ABSTRACT Understanding the competition between the electrochemical CO2 reduction reaction (eCO2RR) and the hydrogen evolution reaction (HER) is crucial for efficient CO2 conversion, particularly under limited reactant availability using diluted CO2. We investigate the interplay between the operating conditions at various differential backpressures concerning CO2 availability, water behavior at the electrified catalyst layer, and how these factors modulate the eCO2RR‐to‐formate selectivity and HER activity. We show that H2 formation can be suppressed while maintaining high formate selectivity even using highly diluted CO2 (2.25%) by increasing the differential backpressure from the gas side of the gas‐diffusion cathode in the flow‐through electrolyzer, which enhances CO2 transport to the catalyst layer through reshaping the triple‐phase boundary. Operando Raman spectroscopy reveals the accumulation and perturbation of water on the catalyst layer at negative potentials, accompanied by weakening of the hydrogen‐bond network. Moreover, even higher differential backpressure minimizes carbonate formation, preserves eCO2RR kinetics, and promotes the displacement of surface water, which in turn minimizes HER.
Proteins exert sophisticated functions through stimulus-responsive conformational changes. Mimicking this structural control and functional integration in synthetic, water-soluble polymers has been a fundamental challenge, largely due to the conflict between charge repulsion and conformational order. Here, we report a peptidomimetic polyelectrolyte that overcomes this limitation, maintaining stable and well-defined helical and sheet-like conformations across the entire physiological pH range despite its high charge density. This intrinsic conformational order creates through-space conjugated carbonyl clusters that function as non-classical chromophores, enabling excitation-dependent near-infrared fluorescence with an exceptionally large Stokes shift. More importantly, the sheet-like conformation enables cooperative, multidentate chelation of Fe3 + ions, which triggers an allosteric transition to a helix and results in nonlinear and ultrasensitive fluorescence quenching. Leveraging this unique "conformational transduction amplification" mechanism, we achieve real-time visualization and tracking of iron ion distribution and metabolic pathways at subcellular and whole-organism levels. This work establishes a paradigm of allosteric control in synthetic polyelectrolytes, opening avenues for the design of intelligent biomimetic materials for advanced sensing and imaging.
A computation-guided investigation of asymmetric donor-acceptor bridged stilbenes reveals structure-property relationships governing conical intersection (CI) accessibility and excited-state deactivation in aggregation-induced emission (AIE) luminogens. Although CIs play a central role in nonradiative decay, the molecular factors governing CI accessibility in AIE systems remain insufficiently understood. Here, we show that asymmetric donor-acceptor placement combined with bridge-controlled structural flexibility strongly influences CI accessibility and excited-state deactivation in push-pull alkylene-bridged stilbenes ([6]/[7]). Quantum-chemical analyses of 30 derivatives reveal substituent-dependent energetic trends associated with CI accessibility that can be rationalized by the relative energetic positions of the Franck-Condon and CI geometries. On the basis of these trends, representative derivatives, including DCBS[6], DCBS[7], DPB[7]C, and DPB[7]N, were synthesized together with reference compounds. Their photophysical properties generally correlate with the computed CI-accessibility trends, indicating that donor-acceptor asymmetry and bridge rigidity cooperatively influence excited-state deactivation and fluorescence suppression in solution. Time-resolved spectroscopy, post-relaxation PES analyses, and CI topology analyses further support the proposed relaxation pathways and suggest that substituent inversion alters CI energetics, topology, and nonadiabatic coupling. These findings provide mechanistic insight into substituent-dependent excited-state deactivation in bridged stilbenes.
Reconstruction of keratinized mucosa (KM) with sufficient dimensions is critical for long-term periodontal and peri-implant health. However, existing biomaterials struggle to recapitulate the complex biophysical and biochemical microenvironment of KM while achieving stable adhesion and integration in the wet and mechanically dynamic oral cavity. Here, we design a photocrosslinkable double-network hydrogel composed of methacrylated platelet-rich fibrin (iPRF-MA), N-hydroxysuccinimide-functionalized alginate (Alg-NHS), and luteolin-loaded epigallocatechin gallate microspheres (Lut@EGCG) to enable KM regeneration in the challenging oral environment through dual microenvironmental modulation. At the material level, the covalent network from iPRF-MA and the supramolecular network based on Alg-NHS work synergistically, resulting in strong wet tissue adhesion and high fatigue resistance, which collectively prevent hydrogel dislodgement under oral dynamic stresses. Biochemically, the hydrogel enables sustained release of growth factors and EGCG, synergistically enhancing angiogenesis and immune regulation, while also redirecting neutrophil phenotype toward a phagocytic state for specific antibacterial activity. Biophysically, the hydrogel provides gingival fibroblasts with a mechanically instructive microenvironment that activates mechanotransduction signaling and accelerates extracellular matrix remodeling. In vivo experiments confirm outstanding KM regeneration following treatment with the double-network hydrogel. This study demonstrates a microenvironment-targeting strategy for KM reconstruction through rational hydrogel design, offering a therapeutic platform for functional KM regeneration.
