Skip to main content

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.

  • Article
  • Published:

Single-cell and spatial transcriptomics reveal somitogenesis in gastruloids

A Publisher Correction to this article was published on 05 March 2020

This article has been updated

Abstract

Gastruloids are three-dimensional aggregates of embryonic stem cells that display key features of mammalian development after implantation, including germ-layer specification and axial organization1,2,3. To date, the expression pattern of only a small number of genes in gastruloids has been explored with microscopy, and the extent to which genome-wide expression patterns in gastruloids mimic those in embryos is unclear. Here we compare mouse gastruloids with mouse embryos using single-cell RNA sequencing and spatial transcriptomics. We identify various embryonic cell types that were not previously known to be present in gastruloids, and show that key regulators of somitogenesis are expressed similarly between embryos and gastruloids. Using live imaging, we show that the somitogenesis clock is active in gastruloids and has dynamics that resemble those in vivo. Because gastruloids can be grown in large quantities, we performed a small screen that revealed how reduced FGF signalling induces a short-tail phenotype in embryos. Finally, we demonstrate that embedding in Matrigel induces gastruloids to generate somites with the correct rostral–caudal patterning, which appear sequentially in an anterior-to-posterior direction over time. This study thus shows the power of gastruloids as a model system for exploring development and somitogenesis in vitro in a high-throughput manner.

This is a preview of subscription content, access via your institution

Access options

Rent or buy this article

Prices vary by article type

from$1.95

to$39.95

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: scRNA-seq and tomo-seq on mouse gastruloids and comparison to embryos.
Fig. 2: Time-lapse imaging and perturbation of the segmentation clock in mouse gastruloids.
Fig. 3: Stainings and time-lapse imaging of somite formation in gastruloids embedded in low percentages of Matrigel.

Similar content being viewed by others

Data availability

All RNA sequencing datasets produced in this study are deposited in the Gene Expression Omnibus (GEO) under accession code GSE123187. All scRNA-seq and tomo-seq data can be explored at https://avolab.hubrecht.eu/MouseGastruloids2020. Source Data for Figs. 1, 2 and Extended Data Fig. 16, 8, 10 are provided with the paper. Any other relevant data are available from the corresponding authors upon reasonable request.

Code availability

All code is available at https://github.com/anna-alemany/mouseGastruloids_scRNAseq_tomoseq and https://github.com/vincentvbatenburg/MouseGastruloids.

Change history

References

  1. van den Brink, S. C. et al. Symmetry breaking, germ layer specification and axial organisation in aggregates of mouse embryonic stem cells. Development 141, 4231–4242 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  2. Turner, D. A. et al. Anteroposterior polarity and elongation in the absence of extra-embryonic tissues and of spatially localised signalling in gastruloids: mammalian embryonic organoids. Development 144, 3894–3906 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  3. Beccari, L. et al. Multi-axial self-organization properties of mouse embryonic stem cells into gastruloids. Nature 562, 272–276 (2018).

    Article  ADS  CAS  PubMed  Google Scholar 

  4. Pijuan-Sala, B. et al. A single-cell molecular map of mouse gastrulation and early organogenesis. Nature 566, 490–495 (2019).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  5. Nowotschin, S. et al. The emergent landscape of the mouse gut endoderm at single-cell resolution. Nature 569, 361–367 (2019).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  6. Kwon, G. S., Viotti, M. & Hadjantonakis, A. K. The endoderm of the mouse embryo arises by dynamic widespread intercalation of embryonic and extraembryonic lineages. Dev. Cell 15, 509–520 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Scotti, M. & Kmita, M. Recruitment of 5′ Hoxa genes in the allantois is essential for proper extra-embryonic function in placental mammals. Development 139, 731–739 (2012).

    Article  CAS  PubMed  Google Scholar 

  8. Koch, F. et al. Antagonistic activities of Sox2 and Brachyury control the fate choice of neuro-mesodermal progenitors. Dev. Cell 42, 514–526 (2017).

    Article  CAS  PubMed  Google Scholar 

  9. Chal, J. & Pourquié, O. Making muscle: skeletal myogenesis in vivo and in vitro. Development 144, 2104–2122 (2017).

    Article  CAS  PubMed  Google Scholar 

  10. Laurent, F. et al. HAND2 target gene regulatory networks control atrioventricular canal and cardiac valve development. Cell Rep. 19, 1602–1613 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Junker, J. P. et al. Genome-wide RNA tomography in the zebrafish embryo. Cell 159, 662–675 (2014).

    Article  CAS  PubMed  Google Scholar 

  12. Chal, J. et al. Differentiation of pluripotent stem cells to muscle fiber to model Duchenne muscular dystrophy. Nat. Biotechnol. 33, 962–969 (2015).

    Article  CAS  PubMed  Google Scholar 

  13. Dequéant, M. L. et al. A complex oscillating network of signaling genes underlies the mouse segmentation clock. Science 314, 1595–1598 (2006).

    Article  ADS  PubMed  Google Scholar 

  14. Oates, A. C., Morelli, L. G. & Ares, S. Patterning embryos with oscillations: structure, function and dynamics of the vertebrate segmentation clock. Development 139, 625–639 (2012).

    Article  CAS  PubMed  Google Scholar 

  15. Sonnen, K. F. et al. Modulation of phase shift between Wnt and Notch signaling oscillations controls mesoderm segmentation. Cell 172, 1079–1090.e12 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Hubaud, A., Regev, I., Mahadevan, L. & Pourquié, O. Excitable dynamics and Yap-dependent mechanical cues drive the segmentation clock. Cell 171, 668–682.e11 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Naiche, L. A., Holder, N. & Lewandoski, M. FGF4 and FGF8 comprise the wavefront activity that controls somitogenesis. Proc. Natl Acad. Sci. USA 108, 4018–4023 (2011).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  18. Dubrulle, J., McGrew, M. J. & Pourquié, O. FGF signaling controls somite boundary position and regulates segmentation clock control of spatiotemporal Hox gene activation. Cell 106, 219–232 (2001).

