Biology subjects

Secrier, M.

Publications and source records attributed to Secrier, M..

At least 19 recordsLinked to original sources

bioRxiv2025

Integrative single-cell and spatial mapping of oxidative stress response uncovers GCLC+ mesenchymal tumor cell state linked with favorable outcomes in triple negative breast cancer

Inhibiting oxidative stress response (OSR) proteins has been suggested as a therapeutic strategy in triple negative breast cancer (TNBC). However, the cell type specificity and spatial distribution of OSR genes and proteins, such as GCLC and NQO1, is unknown. Using single cell and spatial transcriptomics datasets we found that OSR genes were highly expressed in TNBC tumor cells, which localized in spatial clusters. Multiplex immunofluorescence imaging of 345 TNBC samples from 186 patients demonstrated that OSR proteins GCLC and NQO1 exhibit distinct expression profiles across tumor, immune, and stromal cell populations and are elevated in inflamed histological regions. Tumor cell OSR protein expression was associated with the composition of the adjacent cellular neighborhood. Furthermore, we identified GCLC and vimentin positive (GCLC+VIM+) mesenchymal-like tumor cells, residing near immune cells and exhibiting increased proliferation and decreased anastasis signatures, suggesting sensitivity to chemotherapy. Across a panel of thirteen TNBC cell lines, GCLC expression was positively correlated with sensitivity to cisplatin. Cox regression analysis revealed that patients with higher proportions of GCLC+VIM+ tumor cells had a longer overall survival. Collectively, our results demonstrate that individual OSR proteins are expressed in distinct microenvironments and tumor cell states, potentially contributing to patient outcomes.

cancer biology↗

bioRxiv2025

Integrative Spatial Modelling of Cellular Plasticity using Graph Neural Networks and Geostatistics

Cellular plasticity - the ability of cells to change phenotype in response to intrinsic and environmental cues - is central to development, regeneration, and disease, but remains difficult to quantify due to its dynamic, context-dependent nature. Here we introduce a framework that unites AI and geostatistics - graph neural networks and spatial regression models - to both predict and explain cell state variation in spatial transcriptomics data. We formalize state predictability as a quantitative proxy for plasticity, where stable states are predictable and plastic states are not. Applied to epithelial-mesenchymal plasticity (EMP) in breast cancer, the framework shows that mesenchymal states are stabilized by recurrent copy number alterations and microenvironmental cues, whereas hybrid states remain unpredictable and plastic. It also uncovers a long-range influence of myofibroblasts on EMP, demonstrating that stromal remodeling can propagate plasticity across tissue regions. Our framework yields interpretable, scale-aware insights into intrinsic and extrinsic drivers of state transitions, providing a broadly applicable strategy for modelling dynamic cell states in spatial biology.

bioinformatics↗

bioRxiv2025

EnrichMap: Spatially informed enrichment analysis for functional interpretation of spatial transcriptomics

Advances in spatial transcriptomics (ST) allow high-resolution mapping of gene expression within intact tissues, revealing spatially regulated biology. However, accurately quantifying gene set activity in spatial contexts remains inadequate. We present EnrichMap, a Python package designed to compute gene set enrichment scores in ST data by integrating batch correction, smoothing and spatial confounder adjustment. EnrichMap generates enrichment scores for signatures or pathways, with an optional framework for incorporating gene weights. To assess performance, we benchmark EnrichMap against established scoring techniques using spatial autocorrelation metrics such as Morans I. Our results show that EnrichMap improves spatial coherence, producing scores that capture both local and global spatial patterns, while conventional approaches tend to exhibit higher heterogeneity or diminished spatial structure. EnrichMap is demonstrated to be generalisable across a variety of ST platforms, including Visium, Xenium, MERFISH and imaging mass cytometry, maintaining consistent performance despite differences in spatial resolution and detection chemistry. Our method offers a platform-agnostic, spatially and biologically meaningful representation of gene set activity across complex tissue landscapes.

bioinformatics↗

bioRxiv2025

Balancing tumour proliferation and sustained cell cycle arrest through proteostasis remodelling drives immune niche compartmentalisation in breast cancer