Spatial transcriptomics (ST) enables the study of tissue architecture by resolving gene expression in space, but current ST platforms are constrained by limited sequencing depth and indirect single-cell identification. Existing deconvolution methods that integrate single-cell RNA sequencing (scRNA-seq) data with ST often overlook the biological principle that cells in communication with each other tend to be closer spatially. Here we introduce SPADE, a deep learning framework that aligns scRNA-seq data to spatial locations by jointly modeling expression similarity between scRNA-seq and ST data and concordance between the spot distance and cell-cell communication (CCC) patterns. SPADE also enables quantitative characterization of CCC across spots and regions. Evaluations on 55 simulated and real datasets show that SPADE achieves strong performance in recovering region-specific cell-type patterns and enhancing spatial gene expression profiles compared with existing methods. In the breast cancer datasets, SPADE demonstrates a unique advantage in identifying tumor-infiltrating immune cells and tertiary lymphoid structures. In the colorectal cancer liver metastasis dataset, SPADE distinguishes tumor heterogeneity with region-specific CCC events and describes the general CCC landscape in the tissue. Overall, SPADE highlights the key role of spatially constrained CCC in shaping tissue organization and enables biological interpretation of spatial transcriptomics data.
Neutrophils are key inflammatory effector cells and rapidly integrate chemical signals during inflammation. They have long been suspected to use catecholamines (CAs) as immunoregulatory signals, but direct evidence for CA handling in these cells has been lacking due to the absence of suitable real-time detection tools. Here, we combine fluorescent false neurotransmitters (FFNs), near-infrared fluorescent single-walled carbon nanotube (SWCNT)-based catecholamine nanosensors, dual-color Ca2 + imaging, and transcriptomics to resolve neutrophil CA dynamics with high spatiotemporal precision. Using FFNs, we demonstrate VMAT2-dependent vesicular uptake of catecholamines within seconds as well as CA transfer between cells. Serotonin, LPS, and activated platelets trigger calcium (Ca2+) signaling and subsequent fast transient release of CAs from neutrophils, which we directly visualize in real time using SWCNT-based nanosensors. In human experimental endotoxemia, longitudinal transcriptomics reveal coordinated regulation of monoaminergic receptors, synthesis machinery, and transporters, suggesting adaptive tuning of neutrophils to inflammatory CA exposure. CAs suppress NET formation but enhance thrombin-induced platelet aggregation, and serotonin-dependent platelet-neutrophil interactions evoke CA release, establishing a paracrine feedback loop linking inflammation and coagulation. Our integrated imaging and sensing workflow provides direct evidence for rapid vesicular catecholamine communication in human neutrophils and uncovers previously unrecognized mechanistic parallels between neurons and neutrophils. It offers a broadly applicable platform to interrogate monoamine signaling in immune cells.
Polycomb Repressive Complex 2 (PRC2), containing homologous EZH1 or EZH2 as the catalytic subunit, is a conserved methyltransferase complex that represses gene transcription by transferring methyl group from SAM to H3K27. Gain-of-function mutations of EZH2 and aberrant H3K27 methylation have been linked to human cancers. SAM-competitive EZH1 and EZH2 dual inhibitors have been approved by the FDA for treating B-cell lymphoma and other cancers. Here, we characterized a small molecule C36, which potently inhibits EZH2/PRC2, but not EZH1/PRC2, with a novel SAM non-competitive mechanism. Cryo-EM structures revealed that C36 binds to a pocket at the interface of SET-Activation-Loop (SAL), stimulation-responsive motif (SRM), and I-SET domain of EZH2, and WD40 domain of EED. C36 binding induces conformational changes and disrupts allosteric communication between EZH2 and ligand-bound EED. C36 efficiently inhibits H3K27 trimethylation and PRC2 target gene expression in tumor cells and xenograft tumors with low hematotoxicity. Multi-omics analyses employing C36 uncovered the direct regulation of IFNB1 by EZH2/PRC2. The combined treatment of syngeneic LLC lung cancer with C36 and a PD-1 antibody significantly enhances anti-tumor efficacy. Our study identifies a new allosteric mechanism of PRC2 inhibition and paves the way for the development of highly selective EZH2/PRC2 inhibitors for combination therapy.
