DeepProjection: specific and robust projection of curved 2D tissue sheets from 3D microscopy using deep learning.

The efficient extraction of image data from curved tissue sheets embedded in volumetric imaging data remains a serious and unsolved problem in quantitative studies of embryogenesis. Here, we present DeepProjection (DP), a trainable projection algorithm based on deep learning. This algorithm is trained on user-generated training data to locally classify 3D stack content, and to rapidly and robustly predict binary masks containing the target content, e.g. tissue boundaries, while masking highly fluorescent out-of-plane artifacts. A projection of the masked 3D stack then yields background-free 2D images with undistorted fluorescence intensity values. The binary masks can further be applied to other fluorescent channels or to extract local tissue curvature. DP is designed as a first processing step than can be followed, for example, by segmentation to track cell fate. We apply DP to follow the dynamic movements of 2D-tissue sheets during dorsal closure in Drosophila embryos and of the periderm layer in the elongating Danio embryo. DeepProjection is available as a fully documented Python package.

Swimming kinematics and performance of spinal transected lampreys with different levels of axon regeneration.

Axon regeneration is critical for restoring neural function after spinal cord injury. This has prompted a series of studies on the neural and functional recovery of lampreys after spinal cord transection. Despite this, there are still many basic questions remaining about how much functional recovery depends on axon regeneration. Our goal was to examine how swimming performance is related to degree of axon regeneration in lampreys recovering from spinal cord transection by quantifying the relationship between swimming performance and percent axon regeneration of transected lampreys after 11 weeks of recovery. We found that while swimming speeds varied, they did not relate to percent axon regeneration. In fact, swimming speeds were highly variable within individuals, meaning that most individuals could swim at both moderate and slow speeds, regardless of percent axon regeneration. However, none of the transected individuals were able to swim as fast as the control lampreys. To swim fast, control lampreys generated high amplitude body waves with long wavelengths. Transected lampreys generated body waves with lower amplitude and shorter wavelengths than controls, and to compensate, transected lampreys increased their wave frequencies to swim faster. As a result, transected lampreys had significantly higher frequencies than control lampreys at comparable swimming velocities. These data suggest that the control lampreys swam more efficiently than transected lampreys. In conclusion, there appears to be a minimal recovery threshold in terms of percent axon regeneration required for lampreys to be capable of swimming; however, there also seems to be a limit to how much they can behaviorally recover.

Thrust generation during steady swimming and acceleration from rest in anguilliform swimmers.

Escape swimming is a crucial behavior by which undulatory swimmers evade potential threats. The hydrodynamics of escape swimming have not been well studied, particularly for anguilliform swimmers, such as the sea lamprey Petromyzon marinus For this study, we compared the kinematics and hydrodynamics of larval sea lampreys with those of lampreys accelerating from rest during escape swimming. We used experimentally derived velocity fields to calculate pressure fields and distributions of thrust and drag along the body. Lampreys initiated acceleration from rest with the formation of a high-amplitude body bend at approximately one-quarter body length posterior to the head. This deep body bend produced two high-pressure regions from which the majority of thrust for acceleration was derived. In contrast, steady swimming was characterized by shallower body bends and negative-pressure-derived thrust, which was strongest near the tail. The distinct mechanisms used for steady swimming and acceleration from rest may reflect the differing demands of the two behaviors. High-pressure-based mechanisms, such as the one used for acceleration from rest, could also be important for low-speed maneuvering during which drag-based turning mechanisms are less effective. The design of swimming robots may benefit from the incorporation of such insights from unsteady swimming.

How the bending kinematics of swimming lampreys build negative pressure fields for suction thrust.

