Biology subjects

Wirtz, D.

Publications and source records attributed to Wirtz, D..

At least 19 recordsLinked to original sources

bioRxiv2026

Reprogramming CAR-T cell function through synthetic cytokine receptors suppresses metastatic progression of TNBC

Patient T cells modified with chimeric antigen receptors (CARs) have shown remarkable therapeutic success in previously incurable cancers. However, despite its success in hematological malignancies, CAR-T therapy has shown only modest efficacy in solid tumors. The medical field has focused on developing new molecular designs that enhance CAR-T persistence, prevent exhaustion in the tumor microenvironment (TME), and increase the specificity of the CAR-T cell to various tumor antigens. However, much less effort has been placed on enhancing CAR-T cells' invasion into and migration through the stromal matrix of the TME, where they can elicit tumor-directed cytotoxicity. Here, we engineered CAR-T cells with synthetic cytokine receptors targeting inflammatory cytokines produced by the cells themselves, which increase their speed through collagen I matrices. We show that this enhanced motility is driven by increased CAR-T cell translocation through confined microchannels via cellular deformation, together with elevated migratory speeds in proximity to neighboring cells, suggesting improved responsiveness to local cues. These migration-enhanced CAR-T cells significantly reduce both primary tumor size and distant organ metastasis in triple-negative breast cancer (TNBC) mouse models compared to standard CAR-T cells. These results support molecular design strategies that exploit mechanisms to improve cellular therapies for metastatic solid tumors.

bioengineering↗

bioRxiv2026

TiSAM: fusing spatial and histopathology foundation models for whole-slide image segmentation

Semantic segmentation of whole-slide images (WSIs) quantifies tissue composition and architecture, but existing models require extensive annotation and lose accuracy as the number of tissue classes increases or on heterogeneous, cross-scanner datasets. Here, we present the Tissue Segment Anything Model (TiSAM), which fuses SAM3 spatial features with UNI2-h pathology features within a multi-scale decoder. Because public benchmarks contain few tissue classes, we release MegaMouse, comprising 36,833 annotations of 52 tissue classes from 12 mouse organs, and an 11-class human pancreas dataset. TiSAM achieved the highest overall performance across four public datasets compared to existing convolutional neural network and vision transformer models, maintained high accuracy on MegaMouse, and maintained high accuracy in the human pancreas dataset when trained on as few as 5 annotated WSIs where previous integrations of SAM with pathology foundation models did not. TiSAM segmentations were also more consistent across eight slide scanners compared to other models. These results establish TiSAM as a general and highly adaptable framework for tissue segmentation, enabling studies containing many tissue classes, limited annotations, or images from multiple scanners.

bioinformatics↗

bioRxiv2026

VITAL-3D: Volumetric Single-Cell Quantification Reveals Microenvironment-Dependent Drug Responses in Breast Cancer

Preclinical drug evaluation relies heavily on two-dimensional (2D) monolayer assays, which fail to recapitulate the structural and functional complexity of the tumor microenvironment and may therefore misrepresent therapeutic efficacy. Here, we present VITAL (Volumetric Imaging-based Toxicity and Live Analysis), a high-throughput imaging platform that enables direct single-cell quantification of proliferation and cell death in both 2D and three-dimensional (3D) extracellular matrix (ECM) cultures using a 96-well format. By combining volumetric imaging with automated single-cell analysis, VITAL enables dynamic assessment of drug responses beyond conventional viability assays and EC measurements. Using breast cancer cell lines treated with anticancer agents, we systematically compared drug responses between 2D and 3D microenvironments. Although EC values were often comparable between culture formats, growth kinetics and concentrations required to induce complete growth arrest or net cell loss differed substantially in 3D cultures. In particular, drug concentrations required to induce net cell loss were consistently higher in 3D, revealing microenvironment-dependent survival responses that were not captured by EC alone. Furthermore, clinically expected subtype-specific responses, including tamoxifen sensitivity in ER-positive cells and olaparib sensitivity in BRCA1-mutant cells, were more accurately resolved under 3D culture conditions and extended treatment durations. Together, these findings demonstrate that growth-based, single-cell quantification provides a more comprehensive assessment of therapeutic efficacy than conventional endpoint measurements and establish VITAL as a scalable platform for physiologically relevant preclinical drug screening.

