Biology subjects

Ju, L. A.

Publications and source records attributed to Ju, L. A..

9 recordsLinked to original sources

bioRxiv2026

Patient-specific "Physical Twin" artery-on-chip platform reveals complex flow-dependent VWF mechanobiology and guides personalized antithrombotic therapy

Ischemic stroke triggered by carotid atherosclerosis remains unpredictable because the thrombosis embolization is governed by patient-specific vascular geometry and hemodynamics that cannot be recapitulated in conventional models. Here, we engineered "Physical Twin" artery-on-chip platforms that reproduce individualized carotid anatomy with humanized subendothelial matrix composition and arterial endothelial phenotype under physiological flow patterns. We then established thrombotic microenvironments following laser-induced injury. Computational fluid dynamics-guided experiments across distinct patient geometries reveal that local shear dictates a mechanobiological hierarchy: high-shear bifurcations ([~]3,000 s-{superscript 1}) produce von Willebrand factor (VWF) A1-dependent thrombi suppressible by conformationally sensitive inhibitors, whereas low-shear stenoses generate fragile aggregates where VWF-integrin IIb{beta} coupling governs embolization overgrowth. Pulsatile flow suppresses thrombotic growth independent of mean shear. Molecular dynamics simulations reveal why conformationally sensitive nanobodies (caplacizumab) outperform shear-independent aptamers (ARC1172) in high-shear bifurcation flow zones. This Physical Twin platform provides a mechanistic blueprint for geometry-informed, personalized antithrombotic therapy selection.

bioengineering↗

bioRxiv2026

Flow molecular dynamics simulations reveal mechanosensitive regulation of von Willebrand factor through glycan-modulated autoinhibitory modules

Force-induced protein conformational changes govern many essential biological processes, yet their molecular mechanisms remain difficult to resolve. Von Willebrand factor (VWF), a central regulator of haemostasis, is activated by hydrodynamic forces in blood flow, but how mechanical signals propagate across its multidomain architecture is poorly understood. Here, we use flow molecular dynamics (FMD), a simulation framework that applies fluid forces via controlled solvent flow to interrogate mechanosensitive proteins. Using VWF as a model system, we reconstructed the complete mechanomodule (D'D3-A1-A2-A3; 1,109 residues) with native glycosylation by integrating crystallographic data and AlphaFold predictions. FMD simulations capture a force-driven transition from a compact, autoinhibited "birds-nest" ensemble to an extended, activated state, revealing asymmetric autoinhibitory strengths within the N'AIM and C'AIM modules of the A1 domain. By directly linking static structures to dynamic, force-regulated behaviour, this work establishes a generalizable platform for dissecting protein mechanosensitivity and enabling the rational design of force-responsive therapeutics. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=48 SRC="FIGDIR/small/716521v1_ufig1.gif" ALT="Figure 1"> View larger version (16K): org.highwire.dtl.DTLVardef@edba50org.highwire.dtl.DTLVardef@1630df4org.highwire.dtl.DTLVardef@292887org.highwire.dtl.DTLVardef@23cdb7_HPS_FORMAT_FIGEXP M_FIG C_FIG Flow molecular dynamics simulations reveal that GPIb engages the A1 domain only after the disruption of key interdomain and intermodular interactions.

bioinformatics↗

bioRxiv2026

Automated high-throughput fabrication of patient-specific vessel-on-chips enables a generative AI digital twin--Cascade Learner of Thrombosis (CLoT) for personalized thrombosis prediction

We developed an integrated platform combining high-throughput automated biofabrication, systematic patient-derived tissue experiments, and specialized artificial intelligence to enable patient-specific computational "digital twins" for thrombosis prediction. Our automated manufacturing platform fabricates 80 fully assembled, patient-specific vessel-on-chips within 10 hours from clinical imaging--a [~]100-fold improvement over manual methods--achieving sub-micron precision through novel two-stage pneumatic motion control and integrated optical feedback. Using these chips, we systematically captured thrombosis across 491 high-fidelity videos spanning 6 patient-derived vascular geometries, 5 distinct anatomical injury sites, and 14 anticoagulant/antiplatelet interventions, establishing a "physical twin" experimental corpus. We trained CLoT (Cascade Learner of Thrombosis), a conditional video diffusion model efficiently adapted via lightweight Low-Rank Adaptation (LoRA) to generate realistic thrombosis videos conditioned on patient-specific geometry, injury location, and drug treatment. Rigorous benchmarking against state-of-the-art commercial models (Sora, Wan, Kling, Seedance, Hailuo, Hunyuan) reveals CLoT achieves 7.38-fold superior temporal biological consistency and 5.3-fold higher spatial morphological fidelity. Prospective validation on unseen patients demonstrates >90% temporal accuracy. This integrated paradigm--combining automated fabrication with domain-specialized generative AI--establishes proof-of-concept for personalized medicine enabled by digital twins trained on human-derived vascular anatomy, enabling pre-treatment antithrombotic evaluation while providing a replicable template for translating tissue engineering into clinical practice.