Uncontrolled proliferation is a hallmark of cancer, yet tumour cells can enter G0 arrest by halting the cell cycle reversibly (quiescence) or irreversibly (senescence) to survive under stress and in hostile tumour microenvironments (TME). G0 arrested cells contribute to drug tolerance, metastasis and recurrence, but their identification remains challenging due to their rarity and elusive regulatory pathways. Here, we quantify G0 arrest and proliferation decisions in single-cell and spatially profiled breast primary tumours to unveil adaptive responses driving immune compartmentalisation. We identify a genomically-constrained and prolonged G0 arrest state resembling that of dormant precursors of cancer progression. This state featured adaptive transcriptional reprogramming, including unfolded protein response stress, reduced mitochondrial translation and epithelial-mesenchymal plasticity enabled by semaphorin signalling. Spatial transcriptomics revealed G0 arrest pockets encapsulated by APOE+ lipid-associated macrophages, myofibroblastic CAFs and immature perivascular cells, suggestive of an immunosuppressive niche contrasting with cytotoxic environments in proliferative hotspots and displaying distinct drug sensitivities. To facilitate future research, we provide a foundation model capturing G0 arrest and proliferation with 94% accuracy in single cell data, available at https://github.com/secrierlab/G0-LM. Our findings provide new insights into the spatial organisation of cell cycle arrest in breast cancer, highlighting the role of G0 states in tumour heterogeneity and adaptation.

cancer biology↗

bioRxiv2024

Classifying epithelial-mesenchymal transition states in single cell cancer data using large language models

Single-cell foundation models (scFMs) have been widely adopted for cell type annotation, yet their suitability for modelling cellular plasticity, where cells transition along continuous, context-dependent state trajectories, remains unclear. Here, we systematically benchmark state-of-the-art scFMs against conventional machine learning and bioinformatics approaches on epithelial-mesenchymal plasticity (EMP), a prototypical plastic cellular process. We show that naive fine-tuning of scFMs often fails to resolve intermediate states and is strongly influenced by tissue- and stimulus-specific signals. We propose a parameter-efficient dual-task adaptation strategy for EMP foundation models (EMP-FM) that combines discrete classification with pseudotime-guided regression, which improves cell state resolution in controlled settings (up to 85% AUROC), but remains sensitive to domain shifts. Across diverse in vitro and in vivo datasets, scFMs do not consistently outperform conventional methods, which often achieve comparable performance with lower complexity. Together, our results delineate both the potential and current limitations of scFMs for modelling cellular plasticity and support their complementary use alongside established bioinformatics approaches.

bioinformatics↗

bioRxiv2024

Pro-survival roles for p21(Cip1/Waf1) in Non-Small Cell Lung Cancer

Quiescence is a reversible state of proliferative arrest, distinct from senescence. While cancer is a disease of dysregulated proliferation, cancer cells can retain the ability to enter quiescence which confers advantages to tumour cells by protecting them from chemotherapy or by allowing metastasis to distant sites. Multiple mechanisms exist to induce and maintain quiescence that are not yet fully understood. Here, we show that high expression of the CDK inhibitor p21Cip1/Waf1 correlates with a poor prognosis in TP53 wild-type, but not TP53 mutant, non-small cell lung cancer (NSCLC) patients. Using quantitative single-cell imaging of genetically-engineered NSCLC reporter cell lines, we show that TP53 wild-type NSCLC cells can enter a p21-dependent spontaneous quiescent state, downstream of replication stress. Furthermore, p21 expression confers survival advantages to TP53 wild-type NSCLC cells, both under normal proliferation and in response to chemotherapy. We also show that p21 can promote tumour relapse by allowing cells to recover from both G1 and G2 arrest states after drug removal. Together, our data suggest that targeting p21 function in TP53 wild-type tumours could lead to better outcomes for chemotherapy treatment in NSCLC patients. Statement of SignificanceWe show that TP53WT Non-Small Cell Lung Cancer cells can enter a p21-dependent spontaneous quiescent state and that p21 maintains the viability of NSCLC cells, is chemoprotective and can promote tumour relapse.

