Biology subjects

Pan, Y.

Publications and source records attributed to Pan, Y..

At least 37 records · Page 2Linked to original sources

bioRxiv2026

Linker-Length Landscape Mapping Enables Coupling of Diverse Synthetic Chemically Induced Dimerization Systems to Molecular Readouts

Programmable molecular biology increasingly requires strategies for converting engineered recognition or proximity modules into measurable outputs, particularly within transcriptional regulation, RNA imaging, and CRISPR-associated systems. Synthetic chemically induced dimerization (CID) systems provide a class of programmable recognition modules for such applications, yet generalized strategies for coupling structurally diverse CIDs to functional readouts remain limited. Here, we introduce a CID-to-output conversion strategy based on engineering of the linker-mediated coupling interface. Using single-fluorescent-protein sensors as an experimentally tractable optical model readout, we systematically varied paired N- and C-terminal linkers flanking circularly permuted green fluorescent protein (cpGFP) to map coupling landscapes across synthetic CID systems derived from combinatorial selection and computational protein design. The results revealed strong non-additive interactions across paired linkers and suggest that linker length is a first-order determinant of CID-to-output coupling. Across nanobody-, monobody-, and de novo-designed CID architectures, this framework yielded functional sensors with dynamic ranges up to 1270% and robust responses in mammalian cells. Together, this work demonstrates that effective CID-to-output conversion can be achieved by empirically mapping the linker-mediated coupling interface, providing a practical route for adapting synthetic CID to diverse programmable molecular readouts and nucleic-acid-associated synthetic biology systems O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=94 SRC="FIGDIR/small/735888v1_ufig1.gif" ALT="Figure 1"> View larger version (25K): org.highwire.dtl.DTLVardef@1ecc594org.highwire.dtl.DTLVardef@19fb822org.highwire.dtl.DTLVardef@f1e2ddorg.highwire.dtl.DTLVardef@653131_HPS_FORMAT_FIGEXP M_FIG C_FIG

synthetic biology↗

bioRxiv2026

MaxEnt-DTD: Maximum-Entropy Estimation of Diffusion Tensor Distribution for Fiber Orientation and Microstructure Characterization

Diffusion MRI (dMRI) enables noninvasive characterization of white-matter fiber orientations and tissue microstructure, but widely used approaches, such as constrained spherical deconvolution (CSD) and parametric multicompartment models, typically address these features separately. The diffusion tensor distribution (DTD) framework jointly represents fiber orientation and microstructure, but estimating DTD from finite, noisy measurements is severely ill-posed. Existing inversion methods either rely on nonnegativity constrained basis representations, which are challenging to sale to high-dimensional and high-resolution distributions, or use sampling-based approaches with limited reliability. We propose MaxEnt-DTD, a maximum-entropy algorithm for DTD estimation from finite and noisy dMRI data. By deriving the Lagrange dual formulation, we reformulate a constrained infinite-dimensional optimization problem into a finite-dimensional unconstrained convex optimization problem, substantially reducing the parameter space and enabling tractable whole-brain DTD estimation. We evaluate MaxEnt-DTD using both synthetic and in vivo data from the Human Connectome Project protocol and a second dataset using advanced B-tensor diffusion encoding. We compare MaxEnt-DTD-derived fiber orientation distributions with results from CSD and Monte-Carlo inversion methods, and assess fiber-specific microstructure measures and rotation-invariant metrics based on the cumulants of DTD. The results demonstrate that MaxEnt-DTD provides a reliable and efficient framework for joint fiber-orientation and microstructure analysis in dMRI.

neuroscience↗

bioRxiv2026

NF1 deficiency induces metabolic reprogramming and epithelial-mesenchymal transition in glioblastoma

