes from studies of other species of mammals, which have proven to be dissimilar to humans [1 , 3, 10, 11]. As a result, many aspects of extra-embryonic tissue genesis, development and function remain unknown and are the subject of extensive debate. Multiple sources at different developmental stage have been proposed as the origin of extra-embryonic mesoderm and the amnion (Pham et al, Pera & Rossant 2021 , Rossant & Tam 2022). In addition, a complex cellular composition has been reported to originate within the yolk sac. Accordingly, the cellular and molecular drivers that govern the development of these tissues remain unknown and their signalling interactions with the embryo proper poorly understood. Recent studies using human embryonic stem cells (hES cells) have recapitulated different aspects of extra-embryonic tissue formation in vitro. 2D culture systems have allowed for the derivation of several definitive extra-embryonic cell types, such as extra-embryonic mesoderm, amniotic ectoderm, and yolk sac, albeit restricted to individual cell types in a monolayer [12-15]. 3D models forming cyst-like structures and consisting of only amniotic ectoderm cells have successfully captured the early stage of amniogenesis while one of models has partly mimicked the formation of yolk sac-like structures [16-19]. Most recently, integrated 3D embryo models comprising multiple cell types, including some extra- embryonic tissues, were introduced [20-23]. Of these, the most advanced in terms of its complexity and presence of extra-embryonic tissues was the Stem-Cell-Derived Embryo Model (SEMs) which shows remarkable resemblance with the post-implantation embryo

[0022] . Although these recent 3D models represent powerful systems to recapitulate post-implantation embryogenesis, they are restricted to the peri-gastrulation stage of extra-embryonic tissue development. In addition, they lack many of the structural and biomechanical properties of these extra-embryonic tissues, as well as the developmental potential beyond the early stages of gastrulation.

[0006] Thus, there is a need in the art for in vitro models that better recapitulate extra-embryonic structures, particularly at the post-gastrulation stage.

[0007] SUMMARY OF THE INVENTION

[0008] In a first aspect, the invention provides an in vitro three-dimensional amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity.

[0009] In a second aspect, the invention provides a method for producing an in vitro three-dimensional amnion model, the method comprising: a) providing a plurality of cells, b) treating the plurality of cells with a BMP agonist, c) treating the plurality of cells with a WNT agonist, and d) culturing the plurality of cells.

[0010] In a third aspect, the invention provides an in vitro amnion model obtained or obtainable by the methods described herein.

[0011] In a fourth aspect, the invention provides an amniotic membrane derived or obtainable from an in vitro amnion models described herein.

[0012] In a fifth aspect, the invention provides the use of an in vitro amnion model as described herein or an amniotic membrane as described herein in a method of drug screening, a method oftissue engineering, regenerative medicine, a method of assessing amnion permeability, a method of assessing membrane permeability, or a method of assessing drug permeability.

[0013] In a sixth aspect, the invention provides an in vitro amnion model as described herein or an amniotic membrane as described herein for use in methods of wound healing, treating burns, ocular reconstruction, or in regenerative medicine.

[0014] BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1. Engineering post-gastrulation amnioids (PGAs) from primed human embryonic stem cells (A) Histological Carnegie Stage 7 (CS7) section of the human embryo representing day 17-18 post fertilisation. Two-layered amnion composed by amniotic ectoderm and extra-embryonic (Exe) mesoderm is highlighted. Source: EHD virtual embryo. (B) Schematic of the protocol used to generate PGAs. E6 (Essential 6 medium), BMP4 (bone morphogenetic protein 4), CHIR (CHIR99421 : WNT agonist) and Rocki (ROCK inhibitor (Y-27632)). (C) Light-sheet microscopy representative images of PGA development over time from days 1 to 18. Scale bar: 400pm. (D) Light sheet microscopy images (section of z-stack) showing developmental progression of the 2 layered amniotic sac-like structure expressing the amniotic ectoderm marker TFAP2C and the extra-embryonic mesoderm marker GATA6 between days 1 and 18. Scale bar: 100pm. (E) Representative 3D projections of the amniotic sac like structure captured by light-sheet microscopy at days 8, 11 and 18. ROI highlights the two layers of TFAP2C (yellow) and GATA6 (magenta) positive cells within the amnion-like structures. Scale bar: 100pm. (F) PGA measured sizes at days 30, 60 and 90 showing the scale of growth and reproducibility. (G) Quantification of average PGA size and (H) morphological eccentricity at days 1 and 10 for H1 hESCs. (I) Quantification of area and (J) eccentricity for individual PGAs at days 1 and 10 for H1 hESCs. Three independent experiments are shown. (K) Percentage of PGAs that grew over 320pm in 1 and 10 days. Data from 3 independent experiments n = 282. (L) Quantification of area for individual PGAs at days 1 and 10 for RUES2 hESCs. Percentage of PGAs that grew over 320pm is shown. Data from 2 independent experiments n = 192.

[0016] Figure 2. Engineering of post-gastrulation amnioids (PGAs) from primed human embryonic stem cells (A) Representative images of initial 2D culture of H1 human embryonic stem cells before (Day -2) and after (Day -1) BMP4 (5ng / ml) treatment and following CHIR (3uM) treatment (Day 0). ISL1 , GATA6, TFAP2C, HAND1 , GATA3 and CDX2 transcription factors stainings are shown. Scale bar: 100pm. (B) Representative light-sheet images (z-stack sections) of PGAs development from day 1 to day 8 showing ISL1 , GAT A6, TFAP2C, HAND1 , GATA3 and CDX2 transcription factors. Scale bar: 100pm. (C) Projection of representative light light-sheet images at day 8 showing ISL1 , TFAP2C and GATA3 (related to day 8 in figure S1 B). Scale bar: 100pm. (D) Z-stack section and 3D projection of representative day 11 PGA showing ISL1 and GAT A3 staining. Scale bars: 100pm.

[0017] Figure 3. Characterization of PGAs by single cell RNA sequencing

[0018] (scRNAseq) (A) Identification of cell types from scRNA-seq time course. UMAP of Seurat clusters 0-22 (left), final annotations of cell types in UMAP embedding (middle) sankey diagram showing relationship between clusters and cell type annotation (right). (B) Heatmap showing gene expression dynamics through latent time based on RNA-velocity and calculated using scvelo pipeline. (C) UMAPs showing expression of selected lineage specific marker genes during the course of differentiation. (D-F) UMAP of (D) microfluidic model human epiblast and amnion development (Zhen et al 2019) (E) model of human extraembryonic mesoderm cells from naive pluripotent stem cells (Pham et al 2023) and (F) human yolk sac cell atlas (Goh et al 2023) datasets projected on data from this study. (G) Stream plot showing flow of RNA-velocity vectors. (H) Stream plot showing latent time calculated using scvelo pipeline. Figure 4. Characterization of PGAs by single cell RNA sequencing (scRNAseq) (A) UMAP embedding for integrated timepoints, coloured according to annotated cell types. Annotation was based on expression of lineage specific markers. EPI: epiblast, Am: Amnion, ExM: extra-embryonic mesoderm, Ecto: ectoderm, YS: yolk sac. (B) UMAP embedding (left) and alluvial diagram (right) for annotated cell types over time. Alluvial diagram shows proportion of annotated cell types through time. (C) Dot plot showing selected marker genes expressed within the annotated cell types. Dot size shows the percentage of cells expressing the gene within the cluster and the colour intensity represent the mean expression. (D) UMAPs showing expression of lineage specific marker genes projected from all timepoints. Corresponding cell populations are highlighted. (E-G) UMAP of (E) integrated stem cell-based model of the human embryo (Tyser et al 2021) dataset (F) published monkey embryos dataset and (G) human embryo dataset projected on PGAs data from this study. (H) Slingshot trajectories for ExE mesoderm, amnion, and yolk sac lineages. (I) Scaled gene expression dynamics through lineage pseudotime of significant genes associated with each trajectory, as identified by tradeSeq association test.

[0019] Figure 5. GATA3 is necessary and sufficient for amniogenesis (A) UMAP showing cell type annotations with slingshot trajectories to ExE Mesoderm, Amnion and YS Endoderm. Points along trajectories indicate tradeSeq knots used to perform differential dynamics testing. Dashed line reveals ‘early’ bifurcation of ExE mesoderm and Amnion lineage used for EarlyDE analysis. (B) Heatmaps showing expression of n=11805 significant genes with differential dynamics between the amniotic ectoderm and ExE mesoderm lineages. Genes highlighted indicate putative drivers of amniotic ectoderm and ExE mesoderm fates. (C) Binarised regulon specificity score (RSS) of top 15 lineage specific regulons per cell type as calculated by SCENIC. Cell type specific (magenta - ExE mesoderm, yellow - amniotic ectoderm) transcription factors highlighted on right. (D) Schematic of protocol used for siRNA perturbations using PGA platform. (E) Representative images of day 12

[0020] PGAs derived from hES-H2B-mCherry KI cells following siRNA knock-down for indicated genes. Scale bar, 500pm. (F-G) Quantification of (F) size and (G) morphologic features of PGA at day 12, related to figure 3E. Data are from 2 independent experiments n = 12 PGAs for each siRNA. Dashed lines indicated 95% Cl. Perturbations compared to control treatment by performing Mann-Whitney test * p<0.05, ** p<0.005, *** p<0.0005, **** p<0.00005). (H) Schematic of protocol used for PGA generation following dCas9-VRP driven inducible gene expression. (I) Representative images of PGAs derived from hES cells harbouring inducible dCas9-VPR to express the indicated individual genes at day 16. Scale bar: 500pm. (J) Quantification of size and morphological features of PGAs following dCas9-VPR inducible gene expression at days 16, related to figure 3I. Data are from 2 independent experiments n = 12 for each crRNA. Perturbations compared to control treatment by performing Mann-Whitney test * p<0.05, “ p<0.005, “* p<0.0005, ““ p<0.00005). (K) Representative 3D projection crGATA3 induced amniotic sac like structure captured by light-sheet microscopy at day 16 showing two-layered membranes expressing TFAP2C and GATA6, respectively. Scale bar: 100pm.

[0021] Figure 6. Driver gene analysis coupled with genetic perturbation of PGAs reveals early regulators of amniogenesis (A) Predicted top 16 driver genes as identified by tradeSeq earlyDE analysis. Gene expression dynamics through slingshot pseudotime for top 16 putative drivers from clusters 1 , 4 (ExE mesoderm) and 2 (amniotic ectoderm) across the two lineages (Purple shows ExE mesoderm trajectory; yellow shows amnion trajectory). (B) Average SCENIC regulon activity of top 15 regulons per cell type by relative specificity score (RSS). (C) Representative images of PGAs following siRNA knock-down at day 16. 8 replicates for scramble and 4 replicates per siRNA show reproducibility and screening potential of PGAs. Scale bars, 500pm. (D) Representative 3D projection of light sheet microscopy showing PGA at day 12 where GATA3 has been knocked down. Scale bars, 100pm. (E) Representative light sheet image of crGATA3 induced 3D aggregate at day 1 and stained for GAT A3, TBXT and Nanog. Scale bars, 100pm.

[0022] Figure 7. PGAs recapitulates multiple epiblast derived extra-embryonic tissues (A) UMAP of integrated timepoints showing yolk sac endoderm and endothelium clusters. (B) Dot plot showing selected marker genes expressed in yolk sac endoderm and endothelium cell clusters. Dot size shows the percentage of cells expressing the gene within the cluster and the colour intensity represent the mean expression. (C) Representative 3D projection of lightsheet microscopy images showing SOX17+ and CD34+ staining of yolk sac-like compartment at 5 and 8 days. Scale bars, 100pm. (D) UMAPs showing expression of selected genes for late extra-embryonic mesoderm like clusters. Corresponding clusters are highlighted on the left. (E) Representative 3D projection of lightsheet microscopy images of PDGFRA+ and POSTN+ staining showing yolk sac mesoderm-like compartment and the surrounding extra-embryonic mesoderm at day 8. Scale bars, 100pm. (F) Representative 3D projection of lightsheet microscopy images at day 8 showing extra-embryonic lineages including the amniotic sac like (TFAP2C+), yolk sac like (SOX17+), extra-embryonic mesoderm and connecting stalk (GATA6). Scale bars, 100pm. (G) Representative images of PGA development at day 8, 1 1 and 14 and corresponding to illustration of Sagittal sections of human embryo representing around day 20, 24 and 28. Illustration adapted from Sadler, T. W., & Feldkamp, M. L., 2008. Scale bars, 100pm.

[0023] Figure 8. Molecular and morphological characterization of PGAs (A) Representative light-sheet images (z-stack section) of PGAs at day 1 , 3, 5 and 8. stained for SOX17 and CD34 staining is shown (related to figure 4C). Scale bars, 100pm. (B-C) Representative light-sheet images (z-stack sections) of day 8 PGAs stained for (B) PDGFRA and POSTN and (C) KRT8 and GATA6. Scale bars, 100pm. (D) Representative light-sheet images (z-stack sections) of day 8 PGAs showing reproducibility of this model in generating extra-embryonic like tissues including connecting stalk. Scale bars, 100pm. (E) Reproducibility of PGAs made with H1 hES cells at day 14 and generating a fluid-filled amnioic sac-like structure. Scale bars, 500|jm. (F) Reproducibility of PGAs made with RUES2 hES cells at day 14 and generating a fluid-filled amnioic sac-like structure. Scale bars, 500pm. (G) Average gene expression profile per cluster in Exe mesoderm and amniotic-ectoderm. (H) UMAPs showing the expression of cell cycle and proliferation genes (CDK1 , CCNB1 , PLK1 , CDKN1 C, MKI67, ASPM, TOP2A). Exe mesoderm shows high expression of the cell cycle inhibitor CDKN1 C (p21) while epiblast, amnion and intermediate mesoderm show expression of cell cycle drivers such as Cdk1 and Plk1. (I) Representative light-sheet images (z-stack section) and 3D projection of day 8 PGAs stained for CDK1 . Scale bars, 100pm.

[0024] Figure 9. Signals from extra-embryonic lineages induce hES cell differentiation, selforganisation, and amnion formation (A) Schematic of protocol used for co-culture of PGAs (Exe) and primed hES cells (epiblast), wt H1 hES cell colonies were stimulated with 5ng / ml BMP4 for 1 day and 3uM CHIR for another day, as described in Figure 1 A (Exe, PGAs). Exe cells were then dissociated and mixed as single cells with untreated, single-cell dissociated H1-H2B-mCherry hES cells (epiblast cells). (B) Representative images of co-cultured cells showing aggregate formation and developmental progress from days 0 to 224h. Dashed line highlights Exe cells at 24 and 48 hours. Scale bars: 500pm. (C) Representative z-stack section of light-sheet imaging at day 1 , 3, 5 and 8. GATA6 (blue) and TFAP2C (yellow) staining are shown. mCherry represent H1-H2B-mCherry KI hES cells. Corresponding merge 3D projections are highlighted (left). Scale bars, 100pm. n= x number of structures were analysed (D) UMAP embedding of Exe:epiblast cells co-culture for all timepoints, coloured according to cell type annotation. (E) UMAP embedding for annotated cell types through time. (F) Gene expression of selected genes from scRNA-seq of Exe:epiblast cells co-cultured at day 1 post aggregation. (G) Schematic summarising a potential mechanism indicating cross-talk between extraembryonic and epiblast-like cell populations.

[0025] Figure 10. Molecular and morphological characterization of exe:hES cells coculture (A) Representative light-sheet images (z-stack section) of Exe:hES cells at day 1 , 3, 5 and 8. Staining for SOX2, OCT4, GAT A3, CDX2 and TBXT is shown. mCherry represent H1-H2B-mCherry primed hES cells. Scale bar: 100pm. (B) Identifying cell types present in Exe:hES cells co-culture by scRNA-seq. UMAP of Seurat clusters 0-10 (top), sankey diagram showing relationship between clusters and cell type annotation (middle) and final annotations of cell types on UMAP for co-culture populations (bottom). (C) Dot plot showing selected gene markers expressed within the annotated cell types. Dot size shows the percentage of cells expressing the gene within the cell type population and the colour intensity represent the mean expression. (D) UMAPs showing expression of lineage specific marker genes projected from all timepoints in the Exe:hES cells co-culture data. (E) Representative images of the Exe:hES cell co-cultures showing day 8 structures in E6 media or when E6 was switched at day 5 to modified EUCM2 plus 30% serum. Scale bars, 100pm. (F) Quantification of morphological features of Exe:hES cells co-culture at days 8 cultured in E6 media or when E6 was switched at day 5 to modified EUCM2 plus 30% serum, related to figure S5E. n = 12 per condition. (G) UMAP embedding for day 8 samples cultured in E6 media or when E6 was switched at day 5 to modified EUCM2 plus 30% serum. (H) Representative light-sheet images (z-stack sections) of Exe:hESC co-culture at day 8 when media was switched from E6 to modified EUCM2 from day 5. SOX17 and BLIMP1 staining are shown. Zoom images of ROI are shown. Scale bars, 100pm.

[0026] Figure 11 : Amniotic ectoderm cells in PGAs have squamous morphology. 3D projection and Z- stack section of p-catenin staining by light-sheet imaging at day 8. Scale bars: 100pm.

[0027] Figure 12: PGAs have basement membrane between amniotic ectoderm and extraembryonic mesoderm. Z-stack section of Laminin and GATA6 immunofluorescence staining from light-sheet images. Time course shown from day 3 to day 21 -time Scale bars: 100pm.

