Biology subjects

Pentimalli, T. M.

Publications and source records attributed to Pentimalli, T. M..

9 recordsLinked to original sources

bioRxiv2026

InterScale reveals multi-scale cellular interaction programs in spatial transcriptomics

Tissue homeostasis and disease emerge from cell-cell interactions operating across spatial scales: from autocrine and juxtacrine signals within micrometers to paracrine gradients coordinating responses across tissues. While these can be read out from spatial transcriptomics, existing computational methods capture either local adjacency-based or long-range dependencies, but rarely both within a single framework. We introduce InterScale, a graph-transformer approach that jointly models local and global cellular interactions from spatial transcriptomics data. By integrating a Graph Convolutional Network as a local component with a global transformer encoder, InterScale learns multi-scale representations of cellular communication. A downstream workflow enables scale-resolved interpretation of interactions from gene to tissue level. Applied to Sonic Hedgehog morphogen patterning in neural organoids, InterScale resolves spatially restricted neuronal differentiation programs and broader progenitor regulatory states along the morphogen gradient. In a human pancreatic dataset contrasting healthy and type 1 diabetic tissue, it reveals disease-associated spatial reorganization and tissue remodeling. InterScales modular architecture supports diverse spatial transcriptomics platforms and provides a scalable, unbiased, and biologically interpretable framework for studying cellular interactions across scales.

bioinformatics↗

bioRxiv2025

circVDJ-seq for T cell clonotype detection in single-cell and spatial multi-omics

Monitoring T cell repertoires in human tissues provides important insights into immune response mechanisms in cancer, infectious diseases, and autoimmunity. However, retrieving VDJ information from single-cell and spatial transcriptomics workflows with 3-barcoding of cDNA remains resource-intensive or requires specialized sequencing equipment. Here, we introduce circVDJ-seq for simplified and cost-efficient T cell receptor (TCR) profiling from 3-directed workflows like single-cell or single-nucleus RNA sequencing, ATAC+RNA multi-omics, and spatial transcriptomics. Application of circVDJ-seq to freshly resected neuroblastomas and postmortem lymph nodes affected by pneumonia or COVID-19 reveals distinct immune microenvironments and T cell clonality patterns, highlighting broad utility across diverse clinical contexts.

genomics↗

bioRxiv2025

Spatiotemporal dynamics of tumor microenvironment remodelling

During tumorigenesis, interactions between tumor and stromal cells progressively remodel the tumor microenvironment (TME) towards pro-tumoral functions. Understanding early TME remodeling dynamics is therefore crucial for developing interceptive therapies. However, clinical samples typically provide isolated, late tumorigenesis snapshots. To overcome this limitation, we generated triple-negative breast cancer mice that develop multifocal, asynchronous tumors along a continuous luminal-to-basal transdifferentiation trajectory. Ordering spatial transcriptomes from 100+ ducts along this trajectory reveals the spatiotemporal dynamics of TME remodeling and underlying molecular mechanisms. Cancer-associated myofibroblasts (myCAFs) emerge as key players in advanced tumors, where they orchestrate pro-invasive remodeling of the tumor-stromal interface. myCAFs are conserved in patient-derived xenograft models and steer tumor trajectories towards invasive phenotypes when co-injected with tumor cells in syngeneic mice. Our study shows that temporal ordering of spatially-resolved disease snapshots unravels some of the molecular "forces" that, starting from the cell-of-origin, propel cells/microenvironments along a disease trajectory.

cancer biology↗

bioRxiv2025

OpenDVP: An experimental and computational framework for community-empowered deep visual proteomics

Deep visual proteomics (DVP) is an emerging approach for cell type-specific and spatially resolved proteomics. However, its broad adoption has been constrained by the lack of an open-source end-to-end workflow in a community-driven ecosystem. Here, we introduce openDVP, an experimental and computational framework for simplifying and democratizing DVP. OpenDVP integrates open-source software for image analysis, including MCMICRO, QuPath, and Napari, and uses the scverse data formats AnnData and SpatialData for multi-omics integration. It offers two workflows: a fast-track pipeline requiring no image analysis expertise and an artificial intelligence (AI)-powered pipeline with recent algorithms for image pre-processing, segmentation, and spatial analysis. We demonstrate openDVPs versatility in three archival tissue studies, profiling human placenta, early-stage lung cancer, and locally relapsed breast cancer. In each study, our framework provided insights into health and disease states by integrating spatial single-cell phenotypes with exploratory proteomic data. Finally, we introduce deep proteomic profiling of cellular neighborhoods as a scalable approach to accelerate spatial discovery proteomics across biological systems.

systems biology↗

bioRxiv2025

Epiproteomic control of the epitranscriptome drives cortical neurogenesis

Neurological conditions are the leading cause of ill health worldwide. Here, we show that the neurodevelopmental disorder-associated ubiquitin ligase UBE3C regulates the cellular composition of the murine cerebral cortex and human brain organoids, with its loss favoring neurogenesis and suppressing glial fate. Using genetic complementation, we demonstrate that disease-associated UBE3C mutations alter its autoubiquitination activity and disrupt cortical lamination. Proteomic profiling of UBE3C-deficient forebrains and organoids identifies Cbll1 as a UBE3C substrate, and we show that the UBE3C-Cbll1 duo drives N6-methyladenosine (m6A) mRNA methylation. Hyperactivation of m6A writers in UBE3C-deficient neural progenitors impairs cell cycle exit, a defect reversible in vivo by the METTL3 inhibitor STM2457. Our findings uncover an epiproteomic mechanism controlling m6A-mediated gene expression and define a regulatory axis linking ubiquitin signaling to epitranscriptomic control of neural fate. This work provides a mechanistic framework for understanding neurodevelopmental disorders and highlights potential therapeutic strategies.

