Biology subjects

Zhang, J.

Publications and source records attributed to Zhang, J..

At least 73 records · Page 4Linked to original sources

bioRxiv2026

Harmine Plus Exendin-4 Enhances Remission of Recent-Onset Type 1 Diabetes Following Anti-CD3 Therapy

Type 1 diabetes (T1D) results from autoimmune destruction of pancreatic {beta}-cells. While anti-CD3 therapy can delay disease progression and preserve residual {beta}-cell function, disease reversal will likely require both immune modulation and {beta}-cell regeneration. We found that the combination of harmine and exendin-4 (H+E) reduced inflammation-induced human {beta}-cell apoptosis, suppressed cytokine signaling and immunogenicity pathways, and improved {beta}-cell function. Although H+E alone did not reverse diabetes in NOD mice, low-dose anti-CD3 followed by H+E normalized blood glucose, increased insulin levels, improved glucose tolerance, expanded {beta}-cell mass, and enhanced diabetes remission. These effects were associated with reduced pro-inflammatory T-cell responses, increased regulatory T cells, and greater expression of exhaustion-related T-cell markers, without broad lymphocyte depletion. Similar immunomodulatory effects were observed in activated human PBMCs. Transcriptomic analyses identified the lncRNA SNHG6 as a key mediator of H+E action; SNHG6 protected {beta}-cells from cytokine-induced stress, apoptosis, and immunogenicity. Together, these findings demonstrate that H+E promotes {beta}-cell recovery and resilience while reducing {beta}-cell immunogenicity, enabling remission of recent-onset T1D when combined with anti-CD3 therapy. SNHG6 emerges as a novel regulator of {beta}-cell protection during inflammation.

cell biology↗

bioRxiv2026

CalFluxTools: An R package for analysis of high-throughput calcium oscillation screening data

Whole plate detection of calcium flux signals using fluorescence is a powerful technology for monitoring oscillatory activity of excitable cells in a high throughput fashion that sees broad applications in neurotoxicity and cardiotoxicity research. However, analysis of calcium oscillation profiles can output more than 60 peak kinetic parameters, which makes manual data analysis challenging. We have created the CalFluxTools R package, an automated pipeline for parsing plate data for 384 well plates, generating quality control metrics and plots, and performing data analysis, including T-tests, Z-score analysis, and machine learning prediction of compound toxicity values. CalFluxTools requires a plate map and experimental manifest that includes parameter filtration specifications, as well as data in the ".statall" format. CalFluxTools runs in less than a minute on a Macbook Pro with a 2.6 GHz 6-Core Intel Core i7 processor.

bioinformatics↗

bioRxiv2026

A discrete Connexin26+ neural crest lineage emerges in mid-life and mediates enhanced central brainstem responses to elevated CO2 levels for deep breathing.

The precise control of breathing is fundamental to vertebrate survival. Increased PCO2 in blood and brain parenchyma causes an increase in both the frequency and volume of lung ventilation. We have previously demonstrated that CO2 directly binds to connexin26 (Cx26) hemichannels causing them to open and allow release ATP. We now document the role of Cx26 as a direct physiological CO2 sensor in vivo. Here we describe a unique Cx26+ neural crest cell lineage that populates the ventral brainstem in the vicinity of the PreBoetz nucleus/caudal CO2 chemosensory area during middle age and dies again in old age. Ablating Cx26 genetically specifically within this population of about 20 cells by two independent neural crest Cre-driver lines leads to a loss of local ATP release in the posterior chemosensory area as well as a 40% reduction in elevated tidal volume responses specifically in middle age, as measured by whole-body plethysmography. These in vivo effects change over a life-time: they directly mirror the arrival, wiring in middle age and later death of this cell population in old age. This is a first known example of a middle age change in cranial neural crest lineage composition and highlights the significant power of very few cells for global metabolism. Such lineage-dependent middle-age dynamics also impacts upon the evolution of eusociality: altricial naked pups of our common amniote/synapsid ancestors were heated by by the breath of their (middle-aged) carers, sensing elevated CO2 levels in hypercapnic burrows. Such mechanistic exaptation enabled small amniotes to sense and survive the lethal global CO2 spikes during the Permo-Triassic and other extinction events.