[0028] Figure 13: Functional characterization of the proteome and metabolome of PGAs (A) Schematic of protocol used for proteomics analysis. Intra and extra luminal fluid from day 21 PGAs was collected for proteomic analysis. Intra-luminal fluid (Intra), extra-luminal fluid (Extra). (B) PCA analysis comparing intra- and extra-luminal PGA samples. n= 3 replicates per condition. (C) Venn diagram of the proteome identified in intra- and extra-luminal PGA samples. (D) Gene Ontology (GO) terms related to biological process showing pathways enriched in the proteins identified within the intra- and the extra-luminal samples. (E) Heat maps of selected proteins associated with amnion function differently expressed within intra- and extra-luminal PGA samples. (F) Venn diagram of the proteins identified within the intraluminal PGA content compared with published proteome of normal human amniotic fluid (Liu et al., 2019). (G) Schematic of protocol used for metabolomic of PGAs analysis. Intra and extra luminal fluid from day 21 PGAs was collected for metabolomic analysis. Basal media (E6) culture media placed within the same incubator was used as control. Intra-luminal fluid (Intra), extra-luminal fluid (Extra). (H) PCA analysis comparing intra- and extra-luminal PGA fluid and E6 media control. n= 3 replicates per condition. (I) Heat maps showing metabolite abundance in intra-, extra-luminal PGA samples and E6 media control. R1-R3 show different replicates (J) Abundance of selected metabolites in intra-, extraluminal PGA samples and E6 media control. Data are presented as mean ±SEM, n=3. 3’ adenosine monophosphate, hexitol, Hex-2-ulose, creatine, kynurenic acid and folic acid were annotated at Metabolite Standards Initiative (MSI) levels [2, 3, 3, 3, 3 and 1], respectively.

[0029] Figure 14: Functional characterization of PGAs by proteomic and metabolomic analyses (A) Heat map showing levels of all proteins identified within the intra- and the extra-luminal PGA content. (B) Venn diagram of the proteins identified within the intra-luminal PGA content compared with proteomic data set of human and rhesus macaque’s amniotic fluid (Shorey-Kendrick et al., 2023). (C) Abundance of selected metabolites between the intra-, the extra-luminal PGA content and the E6 media control. Data are presented as mean ±SEM, n=3. Allopurinol, L-Aspartic acid, g-Butyrobetaine and Pyruvic acid were annotated at Metabolite Standards Initiative (MSI) levels [3, 1 , 3 and 2], respectively. (D) Heat map of selected proteins associated with folate metabolism and differently expressed within intra- and extraluminal PGA content.

[0030] Figure 15: Development of post-gastrulation amnioids (PGAs) from induced pluripotent stem cells (iPSC) (A) Representative images of PGAs made from KOLF2-iPSCs cultured at day 14. (B) Representative images of PGAs made from MRC5-iPSCs cultured at day 14. (C) Percentage of PGAs made from KOLF2-iPSCs that grew over 320pm in 1 and 14 days. (D) Percentage of PGAs made from MRC5 that grew over 320pm in 1 and 14 days.

[0031] Figure 16: Drug permeation assay using PGAs. PGAs were assessed for drug permeability testing. PGAs were grown for 21 days and incubated with compounds (Propranolol or Chlorothiazide at a final concentration of 10uM). These two compounds were described to have with high (100% for Propranolol) or low (20% for Chlorothiazide) level of intestinal absorption in in vivo studies Kus, M, Ibragimow, I and Piotrowska-Kempisty, H, Pharmaceutics, 2003. Intra and extra-luminal fluids was extracted at various time points (60, 90, 120, 150 and 180 minutes) and analysed using LC-MS.

[0032] DETAILED DESCRIPTION

[0033] The invention provides a 3D, stem cell-derived extra-embryonic model that resembles the post- gastrulation amnion, termed a “post-gastrulation amnioid” (PGA). PGAs recapitulate the epiblast derived definitive extra-embryonic structures of the embryo with high level of reproducibility. It includes a fluid- filled amniotic sac with a transparent membrane consisting of two layers: an outer layer of extra- embryonic mesoderm and an inner layer of amniotic ectoderm, which provide mechanical support for these structures to grow to a critical size in culture for more than 3 months. In addition to the two-layered amnion, PGAs also form yolk sac comprising multiple cell types: endothelium, endoderm, fibroblasts and mesothelium, all surrounded by the extra-embryonic mesoderm that stretches and forms a connecting stalk, altogether resembling in size and structure the supporting embryonic tissues at post- gastrulation. The inventors demonstrate that primed pluripotent stem cells can give rise to a two-layered amnion and accompanying extra-embryonic structures. The high reproducibility of the PGA model suggests it can be an ideal 3D system to study amniogenesis and amniomembryo interactions in-vitro.

[0034] Definitions

[0035] Below are provided certain definitions of terms, technical means, and embodiments used herein.

[0036] In vitro amnion model

[0037] An in vitro amnion model as described herein refers to a composition of cells that reproduces aspects of the amnion in vitro, such as structural characteristics (in particular the two-layered structure), cell lineage characteristics, mechanical characteristics including permeability, among others. The in vitro amnion model may also be described as an amnion organoid, an amnion-like organoid, or a post- gastrulation amnioid (PGA). The in vitro amnion model may be regarded as a specific type of organoid, and provides an “amnion in a dish” to permit the study of the amnion and associated extra-embryonic structures.

[0038] The in vitro amnion model described herein is generally three-dimensional (as opposed to, for example, a monolayer of cells in a 2D culture system). In other words, the cultured cells form spheres, spheroids or aggregates. Amnion models as described herein also generally comprise both an inner cell layer and an outer cell layer, wherein the inner cell layer defines a cavity (an amniotic cavity). This three- dimensional, two-layered structure of the amnion models described herein has not been previously described or achieved in the art.

[0039] Generally, the amnion models described herein are in vitro models, meaning that they comprise cells in culture. They may also be referred to as a “cultured” amnion model.

[0040] The amnion models described herein may be post-gastrulation models, meaning they recapitulate or reproduce the amnion in culture at a post-gastrulation stage of development.

[0041] The in vitro amnion models described herein can recapitulate the early amnion, including a yolk sac and a connecting stalk, or a late amnion, with no yolk sac or in which the yolk sac has regressed.

[0042] The in vitro amnion models described herein are generally mammalian amnion models, for example a human in vitro amnion model. The amnion model may be obtained or obtainable from iPSCs or human embryonic stem cells, by the methods described herein. As is also described herein, the amnion models can be produced by methods which do not result in the destruction of human embryos, for example by obtaining parthenotes or human ESCs obtained via parthenogenesis, according to standard processes known in the art.

[0043] It will be understood that the invention relates only to extra-embryonic structures, and does not model or use an embryo itself. The invention does not rely on the use of, or relate to, fertilized embryos. The in vitro amnion models described herein are not derived from egg donors, for example, and cannot give rise to life. The invention also does not relate to the use of embryos or embryo-derived material ex vivo.

[0044] Inner cell layer

[0045] The amnion models provided herein comprise at least a two-layered structure, including an inner cell layer and an outer cell layer. The inner cell layer generally defines a cavity (i.e. an internal space delimited by the inner cell layer), which recapitulates the amniotic cavity seen in vivo. The amniotic cavity may comprise or contain a fluid. The amniotic cavity may be fluid-filled. The amniotic cavity may comprise or contain amniotic fluid or amniotic fluid-like fluid. The amniotic cavity may be an amniotic sac or amniotic sac-like structure. The amniotic sac (or amniotic sac-like structure) may comprise or contain a fluid. The amniotic sac (or amniotic sac-like structure) may be fluid-filled. The amniotic sac (or amniotic sac-like structure) may comprise amniotic fluid or amniotic fluid-like fluid. The fluid or amniotic fluid-like fluid may recapitulate the amniotic fluid observed in vivo during embryonic development. The term “amniotic fluid-like fluid” is used in the context of this invention to reflect that this is an in vitro model of the amnion, rather than isolated ex vivo extra-embryonic structures.

[0046] The inner cell layer may comprise ectodermal cells. “Ectodermal cells” as used herein refers to cells of the ectoderm, the germ layer that gives rise to the skin, nervous system and sense organs. The inner cell layer may comprise an amniotic ectoderm. The amniotic ectoderm may comprise epithelial cells. In vivo, the amniotic ectoderm is an epithelial layer that lines the amniotic cavity.

[0047] The inner cell layer may express ISL1 , TFAP2C, TFAP2A, GATA3, GABRP and / or VTCN1 . The inner cell layer may be positive for ISL1 , TFAP2C, TFAP2A, GATA3, GABRP and / or VTCN1 . Generally, it is meant that one or more cells within the inner cell layer expresses or is positive for said markers. Expression of said markers or positivity for said markers may be assessed using any suitable method known in the art, for example immunohistochemistry, ELISA, flow cytometry, FACS, Western blot, in situ hybridisation, rt-PCR, sequencing (including single cell sequencing), among others.

[0048] The inner cell layer may express or be positive for ISL1 . The inner cell layer may express or be positive for TFAP2C. The inner cell layer may express or be positive for TFAP2A. The inner cell layer may express or be positive for GATA3. The inner cell layer may express or be positive for one, two, three or all of ISL1 , TFAP2C, TFAP2A, and GATA3. For example, the inner cell layer may express or be positive for ISL1 , TFAP2C, TFAP2A, and GAT A3. Generally, the in vitro amnion model has been cultured for at least 18 hours such that the distinct inner and outer cell layers form. Thereafter, expression at the protein level of ISL1 , TFAP2A and TFAP2C by the inner cell layer is consistent irrespective of time in culture. By “cultured for” and “time in culture”, it is meant that the in vitro amnion model has been maintained in culture for a specified duration since its production or generation (for example since its generation by one of the methods described herein).

[0049] The inner cell layer may express or be positive for GABRP and / or VTCN1. The inner cell layer may express or be positive for GABRP. The inner cell layer may express or be positive for VTCN1. The inner cell layer may express or be positive for GABRP and VTCN1 . In some cases, the in vitro amnion model may comprise an inner cell layer which expresses or is positive for GABRP and / or VTCN1 , wherein the in vitro amnion model has been in culture for at least 6 days, optionally up to 90 days. In some cases, the in vitro amnion model is in culture / is cultured for at least 90 days. In some cases, the inner cell layer does not comprise PGC-like cells. In some cases, the inner cell layer does not comprise primitive streak-like cells. In some cases, the inner cell layer does not comprise PGC-like cells nor primitive streak-like cells. In some cases, the cells of the inner layer do not express BRA. In some cases, the cells of the inner layer do not express SLUG. In some cases, the cells of the inner layer do not express BRA or SLUG (markers of gastrulation / mesoderm i.e., embryonic cells). In some cases, the inner cell layer does not comprise epiblast-like cells. In some cases, the cells of the inner cell layer do not express markers such as Nanog and Sox2. In some cases, the inner cell layer does not comprise epiblast-like cells and the cells of the inner layer do not express markers such as Nanog and Sox2.

[0050] Basement membrane

[0051] In some cases, the amnion model comprises a basement membrane. In some cases, the basement membrane is located between the inner and outer cell layers. In some cases, the basement membrane (or amnion model) comprises proteins associated with extracellular matrix (ECM) synthesis, such as one or more proteins selected from group consisting of: primary collagens, fibronectin, periostin, fibrillins, fibulin-5, lumican and osteopontin. In some cases, the basement membrane (or amnion model) comprises basement membrane assembly proteins, such as one of more proteins selected from the group consisting of laminins, nidogens and heparan sulfate proteoglycan 2. In some cases, the basement membrane is elastic.

[0052] Outer cell layer

[0053] As noted above, the amnion models provided herein comprise at least a two-layered structure, including an inner cell layer and an outer cell layer. The outer cell layer may comprise mesodermal cells. The outer cell layer may comprise an extra-embryonic mesoderm.

[0054] The outer cell layer may express or be positive for GATA6, GATA4, BST2, POSTN, RSPO2, and / or Handl . Generally, it is meant that one or more cells within the outer cell layer expresses or is positive for said markers. Expression of said markers or positivity for said markers may be assessed using any suitable method known in the art, for example immunohistochemistry, ELISA, flow cytometry, FACS, Western blot, in situ hybridisation, rt-PCR, sequencing, among others.

[0055] The outer cell layer may express or be positive for GATA6. In some cases, the in vitro amnion model may comprise an outer cell layer which expresses or is positive for GATA6, wherein the in vitro amnion model has been in culture for at least 6 hours. By “been in culture for”, it is meant that the in vitro amnion model has been maintained in culture for a specified duration since its production or generation (for example since its generation by one of the methods described herein). The outer cell layer may express or be positive for GATA4. In some cases, the outer cell layer may express or be positive for GATA4, wherein the in vitro amnion model has been in culture for at least 12 hours, optionally up to 4 days. The outer cell layer may express or be positive for BST2, POSTN, RSPO2, and / or HAND1. For example, the outer cell layer may express or be positive for BST2, POSTN, RSPO2, and / or HAND1 , wherein the in vitro amnion model has been in culture for at least 1 day. The outer cell layer may express or be positive for BST2, POSTN, RSPO2, and / or HAND1 , wherein the in vitro amnion model has been in culture for at least 2 days.

[0056] In some cases, the inner cell layer comprises ectodermal cells and the outer cell layer comprises mesodermal cells. In some cases, the inner cell layer comprises an amniotic ectoderm and the outer cell layer comprises an extra-embryonic mesoderm. In some cases, the inner cell layer expresses TFAP2C and the outer cell layer expresses GATA6. In some cases, the inner cell layer is positive for TFAP2C and the outer cell layer is positive for GATA6.

[0057] In some cases, the outer cell layer does not comprise PGC-like cells. In some cases, the outer cell layer does not comprise primitive streak-like cells. In some cases, the outer cell layer does not comprise PGC-like cells or primitive streak-like cells. In some cases, the cells of the outer layer do not express BRA. In some cases, the cells of the outer layer do not express SLUG. In some cases, the cells of the outer layer do not express BRA or SLUG (markers of gastrulation / mesoderm i.e., embryonic cells). In some cases, the outer cell layer does not comprise epiblast-like cells. In some cases, the cells of the outer cell layer do not express markers such as Nanog and Sox2. In some cases, the outer cell layer does not comprise epiblast-like cells and the cells of the outer cell layer do not express markers such as Nanog and Sox2.

[0058] Yolk sac

[0059] The amnion models described herein may further comprise a yolk sac or yolk sac-like structure. In vivo, the yolk sac is a small membranous structure, surrounded by extra-embryonic mesoderm, that forms outside the embryo (extra-embryonic) and provides the early embryo with nutrients. During development, the yolk sac reduces in size until it is eventually lost or negligible. The same structure and changes in size are also observed with the in vitro amnion models described herein. For example, an amnion model described herein may further comprise a yolk sac (or yolk sac-like structure), wherein the amnion model has been in culture for at least 24 hours, at least 2 days, at least 3 days, at least 4 days, or at least 5 days. An amnion model described herein may further comprise a yolk sac (or yolk sac-like structure), wherein the amnion model has been in culture for up to 24 days, up to 26 days, up to 28 days, or up to 30 days. The yolk sac may comprise at least one cell type selected from the group consisting of endothelium, endoderm, fibroblasts, and mesothelium. The yolk sac may comprise at least two, at least three, or all cell types selected from the group consisting of endothelium, endoderm, fibroblasts, and mesothelium.

[0060] The yolk sac or yolk sac-like structure may comprise yolk sac endodermal cells. The endodermal cells may express or be positive for SOX17, APOA1 , APOA2, HNF1 B, HNF4A, TTR, and / or FOXA2. The endodermal cells may express or be positive for SOX17, HNF4A, and / or APOA1 .

[0061] The yolk sac or yolk sac-like structure may comprise yolk sac endothelial cells. The yolk sac endothelial cells may express or be positive for SOX17, CD34, PLVAP, ESAM, KIT, KDR, PECAM1 , ACTA2, and / or CDH5. The yolk sac endothelial cells may express or be positive for CD34 and / or PLVAP.

[0062] The yolk sac or yolk sac-like structure may comprise yolk sac mesothelial cells. The yolk sac mesothelial cells may express or be positive for PDPN and / or UPK3B.

[0063] The yolk sac or yolk sac-like structure may comprise yolk sac fibroblast cells. The yolk sac fibroblast cells may express or be positive for PDGFRA and / or VCAN.

[0064] The outer cell layer of the in vitro amnion model may substantially or completely surround the yolk sac (or yolk sac-like structure). For example, the outer cell layer may completely surround the yolk sac.

[0065] Provided herein is an vitro amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity, and wherein the amnion model further comprises a yolk sac, wherein the yolk sac is substantially or completely surrounded by the outer cell layer.

[0066] Connecting stalk (or connecting stalk-like structure)

[0067] The in vitro amnion models described herein may further comprise a connecting stalk. In vivo, the connecting stalk is a thick stalk of extraembryonic membrane that eventually develops into the umbilical cord. The in vitro amnion models described herein can reproduce this extra-embryonic structure.

[0068] The connecting stalk may be formed from the outer cell layer (such as from cells from the outer cell layer).

[0069] The in vitro amnion models described herein may comprise a yolk sac (or yolk sac-like structure) and a connecting stalk (or connecting stalk-like structure).

[0070] Provided herein is an in vitro amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity, and wherein the amnion model further comprises a yolk sac and a connecting stalk, wherein the yolk sac is substantially or completely surrounded by the outer cell layer.

[0071] Any of the in vitro amnion models described herein may be at least 1 mm, at least 2mm, at least 3mm, at least 4mm, at least 5mm, at least 6mm, at least 7mm, at least 8mm, at least 9mm, at least 1 cm, at least 1 .5cm or at least 2cm in diameter.

[0072] Any of the in vitro amnion models described herein may be cultured for at least 1 week, at least 2 weeks, at least 4 weeks, at least 6 weeks, at least 8 weeks, at least 10 weeks, at least 12 weeks, at least 14 weeks, or at least 16 weeks.