    Article  CAS  PubMed  Google Scholar 

  19. Sawada, A. et al. Fgf/MAPK signalling is a crucial positional cue in somite boundary formation. Development 128, 4873–4880 (2001).

    Article  CAS  PubMed  Google Scholar 

  20. Baillie-Johnson, P., van den Brink, S. C., Balayo, T., Turner, D. A. & Martinez Arias, A. Generation of aggregates of mouse embryonic stem cells that show symmetry breaking, polarization and emergent collective behaviour in vitro. J. Vis. Exp. 105, e53252 (2015).

    Google Scholar 

  21. Choi, H. M. T. et al. Third-generation in situ hybridization chain reaction: multiplexed, quantitative, sensitive, versatile, robust. Development 145, dev165753 (2018).

    PubMed  PubMed Central  Google Scholar 

  22. Lauschke, V. M., Tsiairis, C. D., François, P. & Aulehla, A. Scaling of embryonic patterning based on phase-gradient encoding. Nature 493, 101–105 (2013).

    Article  ADS  PubMed  Google Scholar 

  23. Matsumiya, M., Tomita, T., Yoshioka-Kobayashi, K., Isomura, A. & Kageyama, R. ES cell-derived presomitic mesoderm-like tissues for analysis of synchronized oscillations in the segmentation clock. Development 145, dev156836 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  24. Fehling, H. J. et al. Tracking mesoderm induction and its specification to the hemangioblast during embryonic stem cell differentiation. Development 130, 4217–4227 (2003).

    Article  CAS  PubMed  Google Scholar 

  25. Faunes, F. et al. A membrane-associated β-catenin/Oct4 complex correlates with ground-state pluripotency in mouse embryonic stem cells. Development 140, 1171–1183 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Ferrer-Vaquer, A. et al. A sensitive and bright single-cell resolution live imaging reporter of Wnt/ß-catenin signaling in the mouse. BMC Dev. Biol. 10, 121 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Papanayotou, C. et al. A novel Nodal enhancer dependent on pluripotency factors and Smad2/3 signaling conditions a regulatory switch during epiblast maturation. PLoS Biol. 12, e1001890 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  28. Kalmar, T. et al. Regulated fluctuations in Nanog expression mediate cell fate decisions in embryonic stem cells. PLoS Biol. 7, e1000149 (2009).

    Article  PubMed  PubMed Central  Google Scholar 

  29. Turner, D. A. et al. Wnt/β-catenin and FGF signalling direct the specification and maintenance of a neuromesodermal axial progenitor in ensembles of mouse embryonic stem cells. Development 141, 4243–4253 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Turner, D. A., Rué, P., Mackenzie, J. P., Davies, E. & Martinez Arias, A. Brachyury cooperates with Wnt/β-catenin signalling to elicit primitive-streak-like behaviour in differentiating mouse embryonic stem cells. BMC Biol. 12, 63 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  31. Turner, D. A., Trott, J., Hayward, P., Rué, P. & Martinez Arias, A. An interplay between extracellular signalling and the dynamics of the exit from pluripotency drives cell fate decisions in mouse ES cells. Biol. Open 3, 614–626 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  32. van Batenburg, V. et al. Generating gastruloids with somite-like structures from mouse embryonic stem cells. Protoc. Exch. https://doi.org/10.21203/rs.2.18203/v1 (2020).

  33. Muraro, M. J. et al. A single-cell transcriptome atlas of the human pancreas. Cell Syst. 3, 385–394.e3 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Grün, D. et al. Single-cell messenger RNA sequencing reveals rare intestinal cell types. Nature 525, 251–255 (2015).

    Article  ADS  PubMed  Google Scholar 

  35. Wolf, F. A., Angerer, P. & Theis, F. J. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 19, 15 (2018).

    Article  PubMed  PubMed Central  Google Scholar 

  36. Johnson, W. E., Li, C. & Rabinovic, A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 8, 118–127 (2007).

    Article  PubMed  Google Scholar 

  37. Chakraborty, S., Datta, S. & Datta, S. Surrogate variable analysis using partial least squares (SVA-PLS) in gene expression studies. Bioinformatics 28, 799–806 (2012).

    Article  CAS  PubMed  Google Scholar 

  38. Traag, V. A., Waltman, L. & van Eck, N. J. From Louvain to Leiden: guaranteeing well-connected communities. Sci. Rep. 9, 5233 (2019).

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  39. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).

    Article  CAS  PubMed  Google Scholar 

  40. Glynn, E. F., Chen, J. & Mushegian, A. R. Detecting periodic patterns in unevenly spaced gene expression time series using Lomb–Scargle periodograms. Bioinformatics 22, 310–316 (2006).

    Article  CAS  PubMed  Google Scholar 

  41. Richardson, L. et al. EMAGE mouse embryo spatial gene expression database: 2014 update. Nucleic Acids Res. 42, D835–D844 (2014).

    Article  CAS  PubMed  Google Scholar 

  42. van den Brink, S. C. et al. Single-cell sequencing reveals dissociation-induced gene expression in tissue subpopulations. Nat. Methods 14, 935–936 (2017).

