Synthetic optical coherence tomography angiographs for detailed retinal vessel segmentation without human annotations.

Optical coherence tomography angiography (OCTA) is a non-invasive imaging modality that can acquire high-resolution volumes of the retinal vasculature and aid the diagnosis of ocular, neurological and cardiac diseases. Segmenting the visible blood vessels is a common first step when extracting quantitative biomarkers from these images. Classical segmentation algorithms based on thresholding are strongly affected by image artifacts and limited signal-to-noise ratio. The use of modern, deep learning-based segmentation methods has been inhibited by a lack of large datasets with detailed annotations of the blood vessels. To address this issue, recent work has employed transfer learning, where a segmentation network is trained on synthetic OCTA images and is then applied to real data. However, the previously proposed simulations fail to faithfully model the retinal vasculature and do not provide effective domain adaptation. Because of this, current methods are unable to fully segment the retinal vasculature, in particular the smallest capillaries. In this work, we present a lightweight simulation of the retinal vascular network based on space colonization for faster and more realistic OCTA synthesis. We then introduce three contrast adaptation pipelines to decrease the domain gap between real and artificial images. We demonstrate the superior segmentation performance of our approach in extensive quantitative and qualitative experiments on three public datasets that compare our method to traditional computer vision algorithms and supervised training using human annotations. Finally, we make our entire pipeline publicly available, including the source code, pretrained models, and a large dataset of synthetic OCTA images.

The molecular mechanism of load adaptation by branched actin networks.

Branched actin networks are self-assembling molecular motors that move biological membranes and drive many important cellular processes, including phagocytosis, endocytosis, and pseudopod protrusion. When confronted with opposing forces, the growth rate of these networks slows and their density increases, but the stoichiometry of key components does not change. The molecular mechanisms governing this force response are not well understood, so we used single-molecule imaging and AFM cantilever deflection to measure how applied forces affect each step in branched actin network assembly. Although load forces are observed to increase the density of growing filaments, we find that they actually decrease the rate of filament nucleation due to inhibitory interactions between actin filament ends and nucleation promoting factors. The force-induced increase in network density turns out to result from an exponential drop in the rate constant that governs filament capping. The force dependence of filament capping matches that of filament elongation and can be explained by expanding Brownian Ratchet theory to cover both processes. We tested a key prediction of this expanded theory by measuring the force-dependent activity of engineered capping protein variants and found that increasing the size of the capping protein increases its sensitivity to applied forces. In summary, we find that Brownian Ratchets underlie not only the ability of growing actin filaments to generate force but also the ability of branched actin networks to adapt their architecture to changing loads.

Measurement of Retinal Microvascular Blood Velocity Using Erythrocyte Mediated Velocimetry.

Changes in retinal blood flow may be involved in the pathogenesis of glaucoma and other ocular diseases. Erythrocyte mediated velocimetry (EMV) is a novel technique where indocyanine green (ICG) dye is sequestered in erythrocyte ghosts and autologously re-injected to allow direct visualization of erythrocytes for in vivo measurement of speed. The purpose of this study is to determine the mean erythrocyte speed in the retinal microvasculature, as well as the intravisit and intervisit variability of EMV. Data from 23 EMV sessions from control, glaucoma suspect, and glaucoma patients were included in this study. In arteries with an average diameter of 43.11 µm ± 6.62 µm, the mean speed was 7.17 mm/s ± 2.35 mm/s. In veins with an average diameter of 45.87 µm ± 12.04 µm, the mean speed was 6.05 mm/s ± 1.96 mm/s. Intravisit variability, as measured by the mean coefficient of variation, was 3.57% (range 0.44-9.68%). Intervisit variability was 4.85% (range 0.15-8.43%). EMV may represent reliable method for determination of retinal blood speed, potentially allowing insights into the effects of pharmacologic agents or pathogenesis of ocular diseases.

Distinct Retinal Capillary Plexuses in Normal Eyes as Observed in Optical Coherence Tomography Angiography Axial Profile Analysis.

Optical coherence tomography angiography (OCTA) allows the retinal microvasculature to be visualized at various retinal depths. Previous studies introduced OCTA axial profile analysis and showed regional variations in the number and location of axially distinct vascular retinal plexuses. OCTA acquisition and processing approaches, however, vary in terms of their resulting transverse and axial resolutions, and especially the latter could potentially influence the profile analysis results. Our study imaged normal eyes using the Spectralis OCT2 with a full-spectrum, probabilistic OCTA algorithm, that, in marked contrast to split-spectrum approaches, preserves the original high OCT axial resolution also within the resulting OCTA signal. En face OCTA images are generally created by averaging flow signals over a finite axial depth window. However, we assessed regional OCTA signal profiles at each depth position at full axial resolution. All regions had two sharp vessel density peaks near the inner and outer boundaries of the inner nuclear layer, indicating separate intermediate and deep capillary plexuses. The superficial vascular plexus (SVP) separated into two distinct peaks within the ganglion cell layer in the parafoveal zone. The nasal, superior, and inferior perifovea had a deeper SVP peak that was shifted anteriorly compared to the parafoveal zone. Axial vascular density analysis with high-resolution, full spectrum OCTA thus allows healthy retinal vasculature to be precisely reconstructed and may be useful for clinically assessing retinal pathology.

