Skip to main content

Thank you for visiting 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.

Cooperative epithelial phagocytosis enables error correction in the early embryo


Errors in early embryogenesis are a cause of sporadic cell death and developmental failure1,2. Phagocytic activity has a central role in scavenging apoptotic cells in differentiated tissues3,4,5,6. However, how apoptotic cells are cleared in the blastula embryo in the absence of specialized immune cells remains unknown. Here we show that the surface epithelium of zebrafish and mouse embryos, which is the first tissue formed during vertebrate development, performs efficient phagocytic clearance of apoptotic cells through phosphatidylserine-mediated target recognition. Quantitative four-dimensional in vivo imaging analyses reveal a collective epithelial clearance mechanism that is based on mechanical cooperation by two types of Rac1-dependent basal epithelial protrusions. The first type of protrusion, phagocytic cups, mediates apoptotic target uptake. The second, a previously undescribed type of fast and extended actin-based protrusion that we call ‘epithelial arms’, promotes the rapid dispersal of apoptotic targets through Arp2/3-dependent mechanical pushing. On the basis of experimental data and modelling, we show that mechanical load-sharing enables the long-range cooperative uptake of apoptotic cells by multiple epithelial cells. This optimizes the efficiency of tissue clearance by extending the limited spatial exploration range and local uptake capacity of non-motile epithelial cells. Our findings show that epithelial tissue clearance facilitates error correction that is relevant to the developmental robustness and survival of the embryo, revealing the presence of an innate immune function in the earliest stages of embryonic development.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: The embryonic surface epithelium performs efficient phagocytic clearance of apoptotic cells.
Fig. 2: Apoptotic cells acquire fast motility associated with actin-based epithelial arm protrusions.
Fig. 3: Epithelial arms mechanically push apoptotic cells.
Fig. 4: Epithelial cells cooperate through active target spreading to improve the efficiency of tissue clearance.

Data availability

RNA-sequencing data and analyses are available in Supplementary Table 1, and raw data have been deposited in the Gene Expression Omnibus (GEO) under accession number GSE143734. Raw microscopy data are available from the corresponding authors upon request due to their large file sizes. Source data are provided with this paper.

Code availability

RNA-sequencing data were analysed using published processing pipelines as outlined in detail in the Supplementary Methods. Image analysis routines are described in the Supplementary Methods. Image analysis and simulation codes are available on GitHub (


  1. 1.

    Schneider, I. & Ellenberg, J. Mysteries in embryonic development: how can errors arise so frequently at the beginning of mammalian life? PLoS Biol. 17, e3000173 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  2. 2.

    Bolton, H. et al. Mouse model of chromosome mosaicism reveals lineage-specific depletion of aneuploid cells and normal developmental potential. Nat. Commun. 7, 11165 (2016).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  3. 3.

    Juncadella, I. J. et al. Apoptotic cell clearance by bronchial epithelial cells critically influences airway inflammation. Nature 493, 547–551 (2013).

    ADS  CAS  PubMed  Google Scholar 

  4. 4.

    Arandjelovic, S. & Ravichandran, K. S. Phagocytosis of apoptotic cells in homeostasis. Nat. Immunol. 16, 907–917 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  5. 5.

    Doran, A. C., Yurdagul, A. Jr & Tabas, I. Efferocytosis in health and disease. Nat. Rev. Immunol. 20, 254–267 (2020).

    CAS  PubMed  Google Scholar 

  6. 6.

    Akhtar, N., Li, W., Mironov, A. & Streuli, C. H. Rac1 controls both the secretory function of the mammary gland and its remodeling for successive gestations. Dev. Cell 38, 522–535 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  7. 7.

    Artus, J., Kang, M., Cohen-Tannoudji, M. & Hadjantonakis, A. K. PDGF signaling is required for primitive endoderm cell survival in the inner cell mass of the mouse blastocyst. Stem Cells 31, 1932–1941 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  8. 8.

    Akieda, Y. et al. Cell competition corrects noisy Wnt morphogen gradients to achieve robust patterning in the zebrafish embryo. Nat. Commun. 10, 4710 (2019).

    ADS  PubMed  PubMed Central  Google Scholar 

  9. 9.

    Clavería, C., Giovinazzo, G., Sierra, R. & Torres, M. Myc-driven endogenous cell competition in the early mammalian embryo. Nature 500, 39–44 (2013).

    ADS  PubMed  Google Scholar 

  10. 10.

    Zhong, J. X., Zhou, L., Li, Z., Wang, Y. & Gui, J. F. Zebrafish Noxa promotes mitosis in early embryonic development and regulates apoptosis in subsequent embryogenesis. Cell Death Differ. 21, 1013–1024 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  11. 11.

    Vanneste, E. et al. Chromosome instability is common in human cleavage-stage embryos. Nat. Med. 15, 577–583 (2009).

    CAS  PubMed  Google Scholar 

  12. 12.

    Greco, E., Minasi, M. G. & Fiorentino, F. Healthy babies after intrauterine transfer of mosaic aneuploid blastocysts. N. Engl. J. Med. 373, 2089–2090 (2015).

    PubMed  Google Scholar 

  13. 13.

    Elliott, M. R. & Ravichandran, K. S. The dynamics of apoptotic cell clearance. Dev. Cell 38, 147–160 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  14. 14.

    Oltval, Z. N., Milliman, C. L. & Korsmeyer, S. J. Bcl-2 heterodimerizes in vivo with a conserved homolog, Bax, that accelerates programmed cell death. Cell 74, 609–619 (1993).

    Google Scholar 

  15. 15.

    Furniss, K. L. et al. Direct monitoring of the strand passage reaction of DNA topoisomerase II triggers checkpoint activation. PLoS Genet. 9, e1003832 (2013).

    PubMed  PubMed Central  Google Scholar 

  16. 16.

    Palchaudhuri, R. et al. A small molecule that induces intrinsic pathway apoptosis with unparalleled speed. Cell Rep. 13, 2027–2036 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  17. 17.

    Vorselen, D. et al. Microparticle traction force microscopy reveals subcellular force exertion patterns in immune cell–target interactions. Nat. Commun. 11, 20 (2020).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  18. 18.

