bioRxiv Science⌕ Search

Biology subjects

Hong, X.

Publications and source records attributed to Hong, X..

At least 19 recordsLinked to original sources
bioRxiv2026

Genetic accessibility, not superior fitness, determines which anti-HBs antibody escape variants are selected in patients

Background and Aims: Genetic diversity within the hepatitis B virus (HBV) surface antigen (HBsAg) drives the emergence of immune escape variants, but studying this diversity in vivo has been limited by the difficulty of establishing infection from patient isolates. We aimed to perform unbiased in vivo screening of variants in the HBs antigenic loop for antibody escape and fitness, using synthetic DNA launch of HBV in human hepatocyte chimeric mice. Approach and Results: Direct intrahepatic injection (IHI) of barcoded, recombinant cccDNA established infection from up to 32,000 independent launch events of a genotype A strain and successfully launched clinical consensus sequences across the major genotypes. Using IHI, we performed deep mutational scanning of 4,608 codon variants across the HBs antigenic loop. Under control antibody treatment, diverse variants expanded, revealing differences in variant fitness in the absence of adaptive immune pressure. Anti-HBs monoclonal antibody treatment suppressed wild-type viremia while permitting breakthrough of resistant variants from the library. Sequencing identified strong enrichment for substitutions at G145 (G145Q, G145H, G145R). Notably, G145Q, rather than the clinically dominant G145R, showed the greatest fitness advantage, indicating that G145R's clinical prevalence does not reflect superior intrinsic fitness. Conclusions: Unbiased in vivo screening reveals that clinically observed HBV escape mutations reflect which substitutions are genetically accessible by single nucleotide change and compatible with the overlapping polymerase reading frame, rather than which confer optimal fitness. This approach establishes a platform for high-resolution mapping of viral variant fitness under antiviral selection pressure.

microbiology↗
bioRxiv2026

The HBV basal core promoter mutation confers a replicative advantage and transcriptionally reprograms hepatocytes toward HCC subtypes

Background: Hepatitis B virus (HBV) basal core promoter (BCP) and precore (PC) mutations occur during chronic HBV infection with the BCP-mutant being associated with increased hepatocellular carcinoma (HCC) risk. Objectives: The effects of these mutants on viral replication, hepatocyte biology, and carcinogenesis remain poorly defined. To address this, we characterized BCP- and PC-mutants using human hepatocyte chimeric mice and asked whether the resulting infection-induced transcriptomes correspond to subsets of human HBV-related hepatocellular carcinomas (HBV-HCCs). Design: Isogenic wild-type (WT), BCP- and PC-mutants of HBV genotypes A, C, and D (HBV-A, -C, -D) were generated from recombinant covalently-closed circular DNA to infect chimeric mice. HBV-HCC transcriptomic datasets from The Cancer Genome Atlas (TCGA) were used to stratify association of HBV variants with HCC subtypes. Results: WT, BCP-mutant, and PC-mutant HBV sequences remained genetically stable. The BCP mutation, but not the PC mutation, accelerated the rise in serum viremia in HBV-D- and HBV-A-infected chimeras; in HBV-C, acceleration required both (PC+BCP) mutations. Further comparisons of HBV-D variants revealed that the BCP-mutant increased intrahepatic viral DNA, viral protein expression, and upregulated cancer-related pathways, including transcripts associated with a subset of HBV-HCCs. Analysis of HBV-HCC samples from the TCGA revealed that tumors often harbor a mixture of WT and mutant transcripts, and that WT- and BCP-mutant-associated HCCs exhibit distinct transcriptional profiles. Conclusions: The HBV BCP-mutant directly perturbs hepatocyte homeostasis via virus-intrinsic mechanisms, selectively activating cancer-related pathways and defining a molecularly distinct subset of HBV-HCC. These findings suggest that HBV variants form distinct subcategories of HBV-HCCs.

molecular biology↗
bioRxiv2026

Dynamic architecture of the Rixosome reveals mechanism of activation and ITS2 processing