[0073] As noted above, the in vitro amnion models described herein may variously be referred to as a post- gastrulation amnioid (PGA) or an amnion organoid. Accordingly, provided herein is a post-gastrulation amnioid (PGA) comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity. Also provided herein is an amnion organoid comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity.

[0074] Methods

[0075] Provided herein are methods for producing an in vitro three-dimensional amnion model. Generally, the methods comprise providing a plurality of cells, treating the plurality of cells with a Bone morphogenetic protein (BMP) agonist, treating the plurality of cells with a WNT agonist, and culturing the plurality of cells.

[0076] The plurality of cells may comprise stem cells. Said stem cells may comprise iPSCs, embryonic stem cells, or a combination thereof. The stem cells may be human stem cells. Where human embryonic stem cells are used, these are obtained via methods which do not result in the destruction of human embryos according to standard processes known in the art.

[0077] Described herein is a method for producing an in vitro three-dimensional amnion model, the method comprising: e) providing a plurality of cells, f) treating the plurality of cells with a BMP agonist, g) treating the plurality of cells with a WNT agonist, and h) culturing the plurality of cells.

[0078] Steps (b) and (c) (treating the cells with a BMP agonist and a WNT agonist) may be performed sequentially, for example wherein the plurality of cells are first treated with BMP agonist, and subsequently treated with the WNT agonist. Alternatively, steps (b) and (c) may be performed simultaneously.

[0079] The treating steps may each independently be performed for 12-36 hours, for example 20-30 hours, or for 24 hours.

[0080] The BMP agonist may be an BMP agonist or activator of a BMP signalling pathway known in the art. For example, the BMP agonist may be selected from the group consisting of BMP4, BMP2, and GATA3. Preferably, the BMP agonist is BMP4.

[0081] Step (b) may comprise treating the plurality of cells with about 1-10 ng / ml BMP agonist (e.g. BMP4), for example about 3-7 ng / ml BMP agonist, preferably about 5 ng / ml BMP agonist.

[0082] The methods described herein may comprise treating the plurality of cells with about 1-10 ng / ml BMP agonist (e.g. BMP4) for 12-36 hours. Preferably, the methods described herein comprise treating the plurality of cells with about 5 ng / ml BMP4 for 24 hours.

[0083] Step (c) may comprise treating the plurality of cells with about 1 -10 pM WNT agonist, for example about 1 -5 pM WNT agonist, preferably about 3 pM WNT agonist. The methods described herein may comprise treating the plurality of cells with about 1-10 pM WNT agonist for 12-36 hours. Preferably, the methods described herein comprise treating the plurality of cells with about 3 pM WNT agonist for 24 hours.

[0084] The Wnt agonist may be any agent that activates Wnt signalling. Suitable Wnt agonists are known in the art. The Wnt agonist may be selected from the group consisting of a glycogen synthase kinase 3 (GSK3) inhibitor, CHIR99021 , a Wnt ligand, a p-catenin agonist, R-spondin, Norrin, or a combination thereof. In some cases, the Wnt agonist is a GSK3 inhibitor, such as CHIR99021 .

[0085] The methods described herein may additionally comprise a step of dissociating the plurality of cells into single cells following the treatment with the Wnt agonist. The plurality of cells may generally be treated with BMP and WNT agonists while in a 2D monolayer culture. Subsequently, individual cells from the plurality of cells may be dissociated, isolated and separated into single wells for 3D culture. The single cells may be cultured in non-adherent plates, such as U-bottom cell repellent plates, which can facilitate 3D culture. Accordingly, the single cells may be cultured in a three-dimensional culture system.

[0086] The methods described herein may further comprise treating the dissociated single cells or the plurality of cells with a Rho kinase (ROCK) inhibitor, for example following treatment with the Wnt agonist. The ROCK inhibitor may be any suitable ROCK inhibitor known in the art, such as fasudil, ripasudil, netarsudil, and Y27632. The ROCK inhibitor may be Y27632. The methods described herein may comprise treating the plurality of cells or the dissociated single cells with about 1-5 pM ROCK inhibitor, preferably 2-4 pM, more preferably 2.5 pM. The methods described herein may comprise treating the plurality of cells or the dissociated single cells with about 1-5 pM ROCK inhibitor for 12-36 hours. Preferably, the methods described herein comprise treating the plurality of cells or the dissociated single cells with about 2.5 pM ROCK inhibitor for 24 hours.

[0087] “Treating” the cells with an agent as used herein generally refers to culturing the cells in the presence of the agent (for example culturing the cells in the presence of a Wnt agonist or BMP agonist (e.g. BMP4)). Such agents may be added to the culture media. Generally, the culture media used for the methods described herein is basal cell culture media, without the addition of additional differentiation factors. The cell culture media may comprise insulin. Generally, the methods described herein do not comprise the addition of extracellular matrix proteins, such as collagens or laminins. The methods may be carried out in vessels that minimise cell attachment, such as U bottom cell repellent plates. The amnion models described herein can be generated without the use of a scaffold.

[0088] In the methods described herein, the plurality of cells or the dissociated single cells may be cultured for at least 1 week, at least 2 weeks, at least 4 weeks, at least 6 weeks, at least 8 weeks, at least 10 weeks, at least 12 weeks, at least 14 weeks, or at least 16 weeks. The in vitro amnion model will be produced after just 1 day in culture, with differentiation between the inner and outer cell layers apparent. Culturing the amnion model for longer durations allows the amnion model to develop in a pattern similar to that seen in vivo, for example demonstrating formation and subsequent shrinkage / regression of the yolk sac. The step of culturing the plurality of cells or the dissociated single cells may therefore be performed until an amnion model is produced, or until a suitable stage of development of the amnion model is reached.

[0089] “Culturing” the plurality of cells or the dissociated single cells generally means maintaining the plurality of cells under conditions suitable to sustain the cells, for example, providing adequate gas exchange and nutrient delivery via a suitable cell culture media. As noted above, the cell culture media may be a basal media, optionally comprising insulin. The cell culture vessels may be any vessel that prevents cell attachment, such as a cell repellent plate, or U-bottom cell repellent plate.

[0090] The plurality of cells may be mammalian cells, such as mouse, rat, monkey or human cells. Preferably, the plurality of cells comprises human cells.

[0091] In some cases, a method for producing an in vitro three-dimensional amnion model comprises: a) providing a plurality of cells, wherein the plurality of cells comprises iPSCs and / or embryonic stem cells, b) treating the plurality of cells with a BMP agonist at about 3-7 ng / ml, wherein the BMP agonist is BMP4, c) treating the plurality of cells with a WNT agonist at about 1-5 pM, wherein the WNT agonist is CHIR99021 , d) dissociating the plurality of cells into single cells, and e) culturing the dissociated single cells for at least 24 hours; wherein steps (b) and (c) are performed sequentially and are each independently performed for 20-30 hours.

[0092] A method for producing an in vitro three-dimensional amnion model may comprise: a) providing a plurality of cells, wherein the plurality of cells comprises iPSCs and / or embryonic stem cells, b) treating the plurality of cells with a BMP agonist at about 5 ng / ml, wherein the BMP agonist is BMP4, c) treating the plurality of cells with a WNT agonist at about 3 pM, wherein the WNT agonist is CHIR99021 , d) dissociating the plurality of cells into single cells, and e) culturing the dissociated single cells for at least 24 hours; wherein steps (b) and (c) are performed sequentially and are each independently performed for 24 hours.

[0093] In some cases, a method for producing an in vitro three-dimensional amnion model comprises: a) providing a plurality of cells, wherein the plurality of cells comprises iPSCs and / or embryonic stem cells, b) treating the plurality of cells with a BMP agonist at about 3-7 ng / ml, wherein the BMP agonist is BMP4, c) treating the plurality of cells with a WNT agonist at about 1-5 pM, wherein the WNT agonist is CHIR99021 , d) dissociating the plurality of cells into single cells, e) treating the dissociated single cells with a ROCK inhibitor at about 1-5 pM, wherein the ROCK inhibitor is Y27632, and f) culturing the dissociated single cells for at least 24 hours; wherein steps (b) and (c) are performed sequentially, and wherein steps (b), (c) and (e) are each independently performed for 20-30 hours.

[0094] A method for producing an in vitro three-dimensional amnion model may comprise: a) providing a plurality of cells, wherein the plurality of cells comprises iPSCs and / or embryonic stem cells, b) treating the plurality of cells with a BMP agonist at about 5 ng / ml, wherein the BMP agonist is BMP4, c) treating the plurality of cells with a WNT agonist at about 3 pM, wherein the WNT agonist is CHIR99021 , d) dissociating the plurality of cells into single cells, e) treating the dissociated single cells with a ROCK inhibitor at about 2.5 pM, wherein the ROCK inhibitor is Y27632, and f) culturing the dissociated single cells for at least 24 hours; wherein steps (b) and (c) are performed sequentially, and wherein steps (b), (c) and (e) are each independently performed for 24 hours.

[0095] Also provided herein is an in vitro amnion model obtained or obtainable by the methods described herein.

[0096] The methods described herein are highly reproducible, and produce a consistent and predictable in vitro amnion model with a reproducibility rate of greater than 90% (in other words, at least 90% of attempts to put the method into practice result in the in vitro amnion model as described within 10 days). In contrast, methods ofthe art have reproducibility rates closerto 1 %. The methods described herein have a reproducibility rate greater than 90%. In the methods described herein, following step (d) the plurality of cells forms an in vitro amnion model at least 90% of the time, wherein the amnion model is a three- dimensional amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity. This in turn renders the in vitro amnion model a useful platform for research and therapeutic uses, as described below.

[0097] Uses and applications of the in vitro amnion models

[0098] The in vitro amnion models described herein may advantageously be used in a variety of research and therapeutic settings. An amniotic membrane (or membranes) may also be derived or obtainable from the in vitro amnion models described herein, and may also be used in a variety of applications. Accordingly, also provided herein is an amniotic membrane derived or obtainable from an in vitro amnion model described herein. Any of the methods for producing an in vitro three-dimensional amnion model may also comprise a further step of isolating an amniotic membrane from the resulting amnion model (e.g. harvesting or otherwise collecting the amniotic membrane). The amniotic membranes can advantageously be isolated from the three-dimensional amnion model at any point while in culture, and so can be used to recapitulate different stages of development depending on the length of time in culture. The amniotic membrane described herein may comprise one or more features of the amnion model described herein.

[0099] The in vitro amnion model and / or an amniotic membrane may be used in a method of drug screening. For example, they may be used to study the impact of drugs on amnion development, yolk sac changes, cell differentiation and expression patterns, etc. They may also be used to assess drug toxicity and pharmacokinetics. In vitro amnion models and / or amniotic membranes as described herein may also be used to assess membrane permeability, for example permeability to drugs, small molecules, biologies, ions, etc. Generally it will be understood that such methods are practiced in vitro. The in vitro amnion model and / or an amniotic membrane may be used in methods of tissue engineering. Such methods are described in, for example, Elkhenany H et al. Applications of the amniotic membrane in tissue engineering and regeneration: the hundred-year challenge. Stem Cell Research & Therapy. 13, 8 (2022) ; and Fenelon M et al., Applications of Human

[0100] Amniotic Membrane for Tissue Engineering. Membranes. 11 :387 (2021) doi:

[0101] 10.3390 / membranes11060387. By way of example, an amniotic membrane or amnion model as described herein may be used as a cellular scaffold and / or source of growth factors, to support the growth and development of other tissues. Generally it will be understood that such methods are practiced in vitro. In some cases, the amnion model or membrane are cultured in suspension (i.e., free floating) and do not require a matrix or matrix substrate. In some cases, the plurality of cells does not require treatment with growth factors after (initial) treatment with a BMP agonist and a WNT agonist (i.e., after step c in methods described herein).

[0102] The in vitro amnion model and / or an amniotic membrane may be used in regenerative medicine. Such uses are described in, for example, Farhadihosseinabadi B et al., Amniotic membrane and its epithelial and mesenchymal stem cells as an appropriate source for skin tissue engineering and regenerative medicine. Artificial Cells, Nanomedicine and Biotechnology. 46 (2018); Zhang Q & Lai D, Application of human amniotic epithelial cells in regenerative medicine: a systematic review. Stem Cell Research and Therapy. 11 :439 (2020); and Toda A et al., The Potential of Amniotic Membrane / Amnion-Derived Cells for Regeneration of Various Tissues. Journal of Pharmacological Sciences. 103:215-228 (2007). By way of example, an amniotic membrane or amnion model as described herein may be used to support or facilitate the growth of a skin substitute (for example for use in skin grafting), or may even be used as the skin substitute itself. Accordingly, also provided herein is an amniotic membrane or amnion model for use in regenerative medicine. Any of the amniotic membranes or amnion models described herein may also be used as a source of amniotic epithelial cells.

[0103] Described herein are in vitro amnion models and / or amniotic membranes for use in a method of wound healing. The suitability of amniotic membranes or amniotic membrane derivates for wound healing is described in, for example, Schmiedova I et al., Using of Amniotic Membrane Derivatives for the Treatment of Chronic Wounds. Membranes. 11 :941 (2021).

[0104] Described herein are in vitro amnion models and / or amniotic membranes for use in a method of treating burns. The suitability of amniotic membranes for treating burns is described in, for example, Oba J et al., Hyperdry human amniotic membrane application as a wound dressing for a full-thickness skin excision after a third-degree burn injury, Burns & Trauma, 8, (2020). More specifically, described herein are in vitro amnion models and / or amniotic membranes for use in a method of treating ocular surface lesions caused by burns. The suitability of amniotic membranes for treating ocular surface lesions caused by burns is described in, for example, Shimazaki, J. et al. Amniotic Membrane Transplantation for Ocular Surface Reconstruction in Patients with Chemical and Thermal Burns. Ophthalmol. 1997;104:2068-76. Described herein are in vitro amnion models and / or amniotic membranes for use in a method of treating neurotrophic corneal ulcers. The suitability of amniotic membranes for treating neurotrophic corneal ulcers is described in, for example, Chen, H. et al. Amniotic membrane transplantation for sever neurotrophic corneal ulcers. J Opthalmol. 2000;84:826-33.

[0105] Described herein are in vitro amnion models and / or amniotic membranes for use in a method of ocular (surface) reconstruction . The suitability of amniotic membranes for ocular (surface) reconstruction is described in, for example, Walkden A. Amniotic Membrane Transplantation in Ophthalmology: An Updated Perspective. Clin Ophthalmol. 2020;14:2057-2072 and Finger, P. et al. Super-Thick Amniotic Membrane Graft for Ocular Surface Reconstruction. Am J Ophthalmol. 2019;198:45-53.

[0106] Described herein are in vitro amnion models and / or amniotic membranes for use in a method of treating neuropathic corneal pain. The suitability of amniotic membranes for treating neuropathic corneal pain is described in, for example, Morkin, M. Efficacy of self-retained cryopreserved amniotic membrane for treatment of neuropathic corneal pain. The Ocular Surface. 2018;16:132-38.

[0107] Described herein are in vitro amnion models and / or amniotic membranes for use in a method of treating ocular surface disorders. The suitability of amniotic membranes for treating ocular surface disorders is described in, for example, Kotomin, I et al. Sutureless Fixation of Amniotic Membrane for Therapy of Ocular Surface Disorders. PLoS ONE. 2015;10:e0125035.

[0108] The in vitro amnion model and / or an amniotic membrane as described herein may have high throughput capability. Particularly as a result of the high reproducibility of the methods described herein to produce amnion models, they may be used to e.g. screen drugs or assess membrane permeability with a higher throughput capability relative to models known in the art.

[0109] Aspects and embodiments described herein with the term “comprising” may include other features or steps within the scope. It is also understood that aspects and embodiments described as “comprising” also describes aspect and embodiments wherein the term “comprising” is replaced by the term “consisting essentially of’ or “consisting of’.

[0110] The phrase "selected from the group comprising" may be substituted with the phrase "selected from the group consisting of and vice versa, wherever they occur herein.

[0111] It is also understood that the application discloses all combinations of any of the above aspects and embodiments described above with each other, unless the context demands otherwise. Similarly, the application discloses all combinations of the preferred and / or optional features either singly or together with any of the other aspects, unless the context demands otherwise. The invention will now be further described by way of the following Examples, which are meant to serve to assist one of ordinary skill in the art in carrying out the invention and are not intended in any way to limit the scope of the invention, with reference to the Figures.

[0112] EXAMPLES

[0113] Human embryonic stem cell culture

[0114] The majority of experiments were conducted using wild-type H1 human embryonic stem (hES) cell line (XY; WA01 , WiCell; RRID:CVCL_9771) [1], hESC (H1 and RUES2-GLR) and iPS cells were cultured in feeder-free condition on growth factor reduced (GFR) matrigel-coated (1 :100 diluted (Corning)) plates and maintained in chemically defined medium (mTeSRI , STEMCELL

[0115] Technologies) at 37°C with 5% CO2 under humidified condition. hES cells were passaged every 3-4 days by incubating for 6 minutes with gentle dissociation buffer (STEMCELL Technologies) at room temperature and replated in small clumps of cells in mTeSR. hESC were routinely karyotyped and monthly screened for mycoplasma using PCR.

