Biology subjects

Wu, F.

Publications and source records attributed to Wu, F..

At least 55 records · Page 3Linked to original sources

bioRxiv2025

Structure and function of the nairovirus cap-snatching endonuclease

Nairoviruses include several human pathogens such as Crimean-Congo hemorrhagic fever virus (CCHFV) and Kasokero virus (KASV). The cap-snatching endonuclease (EN) domain of the viral polymerase is essential for transcription and represents a promising antiviral target. However, the structural and functional mechanisms of nairovirus ENs remain poorly understood. Here, we describe biochemical and structural studies of the ENs from CCHFV and KASV. Biochemical assays demonstrate that the RNA endonuclease activity of both ENs is activated by manganese ions and exhibits a preference for uridine-rich RNA substrates. This activity is inhibited by three metal-chelating inhibitors (DPBA, L-742,001, and BXA), with BXA displaying the highest binding affinity and inhibitory potency. We further determine nine crystal structures of CCHFV and KASV ENs in apo, metal ion-bound, and inhibitor-bound states. Comparative structural analysis uncovers a two-metal-ion binding mode unique to nairovirus ENs, in which conserved residues coordinate two manganese ions via bridging water molecules. In the inhibitor-bound structures of KASV EN, BXA forms additional stabilizing interactions with the enzyme, explaining its superior inhibitory effect. Functional assays further confirm that the two-metal-ion mechanism is critical for viral transcription. These findings provide a structural foundation for the rational design of antivirals against CCHFV and related pathogens.

microbiology↗

bioRxiv2025

Targeted Transcriptional Repression by Induced Proximity

Most cancer-driving proteins remain "undruggable" due to the absence of ligandable pockets and their reliance on intrinsically disordered or protein-DNA/protein-protein interactions. Transcription factors, which orchestrate oncogenic gene expression programs, are particularly challenging: they turn over rapidly, evade durable pharmacological inhibition, and resist even emerging targeted protein degradation strategies. Here, we describe a new induced-proximity therapeutic modality, Transcriptional Regulation via Active Control of Epigenetic Reprogramming (TRACER), that enforces locus-specific transcriptional silencing by recruiting endogenous corepressor complexes to transcription factor binding sites. We developed small-molecule TRACERs that tether methyl-CpG binding domain protein 2 (MBD2) and the Nucleosome Remodeling and Deacetylase (NuRD) complex to transcription factor-directed ligands. An estrogen receptor (ER) TRACER potently suppressed ER transcriptional activity in breast cancer cells, downregulated canonical ER target genes, and required MBD2 and histone deacetylase (HDAC1/2) for activity, confirming on-target epigenetic repression. Extending this approach to prostate cancer, an androgen receptor (AR) TRACER transcriptionally repressed both full-length AR and the drug-resistant truncation variant AR-V7, achieving >90% inhibition of AR transcriptional activity in androgen-independent prostate cancer cells with locus-specific gene repression. We further show that TRACERs can be modularly reprogrammed to recruit alternative repressors, including PRC2. Collectively, these findings establish TRACERs as a generalizable modality to pharmacologically silence undruggable transcription factors through targeted epigenetic reprogramming, offering a powerful new strategy for treating cancers refractory to existing therapies.

biochemistry↗

bioRxiv2025

A deep reinforcement learning platform for antibiotic discovery

Antimicrobial resistance (AMR) is projected to cause up to 10 million deaths annually by 2050, underscoring the urgent need for new antibiotics. Here we present ApexAmphion, a deep-learning framework for de novo design of antibiotics that couples a 6.4-billion-parameter protein language model with reinforcement learning. The model is first fine-tuned on curated peptide data to capture antimicrobial sequence regularities, then optimised with proximal policy optimization against a composite reward that combines predictions from a learned minimum inhibitory concentration (MIC) classifier with differentiable physicochemical objectives. In vitro evaluation of 100 designed peptides showed low MIC values (nanomolar range in some cases) for all candidates (100% hit rate). Moreover, 99 our of 100 compounds exhibited broad-spectrum antimicrobial activity against at least two clinically relevant bacteria. The lead molecules killed bacteria primarily by potently targeting the cytoplasmic membrane. By unifying generation, scoring and multi-objective optimization with deep reinforcement learning in a single pipeline, our approach rapidly produces diverse, potent candidates, offering a scalable route to peptide antibiotics and a platform for iterative steering toward potency and developability within hours.

