Biology subjects

Coskun, A. F.

Publications and source records attributed to Coskun, A. F..

11 recordsLinked to original sources

bioRxiv2026

3D Spatial Interactomics Maps the Dynamics of NF-κB Multiprotein Signalosomes in Single Cells

NF{kappa}B signaling drives inflammatory responses by rapidly assembling membrane-proximal multiprotein supercomplexes, yet how these assemblies are organized in space and time within the 3D interior of a single cell has remained uncharacterized. We addressed this by profiling endogenous NF{kappa}B protein-protein interactions with an intelligent sequential proximity ligation assay (iseqPLA), read out by spinning disk confocal microscopy and 3D reconstruction. Each detected protein-protein proximity event is represented by a rolling-circle amplification product, and we treat clusters of co-localized puncta as a measure of supercomplex spatial organization. Across NIH-3T3 mouse fibroblasts, cystic fibrosis (CF) patient-derived macrophage co-cultures with IMR-90 human fibroblasts, and an independent set of healthy- and CF-donor monocyte-fibroblast co-cultures profiled by 3D iseqPLA, we tracked supercomplex dissociation, p65 nuclear translocation, and negative-feedback engagement across cytokine time courses. Three findings emerge: 3D volumetric quantification reduces the variance in nuclear-to-cytoplasmic ratio measurements relative to 2D projections, the choice of extracellular matrix coating shapes the fraction of NF{kappa}B-responsive cells, and CF airway-conditioned macrophages amplify paracrine NF{kappa}B signaling in neighboring fibroblasts in a CF model. A single-cell Generative Pretrained Transformer (scGPT) foundation model, fine-tuned on curated transcriptomic datasets, further places our NF{kappa}B gene panel within an inflammation-relevant feature space. Together, these results establish a 3D spatial interactomics workflow for dissecting supercomplex dynamics in health and disease.

bioengineering↗

bioRxiv2026

Morphomechanically-Informed Spatial Curvature Sequencing in Prostate Cancer

Morphological changes in prostate glands, assessed by Gleason grading, remain the gold standard for diagnosing prostate cancer, yet molecular biomarkers associated with gland shape are not well understood. Here, we introduce CurvSeq, a mechanomorphology-informed framework for spatial sequencing data, and CurvSee, its complementary version for proteomic and imaging datasets. These methods integrate gland boundary curvature, pocket architecture, microenvironmental composition, and molecular profiles to study morphomechanical relationships in prostate adenocarcinoma. Using five independent spatial transcriptomic and multiplexed imaging datasets, we segmented individual prostate glands, extracted gland contours, quantified local curvature and pocket-like concavities, and projected these features onto spatially resolved gene and protein measurements. In Xenium data, CurvSeq distinguished benign and GG1 glands, identifying cancer-associated genes such as PCA3 and AMACR in GG1 glands and basal, basement membrane, and mechanotransduction-associated programs in benign glands. In Visium data, a diffusion-based morphomechanical score ordered benign glands by area, circularity, pocket number, smooth muscle abundance, immune-cell proximity, and remodeling-associated genes including MMP7. In GG4 glands, CurvSeq identified neuroendocrine-like boundary regions associated with MMP7 expression, COL1A1-rich adjacent stroma, and immune-cell accumulation. Finally, CurvSee extended this framework to multiplexed protein imaging, where combined morphology and protein-expression features distinguished Gleason-associated gland states. Together, CurvSeq and CurvSee provide a quantitative framework for linking gland architecture, local microenvironment, and molecular state, showing that prostate gland morphology can be integrated with spatial omics to identify morphomechanical niches associated with cancer progression.

systems biology↗

bioRxiv2026

Interleukin-6 restricts pre-thymic T cell lineage commitment of progenitors driving loss of SIV control