cancer biology↗

bioRxiv2026

Distinct senescent β-cell senotypes differentially drive islet aging and dysfunction

Biological aging greatly impacts the bodys ability to handle glucose, and represents a major risk factor the development and progression of type 2 diabetes (T2D). Nonetheless, despite advances in cellular senescence research and the development of new senolytic therapies, the heterogeneity of cellular senescence in the human endocrine pancreas, as well as its roles in normal aging, remains to be elucidated at the single-cell level. Here, we performed single-cell-resolved spatial proteomics and transcriptomics on intact pancreas from 26 donors (ages 20-80) and multiplexed single-cell RNA sequencing and functional assays on dispersed islets from 14 donors (ages 34-69). We identify two discrete SnC subpopulations distinguished by relative expression of CDKN1A and CDKN2A. CDKN1A senescent cells (SnCs) exhibit loss of {beta}-cell identity, impaired insulin secretion, and a proinflammatory SASP associated with increased islet immune infiltration. In contrast, CDKN2A SnCs retain transcriptional identity and functional competence, with lower inflammatory signaling. Together, these findings identify heterogeneous and functionally divergent senotypes in the human pancreas, distinguishing an adaptive (CDKN2A) from a maladaptive (CDKN1A) senescence program, thus providing a mechanism-guided framework for senescence-targeted therapies in T2D.

cell biology↗

bioRxiv2026

Spatial Regulation of CAR Signaling Enables Logic-Gated Activity

Chimeric antigen receptors (CARs) can induce T cells to kill cancer cells but also to kill normal cells that express the same antigens1. Designing CARs to recognize combinations of antigens, via Boolean logic, can simultaneously expand the scope of targetable antigens and make CAR T cells more specific to cancer2. For example, one antigen may be expressed on cancer cells and normal bone marrow cells, while a second antigen may be present on the same cancer cells but only in normal lungs. If recognition of both antigens is required for T cell activation, only the cancer cells will be killed. Creating such AND-gated CAR T cells has been challenging given the need to engineer non-natural signaling mechanisms that integrate two ligand binding events into a single T cell activation stimulus3-6. Here, we design a fundamentally new AND-gated receptor called Multi-ANtigen Triggered Immune Synapse (MANTIS), which leverages differences in extracellular receptor dimensions to regulate CAR signaling. MANTIS initially prevents CAR activity by steric blocking with a bulky extracellular domain. Upon engagement of the first antigen, MANTIS sheds this blocking domain, releasing a free CAR that can bind a second antigen and activate the T cell in an AND-gated manner. This work demonstrates how differences in extracellular receptor size can be leveraged to spatially regulate intracellular signaling pathways in response to antigen patterns, paving the way for new applications in synthetic biology and cell engineering. One sentence summaryDifferences in extracellular size can be leveraged to regulate CAR T cell activity for precise recognition of antigen patterns.

synthetic biology↗

bioRxiv2026

Virtual multiplex staining of the pancreatic islets across type 1 diabetes progression using a Schroedinger bridge