Single-molecule enzymes serve as molecular motors for long-read sequencing, where laser tolerance under high photon flux is a critical limiting factor for ultra-long reads. However, elucidating the mechanism of laser-induced enzyme inactivation remains a technical bottleneck due to the lack of long-term, high-throughput single-molecule evaluation methods to decouple intrinsic heterogeneity from photodamage. Here, a digital Single-Molecule Activity Tracker (dSMAT) is presented, combining deep learning with high-throughput digital microfluidics to enable the precision tracking of thousands of compartmentalized single-molecule reactions for 15 h. This strategy reveals a distinct photoinactivation mechanism designated as oxidative scarring through comparative tracking of individual polymerases before and after laser irradiation. This process is driven by the stochastic accumulation of photochemical lesions on redox-sensitive residues (specifically Methionine, Tryptophan, and Cysteine) within functionally accessible pathways, creating a kinetically disordered subpopulation. A synergistic reductive-antioxidant buffer system is engineered to mitigate this effect and rescue kinetic homogeneity. Quantitative cross-platform validation via single-molecule real-time sequencing confirms that dSMAT-derived kinetic metrics-including catalytic rate, heterogeneity, and temporal stability-deterministically govern sequencing read limits. This work establishes a mechanistically sound biophysical framework for the rational design of photostable molecular motors, offering a generalizable strategy for enhancing high-photon-flux enzymology across genomic and biotechnological applications.
Japanese encephalitis virus (JEV) is a neurotropic flavivirus that causes a substantial threat to human health and livestock; however, the epitranscriptomic mechanisms that support its replication remain poorly defined. Here, we identify a proviral host factor C2H2 zinc-finger protein ZNF33B that promotes JEV infection through coupling N6-methyladenosine (m6A) RNA modification to autophagy regulation. Mechanistically, ZNF33B recruits METTL14 to stabilize the METTL3-METTL14 methyltransferase complex, thereby increasing global m6A deposition. Multi-omics analyses reveal that ZNF33B selectively binds m6A-modified sites within the antiviral transcript Trim25 (c.1567 and c.1669 bp) to accelerate its decay. We further demonstrate that TRIM25 functions as an E3 ubiquitin ligase that catalyzes K48-linked ubiquitination of ATG7 at lysines 389 and 423, leading to its proteasomal degradation and ultimately suppressing autophagic flux. In contrast, ZNF33B-mediated Trim25 degradation counteracts its inhibitory effect on autophagy, creating a favorable environment for viral replication. In vivo, adeno-associated virus (AAV)-mediated ZNF33B delivery increases mouse brain m6A levels, decreases TRIM25 expression, elevates ATG7 abundance, exacerbates JEV-induced neuropathology, and accelerates mouse mortality. Together, these findings reveal a previously uncharacterized ZNF33B-m6A-TRIM25-autophagy axis that JEV hijacks to evade host antiviral responses, providing new insights into flaviviral pathogenesis and potential therapeutic targets.
Liver metastasis remains the primary cause of death in pancreatic cancer. Collagen deposition by activated hepatic stellate cells (HSCs) generates a stiff fibrotic niche that favors metastatic colonization, yet the underlying mechanisms remain incompletely understood. Using stiffness-tunable hydrogels, it is shown that elevated substrate stiffness activates HSCs and establishes a self-reinforcing loop of matrix stiffening. Mechanistically, stiffness triggers Piezo1-mediated Ca2 + influx, induces endoplasmic reticulum stress (ERS), and activates the IRE1α-XBP1 pathway to upregulate glia maturation factor gamma (GMFG) transcription and secretion. GMFG is transported into pancreatic cancer cells where it binds to intracellular tensin-4 (TNS4), promoting FAK/AKT phosphorylation and coordinating two programs critical for metastatic outgrowth: enhanced cell-ECM adhesion and increased de novo fatty acid synthesis. In mice with graded liver stiffness, pharmacological inhibition of mechanosensitive cation channels reduces metastatic burden and dampens GMFG-associated epithelial and lipogenic features, while targeting the GMFG-TNS4 axis suppresses early hepatic micrometastatic seeding and long-term liver metastasis burden. Together, these findings define a mechano-ER stress-paracrine cascade linking fibrotic stiffness to pro-colonization signaling, highlighting the Piezo1-GMFG-TNS4 pathway as a therapeutic vulnerability in PDAC liver metastasis.
