Science day – April 9, 2026

1 – What is the role of the glucocorticoid receptor in skeletal muscles in metabolic and pancreatic adaptations to insulin resistance?

Emma Rousseau

Team: Lipodystrophies, Metabolic and Hormonal Adaptations, and Aging UMRS_938, Saint-Antoine Research Center (CRSA)

E.Rousseau, N.Roblot, A.Liboz, M.Mendès, C.Beaupère, B.Fève, G.Guillemain, B.Blondeau

Introduction and objectives: Pancreatic beta cells possess an adaptive capacity thatallows them to increase their number and function in order to produce more insulin to meet increased demand, such as that observed in insulin resistance. However, the mechanisms involved are poorly understood. To study the signals enabling this plasticity, the team developed a mouse model of insulin resistance induced by a
glucocorticoid, corticosterone (CORT). In this model, an increase in mass, through proliferation and neogenesis, and in beta cell function was observed. Data from the laboratory suggest that one or more factors derived from skeletal muscle may be involved in this adaptation.

Material and methods: To understand the role of skeletal muscle in these adaptations, the team developed a mouse model with a glucocorticoid receptor that is specifically and inducibly knocked out in skeletal muscle. The aim of my project is therefore to characterize, in control and knockout mice, whether treated or untreated: 1) the metabolic phenotype, 2) the beta cell phenotype.

Results: We demonstrated that in response to CORT, inactivation of the glucocorticoid receptor in skeletal muscle:

i) improves glucose tolerance in mice and only partially improves insulin sensitivity;

ii) partially affects the adaptation of beta cell mass and function in response to CORT;

iii) leads to the loss of beta cell neogenesis.

Conclusion: We identify the glucocorticoid signaling pathway in skeletal muscle as a major regulator of systemic insulin sensitivity and as an essential mediator of beta cell adaptation in response to chronic glucocorticoid treatment.

2 – Quality of life of non-HIV-associated lipodystrophy syndromes patients: A systematic review

Inès BELALEM

Team: Lipodystrophies, Metabolic and Hormonal Adaptations, and Aging UMRS_938, Saint-Antoine Research Center (CRSA)

Belalem Inès1,2,3,4, Rivet Clémence, Vuillaume Philippine4, Vatier Camille2,3,4, Leblanc Judith1,5
1 Sorbonne University, INSERM, Pierre Louis Institute of Epidemiology and Public Health
2 Assistance Publique – Hôpitaux de Paris (AP-HP), Hôpital Saint-Antoine, Department of Endocrinology
3 Insulino-secretion and insulino-sensitivity rare disease reference center, European Reference Network for Rare Endocrine Conditions (Endo-ERN Assistance – Publique Consortium), Paris, France
4 Sorbonne University, INSERM, Saint Antoine Research Center
5 Assistance Publique – Hopitaux de Paris, Hôpital Saint Antoine, Clinical Research Platform Paris-East, Paris, France

Introduction and objectives: Background: Non-HIV-lipodystrophy syndromes (LS) are rare diseases characterized by a partial or total loss of adipose tissue and are associated with significant cardiometabolic, morphological and reproductive complications. Recently, quality of life (QoL) in LS has gained increasing attention in clinical research and practice.

This systematic review aims to provide a comprehensive overview of the impact of LS on patients’ QoL , to identify interventions proposed to improve QoL, and to describe the tools and measures used to assess the QoL in this population.

Methods: A literature search was conducted in Pubmed, Embase, Cochrane, CINAHL and Web of Science. Eligible articles were published in English or French between January 1, 1990 to September 12, 2025. Two authors independently screened titles, abstracts, and full texts. Discrepancies were resolved by a third author. Data were extracted using a standardized form. A narrative synthesis will be performed, and the risk of bias will be assessed using the Joanna Briggs Institute critical appraisal tools.

Preliminary results: The search identified 4,271 records. After duplicate removal and screening, 10 studies met the inclusion criteria (3 cohort studies, 3 cross-sectional studies, 1 case-control study and 3 qualitative studies).

QoL domains assessed included global health-related QoL, mental health, social and psychosocial well-being, pain and fatigue, body image, and sexual health.

Global health-related QoL was impaired in patients with LS compared to general population (n=3). Mental health outcomes revealed frequent anxiety and depressive symptoms, with negative self-perception and unmet mental needs (n=6). Major social impairments were reported, including limited access to work or school, social isolation and experiences of discrimination (n=4). Pain and fatigue were assessed in half of the studies (n=5), and described as chronic and severe, representing a major burden. Body image dissatisfaction was reported in nearly all studies (n=9), often associated with stigmatization and discrimination. Sexual health was addressed in only one study, in which 42% of respondents reported severe dissatisfaction (n=1).

