HAL - RIIP
Not a member yet
17635 research outputs found
Sort by
Campylobacter, Salmonella et Blastocystis : contaminations et associations chez le poulet de chair en France
No full textInternational audienc
Physicochemical stability of a polysorbate-80-containing solvent compounded in the hospital pharmacy and used to reconstitute a biologic for nebulisation
No full textInternational audienc
Identification of metabolite biomarkers for pancreatic neuroendocrine tumours using a metabolomic approach
No full textInternational audienceAbstract Importance Metabolic flexibility, a key hallmark of cancer, reflects aberrant tumour changes associated with metabolites. The metabolic plasticity of pancreatic neuroendocrine tumours (pNETs) remains largely unexplored. Notably, the heterogeneity of pNETs complicates their diagnosis, prognosis, and therapeutic management. Objective Here, we compared the plasma metabolomic profiles of patients with pNET and non-cancerous individuals to understand metabolic dysregulation. Design, setting, participants, intervention and measure Plasma metabolic profiles of 76 patients with pNETs and 38 non-cancerous individuals were analyzed using LC-MS/MS and FIA-MS/MS (Biocrates AbsoluteIDQ p180 kit). Statistical analyses, including univariate and multivariate methods, were performed along with the generation of receiver operating characteristic (ROC) curves for metabolomic signature identification. Results Compared with non-cancerous individuals, patients with pNET exhibited elevated levels of phosphoglyceride metabolites and reduced acylcarnitine levels, indicating an upregulation of fatty acid oxidation (FAO), which is crucial for the energy metabolism of pNET cells and one-carbon metabolism metabolites. Elevated glutamate levels and decreased lipid metabolite levels have been observed in patients with metastatic pNETs. Patients with the germline MEN1 mutations showed lower amino acid metabolites and FAO, with increased metabolites related to leucine catabolism and lipid metabolism, compared to non-MEN1 mutated patients. The highest area under the ROC curve was observed in patients with pNET harbouring MEN1 mutations. Conclusion and relevance This study highlights the distinct plasma metabolic signatures of pNETs, including the critical role of FAO and elevated glutamate levels in metastasis, supporting the energy and biosynthetic needs of rapidly proliferating tumour cells. Mapping of these dysregulated metabolites may facilitate the identification of new therapeutic targets for pNETs management
Conception d'espaces latents interprétables pour la médecine de précision : applications en neuroimagerie multimodale
No full textThe rise of deep learning and the increasing availability of large-scale medical data—diagnoses, treatments, genetic information, clinical histories, imaging, and textual reports—are profoundly transforming biomedical research. This abundance opens unprecedented opportunities for precision medicine: identifying patients with similar profiles, anticipating disease progression, and tailoring individualized care. However, effectively leveraging these data remains a major challenge. Their heterogeneity and complexity require the design of representations capable of accurately reflecting similarities and differences between individuals, while remaining clinically actionable. In this context, latent spaces provide a particularly relevant methodological framework. By projecting complex medical data into lower-dimensional spaces, they allow information to be condensed while preserving its essential structure. The development of structured and interpretable latent spaces, specifically adapted to the medical domain, constitutes the core of this doctoral research. Moreover, as medical data are inherently heterogeneous, the progressive integration of multimodal sources represents a natural extension of this approach. The first part introduces the conceptual foundations of latent spaces and the main methods for their construction, initially in a unimodal and then in a multimodal framework. Particular attention is given to their structuring, which is essential to balance fidelity, generalization, and clinical interpretability. A brief review of existing applications in neuroimaging complements this context. The second part builds on the BrainAGE model to demonstrate that a supervisedly learned latent space can encode clinically relevant dimensions. This approach enabled the identification of patient subgroups in early onset Alzheimer’s disease independently of conventional clinical features, highlighting the potential of latent representations to capture the complexity of phenotypic dimensions. The third and fourth parts present PatientSpace, a methodological framework for constructing interpretable latent spaces in the context of neurodegenerative diseases. Initially applied to magnetic resonance imaging (MRI) of patients with frontotemporal dementia (FTD), PatientSpace enabled the identification of coherent subgroups while providing clinical interpretability of the