Tandem duplication of tailoring enzymes allows evolutionary innovation that diversifies plant specialized metabolism. Here, we present an interesting example of how tandem duplicated UDP-glycosyltransferases undergo neofunctionalization and shape the chemical diversity of triterpenoid saponins in the Cucurbitaceae family. A chromosome-level genome of Siraitia grosvenorii was assembled and aligned with multiple cucurbit genomes, revealing a specific UGT73AM tandem duplication responsible for regio-selective glycosylation (e.g. the rare 1,4-linked disaccharide) of diverse saponins such as mogrosides, ginsenosides, and momordicines. Comparative genomics depicted the evolutionary trajectory of a universal saponin-biosynthesizing UGT73 tandem arrays syntenously preserved across core eudicots, where lineage-specific UGT copies contribute to distinct metabolic phenotypes. A crystal structure of SgUGT73AM30 (mogrol 25-O-glycosyltransferase) in complex with UDP and mogrol was obtained to elucidate the molecular basis of the regio-specific decoration on vicinal diol of the substrates. Altogether, these findings provide insights into tandem duplication-driven diversification of glycosyltransferases and lay the foundation for engineered glycosylation of valuable triterpenoid saponins.
High-voltage organic cathodes based on stable nitroxyl radicals are promising candidates for sustainable energy storage. However, porous TEMPO-based frameworks remain underdeveloped compared to linear polymers, and the role of electrolyte anions in governing their performance is poorly understood. Herein, two imine-linked covalent organic frameworks (COFs) are post-synthetically functionalized with N3-TEMPO via click chemistry, affording crystalline, porous TEMPO-TB and TEMPO-TP COFs with uniformly distributed redox-active sites. When evaluated as cathodes for Li-organic batteries, both materials exhibit reversible p-type redox activity at ∼3.6-3.7 V vs. Li/Li+. A systematic comparison of LiX electrolytes (X = PF6 -, ClO4 -, BF4 -, DFOB-, and TFSI-) in carbonate-based media reveals strong anion-dependent electrochemical behavior. Among the electrolytes studied, LiDFOB provides the best balance of capacity, rate capability, and cycling stability, attributed to favorable anion-coupled charge storage, interfacial charge-transfer behavior, and pseudocapacitive contributions. Binder-free buckypaper electrodes enable up to 80 wt.% active material and mass loadings of 40 mg cm-2, while maintaining ∼1.3 mAh cm-2. This represents the highest reported mass loading and areal capacity for TEMPO-based cathodes in lithium batteries. These results establish framework-electrolyte matching as a key design principle for high-voltage TEMPO-based cathodes and provide new guidance for the development of radical-functionalized organic battery materials.
Indoor humidity strongly affects both building energy demand and occupant health in hot and humid climates, but it remains largely controlled by energy-intensive mechanical systems and is seldom exploited as a source of freshwater. Here, we demonstrate a passive humidity regulation and freshwater harvesting strategy enabled by cost-effective adsorption films (18.79$·m-2) that couple moisture capture, storage, and release within indoor environments. The films rapidly reduce relative humidity from 90.7% to 21.6% within one hour while maintaining long-term operational stability and release the captured moisture to produce freshwater at a rate of approximately 1.1 kg·m-2·day-1 under ambient conditions. Leveraging the harvested water, an autonomous plant irrigation system is achieved, enabling sustained regulation of indoor CO2 concentration (about 920.2 ppm) without manual intervention. Global-scale projections indicate that this passive approach could reduce building energy consumption by up to 29.9 kWh·year-1·m-2 and associated carbon emissions by 16.5 kg·year-1·m-2, with an exceptionally short payback period of 48 days. This work reframes indoor humidity from a latent load to a recoverable resource, offering an integrated route towards water harvesting, energy reduction and healthier indoor environments in sustainable buildings.