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

This work was supported by an European Research Council Advanced grant (ERC-AdG 742225-IntScOmics), a Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO) TOP award (NWO-CW 714.016.001) and the Foundation for Fundamental Research on Matter, financially supported by NWO (FOM-14NOISE01) to S.C.v.d.B., A.A., V.v.B., M.B., J.V. and A.v.O., a Biotechnology and Biological Sciences Research Council (no. BB/P003184/1), Newton Trust (INT16.24b) and Medical Research Council (MR/R017190/1) grant to A.M.A., a Newnham College Cambridge Junior Research Fellowship to N.M. and a studentship from the Engineering and Physical Sciences Research Council to P.B.-J. The Cambridge Stem Cell Institute is supported by core funding from the Wellcome Trust and Medical Research Council; J.N. was funded by the University of Cambridge and K.F.S. by core funding from the Hubrecht Institute. This work is part of the Oncode Institute, which is partly financed by the Dutch Cancer Society. We thank A. Ebbing and M. Betist for the robotized tomo-seq protocol; G. Keller for the Brachyury-GFP cell line; J. Collignon for the Nodal-YFP line; K. Hadjantonakis for the TCF/LEF-mCherry line; S. van den Brink and E. R. Maandag for the E14-IB10 cells; J. Kress and A. Aulehla for the LfngT2AVenus mouse ES cell line; I. Misteli Guerreiro, J. Peterson-Maduro and J. Hoeksma for suggestions for in situ hybridization experiments; W. Thomas, Y. el Azhar, J. Juksar and J. Beumer for reagents and inhibitors; A. de Graaff and A. Stokkermans for help with multiphoton microscopy and analysis of the microscopy data; D. A. Turner for microscopy panels that were used for tomo-seq validation; J. Korving for help with the somite-size measurements in embryos; the Hubrecht FACS facility and R. van der Linden for FACS experiments; Single Cell Discoveries for 10x Genomics scRNA-seq; the Utrecht Sequencing facility for sequencing; and P. Zeller, H. Viñas Gaza, M. Vaninsberghe, V. Bhardwaj and all members of the van Oudenaarden, Sonnen and Martinez Arias laboratories for discussions.

Author information

Authors and Affiliations

Authors

Contributions

S.C.v.d.B. and A.v.O. conceived and designed the project. S.C.v.d.B. and V.v.B. generated gastruloids, and S.C.v.d.B., M.B. and J.V. performed scRNA-seq experiments. Embedding of mouse gastruloids for tomo-seq was done by S.C.v.d.B. N.M. and P.B.-J. embedded mouse embryos for tomo-seq, with help from J.N. S.C.v.d.B. cryosectioned gastruloids and embryos, and performed tomo-seq experiments. J.V. developed the robotized tomo-seq protocol. A.A. performed the mapping and analysis, including comparisons with embryonic datasets, of the scRNA-seq and tomo-seq data. A.v.O. performed the linearized UMAP analysis. S.C.v.d.B., M.B., A.A., N.M. and A.M.A. interpreted the sequencing datasets. P.B.-J. performed the first Matrigel-embedding pilot experiments. V.v.B. performed time-lapse imaging experiments, in situ hybridization and HCR stainings, with help from S.C.v.d.B. and K.F.S. V.v.B. analysed the microscopy data, with support from K.F.S. V.v.B., S.c.v.d.B., A.v.O. and K.F.S. interpreted the imaging results. S.C.v.d.B., A.A., V.v.B. and A.v.O. wrote the manuscript with support from K.F.S. and A.M.A., and A.M.A. and A.v.O. guided the project.

Corresponding authors

Correspondence to Susanne C. van den Brink or Alexander van Oudenaarden.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature thanks Jianping Fu and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

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

Extended data figures and tables

Extended Data Fig. 1 scRNA-seq on 120-h mouse gastruloids and comparison to embryos.

