Biology subjects

Wu, H.

Publications and source records attributed to Wu, H..

At least 145 records · Page 8Linked to original sources

bioRxiv2025

Integrated Single-Cell and Spatial Transcriptomic Analysis Reveals an Aging-Associated Fibroblast Subtype Linked to Tumor Progression in Human Skin

Aging is a major risk factor for the development of many cancers, yet the mechanisms underlying this increased susceptibility remain incompletely understood. Traditionally, the accumulation of genetic mutations over time has been considered a primary driver of age-related tumorigenesis. However, emerging evidence highlights the critical role of the aging tissue microenvironment--including changes in immune, stromal, and epithelial compartments--in shaping cancer initiation and progression. In this study, we employed integrated single-cell and spatial transcriptomics to systematically characterize age-associated alterations in human skin and skin cancers. Our analysis uncovered widespread, cell type-specific transcriptional reprogramming with age, revealing key pathways and cellular populations that may contribute to a pro-tumorigenic environment. Notably, we identified a previously unrecognized fibroblast subtype marked by SFRP2 expression, which expands with age and is associated with poor prognosis in basal cell carcinoma. These fibroblasts appear to enhance Wnt signaling within the tumor niche, suggesting a potential mechanism by which the aging stroma supports malignancy. Collectively, our findings shed light on how age-related changes in tissue ecosystems may predispose to cancer and point to novel therapeutic opportunities for targeting the aged tumor microenvironment.

cancer biology↗

bioRxiv2025

A Novel FNDC1-NAMPT-NAD axis is Implicated in Small and Large-vessel Arterial Disease and Drives Vascular Calcification

Vascular calcification represents a convergent pathological feature of diverse cardiovascular diseases, yet the upstream molecular programs orchestrating this process remain poorly defined. Here, we uncover fibronectin type III domain-containing 1 (FNDC1) as a previously unrecognized regulator of vascular calcification across both microvascular and macrovascular beds. Integrative transcriptomic profiling of human calciphylaxis lesions and atherosclerotic coronaries identified FNDC1 as one of the most significantly upregulated genes. In primary human vascular smooth muscle cells, FNDC1 drove osteogenic phenotype switch and vascular calcification through activation of PI3K/AKT signaling and metabolic reprogramming. Mechanistically, FNDC1 directly binds to nicotinamide phosphoribosyltransferase (NAMPT) resulting in elevated intracellular NAD levels, thus coupling vascular signaling to control of NAD biosynthesis. In murine models, genetic deletion of Fndc1 or pharmacologic inhibition of NAMPT suppressed arterial calcification and prolonged survival. Clinically, circulating FNDC1 levels were elevated in patients with both calciphylaxis and coronary artery disease and independently predicted cardiovascular risk in 42,687 UK Biobank participants. Together, these findings establish FNDC1 as a central mediator of vascular pathology and highlight the FNDC1- NAMPT-NAD+ axis as a promising target for therapeutic intervention.

cell biology↗

bioRxiv2025

A k-mer-based estimator of the substitution rate between repetitive sequences

K-mer-based analysis of genomic data is ubiquitous, but the presence of repetitive k-mers continues to pose problems for the accuracy of many methods. For example, the Mash tool (Ondov et al 2016) can accurately estimate the substitution rate between two low-repetitive sequences from their k-mer sketches; however, it is inaccurate on repetitive sequences such as the centromere of a human chromosome. Follow-up work by Blanca et al. (2021) has attempted to model how mutations affect k-mer sets based on strong assumptions that the sequence is non-repetitive and that mutations do not create spurious k-mer matches. However, the theoretical foundations for extending an estimator like Mash to work in the presence of repeat sequences have been lacking. In this work, we relax the non-repetitive assumption and propose a novel estimator for the mutation rate. We derive theoretical bounds on our estimators bias. Our experiments show that it remains accurate for repetitive genomic sequences, such as the alpha satellite higher order repeats in centromeres. We demonstrate our estimators robustness across diverse datasets and various ranges of the substitution rate and k-mer size. Finally, we show how sketching can be used to avoid dealing with large k-mer sets while retaining accuracy. Our software is available at https://github.com/medvedevgroup/Repeat-Aware_Substitution_Rate_Estimator.