[0116] Generation of H1 H2B-mCherry Knock-in line

[0117] H1 hES cells were used to generate a monoclonal H2B-mCherry knock in (KI) line using CRISPR- Cas9 technology. A homology donor plasmid was generated containing a 1000bp left homology arm sequence upstream of stop codon, a cassette featuring mCherry sequence and a 1000bp right homology arm down stream of stop codon. PCR was carried out using Q5 High Fidelity

[0118] Polymerase and the H2B donor plasmid as template to generate a sequence covering around 500bp of left and right homology arms. The amplified DNA fragments were purified using Wizard

[0119] PCR Clean-up system (Promega, A9282). 500ng of purified dsDNA template along with 1500ng

[0120] HiFi SpCas9 (Alt-R Sp HiFi Cas9 Nuclease V3; IDT, 1081061), 10pmol crRNA: hH2B KI01 (Target 5’- GCGCTAAGTAAACAGTGAGT-3’) (Dharmacon) and 10pmol trRNA (Dharmacon) with 80K cells per 10ul transfection using Neon electroporation system (ThermoFisher Scientific).

[0121] Subsequently, monoclonal lines were generated by cell sorting in mTeSRI with CloneR2 supplement, as per manufacturer’s recommendations (STEMCELL Technologies). Homozygous clone was identified by PCR screening and Sanger sequencing. No unwanted mutation was detected. Low pass sequencing validated the chromosomal integrity of the selected clone.

[0122] Generation of 3D extra-embryonic PGA model

[0123] To generate 3D extra-embryonic PGA model, hES cells were grown to confluency in a 6-well matrigel-coated plate in mTeSR and plated in very small clumps of 15-20 cells. Following 24h incubation (37°C, 5% CO2) cells were treated with 5ng / ml BMP4 in mTeSR for 24h and subsequently treated with 3pM WNT agonist CHIR99021 in NutriStem media for further 24h, which the inventors refer here as extra-embryonic progenitor state. Cells were dissociated for 10 minutes at 37°C with gentle dissociation buffer to single cells and washed with PBS. Cells were resuspended in E6 media (STEMCELL Technologies), counted and diluted at 10 cell / pL of E6 media containing ROCK inhibitor Y27632 (5 pM). 40 pL of cell suspension was plated in each well of Cell Repellent-Treated U-Shaped-Bottom CELLSTAR 96-well Microplate (Greiner Bio- One Ltd, 650970). The plate was centrifuge at 700rpm for 2 minutes at room temperature and returned to the incubator (37°C, 5% CO2) for 24 hours to allow aggregation take place. 150pL of E6 medium was added to each well and replaced again after 24h with 150pL of E6 medium and maintained in the same media for up to 16 days. For prolonged culture, the 3D structures were transferred into 24-well plates at day 18 and subsequently to larger vessels thereafter.

[0124] Media was equilibrated for 30 minutes in incubator (37°C, 5% CO2) prior to exchange. For coculture experiments extra-embryonic progenitors were combined with undifferentiated H1 ES cells at ratio of 1 :1 and plated in 40pL aliquots at 10 cell / pL of E6 media containing ROCK inhibitor Y27632 (5pM) and maintained in E6 media as mentioned above. In some occasions, day 5 chimeric structures were cultured in modified version of EUCM2 [2] by adding the following compositions to E6 media: 0.5% CMRL media (GIBCO), 1x of ITS-X supplement (Thermo Fisher Scientific), 100nM T3 (Sigma Aldrich), 8nM B-estradiol (Sigma-Aldrich), 200ng / mL progesterone (Sigma-Aldrich) and 30% FBS (Sigma Aldrich).

[0125] Knock-downs with siRNA

[0126] For siRNA knock-downs, hES cells grown in mTeSR were dissociated for 10 minutes at 37°C with gentle dissociation buffer to single cells. Cells were counted and diluted to a density of 1 .25x105 cells / mL in mTeSR containing ROCK inhibitor Y27632 (10pM) and reverse transfected. Transfection was carried out using Lipofectamine RNAiMAX (Invitrogen) and ON-TARGET-Plus pool siRNAs (Dharmacon) at a final concentration of 25nM. After an over-night incubation media was replaced with fresh mTeSR and incubated for further 8h for the cells to form small colonies. Subsequently cells were treated with 5ng / ml BMP4 in mTeSR for 24h, 3pM CHIR99021 in Nutristem media for further 24h and used to make PGAs.

[0127] Generation of H1-Cas9-VPR and crRNA-induced gene activation

[0128] H1 hES cells were used to generate a monoclonal line with an integrated doxycycline-inducible gene activation system. ES cells were transfected with plasmids containing cassette featuring an inducible nuclease-dead Cas9 fused with VP64-p65-Rta transactivation domains (ndCas9-VPR) to the AAVS1 (GeneCopoeia) using Fugene HD (Promega) and selected with Puromycin (100ng / mL). Subsequently, monoclonal lines were generated by cell sorting in mTeSRI with CloneR2 supplement as per manufacturer’s recommendations (STEMCELL Technologies). Expression of a gene of interest was induced by transfecting the H1-Ca9-VPR with a specific crRNAa (CRISPRa RNA) designed against a target promoter, along with a tracrRNA (Trans-activating CRISPR RNA). To transfect cells, crRNA (50nM) and tracrRNA (50nM) (Dharmacon) were first combined in HBSS (Hanks Buffered Saline Solution), mixed with a Lipofectamine RNAiMAX (1 :50) master mix and incubated for 30 minutes. hES cells were dissociated for 10 minutes at 37°C with gentle dissociation buffer to single cells. Cells were counted and diluted into a density of 1 .25x105 cells / mL in mTeSR containing ROCK inhibitor Y27632 (10pM) and Doxycycline (1 pg / mL) and reverse transfected. Following an over-night incubation media was replaced with fresh mTeSR containing Doxycycline (1 pg / mL) in mTeSR for further 24 h and used for PGA generation.

[0129] Immunofluorescence staining

[0130] Cells were fixed with formalin solution neutral buffered, 10% (Merck) for 30 minutes at room temperature. After 3x washes in PBS samples were transferred to Eppendorf tubes using a 1 ml pipettes with cut tip. Cells were permeabilised with 0.3% Triton X-100 in PBS for 30 minutes, blocked with 10%FBS / 3%BSA in PBS for 1 h and incubated with the primary antibodies in 0.3% Triton X-100 / 3%BSA in PBS at 4°C overnight. Samples were then washed 3 times with 1 % Tween in PBS, incubated with the secondary antibodies at room temperature for 2h followed by DAPI, 7-aminoactinomycin D (7-AAD) or TO-PRO™-3 dye (Thermo Fisher Scientific) staining.

[0131] Light-sheet microscopy

[0132] For samples ranging from days 1 to 8 imaging was carried out on an LS1 light sheet microscope system by Viventis. Samples from day 1-7 were mounted in 1 % low-melting temperature agarose (SERVA 11384.01) within chamber of multi-well FEP holders (Viventis). Day 8 Samples with spherical amnion-like structure were imaged in water to avoid collapse of spherical shape. LS1

[0133] Live Light sheet microscope was eguipped with two opposing Nikon Plan Fluor 10X objectives, a Nikon 25X W NA 1 .1 detection lens and Hamamatsu ORCA-Fusion Camera. Image volumes were acguired using the following combination of laser lines and filters: Laser 488nm, 60 mW diode laser, Semrock FF03-525 / 50 or Semrock FF01 -523 / 610 or Chroma ZET405 / 488 / 561 / 640mv2 and Laser 561 nm, 50mW DPSS laser, Semrock FF01-523 / 610 or Chroma ZET405 / 488 / 561 / 640mv2. Z-step between planes was set to 2pm with a thickness of the light-sheet set at 3.3 and the final magnification at 18.7x. Images were processed using FIJI software [3]. 3D image processing was carried out using imaging software Imaris.

[0134] For samples larger than 700pm (after day 8) imaging was performed using a Bruker-Luxendo MuVi- SPIM light-sheet microscope. Samples were placed in capillary containing water to avoid the collapse of the amniotic cavity. Bruker-Luxendo MuVi- SPIM light-sheet microscope was equipped with two 20x objectives for detection (Olympus XLUMPLFLN 20XW NA 1.0), two 10x objectives for illumination (Nikon CFI Plan Fluor 10X W NA 0.3) and two CMOS cameras (Hamamatsu Orca Flash 4.0 v3). Image volumes were acquired using the following combination of laser lines and filters: 1) 405 nm and BP 418-462; 2) 488 nm and BP 499-569; 3) 594 nm and BP 610-651 ; z-step between 8 planes has been set to 1 pm and a magnification changer was used giving a final magnification of 16.6x and a voxel size of 0.392 x 0.329 x 1 m. Thickness of the lightsheet was 4.1 pm and the “line mode” (virtual confocal slit of 20 pixels) has been used to increase the image contrast. For samples bigger than the field of view tiling and stitching of multiple volume have been performed. Final image volumes are the result of the fusion of the 2 opposite camera views. Registration and fusion of volumes was performed using the Luxendo processing software (GUI 2.9.0). 3D image processing was carried out using imaging software Imaris.

[0135] 2D microscopy

[0136] For imaging in monolayer, hES cells were seeded into 96-well glass-bottom imaging plates (Greiner Bio-One Ltd). Following fixation and immunofluorescence staining cells were imaged using ScanR automated inverted epifluorescence microscope system IX83 (Olympus). ScanR images were acquired with a 20x plan (UCPLFLN) fluorescence objective (NA 0.7) and a sCMOS (Orca Flash 4.0, Hamamatsu) camera. LED-based illumination (SpectraX LED, Lumenco) was used for excitation. Excitation (ex) and emission (em) filters were as follows: DAPI ex: 391 / 20nm, em:440 / 521 / 607 / 700nm; GFP / Alexa 488 ex: 474 / 27nm, em: 440 / 521 / 607 / 700nm and mCherry ex: 554 / 23nm, em: 440 / 521 Z607 / 700nm.

[0137] Live cell imaging was carried out on an IncuCyte Zoom® (Essen BioScience) equipped with temperature, humidity and CO2 levels control to an optimal cell culture condition. IncuCyte Zoom images were acquired with either 4x or 10x plan fluorescence objectives and a CCD camera. Fluorescence excitation (ex) and emission (em) filters were as follows: Green channel ex: 440-480nm, em: 504-544nm; Red channel ex: 565-605nm em: 625-705nm. False colour and merged- channel images were generated using Fiji [3]. To quantify the morphology of structures using brightfield images, the inventors trained a bespoke neural network using the ZEISS APEER image analysis toolkit. Custom neural network was trained using manually annotated training images and the trained network was deployed for segmentation and quantification using the napari CZANN segmentation plugin and scikit-image in python [4, 5]. To measure the reproducibility of the system the inventors set a manual area threshold based on visual inspection of whether structures successfully expanded into mature amniotic structures. For both characterisation and screening studies, pairwise comparisons were performed using pairwise Mann-Whitney tests corrected for multiple testing.

[0138] Single cell RNA sequencing (scRNA-seq) scRNA-seq was carried out in two biological replicates per time point. Cell dissociation was carried out using Trypsin-EDTA (0.25%) (Thermo Fisher) at 37°C for all samples with varying incubation time for samples at different stages. For monolayer hES cells, 1 well of a 6 well plate was used per biological replicate and dissociated for 5 minutes. To dissociate 3D samples, they were transferred from 96 well plates to 15 ml conical tube, washed with PBS and incubated at 37°C in 500 pL of Trypsin-EDTA (0.25%) with interval gentle pipetting. For day 1 , 250 were collected per biological replicate and incubated for 5 minutes. For day 3, 150 were collected per biological replicate and incubated for 8 minutes. For day 5, 90 were collected per biological replicate and incubated for 10 minutes. For day 8, 45 were collected per biological replicate and incubated for 20 minutes. Single cell dissociated samples were subsequently washed 5 min at 1000 RPM with PBS and re-suspended in 1 % BSA in PBS. Duplicate samples from biological replicates were multiplexed together using 10x genomics 3' Cellplex kit set A (PN-1000261) according to the manufacturer’s instructions. Labelled samples were assessed individually for cell number and viability using a LUNA-FX7™ Automated Cell Counter. Equal number of cells per labelled replicate were mixed together and cell number were counted again. 2 x biological replicates per condition were pooled and 20K cells were processed using 10x Chromium 3’ mRNA-Seq kit version 3.1 (PN-1000121) following manufacturer’s instructions with an estimated recovery of 10K cells. Libraries were sequenced using NovaSeq 6000 with an average of 50K reads per cell for the mRNA libraries and 5K reads / cell for the cell plex library.

[0139] Quantification and statistical analysis

[0140] Pre-processing, integration and clustering scRNAseq data

[0141] Cell Ranger multi (v7.1.0, 10x Genomics) was used to demultiplex the replicates and quantify gene expression, specifying chemistry = SC3Pv3 argument. Reads were aligned to a custom reference genome built by adding the mCherry sequence to the prebuilt human reference GRCh38-2020-A (10x Genomics). All subsequent analyses were carried out using Seurat (v4) package [6, 7], with default parameters unless specified, in R-4.2.0 (R Core Team (2022) https: / / www.R-project.org / ). Primary filtering was performed using sample-specific thresholds identified using median absolute deviation (MAD) measures for cells expressing > 3 MAD of mitochondrial genes along with < 3 MAD of total number of detected features. Each individual dataset was “LogNormalize” and the top 2,000 highly variable genes were identified using the

[0142] “FindVariableFeatures” function based on “vst” method. The two replicates per condition were integrated (“IntegrateData” function) using the Seurat standard CCA integration workflow, after identifying 2,000 anchor features. The inventors run PCA selecting the first 50 PCs, constructed a shared nearest neighbour graph using the “FindNeighbors” function and clustered the cells using Louvain clustering (“FindClusters” function) at resolution 0.8. Additional filtering of low-quality cells was performed at this stage and clusters with high average number of mitochondrial content (percent.mt > 20) or low average number of molecules (nCount_RNA < 15,000) were removed. Quality-controlled individual replicates were then reprocessed and integrated again as previously described to create time point condition datasets, and further integrated into timecourse datasets. The resulting integrated datasets were visualized on the UMAP space using the first 50 PCs. Integrated clusters at resolution 1 for the PGA dataset and resolution 0.2 for the PGA co- culture dataset were manually annotated based on a combination of the expression of well- known cell-type specific marker genes, and top cluster specific gene markers identified using the “FindAIIMarkers” function (with the settings only.pos = TRUE, logfc.threshold = 0.25, min. pct = 0.2).

[0143] Comparison and integration with published datasets

[0144] For benchmarking cell-type annotation, labels from published single-cell references from both in vitro model systems and in vivo human and primate embryonic data [8-13] were projected into our PGA time-course data using semap (v1 .20.2)

[0014] . Similarities were calculated based on the default 500 most informative features within each reference dataset, the “scmapCIuster” method was used with a similarity threshold of 0.5, and predicted annotations were visualized onto UMAP embedding. The following available reference datasets with published cell type annotations were analysed: human in vivo embryo data from Tyser et al [8] (ArrayExpress accession E-MTAB- 9388); cynomolgus monkey in vivo embryo data from Ma et al [9] (Gene Expression Omnibus (GEO) accession GSE1301 14), and human in vitro data from Oldak et al

[0010] (Gene Expression Omnibus (GEO) accession GSE229578), Zheng et al

[0011] (GEO accession GSE134571), Pham et al

[0012] (GEO accession GSE191285) and Goh et al

[0013] (ArrayExpress accession E-MTAB-10552). Gene symbols from cynomolgus monkey to human were converted using Ensembl biomaRt.

[0145] Trajectory inference and differential gene expression Developmental cell trajectories were inferred using Slingshot (v2.6.0)

[0015] . The analysis was performed on the UMAP space and the EPI cluster was specified as the starting point.

[0146] Differentially expressed genes along the inferred trajectories were further investigated using tradeSeg (v1 .14.0)

[0016] . All genes with at least 2 counts in 1 % of the cells were kept for the analysis. For each gene, a negative binomial generalized additive model (NB-GAM) was fitted with the “fitGAM” function and 8 knots, as determined using the “evaluateK” function. The inventors used the “associationTest” to identify genes with expression patterns associated with pseudotime along each lineage, and the “earlyDETest” at the bifurcation point (between knots 2 and 3) to characterize potential early drivers of the differentiation towards extra-embryonic mesoderm and amnion lineages. All tests were performed against a fold change of Iog2(2), and only genes with a false discovery rate < 0.05 were considered. Heatmaps were generated using the smoothed scaled expression of significant differential genes and the cells were ranked by increasing pseudotime. The Ward.D2 method was used for hierarchical clustering and correlation to determine the distance. Pathway analysis was performed using the “enricher” function from ClusterProfiler (v4.6.2)

[0017] for C2-CP:REACTOME (MSigDB v2023.1), and only pathways with an adjusted p-value lower than 0.05 were considered statistically significant. Gene regulatory network inference

[0147] The SCENIC pipeline (pySCENIC vO.12.1) [18, 19] was used to infer transcription factor (TF) regulatory target networks or regulons. The normalized matrix filtered to kept all genes with at least 2 counts in 1% of the cells was used as input. Briefly, co-expression modules between TFs and candidate target genes were inferred using GRNBoost2

[0020] and indirect targets were pruned from these modules using cis-regulatory motif discovery (v10 human motif collection and hg38 databases from cisTarget). The activity of each resulting regulon in each cell was quantified using the AUCell algorithm, and the default thresholds were applied to binarize the activity in an ‘on’ or ‘off state. Cell-type specific regulons were identified based on the Regulon Specificity Score (RSS)

[0021] .

[0148] RNA velocity

[0149] RNA velocity analysis was performed using scVelo (vO.2.4)

[0022] . Spliced and unspliced count matrices were quantified from Cell Ranger aligned outputs for each sample with velocyto

[0150] (vO.17.17)

[0023] , combined for all samples with loompy (v3.0.6), and further merged with the integrated Seurat dataset previously converted to an h5ad object. Genes with less than 20 spliced and unspliced counts were filtered out, the top 2,000 highly variable genes were selected, and first- and second-order moments were computed among 30 nearest neighbours in the PCA space (first 30 PCs). Velocities were estimated using the generalized dynamical model and visualised as streamlines in the UMAP embedding. The latent time was computed using the initial and terminal states identified with CellRank v1.5.0

[0024] .

