Biology subjects

Choi, M.

Publications and source records attributed to Choi, M..

At least 19 recordsLinked to original sources

bioRxiv2026

Brain-wide Cerebrospinal Fluid Recirculation Collapse in Alzheimers Disease

Cerebrospinal fluid (CSF) flow and the glymphatic system are essential for maintaining brain homeostasis, yet the precise mechanisms driving their profound failure in Alzheimers disease (AD) remain largely elusive. Here, we reveal that macroscopic CSF propagates dorsally via the perivascular spaces (PAS) of major ventral arteries--the Circle of Willis (CW), middle cerebral artery (MCA), and olfactory artery (OlfA)--which undergo severe structural degradation in an AD mouse model. High-resolution imaging of the dorsal cortex demonstrates that this structural collapse induces a severe influx blockade; in AD, CSF tracers form massive aggregates within the PAS, physically decoupling the gliovascular interface, while tracers entering the parenchyma are pathologically sequestered by amyloid plaques. Furthermore, we identify an active, cholesterol-mediated clearance mechanism responding to exogenous CSF tracer infiltration. While healthy brains rapidly synthesize and efficiently clear this cholesterol via vascular efflux within 120 minutes, the AD model exhibits blunted cholesterol mobilization and a complete failure of capillary efflux. This terminal bottleneck forces uncleared cholesterol to retrogradely accumulate into pathological aggregates within perivascular pericytes and neuronal somata. Collectively, our findings demonstrate that AD dismantles the entire CSF recirculation loop--from macroscopic ventral entry to terminal microvascular efflux--highlighting pericytic and neuronal lipidosis as fundamental drivers of brain fluid stagnation. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=114 SRC="FIGDIR/small/755971v1_ufig1.gif" ALT="Figure 1"> View larger version (48K): org.highwire.dtl.DTLVardef@1c62a5eorg.highwire.dtl.DTLVardef@107919eorg.highwire.dtl.DTLVardef@1e921baorg.highwire.dtl.DTLVardef@a549ad_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2026

A scalable human neuron model of Alzheimer's disease relevant tauopathy reveals mechanisms linking Tau fibrillization to synaptic dysfunction

Tauopathies, including Alzheimer's disease, are driven by pathological aggregation of hyperphosphorylated Tau, which disrupts synaptic integrity, impairs neuronal communication, and contributes to cognitive decline. To dissect tauopathy pathogenesis and enable therapeutic discovery, reliable and scalable human iPSC-neuron models are essential. Here, we developed two complementary iPSC-derived neuron models: an endogenous Tau seeding model, in which neurons are challenged with pre-formed Tau fragments that form paired helical filament (PHF)-consistent structures, and a Tau-0N3R overexpression seeding model to accelerate pathology. Both models recapitulate hallmark features of tauopathy, including the progressive formation of intracellular, hyperphosphorylated, sarkosyl-insoluble, and conformationally altered Tau aggregates (AT8, MC1 positive), along with synaptic and neuronal dysfunction. Cryogenic electron tomography (cryo-ET) further revealed the morphology of Tau fibrils within cells, as well as the ultrastructure of Tau fibrils trapping synaptic vesicles in situ. Using this platform, we performed integrated phosphoproteomics, high-content screening, and functional validation to identify key pathways driving Tau aggregation. MARK2-mediated phosphorylation within Tau's microtubule-binding domain emerged as an early trigger of aggregation, confirmed by site-specific mutagenesis. In parallel, small molecules targeting the PI3K/mTOR/GSK3 pathway reduced aggregation and restored synaptic function, with GSK3 inhibition lowering phosphorylation at critical aggregation-driving sites on Tau. Together, these findings establish a physiologically relevant, scalable platform for therapeutic screening that connects Tau seed uptake, site-specific phosphorylation, fibril formation, and synaptic disruption, ultimately identifying mechanistically separable intervention points along the aggregation cascade.

cell biology↗

bioRxiv2026

A modular cranial window enabling maintainable widefield optical access in non-human primates