bioengineering↗

bioRxiv2025

Fluorescently Labeled Gradient Hydrogels Reveal Matrix-Dependent Cell Responses to Substrate Stiffness

Microfabricated stiffness gradient hydrogels hold significant value for advancing mechanobiology, tissue engineering, and in vitro tissue models. However, it remains challenging to design these materials given their broad processing parameter space. The continuum of stiffness values also makes it difficult to precisely correlate the local substrate properties and observed biological responses, often relying on cumbersome characterization methods such as atomic force microscopy. To address these bottlenecks, we present a straightforward thermophoresis-based fabrication strategy to pattern stiffness gradients in fluorescein isothiocyanate-labeled hydrogel network, which displays a polymer concentration-dependent fluorescence readout. This approach enables quantitative assessment of the gradient formation process and contactless stiffness mapping via standard microscopy imaging. Using gelatin methacryloyl and Gellan gum as model systems, it is shown that substrate stiffness and extracellular matrix protein composition work together to affect 3T3-L1 fibroblast cell morphology and migration, with the underlying hydrogel type also affecting the outcome. By offering a simple and reliable approach for characterizing stiffness gradient hydrogels, this work advances the thermophoretic fabrication platform, opening avenues for new biomaterial systems for understanding and controlling the cell-material interplay.

bioengineering↗

bioRxiv2025

DNA Partitioning Modulates Liquid-to-Solid Transitions and the Internal Microstructure of FUS Condensates

Protein liquid-liquid phase separation has recently been recognized as an essential process involved in cellular functions, including transcription, translation, and DNA damage repair. However, a further liquid-to-solid transition (LST) of condensates due to mutation or external stimuli can result in aggregation, sometimes pathological. The unique ability of protein condensates to concentrate and sequester biomolecules is at the heart of the regulatory mechanism, controlling the dynamics and function of the condensates. While the recruitment of essential biomolecules, such as RNA has been studied to have the impact to the condensate formation, how DNA partition can affect the phase behaviour and dynamics of protein condensates is not fully elucidated. In this study, we investigate both the short-term and long-term kinetics of double-stranded DNA partitioning into preformed fresh and aged FUS protein condensates. Confocal imaging shows that DNA partition follows the core-shell diffusion pattern within the condensates. LST slows down and reduces FUS condensates ability to recruit DNA but stabilizes the DNA-FUS condensate complex due to the heterogeneous solid network formation. Using the optical technique of Spatial Dynamic Mapping (SDM), we find that DNA partition promotes coalescence and alters the characteristics of the condensates. The partition made the condensates more dynamic in the short term (within minutes) but accelerated LST in long-term incubation (within hours), ultimately leading to an irreversible porous core-shell structure of FUS condensates. Our findings reveal the kinetics of DNA partition during aging and its impact on LST, underlining the modulation of condensate properties by molecule sequestration, shedding light on possible regulation of disease-related LST of biomolecular condensates.

biophysics↗

bioRxiv2024

Multi-parametric thrombus profiling microfluidics detects intensified biomechanical thrombogenesis associated with hypertension and aging

Arterial thrombosis, which represents a critical complication of cardiovascular diseases, is a leading cause of death and disability worldwide with no effective bioassay for clinical prediction. As a symbolic feature of arterial thrombosis, severe stenosis in the blood vessel creates a high-shear, high-gradient flow environment that effectively facilitates platelet aggregation towards vessel occlusion even with platelet amplification loops inhibited. However, no approach is currently available to comprehensively characterize the size, composition and platelet activation status of thrombi forming under this biorheological condition. Here, we present a thrombus profiling assay that monitors the multi-dimensional attributes of thrombi forming in conditions mimicking the physiological scenario of arterial thrombosis. Using this platform, we demonstrate that different receptor-ligand interactions contribute distinctively to the composition and activation status of the thrombus. Our investigation into hypertensive and older individuals reveals intensified biomechanical thrombogenesis and multi-dimensional thrombus profile abnormalities, demonstrating a direct contribution of mechanobiology to arterial thrombosis and endorsing the diagnostic potential of the assay. Furthermore, we identify the hyperactivity of GPIb-integrin IIb{beta}3 mechanosensing axis as a molecular mechanism that contributes to hypertension-associated arterial thrombosis. By studying the interactions between anti-thrombotic inhibitors and hypertension, and the inter-individual variability in personal thrombus profiles, our work reveals a critical need for personalized anti-thrombotic drug selection that accommodates each patients pathological profile.