molecular biology↗

bioRxiv2025

3D multi-omic mapping of whole nondiseased human fallopian tubes at cellular resolution reveals a large incidence of ovarian cancer precursors

Uncovering the spatial and molecular landscape of precancerous lesions is essential for developing meaningful cancer prevention and early detection strategies. High-Grade Serous Carcinoma (HGSC), the most lethal gynecologic malignancy, often originates from Serous Tubal Intraepithelial Carcinomas (STICs) in the fallopian tubes, yet their minute size and our historical reliance on standard 2D histology contribute to their underreporting. Here, we present a spatially resolved, multi-omics framework that integrates whole-organ 3D imaging at cellular resolution with targeted proteomic, metabolomic, and transcriptomic profiling to detect and characterize microscopic tubal lesions. Using this platform, we identified a total of 99 STICs and their presumed precursors that harbor TP53 mutations in morphologically unremarkable tubal epithelium in all five specimens obtained from cancer-free organ donors with average-risk of developing ovarian cancer. Although these lesions comprised only 0.2% of the epithelial compartment, they displayed geographic diversity, immune exclusion, metabolic rewiring, and DNA copy number changes among lesions and normal fallopian tube epithelium discovered alterations in STIC-associated genes and the pathways they control. In sum, this platform provides a comprehensive 3D atlas of early neoplastic transformation, yielding mechanistic insights into tumor initiation and informing clinical screening strategies for detecting cancer precursors in whole organs at cellular resolution.

bioinformatics↗

bioRxiv2025

A Generative Foundation Model for Antibody Design

Antibodies are indispensable components of the immune system, yet the design of high-affinity antibodies remains a time-consuming and experimentally intensive process. To address this challenge, we present IgGM, a novel generative foundation model designed to accelerate high-affinity antibody engineering. IgGM learns the complex relationships underlying the binding interactions between antigens and antibodies, as well as the mapping between antibody sequences and structures. By conditioning on different inputs, IgGM supports a wide range of antibody design tasks, including complex structure prediction, inverse design, affinity maturation, framework optimization, humanization, and de novo antibody design. It is compatible with both conventional antibodies and nanobodies, and allows user-defined CDR loop lengths for flexible design. To prioritize candidates, we introduce a frequency-based computational screening strategy that enhances design efficiency. Extensive evaluation through both in silico benchmarks and in vitro experiments across diverse antigens such as PD-L1, Protein A, TNF-, IL-33, SARS-CoV-2 RBD and its variants demonstrates that IgGM consistently generates antibodies or nanobodies with high measured affinity. These results underscore IgGMs versatility and effectiveness as a powerful tool for next-generation antibody discovery and optimization.

bioinformatics↗

bioRxiv2025

High-density extracellular recordings from the interior of intact brain organoids enable automated high-throughput functional assay

Bridging the gap between preclinical screening and clinical outcomes remains a major challenge in drug development for neurological disorders. Brain organoids, derived from human induced pluripotent stem cells, offer a scalable and physiologically relevant platform to model human neural circuits. We develop a fully automated system to record neural activity from the interior of intact human cortical organoids using a high-density microfabricated probe. The robotic system completes insertion within minutes while preserving organoid integrity and enables immediate recording of spontaneous spikes. We extract physiologically grounded and deterministic spike features, and train a long short-term memory classifier to distinguish between organoids derived from healthy individuals and those harboring familial Alzheimers disease (AD) mutations in the amyloid precursor protein. Despite intra-class variability, the classifier differentiates between organoid classes with 100% accuracy. The model classifies AD organoids treated with a drug candidate that reduces amyloid-{beta} levels as retaining an AD-like electrophysiological phenotype, demonstrating that functional readout can contradict molecular markers. The findings establish a high-throughput functional framework that may complement and extend existing drug screening assays.