    Patel, P. C. & Harrison, R. E. Membrane ruffles capture C3bi-opsonized particles in activated macrophages. Mol. Biol. Cell 19, 4628–4639 (2008).

    CAS  PubMed  PubMed Central  Google Scholar 

  19. 19.

    Schlam, D. et al. Phosphoinositide 3-kinase enables phagocytosis of large particles by terminating actin assembly through Rac/Cdc42 GTPase-activating proteins. Nat. Commun. 6, 8623 (2015).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  20. 20.

    Jaumouillé, V., Cartagena-Rivera, A. X. & Waterman, C. M. Coupling of β2 integrins to actin by a mechanosensitive molecular clutch drives complement receptor-mediated phagocytosis. Nat. Cell Biol. 21, 1357–1369 (2019).

    PubMed  PubMed Central  Google Scholar 

  21. 21.

    Fadok, V. A. et al. Exposure of phosphatidylserine on the surface of apoptotic lymphocytes triggers specific recognition and removal by macrophages. J. Immunol. 148, 2207–2216 (1992).

    CAS  PubMed  Google Scholar 

  22. 22.

    Vieira, O. V. et al. Modulation of Rab5 and Rab7 recruitment to phagosomes by phosphatidylinositol 3-kinase. Mol. Cell. Biol. 23, 2501–2514 (2003).

    CAS  PubMed  PubMed Central  Google Scholar 

  23. 23.

    Park, H. & Cox, D. Cdc42 regulates Fc gamma receptor-mediated phagocytosis through the activation and phosphorylation of Wiskott–Aldrich syndrome protein (WASP) and neural-WASP. Mol. Biol. Cell 20, 4500–4508 (2009).

    CAS  PubMed  PubMed Central  Google Scholar 

  24. 24.

    Gold, E. S. et al. Dynamin 2 is required for phagocytosis in macrophages. J. Exp. Med. 190, 1849–1856 (1999).

    CAS  PubMed  PubMed Central  Google Scholar 

  25. 25.

    Deng, Q., Ramsköld, D., Reinius, B. & Sandberg, R. Single-cell RNA-seq reveals dynamic, random monoallelic gene expression in mammalian cells. Science 343, 193–196 (2014).

    ADS  CAS  PubMed  Google Scholar 

  26. 26.

    Petropoulos, S. et al. Single-cell RNA-seq reveals lineage and X chromosome dynamics in human preimplantation embryos. Cell 165, 1012–1026 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  27. 27.

    van Ham, T. J., Kokel, D. & Peterson, R. T. Apoptotic cells are cleared by directional migration and elmo1-dependent macrophage engulfment. Curr. Biol. 22, 830–836 (2012).

    PubMed  PubMed Central  Google Scholar 

  28. 28.

    Yamaguchi, Y. et al. Live imaging of apoptosis in a novel transgenic mouse highlights its role in neural tube closure. J. Cell Biol. 195, 1047–1060 (2011).

    CAS  PubMed  PubMed Central  Google Scholar 

  29. 29.

    Heasman, S. J. & Ridley, A. J. Mammalian Rho GTPases: new insights into their functions from in vivo studies. Nat. Rev. Mol. Cell Biol. 9, 690–701 (2008).

    CAS  PubMed  Google Scholar 

  30. 30.

    Welch, M. D., Iwamatsu, A. & Mitchison, T. J. Actin polymerization is induced by Arp2/3 protein complex at the surface of Listeria monocytogenes. Nature 385, 265–269 (1997).

    ADS  CAS  PubMed  Google Scholar 

  31. 31.

    Lämmermann, T. et al. Neutrophil swarms require LTB4 and integrins at sites of cell death in vivo. Nature 498, 371–375 (2013).

    ADS  PubMed  Google Scholar 

  32. 32.

    Zent, C. S. & Elliott, M. R. Maxed out macs: physiologic cell clearance as a function of macrophage phagocytic capacity. FEBS J. 284, 1021–1039 (2017).

    CAS  PubMed  Google Scholar 

  33. 33.

    Armitage, E. L., Roddie, H. G. & Evans, I. R. Overexposure to apoptosis via disrupted glial specification perturbs Drosophila macrophage function and reveals roles of the CNS during injury. Cell Death Dis. 11, 627 (2020).

    CAS  PubMed  PubMed Central  Google Scholar 

  34. 34.

    Evans, I. R., Ghai, P. A., Urbančič, V., Tan, K. L. & Wood, W. SCAR/WAVE-mediated processing of engulfed apoptotic corpses is essential for effective macrophage migration in Drosophila. Cell Death Differ. 20, 709–720 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  35. 35.

    Voelz, K., Gratacap, R. L. & Wheeler, R. T. A zebrafish larval model reveals early tissue-specific innate immune responses to Mucor circinelloides. Dis. Model. Mech. 8, 1375–1388 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  36. 36.

    Lee, C. S. et al. Boosting apoptotic cell clearance by colonic epithelial cells attenuates inflammation in vivo. Immunity 44, 807–820 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  37. 37.

    Serizier, S. B. & McCall, K. Scrambled eggs: apoptotic cell clearance by non-professional phagocytes in the Drosophila ovary. Front. Immunol. 8, 1642 (2017).

    PubMed  PubMed Central  Google Scholar 

  38. 38.

    Penberthy, K. K., Lysiak, J. J. & Ravichandran, K. S. Rethinking phagocytes: clues from the retina and testes. Trends Cell Biol. 28, 317–327 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  39. 39.

    Zhu, Y. et al. Migratory neural crest cells phagocytose dead cells in the developing nervous system. Cell 179, 74–89 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  40. 40.

    Wood, W. et al. Mesenchymal cells engulf and clear apoptotic footplate cells in macrophageless PU.1 null mouse embryos. Development 127, 5245–5252 (2000).

    CAS  PubMed  Google Scholar 

  41. 41.

    Ellis, S. J. et al. Distinct modes of cell competition shape mammalian tissue morphogenesis. Nature 569, 497–502 (2019).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  42. 42.

