Biology subjects

Tian, L.

Publications and source records attributed to Tian, L..

At least 19 recordsLinked to original sources

bioRxiv2026

Targeting the sinusoidal-to-capillarized transition to normalize vessels in breast cancer liver metastasis

Liver metastases exhibit marked resistance to anti-VEGF anti-angiogenic therapies, yet the underlying mechanisms remain poorly understood. By combining single-cell RNA sequencing of human breast cancer liver metastases with a liver sinusoidal endothelial cell (LSEC) lineage-tracing mouse model, we show that liver metastasis-associated vessels originate from Lyve1+ sinusoidal endothelial cells but undergo a sinusoidal-to-capillarized transition as they grow into the metastatic nodule, characterized by enhanced inflammatory signaling, loss of sinusoidal markers, and acquisition of capillarized markers. We functionally perturbed each of these abnormalities individually. Unexpectedly, ectopic Lyve1 overexpression paradoxically enhanced hyaluronan and other extracellular matrix deposition in the tumor microenvironment, suggesting a compensatory mechanism. Similarly, targeting TNF signaling, a hallmark inflammatory pathway in liver metastasis, dampened the inflammatory state of metastasis-associated vessels but promoted metastatic outgrowth, indicating that TNF blockade is not a viable strategy for metastasis treatment. Strikingly, inhibition of the capillary tip cell marker Esm1 improved vessel perfusion and enhanced chemotherapeutic drug delivery, two defining features of vascular normalization. Pan-cancer analysis further revealed that Esm1 expression is restricted to tip cells and correlates with poor patient prognosis, positioning Esm1 as a candidate target for promoting vascular normalization. Collectively, our findings reveal the vascular plasticity of liver metastasis and, although not all abnormal features are actionable, we propose targeting Esm1 to inhibit sprouting angiogenesis and to normalize dysfunctional liver metastasis vessels, offering a promising therapeutic avenue for liver malignancies.

cancer biology↗

bioRxiv2026

Cause-excluded Effect Evaluation: a framework for the faithful assessment of genetic perturbation target identification

Identifying which gene drives a desired cell-state transition is a foundational inverse problem in target discovery and cell-state engineering, yet far less developed than forward prediction. We show that prevailing benchmarks inflate performance through two artefacts: training leakage, where test-perturbation cells are seen during fitting, and evaluation leakage, where the perturbed gene's own expression change exposes the answer. Here, we introduce Cause-excluded Effect Evaluation (CEE), a framework that separates train and test perturbation targets, removes target-gene self-effects from expression profiles and scores candidates in downstream response space. With CEE, we find that simple structure-based methods consistently outperform complex deep learning models across 16 single-cell perturbation datasets, revealing that the recoverable signal is governed by low-rank and additive structure in perturbation responses rather than by model complexity. Our work disentangles evaluation artefacts from genuine predictive capacity and offers a design principle for developing perturbation target identification methods.

bioinformatics↗

bioRxiv2026

Restoration of deficient IL-15 signaling enhances innate lymphoid cell immunity in brain metastasis

Natural killer (NK) cells and type 1 innate lymphoid cells (ILC1s), collectively known as group 1 innate lymphoid cells (ILCs), are critical mediators of anti-tumor immunity, yet their role in brain metastasis remains poorly defined. Using single-cell RNA sequencing and functional depletion models, we show that group 1 ILCs fail to provide adequate tumor control in breast cancer brain metastasis (BCBM). In contrast to the liver, where diverse, functionally competent group 1 ILC populations facilitate excellent metastatic control, group 1 ILCs in BCBM show limited diversity and profound tissue-specific dysfunction such that their depletion had no effect on tumor burden despite modest parenchymal infiltration. We attribute this group 1 ILC dysfunction to severely restricted IL-15 availability, with IL-15R protein detected at 500-fold lower concentration in the brain than in the liver. Local IL-15 complex administration restored group 1 ILC accumulation, cytotoxicity, and tissue residency, significantly reducing brain metastatic burden independent of T cells. Intracranial delivery of IL-15-secreting induced pluripotent stem-cell-derived (iPSC)-NK cells, already in clinical testing for non-small cell lung cancer, significantly reduced tumor burden. Collectively, these findings reveal how tissue-specific differences in cytokine availability shape group 1 ILC population diversity and disease response, establishing localized IL-15 signaling as a compelling strategy to unlock innate immunity against intracranial malignancies.