Classical hematoxylin and eosin (H&E) staining enables review of tissue morphology but lacks information regarding the molecular state of cells. Immunohistochemical (IHC) techniques label specific proteins in tissue, allowing differentiation of relevant structures that may go undetectable in H&E. However, the IHC process is complex, expensive, and time-consuming, especially for multiplex IHC (mIHC) limiting its use in large cohorts. Stain conversion of H&E to IHC using generative artificial intelligence models such as generative adversarial networks (GANs) represent one solution to this problem. However, GANs are unstable during out of distribution sampling and are prone to hallucinations or mode collapse, limiting their accuracy in challenging image conversion tasks. To address this, the field has recently turned to diffusion models. Here, we introduce Schrodinger-bridge for Multiplex ImmunoLabel Estimation (SMILE). Unlike conventional diffusion models that map from source to target through an intermediate Gaussian noise, Schrodinger-bridge diffusion models skip this step and have been shown to better preserve structures during image translation. To test the performance of SMILE, we generated a large cohort of high-fidelity H&E-mIHC image pairs from pancreatic organ donors, targeting insulin, glucagon, and CD3. Our dataset well-sampled across type-1 diabetes status, pancreas anatomical location, age, and sex. Using this cohort, we demonstrate the superiority of SMILE compared to GANs via a comprehensive evaluation framework incorporating texture, distribution, and antibody-specific metrics, as well as blinded pathologist reviews. We further confirmed the ability of SMILE to generate accurate mIHC images from H&Es generated at an external site, to perform whole slide image conversion, and to generate realistic three-dimensional maps of the pancreatic islets in non-diabetic, auto-antibody positive, and type-1 diabetic donor tissue. Finally, we performed stain conversion of paired H&E to HER2 and Ki67 images in breast cancer, confirming the superiority of SMILE in diverse stain conversion applications. Collectively, this framework provides a scalable pipeline for high-throughput proteomic inference from archival H&Es, providing transformative potential for pancreatic research and digital pathology.

bioinformatics↗

bioRxiv2026

Whole organism 3D mapping reveals universal branching topology and biophysical optimization governs vascular and nervous system development

The vascular and nervous systems are transport networks essential for life, yet whether universal geometric and topological principles govern their formation remains unclear. The organism-wide molecular and biophysical coordination required to build and maintain these networks during embryonic development has inspired decades of theoretical work, offering predictions about their expected organization. However, the lack of complete three-dimensional (3D) data has limited validation to isolated structures, leaving whole-organism networks unexplored. Here, we developed a computational pipeline for whole-organism 3D imaging to reconstruct the complete vascular and nervous systems of rhesus macaque, mouse, and turtle embryos. Our analysis reveals that both networks share structural principles, including binary branching and scale-invariant bifurcation geometry, maintained across species and throughout development. Yet, from these shared rules emerge fundamentally different architectures. Vasculature exhibits fractal topology with a fractal dimension [~]3, forming space-filling trees that prioritize proximity to every cell in the body. Nervous system networks exhibit a fractal dimension [~]2, forming sheet-like arbors that prioritize electrical signal transmission. This architectural divergence originates from distinct biophysical constraints operating in bifurcations, where vascular junctions minimize energy expenditure while conserving fluid flow and nerve junctions maximize conduction velocity while conserving electrical current. These local optimization rules, iterated across generations, construct organism-wide networks governed by distinct physical constraints, revealing how evolution generated different solutions for fluid versus electrical transport.

developmental biology↗

bioRxiv2026

Deep Learning Enabled 3D Multi-Omic Analysis Reveals Molecular Signatures of Heterogeneous Response to Chemotherapy in Pancreatic Cancer

Resistance to systemic therapy is a major unmet challenge in pancreatic cancer. To identify potential mechanisms of resistance, we developed a novel 3D pipeline in clinical samples that uses deep learning to classify sensitive and persistent tumor cell populations based on morphological features, enabling subsequent molecular characterization of intratumoral heterogeneity. We applied this automated 3D pipeline to a cohort of human pancreatic cancer samples treated with neoadjuvant chemotherapy, identifying heterogeneity in response to therapy both between and within tumors. Application of spatial proteomics to these sensitive and persistent regions identified enhanced epithelial-to-mesenchymal transition and non-classical cell states in persistent cells, confirming our morphological classification. Integration of spatial transcriptomics in multiple pancreatic cancer cohorts associated fibroblast-cancer crosstalk via syndecans with resistance to cytotoxic therapy. Our validated 3D multi-omic pipeline is now poised for application to clinical trials, enabling discovery of resistance mechanisms and design of new therapeutic combinations to circumvent resistance. Statement of significanceWe developed a novel 3D multi-omic pipeline to identify mechanisms of resistance to chemotherapy in clinical samples. This approach associated fibroblast-cancer crosstalk via syndecans with resistance to cytotoxic therapy and is poised for broader application in neoadjuvant clinical trials.