    Rasmussen, J. P., Sack, G. S., Martin, S. M. & Sagasti, A. Vertebrate epidermal cells are broad-specificity phagocytes that clear sensory axon debris. J. Neurosci. 35, 559–570 (2015).

    PubMed  PubMed Central  Google Scholar 

  43. 43.

    Fazeli, G., Stetter, M., Lisack, J. N. & Wehman, A. M. C. elegans blastomeres clear the corpse of the second polar body by LC3-associated phagocytosis. Cell Rep. 23, 2070–2082 (2018).

    CAS  PubMed  Google Scholar 

  44. 44.

    Gumienny, T. L., Lambie, E., Hartwieg, E., Horvitz, H. R. & Hengartner, M. O. Genetic control of programmed cell death in the Caenorhabditis elegans hermaphrodite germline. Development 126, 1011–1022 (1999).

    CAS  PubMed  Google Scholar 

  45. 45.

    Nakanishi, N., Sogabe, S. & Degnan, B. M. Evolutionary origin of gastrulation: insights from sponge development. BMC Biol. 12, 26 (2014).

    PubMed  PubMed Central  Google Scholar 

  46. 46.

    Silva, J. R. The onset of phagocytosis and identity in the embryo of Lytechinus variegatus. Dev. Comp. Immunol. 24, 733–739 (2000).

    CAS  PubMed  Google Scholar 

  47. 47.

    Lim, J. J., Grinstein, S. & Roth, Z. Diversity and versatility of phagocytosis: roles in innate immunity, tissue remodeling, and homeostasis. Front. Cell. Infect. Microbiol. 7, 191 (2017).

    PubMed  PubMed Central  Google Scholar 

  48. 48.

    Arienti, S., Barth, N. D., Dorward, D. A., Rossi, A. G. & Dransfield, I. Regulation of apoptotic cell clearance during resolution of inflammation. Front. Pharmacol. 10, 891 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  49. 49.

    Shinoda, H., Shannon, M. & Nagai, T. Fluorescent proteins for investigating biological events in acidic environments. Int. J. Mol. Sci. 19, 1548 (2018).

    PubMed Central  Google Scholar 

  50. 50.

    Delamarre, L., Pack, M., Chang, H., Mellman, I. & Trombetta, E. S. Differential lysosomal proteolysis in antigen-presenting cells determines antigen fate. Science 307, 1630–1634 (2005).

    ADS  CAS  PubMed  Google Scholar 

Download references


We thank T. Graf, B. Lehner and V. Malhotra for discussions and reading of the manuscript. Q.T.-R. acknowledges a grant funded by ‘The Ministerio de Ciencia, Innovación y Universidades and Fondo Social Europeo (FSE)’ (PRE2018-084393). M.I. acknowledges funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (ERC-StG-LS2-63759) and the Spanish Ministry of Economy and Competitiveness (BFU2014-55076-P). S.W. acknowledges support from the Government of Spain (MINECO’s Plan Nacional (PGC2018-098532-A-I00), Severo Ochoa (CEX2019-000910-S)), Fundació Cellex, Fundació Mir-Puig and Generalitat de Catalunya (CERCA, AGAUR). V.R. acknowledges support from the Spanish Ministry of Science and Innovation to the EMBL partnership, the Centro de Excelencia Severo Ochoa, MINECO’s Plan Nacional (BFU2017-86296-P), Generalitat de Catalunya (CERCA) and support from the CRG Core Facilities for Genomics, Biomolecular Screening and Protein Technologies, Flow Cytometry and Advanced Light Microscopy.

Author information




E.H. designed research, performed experiments and analysed data. H.-M.H. contributed to in vivo experiments and data analysis. Q.T.-R. contributed to in vitro experiments. S.J.-D. supported cloning and mRNA injections. C.W. performed sequencing analysis. M.M.-C. contributed to mouse blastocyst experiments. M.I. supervised C.W. and M.M.-C. and performed sequencing analysis. A.C.-J. performed theoretical modelling. S.W. analysed data and performed Monte Carlo simulations. V.R. designed research, performed data analysis and supervised the project. E.H. and V.R. wrote the manuscript with contributions from all of the authors.

Corresponding authors

Correspondence to Esteban Hoijman, Stefan Wieser or Verena Ruprecht.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Peer review information Nature thanks Jean-Léon Maître 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 Endogenous and induced cell death events occur by apoptosis and are cleared by the surface epithelium.

a, Endogenous cell (white arrowhead) undergoing shape loss (magenta arrowheads), fragmentation (yellow arrowheads) and PS exposure (27–40 min) as detected by ubiquitous expression of a secreted version of annexin-V–YFP. b, An endogenous cell captured during abnormal mitosis (35–38 min), apoptosis (99 min) and its phagocytic uptake (379 min). c, Caspase-3 activation dynamics in an endogenously dying cell (yellow arrowheads) visualized by expression of the Flip-Casp3–GFP reporter. White arrowhead indicates uptake by an epithelial cell. d, Bax+ cells co-expressing Lyn–tdTomato (PM; white arrowheads point to example Bax+ cells) expose PS (magenta arrowheads) as detected by injection of annexin-V–AlexaFluor488 in a live embryo. Yellow arrowheads highlight the co-localization of both signals. e, In vivo detection of Caspase-3 activation by expression of the Flip-Casp3–GFP reporter in Bax+ cells co-expressing H2A–mcherry (nucleus). Yellow arrowheads indicate cells showing both signals. e′, Caspase-3 activation dynamics during nuclear fragmentation in a Bax+ cell. f, Mosaic overexpression of tamoxifen-activated Caspase-8ERT2 leads to epithelial uptake (arrowheads) similar to the one observed for Bax+ cells. g, Engulfment of apoptotic particles by the EVL or by progenitor cells in the embryo interior at 6 hpf (n = 6 embryos). Box plot shows the maximum and minimum (whiskers), medians (lines), 25th and 75th percentiles (boxes) and means (crosses). Paired two-sided t-test; ****P = 1.04 × 10−7. hn, Characteristic morphological changes in apoptotic cells (hj, white arrowheads) and phagocytic uptake by epithelial cells (kn, yellow arrowheads) of cells overexpressing a mutant topoisomerase II (TopoII-mut+) and co-expressing Lyn–tdTomato and H2A–mCherry (h, k), transplanted UV-irradiated cells expressing Lyn-tdTomato (i, l, n) and cells from Raptinal-treated embryos (j, m). In n, a small number of UV-irradiated apoptotic cells were transplanted, the white arrowhead indicates a dying cell that is later ingested by the epithelium as indicated by the yellow arrowheads. Dashed lines indicate the position of the transverse sections. o, Induced programmed cell death by two-photon illumination in a single progenitor cell (arrowheads) leading to nuclear fluorescence bleaching (0 min), cell death (120 min), cell fragmentation (127 min) and epithelial uptake (135 min). Embryos are Tg(actb1:Lifeact-GFP) in (b, d, f, h, i, k, l, n) and Tg(actb1:Myl12.1-eGFP) in (j, m, o). Lyn–tdTomato (PM marker; ac, j, m, o) and H2A–mCherry (nucleus; b, j, m, o) were co-expressed in the whole embryo. Scale bars, 20 μm (a, b, e, f, h, m, o), 10 μm (c), 50 μm (d) and 100 μm (n).