immunology↗

bioRxiv2026

A closed-loop language-model agent for target-specific multi-objective hit-to-lead optimization

Hit-to-lead optimization is a multi-objective molecular-design problem in which binding-related scores must be considered together with drug-likeness and synthetic accessibility. We developed a closed-loop language-model medicinal-chemistry agent that combines target and starting-hit context, PubMed retrieval, molecular-property tools, docking, and feedback from previously scored analogues. In a matched computational benchmark comprising six kinase targets, six methods, a 40-candidate budget, and six independent campaigns per method, the agent achieved the highest mean aggregate score under a pre-specified composite objective on four targets. Re-analysis of the same candidate sets with drug-likeness-gated docking and docking-drug-likeness hypervolume produced different target-level winners, showing that evaluation rules can change the comparative interpretation of an optimization campaign. Candidate distributions and component ablations indicated that the agent's advantage was associated primarily with retention of scored-candidate feedback and higher-QED regions rather than with docking alone. An EGFR case study illustrates a scaffold-preserving in-silico optimization trajectory from erlotinib. The study is a computational benchmark and evaluation framework; the prioritized molecules are hypotheses for subsequent experimental testing, not experimentally validated leads.

bioinformatics↗

bioRxiv2026

Benchmarking Uncertainty and Improving Risk Ranking in Multi-task Bioactivity Prediction

Multi-task bioactivity prediction transfers information across assays, but experimental prioritization also requires uncertainty estimates that identify unreliable predictions. We benchmarked prediction accuracy and uncertainty calibration for five multi-task predictors on 100 ChEMBL27 test assays and 100 CHEMBL37 assays under the random split and the clustering split. Calibration used native, Deep Ensemble and MC Dropout uncertainty; risk ranking was compared for two Gaussian process (GP) backbones. Adaptive deep kernel fitting (ADKF) achieved the best mean calibration results, while deep kernel transfer (DKT) ranked errors more effectively than ADKF with native uncertainty. We also introduce Influence Calibrated Support Reconstruction (ICSR), a risk score for kernel-based predictors. ICSR combines measured support reconstruction errors with query-specific influence and stabilizes their weighted average toward the full-support mean to adjust Gaussian process uncertainty while preserving predicted activities. With DKT and ADKF, ICSR improved all three mean risk-ranking metrics over native uncertainty and an adapted neighborhood comparator in every panel and split. ICSR achieved a mean half-query MAE reduction (R50) of 14.2-19.6%, where R50 measures the percentage decrease in mean absolute error (MAE) after retaining the lowest-risk half of the queries. A retrospectively selected SARS-CoV-2 main protease case further illustrated its use for selecting more reliable predictions. The benchmark supports joint assessment of calibration and error ranking, while ICSR improves selective use of kernel-based bioactivity predictions.

bioinformatics↗

bioRxiv2026

MINT infers the latent single-cell spatial transcriptome from paired histology

Sequencing-based spatial transcriptomics captures transcriptome-wide molecular information, yet the corresponding cell-resolved tissue map remains incomplete. Histology records dense cellular morphology, tissue architecture and local context. Here, we present MINT (Morphological Integrative Network for Transcriptomics), which combines this latent cellular organisation with sample-specific spatial measurements to infer tissue-wide, cell-resolved transcriptomes without an external single-cell reference. MINT reconstructs expression from sparse Slide-tags profiles across unmatched adjacent sections; its shared nucleus-centred representation also resolves pooled Visium measurements into cell-level profiles. MINT improved Slide-tags alignment and expression recovery under controlled nuclear loss and cross-section perturbation. It also outperformed reference-free Visium disaggregation methods against Xenium ground truth and was validated with Visium HD sequencing. In human lung adenocarcinoma, MINT reconstructed the tumour immune microenvironment cell by cell and resolved tertiary lymphoid structures and their maturity. These results establish paired histology as a quantitative, sample-intrinsic complement to molecularly rich but cellularly incomplete spatial-transcriptomic assays.

bioinformatics↗

bioRxiv2026

Dual-Mode Bio-CM2: Multimodal Computational Miniature Mesoscope for Fluorescence and Label-Free Imaging