neuroscience↗

bioRxiv2025

Pleistocene demographic histories dominate contemporary genomic diversity in a continental radiation of Himalayan-Hengduan songbirds

While population genetic theory expects genetic diversity to scale predictably with population size, mounting empirical evidence has challenged this fundamental prediction. This striking inconsistency has prompted a paradigm shift in our understanding of the determinants of genetic diversity, driving increased efforts to disentangle the relative contributions of Pleistocene demography and linked selection. Here, based on systematic sampling and a unified analytical pipeline, we de novo assembled genomes for 120 songbird species breeding in the Himalayas-Hengduan Mountains (HHMs) and conducted population genomic analysis to examine the drivers of their genomic diversity. We observed a 6.5-fold variation in genome-wide heterogeneity and a 16.4-fold variation nucleotide diversity and across species. Notably, these measures of genomic diversity showed no correlation with recent population dynamics, current population size, or other contemporary factors--such as natural selection, elevational distribution, or life-history traits. Instead, historical demography strongly predicted genetic diversity, with ancestral population size during the late Pleistocene emerging as the sole correlate: larger ancestral sizes consistently coincided with higher diversity. These findings underscore the critical influence of historical demography on contemporary genetic diversity in natural populations--an insight essential for designing effective conservation strategies.

genetics↗

bioRxiv2025

A trade-off mechanism underpins the evolution of a young two-gene sex-determining system in plants

Sexual systems in animals and plants are remarkably diverse, with dioecy having evolved independently in numerous lineages. In plants, dioecy often evolved more recently than in the best-studied animal systems, making plants especially important for understanding how separate sexes evolved independently from functionally hermaphrodite ancestors. Despite long-standing theories of developmental trade-offs in sex allocation, the underlying genetic mechanisms remain elusive. Here, we show that the XY sex determination system in the dioecious plant species Eurycorymbus cavaleriei in Sapindaceae involves two Y-linked mutations that act jointly within the developmental male-female trade-off: YUN{Delta}, a truncated allele that lowers the dosage of the D-class MADS-box gene YUN, and SUNMAO, a novel sRNA locus that silences the X-linked SUN allele. In females, SUN stabilizes the HD-ZIP transcription factor KUN, which is a known sex determinant in another dioecious plant, thereby promoting femaleness by increasing YUN expression; loss of SUN expression, together with the effect of YUN{Delta}, shifts development toward males. Two interlocking regulatory loops in this "SKY" module (SUN-KUN-YUN) fine-tunes YUN dosage. This dioecious system in E. cavaleriei likely evolved by sequential mutations in genes acting in the predicted male-female trade-off system, with their close linkage reflecting a translocation, and later recombination-suppressing inversions.

plant biology↗

bioRxiv2025

Netrin-1 promotes pancreatic tumorigenesis and innervation through NEO1.