bioinformatics↗

bioRxiv2025

Hierarchical microstructural tissue growth of the gray and white matter of human visual cortex during the first year of life

Development of gray and white matter tissue microstructure is critical for the emergence of sensory and cognitive functions. However, it is unknown how microstructural tissue properties of the human visual system develop in the first year of human life. Here, we use tissue relaxation rate (R1) obtained using quantitative MRI to measure the longitudinal development of gray and white matter in brain areas spanning three visual processing streams: dorsal, lateral, and ventral, during the first year of life. R1 in gray and white matter of all visual regions in the three processing streams increases postnatally, indicating microstructural tissue growth. R1 increases faster between 0-6 months than 6-12 months, and faster in white matter than gray matter, with white matter R1 surpassing that of gray matter after two months of age. Strikingly, this microstructural growth is hierarchical: across all streams, early visual areas are more mature at birth than higher-level areas but develop more slowly postnatally than higher-level areas. The exception is TO1 (MT) which is similar to V1: it is microstructurally more mature at birth and develops slower than neighboring areas. Overall, our findings provide the first comprehensive measurement of microstructural tissue growth in infancy across three visual processing streams and propose a new hypothesis that functional development of the visual cortex may be coupled with microstructural development and follows a similar hierarchical trajectory.

neuroscience↗

bioRxiv2025

KDM5A enhances MHC-I antigen presentation and potentiates antitumor immunity unleashed by immune checkpoint blockade

While some studies have implicated the histone 3 lysine 4 demethylase KDM5A in tumorigenesis and promotion of metastasis, KDM5A has also been shown to boost immunotherapy responses in melanoma. The distinct functional role of KDM5A in the context of immunotherapy and the underlying mechanisms remain largely unknown. Here, we report that higher KDM5A expression strongly correlates with higher MHC-I expression, CD8+ T cell infiltration, and prolonged survival in colorectal cancer and gastric cancer patients receiving anti- PD-1 therapy. KDM5A promotes MHC-I-mediated antigen presentation and CD8+ T cell-mediated tumor killing across multiple murine cancer models. Mechanistically, KDM5A upregulates MHC-I expression through the SOCS1/IFN-{gamma}/STAT1 signaling and improves tumor cell-intrinsic antigen presentation capacity by inhibiting lysosomal proteases in a demethylase-dependent manner. Moreover, KDM5A directly represses lysosomal cathepsins in dendritic cells and promotes the cross-priming activity. Through a high-throughput chemical compound library-based screen, we identified Clomiphene, an FDA-approved small molecule, which significantly elevates KDM5A expression, and enhances antitumor immunity during anti-PD-1 immunotherapy in mouse models of melanoma and colon cancer. Our findings that KDM5A functions in MHC-I-mediated immune activation during anti-PD-1 therapy present an opportunity for developing KDM5A-enhancing therapies to increase tumor immunogenicity, sensitize solid tumors to immunotherapy and exclude immune evasion.

immunology↗

bioRxiv2025

Highly sensitive and scalable time-resolved RNA sequencing in single cells with scNT-seq2

Understanding gene expression dynamics requires resolving newly synthesized RNAs from pre-existing pools at single-cell resolution. Here, we present scNT-seq2, a highly sensitive and scalable method for time-resolved single-cell RNA sequencing. By systematically optimizing the second-strand cDNA synthesis (2nd SS) step, we substantially improved read alignment rates, reduced background mutations, and enhanced library complexity compared to the original scNT-seq 1. Benchmarking in 4sU-labeled K562 cells demonstrated that scNT-seq2 accurately quantifies newly synthesized transcripts and preserves the gene level RNA turnover. The enhanced sensitivity enables robust detection of dynamic, cell-cycle state specific genes, such as S-phase regulators. Together, scNT-seq2 provides an efficient and versatile tool for dissecting transcriptional dynamics across diverse biological systems at single-cell resolution.

