Nuclear instance segmentation and tracking for preimplantation mouse embryos.

For investigations into fate specification and morphogenesis in time-lapse images of preimplantation embryos, automated 3D instance segmentation and tracking of nuclei are invaluable. Low signal-to-noise ratio, high voxel anisotropy, high nuclear density, and variable nuclear shapes can limit the performance of segmentation methods, while tracking is complicated by cell divisions, low frame rates, and sample movements. Supervised machine learning approaches can radically improve segmentation accuracy and enable easier tracking, but they often require large amounts of annotated 3D data. Here we first report a novel mouse line expressing near-infrared nuclear reporter H2B-miRFP720. We then generate a dataset (termed BlastoSPIM) of 3D images of H2B-miRFP720-expressing embryos with ground truth for nuclear instances. Using BlastoSPIM, we benchmark seven convolutional neural networks and identify Stardist-3D as the most accurate instance segmentation method. With our BlastoSPIM-trained Stardist-3D models, we construct a complete pipeline for nuclear instance segmentation and lineage tracking from the 8-cell stage to the end of preimplantation development (>100 nuclei). Finally, we demonstrate BlastoSPIM's usefulness as pre-train data for related problems, both for a different imaging modality and for different model systems.

Temporal variability and cell mechanics control robustness in mammalian embryogenesis.

How living systems achieve precision in form and function despite their intrinsic stochasticity is a fundamental yet ongoing question in biology. We generated morphomaps of preimplantation embryogenesis in mouse, rabbit, and monkey embryos, and these morphomaps revealed that although blastomere divisions desynchronized passively, 8-cell embryos converged toward robust three-dimensional shapes. Using topological analysis and genetic perturbations, we found that embryos progressively changed their cellular connectivity to a preferred topology, which could be predicted by a physical model in which actomyosin contractility and noise facilitate topological transitions, lowering surface energy. This mechanism favored regular embryo packing and promoted a higher number of inner cells in the 16-cell embryo. Synchronized division reduced embryo packing and generated substantially more misallocated cells and fewer inner-cell-mass cells. These findings suggest that stochasticity in division timing contributes to robust patterning.

The cell cycle oscillator and spindle length set the speed of chromosome separation in Drosophila embryos.

Anaphase is tightly controlled in space and time to ensure proper separation of chromosomes. The mitotic spindle, the self-organized microtubule structure driving chromosome segregation, scales in size with the available cytoplasm. Yet, the relationship between spindle size and chromosome movement remains poorly understood. Here, we address how the movement of chromosomes changes during the cleavage divisions of the Drosophila blastoderm. We show that the speed of chromosome separation gradually decreases during the 4 nuclear divisions of the blastoderm. This reduction in speed is accompanied by a similar reduction in the length of the spindle, thus ensuring that these two quantities are tightly linked. Using a combination of genetic and quantitative imaging approaches, we find that two processes contribute to controlling the speed at which chromosomes move at mitotic exit: the activity of molecular motors important for microtubule depolymerization and the cell cycle oscillator. Specifically, we found that the levels of Klp10A, Klp67A, and Klp59C, three kinesin-like proteins important for microtubule depolymerization, contribute to setting the speed of chromosome separation. This observation is supported by quantification of microtubule dynamics indicating that poleward flux rate scales with the length of the spindle. Perturbations of the cell cycle oscillator using heterozygous mutants of mitotic kinases and phosphatases revealed that the duration of anaphase increases during the blastoderm cycles and is the major regulator of chromosome velocity. Thus, our work suggests a potential link between the biochemical rate of mitotic exit and the forces exerted by the spindle. Collectively, we propose that the cell cycle oscillator and spindle length set the speed of chromosome separation in anaphase.

Robust cytoplasmic partitioning by solving an intrinsic cytoskeletal instability.