cancer biology↗

bioRxiv2024

TDP-43 loss induces extensive cryptic polyadenylation in ALS/FTD

Nuclear depletion and cytoplasmic aggregation of the RNA-binding protein TDP-43 is the hallmark of ALS, occurring in over 97% of cases. A key consequence of TDP-43 nuclear loss is the de-repression of cryptic exons. Whilst TDP-43 regulated cryptic splicing is increasingly well catalogued, cryptic alternative polyadenylation (APA) events, which define the 3 end of last exons, have been largely overlooked, especially when not associated with novel upstream splice junctions. We developed a novel bioinformatic approach to reliably identify distinct APA event types: alternative last exons (ALE), 3UTR extensions (3Ext) and intronic polyadenylation (IPA) events. We identified novel neuronal cryptic APA sites induced by TDP-43 loss of function by systematically applying our pipeline to a compendium of publicly available and in house datasets. We find that TDP-43 binding sites and target motifs are enriched at these cryptic events and that TDP-43 can have both repressive and enhancing action on APA. Importantly, all categories of cryptic APA can also be identified in ALS and FTD post mortem brain regions with TDP-43 proteinopathy underlining their potential disease relevance. RNA-seq and Ribo-seq analyses indicate that distinct cryptic APA categories have different downstream effects on transcript and translation. Intriguingly, cryptic 3Exts occur in multiple transcription factors, such as ELK1, SIX3, and TLX1, and lead to an increase in wild-type protein levels and function. Finally, we show that an increase in RNA stability leading to a higher cytoplasmic localisation underlies these observations. In summary, we demonstrate that TDP-43 nuclear depletion induces a novel category of cryptic RNA processing events and we expand the palette of TDP-43 loss consequences by showing this can also lead to an increase in normal protein translation.

neuroscience↗

bioRxiv2023

SpottedPy quantifies relationships between spatial transcriptomic hotspots and uncovers new environmental cues of epithelial-mesenchymal plasticity in cancer

Spatial transcriptomics is revolutionising the exploration of intratissue heterogeneity in cancer, yet capturing cellular niches and their spatial relationships remains challenging. We introduce SpottedPy, a Python package designed to identify tumour hotspots and map spatial interactions within the cancer ecosystem. Using SpottedPy, we examine epithelial-mesenchymal plasticity in breast cancer and highlight stable niches associated with angiogenic and hypoxic regions, shielded by CAFs and macrophages. Hybrid and mesenchymal hotspot distribution followed transformation gradients reflecting progressive immunosuppression. Our method offers flexibility to explore spatial relationships at different scales, from immediate neighbours to broader tissue modules, providing new insights into tumour microenvironment dynamics.

bioinformatics↗

bioRxiv2023

BDNF controls phosphorylation and transcriptional networks governing cytoskeleton organization and axonal regeneration

The cell-intrinsic capacity of neurons to regenerate axons requires widespread coordination of the transcriptome, activation of multiple kinases, and reorganization of the cytoskeleton. Axonal repair is also influenced by extrinsic activating factors, such as neurotrophins. Here, we reveal that brain-derived neurotrophic factor (BDNF) amplifies multiple neuron-intrinsic programs to foster axonal regeneration in human motor neurons. Through metabolic RNA sequencing and phosphoproteomic profiling, we elucidate BDNF signalling and its role in axonal regeneration. We discover that BDNF controls RNA stability and transcriptional programs that converge with regeneration-associated gene (RAG) sets. We further unveil that BDNF governs the phosphorylation of multiple proteins essential for cytoskeletal dynamics, a major determinant of effective nerve regeneration. Using compartmentalized neuronal cultures, we demonstrate that the regeneration driven by BDNF depends on the axon-specific activation of ERK/RSK/S6K kinase pathway. We propose a model in which BDNF augments neuron-intrinsic pathways to drive axonal regeneration in human motor neurons. TeaserBDNF aids nerve repair by fine-tuning the metabolism of RNA and by changing the building blocks of the nerve cell cytoskeleton.

molecular biology↗

bioRxiv2023

Multi-scale characterisation of homologous recombination deficiency in breast cancer