Mitochondrial damage in dorsal root ganglion (DRG) neurons contributes to the pathogenesis of paclitaxel (PTX)-induced peripheral neuropathic pain (PIPNP). Fibroblast growth factor 13 (FGF13), abundantly expressed in DRG neurons, is crucial for the regulation of somatosensation; however, its role in PIPNP remains unclear. Here, we demonstrated that FGF13 expression is upregulated in DRG neurons of PIPNP model mice. Conditional knockout of Fgf13 in DRG neurons effectively alleviates PTX-induced mitochondrial damage and neuropathic pain. RNA sequencing analysis revealed that mitophagy mediates the regulatory effects of FGF13 in PIPNP. Mechanistically, FGF13 physically interacts with vasohibin 1 (VASH1), regulating the binding of VASH1 to microtubules and promoting microtubule detyrosination. FGF13 ablation disrupts assembly of the FGF13-VASH1-α-tubulin ternary complex, impairing VASH1-mediated microtubule detyrosination and increasing microtubule tyrosination. The resulting accumulation of tyrosinated microtubules facilitates kinesin-3 (KIF1A)-driven lysosomal trafficking, which in turn promotes mitophagy activation and ultimately ameliorates PTX-induced mitochondrial damage and PIPNP. Furthermore, VASH1 overexpression in DRG neurons reversed the alleviating effects of FGF13 deficiency on PTX-induced mitochondrial damage and PIPNP. In summary, our findings demonstrate that FGF13 deficiency alleviates mitochondrial dysfunction and PIPNP by suppressing VASH1-dependent microtubule detyrosination and subsequently activating mitophagy. Targeting FGF13 may be a promising therapeutic strategy for PIPNP.
Asymmetric supercapacitors (ASCs) comprising two different pseudocapacitive electrodes offer a promising route toward higher energy density, yet the serious self-discharge behavior and poor cycle life hinder their wider applications. This work proposes a controllable carbon shell encapsulation strategy based on rapid Joule heating calcination to construct high-performance ASC with suppressed self-discharge and robust cycling stability. As a result, the as-assembled ASC (H-Fe3O4@C-15//H-NiCo2S4@C-40) exhibits a maximum energy density of 105.6 W h kg-1 at a power density of 749 W kg-1, and long cycling lifespan with 93.7% capacitance retention after 25 000 cycles. Furthermore, current ASC also demonstrates moderated self-discharge, with its open-circuit voltage decaying from 1.48 to 0.75 V over 31102 s. Theoretically, the density functional theory (DFT) results adequately uncover that the significantly improved self-discharge performance should originate from the increased adsorption energy between the electrode and the electrolyte ions. Meaningfully, this controllable carbon shell encapsulation strategy represents a universal and feasible approach to effectively suppress self-discharge and extend the cycle life of ASC.
Energy-efficient and adaptive neuromorphic hardware requires material platforms that can intrinsically integrate transient neural dynamics with stable long-term memory within a single device architecture. Here, a cross-point nanoporous SiO2 memristor that unifies volatile and nonvolatile switching behaviors within a single material platform is reported. The engineered nanoporous framework provides well-defined ion migration pathways, enabling controlled modulation of conductive filaments and reversible transitions between short-term plasticity (STP) and long-term plasticity (LTP) through simple compliance-current tuning. Leveraging this dual-mode functionality, the volatile dynamics of the nanoporous SiO2 memristors are employed directly as a physical reservoir, while the nonvolatile conductance states serve as synaptic weights in the readout layer. Using a conductance-aware training scheme, reservoir computing (RC) is demonstrated on the same device platform, achieving 93.7% accuracy in MNIST handwritten-digit recognition. Beyond standard benchmark datasets, the system further enables ECG temporal biosignal classification, reaching over 88% accuracy in distinguishing normal and abnormal heartbeat patterns. These results establish a single-material, CMOS-compatible neuromorphic platform capable of integrating dynamic processing with persistent memory, offering a scalable and low-power pathway toward compact intelligent edge computing hardware.