Understanding biological function often requires complementary information on molecular identity, morphology, behavior, and physiology acquired simultaneously from the same specimen. Achieving this capability in miniature microscopes remains challenging because integrating multiple imaging contrasts typically increases optical complexity while compromising field of view (FOV), spatial resolution, or form factor. Here we present dual-mode Bio-CM2, a multimodal computational miniature mesoscope that simultaneously captures fluorescence and label-free reflectance through a shared optical architecture based on distributed computational optics. The system enables frame-interleaved, co-registered dual-mode imaging over a [~]7.5 x 10 mm2 FOV with [~]6 m lateral resolution. We demonstrate the versatility of dual-mode Bio-CM2 across diverse biological systems. In freely moving Caenorhabditis elegans, reflectance captures body posture and locomotor behavior that contextualize fluorescently labeled protein aggregates. In freely swimming larval zebrafish, reflectance captures whole-body morphology and swimming behavior, while fluorescence visualizes cardiac activity. In head-fixed mice, simultaneous fluorescence and reflectance imaging across the bilateral dorsal cortex provides complementary measurements of neuronal calcium activity and intrinsic hemodynamic signals. By integrating molecularly specific fluorescence with complementary reflectance contrast, dual-mode Bio-CM2 extends distributed computational optics into a scalable platform for multimodal miniature imaging.

bioengineering↗

bioRxiv2026

Spatially confined dopamine from locus coeruleus axons selectively shapes hippocampal dynamics and action timing

Neuromodulatory systems project broadly, yet computations engage specific neuronal populations. Whether neuromodulation can selectively influence task-engaged neurons remains unclear. During goal-directed navigation, neurons in the locus coeruleus (LC), the brain's principal norepinephrine source, responded at navigation onset. Concurrently, dopamine transients arose in micrometer-scale domains around LC axons in CA1. These transients preferentially enhanced nearby CA1 neurons with ramping dynamics, whose higher activity predicted later reward-anticipatory action. Brief LC activation evoked local dopamine, enhanced ramping dynamics, and delayed action initiation seconds later. Blocking D1-like, but not adrenergic, receptors weakened ramping dynamics and impaired reward anticipation. Modeling showed that selectivity can emerge when localized dopamine coincides with strong ongoing neuronal activity, without dedicated wiring. Thus, a widely projecting system can selectively modulate task-engaged neurons and shape behavior.

neuroscience↗

bioRxiv2026

Unlocking Sensitive Data with SPHERE in the Age of AI

Sensitive human data underpin discoveries across medicine, biology and the social sciences, yet privacy regulation often prevents sharing them with collaborators or artificial intelligence (AI) systems. We introduce SPHERE, a model-free method that makes sensitive datasets directly usable by AI and shareable for open science as a synthetic twin, while the original records never leave the local environment. Across 33 datasets spanning five scientific domains, SPHERE protects individual privacy against adversarial re-identification attacks while preserving the data's statistical structure: means, variances and correlations are reproduced exactly, effect size and P value in linear statistical analysis is numerically identical, nonlinear machine-learning utility is retained, and each twin is generated in seconds on a laptop. Frontier AI agents running on the twin reach the same scientific conclusions as on the original records. Analyses of the twin reproduce genome- and proteome-wide results at UK Biobank scale and recover the findings of landmark studies across three independent cohorts and consortia. The approach also extends to deep-learning embeddings across language, vision and time-series, with minimal utility loss. We make the Stanford Alzheimer's Disease Research Center cohort openly available for the first time, as a SPHERE twin spanning nine modalities that any registered researcher can analyze without an approval process. We release SPHERE with certification of each twin's privacy and fidelity, and an AI agent that autonomously executes research tasks on sensitive data without ever accessing it. Sensitive datasets that are currently closed to research could thus become routine inputs to open science and AI to enable key discoveries.

bioinformatics↗

bioRxiv2026

Combined Image-Based Profiling and Biochemical Analysis of GCaMP Overexpression Effects on Mammalian Cells