Eukaryotic ribosome assembly requires the coordinated processing and extensive remodeling of pre-rRNAs. During late nuclear maturation of the 60S subunit, sequential removal of the internal transcribed spacer 2 (ITS2) is initiated by endonucleolytic cleavage at site C2 by the conserved Las1 nuclease. Las1 acts together with the kinase Grc3 and the Rix1 complex to form the Rixosome, which also functions in transcriptional regulation. However, the assembly of the Rixosome, its recruitment to pre-ribosomes, and its activation for ITS2 cleavage remain unclear. Here, we present cryo-EM structures of the human LAS1 complex, two structures of the isolated Rixosome and nine transition states of Rix1-bound pre-60S particles from Schizosaccharomyces pombe. These structures reveal a dynamic Rixosome architecture in which the heterotetrameric Las1 complex engages one or two copies of the Rix1 complex. Rix1 binding is highly flexible in the human Rixosome but rigid in the yeast complex. The isolated yeast Rixosome remains inactive, but binding to the pre-60S particle triggers a structural rearrangement that allows for substrate engagement and activation of the nuclease. Together, our results define the dynamic architecture of the Rixosome and provide a structural framework for ITS2 processing during nuclear maturation of the eukaryotic 60S ribosomal subunit.

biochemistry↗
bioRxiv2026

The HES1-SOD1 Antagonism Shapes Senescence Heterogeneity and Impacts Metastatic Relapse of Circulating Tumor Cells

Circulating tumor cells (CTCs) encounter multiple challenges within the blood microenvironment, including oxidative stress, flow shear forces, and immune surveillance, often leading to anoikis. Recently, a transitional state of CTC senescence has been identified, contributing to metastatic inefficiency. However, the molecular mechanisms linking senescent CTCs to disease relapse remain to be defined. By integrating a genetic model of cortactin knockdown-induced CTC senescence with single-cell multi-omic analyses, we revealed two distinct senescent CTC subpopulations marked by HES1 expression levels. These HES1low and HES1high subpopulations exhibited differential evolutionary trajectory dynamics and unique molecular and metabolic signatures, which were significantly correlated with adverse clinical outcome across several patient cohorts. HES1low senescent CTCs displayed enhanced mitochondrial fitness, oxidative phosphorylation, and ROS-detoxifying capabilities, resulting in more efficient tumor regrowth with a pro-inflammatory and thrombotic phenotype when compared to the HES1high group. Mechanistically, HES1 directly bound to the Sod1 promoter and repressed its expression, leading to redox imbalance and mitochondrial dysfunction that were linked to weakened tumor regrowth capacity. Both senescent CTC subpopulations were broadly resistant to cytotoxic and targeted therapies, yet they showed elevated dependency on anti-apoptosis programs that make them susceptible to dual blockade by SOD1 inhibitor and the anti-senolytic drug ABT737 in vivo. Finally, in a prospective cohort of on-treatment melanoma patients, HES1 senescent CTCs were highly enriched in patients with progressive disease. Thus, the HES1-SOD1 antagonism shapes CTC senescence heterogeneity and contributes to differential tumor relapse, which can be therapeutically explored for eliminating residual metastatic disease.

cancer biology↗
bioRxiv2026

Orthogonal CRISPR screens and human liver chimeric mice identify hepatitis B virus host factors

Hepatitis B virus (HBV) chronically infects approximately 250 million people worldwide, and reliable curative therapies are lacking. A broader understanding of viral-host interactions could accelerate efforts to find new host-centric therapeutic targets. However, inefficient cell culture systems and limited replication markers compatible with pooled screening have precluded the widespread use of genetic perturbation screens. Here, we performed the first pooled, genome-wide CRISPR-Cas9 knockout (KO) screen with authentic HBV infection and integrated these results with two orthogonal pooled screens to identify host factors. We selected 72 genes for a multi-step assessment that included arrayed validation assays using both HBV infection and pgRNA transfection. We then independently tested thirteen genes using high-efficiency bulk KO experiments to guide further investigations of both antiviral and proviral factors. In both KO and siRNA-mediated knockdown experiments, depletion of the top antiviral factor, EXOC1, enhanced multiple HBV replication markers, and transcriptomic analysis revealed activation of hypoxia- and HIF-1 gene signatures. Three proviral factors, IRF2, WDR48, and ZCCHC14, were investigated in vivo using a human liver chimeric mouse model, which demonstrated that ZCCHC14 KO greatly reduced HBV replication and spread. Together, these complementary in vitro and in vivo platforms expand the catalog of HBV host factors and provide a scalable framework for host target discovery.