[0151] Data Availability

[0152] All raw and processed (scRNA) sequencing data generated in this study, including normalized counts and metadata annotations, have been deposited at NCBI Gene Expression Omnibus (GEO) under accession number: GSE250235.

[0153] References 1

[0154] . Thomson, J.A., et al., Embryonic stem cell lines derived from human blastocysts, science, 1998. 282(5391): p. 1145-1147.

[0155] 2. Tarazi, S., et al., Post-gastrulation synthetic embryos generated ex utero from mouse naive ESCs. Cell, 2022. 185(18): p. 3290-3306. e25.

[0156] 3. Schindelin, J., et al., Fiji: an open-source platform for biological-image analysis. Nature methods, 2012. 9(7): p. 676-682.

[0157] 4. Contributors, N., napari: a multi-dimensional image viewer for Python. Zenodo https: / / doi. org / 10.5281 / zenodo, 2019. 3555620.

[0158] 5. van der Walt, S., et al., T. scikit-image contributors, “sci kit-image: image processing in {P} ython,” PeerJ, vol. 2, p. e453, 2014. 6. Butler, A., et al., Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nature biotechnology, 2018. 36(5): p. 411-420.

[0159] 7. Stuart, T., et al., Comprehensive integration of single-cell data. Cell, 2019. 177(7): p. 1888-1902. e21 .

[0160] 8. Tyser, R.C., et al., Single-cell transcriptomic characterization of a gastrulating human embryo. Nature, 2021 . 600(7888): p. 285-289.

[0161] 9. Ma, H., et al., In vitro culture of cynomolgus monkey embryos beyond early gastrulation. Science, 2019. 366(6467): p. eaax7890.

[0162] 10. Oldak, B., et al., Complete human day 14 post-implantation embryo models from naive ES cells. Nature, 2023: p. 1-3.

[0163] 11 . Zheng, Y., et al., Controlled modelling of human epiblast and amnion development using stem cells. Nature, 2019. 573(7774): p. 421-425.

[0164] 12. Pham, T.X.A., et al., Modeling human extraembryonic mesoderm cells using naive pluripotent stem cells. Cell stem cell, 2022. 29(9): p. 1346-1365. e10.

[0165] 13. Goh, L, et al., Yolk sac cell atlas reveals multiorgan functions during human early development. Science, 2023. 381 (6659): p. eadd7564.

[0166] 14. Kiselev, V.Y., A. Yiu, and M. Hemberg, semap: projection of single-cell RNA-seq data across data sets. Nature methods, 2018. 15(5): p. 359-362.

[0167] 15. Street, K., et al., Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC genomics, 2018. 19: p. 1-16.

[0168] 16. Van den Berge, K., et al., Trajectory-based differential expression analysis for single-cell sequencing data. Nature communications, 2020. 11 (1): p. 1201.

[0169] 17. Wu, T., et al., clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The innovation, 2021. 2(3).

[0170] 18. Aibar, S., et al., SCENIC: single-cell regulatory network inference and clustering. Nature methods,

[0171] 2017. 14(11): p. 1083-1086.

[0172] 19. Van de Sande, B., et al., A scalable SCENIC workflow for single-cell gene regulatory network analysis. Nature Protocols, 2020. 15(7): p. 2247-2276.

[0173] 20. Moerman, T., et al., GRNBoost2 and Arboreto: efficient and scalable inference of gene regulatory networks. Bioinformatics, 2019. 35(12): p. 2159-2161.

[0174] 21. Suo, S., et al., Revealing the critical regulators of cell identity in the mouse cell atlas. Cell reports,

[0175] 2018. 25(6): p. 1436-1445. e3.

[0176] 22. Bergen, V., et al., Generalizing RNA velocity to transient cell states through dynamical modeling. Nature biotechnology, 2020. 38(12): p. 1408-1414.

[0177] 23. La Manno, G., et al., RNA velocity of single cells. Nature, 2018. 560(7719): p. 494-498.

[0178] 24. Lange, M., et al., CellRank for directed single-cell fate mapping. Nature methods, 2022. 19(2): p. 159-170.

[0179] Results

[0180] Formation of a fluid filled, two-layered amniotic sac from primed hES cells In order to establish a 3D embryonic stem based model that could accurately capture early amniogenesis and give rise to a fully formed, amnion-like structure, the inventors treated 2D cultures of H1 , human embryonic stem (hES) cells, in a stepwise manner, starting with exposure to low levels of BMP4 to promote the induction of amniotic ectoderm, followed by a WNT agonist to induce the mesodermal lineage (Figure 1 B). The initial 24-hour pulse of stimulation led to the expression of GAT A3 at the periphery of hES cell colonies, and the subsequent 24-hour pulse resulted in a mixed population of cells expressing both amniotic ectoderm and mesoderm markers (Figure 2A).

[0181] Next, the inventors generated cell aggregates in Essential-6 media (E6) using U-bottom low-adherence plates, which grew until day 18 (Figure 1 C). Within the first 24 hours, the aggregates formed spheroids that self-organized into two distinct populations (Figure 1 D-E and 2B-C). One represents an inner cell compartment, positive for ISL1 , TFAP2C, TFAP2A, and GATA3, resembling amniotic ectoderm. Surrounding this is a layer of GATA6, Handl -positive mesodermal cells (Figure 1 D and 2B). These two-layers resemble the amnion’s outer layer of extra- embryonic amniotic mesoderm and its inner layer of amniotic ectoderm (Figure 1A, 1 D-E). Over the following 24 hours, the amniotic ectoderm domain begins to form an amniotic cavity, gradually expanding in size over the coming days of culture (Figures 1 D-E and 3B-D) acquiring epithelial-like morphology (Figure 12). By day 18, the inventors observed the formation of fluid-filled amniotic-sac like structures measuring approximately 2-3 mm in diameter (Figure 1 D-E). The duallayered structure provided mechanical support, enabling it to grow in culture for more than 3 months and reach a size of over 1 .5 cm in diameter (Figure 1 F). The dual-layered cellular structure produced a basement membrane that thickened as the PGA expanded (Figure 13), providing structural support and enabling PGAs to grow in culture for more than 3 months and reach a size of over 1 .5 cm in diameter (Figure 1 F).

[0182] Notably, the generation of these 3D amnioid structures did not require the addition of extracellular matrix for structural support nor external stimuli to direct self-organization and specification. Except for the first 48 hours of 2D culture it was entirely driven by self-organization in E6 basal media (Figure 1 B), suggesting that early signals can drive late lineage specification and highlight the extraordinary self-organization capacity of pluripotent stem cells.

[0183] Morphological analysis through time revealed that the amnion-like structures gradually expanded to an area of 0.84 mm2 over the course of 10 days (Figure 1 G-J). The inventors observed changes in the eccentricity of structures through time as they transition from round aggregates to more elongated structures before the expansion of the amniotic cavity (Figure 1 H, 1J). Furthermore, the inventors see a high level of reproducibility between independent experiments, with approximately 92% of aggregates forming fluid filled amniotic sac structures within 10 days (Figure 1 K). Similar levels of reproducibility were seen for other cell lines (Figure 1 L) and, to a lower extend, with iPSC lines (Figure 16).

[0184] Altogether, this data shows that primed embryonic stem cells can give rise to two-layered amnion like sacs and suggests that these structures may bear morphological resemble to the human amnion. PGAs recapitulate the developmental progression and transcriptional composition of the human amnion

[0185] To validate the observations from imaging and provide a comprehensive molecular characterization of cell types formed at the single-cell level, the inventors conducted a single-cell RNA- sequencing (scRNA-seq) on pooled samples at specific time points (-2, 0, 1 , 3, and 8 days) using the 10x Chromium platform. All samples were integrated, projected into a common Uniform Manifold Approximation and Projection (UMAP) space, and clustered. 22 transcriptionally distinct clusters across the dataset were identified (Figure 3A).

[0186] Using the expression of curated marker genes identified in the literature, the inventors annotated clusters across the time course as belonging to one of 9 specific cell types (Figure 4A-D and 3A). the inventors observed a population of epiblast-like cells defined by the expression of primed pluripotency markers: NANOG, SOX2 and POU5F1 (Figure 4C-D). the inventors also saw two populations of amniotic ectoderm cells marked by the expression of ISL1 , TFAP2A, TFAP2C, GABRP, and VTCN1 (Figure 4C-D). The distribution of these two populations through time (Figure 4B), made us hypothesise that these two populations could represent different stages of amnion progression (Figure 4B-C, 3C). Early markers such as ISL1 , TFAP2A, and TFAP2C showed high expression at day 1 and were gradually downregulated by day 8, in exchange for a gain in expression of late markers such as GABRP and VTCN1 (Figure 3C). In addition to amniotic ectoderm, the inventors observed four populations expressing extra-embryonic mesoderm associated genes (Figure 4A-D). Extra-embryonic mesoderm clusters formed a progression of cell fates from epiblast to late mesoderm. Early extra-embryonic mesoderm cells were defined by the expression of markers such as MESP1 , MIXL1 , TBXT, MESP2, FOXC1 , CER1 , and GATA6, intermediate mesoderm by the expression of APLNR, GATA4, BMP4, and LEF1 and finally a more mature ‘late’ mesoderm state expressing BST2, POSTN, RSPO2, COL6A3 (Figure 4C-D and 3C) that further divided into two sub-populations (Figure 4A-C). Finally, the analysis unveiled the existence of two yolk sac-like populations closely resembling yolk-sac endoderm cells expressing and yolk-sac endothelial cells (Figure 4A-D). Yolk-sac endoderm is defined by the expression of SOX17, HNF4A and APOA1 , while yolk-sac endothelial cells express canonical markers CD34 and PLVAP (Figure 4A-C).

[0187] To independently confirm marker-based annotation of cell types, the inventors utilised semap, an unsupervised label transfer method

[0024] , to compare the profile cell types identified in this study with published reference datasets from both in vitro model systems and in vivo embryonic data [14, 18, 22, 25-27]. When comparing with the recently published integrated post-implantation stem cell-derived human embryo model (Oldak, B., et al., Complete human day 14 post-implantation embryo models from naive ES cells. Nature, 2023: p. 1-3.), the inventors observed a high degree of concordance between our marker-based fate assignment and predicted cell fates from semap comparison (Figure 4E). For example, the inventors found that amniotic ectoderm, extra-embryonic mesoderm, and epiblast populations matched closely to equivalent populations in this dataset. In addition, when comparing scRNA-seq from direct differentiation to extra-embryonic mesoderm as well as a microfluidic model of amnion induction, the inventors observed close agreement between cell type annotations [14, 18, 22, 25] (Figure 3D-E).

[0188] As well as performing comparisons with in-vitro models the inventors also compared the cell types formed to stage appropriate in-vivo datasets from both non-human primate embryos and gastrulating human embryos at Carnegie Stage 7 (CS7) [26, 27]. the inventors observed a striking transcriptional similarity in the annotated cell types within our sc-RNAseq data when compared with data from both human and cynomolgus monkey embryos (Figure 4F-G). These included annotations to epiblast, amniotic ectoderm, endoderm (yolk sac), nascent, advanced, and extra-embryonic mesoderm (yolk sac mesoderm). The yolk-sac endothelial cell cluster did not annotate to yolk sac-specific clusters from both the human embryo and embryo model. Given the distinctly specific expression profile exhibited by this cluster for yolk sac endothelium, the inventors postulated that it might represent a cell type associated with later stages of yolk sac development. Indeed, our yolk sac endothelial cell type aligned well with the transcriptional profile of the recently published yolk sac cell atlas

[0025] (Figure 3F).

[0189] Taken together, this data validates that PGAs recapitulate the developmental progression and transcriptional composition of cell types of human extra-embryonic tissues.

[0190] Inferred trajectories from epiblast to extra-embryonic lineages in PGAs

[0191] The progression of cell states and temporal dynamics of cell populations suggest that PGAs may capture the state changes and lineage specification events of early extra-embryonic mesoderm and amniotic ectoderm induction (Figure 3B-C). the inventors combined two different trajectory inference methods

[0028] to explore potential state transitions at the single-cell level. Lineage inference using slingshot

[0029] , revealed three distinct trajectories traversing our scRNA-seq dataset, originating in the epiblast population, and terminating in late extra-embryonic mesoderm, amniotic ectoderm, and yolk-sac endoderm respectively (Figure 4H) - these trajectories were independently supported by RNA-velocity vectors indicating changes in gene-expression state (Figure 3G-H). This suggests that PGAs may capture the early stages of extra-embryonic mesoderm and amniotic ectoderm specification, indicating a progression of cell fates and dynamic changes in gene expression. Indeed, exploring gene expression trends across pseudotime for the amniotic ectoderm, extra embryonic mesoderm and yolk-sac lineages, the inventors see clusters of gene sets with distinct dynamics associated with different stages of lineage induction and fate specification (Figure 4I).

[0192] This data suggests that the epiblast can give rise to amniotic ectoderm, extra embryonic mesoderm and yolk-sac lineages and uncovers the transcriptional trajectories that may be underlying these fate transitions.

[0193] PGAs recapitulate functional aspects of the human amnion

[0194] The amnion is known to produce numerous proteins and metabolites required for supporting fetal growth and health during pregnancy. Since PGAs closely resemble the morphology and transcriptional profile of the human amniotic sac, we set out to examine whether it can also recapitulate the functional characteristics of the amniotic sac.

[0195] Proteomics analysis was performed on the fluid extracted from the inside (intra-luminal) and the outside (extra-luminal) of PGAs, grown for 21 days (Figure 14A). There was a clear separation between protein content from the intra- and extra-luminal samples (Figure 14B and 15A). A total of 2915 protein were identified, of which 2370 were unique to the inside amnion-like fluid, 524 were shared between both the intra- and extra-luminal samples, while 21 were unique to the extra-luminal fluid (Figure 14C). Gene ontology (GO) enrichment analysis showed that pathways associated with secretion and metabolic processes were enriched within the intra-luminal samples, while the extra-luminal content was enriched for pathways associated with extra-cellular matrix (ECM) organization (Figure 14D). Notably, we observe that the intra-luminal PGA fluid is enriched in human amniotic fluid factors known to be involved in fetal health and growth such as heat shock proteins, HGF, G6PD, TGFB2 and APOE Mamede, A.C. etal. Cell and tissue research 349, 447-458 (2012) and Jee, B.,et al. Frontiers in Cell and Developmental Biology 9, 648463 (2021) (Figure 14E).

[0196] A key function of the amnion is to provide mechanical support to the fetus. To achieve this function, the amnion secretes proteins that form a thick layer of basement membrane and collagenous matrix. This provides structural integrity to amniotic sac while allowing the transport of amniotic macromolecules Mamede, A.C. et al. Cell and tissue research 349, 447-458 (2012) and Alitalo, K., Kurkinen, M., Vaheri, A., Krieg, T. & Timpl, R. Cell 19, 1053-1062 (1980). In line with this, and in agreement with the data on PGA morphology (Figures 14 and 15), protein groups associated with ECM synthesis, such as primary collagens, fibronectin, periostin, fibrillins, fibulin-5, lumican, osteopontin and basement membrane assembly proteins such as laminins, nidogens, and heparan sulfate proteoglycan 2 content [Alitalo, K., Kurkinen, M., Vaheri, A., Krieg, T. & Timpl, R. Cell 19, 1053-1062 (1980) and Mouw, J.K., Ou, G. & Weaver, V.M. Nature reviews Molecular cell biology 15, 771-785 (2014). were present both within both intra- and extra-luminal PGA (Figure 14E). The abundance of fibrous proteins (collagens) and glycoproteins (fibronectin, laminins and heparan sulfate proteoglycan) were higher within extra and intraluminal fluids, respectively, suggesting possible specific roles in amniotic membrane assembly (Figure 14E).

[0197] Next, we compared the proteins identified with two available proteomic datasets from human and rhesus macaque amniotic fluid Liu, X., et al . Journal of proteomics 192, 1-9 (2019) and Shorey-Kendrick, L.E. et al. Scientific Reports 13, 17039 (2023) We observed a large degree of overlap between proteins identified within the intra-luminal PGA samples and the published amniotic fluid data (Figure 14F and 15B). The differences observed between intra-luminal fluid and human amniotic fluid are expected and can be multi-factorial. On one hand, the proteome from the amniotic fluid is not static, but is known to change across gestational age in humans and primates Bhatti, G. et al. Scientific reports 12, 601 (2022). . On the other hand, the available amniotic fluid data has been collected at late fetal development, during the second or third trimester of pregnancy, much later developmental stage than PGA can model. More importantly, the composition of the amniotic fluid proteome is likely to be further contributed to by proteins released from the fetus itself or as a consequence of maternal fetal interactions. However, the high number of shared proteins suggests that PGAs can be useful models to study amniotic fluid content in isolation from embryonic and maternal content.