Longitudinal optical imaging of the primate cortex requires stable yet maintainable cortical access, but existing cranial window approaches remain limited by mechanical instability and tissue responses that progressively degrade optical clarity. These challenges are amplified in primates, where large craniotomies and physiological variability complicate chronic experiments. Moving beyond conventional fixed cranial windows, we present the PRIME cranial window (Primate Reconfigurable Interchangeable Modular Enclosure), a modular platform enabling maintainable, long-term, and widefield optical access in awake macaques. Built on a modular architecture, PRIME supports non-invasive adjustment, repeated cortical access, and on-demand maintenance, including tissue removal, without surgical re-entry, while a dual-ring sealing mechanism prevents cerebrospinal fluid leakage and contamination. Combined with optimized viral delivery, implantation, and maintenance strategies, PRIME preserves cortical integrity and optical clarity for long-term optical imaging. PRIME establishes a versatile, re-accessible experimental framework for chronic, large-area optical studies of the primate brain.

neuroscience↗

bioRxiv2026

Divergent transcriptomic transition programs in Alzheimer disease: immune priming and synaptic collapse in females versus organelle membrane bifurcation in males

Alzheimers disease (AD) disproportionately affects women, who account for approximately two-thirds of prevalent cases. Despite decades of research, the mechanistic basis for this profound sex disparity remains poorly resolved. Prior transcriptomic studies have predominantly used pooled or female-enriched cohorts, obscuring whether the transition from normal cognition (NCI) to AD follows a common molecular program across sexes or reflects fundamentally divergent biology. Here, we analyzed bulk RNA-sequencing data from the dorsolateral prefrontal cortex of 624 individuals (401 females, 223 males) in the ROSMAP cohort using a transition-aware framework that separates variance instability along the disease axis from mean expression changes. We demonstrate that female and male brains exhibit structurally distinct transcriptomic transition programs. Females display a sequential, multi-tier architecture: interferon and immune variance priming (628 genes) is detectable early at the NCI-to-mild cognitive impairment (MCI) interval, which structurally precedes a massive mean-level synaptic and neuropeptide loss (8,935 genes) in AD. The female NCI-to-MCI interval alone produces 1,249 variance bifurcation events entirely absent in males. Conversely, males exhibit no early immune priming or powered mean-level changes. Instead, they collapse into a single large variance bifurcation pool (8,237 genes) heavily enriched for system-wide organelle membrane fusion, post-translational modification, and intracellular trafficking. These findings reveal that the AD transcriptomic transition is not a unitary program quantitatively modified by sex, but two distinct biological trajectories. This fundamental divergence motivates sex-stratified mechanistic models and independent biomarker development for each sex, cautioning against analytical pooling in transition-stage cohorts.

bioinformatics↗

bioRxiv2026

Mtf2 Safeguards Naïve Pluripotency by Restricting Trophectoderm Competence

The first fate decision in mammalian embryos segregates embryonic and extraembryonic lineages, establishing a largely irreversible transition from totipotency to pluripotency. The mechanisms that enforce this unidirectionality remain incompletely understood. Here, we identify the PRC2.1 subunit Mtf2 as a key epigenetic barrier that safeguards lineage fidelity in naive embryonic stem cells (ESCs). Despite globally elevated H3K27me3 levels in naive ESCs, loss of Mtf2 derepresses trophectoderm (TE) regulators, enhances cellular plasticity, generating both ICM- and TE-like populations. Mtf2-deficient cells readily acquire expanded potential with extraembryonic competence and form blastoids without trophoblast stem cell supplementation, whereas loss of the PRC2.2 subunit Jarid2 has minimal effects. Integrated CUT&Tag and transcriptomic analyses reveal that Mtf2 directly represses TE-associated genes in naive ESCs. Thus, reduced Mtf2 occupancy permits their activation in expanded potential states. Collectively, our findings establish Mtf2 as a central epigenetic barrier that stabilizes naive pluripotency and restricts inappropriate acquisition of extraembryonic fate.