microbiology↗
bioRxiv2026

Autophagy suppresses microglial activation and enhances M2 polarization via the mTOR/ULK1 pathway after optic nerve crush

PurposeTo investigate whether rapamycin can regulate microglial activation and polarization via mTOR and its downstream signals via autophagy both in vivo and in vitro. MethodsThe in vivo study used wild type C57BL/6 mice that were intraperitoneally injected with rapamycin (2 mg/kg) plus ONC. The BV2 cell line was used in the in vitro study and the cells were incubated with rapamycin (50 nM) or transfected with a specific mTOR-targeting small interfering RNA (si-mTOR). Immunohistochemical staining was used to observe the changes in the morphology and cell surface area of microglia and Weste blotting analysis was used for detection of the changes in the proteins related autophagy, microglia polarization and mTOR pathway after the retinal tissue or the cell samples were collected. ResultsThese results indicate that rapamycin increases autophagy and M2 polarization by inhibiting p-mTOR in wild-type C57BL/6 mice in vivo. In the BV2 cell line, rapamycin and si-mTOR can enhance autophagy and promote M2 polarization by inhibiting the p-mTOR/p-Unc-51-like kinase 1 (p-ULK1) pathway. ConclusionsIn conclusion, this work contributes to the understanding of the complex interplay among rapamycin, autophagy and microglial activation/polarization, highlights the downstream signaling pathway of mTOR, and highlights the potential therapeutic effects of autophagy-modulating drugs in retinal neuroinflammation and neurodegeneration after TON.

neuroscience↗
bioRxiv2026

dCas allele sequestration (das-CRISPR): A Versatile New Method to Achieve Monoallelic Gene Editing in Mouse Embryos and in cell culture.

CRISPR-Cas9 technology is a powerful tool extensively used for genome editing in mouse and many other species. Streptococcus pyogenes Cas9 efficiently cuts both alleles in mouse zygotes leaving many edited embryos without a functional protein that might be needed to sustain development, to survive postnatally or to reproduce, thus complicating its overwhelmingly advantageous use in making gene modifications. About 25% of mouse genes are essential for embryonic development and another 7% are necessary for fertility, thus for these genes it is desirable to maintain a functional allele to establish viable lines from CRISPR-Cas9 edited mouse embryos. However, exclusive monoallelic editing is challenging to achieve with current CRISPR methods. Controlling the activity of Cas9 in genome editing is an ongoing research field focused on developing new methods to curtail its damage caused by excess of on-target and off-target editing. In this study we describe a novel and a simple method, we termed das-CRISPR, for dCas allele sequestration in combination with CRISPR system, that allows monoallelic editing of targeted allele in mouse and in cultured cell lines. This method incorporates the use of a nuclease deficient deadCas9 (dCas9) present at higher levels than an active Cas9, both complexed with the same single guide RNA (sgRNA) sequence. We showed the delivery of the two proteins as ribonucleoprotein complexes (RNP) into mouse zygotes leads to the generation of viable and fertile mice carrying lethal mutations in an essential gene. We found that greater amounts of dCas9 RNPs bind and protect a target site while the lower amount of functional Cas9 RNPs accessed the unoccupied target site resulting in higher frequency of monoallelic gene editing, compared to using just Cas9 alone. We also showed this method can mitigate and control the activity of Cas9 in mouse NIH3T3 cells in culture to achieve monoallelic editing. This method is a versatile approach to controlling excessive Cas9 activity on-target and off-target both in vitro and in vivo.