BackgroundMetabolic reprogramming is a common occurrence in tumor cells, where enhanced glycolysis promotes cell growth, invasion and migration. NF1 is tumor suppressor gene that downregulates the encoded neurofibromin protein. However, the effects of NF1 on energy metabolism and epithelial-mesenchymal transition (EMT) in glioblastoma multiforme (GBM), as well as the underlying molecular mechanisms, remain unclear. MethodsCRISPR/Cas9 gene editing technology was employed to construct GBM cell lines with NF1 gene mutations. Metabolomics was utilized to examine the impact of NF1 on metabolic remodeling in GBM. The Seahorse XF24 extracellular flux analyzer was used to detect the effect of NF1 knockdown on glycolysis and mitochondrial oxidative phosphorylation in GBM cells. Wound healing assay and Transwell chamber assay were utilized to detect the effect of NF1 on GBM cell invasion. Orthotopic tumor model in nude mice was established to explore the role of NF1 in vivo. In addition, Co-IP, western blotting, and immunofluorescence were used to explore the changes of key enzymes in glycolysis and mitochondrial oxidative phosphorylation and the relationship between NF1 and MFN1. ResultsThe expression of NF1 is decreased in glioma tissues and is significantly correlated with patient prognosis. NF1 knockdown may promote the invasion, migration, and EMT of GBM cells. At the same time, the activation of the AKT/mTOR signaling pathway promotes aerobic glycolysis in GBM cells, promotes mitochondrial division through targeted regulation of MFN1, and inhibits mitochondrial oxidative phosphorylation. NF1 deficiency promotes EMT in GBM cells by enhancing aerobic glycolysis and mitochondrial division. ConclusionNF1 deficiency promotes GBM glycolysis by activating the AKT/mTOR signaling pathway and inhibits the mitochondrial oxidative phosphorylation by regulating MFN1; NF1 deletion promotes GBM EMT by remodeling the pattern of energy metabolism.

Cancer Biology↗

bioRxiv2026

Chromosome gigantism and auxin deconjugation underpin gall induction in a horned gall aphid

Insect galls are extended phenotypes built through sustained reprogramming of host development, yet the genomic innovations and effector mechanisms that enable this process remain poorly resolved. Here we generated chromosome-level genomes for the horned gall aphid Schlechtendalia chinensis and its host tree Rhus chinensis, and integrated these resources with morph-resolved salivary proteomics, metabolomics and functional assays. The aphid genome contains an anomalously expanded chromosome 1 spanning 93.61 Mb that carries 33.08% of anchored genes and concentrates 70.47% of genome-wide segmental duplications, together with elevated transposable-element load and multiple tandem gene clusters. We identified a salivary M20-family enzyme, ACY1, that consistent with an auxin-conjugate hydrolase. Loss- and gain-of-function assays show that ACY1 hydrolyzes stored, inactive auxin-amino acid conjugates in host tissue, increasing local active auxin levels. This hormone manipulation reprograms plant cell development to form galls. Our findings demonstrate how concentrated genomic duplication can drive the evolution of an effector that controls host physiology, suggesting a potential principle in the evolution of parasitic traits.

molecular biology↗

bioRxiv2026

A U1-U3 snRNA-snoRNA interaction couples SF3B1 mutation to chromatin-state rewiring and genome instability

Mutations in spliceosome factors such as SF3B1 are recurrent across human diseases, including myelodysplastic syndromes and leukemia1-4, yet splicing defects alone do not fully explain the widespread chromatin alterations and genome instability in mutant cells5. Here, by comprehensively mapping snRNA-directed RNA-RNA interactions, we identify two previously unrecognized interaction motifs in U1 snRNA beyond canonical 5' splice-site pairing6,7. These motifs enable U1 RNA to i) bind intronic and other chromatin-associated RNA (caRNA) regions outside of splice sites, and ii) base pair specifically with snoRNA. We uncover a U1-U3 snRNA-snoRNA interaction that recruits the H3K36 methyltransferase SETD2 to caRNA, promoting gene-body H3K36me3 and antagonizing H3K27me3 to modulate chromatin accessibility. The snRNA-snoRNA interface is essential for this previously unrecognized layer of chromatin and transcriptional regulation mediated through SETD2. SF3B1 mutation enhances U1-U3 binding and increases the association of the U1-U3 complex with caRNA, driving chromatin-accessibility remolding, R-loop formation, DNA damage, and copy-number abnormalities that promote tumorigenesis. A U1-specific 2'-O-methoxyethyl antisense oligonucleotide that selectively blocks U1-U3 pairing suppresses these genomic abnormalities, reduces leukemic infiltration, and prolongs survival in xenograft and patient-derived models, establishing pathological snRNA-snoRNA rewiring as a critical driver of SF3B1-mutant leukemogenesis.

cell biology↗

bioRxiv2026

Alginate-Coated Collagen Hydrogel Tubes for Scalable Cell and Biotherapeutic Particle Manufacturing