Nerves have been shown to regulate cancer progression. However, a clear demonstration of a role for axon guidance molecules in pancreatic tumorigenesis, innervation, and metastasis has been lacking. Using murine KrasG12D-mutant pancreatic organoids, we screened axon guidance molecules by qRT-PCR, identified Ntn1 upregulation, and then verified its in vivo upregulation during pancreatic tumorigenesis in humans and mice. NTN1 and its receptor NEO1 were upregulated in epithelial cells by the Kras mutation and {beta}-adrenergic signaling, in part, through the MAPK pathway. Ex-vivo culture of celiac ganglia showed that NTN1 promoted the axonogenesis of sympathetic neurons through the nerve NEO1 receptor. In the Pdx1-Cre;LSL-KrasG12D/+ model, Ntn1 knockout decreased sympathetic innervation and the development of pancreatic intraepithelial neoplasia. Treatment of pancreatic tumor organoids with recombinant NTN1 enhanced cell growth, epithelial-mesenchymal transition (EMT), and cancer stemness with the upregulation of ZEB1 and SOX9 through NEO1-mediated activation of focal adhesion kinase (FAK). In Pdx1-Cre;LSL-KrasG12D/+;LSL-Trp53R172H/+mice, Ntn1 knockout reduced innervation, FAK phosphorylation, and the features of EMT and stemness to extend mouse survival. In a liver metastasis model of PDAC (pancreatic ductal adenocarcinoma), treatment with a NTN1-neutralizing antibody or tumoral knockdown of Neo1 reduced ZEB1 and SOX9 and decreased tumor progression. In contrast, Ntn1 overexpression promoted innervation and the progression of PDAC liver metastasis. These data suggest that the NTN1/NEO1 axis is a key regulator of PDAC progression, directly influencing cancer cell stemness and EMT, while indirectly promoting tumor growth through nerves. Inhibiting the NTN1/NEO1 axis could represent a potential therapeutic approach for PDAC. Statement of SignificanceNTN1 promotes pancreatic tumorigenesis and metastasis directly and indirectly through nerves, highlighting the importance of tumor cell-nerve crosstalk in cancer. NTN1 blockade could represent a promising strategy for treating PDAC liver metastasis. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=88 SRC="FIGDIR/small/666009v1_ufig1.gif" ALT="Figure 1"> View larger version (19K): org.highwire.dtl.DTLVardef@ca611corg.highwire.dtl.DTLVardef@ad9264org.highwire.dtl.DTLVardef@1661c00org.highwire.dtl.DTLVardef@b86798_HPS_FORMAT_FIGEXP M_FIG C_FIG

cancer biology↗

bioRxiv2025

Deep sea anaerobic microbial community couples the degradation of insoluble chitin to extracellular electron transfer

Chitin is a major structural component of arthropod exoskeletons, and an important carbon and nitrogen source in marine environments. In anoxic sediments, its degradation generates chitooligosaccharides and N-acetylglucosamine (GlcNAc), which are fermented into smaller organic molecules and oxidized anaerobically using soluble electron acceptors or insoluble ones such as metal oxides. To date, many aspects of chitin degradation in deep-sea anoxic sediments have been overlooked, including the potential coupling of insoluble chitin degradation to metal oxide reduction, the involvement of extracellular electron transfer (EET), and the spatial organization of the microorganisms involved. Using anoxic deep-sea sediments recovered from a whale fall site, we developed an innovative workflow based on electrochemical reactors, to characterize chitin degradation in these environments. Sediment samples enriched on poorly crystalline iron oxides, and subsequently transferred into an electrochemical reactor poised at +0.22 V vs SHE, showed active anodic current production when supplied with chitin, which increased 2-fold when amended with GlcNAc. Chitin reactors were dominated by Vallitalea (Firmicutes), Spirochaetota, Gammaproteobacteria and Desulfobacterota. Exoenzyme activity assays, metabolite profiling, and continued anodic current production confirmed ongoing chitin degradation linked to EET. We observed metabolic associations between chitin degraders and secondary consumers using in situ imaging (16S rRNA gene FISH coupled with BONCAT and nanoSIMS). These microbial partners, within the electrode-attached community, required close proximity to the poised electrode ([≤] 10 {micro}m) to remain metabolically active. Supporting these observations, cultured isolates of Vallitalea sp. and Trichloromonas sp. recovered from the whale fall site exhibited chitin degradation and electrochemical activity, respectively. When co-cultured in an bioelectrochemical reactor, the acetate produced by Vallitalea sp. during chitin degradation fueled Trichloromonas sp., which facilitated EET, hereby demonstrating that syntrophic interactions are used to couple anoxic chitin degradation to EET in deep-sea sediments. These findings exemplify the interspecies interactions and resource optimization occurring in hard-to-reach and largely unknown deep-sea ecosystems.