BackgroundHomologous recombination is a robust, broadly error-free mechanism of double strand break repair, and deficiencies cause reliance on alternative repair processes, leading to sensitivity to PARP inhibitors. Patients displaying homologous recombination deficiency can be identified using mutational signatures. However, these patterns are difficult to reliably infer from exome sequencing. Additionally, as mutational signatures are a historical record of mutagenic processes, this limits their utility in describing the current status of a tumour. ResultsHere, we apply two methods for classifying homologous recombination deficiency in breast cancer to explore the features and heterogeneity associated with this phenotype. Firstly, we apply a likelihood-based method which leverages small insertions and deletions to improve classification of homologous recombination deficiency in exome sequenced breast cancers. We find that whilst BRCA+ homologous recombination deficient samples display strong similarities to those harbouring BRCA defects, they deviate in microenvironmental features such as hypoxic signalling. Secondly, using this classification we develop a 228-gene transcriptional signature which simultaneously characterises homologous recombination deficiency and BRCA1/2-defect status, and is linked with PARP inhibitor response. Finally, we apply this signature to single-cell RNA-sequenced breast cancer cohorts to study homologous recombination deficiency at single cell resolution and demonstrate that these cells present a distinct milieu of interactions with immune cells in their microenvironment compared to their HR proficient counterparts, which could inform checkpoint blockade for HRD tumours. ConclusionsWe apply multi-scale approaches to characterise homologous recombination deficiency in breast cancer through the development of mutational and transcriptional signatures. We show that indels, even at low levels, can improve homologous recombination deficiency classification. Additionally, we demonstrate the heterogeneity of homologous recombination deficiency, especially in relation to BRCA status, and show that indications of this feature can be captured at a single-cell level, enabling further investigations into interactions between DNA repair deficient cells and their tumour microenvironment.

genomics↗

bioRxiv2023

HistoMIL: a Python package for training Multiple Instance Learning models on histopathology slides

Haematoxilin and eosin (H&E) stained slides are commonly used as the gold standard for disease diagnosis. Remarkable progress in the deep learning field in recent years has enabled the detection of complex molecular patterns within such histopathology slides, suggesting automated approaches could help inform pathologists decisions. In this context, Multiple Instance Learning (MIL) algorithms have been shown to outperform Transfer Learning (TL) based methods for a variety of tasks. However, there is still a considerable complexity to implementing and using such methods for computational biology research and clinical practice. We introduce HistoMIL, a Python package designed to simplify the implementation, training, and inference process of MIL-based algorithms for computational pathologists and biomedical researchers. In HistoMIL, we have integrated a self-supervised learning-based module to train the feature encoder, a full pipeline encompassing TL as well as three MIL algorithms, namely ABMIL (1), DSMIL (2), and TransMIL (3). By utilising the PyTorch Lightning framework (4), HistoMIL enables effortless customization of training intricacies and implementation of novel algorithms. We illustrate the capabilities of HistoMIL by building predictive models for 2,487 cancer hallmark genes on breast cancer histology slides from The Cancer Genome Atlas, on which we demonstrate AUROC performances of up to 85%. Cell proliferation processes were most easily detected, shedding light on the opportunities but also limitations of applying deep learning for gene expression detection. The HistoMIL package is proposed as a tool to simplify the implementation and usage of deep learning tasks for researchers.

pathology↗

bioRxiv2023

Mis-spliced transcripts generate de novo proteins in TDP-43-related ALS/FTD

Functional loss of TDP-43, an RNA-binding protein genetically and pathologically linked to ALS and FTD, leads to inclusion of cryptic exons in hundreds of transcripts during disease. Cryptic exons can promote degradation of affected transcripts, deleteriously altering cellular function through loss-of-function mechanisms. However, the possibility of de novo protein synthesis from cryptic exon transcripts has not been explored. Here, we show that mRNA transcripts harboring cryptic exons generate de novo proteins both in TDP-43 deficient cellular models and in disease. Using coordinated transcriptomic and proteomic studies of TDP-43 depleted iPSC-derived neurons, we identified numerous peptides that mapped to cryptic exons. Cryptic exons identified in iPSC models were highly predictive of cryptic exons expressed in brains of patients with TDP-43 proteinopathy, including cryptic transcripts that generated de novo proteins. We discovered that inclusion of cryptic peptide sequences in proteins altered their interactions with other proteins, thereby likely altering their function. Finally, we showed that these de novo peptides were present in CSF from patients with ALS. The demonstration of cryptic exon translation suggests new mechanisms for ALS pathophysiology downstream of TDP-43 dysfunction and may provide a strategy for novel biomarker development. One Sentence SummaryLoss of TDP-43 function results in the expression of de novo proteins from mis-spliced mRNA transcripts.