molecular biology↗
bioRxiv2026

GalaxyVS: Exploring 100-Billion Compounds in Seconds

We present GalaxyVS, a hardware-software co-designed virtual screening framework built to explore the 100-billion commercially accessible chemical space in seconds, deployed at the National Supercomputing Center in Tianjin. Built upon the dense vector retrieval paradigm of DrugCLIP, GalaxyVS bypasses the structural dependencies and computational overhead of classical docking to enable rapid screening against experimentally determined as well as geometrically feasible pockets on AlphaFold-predicted structures. To scale this paradigm to the 100-billion level, the system must overcome the significant computational burden of offline representation encoding, critical memory and I/O bottlenecks during online retrieval, and the risks of diversity collapse and precision loss within final screening results. Utilizing the heterogeneous supercomputing infrastructure, GalaxyVS accelerates the offline encoding through deep operator adaptations and resolves online retrieval bottlenecks via disk-native vector indexing coupled with in-memory staging to ensure both broad accessibility and high throughput. Concurrently, a two-stage refinement protocol effectively mitigates diversity collapse and ensures high-fidelity affinity ranking. Consequently, GalaxyVS achieves a daily scoring throughput of 1.5 x 1016 target-ligand pairs, representing a six-orders-of-magnitude leap over previous supercomputing records. Driven by this throughput, we screened nearly 100,000 protein structures across six species against the 100-billion compound library in just 16 hours. The resulting comprehensive cross-species interaction landscape, GalaxyDB, will be openly released at https://galaxyvs.drugclip.com.

bioinformatics↗
bioRxiv2026

Candida albicans reprograms host inosine metabolism to drive immunosuppressive macrophage polarization and gastric cancer carcinogenesis

Gastric cancer (GC) is strongly associated with changes in the gastric microbiome. However, the direct contribution of individual pathogenic species to tumor initiation and progression remains poorly defined. Here, we identified Candida albicans (C. albicans), a fungus that ranks as the most significantly enriched taxon in GC tissues, can directly promote gastric carcinogenesis. We found that when C. albicans alone was colonized into the gastric mucosa of mice, it induced the complete sequence of GC-associated preneoplastic lesions, including chronic gastritis, mucinous metaplasia, and atypical hyperplasia, with 16% of colonized animals progressing to low-grade intraepithelial dysplasia. In K19-C2mE transgenic mice, a model of spontaneous GC formation, infection with C. albicans significantly accelerated GC formation, with synergistic effects observed upon co-infection with the well-known gastric carcinogen, Helicobacter pylori. Mechanistically, C. albicans upregulated inosine levels in both gastric tissues and circulation. This inosine promoted polarization of macrophages toward an M2-like phenotype by activating the adenosine A2a receptor (A2aR), creats an immunosuppressive microenvironment that protects GC cells from the tumoricidal effects of anti-tumor immunity. Depletion of M2 macrophages or pharmacological inhibition of A2aR attenuated C. albicans-mediated tumorigenesis. Clinically, elevated inosine levels correlated with prior C. albicans exposure and poorer outcomes of GC patients. Our findings establish C. albicans as a direct pro-tumorigenic pathogen in GC and highlight the inosine-A2aR-M2 macrophage axis as a promising target for therapeutic intervention.

cancer biology↗
bioRxiv2026

Decoding Immune Reconstitution Failure in People Living with HIV through Single-cell Genomics

Immune non-responders (INRs), a subset of people living with HIV (PLWH), fail to achieve full immune reconstitution and remain at increased risk of morbidity, mortality, and non-AIDS-related illnesses. As the mechanisms underlying this impaired immune recovery remain poorly understood, we performed single-cell multi-omics profiling on peripheral blood mononuclear cells from 43 INRs, 47 immune responders and 53 healthy donors. Our dataset comprises 2,744,009 transcriptomes and 1,226,658 chromatin-accessibility profiles across 58 identified immune cell types. INRs exhibited markedly elevated inflammatory signaling, increased apoptotic activity, and dysregulation of immune-activating ligand-receptor interactions. Additionally, we identified 2,996 interaction cis-eGenes and 5,938 cis-caPeaks, and validated key cell type-specific cis-xQTL effects associated with immune reconstitution failure. Furthermore, we identified 3,551 sc-eQTLs and 872 cell-state interaction eQTLs (ieQTLs) in CD4+ T cells. A genotype-specific upregulation of OAS3 in INRs, likely regulated by STAT1, was identified based on these results. We developed scPRISM (single cell Predictive Reconstitution Immune Status Model), a multi-modal framework that integrates scRNA-seq and scATAC-seq data to perform the dual tasks of differentiating disease states and predicting gene expression from chromatin accessibility, thereby deciphering the cell type-specific cis-regulatory basis of the disease. Our study delineates the transcriptional and epigenomic landscape of immune dysregulation in INRs and provides new genetic insights into HIV-associated immune reconstitution failure. These findings offer a foundation for developing biomarkers and precision therapies to restore immune function in PLWH.

