Biology subjects

Wei, K.

Publications and source records attributed to Wei, K..

At least 19 recordsLinked to original sources

bioRxiv2026

Conditions and properties of non-Mendelian transmission of recombinants in female meiosis

In female meiosis, each of the four chromatids in a tetrad has a 25% chance to be selected for transmission through the pronucleus, realizing Mendel's second law. Here, we characterize a cheating but unselfish behaviour that has documented cases across diverse animal taxa including humans and flies, whereby chromatids with crossovers (COs) have a transmission advantage, resulting in increased production of recombinant offspring. Taking advantage of Drosophila ovarian physiology, we show that this form of meiotic cheating which we call recombinant drive occurs on the autosomes when females are under nutrient stress, potentially as a mechanism for recombination plasticity, while curiously, the effect is suppressed on the X by the distributive system. We explored recombinant drive using simulations and identified unique quantitative signatures that are at odds with several intrinsic properties of linkage. One violation is the production of more offspring with recombinant versus parental allele combinations, which we empirically demonstrated to be possible. Further, we show that recombinant drive can appear to modify CO patterning effectively acting as assurance and interference mechanisms, even when it has no influence on and acts downstream of CO spacing. We discuss the potential benefits and consequences of having a conserved method that can rapidly increase recombination in response to stress, and speculate on a mechanism for the preferential transmission of COs at meiosis II. Overall, our study revealed distinct properties and behaviours of a poorly understood, but potentially widespread, conserved phenomenon and offers avenues for broad detection and mechanistic dissection.

genetics↗

bioRxiv2026

Cancer Cell Line Heterogeneity Imposes a Primary Bottleneck for Virtual Perturbation Screening at Scale

UniPert-G2CP (Li et al., Cell, 2026) bridges genetic and chemical screens from molecular representation to phenotype modeling across five cancer cell lines [1]. Here we extend this architecture to 162 cell lines, 32,039 compounds, and genome-wide output (12,328 genes), and find that the resulting prediction platform reveals a striking performance divide: normal and primary cell lines achieve substantially higher prediction fidelity (mean per-cell-line Pearson Correlation Coefficient (PCC) = 0.480, median 0.502) than cancer lines (0.296, median 0.324; Mann-Whitney p = 3.93e-11), identifying cancer cell line heterogeneity as a primary bottleneck for virtual perturbation screening at scale. To understand the sources of this divide, we analyzed per-cell-line performance, genome-wide directional accuracy, mechanism-clustering (SMD), and compound-protein interaction (CPI) enrichment. Directional accuracy on top-5% effect-size genes reaches 73.8% (genome-wide 60.8%), with pathway-dependent recovery: of four literature-supported perturbation-gene pairs queried across three cell lines, two were recapitulated (dexamethasone-TSC22D3/NFKBIA/FKBP5; bortezomib-BAG3/DNAJB1/HSPA1A), one was absent (CD36 depletion-PPARG/CEBPA in ASC), and one was partially recapitulated (metformin-SLC7A5 in HEPG2). Mechanism-clustering SMD of the learned embedding reached 1.636 (vs. original 1.85; 88.5% retention at 32x cell-line coverage), exceeding the ECFP4 fingerprint baseline (1.613), while self-consistency Mantel rho=0.852 confirmed the model retains compound mechanism structure internally. Overall held-out performance: genetic perturbation PCC=0.442 (978-gene subset); novel drug PCC=0.3047 (genome-wide). Analysis of CPI enrichment reveals that training-data overlap inflates apparent performance: 54.9% of Touchstone evaluation pairs overlap with our ChEMBL-derived training CPI pairs, reducing effective EF from 139 to 109 at top 0.5% yet remaining far above random (1.0). These findings establish the first large-scale characterization of cell-type-dependent generalization in perturbation-to-phenotype prediction. The observed performance stratification between normal and cancer lines generates testable hypotheses for why virtual cell models degrade on heterogeneous cancer contexts, and provides a diagnostic framework for identifying where and why such models fail--informing future architecture improvements targeting the CPI vocabulary gap, protein encoder design, and cell-type-aware training strategies.

