Biology subjects

Deng, J.

Publications and source records attributed to Deng, J..

At least 19 recordsLinked to original sources

bioRxiv2026

Evolution-Guided Engineering of Non-conserved N-Glycosylation Sites Improves Heterologous Production of Sulfotransferases

Escherichia coli features low cost, mature fermentation technology and high recombinant protein output, but the expression of glycosylation-dependent eukaryotic proteins therein is severely limited. Many glycosylated proteins are key tool enzymes for the biosynthesis of various pharmaceutical molecules, such as glycosaminoglycan-related synthases, but there is a lack of universal protein engineering methods to enable the adaptive evolution of these enzymes in E. coli. We established an evolution-guided optimization strategy that targets non-conserved N-glycosylation sites (EGO-NGly). The core of this strategy lies in the phylogeny-guided selective reversion of evolutionarily acquired N-glycosylation sequons to naturally occurring non-glycosylated residues from early-divergent orthologs, thereby improving heterologous production and residual activity stability for selected sulfotransferases in E. coli. Using heparin 3-O-sulfotransferase 1 (3OST-1) as a model enzyme apparatus in eukaryotic organisms, EGO-NGly extended the residual activity half-life by 4.5-fold. Structural analysis and molecular dynamics simulations indicate that greater surface rigidity and remodeling of intramolecular interaction networks contribute to the improved conformational stability of M7 in an aglycosylated host. EGO-NGly was further applied to a panel of 3OST and 6OST homologs, demonstrating its applicability to a subset of the tested enzymes. Furthermore, EGO-NGly can be orthogonally integrated with other complementary protein engineering strategies, as exemplified by its combination with the machine learning-aided Protein Repair One-Stop Service, generating variants with 8-fold and 17-fold extensions in residual activity half-life. Collectively, these results establish EGO-NGly as a practical and potentially cost-effective engineering strategy for improving the prokaryotic expression and functional performance of selected glycosylation-dependent eukaryotic sulfotransferases.

biochemistry↗

bioRxiv2026

Establishing toxicologically relevant primary alveolar macrophage models to enable a realistic assessment of nanomaterials safety

Alveolar macrophages (AMs) are key sentinels of the distal lung, maintaining surfactant homeostasis and clearing inhaled particles through continuous surveillance and phagocytosis. Mechanistic studies of particle-cell interactions, important to assess the respiratory safety or toxicity of particles, however, are limited by the lack of suitable in vitro models that preserve the core features of resident AMs. Here, we benchmarked six murine AM models, including short-term cultured primary AMs (AMs 3 h, AMs 24 h), primary AMs maintained with GM-CSF, TGF-{beta}, and the PPAR-{gamma} agonist rosiglitazone (AMs GTR), bone marrow-derived AM-like cells (BM-AMs), the immortalized AM cell line MH-S, and the non-transformed fetal liver-derived AM-like cell line MPI, against in vivo AMs based on surface phenotype, transcriptomic identity, and their responses to three carbon-based nanomaterials (CBNs). Both cell lines showed limited AM identity, characterized by low CD11c and SiglecF but high CD11b surface expression. In contrast, AMs GTR and BM-AMs more closely preserved resident AM-like surface marker profiles. Transcriptomic features, although differing for alterations in metabolism-related pathways, particularly oxidative phosphorylation, revealed most in vivo-like for the primary AM models. Functional, toxicological assessment using soot-like carbon nanoparticles (CNP), double-walled carbon nanotubes (DWCNT) and multi-walled carbon nanotubes (MWCNT) demonstrated strong model-dependent responses. Consistent with MWCNT-induced reduction in AM numbers observed in vivo, AMs GTR and BM-AMs exhibited the strongest in vitro responses, including cytotoxicity and inflammatory cytokine release. MH-S cells in contrast stood out for their cytotoxic sensitivity supporting their use for high throughput cytotoxicity tests. Overall, these findings identify AMs GTR and BM-AMs as the closest models for studying AMs biology and AM-intrinsic responses to poorly soluble nanoparticles like CBNs, and with regard to availability of higher cell numbers, BM-AMs might provide a scalable platform for mechanistic and higher-throughput nanotoxicology studies.