microbiology↗

bioRxiv2025

HILIC-Enabled Mass Spectrometric Discovery of Novel Endogenous and Glycosylated Neuropeptides in the American Lobster Nervous System

Neuropeptides are a highly conserved and diverse class of intercellular signaling molecules that regulate a broad range of neural and hormonal processes across animal phyla. The American lobster, Homarus americanus, has long served as a powerful invertebrate model for the discovery and functional investigation of neuropeptides. Among common post-translational modifications (PTMs) found in neuropeptides, glycosylation remains underexplored due to the inherently low in vivo abundance and intrinsically complex structural heterogeneity. In this study, we employed hydrophilic interaction liquid chromatography (HILIC) enrichment coupled with oxonium-ion triggered EThcD fragmentation strategy to simultaneously profile novel endogenous and glycosylated neuropeptides across eight distinct neural tissues and neuroendocrine organs of Homarus americanus. This integrative mass spectrometry-based approach led to the identification of 154 endogenous neuropeptides derived from 25 families, approximately one-third of which are newly reported, and uncovered 24 glycosylated neuropeptides in this species for the first time. These peptides exhibit strong tissue-specific expression, distinct proteolytic cleavage patterns, and confidently localized glycosylation sites. Our results highlight the utility of integrated sampling enrichment and hybrid fragmentation strategies for deep neuropeptidomic profiling and provide a valuable resource for future studies on the functional roles of newly identified neuropeptides and glycosylation in crustacean neuromodulation and peptidergic signaling. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=108 SRC="FIGDIR/small/661634v2_ufig1.gif" ALT="Figure 1"> View larger version (20K): org.highwire.dtl.DTLVardef@9a1745org.highwire.dtl.DTLVardef@6decc4org.highwire.dtl.DTLVardef@d34256org.highwire.dtl.DTLVardef@1a83b4c_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2025

NKX2-5 congenital heart disease mutations show diverse loss and gain of epigenomic, biochemical and chromatin search functions underpinning pathogenicity

Congenital heart defects (CHD) occur in [~]1% of live births, with inherited and acquired mutations and environmental factors contributing to causation. However, network perturbations in CHD remain poorly understood. We report an integrated functional-epigenomics approach to CHD, focusing on the cardiac homeodomain (HD) transcription factor NKX2-5, mutations which cause diverse heart structural and conduction defects. We selected twelve NKX2-5 CHD variants affecting different residue classes - homeodomain DNA base-contacting, backbone-contacting and helix-stabilizing, and those affecting other conserved protein:protein interaction (PPI) domains. In HL-1 cardiomyocytes, we profiled genome-wide DNA targets of NKX2-5 wild type (WT) and variants, their DNA binding affinity and specificity, PPI with known NKX2-5 cofactors and chromatin search dynamics. Variants showed diverse yet class-specific behaviours. All variants failed to bind many WT targets but retained binding to a subset of core cardiomyocyte-related targets, and bound hundreds of unique "off-targets" via changes to DNA binding site specificity, homodimerization, cofactor interactions and chromatin search functions. Our data suggest that complex residue-by-residue scale epigenomic, biochemical and chromatin search defects involving both loss-and gain-of-function contribute to CHD. These findings may inform precision molecular therapeutic approaches in patients with CHD.

genomics↗

bioRxiv2025

Multiplex Proteomics of Lewy Body Dementia Reveals Cerebrospinal Fluid Biomarkers of Disease Pathology and Progression