Source data

Extended Data Fig. 2 Phagocytic dynamics in epithelial cells.

a, A phagocytic event of an apoptotic fragment visualized by slices parallel (top, xy) and transverse (bottom, xz) to the surface, in embryos with mosaic Lifeact–GFP staining of EVL cells. Arrowheads point to the formation of a phagocytic cup by the extension of characteristic actin-rich pseudopods. After engulfment is completed, the height (z-dimension) of the epithelial cell remains increased due to the size of the ingested particle. bb‴, F-actin accumulates in a phagocytic ring (b, arrowheads) depleting the cytosolic actin pool. As phagocytosis is being completed (finalized at 800 s), actin replenished in the cytosol (b′–b‴). Quantification of the cytosolic actin pool dynamics is depicted on the bottom. Single planes (b) and z-projections (b′–b‴) are shown. c, Shape deformation of an apoptotic cell during epithelial uptake (left) indicated by curvature vectors normal to the cell perimeter (centre) and corresponding normalized scalar bending values (right). d, Aspect ratio of the apoptotic cell before, during and after uptake (n = 6 cells). Paired t-test with Bonferroni adjusted P values for multiple comparisons; **P = 0.0048; ##P = 0.0047. e, A region of an epithelial cell (dashed line) showing membrane ruffles (arrowheads) with actin accumulations in the front of the ruffles. f, Phagocytic cup formation leading to successful target uptake (23–36 min) originating from a membrane ruffle (17 min). Arrowheads indicate membrane ruffles. g, Coordinated apoptotic cell fragmentation and phagocytosis by different epithelial cells. A single apoptotic cell (green arrowhead) is contacted by an actin protrusion (orange arrowhead) splitting it into two particles (magenta and yellow arrowheads, 2 min). Whereas the left particle (magenta arrowhead) is rapidly phagocytosed by one epithelial cell (from 4.5 min to 7.5 min), the other particle (yellow arrowhead) is only later contacted by another epithelial cell (7.5 min) and then ingested (from 9 min to 13 min). White arrowhead indicates a previously phagocytosed apoptotic cell. h, Mosaic EVL-specific expression of dnRac1–GFP driven by the krt18 promoter in embryos with mosaic expression of Bax+ cells. dnRac1–GFP epithelial cells do not ingest apoptotic particles, whereas control neighbouring epithelial cells do (arrowheads). i, Rac1 involvement in epithelial phagocytic uptake analysed in embryos with two mosaic cell populations: (1) cells expressing H2A-mCherry only (top) or co-expressing dnRac1 (bottom) and (2) Bax+ cells. White arrowheads indicate apoptotic fragments ingested in the EVL, showing that epithelial cells expressing dnRac1 are not able to phagocytose particles compared to control cells. Dashed lines indicate EVL cell contours. Yellow arrowheads indicate red nuclear staining. Quantification of uptake rates (right; n = 6 and 7 control and dnRac1 embryos, respectively). Red bar indicates the mean. j, k, Epithelial phagocytosis of Bax+ cells in the presence of the PI3K inhibitor LY294 or a DMSO vehicle control (n = 8 and 6 embryos for DMSO and LY294, respectively). Unpaired two-sided t-test; **** P = 3.34 × 10−8. l, m, Two independent double mosaic stainings enable the identification of the origin of membranes in phagocytic vesicles. In l, Bax+ cells express Lifeact–GFP and the epithelial cell expresses Lyn–tdTomato (PM); in m, Bax+ cells express Lyn–tdTomato (PM) and the epithelial cell expresses Lifeact–GFP. n, Dynamics of superoxide generation in epithelial phagosomes (top) and quantification in a single phagosome (bottom) using the DHE reporter. or, Degradation of the phagocytic vesicle content measured by double mosaic staining. Fluorescence signals of apoptotic cells expressing Lifeact–GFP (o) or Lyn–tdTomato (pr) were monitored after phagocytosis (epithelial cells express Lyn–tdTomato (o) or Lifeact–GFP (p)). o′, GFP fluorescence decreases over time (n = 7 phagosomes from 7 epithelial cells), known to be associated to both pH-dependent quenching and protein degradation49,50. Data are mean ± s.e.m. pr, Change in spatial distribution of Lyn–tdTomato fluorescence, indicative of lytic activity degrading (or releasing the protein from) the plasma membrane of the ingested apoptotic cell (DsRed-derived fluorescent proteins are resistant to degradation by lysosomal enzymes49). r, Quantification of the coefficient of variation of the tdTomato signal in the whole phagosome over time (each line corresponds to a single phagosome in 6 epithelial cells). Redistribution begins 55.3 ± 9.5 min after internalization and takes 24.8 ± 3.4 min to be completed. s, The Itga5–GFP protein decorates epithelial phagosomes. Box plots (d, k) show the maximum and minimum (whiskers), medians (lines), 25th and 75th percentiles (boxes) and means (crosses). Embryos are Tg(krt18:Gal4FF) injected with a Tol2-UAS-Lifeact-GFP plasmid (ac, eg) or with a Tol2-UAS-dnRac1-GFP plasmid (h), Tg(actb1:Lifeact-GFP) (i, j, n) and Tg(actb1:Lifeact-RFP) (s). Scale bars, 20 μm (a, bb‴, fj), 10 μm (lp, s) and 5 μm (c, e, q).