neuroscience↗

bioRxiv2022

A deep learning and graph-based approach to characterise the immunological landscape and spatial architecture of colon cancer tissue

Tumour immunity is key for the prognosis and treatment of colon adenocarcinoma, but its characterisation remains cumbersome and expensive, requiring sequencing or other complex assays. Detecting tumour-infiltrating lymphocytes in haematoxylin and eosin (H&E) slides of cancer tissue would provide a cost-effective alternative to support clinicians in treatment decisions, but inter- and intra-observer variability can arise even amongst experienced pathologists. Furthermore, the compounded effect of other cells in the tumour microenvironment is challenging to quantify but could yield useful additional biomarkers. We combined RNA sequencing, digital pathology and deep learning through the InceptionV3 architecture to develop a fully automated computer vision model that detects prognostic tumour immunity levels in H&E slides of colon adenocarcinoma with an area under the curve (AUC) of 82%. Amongst tumour infiltrating T cell subsets, we demonstrate that CD8+ effector memory T cell patterns are most recognisable algorithmically with an average AUC of 83%. We subsequently applied nuclear segmentation and classification via HoVer-Net to derive complex cell-cell interaction graphs, which we queried efficiently through a bespoke Neo4J graph database. This uncovered stromal barriers and lymphocyte triplets that could act as structural hallmarks of low immunity tumours with poor prognosis. Our integrated deep learning and graph-based workflow provides evidence for the feasibility of automated detection of complex immune cytotoxicity patterns within H&E-stained colon cancer slides, which could inform new cellular biomarkers and support treatment management of this disease in the future.

cancer biology↗

bioRxiv2022

Immune evasion impacts the selective landscape of driver genes during tumorigenesis

Carcinogenesis is an evolutionary process fueled by the interplay of somatic mutations and the local microenvironment. In recent years, hundreds of cancer related genes have been discovered using cancer cohorts. However, these cohorts are heterogenous mixtures of different molecular phenotypes, which hampers the identification of driver genes associated to a specific cancer hallmark or microenvironment. Here, we compared the landscape of positively selected somatic mutations in immune-escaped (escape+) versus non-escaped (escape-) tumors. We applied the ratio of non-synonymous to synonymous mutations (dN/dS) to 9896 individuals from 31 primary tumor tissues from the Cancer Genome Atlas (TCGA) separated by escape status. Altogether, we found 85 driver genes, including 27 and 16 novel driver genes in escape- and escape+ tumors, respectively. Overall, driver dN/dS of escape+ tumors (dN/dS=1.23) was significantly lower and closer to neutrality than driver dN/dS of escape-tumors (dN/dS=1.62), suggesting a relaxation of positive selection in driver genes, a relaxation of negative selection on immunogenic driver sites, or a combination of both fueled by immune escape. We also found that the proportion of unique sites mutated in escape+ tumors is almost double than in escape-tumors, and that immune evasion allows for a more diverse repertoire of mutational signatures. We also identified that strong immunoediting in the absence of escape leads to a better overall survival in tumors enriched by an inflamed phenotype. Ultimately, our findings reveal differences in the evolutionary strategies used by cancer cells to establish tumorigenesis and highlight the need for better patient stratification to develop tailored treatments based on molecular targets.

genomics↗

bioRxiv2021

Genomic hallmarks of cellular dormancy in cancer and therapeutic implications

Therapy resistance in cancer is often driven by a subpopulation of cells that are temporarily arrested in a non-proliferative G0 state, which is difficult to capture and whose mutational drivers remain largely unknown. We developed methodology to robustly identify this state from transcriptomic signals and characterised its prevalence and genomic constraints in solid primary tumours. We show that G0 arrest preferentially emerges in the context of more stable, less mutated genomes which maintain TP53 integrity and lack the hallmarks of DNA damage repair deficiency, while presenting increased APOBEC mutagenesis. We employ machine learning to uncover novel genomic dependencies of this process and validate the role of the centrosomal gene CEP89 as a modulator of proliferation/G0 arrest capacity. Lastly, we demonstrate that G0 arrest underlies unfavourable responses to various therapies exploiting cell cycle, kinase signalling and epigenetic mechanisms in single cell data, and propose a G0 arrest transcriptional signature that is linked with therapeutic resistance and can be used to further study and clinically track this state.