clusters. The approach was then extended to multiple dementias, including Alzheimer’s and FTD, and enriched with a multimodal dimension by integrating both MRI and positron emission tomography (PET) data. These studies demonstrated the robustness of the method and its ability to capture inter-individual variability while remaining clinically interpretable. Finally, the fifth part explores a more heterogeneous multimodal setting, applied to predicting three-month functional outcomes after stroke in patients eligible for thrombectomy. By combining MRI sequences (FLAIR and DWI), radiology reports, and clinical data, latent spaces were structured to highlight the role and interactions of each modality in predicting the three-month modified Rankin Scale (mRS). Integration of these sources through attention and gating mechanisms not only improved predictive performance but also allowed quantification of the specific contribution of each modality, providing interpretations directly actionable in clinical practice.Overall, these works advance the exploitation and structuring of latent spaces in medical contexts, paving the way for interpretable and clinically relevant representations, and supporting the development of novel approaches for diagnosis, prognosis, and personalized medicine.L’essor du deep learning et l’accès croissant à des volumes massifs de données médicales — diagnostics, traitements, informations génétiques, antécédents cliniques, imagerie ou encore comptes rendus textuels — transforment profondément la recherche biomédicale. Cette abondance ouvre des perspectives inédites pour la médecine de précision : identifier des patients présentant des profils similaires, anticiper l’évolution des pathologies et adapter les prises en charge de façon individualisée. Néanmoins, exploiter efficacement ces données reste un défi majeur. Leur hétérogénéité et leur complexité exigent de concevoir des représentations capables de refléter fidèlement les similarités et dissimilarités entre individus, tout en restant exploitables dans un cadre clinique.Dans ce contexte, les espaces latents offrent un cadre méthodologique particulièrement pertinent. En projetant des données médicales complexes dans des espaces de dimension réduite, ils permettent de condenser l’information tout en préservant sa structure essentielle. Le développement d’espaces latents structurés et interprétables, adaptés au domaine médical, constitue le coeur de cette thèse. De plus, comme les données médicales sont par nature hétérogènes, l’intégration progressive de la multimodalité en représente un prolongement naturel. La première partie introduit les fondements conceptuels des espaces latents et les principales méthodes permettant de les construire, d’abord dans un cadre unimodal puis multimodal. Une attention particulière est portée à leur structuration, une condition essentielle pour concilier fidélité, généralisation et interprétation clinique. Une brève revue des applications existantes en neuroimagerie vient compléter cette mise en contexte. La deuxième partie s’appuie sur le modèle BrainAGE afin de démontrer qu’un espace latent appris de manière supervisée peut encoder des dimensions cliniquement pertinentes. Cette approche a permis d’identifier des sous-groupes de patients dans la maladie d’Alzheimer à début précoce, indépendamment des caractéristiques cliniques conventionnelles, mettant en évidence le potentiel des représentations latentes à saisir la complexité des dimensions phénotypiques. Les troisième et quatrième parties présentent le PatientSpace, un cadre méthodologique destiné à construire des espaces latents interprétables dans le contexte des maladies neurodégénératives. Appliqué initialement à l’imagerie par résonance magnétique (IRM) de patients atteints de démences fronto-temporales (DFT), le PatientSpace a permis d’identifier des sous-groupes cohérents tout en fournissant une explicabilité clinique des regroupements. L’approche a ensuite été étendue à plusieurs démences, notamment Alzheimer et DFT, et enrichie d’une dimension multimodale en intégrant à la fois des données IRM et de tomographie par émission de positons (TEP). Ces travaux ont démontré la robustesse de la méthode et sa capacité à capturer la variabilité interindividuelle tout en restant exploitable dans un cadre clinique. Enfin, la cinquième partie explore un cadre multimodal plus hétérogène, appliqué à la prédiction du pronostic fonctionnel à trois mois après un AVC chez des patients candidats à une thrombectomie. En combinant des séquences IRM (FLAIR et DWI), des comptes rendus radiologiques et des données cliniques, les espaces latents ont été structurés de manière à mettre en évidence le rôle et les interactions de chaque modalité dans la prédiction du mRS à 3 mois. L’intégration de ces sources via des mécanismes d’attention et de gating a non seulement amélioré la performance prédictive, mais aussi permis de quantifier l’apport spécifique de chaque modalité, fournissant ainsi une interprétation directement exploitable en pratique clinique [...