Early development across vertebrates and insects critically relies on robustly reorganizing the cytoplasm of fertilized eggs into individualized cells. This intricate process is orchestrated by large microtubule structures that traverse the embryo, partitioning the cytoplasm into physically distinct and stable compartments. Despite the robustness of embryonic development, here we uncover an intrinsic instability in cytoplasmic partitioning driven by the microtubule cytoskeleton. We reveal that embryos circumvent this instability through two distinct mechanisms: either by matching the cell cycle duration to the time needed for the instability to unfold or by limiting microtubule nucleation. These regulatory mechanisms give rise to two possible strategies to fill the cytoplasm, which we experimentally demonstrate in zebrafish and Drosophila embryos, respectively. In zebrafish embryos, unstable microtubule waves fill the geometry of the entire embryo from the first division. Conversely, in Drosophila embryos, stable microtubule asters resulting from reduced microtubule nucleation gradually fill the cytoplasm throughout multiple divisions. Our results indicate that the temporal control of microtubule dynamics could have driven the evolutionary emergence of species-specific mechanisms for effective cytoplasmic organization. Furthermore, our study unveils a fundamental synergy between physical instabilities and biological clocks, uncovering universal strategies for rapid, robust, and efficient spatial ordering in biological systems.

A novel ground truth dataset enables robust 3D nuclear instance segmentation in early mouse embryos.

For investigations into fate specification and cell rearrangements in live images of preimplantation embryos, automated and accurate 3D instance segmentation of nuclei is invaluable; however, the performance of segmentation methods is limited by the images' low signal-to-noise ratio and high voxel anisotropy and the nuclei's dense packing and variable shapes. Supervised machine learning approaches have the potential to radically improve segmentation accuracy but are hampered by a lack of fully annotated 3D data. In this work, we first establish a novel mouse line expressing near-infrared nuclear reporter H2B-miRFP720. H2B-miRFP720 is the longest wavelength nuclear reporter in mice and can be imaged simultaneously with other reporters with minimal overlap. We then generate a dataset, which we call BlastoSPIM, of 3D microscopy images of H2B-miRFP720-expressing embryos with ground truth for nuclear instance segmentation. Using BlastoSPIM, we benchmark the performance of five convolutional neural networks and identify Stardist-3D as the most accurate instance segmentation method across preimplantation development. Stardist-3D, trained on BlastoSPIM, performs robustly up to the end of preimplantation development (> 100 nuclei) and enables studies of fate patterning in the late blastocyst. We, then, demonstrate BlastoSPIM's usefulness as pre-train data for related problems. BlastoSPIM and its corresponding Stardist-3D models are available at: blastospim.flatironinstitute.org.

Spiny and soft-rayed fin domains in acanthomorph fish are established through a BMP-gremlin-shh signaling network.

With over 18,000 species, the Acanthomorpha, or spiny-rayed fishes, form the largest and arguably most diverse radiation of vertebrates. One of the key novelties that contributed to their evolutionary success are the spiny rays in their fins that serve as a defense mechanism. We investigated the patterning mechanisms underlying the differentiation of median fin Anlagen into discrete spiny and soft-rayed domains during the ontogeny of the direct-developing cichlid fish Astatotilapia burtoni Distinct transcription factor signatures characterize these two fin domains, whereby mutually exclusive expression of hoxa13a/b with alx4a/b and tbx2b marks the spine to soft-ray boundary. The soft-ray domain is established by BMP inhibition via gremlin1b, which synergizes in the posterior fin with shh secreted from a zone of polarizing activity. Modulation of BMP signaling by chemical inhibition or gremlin1b CRISPR/Cas9 knockout induces homeotic transformations of spines into soft rays and vice versa. The expression of spine and soft-ray genes in nonacanthomorph fins indicates that a combination of exaptation and posterior expansion of an ancestral developmental program for the anterior fin margin allowed the evolution of robustly individuated spiny and soft-rayed domains. We propose that a repeated exaptation of such pattern might underly the convergent evolution of anterior spiny-fin elements across fishes.