genetics↗
bioRxiv2026

Faster science, penalties in evaluation, and concerns on quality and impact: Researchers' use and perceptions of preprints

The preprint ecosystem has expanded rapidly over the past decade, fundamentally altering science communication. Yet, the scholarly community's attitudes toward this shift remain underexplored. Through a large-scale survey of US and Canadian biomedical scholars, we provide a comprehensive analysis of preprint utilization, perceived impact, and integration into academic credit systems. We find robust engagement across reading, citing, and submitting preprints; however, this activity is driven primarily by a desire for rapid dissemination rather than a foundational commitment to open science. Furthermore, while preprints are valued as networking assets, perceived career penalties during formal academic evaluations stifle broader cultural adoption. Crucially, to navigate the absence of formal peer review, scholars report a heavy reliance on author reputation as a primary heuristic for evaluating a preprint's credibility and guiding their reading and citation decisions. Notably, despite acknowledging preprints' role in accelerating knowledge sharing, scholars express significant concerns regarding fraud and misinformation, particularly amid declining public trust in science and emerging threats to scientific integrity from artificial intelligence. To resolve these tensions, the preprint ecosystem must evolve beyond prioritizing speed to foster genuine academic dialogue. Simultaneously, evaluation frameworks must adapt to the realities of preprinting, and innovative quality-control mechanisms are urgently needed to balance rapid dissemination with rigorous scientific integrity.

scientific communication and education↗
bioRxiv2026

Beyond Lipschitz: Ranking Binding Affinity in Hyperbolic Space

Despite the importance of protein-ligand affinity ranking in drug discovery, existing deep learning models struggle to distinguish hard inactives that are structurally similar to active compounds but biologically inactive. We theoretically show this failure stems from the Lipschitz continuity constraint in Euclidean space, which makes neural networks locally insensitive to subtle yet critical structural perturbations. To overcome this, we propose AlphaRank, which demonstrates that scoring affinity as the negative hyperbolic geodesic distance allows subtle tangential variations to bypass Lipschitz constraints, intuitively corresponding to pivotal binding mode factors such as conformational fit. Meanwhile, AlphaRank employs joint optimization to simultaneously ranking affinity scores among actives in a proximity-aware manner and separate actives from decoy inactives through a geometric cone constraint. Experiments demonstrate that AlphaRank outperforms state-of-the-art models like Boltz-2 in both affinity ranking and active-inactive discrimination, while providing a interpretable representation space.

bioinformatics↗
bioRxiv2026

Glucose is a ligand-like trigger for secretion of postprandial GLP-1

Glucagon-like peptide 1 (GLP-1) is secreted from the intestinal L cells in a glucose-dependent manner. However, the mechanism underlying how glucose stimulates the secretion remains unclear. Here, we found that glucose at postprandial concentrations near the Km of GK for glucose ([~]10 mM), as well as nonmetabolizable glucose analogs, triggers GLP-1 secretion by intestinal L cells through directly binding to glucokinase (GK), without increasing intracellular ATP levels. We also show that in postprandially high glucose, glucose-bound GK interacts and inhibits KATP, as assessed by immunoprecipitation and FRET-FLIM analyses, as well as patch clamping. In addition, we have utilized GK mutants identified in diabetics, demonstrating that those mutants defective in binding to glucose impair GLP-1 secretion in however high glucose, whereas mutants constitutively bound to glucose with defects in catalysis, promote GLP-1 secretion even in low glucose. Furthermore, the KATP mutant unable to bind to GK blocks GLP-1 secretion in high glucose. We have also provided evidence showing that sodium-glucose cotransporter 1 (SGLT1) plays an important role in the basal secretion of GLP-1, but not in high glucose. We have thus revealed that glucose the molecule itself and GK act as a ligand and the receptor, respectively, to induce GLP-1 secretion in response to high glucose, providing new therapeutic avenues targeting metabolic diseases.