[0198] In addition to the proteomics analysis, LC-MS (metabolomics) measurements were performed on polar extracts of both intra- and extra-luminal PGA samples grown for 21 days. Basal E6 media also incubated for 21 days was used as control (Figure 14G). PCA analysis showed a clear separation between the two sets of samples and high reproducibility between replicates (Figure 14H). Annotated metabolites with differential abundance between the intra and the extra-luminal samples included 3’-adenosine monophosphate, Hex-2-ulose (myo-inositol, fructose or glucose), creatine, kynurenic acid, hexitol, L- aspartic acid, folic acid and pyruvic acid (Figure 141-J and 15C). Notably, metabolites highly abundant within the intra-luminal content such as Hex-2-ulose (glucose and myo-inositol), kynurenic acid, creatine, hexitol and derivatives of adenosine monophosphate have all been reported to be present in the human amniotic fluid Kolvatzis, C. et al. Metabolites 13, 1147 (2023) and Menon, R. et al. Amniotic fluid metabolomic analysis in spontaneous preterm birth. Reproductive Sciences 21 , 791-803 (2014). In contrast, folic acid was reduced in the intra-luminal fluid compared with both the extra-luminal samples and control media, suggesting folic acid metabolic conversion. In line with that, we find a high abundance of proteins associated with folate biosynthesis and downstream metabolic processes, such as Dihydrofolate Reductase (DHFR), Methylenetetrahydrofolate Reductase (MTHFR), C-1- Tetrahydrofolate Synthase (MTHFD1), and Methionine Synthase (MTR) in the intra luminal samples (Figure 15D). Folate is widely known to be involved in fetal health and development during early pregnancy, but only when converted into its active form, methyltetrahydrofolate, by the liver Pannia, E., et al Nutrition Reviews 80, 2178-2197 (2022). Dysregulated folate metabolic pathways have been implicated in neural tube defects Blom, H.J., et al., Nature Reviews Neuroscience 7, 724-731 (2006) and Mills, J.L. et al. The Lancet 345, 149-151 (1995). This data suggests that the amniotic sac might be involved in the metabolic processing of folate during early pregnancy before the placenta is formed.

[0199] Taken together, proteomics and metabolomics characterization of PGAs indicates that in addition to structural and transcriptional similarities, PGAs produce secreted proteins, metabolites and structural proteins that resemble a functional human amniotic sac.

[0200] GAT A3 is necessary and sufficient for amnioqenesis Given that PGAs faithfully captured the morphological and transcriptional development of human amnion formation, the inventors next explored potential molecular mechanisms involved in amniogenesis. To identify drivers of fate specification the inventors explored the early gene expression changes and transcription factor activities that could be driving the extra-embryonic mesoderm and amniotic ectoderm trajectories. Using trajectory-based differential gene-expression analysis (tradeSeq)

[0030] , the inventors explored the earliest difference in gene expression dynamics (earlyDETest) that coincide with the bifurcation of extra-embryonic mesoderm and amnion (Figure 5A). By clustering significant genes by pseudotemporal dynamics, the inventors identified two dynamic clusters of putative extra-embryonic mesoderm drivers and one amniotic ectoderm driver cluster (Figure 5B and 6A). Amongst predicted drivers of extra-embryonic mesoderm were signalling (RSPO3, BMP5, PLXNA2 and PDGFRa), structural (COL5A1 , CDH11 , COLC12 and FREM1) and transcription (GATA6, HAND1 and LEF1) factors (Figure 5B and 6A). Putative amniotic ectoderm drivers also included signalling and structural proteins (EPCAM, CLDN6, FGF13, CDH1 and TAGLN) and transcription factors (GRHL2, GATA3, GATA2, TFAP2A, TFAP2C, ISL1 , MSX2) (Figure 5B and 6A).

[0201] Since many drivers of developmental fate decisions include transcription factors, the inventors additionally used single-cell regulatory network inference and clustering method (SCENIC)

[0031] , to predict active transcription factors during the bifurcation of these extra-embryonic mesoderm and amniotic ectoderm (Figure 5C and 6B). By scoring regulon activity per cluster, the inventors identified the top 15 contributing transcription factors for the two different cell states (Figure 5C and 6B). the inventors found three regulon clusters associated with early, intermediate, and late stages of extra- embryonic mesoderm induction (Figure 5C). Early extra-embryonic mesoderm transcription factors included HOXB7, GATA6 and RXRG, which were followed by a way of intermediate transcription factors including LEF1 , FOXF1 and FOXC1 (Figure 5C). Finally late extra-embryonic mesoderm associated transcription factors included IRF1 , HOXC9 and HOXA11 (Figure 5C). the inventors observed two distinct sets of predicted amniotic ectoderm transcription factors, these included GATA3, GATA2, PPARG, CDX4, DLX2, DLX5 and TFAP2C (Figure 5C and 6B).

[0202] To test the TradeSeq and SCENIC predictions and validate potential early drivers of amniogenesis, the inventors transfected H2B-mCherry hES cells with 19 different siRNAs 24 hours prior to generating PGAs and assessed the morphological differences in amniotic sac structure on day 12 (Figure 5D). the inventors observed a high level of reproducibility within each siRNA group (Figure 6C).

[0203] Two opposing morphological behaviours were evident with gene knockdowns. Downregulations that led to impairment of amniogenesis and / or of PGA growth. This was prominently observed with GAT A3, and to a lesser extent, MSX2, POU5F1 or other GATA family members, strongly suggesting a role for these genes in the induction of amniogenesis (Figure 5E-G and 6C).

[0204] Conversely, downregulations leading to accelerated amniogenesis and / or the formation of larger, more mature PGAs, such as with pluripotency factors SOX2 and NANOG (Figure 5E-G and 6C). Interestingly, known key markers of amniotic ectoderm such as ISL1 and TFAP2C were dispensable for early amniogenesis, while perturbation of TFAP2A led to the formation of multiple sub-structures as opposed to a singular amniotic sac-like structure (Figure 5E-G and 6C).

[0205] To determine whether the genes identified as necessary for proper amniogenesis are also sufficient in driving it, the inventors made use of an inducible activated CRISPR / Cas9 (ndCas9-VPR) knock in cell line to up-regulate the expression of GATA2, GAT A3, GATA6, and MSX2 in hES cells. After 48 hours of induction, the inventors aggregated the cells in the absence BMP4 or Wnt agonist and cultured the aggregates for 16 days in E6 media (Figure 5H). Strikingly, the inventors saw that in the absence of any external signals activation of GATA3 alone resulted in the formation of a fluid-filled amniotic sac-like structure, with a high degree of reproducibility (Figure 5I-K). Immunofluorescence staining further revealed loss Nanog and gain of Brachyury after day 1 of aggregation in GAT A3 induced cells (Figure 6D) and confirmed the presence the two (TFAP2C, GATA6) distinct amniotic sac layers at day 16 (Figure 5K).

[0206] Overall, these findings strongly suggest that GATA3 is both necessary and sufficient for proper amniogenesis and confirm the suitability of our model for genetic studies.

[0207] PGAs recapitulates multiple epiblast derived extra-embryonic tissues

[0208] The yolk sac is a multifunctional extra-embryonic tissue that develops simultaneously and adjacent to the amnion. Due to its functional complexity, ranging from nutritional support to hematopoiesis, the yolk sac is composed of multiple cell types, including extra-embryonic mesoderm, endothelial and endoderm cells

[0025] . Integration of our PGAs scRNAseq data with the human embryo and the human yolk sac atlas datasets suggested the presence of multiple yolk sac cell types within our model (Figure 4A, and Figure 7A)

[0025] .

[0209] Two clusters shared SOX17 as a common marker of primary yolk sac but diverged into two distinct cell types suggestive of yolk sac endoderm, with marked expression of APA1 , APOA2, HNF1 B, HNF4A, TTR and FOXA2, and yolk sac endothelial cells expressing a unique signature of genes including CD34, PLVAP, ESAM, KIT, KDR, PECAM1 , ACTA2, CDH5 (Figures 7B). Indeed, immunofluorescence staining confirmed the presence of a cell population expressing SOX17+ and CD34+ that formed structures closely resembling the yolk sac (Figures 7C and 8A). In line with the transcriptional analysis, the inventors also observed a temporal expression hierarchy with SOX17 preceding CD34, indicating developmental progression towards a more definitive yolk sac-like phenotype (Figure 8A).

[0210] The yolk sac is also composed of mesodermal lineages originating from the extra-embryonic mesoderm that are thought to primarily form the yolk sac lining. Clusters of late extra-embryonic cells expressed specific markers for yolk sac mesothelium (including PDPN and UPK3B) and yolk sac fibroblast (such as PDGFRA and VCAN)

[0025] , as well as, common markers with other mesodermal clusters (including POSTN and GATA6) (Figures 7D). These suggest that the late extra-embryonic mesoderm lineages may develop to give rise to yolk sac mesoderm lineages.

[0211] Recently, induction of GATA6 expression in stem cells was shown to induce a yolk sac identity marked by PDGFRA expression.

[0032] . In line with this, immunofluorescence, revealed a PDGFRA+ cell population that exclusively integrated within the yolk sac compartment that also shared common mesodermal marker POSTN but at lower level compared to surrounding mesoderm cells (Figures 7E and 8B). On the other hand, the extra-embryonic mesoderm layer covering the yolk sac and amnion-like structures expressed common markers such as GATA6 along with KRT8. This lining cell population further extended to form a structure closely resembling the connecting stalk (Figures 8C).

[0212] Overall, this data suggests that PGAs are a highly reproducible model of epiblast derived extra- embryonic tissues. At day 8, PGAs consist of an amnion-like sac compartment located adjacent to a complex yolk sac like structure, both covered by extra-embryonic mesoderm that also forms the connecting stalk (Figures 7F). A connective stalk was observed in approximately 89% of the structures (Figures 8D). This closely recapitulates aspects of day 20 of extra-embryonic development (Figures 7F), making this a very reproducible post-gastrulation model (Figure 8E-F).

[0213] In pregnancy, the human yolk sac grows in size during 5-to-10-week post fertilisation, following a rapid reduction in size to a negligible level [33, 34]. The loss of the yolk sac has been attributed to atrophy, but the underlying mechanism remains unknown. Interestingly, the inventors see that the two-layered amniotic sac continues to rapidly expand and accumulate amniotic fluid, while the yolk sac-like structure ceases growth and shrinks in size, resembling an in-utero scenario (Figures 7G, and 8E).

[0214] Finally, the inventors see that cell cycle dramatically change in PGAs with developmental stage in a fate specific manner. Within the amnion, the cell cycle showed a similar pattern of gene expression at early and late stage of PGA development (Figure 8G-I). However, the inventors see that early / intermediate and mature extra embryonic mesoderm show distinct cell cycle profiles. While early / intermediate mesoderm is enriched for G2 / M phase genes such as CDK1 , CCNB1 , PLK1 , late extraembryonic mesoderm including yolk sac mesoderm shows elevated CDKN1 C (p21) expression coupled with low expression of G2 / M proliferation genes, namely MKI67, ASPM, TOP2A, CDK1 , CCNB1 , PLK1

[0035] (Figure 8G-I). Live imaging confirmed that after day 7, the amniotic like cavity expanded faster compared to the rest of the other lineages.

[0215] Taken together, our data indicates that cell cycle remodelling during yolk sac maturation leads to quiescence, possibly contributing to the regression of the definitive yolk sac in later stages of pregnancy. Signals from extra-embryonic lineages induce hES cell differentiation and self-organisation Extra-embryonic tissues have been postulated to serve as signalling centres, orchestrating patterning and morphogenesis during early human development [2, 36]. To understand this exe:epiblast interaction further, the inventors co-cultured wt H1 hES cells pre-treated with BMP and CHIR (in order to make PGAs, as shown in Figure 1 B) with untreated H2B-mCherry H1 hES cells, the inventors combined a total of 400 cells consisting of extra-embryonic progenitors (PGAs at day 0, see Figure 1 B) with pluripotent cells at ratio of 1 :1 in E6 media (Figure 9A). Aggregates of the mixed cell population formed within the first 8 hours. Notably, however, by 16 hours, the mCherry-positive and negative populations began to segregate from each other. Within 24 hours, they selforganized into two distinct cell populations expressing SOX2 and OCT or GATA6 respectively (Figures 9B, and 10A). By 72 hours, these aggregates assumed elongated 3D structures featuring two well-defined domains. Cells within the mCherry-negative domain were confined to the extra- embryonic mesoderm lineage, indicated by a high level of GATA6 (Figures 9C). Strikingly, mCherry -positive cells formed structures resembling amnion cavities (Figures 9C). the inventors saw that at the interface with the extra-embryonic mesoderm, epiblast cells started to form amniotic-like cavity structures, co-expressing early amnion cell markers GATA3, TFAP2C, and OCT4 (Figures 9C). In addition to this, a small pocket of TBXT-positive cells was present adjacent to the amniotic structure, reminiscent of a gastrulation-like event (Figures 9C, and 10A). Subsequently, the inventors see that the amniotic-like cavity rapidly expanded, forming a two-layered fluid filled amniotic sac structure (Figures 9C). Co-immunostaining revealed the distinct origins of the two amniotic layers: the inner layer of amniotic ectoderm originated from the mCherry-positive pluripotent population, while the outer extra-embryonic mesoderm was formed by the wt, mCherry -negative cells (Figure 9C).

[0216] This data suggests that the signals from the extra embryonic lineages are important for amniogenesis and reveal the potential for auto-regulation in the development of the amnion.

[0217] To gain insights into the cellular composition and to characterize transcriptional profile of the PGA:hES cell co-culture, the inventors conducted scRNA-seq on pooled samples at various time points during PGA:hES cell development (-2, 0, 1 , 3, and 8 days), employing two biological replicates per sample. Annotation based on the presence of lineage-specific markers revealed clusters similar to those observed in PGAs, exhibiting matching temporal resolution across the time points (Figures 9D-E, and 10B-D). However, in contrast to PGAs alone, PGA:hES cells from day 1 formed distinctly separate populations. These were characterized by the expression of SOX2 and NANOG, indicating epiblast-like cells, and a cluster of extra-embryonic-like cells expressing GATA6 and GATA4 (Figure 9E). This observation closely aligns with our previous observations with imaging approaches (Figures 9C and 9E). Furthermore, the inventors identified a smaller cluster exhibiting a signature similar to amnion-like cells, marked by the expression of TFAP2C and ISL1 and two clusters associated with the emergence of yolk-sac-like cells appeared later in time (Figures 9D- E and 10B-D).

[0218] To investigate whether paracrine signalling from either lineage underlies the self-organization and lineage specification of hES epiblast-like cells, in the absence of exogenous signals, the inventors assessed the expression of key signalling factors (BMP4, WNT, and NODAL) known to regulate embryogenesis. Notably, cells belonging to both the extra-embryonic mesoderm like and amnion like clusters exhibited high expression levels of BMP4 and WNT5A genes. In contrast, the pluripotent population was enriched for the NODAL and exhibited low levels of BMP4 and WNT5A expression (Figures 9E).

[0219] Together, our results suggest potential paracrine signalling from extra-embryonic mesoderm and amnion-like cells, influencing the epiblast-like population to drive the specification of amnion, mesoderm, and yolk sac-like lineages, and orchestrating these cells spatiotemporal organization (Figures 9F).

[0220] Recently, the embryo-specific culture medium "ex utero culture medium 2" (EUCM2) has been used to support and enhance mouse and human stem cell-based embryo models (REF). Alongside the E6 culture, the inventors also cultured the PGAs:hES cells structures in a modified EUCM2 containing 30% FBS from day 5. Following three days of incubation in modified EUCM2, the inventors observed a reduced rate of amniotic cavitation and an increased variation in the morphology of the structures formed, while maintaining similar cellular configuration, as shown by scRNA-seq

[0221] (Figure S5E-G). Using scRNA-seq the inventors identified a small cell population of primordial germ celllike cells (PGC) co-expressing markers such as BLIMP1 , SOX17, TFAP2C, POU5F1 , and NANOS. Indeed, immunostaining confirmed that PGC-like cells co-expressing BLIMP1 and SOX17 were localized to a section of the amnion-like structure at day 8 (Figures 10E).

[0222] Altogether this data suggests that extra-embryonic (PGA) progenitors can provide the signalling niche necessary for self-organisation and differentiation. And poses the interesting hypothesis that PGC-like cells might arise in part from signals from the amnion. iPSCs can also be used to generate in vitro three-dimensional amnion model

[0223] Two different iPSC cell lines (KOLF2 and MRC5) were cultured according to the methods described above in order to determine whether this precursor cell type could also be used to generate in vitro amnion models. As shown in Figure 16, amnion models were successfully generated using both iPSC lines. The three-dimensional and two-layered structure could be clearly resolved in IPSC-derived amnion models.

[0224] Drug permeation assay using PGAs PGAs were assessed for drug permeability testing (Figure 17). PGAs were grown for 21 days and incubated with compounds (Propranolol or Chlorothiazide at a final concentration of 10uM). These two compounds were described to have with high (100% for Propranolol) or low (20% for Chlorothiazide) level of intestinal absorption in in vivo studies Kus, M, Ibragimow, I and Piotrowska-Kempisty, H, Pharmaceutics, 2003. Intra and extra-luminal fluids was extracted at various time points (60, 90, 120, 150 and 180 minutes) and analysed using LC-MS. There was a close correlation between the drug permeability obtained from PGAs experimental data and human intestinal absorption.

[0225] Discussion

[0226] The inventors have used primed pluripotent stem cells to establish a 3D model that faithfully recapitulates various developmental stages of extra-embryonic tissues. While recent studies have developed embryo models that reconstructed the peri-gastrulation developmental stage around 13-14 days post-fertilization and mimicked aspects of extra-embryonic tissue development with varying accuracy [20-23], our model extends these to form extra-embryonic like tissues that resemble later stages in human development. By day 8, it closely mirrors the formation of extra- embryonic tissues observed at day 20 of embryo development, comprising an amnion-like sac, a yolk sac-like structure surrounded by extra-embryonic mesoderm that extends to form the connecting stalk with a high level of reproducibility

[0037] . PGAs progresses through development as the amniotic sac like compartment continues its expansion and accumulate amniotic fluid, while the yolk sac like compartment gradually undergoes regression. In contrast to other embryo models and other 3D amnion models where the amnion compartment is formed of a ring of cells representing the cavitation stage [18, 38], PGAs develop into advanced amniotic sac-like structures consisting of a fluid-filled transparent membrane composed of two layers: inner amniotic ectoderm and outer extra-embryonic mesoderm, attaining a diameter exceeding 1 .5 centimetres. Similarly, the yolk sac-like compartment shows a good level of complexity, comprising diverse cell types including endothelium, endoderm, and mesothelium cells, surrounded by layers of extra-embryonic mesoderm cells.