Source data

Extended Data Fig. 3 Molecular pathways involved in epithelial phagocytosis.

a, b, Representative images (a) and quantification (b) of epithelial clearance in annexin-V–AlexaFluor488-injected embryos. Arrowheads indicate Bax+ cell fragments co-labelled with Lyn–tdTomato (PM) and annexin-V–AlexaFluor488 (overlay in white) at 6 hpf (n = 6 and 8 embryos for control and annexin-V–AlexaFluor488, respectively). Unpaired two-sided t-test; ****P = 1.84 × 10−5 (b). c, Epithelial clearance of surrogate apoptotic targets formed by lipid aggregates containing PS, PC and TexasRed–DHPE (PS+ targets) injected into the embryo. Inset shows a magnified view of a PS+ surrogate apoptotic target. d, Dynamics of a single phagocytic event of a PS+ target. Note the formation of the actin ring in the phagocytic cup (3.5 min and 4.5 min) and actin accumulation at the phagocytic cup closure site (5 min) during particle uptake (arrowheads). Slices parallel (top, xy) and transverse (bottom, xz) to the surface are shown. e, PS surrogate targets are not cleared. f, Death rate of embryos injected at 4 hpf with a large number of surrogate targets (n = 250 and 257 embryos for PS+ and PS, respectively, from 3 independent experiments). Paired two-sided t-test; *P = 0.0159 (8 h), #P = 0.013 (24 h) compared to dead embryos (percentage PS+ or PS). g, h, Epithelial phagocytosis of apoptotic UV-irradiated human Jurkat T cells transplanted into a zebrafish embryo. BODIPY-FL–C5–Ceramide membrane (g) and Caspase-3 chemical reporter (h) staining of transplanted apoptotic Jurkat T cells, showing the process of internalization (h). i, j, Top enriched gene ontology terms for zebrafish EVL cells versus progenitor cells (i; 3 independent experiments with 1 × 104 cells analysed) and mouse trophoblast cells versus the inner cell mass (j), showing raw P values for biological process (left), cellular component (middle) and KEGG pathway (right). Blue bars represent immune-related functions and orange bars phagocytosis-related terms. Embryos are Tg(actb1:Lifeact-GFP) (a, c, e), Tg(krt18:Gal4FF/UAS:Lifeact-GFP) (d), Tg(actb1:Lifeact-RFP) (g, h). Scale bars, 50 μm (a, c, e), 10 μm (a, c, insets) and 20 μm (d, g, h). Box plots show the maximum and minimum (whiskers), medians (lines), 25th and 75th percentiles (boxes) and means (crosses).

Source data

Extended Data Fig. 4 Conservation of epithelial phagocytosis of apoptotic targets in mouse embryos.

a, Self-organizing blastula cell aggregates recapitulate the de novo formation of a surface epithelium with phagocytic capacity. Maximum z-projection (top) and single slice (bottom) are shown, revealing epithelial-specific krt18 promoter activity, squamous epithelial morphology, junction formation in surface cells, and their ability to engulf particles (inset). Cells were obtained from Tg(krt18:krt18-GFP) embryos co-expressing Lyn–tdTomato (PM). b, b′, Time-lapse imaging of DsRed-expressing blastocysts. b, A cell from the inner cell mass (dotted line) is released to the cavity (0 min), phagocytosed (50 and 60 min) and remains inside the trophoblast cell during blastocyst expansion (255 min). Note the deformation of the particle during uptake similar to the one occurring during phagocytosis in the zebrafish epithelium (insets). b′, A second example of the same process described in b. c, Injection of annexin-V-AlexaFluor488 into the cavity of unperturbed live blastocysts showing an apoptotic cell. d, Dynamics of apoptotic mouse embryonic stem (ES) cells injected into the blastocyst cavity and their localization inside trophoblast (arrowheads) 30 h after injection. A detailed view of the uptake process is shown in d′: an ES cell appears inside an epithelial cell (arrowhead), similar to epithelial uptake in zebrafish embryos. eg, Representative images (e) and quantitative analysis (f, g) of the final localization of live or apoptotic mouse ES cells injected into the blastocyst cavity. f, Most live mouse ES cells were incorporated into the inner cell mass (n = 16 and 23 embryos for live and apoptotic samples, respectively). g, Apoptotic mouse ES cell fragments were mainly found inside the trophoblast (yellow arrowheads, n = 23 embryos). Mouse ES cells express glycosylphosphatidylinositol (GPI)–GFP (plasma membrane) and histone 2B (H2B)–mCherry (nucleus). In f, g, bars represent the means. f, Unpaired two-sided t-test; ****P > 1 × 10−15. g, Paired Friedman test and Dunn’s multiple comparison test; ****P = 4.97 × 10−7; **P = 0.0096. h, PS surrogate targets (glass spheres coated with PC and TexasRed–DHPE) are not ingested by the trophoblast and remain in the blastocyst cavity. Scale bars, 25 μm (a), 10 μm (a, inset), 20 μm (b, b′, c, d, d′, h) and 30 μm (e).