Size-dependent protein segregation at membrane interfaces.

Membrane interfaces formed at cell-cell junctions are associated with characteristic patterns of membrane protein organization, such as E-cadherin enrichment in epithelial junctional complexes and CD45 exclusion from the signaling foci of immunological synapses. To isolate the role of protein size in these processes, we reconstituted membrane interfaces in vitro using giant unilamellar vesicles decorated with synthetic binding and non-binding proteins. We show that size differences between binding and non-binding proteins can dramatically alter their organization at membrane interfaces in the absence of active contributions from the cytoskeleton, with as little as a ~5 nm increase in non-binding protein size driving its exclusion from the interface. Combining in vitro measurements with Monte Carlo simulations, we find that non-binding protein exclusion is also influenced by lateral crowding, binding protein affinity, and thermally-driven membrane height fluctuations that transiently limit access to the interface. This simple, sensitive, and highly effective means of passively segregating proteins has implications for signaling at cell-cell junctions and protein sorting at intracellular contact points between membrane-bound organelles.

The More the Tubular: Dynamic Bundling of Actin Filaments for Membrane Tube Formation.

Tubular protrusions are a common feature of living cells, arising from polymerization of stiff protein filaments against a comparably soft membrane. Although this process involves many accessory proteins in cells, in vitro experiments indicate that similar tube-like structures can emerge without them, through spontaneous bundling of filaments mediated by the membrane. Using theory and simulation of physical models, we have elaborated how nonequilibrium fluctuations in growth kinetics and membrane shape can yield such protrusions. Enabled by a new grand canonical Monte Carlo method for membrane simulation, our work reveals a cascade of dynamical transitions from individually polymerizing filaments to highly cooperatively growing bundles as a dynamical bottleneck to tube formation. Filament network organization as well as adhesion points to the membrane, which bias filament bending and constrain membrane height fluctuations, screen the effective attractive interactions between filaments, significantly delaying bundling and tube formation.

Force Feedback Controls Motor Activity and Mechanical Properties of Self-Assembling Branched Actin Networks.

Branched actin networks--created by the Arp2/3 complex, capping protein, and a nucleation promoting factor--generate and transmit forces required for many cellular processes, but their response to force is poorly understood. To address this, we assembled branched actin networks in vitro from purified components and used simultaneous fluorescence and atomic force microscopy to quantify their molecular composition and material properties under various forces. Remarkably, mechanical loading of these self-assembling materials increases their density, power, and efficiency. Microscopically, increased density reflects increased filament number and altered geometry but no change in average length. Macroscopically, increased density enhances network stiffness and resistance to mechanical failure beyond those of isotropic actin networks. These effects endow branched actin networks with memory of their mechanical history that shapes their material properties and motor activity. This work reveals intrinsic force feedback mechanisms by which mechanical resistance makes self-assembling actin networks stiffer, stronger, and more powerful.

Investigating the role of F-actin in human immunodeficiency virus assembly by live-cell microscopy.

Human immunodeficiency virus type 1 (HIV-1) particles assemble at the plasma membrane, which is lined by a dense network of filamentous actin (F-actin). Large amounts of actin have been detected in HIV-1 virions, proposed to be incorporated by interactions with the nucleocapsid domain of the viral polyprotein Gag. Previous studies addressing the role of F-actin in HIV-1 particle formation using F-actin-interfering drugs did not yield consistent results. Filamentous structures pointing toward nascent HIV-1 budding sites, detected by cryo-electron tomography and atomic force microscopy, prompted us to revisit the role of F-actin in HIV-1 assembly by live-cell microscopy. HeLa cells coexpressing HIV-1 carrying fluorescently labeled Gag and a labeled F-actin-binding peptide were imaged by live-cell total internal reflection fluorescence microscopy (TIR-FM). Computational analysis of image series did not reveal characteristic patterns of F-actin in the vicinity of viral budding sites. Furthermore, no transient recruitment of F-actin during bud formation was detected by monitoring fluorescence intensity changes at nascent HIV-1 assembly sites. The chosen approach allowed us to measure the effect of F-actin-interfering drugs on the assembly of individual virions in parallel with monitoring changes in the F-actin network of the respective cell. Treatment of cells with latrunculin did not affect the efficiency and dynamics of Gag assembly under conditions resulting in the disruption of F-actin filaments. Normal assembly rates were also observed upon transient stabilization of F-actin by short-term treatment with jasplakinolide. Taken together, these findings indicate that actin filament dynamics are dispensable for HIV-1 Gag assembly at the plasma membrane of HeLa cells. Importance: HIV-1 particles assemble at the plasma membrane of virus-producing cells. This membrane is lined by a dense network of actin filaments that might either present a physical obstacle to the formation of virus particles or generate force promoting the assembly process. Drug-mediated interference with the actin cytoskeleton showed different results for the formation of retroviral particles in different studies, likely due to general effects on the cell upon prolonged drug treatment. Here, we characterized the effect of actin-interfering compounds on the HIV-1 assembly process by direct observation of virus formation in live cells, which allowed us to measure assembly rate constants directly upon drug addition. Virus assembly proceeded with normal rates when actin filaments were either disrupted or stabilized. Taken together with the absence of characteristic actin filament patterns at viral budding sites in our analyses, this indicates that the actin network is dispensable for HIV-1 assembly.