biochemistry↗
bioRxiv2025

Hepatic cytochrome P450 endoplasmic reticulum-associated degradation (ERAD): Topological determinants and cellular partnerships that dictate the preferential P450 proteolytic sorting into macroautophagy rather than UPS.

Many N-terminally endoplasmic reticulum (ER)-anchored cytochrome P450 proteins (P450s) turn over proteolytically via ER-associated degradation (ERAD), others via ER-to-lysosomal-associated degradation (ERLAD), and yet others via both pathways. What precisely dictates their differential proteolytic turnover is unknown. Herein, we employed rabbit liver CYPs 1A1 and 1A2, which reportedly reside in liquid-disorded (ld)- and detergent-resistant, liquid-ordered (lo)-ER-microdomains, respectively, governed by their specific N-terminal (NT) signal-anchor (SA) subdomains. We now report that this precise SA-dependent ER-topology not only determines the proclivity of CYP1A1 towards ERAD and CYP1A2 towards ERLAD, but also their differential lifespans. We further document that the detergent-resistant lo-ER-membranes (DRMs) are morphologically quite similar to mitochondria-associated ER-membranes (MAMs), documented cellular platforms for autophagic-initiation complexes. DRMs and MAMs, composed of saturated fatty-acids, glycosphingolipids and cholesterol, harbor many common morphological markers including the ER-specific prohibitin, erlin-1. Herein employing SURF, a split fluorogenic bifunctional complementation assay, we show that intracellularly, erlin-1 and CYP1A2 interact closely via their ER NT-SAs. siRNA-knockdown (KD) of erlin-1 in HepG2-cells, not only relocated CYP1A2 from DRMs to non-DRMs, but also impaired its ERLAD, resulting in insoluble cellular CYP1A2 aggregates. Upon erlin-1 KD, CYP1A2 ERLAD could be rescued by co-expression of a siRNA-resistant intact erlin-1 or just its NT-1-30 residue SA-domain. Our findings are the first to reveal that the CYP1A2 lifespan and preferential proclivity towards ERLAD is determined by its close association with erlin-1 within DRMs/MAMs. As proof of concept, we document that the ERLAD-proclivity of CYP2B1 is also similarly dependent upon erlin-1-DRM-association. SIGNIFICANCE STATEMENTThe endoplasmic reticulum (ER)-anchored cytochromes P450 (P450s) incur ER-associated degradation (ERAD) and/or ER-to-lysosomal-associated degradation (ERLAD). What determines their preferential proteolytic turnover is unknown. Here, employing P450s CYP1A1 and CYP1A2 that reside in their N-terminal signal anchor-determined liquid-disordered (ld)- and detergent-resistant, liquid-ordered (lo)-ER-microdomains, respectively, we documented that in HepG2-cells these ER-microdomains determine the CYP1A proteolytic preferences for ERAD vs ERLAD as well as their lifespans. More importantly, we discovered that its intimate interaction with the ER-specific prohibitin erlin-1 colocalized in these lo-ER-microdomains, specifically dictates CYP1A2s ERLAD-preference. Accordingly, siRNA-elicited erlin-1-knockdown disrupted CYP1A2-ERLAD, which was rescued upon coexpression of either a siRNA-resistant erlin-1 or just its N-terminal 1-30 residues. As proof of concept, we document similar characteristics for CYP2B1, another ERLAD-targeted P450.

cell biology↗
bioRxiv2025

ESM-LoRA-Gly: Improved prediction of N- and O-linked glycosylation sites by tuning protein language models with low-rank adaptation (LoRA)