Lewy body dementia (LBD), which encompasses Parkinsons disease dementia (PDD) and Dementia with Lewy bodies (DLB), lacks established biofluid markers of its complex clinical and neuropathological heterogeneity. Multiplex proteomic tools, such as the recently developed NUcleic acid Linked ImmunoSandwich Assay (NULISATM), can assess a diverse array of neurodegenerative targets and address these biomarker gaps. We used this platform to analyze 852 baseline and longitudinal cerebrospinal fluid (CSF) samples from control, PD, and DLB subjects from the Parkinsons Disease Biomarker Program (PDBP). We then examined LBD protein signatures across different clinical diagnoses, patterns of cognitive progression, and underlying neuropathologies. Of the 124 markers analyzed, 54 were significantly altered in PD with cognitive impairment and/or DLB compared to controls, many serving as informative clinical classifiers in machine learning models. We also identified 15 proteins with significant baseline and/or longitudinal changes linked to cognitive progression. To identify neuropathological signatures, we stratified 209 cases by the presence of underlying synuclein and amyloid-beta (A{beta}) pathology using CSF alpha-synuclein seed amplification assay (Syn-SAA) results and NULISA pTau-217 levels, respectively. Select synaptic markers were decreased relative to synucleinopathy but increased relative to A{beta} accumulation, complicating their interpretation in cases with co-pathology. These neuropathological analyses were replicated in an independent cohort of 917 cases from the Alzheimers Disease Neuroimaging Initiative (ADNI). Overall, these results highlight the utility of multiplex proteomic analysis in the identification of candidate CSF biomarkers of disease heterogeneity in LBD. One Sentence SummaryThis multiplex proteomic analysis using the recently released NULISA platform revealed unique targets of Lewy body dementia progression and neuropathology.

systems biology↗

bioRxiv2025

Habitat use of juvenile green turtles along the North Pacific coast of Costa Rica

Understanding how threatened marine species use coastal areas and the extent of connectivity across different spatial and temporal scales is important for identifying critical habitats that can enhance conservation efforts in other regions of their distribution. In this study, we investigated the site fidelity and habitat use of juvenile yellow morphotype green sea turtles (Chelonia mydas) in the Gulf of Santa Elena, north Pacific coast of Costa Rica. Fifteen juvenile green sea turtles (49 - 83 cm curved carapace length; CCL) were monitored for 19 - 629 days with an array of 11 acoustic receivers placed within 5 main habitat types: muddy areas, reef patches, macroalgae, rocky reefs, and mangroves. Daily and seasonal locations of the turtles were estimated using a positioning estimation algorithm to determine the site fidelity and habitat use within the study area. Large juveniles ([≥] 65 CCL) were detected disproportionally in the macroalgae habitat during the upwelling season (43.9% detections) from December - April, and in the reef patch habitat during the non-upwelling season (36.0%) from May to November. Small juveniles (< 65 cm CCL) had more detections in the reef patch habitat during both seasons (Upwelling: 35.4%, non-upwelling: 45.4%) relative to other habitats sampled. More than half of the turtles monitored (67%) showed strong site fidelity to the study area, ranging from 68.9 to 100% from October 2020 to September 2022. Our results suggest that juvenile green sea turtles, particularly larger individuals, have a strong association with coral / rocky reef patches and macroalgae habitats, similar to adults. Our finding suggested that protecting similar habitats in other areas along the Central American coast will help rebuild the Eastern Pacific Ocean green sea turtle population.

ecology↗

bioRxiv2025

Electroacupuncture promotes post-stroke motor recovery through miR-124-3p-mediated regulation of NRG1/ErbB4 signaling pathway