The visualization of mechanical stress in soft materials is highly desirable; however, real-time optical readouts using conventional sensing approaches are problematic because mechanophore-based systems typically require strong threshold-type activation with slow recovery. Herein, we report supramolecular hydrogels that enable the continuous and reversible visualization of mechanical stress in real time via stretch-induced phase separation. Supramolecular switching mechanotransduction (SSM) has been proposed as the key mechanism. Mechanical stimuli are transduced into a distinct network state transition through host-guest complexes between β-cyclodextrin and adamantane as supramolecular switches. In this design, guest-functionalized polymers undergo on-off transition between the hydrated and dehydrated states via host-guest complexation and decomplexation. Responsive polymers, incorporated into the hydrogel network via supramolecular bonds, function as reversible cross-links and switches. Upon stretching, the hydrogels macroscopically transition from transparent to opaque owing to the dehydration-induced heterogeneity within the responsive domains. The linear and reversible changes in the opacity with applied stress-attributable to sacrificial and reversible supramolecular switching-enable the visualization of stress distributions. This design principle offers a platform for spatiotemporally resolved mapping of the mechanical states in hydrogels with an intuitive, instrument-free readout, and lays the foundation for monitoring, timely intervention, and safer operation of soft-material systems.
Hierarchical composites composed of covalent-organic frameworks (COFs) and metal-organic frameworks (MOFs) (COF@MOF), exhibiting superior tunability in terms of pore structure and electronic distribution, have gained increasing attention in various fields. However, the development of a universal strategy for controllable assembly of COF@MOF composites, enabling precise tuning of composition and structure, remains a significant challenge. Here, a flexible and adjustable COF@MOF synthesis strategy (metal pre‑fixation, MPF) is proposed, which facilitates the extensive growth of MOFs with varying morphologies and particle sizes on the surface of PY-COF-COOH. By employing a hypothesis-deduction approach alongside experimental and density functional theory (DFT) calculations, systematically elucidated that the MPF strategy enables controllable modulation of morphology, size, and coordination structure through ion distribution-guided synthesis (IDGS), a proposed working hypothesis of crystal-facet shielding (CFS), and aperture synergistic regulation (ASR) effects. Furthermore, the MPF-derived PY-COF-COOH@MOFs exhibit significantly enhanced CH4/C2H6/C3H8 separation performance and sustained H2O2 activation efficiency. These results demonstrate significant potential in designing high-performance COF@MOF composites and offer a comprehensive application guide for MPF-based design.
Protein aggregation drives major neurodegenerative diseases, yet most computational predictors collapse assembly into static risk scores and do not resolve the distinct structural determinants of nucleation and elongation. Here, we present SKALE 2.0, a phase-resolved geometric deep learning framework that represents proteins as multimodal structural graphs and learns mutation-induced aggregation phenotypes directly from three-dimensional topology. Across SOD1, TDP-43, MAPT, and PRNP, SKALE 2.0 recovered a conserved latent transition from nucleation to elongation while resolving distinct mutation-specific phase sensitivities. Representative protein language model, AlphaFold-derived feature, and non-phase-aware structural baselines failed to recover both phase-dependent mutation modulation and phase separability, indicating that explicit phase conditioning is essential. The learned geometry showed that nucleation is preferentially coupled to buried hydrophobic perturbations, whereas elongation is shaped by solvent-accessible interfaces that support fibril propagation. This framework explains how pathogenic variants can remain globally folded yet acquire aggregation competence through localized structural rewiring. Recombinant SOD1 experiments validated predicted suppressor, enhancer, and phase-switch mutations, demonstrating that initiation and propagation can be tuned independently. SKALE 2.0 links atomic topology to phase-specific assembly kinetics and enables a constraint-aware design of aggregation suppressors.
ABSTRACT The potential for fibrosis across most organ systems may stem from connections to wound healing and the widespread presence of vascular endothelium. Endothelial cells (ECs) and angiogenesis have been heavily implicated in many organ‐specific fibrotic conditions, but little has been established in terms of how EC phenotype governs tissue healing vs. fibrosis. Here, we examined a murine lung injury model enabling EC lineage tracing and observed the invasion of aberrant ECs from the bronchial microvasculature following injury, along with concurrent densification of surrounding extracellular matrix fibers. To investigate mechanisms governing their appearance, we established a microphysiological system of human microvessels embedded within a tunable stromal matrix and found that heightened fiber density drives endothelial to mesenchymal transition to promote aberrant tip EC (ATEC) invasion into the matrix. ATECs remained adherent to fibrotic matrix and possessed a pro‐inflammatory phenotype that secretes TGF‐β2. Mechanistically, we identify ATEC formation was gated by destabilization of EC adherens junctions upon adhesion to fibrous matrix and associated regulation of TGF‐β signaling through a novel VE‐cadherin – TGF‐βR2 interaction. Altogether, this work identifies how enhanced fiber density associated with fibrogenesis regulates EC phenotype to generate pro‐inflammatory ATECs and suggests new contributions of ECs to fibrotic progression.