bioinformatics↗

bioRxiv2026

A Modular Bio-Hybrid Skin Model for Optical Testing Applications

Synthetic optical skin models offer reproducible, tunable optical properties but lack biological responsiveness, while tissue engineered skin models provide cellular authenticity but suffer from optical variability and limited controllability. The growing demand for alternatives to animal models in the development and validation of optical biomedical technologies highlights the need for a new class of test system that combines the strengths of both approaches while addressing their respective limitations. Here, we introduce the concept of a modular biohybrid skin model, a new testing concept that integrates an optically defined artificial epidermal layer, incorporating polydopamine nanoparticles for changes in skin tone, with living human keratinocytes in two and three-dimensional configurations. In the Optical Protection Model, UV-B-induced apoptosis in primary keratinocytes is quantitatively modulated by model pigmentation level, demonstrating a relationship between optical attenuation and caspase 3/7 activity across three artificial skin tone conditions. In a Structured Dermal Model, keratinocytes seeded onto a hydrogel scaffold localize within follicle-like microcavities, as confirmed by live/dead staining and confocal z-stack imaging. Together, these experiments lead to a new category of test system in the space between inert optical models and variable tissue models that may contribute to reducing the reliance on animal models in biomedical optics.

bioengineering↗

bioRxiv2026

A dermal-epidermal junction-inclusive skin model enabled by controllable hydrogel swelling

Tissue engineered skin models are important tools for the in vitro study of physiological and pathophysiological processes as well as the evaluation of therapeutic strategies and the efficacy of pharmaceutical and cosmetic compounds. Replicating the functional anatomy of cutaneous tissue is a crucial aspect in ensuring that observations made using these models are translatable to the actual situation in native skin. However, most contemporary full-thickness skin models neglect the reconstruction of the undulated microtopography of the dermal-epidermal junction (DEJ), which not only contributes to the biological functionality of the skin (e.g. stem cell niches), but also affects tissue mechanics and drug diffusion. Herein, we fabricated bilayer skin models with DEJ-like microtopographies introduced by interfacial wrinkling between a hydrogel and a nanofibrous membrane through a controllable swelling-deswelling approach. The interfacial wrinkles contributed to the structural integrity of the bilayer models. Their formation could be induced in the presence of living cells through mechanical stress-driven buckling instabilities, thus differentiating the process from commonly used pre-patterning techniques. Bilayer models supported the co-culture of human dermal fibroblasts and human epidermal keratinocytes, and the formation of stratified epithelia. Our findings provide a potential alternative method to introduce DEJ-like anatomical features into full-thickness skin tissue models.

bioengineering↗

bioRxiv2026

SPP1hi macrophages in fibrin niches promote hyperplastic tissue remodeling in rheumatoid arthritis synovium

In chronic inflammatory diseases, maladaptive tissue remodelling is driven by a complex interplay of resident cells, immune filtrates and the extracellular matrix. In the autoimmune disorder rheumatoid arthritis (RA), synovial tissue undergoes assive expansion to form an invasive pannus that drives the erosion of cartilage and bone. The mechanisms mediating this ggressive growth are incompletely defined. Using spatial transcriptomics profiling of patient tissue, we detected an bundance of proliferating fibroblasts near the synovial tissue lining surface and adjacent to SPP1hi macrophages. Notably, ese synovial lining regions were also distinctly marked by deposits of the clot-forming protein fibrin. While the SPP1hi acrophages phenotypically resemble pro-fibrotic macrophages that drive lung and liver fibrosis, these niches were devoid f the dense highly ordered collagen that marks fibrosis. Functionally, we found that SPP1hi macrophages degrade and hagocytose fibrin matrices and promote fibroblast proliferation. As fibrin provides transient matrices for de novo tissue eneration in the context of wound healing, these data support a model of hyperplastic tissue outgrowth involving SPP1hi acrophages, fibroblasts and fibrin matrices adhered to the exterior synovial tissue surface. While current RA therapies rimarily aim to dampen pro-inflammatory responses, our findings provide rationale for targeting pro-generative pathways nd SPP1hi macrophages. Once Sentence SummarySPP1hi macrophages in RA synovial fibrin deposits promote tissue hyperplasia.

immunology↗

bioRxiv2026

Connections across regional glymphatic clearance, neural activity and amyloid-β deposition in cortex