pharmacology and toxicology↗

bioRxiv2026

SMORE: joint dimension reduction and cell population discovery on single-cell methylome data

Single-cell DNA methylation profiling technology captures novel epigenetic data modality but are challenging to analyze because of their heterogeneity, high dimensionality, and ultra-sparsity. Here we present SMORE (Single-cell MethylOme Reduction and Embedding), a computational method for joint dimensionality reduction and cell population discovery dedicated to single-cell DNA methylation data. SMORE operates on a Bayesian framework that converts methylation proportions into ordered methylation states and jointly infers a low-dimensional representation, cell populations and their number. By using low-rank latent Gaussian factorization and adopting a mixture-of-finite-mixtures prior on latent cell scores, SMORE infers cell assignments without requiring a prespecified cluster number, and propagates uncertainty from methylation measurements to cell assignments. Across simulations spanning varying sample sizes, population imbalance, signal strengths and model misspecification, SMORE accurately recovered latent population structure and outperformed existing methods. Applied to human single-cell methylation datasets from lung, peripheral blood and primary motor cortex, SMORE recovered biologically supported cell population structures. SMORE provides an uncertainty-aware framework for dimension reduction and population discovery for single-cell methylomes.

bioinformatics↗

bioRxiv2026

Architectural trade-offs between environmental stability and genomic redundancy reveal divergent pneumoviral entry strategies

Human respiratory syncytial virus (RSV) and human metapneumovirus (hMPV) exhibit distinct seasonal epidemiology, with RSV circulating in early autumn and hMPV peaking in midwinter, yet the structural basis for this niche partitioning remains undefined. Here, we integrate in situ cryo-electron tomography and functional virology to decode the architectural logic governing their entry dynamics. RSV employs a matrix (M)-regulated prefusion F (pre-F) organization, partitioning trimers into stabilizing hexagonal superlattices and fusion-competent pools to maintain superior thermotolerance. By contrast, hMPV compensates for its intrinsically unstable, monomeric pre-F with extreme ribonucleoprotein polyploidy, packaging [~]4-fold more genome equivalents to ensure productive infection. Fusion events localize exclusively to M-depleted, non-arrayed membrane regions, establishing a spatial checkpoint for activation. These findings reveal a conserved trade-off between environmental resilience and genomic redundancy that dictates divergent pneumoviral entry strategies, explaining their distinct seasonal ecological niches.

microbiology↗

bioRxiv2026

A hierarchical activation threshold shapes RSV F antibody repertoires from toddler vulnerability to adult protection

Respiratory syncytial virus (RSV) causes severe infant morbidity, yet the mechanisms underlying their suboptimal immunity and the failure of prefusion F-stabilized adult vaccines in infants remain unclear. Here we integrate deep mutational scanning of 761 preF-binding antibodies, repertoire profiling of 102 repeatedly exposed pediatricians and 61 RSV-experienced toddlers, structural analysis, and >40,000 viral genomes to decode RSV F immunity. We resolved 12 immunologically distinct functional antibody subclasses and uncovered a hierarchical activation threshold: apical epitope-targeting antibodies with superior neutralizing potency demand extensive somatic hypermutation and cooperative CDR networks, whereas central-to-basal epitope-directed, marginal or non-neutralizing antibodies engage germline-encoded antibodies through minimal mutations (e.g., S31G). Toddler repertoires are confined to low-threshold, non-productive sites; adult repertoires enrich for apical elite neutralizers whose immune pressure drives contemporary RSV-B evolution. These reframe pediatric RSV vulnerability as a threshold-gated repertoire deficit and prescribe vaccine strategies that actively redirect immunodominance from permissive epitopes toward high-barrier apical targets.

immunology↗

bioRxiv2026

Rational antigen engineering and mucosal delivery design for next-generation RSV vaccines