genomics↗

bioRxiv2025

Peptides alleviate cognitive impairment by inhibiting and disassembling amyloid-beta aggregates in Alzheimer disease

Alzheimers disease (AD) is a devasting neurodegenerative disorder characterized by {beta}-amyloid formation, further exacerbated by RIPK1/RIPK3 necrosome-induced programmed necrosis (necroptosis). We previously showed that the RIPK1/RIPK3 necrosome forms a functional amyloid complex using its RIP homotypic interaction motifs (RHIMs). Here, we discovered that the core RIPK1/RIPK3 necrosome shares strikingly structural similarity to the C-terminal region of {beta}-amyloid (A{beta}42), and the RHIM-derived tetrapeptides (IQIG or VQVG) directly inhibit A{beta} aggregation, disassemble preformed A{beta} fibrils (PFFs), and reduce RIPK1 polymerization. Also, the peptides exhibit effective membrane permeability and could reduce A{beta}40 or A{beta}42-induced neural death and TNF-induced necroptosis in SH-SY5Y cells. IQIG and VQVG injected by ICV increase learning and memory abilities by reducing A{beta} plaques and hyperphosphorylated tau in the cortex and hippocampus of APP/PS1 double-transgenic mice. Mechanistically, the peptides directly interact with A{beta} to block A{beta} aggregation and alleviate microglia-mediated neuroinflammation. Strikingly, single-cell RNA sequencing revealed that the peptides-treated AD transgenic mice restored neuronal homeostasis with increased GABAergic neurons and decreased glutamatergic neurons. Furthermore, total cell-cell interaction strength increased while the AD risk gene Apoe expression decreased in the specific oligodendrocyte subtype of peptides-treated AD mice. Thus, our findings revealed that the peptides could improve cognition and memory capabilities and serve as promising structural templates for potential drugs against AD.

neuroscience↗

bioRxiv2025

Automated SP3 Workflow Enables Robust Sample Preparation on Proteonano Ultraplex Proteomics Platform

The efficiency and reproducibility of sample preparation are critical prerequisites for obtaining reliable results in bottom-up proteomics. To conclude, we systematically evaluated three commonly used sample preparation methods: automated single-pot solid-phase-enhanced sample preparation (SP3), precipitation by organic solvents, and filter-aided sample preparation method (FASP). When processing varying input amounts of HeLa cell lysate (10, 20, 50, 100, and 150 g), the SP3 method demonstrated strong linearity (R2 = 0.95), indicating excellent quantitative consistency across a broad protein input range. A comparison between manual SP3 and manual precipitation protocols revealed comparable protein identification numbers, with a correlation coefficient of approximately 0.97. We further developed an automated version of the SP3 method based on the Proteonano Ultraplex Proteomics platform (autoSP3), which yielded highly consistent results with the manual approach, achieving a protein identification overlap of 98.5% (6,949 out of 7,056). These findings support the suitability of the autoSP3 workflow as a robust and scalable approach for high-throughput and reproducible sample preparation in proteomics sample preparation.

molecular biology↗

bioRxiv2025

TIP-seq: A Single-cell Multiomics Approach for Simultaneous Transcriptome and Intracellular Protein Profiling