[0227] The embryonic origin of the yolk sac is thought to be the hypoblast, that segregates from the epiblast prior to embryo implantation [4], while the origin of extra-embryonic mesoderm remains unclear and attributed to several sources [14, 39, 40]. Human naive epiblast cells possess a potential to generate these lineages in vitro. [14, 41]. Hence, current human embryo models have been developed based on naive, semi-naive pluripotent stem cells, or expanded potential stem cells (EPSCs), relying on the expanded potential of these cells to serve as the source of extra- embryonic tissues, implying that the capture of multi-lineage extra-embryonic cells requires an extended capacity absent in primed pluripotent cells [20-23]. Here, the inventors demonstrate that primed pluripotent stem cells, when subjected to the right signalling environment, undergo differentiation into extra-embryonic mesoderm, yolk sac, and amnion lineages and self-organize into structures morphologically similar to in vivo extra-embryonic tissues.

[0228] The transcriptional profile of the identified lineages closely aligns with reference datasets sourced from human and primate embryos, as well as directed differentiation and integrated in vitro embryo models derived from human naive pluripotent cells, validating the identity and complex nature of cell types within our model. Integrated stem cell-derived embryo models combine naive ES-derived lineages to form various compartments of an embryo and rely on interactions and signalling from extra-embryonic compartments such as trophoblast or hypoblast to promote self-organization of naive ES cells into gastrulating cells that form the amnion[21 , 22]. Furthermore, both the hypoblast and the amnion have been suggested as signalling centres governing post-implantation embryogenesis [2, 36]. When the inventors combined PGAs with primed ES cells, the inventors observed a remarkable capacity for the hES cells to self-organize despite the absence of external signalling cues.

[0229] Surrounded by extra-embryonic like mesoderm layers, an amniotic like ectoderm emerged that progressively developed into a definitive amniotic sac-like structure juxtaposed to a yolk sac-like compartment. These structures originated from primed pluripotent cells population, as did a fraction of the extra-embryonic mesoderm. Thus, our data suggests that extra-embryonic mesoderm-like cells can also serve as signalling hubs for the formation of several extra- embryonic lineages from primed pluripotent cells. BMP and WNT signalling have been implicated as key drivers of embryogenesis and used as inducers of gastrulation in 2D and 3D cultures of hES cells [18, 42-45]. the inventors see high expression of BMP4 and WNT5A in the early extra-embryonic- like population. This observation leads us to postulate that this population may serve as a source for the induction of down-stream signalling.

[0230] In summary, here the inventors engineered a model of extra-embryonic amniotic tissues, that allows for investigations of the cellular complexity, developmental trajectories, and morphological attributes of these supportive tissues, beyond gastrulation. Due to its high reproducibility, efficiency and methodological simplicity, our model stands as a promising, scalable platform for conducting high-throughput functional assays in biomedical studies. A proof-of-concept genetic manipulation assay validated such potential and led to the identification of GATA3 as an early gene, indispensable and sufficient for the proper formation of the amniotic sac. Given that our model yields fully formed, fluid-filled amniotic sac structure, it presents a unique possibility for investigations into the structural integrity and biomechanical properties of the amnion. The accurate cellular constitution of the amniotic sac-like membrane implies a faithful functional capacity concerning fluid exchange and the supply of a barrier to safeguard the embryo. Thus, given their high reproducibility, PGAs can be seen as a useful tool for drug permeation and teratogenicity screening in biomedical research.

[0231] References 1 . Dobreva, M.P., et al., On the origin of amniotic stem cells: of mice and men. International Journal of Developmental Biology, 2010. 54(5): p. 761-777.

[0232] 2. Yang, R., et al., Amnion signals are essential for mesoderm formation in primates. Nature communications, 2021. 12(1): p. 5126.

[0233] 3. Chuva de Sousa Lopes, S.M., et al., The development of the amnion in mice and other amniotes. Philosophical Transactions of the Royal Society B, 2022. 377(1865): p.

[0234] 20210258.

[0235] 4. Ross, C. and T.E. Boroviak, Origin and function of the yolk sac in primate embryogenesis. Nature communications, 2020. 11 (1): p. 3760.

[0236] 5. Cindrova-Davies, T., et al., RNA-seq reveals conservation of function among the yolk sacs of human, mouse, and chicken. Proceedings of the National Academy of Sciences, 2017.

[0237] 114(24): p. E4753-E4761.

[0238] 6. Hyun, L, A. Wilkerson, and J. Johnston, Embryology policy: Revisit the 14-day rule. Nature, 2016. 533(7602): p. 169-171.

[0239] 7. Cavaliere, G., A 14-day limit for bioethics: the debate over human embryo research. BMC medical ethics, 2017. 18: p. 1-12.

[0240] 8. Lovell-Badge, R., et al., ISSCR guidelines for stem cell research and clinical translation: the 2021 update. Stem cell reports, 2021. 16(6): p. 1398-1408.

[0241] 9. Rugg-Gunn, P.J., N. Moris, and P.P. Tam, Technical challenges of studying early human development. Development, 2023. 150(11): p. dev201797.

[0242] 10. Shahbazi, M.N. and M. Zernicka-Goetz, Deconstructing and reconstructing the mouse and human early embryo. Nature cell biology, 2018. 20(8): p. 878-887.

[0243] 11. Perry, J., The mammalian fetal membranes. Reproduction, 1981. 62(2): p. 321-335.

[0244] 12. Mackinlay, K.M., et al., An in vitro stem cell model of human epiblast and yolk sac interaction. Elife, 2021. 10: p. e63930.

[0245] 13. Atkins, M.H., et al., Modeling human yolk sac hematopoiesis with pluripotent stem cells. Journal of Experimental Medicine, 2021 . 219(3): p. e20211924.

[0246] 14. Pham, T.X.A., et al., Modeling human extraembryonic mesoderm cells using naive pluripotent stem cells. Cell stem cell, 2022. 29(9): p. 1346-1365. e10.

[0247] 15. Rostovskaya, M., et al., Amniogenesis occurs in two independent waves in primates. Cell Stem Cell, 2022. 29(5): p. 744-759. e6.

[0248] 16. Tamaoki, N., et al., Self-organized yolk sac-like organoids allow for scalable generation of multipotent hematopoietic progenitor cells from induced pluripotent stem cells. Cell Reports Methods, 2023. 3(4).

[0249] 17. Shao, Y., et al., A pluripotent stem cell-based model for post-implantation human amniotic sac development. Nature communications, 2017. 8(1): p. 208.

[0250] 18. Zheng, Y., et al., Controlled modelling of human epiblast and amnion development using stem cells. Nature, 2019. 573(7774): p. 421-425.

[0251] 19. Shao, Y., et al., Self-organized amniogenesis by human pluripotent stem cells in a biomimetic implantation-like niche. Nature materials, 2017. 16(4): p. 419-425. 20. Weatherbee, B.A., et al., Pluripotent stem cell-derived model of the post-implantation human embryo. Nature, 2023: p. 1-10.

[0252] 21. Pedroza, M., et al., Self-patterning of human stem cells into post-implantation lineages. Nature, 2023: p. 1-10.

[0253] 22. Oldak, B., et al., Complete human day 14 post-implantation embryo models from naive ES cells. Nature, 2023: p. 1-3.

[0254] 23. Liu, L., et al., Modeling post-implantation stages of human development into early organogenesis with stem-cell-derived peri-gastruloids. Cell, 2023. 186(18): p. 3776-3792. e16.

[0255] 24. Kiselev, V.Y., A. Yiu, and M. Hemberg, semap: projection of single-cell RNA-seq data across data sets. Nature methods, 2018. 15(5): p. 359-362.

[0256] 25. Goh, L, et al., Yolk sac cell atlas reveals multiorgan functions during human early development. Science, 2023. 381 (6659): p. eadd7564.

[0257] 26. Tyser, R.C., et al., Single-cell transcriptomic characterization of a gastrulating human embryo. Nature, 2021 . 600(7888): p. 285-289.

[0258] 27. Ma, H., et al., In vitro culture of cynomolgus monkey embryos beyond early gastrulation. Science, 2019. 366(6467): p. eaax7890.

[0259] 28. Bergen, V., et al., Generalizing RNA velocity to transient cell states through dynamical modeling. Nature biotechnology, 2020. 38(12): p. 1408-1414.

[0260] 29. Street, K., et al., Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC genomics, 2018. 19: p. 1-16.

[0261] 30. Van den Berge, K., et al., Trajectory-based differential expression analysis for single-cell sequencing data. Nature communications, 2020. 11 (1): p. 1201.

[0262] 31. Aibar, S., et al., SCENIC: single-cell regulatory network inference and clustering. Nature methods, 2017. 14(11): p. 1083-1086.

[0263] 32. Hislop, J., et al., Modeling post-implantation human development to yolk sac blood emergence. Nature, 2023: p. 1-3.

[0264] 33. Tan, S., M.K. Pekta§, and H. Arslan, Sonographic evaluation of the yolk sac. Journal of Ultrasound in Medicine, 2012. 31 (1): p. 87-95.

[0265] 34. Detti, L., et al., Early pregnancy ultrasound measurements and prediction of first trimester pregnancy loss: A logistic model. Scientific reports, 2020. 10(1): p. 1545.

[0266] 35. Guo, L., et al., Screening and identification of genes associated with cell proliferation in cholangiocarcinoma. Aging (Albany NY), 2020. 12(3): p. 2626.

[0267] 36. Okubo, T., et al., Hypoblast from human pluripotent stem cells regulates epiblast development. Nature, 2023: p. 1-3.

[0268] 37. Sadler, T. and M.L. Feldkamp. The embryology of body wall closure: relevance to gastroschisis and other ventral body wall defects, in American Journal of Medical Genetics Part C: Seminars in Medical Genetics. 2008. Wiley Online Library.

[0269] 38. Moris, N., Stem cells used to model a two-week-old human embryo. 2023, Nature Publishing Group UK London. 39. Pera, M.F. and J. Rossant, The exploration of pluripotency space: Charting cell state transitions in peri-implantation development. Cell Stem Cell, 2021. 28(11): p. 1896-1906. 40. Rossant, J. and P.P. Tam, Early human embryonic development: blastocyst formation to gastrulation. Developmental cell, 2022. 57(2): p. 152-165.

[0270] 41. Guo, G., et al., Human naive epiblast cells possess unrestricted lineage potential. Cell stem cell, 2021. 28(6): p. 1040-1056. e6.

[0271] 42. Linneberg-Agerholm, M., et al., Naive human pluripotent stem cells respond to Wnt, Nodal and LIF signalling to produce expandable naive extra-embryonic endoderm. Development, 2019. 146(24): p. dev180620.

[0272] 43. Martyn, I., et al., Self-organization of a human organizer by combined Wnt and Nodal signalling. Nature, 2018. 558(7708): p. 132-135.

[0273] 44. Moris, N., et al., An in vitro model of early anteroposterior organization during human development. Nature, 2020. 582(7812): p. 410-415.

[0274] 45. Simunovic, M., et al., A 3D model of a human epiblast reveals BMP4-driven symmetry breaking. Nature cell biology, 2019. 21 (7): p. 900-910.

[0275] CLAUSES

[0276] The invention may also be understood by reference to the following numbered embodiments.

[0277] 1 . An in vitro three-dimensional amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity.

[0278] 2. The amnion model of clause 1 , wherein the inner cell layer comprises ectodermal cells.

[0279] 3. The amnion model of clause 1 or 2, wherein the inner cell layer comprises an amniotic ectoderm.

[0280] 4. The amnion model of any preceding clause, wherein the inner cell layer expresses ISL1 , TFAP2C, TFAP2A, GATA3, GABRP and / or VTCN1.

[0281] 5. The amnion model of any preceding clause, wherein the inner cell layer expresses GAT A3.

[0282] 6. The amnion model of any preceding clause, wherein the inner cell layer expresses ISL1 , TFAP2C, and / or TFAP2A.

[0283] 7. The amnion model of any preceding clause, wherein the inner cell layer expresses GABRP and / or VTCN1.

[0284] 8. The amnion model of any preceding clause, wherein the inner cell layer is positive for ISL1 , TFAP2C, TFAP2A, GATA3, GABRP and / or VTCN1. The amnion model of any preceding clause, wherein the inner cell layer is positive for GATA3. The amnion model of any preceding clause, wherein the inner cell layer is positive for ISL1 , TFAP2C, and / or TFAP2A. The amnion model of any preceding clause, wherein the inner cell layer is positive for GABRP and / or VTCN1. The amnion model of any preceding clause, wherein the outer cell layer comprises mesodermal cells. The amnion model of any preceding clause, wherein the outer cell layer comprises an extra- embryonic mesoderm. The amnion model of any preceding clause, wherein the outer cell layer expresses GATA6, GATA4, BST2, POSTN, RSPO2, and / or Handl . The amnion model of any preceding clause, wherein the outer cell layer expresses GATA6. The amnion model of any preceding clause, wherein the outer cell layer expresses GATA4. The amnion model of any preceding clause, wherein the outer cell layer expresses BST2,

[0285] POSTN, RSPO2, and / or Handl . The amnion model of any preceding clause, wherein the outer cell layer is positive for GATA6, GATA4, BST2, POSTN, RSPO2 and / or HAND1. The amnion model of any preceding clause, wherein the outer cell layer is positive for GATA6. The amnion model of any preceding clause, wherein the outer cell layer is positive for GATA4. The amnion model of any preceding clause, wherein the outer cell layer is positive for BST2,

[0286] POSTN, RSPO2, and / or HAND1. The amnion model of any preceding clause, wherein the model has been cultured for at least 18 hours. The amnion model of any preceding clause, wherein the inner cell layer expresses TFAP2C and the outer cell layer expresses GATA6. 24. The amnion model of any preceding clause, wherein the inner cell layer comprises ectodermal cells and the outer cell layer comprises mesodermal cells.

[0287] 25. The amnion model of any preceding clause, wherein the inner cell layer comprises an amniotic ectoderm and the outer cell layer comprises an extra-embryonic mesoderm.

[0288] 26. The amnion model of any preceding clause, wherein the amniotic cavity is an amniotic sac-like structure.

[0289] 27. The amnion model of any preceding clause, wherein the amniotic cavity is an amniotic sac.

[0290] 28. The amnion model of any preceding clause, wherein the amniotic cavity comprises fluid.

[0291] 29. The amnion model of any preceding clause, wherein the amniotic cavity is fluid-filled.

[0292] 30. The amnion model of any preceding clause, wherein the amniotic cavity comprises amniotic fluid or amniotic fluid-like fluid.

[0293] 31. The amnion model of any preceding clause, wherein the amnion model is a post-gastrulation model.

[0294] 32. The amnion model of any preceding clause, further comprising a yolk sac.

[0295] 33. The amnion model of any of clauses 1-31 , further comprising a yolk sac-like structure.

[0296] 34. The amnion model of any preceding clause, wherein the amnion model has been in culture for at least 24 hours, at least 2 days, at least 3 days, at least 4 days, or at least 5 days.

[0297] 35. The amnion model of any of clauses 32-34, wherein the amnion model has been in culture for up to 24 days, up to 26 days, up to 28 days, or up to 30 days.

[0298] 36. The amnion model of any of clauses 32-35, wherein the yolk sac or yolk sac-like structure comprises at least one cell type selected from the group consisting of endothelium, endoderm, fibroblasts, and mesothelium.

[0299] 37. The amnion model of any of clauses 32-36, wherein the yolk sac or yolk sac-like structure comprises yolk sac endodermal cells. 38. The amnion model of clause 37, wherein the yolk sac endodermal cells express SOX17, APOA1 , APOA2, HNF1 B, HNF4A, TTR, and / or FOXA2.

[0300] 39. The amnion model of clause 38, wherein the yolk sac endodermal cells express SOX17, HNF4A, and / or APOA1.

[0301] 40. The amnion model of any of clauses 32-39, wherein the yolk sac or yolk sac-like structure comprises yolk-sac endothelial cells.

[0302] 41. The amnion model of clause 40, wherein the yolk sac endothelial cells express SOX17, CD34, PLVAP, ESAM, KIT, KDR, PECAM1 , ACTA2, and / or CDH5.

[0303] 42. The amnion model of clause 41 , wherein the yolk sac endothelial cells express CD34 and / or PLVAP.

[0304] 43. The amnion model of any of clauses 32-42, wherein the yolk sac or yolk sac-like structure comprises yolk sac mesothelial cells.

[0305] 44. The amnion model of clause 43, wherein the yolk sac mesothelial cells express PDPN and / or UPK3B.

[0306] 45. The amnion model of any of clauses 32-44, wherein the yolk sac or yolk sac-like structure comprises yolk sac fibroblast cells.

[0307] 46. The amnion model of clause 45, wherein the yolk sac fibroblast cells express PDGFRA and / or VCAN.

[0308] 47. The amnion model of any of clauses 32-46, wherein the outer cell layer substantially or completely surrounds the yolk sac or yolk sac-like structure.

[0309] 48. The amnion model of any preceding clause, wherein the amnion model further comprises a connecting stalk.

[0310] 49. The amnion model of clause 48, wherein the connecting stalk is formed from the outer cell layer.

[0311] 50. The amnion model of any preceding clause, wherein the amnion model is at least 1 mm, at least 2mm, at least 3mm, at least 4mm, at least 5mm, at least 6mm, at least 7mm, at least 8mm, at least 9mm, at least 1cm, at least 1 .5cm or at least 2cm in diameter. 51. The amnion model of any preceding clause, wherein the amnion model is at least 5 mm in diameter.