Neural activity inevitably produces waste, which promotes neurodegeneration with topographic features. The glymphatic system is important for waste clearance. However, the spatial characteristics of glymphatic clearance across cortex and whether it interplays with neural activity in contribution to amyloidosis in human remain unexplored. Here, by intrathecal administration of gadolinium-based contrast agents, glymphatic influx and clearance patterns across cortex in 96 participants are depicted via Glymphatic MRI. Analyses integrating post-mortem transcriptomic profiles from Allen Human Brain Atlas indicate that, genes related with excitatory and inhibitory neurons, and pathways engaging in synaptic function were enriched in regions with faster glymphatic clearance. FALFF was calculated from resting-state fMRI to represent neural activity. At the regional level, based on a subgroup with rs-fMRI (N = 15), regional glymphatic clearance was positively coupled with spontaneous neural activity. Mismatch index, reflecting decoupling between spontaneous neural activity and glymphatic clearance function, turned out to be positively associated with regional severity of amyloidosis using open-source 11C-PiB dataset. Together, this study for the first time demonstrates the intricate interplays between neural activity and glymphatic dynamics from transcriptional to physiological level. The mismatch between these two processes may serve as an undescribed comprehensive mechanism promoting regional vulnerability to proteopathy and subsequent neurodegeneration in cortex.

neuroscience↗

bioRxiv2026

Protein Language Models Encode Evolutionary Grammar but Conflate Topological and Thermodynamic Phases

How a one-dimensional sequence encodes a dynamic thermodynamic ensemble rather than a static geometry remains a fundamental biophysical challenge. While protein language models achieve high structural accuracy from single sequences, it remains unclear whether they internalize physical folding principles or merely capture evolutionary statistical correlations. We investigate this representational capacity by analyzing the latent manifold of ESM-2 across 11068 proteins, focusing on fundamental exceptions to the Anfinsen assumption: intrinsically disordered, fold-switching, and knotted proteins. We demonstrate that ESM-2 integrates out microscopic backbone geometry during feature extraction. It instead forms a macroscopic sequence grammar manifold that separates random sequences while ordering biological proteins along physicochemical composition gradients. Consequently, the model exhibits topological aliasing. It conflates these physically distinct categories because they frequently share overlapping sequence statistics despite divergent three-dimensional topologies. A region-replacement experiment establishes this aliasing as an intrinsic characteristic rather than an artifact of sequence dilution, and a structure-aware control model (SaProt) demonstrates that explicit structural tokens partially alleviate topological blindness for static anomalies while failing to resolve multi-state thermodynamic phases. Furthermore, local curvature analysis reveals class-invariant geometric turbulence that homogenizes microscopic details while preserving macroscopic distinguishability. ESM-2 therefore functions as an evolutionary grammar compressor rather than a microscopic threedimensional geometry encoder, imposing defined boundaries on tasks requiring high-precision geometric awareness and necessitating integration with explicit physical constraints.

biophysics↗

bioRxiv2026

Fibroblasts neurotrophin signaling sustains pathological vascular maturation in rheumatoid arthritis.

Treatment failures in rheumatoid arthritis (RA) leads to undesirable morbidity associated with immunosuppression. Recent studies of synovial tissue from refractory RA patients highlight the role of synovial fibroblasts and vascular endothelium in driving treatment failure. Utilizing high-dimensional spatial transcriptomics, we uncovered a crucial role for neurotrophin signaling in driving abnormal vascular maturation in RA synovia. Neurotrophins, including nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and neurotrophin-3 (NT3), induce differentiation of synovial fibroblasts into mural cells - pericytes and vascular smooth muscle cells. Mechanistically, NOTCH3 signaling activates a cascade of neurotrophin signaling through transcriptional induction of NGFR, a co-receptor for NGF. In RA synovial tissue explants, stimulation with NGF, BDNF, or NT3 leads to a dramatic increase in maturation of synovial tissue vasculature. Conversely, pharmacologic inhibition with neurotrophin inhibitors drastically abolished maturation of vascularization in RA synovial explants. Notably, the FDA-approved tropomyosin receptor kinase (TRK) inhibitors larotrectinib and entrectinib effectively reverse synovial vascular maturation in human RA tissue explants.Our findings suggest that fibroblast-derived neurotrophin signaling is a critical pathway in sustaining mature blood vessels in RA synovia, and that neurotrophin inhibitors reverse abnormal vascular maturation in RA. One Sentence SummaryIn rheumatoid arthritis, fibroblast neurotrophin signaling drives abnormal vascular maturation by inducing differentiation of fibroblasts into vascular mural cells.