Respiratory syncytial virus (RSV) prefusion F (preF) vaccines have transformed adult prophylaxis, yet unmet needs in antigen stability, pediatric safety, and mucosal protection persist. Here, we develop an integrated structure-guided RSV vaccine design platform that couples allosteric stabilization, epitope-focused immunogen engineering, and route-specific mRNA delivery for systemic and mucosal immune activations. By mapping prefusion F "breathing" motions and applying a ThermoNet-and Rosetta-guided screening funnel, we identified R296, a stabilized prefusion F immunogen that reinforces the 1-5 hinge and interprotomer interfaces while preserving key neutralizing epitopes. Cryo-EM confirmed that R296 retains a native-like prefusion architecture. And mRNA-LNP vaccination elicited potent, durable, and broadly protective neutralizing responses in mice, rats, and cotton rats, with clearance of detectable infectious virus and no evidence of Th2-skewed enhanced respiratory disease. To address pediatric safety, we designed a stalkless nanoparticle immunogen, Head38-50AB-3, which enriches high-potency apical epitopes while excluding stalk regions associated with low-potency or non-protective responses, conferring protection without VAERD-like pathology. Finally, we engineered an intranasal-delivered LNP that enables intranasal R296 mRNA delivery, inducing systemic neutralization together with robust nasal and bronchoalveolar secretory IgA (sIgA). R296 has now advanced to Phase 1 clinical trials. These results establish a modular framework for next-generation RSV vaccines.

microbiology↗

bioRxiv2026

Noncanonical Circular RNAs and Potential Functions

Circular RNAs (circRNAs) are ubiquitous in eukaryotes; dysregulated circRNA expression is linked to diseases, including lung cancer. In contrast to canonical circRNAs arising from exon-intron boundaries, noncanonical circRNAs originating within exonic, intronic, and intergenic regions have typically been dismissed as transcriptional noise or technical artifacts. To explore circRNA diversity and appreciate their functions, we developed an algorithm to identify both canonical and noncanonical circRNAs without relying on genome annotation, enabling the identification of circRNAs of all types and in newly sequenced or poorly annotated species. Results from lung cancer cells revealed that noncanonical circRNAs constituted over two-thirds of the circRNA population and were expressed more abundantly than canonical circRNAs, and genes with fewer and shorter exons were hotspots for noncanonical circRNA and circRNA isoform production. Further analyses showed that many noncanonical circRNAs were indeed endogenous circRNAs transcribed within cells rather than experimental artifacts, were potentially translated into proteins or peptides, and were conserved across species. Moreover, we validated 65 noncanonical circRNAs in NCI-H23 cells using multiple bioassays and demonstrated that both exonic and intergenic noncanonical circRNAs influenced cell viability. CircRNA profiles in tumor and tumor-adjacent tissues of lung cancer patients revealed tissue-specific expression and differentially expressed canonical and noncanonical circRNAs from cognate genes involved in cancer-related pathways, indicating their potential clinical relevance. This study confirmed the authenticity of noncanonical circRNAs and provided the first experimental evidence that noncanonical circRNAs influence cancer cell phenotypes. These findings broaden our understanding of circRNA biology, highlighting their widespread genomic distribution, diverse functions, and potential clinical relevance.

molecular biology↗

bioRxiv2026

Gut Virome Diversity and Health Status in Rescued Pangolins: Metaviromics Reveals Parvovirus as a Key Virus Associated with Disease

BackgroundPangolins are critically endangered mammals that suffer from high rates of gastrointestinal disease during captivity, yet the role of the gut virome in their health remains unexplored. This study presents the first comprehensive characterization of the gut DNA virome in Malayan (Manis javanica) and Chinese (M. pentadactyla) pangolins across different health states. ResultsMetaviromic sequencing of 16 fecal samples from healthy, diarrheal, pneumonic, free-ranging, and deceased pangolins generated 7.2-11.8 Gb clean data per sample. A total of 12 viral phyla, 26 families, 219 genera, and 1,132 species were identified. Caudovirales phages (Siphoviridae, Myoviridae, and Podoviridae) dominated the gut virome of healthy individuals, with phage content exceeding 90% in most healthy samples. However, diseased and deceased individuals exhibited a significant reduction in phage proportion (94.5% vs. 53.3%, P = 0.02), accompanied by a dramatic increase in eukaryotic viruses--particularly Parvoviridae, which accounted for 66% of the virome in deceased Malayan pangolins. Iridoviridae (38%) and Polydnaviridae (34%) dominated in deceased Chinese pangolins. SIMPER analysis identified Parvoviridae as the primary contributor to virome dissimilarity between healthy and diseased Malayan pangolins, whereas Iridoviridae and Polydnaviridae were the key contributors for Chinese pangolins. LEfSe analysis revealed 12 biomarker viruses in free-ranging pangolins (predominantly Staphylococcus and Escherichia phages), six in healthy Chinese pangolins (predominantly Streptococcus phages and Lactobacillus viruses), and six in deceased Malayan pangolins (predominantly vertebrate viruses). KEGG functional annotation revealed that genes related to DNA replication, repair, and recombination were the most abundant, suggesting frequent genomic recombination within the pangolin gut virome. ConclusionsThe gut virome of pangolins is closely associated with health status. Disease induces a shift from a phage-dominated to a eukaryotic virus-dominated virome, with Parvoviridae and Iridoviridae emerging as candidate pathogenic viruses in Malayan and Chinese pangolins, respectively. These findings provide molecular evidence for virome monitoring in pangolin conservation and highlight the need for targeted surveillance of ImportancePangolins are the most trafficked mammals in the world, and all eight species are critically endangered. During rescue and captive care, these animals suffer from high rates of gastrointestinal disease, yet the role of viruses in their gut health has never been studied. This research provides the first comprehensive map of the gut virus community in pangolins, revealing that healthy individuals carry a virus community dominated by beneficial bacterial viruses called phages, while sick and dying animals show a dramatic shift toward disease-causing viruses. Notably, parvoviruses emerge as potential pathogens in Malayan pangolins, while iridoviruses dominate in deceased Chinese pangolins. These findings offer a new tool for monitoring pangolin health in rescue centers: by tracking the balance between beneficial phages and harmful viruses, veterinarians may be able to detect illness earlier and improve survival rates. The study also highlights the importance of environmental exposure for maintaining a healthy gut virus community, supporting the creation of semi-natural enclosures for rescued pangolins. Beyond conservation, this work contributes to our understanding of how viruses move between wildlife and humans, which is critical for preventing future disease outbreaks.

microbiology↗

bioRxiv2026

LARP4 is a B cell-specific metabolic checkpoint for plasma cell differentiation and a therapeutic target in systemic lupus erythematosus

RNA-binding protein LARP4 plays an important role in T cell activation and differentiation, but its role in B cell biology and the pathogenesis of systemic lupus erythematosus (SLE) remains unclear. This study found that LARP4 was specifically highly expressed in B cells of SLE patients and was positively correlated with disease activity. By constructing T cell-specific and B cell-specific conditional knockout mice, we found that deletion of LARP4 in B cells, but not in T cells, significantly alleviated pristane-induced and Bm12-induced lupus nephritis. Further analysis showed that LARP4 deletion selectively inhibited B cell differentiation into plasma cells, but did not affect germinal center B cell formation. Integrated transcriptomic and metabolomics analyses revealed that this effect is due to reduced phosphatidic acid synthesis and decreased mTORC1 activity caused by mitochondrial oxidative phosphorylation dysfunction. Furthermore, we used LIPEP, a LARP4 inhibitory peptide that effectively mimicked the therapeutic effects of LARP4 gene knockout in the MRL/lpr spontaneous lupus model and outperformed cyclophosphamide in reducing glomerular immune complex deposition and improving extrarenal dermatitis. These results indicates that LARP4 is a key metabolic checkpoint regulating B cell differentiation into Plasma cells and suggest that it may be a potential therapeutic target for SLE.

immunology↗

bioRxiv2026

MAERM: Predicting Enzyme-Reaction Matching Relationships with a Mixed-Attention Model

Harnessing enzyme specificity requires a thorough understanding of enzyme promiscuity, which determines enzymes catalytic scope; however, measuring this scope still relies heavily on labor-intensive analytical approaches. While data-driven approaches have emerged to predict the catalytic scope of enzymes, these methods continue to face challenges such as restricted datasets and insufficient integration of enzyme structural information and reaction transformations. Here, we introduce MAERM, an innovative mixed-attention model designed to predict enzyme-reaction matching relationships. Built on our MAERM-DB, a dataset with broad coverage of validated and chemoenzymatic catalysis data, MAERM utilizes a local-global attention module to integrate multimodal enzyme information with fine-grained reaction representations, thereby predicting enzyme-reaction matching probabilities. Results show that MAERM consistently outperforms all baselines, with an average F1-score of 0.984. Notably, on challenging test samples with less than 40% sequence identity to the training set, MAERM outperforms the second-ranked model by 5.9% in F1-score. In addition, MAERM achieves the highest top-10 success rate of 51.7% on Enzyme-405 and the highest balanced accuracy of 0.697 on BioCat-547, further supporting its generalizability in enzyme screening and chemoenzymatic catalysis. Finally, MAERM can serve as an efficient scoring module. When integrated with ProteinMPNN, MAERM has successfully guided novel enzyme design for two carbonyl reduction reactions, resulting in enhanced catalytic potential for the native substrate and demonstrating broad compatibility. Overall, MAERM has the potential to reduce the experimental cost of measuring enzymes catalytic scope, facilitate enzyme design, and ultimately accelerate the design-build-test-learn cycle in enzyme engineering.

bioinformatics↗

bioRxiv2026

Biogenesis and antigen presentation properties of dendritic cell-derived apoptotic bodies

Dendritic cells (DCs) are an important type of antigen presentation cell that regulate immunity by initiating antigen-specific immunity and tolerance through T cell activation. The interaction between DCs and T cells can be mediated through direct cell-cell contact or via the release of extracellular vesicles (EVs) from DCs that harbour antigen presentation machineries. Although small EVs (<200 nm in diameter) such as exosomes released by DCs have been shown to regulate immunity, whether other EV subtypes, in particular those that are released by dying DCs due to homeostatic turnover or following infection, can modulate immune responses is not defined. In this study, we demonstrated that DCs undergoing apoptosis can generate a subclass of large EVs ([~]1,000-5,000 nm in diameter) known as apoptotic bodies (ApoBDs) via distinct morphological steps. Mechanistically, ApoBD formation by apoptotic DCs is regulated by Rho-associated kinase 1 and T-type calcium channels. Functionally, DC-derived ApoBDs were found to mediate direct antigen presentation. These data demonstrate a novel function of ApoBDs and highlight the ability of apoptotic materials derived from dying DCs to continue mediating intercellular communication and regulating immune responses.

cell biology↗

bioRxiv2026

Insect COI barcoding data as an untapped resource for surveying Wolbachia symbioses

O_LIDNA barcoding of the mitochondrial cytochrome c oxidase I (COI) gene is widely used to characterise insect diversity and distributions; however, its potential to reveal information on species interactions, including host-symbiont associations, remains largely unexplored. Here, we assess whether COI amplicon data can be used to identify Wolbachia - one of the most widely distributed bacterial symbionts known to profoundly affect their hosts biology. C_LIO_LIWe demonstrate that several commonly used invertebrate COI primer sets perfectly match many reference Wolbachia genomes, leading to frequent co-amplification. C_LIO_LIBy screening 7,901 individual-insect COI amplicon libraries obtained with the popular BF3-BR2 primer set, we detected Wolbachia sequences in over 35% of samples, revealing that co-amplification is indeed widespread. After removing low-abundance reads, Wolbachia detection based on COI amplicons showed over 90% agreement with simultaneously generated 16S-V4 rRNA amplicon data from the same specimens. The degree of agreement, however, varied depending on the thresholds used, among datasets and insect clades. C_LIO_LIFurther, we show that Wolbachia abundance inferred from COI amplicons correlated with their abundance in metagenomic datasets for 152 specimens, supporting the quantitative relevance of the signal. C_LIO_LIFinally, we find that Wolbachia COI sequences provide greater phylogenetic resolution than 16S-V4 rRNA data (mean pairwise genetic distance of COI sequences - 9.6%, 16S-V4 rRNA - 2.8%), and the reconstructed Wolbachia COI-based genotype network largely agrees with genome-based phylogenies. C_LIO_LICollectively, our results demonstrate that off-target Wolbachia sequences recovered from standard insect COI barcoding data may reliably detect symbiont presence, provide phylogenetic insight, and guide sample selection for metagenomics. Given the rapid expansion of global insect barcoding initiatives, these findings highlight an opportunity for cost-effective monitoring of their most prevalent bacterial symbionts, offering new perspectives on how host-microbe interactions may shape insect communities. C_LI

ecology↗

bioRxiv2026

Multi-stage physics-informed neural networks for JAK--STAT5 signaling and ultradian insulin--glucose dynamics: latent-species identifiability and suppression of parameter-induced divergence

Coupled diffusion-reaction partial differential equations (PDEs) describe biochemical network dynamics but are difficult to solve for realistic multi-species systems without combining mechanism and data. We present a multi-stage physics-informed neural network (PINN) for multi-species diffusion-reaction PDEs and apply it to two ordinary-differential-equation (ODE) reference systems: the Boehm et al. JAK-STAT5 signaling pathway and the Sturis ultradian insulin-glucose model. For STAT5 we pose a latent-species identifiability test: given sparse observations of eight species, a ten-species model that retains two deliberately withheld but mechanistically standard components--an active receptor-JAK complex and the SOCS negative-feedback inhibitor--recovers the reference trajectory and reduces mean root-mean-square error 3.1-fold relative to an eight-species model that omits them, whereas a PDE-only solution without data anchoring diverges. Because the reference is itself ODE-generated, this demonstrates identifiability against synthetic data, not the discovery of new biology. For the insulin-glucose model the same framework reproduces the [~]120-minute oscillation to 1.0% mean relative error as a benchmark on a stiff, multi-timescale oscillator; its spatial dimension is treated as a numerical construct, not a physical transport setting. A Lyapunov analysis of the STAT5 ODE returns a maximal exponent statistically indistinguishable from zero ({lambda}max {approx} 3.61 x 10-5 min-1, 5/8 trials positive; Lyapunov time [~]1.9 x 104 min, far exceeding the 240-720 min horizon), so the system is effectively non-chaotic and the relevant instability is a bounded, parameter-induced trajectory divergence. Anchoring the solution to baseline data suppresses this divergence, with the reduction growing monotonically with sampling density--from [~]15-19% at eight time points to [~]88-97% at sixty-four, depending on perturbation magnitude. The framework thus offers a data-anchored route to latent-species identifiability and divergence suppression in biochemical ODE/PDE systems, demonstrated here against synthetic reference data. Inside cells, a three-dimensional chemistry of diffusing, reacting molecules drives signaling and rhythm--dynamics that, for realistic networks, strain conventional solvers. Here a multi-stage physics-informed neural network--machine learning constrained by the governing equations--solves stiff, multi-species reaction systems from sparse data. In the JAK-STAT5 signaling pathway, a model that retains two standard but unobserved components (an active receptor complex and a negative-feedback brake) recovers a reference trajectory that a reduced model cannot--a controlled test of whether sparse data can pin down withheld pecies, not a claim of new biology. The same framework reproduces the roughly two-hour insulin-glucose rhythm to within 1% as a benchmark on a stiff oscillator. And anchoring the solution to a few dozen baseline measurements collapses parameter-induced trajectory divergence, turning a parametrically sensitive simulation into a stable one. Where mechanism and data meet, sparse measurements can constrain the structure a model would otherwise leave undetermined.

biochemistry↗

bioRxiv2026

Supervised Domain Adaptation Mitigates Cross-Ethnicity Prediction Error in Neuroimaging Based Cognitive Prediction

Machine learning models are increasingly used to predict cognitive and clinical outcomes from neuroimaging data, yet challenges in fairness and generalizability remain. Large scale datasets are often racially and ethnically imbalanced, leading to systematic performance disparities, with models typically achieving higher accuracy for majority populations represented in the training data. In this study, we evaluated whether supervised domain adaptation methods including balanced weighting, two-stage TrAdaBoost, feature augmentation with SrcOnly prediction, and linear interpolation can mitigate these biases. Using the ABCD dataset, we assessed whether models trained on 80 MRI measures from White American participants could generalize more effectively to African American participants. All domain adaptation methods reduced prediction error for African American participants, particularly for MRI modalities with large baseline disparities (e.g., structural MRI), while offering limited improvements where initial gaps were smaller (e.g., functional connectivity). Among the approaches, balanced weighting performed best and remained stable and beneficial even when only 10 African American participants were used to adapt the original model trained exclusively on White American participants. These findings suggest that simple, low-cost strategies can effectively reduce cross-ethnic performance gaps and improve equity in predictive neuroimaging, offering a practical path forward for future neuroimaging predictive biomarkers.

neuroscience↗

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↗

bioRxiv2026

Micro-to-Macro Scale Hydrogel Microchannel Networks by Twisted Wire Templating

The human vasculature is a complex, multiscale system comprising hierarchical networks of macroscale to microscopic vessels. Existing in vitro fabrication techniques often fail to bridge these disparate scales, as high-resolution methods like multiphoton ablation are too slow for replicating larger vessels, while 3D printing lacks the resolution for fine microscale features. Here, we report a "twisted wire templating" strategy capable of generating perfusable bifurcating hydrogel networks that seamlessly transition from the macro- to the micro-scale (2.3 mm to 140 {micro}m) through seven orders of bifurcations. By optimizing wire-twisting geometries and polyurethane dip-coating, we overcame instability-driven bead formation to ensure replication fidelity across the networks. Fabrication rigs were reconfigured from existing 2D planar layouts to 3D reconfigurable architectures to better replicate 3D vessel geometries which simultaneously reducing the laboratory footprint and fabrication times by 47%. Using a Taguchi orthogonal array, we further optimized surface chemistry and hydrogel composition to inhibit structural failure during template extraction, resulting in fully patent, perfusable networks. This method provides a robust, low-cost, and scalable foundation for creating physiologically representative vascular models for investigating multiscale disease mechanisms and organ-level tissue engineering.

bioengineering↗

bioRxiv2026

Structural Complementarity Maximizes Feasibility and Stability in Microbial Community Coalescence

Microbial communities frequently coalesce through dispersal, disturbance, or deliberate transplantation, yet the dynamical consequences of such coalescence remain poorly understood. Here, we show that coalescence can function as a structural design mechanism to enhance microbial community robustness. Using a mechanistic consumer-resource model in which the balance between competition and metabolic cooperation is explicitly tunable, we quantify how interaction structure shapes both feasibility, namely the environmental domain supporting coexistence, and dynamical stability. Cooperation-dominated communities exhibit greater but more variable feasibility and intrinsic stability than competition-dominated communities. Strikingly, coalescing communities with maximally distinct interaction structures consistently maximizes both feasibility and stability by reducing alignment among interaction vectors and strengthening effective self-regulation in the resulting assemblage. Heterogeneous coalescence balances reduced facilitation, moderated interspecific effects, and stronger self-regulation. These results identify structural complementarity as a general principle for assembling robust microbial ecosystems and provide a theoretical foundation for microbiome engineering strategies that enhance persistence and functional stability.

ecology↗

bioRxiv2026

Integrated single-cell and spatial transcriptomic analysis of T cell exhaustion and immunometabolic remodeling in HPV-positive oropharyngeal squamous cell carcinoma

HPV-positive oropharyngeal squamous cell carcinoma (OPSCC) harbors dense lymphocytic infiltration yet responds poorly to immune checkpoint blockade (ICB), a paradox whose mechanistic basis remains unresolved. Here we constructed an integrated single-cell and spatial transcriptomic atlas of HPV OPSCC (23,119 cells, 24 cell types), complemented by CRISPR-Cas9 functional validation in two HPV cell lines and multiplex immunofluorescence across independent cases. Pseudotime trajectory analysis placed TOXCD8 T cells, rather than canonically defined exhausted cells, at the terminal exhaustion endpoint, and uncovered a previously uncharacterized IEGCD8 effector memory population at a fate-decision juncture between activation and irreversible exhaustion. Actively cycling Tregs with broad co-stimulatory and co-inhibitory receptor expression underwent intratumoral immunosuppressive differentiation. Compartmentalized LDHA/MCT4 expression in cancer cells versus LDHB/MCT1 in T cells established a directional lactate flux model, and spatial deconvolution confirmed peripheral immune exclusion. We identified a KMT2D-KLF7-PD-L1 regulatory axis: KMT2D sustains KLF7 transcription through H3K4me1-dependent enhancer activation; KLF7 concurrently drives neural crest differentiation and upregulates PD-L1, thereby linking epigenetic remodeling to both neural programming and immune checkpoint expression. These findings define converging multi-scale mechanisms underlying ICB resistance in HPV OPSCC and nominate the KMT2D-KLF7-PD-L1 axis as a combinatorial therapeutic target.

cancer biology↗