neuroscience↗

bioRxiv2026

Maleic acid adjuvates BCG to induce CD4+CD8+ T cells and sterilize tuberculosis

Anti-tuberculosis (TB) immunity typically emphasizes CD4+ T and CD8+ T cells that only provide limited protection against Mycobacterium tuberculosis (Mtb) infection. The anti-TB function of CD4+CD8+ T cells remains unexplored, mostly due to the dogma that ThPOK and Runx3 antagonize each other to block their development and their scarcity. Here, we show that immunization of BCG (Bacillus Calmette-Gu&eacuterin) with Complete Freund's Adjuvant (CFA) markedly induced CD4+CD8+ T cells through the gut microbiota Clostridium butyricum or its metabolite maleic acid. Mtb-specific CD4+CD8+ T cells exhibited a broad T-cell receptor repertoire and sterilized Mtb infection by inducing SLAMF1 and neutrophil responses. Mechanistically, maleic acid interacted with PAK2 to promote STAT5 phosphorylation at Ser127, thus upregulating Ets1 expression. Direct binding of Ets1 to the Runx3 and ThPOK promoters orchestrated their expression to redirect the CD8+ T cell differentiation towards the CD4+CD8+ lineage. Moreover, immunization of BCG plus Clostridium butyricum or maleic acid sterilized Mtb infection in mouse models. These findings establish CD4+CD8+ T cells as superior immune effectors against Mtb, a concept that could fundamentally redirect the current paradigm for anti-TB vaccine development.

microbiology↗

bioRxiv2026

A reference genome without a virus: cDNA reconstruction reveals the provenance and function of the MS2 phage sequence

Reference genomes are often treated as faithful representations of experimentally validated viral genomes, yet the relationship between historically curated reference sequences and infectivity is rarely tested experimentally. Here, we developed a cDNA-based reconstruction platform for the canonical RNA phage MS2 and used it to compare the current NCBI reference genome (RefSeq) with closely related published isolate sequences. We found that isolate-derived sequences reproducibly yielded infectious phage, whereas the current MS2 RefSeq-derived construct did not, showing that the present reference does not represent a single experimentally validated infectious genome but instead reflects sequence curation across multiple studies. We then compared conventional and AI-enabled approaches to identify minimal changes that restore infectivity to MS2 RefSeq; a human experimentalist correctly prioritized corrective changes, whereas the genome language model Evo2 did not. We also observed that closely related corrected reference-derived constructs showed a ~4-log difference in phage output, and subsequent analysis indicated that this difference was associated with an apparent replicase frameshift in the lower-output background. This suggests that the low output construct class represents rare mutations from genomes that are one mutational step away from true function, rather than uniform function of the dominant construct population. A complementary cell-free assay provided a lower-background orthogonal readout of construct-level function, yielding ~1 x106 PFU/mL from the high-output background within 2 hours while showing no detectable recovery from the low-output background. Together, these results establish a robust platform for RNA phage reconstruction and raise the possibility that historical reference genomes, especially for RNA viruses, may not always remain faithful to experimentally validated biological function. More broadly, these findings underscore the need to verify the infectivity of reference genomes, particularly when they were assembled non-contiguously or shaped by cumulative human curation. They also highlight the importance of clearly distinguishing historically curated reference sequences from experimentally validated infectious genomes when such data are used to train or evaluate AI/ML models.

microbiology↗

bioRxiv2026

Exposure to a mixture of long-chain PFAS disrupts ovarian follicle development, ovulation, and luteinization in female mice