molecular biology↗

bioRxiv2026

DemuxHMM: Large-Scale Single-Cell Embryo Profiling via Recombination Barcoding

High-resolution developmental time-courses with single-cell RNA sequencing (scRNA-seq) increasingly target trajectory inference and other analyses in the study of development and disease [1-6]. These datasets are often generated by pooling individuals and inferring cell-to-individual mappings after sequencing, in a process called demultiplexing. Existing demultiplexing methods are limited in the number of timepoints they can support, due to either the need for individual-by-individual processing or reduced accuracy at large numbers of individuals. To address these limitations, we introduce a combined experimental and computational framework for creating large-scale, individual-resolved datasets. Our framework couples a simple breeding scheme that creates contiguous SNP patterns (recombination barcodes) with a recombination-aware demultiplexing method, DemuxHMM, that explicitly models this structure with a Hidden Markov Model (HMM). We demonstrate substantial performance and scalability gains from this combined approach on simulated data, highlighting its potential to enable the creation of large-scale single-cell time series.

bioinformatics↗

bioRxiv2026

Structural and Evolutionary Divergence of the RAG1/1L N-Terminal Zinc-Finger Domain

RAG1/2 catalyzes V(D)J recombination to assemble antigen receptor genes, yet the zinc-finger element at the RAG1 N terminus has remained structurally and functionally uncharacterized. Using NMR spectroscopy, we determined the structure of the RAG1 N-terminal zinc-finger domain (NZD; residues 89-223). NZD adopts a compact, zinc-dependent fold comprising four -helices organized into two interdigitated zinc-finger modules, ZFa and ZFb. Structural similarity searches across protein-structure databases revealed no solved homologs, establishing NZD as a previously undescribed zinc-finger fold. Comparative analyses show that NZD is broadly conserved across RAG1 and RAG1-like (RAG1L) proteins but has undergone lineage-specific remodeling, including the jawed vertebrate-specific acquisition of an additional -helix (H2). Guided by structure prediction, we identified two NZD-like types in Chapaev transposases; one is highly similar to the RAG1L NZD, supporting an evolutionary link between RAG1/RAG1L and Chapaev NZDs. Similarity between ZFa and the single-zinc ZAD further suggests that NZD originated from an ancestral single-zinc fold. Together, these findings provide a structural framework for mechanistic and evolutionary analyses of RAG1.

biochemistry↗

bioRxiv2025

Use-dependent learning biases the initial state but not the learning dynamics of implicit adaptation

Motor adaptation arises from multiple learning mechanisms, including use-dependent learning (UDL) driven by repetition and implicit error-based learning (EBL) driven by motor prediction errors. Although both mechanisms contribute implicitly to shaping movement execution, whether these two mechanisms interact remains unclear. The present study used an error-clamp (EC) task to isolate implicit EBL and directly examined whether UDL affects subsequent implicit adaptation. Participants first performed extensive single-target repetitive reaches to induce use-dependent biases and then immediately transitioned to an EC task. We manipulated whether the UDL bias was aligned with, opposed to, or neutral toward the to-be-adapted direction in three separate groups of participants. Results demonstrate that UDL robustly shifted the initial state of adaptation, but did not alter the learning dynamics: all groups showed comparable trial-by-trial adaptation and converged to similar asymptotic levels despite their different initial states. These findings support an independent rather than an interactive relationship between UDL and implicit EBL, highlighting that execution-level motor learning components are additive.

animal behavior and cognition↗

bioRxiv2025

Fibroblasts sense spatial proximity via an EGFR CREB5 axis to restore quiescent synovial lining in remission rheumatoid arthritis

Rheumatoid arthritis (RA) is a chronic inflammatory disease where the synovial lining membrane undergoes pathological changes resulting in joint destruction. In healthy joints, the synovial lining is essential for joint homeostasis, forming a selective barrier and secreting lubricating molecules, yet the mechanisms that restores homeostatic synovial lining during RA remission remains poorly understood. Here, we applied spatial transcriptomics to examine biopsies of RA patients in remission to identify a mechanism that orchestrates a phenotypic switch specifying synovial quiescent lining fibroblast differentiation. Spatial transcriptomics revealed a proximity-sensing program where at low cell-density, fibroblasts adopt proliferative and fibrotic transcriptional state characterized by expression of MKI67, COL1A2 and COL6A2, whereas at high cell-density, fibroblasts induce a quiescent lining fibroblast transcriptional program characterized by PRG4, CLU and PDPN. Mechanistically, fibroblasts sense spatial proximity through HB-EGF-EGFR signaling, which leads to phosphorylation of transcription factor CREB5. Perturbation of the EGFR-CREB5 axis abolishes fibroblast proximity-sensing and blocks synovial lining fibroblast differentiation. Conversely, EGFR activation by the ligand HB-EGF or pharmacologic activation of CREB5 is sufficient to induce synovial lining fibroblast differentiation. Together, our findings define a novel spatial proximity-sensing pathway underlying a return to homoeostatic fibroblast function during RA remission. By sensing their spatial proximity to neighboring fibroblasts, synovial fibroblasts translate these positional cues into signals that lead to restoration of normal, steady-state synovial lining membrane.