Unifying autocatalytic and zeroth-order branching models for growing actin networks.

The directed polymerization of actin networks is an essential element of many biological processes, including cell migration. Different theoretical models considering the interplay between the underlying processes of polymerization, capping, and branching have resulted in conflicting predictions. One of the main reasons for this discrepancy is the assumption of a branching reaction that is either first order (autocatalytic) or zeroth order in the number of existing filaments. Here we introduce a unifying framework from which the two established scenarios emerge as limiting cases for low and high filament numbers. A smooth transition between the two cases is found at intermediate conditions. We also derive a threshold for the capping rate above which autocatalytic growth is predicted at sufficiently low filament number. Below the threshold, zeroth-order characteristics are predicted to dominate the dynamics of the network for all accessible filament numbers. Together, these mechanisms allow cells to grow stable actin networks over a large range of different conditions.

Reconstructing the orientation distribution of actin filaments in the lamellipodium of migrating keratocytes from electron microscopy tomography data.

Migration of motile cells on flat substrates is usually driven by the polymerization of a flat actin filament network. Theoretical models have made different predictions regarding the distribution of the filament orientation in the lamellipodium with respect to the direction of motion. Here we show how one can automatically reconstruct the orientation distribution of actin filaments in the lamellipodium of migrating keratocytes from electron microscopy tomography data. We use two different image analysis methods, an algorithm which explicitly extracts an abstract network representation and an analysis of the gray scale information based on the structure tensor. We show that the two approaches give similar results, both for simulated data and for electron microscopy tomography data from migrating keratocytes. For the lamellipodium at the leading edge of fast moving cells, we find an orientation distribution that is peaked at +35/-35 degrees. For the lamellipodium at the leading edge of slow moving cells as well as for the lamellipodium at the flanks of fast moving cells, one broad peak around 0 degree dominates the distribution.

Two competing orientation patterns explain experimentally observed anomalies in growing actin networks.

The lamellipodium of migrating animal cells protrudes by directed polymerization of a branched actin network. The underlying mechanisms of filament growth, branching, and capping can be studied in in vitro assays. However, conflicting results have been reported for the force-velocity relation of such actin networks, namely both convex and concave shapes as well as history dependencies. Here we model branching as a reaction that is independent of the number of existing filaments, in contrast to capping, which is assumed to be proportional to the number of existing filaments. Using both stochastic network simulations and deterministic rate equations, we show that such a description naturally leads to the stability of two qualitatively different stationary states of the system, namely a +/- 35 degrees and a +70/0/-70 degrees orientation pattern. Changes in network growth velocity induce a transition between these two patterns. For sufficiently different protrusion efficiency of the two network architectures, this leads to hysteresis in the growth velocity of actin networks under force. Dependent on the history of the system, convex and concave regimes are obtained for the force-velocity relation. Thus a simple generic model can explain the experimentally observed anomalies, with far reaching consequences for cell migration.

A quantitative measure for alterations in the actin cytoskeleton investigated with automated high-throughput microscopy.

The actin cytoskeleton modulates a large variety of physiological and disease-related processes in the cell. For example, actin has been shown to be a crucial host factor for successful infection by HIV-1, but the underlying mechanistic details are still unknown. Automated approaches open up the perspective to clarify such an issue by processing many samples in a high-throughput manner. To analyze the alterations in the actin cytoskeleton within an automated setting, large-scale image acquisition and analysis were established for JC-53 cells stained for actin. As a quantitative measure in such an automated approach, we suggest a parameter called image coherency. We successfully benchmarked our analysis by calculating coherency for both a biophysical model of the actin cytoskeleton and for cells whose actin architecture had been disturbed pharmacologically by latrunculin B or cytochalasin D. We then tested the influence of HIV-1 infection on actin coherency, but observed no significant differences between uninfected and infected cells.