Perspectives: This systematic review highlights the multidimensional impact of LS on patients’ QoL. Further steps are ongoing to complete this study: data synthesis and risk of bias assessment.

3 – miR-100-5p controls postprandial triglyceride response by targeting PCSK9

Amandine VANDUYSE

Team: Cellular and Systemic Lipid Metabolism in Cardiometabolic Diseases UMRS_1166-ICAN, Cardiovascular and Metabolic Diseases

Amandine VANDUYSE1, Alexandre MOTTE1, Carolina NEVES1, Rita DACLAT1, Sophie GALIER1, Olivier BLUTEAU1,2, Clément MATERNE1, Eric FRISDAL1, Hervé DURAND1, Philippe GIRAL1, Joe-Elie SALEM1,3, Jean-Marc LACORTE1,4, Cedric LE MAY5, Wilfried LE GOFF1, Philippe LESNIK1, Maryse GUERIN1
1 Sorbonne University, Inserm, UMR_S1166, Research Institute of Cardiovascular Disease, Metabolism 9 and Nutrition, Faculté de Médecine – Hôpital Pitié-Salpêtrière, Paris, France
2 Hôpital Pitié-Salpêtrière, Assistance Publique Hôpitaux de Paris (AP-HP), Department of Biochemistry, Obesity and Dyslipidemia Genetics Unit, Paris, France
3 Centre d’Investigation Clinique Paris-Est CIC-1901, Hôpital de la Pitié-Salpêtrière, AP-HP, Paris, France
4 Hôpital Pitié-Salpêtrière, AP-HP, Service de Biochimie Endocrinienne et Oncologique, Paris
5 Nantes Université, CHU Nantes, CNRS, INSERM, l’institut du thorax, Nantes, France

Background: Elevated postprandial hypertriglyceridemia (PP-HTG) is a significant risk factor for development of cardiovascular diseases, however, the mechanisms underlying its exaggerated rise remains poorly understood. MicroRNAs (miRs) are known to be implicated in the regulation of lipid metabolism, thus identifying them as potential key players. We presently investigated whether miRs may control postprandial triglyceride (PP-TG) response.

Methods: Postprandial changes in circulating miR expression as a function of the degree of postprandial TG response were evaluated in non-dyslipidemic healthy subjects (n=32). The impact of miR-100-5p on hepatic gene expression was evaluated in differentiated Caco2 and HepG2 cells by analysis of hepatic transcriptome (RNAseq), western blot and ELISA. In vivo studies were conducted in C57BL/6J mice overexpressing mimic miR-100-5p.

Results: Postprandial variation in circ-miR-100-5p levels inversely correlate with PP-TG response. Cir-miR-100-5p was preferentially associated with TGRL particles of intestinal origin in subjects exhibited a low PP TG response. Differential analysis of transcriptome from HepG2 cells transfected with a mimic miR-100-5p or scramble mimic miR as control allowed us to identify PCSK9 as a down regulated gene. Overexpression of miR-100-5p in HepG2 cells significantly decreased PCSK9 mRNA levels by 52% (p<0.0001), cellular protein content by 28 % (p<0.0001) as well as PCSK9 secretion by 39% (p<0.0001). In vivo systemic delivery of mimic miR-100-5p induced a two-fold reduction (p<0.0001) on PP-TG in mice, such effect being abolished by blocking the circulating form of PCSK9 with alirocumab. Finally, we revealed a significant inverse relationship between circulating miR-100-5p expression levels and both PCSK9 levels and the magnitude of postprandial hypertriglyceridemia.

Conclusion: Taken together, our observations reveal that miR-100-5p regulates postprandial hypertriglyceridemia by targeting PCSK9, thus enhancing hepatic triglyceride-rich lipoproteins (TGRL) uptake. Our findings allow us to propose circ-miR-100-5p as a potential biomarker for early identification of subjects at high cardiovascular risk, prior to appearance of classical clinical features of metabolic disorders.

4 – The winding path towards cellular models of imprinting defects

Eloïse GIABICANI

Team: Pathophysiology of Fetal Growth: The IGF System and Parental Imprinting UMRS_938, Saint-Antoine Research Center (CRSA)

1-Sorbonne Université, INSERM, Centre de Recherche Saint Antoine, 75012, Paris, France.

2-APHP, Hôpital Armand Trousseau, Explorations Fonctionnelles Endocriniennes, 75012, Paris, France.