Per- and polyfluoroalkyl substances (PFAS) are persistent environmental contaminants associated with adverse female reproductive outcomes. However, most prior studies examined individual PFAS, despite that humans are exposed to complex PFAS mixtures. Here, we investigated the ovarian effects of a mixture of five commonly detected long-chain PFAS, including PFOS, PFOA, PFNA, PFDA, and PFHxS, using complementary in vivo mouse and in vitro 3D ovarian follicle culture models. Young adult female mice were exposed to a range of concentrations of the PFAS mixture (PFASmix) through daily drinking water for 8 weeks. PFASmix at high concentrations (PFASmix-high) disrupted mouse estrous cyclicity, decreased circulating estradiol, increased testosterone, and promoted follicular atresia. Antral follicles from exposed mice showed decreased Lhcgr and increased Amh expression. Although the number of ovulated oocytes was unchanged following superovulation, PFASmix-high reduced oocyte size and polar body extrusion and altered oocyte cytoskeletal characteristics. In the 3D follicle culture model, direct PFASmix exposure impaired follicle growth, ovulation, and oocyte maturation in a concentration-dependent manner. PFASmix-high suppressed the periovulatory expression of Areg, Ereg, Tnfaip6, and Adamts1 and impaired corpus luteal (CL) development and function, as evidenced by reduced CL spheroid growth and cell survival, decreased expression of Star, Cyp11a1, Hsd3b1, and Lhcgr, and reduced progesterone secretion. PFASmix-high exposure additionally altered the expression of multiple PPAR{gamma}-associated genes in cultured ovarian follicles. Benchmark dose modeling identified estrous cyclicity and follicle rupture as sensitive in vivo and in vitro endpoints, respectively. Together, these findings demonstrated that exposure to the mixture of long-chain PFAS disrupts multiple stages of ovarian function, from follicular development, oocyte maturation to ovulation and luteal development and function and provided mechanistic and quantitative evidence for the reproductive toxicity of PFAS mixtures.

pharmacology and toxicology↗

bioRxiv2026

Iron as a principal mediator of dysbiosis prevention in the phyllosphere

Multicellular organisms require a correct microbiota composition for optimal health and fitness. Dysbiosis could lead to serious health consequences in humans and plants, and is pervasive during pathogen infections across systems. A recent study shows that carbon source availability plays a role in bacterial community assembly in Arabidopsis leaves [1]. However, whether carbon sources or other molecules are critical for leaf microbiota homeostasis remains unknown. Here, an unbiased in planta metatranscriptomic analysis of endophytic microbiota in leaves of wild-type and dysbiotic min7 fls2 efr cerk1 (mfec) mutant, which is defective in pattern-recognition receptor (PRR) signaling and MIN7-dependent vesicle trafficking, revealed that bacteria inside Arabidopsis leaves exhibit taxa-specific transcriptional responses. Most notably, iron starvation response was detected in Stenotrophomonas strains, a group of endophytic bacteria that cause dysbiotic symptoms. Elevated expression of canonical iron starvation genes confirmed low iron availability in mfec plants. This low iron environment was associated with a reduced level of the plasma membrane H-ATPase AHA2, resulting in a higher apoplastic pH that favors less-bioavailable ferric iron over soluble ferrous iron. Supplementing iron in mfec plants was sufficient to restore microbiota homeostasis and alleviate dysbiosis phenotypes. The role of iron in dysbiosis extends to Pseudomonas syringae foliar infection, which drives leaf dysbiosis through two iron-dependent mechanisms. Iron thus emerges as a surprising, key switch between healthy and dysbiotic microbiota in the endophytic spaces of plants.

plant biology↗

bioRxiv2026

Genome-wide characterization of host factors involved in single-stranded RNA and DNA phage infection pathways

Single stranded RNA (ssRNA) and single stranded DNA (ssDNA) bacteriophages represent a key component of the global virome, yet the host genetic networks supporting their infection cycles remain poorly understood. Here, we present a comprehensive, genome-wide mapping of the genetic landscape regulating infection cycles for F pilus-dependent ssRNA and ssDNA phages in Escherichia coli. Genetic screens across ssRNA phages spanning all four genogroups of the Leviviricetes revealed a highly conserved network of host dependencies, with the notable exception of the F plasmid gene traD. While primary structural receptor components and dsbA mediated disulfide bond formation are universally required across all lineages to ensure F pilus integrity, traD exhibits a strict genogroup-specific requirement during entry, showing variable essentiality across different viral groups despite sharing an identical primary receptor. Our gene dosage screens revealed that an elevated copy number of the hslU protease or the RNA chaperone stpA restricts infection, identifying clear genetic barriers that can perturb the viral life cycle. Parallel assays with filamentous ssDNA phages produced host factor profiles consistent with published literature, while revealing additional variations in host dependency. These screens confirmed that ssDNA phages strictly rely on the host TolQRA complex for entry downstream of pilus engagement. The assays tracked prominent negative fitness signatures across homeostatic clusters, highlighting how the physiological burden of continuous, non-lytic virion extrusion strains the host envelope. Finally, this comparative approach traced the selectivity of our isolation host (E. coli HSF) to a horizontally acquired capsule architecture from Klebsiella. This surface shield excludes a large panel of double stranded DNA phages isolated on diverse E. coli strains, while allowing virions from ssDNA and ssRNA phages to engage the extended F pilus and bypass the barrier via native pilus retraction. Together, this work provides a systematic, class-wide map of single stranded phage-host interactions, bridging classical genetics with modern viral discovery while establishing a robust host platform to access uncultured viral diversity and a functional blueprint to design next generation diagnostics, protein antibiotics, and biocontrol tools to halt horizontal gene transfer.