Single-cell multiomics technologies have significantly advanced our understanding of cellular heterogeneity; however, the concurrent profiling of mRNA transcripts and target proteins continues to pose a substantial challenge. To bridge the gap between single-cell sequencing and proteomic methods, we developed an innovative single-cell multiomics methodology, termed TIP-seq, which facilitates the simultaneous profiling of the transcriptome and intracellular proteins without prior cell manipulation. In TIP-seq, microfluidic technology is leveraged to integrate oligonucleotide-labeled antibodies with hydrogel beads, allowing for the capture of mRNAs and proteins from individual cells within droplets, thereby enabling high-throughput and precise dual-omics data acquisition. When applied to lung cell lines and non-small cell lung cancer (NSCLC) tissues, TIP-seq analysis revealed notable cellular heterogeneity and molecular dynamics, emphasizing the distinct immune cell interactions within NSCLC tissues. Key immune checkpoint interactions in NSCLC, such as SPP1-CD44, NECTIN2-TIGIT, and NECTIN2-CTLA4, along with functional alterations in tumor-associated dendritic cells (DCs) and T cells, were identified via TIP-seq, underscoring their pivotal roles in mediating immune suppression within the tumor microenvironment. Collectively, TIP-seq represents a powerful methodology for identifying novel therapeutic targets and biomarkers, thereby holding significant potential for the advancement of precision medicine in the treatment of lung cancer and other complex diseases.

biochemistry↗

bioRxiv2025

FoldBench: An All-atom Benchmark for Biomolecular Structure Prediction

Accurate prediction of biomolecular complex structures is fundamental for understanding biological processes and rational therapeutic design. Recent advances in deep learning methods, particularly all-atom structure prediction models, have significantly expanded their capabilities to include diverse biomolecular entities, such as proteins, nucleic acids, ligands, and ions. However, comprehensive benchmarks covering multiple interaction types and molecular diversity remain scarce, limiting fair and rigorous assessment of model performance and generalizability. To address this gap, we introduce FoldBench, an extensive benchmark dataset consisting of 1,522 biological assemblies categorized into nine distinct prediction tasks. Our evaluations reveal critical performance dependencies, showing that ligand docking accuracy notably diminishes as ligand similarity to the training set decreases, a pattern similarly observed in protein-protein interaction modeling. Furthermore, antibody-antigen predictions remain particularly challenging, with current methods exhibiting failure rates exceeding 50%. Among evaluated models, AlphaFold 3 consistently demonstrates superior accuracy across the majority of tasks. In summary, our results highlight significant advancements yet reveal persistent limitations within the field, providing crucial insights and benchmarks to inform future model development and refinement.

bioinformatics↗

bioRxiv2025

Diffusion Neuroimaging of Speech Acquisition in Infants

How multisystem white-matter maturation supports early speech acquisition in infancy remains unclear. Using longitudinal diffusion MRI from the Baby Con-nectome Project, we quantified neurite density maturation from birth to 24 months across 33 white-matter tracts spanning motor, auditory, visual, ventral-language, and dorsal-language systems. Maturation was coordinated across systems rather than confined to canonical language pathways: dorsal-language association tracts began relatively immature near birth but showed steep postnatal growth, pro-gressively rebalancing with motor and sensory systems. Tract-behavior coupling was likewise distributed across developmental domains, with broadly positive age-adjusted associations for motor outcomes, a sparser, partly motor-like profile for gesture-related phrase comprehension, and predominantly negative associa-tions for lexical speech-related outcomes. Exploratory follow-up analyses further suggested motor and social precursor capacities that may help mediate these in-direct developmental routes from structural maturation to later speech-related outcomes. These findings support a multisystem developmental account in which early speech acquisition emerges not from an isolated language pathway, but from coordinated maturation across multiple systems and the precursor skills that link them.

neuroscience↗

bioRxiv2025

CCN1 forms a complex with GPC4 and heparin to fine-tune signaling activities for embryonic cortical neural stem cell maintenance

