Biology subjects

Roncaroli, F.

Publications and source records attributed to Roncaroli, F..

7 recordsLinked to original sources

bioRxiv2026

Proteomics-constrained deconvolution reveals spatial cell-type programs in tumours

Accurately resolving cell-type mixtures in spatial transcriptomics remains challenging, particularly in heterogeneous tumours where cell populations are intermixed and matched single-cell references may be unavailable or poorly aligned. Current deconvolution approaches either require high-quality scRNA-seq references, suffer from scalability limitations, or lack interpretability. We introduce PISTACHIO, a proteomics-informed spatial transcriptomics deconvolution framework based on constrained non-negative matrix factorization with a negative-binomial likelihood. Rather than using probabilistic priors, PISTACHIO incorporates spatial cell-type constraints derived from paired Imaging Mass Cytometry, enforcing biologically grounded sparsity and explicit spatial feasibility of cell-type presence. PISTACHIO improved recovery of spatial cell-type distributions compared with Cell2location and STdeconvolve across synthetic and real tumour datasets. Our approach remains robust under cell-type assignment errors, maintaining high correlation with ground-truth under moderate noise, and achieves fast runtime on standard hardware, enabling practical large-scale deployment.

bioinformatics↗

bioRxiv2025

Disruption of the Wnt-antagonist Apc in the pituitary stem cells drives the development of adamantinomatous craniopharyngioma.

Adamantinomatous craniopharyngiomas (aCPs) are complex intracranial neoplasms that generally arise in the sellar and suprasellar region of the brain which affect the endocrine and nervous systems causing severe sequelae. Activating mutations resulting in degradation-resistant forms of {beta}-catenin (CTNNB1) have been shown to be the main driver for many of these neoplasms. However, the underlying genetic driver for a proportion of these tumours is still unknown. Using murine transgenic models, we show that genetic disruption of the Wnt-antagonist and tumour suppressor, adenomatous polyposis coli (Apc), within the pituitary progenitors/stem cells, leads to pituitary tumours that closely resemble human aCPs. These tumours present classic histopathological hallmarks of aCPs, such as large deposits of wet keratin, stellar-reticular-like cells, large cystic components and clusters of accumulating nucleo-cytoplasmic {beta}-catenin that are slow-dividing and exhibit a secretory phenotype. We show that a hypomorphic allele of Apc is sufficient for tumour development, indicating that disruption of Apc function can lead to aCP formation. Moreover, we identify that bi-allelic loss of Apc in the Sox2+ve pituitary stem cells is sufficient to initiate tumour formation, indicating that Sox2+ve stem cells are the cell origin of these Apc-driven aCPs. Transcriptomic analyses of early tumour-initiating cells revealed that Apc-driven clusters of accumulating {beta}-catenin undergo senescence-associated secretory phenotype (SASP), which is p21-mediated and results in secretion of inflammasome, angiosome and developmental growth factors. Our data unequivocally show a causal role for the disruption of the tumour suppressor Apc as a primary driver of aCPs independent of mutations in {beta}-catenin. We provide murine models representing a novel genetic subtype of human aCPs offering insights into aCP pathogenesis. Our work reinforces the importance of genetic testing for mutations in APC in patients with aCPs and identifies a potential need to screen patients with familial adenomatous polyposis (FAP) or spectrum of APC-pathogenic syndromes for aCPs in early life.

cancer biology↗

bioRxiv2025

DeepPathway: Predicting Pathway Expression from Histopathology Images

Spatial transcriptomics (ST) technologies provide spatially resolved gene expression along with image data, allowing the integrative analysis of complex tissue microenvironments. Despite their potential, the widespread adoption of ST remains limited due to high costs, and methodological challenges in data acquisition. Thus, there have been recent efforts to develop deep learning methods capable of inferring spatial gene expression from the much cheaper and easily available haematoxylin and eosin (H&E) images. These methods demonstrate promising results in reconstructing transcriptomic landscapes within tissue sections. While existing approaches predominantly focus on gene-level predictions, biological processes are often regulated at the pathway level through coordinated activity among functionally related genes. We present DeepPathway, a contrastive learning-based approach trained on ST data to predict pathway expression from H&E-stained sections. We compute input pathway expression by summarizing the expression of constituent genes using established pathway definitions. We evaluate the performance of our method on two prostate cancer datasets and validate our approach on the H&E images acquired from The Cancer Genome Atlas (TCGA) clearly differentiating between normal and tumour tissues. Finally, we apply our method to predict hypoxia signatures using H&Es of brain tumour samples where hypoxia staining with pimonidazole was available as ground truth. Implementation code for DeepPathway is available at https://github.com/aahsan045/DeepPathway.

bioinformatics↗

bioRxiv2024

Mitochondrial sites of contact with the nucleus aid in chemotherapy evasion of glioblastoma cells