[0312] 52. The amnion model of any preceding clause, wherein the amnion model can be cultured for at least 2 weeks, at least 4 weeks, at least 6 weeks, at least 8 weeks, at least 10 weeks, at least 12 weeks, at least 14 weeks, or at least 16 weeks.

[0313] 53. The amnion model of any preceding clause, wherein the amnion model is a mammalian amnion model.

[0314] 54. An in vitro amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity, and wherein the amnion model further comprises a yolk sac, wherein the yolk sac is substantially or completely surrounded by the outer cell layer.

[0315] 55. An in vitro amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity, and wherein the amnion model further comprises a yolk sac and a connecting stalk, wherein the yolk sac is substantially or completely surrounded by the outer cell layer.

[0316] 56. A post-gastrulation amnioid (PGA) comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity.

[0317] 57. An amnion organoid comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity.

[0318] 58. A method for producing an in vitro three-dimensional amnion model, the method comprising: i) providing a plurality of cells, j) treating the plurality of cells with a BMP agonist, k) treating the plurality of cells with a WNT agonist, and l) culturing the plurality of cells.

[0319] 59. The method of clause 58, wherein the plurality of cells comprises stem cells.

[0320] 60. The method of clause 59, wherein the stem cells comprise iPSCs, embryonic stem cells, or a combination thereof.

[0321] 61. The method of clause 59 or 60, wherein the stem cells are human stem cells.

[0322] 62. The method of any of clauses 58-61 , wherein steps (b) and (c) are performed sequentially. 63. The method of any one of clauses 58-61 , wherein steps (b) and (c) are performed simultaneously.

[0323] 64. The method of any of clauses 58-63, wherein steps (b) and (c) are each independently performed for 12-36 hours.

[0324] 65. The method of any of clauses 58-64, wherein steps (b) and (c) are each independently performed for 20-30 hours.

[0325] 66. The method of any of clauses 58-65, wherein steps (b) and (c) are each independently performed for 24 hours.

[0326] 67. The method of any of clauses 58-66, wherein step (b) comprises treating the plurality of cells with about 1-10 ng / ml BMP agonist.

[0327] 68. The method of any of clauses 58-67, wherein step (b) comprises treating the plurality of cells with about 3-7 ng / ml BMP agonist.

[0328] 69. The method of any of clauses 58-68, wherein step (b) comprises treating the plurality of cells with about 5 ng / ml BMP agonist.

[0329] 70. The method of any of clauses 58-69, wherein the BMP agonist is selected from the group consisting of BMP4, BMP2, and GATA3.

[0330] 71. The method of any of clauses 58-70, wherein the BMP agonist is BMP4.

[0331] 72. The method of any of clauses 58-71 , wherein step (c) comprises treating the plurality of cells with about 1-10 pM WNT agonist.

[0332] 73. The method of any of clauses 58-72, wherein step (c) comprises treating the plurality of cells with about 1-5 pM WNT agonist.

[0333] 74. The method of any of clauses 58-73, wherein step (c) comprises treating the plurality of cells with about 3 pM WNT agonist.

[0334] 75. The method of any of clauses 58-74, wherein the WNT agonist is selected from the group consisting of a GSK3 inhibitor, CHIR99021 , a Wnt ligand, a p-catenin agonist, R-spondin, Norrin, or a combination thereof.

[0335] 76. The method of any of clauses 58-75, wherein the WNT agonist is CHIR99021. 77. The method of any of clauses 58-76, wherein steps (a)-(c) are performed in a two-dimensional monolayer.

[0336] 78. The method of any of clauses 58-77, further comprising a step of dissociating the plurality of cells into single cells following step (c).

[0337] 79. The method of clause 78, wherein the single cells are cultured in non-adherent plates.

[0338] 80. The method of clause 78 or 79, wherein the single cells are cultured in a three-dimensional culture system.

[0339] 81. The method of any of clauses 58-80, wherein the method further comprises treating the plurality of cells or the dissociated single cells with a ROCK inhibitor following step (c).

[0340] 82. The method of any of clauses 58-81 , wherein the plurality of cells or the dissociated single cells are treated with the ROCK inhibitor at about 1-5 pM.

[0341] 83. The method of any of clauses 58-82, wherein the plurality of cells or the dissociated single cells are treated with the ROCK inhibitor at about 2.5 pM.

[0342] 84. The method of any of clauses 81-83, wherein the ROCK inhibitor is selected from the group consisting of fasudil, ripasudil, netarsudil, and Y27632.

[0343] 85. The method of clause 84, wherein the ROCK inhibitor is Y27632.

[0344] 86. The method of any of clauses 58-85, using basal cell culture media.

[0345] 87. The method of any of clauses 58-86, wherein the cell culture media comprises insulin.

[0346] 88. The method of any of clauses 58-87, wherein the method does not comprise the addition of extracellular matrix proteins.

[0347] 89. The method of any of clauses 58-88, wherein the method does not comprise the addition of collagen or laminin.

[0348] 90. The method of any of clauses 58-89, wherein step (d) comprises culturing the plurality of cells or the dissociated single cells for at least 1 week, at least 2 weeks, at least 4 weeks, at least 6 weeks, at least 8 weeks, at least 10 weeks, at least 12 weeks, at least 14 weeks, or at least 16 weeks.

[0349] 91. The method of any of clauses 58-90, wherein the plurality of cells comprises human cells. 92. The method of any of clauses 58-91 , comprising: a. providing a plurality of cells, wherein the plurality of cells comprises iPSCs and / or embryonic stem cells, b. treating the plurality of cells with a BMP agonist at about 3-7 ng / ml, wherein the BMP agonist is BMP4, c. treating the plurality of cells with a WNT agonist at about 1-5 pM, wherein the WNT agonist is CHIR99021 , d. dissociating the plurality of cells into single cells, e. treating the dissociated single cells with a ROCK inhibitor at about 1-5 pM, wherein the ROCK inhibitor is Y27632, and f. culturing the dissociated single cells for at least 24 hours; wherein steps (b) and (c) are performed sequentially, and wherein steps (b), (c) and (e) are each independently performed for 20-30 hours.

[0350] 93. The method of any of clauses 58-92, wherein the method has a reproducibility rate greater than 90%.

[0351] 94. The method of any of clauses 58-93, wherein following step (d), the plurality of cells or the dissociated single cells form an in vitro amnion model at least 90% of the time, wherein the amnion model is a three-dimensional amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity.

[0352] 95. The method of any of clauses 58-95, further comprising isolating an amniotic membrane from the resulting in vitro three-dimensional amnion model.

[0353] 96. An in vitro amnion model obtained or obtainable by the method of any of clauses 58-95.

[0354] 97. Use of the in vitro amnion model according to any of clauses 1-57 or 96 in a method of drug screening.

[0355] 98. Use of the in vitro amnion model according to any of clauses 1-57 or 96 in a method of tissue engineering.

[0356] 99. Use of the in vitro amnion model according to any of clauses 1-57 or 96 in regenerative medicine.

[0357] 100. Use of the in vitro amnion model according to any of clauses 1-57 or 96 in a method of assessing amnion permeability.

[0358] 101 . Use of the amnion model according to any of clauses 1 -57 or 96 in a method of assessing membrane permeability. 102. Use of the amnion model according to clause 1-57 or 96 in a method of assessing drug permeability.

[0359] 103. An amniotic membrane derived or obtainable from the in vitro amnion model of any of clauses 1-57 or 96.

[0360] 104. The amniotic membrane of clause 95 or 103, for use in a method of wound healing.

[0361] 105. The amniotic membrane of clause 95 or 103, for use in a method of treating burns.

[0362] 106. The amniotic membrane for use of clause 105, wherein the method comprises treating ocular surface lesions caused by burns.

[0363] 107. The amniotic membrane of clause 95 or 103, for use in a method of ocular reconstruction.

[0364] 108. The amniotic membrane for use of clause 107, wherein the ocular reconstruction is ocular surface reconstruction.

[0365] 109. The amniotic membrane of clause 95 or 103, for use in regenerative medicine.

[0366] 110. The amniotic membrane of clause 95 or 103, for use in a method of treating neurotrophic corneal ulcers.

[0367] 111. The amniotic membrane of clause 95 or 103, for use in a method of treating neurotrophic corneal pain.

[0368] 112. The amniotic membrane of clause 95 or 103, for use in a method of treating ocular surface disorders.

[0369] 113. Use of the amniotic membrane according to clause 95 or 103 in a method of tissue engineering.

[0370] 114. Use of the amniotic membrane according to clause 95 or 103 in a method of drug screening.

[0371] 115. Use of the amniotic membrane according to clause 95 or 103 in a method of assessing membrane permeability. 116. Use of the amniotic membrane according to clause 95 or 103 in a method of assessing drug permeability.

[0372] 117. Use of the amniotic membrane according to clause 95 or 103 in regenerative medicine.

[0373] 118. The in vitro amnion model of any of clauses 1 -57 or the amniotic membrane of clause 95 or 103, having high throughput capability.

[0374] 119. The in vitro amnion model of any of clauses 1 -57 or the amniotic membrane of clause 95 or 103, further comprising a basement membrane.

[0375] 120. The in vitro amnion model or amniotic membrane of clause 1 19, wherein the basement membrane is located between the inner and outer cell layers.

[0376] 121 . The in vitro amnion model or amniotic membrane of clause 1 19 or 120, wherein the basement membrane comprises proteins associated with extracellular matrix (ECM) synthesis.

[0377] 122. The in vitro amnion model or amniotic membrane of clause 121 , wherein proteins associated with extracellular matrix (ECM) synthesis comprise one or more proteins selected from group consisting of: primary collagens, fibronectin, periostin, fibrillins, fibulin-5, lumican and osteopontin.

[0378] 123. The in vitro amnion model or amniotic membrane according to any one of clauses 119 to 122, wherein the basement membrane comprises basement membrane assembly proteins.

[0379] 124. The in vitro amnion model or amniotic membrane of clause 123, wherein the basement membrane assembly proteins comprise one of more proteins selected from the group consisting of laminins, nidogens and heparan sulfate proteoglycan 2.

[0380] EQUIVALENTS AND SCOPE

[0381] Those skilled in the art will appreciate that the present invention is defined by the appended claims and not by the Examples or other description of certain embodiments included herein.

[0382] Similarly, the singular forms “a”, “an”, and “the” include plural referents unless the context clearly dictates otherwise.

[0383] Unless defined otherwise above, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Any methods and materials similar or equivalent to those described herein can also be used in the practice or testing of the present invention. Generally, nomenclatures used in connection with, and techniques of, cell and tissue culture, molecular biology, immunology, genetics and protein and nucleic acid chemistry described herein are those well-known and commonly used in the art, or according to manufacturer’s specifications.

[0384] All publications mentioned herein are incorporated herein by reference to disclose and describe the methods and / or materials in connection with which the publications are cited.

Claims

CLAIMS1 . An in vitro three-dimensional amnion model comprising an inner cell layer and an outer cell layer, wherein the inner cell layer defines an amniotic cavity.

2. The amnion model of claim 1 , wherein the inner cell layer comprises ectodermal cells.

3. The amnion model of any preceding claim, wherein the inner cell layer expresses ISL1 , TFAP2C, TFAP2A, GATA3, GABRP and / or VTCN1.

4. The amnion model of any preceding claim, wherein the outer cell layer comprises mesodermal cells.

5. The amnion model of any preceding claim, wherein the outer cell layer expresses GATA6, GATA4, BST2, POSTN, RSPO2, and / or Handl .

6. The amnion model of any preceding claim, wherein the inner cell layer expresses TFAP2C and the outer cell layer expresses GATA6.

7. The amnion model of any preceding claim, wherein the inner cell layer comprises ectodermal cells and the outer cell layer comprises mesodermal cells.

8. The amnion model of any preceding claim, wherein the amniotic cavity is an amniotic sac or amniotic sac-like structure.

9. The amnion model of any preceding claim, wherein the amniotic cavity comprises amniotic fluid or amniotic fluid-like fluid.

10. The amnion model of any preceding claim, further comprising a yolk sac or yolk sac-like structure.

11. The amnion model of any preceding claim, wherein the amnion model further comprises a connecting stalk.

12. The amnion model of any preceding claim, wherein the amnion model is at least 1 mm, at least 2mm, at least 3mm, at least 4mm, at least 5mm, at least 6mm, at least 7mm, at least 8mm, at least 9mm, at least 1cm, at least 1 .5cm or at least 2cm in diameter.

13. The amnion model of any preceding claim, wherein the amnion model can be cultured for at least 2 weeks, at least 4 weeks, at least 6 weeks, at least 8 weeks, at least 10 weeks, at least 12 weeks, at least 14 weeks, or at least 16 weeks.

14. A method for producing an in vitro three-dimensional amnion model, the method comprising: a) providing a plurality of cells, b) treating the plurality of cells with a BMP agonist, c) treating the plurality of cells with a WNT agonist, and d) culturing the plurality of cells.

15. The method of claim 14, wherein the plurality of cells comprises stem cells.

16. The method of claim 15, wherein the stem cells comprise iPSCs, embryonic stem cells, or a combination thereof.

17. The method of any of claims 14-16, wherein steps (b) and (c) are performed sequentially.

18. The method of any of claims 14-17, wherein steps (b) and (c) are each independently performed for 12-36 hours.

19. The method of any of claims 14-18, wherein steps (b) and (c) are each independently performed for 24 hours.

20. The method of any of claims 14-19, wherein step (b) comprises treating the plurality of cells with about 1-10 ng / ml BMP agonist.

21. The method of any of claims 14-20, wherein the BMP agonist is selected from the group consisting of BMP4, BMP2, and GATA3.

22. The method of any of claims 21 , wherein the BMP agonist is BMP4.

23. The method of any of claims 14-22, wherein step (c) comprises treating the plurality of cells with about 1-10 pM WNT agonist.

24. The method of any of claims 14-23, wherein the WNT agonist is selected from the group consisting of a GSK3 inhibitor, CHIR99021 , a Wnt ligand, a p-catenin agonist, R-spondin, Norrin, or a combination thereof.

25. The method of any of claims 24, wherein the WNT agonist is CHIR99021 .

26. The method of any of claims 14-25, wherein steps (a)-(c) are performed in a two-dimensional monolayer.

27. The method of any of claims 14-26, further comprising a step of dissociating the plurality of cells into single cells following step (c).

28. The method of claim 28, wherein the single cells are cultured in non-adherent plates.

29. The method of any of claims 14-28, wherein step (d) comprises culturing the plurality of cells or the dissociated single cells for at least 1 week, at least 2 weeks, at least 4 weeks, at least 6 weeks, at least 8 weeks, at least 10 weeks, at least 12 weeks, at least 14 weeks, or at least 16 weeks.

30. The method of any of claims 14-29, comprising: a. providing a plurality of cells, wherein the plurality of cells comprises iPSCs and / or embryonic stem cells, b. treating the plurality of cells with a BMP agonist at about 3-7 ng / ml, wherein the BMP agonist is BMP4, c. treating the plurality of cells with a WNT agonist at about 1-5 pM, wherein the WNT agonist is CHIR99021 , d. dissociating the plurality of cells into single cells, e. treating the dissociated single cells with a ROCK inhibitor at about 1-5 pM, wherein the ROCK inhibitor is Y27632, and f. culturing the dissociated single cells for at least 24 hours; wherein steps (b) and (c) are performed sequentially, and wherein steps (b), (c) and (e) are each independently performed for 20-30 hours.

31. The method of any of claims 14-30, further comprising isolating an amniotic membrane from the resulting in vitro three-dimensional amnion model.

32. An in vitro amnion model obtained or obtainable by the method of any of claims 14-31 .

33. An amniotic membrane derived or obtainable from the in vitro amnion model of any of claims 1-13 or 32.

34. Use of the in vitro amnion model according to any of claims 1-13 or 32 or the amniotic membrane of claim 33 in: a) a method of drug screening; b) a method of tissue engineering; c) regenerative medicine;d) a method of assessing amnion permeability; e) a method of assessing membrane permeability; or f) a method of assessing drug permeability.

35. The in vitro amnion model of any of claims 1-13 or 32, or the amniotic membrane of claim 33, for use in: a) a method of wound healing; b) a method of treating burns; optionally wherein the method comprises treating ocular surface lesions caused by burns; c) a method of ocular reconstruction; optionally ocular surface reconstruction; d) regenerative medicine; e) a method of treating neurotrophic corneal ulcers; f) a method of treating neuropathic corneal pain; or g) a method of treating ocular surface disorders.

36. The in vitro amnion model of any of claims 1-13 or 32, or the amniotic membrane of claim 33, further comprising a basement membrane.

37. The in vitro amnion model or amniotic membrane of claim 36, wherein the basement membrane is located between the inner and outer cell layers.

38. The in vitro amnion model or amniotic membrane of claims 36 or 37, wherein the basement membrane comprises proteins associated with extracellular matrix (ECM) synthesis.

39. The in vitro amnion model or amniotic membrane of claim 38, wherein proteins associated with extracellular matrix (ECM) synthesis comprise one or more proteins selected from group consisting of: primary collagens, fibronectin, periostin, fibrillins, fibulin-5, lumican and osteopontin.

40. The in vitro amnion model or amniotic membrane according to any one of claims 36 to 39, wherein the basement membrane comprises basement membrane assembly proteins.

41. The in vitro amnion model or amniotic membrane of claim 40, wherein the basement membrane assembly proteins comprise one of more proteins selected from the group consisting of laminins, nidogens and heparan sulfate proteoglycan 2.

Images