a, FACS gating strategy before scRNA-seq. Live cells were selected on the basis of DAPI staining. Four sequential gates (P1–P4) were used; cells from gate P4 were used for scRNA-seq. SSC, side scatter; FSC, forward scatter; H, height; W, width; A, area. b, Box plot showing the median number of transcripts (left) and genes (right) detected per cell for SORT-seq experiments on E14-IB10 (E14-S, n = 5,951 cells from 26 biologically independent samples) and LfngT2AVenus gastruloids (Lfng-S, n = 4,592 cells from 74 biologically independent samples), and for 10x Genomics experiments on LfngT2AVenus gastruloids (Lfng-10x, n = 14,659 cells from 74 biologically independent samples). SORT-seq and 10x Genomics analyses were performed in parallel on the same 74 biologically independent LfngT2AVenus gastruloids; all cells extracted from these gastruloids were pooled and split into two tubes, of which one was used for SORT-seq and the other for 10x Genomics. The box extends from the lower to the upper quartile; whiskers are 1.5× the interquartile range; flier points are those past the end of the whiskers. c, UMAP plot for each experiment separately (n = 5,883, 4,589 and 14,636 cells for E14-S, Lfng-S and Lfng-10x, respectively; Methods). The E14-S cells (n = 5,883) were extracted from n = 26 biologically independent samples; the Lfng-S and Lfng-10x cells (n = 4,589 and 14,636, respectively) were extracted from n = 74 biologically independent samples that were pooled and then split into one tube for SORT-seq and one tube for 10x Genomics. The colour of each cell is the same as the colour of that particular cell in Fig. 1a. d, UMAP plot obtained by analysing all the cells from the different experiments together (n = 25,202 cells from 100 biologically independent samples), in which cells are coloured according to their batch (Methods, Supplementary Table 1). The black line indicates the symmetry line in clusters 1–8 used to generate the linearized UMAP plot in Extended Data Fig. 2d (Methods). e, Fraction of E14-IB10 (n = 26 biologically independent samples) and LfngT2AVenus (n = 74 biologically independent samples) cells in each scRNA-seq cluster from Fig. 1a. Blue, green and black numbers, number of E14-IB10, LfngT2AVenus and total cells in each cluster, respectively (Supplementary Tables 1, 4). f, Fraction of cells for each cell type in each plate in SORT-seq experiments (Lfng-S, n = 19 plates containing cells from n = 74 biologically independent gastruloids; E14-S, n = 30 plates containing cells from n = 26 biologically independent gastruloids), and in each experimental batch in 10x Genomics experiments (Lfng-10x, n = 2 independent batches containing cells extracted from n = 44 and 30 biologically independent gastruloids, respectively, with 2 technical replicates each). In the box plots, centre line is median; box limits are the 1st and 3rd quartiles; and whiskers denote the range. g, Fraction of cells detected in the E8.5 mouse embryo scRNA-seq dataset4 with which we compared our gastruloid scRNA-seq data. Exact numbers in each cluster are indicated. h, Dot plot showing the number of overlapping genes between significantly upregulated genes (n = 79, 87, 84, 22, 84, 66, 82, 78, 100, 97, 100, 96 and 90 genes for clusters 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 and 13 (respectively) in the gastruloid dataset, determined using the two-side t-test, followed by selection of genes with fold change above 1.01 and P value below 0.01; n = 7, 20, 21, 35, 200, 39, 23, 200, 200, 95, 54, 21, 58, 57, 200, 81, 135, 28, 200 and 200 genes for the embryonic-cell types reported in the x axis, determined in ref. 4 and selecting genes with P value below 0.01) for each gastruloid cluster (n = 25,202 cells extracted from 100 biologically independent samples) and each E8.5 mouse embryonic-cell type4. Dot colour indicates the probability of finding such a number of overlapping genes between the two sets by random chance (P value determined by binomial testing, one-sided, no adjustments for multiple corrections were made). Dot size represents the number of overlapping genes. i, Dot plot showing overlapping genes between significantly upregulated genes for each gastruloid scRNA-seq cluster (n = 79, 87, 84, 22, 84, 66, 82, 78, 100, 97, 100, 96 and 90 genes for clusters 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 and 13, respectively (Supplementary Table 2); scRNA-seq dataset obtained for 25,202 cells that were extracted from n = 100 biologically independent gastruloids), and upregulated genes for each E7.0–E8.5 mouse embryonic-cell type4. Dot colour indicates the probability of finding such a number of overlapping genes between the two sets by random chance (P value determined by binomial testing, one-sided, no adjustments for multiple corrections were made), and dot size represents the number of overlapping genes. Blue, embryonic stage. 10x, 10x Genomics; Ant, anterior; EnD, endoderm; haemato, haemato-endothelial; prog, progenitors; S, SORT-seq33.

Source Data

Extended Data Fig. 2 Expression of relevant markers in the gastruloid scRNA-seq dataset.

a, Mean log expression of relevant markers of outlier populations (clusters 9–13) plotted on the UMAP plot from Fig. 1a. Olfr129 and Onecut1, head mesenchyme (cluster 9); Etv2, haemato-endothelial progenitors (bottom part of cluster 10); Kdr, haemato-endothelial progenitors and endothelium (cluster 10); Cdh5 and Tie1, endothelium (top part of cluster 10); Tbx4, Hoxa11, Ass1 and Bmp7, allantois (cluster 11); Ephx2, Mt1, Utf1 and Pou5f1, primordial-germ-cell-like or extra-embryonic ectoderm (cluster 12); Col4a1, Epcam and Sox17, endoderm (cluster 13). b, Mean log normalized expression of relevant markers of clusters 1–8 plotted on the UMAP plot from Fig. 1a. Hand2 and Gata6, heart (cluster 1); Meox2 and Pax3, differentiated somite (cluster 3); Aldh1a2 and Uncx4.1, somite (cluster 4); Lfng, Mesp2, Ripply2 and Dll1, differentiation front (cluster 5); Hes7 and Tbx6, presomitic mesoderm (cluster 6); Wnt3a, Fgf17, Fgf8, Cyp26a1, Nkx1-2 and T, tail bud containing neuromesodermal progenitors (cluster 7); Pax6, Sox1, Hes3 and Sox2, differentiated neural cells (spinal cord; cluster 8). Expression was first count-normalized to 10,000 for each cell (Methods), and then log-transformed. Additional markers of all clusters are provided in Supplementary Table 2. c, Percentage of total unique transcripts per cell corresponding to stress genes42 plotted on the UMAP plot from Fig. 1a. d, Linearized UMAP plot of clusters 1–8 (top, n = 24,148 cells, isolated from 100 biologically independent gastruloids during n = 6 independent experiments) and expression profiles of genes related to neural and mesodermal differentiation8,9 (bottom). Green and grey shades indicate location of cardiac cells and neuromesodermal progenitors, respectively. The position of each cell along the x-axis relates to its differentiated state towards a neural or mesodermal fate.

Source Data

Extended Data Fig. 3 The number of genes and reads in gastruloid and embryo tomo-seq datasets, and comparison to microscopy data.

ac, Number of unique transcripts and genes detected in 3 E14-IB10 120-h mouse gastruloids that were sectioned using 20-μm sections and 2 E14-IB10 120-h mouse gastruloids that were sectioned using 8-μm sections (a); in 3 LfngT2AVenus 120-h mouse gastruloids that were sectioned using 20-μm sections (b); and in 3 E8.5 mouse embryos that were sectioned using 20-μm sections (c). Owing to their lengths, embryo sections were collected in 2 or 3 sequential 96-well plates. d, Validation of tomo-seq data with microscopy. Top panels, Brachyury-GFP, WNT signalling activity (as reported using a TCF/LEF-mCherry mouse ES cell line) and Nodal-YFP expression in 120-h mouse gastruloids as measured by microscopy (Methods). With each reporter line, we obtained similar results in n = 5 independent experiments. Bar plots show the normalized expression levels of Brachyury, Wnt3a and Nodal in 120-h E14-IB10 gastruloids, 120-h LfngT2AVenus gastruloids and E8.5 mouse embryos as determined by tomo-seq (Methods), and in the posterior mesoderm of E9.5 mouse embryos as determined by microarray12. e, Scaled average z-score of significantly upregulated genes (P value < 0.01 and log2-transformed fold change > 1.01) detected in each single-cell cluster from Fig. 1a (Supplementary Table 2) as measured in the averaged LfngT2AVenus tomo-seq gastruloid. Scale bar, 100 μm.