Effective T cell reconstitution in people living with HIV is central to durable immune control and cure strategies. Sustained thymic output underpins T cell recovery and requires continuous seeding by T cell-committed progenitors originating in the bone marrow (BM). Using the SIV/rhesus macaque model, we identified a thymus-seeding progenitor (TSP; CD4-CD8-CD34CD38-CD7) in BM declining rapidly following SIV infection. This loss closely associated with reduction in T cell lineage committed differentiation of BM-derived hematopoietic stem and progenitor cells (HSPCs). Importantly, both the decline in TSPs and the impairment of pre-thymic T cell potential were strongly associated with early loss of viral control, independent of peripheral T cell dynamics. Plasma interleukin-6 (IL-6) levels robustly predicted the magnitude of TSP loss and the restriction of T cell-biased HSPC differentiation. Integrated transcriptomic and proteomic analyses revealed inflammatory imprinting of HSPCs characterized by activation of the IL-6-JAK-STAT axis, inflammasome engagement, and coordinated suppression of key T cell specification factors, including RUNX1, FYN, and ZAP70. In a nonanimal model of thymopoiesis, IL-6 exposure of rhesus macaque and human HSPCs inhibited their transition from DN1 (CD38-) to DN2 (CD38) TSP states, indicating an early block in T cell lineage commitment. Conversely, ex vivo IL-6 receptor blockade restored thymocyte differentiation to levels comparable to untreated controls. Collectively, these findings demonstrate that pathogenic inflammation restricts pre-thymic T cell development early after infection, directly contributing to loss of viral control. These findings have important implications for understanding the mediators of anti-viral T cell immunity and HIV cure.

immunology↗

bioRxiv2026

CELLWHISPER: INFERENCE OF DIRECT CELL-CELL COMMUNICATION FROM SPATIAL TRANSCRIPTOMICS

AbstractDirect, contact-dependent cell-cell communication shapes tissue physiology and disease. Spatial transcriptomics offers a unique opportunity to infer such communication mechanisms, but the few existing tools for such inference are prone to high error rates due to spatial structure in distribution of cell types and gene expression. We present CellWHISPER, a statistical framework for inferring contact-mediated cell-cell communication, including gap junctions and ligand-receptor mechanisms, from single-cell resolution spatial data. CellWHISPER achieves strict error control via a specialized hypothesis-testing procedure that accounts for the confounding effects of cell- type-specific expression and spatial organization. Its use of an exact closed-form significance score obviates expensive permutation testing and enables scalable analysis of large tissues an extensive signaling compendia. To uncover salient communication patterns, CellWHISPER also includes a latent variable model that distills recurrent patterns into maps of mutual preferences among cell types and signaling genes. Applied to mouse brain STEREO-seq and Xenium datasets, CellWHISPER identifies cell type-specific gap-junction coupling and yields a comprehensive connexin-interaction map. We experimentally confirmed Connexin 43-mediated interfaces among astrocytes, microglia and endothelial cells, which were among the highest- ranked predictions; in contrast, existing approaches fail to prioritize these confirmed interactions. Differential analysis of an Alzheimers disease model versus wild-type identifies preserved astroglial coupling but increased microglia-associated gap-junction communication. In summary, CellWHISPER disentangles direct cell-cell communication from structural proximity, enabling statistically robust and computationally scalable inference from spatial transcriptomics, leading to the discovery of tissue- and disease-associated contact-mediated signaling programs.

bioinformatics↗

bioRxiv2025

MetaboFM: A Foundation Model for Spatial Metabolomics

Mass spectrometry imaging (MSI) provides molecularly resolved maps of metabolites and lipids across tissues, yet the lack of large-scale, unified representation learning frameworks limits its potential for generalization and downstream analysis. Here, we introduce MetaboFM, a foundation model for spatial metabolomics that consolidates thousands of public MSI datasets into standardized spatial-spectral tensors and extracts transferable embeddings using pretrained Vision Transformers. We curated and standardized around 4000 publicly available MSI datasets from the METASPACE repository, spanning multiple organisms, tissue types, ionization modes, and instruments. Across six metadata prediction tasks--encompassing organism, ionization polarity, tissue type, condition, analyzer type, and ionization source embeddings from pretrained MetaboFM encoders achieved a mean macro-F1 of 0.74 and accuracy of 0.80 with linear probes, demonstrating substantially higher discriminative power than classical principal component analysis (PCA) or randomly initialized baselines by over 20 percentage points. To interpret the learned representations, we mapped embedding directions back to the m/z domain, revealing distinct spectral regions that drive class separation across tissues, conditions, and ionization sources. A multimodal visual question answering (VQA) extension further links MSI embeddings with natural-language queries through a cross-attention fusion module, attaining an average macro-F1 of 0.61 {+/-} 0.05 across tasks. Finally, an interactive Gradio interface enables users to visualize MSI patches and query sample metadata in free-form language. Together, MetaboFM establishes a scalable foundation model paradigm for MSI, unifying representation learning, spectral interpretability, and multimodal interaction within a single framework for spatial metabolomics.

