Atlas Of Human Embryology From Oocytes To Preimplantation Embryos
Arleen Smelko
Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
The Human Developmental Cell Atlas (HDCA) initiative, which is part of the Human Cell Atlas, aims to create a comprehensive reference map of cells during development. This will be critical to understanding normal organogenesis, the effect of mutations, environmental factors and infectious agents on human development, congenital and childhood disorders, and the cellular basis of ageing, cancer and regenerative medicine. Here we outline the HDCA initiative and the challenges of mapping and modelling human development using state-of-the-art technologies to create a reference atlas across gestation. Similar to the Human Genome Project, the HDCA will integrate the output from a growing community of scientists who are mapping human development into a unified atlas. We describe the early milestones that have been achieved and the use of human stem-cell-derived cultures, organoids and animal models to inform the HDCA, especially for prenatal tissues that are hard to acquire. Finally, we provide a roadmap towards a complete atlas of human development.
Human development begins with a fertilized oocyte that divides and differentiates through preimplantation, embryonic and fetal stages (Fig. 1). Early studies began with morphometric and qualitative assessments of human embryos, leading to development of the Carnegie staging system15 (Fig. 1). Advances in imaging, cytometry and genomics technologies have provided further insights into the complex spatiotemporal changes during organogenesis16. Recent progress in single-cell profiling technologies has revolutionized our ability to study human development at an unprecedented resolution17. Leveraging these advances to build a comprehensive atlas of human development (from the fertilized oocyte to birth) at cellular resolution is an ambitious endeavour that is similar in scale to the Human Genome Project, which required multidisciplinary scientific expertise from disparate fields working together collaboratively. Such a community has arisen from the grassroots assembly of global researchers who are working as part of the Human Cell Atlas (HCA)18 initiative. As with the Human Genome Project, the HCA will be a foundational scientific resource, composed of diverse data types and available freely through browsable and searchable web portals that visualize cells across anatomical space and developmental time.
The HDCA is a strategic focus of HCA19, and is pursued by scientists from individual laboratories as well as large national and international research consortia (Supplementary Table 1); the HDCA is open to all who adhere to its mission and open science values20. The HDCA aims for equity, inclusivity and diversity both in terms of scientific participation and the representation of human tissue samples. We encourage any interested researcher to become a member, participate, register their study and contribute their data and publications to the HDCA and HCA21.
The successful construction of a HDCA poses substantial scientific challenges in terms of experimental measurement technologies, computational analysis and visualization algorithms (Fig. 2). In particular, the dynamic nature of gestation creates challenges for designing a sampling strategy, especially to capture transient morphological changes in the first eight weeks. A major endeavour for the HDCA will be to develop the conceptual and computational framework to capture development with respect to cellular and morphological changes. The HDCA, through coordination with the HCA Organoid Network22, will incorporate data from in vitro culture model and organoid systems23 to cautiously infer development between seven days and four weeks after conception (a period in which samples are difficult to obtain) (Fig. 1b, c).
Accessing human developmental samples is constrained by general and geographically specific ethical and legal challenges. These include issues relating to donation, access to and research use of legally defined developing human tissue material, regulatory approvals processes and cultural sensitivities. Research on human embryos and fetuses is supported within European and national regulations, such as the UK National Research Ethics Service (NRES) and the French Agence de Biomdecine. In the UK, studies on preimplantation human embryos (up to 14 days after conception) are governed by the Human Fertilisation and Embryology Authority and a research ethics committee (such as NRES). However, in the USA, research on donated human embryonic and fetal materials has increasingly been restricted over the past two decades, despite the existence of similar regulatory oversight.
The international sharing of genomic sequencing and clinical data derived from prenatal or paediatric tissue samples is subject to data protection regulation that considers live versus deceased status, consent regarding research data use and confidentiality. Data from living donors are shared under appropriate access controls. The HCA Ethics Working Group is developing tools, guidance notes26, consent-form templates and sampling information for embryonic, fetal and paediatric tissue material, and international data-sharing guidance for the HDCA.
Development is intricately orchestrated in three spatial dimensions and gestation time. Human embryogenesis cannot be easily assessed at high resolution in vivo27. Time-lapse studies are limited to in vitro preimplantation embryos. The application of high-throughput genomics technologies to dissociated cells and tissue sections in situ is beginning to provide data of unprecedented resolution (Fig. 3, Table 1).
a, Organ or anatomical unit profiling of a prenatal embryo derived from multiple germ layers. b, Single-cell atlas technologies by relative resolution and genome scale. CODEX, co-detection by indexing; H&E, haematoxylin and eosin staining; IF, immunofluorescence; ISS, in situ sequencing; MERFISH, multiplexed error-robust FISH; MIBI, multiplexed ion beam imaging; seqFISH, sequential FISH; smFISH, single-molecule FISH; STARmap, spatially resolved transcript annotation readout mapping. c, Integration of datasets from different technologies (for example, spatial transcriptomics, scRNA-seq and targeted ISS) to profile organs or whole embryos.
Single-cell molecular profiles based on RNA, chromatin accessibility, methylation or select protein signatures have enabled a more nuanced definition of cell types and states. The data underpinning such definitions are increasingly derived from single-cell RNA sequencing (scRNA-seq), barcoded antibodies and accessible chromatin sequencing of dissociated cells28,29. Resolving cell types and trajectories at high granularity is aided by full-length scRNA-seq, but is primarily performed by profiling large numbers of cells. Cell-type definition is currently guided by existing knowledge from model organisms and adult cellular profiles, which may not faithfully reflect prenatal cell types, transient cell types that are present only during development and transitional states of differentiation.
To overcome these challenges, many time points need to be profiled and defined cell states need to be mapped back into their 3D space over time and functionally characterized. High levels of multiplexing can attain this level of granularity at an affordable cost for a complete human developmental cell atlas30,31. Molecular profiles, morphology, functional assessment and other features can reflect the multifaceted state of a cell. For example, the transcriptome reflects the present and potential future of a cell; protein expression captures the immediate past and present state of a cell; chromatin profiles reveal its invariant type and potential for future differentiation; and ontogeny reveals its history.
The field of developmental biology has traditionally drawn on ontogenic relationships to define cell types, but this is challenging in humans for whom information is captured as snapshots across gestation. CRISPR scarring is applicable only in stem cells, organoid systems and short-term explants32,33. The tracking of somatic mutations is the only available technology to definitively determine ontogeny, but is limited by its current lack of scalability34,35. Recent methods that rely on the simultaneous measurement of mitochondrial DNA and RNA, transcriptome and open chromatin may overcome this challenge36,37. We anticipate the field moving towards a consensus cell ontology that integrates multimodal single-cell profiling data as well as legacy knowledge of embryonic cell-type definitions augmented by information from diverse animal models.
Spatial genomics methods to measure RNA in tissue sections typically offer a trade-off: high-resolution (single-cell and subcellular) methods that typically measure hundreds of transcripts or whole transcriptome profiles at a multicellular level38,39. This trade-off can be mitigated by integration with single-cell profiles from dissociated cells, expanding the genomic coverage by predicting the spatial expression of unmeasured genes or enhancing resolution by deconvolution of multicellular measurements. Tissue clearing methods to render organs transparent40, combined with whole-mount protein immunostaining and RNA single-molecule fluorescence in situ hybridization (FISH)41,42, can now provide 3D molecular profiling at cellular or subcellular resolution using light-sheet microscopy43,44,45. Increasing multiplex capacity and use of artificial intelligence and machine learning algorithms to overcome data analytical challenges have successfully been deployed to image whole-organismal vasculature following tissue clearing46,47.