Source Data

Extended Data Fig. 4 Individual replicates of gastruloids, E8.5 embryo tomo-seq and E9.5 posterior mesoderm datasets, and comparison to gastruloid and E8.5 embryonic scRNA-seq datasets.

a, Heat maps showing the anterior–posterior expression patterns of 1,199 genes as detected by tomo-seq11 in individual replicates of 120-h E14-IB10 gastruloids (n = 3 gastruloids, 20-μm sections and n = 2 gastruloids, 8-μm sections) that were cultured in standard1,20 (non-Matrigel-based) conditions; average heat map of the five replicates; average expression of genes found in each tomo-seq domain in the E14-IB10 tomo-seq dataset, projected in the UMAP plot from Fig. 1a; dot plot showing overlapping genes between genes detected in each tomo-seq domain in the E14-IB10 tomo-seq dataset, and upregulated genes for each E8.5 mouse embryonic-cell type4. Dot colour represents the probability of finding such a number of overlapping genes between the two sets by random chance (Methods), and dot size represents the number of overlapping genes. Only genes that were reproducible between replicates are shown (Methods). Genes are clustered on the basis of their anterior–posterior expression pattern (Methods); bars marked with Roman numerals represent tomo-seq clusters. Clusters I, II, III, IV, V, VI, VII, VIII and IX contain n = 58, 231, 72, 138, 38, 43, 165, 90 and 364 genes, respectively. b, Similar to a, but for 1,456 genes in 120-h LfngT2AVenus15 (n = 3 gastruloids, 20-μm sections) gastruloids that were cultured in standard1,20 (non-Matrigel-based) conditions. Clusters I, II, III, IV, V, VI, VII, VIII, IX, X and XI contain n = 74, 235, 47, 235, 99, 47, 264, 287, 99, 14 and 55 genes, respectively. c, Similar to a, but for 1,553 genes in E8.5 embryos (n = 3 embryos, 20-μm sections). Clusters I, II, III, IV, V, VI, VII, VIII, IX, X, XI, XII, XIII, XIV, XV, XVI, XVII, XVIII and XIX contain n = 186, 179, 63, 25, 114, 103, 48, 65, 28, 65, 111, 89, 61, 44, 134, 155, 79, 2 and 2 genes, respectively. d, Similar to a, but for 1,989 genes in an E9.5 mouse embryo posterior-mesoderm dataset (tail bud to newly formed somite) (n = 3 embryos; previously published microarray data; approximately 100-μm sections12). Clusters I, II, III, IV, V and VI contain n = 294, 512, 226, 165, 181 and 611 genes, respectively. All genes are shown in Supplementary Table 6. AP, anterior–posterior; FMH, fore-, mid- and hind-.

Source Data

Extended Data Fig. 5 Comparisons between mouse gastruloid and mouse embryo datasets, including genes that are reproducible in at least one system.

a, Heat map showing the average anterior–posterior expression pattern of 2,065 genes as detected by tomo-seq11 in 120-h mouse gastruloids that were generated from E14-IB10 and LfngT2AVenus15 mouse ES cells and that were cultured in standard1,20 (non-Matrigel-based) conditions; average expression of genes found in each tomo-seq domain in the E14-IB10–LfngT2AVenus comparison heat map, projected in the UMAP plot from Fig. 1a; dot plot showing overlapping genes between genes detected in each tomo-seq domain in a, and upregulated genes for E8.5 mouse embryonic-cell types4. Dot colour represents the probability of finding such a number of overlapping genes by random chance (Methods), and dot size represents the number of overlapping genes. In contrast to the heat maps in Fig. 1, this heat map contains genes that were reproducible in either E14-IB10 (n = 3 biologically independent gastruloids, 20-μm sections and n = 2 biologically independent gastruloids, 8-μm sections) or LfngT2AVenus (n = 3 biologically independent gastruloids, 20-μm sections) gastruloids (Extended Data Fig. 4, Methods, Supplementary Tables 5, 6). This means that genes that are reproducible in E14-IB10 replicates but not in LfngT2AVenus replicates—and vice versa—are included. Genes are clustered on the basis of their anterior–posterior expression pattern (Methods); bars marked by Roman numerals represent tomo-seq clusters, which are also indicated with the grey–black bar plot. The red-to-white bar plots indicate the P value of reproducibility of each gene in each heat map. The order of these bar plots corresponds to the order of the heat maps. Clusters I, II, III, IV, V, VI and VII contain n = 377, 398, 259, 124, 145, 395 and 367 genes, respectively. b, Similar to a, but for 2,894 genes that were reproducible in E14-IB10 (n = 3 gastruloids, 20-μm sections and n = 2 gastruloids, 8-μm sections) or LfngT2AVenus (n = 3 gastruloids, 20-μm sections) or E8.5 mouse embryos (n = 3 embryos, 20-μm sections). Clusters I, II, III, IV, V, VI, VII and VIII contain n = 477, 618, 302, 161, 230, 527, 179 and 400 genes, respectively. c, Similar to a, but for 3,086 genes that were reproducible in E14-IB10 (n = 3 gastruloids, 20-μm sections and n = 2 gastruloids, 8-μm sections) or LfngT2AVenus (n = 3 gastruloids, 20-μm sections) or the E9.5 mouse embryo posterior-mesoderm dataset (tail bud to newly formed somite) (n = 3 embryos; previously published microarray data; approximately 100-μm sections12). Clusters I, II, III, IV, V, VI, VII, VIII and IX contain 419, 325, 392, 337, 114, 279, 602, 220 and 398 genes, respectively. Gene lists are provided in Supplementary Table 8.