bioinformatics↗

bioRxiv2024

Super Resolved Single-Cell Spatial Metabolomics from Multimodal Mass Spectrometry Imaging guided by Imaging Mass Cytometry

Mass spectrometry imaging (MSI) is a powerful technique for spatially resolved analysis of metabolites and other biomolecules within biological tissues. However, the inherent low spatial resolution of MSI often limits its ability to provide detailed cellular-level information. To address this limitation, we propose a guided super-resolution (GSR) approach that leverages high-resolution Imaging Mass Cytometry (IMC) images to enhance the spatial resolution of low-resolution MSI data. By using these detailed IMC images as guides, we improve the resolution of MSI images, creating high-resolution metabolite maps. This enhancement facilitates more precise analysis of cellular structures and tissue architectures, providing deeper insights into super-resolved spatial metabolomics at the single-cell level.

systems biology↗

bioRxiv2024

Multi-Niche Human Bone Marrow On-A-Chip for Studying the Interactions of Adoptive CAR-T Cell Therapies with Multiple Myeloma

Multiple myeloma (MM), a cancer of bone marrow plasma cells, is the second-most common hematological malignancy. However, despite immunotherapies like chimeric antigen receptor (CAR)-T cells, relapse is nearly universal. The bone marrow (BM) microenvironment influences how MM cells survive, proliferate, and resist treatment. Yet, it is unclear which BM niches give rise to MM pathophysiology. Here, we present a 3D microvascularized culture system, which models the endosteal and perivascular bone marrow niches, allowing us to study MM-stroma interactions in the BM niche and model responses to therapeutic CAR-T cells. We demonstrated the prolonged survival of cell line-based and patient-derived multiple myeloma cells within our in vitro system and successfully flowed in donor-matched CAR-T cells. We then measured T cell survival, differentiation, and cytotoxicity against MM cells using a variety of analysis techniques. Our MM-on-a-chip system could elucidate the role of the BM microenvironment in MM survival and therapeutic evasion and inform the rational design of next-generation therapeutics. TEASERA multiple myeloma model can study why the disease is still challenging to treat despite options that work well in other cancers.

cancer biology↗

bioRxiv2024

Spatial Morphoproteomic Features Predict Uniqueness of Immune Microarchitectures and Responses in Lymphoid Follicles

Multiplex imaging technologies allow the characterization of single cells in their cellular environments. Understanding the organization of single cells within their microenvironment and quantifying disease-status related biomarkers is essential for multiplex datasets. Here we proposed SNOWFLAKE, a graph neural network framework pipeline for the prediction of disease-status from combined multiplex cell expression and morphology in human B-cell follicles. We applied SNOWFLAKE to a multiplex dataset related to COVID-19 infection in humans and showed better predictive power of the SNOWFLAKE pipeline compared to other machine learning and deep learning methods. Moreover, we combined morphological features inside graph edge features to utilize attribution methods for extracting disease-relevant motifs from single-cell spatial graphs. The underlying subgraphs were further analyzed and associated with disease status across the dataset. We showed that SNOWFLAKE successfully extracted significant low dimensional embedding from subgraphs with a clear separation between disease status and helped characterize unique cellular interactions in the subgraphs. SNOWFLAKE is a generalizable pipeline for the analysis of multiplex imaging data modality by extracting disease-relevant subgraphs guided by graph-level prediction.