microbiology↗

bioRxiv2026

PlantRegMoD: An integrative and AI-driven multi-omics database for plant regeneration research

Plant regeneration underpins plant developmental plasticity, tissue culture, genetic transformation and crop improvement. Although numerous related omics datasets have been generated, specialized multi-omics databases for this field are still scarce. Here, we constructed PlantRegMoD, an AI-powered integrated database dedicated to plant regeneration. This platform hosts 20.54 TB standardized multi-omics data from 147 projects across 32 plant species and 2,593 samples. We established a unified hierarchical classification system covering five major categories and nine regeneration models, and curated 236 regeneration genes as well as their 28,190 homologs across 58 representative plant species. It contains extensive transcriptomic resources across all regeneration models, together with 196,423 single cells and over 8.81 million epigenetic peaks to dissect cellular heterogeneity and multi-layered epigenetic regulation. Equipped with nine online omics-related tools and a RAG-based intelligent Q&A system, PlantRegMoD greatly reduces bioinformatic barriers and serves as a robust resource for mechanistic, functional and evolutionary studies of plant regeneration.

plant biology↗

bioRxiv2026

A quantitative redox proteome of the human muscle response to exercise

Reactive oxygen species (ROS) regulate protein function through reversible cysteine oxidation. In human skeletal muscle, exercise-induced ROS initiates adaptations such as mitochondrial biogenesis, increased insulin sensitivity, and hypertrophy. However, specific protein targets of ROS regulation during exercise remain unclear owing to longstanding challenges in analyzing redox proteomes in vivo. We applied cysteine derivatization and multiplexed proteomics to map muscle protein cysteine oxidation in humans during exercise. The OxiMuscle dataset quantifies reversible modifications across 9,177 unique cysteine sites on 2,782 proteins, comprising 17,492 individual cysteine site measurements in young men undergoing three types of exercise, providing the first comprehensive, site-resolved and quantitative analysis of the exercise-regulated redox cysteine proteome in humans. We systematically define cysteine oxidation targets regulated by at least one form of exercise, many of which reside in proteins with established roles in muscle physiology. Among these sites is a redox-regulated cysteine on the autophagy receptor protein p62. We demonstrate that reversible oxidation of this cysteine regulates p62-mediated autophagy upon myotube contraction and mouse muscle adaptation to exercise in vivo. Together, these results define a redox-driven mechanism linking exercise-induced autophagy to muscle adaptation. More broadly, our findings offer a comprehensive resource on redox-signaling networks in human muscle, accessible at http://oximuscle-alb-1899330623.us-east-1.elb.amazonaws.com/.

cell biology↗

bioRxiv2026

Ourotide: decoding the hierarchical peptide recognition for generative design

The historical dichotomy between small-molecule pocket and extended protein interfaces misrepresents the physical reality of peptide recognition.1,2 Here we show that peptide binding is not a simple structural intermediate but a distinctly multimodal landscape comprising small-molecule-like pockets, protein-like interfaces, and a previously unrecognized third mode. This third regime is governed by a hierarchical subpocket architecture where a flattened surface achieves near-complete peptide engagement through spatially partitioned hydrophobic components. To maintain stability, an incompletely enclosed dominant anchor cooperates with highly hydrated auxiliary subpockets and an asymmetric receptor coupling mechanism that concentrates energy in an adjacent continuous water network. Because this unique binding mode suffers from extreme data scarcity, standard deep learning models fail to capture its physics.3,4 To resolve this, we mapped these specific geometric signatures to mine structurally faithful training distributions from global protein interactomes. Based on these data, we trained Ourotide, a deep learning framework coupling conditional geometric flow matching with interface-aware affinity learning. Evaluated across peptides up to 65 residues, Ourotide outperforms generalist models in backbone accuracy, interface recovery, and affinity prediction. This approach suggests that overcoming data scarcity in the physical sciences requires physics-guided data augmentation rather than naive statistical scaling.