bioengineering↗

bioRxiv2023

Force-regulated spontaneous conformational changes of integrins α5β1 and αVβ3

Force can modulate the properties and functions of macromolecules by inducing conformational changes, such as coiling/uncoiling, zipping/unzipping, and folding/unfolding. Here we compared force-modulated bending/unbending of two purified integrin ectodomains, 5{beta}1 and V{beta}3, using single-molecule approaches. Similar to previously characterized mechano-sensitive macromolecules, the conformation of 5{beta}1 is determined by a threshold head-to-tail tension, suggesting a canonical energy landscape with a deep energy well that traps the integrin in the bent state until sufficient force tilts the energy landscape to accelerate transition to the extended state. By comparison, V{beta}3 exhibits bi-stability even without force and can spontaneously transition between the bent and extended conformations in a wide range of forces without energy supplies. Molecular dynamics simulations revealed consecutive formation and disruption of 7 hydrogen bonds during V{beta}3 bending and unbending, respectively. Accordingly, we constructed an energy landscape with hexa-stable intermediate states to break down the energy barrier separating the bent and extended states into smaller ones, making it possible for the thermal agitation energy to overcome them sequentially and to be accumulated and converted into mechanical work required for V{beta}3 to bend against force. Our study elucidates the different inner workings of 5{beta}1 and V{beta}3 at the sub-molecular level, sheds lights on how their respectively functions are facilitated by their distinctive mechano-sensitivities, helps understand their signal initiation processes, and provides critical concepts and useful design principles for engineering of protein-based biomechanical nanomachines.

biophysics↗

bioRxiv2022

3D spheroid-microvasculature-on-a-chip for tumor-endothelium mechanobiology interplay

In the final step of cancer metastasis, tumor cells become lodged in a distant capillary bed, where they can undergo extravasation and form a secondary tumor. While increasing evidence suggests blood/lymphatic flow and shear stress play a critical role in the tumor extravasation process, there is a lack of systematic and biomechanical approaches to recapitulate sophisticated 3D microtissue interactions within the controllable hydrodynamic microenvironment. Here, we report a simple-to-use 3D spheroid-microvasculature-on-a-chip (SMAC) model. Under static and controlled flow conditions, the SMAC recapitulates the biomechanical crosstalk between heterogeneous tumor spheroids and the endothelium in a high-throughput and quantitative manners. As an in vitro metastasis mechanobiology model, we discover 3D spheroid-induced endothelial compression and cell-cell junction degradation in the process of tumor migration and expansion. Lastly, we examine the shear stress effects on the endothelial orientation, polarization as well as the tumor spheroid expansion. Taken together, our SMAC model offers a miniaturized, cost-efficient and versatile platform for future investigation on metastasis mechanobiology, enhanced permeability and retention effect and even personalized therapeutic evaluation.

bioengineering↗

bioRxiv2020

Modified N-linked glycosylation status predicts trafficking defective human Piezo1 channel mutations

Mechanosensitive channels are integral membrane proteins that sense mechanical stimuli. Like all membrane proteins, they pass through biosynthetic quality control in the endoplasmic reticulum and Golgi that results in them reaching their destination at the plasma membrane. Here we show that N-linked glycosylation of two highly conserved asparagine residues in the cap region of mechanosensitive Piezo1 channels are necessary for the mature protein to reach the plasma membrane. Both mutation of these asparagines (N2294Q/N2331Q) and treatment with an enzyme that hydrolyses N-linked oligosaccharides (PNGaseF) eliminates the fully glycosylated mature Piezo1 protein. The N-glycans in the cap are a pre-requisite for higher-order glycosylation in the propeller regions, which are present in loops that are essential for mechanotransduction. Importantly, trafficking-defective Piezo1 variants linked to generalized lymphatic dysplasia and bicuspid aortic valve display reduced fully N-glycosylated protein. The higher order glycosylation status in vitro correlates with efficient membrane trafficking and will aid in determining the functional impact of Piezo1 variants of unknown significance.

biophysics↗