Developmental Biology↗

bioRxiv2026

Transcriptomic phase transitions along the Alzheimer's disease continuum using riemmaninan tensor model: a bifurcation-intermediate variance framework on the ROSMAP cohort

Alzheimers disease (AD) progression involves systemic network transitions. To capture these using ROSMAP bulk RNA-seq (n=624), we focused on the geometry of the covariance structure, performing a Riemannian (Log-Euclidean) analysis of stage-wise covariance matrices as points on the manifold of symmetric positive-definite (SPD) matrices. On the SPD manifold the three stages were non-collinear: geodesic distances were non-uniform and MCI was displaced from the NCI-AD chord, while the von Neumann entropy of the covariance structure dipped at MCI (S = 2.760, 2.639, 2.647 for normal cognitive intact NCI, mild cognitive impairment MCI, AD) and the path-curvature profile reached a minimum there -- together identifying MCI as a saddle/bifurcation state. The differential covariance spectrum (CAD - CNCI) separated AD-amplified ("structural collapse") from AD-suppressed ("protective loss") modes. Ultimately, second-order statistics analyzed through Riemannian geometry, rather than Euclidean summaries, reveal AD progression structure invisible to mean-level analysis.

bioinformatics↗

bioRxiv2026

Comparative connectomics reveals stage-specific gap junction rewiring that reshapes avoidance behavior

Environmental stress can remodel neural circuits, yet how such rewiring generates behavioral adaptation remains poorly understood. The dauer stage of Caenorhabditis elegans provides a unique model to address this question. Here, we investigated how dauer-specific circuit remodeling alters nociceptive behavior. We found that dauers exhibit markedly shorter avoidance durations than adults. Comparative connectomics revealed substantial expansion of gap junctions within the dauer nociceptive circuit. Calcium imaging further showed that dauer neurons exhibit faster and more transient activity dynamics throughout the circuit. Introducing synthetic dauer-specific gap junctions into adults was sufficient to recapitulate dauer-like neuronal activity patterns and significantly shorten avoidance duration. Despite extensive circuit rewiring, however, avoidance initiation remained preserved across developmental stages. Together, our findings demonstrate that the dauer connectome is selectively rewired through gap junction remodeling to tune behavioral persistence while robustly preserving behavioral initiation, revealing how developmental circuit reorganization balances flexibility and stability for survival under stress.

neuroscience↗

bioRxiv2026

Direct RNA Sequencing reveals epitranscriptomic regulation of brain cells and Alzheimer's Disease pathology

Alternative mRNA splicing and post-transcriptional RNA modification are key mechanisms that regulate transcript function; however, their role in neuronal activity and neurodegenerative disease remains poorly defined. In this study, we evaluated two nanopore-based long-read sequencing (LR-seq) formats: cDNA-PCR sequencing (CPS) and direct RNA sequencing (DRS). We then applied DRS to profile both full-length isoforms and RNA modifications in major brain cell types derived from induced pluripotent stem cells (iPSCs) and post-mortem Alzheimers disease (AD) brains. Relative to CPS, DRS achieved higher accuracy and sensitivity for transcript quantification, de novo transcript model construction, and open reading frame (ORF) annotation across neuropathological gene sets. Focusing on iPSC-derived neurons, we built a multi-omic atlas to connect transcriptional output with translational engagement and protein abundance, by integrating DRS-based mRNA abundance, N6-methyladenosine (m6A) status and poly(A) tail length with ribosome profiling (Ribo-seq) and mass spectrometry (MS). The combination of DRS and Ribo-seq data demonstrated synergism in predicting protein abundance. This analysis also uncovered a significant inverse relationship between m6A modification and mRNA abundance, which was dependent on the engagement of the ribosomal A-site. Lastly, we applied DRS to the epitranscriptomic analysis of AD brain samples, demonstrating that m6A profiles can be used to distinguish early-versus late-stage disease.

neuroscience↗

bioRxiv2026

Riemannian Tensor Analysis Identifies a Bifurcation State in the Single-Cell Transcriptomic Landscape of Parkinson's Disease