bioinformatics↗

bioRxiv2026

Unmasking genetic load: Genomic inbreeding drives growth depression in captive critically endangered Chinese Bahaba (Bahaba taipingensis)

Chinese bahaba (Bahaba taipingensis) is a critically endangered marine fish that has been commercially extinct in the wild. Conservation-oriented captive breeding has become an effective strategy for rescuing such endangered species, but concerns about inbreeding depression often remain. Here, after merely two generations of closed breeding from 18 wild founders, Chinese bahaba exhibited severe growth depression, with body length and weight diverging by more than three-fold. To further investigate the early genomic consequences of captive breeding, we analyzed whole-genome sequences from 348 F2 individuals together with 19 wild individuals from the same marine region. While genome-wide genetic diversity was overall preserved, closed breeding rapidly partitioned the F2 population into eight distinct genetic lineages. A significant elevation in genomic inbreeding was observed to facilitate the homozygous exposure of recessive deleterious mutations, which was strongly correlated with growth decline. An integrative framework combining GWAS, gene-level burden test, and ROH hotspots was applied to dissect the genetic architecture underlying growth decline, which converged predominantly on skeletal scaffolding and immune regulation. This was evidenced by candidate genes related to skeleton development and homeostasis (e.g. fam20b, c-fos, trpv5, and il23r), the GH/IGF axis (e.g. socs1a), and immune/inflammatory pathways (e.g. il23r, nlrp7, cd163, il12rb2), as well as calcium ion transport pathway. Together, our findings provide novel insights into the genomic consequences and genetic architecture of growth decline under captive conditions in Chinese bahaba, and offer critical guidance for genetic management in captive conservation programs of endangered species.

genetics↗

bioRxiv2026

Isoleucine and asparagine influence the growth of staphylococcal strains in atopic dermatitis

Atopic dermatitis (AD) is a multifactorial, chronic skin disease associated with microbial factors. It is known that staphylococcal skin microbes play a role in the pathophysiology of AD, notably S. aureus. However, how staphylococcal strain variation affects disease aetiology and progression remains unclear. Using a culture-dependent approach, we analyzed the genomic and phenotypic profile of staphylococcal strains isolated from a Hong Kong children cohort (n = 70). Presence of S. aureus was positively associated with AD skin, whilst S. epidermidis and S. hominis were positively associated with healthy skin, but no overall strain type was associated with AD. In vitro experimentation revealed that there is demarcation in S. aureus strain growth, biofilm biomass formation, and metabolite production by disease severity. While isoleucine and asparagine promoted S. aureus growth, these amino acids suppressed S. hominis growth. These collective findings highlight the importance of strain-level and metabolic interactions of staphylococci in AD pathophysiology.

microbiology↗

bioRxiv2026

Distinct Spatial Immune Architectures in Tumor and Tumor-Adjacent Tissues of Early-Stage Non-Small Cell Lung Cancer