Source Data

Extended Data Fig. 6 Gene-expression profiles in gastruloid and embryo tomo-seq datasets.

Line plots for the normalized anterior–posterior expression of genes emphasized in Fig. 1b, d, e for the E14-IB10 and LfngT2AVenus gastruloids, and for the E8.5 mouse embryo, as measured by tomo-seq11. Each colour denotes a different replicate.

Source Data

Extended Data Fig. 7 Kymographs of time-lapse experiments performed on LfngT2AVenus gastruloids that were embedded in 100% Matrigel at 96 h.

ad, Kymographs (space–time plots) of bright field channel and LfngT2AVenus signal along the anterior–posterior axis of all replicates from all time-lapse experiments (experiments 1–4) that are presented in Fig. 2e, Extended Data Fig. 8e, f. These gastruloids were embedded in 100% Matrigel (Methods) to stabilize them during imaging, and subsequently imaged for at least 17 h (Supplementary Videos 1, 2, 4, 5). Inhibitors were added at the start of the time lapse (Methods) and are indicated above the kymographs, together with their concentration. Asterisks refer to gastruloids used to generate Fig. 2e, Extended Data Fig. 8b–d. Similar results were obtained in n = 5 and 4 independent experiments for a and bd, respectively. e, Imaging of a LfngT2AVenus gastruloid that was embedded in 100% Matrigel at 96 h, and to which the Notch inhibitor DAPT was added at 96.5 h (Supplementary Video 2); the Lfng signal disappears about 6 h after addition of DAPT. Corresponding kymographs in a. Similar results were obtained in n = 5 independent experiments. Scale bar, 200 μm.

Extended Data Fig. 8 Detrending procedure and Lomb–Scargle analysis of replicates, as well as measurements of elongation and differentiation front speed in small panel screening, and upon BGJ389 and PD17 treatment.

Replicates subjected to detrending and Lomb–Scargle analysis are from Fig. 2. a, Black line, measured intensity of the Lfng signal along the white dashed line in Fig. 2c; blue line, trend (Methods) of this signal and periodogram of the Lfng oscillations in Fig. 2d, as determined by Lomb–Scargle decomposition. b, As in a, but then for the 13 DMSO-control LfngT2AVenus gastruloid replicates shown in Extended Data Fig. 7c, d. c, Cyclical component of the scaled intensity of the LfngT2AVenus oscillations relative to the trend line shown in b. A.U., arbitrary units. d, Periodogram of the Lfng oscillations in c, as determined by Lomb–Scargle decomposition (Methods). Gastruloids used for this experiment were embedded in 100% Matrigel at 96 h, and subsequently imaged for at least 17 h. e, f, Speed of posterior gastruloid elongation (VPSM) and speed of posteriorly moving differentiation front (VDIFF) (see explanation in Extended Data Fig. 9a) in LfngT2AVenus gastruloids treated with DMSO (control) or with various inhibitors (Supplementary Videos 3, 5). Points refer to replicates; n = 14, 3, 3, 3, 1, 2 and 3 (from left to right) and 9, 3, 6 and 6 (from left to right) replicates for e and f, respectively. Kymographs of replicates are shown in Extended Data Fig. 7. In the box plots, centre line is median; box limits are the 1st and 3rd quartiles; and whiskers denote the range.

Source Data

Extended Data Fig. 9 Explanation for how elongation and differentiation-front speed were measured, and HCR stainings and kymographs of gastruloids embedded in 10% Matrigel.

a, Kymographs (space–time plots) of bright field channel and LfngT2AVenus signal along the anterior–posterior axis of a DMSO-treated (control) and a PD03 (MEKinhibitor)-treated LfngT2AVenus gastruloid. Gastruloids were embedded in 100% Matrigel at 96 h; DMSO or PD03 (66.7 μM) was added at 96.5 h. Kymographs were used to measure the elongation speed of the gastruloid (angle of blue dashed line; VPSM) (Methods) and the speed of the differentiation front (angle of red dashed line; VDIFF). Similar results were obtained in n = 4 independent experiments. b, Illustration explaining the effect of FGF inhibition, which increases the speed of the differentiation front (red arrows, VDiff) without altering the elongation rate (blue arrows, VPSM) of gastruloids. Three time points (t1, t2 and t3) are depicted. White tissue, nondifferentiated tissue (presomitic mesoderm); grey tissue, differentiated tissue. c, In situ hybridization staining for Uncx4.1 on 120-h LfngT2AVenus gastruloids that were not embedded in Matrigel (0%) (standard, previously published protocol1,20) or that were embedded in 25% or 100% Matrigel at 96 h. Numbers below the panels indicate the number of gastruloids in which stripy Uncx4.1 expression patterns were observed. Similar results were obtained in n = 3 independent experiments. d, LfngT2AVenus gastruloids that were embedded in 10% Matrigel (Methods) at 96 h and stained for Uncx4.1 using HCR21 at 120 h. A magnified view of the left gastruloid is shown in Fig. 3a. Similar results were obtained in n = 5 independent experiments. e, Kymographs of LfngT2AVenus signal and bright field channel along the anterior–posterior axis of gastruloids that were embedded in 10% Matrigel at 96 h, and subsequently imaged for 20 h (Supplementary Video 6). Top kymograph belongs to the gastruloid that is shown in Fig. 3b. Similar results were obtained in n = 2 independent experiments. Scale bars, 200 μm.

