The clock that controls spine development modelled in a dish

What do the flashes of a firefly and the chirpings of a cricket have in common? Both occur in a regular rhythm, which is controlled by an oscillating biological clock1. Another oscillating genetic clock controls the development of embryonic structures called somites, which give rise to the vertebrae that protect the spinal cord. Our knowledge of this segmentation clock stems almost entirely from research on animals2,3, because technical and ethical considerations limit the study of human embryos in culture. Writing in Nature, Diaz-Cuadros et al.4 and Matsuda et al.5 now report a breakthrough that enables studies of the human segmentation clock in vitro. In addition, Yoshioka-Kobayashi et al.6 use sophisticated techniques in mice to provide insights into the mechanisms that control the mammalian segmentation clock.

Somites arise from a tissue called the presomitic mesoderm (PSM). During somite formation, temporally and spatially controlled oscillations in transcription yield gene-expression waves that propagate through the PSM along the embryo’s head-to-tail axis. The result is a striped pattern of somites that forms the blueprint for the spine. Although the molecular components of the segmentation clock are highly evolutionarily conserved across vertebrates, new somites form with different rhythms in each species. For instance, gene oscillations have a period of 30 minutes in zebrafish and 2 hours in mice. Oscillations have been estimated to occur every 4 to 5 hours in humans2 — although until now they have never been directly observed.

Diaz-Cuadros et al. and Matsuda et al. set out to model the human clock using induced pluripotent stem cells (iPSCs) — cells that are generated in vitro from differentiated human cells and, similarly to embryonic stem cells, can give rise to every cell type in the body. The groups used established protocols79 to convert iPSCs into PSM in vitro.

To visualize and monitor the dynamic oscillations of clock genes in the cultured PSM in real time, each group used a different ‘reporter’ protein. Matsuda and colleagues used a reporter in which a key segmentation-clock gene10, Hes7, drives production of the bioluminescent enzyme luciferase. As Hes7 expression oscillates, levels of the reporter increase and decrease. Diaz-Cuadros et al. used an engineered version of Hes7 fused to a gene that encodes Achilles, which is a more rapidly generated variant of yellow fluorescent protein developed by Yoshioka-Kobayashi and colleagues. The use of Achilles enabled Diaz-Cuadros and co-workers to track fluorescent waves of Hes7 expression at the single-cell level4 — a resolution not possible with the luciferase reporter. Analyses using both reporters provide the first definitive evidence that the human segmentation clock has a period of approximately 5 hours (Fig. 1a).

Figure 1 | Modelling embryonic segmentation in vitro. A tissue called the presomitic mesoderm (PSM) gives rise to somites — embryonic precursors of vertebrae. This process involves a ‘segmentation clock’ that drives rhythmic oscillations of gene expression, including that of the gene Hes7. Three groups have developed systems to analyse the clock in culture using live-cell imaging. a, Diaz-Cuadros et al.4 and Matsuda et al.5 directed wild-type (WT) human induced pluripotent stem cells (iPSCs) to become PSM cells. The iPSCs had been engineered to express a version of Hes7 that drives expression (arrow) of genes encoding the fluorescent molecule Achilles4 or the luminescent molecule luciferase5. Monitoring the oscillations of these genes in PSM cells revealed that the human segmentation clock has a period of about 5 hours. b, Matsuda et al. performed the same experiment using iPSCs in which Hes7 is mutated, as in the skeletal disorder spondylocostal dysostosis, and found a lack of oscillations. c, Yoshioka-Kobayashi et al.6 isolated the PSM from mouse embryos carrying a Hes7–Achilles reporter, and monitored oscillations, which have a 2-hour period.

Three key signalling pathways — the Notch, Wnt and FGF pathways — act in sequential negative feedback loops to regulate oscillating gene expression during somite formation2,3,11,12. Diaz-Cuadros and colleagues used their culture system to investigate these pathways in detail. They confirmed the roles of these pathways in PSM cells taken from mouse embryos, and then showed that similar pathways govern segmentation in human PSM differentiated from iPSCs, with oscillations dependent on Notch signalling and another pathway, mediated by a protein called YAP. They found that FGF signalling not only determines the positions along the body axis at which oscillations stop, as previously reported2, but also regulates the complex dynamics of the oscillations — their period, phase and amplitude.

Matsuda and colleagues used their culture protocol to study a human genetic disease, congenital spondylocostal dysostosis, in which defects in segmentation of the vertebrae lead to skeletal anomalies13,14. The authors generated PSM from iPSCs derived from two people with the disease, who each had mutations in a different gene of the Notch signalling pathway. Surprisingly, despite these mutations and differences in overall gene expression, the authors observed normal oscillations in the PSM. By contrast, when the authors produced PSM from cells genetically engineered to carry a Hes7 mutation that had previously been identified as a cause of spondylocostal dysostosis15, they observed a dramatic loss of oscillations (Fig. 1b). This work highlights the potential of using iPSC-derived PSM to determine the relative roles of various clock components in development.