Background: Lung cancer remains the leading cause of cancer-related deaths in the United States with over 124,000 estimated deaths for 2026. Previous studies have found that tumor-adjacent lung tissues may provide additional insight into the immune microenvironment of early-stage NSCLC. Methods: Multiplex immunofluorescence (mIF) imaging was performed on 192 tissues from 101 early-stage non-small cell lung cancer patients including 91 matched tumor-adjacent pairs, using three mIF panels. Spatial analyses were performed to identify distinctions between tissues and identify associations with clinical and genomic features. Results: Tumor tissue showed significantly higher densities of T cells, macrophages, B cells, and memory/regulatory populations than adjacent tissue (p<0.001). Tumors exhibited greater spatial heterogeneity, with higher prevalence of spatial patterning (47.7% vs. 31.2%) and more consistently localized organization, whereas adjacent tissue showed stronger individual-cell clustering. Pairwise colocalization (Ripley's L-cross) revealed selective spatial segregation in tumors (including B-cells from memory/regulatory cells and among cytotoxic T-cell subsets) and reduced immune proximity to malignant cells relative to the strong immune-epithelial association in adjacent tissue. Spatial features were linked to genomic features or patient outcomes: tumor neoantigen burden exhibited a positive association with CD3+ T cells in the tumor, while colocalization between CD45RO+CD57+GZMB+ cells and CD57+GZMB+ cells was negatively associated with overall survival outside the tumor. Conclusion: Tumor and adjacent tissues harbor distinct spatial immune architectures, with spatial features displaying associations with genomic features or patient outcomes. These findings highlight immune microenvironment reorganization and the importance of incorporating spatial context from both compartments into risk assessment in NSCLC.

immunology↗

bioRxiv2026

Predicting Capsid Protein Binding Sites in Single-Stranded RNA Viruses Using Machine Learning from Local Geometric Features

Selective recognition of viral RNA by capsid proteins is essential for genome packaging during the assembly of single-stranded RNA (ssRNA) viruses. However, identification of capsid protein binding sites in the RNA genome remains challenging because current experimental techniques are labor-intensive and low-throughput, motivating the development of computational approaches. Here, we present a sequence-based framework that integrates RNA tertiary structural modeling, local geometric feature extraction, and machine learning to predict capsid protein binding sites. Using the Qbeta; bacteriophage as a proof-of-concept system, we constructed a benchmark dataset of experimentally identified binding and non-binding RNA fragments. We designed a set of geometric descriptors to characterize the local structural features of the RNA backbone. When repeatedly trained and tested with these geometric descriptors on different subsets of the benchmark dataset, the neural network showed a strong ability to distinguish capsid protein binding sites from non-binding RNA fragments, achieving an area under the receiver operating characteristic curve (AUC) of 0.88 in a 5-fold cross-validation. To evaluate whether the model can predict RNA binding sites without experimental structures, we applied it to local geometric features derived from AlphaFold-predicted RNA structures. Despite substantial structural differences between predicted and experimentally determined models, the classifier retained considerable predictive performance (AUC = 0.75), indicating that approximate RNA tertiary structures may still preserve biologically meaningful information for capsid binding site prediction. Furthermore, the failed predictions suggest that viral genome packaging is not only governed by intrinsic RNA structural features, but also by additional dynamic factors beyond static RNA conformations. In summary, our findings provide new mechanistic insights into RNA-capsid interactions and establish a foundation for extending this approach to diverse ssRNA viruses.

biophysics↗

bioRxiv2026

AutoScreen: AI Co-Scientist System for Target Discovery in Functional Genomics

Target discovery in functional genomics remains largely manual and time-consuming, lacking systematic tools for efficient and reproducible gene-level hypothesis generation. We introduce AutoScreen, an AI co-scientist system supporting target discovery through both Pre-screen Design, which constructs perturbation libraries de novo from free-text research descriptions, and Post-screen Analysis, which re-ranks experimental screen hits by integrating statistical scores with biological context. AutoScreen leverages the complementary strengths of multiple specialist agents that perform deep research into multi-modal evidence, information restructuring, parallel searches across >26 biomedical databases, evidence synthesis, and target review to provide transparent, explainable gene prioritization with pipeline provenance. Across 320 genome-scale CRISPR screens as expert-curated benchmarks, AutoScreen achieved a ~19% increase in validated-hit recovery among its top 100 predictions, relative to the strongest agent baseline, and required a 1.2-fold smaller library to recover the same number of hits at the top-500 reference point. AutoScreen reached mean average precision more than two orders-of-magnitude above random baseline. Further, we validated AutoScreen in cancer immune-evasion case studies focusing on natural killer (NK) and T-cell therapeutics. AutoScreen recovered NK-resistance genes that were initially lower-ranked in a leukemia screen, moving MUC1, PDPN, and LRRC15 from raw ranks of 118, 81, 1384 to 5, 44, 659, respectively. In follow-up tumor killing assay with primary human NK cells, individual perturbation validated all three hits successfully. Next, in prospective benchmarking across two cytotoxic T-cell-killing screens, AutoScreen recovered 77.1% of ground-truth hits called by the consensus of two gold-standard analysis pipelines (FDR<0.1), achieving 8% improvement over the strongest general-purpose LLM agent baseline. Finally, we re-analyzed >500 public datasets to construct the AutoScreen Resource Hub, a growing knowledge base of pre-computed, annotated reports for CRISPR screens, RNA-seq differential expression, and gene-level UK Biobank genome-wide association studies. This agentic AI approach enables real-time genomics benchmark construction that are continuously updated, expanded, for evaluating frontier AI co-scientists. Overall, AutoScreen enables AI-powered target discovery to be more auditable, scalable, extensible, and reproducible.