Source data

Extended Data Fig. 5 F-actin dynamics in epithelial arm protrusions.

a, b, External F-actin localizes to the rear part of the apoptotic cell. Arrowheads represent the direction of movement (a). External F-actin distribution along the normalized apoptotic cell perimeter (b; x = 0.5 corresponds to the rear of the apoptotic cell, n = 15 cells from 7 embryos). c, Lifeact–GFP expression driven by the krt18 promoter to visualize F-actin specifically in the epithelial tissue. z-projection (top), single-slice z-section (middle) and transverse sections at the location of the dashed lines (bottom and lateral) are shown. An apoptotic particle (arrowheads) starts to move in close association with epithelial actin enrichment. d, Epithelial arm protrusion retracting after extension (related to Fig. 2i). ee″, F-actin enrichment is evident at the tip of the epithelial protrusion, where the protrusion contacts the apoptotic cell. e′, e″, Quantification of Lifeact–GFP fluorescence intensity along the yellow line. f, F-actin flow in the external actin accumulation in contact with the apoptotic cell rear. The kymograph (right) was created from the yellow box. The arrow indicates the direction of movement of the apoptotic particle. Embryos are Tg(actb1:Lifeact-GFP) (a, f), Tg(krt18:Gal4FF/UAS:Lifeact-GFP) (c). Mosaic Lifeact–GFP was used in d and e. Scale bars, 20 μm (ae′) and 5 μm (f).

Source data

Extended Data Fig. 6 Apoptotic motility and epithelial arms are present in various apoptotic contexts.

a, Endogenous apoptotic fragment (arrowhead) revealing fast motility. Yellow line indicates the total apoptotic path length over the indicated time period. The path of a live cell (white dots and line) is shown for comparison. b, Epithelial arm formation in contact with an endogenous apoptotic cell. F-actin (yellow arrowheads) accumulates in an epithelial cell (white arrowheads) in contact with an apoptotic cell (magenta arrowhead) after its nuclear fragmentation. Epithelial actin (yellow arrowheads) then polarizes to the rear of the apoptotic cell, which starts to move, 10 min after the initial contact (32.5 min). ch, Motility of Caspase-8ERT2 (ce) or TopoII-mut+ (fh) apoptotic cells. Representative images of actin-rich arms in association with apoptotic cells (c, f), representative xy trajectories (d, g) and apoptotic cell speed (e, h; Caspase-8+ cells: vmean = 7.5 ± 0.3 μm min−1, vmax = 25.5 ± 1.1 μm min−1; n = 16 apoptotic cells from 3 embryos; TopoII-mut+ cells: vmean = 6.8 ± 0.1 μm min−1, vmax = 21.3 ± 1.3 μm min−1; n = 15 cells from 4 embryos). Box plots show the maximum and minimum (whiskers), medians (lines), 25th and 75th percentiles (boxes) and means (crosses). Data are mean ± s.e.m. i, UV-irradiated human Jurkat T cells show motility and association with epithelial arms. j, UV-irradiated and transplanted cells expressing Lyn–tdTomato showing cell death (yellow arrowhead), fragmentation (white arrowheads) and motility (yellow tracks). Embryos are Tg(actb1:Myl12.1-eGFP) (a), Tg(actb1:Lifeact-GFP) (b, c, f, j) and Tg(actb1:Lifeact-RFP) (i). Embryos (a, b) or TopoII-mut+ cells (f) co-express Lyn–tdTomato (PM) and H2A–mCherry (nucleus); in c, Caspase-8 cells co-express Lyn–tdTomato (PM); in i, membranes of Jurkat cells were stained with BODIPY-FL–C5–Ceramide before transplantation. Scale bars, 20 μm (ac, i), 10 μm (f) and 50 μm (j).

Source data

Extended Data Fig. 7 Possible factors underlying apoptotic motility.

a, b, In vivo high-speed single-cell imaging of F-actin and myosin-II dynamics in motile apoptotic cells. a, Apoptotic cells show a static internal F-actin distribution (arrowhead) during cell movement. Kymograph on the right was generated from the yellow box region; blue arrowhead highlights the static actin localization during apoptotic cell movement (progression of cell front and rear is indicated by black arrowheads). b, Apoptotic cells show a characteristic accumulation of myosin-II in static rods. c, d, Analysis of apoptotic cell speed (c) and phagocytic uptake (d) in Bax+ cells co-expressing the indicated dominant negative proteins (c, n = 28, 19, 43, 23 cells from 3 control, 3 dnRac1, 5 dnRhoA, 3 dnRock2a embryos, respectively; d, n = 11 control, 11 dnRac1, 9 dnRhoA, 9 dnRock2a embryos, respectively). e, f, Apoptotic cell motility in annexin-V-injected embryos. Bax+ cells co-expressing Lyn–tdTomato (PM, left) and co-labelled with annexin-V–AlexaFluor488 (right, overlay in white) are shown. Yellow paths correspond to a period of 30 min of tracking (n = 15 and 20 cells from 3 and 5 embryos for control and annexin-V–AlexaFluor488). g, Actin arms associated with apoptotic motility in embryos with Lifeact–GFP expressed specifically in the EVL tissue. Yellow line shows the total apoptotic cell movement path over the indicated time period. h, Angular histogram of apoptotic particle movement directions (tlag = 60 s) measured as the angular difference relative to the peak of F-actin density localization from epithelial arm protrusions (oriented to −270°, n = 15 cells, 6 embryos). The histogram indicates that whenever an apoptotic particle undergoes directional movement the F-actin (EVL specific labelling) accumulates in contact with the back of the apoptotic fragment. i, j, Cell motility analysis in an in vitro co-culture of apoptotic (Bax+) and live cells in vitro. Bax+ cells were tracked before (magenta arrowheads) and after apoptosis (yellow arrowheads). White lines show examples of apoptotic tracks (n = 12 and 14 cells for live and apoptotic phases). k, l, Analysis of epithelial phagocytic uptake under conditions of dnRac1 expression in the whole embryo (n = 4 embryos). Red bar indicates the mean. Arrowheads indicate apoptotic cells. Embryos (and cells obtained thereof) are Tg(actb1:Lifeact-GFP) (a, e, i, k), Tg(actb1:Myl12.1-eGFP) (b) and Tg(krt18:Gal4FF/UAS:Lifeact-GFP) (g). Scale bars, 20 μm (a, b, g) and 50 μm (e, i, k). Box plots show the maximum and minimum (whiskers), medians (lines), 25th and 75th percentiles (boxes) and means (crosses). Unpaired Kruskal–Wallis test, Dunn multiple comparison test; P > 0.9999 (not significant) (c, control (ctrl) versus dnRac1 and dnRhoA; d, control versus dnRac1, dnRhoA and dnRock2a), P = 0.4127 (c, control versus dnRock2a). Unpaired two-sided t-test; ****P < 1 × 10−15 (f); ***P = 1.2 × 10−5 (j).