Protein-based fluorescent sensors are a powerful addition to the biology toolbox for their ability to be stably expressed within living organisms, tissues, cells, and subcellular compartments, with the capacity to report on the presence of specific target molecules or other analytes. At the same time, sensor components will unavoidably present opportunities for unintended interaction with endogenous cellular machinery, potentially confounding both sensor function and cell health. Interactions with host components may not be readily predictable during the sensor design process, especially when simultaneously optimizing many other sensor parameters such as fluorescence response, dynamic range, and kinetics. Characterizing effects of sensor expression on cells is currently a laborious ad hoc process; new methods to characterize the cell expression effects of sensors and their variants could dramatically improve sensor design pipelines, laying the groundwork to recognize potentially problematic expression side effects earlier in the iterative design and testing workflow. Here, we take a dual high-content imaging-based and biochemical approach to examine sensor interactions with native cell biology, focusing on the widely used GCaMP calcium sensor. We identify a morphology-based signature of the cellular effects of high sensor expression in a neuroblastoma cell line. Subsequently, we identify biochemical interactions between GCaMP and a component of the mammalian cytoskeleton and track morphological features in sensor-expressing cells that lack these structural components. Our findings present an entry point for engineering new minimally cross-reactive sensor versions given a contextual biological understanding of sensor overexpression. We anticipate that as this and related workflows are incorporated into sensor engineering pipelines, bioorthogonality can be more systematically assessed and prioritized in diverse sensor scaffolds.

cell biology↗

bioRxiv2026

Genetic Disruption at the CIP2A Locus Modulates T Cell Responses and Attenuates Experimental Autoimmune Encephalomyelitis

Multiple sclerosis (MS) is a chronic autoimmune disease of the central nervous system (CNS) driven by pathogenic T cell-mediated inflammation. Fingolimod (FTY720), an approved therapy for MS, is an established activator of protein phosphatase 2A (PP2A). However the contribution of PP2A in autoimmune neuroinflammation remains incompletely understood. Here, we addressed this question using experimental autoimmune encephalomyelitis (EAE), a murine model of MS, in mice carrying a genetic disruption of the locus encoding cancerous inhibitor of protein phosphatase 2A (CIP2A), an endogenous inhibitor of PP2A. Mice with disruption of the CIP2A locus, the knock out (KO) mice, exhibited attenuated EAE severity compared with wild-type (WT) controls. Histological and flow-cytometric analyses revealed markedly reduced infiltration of mononuclear cells, including CD4 and CD4CXCR6 encephalitogenic T cells, in the CNS of diseased KO mice. Reduced numbers of these T cell populations were also observed in peripheral lymphoid organs of the Cip2a-deficient mice during EAE, while T cell abundance was comparable under steady-state conditions, suggesting impaired activation-induced expansion rather than altered homeostasis or migration. Single-cell RNA sequencing of CNS and lymph node immune cells revealed changes in cell-type abundance and gene expression. Notably, Il17a expression was reduced in CNS CD8+ T cells and showed a similar trend in {gamma}{delta} T cells. Together, our findings reveal that genetic disruption at the CIP2A locus attenuates EAE, possibly by limiting the expansion and accumulation of encephalitogenic T cell populations in CNS. These results identify the CIP2A locus as a previously unrecognized regulator of T cell-driven autoimmune neuroinflammation and provide new insights into mechanisms that restrain pathogenic T cell responses during EAE.

immunology↗

bioRxiv2026

Virtual-cell verification enables self-auditing AI discovery for immune rejuvenation

Artificial-intelligence agents propose drug-discovery hypotheses faster than experiments can test them, yet their conclusions are rarely verified, against the underlying biology, the predicted perturbation, or the agent's own scoring logic. We close this verification gap with an agentic framework built on three verifiers. First, PACE, a phenotype verifier, resolves immune aging into ten directionally scored, cell-type-resolved gene-set modules, selected for cross-cohort stability across four PBMC cohorts, and outperforms five established aging clocks in an independent in-house aging cohort of 434 elderly donors. Second, CellQ, a virtual-cell verifier built with multi-modal LLM, compresses each single-cell transcriptome into eight discrete tokens aligned to a language model's vocabulary through residual vector quantization; it attains state-of-the-art perturbation prediction and uniquely resolves the weak, module-level shifts that differential-expression recovery misses. Third, an Analyzer-Planner-Auditor agent verifies its own scoring logic: screening 110 compounds in primary human PBMCs, it found aged-down modules more reversible than aged-up modules and revised its objective from an equal-weight mean to a balance-constrained minimum, a self-correction that generalized to an independent 13-compound T-cell assay. By verifying its predictions and its own objective against experiment, the framework points beyond hypothesis-generating AI toward self-correcting AI scientists whose objectives could continuously evolve.