Summary: There is no need to prove that disorders models, either animal and/or cellular, are a substantial help
in understanding pathophysiology of human genetic defects together with improving therapeutic
management of patients carrying these disorders. In imprinting disorders in whom epigenetic defects
are involved, developing models with manipulation of epigenome is challenging. Nevertheless,
recently different teams have succeeded using adaptations of CRISPR/Cas9 technology recruiting
demethylases for epigenome editing. To date, the relevance of these models is not perfect due to offtarget demethylation and phenotype not fully concordant with the one observed in human. In our
team we are interested in imprinting disorders impacting foetal growth in whom loss of methylation
at imprinting control regions (ICR) is a frequent, namely Silver-Russell syndrome (SRS), Temple
syndrome (TS14) and Beckwith-Wiedemann syndrome (BWS). Since a decade, stem cells have been largely and successfully used to model genetic defects. In line with this trend, we aimed at developing cellular models of these syndromes from human induced pluripotent stem cells (iPSCs) and dental pulp stem cells (DPSCs).
IPSCs have the capacity to differentiate into all somatic lineages, however, concerning imprinting
disorders aberrant hypermethylation at ICR can occur during culture. We therefore developed culture
conditions based on the synergistic effect of vitamin C and hypoxia to obtain iPSCs lines with balanced
methylation at most ICRs. We are now differentiating these iPSCs in cells of interest in SRS, TS14 and
BWS such as chondrocyte of the growth plate or adipocyte.
Unlike iPSCs, DPSC express a much more stable methylation at ICRs in culture and during
differentiation. As multipotent mesenchymal stem cells they theoretically be differentiated in vitro
into different cell lines of interest such as chondrocyte and adipocytes. We first confirmed the potential of such cells in modelling imprinting defect by osteogenic differentiation and we are now targeting adipogenic and chondrogenic differentiation. Nevertheless, the capacity of differentiation in cell-types distant from osteogenic lineage is often harder to achieve with DPSC than iPSCs.
This series of experiments inclines us to think that DPSCs and iPSCs are two promising cellular models
of imprinting disorders but are still demanding in term of cost, lab work and expertise.

5 – Novo Nordisk: Innovations and outlook

Melissa AKERIB

NOVO NORDISK

1 – From myocarditis to inherited cardiomyopathy: when inflammation unmasks genetic vulnerability

Matthieu PROUST

Department : Cardiology, Pitié-Salpêtrière Hospital

Matthieu Proust, MD1,2

1- Sorbonne Université, Assistance Publique – Hôpitaux de Paris (AP-HP), Hôpital La Pitié-Salpêtrière, Département de Cardiologie, Paris, France

2- INSERM UMRS 1166-ICAN (Institute of CardioMetabolism and Nutrition), Sorbonne Université, Paris, France.

Summary: Inherited cardiomyopathies are genetic myocardial disorders characterized by incomplete penetrance and marked phenotypic heterogeneity, encompassing entities such as arrhythmogenic (ACM) and dilated cardiomyopathies (DCM). Emerging data suggest that their clinical course may be punctuated by inflammatory flares, corresponding to episodes of acute myocarditis (AM). Environmental triggers, including viral infections and intense physical activity, appear to play a contributory role in disease expression and progression. In addition, immune-mediated mechanisms may be intricately involved in their pathophysiology.

These studies aim to characterize the clinical features, phenotypic spectrum, and prognostic implications of patients presenting with overlapping entities at the intersection of inflammation and inherited cardiomyopathies. This integrative approach is structured around three principal pathophysiological axes: infectious, immunologic, and genetic determinants.

The analysis was based on two ambispective observational cohorts of patients hospitalized at La Pitié-Salpêtrière Hospital: the ACORE registry, comprising 140 patients with ACM, and the AMPHIBIA registry, including 590 patients hospitalized for AM. Comprehensive data collection encompassed clinical characteristics, laboratory and immunologic parameters, multimodality imaging, and genetic profiles.

Analysis of the genetic architecture of AM revealed the presence of pathogenic or likely pathogenic variants in cardiomyopathy-associated genes in 13% of cases, increasing to up to 18% among patients with recurrent AM, predominantly involved genes encoding desmosomal and sarcomeric proteins. These forms are associated with distinct clinical presentations and cardiac magnetic resonance patterns. dysfunction. Comprehensive etiological evaluation of AM highlights the existence of overlapping forms involving infection-related, immune-mediated, and genetically predisposed mechanisms. Nevertheless, despite extensive diagnostic workup, nearly 50% of AM cases remain truly idiopathic.