genomics↗

bioRxiv2025

Sequence-context-aware decoding enables robust reconstruction of protein dynamics from crystallographic B-factors

While X-ray crystallography remains the primary source of protein structures, the B-factors associated with these coordinates are frequently obscured by crystal packing and refinement artifacts that limit their utility for quantifying solutionstate dynamics. We address the disparity between static structural abundance and dynamic information scarcity by introducing the B-Factor Corrector. This fine-tuned protein language model redefines B-factor analysis as a sequenceto-dynamics translation task rather than a purely physical calculation. By leveraging deep contextual embeddings to decouple intrinsic flexibility from lattice constraints, our model recovers ground-truth conformational fluctuations derived from structural ensembles and achieves a Pearson correlation of 0.80. We further demonstrate the utility of these sequence-inferred dynamics in structural biology applications where the model identifies binding-competent conformations for flexible molecular docking, unmasks cryptic mechanical hinges essential for chaperone function, and guides the energetic refinement of antibody interfaces to resolve steric conflicts. These results suggest that static crystallographic data can be effectively repurposed to decode accurate dynamic patterns latent within the Protein Data Bank.

biophysics↗

bioRxiv2025

Spatial Transcriptomics Identify T Cell-Driven Mechanisms of Kidney Damage in Immune Checkpoint Inhibitor-Associated Acute Interstitial Nephritis

IntroductionImmune checkpoint inhibitor-associated acute interstitial nephritis (ICI-AIN) is the most common finding on histopathology among patients with ICI-associated acute kidney injury (ICI-AKI). Patients with ICI-AIN often have T cell-dominant infiltration of the kidney and high tissue levels of CXCR3 ligands like CXCL9, 10, and 11; however, the mechanisms of inflammation in ICI-AIN are not well-understood. MethodsWe applied a sub-cellular spatial transcriptomics platform (Xenium Prime 5K) to compare the cellular composition of kidney biopsy tissue from patients with ICI-AIN with ICI-treated patients with acute tubular necrosis (ICI-ATN). ResultsAcross 8 kidney biopsy specimens (4 with ICI-AIN, 4 with ICI-ATN), we analyzed 332,000 cells, comprising kidney parenchymal cells and infiltrating immune cells. Using a spatially-aware cellular neighborhood-based classification, we identified cellular niches corresponding to each part of the nephron, in addition to unique fibrotic and inflammatory niches. Gene pathway analysis identified interferon-gamma (IFN-{gamma})/STAT1 signaling as strongly increased in ICI-AIN compared to ICI-ATN. While all inflammatory niches were overrepresented in ICI-AIN, CD8+ T cell infiltration and proinflammatory myeloid cells were the dominant immune niches. Spatial niche crosstalk analysis revealed that CD8+ T cell-derived IFN-{gamma} likely induced a proinflammatory program in myeloid cells, with increased production of CXCL9, 10, and 11. Furthermore, IFN-{gamma} signaling in ICI-AIN was associated with reduced oxidative phosphorylation in kidney tubular niches. ConclusionsSpatial transcriptomics reveal novel insights into key differences in the pathophysiology of ICI-AIN versus ICI-ATN. IFN-{gamma}-producing CD8+ T cells are likely key drivers of ICI-AIN and should be investigated as future therapeutic targets. Translational StatementSpatial transcriptomics may provide insight into differences in the cellular and spatial composition of immune checkpoint inhibitor-associated acute interstitial nephritis (ICI-AIN) and ICI-associated acute tubular necrosis (ICI-ATN). Using the novel Xenium 5K platform, we demonstrate that IFN-{gamma}-producing CD8+ T cells are central to the pathogenesis of ICI-AIN, and that CD8+ T cell-derived IFN-{gamma} likely induces a proinflammatory state in myeloid cells, with increased tissue production of CXCL9, 10, and 11. We then show that a reduction in oxidative phosphorylation and an IFN-{gamma}-driven influx of proinflammatory myeloid cells may further drive kidney damage in ICI-AIN. Many of these pathways represent potential druggable targets, and our findings may therefore inform future therapeutic approaches for ICI-AIN.

immunology↗

bioRxiv2025

The ontogeny of myeloid-stromal synovial tissue niches in rheumatoid arthritis.