ObjectiveElectroacupuncture has demonstrated beneficial effects in post-stroke motor dysfunction, yet the molecular mechanisms underlying its therapeutic efficacy remain incompletely understood. This study aims to investigate whether electroacupuncture promotes post-stroke motor function recovery by modulating the interaction between miR-124-3p and the NRG1/ErbB4 signaling pathway, specifically exploring whether miR-124-3p directly targets NRG1 to regulate neural plasticity in a focal cerebral ischemia rat model. MethodsThe ischemic stroke model was established by middle cerebral artery occlusion/reperfusion (MCAO/R) in adult rats. These rats were randomly divided into sham, model, electroacupuncture and model plus miR-124-3p inhibitor groups. The model group, electroacupuncture group and model plus miR-124-3p inhibitor group received EA intervention 24 h after modelling for 7 consecutive days. Behavioural function was assessed by Zea Longa score and mechanical pain rating. Hippocampal damage was detected by HE staining and neuronal apoptosis was observed by TUNEL staining. IL-1{beta} and IL-18 levels were measured by ELISA. PCR and Western blotting were used to detect the expression of miR-124-3p and inflammatory pathway proteins. The interaction between miR-124-3p and TLR4 was verified by dual-luciferase reporter assay. ResultsElectroacupuncture improved motor function in rat model of MCAO/R, as evidenced by improved Zea Longa scores and decreased mechanical withdrawal thresholds. Electroacupuncture significantly attenuated neuronal damage and also inhibited the inflammatory response by decreasing IL-18 and IL-1{beta} levels (P < 0.001). Notably, electroacupuncture upregulated miR-124-3p expression (P < 0.0001) and activated the NRG1/ErbB4 signaling pathway in the hippocampus. When miR-124-3p was inhibited, NRG1 protein expression decreased while GABA expression tended to increase. Dual-luciferase reporter assays confirmed that miR-124-3p directly targets the 3UTR of NRG1 mRNA and regulates its expression at the translational level. These findings suggest that electroacupuncture may alleviate neuronal and axonal damage by modulating miR-124-3p/NRG1/ErbB4 signalling and regulating GABA release. ConclusionElectroacupuncture can ameliorate motor dysfunction induced after brain I/R injury by targeting and modulating the NRG1-ErbB4 signaling pathway via miR-124-3p. These data are expected to provide new insights into the mechanisms of electroacupuncture for the prevention of potential targets for the recovery of motor dysfunction after stroke.

animal behavior and cognition↗

bioRxiv2025

A synergistic generative-ranking framework for tailored design of therapeutic single-domain antibodies

Single-domain antibodies (sdAbs) have emerged as powerful therapeutic agents due to their small size, high stability, and superior tissue penetration. However, unlike conventional monoclonal antibodies (mAbs), sdAbs lack an Fc domain, limiting their functional versatility and manufacturability. To address this challenge, we developed TFDesign-sdAb, a deep learning-based generative-ranking framework that enables rational engineering of sdAbs with tailored functionalities. Our framework integrates a structure-aware diffusion model (IgGM) for large-scale candidate generation and a fine-tuned sorter (A2binder) that evaluates and prioritizes them based on predicted functionality. Unlike traditional CDR-focused approaches, TFDesign-sdAb optimizes both complementarity-determining regions (CDRs) and framework regions (FRs), allowing sdAbs to acquire new functional properties while maintaining antigen specificity. We validated our approach by conferring Protein A binding to human VHs and nanobodies that originally lacked this feature, achieving high expression rates, strong binding affinities, and successful purification via industry-standard Protein A affinity chromatography. High-resolution structural characterization (1.49 [A] and 2.0 [A]) of the redesigned sdAb-Protein A complexes revealed conserved FR-mediated binding motifs that recapitulate natural Fc-Protein A interactions, validating the accuracy of our model. Furthermore, our pipeline streamlined the antibody engineering process, achieving a 100% success rate in generating Protein A-binding sdAbs while maintaining their original antigen-binding affinity. This work demonstrates the power of AI-driven design in overcoming long-standing limitations in antibody engineering and presents a scalable, generalizable solution for enhancing sdAb functionality.

bioinformatics↗

bioRxiv2025

Proviruses in CD4+ T cells reactive to autologous antigens contribute to nonsuppressible HIV-1 viremia