Impact of temperature on survival, development and longevity of Aedes aegypti and Aedes albopictus (Diptera: Culicidae) in Phnom Penh, Cambodia
No full textInternational audienceBackgroundAedes aegypti and Ae. albopictus are primary vectors of dengue virus in Cambodia, distributed throughout the country. Climate change is predicted to affect the relative density of these two species, but there is a lack of studies evaluating the impact of temperature on populations of these two species in this region. This study investigates the impact of temperature on the survival, development and longevity of Ae. aegypti and Ae. albopictus from populations collected in Phnom Penh, Cambodia.MethodsAedes aegypti and Ae. albopictus populations were collected in Phnom Penh. The experiment was conducted in a climatic chamber with temperatures ranging from 15 °C to 40 °C, with a 5 °C increment between each treatment. Bionomic parameters from the F2 egg hatching rate to the number of F3 eggs produced at each temperature treatment were measured.ResultsTemperature significantly influenced all life history traits of Ae. aegypti and Ae. albopictus. The highest egg hatching rates were observed at 25 °C for Ae. aegypti (97.97%) and 20 °C for Ae. albopictus (90.63%). Larvae of both species could not survive beyond the first stage at 40 °C. During immature stages, development time decreased at higher temperature (35 °C), but mortality was increased. Female longevity peaked at 25 °C for Ae. aegypti (66.7 days) and at 20 °C for Ae. albopictus (22.6 days), with males having significantly shorter lifespans. In addition, the optimal temperature for female survival is predicted higher in Ae. aegypti than in Ae. albopictus, at 27.1 °C and 24.5 °C, respectively. Wing length increased at lower temperatures, with Ae. aegypti consistently longer than Ae. albopictus at 15 °C and 35 °C. Blood-feeding rates were highest at 30 °C for Ae. aegypti (61.0%) and at 25 °C for Ae. albopictus (52.5%).ConclusionAedes albopictus appears better adapted to lower temperatures, whereas Ae. aegypti is better adapted to higher temperatures. Warmer temperatures accelerate mosquito development but also increased mortality and reduced adult longevity, which could influence their ability to transmit pathogens. These findings highlight the critical role of temperature in mosquito biology and emphasize the potential impact of climate change on dengue transmission dynamics in the future
Ultrasound‐Assisted Synthesis of Pyrazoline Derivatives as Potential Antagonists of RAGE‐Mediated Pathologies: Insights from SAR Studies and Biological Evaluations
No full textInternational audienceIn the context of age‐related disorders, the receptor of advanced glycation end products (RAGE), plays a pivotal role in the pathogenesis of these conditions by triggering downstream signaling pathways associated with chronic inflammation and oxidative stress. Targeting this inflammaging phenomenon with RAGE antagonists holds promise for interventions with broad implications in healthy aging and the management of age‐related conditions. This study explores the structure‐activity relationship (SAR) of pyrazoline‐based RAGE antagonists synthesized using an ultrasound‐assisted green one‐pot two‐steps methodology. Our investigation identifies phenylurenyl‐pyrazoline 2 g as a promising candidate, demonstrating superior efficiency compared to the reference antagonist Azeliragon (IC 50 =13 μM). Compound 2 g exhibits potent inhibition of the AGE2‐BSA/sRAGE interaction (IC 50 =22 μM) and favorable affinity in Microscale Thermophoresis (MST) assays (K d =17.1 μM), along with a favorable safety profile, with no apparent cytotoxicity observed in vitro in the MTS assay. These findings underscore the potential of pyrazoline‐derived RAGE antagonists as therapeutic agents for addressing age‐related disorders
A gut microbiome-kidney-heart axis predictive of future cardiovascular diseases.
No full textAbstract Cardiometabolic diseases (CMD) are on the rise globally with one billion people expected to suffer from obesity and 643 million from type 2 diabetes by 2030, of which one-third will likely develop chronic kidney disease and two-thirds will die from cardiovascular disease (CVD). However, the mechanistic and molecular drivers of the transition from health to disease remain elusive. Here, in 275 metabolically healthy individuals recruited to the MetaCardis study, we identify a gut microbiome-kidney-heart axis that is predictive of future cardiovascular events. This axis, as evidenced by the associations between gut microbial metabolism of phenylalanine and tyrosine with variations in both kidney functon(as measured by estimated glomerular filtration rate) and circulating pro-atrial natriuretic peptide concentration, shows a depletion pattern in metabolically unhealthy participants of the MetaCardis study (n = 1,602) indicating a loss of health-sustaining microbiome features with CMD progression. We then validate that microbial compounds from the phenylalanine and tyrosine pathways and their host co-metabolites act as mediators of the gut microbiome-kidney associations. Moreover, Mendelian Randomization analysis adds genetic evidence to suggest that the microbial mediator metabolites regulate host kidney function and vice versa. Finally, we demonstrate that plasma metabolites derived from the microbial metabolism of phenylalanine and tyrosine associate with incident CVD in the Canadian Longitudinal Study on Aging (n = 8,669). Collectively, our results depict the presence of a gut microbiome-kidney-heart axis in metabolically healthy individuals. Major aberrations of the gut microbiome as part of this axis throughout life may increase risk of CVD
Histo-radiological correlations in an early-stage model of a Parkinson's disease.
No full textInternational audienc
RETRACTED: Bilyy et al. Rapid Generation of Coronaviral Immunity Using Recombinant Peptide Modified Nanodiamonds. Pathogens 2021, 10, 861
No full textThe Journal retracts the article “Rapid Generation of Coronaviral Immunity Using Recombinant Peptide Modified Nanodiamonds” [...
Leveraging Sortase A Electrostatics for Powerful Transpeptidation Reactions
No full textInternational audienceSortase‐mediated transpeptidation is a powerful biochemical reaction to perform protein engineering. In this work, we leverage the unique electrostatic profile of sortase A pentamutant (SrtA‐5M) to improve SrtA‐5M‐mediated transpeptidations by incorporating short, charged peptidic modules into the substrates. Importantly, the reaction proceeds with a minimal excess of nucleophile and is fast and highly efficient in the low micromolar substrate concentration range. Electrostatic assistance eliminates the need for additives or complex substrate engineering strategies, thereby giving it a broad scope. Our findings also provide fundamental insights into the influence of substrate charge on SrtA‐5M activity, paving the way for further optimization of sortase A‐catalyzed transpeptidation reactions