Recent single-cell multi-omic and spatial analyses of synovial biopsies have transformed our understanding of myeloid cell-driven mechanisms underlying human joint pathology and tissue homeostasis in Rheumatoid arthritis (RA). However, the developmental trajectories of synovial tissue macrophage (STM) subsets in humans remain poorly understood, due in part to the lack of models that faithfully replicate synovial tissue niches. This hinders the exploration of the therapeutic potential of targeting specific synovial myeloid cell clusters. Using multi-omics analyses of synovial tissue from an allogeneic bone marrow transplant recipient, we show that joint-specific tissue-resident STM subsets, including both health- and disease-associated clusters, can derive from peripheral blood monocytes. Analysis of embryonic synovial joints revealed that macrophage localization and maturation in the joints are preceded by local stromal niche specialisation, indicating that synovial fibroblasts (FLS) provide tissue-specific instructive cues to STM precursors. To elucidate human STM developmental trajectories, we established a SNP-based fate-tracking human synovial organoid system by embedding distinct blood-derived myeloid precursors, together with FLS clusters from RA synovial biopsies and endothelial cells, into 3D structures. These organoids reproduced key synovial tissue features, including lining and sublining architecture and stromal-myeloid cell cluster composition. Importantly, they supported differentiation of all resident STM subsets: homeostatic lining TREM2pos macrophages, their pathogenic TREM2lowSPP1pos counterparts that characterize the RA hyperplastic lining, and both homeostatic and RA-associated perivascular LYVE1pos STM clusters, all traced to monocytic precursors. In summary, we show that development of STM subsets is driven by fibroblast-conditioned spatial niches. We have established a novel, tractable ex vivo platform to dissect the niche-specific cues driving homeostatic versus pathogenic phenotypic clusters. One Sentence SummaryHuman tissue-resident STMs develop from monocytes under the guidance of cues from FLS within discrete spatial locations.

immunology↗

bioRxiv2025

From Signal to Symphony: Exploring 2D Sequence Representations for Protein Function Prediction

Predicting protein function from its primary sequence is a fundamental challenge in computational biology. While deep learning has excelled, the optimal representation of sequence data remains an open question. This study explores protein sonification--the conversion of amino acid sequences into 2D spectrograms--as a representation for this task. To facilitate this investigation, we developed a benchmark dataset of 18,000 sequences spanning 12 functionally diverse protein classes. Our systematic evaluation suggests that the structural transformation from a 1D sequence to a 2D spectrogram may be a key contributor to the models predictive performance. This observation is supported by ablation studies where models using either purely visual or acoustic features from the spectrogram demonstrated effective standalone performance, suggesting that the representation itself is a key source of this capability. For instance, a model using a sonification map without explicit biophysical meaning achieved 81.08% accuracy, while our biophysically-informed model reached 84.00%, indicating that such domain knowledge may offer a modest performance benefit. When trained from scratch on our dataset, our fusion model achieved performance comparable to or slightly exceeding that of standard transformer architectures like ESM-2 and ProtBERT, suggesting its potential for data efficiency in this specific context. The models potential for generalizability was further supported by its performance on the external CARE enzyme classification benchmark, where it achieved 90.44% accuracy. Finally, as a proof-of-concept, we explore the utility of our encoding to guide a diffusion model in generating novel GFP variants, which were assessed for structural viability using computational methods. Our work provides evidence suggesting that the utility of sonification in this context may stem largely from its representational structure, offering a perspective on feature engineering for biological sequences.

bioinformatics↗

bioRxiv2025

A Spatially Coordinated Keratinocyte-Fibroblast Circuit Recruits MMP9+ Myeloid Cells to Drive IFN-I-Driven Inflammation in Photosensitive Autoimmunity