cancer biology↗

bioRxiv2026

3D, multi-omic imaging reveals molecular biomarkers of the pre-metastatic niche in lung cancer

The recurrence rate following complete surgical resection of primary non-small cell lung cancer is as high as 55%, yet no approach currently exists to evaluate the risk of local recurrence. The premetastatic paradigm is the recognition that metastasis is preceded by reprogramming naive tissues to prime a microenvironment for tumor cell survival and subsequent reactivation. Identification of biomarkers of the pre-metastatic niche would allow us to evaluate a patients risk of local relapse in the normal lung parenchyma surrounding the resected tumor. We designed a workflow incorporating in vivo modelling, radiology, and deep learning-guided three-dimensional (3D) imaging, spatial proteomics, and transcriptomics to identify previously unreported signals associated with the early transformation of the lung parenchyma announcing regional metastasis. We curated biorepository spanning timepoints before and after resection of primary Lewis Lung Carcinoma (LLC) tumors. Using radiology and cellular resolution 3D histology, we calculated the number and distribution of metastases in mouse lungs and developed an algorithm to guide placement of spatial proteomics and transcriptomics to regions containing early micro-metastases and the pre-metastatic microenvironment. Molecular and tissue features associated with presence, size, and location of metastases guided the identification of both myeloid (F4/80) and senescent (p16/p21) cell signatures in the premetastatic and metastatic environments. Finally, multiparametric flow cytometry of metastatic lungs in a senescence reporter GEMM (tdTomato-p16 INKA mice) resolved senescent cells including alveolar macrophages as the cellular phenotypes associated with these early premetastatic signatures. Altogether, this work highlights a novel AI-assisted approach for detection of biomarkers of tissue remodeling during lung cancer invasion.

bioengineering↗

bioRxiv2025

Senescence-Linked Fibrosis in the Aging Human Ovary Revealed by p16-Based Histological Profiling and Spatial Transcriptomics

Cellular senescence is implicated as a driver of ovarian aging, but senescent cells in the human postmenopausal ovary remain poorly defined. Using spatially resolved p16INK4a protein expression, a canonical senescence marker, we identified and mapped senescent cells in postmenopausal ovaries. We integrated p16 immunohistochemistry, multiplexed immunofluorescence, spatial transcriptomics, and AI-guided digital pathology to map senescent microenvironments. p16-positive cells formed discrete stromal, vascular, and cyst-associated clusters that increased with age and were enriched for macrophages and myofibroblast-like cells. Wholetranscriptome profiling of 92 spatial regions uncovered a 32-gene p16-associated signature, BuckSenOvary, that distinguished p16-positive regions across cortex and medulla. BuckSenOvary is characterized by suppression of cell-cycle regulators and activation of inflammatory and extracellular-matrix remodelling genes. AI-based collagen matrix analysis confirmed that p16-positive regions exhibit more architecturally complex collagen, demonstrating that focal senescent microenvironments are fibro-inflammatory. These findings position senescent ovarian niches as therapeutic targets to preserve ovarian function.

cell biology↗

bioRxiv2025

Precision targeting of autoreactive B cells in systemic lupus erythematosus using anti-9G4 idiotope synthetic immune receptor T cells

Chimeric antigen receptor (CAR)-T cell therapies can induce drug-free remission in systemic lupus erythematosus (SLE), but indiscriminate B cell targeting causes immunosuppression, unnecessary infections, and cytokine toxicities that preclude widespread use. Here, we overcome this by targeting the 9G4 idiotope, a shared structural feature of pathogenic B cell receptors encoded by the IGHV4-34 gene. We engineered anti-9G4 CAR-T cells and chimeric TCR-T cells to selectively eliminate autoreactive B cells while preserving protective immunity. Both platforms eradicated autoreactive B cells and autoantibodies in vitro and in vivo, spared normal B cells, and markedly reduced cytokine release compared to conventional CAR-T cells. This precision extended to cold agglutinin disease and lymphoma. These findings establish a framework for IGHV idiotope-directed cellular therapies for treating autoimmune and neoplastic diseases while preserving immune competence.