Efficient, scalable, and cost-effective production of mammalian cells and biotherapeutic particles remains a major challenge for both research and clinical applications. Conventional 2D and 3D culture systems suffer from low volumetric yields, poor scalability, and high costs. Previously, we developed collagen hydrogel tube microbioreactors (ColTubes) that support high-density, high-viability cell culture by preventing excessive cell aggregation and minimizing hydrodynamic stress. However, ColTubes exhibit adhesion to culture vessels and to each other, and leaked cells frequently attach to outer tube surfaces -- behaviors that would limit scalability. Here we introduce AlgColTubes: collagen hydrogel tubes coated with a thin, ionically crosslinked alginate layer to overcome these limitations. Scanning electron microscopy confirms that alginate penetrates the collagen wall and forms a stable interpenetrating hydrogel network, whose depth can be tuned by coating concentration and duration. The alginate coating remains structurally intact under static and dynamic culture conditions without impairing nutrient transport or cell growth. AlgColTubes eliminate tube-tube and tube-vessel adhesion, and prevent exogenous cell attachment to the outer surface, while maintaining cell viability and proliferation comparable to uncoated ColTubes. Their unique architecture -- an adhesive collagen interior and non-adhesive alginate exterior -- further enables a truncated-tube format for continuous release of biotherapeutic particles through open tube ends. We demonstrate that lentivirus is released from truncated AlgColTubes in a segment length-dependent manner, reaching ~100% release efficiency at 1-mm segment lengths. AlgColTubes provide a scalable, cost-effective platform for high-yield cell and particle manufacturing, with broad potential across basic research, translational studies, and industrial bioprocessing.

bioengineering↗

bioRxiv2026

Trustworthy ML/AI for Aging Clocks: Preventing Systematic Prediction Bias in Biological Age Estimation

Machine learning (ML)- and artificial intelligence (AI)-based aging clocks are increasingly used to quantify physiological and molecular aging from omics and medical imaging data as distinct from chronological age. Here, we characterize a fundamental but underappreciated statistical limitation of commonly used ML/AI regression models for continuous outcomes: systematic prediction bias and its propagation to downstream association estimates. This issue becomes more challenging when the true outcome, biological age, is latent and therefore unobserved during ML/AI model training. We demonstrate that systematic prediction bias can distort and, in some cases, even reverse downstream association analyses that use aging clocks as ML/AI-predicted outcomes to assess their associations with exposures or clinical factors. For example, it can produce spurious associations suggesting that older predicted brain age is linked to better cognitive performance, or that older epigenetic age is associated with better kidney function. To address this problem, we introduce a principled and broadly applicable ML/AI regression framework based on constrained optimization, yielding better calibrated aging-clock estimates and valid downstream inference.

bioinformatics↗

bioRxiv2026

Behavioural plasticity in Caenorhabditis elegans navigating dynamic granular environments

The mechanisms that steer behavioural plasticity in complex environments remain poorly understood. Here, we introduce Peb2les, a quasi-two-dimensional granular arena, for the behavioural assessment of Caenorhabditis elegans nematodes, as they interact with deformable, dynamic terrains. Using customized deep-learning-based methods, we track both nematodes and particles to characterize coupled animal-environment dynamics. We find that nematodes locomotor performance depends on particle density and surface properties, and is largely independent of mechanosensation, whereas functional touch sense is required for the initial decision to enter rough-textured arenas. In addition, we identify a previously undescribed touch-seeking behaviour that partially depends on mechanosensation, in which C. elegans actively return to the granular arena after exiting. In parallel, locomoting nematodes rearrange and stack surrounding particles, continuously modifying the granular terrain. Together, these findings establish Peb2les as a versatile platform for probing the mechanism and limits of behavioural plasticity and locomotory adaptability of nematodes in sensory-enriched, soil-like environments, while also providing insight into the potential evolutionary origins of adaptive touch-seeking behaviours in animals.

animal behavior and cognition↗

bioRxiv2026

Alignment-free phylogenetic inference via hyperbolic protein language models

Conventional phylogenetic methods rely on multiple sequence alignments which are computationally intensive and often fail for highly divergent lineages. Here, we introduce LucaPhylo, an alignment-free framework that infers evolutionary relationships directly from unaligned sequences. Through a cascaded learning strategy LucaPhylo integrates protein language models with hyperbolic geometry, a representation space naturally suited to hierarchical branching, to capture deep evolutionary constraints without explicit homology matching. Using highly divergent RNA virosphere as a test case, LucaPhylo places unaligned sequences into phylogenetic trees with an accuracy comparable to leading alignment-based tree construction tools, while retaining divergent sequences that conventional pipelines frequently discard. It further enables the integration of divergent viral lineages into phylogenetic trees, thereby expanding the evolutionary landscape of RNA viruses. Together, LucaPhylo establishes an AI-driven, alignment-free paradigm for phylogenetic inference and provides a robust computational foundation for resolving deep evolutionary relationships among RNA viruses and other biological systems.