It is known that, although individual PSM cells show autonomous oscillations, Notch signalling between cell neighbours synchronizes these oscillations1,16 to produce gene-expression waves at the population level. Yoshioka-Kobayashi et al. set out to examine this role for Notch signalling in detail. The authors engineered mice to carry a Hes7–Achilles reporter, and to lack a protein called Lunatic fringe that modulates Notch signalling. They then isolated the entire PSM from embryos that lacked Lunatic fringe and from controls that did not, and made use of optogenetics, a light-triggered gene-expression system, to visualize somite development in culture by tracking Hes7 oscillations over time (Fig. 1c). Although the autonomous oscillations of single PSM cells were unaffected by loss of Lunatic fringe, the researchers observed oscillation defects at the population level.

Notch signalling involves the release of the protein DLL1 from one cell and its binding by Notch receptors on another. This interaction triggers a downstream signalling cascade in the receiving cell that causes increases in the expression of various genes, including Hes117. This sender–receiver system can be modulated using a genetically engineered optogenetic variant of the Dll1 gene that is expressed in response to stimulation by light18. The authors stimulated Dll1, and compared how long it took for neighbouring cells to exhibit Hes1 upregulation in mice lacking Lunatic fringe with the time it took in controls. The study revealed that Lunatic fringe controls population-level oscillations by regulating the timing and amplitude of the signal-sending and signal-receiving process in adjacent cells. This work underscores the intricate role of Notch components in the cell–cell interactions that control clock oscillations.

Together, the current studies provide a remarkable demonstration that simple iPSC culture systems can be used for in-depth analysis of the oscillatory gene expression associated with somite segmentation at single-cell resolution. However, they also have limitations. For instance, Diaz-Cuadros et al. and Matsuda et al. did not observe final stages of somite development and vertebra formation in their human culture systems. Nonetheless, their protocols will undoubtedly help to advance our understanding of the molecular basis of normal segmentation and to reveal the genes that, when mutated, lead to the development of disorders of the spine.

More broadly, gene-regulatory networks are highly conserved between mammals, regardless of the animals’ size or whether they are bipedal or quadrupedal. This is in stark contrast to the species-specific timing of gene oscillations, which is fundamental to body-plan development. What causes these crucial differences in timing remains an enigma — but one that can now begin to be unravelled.

Nature 580, 32-34 (2020)

doi: https://doi.org/10.1038/d41586-020-00322-y


  1. 1.

    Herrgen, L. et al. Curr. Biol. 20, 1244–1253 (2010).

    PubMed  Article  Google Scholar 

  2. 2.

    Hubaud, A. & Pourquié, O. Nature Rev. Mol. Cell Biol. 15, 709–721 (2014).

    PubMed  Article  Google Scholar 

  3. 3.

    Oates, A. C., Morelli, L. G. & Ares, S. Development 139, 625–639 (2012).

    PubMed  Article  Google Scholar 

  4. 4.

    Diaz-Cuadros, M. et al. Nature 580, 113–118 (2020).

    Article  Google Scholar 

  5. 5.

    Matsuda, M. et al. Nature 580, 124–129 (2020).

    Article  Google Scholar 

  6. 6.

    Yoshioka-Kobayashi, K. et al. Nature 580, 119–123 (2020).

    Article  Google Scholar 

  7. 7.

    Chal, J. et al. Nature Biotechnol. 33, 962–969 (2015).

    PubMed  Article  Google Scholar 

  8. 8.

    Loh, K. M. et al. Cell 166, 451–467 (2016).

    PubMed  Article  Google Scholar 

  9. 9.

    Xi, H. et al. Cell Rep. 18, 1573–1585 (2017).

    PubMed  Article  Google Scholar 

  10. 10.

    Bessho, Y. et al. Genes Dev. 15, 2642–2647 (2001).

    PubMed  Article  Google Scholar 

  11. 11.

    Hofmann, M. et al. Genes Dev. 18, 2712–2717 (2004).

    PubMed  Article  Google Scholar 

  12. 12.

    Wahl, M. B., Deng, C., Lewandoski, M. & Pourquié, O. Development 134, 4033–4041 (2007).

    PubMed  Article  Google Scholar 

  13. 13.

    Sparrow, D. B. et al. Am. J. Hum. Genet. 78, 28–37 (2006).

    PubMed  Article  Google Scholar 

  14. 14.

    Whittock, N. V. et al. Am. J. Hum. Genet. 74, 1249–1254 (2004).

    PubMed  Article  Google Scholar 

  15. 15.

    Sparrow, D. B., Guillén-Navarro, E., Fatkin, D. & Dunwoodie, S. L. Hum. Mol. Genet. 17, 3761–3766 (2008).

    PubMed  Article  Google Scholar 

  16. 16.

    Jiang, Y.-J. et al. Nature 408, 475–479 (2000).

    PubMed  Article  Google Scholar 

  17. 17.

    Jouve, C. et al. Development 127, 1421–1429 (2000).

    PubMed  Google Scholar 

  18. 18.

    Isomura, A., Ogushi, F., Kori, H. & Kageyama, R. Genes Dev. 31, 524–535 (2017).

    PubMed  Article  Google Scholar 

Download references


Nature Careers


Nature Briefing

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.

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