Extended Data Fig. 10 Uncx4.1 Tbx18 Ripply2 stainings and somite-size measurements.

a, HCR21 double staining for Uncx4.1 (cyan) and Tbx18 (magenta) on 120-h LfngT2AVenus gastruloids embedded in 10% Matrigel at 96 h. For replicate 4, 1.3 μM of PD03 was added at 96.5 h. Similar results were obtained in n = 4 independent experiments. b, Similar to a, but for Uncx4.1 (cyan) and Ripply2 (yellow). Similar results were obtained in n = 2 independent experiments. c, Intensity of Uncx4.1 and Tbx18 signal along the anterior–posterior axis of the gastruloids in a. Peaks (circles) are called on the smoothened Uncx4.1 profile (dark blue) (Methods). d, Similar to c, but for the Uncx4.1- and Ripply2-stained gastruloids from b. e, Distance between Uncx4.1 peaks in the 120-h LfngT2AVenus gastruloids (n = 7) from replicates 1–6 in ad and in replicate 7 (which is shown in Fig. 3c). Replicate 8 was excluded from quantification and both replicate 4 and replicate 7 were incubated in 1.3 μM PD03 from 96–120 h. Scale bars, 200 μm.

Source Data

Supplementary information

Supplementary Methods

Reporting Summary

Supplementary Table 1

Metadata for all the cells from the 120 h mouse gastruloid scRNA-seq dataset that passed our quality-control checks. Table provides UMAP coordinates obtained for the n = 25,202 cells isolated from 120 h gastruloids (n = 26 and 74 biologically independent gastruloids grown using E14-IB10 and LfngT2AVenus ESC lines, respectively, from n = 6 independent experiments). CellID, cell id; u1, UMAP coordinate 1; u2, UMAP coordinate 2; n_counts, number of unique reads detected in this cell; n_genes, number of genes detected in this cell. E14_sort corresponds to SORT-seq experiments performed on the E14-IB10 cell line; Lfng_sort corresponds to SORT-seq experiments performed on the LfngT2AVenus cell line; 10x correspond to 10x Genomics scRNA-seq experiments performed on the LfngT2AVenus cell line.

Supplementary Table 2

Differentially expressed genes detected for each cluster of cells in scRNA-seq dataset of 120 h mouse gastruloids, determined using the two-side t-test and adjusted for multiple testing with Benjamini-Hochberg correction.

Supplementary Table 3

The number of overlapping genes between significantly upregulated genes (n = 79, 87, 84, 22, 84, 66, 82, 78, 100, 97, 100, 96, 90 genes for clusters 1-13 in the gastruloid dataset, determined using the two-side t-test, followed by selection of genes with fold change above 1.01 and P value below 0.01; n = 7, 20, 21, 35, 200, 39, 23, 200, 200, 95, 54, 21, 58, 57, 200, 81, 135, 28, 200, 200 genes for the embryonic cell types reported in the x axis, determined in reference 4 and selecting genes with P value below 0.01) for each gastruloid cluster (n = 25,202 cells extracted from 100 biologically independent samples) and each E8.5 mouse embryonic cell type4. P value was determined by binomial testing, one-sided, no adjustments for multiple corrections were made. PS, Pijuan-Sala dataset4.

Supplementary Table 4

Number of cells detected for each cluster in each experimental 120 h mouse gastruloid scRNA-seq batch. E14_sort corresponds to SORT-seq experiments performed on the E14-IB10 cell line; Lfng_sort corresponds to SORT-seq experiments performed on the LfngT2AVenus cell line; Lfng_10x correspond to 10x Genomics scRNA-seq experiments performed on the LfngT2AVenus cell line.

Supplementary Table 5

For each gene, the Pearson correlation value of the expression profiles detected in each pair-wise comparison of the different replicates is provided for all the tomo-seq experiments performed with E14-IB10 and LfngT2AVenus gastruloids (n = 5 and n = 3 biological replicates, respectively) and with E8.5 mouse embryos (n = 3 biological replicates). Additionally, the same analysis is performed on the posterior mesoderm (n = 3 biological replicates, tail bud to newly formed somite; previously published microarray data12). Corr(x,x) refers to the correlation value between expression patterns of a gene in the two replicates that are being compared in each case. P values and adjusted P values have been obtained as described in the Methods.

Supplementary Table 6

Clusters found for reproducible genes identified in E14-IB10 and LfngT2AVenus gastruloid tomo-seq data, E8.5 mouse embryos tomo-seq data, and posterior mesoderm of E9.5 mouse embryo (tail bud to newly formed somite; previously published microarray data12) (Methods, Extended Data Fig. 4).

Supplementary Table 7

Clusters found for the comparison between gene expression patterns of the E14-IB10 (n = 5 biological replicates) and LfngT2AVenus gastruloids (n = 3 biological replicates), the E14-IB10 and LfngT2AVenus gastruloids and the E8.5 mouse embryos (n = 3 biological replicates), and the E14-IB10 and LfngT2AVenus gastruloids and the posterior mesoderm of E9.5 embryos (n = 3 biological replicates; tail bud to newly formed somite; previously published microarray data12) (Methods). Gene selection here took the genes that are reproducible in both of the systems that are being compared. This table is associated with Fig. 1b, d, and e.

Supplementary Table 8

Clusters found for the comparison between gene expression patterns of the E14-IB10 (n = 5 biological replicates) and LfngT2AVenus gastruloids (n = 3 biological replicates), the E14-IB10 and LfngT2AVenus gastruloids and the E8.5 mouse embryos (n = 3 biological replicates), and the E14-IB10 and LfngT2AVenus gastruloids and the posterior mesoderm of a E9.5 embryo (n = 3 biological replicates; tail bud to newly formed somite; previously published microarray data12; Methods, Extended Data Fig. 5). Gene selection here took the union of genes that are reproducible in at least one of the systems that are being compared. This table is associated with Extended Data Fig. 5.

