MiCellAnnGELo: annotate microscopy time series of complex cell surfaces with 3D virtual reality.

SUMMARY: Advances in 3D live cell microscopy are enabling high-resolution capture of previously unobserved processes. Unleashing the power of modern machine learning methods to fully benefit from these technologies is, however, frustrated by the difficulty of manually annotating 3D training data. MiCellAnnGELo virtual reality software offers an immersive environment for viewing and interacting with 4D microscopy data, including efficient tools for annotation. We present tools for labelling cell surfaces with a wide range of applications, including cell motility, endocytosis and transmembrane signalling. AVAILABILITY AND IMPLEMENTATION: MiCellAnnGELo employs the cross-platform (Mac/Unix/Windows) Unity game engine and is available under the MIT licence at https://github.com/CellDynamics/MiCellAnnGELo.git, together with sample data. MiCellAnnGELo can be run in desktop mode on a 2D screen or in 3D using a standard VR headset with a compatible GPU. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Making cups and rings: the 'stalled-wave' model for macropinocytosis.

Macropinocytosis is a broadly conserved endocytic process discovered nearly 100 years ago, yet still poorly understood. It is prominent in cancer cell feeding, immune surveillance, uptake of RNA vaccines and as an invasion route for pathogens. Macropinocytic cells extend large cups or flaps from their plasma membrane to engulf droplets of medium and trap them in micron-sized vesicles. Here they are digested and the products absorbed. A major problem - discussed here - is to understand how cups are shaped and closed. Recently, lattice light-sheet microscopy has given a detailed description of this process in Dictyostelium amoebae, leading to the 'stalled-wave' model for cup formation and closure. This is based on membrane domains of PIP3 and active Ras and Rac that occupy the inner face of macropinocytic cups and are readily visible with suitable reporters. These domains attract activators of dendritic actin polymerization to their periphery, creating a ring of protrusive F-actin around themselves, thus shaping the walls of the cup. As domains grow, they drive a wave of actin polymerization across the plasma membrane that expands the cup. When domains stall, continued actin polymerization under the membrane, combined with increasing membrane tension in the cup, drives closure at lip or base. Modelling supports the feasibility of this scheme. No specialist coat proteins or contractile activities are required to shape and close cups: rings of actin polymerization formed around PIP3 domains that expand and stall seem sufficient. This scheme may be widely applicable and begs many biochemical questions.

The Amoebal Model for Macropinocytosis.

Macropinocytosis is a relatively unexplored form of large-scale endocytosis driven by the actin cytoskeleton. Dictyostelium amoebae form macropinosomes from cups extended from the plasma membrane, then digest their contents and absorb the nutrients in the endo-lysosomal system. They use macropinocytosis for feeding, maintaining a high rate of fluid uptake that makes assay and experimentation easy. Mutants collected over the years identify cytoskeletal and signalling proteins required for macropinocytosis. Cups are organized around plasma membrane domains of intense PIP3, Ras and Rac signalling, proper formation of which also depends on the RasGAPs NF1 and RGBARG, PTEN, the PIP3-regulated protein kinases Akt and SGK and their activators PDK1 and TORC2, Rho proteins, plus other components yet to be identified. This PIP3 domain directs dendritic actin polymerization to the extending lip of macropinocytic cups by recruiting a ring of the SCAR/WAVE complex around itself and thus activating the Arp2/3 complex. The dynamics of PIP3 domains are proposed to shape macropinocytic cups from start to finish. The role of the Ras-PI3-kinase module in organizing feeding structures in unicellular organisms most likely predates its adoption into growth factor signalling, suggesting an evolutionary origin for growth factor signalling.

Mathematical modelling in cell migration: tackling biochemistry in changing geometries.

Directed cell migration poses a rich set of theoretical challenges. Broadly, these are concerned with (1) how cells sense external signal gradients and adapt; (2) how actin polymerisation is localised to drive the leading cell edge and Myosin-II molecular motors retract the cell rear; and (3) how the combined action of cellular forces and cell adhesion results in cell shape changes and net migration. Reaction-diffusion models for biological pattern formation going back to Turing have long been used to explain generic principles of gradient sensing and cell polarisation in simple, static geometries like a circle. In this minireview, we focus on recent research which aims at coupling the biochemistry with cellular mechanics and modelling cell shape changes. In particular, we want to contrast two principal modelling approaches: (1) interface tracking where the cell membrane, interfacing cell interior and exterior, is explicitly represented by a set of moving points in 2D or 3D space and (2) interface capturing. In interface capturing, the membrane is implicitly modelled analogously to a level line in a hilly landscape whose topology changes according to forces acting on the membrane. With the increased availability of high-quality 3D microscopy data of complex cell shapes, such methods will become increasingly important in data-driven, image-based modelling to better understand the mechanochemistry underpinning cell motion.

QuimP: analyzing transmembrane signalling in highly deformable cells.

Summary: Transmembrane signalling plays important physiological roles, with G protein-coupled cell surface receptors being particularly important therapeutic targets. Fluorescent proteins are widely used to study signalling, but analyses of image time series can be challenging, in particular when cells change shape. QuimP software semi-automatically tracks spatio-temporal patterns of fluorescence at the cell membrane at high spatial resolution. This makes it a unique tool for studying transmembrane signalling, particularly during cell migration in immune or cancer cells for example. Availability and implementation: QuimP (http://warwick.ac.uk/quimp) is a set of Java plugins for Fiji/ImageJ (http://fiji.sc) installable through the Fiji Updater (http://warwick.ac.uk/quimp/wiki-pages/installation). It is compatible with Mac, Windows and Unix operating systems, requiring version >1.45 of ImageJ and Java 8. QuimP is released as open source (https://github.com/CellDynamics/QuimP) under an academic licence. Supplementary information: Supplementary data are available at Bioinformatics online.

How blebs and pseudopods cooperate during chemotaxis.

Two motors can drive extension of the leading edge of motile cells: actin polymerization and myosin-driven contraction of the cortex, producing fluid pressure and the formation of blebs. Dictyostelium cells can move with both blebs and actin-driven pseudopods at the same time, and blebs, like pseudopods, can be orientated by chemotactic gradients. Here we ask how bleb sites are selected and how the two forms of projection cooperate. We show that membrane curvature is an important, yet overlooked, factor. Dictyostelium cells were observed moving under agarose, which efficiently induces blebbing, and the dynamics of membrane deformations were analyzed. Blebs preferentially originate from negatively curved regions, generated on the flanks of either extending pseudopods or blebs themselves. This is true of cells at different developmental stages, chemotaxing to either folate or cyclic AMP and moving with both blebs and pseudopods or with blebs only. A physical model of blebbing suggests that detachment of the cell membrane is facilitated in concave areas of the cell, where membrane tension produces an outward directed force, as opposed to pulling inward in convex regions. Our findings assign a role to membrane tension in spatially coupling blebs and pseudopods, thus contributing to clustering protrusions to the cell front.

Phase locking and multiple oscillating attractors for the coupled mammalian clock and cell cycle.

Daily synchronous rhythms of cell division at the tissue or organism level are observed in many species and suggest that the circadian clock and cell cycle oscillators are coupled. For mammals, despite known mechanistic interactions, the effect of such coupling on clock and cell cycle progression, and hence its biological relevance, is not understood. In particular, we do not know how the temporal organization of cell division at the single-cell level produces this daily rhythm at the tissue level. Here we use multispectral imaging of single live cells, computational methods, and mathematical modeling to address this question in proliferating mouse fibroblasts. We show that in unsynchronized cells the cell cycle and circadian clock robustly phase lock each other in a 1:1 fashion so that in an expanding cell population the two oscillators oscillate in a synchronized way with a common frequency. Dexamethasone-induced synchronization reveals additional clock states. As well as the low-period phase-locked state there are distinct coexisting states with a significantly higher period clock. Cells transition to these states after dexamethasone synchronization. The temporal coordination of cell division by phase locking to the clock at a single-cell level has significant implications because disordered circadian function is increasingly being linked to the pathogenesis of many diseases, including cancer.

A novel human receptor involved in bitter tastant detection identified using Dictyostelium discoideum.

Detection of substances tasting bitter to humans occurs in diverse organisms including the social amoeba Dictyostelium discoideum. To establish a molecular mechanism for bitter tastant detection in Dictyostelium, we screened a mutant library for resistance to a commonly used bitter standard, phenylthiourea. This approach identified a G-protein-coupled receptor mutant, grlJ(-), which showed a significantly increased tolerance to phenylthiourea in growth, survival and movement. This mutant was not resistant to a structurally dissimilar potent bitter tastant, denatonium benzoate, suggesting it is not a target for at least one other bitter tastant. Analysis of the cell-signalling pathway involved in the detection of phenylthiourea showed dependence upon heterotrimeric G protein and phosphatidylinositol 3-kinase activity, suggesting that this signalling pathway is responsible for the cellular effects of phenylthiourea. This is further supported by a phenylthiourea-dependent block in the transient cAMP-induced production of phosphatidylinositol (3,4,5)-trisphosphate (PIP3) in wild-type but not grlJ(-) cells. Finally, we have identified an uncharacterized human protein γ-aminobutyric acid (GABA) type B receptor subunit 1 isoform with weak homology to GrlJ that restored grlJ(-) sensitivity to phenylthiourea in cell movement and PIP3 regulation. Our results thus identify a novel pathway for the detection of the standard bitter tastant phenylthiourea in Dictyostelium and implicate a poorly characterized human protein in phenylthiourea-dependent cell responses.

Frequency modulated translocational oscillations of Nrf2, a transcription factor functioning like a wireless sensor.

The discovery that nuclear factor erythroid 2-related factor 2 (Nrf2) undergoes translocational oscillations from cytoplasm to nucleus in human cells with frequency modulation linked to activation of a stress-stimulated cytoprotective response raises the prospect that the Nrf2 works mechanistically analogous to a wireless sensor. Herein, we consider how this new model of Nrf2 oscillation resolves previous inexplicable experimental findings on Nrf2 regulation and why it is fit-for-purpose. Further investigation is required to assess how generally applicable the oscillatory mechanism is and if characteristics of this regulatory control can be found in vivo. It suggests there are multiple, potentially re-enforcing receptors for Nrf2 activation, indicating that potent Nrf2 activation for improved health and treatment of disease may be achieved through combination of Nrf2 system stimulants.

Image based validation of dynamical models for cell reorientation.

A key feature of directed cell movement is the ability of cells to reorient quickly in response to changes in the direction of an extracellular stimulus. Mathematical models have suggested quite different regulatory mechanisms to explain reorientation, raising the question of how we can validate these models in a rigorous way. In this study, we fit three reaction-diffusion models to experimental data of Dictyostelium amoebae reorienting in response to alternating gradients of mechanical shear flow. The experimental readouts we use to fit are spatio-temporal distributions of a fluorescent reporter for cortical F-actin labeling the cell front. Experiments performed under different conditions are fitted simultaneously to challenge the models with different types of cellular dynamics. Although the model proposed by Otsuji is unable to provide a satisfactory fit, those suggested by Meinhardt and Levchenko fit equally well. Further, we show that reduction of the three-variable Meinhardt model to a two-variable model also provides an excellent fit, but has the advantage of all parameters being uniquely identifiable. Our work demonstrates that model selection and identifiability analysis, commonly applied to temporal dynamics problems in systems biology, can be a powerful tool when extended to spatio-temporal imaging data.

The distribution of Dishevelled in convergently extending mesoderm.

Convergent extension (CE) is a conserved morphogenetic movement that drives axial lengthening of the primary body axis and depends on the planar cell polarity (PCP) pathway. In Drosophila epithelia, a polarised subcellular accumulation of PCP core components, such as Dishevelled (Dvl) protein, is associated with PCP function. Dvl has long been thought to accumulate in the mediolateral protrusions in Xenopus chordamesoderm cells undergoing CE. Here we present a quantitative analysis of Dvl intracellular localisation in Xenopus chordamesoderm cells. We find that, surprisingly, accumulations previously observed at mediolateral protrusions of chordamesodermal cells are not protrusion-specific but reflect yolk-free cytoplasm and are quantitatively matched by the distribution of the cytoplasm-filling lineage marker dextran. However, separating cell cortex-associated from bulk Dvl signal reveals a statistical enrichment of Dvl in notochord-somite boundary-(NSB)-directed protrusions, which is dependent upon NSB proximity. Dvl puncta were also observed, but only upon elevated overexpression. These puncta showed no statistically significant spatial bias, in contrast to the strongly posteriorly-enriched GFP-Dvl puncta previously reported in zebrafish. We propose that Dvl distribution is more subtle and dynamic than previously appreciated and that in vertebrate mesoderm it reflects processes other than protrusion as such.

3D time series analysis of cell shape using Laplacian approaches.

BACKGROUND: Fundamental cellular processes such as cell movement, division or food uptake critically depend on cells being able to change shape. Fast acquisition of three-dimensional image time series has now become possible, but we lack efficient tools for analysing shape deformations in order to understand the real three-dimensional nature of shape changes. RESULTS: We present a framework for 3D+time cell shape analysis. The main contribution is three-fold: First, we develop a fast, automatic random walker method for cell segmentation. Second, a novel topology fixing method is proposed to fix segmented binary volumes without spherical topology. Third, we show that algorithms used for each individual step of the analysis pipeline (cell segmentation, topology fixing, spherical parameterization, and shape representation) are closely related to the Laplacian operator. The framework is applied to the shape analysis of neutrophil cells. CONCLUSIONS: The method we propose for cell segmentation is faster than the traditional random walker method or the level set method, and performs better on 3D time-series of neutrophil cells, which are comparatively noisy as stacks have to be acquired fast enough to account for cell motion. Our method for topology fixing outperforms the tools provided by SPHARM-MAT and SPHARM-PDM in terms of their successful fixing rates. The different tasks in the presented pipeline for 3D+time shape analysis of cells can be solved using Laplacian approaches, opening the possibility of eventually combining individual steps in order to speed up computations.

Quantitative analysis of human ras localization and function in the fission yeast Schizosaccharomyces pombe.

Ras signalling is central to fundamental and diverse cellular processes. In higher eukaryotes ras signalling is highly complex, involving multiple isoforms, regulatory proteins and effectors. As a consequence, the study of ras activity in mammalian systems presents a number of technical challenges. The model organism Schizosaccharomyces pombe has previously proved a key system for the study of human signalling components and provides an ideal model for the study of ras, as it contains just one ras protein (Ras1p), which is non-essential and controls a number of downstream processes. Here we present data demonstrating the quantitative analysis of three distinct Ras1-related signalling outputs, utilizing the three most abundant human ras isoforms, H-Ras, N-Ras and K-Ras4B, in Sz. pombe. Further, we have characterized the localization of these three human ras isoforms in Sz. pombe, utilizing quantitative image analysis techniques. These data indicate that all three human ras isoforms are functional in fission yeast, displaying differing localization patterns which correlate strongly with function in the regulation of pheromone response and cell shape. These data demonstrate that such yeast strains could provide powerful tools for the investigation of ras biology, and potentially in the development of cancer therapies.

Formation and closure of macropinocytic cups in Dictyostelium.

Macropinocytosis is a conserved endocytic process by which cells engulf droplets of medium into micron-sized vesicles. We use light-sheet microscopy to define an underlying set of principles by which macropinocytic cups are shaped and closed in Dictyostelium amoebae. Cups form around domains of PIP3 stretching almost to their lip and are supported by a specialized F-actin scaffold from lip to base. They are shaped by a ring of actin polymerization created by recruiting Scar/WAVE and Arp2/3 around PIP3 domains, but how cups evolve over time to close and form a vesicle is unknown. Custom 3D analysis shows that PIP3 domains expand from small origins, capturing new membrane into the cup, and crucially, that cups close when domain expansion stalls. We show that cups can close in two ways: either at the lip, by inwardly directed actin polymerization, or the base, by stretching and delamination of the membrane. This provides the basis for a conceptual mechanism whereby closure is brought about by a combination of stalled cup expansion, continued actin polymerization at the lip, and membrane tension. We test this through the use of a biophysical model, which can recapitulate both forms of cup closure and explain how 3D cup structures evolve over time to mediate engulfment.

The Diaphanous-related formin dDia2 is required for the formation and maintenance of filopodia.

Formins have important roles in the nucleation of actin and the formation of linear actin filaments, but their role in filopodium formation has remained elusive. Dictyostelium discoideum Diaphanous-related formin dDia2 is enriched at the tips of filopodia and interacts with profilin II and Rac1. An FH1FH2 fragment of dDia2 nucleated actin polymerization and removed capping protein from capped filament ends. Genetic studies showed that dDia2 is important for cell migration as well as the formation, elongation and maintenance of filopodia. Here we provide evidence that dDia2 specifically controls filopodial dynamics by regulating actin turnover at the barbed ends of actin filaments.

A Curvature-Enhanced Random Walker Segmentation Method for Detailed Capture of 3D Cell Surface Membranes.

High-resolution 3D microscopy is a fast advancing field and requires new techniques in image analysis to handle these new datasets. In this work, we focus on detailed 3D segmentation of Dictyostelium cells undergoing macropinocytosis captured on an iSPIM microscope. We propose a novel random walker-based method with a curvature-based enhancement term, with the aim of capturing fine protrusions, such as filopodia and deep invaginations, such as macropinocytotic cups, on the cell surface. We tested our method on both real and synthetic 3D image volumes, demonstrating that the inclusion of the curvature enhancement term can improve the segmentation of the aforementioned features. We show that our method performs better than other state of the art segmentation methods in 3D images of Dictyostelium cells, and performs competitively against CNN-based methods in two Cell Tracking Challenge datasets, demonstrating the ability to obtain accurate segmentations without the requirement of large training datasets. We also present an automated seeding method for microscopy data, which, combined with the curvature-enhanced random walker method, enables the segmentation of large time series with minimal input from the experimenter.

Interactive segmentation of clustered cells via geodesic commute distance and constrained density weighted Nyström method.

An interactive method is proposed for complex cell segmentation, in particular of clustered cells. This article has two main contributions: First, we explore a hybrid combination of the random walk and the geodesic graph based methods for image segmentation and propose the novel concept of geodesic commute distance to classify pixels. The computation of geodesic commute distance requires an eigenvector decomposition of the weighted Laplacian matrix of a graph constructed from the image to be segmented. Second, by incorporating pairwise constraints from seeds into the algorithm, we present a novel method for eigenvector decomposition, namely a constrained density weighted Nyström method. Both visual and quantitative comparison with other semiautomatic algorithms including Voronoi-based segmentation, grow cut, graph cuts, random walk, and geodesic method are given to evaluate the performance of the proposed method, which is a powerful tool for quantitative analysis of clustered cell images in live cell imaging.

Analysis of cell movement by simultaneous quantification of local membrane displacement and fluorescent intensities using Quimp2.

The use of fluorescent markers in living cells has increased dramatically in the recent years. The quantitative analysis of the images requires specific analysis software. Previously, the program Quimp was launched for quantitating fluorescent intensities at the membrane or the cortex of the cell. However, Quimp is not well suited to quantitate local membrane displacement. Here we present Quimp2 that is capable of tracking membrane subregions in time, which enables the simultaneous quantification of fluorescent intensities and membrane movement. Quimp2 has two new tools, (i) conversion filters to analyze movies obtained with fluorescent, DIC and phase contrast different microscopes, and (ii) a macro that calculates the local membrane displacement and provides various options to display the results. Quimp2 is used here to investigate the molecular mechanism of cell movement by correlating the dynamics of local membrane movement with the local concentration of myosin and F-actin.

The three-dimensional dynamics of actin waves, a model of cytoskeletal self-organization.

Actin polymerization is typically initiated at specific sites in a cell by membrane-bound protein complexes, and the resulting structures are involved in specialized cellular functions, such as migration, particle uptake, or mitotic division. Here we analyze the potential of the actin system to self-organize into waves that propagate on the planar, substrate-attached membrane of a cell. We show that self-assembly involves the ordered recruitment of proteins from the cytoplasmic pool and relate the organization of actin waves to their capacity for applying force. Three proteins are shown to form distinct three-dimensional patterns in the actin waves. Myosin-IB is enriched at the wave front and close to the plasma membrane, the Arp2/3 complex is distributed throughout the waves, and coronin forms a sloping layer on top of them. CARMIL, a protein that links myosin-IB to the Arp2/3 complex, is also recruited to the waves. Wave formation does not depend on signals transmitted by heterotrimeric G-proteins, nor does their propagation require SCAR, a regulator upstream of the Arp2/3 complex. Propagation of the waves is based on an actin treadmilling mechanism, indicating a program that couples actin assembly to disassembly in a three-dimensional pattern. When waves impinge on the cell perimeter, they push the edge forward; when they reverse direction, the cell border is paralyzed. These data show that force-generating, highly organized supramolecular networks are autonomously formed in live cells from molecular motors and proteins controlling actin polymerization and depolymerization.

Reversal of cell polarity and actin-myosin cytoskeleton reorganization under mechanical and chemical stimulation.

To study reorganization of the actin system in cells that invert their polarity, we stimulated Dictyostelium cells by mechanical forces from alternating directions. The cells oriented in a fluid flow by establishing a protruding front directed against the flow and a retracting tail. Labels for polymerized actin and filamentous myosin-II marked front and tail. At 2.1 Pa, actin first disassembled at the previous front before it began to polymerize at the newly induced front. In contrast, myosin-II slowly disappeared from the previous tail and continuously redistributed to the new tail. Front specification was myosin-II independent and accumulation of polymerized actin was even more focused in mutants lacking myosin-II heavy chains. We conclude that under mechanical stimulation, the inversion of cell polarity is initiated by a global internal signal that turns down actin polymerization in the entire cell. It is thought to be elicited at the most strongly stimulated site of the cell, the incipient front region, and to be counterbalanced by a slowly generated, short-range signal that locally activates actin polymerization at the front. Similar pattern of front and tail interconversion were observed in cells reorienting in strong gradients of the chemoattractant cyclic AMP.