Radial glial cells (RGCs) are the neural stem cells (NSCs) in the developing brain, capable of self-renewing and generating neurons and glia in order. The developmental behaviors of NSCs are determined by extrinsic factors from the NSC niche, together with intrinsic factors. Here, we show that Cellular Communication Network 1 (CCN1), a secreted matricellular protein, is an autocrine factor derived from RGCs in the embryonic telencephalon. Loss of Ccn1 in RGCs leads to accelerated lineage progression of RGCs, and causes premature cell cycle exit and neuronal differentiation, resulting in a reduction in NSC pool size, and in turn, at the expense of the perinatal gliogenesis. Mechanistically, CCN1 interacts with an RGC-expressed membrane protein GPC4, one of the heparan sulfate proteoglycans (HSPGs), and is required for GPC4 to maintain NSCs through the Sonic Hedgehog (Shh) signaling pathway, depending on the binding of heparin to CCN1. Our work provides a potential mechanism for the dynamic and sophisticated interconnections of niche components that meet the needs of context-dependent maintenance, differentiation, and fate determination of NSCs.

developmental biology↗

bioRxiv2025

Human Messenger RNA Harbors Widespread Noncoding Splice Isoforms

Pre-messenger RNA transcribed from canonical protein-coding genes is frequently alternatively spliced into diverse mature mRNA isoforms, thus translating into various protein products. However, it remains unknown whether human mRNA from a protein-coding gene harbors noncoding splice isoforms in the genome. Herein, we discovered 15,836 mRNA noncoding splice isoforms across 7,298 protein-coding genes in human. mRNA noncoding isoforms are mainly produced by alternative splicing and polyadenylation, which display tissue-specific distributions and involve in RNA processing pathways. Notably, mRNA noncoding isoforms are frequently upregulated in cancer. Differentially expressed mRNA noncoding isoforms are associated with the cancer hallmarks and can independently predict patient survival. These findings unraveled human mRNA harbors widespread noncoding splice isoforms, and highlight the dual characters of mRNA including protein-translation isoforms and noncoding splice isoforms, providing a new dimension to our understanding of mRNA functional property.

genomics↗

bioRxiv2025

Breaking the Link: Neural Dissociation of Attention and Working Memory through Inhibitory Control

Attention and working memory (WM) encoding have traditionally been considered inseparable processes with shared neural mechanisms. Here, using an innovative experimental design and a multimodal approach, we provide the first direct neural evidence that attention and WM encoding are dissociable. Functional MRI identifies the supramarginal gyrus (SMG) as the key region enabling this dissociation, while dynamic causal modeling reveals the neural circuitry through which the SMG exerts inhibitory control over attentional representations, regulating their integration into WM. Furthermore, neuromodulation via transcranial direct current stimulation (tDCS) demonstrates that enhancing SMG activity strengthens this inhibitory control, providing causal evidence for the dissociation mechanism. A second tDCS experiment with varied stimuli confirms the generalizability of this mechanism and reinforces the robustness of our results. These findings challenge the long-standing view that attention and WM encoding form a continuous process, demonstrating instead that they constitute two dissociable neural processes of information selection.

neuroscience↗

bioRxiv2025

How infant brains fold: Sulcal deepening is linked to development of sulcal span, thickness, curvature, and microstructure

Cortical folding begins in utero as sulci emerge and continues postnatally as sulci deepen. However, the timeline and mechanisms underlying postnatal sulcal development remain unknown. Using structural and quantitative magnetic resonance imaging in infants from birth to one year of age, we longitudinally measured macroanatomical and microstructural development in major sulci that emerge in utero between the 16th and 31st gestational weeks. We find that sulci that emerge earlier in utero are deeper at birth and deepen at a slower rate postnatally than later emerging sulci. Sulci also become wider, thicker, and microstructurally denser, while their curvature decreases. Notably, mean sulcal depth is predicted by a weighted sum of sulcal span, thickness, curvature, and tissue microstructure, with differential weights across sulci. Analysis of local depth along the sulcus also reveals that deeper portions of sulci (fundi) have higher curvature and higher microstructural density than the surrounding sulcal walls. These data reveal that postnatal sulcal deepening is nonuniform and depends on the time of emergence in utero, tissue microstructure, and multiple macroanatomical factors. Together, these findings have important ramifications for theories of cortical folding and elucidating neurodevelopmental disorders and delays.