Glycosylation associates with many diseases ranging from cancer to neurodegeneration and understanding these disease mechanisms requires the precise identification of glycosylation sites. Computational prediction of glycosylation sites has been useful to complement laborious experimental methods, while existing tools lack sufficient accuracy and scalability. Here, we introduce ESM-LoRA-Gly, a method that employs Low-Rank Adaptation (LoRA) to fine-tune the ESM2-3B protein language model for predicting both N- and O-linked glycosylation sites. According to the evaluation on the benchmark datasets, ESM-LoRA-Gly outperforms existing state-of-the-art techniques. The improvement is particularly significant (>100% in Matthews correlation coefficient) for the O-linked dataset. By substantially reducing trainable parameters while maintaining predictive power, ESM-LoRA-Gly enables computationally efficient proteome-scale predictions. This approach should be instrumental for advancing glycoproteomic research and accelerating therapeutic discovery for glycosylation-related diseases.

bioinformatics↗
bioRxiv2025

RESM: Capturing sequence and structure encoding of RNAs by mapped transfer learning from ESM (evolutionary scale modeling) protein language model

RNA sequences exhibit lower evolutionary conservation than proteins due to their informationally constrained four-letter alphabet, compared to the 20-letter code of proteins. More limited information makes unsupervised learning of structural and functional evolutionary patterns more challenging from single RNA sequences. We overcame this limitation by mapping RNA sequences to pseudo-protein sequences to allow effective transfer training from a protein language model (protein Evolution-Scale Model 2, protESM-2). The resulting RNA ESM (RESM) outperforms 12 existing RNA language models in zero-shot prediction, not only in sequence classification but also in RNA secondary structure and RNA-RNA interaction prediction. Further supervised fine-tuning demonstrates RESMs generalizability and superior performance over the existing models compared across multiple downstream tasks, including mRNA ribosome loading efficiency and gene expression prediction, despite RESM being trained exclusively on noncoding RNAs. Moreover, RESM can generalize to unseen sequences beyond its 1,024-nucleotide training limit, achieving 81.3% improvement over state-of-the-art methods in supervised secondary structure prediction for RNAs up to 4,000 nucleotides, limited only by the available GPU memory, while providing >1000-fold speedup compared to MSA-based approaches. RESM provides a robust foundation for deciphering RNA sequence-structure-function relationships, with broad implications for RNA biology.

bioinformatics↗
bioRxiv2025

APOE knockout attenuates vascular graft fibrosis by limiting profibrotic macrophage formation through low-density lipoprotein receptor related protein 1

Vascular graft fibrosis can cause a decrease in cellular infiltration and capillary ingrowth in vascular walls and vascular stiffening. As such, there are still no vascular grafts that can be used in blood vessels where their diameters are less than 6 mm in patients. Although various approaches have been evaluated to mitigate implant-associated fibrosis, effective treatments remain quite limited. In this study, we demonstrated that APOE was significantly increased during vascular regeneration after graft implantation in vivo. APOE knockout (KO) increased compliance of regenerated aortas and reduced extracellular matrix (ECM) deposition in adventitia of the regenerated aortas. Using single cell RNA sequencing (scRNA-seq), a subset of profibrotic macrophages was found to be involved in graft fibrosis and APOE KO limited the formation of profibrotic macrophage formation during vascular regeneration. The interaction between APOE and low-density lipoprotein receptor related protein 1 (LRP1) partially mediated fibrotic differentiation of the macrophages. Profibrotic macrophages promoted graft fibrosis mainly through secretion of insulin-like growth factor-1 (IGF-1) that could support proliferation of fibroblasts. Finally, we showed that APOE knockdown in vivo using adeno-associated virus (AAV) improved the compliance of regenerated aortas and reduced ECM deposited in the adventitial areas by limiting formation of profibrotic macrophages. Collectively, these data indicate that APOE promotes the profibrotic transition of macrophages partially through LRP1, and the profibrotic macrophages increase the proliferation of fibroblasts via IGF-1. Inhibition of APOE by AAV can alleviate graft fibrosis occurring during vascular regeneration.

bioengineering↗