bioengineering↗

bioRxiv2026

Targeting Macrophage Connexin 43 Hemichannels Suppresses Synovial Inflammation and Osteoarthritis

Osteoarthritis (OA) is the most common chronic joint disease and a leading cause of disability worldwide, yet no disease-modifying therapies are available. Although OA has traditionally been viewed as a cartilage-centered degenerative disease, growing evidence suggests that synovial inflammation is a critical driver of disease progression. However, the upstream mechanisms and actionable targets that govern this inflammatory response remain poorly defined. Connexin 43 (Cx43), a key regulator of musculoskeletal homeostasis that forms both gap junctions and hemichannels (HCs), is upregulated in osteoarthritic cartilage and synovium, but the specific contribution of Cx43 channels remains unclear. Here, we show that selective inhibition of Cx43 HCs attenuates OA progression in a murine post-traumatic OA model. Using adeno-associated virus (AAV)-mediated gene deletion, we found that Cx43 knockdown in joint tissues reduced cartilage erosion, synovial hyperplasia, subchondral bone sclerosis, and OA-associated pain. To dissect HC-specific functions, we developed a monoclonal antibody, Cx43(M1), that blocks Cx43 HCs without affecting junctional communication. Cx43(M1) preserved cartilage integrity, suppressed synovial inflammation, and alleviated pain behaviors, even when treatment was initiated after established cartilage damage. Mechanistically, the antibody preferentially localized to synovial macrophages, where it inhibited HC activity, reduced ATP and PGE2 release, and suppressed pro-inflammatory gene expression. Macrophage-specific Cx43 deletion recapitulated these protective effects, and Cx43 HC blockade in primary human OA synovial macrophages similarly inhibited inflammatory and catabolic responses. These findings identify synovial macrophage Cx43 HCs as upstream drivers of OA and support a paradigm shift from a predominantly chondrocyte-centered view toward synovial macrophage directed intervention.

immunology↗

bioRxiv2026

Diffusion MR Microscopy of the Human Hippocampus at 50 μm Resolution at 14.1 T for 3D Virtual Histology

Bridging the gap between macroscopic neuroimaging and microscopic histology remains a challenge in neuroscience. Here, we present a technical feasibility study of diffusion MR microscopy (dMRM) in the ex vivo human hippocampus using a 14.1 T ultrahigh-field MRI system. Mid-hippocampal blocks from control and Alzheimers disease (AD) patients were acquired at 50 x 50 x 100 {micro}m3, representing, to our knowledge, the highest resolution reported for human brain diffusion MRI. At a spatial scale approaching cellular dimensions, dMRM maps and tractography showed close correspondence with immunohistological markers of axons and dendrites. The acquired data delineated six hippocampal laminae, distinguished the inner and outer molecular layers, and preserved short, lamina-specific trajectories that became attenuated or blurred at 100 {micro}m resolution. Tractography further depicted the major components of the hippocampal tri-synaptic circuit--the perforant path, mossy fibers, and Schaffer collaterals in 3D. In the AD specimen, dMRM revealed reduced laminar contrast, local discontinuities, and altered radial and tangential fibers that were paralleled by pathological staining. With the 3D nondestructive characterization of human brain microstructure and circuitry, ex vivo dMRM showed its potential toward the goal of 3D virtual histology.

neuroscience↗