evolutionary biology↗

bioRxiv2026

Bright Probes, Blurred Metabolism: Navigating Fluorescent Protein Cross-Excitation in NADH FLIM

Fluorescence lifetime imaging microscopy (FLIM) of endogenous NAD(P)H enables the label-free assessment of cellular metabolic state. Although metabolic imaging is increasingly combined with fluorescent protein (FP) reporters to enhance biological specificity, the potential cross-talk between the intrinsic and extrinsic labels remain ill-defined. Here, we systematically evaluate cross-talk from FPs in metabolic FLIM using phasor analysis of two-photon fluorescence microscopy. The results clearly show that many widely used fluorescent proteins are excited under the conditions used for NADH imaging; they emit blue-shifted, short-lifetime fluorescence that can interfere with imaging NADH metabolic signatures. This overlap persists across excitation wavelengths and FP classes, posing a significant challenge for multiplexed metabolic imaging. This cautionary tale argues against unvalidated multiplexing strategies in metabolic FLIM studies. Our study aims to identify acceptable imaging partners, offer a pipeline for assaying potential cross-talk, and provide practical guidance for experimental design. SignificanceFluorescence lifetime imaging of NADH autofluorescence is a powerful, label-free approach to map cellular metabolism in living tissues. A growing number of studies combine NADH imaging with fluorescent protein (FP) reporters to simultaneously identify specific cell types or subcellular compartments. This study reveals that many FPs, spanning the visible spectrum, are unexpectedly excited under NADH conditions. Commonly used green, yellow, and red variants produce short-lifetime, blue-shifted fluorescence that directly overlaps with metabolic NADH signals. This cross-excitation can be falsely interpreted as a shift in cellular metabolic state, posing a significant risk for multiplexed metabolic imaging studies. Our studies establish a pipeline to assess and manage this risk. We identify StayGoldE138D and mNeonGreen as the most compatible FPs for co-imaging with NADH, and provide a practical framework to guide experimental design and control strategies for multiplexed metabolic FLIM.

biophysics↗

bioRxiv2026

Translational Quantitative Proteomic Assay for Bacteriophages: A New Frontier in Phage Pharmaceutical Development

Accurate quantitation of therapeutic bacteriophages (phages) remains a challenge for clinical development. Plaque-based enumeration is the current standard but is laborious, host-dependent, and variable, particularly when distinguishing individual phages in cocktails. Targeted mass spectrometry of virion structural proteins offers an orthogonal, structure-based approach amenable to reproducible and scalable phage quantitation. Here, we describe a targeted proteomic liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay for host-independent quantitation of the Pseudomonas aeruginosa podovirus LUZ19. Proteomic characterization was performed on an LTQ Orbitrap XL to assess sequence coverage and select surrogate peptide candidates based on specificity and sensitivity. High-resolution peptide mapping identified multiple structural proteins of LUZ19 and provided 55% sequence coverage for the major head protein (YP_001671977.1). Fifteen peptides were detected and evaluated, from which the tryptic peptide EVAELDGQELAR was selected based on abundance, stability, and chromatographic performance. Quantitative analysis was conducted on a QTRAP 7500+ using optimized multiple reaction monitoring transitions for targeted peptide detection. Back-calculated concentrations met accuracy criteria across a validated range of 0.008 to 80 pg/mL, with bias spanning -8.2 to 8.2%, intra-day precision ranging from 0.5 to 9.8%, and inter-day precision ranging from 6.3 to 9.7%. Peptide concentrations from digested lysate samples were related to phage concentrations determined by double layer agar assay, yielding an estimated three copies of the major head protein per virion. ImportanceBacteriophages are the most abundant biological entities on the planet and represent a promising therapeutic class for combating drug-resistant bacterial infections. Realizing the clinical potential of bacteriophage therapy requires analytical methods capable of meeting the standards of modern drug development. Targeted mass spectrometry offers unmatched specificity and resolution for precise quantitation of individual bacteriophages within complex biological samples, a capability that conventional enumeration methods cannot match. Only one prior study has applied mass spectrometry to bacteriophage quantitation, using a well-characterized model bacteriophage at a single concentration without calibration or a validated analytical range. Using Pseudomonas aeruginosa podovirus LUZ19, we present the first targeted mass spectrometry-based bacteriophage quantitation assay developed and validated following FDA bioanalytical guidance. This work establishes a rigorous analytical foundation that moves bacteriophage therapy closer to the standards required for informed dose selection, candidate evaluation, and clinical development.

pharmacology and toxicology↗

bioRxiv2026

Elimination of senescent cells by mechanical cell competition

Cellular senescence, a hallmark of aging, leads to the accumulation of apoptosis-resistant cells that compromise tissue homeostasis. While senescent cells are known to influence neighboring cells through the senescence-associated secretory phenotype (SASP), the precise nature of the interactions between senescent and normal cells remains elusive. Here we show that progerin-induced senescent cells undergo apoptosis when co-cultured with normal cells. This elimination requires direct cell-cell contact and is mediated by the JNK and p38-MAPK pathways, leading to p53 upregulation and p21 downregulation in progerin-expressing cells. Furthermore, neighboring normal cells exert persistent mechanical compression on progerin-expressing cells prior to their elimination, consistent with mechanical cell competition. In contrast, p16-induced senescent cells resist elimination under the same co-culture conditions, maintaining high p21 levels. Our findings reveal a non-cell-autonomous mechanism for senescent cell clearance, providing new insights into the maintenance of tissue homeostasis during aging.

cell biology↗

bioRxiv2026

Morphological fingerprints enable machine learning based inference of neuroblastoma cell states without transcriptomics

Inference of cancer cell states is essential for understanding oncogenic mechanisms and predicting clinical outcomes, yet current reliance on transcriptomic profiling limits scalability and real-time monitoring. Here, we show that cell morphology provides a low-dimensional, observable representation of cellular identity and its dynamics. Using neuroblastoma (NB) as a model system, we establish a machine learning- morphology profiling framework that infers adrenergic (ADRN) and mesenchymal (MES) cell states directly from high-dimensional morphological fingerprints without reliance on transcriptomic measurements. By benchmarking against single-cell RNA sequencing (scRNA-seq), we demonstrate that morphology-defined states closely align with transcriptomic profiles at single-cell resolution. We further show that cell state transitions are represented as continuous trajectories within a morphology-defined state space. Perturbations targeting distinct regulatory layers, including ROCK signaling and epigenetic regulation via EZH2, drive convergent trajectories along a shared phenotypic axis. Together, these results establish cell morphology as a scalable and non-destructive readout of cell state with machine learning providing a unified framework for high-throughput phenotyping and real-time tracking of cancer cell state plasticity.

cell biology↗

bioRxiv2026

Task-space dimensions guide human exploration in complex environments

Efficient exploration is essential for rapid learning in high-dimensional environments, where rewards depend on a sparse and unknown subset of task dimensions. Here, we developed a multidimensional value-learning task in which participants explored without explicit knowledge of the underlying structure. Across multiple task designs, participants consistently demonstrated successful learning under structural uncertainty. Behavioral analyses revealed that over half of participants adopted a structured exploration strategy--Full Dimension Scan (FDS)--with three key properties. First, participants systematically isolated and sampled one candidate dimension at a time, effectively reducing search complexity. Second, FDS was temporally persistent within dimensions while maintaining diverse sampling of feature values, enabling efficient local inference. Third, participants sequentially deployed FDS across dimensions, achieving broad coverage of the task space. Importantly, this strategy transferred across tasks: prior experience in lower-dimensional environments promoted stronger FDS use and improved performance in higher-dimensional settings. Consistent with these findings, a dimension-guided exploration model incorporating dimension-level attentional control best captured human behavior. Together, these results identify dimension-guided exploration as a core principle that transforms high-dimensional search into tractable, structured learning, providing new insight into adaptive behavior in complex environments.

animal behavior and cognition↗

bioRxiv2026

Transformed reactivation of latent working memory enables hierarchical language processing