cancer biology↗

bioRxiv2021

Genomic events shaping epithelial-to-mesenchymal trajectories in cancer

The epithelial to mesenchymal transition (EMT) is a key cellular process underlying cancer progression, with multiple intermediate states whose molecular hallmarks remain poorly characterized. To fill this gap, we explored EMT trajectories in 7,180 tumours of epithelial origin and identified three macro-states with prognostic and therapeutic value, attributable to epithelial, hybrid E/M and mesenchymal phenotypes. We show that the hybrid state is remarkably stable and linked with increased aneuploidy and APOBEC mutagenesis. We further employed spatial transcriptomics and single cell datasets to show that local effects impact EMT transformation through the establishment of distinct interaction patterns with cytotoxic, NK cells and fibroblasts in the tumour microenvironment. Additionally, we provide an extensive catalogue of genomic events underlying distinct evolutionary constraints on EMT transformation. This study sheds light on the aetiology of distinct stages along the EMT trajectory, and highlights broader genomic and environmental hallmarks shaping the mesenchymal transformation of primary tumours.

cancer biology↗

bioRxiv2021

Pan-cancer survey of tumour mass dormancy and underlying mutational processes

Tumour mass dormancy is the key intermediate step between immune surveillance and cancer progression, yet due to its transitory nature it has been difficult to capture and characterise. Little is understood of its prevalence across cancer types and of the mutational background that may favour such a state. While this balance is finely tuned internally by the equilibrium between cell proliferation and cell death, the main external factors contributing to tumour mass dormancy are immunological and angiogenic. To understand the genomic and cellular context in which tumour mass dormancy may develop, we comprehensively profiled signals of immune and angiogenic dormancy in 9,631 cancers from the Cancer Genome Atlas and linked them to tumour mutagenesis. We find evidence for immunological and angiogenic dormancy-like signals in 16.5% of bulk sequenced tumours, with a frequency of up to 33% in certain tissues. Mutations in the CASP8 and HRAS oncogenes were positively selected in dormant tumours, suggesting an evolutionary pressure for controlling cell growth/apoptosis signals. By surveying the mutational damage patterns left in the genome by known cancer risk factors, we found that ageing-induced mutations were relatively depleted in these tumours, while patterns of smoking and defective base excision repair were linked with increased tumour mass dormancy. Furthermore, we identified a link between APOBEC mutagenesis and dormancy, which comes in conjunction with immune exhaustion and may partly depend on the expression of the angiogenesis regulator PLG as well as interferon and chemokine signals. Tumour mass dormancy also appeared to be impaired in hypoxic conditions in the majority of cancers. The microenvironment of dormant cancers was enriched in cytotoxic and regulatory T cells, as expected, but also in macrophages and showed a reduction in inflammatory Th17 signals. Finally, tumour mass dormancy was linked with improved patient survival outcomes. Our analysis sheds light onto the complex interplay between dormancy, exhaustion, APOBEC activity and hypoxia, and sets directions for future mechanistic explorations.

cancer biology↗

bioRxiv2021

Cell environment shapes TDP-43 function: implications in neuronal and muscle disease

TDP-43 aggregation and redistribution have been recognised as a hallmark of amyotrophic lateral sclerosis, frontotemporal dementia and other neurological disorders. While TDP-43 has been studied extensively in neuronal tissues, TDP-43 inclusions have also been described in the muscle of inclusion body myositis patients, highlighting the need to understand the role of TDP-43 beyond the central nervous system. Using RNA-seq we performed the first direct comparison of TDP-43-mediated transcription and alternative splicing in muscle (C2C12) and neuronal (NSC34) mouse cells. Our results clearly show that TDP-43 displays a tissue-characteristic behaviour targeting unique transcripts in each cell type. This is not due to variable transcript abundance but rather due to cell-specific expression of RNA-binding proteins, which influences TDP-43 performance. Among splicing events commonly dysregulated in both cell lines, we identified some that are TDP-43-dependent also in human cells and show that inclusion levels of these alternative exons appear to be differentially altered in affected tissues of FTLD and IBM patients. We therefore propose that TDP-43 dysfunction, reflected in aberrant splicing, contributes to disease development but it does so in a tissue- and disease-specific manner.

neuroscience↗