Photosensitivity is a hallmark of cutaneous lupus erythematosus (CLE) and dermatomyositis (DM), yet the mechanisms linking ultraviolet B (UVB) exposure to tissue-specific autoimmunity remain incompletely defined. Here, we use an integrative human-based approach, including single-cell RNA sequencing, spatial transcriptomics (seqFISH+), in vivo UVB provocation, and in vitro modeling, to uncover a spatially coordinated inflammatory circuit that underlies interferon-I (IFN-I)-amplified skin pathology. We identify MMP9+ CD14+ myeloid cells as central effectors of photosensitivity in both CLE and DM. These cells are markedly expanded in lesional skin, serve as the dominant source of IFN-{beta}, and colocalize with cytotoxic CD4+ T cells at the dermal-epidermal junction. Spatial transcriptomics further reveals a keratinocyte-fibroblast-myeloid axis, wherein keratinocytes activate discrete subsets of pro-inflammatory fibroblasts in the superficial dermis to produce monocyte-attracting chemokines, including CCL2, CCL19, CCL7, CCL8, and CXCL12, directing MMP9+ CD14+ cell recruitment toward the interface. In our in-vitro model, IFN-I-primed basal keratinocytes undergo heightened UVB-induced cell death and release membrane-associated cytokines such as TNF-, IL-1, which activate monocyte-derived dendritic cells (moDCs) and induce transcriptional programs mirroring those of MMP9+ CD14+ cells in vivo. In vivo, UVB irradiation of non-lesional DM skin, but not healthy controls, elicits rapid infiltration of these myeloid cells, confirming their disease-specific responsiveness to UVB. Finally, in a proof-of-concept clinical study, treatment with anifrolumab (anti-IFN-I receptor) blocked UVB-induced MMP9+ CD14+ infiltration and attenuated photosensitivity in CLE. Together, these findings define a multicellular inflammatory cascade linking keratinocyte injury, fibroblast chemotactic programming, and myeloid effector function in IFN-I-driven skin autoimmunity and nominate MMP9+ CD14+ cells as actionable targets in photosensitive dermatoses. Photosensitivity is central to cutaneous lupus erythematosus (CLE) and dermatomyositis (DM), but the mechanisms linking UVB exposure to tissue-specific autoimmunity are poorly defined. Using single-cell RNA sequencing, spatial transcriptomics, UVB provocation, and in vitro modeling, we identify MMP9+ CD14+ myeloid cells as critical mediators of photosensitivity. These cells expand significantly in lesional skin, produce IFN-{beta}, and colocalize with cytotoxic CD4+ T cells at the dermal-epidermal junction. Keratinocytes activate fibroblasts in the superficial dermis, prompting them to release chemokines (CCL2, CCL19, CCL7, CCL8, CXCL12) that recruit MMP9+ CD14+ cells. IFN-I-primed keratinocytes exposed to UVB release cytokines activating dendritic cells, mirroring in vivo responses. UVB irradiation of non-lesional DM skin rapidly recruits these myeloid cells. In a clinical proof-of-concept study, anti-IFN-I treatment with anifrolumab prevented UVB-induced myeloid infiltration and reduced photosensitivity. Thus, targeting MMP9+ CD14+ cells may offer therapeutic potential for managing photosensitive autoimmune skin conditions.

immunology↗

bioRxiv2025

Modeling Post-Gastrula Development via Bidirectional Pluripotent Stem Cells

The absence of stem cells capable of efficiently generating both trophoblast and epiblast lineages has hindered precise recapitulation of embryonic development. Through high-content chemical screening, we established an AL medium to generate mouse and human bidirectional pluripotent stem cells (BPSCs) characterized by concurrent OCT4/CDX2 expression. Mouse BPSCs demonstrated high plastic differentiation into trophoblast, epiblast and primitive endoderm lineages in vitro within 48 hours without exogenous induction factors and efficiently contributed to the embryo and extraembryonic tissues in vivo. Mechanistically, hyperactivation of the Wnt signaling pathway breaks the early lineage differentiation barrier by initiating a Lef1-dependent bypass. Remarkably, BPSCs can efficiently generated E8.5 embryoids that completed gastrulation and displayed advanced features such as brain development, a closed neural tube, a beating heart, somite formation, and primordial germ cells. These findings highlight BPSCs as a powerful tool for investigating early lineage specification and post-gastrulation embryonic development, with potential applications across multiple species.

developmental biology↗