Human language comprehension requires tracking words across intervening material to construct grammatical structures, yet two fundamental questions remain unresolved: whether maintenance of earlier words relies on sustained neural activity or activity-silent working memory mechanisms, and whether memory retrieval during integration engages domain-general or language-selective networks. Using magnetoencephalography with time-resolved decoding, we tracked main-clause subjects as participants heard sentences with embedded clauses (e.g., "The dog, who chases the cat, jumps over the mud."). The subject representation ("dog") decayed to baseline during the embedded clause but reactivated after the main-clause verb ("jumps"), with transformed rather than reinstated neural codes. Critically, reactivation emerged first in right dorsolateral prefrontal cortex before engaging frontotemporal language regions, and reactivation strength was modulated by syntactic structure and predicted comprehension accuracy. These findings demonstrate activity-silent working memory maintenance, structure-dependent retrieval, and cooperative function of domain-general and language-selective networks during hierarchical language processing.

neuroscience↗

bioRxiv2026

SpatialBench: Comparative cross-platform benchmarking of high-resolution spatial transcriptomics using matched mouse lymphoid tissue

Spatial transcriptomics (ST) has rapidly expanded with the introduction of multiple high-resolution platforms, yet cross-platform benchmarking remains limited and largely focused on technical performance. Here we present SpatialBench, a matched multi-platform resource comprising Visium HD, Xenium and MERSCOPE data together with single-cell and single-nucleus references from a malaria-challenged wild-type and B cell-specific Tbx21 knockout mouse spleen model. In this system, loss of T-bet in B cells disrupts germinal center (GC) polarization and antibody maturation, providing a biologically grounded benchmark for technology comparison. We leveraged this system to systematically evaluate ST platform performance using technical and biological readouts. Across platforms, immune organization and Tbx21-associated programs were consistently recovered, indicating robustness of major biological signals. Platforms instead differed in the level of biological resolution accessible. Visium HD enabled transcriptome-scale GC characterization and, together with Xenium, resolved dark and light zone organization, whereas GC zonation was not resolved in MERSCOPE, consistent with differences in transcript detection sensitivity. SpatialBench provides a biologically defined reference dataset for evaluation of ST technologies, method development, computational benchmarking, and studies of GC spatial organization in lymphoid tissue.

bioinformatics↗

bioRxiv2026

Advancing ab initio genome annotation with OrionGeno

The rapid expansion of eukaryotic genome sequencing has created an urgent demand for accurate and scalable genome annotation. Existing ab initio methods often struggle to reconstruct complex gene architectures and generalize across distant lineages, limiting their use for large-scale annotation. Here we present OrionGeno, a phylogeny-aware deep learning model for end-to-end eukaryotic genome annotation. OrionGeno integrates phylogenetic context, long-range sequence modeling and joint prediction of gene structures and repetitive elements to annotate exons, introns, untranslated regions and repeats directly from genomic sequences. Applied to chromosome-level eukaryotic genomes from NCBI that lack annotations, OrionGeno generates annotations for more than 5,300 genomes, substantially expanding public annotation resources. Across diverse eukaryotic lineages, OrionGeno outperforms state-of-the-art methods at the exon, gene, protein-sequence, and protein-structural levels. It also identifies candidate protein-coding loci absent from reference protein-coding annotations in well-curated genomes. Together with a web platform and integrated annotation database, OrionGeno provides a scalable and accessible framework for translating genome assemblies into functional biological resources and supporting large-scale biodiversity initiatives such as the Earth BioGenome Project.

bioinformatics↗

bioRxiv2026

Extracellular Vesicles Enable CircRNA Delivery via in situ Biogenesis and Sorting

Extracellular vesicles (EVs) are promising vehicles for nucleic acid delivery, yet efficient delivery of circular RNA (circRNA) remains challenging due to inefficient loading and limited intracellular expression. Here, we establish an EV-based platform that enables efficient circRNA delivery via in situ biogenesis and sorting. By optimizing intracellular circularization and translation through vector design, we markedly enhance circRNA expression. By combining this with Snu13-mediated EV sorting and enhanced vesicle biogenesis, we achieve efficient packaging of circRNA without compromising vesicle integrity. This integrated strategy enables robust and sustained protein expression following EV-based circRNA delivery. By leveraging this platform, we demonstrate a dendritic cell-targeting circRNA vaccine that elicits strong antigen-specific CD8+ T cell responses and antitumor efficacy. We further show that systemic delivery of BNP-encoding circRNA attenuates doxorubicin-induced myocardial fibrosis. Together, this work establishes a generalizable platform for circRNA therapeutics by overcoming key barriers in circRNA expression and EV-mediated delivery. TeaserEngineered EVs enable efficient circRNA delivery for sustained protein expression and therapy.

bioengineering↗