Recurrent AM was also specifically investigated. Although overall prognosis does not appear to differ significantly from nonrecurrent forms, distinct phenotypic patterns emerge. The risk of recurrence increases with the number of prior episodes, rising from 13% after the first episode to as high as 45% after the fourth. Notably, recurrent episodes tend to mirror the initial presentation in terms of severity. Identified predictors of recurrence include a family history of cardiomyopathy, prior episodes of AM or pericarditis, immunological forms, and the presence of septal or anterior LGE distribution. Recurrent AM may fall within distinct nosological frameworks, including cardiac involvement of systemic diseases (e.g., connective tissue disorders or sarcoidosis), inflammatory cardiomyopathies, or ACM.

The prognostic implications of these infectious, immunologic, and genetic phenotypes were evaluated. Immunologic and infection-related forms were associated with a higher risk of hemodynamic compromise, whereas genetic forms carried an increased risk of ventricular arrhythmias. Recurrent AM were more frequent in both immunologic and genetically predisposed myocarditis.

Finally, unsupervised phenotypic clustering further delineated distinct patient groups: uncomplicated AM in young men, immunologic and infection-related AM associated with hemodynamic complications, and AM with arrhythmic manifestations, often linked to an underlying genetic predisposition.

In conclusion, these studies highlight a continuum between myocardial inflammation and inherited cardiomyopathy, driven by three principal mechanisms – infection, immune-mediated processes, and genetic susceptibility – resulting in divergent clinical trajectories and prognoses. Identification of these phenotypes is critical to enable personalized risk stratification and guide targeted therapeutic strategies.

2 – Early-Life Cardiac Remodeling Induced by Maternal Obesity: From Postnatal Maturation to Adult Dysfunction

Dounia FARHI

Team: Cellular and molecular plasticity in cardiovascular diseases UMRS_1166-ICAN, Cardiovascular and Metabolic Diseases

Introduction: Cardiovascular diseases remain the leading cause of death worldwide, responsible for nearly one-third of al deaths. Obesity, a major risk factor, has increased dramatica ly among women of reproductive age, raising concerns about its impact on maternal and offspring health. Children born to obese mothers exhibit higher rates of cardiovascular events and mortality in adulthood. Maternal obesity has been associated with congenital heart defects, cardiometabolic disorders, and abnormal placental vascularization. While fetal cardiac development is wel characterized, cardiomyocyte maturation continues after birth. This process involves the structural organization of intercalated discs and lateral membranes, which are essential for celular cohesion. Disruptions during this prolonged maturation period can lead to arrhythmias and cardiac dysfunction. Given the rise in maternal obesity, it is critical to investigate how obesogenic environments affect postnatal heart development.

Objective: Our goal was to define how maternal obesity reshapes the postnatal heart— assessing its impact on cardiac structure, function, and gene expression to uncover the long-term consequences for offspring health.

Method: To investigate this, we established a murine model in which female were fed a high-fat diet (HFD) for two months prior to conception and throughout gestation and lactation. Offspring were transitioned to a standard diet at weaning (P20) to closely recapitulate the human postnatal nutritional environment.

Results: Our data reveal that offspring of obese mothers initialy exhibit enhanced cardiac function during early development. However, this early advantage reverses in adulthood, where impaired heart function emerges. Structural analysis identified developmental defects in cardiomyocytes—particularly at the intercalated discs— alongside premature hypertrophy and metabolic dysregulation. These findings suggest that early metabolic stress may accelerate cardiomyocyte maturation, ultimately contributing to long-term cardiac dysfunction.

Conclusion: Maternal obesity induces early metabolic and structural alterations in cardiomyocytes, leading to transiently enhanced cardiac function during development. This accelerated maturation is accompanied by intercalated disc abnormalities, hypertrophy, and metabolic remodeling in adulthood. Colectively, these early adaptations predispose offspring to cardiac dysfunction in adulthood.

3 – Long-Term Prediction of Atrial Dysfunction Severity from Echocardiography Using Spatio-Temporal Deep Learning

Kimia SADDREDINI

Team: Platform I/O

Summary: Predicting long-term cardiac rhythm outcomes from baseline imaging remains a major challenge in atrial fibrillation (AF) management. In this study, we propose a neural network architecture, « AF-Echo », based on spatio-temporal convolution and video self-attention, to predict cardiac rhythm status two years after initial diagnosis using transthoracic echocardiography (TTE) videos. We analyze a private, fully anonymized cohort of patients hospitalized for AF and followed over a two-year period. For each patient, baseline four-chamber (4CH) TTE videos in DICOM format were collected alongside comprehensive clinical and echocardiographic parameters. The binary ground-truth label was defined by expert electrocardiogram (ECG) assessment at the 24-month follow-up (M24), distinguishing sinus rhythm (healthy) from AF (non-healthy). The model is trained in a multi-task learning setting, jointly predicting rhythm outcome and auxiliary cardiological parameters. Our quantitative results demonstrate the feasibility of long-term rhythm prediction from baseline echocardiography, highlighting the potential of deep video networks for early risk prediction in AF management.