Antiretroviral therapy (ART) halts HIV-1 replication, reducing plasma virus levels to below the limit of detection, but it is not curative due to a reservoir of latently infected CD4+ T cells. In some people living with HIV-1 (PLWH), plasma HIV-1 RNA becomes persistently detectable despite optimal ART. This nonsuppressible viremia (NSV) is characterized by identical, non-evolving HIV-1 RNA variants expressed from infected CD4+ T cell clones. The mechanisms driving persistent virus production from a specific population of infected cells are poorly understood. We hypothesized that proviruses in cells responding to chronic immunologic stimuli, including self-associated antigens, may drive viral gene expression and NSV. Here, we demonstrate that stimulation of CD4+ T cells with autologous cell lysates induces virus production in an MHC-II-dependent manner. In 7 of 8 participants with NSV, we recovered viral RNA released ex vivo in response to autologous cell lysates that matched plasma virus. This process involves both defective and replication-competent proviruses residing in conventional T cells, and is also observed in PLWH with undetectable viremia. These findings suggest that recognition of self-associated antigens is an important cause of HIV-1 reservoir expression, which can contribute to persistent systemic inflammation and potential rebound upon ART interruption. One sentence summaryHIV-1 viremia not suppressed by effective ART can be caused by proviruses in CD4+ T cells reactive to autologous antigens.

immunology↗

bioRxiv2025

Tff2 marks gastric corpus progenitors that give rise to pyloric metaplasia/SPEM following injury

In BriefTu et al. show that Tff2+ corpus isthmus cells are TA progenitors, and they, not chief cells, are the primary source of SPEM following injury. Upon Kras mutation, these progenitors directly progress to dysplasia, bypassing metaplasia, highlighting them as a potential origin of gastric cancer. HighlightsO_LITff2+ corpus cells are TA progenitors that give rise to secretory cells. C_LIO_LITff2+ progenitors, not chief cells, are the primary source of SPEM after injury. C_LIO_LIKras-mutant Tff2+ progenitors progress directly to dysplasia, bypassing metaplasia. C_LIO_LIMulti-omics analysis reveals distinct trajectories for SPEM and gastric cancer. C_LI O_FIG O_LINKSMALLFIG WIDTH=191 HEIGHT=200 SRC="FIGDIR/small/647847v1_ufig1.gif" ALT="Figure 1"> View larger version (46K): org.highwire.dtl.DTLVardef@48a2d1org.highwire.dtl.DTLVardef@b96642org.highwire.dtl.DTLVardef@1e89a4dorg.highwire.dtl.DTLVardef@1ea6a86_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical abstractC_FLOATNO C_FIG Pyloric metaplasia, also known as spasmolytic polypeptide-expressing metaplasia (SPEM), arises in the corpus in response to oxyntic atrophy, but its origin and role in gastric cancer remain poorly understood. Using Tff2-CreERT knockin mice, we identified highly proliferative Tff2+ progenitors in the corpus isthmus that give rise to multiple secretory lineages, including chief cells. While lacking long-term self-renewal ability, Tff2+ corpus progenitors rapidly expand to form short-term SPEM following acute injury or loss of chief cells. Genetic ablation of Tff2+ progenitors abrogated SPEM formation, while genetic ablation of GIF+ chief cells enhanced SPEM formation from Tff2+ progenitors. In response to H. pylori infection, Tff2+ progenitors progressed first to metaplasia and then later to dysplasia. Interestingly, induction of KrasG12D mutations in Tff2+ progenitors facilitated direct progression to dysplasia in part through the acquisition of stem cell-like properties. In contrast, Kras-mutated SPEM and chief cells were not able to progress to dysplasia. Tff2 mRNA was downregulated in isthmus cells during progression to dysplasia. Single-cell RNA sequencing and spatial transcriptomics of human tissues revealed distinct differentiation trajectories for SPEM and gastric cancer. These findings challenge the conventional interpretation of the stepwise progression through metaplasia and instead identify Tff2+ progenitor cells as potential cells of origin for SPEM and possibly for gastric cancer.

cell biology↗