The transition from healthy aging to Parkinsons disease (PD) involves highly volatile transcriptomic rewiring that remains invisible to conventional mean-based analyses. To decode this critical tipping point, we integrated single-cell RNA-sequencing with a novel Log-Euclidean Riemannian tensor framework. By conceptualizing distinct transcriptomic states as symmetric positive definite (SPD) covariance tensors, we bypassed Euclidean geometric artifacts to accurately map the macroscopic network architecture of the human prefrontal cortex. Our analysis identified a highly unstable, intermediate bifurcation (BIF) state. Thermodynamic and topological validation demonstrated that this BIF state operates as a definitive geometric saddle point, characterized by maximal von Neumann entropy and positioned perfectly equidistant between the healthy (HC) and pathological (PD) Frechet means on the non-Euclidean manifold. Furthermore, spectral deconstruction of differential covariance networks [Formula] isolated the exact topological drivers of this transition--revealing the catastrophic structural collapse of core synaptic and electrophysiological hubs prior to overt pathological commitment. Ultimately, this ab initio geometric framework fundamentally redefines PD progression, providing a quantitative blueprint to intercept neurodegenerative trajectories during their final reversible window.

systems biology↗

bioRxiv2026

Allosteric Biosensors Unravel GTPase-Effector Feedback

Fluorescent biosensors that report protein conformation in vivo have been invaluable for understanding how the spatio-temporal dynamics of signaling controls cells. However, for GTPases these biosensors report the activated conformation using reagents that block the binding of downstream proteins, generating dominant negative effects and altering normal cell physiology. We present here a generalizable design to make GTPase biosensors (AlloRac1 and AlloCdc42), in which a circularly permuted fluorescent protein is inserted into a conserved loop allosterically connected to the effector binding site, generating activity-dependent fluorescence without blocking ligand interactions. The Rac1 biosensor showed that effector interactions led to increased Rac1 activation, indicating an auto-regulatory positive feedback made visible by the new biosensor design. This feedback regulated the kinetics and localization of Rac1 activity, including Rac1 activity gradients that controlled motility. Feedback was generated through Rac1 interaction with the effector Pak1, which led to further activation of Rac1 by the guanine exchange factor {beta}-Pix. The new biosensor approach enables quantitative imaging of previously obscure spatio-temporal dynamics in GTPase regulation.

cell biology↗

bioRxiv2026

Hormone signaling and immune programs define differential endocrine responsiveness in high-risk breast tissue

Hormone therapies are frequently used to reduce breast cancer risk in individuals at increased risk for primary or subsequent disease; however, tissue-level responses to these therapies are heterogeneous and incompletely understood. Background parenchymal enhancement (BPE) on breast magnetic resonance imaging (MRI) provides a non-invasive radiologic readout of breast tissue features associated with endocrine responsiveness and cancer risk. Although BPE is associated with hormonal exposure, a subset of patients with BPE do not show a response to preventive endocrine therapy and therefore may remain at increased breast cancer risk. In this study, we integrated single-nucleus RNA sequencing and spatial transcriptomics to define the determinants of endocrine responsiveness in the setting of BPE. We identify hormone-driven epithelial cells with high levels of estrogen signaling and endocrine responsiveness, together with immune-associated epithelial programs characterized by diminished luminal identity and increased expression of immune-modulatory pathways, including major histocompatibility complex (MHC) class II and CD74. Functional organoid assays validate that these epithelial states exhibit differential sensitivity to tamoxifen and demonstrate that inflammatory signals can induce immune-modulatory epithelial programs. Together, our findings identify hormone signaling and immune programs as key determinants of endocrine responsiveness in breast tissue and provide a biological basis for interpreting radiologic markers relevant to cancer prevention.

cancer biology↗

bioRxiv2025

Ataxin-2 knockdown is neuroprotective via cell-autonomous and non-cell-autonomous mechanisms