Supplementary Table 9

Number of replicates, number of genes detected and number of reproducible genes found in the different tomo-seq experiments (n = 5 biological replicates of E14-IB10 gastruloids, n = 3 biological replicates of LfngT2AVenus gastruloids and n = 3 biological replicates of E8.5 mouse embryos) and in the posterior mesoderm of E9.5 mouse embryo (n = 3 biological replicates; tail bud to newly formed somite; previously published microarray data12); number of genes being simultaneously reproducible in different tomo-seq experiments (labelled as intersection), and number of genes that are reproducible in at least one of the tomo-seq experiments being compared (labelled as union) between E14-IB10 and LfngT2AVenus gastruloids, between E14-IB10 gastruloids, LfngT2AVenus gastruloids and E8.5 mouse embryos, and between E14-IB10 gastruloids, LfngT2AVenus gastruloids and the posterior mesoderm of E9.5 mouse embryo12. This table explains how gene selection was performed in Supplementary Tables 6-8, Fig. 1b, d, and e, and Extended Data Figs. 4-5.

Supplementary Video 1 | Time-lapse imaging of LfngT2AVenus mouse gastruloids that are embedded in 100% Matrigel at 96 h

Gastruloids were mounted in 100% Matrigel at 96 h to stabilize them during imaging (Methods). Multi-photon imaging was started directly after embedding, and a z-stack was taken every 15 minutes for 17 hours. In red we show the summed projection of the fluorescence intensity (Methods). All images were rotated such that their AP axis is approximately oriented from left to right. Similar results obtained in N = 9 independent experiments. Scale bar, 200 μm.

Supplementary Video 2 | Time-lapse imaging of LfngT2AVenus mouse gastruloids treated with DMSO or DAPT

Gastruloids were mounted in 100% Matrigel at 96 h to stabilize them during imaging (Methods). Top row, DMSO treated control gastruloids; bottom row, gastruloids treated with the Notch-inhibitor DAPT (27 μM). DMSO/inhibitor was added at 97 h (Methods). Multi-photon imaging was started directly after embedding, and a z-stack was taken every 15 minutes for 16 hours. Red, summed projection of the fluorescence intensity (Methods). All images were rotated such that their AP axis is approximately oriented from left to right. Similar results obtained in N = 5 independent experiments. Scale bar, 200 μm.

Supplementary Video 3 | Time-lapse imaging of LfngT2AVenus mouse gastruloids treated with agonists or inhibitors of FGF, WNT and BMP signalling pathways

Gastruloids were mounted in 100% Matrigel at 96 h (Methods). DMSO (control), Chiron (CHI99021; WNT-agonist; 10 μM), FGF1 (0.02 μg/ml), FGF10 (0.2 μg/ml), IWP (IWP-2; WNT-antagonist; 2 μM), IWR (IWR-1; WNT-antagonist; 10 μM), LDN (LDN193189; BMP-inhibitor; 0.2 μM), or PD03 (PD0325901; MEK/ERK inhibitor; 13.3 μM) was added at 97 h (Methods). Imaging was started directly after embedding, and z-stacks were taken every 15 minutes for 17 hours. Red, summed intensity projection (Methods). All gastruloids were oriented with their AP axis approximately from left to right. Similar results obtained in N = 2 independent experiments.

Supplementary Video 4 | Time-lapse imaging of LfngT2AVenus mouse gastruloids treated with 1.3-66.7 μM PD03

Gastruloids were mounted in 100% Matrigel at 96 h to stabilize them during imaging (Methods). DMSO (control; top row) or PD03 (PD0325901; MEK/ERK (downstream of FGF signalling) inhibitor; concentrations between 1.3-66.7 μM as indicated in the video) was added at 97 h (Methods). Imaging was started directly after embedding, and a z-stack was taken every 15 minutes for 17 hours. Red, summed projection of the fluorescence intensity (Methods). All images were rotated such that their AP axis is approximately oriented from left to right. Similar results obtained in N = 4 independent experiments. Scale bar, 200 μm.

Supplementary Video 5 | Time-lapse imaging of LfngT2AVenus mouse gastruloids treated with DMSO, BGJ398 or PD17

Gastruloids were mounted in 100% Matrigel at 96 h to stabilize them during imaging (Methods). DMSO (control; top row), BGJ396 (FGF receptor inhibitor; 0.2 μM; middle row) or PD17 (PD 173074; FGF receptor inhibitor; 0.5 μM; bottom row) was added at 97 h (Methods). Imaging was started directly after embedding, and a z-stack was taken every 15 minutes for 17 hours. Red, summed projection of the fluorescence intensity (Methods). All images were rotated such that their AP axis is approximately oriented from left to right. Similar results obtained in N = 2 independent experiments. Scale bar, 200 μm.

Supplementary Video 6 | Time-lapse imaging of LfngT2AVenus mouse gastruloids that are embedded in 10% Matrigel at 96 h

Gastruloids were embedded in 10% Matrigel at 96 h (Methods). Multi-photon imaging was started directly after embedding, and a z-stack was taken every 15 minutes for 20 hours. In red we show the summed projection of the fluorescence intensity (Methods). All images were rotated such that their AP axis is approximately oriented from left to right. Similar results obtained in N = 2 independent experiments. Scale bar, 200 μm.

Source data

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

van den Brink, S.C., Alemany, A., van Batenburg, V. et al. Single-cell and spatial transcriptomics reveal somitogenesis in gastruloids. Nature 582, 405–409 (2020). https://doi.org/10.1038/s41586-020-2024-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41586-020-2024-3

This article is cited by

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing