Biology subjects

Wu, J.

Publications and source records attributed to Wu, J..

At least 253 records · Page 14Linked to original sources

bioRxiv2025

The Mn-motif protein MAP6d1 assembles ciliary doublet microtubules

Most eukaryotic cells have cilia that serve vital functions in sensing, signaling, motility. The core architecture of cilia is an array of microtubule doublets, which consist of a complete A-tubule and an incomplete B-tubule. The mechanisms governing the assembly of this complex structure remain poorly understood. Here, using total internal reflection fluorescence microscopy and cryo-electron tomography, we investigate the role of MAP6d1, a brain-specific protein containing microtubule lumen-targeting Mn-motifs. We show that MAP6d1 assembles stable microtubule doublets by recruiting tubulin dimers onto the lattice of the A-tubule to initiate the nucleation of the B-tubule. MAP6d1 also promotes the formation of luminal protofilaments in singlet and doublet microtubules, a previously undescribed phenomenon that likely enhances microtubule stability. In neurons, MAP6d1 localises to the proximal part of primary cilia via its Mn-motif, with its loss resulting in shortened cilia, a characteristic of ciliopathies. MAP6d1 is thus the first microtubule-associated protein found to assemble microtubule doublets, uncovering new functions for Mn-motif proteins in neurons.

cell biology↗

bioRxiv2025

Integrating multi-modal transcriptomics identifies cellular subtypes with distinct roles in PDAC progression

BackgroundPancreatic ductal adenocarcinoma (PDAC) features a complex tumor microenvironment (TME) that significantly influences patient outcomes. Understanding the TMEs cellular composition and interactions is crucial for identifying therapeutic targets to improve treatment. MethodsWe performed an integrative analysis combining 88 single-cell RNA sequencing (scRNA-seq) samples with 187,520 cells, 20 Visium spatial transcriptomics (ST-seq) samples with 67,933 spots, and 1383 bulk RNA-seq samples and 2 Xenium high-resolution ST-seq samples with 307,679 cells, to delineate and characterize distinct subpopulations of fibroblasts, macrophages, T/NK cells, and B/plasma cells within the PDAC microenvironment. Correlations among major cell types across 12 PDAC datasets were assessed through gene set variation analysis (GSVA). Kaplan-Meier survival analysis was used to evaluate the prognostic significance of cell subtypes. Furthermore, multiplex immunohistochemistry (mIHC) and the Xenium platform was utilized to validate cellular interactions in human PDAC tissues at both protein and RNA levels. ResultsSix fibroblast subtypes and eight macrophage/monocyte subtypes were identified. POSTN high fibroblasts and SPP1 high macrophages highly infiltrated tumor tissues and were associated with poor prognosis. Most immune cell subtypes mediate adverse prognoses, except for CCL4 high CD8+ T EFF cells and IGHG1 high plasma cells, which are linked to favorable outcomes. ST-seq revealed spatial colocalization of POSTN high Fibro and SPP1 high Macro cells, as well as CCL4 high CD8+ T EFF and IGHG1 high plasma cells. These findings were corroborated by mIHC and validated using Xenium spatial transcriptomics with single-cell resolution, which confirmed the expression and spatial proximity of these markers at both the gene and protein levels. ConclusionThis integrative analysis of the PDAC TME underscores the prognostic importance of POSTN high fibroblasts and SPP1 high macrophages, while also highlighting the protective roles of CCL4 high CD8+ T EFF cells and IGHG1 high plasma cells. These insights could improve PDAC treatment strategies.

cancer biology↗

bioRxiv2025

Infectome Analysis of Small Mammals in Southern China Reveals Ecological Associations and Emerging Threats from Diverse Pathogens

Small mammals harbor a diverse array of zoonotic pathogens. To date, however, metagenomic surveys of these species have primarily focused on viral diversity, with limited attention paid to bacteria and eukaryotic pathogens. Additionally, the ecological determinants of pathogen diversity within these mammals have not been systematically examined. Herein, we employed a metatranscriptomics approach to survey the pathogen infectome of 2,408 individual samples, representing lung, liver, and gut tissues from 858 animals collected throughout Guangdong province, China, with a study design that accounted for host species, tissue, season, and geographic location. We identified 76 pathogen species, including 29 RNA viruses, 12 DNA viruses, 5 bacteria, and 30 eukaryotic pathogens, 33 of which were newly discovered. Tissue distribution analysis revealed distinct organotropisms, suggesting varied transmission routes, while host distribution analysis showed that each animal carried an average of one pathogen, with 10 pathogens widely distributed among mammalian orders. Our characterization of the geographic and seasonal patterns revealed that pathogen richness was primarily influenced by region, host species, and season, while pathogen composition was largely shaped by host genetic distance. Collectively, these data provide the first comprehensive insight into the dynamics of the pathogen infectome in these key mammalian disease reservoirs, highlighting the major factors driving pathogen diversity and transmission.

microbiology↗

bioRxiv2025

scKAN: Interpretable Single-cell Analysis for Cell-type-specific Gene Discovery and Drug Repurposing via Kolmogorov-Arnold Networks

Single-cell analysis has revolutionized our understanding of cellular heterogeneity, yet current approaches face challenges in efficiency and interpretability. In this study, we present scKAN, a framework that leverages Kolmogorov-Arnold Networks for interpretable single-cell analysis through three key innovations: efficient knowledge transfer from large language models through a lightweight distillation strategy; systematic identification of cell-type-specific functional gene sets through KANs learned activation curves; and precise marker gene discovery enabled by KANs importance scores with potential for drug repurposing applications. The model achieves superior performance on cell-type annotation with a 6.63% improvement in macro F1 score compared to state-of-the-art methods. Furthermore, scKANs learned activation curves and importance scores provide interpretable insights into cell-type-specific gene patterns, facilitating both gene set identification and marker gene discovery. We demonstrate the practical utility of scKAN through a case study on pancreatic ductal adenocarcinoma, where it successfully identified novel therapeutic targets and potential drug candidates, including Doconexent as a promising repurposing candidate. Molecular dynamics simulations further validated the stability of the predicted drug-target complexes. Our approach offers a comprehensive framework for bridging single-cell analysis with drug discovery, accelerating the translation of single-cell insights into therapeutic applications.

bioinformatics↗

bioRxiv2025

EpiFoundation: A Foundation Model for Single-Cell ATAC-seq via Peak-to-Gene Alignment

Foundation models exhibit strong capabilities for downstream tasks by learning generalized representations through self-supervised pre-training on large datasets. While several foundation models have been developed for single-cell RNA-seq (scRNA-seq) data, there is still a lack of models specifically tailored for single-cell ATAC-seq (scATAC-seq), which measures epigenetic information in individual cells. The principal challenge in developing such a model lies in the vast number of scATAC peaks and the significant sparsity of the data, which complicates the formulation of peak-to-peak correlations. To address this challenge, we introduce EpiFoundation, a foundation model for learning cell representations from the high-dimensional and sparse space of peaks. Epi-Foundation relies on an innovative cross-modality pre-training procedure with two key technical innovations. First, EpiFoundation exclusively processes the non-zero peak set, thereby enhancing the density of cell-specific information within the input data. Second, EpiFoundation utilizes dense gene expression information to supervise the pretraining process, aligning peak-to-gene correlations. EpiFoundation can handle various types of downstream tasks, including cell-type annotation, batch correction, and gene expression prediction. To train and validate EpiFoundation, we curated MiniAtlas, a dataset of 100,000+ single cells with paired scRNA-seq and scATAC-seq data, along with diverse test sets spanning various tissues and cell types for robust evaluation. EpiFoundation demonstrates state-of-the-art performance across multiple tissues and diverse downstream tasks.

genomics↗

bioRxiv2025

RFdiffusion Exhibits Low Success Rate in De Novo Design of Functional Protein Binders for Biochemical Detection

The design of high-affinity protein binders is critical for biochemical detection, yet traditional methods remain labor-intensive. AI-driven tools like RFdiffusion, a RoseTTAFold-based diffusion model, offer promising alternatives for generating protein structures with tailored binding interfaces. This study evaluates RFdiffusions efficacy in designing de novo binders for six targets: Strep-Tag II (a peptide tag) and five eukaryotic proteins (STAT3, FGF4, EGF, PDGF-BB, and CD4). Five binders were designed for each target and experimentally validated. While two Strep-Tag II binders outperformed streptavidin in Western blot assays, none matched the sensitivity of anti-Strep-Tag II antibodies. Binders for the other targets failed due to low expression, nonspecific binding, or undetectable affinity. Despite generating structurally diverse candidates, RFdiffusions success rate was limited by low-affinity designs and inconsistent recombinant expression. These results underscore the need for further optimization of AI-driven protein design tools for practical biochemical applications.

bioengineering↗

bioRxiv2025

DAZAP2 functions as a pan-coronavirus restriction factor by inhibiting viral entry and genomic replication

The SARS-CoV-2 pandemic and the emergence of novel variants underscore the need to understand host-virus interactions and identify host factors that restrict viral infection. Here, we perform a genome-wide CRISPR knockout screen to identify host restriction factors for SARS-CoV-2, revealing DAZAP2 as a potent antiviral gene. DAZAP2, previously implicated in SARS-CoV-2 restriction, is found to inhibit viral entry by blocking virion fusion with both endolysosomal and plasma membranes. Additionally, DAZAP2 suppresses genomic RNA replication without affecting the primary translation of viral replicases. We demonstrate that DAZAP2 functions as a pan-coronavirus restriction factor across four genera of coronaviruses. Importantly, knockout of DAZAP2 enhances SARS-CoV-2 infection in mouse models and in human primary airway epithelial cells, confirming its physiological relevance. Mechanistically, antiviral activity of DAZAP2 appears to be indirect, potentially through the regulation of host gene expression, as it primarily localizes to the nucleus. Our findings provide new insights into the host defense system against coronaviruses and highlight DAZAP2 as a potential target for host-directed antiviral therapies. IMPORTANCEDuring viral infection, the host defense response is mediated by a variety of host factors through distinct mechanisms that have yet to be fully elucidated. Although DAZAP2 was previously implicated in SARS-CoV-2 restriction, its mechanisms of action and in vivo relevance remain unclear. In this study, we identify the DAZAP2 as a potent pan-coronavirus restriction factor that inhibits viral infection through dual mechanisms: blocking virion fusion with both endolysosomal and plasma membranes, and suppressing genomic RNA replication. We confirm its physiological relevance in host defense using mouse models and primary cell cultures. This study advances our understanding of host-pathogen interactions. Targeting DAZAP2 or its regulatory pathways could provide a new approach to enhance host defense against current and future coronavirus threats.

microbiology↗

bioRxiv2025

Master control genes in the regeneration of rod photoreceptors from endogenous progenitor cells in zebrafish retina

Retinitis Pigmentosa is a chronic retinal degenerative disease characterized by the gradual loss of rod, and later, cone photoreceptors until the individual is completely blind. Regeneration of photoreceptors from endogenous progenitor cells is a possible therapeutic approach, but mammals do not do this naturally. Mammalian models can be induced to generate retinal progenitors from Muller glial cells, but there has been limited success in rod photoreceptor specific regeneration. Unlike mammals, zebrafish have the natural ability to regenerate neurons after injury or disease and can provide insight into the molecular mechanisms of regeneration. In this study, we used a zebrafish model of Retinitis Pigmentosa to investigate the class of progenitors responsible for rod photoreceptor regeneration in the context of chronic disease. Using bioinformatic analyses of single-cell RNA sequencing datasets, we identified master regulator genes responsible for proliferation of retinal progenitors, differentiation of progenitors into rod photoreceptors, and maturation of the new rod photoreceptors. Using transient knockdown of gene expression in adult regenerating retina we determined that e2f1, e2f2, e2f3 and aurkb are critical for proliferation of progenitors, and prdm1a is critical for differentiation of progenitors into rod photoreceptors. This study provides a list of master regulators responsible for the specific regeneration of rod photoreceptors during chronic retinal degeneration. Impact StatementIdentification of master regulating genes that drive the proliferation of progenitor cells and their differentiation specifically into rod photoreceptors provides insight that can be used to develop regenerative therapies for retinal degenerative diseases.

neuroscience↗

bioRxiv2025

Deletion of PIEZO1 in adult cardiomyocytes accelerates cardiac aging and causes premature death

Mechanosensitive PIEZO1 channels have emerged as key transducers of mechanical forces in the cardiovascular system. In cardiomyocytes, we previously showed that PIEZO1 decodes mechanical cues driving pressure-overload induced hypertrophy. However, conflicting reports exist on the influence of PIEZO1 on baseline cardiac function. Here we show that conditional deletion of Piezo1 from cardiomyocytes in adult mice results in premature mortality. The hearts from these mice exhibited signs of accelerated aging, including elevated markers of the senescence associated secretory phenotype, with significant blunting of the normal cardiac hypertrophic response to aging, associated with a reduction in the activation of the pro-hypertrophic Ca2+/calmodulin-dependent protein kinase II (CaMKII). Functionally, aged-Piezo1 KO mice exhibited impaired cardiac relaxation due to altered cellular Ca2+ handling kinetics. Young adult Piezo1 KO mice exhibited a normal resting heart rate but developed significant progressive sinus bradycardia and cardiac fibrotic remodelling with aging, which was most prominent in the right atrium, where Piezo1 expression is highest in the healthy heart. Mechanistically, loss of PIEZO1 was associated with a marked reduction in the anti-fibrotic molecule, atrial natriuretic peptide (ANP). Moreover, in vivo, ANP release instigated by atrial stretch was markedly blunted in conditional Piezo1 KO mice, providing a plausible and long sought-after mechanism for the link between mechanical stretch and ANP release. Taken together, our data show that PIEZO1 is a crucial homeostatic molecule during cardiac aging, enabling adaptation to an aging tissue microenvironment.

biophysics↗

bioRxiv2025

Avoidance of fratricide and improved efficacy of CD33-MSLN CAR-iNK cells in treating acute myeloid leukemia

Acute myeloid leukemia (AML) patients are often older, which brings challenges of endurance and persistent efficacy of autologous CAR-T cell therapies. Allogenic CAR-NK cell therapies may offer reduced toxicities and enhanced anti-leukemic potential against AML. In this study, we designed a novel CD33-mesothelin loop CAR (Loop CAR) and evaluated its anti-tumor efficacy in human umbilical cord blood-derived NK (UCB-NK) cells and human pluripotent stem cell-derived NK (hPSC-iNK) cells. The Loop CAR exhibited superior cytotoxicity against dual-antigen-positive tumor cell lines and primary AML cells. To further avoid fratricide caused by endogenous CD33 expression in NK cells, we established a hPSC-derived cell line via knockout of CD33 gene (CD33KO) and engineered Loop CAR. Rather than enforced expression of exogenous CD16, we generated abundant mature CD33KO-Loop CAR-iNK cells highly expressing endogenous CD16 via an organoid induction approach. This innovative strategy effectively mitigated NK cell fratricide and significantly enhanced CD33 and mesothelin-mediated specific cytotoxicity. Moreover, the CD33KO-Loop CAR-iNK cells demonstrated superior tumor-killing activity in AML xenograft mice and significantly prolonged survival. hPSC-derived CD33KO-Loop CAR-iNK cells possess unique advantages and translational potential for treating AML.

cancer biology↗

bioRxiv2025

Engineered CRO-CD7 CAR-NK cells derived from pluripotent stem cells avoid fratricide and efficiently suppress human T-cell malignancies

BackgroundT cell malignancies are highly aggressive hematological tumors with limited effective treatment options. CAR-NK cell therapy targeting CD7 has emerged as a promising approach for treating T-cell malignancies. However, conventional CAR-NK cell therapy faces the challenges of cell fratricide due to CD7 expression on both malignant cells and normal NK cells. Additionally, engineering CARs into human tissue-derived NK cells demonstrates heterogeneity, low transduction efficiency, and high manufacturing costs. MethodsThe human pluripotent stem cells (hPSCs) were genetically modified by knocking out the CD7 gene and introducing the CD7 CAR expression cassette to generate CD7 KO-CD7 CAR-hPSCs. These modified hPSCs were subsequently differentiated into CD7 KO-CD7 CAR-iNK cells using an efficient organoid induction method. The cytotoxicity of CD7 KO-CD7 CAR-iNK cells against CD7+ tumor cells was evaluated. Furthermore, we overexpressed the CXCR4 gene in CD7 KO-CD7 CAR-hPSCs and derived CXCR4-expressing CD7 KO-CD7 CAR-iNK (CRO-CD7 CAR-iNK) cells. The dynamics of CRO-CD7 CAR-iNK cells in vivo were tracked, and their therapeutic efficacy was assessed using human T-cell acute lymphoblastic leukemia (T-ALL) xenograft models. ResultsThe CD7 KO-CD7 CAR-iNK cells derived from CD7 KO-CD7 CAR-hPSCs effectively avoided fratricide, demonstrated normal expansion, and exhibited potent and specific anti-tumor activity against CD7+ T-cell tumor cell lines and primary T-ALL cells. CXCR4 overexpression in CRO-CD7 CAR-iNK cells improved their homing capacity and extended their persistence in vivo. The CRO-CD7 CAR iNK cells significantly suppressed tumor growth and prolonged the survival of T-ALL tumor-bearing mice. ConclusionsOur study provides a reliable strategy for the large-scale generation of fratricide-resistant CD7 CAR-iNK cells with robust anti-tumor effects from hPSCs, offering a promising cell product to treat T-cell malignancy.

cancer biology↗

bioRxiv2025

Unsupervised multi-animal tracking for quantitative ethology

Quantitative ethology necessitates accurate tracking of animal locomotion, especially for population-level analyses involving multiple individuals. However, current methods rely on laborious annotations for supervised training and have restricted performance in challenging conditions. Here we present an unsupervised deep-learning method for multi-animal tracking (UDMT) that achieves state-of-the-art performance without requiring human annotations. By synergizing a bidirectional closed-loop tracking strategy, a spatiotemporal transformer network, and three sophisticatedly designed modules for localization refining, bidirectional ID correction, and automatic parameter tuning, UDMT can track multiple animals accurately in various challenging conditions, such as crowding, occlusion, rapid motion, low contrast, and cross-species experiments. We demonstrate the versatility of UDMT on five different kinds of model animals, including mice, rats, Drosophila, C. elegans, and Betta splendens. Combined with a head-mounted miniaturized microscope, we illustrate the power of UDMT for neuroethological interrogations to decipher the correlations between animal locomotion and neural activity. UDMT will facilitate advancements in ethology by providing a high-performance, annotation-free, and accessible tool for multi-animal tracking.

animal behavior and cognition↗

bioRxiv2025

Benchmarking and optimizing Perturb-seq in differentiating human pluripotent stem cells

Perturb-seq is a powerful approach to systematically assess how genes and enhancers impact the molecular and cellular pathways of development and disease. However, technical challenges have limited its application in stem cell-based systems. Here, we benchmarked Perturb-seq across multiple CRISPRi modalities, on diverse genomic targets, in multiple human pluripotent stem cells, during directed differentiation to multiple lineages, and across multiple sgRNA delivery systems. To ensure cost-effective production of large-scale Perturb-seq datasets as part of the Impact of Genomic Variants on Function (IGVF) consortium, our optimized protocol dynamically assesses experiment quality across the weeks-long procedure. Our analysis of 1,996,260 sequenced cells across benchmarking datasets reveals shared regulatory networks linking disease-associated enhancers and genes with downstream targets during cardiomyocyte differentiation. This study establishes open tools and resources for interrogating genome function during stem cell differentiation.

genomics↗

bioRxiv2025

Super-omics and PCAS reveal rheumatoid arthritis as a tumor-like disease

Rheumatoid arthritis (RA) is a debilitating systemic autoimmune disorder that significantly impairs quality of life. To elucidate the molecular alterations in RA progression, we performed comprehensive super-omics analyses on synovial tissues from patients with joint trauma, arthritis, and RA. These analyses included transcriptomics, proteomics, metabolomics, microbiomics, and fourteen PTMs (phosphorylation, acetylation, lactylation, O-GlcNAc glycosylation, arginine monomethylation, lysine monomethylation, lysine dimethylation, lysine trimethylation, succinylation, malonylation, glutarylation, tyrosine nitration, N-glycosylation, and O-glycosylation). Additionally, we developed a protein-centered association study (PCAS) method to integrate these complex datasets. Using this approach, we identified key proteins, such as STK17B, and its interactors, which may be crucial in RA pathogenesis. Tumor-like features, including aberrant angiogenesis and epithelial-mesenchymal transition, were observed in RA, alongside the potential involvement of oncogenes and tumor suppressor genes. Finally, we constructed a molecular interaction blueprint of RA, providing a comprehensive framework for advancing RA pathogenesis, diagnosis, and therapy.

immunology↗

bioRxiv2025

CPI-Pred: A deep learning framework for predicting functional parameters of compound-protein interactions

Recent advancements in deep learning have enabled functional annotation of genome sequences, facilitating the discovery of new enzymes and metabolites. However, accurately predicting compound-protein interactions (CPI) from sequences remains challenging due to the complexity of these interactions and the sparsity and heterogeneity of available data, which constrain the generalization of patterns across their solution space. In this work, we introduce CPI-Pred, a versatile deep learning model designed to predict compound-protein interaction function. CPI-Pred integrates compound representations derived from a novel message-passing neural network and enzyme representations generated by state-of-the-art protein language models, leveraging innovative sequence pooling and cross-attention mechanisms. To train and evaluate CPI-Pred, we compiled the largest dataset of enzyme kinetic parameters to date, encompassing four key metrics: the Michaelis-Menten constant (KM), enzyme turnover number (kcat), catalytic efficiency (kcat/KM), and inhibition constant (KI).These kinetic parameters are critical for elucidating enzyme function in metabolic contexts and understanding their regulation by compounds within biological networks. We demonstrate that CPI-Pred can predict diverse types of CPI using only the amino acid sequence of enzymes and structural representations of compounds, outperforming state-of-the-art models on unseen compounds and structurally dissimilar enzymes. Over workflow provides a valuable tool for tackling a range of metabolic engineering challenges, including the designing of novel enzyme sequences and compounds, such as enzyme inhibitors. Additionally, the datasets curated in this study offer a valuable resource for the scientific community, serving as a benchmark for machine learning models focused on enzyme activity and promiscuity prediction.

bioinformatics↗

bioRxiv2025

A Novel Optimized Silver Nitrate Staining Method for Visualizing the Osteocyte Lacuno-Canalicular System

A 50% silver nitrate solution is commonly used to stain and characterize the osteocyte lacuno-canalicular system (LCS) in bone biology research. However, variability in reagent concentrations and types, along with inconsistent staining procedures, have limited the broader application of this method in osteocyte research. In this study, we present a novel optimized silver nitrate staining technique aimed at addressing these limitations. This new method utilizes a 1 mol/L silver nitrate solution in combination with a type-B gelatin-formic acid solution at various concentrations (0.05%-0.5% gelatin and 0.05%-5% formic acid, or 1%-2% gelatin and 0.1%-2% formic acid) in volume ratios of 4:1, 2:1, or 1:1, or a 0.5 mol/L silver nitrate solution at a 4:1 ratio. The staining process is carried out for 1 hour under ultraviolet light or 90 minutes under regular room light (or dark), followed by washing with Milli-Q water to terminate the reaction. We applied this new method to stain the osteocyte LCS in bone samples from different species and pathological bone models. The technique consistently produced clear, distinct staining patterns across all samples. Moreover, our novel method revealed a greater number of LCS compared to the traditional 50% silver nitrate solution. This suggests that the commonly used 50% silver nitrate method may disrupt or inadequately reveal the LCS in certain species, potentially leading to an underestimation of LCS density and number. In conclusion, our novel silver nitrate staining method provides a simpler and more cost-effective alternative to the traditional technique. By offering a more accurate and comprehensive analysis of the LCS across species, this approach has the potential to advance research on osteocyte morphogenesis, as well as the functional and evolutionary adaptations of the osteocyte LCS across different taxa.

pathology↗

bioRxiv2025

Age-Associated Decline in Autophagy Pathways in Retinal Pigment Epithelium and Protective Effects of Topical Trehalose in Light-induced Outer Retinal Degeneration in Mice

Age is a primary risk factor for chronic conditions, including age-related macular degeneration (AMD). Impairments in autophagy processes are implicated in AMD progression, but the extent of autophagys contribution and its therapeutic potential remain ambiguous. This study investigated age-associated transcriptomic changes in autophagy pathways in the retinal pigment epithelium (RPE) and evaluated the protective effects of topical trehalose, an autophagy-enhancing small molecule, against light-induced outer retinal degeneration in mice. Transcriptomic analysis of human RPE/choroid and mouse RPE revealed consistent downregulation of autophagy pathways with age, alongside variable changes as AMD severity progressed. Given the age- and AMD-associated perturbation of autophagy pathways, we examined trehalose treatment in vitro, which enhanced autophagic flux and restored mitochondrial respiratory function in primary murine RPE cells exposed to oxidative stress. In vivo, topical trehalose improved autophagy-lysosome activity in mouse RPE, demonstrated by elevated LC3B turnover and SQSTM1/p62 degradation. Furthermore, trehalose eyedrops protected mice from light-induced damage to the RPE and photoreceptors, preserving outer nuclear layer thickness, RPE morphology, and junctional F-actin organization. Taken together, the data support that age-related decline and severe dysregulation in autophagy contributed to AMD progression. By restoring autophagic flux, topical trehalose demonstrates therapeutic potential to address early autophagy-related pathological changes in AMD.

pathology↗

bioRxiv2025

Exploring the Suitability of Piecewise-Linear Dynamical System Models for Cognitive Neural Dynamics

Dynamical system models have proven useful for decoding the current brain state from neural activity. So far, neuroscience has largely relied on either linear models or nonlinear models based on artificial neural networks. Piecewise linear approximations of nonlinear dynamics have proven useful in other technical applications, providing a clear advantage over network-based models, when the dynamical system is not only supposed to be observed, but also controlled. Here we explore whether piecewise-linear dynamical system models (recurrent Switching Linear Dynamical System or rSLDS models) could be useful for modeling brain dynamics, in particular in the context of cognitive tasks. We first generate artificial neural data based on a nonlinear computational model of perceptual decision-making and demonstrate that piecewise-linear dynamics can be successfully recovered from these observations. We then demonstrate that the piecewise-linear model outperforms a linear model in terms of predicting future states of the system and associated neural activity. Finally, we apply our approach to a publicly available dataset recorded from monkeys performing perceptual decisions. Much to our surprise, the piecewise-linear model did not provide a significant advantage over a linear model for these particular data, although linear models that were estimated from different trial epochs showed qualitatively different dynamics. In summary, we present a dynamical system modeling approach that could prove useful in situations, where the brain state needs to be controlled in a closed-loop fashion, for example, in new neuromodulation applications for treating cognitive deficits. Future work will have to show under what conditions the brain dynamics are sufficiently nonlinear to warrant the use of a piecewise-linear model over a linear one.

neuroscience↗