Ataxin-2 (ATXN2) is a genetic modifier of TDP-43 toxicity and a promising therapeutic target in amyotrophic lateral sclerosis (ALS). However, the mechanisms underlying its neuroprotective effects remain poorly understood. Here we show that ataxin-2 reduction confers neuroprotection by engaging adaptive metabolic programs in both neuronal and glial cells. Using global proteomic profiling in yeast and mouse TDP-43 models, we establish that pbp1/ataxin-2 (pbp1 is the yeast ataxin-2 ortholog) activates orthogonal stress-adaptive programs that enable alternative energy production and augment trophic support, rather than simply reversing TDP-43-induced damage. In neurons, ataxin-2 downregulation activates glycolysis and reductive glutamine carboxylation driven by IDH1, restoring ATP production independently of impaired mitochondria. In astrocytes, ataxin-2 downregulation upregulates cholesterol biosynthesis via HMGCS1, enhancing trophic support to neurons in a non-cell-autonomous manner. Full neuroprotection requires both mechanisms: neuronal survival is only completely rescued when ataxin-2 is reduced in the context of neuron-astrocyte co-cultures. Importantly, ataxin-2 downregulation protects against both TDP-43 gain-of-function and loss-of-function toxicity, broadening its therapeutic relevance. Collectively, our findings reveal novel mechanisms by which ataxin-2 orchestrates adaptive metabolic resilience and establish ataxin-2 as a regulator of stress-adaptive response in brain cells under the conditions of stress.

neuroscience↗

bioRxiv2025

BoltzGen: Toward Universal Binder Design

We introduce BoltzGen, an all-atom generative model for designing proteins and peptides across all modalities to bind a wide range of biomolecular targets. BoltzGen builds strong structural reasoning capabilities about target-binder interactions into its generative design process. This is achieved by unifying design and structure prediction, resulting in a single model that also reaches state-of-the-art folding performance. BoltzGens generation process can be controlled with a flexible design specification language over covalent bonds, structure constraints, binding sites, and more. We experimentally validate these capabilities in eight diverse design campaigns with functional and affinity readouts across 26 targets. In our experiments, binder modalities span from nanobodies to disulfide-bonded peptides, and targets from disordered proteins to small molecules. In particular, we identify nanobody binders for novel targets with low similarity to proteins with already known bound structures. We release model weights, data, and both inference and training code at: https://github.com/HannesStark/boltzgen.

bioengineering↗

bioRxiv2025

Preference-Based Fine-Tuning of Genomic Sequence Models for Personal Expression Prediction with Data Augmentation

Despite substantial progress in genomic foundation models, accurately predicting inter-individual variation in gene expression from DNA sequence alone remains a major challenge. Current sequence-based models, such as Enformer and Borzoi, trained exclusively on the reference genome, cannot capture the effects of individual-specific regulatory variants. Moreover, the acquisition of paired whole-genome and transcriptome data required for personalized modeling is hindered by privacy and data-sharing constraints. To address this limitation, we integrate genomic data synthesis with established statistical frameworks. Our approach generates thousands of synthetic training samples by simulating genetic variation from the 1000 Genomes Project and assigning pseudo-expression labels using PrediXcan, a validated eQTL-based predictor. Because simulated and real expression values differ in scale and distribution, we introduce a preference-based objective that models relative rather than absolute expression patterns. Fine-tuning Enformer through alternating cycles of real-data regression and synthetic-data preference optimization enables efficient learning from both real and synthesized data. Using the GEUVADIS dataset, our framework outperforms AlphaGenome, PrediXcan, and Enformer fine-tuned without synthesized data, demonstrating that simulation-based integration of population-level regulatory knowledge can effectively mitigate data scarcity and improve cross-individual generalization in sequence-based gene expression prediction. Availability and implementationCode and data are available at https://github.com/pacifiic/augment-finetune-genomics.

bioinformatics↗

bioRxiv2025

Isolation of Pure Disease Specific Aging Trajectories in Spatial Transcriptomics via the Delta Delta Method

Disentangling normal aging from disease driven transcriptional change remains a major obstacle for spatial genomics. We introduce the Delta-Delta ({Delta}{Delta}) Method, a contrastive trajectory framework that resolves a four-dimensional progression (genes x cell types x brain regions x time) by subtracting the wild type (WT) aging trajectory from the transgenic (TG) trajectory to yield a pure disease trajectory ({Delta}{Delta}log2FC). The method is platform agnostic, integrates with common spatial transcriptomics workflows, and outputs direction and speed of change summaries, enriched pathways, and region and cell type specific maps. In a demonstration using G2-3 -synuclein TG mice and age matched WT controls at 6 and 10 months across hippocampus and midbrain, {Delta}{Delta} uncovered opposite regional dynamics in glutamatergic neurons and a convergent enrichment of RNA splicing pathways, corroborated by alternative splicing analyses. By explicitly modeling time while controlling for aging within each region and cell type, the {Delta}{Delta} Method isolates disease specific molecular programs that are obscured in conventional bulk or single cell analyses, and provides a generalizable framework for trajectory aware mechanistic target prioritization in neurodegeneration and other progressive conditions.