neuroscience↗

bioRxiv2025

Enhance Memory CD8+ T Cell Formation via Downregulation of Hematopoietic Progenitor Kinase 1 and Sustained Mitochondrial Fitness

Memory CD8+ T cells are pivotal for long-term anti-tumor immunity due to their longevity and rapid response upon encountering tumor cells. Hematopoietic progenitor kinase 1 (HPK1), a member of the Ste20 family of kinases, restricts proximal T cell receptor (TCR) signaling and plays critical roles in T cell priming and activation. However, the impact of HPK1 on CD8+ T cell differentiation is not well understood. Here, we demonstrate that chimeric antigen receptor (CAR) T cells derived from patients lacking HPK1 exhibit a stronger memory phenotype. HPK1 deletion promotes the formation of precursor and central memory CD8+ T cells, resulting in superior and long-lasting antitumor activity in murine cancer models compared to WT mice. Additionally, HPK1 knockout induces metabolic reprogramming by enhancing oxidative phosphorylation (OXPHOS) and mitochondrial respiration complex activity. Mechanistically, HPK1 deletion downregulates mTOR signaling by reducing the phosphorylation of mTOR and S6. Increased mitophagy activity is observed in HPK1-depleted cells, with upregulation of key mitophagy regulators, including Pink1 and Bnip3l, maintaining mitochondrial fitness. Our study reveals that HPK1 regulates CD8+ T cell metabolic reprogramming, guiding cell differentiation and antitumor responses.

immunology↗

bioRxiv2025

Autism-associated Scn2a haploinsufficiency disrupts in vivo dendritic signaling and impairs flexible decision-making

SCN2A is a high-confidence risk gene for autism spectrum disorder. Loss-of-function mutations in Scn2a reduce dendritic excitability in neocortical pyramidal cells. However, the impact of Scn2a haploinsufficiency on dendritic signaling in vivo, particularly during behavior, is unknown. In this study, we used two-photon microscopy to image dendritic calcium transients in deep layer pyramidal cells in the mouse medial frontal cortex. Scn2a+/- mice had diminished coupling between apical and proximal dendritic compartments. Pyramidal tract neurons had abnormal event rates, while intratelencephalic neurons had compartment-specific alterations indicative of diminished dendritic integration. In a matching pennies task, Scn2a+/- mice were inflexible in the face of changing competitive pressure. Apical dendritic tuft in IT neurons typically encoded reward and strategy, but these task-specific representations were altered in Scn2a+/- mice. Collectively, the findings demonstrate that Scn2a haploinsufficiency weakens dendritic integration in vivo and disrupts the dendritic encoding of decision variables, potentially contributing to the cognitive rigidity in autism spectrum disorder.

neuroscience↗

bioRxiv2025

Integrated single-cell and spatial multiomic analysis reveals widespread reactivation of developmental programs in diseased human hearts

The first organ to develop in utero, the human heart undergoes significant changes during development and must sustain its function over a lifetime. To better characterize molecular changes in human cardiac cell-types across sex, aging, developmental and disease, we analyzed single nucleus RNA sequencing (snRNA-seq) datasets from 299 donors, identifying many more differentially expressed genes (DEGs) across developmental and disease states than by sex and age. In cardiomyocytes and most non-cardiomyocyte cell types, developmental and disease DEGs showed significant overlap. Cardiac development and disease were associated with convergent changes in non-cardiomyocyte intercellular communication, including TGF{beta} signaling, but differences in cell-type proportions. By integrating snRNA-seq with 106 snATAC-seq datasets, we reveal potential transcriptional factors driving fetal reactivation in disease. Finally, using spatial transcriptomics data, we identify that fetal reactivation is highly localized in niches. This work offers the largest multimodal, cell-type resolved interrogation of the human heart, providing insights into convergence in development and disease.

genomics↗