immunology↗

bioRxiv2025

Optimization of velocity receptor transduction in CAR T cells

Limited infiltration capacity significantly limits the effectiveness of chimeric antigen receptor (CAR) T cells for solid tumors. We have recently developed a large family of highly modular synthetic cytokine receptors termed velocity receptors (VRs), capable of binding key inflammatory cytokines, such as IL5, IL8, and TNF, which drive CAR T cells into an elevated motility state. These new CAR T cells sense and amplify these autocrine secreted cytokines, thereby maintaining a self-propelled, high migratory state, facilitating penetration into dense tumor cores. In this study, we systematically evaluated key factors influencing VR transduction in order to improve their stable integration and expression. We established a dual-fluorescence reporter system to allow simultaneous monitoring of both VR and CAR constructs, and while evaluating modifications to the vector construct and generating standardized infectious unit (IFU) curves under various conditions. Our results demonstrate that the attempt to reduce overall vector size by eliminating non coding sections upstream of the central polypurine tract (cPPT) do not yield better transduction efficiency, though it is unclear if the effect is due to viral production or integration impairment. We also observed a log-linear relationship between viral dose and transduction efficiency for a subset of VRs previously tested in various mouse models of human cancer, with VR5IL8 and VR5TNF VRs consistently outperforming VR5IL5 and V5 (full length native IL5 receptor). Overall, these findings establish an optimized and reproducible framework that offers valuable guidance for the future development and functional study of VR-CAR T cells in cellular therapies for solid tumors.

bioengineering↗

bioRxiv2025

NLTD 2.0: A Nonlinear Framework for Robust and Customizable Color Deconvolution in Histopathology

Advancements in computational approaches have enabled robust utilization of histological tissue data. A crucial step in the development of computational tools for the objective and quantitative analysis of tissue sections has been color deconvolution. Color deconvolution functions by separating the absorption of colors corresponding to stained molecular or tissue compartments. The most widely used color deconvolution method in digital pathology, linear color deconvolution as described in Ruifrok 2001 et al., decomposes color images according to the absorbance values for individual stains. However, linear color deconvolution assumes that stains are linearly decomposable, and it relies heavily upon identifying optimal color vectors of stains, which is often challenging. Furthermore, linear deconvolution methods cannot deconvolve the image with more than three stains, further limiting their broader applicability. To combat the limitations of previous methods, we developed an intuitive and robust color deconvolution method that effectively and accurately separates more than three stain signals, does not rely on predetermined color vectors, and doesnt rely on identifying optimal stain vectors. The proposed method, NLTD 2.0, presents a robust and efficient solution to tackle color variations in histopathology images, enhancing the reliability and precision of computational pathology. Additionally, incorporating the method as an ImageJ plugin amplifies accessibility and usability, enabling researchers and pathologists to leverage its capabilities without specialized programming skills. The intuitive interface streamlines the application, fostering broader acceptance within the computational pathology community.

bioinformatics↗

bioRxiv2025

Hyper-migratory CAR T cells alleviate ovarian cancer metastatic burden and improve prognosis

Cellular immunotherapy has shown remarkable efficacy in hematological malignancies but remains limited by infiltration issues in solid tumors, leading to poor treatment efficacy. We have recently shown that mesothelin-targeting CAR T cells co-transduced with cytokine-binding synthetic velocity receptors (VRs, referred to as CAR TV cells) demonstrate increased motility, improved infiltration in solid primary tumors, and lead to a better anti-tumor effect compared to CAR T cells that do not express VRs. However, it is metastasis that causes the vast majority of cancer related deaths and is difficult to target clinically, indicating an urgent unmet need. We show that these CAR T cells engineered to be hyper-migratory using VRs are highly effective against liver metastasis of ovarian cancer along different stages of the metastatic cascade. Mesothelin-targeting CAR TV cells expressing synthetic or native receptors responsive to the cytokine Interleukin-5 improved the survival of mice bearing an extremely high or terminal level of metastatic burden compared to CAR T cells that did not express VRs. Against newly established metastatic lesions and lesions undergoing metastatic outgrowth, CAR TV cells showed a robust anti-cancer effect resulting in an improved prognosis compared to control CAR T cells. Histopathological assessments showed a substantial reduction of metastasis number and lesion size with CAR TV treatment, concomitant with increased immune cell infiltration in the metastatic regions. Our work demonstrates the efficacy of high-motility CAR T cells in a metastatic setting and extends their scope to the treatment of metastasis of solid tumors.