cell biology↗

bioRxiv2024

Deep learning-driven neuromorphogenesis screenings identify repurposable drugs for mitochondrial disease

Mitochondrial disease encompasses untreatable conditions affecting tissues with high energy demands. A severe manifestation of mitochondrial disease is Leigh syndrome (Leigh), which causes defects in basal ganglia and midbrain regions, psychomotor regression, lactic acidosis, and early death. We previously generated isogenic pairs of Leigh cerebral organoids and uncovered defects in neuromorphogenesis. Here, we leveraged on this disease feature to devise drug discovery pipelines. We developed a deep learning algorithm tailored for cell type-specific drug repurposing to identify drugs capable of promoting neuronal commitment. In parallel, we performed a survival drug screen in yeast and validated the repurposable hits on branching capacity in Leigh neurons. The two approaches independently highlighted azole compounds, Talarozole and Sertaconazole, both of which lowered lactate release and improved neurogenesis and neurite organization in Leigh midbrain organoids. Hence, targeting neuromorphogenesis has led to identify potential new drugs for mitochondrial disease and could prove an effective strategy for further drug discovery.

neuroscience↗

bioRxiv2023

Open-ST: High-resolution spatial transcriptomics in 3D

Spatial transcriptomics (ST) methods have been developed to unlock molecular mechanisms underlying tissue development, homeostasis, or disease. However, there is a need for easy-to-use, high-resolution, cost-efficient, and 3D-scalable methods. Here, we report Open-ST, a sequencing-based, open-source experimental and computational resource to address these challenges and to study the molecular organization of tissues in 3D. In mouse brain, Open-ST captured transcripts at subcellular resolution and reconstructed cell types. In primary tumor and patient-matched healthy/metastatic lymph nodes, Open-ST captured the diversity of immune, stromal and tumor populations in space. Distinct cell states were organized around cell-cell communication hotspots in the tumor, but not the metastasis. Strikingly, the 3D reconstruction and multimodal analysis of the metastatic lymph node revealed spatially contiguous structures not visible in 2D and potential biomarkers precisely at the 3D tumor/lymph node boundary. We anticipate Open-ST to accelerate the identification of spatial molecular mechanisms in 2D and 3D.

systems biology↗

bioRxiv2023

High-resolution molecular atlas of a lung tumor in 3D

Cells live and interact in three-dimensional (3D) cellular neighborhoods. However, histology and spatial omics methods mostly focus on 2D tissue sections. Here we present a 3D spatial atlas of a routine clinical sample, an aggressive human lung carcinoma, by combining in situ quantification of 960 cancer-related genes across [~]340,000 cells with measurements of tissue-mechanical components. 3D cellular neighborhoods subdivided the tumor microenvironment into tumor, stromal, and immune multicellular niches. Interestingly, pseudotime analysis suggested that pro-invasive epithelial-to-mesenchymal transition (EMT), detected in stroma-infiltrating tumor cells, already occurred in one region at the tumor surface. There, myofibroblasts and macrophages specifically co-localized with pre-invasive tumor cells and their multicellular molecular signature identified patients with shorter survival. Moreover, cytotoxic T-cells did not infiltrate this niche but colocalized with inhibitory dendritic and regulatory T cells. Importantly, systematic scoring of cell-cell interactions in 3D neighborhoods highlighted niche-specific signaling networks accompanying tumor invasion and immune escape. Compared to 2D, 3D neighborhoods improved the characterization of immune niches by identifying dendritic niches, capturing the 3D extension of T-cell niches and boosting the quantification of niche-specific cell-cell interactions, including druggable immune checkpoints. We believe that 3D communication analyses can improve the design of clinical studies investigating personalized, combination immuno-oncology therapies.

systems biology↗

bioRxiv2021

Neurodegeneration in human brain organoids infected with herpes simplex virus type 1

Herpes simplex virus type 1 (HSV-1) infection of the nervous system may lead to brain damage, including neurodegeneration. However, lack of suitable experimental models hinders understanding molecular mechanisms and cell-type-specific responses triggered by HSV-1. Here, we infected human brain organoids with HSV-1. Known features of HSV-1 infection such as alteration of neuronal electrophysiology and induction of antisense transcription were confirmed. Full-length mRNA-sequencing revealed aberrant 3 end formation and poly(A)-tail lengthening. Single-cell RNA-seq and spatial transcriptomics uncovered changes in the cellular composition of the infected organoids caused by viral replication and dysregulation of molecular pathways in cell-type specific manner. Furthermore, hallmarks of early neurodegeneration were observed, namely extracellular matrix disruption, STMN2 and TARDBP/TDP43 downregulation, and upregulation of the AD-related non-coding RNA BC200/BCYRN1. These hallmarks were weaker/absent when infecting with a mutant HSV-1 control. Together, our data indicate that brain organoids serve as a powerful model to study mechanisms of HSV-1-driven neurodegeneration.

neuroscience↗