Glioblastoma (GBM) is the most common form of a malignant primary brain tumour in adults for which therapeutic options are minimal. The rapid onset of the resistance mechanisms against the chemotherapeutic agent Temozolomide (TMZ), the first line of pharmacological care for patients, prevents the long-term validity of this approach. The underpinning biology for this remains poorly understood thus compromising the efficacy of this approach. The Translocator Protein (TSPO) is an 18kDa ubiquitous cholesterol-binding molecule on the outer membrane of mitochondria (OMM). Upregulated in cancers TSPO is required to form contacts between mitochondria and the nucleus termed: Nucleus Associated Mitochondria (NAM). In GBM tissues as well as in 2D and 3D cell cultures we assayed patterns of TSPO expression (i), autophagy/mitophagy (ii), transcription factors (iii) and susceptibility to TMZ-induced demise (iv). Confocal and ultrastructural imaging detailed the organization and redistribution of the mitochondrial network (v). Our findings show that TMZ exploits mitochondria via TSPO to aid the formation of NAM which couples the expression of the nuclear transcription factor Sterol regulatory element-binding transcription factor 1 (SREBP1) and the stabilization of YAP/TAZ. Pharmacological modulation of TSPO counteracts all the above and re-instates susceptibility to TMZ-induced demise. NAM is therefore proposed as a variable in the engagement and execution of pro-survival mechanisms in GBM thus offering a means to both insight into the pathophysiology of this disease and offer novel therapeutic strategies. Key PointsO_LITMZ exploits TSPO to curb mitochondrial quality control in glioblastoma cells. C_LIO_LITMZ-mediated MRR is associated with the relocation of mitochondria to the nucleus and modulation of transcriptional factors involved in cholesterol metabolism and adaptation to aggressive growth. C_LIO_LITSPO represents a pharmacological target to revert chemoresistance in glioblastoma cells. C_LI Importance of the StudyThis study elucidates a mitochondrion-driven mechanism of chemoresistance in human glioblastoma cells, which depends on the mitochondrial translocator protein TSPO. The administration of TSPO ligands restores susceptibility to TMZ by influencing the dynamics of transcriptional factors associated with cholesterol metabolism and mechanical transduction.

cancer biology↗

bioRxiv2023

Hypoxia coordinates the spatial landscape of myeloid cells within glioblastoma to affect outcome

Myeloid cells are highly prevalent in glioblastoma (GBM), existing in a spectrum of phenotypic and activation states. We currently have limited knowledge of the tumour microenvironment (TME) determinants that influence the localisation and the functions of the diverse myeloid cell populations in GBM. Here we have utilised orthogonal imaging mass cytometry with single cell and spatial transcriptomics approaches to identify and map the various myeloid populations in the human GBM tumour microenvironment (TME). Our results show that different myeloid populations have distinct and reproducible compartmentalisation patterns in the GBM TME that is driven by tissue hypoxia, regional chemokine signalling, and varied homotypic and heterotypic cellular interactions. We subsequently identified specific tumour sub-regions in GBM, based upon composition of identified myeloid cell populations, that were linked to patient survival. Our results provide new insight into the spatial organisation of myeloid cell sub populations in GBM, and how this is predictive of clinical outcome. TeaserMulti-modal mapping reveals that the spatial organisation of myeloid cells in glioblastoma impacts disease outcome.

cancer biology↗

bioRxiv2022

Novel Laser Capture Microdissection-Proteomic Analysis Identifies Spatially Distinct Extracellular Matrix Signatures in the Core and Infiltrating Edge of Human Glioblastoma.

BackgroundGlioblastoma is the most common and aggressive primary brain tumour in adults. Hallmarks of glioblastoma include its intra-tumoural heterogeneity and extensive invasion of the surrounding brain. Glioblastoma is known to remodel the extracellular matrix (ECM) of the brain, resulting in altered mechanical properties and the establishment of a tumour-promoting microenvironment. How changes in the expression and spatial distribution of ECM constituents within glioblastoma contribute to invasion and disease progression is still unclear. MethodsHere we report on a protocol for laser-capture microdissection coupled with mass spectrometry (LCM-proteomics) that allowed a spatially resolved and unbiased analysis of the regional ECM proteome (matrisome) in formalin-fixed and paraffin-embedded (FFPE) samples of human glioblastoma. We investigated five molecularly characterised hemispheric adult glioblastomas where the brain/tumour interface and tumour epicentre were represented in the surgical specimens and snap-frozen tissue was available. LCM-proteomic analysis was validated with immunohistochemistry. ResultsLCM-proteomics identified 53 matrisome proteins in FFPE tissue, demonstrating comparable performance with conventional analysis of snap-frozen tissue. The analysis revealed distinct matrisome components in the brain/tumour interface versus the tumour epicentre. Guided by data from LCM-proteomic analysis, immunostaining for tenascin-R confirmed greater staining in the brain/tumour interface, whilst expression of fibronectin was higher in the tumour epicentre. ConclusionThe protocol described in this work allowed for accurate, spatially resolved analysis of ECM in FFPE tissues, with performance comparable to analysis of snap-frozen tissue. While the focus for this work was on the regional ECM composition of glioblastoma, we found that the LCM-proteomics protocol is also applicable to the study of the wider proteome, and represents a valuable tool for investigating tumour/tissue heterogeneity. This protocol opens the possibility to apply LCM-proteomics to retrospective studies with the advantage of accessing clinical history and follow-up information, providing a valuable resource for translational research in glioblastoma.

cancer biology↗

bioRxiv2021

Glioblastoma cell fate is differentially regulated by the microenvironments of the tumour bulk and infiltrative margin

Glioblastoma recurrence originates from invasive cells at the tumour margin that escape surgical debulking, but their biology remains poorly understood. Here we generated three somatic mouse models recapitulating the main glioblastoma driver mutations to characterise margin cells. We find that, regardless of genetics, tumours converge on a common set of neural- like cellular states. However, bulk and margin display distinct neurogenic patterns and immune microenvironments. The margin is immune-cold and preferentially follows developmental-like trajectories to produce astrocyte-like cells. In contrast, injury-like programmes dominate in the bulk, are associated with immune infiltration and generate lowly-proliferative injured neural progenitor-like (iNPCs) cells. In vivo label-retention approaches further demonstrate that iNPCs account for a significant proportion of dormant glioblastoma cells and are induced by interferon signalling within T-cell niches. These findings indicate that tumour region is a major determinant of glioblastoma cell fate and therapeutic vulnerabilities identified in bulk may not extend to the margin residuum.

cancer biology↗