Source data

Extended Data Fig. 8 Simulations and modelling of apoptotic dispersal and epithelial uptake.

a, Representative images showing the transplantation of Bax+ cells into a Tg(actb1:Lifeact-GFP) embryo. White lines outline the areas of cell dispersion over time. Yellow lines indicate apoptotic cell trajectories. The differential increase in the effective spreading area of apoptotic cells is indicated on the right. b, Schematic illustration of the experimental condition (top) showing a local source of apoptotic targets (red) underneath an epithelial layer (EVL). Individual epithelial cells can form two types of basal protrusions, phagocytic cups (formed at a rate pup) mediating target ingestion, and epithelial arms (formed at a rate pv), which actively push apoptotic targets with velocity v and induce target dispersal underneath the epithelial layer. Each epithelial cell can engulf a maximum number of targets Nmax. Schematic of the Monte Carlo simulation (bottom), and the minimal set of parameters included therein (right). c, d, Kymographs obtained from Monte Carlo simulations of apoptotic target cell spreading (bottom) and epithelial uptake (top) across a one-dimensional epithelial layer (x) over time (t) for two conditions: (1) high pushing speed v and low maximum uptake capacity Nmax per epithelial cell (v = 15 μm min−1; Nmax = 5) (c) or (2) lower pushing rate and higher uptake capacity (v = 3 μm min−1; Nmax = 10) (d). Scale bars, 25 min (t) and 100 μm (x). e, Plot of clearance time versus pushing velocity for variable Ntarget numbers with Nmax = 4 (Supplementary equation (1), Supplementary Note 1). f, Clearance time as a function of pushing speed v for various ratios of Ntarget/Nmax. The clearance time varies with 1/v2 and (Ntarget/Nmax)2 (Supplementary equation (1), Supplementary Note 1). g, Tissue clearance time as a function of pushing speed v for various clearance efficiencies. The analytically derived clearance time (Supplementary equation (1), Supplementary Note 1) matches the clearance time from Monte Carlo simulations at approximately 98% clearance efficiency. Scale bar, 50 μm (a).

Extended Data Fig. 9 In vivo analysis of epithelial uptake efficiency.

a, Annexin-V–AlexaFluor488 staining of Raptinal-treated (or control) Jurkat cells before transplantation (percentage of annexin-V+ cells: 97.4 ± 0.4% for Raptinal treatment versus 10.3 ± 0.8% for control conditions; n = 120 cells per sample). b, Target localization in the embryo interior and uptake by the EVL after transplantation of apoptotic Jurkat cells (left), at the moment of reaching saturation (centre) and 1 h after saturation (right). Inset highlights the morphology of a non-fragmented apoptotic Jurkat cell. c, Representative images of a single EVL cell sequentially ingesting three apoptotic targets (yellow arrowheads indicate the uptake of a new particle). In the transverse section, the presence of ingested (magenta arrowhead) and non-ingested (white arrowhead) targets is highlighted. The section underneath the EVL shows multiple non-ingested targets in contact with the saturated EVL cell (white polygon). d, Mean number of ingested apoptotic targets per EVL cell over time (n = 120 cells from 6 embryos). e, Quantification of the external apoptotic targets that were not ingested and remained in contact with single EVL cells at or 1 h after saturation (n = 69 cells (at saturation) and 58 cells (1 h after saturation) from 6 embryos). Box plots show the maximum and minimum (whiskers), medians (lines), 25th and 75th percentiles boxes) and means (crosses). Unpaired two-sided t-test; n.s., P = 0.9629. f, Neighbouring epithelial cells (cyan) in the vicinity of saturated epithelial cells (orange) are still able to ingest apoptotic targets (green). g, Mean volume of apoptotic targets ingested per EVL cell over time (n = 120 cells from 6 embryos). h, Distribution of the total volume ingested by single EVL cells at saturation (n = 6 embryos). In b, c, f, Jurkat cells are stained with FM 4-64FX and embryos are Tg(actb1:Lifeact-GFP). Scale bars, 20 μm (a, c, h), 50 μm (b) and 5 μm (b, inset). Data are mean ± s.e.m. (open circles (d, g) and bars (h)) from 6 embryos (red dots (d, g) and circles (h)).

Source data

Extended Data Fig. 10 Experimental schemes.

Details of experimental staining and handling procedures and the corresponding figure panels.

Supplementary information

Supplementary Information

This file contains Supplementary Table 1, Supplementary Figure 1, Supplementary Methods, Supplementary Note 1, Supplementary Video Legends 1-14 and Supplementary References.

Reporting Summary

Supplementary Table 1

Differential gene expression for zebrafish EVL versus progenitor cells and mouse trophoblast versus inner cell mass cells1. Gene Ontology (GO) analysis for both species is also shown. Published data2 from human GO analysis are also included.

Video 1

Single cell dynamics of endogenous apoptotic events and epithelial uptake. The enveloping epithelial layer (EVL) uptakes apoptotic cells as identified by morphological features (I, yellow lines indicate apoptotic cell fragment paths), Annexin-V-YFP (II, cyan) binding, or activated Capase-3 (III, cyan; yellow and white lines indicate apoptotic and live cell paths, respectively). Myosin II-EGFP (I, cyan), plasma membrane (PM, Lyn-tdTomato, red, I,II,III) and nucleus (H2A-mCherry, red, I) are shown. Yellow arrowheads indicate apoptotic cells. White, magenta and green arrowheads indicate apoptotic fragments before and after epithelial uptake (note that fragments from a single apoptotic cell are uptaken by different epithelial cells).

Video 2

Errors in cell division (I), associated to cell death (II), and subsequent epithelial uptake (III). Lifeact-GFP (I and II, cyan), Myosin II-EGFP (III, cyan), plasma membrane (Lyn-tdTomato, red, I,II,III) and nucleus (H2A-mCherry, red, I, II, III) are shown.