bioinformatics↗

bioRxiv2023

Microbial Changes occurring during oronasal fistula wound healing

The oral microbiome is a complex community that matures with dental development while oral health is also a recognized risk factor for systemic disease. Despite the oral cavity having a substantial microbial burden, healing of superficial oral wounds occurs quickly and with little scarring. By contrast, creation of an oro-nasal fistula (ONF), often occurring after surgery to correct a cleft palate, is a significant wound healing challenge that is further complicated by a connection of the oral and nasal microbiome. In this study, we characterized the changes in the oral microbiome of mice following a freshly inflicted wound in the oral palate that results in an open and unhealed ONF. Creation of an ONF in mice significantly lowered oral microbiome alpha diversity, with concurrent blooms of Enterococcus faecalis, Staphylococcus lentus, and Staphylococcus xylosus in the oral cavity. Treatment of mice with oral antibiotics one week prior to ONF infliction resulted in a reduction in the alpha diversity, prevented E. faecalis and S. lentus, and S. xylosus blooms, but did not impact ONF healing. Strikingly, delivery of the beneficial microbe Lactococcus lactis subsp. cremoris (LLC) to the wound bed of the freshly inflicted ONF via a PEG-MAL hydrogel vehicle resulted in rapid healing of the ONF. Healing of the ONF was associated with the maintenance of relatively high microbiome alpha diversity, and limited the abundance of E. faecalis and S. lentus, and S. xylosus in the oral cavity. These data demonstrate that a freshly inflicted ONF in the murine palate is associated with a dysbiotic oral microbiome state that may prevent ONF healing, and a bloom of opportunistic pathogens. The data also demonstrate that delivery of a specific beneficial microbe, LLC, to the ONF can boost wound healing, can restore and/or preserve oral microbiome diversity, and inhibit blooms of opportunistic pathogens.

pathology↗

bioRxiv2022

Subcellular spatially resolved gene neighborhood networks in single cells

Mesenchymal stem cell (MSC)-based therapies have offered promising treatments against several disorders. However, the clinical efficacy and consistency remain underdeveloped. Single-cell and bulk molecular analyses have provided considerable heterogeneity of MSCs due to origin, expansion, and microenvironment. Image-based cellular omics methods elucidate ultimate variability in stem cell colonies, otherwise masked by bulk omics approaches. Here, we present a spatially resolved Gene Neighborhood Network (spaGNN) method to produce transcriptional density maps and analyze neighboring RNA distributions in single human MSCs and chondrocytes cultured on 2D collagen-coated substrates. This proposed strategy provides cell classification based on subcellular spatial features and gene neighborhood networks. Machine learning-based clustering of resultant data yields subcellular density classes of 20-plex biomarkers containing diverse transcript and protein features. The spaGNN reveals tissue-source-specific MSC transcription and spatial distribution characteristics. Multiplexed spaGNN analysis allows for rapid examination of spatially resolved subcellular features and activities in a broad range of cells used in pre-clinical and clinical research.

systems biology↗

bioRxiv2019

Ion beam subcellular tomography

Multiplexed ion beam imaging (MIBI) has been previously used to profile multiple parameters in two dimensions in single cells within tissue slices. Here, a mathematical and technical framework for three-dimensional subcellular MIBI is presented. We term the approach ion beam tomography (IBT) wherein ion beam images are acquired iteratively across successive, multiple scans and later compiled into a 3D format. For IBT, cells were imaged at 0.2-4 pA ion current across 1,000 axial scans. Consecutive subsets of ion beam images were binned over 3 to 20 slices (above and below) to create a resolved image, wherein binning was incremented one slice at a time to yield an enhanced multi-depth data without loss of depth resolution. Algorithmic deconvolution, tailored for ion beams, was then applied to the transformed ion image series using a hybrid deblurring algorithm and an ion beam current-dependent point-spread function. Three-dimensional processing was implemented by segmentation, mesh, molecular neighborhoods, and association maps. In cultured cancer cells and tissues, IBT enabled accessible visualization of three-dimensional volumetric distributions of genomic regions, RNA transcripts, and protein factors with 65-nm lateral and 5-nm axial resolution. IBT also enabled label-free elemental mapping of cells, allowing "point of source" cellular component measurements not possible for most optical microscopy targets. Detailed multiparameter imaging of subcellular features at near macromolecular resolution should now be made possible by the IBT tools and reagents provided here to open novel venues for interrogating subcellular biology.

systems biology↗