bioinformatics↗

bioRxiv2026

Emergence of tunneling nanotube-like intervesicular connections

Tunneling nanotubes (TNTs) are actin-supported membrane bridges that mediate long-range intercellular communication and direct transfer of signaling molecules, organelles, and pathogenic cargo, yet the physicochemical mechanisms underlying their formation and organization remain poorly understood. Here we show that TNT-like intervesicular connections emerge from minimal physicochemical interactions between actin filaments and lipid membranes. Upon Mg2+ exposure, actin-encapsulating vesicles spontaneously generated actin bundle-embedded lipid nanotubes (AT-LNTs) that formed stable intervesicular networks. Mg2+ simultaneously induced actin polymerization, filament bundling, and electrostatic recruitment of F-actin to phosphatidylcholine membranes, enabling membrane tubulation without actin-binding proteins. Systematic perturbation of membrane phase, membrane tension, Mg2+ concentration, ionic strength, and actin concentration revealed that AT-LNT formation occurs only within a narrow physicochemical regime where membrane deformation and actin-membrane coupling are simultaneously permissive. The resulting AT-LNTs reproduced key structural and dynamic features of cellular TNTs, including bundled organization, helical unwinding, lumenal diffusion, and spontaneous bridging between synthetic vesicles and living cells. These findings establish a minimal biophysical framework for understanding the emergence of intercellular membrane connections in living systems and provide a foundation for engineering communication between synthetic and living cell.

synthetic biology↗

bioRxiv2026

Bio-CM{superscript 2}: Distributed computational optics for cortex-widecellular imaging

Understanding distributed biological systems, particularly neural circuits, requires simultaneous cellular-resolution imaging across millimeter-scale fields of view (FOV). Existing miniature microscopes remain fundamentally constrained by trade-offs among FOV, spatial resolution, and optical complexity, limiting their ability to bridge cellular microscopy with cortex-scale imaging. Here we introduce distributed computational optics, a framework that distributes image formation across coordinated optical modules and computationally integrates their measurements into a unified image. We realize this framework in Bio-CM2, a computational miniature mesoscope that partitions the imaging field across four optical modules while converging their measurements onto a common image sensor. This architecture overcomes the aberration-scaling limitations of conventional miniature optics while avoiding the hardware complexity of multi-camera systems and the contrast degradation associated with optical multiplexing. Bio-CM2 achieves a 7.5 x 10 mm2 FOV while enabling cellular-resolution in vivo imaging at video rates. We demonstrate its utility through two complementary imaging modalities in head-fixed mice: cortex-wide functional vascular imaging, enabling simultaneous quantification of pial arteriole vasomotion and mesoscale hemodynamic functional connectivity, and cellular-resolution calcium imaging, resolving the activity of over 3,000 neurons together with mesoscale neuronal functional connectivity. We further demonstrate the versatility of the platform through cellular-resolution imaging of entire coronal mouse brain sections, population-scale imaging of freely behaving Caenorhabditis elegans, and odor-evoked calcium imaging of the main olfactory bulb in head-fixed mice, highlighting its broad applicability across diverse biological systems and imaging modalities. By overcoming the conventional trade-off between FOV and spatial resolution in a compact miniature platform, Bio-CM2 establishes distributed computational optics as a scalable framework for multiscale biological imaging.

bioengineering↗

bioRxiv2026

Enkephalin constrains fear learning via volume transmission to the lateral amygdala

Fear learning involves the formation of associations between cues and aversive outcomes, a process that must be tightly regulated to prevent excessive or generalised fear. Dopamine release in the lateral amygdala (LA) drives fear acquisition, whereas endogenous opioids constrain it. However, whether opioids are dynamically released within the amygdala circuits during learning, and how they exert this control remain unclear. Here we show that met-enkephalin is locally released within the amygdala during auditory fear conditioning, with signals shifting from the aversive outcome to its predictive cue as learning progresses. The amygdalo-striatal transition zone (ASt), is the principal source of this enkephalin, released from medium spiny neurons receiving strong auditory thalamic input. This enkephalin spreads from the ASt to the LA via volume transmission. Selective knockdown in the ASt abolished opioid signals and enhanced fear learning, demonstrating that this diffuse signal constrains fear memory formation. We further show that enkephalin suppresses dopamine release in both the ASt and LA via {micro}-opioid receptors, identifying the ASt as a neuromodulator hub coordinating opioid and dopaminergic signalling across amygdala fear circuits. Although demonstrated here for auditory fear learning, the ASt receives multimodal sensory input, suggesting a broader mechanism through which sensory experience recruits enkephalin release to gate associative learning