Video 3

EVL clearance dynamics of Bax+ apoptotic cells (I, co-expressing the PM marker LyntdTomato, red). Embryo expresses Lifeact-GFP (cyan). (II) Dynamics of Caspase-3 reporter activation (cyan) in Bax+ cells during nuclear (H2A-mCherry, red) fragmentation.

Video 4

Epithelial clearance under stress conditions. Mosaic Caspase-8 activation (I), mosaic mutant Topo-II expression (II), whole embryo Raptinal treatment (III) and single cell 2-photon induced programmed cell death (IV). PM (Lyn-tdTomato, red) alone or together with nucleus (H2A-mCherry, red) in apoptotic cells (I, II) or in the whole embryo (III, IV), F-actin (Lifeact-GFP, cyan, I, II) and Myosin II-EGFP (cyan, III and IV) are shown.

Video 5

Single cell visualization of epithelial phagocytosis of Bax+ cells. (I) Two examples of zprojections of phagocytic cups showing an actin ring, and single slices with sections parallel and transverse to the epithelial surface showing the dynamics of phagocytic pseudopods. (II) Phagocytic cup formed from epithelial ruffles. F-actin (Lifeact-GFP, cyan) and PM of apoptotic cells (Lyn-tdTomato, red) are shown (I, II). (III) Ruffles in an embryo expressing Lifeact-GFP (cyan) and showing Bax+ apoptotic cells (PM, Lyn-tdTomato, red). (IV) Degradation of the apoptotic debris after uptake. Double mosaic staining showing F-actin (Lifeact-GFP, cyan) in an epithelial cell and PM of apoptotic cells (Lyn-tdTomato, red).

Video 6

Epithelial clearance of PS+ surrogate apoptotic targets. Global (I, parallel and transverse sections) or high magnification epithelial-specific staining (II) for F-actin (Lifeact-GFP, cyan) are shown. PM of Bax+ cells is shown (Lyn-tdTomato, red).

Video 7

Mouse trophoblast ingests apoptotic particles. Uptake of an apoptotic cell released from the inner cell mass (I), uptake of an in vivo stained apoptotic cell in the cavity by Annexin-V-AlexaFluor488 injection (cyan, II), and uptake detail of an apoptotic ES cell injected into the cavity (III). Blastocysts express DsRed (red), and mES cells are stained with Bodipy-FL-C5 ceramide (cyan).

Video 8

In vivo apoptotic motility in association with actin arms. Whole embryo (I) or detailed view (II) of Bax+ cells (co-expressing Lyn-tdTomato, red) moving fast and non-directionally along the embryo in association with external F-actin (Lifeact-GFP, cyan). An embryo with mosaic PM (Lyn tdTomato) staining of motile live cells with slow and directional movement in comparison to apoptotic cells is shown on the right (I). Tracks are depicted by yellow (apoptotic cells) or white (live cells) lines.

Video 9

High speed imaging of in vivo actin filament dynamics (Lifeact-GFP, cyan) in association with the external rear surface of single Bax+ apoptotic cells (PM, Lyn-tdTomato, red).

Video 10

Dynamics of epithelial arms. (I) Single cell epithelial arm visualization (Lifeact-GFP, cyan) in association with a Bax+ apoptotic cell (PM, Lyn-tdTomato, red).

Video 11

Arm-mediated motility of endogenous apoptotic cells. (I) Epithelial actin contacting a dead cell and polarizing to its rear to initiate motility and (II) Phagocytic uptake and epithelial arms associated to motility of an endogenous cell dying after an error in mitosis. Lifeact-GFP (cyan), PM (Lyn-tdTomato, red), and H2A-mCherry (nucleus, red) are shown.

Video 12

Epithelial arm dynamics in multiple stress conditions. (I) Epithelial arms (Lifeact-RFP, red) formed in contact with apoptotic Jurkat T cells (Bodipy-FL-C5-Ceramide, cyan). (II) Apoptotic motility after transplantation of few UV-irradiated cells (Lifeact-GFP, cyan; PM, Lyn-tdTomato, red, in apoptotic cells). (III) In vivo single cell Myosin II dynamics (Myosin II-EGFP, cyan) in a single Bax+ apoptotic cell (PM, Lyn-tdTomato, red). (IV) Epithelial specific actin (cyan) associates with motile Bax+ apoptotic cells (PM, Lyn-tdTomato, red). Embryos are Tg(Krt18:Gal4FF/UAS:Lifeact-GFP).

Video 13

Epithelial arms propel apoptotic targets. (I) Loss of epithelial arm protrusions and apoptotic cell motility (Bax and Lyn-tdTomato, red plasma membrane) in dnRac1-expressing (H2A-mCherry, red nucleus) embryos. F-actin is shown (Lifeact-GFP, cyan). (II) Epithelial arms (Lifeact-GFP, cyan) push PS+ surrogate targets (Texas Red-DHPE, red). (III) CK-666- mediated Arp2/3 inhibition blocks apoptotic motility, arm formation, and phagocytic uptake. (IV) SMIFH2-mediated Formin inhibition blocks phagocytic uptake but does not affect epithelial arm formation and apoptotic motility. In III and IV embryos express Lifeact-GFP (cyan) and apoptotic cells express Bax and Lyn-tdTomato (red).

Video 14

Apoptotic spreading extends the pool of epithelial phagocytes. (I) Consecutive pushing of an apoptotic cell by arms coming from different epithelial cells. F-actin (cyan) and a Bax+ cell (co-expressing Lyn-tdTomato, red) are shown. (II) Tracking of Bax+ apoptotic cells (co-expressing Lyn-tdTomato, yellow dots) after transplantation into a localized region underneath the EVL of an acceptor Tg(actb1:Lifeact-GFP) embryo. A pool of epithelial cells initially in close proximity to the transplanted cells are depicted in blue, and cells for the expanded epithelial area after apoptotic spreading are shown in magenta.

Source data

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Hoijman, E., Häkkinen, HM., Tolosa-Ramon, Q. et al. Cooperative epithelial phagocytosis enables error correction in the early embryo. Nature 590, 618–623 (2021).

Download citation

Further reading


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.


Quick links