Swimming animals commonly bend their bodies to generate thrust. For undulating animals such as eels and lampreys, their bodies bend in the form of waves that travel from head to tail. These kinematics accelerate the flow of adjacent fluids, which alters the pressure field in a manner that generates thrust. We used a comparative approach to evaluate the cause-and-effect relationships in this process by quantifying the hydrodynamic effects of body kinematics at the body-fluid interface of the lamprey, Petromyzon marinus, during steady-state swimming. We compared the kinematics and hydrodynamics of healthy control lampreys to lampreys whose spinal cord had been transected mid-body, resulting in passive kinematics along the posterior half of their body. Using high-speed particle image velocimetry (PIV) and a method for quantifying pressure fields, we detail how the active bending kinematics of the control lampreys were crucial for setting up strong negative pressure fields (relative to ambient fields) that generated high-thrust regions at the bends as they traveled all along the body. The passive kinematics of the transected lamprey were only able to generate significant thrust at the tail, relying on positive pressure fields. These different pressure and thrust scenarios are due to differences in how active versus passive body waves generated and controlled vorticity. This demonstrates why it is more effective for undulating lampreys to pull, rather than push, themselves through the fluid.

Notochord segmentation in zebrafish controlled by iterative mechanical signaling.

In bony fishes, patterning of the vertebral column, or spine, is guided by a metameric blueprint established in the notochord sheath. Notochord segmentation begins days after somitogenesis concludes and can occur in its absence. However, somite patterning defects lead to imprecise notochord segmentation, suggesting that these processes are linked. Here, we identify that interactions between the notochord and the axial musculature ensure precise spatiotemporal segmentation of the zebrafish spine. We demonstrate that myoseptum-notochord linkages drive notochord segment initiation by locally deforming the notochord extracellular matrix and recruiting focal adhesion machinery at these contact points. Irregular somite patterning alters this mechanical signaling, causing non-sequential and dysmorphic notochord segmentation, leading to altered spine development. Using a model that captures myoseptum-notochord interactions, we find that a fixed spatial interval is critical for driving sequential segment initiation. Thus, mechanical coupling of axial tissues facilitates spatiotemporal spine patterning.

Axial segmentation by iterative mechanical signaling.

In bony fishes, formation of the vertebral column, or spine, is guided by a metameric blueprint established in the epithelial sheath of the notochord. Generation of the notochord template begins days after somitogenesis and even occurs in the absence of somite segmentation. However, patterning defects in the somites lead to imprecise notochord segmentation, suggesting these processes are linked. Here, we reveal that spatial coordination between the notochord and the axial musculature is necessary to ensure segmentation of the zebrafish spine both in time and space. We find that the connective tissues that anchor the axial skeletal musculature, known as the myosepta in zebrafish, transmit spatial patterning cues necessary to initiate notochord segment formation, a critical pre-patterning step in spine morphogenesis. When an irregular pattern of muscle segments and myosepta interact with the notochord sheath, segments form non-sequentially, initiate at atypical locations, and eventually display altered morphology later in development. We determine that locations of myoseptum-notochord connections are hubs for mechanical signal transmission, which are characterized by localized sites of deformation of the extracellular matrix (ECM) layer encasing the notochord. The notochord sheath responds to the external mechanical changes by locally augmenting focal adhesion machinery to define the initiation site for segmentation. Using a coarse-grained mathematical model that captures the spatial patterns of myoseptum-notochord interactions, we find that a fixed-length scale of external cues is critical for driving sequential segment patterning in the notochord. Together, this work identifies a robust segmentation mechanism that hinges upon mechanical coupling of adjacent tissues to control patterning dynamics.

Superresolution microscopy reveals actomyosin dynamics in medioapical arrays.

Arrays of actin filaments (F-actin) near the apical surface of epithelial cells (medioapical arrays) contribute to apical constriction and morphogenesis throughout phylogeny. Here, superresolution approaches (grazing incidence structured illumination, GI-SIM, and lattice light sheet, LLSM) microscopy resolve individual, fluorescently labeled F-actin and bipolar myosin filaments that drive amnioserosa cell shape changes during dorsal closure in Drosophila. In expanded cells, F-actin and myosin form loose, apically domed meshworks at the plasma membrane. The arrays condense as cells contract, drawing the domes into the plane of the junctional belts. As condensation continues, individual filaments are no longer uniformly apparent. As cells expand, arrays of actomyosin are again resolved-some F-actin turnover likely occurs, but a large fraction of existing filaments rearrange. In morphologically isotropic cells, actin filaments are randomly oriented and during contraction are drawn together but remain essentially randomly oriented. In anisotropic cells, largely parallel actin filaments are drawn closer to one another. Our images offer unparalleled resolution of F-actin in embryonic tissue, show that medioapical arrays are tightly apposed to the plasma membrane and are continuous with meshworks of lamellar F-actin. Medioapical arrays thereby constitute modified cell cortex. In concert with other tagged array components, superresolution imaging of live specimens will offer new understanding of cortical architecture and function.

Identifying Key Genetic Regions for Cell Sheet Morphogenesis on Chromosome 2L Using a Drosophila Deficiency Screen in Dorsal Closure.

Cell sheet morphogenesis is essential for metazoan development and homeostasis of animal form - it contributes to developmental milestones including gastrulation, neural tube closure, heart and palate formation and to tissue maintenance during wound healing. Dorsal closure, a well-characterized stage in Drosophila embryogenesis and a model for cell sheet morphogenesis, is a remarkably robust process during which coordination of conserved gene expression patterns and signaling cascades regulate the cellular shape changes and movements. New 'dorsal closure genes' continue to be discovered due to advances in imaging and genetics. Here, we extend our previous study of the right arm of the 2nd chromosome to the left arm of the 2nd chromosome using the Bloomington deficiency kit's set of large deletions, which collectively remove 98.9% of the genes on the left arm of chromosome two (2L) to identify 'dorsal closure deficiencies'. We successfully screened 87.2% of the genes and identified diverse dorsal closure defects in embryos homozygous for 49 deficiencies, 27 of which delete no known dorsal closure gene. These homozygous deficiencies cause defects in cell shape, canthus formation and tissue dynamics. Within these deficiencies, we have identified pimples, odd-skipped, paired, and sloppy-paired 1 as dorsal closure genes on 2L that affect lateral epidermal cells. We will continue to identify novel 'dorsal closure genes' with further analysis. These forward genetic screens are expected to identify new processes and pathways that contribute to closure and links between pathways and structures already known to coordinate various aspects of closure.

Regenerative capacity in the lamprey spinal cord is not altered after a repeated transection.

The resilience of regeneration in vertebrates is not very well understood. Yet understanding if tissues can regenerate after repeated insults, and identifying limitations, is important for elucidating the underlying mechanisms of tissue plasticity. This is particularly challenging in tissues, such as the nervous system, which possess a large number of terminally differentiated cells and often exhibit limited regeneration in the first place. However, unlike mammals, which exhibit very limited regeneration of spinal cord tissues, many non-mammalian vertebrates, including lampreys, bony fishes, amphibians, and reptiles, regenerate their spinal cords and functionally recover even after a complete spinal cord transection. It is well established that lampreys undergo full functional recovery of swimming behaviors after a single spinal cord transection, which is accompanied by tissue repair at the lesion site, as well as axon and synapse regeneration. Here we begin to explore the resilience of spinal cord regeneration in lampreys after a second spinal transection (re-transection). We report that by all functional and anatomical measures tested, lampreys regenerate after spinal re-transection just as robustly as after single transections. Recovery of swimming, synapse and cytoskeletal distributions, axon regeneration, and neuronal survival were nearly identical after spinal transection or re-transection. Only minor differences in tissue repair at the lesion site were observed in re-transected spinal cords. Thus, regenerative potential in the lamprey spinal cord is largely unaffected by spinal re-transection, indicating a greater persistent regenerative potential than exists in some other highly regenerative models. These findings establish a new path for uncovering pro-regenerative targets that could be deployed in non-regenerative conditions.

Identifying Genetic Players in Cell Sheet Morphogenesis Using a Drosophila Deficiency Screen for Genes on Chromosome 2R Involved in Dorsal Closure.

Cell sheet morphogenesis characterizes key developmental transitions and homeostasis, in vertebrates and throughout phylogeny, including gastrulation, neural tube formation and wound healing. Dorsal closure, a process during Drosophila embryogenesis, has emerged as a model for cell sheet morphogenesis. ∼140 genes are currently known to affect dorsal closure and new genes are identified each year. Many of these genes were identified in screens that resulted in arrested development. Dorsal closure is remarkably robust and many questions regarding the molecular mechanisms involved in this complex biological process remain. Thus, it is important to identify all genes that contribute to the kinematics and dynamics of closure. Here, we used a set of large deletions (deficiencies), which collectively remove 98.5% of the genes on the right arm of Drosophila melanogaster's 2nd chromosome to identify "dorsal closure deficiencies". Through two crosses, we unambiguously identified embryos homozygous for each deficiency and time-lapse imaged them for the duration of closure. Images were analyzed for defects in cell shapes and tissue movements. Embryos homozygous for 47 deficiencies have notable, diverse defects in closure, demonstrating that a number of discrete processes comprise closure and are susceptible to mutational disruption. Further analysis of these deficiencies will lead to the identification of at least 30 novel "dorsal closure genes". We expect that many of these novel genes will identify links to pathways and structures already known to coordinate various aspects of closure. We also expect to identify new processes and pathways that contribute to closure.

Reducing synuclein accumulation improves neuronal survival after spinal cord injury.

Spinal cord injury causes neuronal death, limiting subsequent regeneration and recovery. Thus, there is a need to develop strategies for improving neuronal survival after injury. Relative to our understanding of axon regeneration, comparatively little is known about the mechanisms that promote the survival of damaged neurons. To address this, we took advantage of lamprey giant reticulospinal neurons whose large size permits detailed examination of post-injury molecular responses at the level of individual, identified cells. We report here that spinal cord injury caused a select subset of giant reticulospinal neurons to accumulate synuclein, a synaptic vesicle-associated protein best known for its atypical aggregation and causal role in neurodegeneration in Parkinson's and other diseases. Post-injury synuclein accumulation took the form of punctate aggregates throughout the somata and occurred selectively in dying neurons, but not in those that survived. In contrast, another synaptic vesicle protein, synaptotagmin, did not accumulate in response to injury. We further show that the post-injury synuclein accumulation was greatly attenuated after single dose application of either the "molecular tweezer" inhibitor, CLR01, or a translation-blocking synuclein morpholino. Consequently, reduction of synuclein accumulation not only improved neuronal survival, but also increased the number of axons in the spinal cord proximal and distal to the lesion. This study is the first to reveal that reducing synuclein accumulation is a novel strategy for improving neuronal survival after spinal cord injury.

Cyclic AMP promotes axon regeneration, lesion repair and neuronal survival in lampreys after spinal cord injury.

Axon regeneration after spinal cord injury in mammals is inadequate to restore function, illustrating the need to design better strategies for improving outcomes. Increasing the levels of the second messenger cyclic adenosine monophosphate (cAMP) after spinal cord injury enhances axon regeneration across a wide variety of species, making it an excellent candidate molecule that has therapeutic potential. However, several important aspects of the cellular and molecular mechanisms by which cAMP enhances axon regeneration are still unclear, such as how cAMP affects axon growth patterns, the molecular components within growing axon tips, the lesion scar, and neuronal survival. To address these points, we took advantage of the large, identified reticulospinal (RS) neurons in lamprey, a vertebrate that exhibits robust axon regeneration after a complete spinal cord transection. Application of a cAMP analog, db-cAMP, at the time of spinal cord transection increased the number of axons that regenerated across the lesion site. Db-cAMP also promoted axons to regenerate in straighter paths, prevented abnormal axonal growth patterns, increased the levels of synaptotagmin within axon tips, and increased the number of axotomized neurons that survived after spinal cord injury, thereby increasing the pool of neurons available for regeneration. There was also a transient increase in the number of microglia/macrophages and improved repair of the lesion site. Taken together, these data reveal several new features of the cellular and molecular mechanisms underlying cAMP-mediated enhancement of axon regeneration, further emphasizing the positive roles for this conserved pathway.