neuroscience↗

bioRxiv2026

First community challenge for automated virus taxonomy

The rapid rate of virus discovery renders manual curation by taxonomy experts increasingly impractical, creating a need for reliable software that can reproducibly assign viral contigs to taxa at all fifteen ranks of the virus taxonomy. We led an open community challenge for the computational taxonomic classification of viruses and assembled a dataset of virus sequences combining expert-curated and metagenomic sequences. Seventeen teams contributed a total of thirty-four automated, fully reproducible classification pipelines. Most tools correctly assigned viruses belonging to established species, genera, or families, but viruses that are unclassified at those lower ranks remain challenging. This study provides datasets, open-source software, novel approaches, and recommendations to benchmark computational taxonomic classification of viruses, and support organizing the many viruses discovered in big omics data.

microbiology↗

bioRxiv2026

Long-read genome sequencing reveals complex variability in lentiviral provirus insertions in deeply characterized Clonal CD19 CAR-T vector copy number reference cell lines

Chimeric antigen receptor (CAR)-T cell therapy is an important therapy involving provirus insertions in the genome. Characterizing these insertions is important for understanding the safety and efficacy of cell products, but the sequence of these insertions is not fully characterized. We generate clonal CD19 CAR-T cell lines with one to five copies of the lentiviral provirus insertions. Vector copy number (VCN) was determined by droplet digital PCR (ddPCR), which revealed that most of the elements (LTR, Psi, RRE, CD19, and WPRE) were 1 to 5 or 6 copies per cell. DdPCR data also revealed that there was an additional copy of eGFP gene in VCN4 and VCN5 cell lines. To fully characterize the sequences and locations of these insertions, we use short- and long-read whole genome sequencing as well as digital PCR and flow cytometry. Long-reads enable full resolution of each insertion, and we find that of 10 insertion events, 3 have the expected insertion sequence, 2 differ from the expected only in small variants, 3 have structural abnormalities, and 2 are small partial insertions missed by most other approaches. One particularly important structural abnormality resolved only by long-reads is a 724 bp deletion of the EF1 promoter disrupting expression of the CD19 CAR. Standard short-read and ddPCR approaches miss this deletion due to this commonly used promoter being in the unengineered human genome. These results demonstrated that these cell lines are suitable VCN reference standards for 1 to 5 or 6 copies and highlight the utility of long-read sequencing in characterizing both quantity and quality of insertions in lentiviral-engineered cells.

genomics↗

bioRxiv2026

Hypoxia Promotes Wound Healing via Dynamical-Mechanical Balance and Adhesion Remodeling

Wound healing is a tightly orchestrated physiological process governed by dynamic cell-cell and cell-matrix interactions, yet how hypoxic microenvironments regulate migratory behavior in cells with latent lineage plasticity remains fully elucidated. Here, utilizing human embryonic kidney (HEK293T) and Madin-Darby Canine Kidney (MDCK) cells as a genetically tractable model, we investigate the cellular and molecular mechanisms driving hypoxia-accelerated collective wound repair. Time-lapse live-cell imaging and morphometric analyses reveal that hypoxic exposure significantly accelerates migration, shifts cell cycle dynamics toward the S/G2/M proliferative phases, and induces pronounced morphological spreading. Mechanistically, hypoxia induces a persistent, time-dependent downregulation of the desmosomal cadherin desmoglein-2 (DSG2), thereby weakening intercellular cohesion. Concurrently, the cell-matrix adhesion molecule integrin {beta}3 (ITGB3) exhibits a distinctive biphasic kinetic response--an initial sharp upregulation followed by a sustained decline-which serves to optimize focal adhesion traction and subsequent trailing-edge detachment. Transcriptomic profiling further corroborates these phenotypic transitions, demonstrating a global enrichment of gene networks associated with plasma-membrane adhesion organization, receptor activity, and ion homeostasis that independently mirrors the altered junctional dynamics and accelerated cellular responses. Collectively, our findings uncover a novel cooperative mechanism by which hypoxic stress coordinates cell-cell and cell-matrix adhesion remodeling to facilitate efficient tissue repair, highlighting the valuable utility of plastic cellular models in decoding microenvironmental stress responses.

Cell Biology↗