bioengineering↗

bioRxiv2025

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

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

bioinformatics↗

bioRxiv2025

PIVOT: an open-source tool for multi-omic spatial data registration

Advances in spatial profiling have resulted in the generation of multi-omic atlases that span biological scales. In general, multiple workflows are required for image registration, coordinate registration, and spot deconvolution to integrate modalities. To improve the throughput of registration of multi-omic cohorts, we introduce PIVOT, a user-friendly and open-source interface for streamlined nonlinear registration. We demonstrate PIVOTs strengths through registration of three multi-omic datasets, and show comparison of its performance to existing workflows.

bioinformatics↗

bioRxiv2025

CODAvision: best practices and a user-friendly interface for rapid, customizable segmentation of medical images

Image-based machine learning tools have emerged as powerful resources for analyzing medical images, with deep learning-based semantic segmentation commonly utilized to enable spatial quantification of structures in images. However, customization and training of segmentation algorithms requires advanced programming skills and intricate workflows, limiting their accessibility to many investigators. Here, we present a protocol and software for automatic segmentation of medical images guided by a graphical user interface (GUI) using the CODAvision algorithm. This workflow simplifies the process of semantic segmentation of microanatomical structures by enabling users to train highly customizable deep learning models without extensive coding expertise. The protocol outlines best practices for creating robust training datasets, configuring model parameters, and optimizing performance across diverse biomedical image modalities. CODAvision enhances the usability of the CODA algorithm (Nature Methods, 2022) by streamlining parameter configuration, model training, and performance evaluation, automatically generating quantitative results and comprehensive reports. We expand beyond the original implementation of CODA to serial histology by demonstrating robust performance across numerous medical image modalities and diverse biological questions. We provide sample results in data types including histology, magnetic resonance imaging (MRI), and computed tomography (CT). We demonstrate the diverse use of this tool in applications including quantification of metastatic burden in in vivo models and deconvolution of spot-based spatial transcriptomics datasets. This protocol is designed for researchers with interest in rapid design of highly customizable semantic segmentation algorithms and a basic understanding of programming and anatomy.

bioinformatics↗

bioRxiv2025

Integration of nuclear morphology and 3D imaging to profile cellular neighborhoods

Nuclear morphology is an indicator of cellular function and disease states, as changes in nuclear size, shape, and texture often reflect underlying disease-related genetic, epigenetic, and microenvironmental alterations. For disease diagnosis, nuclear segmentation performed in 2D hematoxylin and eosin (H&E)-stained tissue sections has long represented the gold standard. However, recent advances in three-dimensional (3D) histology, which provide a more biologically accurate representation of the spatial heterogeneity of human microanatomy, has led to improved understandings of disease pathology. Yet challenges remain in the development of scalable and computationally efficient pipelines for extracting and interpreting nuclear features in 3D space. 2D histology neglects crucial spatial information, such as 3D connectivity, morphology, and rare events missed by sparser sampling. Here, through extension of the CODA platform, we integrate 3D imaging with nuclear segmentation to analyze nuclear morphological features in human tissue. Analysis of 3D tissue microenvironments uncovered critical changes in 3D morphometric heterogeneity. Additionally, it enables the spatial characterization of immune cell distribution in relation to tissue structures, such as variations in leukocyte density near pancreatic ducts and blood vessels of different sizes. This approach provides a more comprehensive understanding of tissue and nuclear structures, revealing spatial patterns and interactions that are critical for disease progression.

bioinformatics↗