bioinformatics↗

bioRxiv2025

Recent social experience alters song behavior in Drosophila

Innate behaviors are hardwired into the nervous system, allowing animals to produce them instinctively and without prior learning. This includes social behaviors (e.g., aggression, courtship) in most animals. If and how learning shapes such innate behaviors has not been systematically explored. Here, we investigate learning from social feedback in Drosophila. We use closed-loop optogenetics to systematically perturb the social experience of individual flies during courtship, and find that altered feedback changes the male flys innate courtship strategy, opening the door for investigating the neural and molecular basis of social plasticity in flies. TeaserMale flies adjust their innate communication strategy in response to systematic perturbation of social experience.

neuroscience↗

bioRxiv2025

Targeted degradation of SARS-CoV-2 via the autophagy-lysosome system using chemical mimetics of the N-degron pathway

In the N-degron pathway, ATE1 transfers the amino acid L-arginine (L-Arg) from Arg-tRNAArg to N-terminal (Nt) residues of cellular proteins. The resulting Arg/N-degrons bind the autophagic receptor p62/SQSTSM-1/Sequestosome-1 to induce lysosomal degradation of various biomaterials. Here, we demonstrate that the chemical mimetics of Arg/N-degrons, termed autophagy-targeting ligands (ATLs), can induce lysosomal degradation of SARS-CoV-2 (severe acute respiratory syndrome coronavirus-2) via p62-mediated macroautophagy. In Vero E6 cells infected with SARS-CoV-2, ATLs promoted p62 self-polymerization and enhanced LC3 synthesis and lipidation, leading to viral sequestration within autophagosomes for lysosomal degradation. In transgenic mice overexpressing human angiotensin-converting enzyme 2 (ACE2), oral administration of ATL1014 inhibited viral replication and increased viability. In a Syrian hamster model, ATL1014 attenuated viral replication in the lungs and demonstrated efficacy in inflammatory lesions and pulmonary congestions. These results identify the N-degron pathway as a potential target for a host-targeting strategy (HTS) against a broad spectrum of viruses.

microbiology↗

bioRxiv2025

Functional dissection of SPOP on the amino acid level reveals a comprehensive functional landscape of variants during tumorigenesis

Numerous proteins display pleiotropic functions in different clinical contexts. However, molecular mechanism underlying such effects is rarely understood. Speckle-type POZ protein (SPOP) is a typical example, exhibiting tumor-suppressing or -promoting effects in different tumor types in accordance with different amino acid changes; specifically, two distinct sets of variants in SPOP are commonly found in subsets of prostate cancer and endometrial cancer patients. To comprehensively characterize the functional landscape of SPOP alteration, we performed a deep mutational screening (DMS), elucidating the functionality of 7,933 out of 8,228 possible single amino acid changes (96.4% coverage). Leveraging the observation that overexpression of human SPOP leads to yeast growth arrest, we assessed the functionality of each variant using a yeast proliferation assay. In addition, our approach combined long-read and short-read sequencing. Finally, our DMS model enables a clear distinction of likely-loss-of-function (LoF) variants that are enriched in prostate cancers and reveals their differential characteristics in both protein structure and genetic assessments. These results demonstrate the utility of our approach in high resolution mapping and amino-acid-level interpretation of protein function. Significance statementGenetic mutations often play different roles in cancer, driving or suppressing tumor growth depending on their molecular context. The protein SPOP is a striking example that acts as either a tumor-suppressor or oncogene depending on the specific mutation and tissue. However, understanding how thousands of possible mutations alter its function has been a major challenge. Here, we applied deep mutational scanning to experimentally measure the functional effects of nearly every possible amino acid change in SPOP. This work provides the first comprehensive functional map of SPOP variants, not only advancing basic knowledge of cancer biology but also establishing a framework for interpreting patient mutations in precision medicine.

genetics↗