4 – Prevalence and prognostic significance of clonal hematopoiesis of undetermined significance in premature coronary artery disease

Ghilas RAHOUAL

Team: Atherothrombosis and Applied Pharmacology UMRS_1166-ICAN, Cardiovascular and Metabolic Diseases

Summary: Premature coronary artery disease, which manifests before the age of 45, is an aggressive chronic condition characterized by rapid progression to multivessel disease and an associated 10-year mortality rate of 20%. The incidence of premature coronary artery disease is steadily increasing. In fact, the overall proportion of hospital admissions for myocardial infarction among young patients has risen steadily, reaching 32% between 2010 and 2014. This serious condition is responsible for a high mortality rate among the young population, which is the only demographic group in which no clear decline in cardiovascular mortality has been observed over the past twenty years. Current treatments appear to have limited efficacy in managing the disease’s progression.

Recent advances in high-throughput sequencing technology have enabled the identification, in individuals without obvious hematological abnormalities, of clonal hematopoiesis of undetermined potential (CHIP). This clonal proliferation, common in the elderly (>20% in those over 70 years of age), constitutes an independent risk factor for atherosclerosis, thrombosis, and heart failure. Experimental data suggest that acquired mutations in the DNMT3A, TET2, ASXL1, and JAK2 genes promote systemic inflammation and are causally linked to the accelerated development of atherosclerosis.

Changes in the inflammatory environment mediated by CHIP that affect the formation and/or stability of atherosclerotic plaques may play a role in the onset and progression of premature coronary artery disease.

The creation of a DNA bank within the world’s largest cohort of patients with premature coronary artery disease, with a 20-year follow-up, allows us to explore the mechanisms leading to the genesis of early atherothrombosis, the pathophysiology of which remains poorly understood. We hypothesize that a multimodal assessment of residual risk driven by inflammation will improve risk stratification in patients with

premature coronary artery disease and facilitate the exploration of new therapeutic approaches.

The results of this study could lead to a better understanding of the mechanisms leading to early atherosclerosis and thrombosis and to the phenotyping of different disease manifestations. The potential pathophysiological role of CHIP in this disease, via a pro-inflammatory action, could provide a pathophysiological basis for the development of

new targeted, anti-inflammatory therapeutic approaches aimed at improving the prognosis and quality of life of these young patients, whose theoretical life expectancy is 40 years at disease onset.

This study could support the recent findings demonstrating the benefits of anti-inflammatory treatment targeting IL-1β and IL-6 in coronary artery disease.

5 – Quantitative imaging of the left atrium in atrial fibrillation

Mikael LAREDO and Jonas LEITE

Team: LIB

1 – Hepatocyte Ballooning and Senescence in Metabolic Steatohepatitis: Multi-Omic Characterization and Therapeutic Targeting

Antoine LEFEBVRE

Team: Biliary and Fatty Liver Diseases – UMR 938, Saint-Antoine Research Center (CRSA

Antoine Lefebvre^1,2, Valeria Pistorio^1,2, Virginie Steunou^1,2, Vlad Ratziu^2,3, Jérémie Gautheron^1,2

¹Inserm UMR_S938, Saint-Antoine Research Center, Sorbonne University, Paris, France;

²Foundation for Innovation in Cardiometabolism And Nutrition, IUH-ICAN, F-75013 Paris, France;

³Inserm, UMR_S1166, F-75013, Sorbonne University, Paris, France.

Summary: Metabolic dysfunction-associated steatohepatitis (MASH) is a major public health concern. It represents the inflammatory and progressive form of metabolic dysfunction-associated steatotic liver disease (MASLD), which is associated with metabolic disorders such as overweight, diabetes, and dyslipidemia. This disease affects up to 25% of the population, and a significant proportion of patients will progress to MASH, which can lead to fibrosis, cirrhosis, and liver cancer. To date, no effective treatments are available for this condition, underscoring the need to identify new biological targets.

In patients with MASH, abnormally enlarged liver cells, known as “ballooned” hepatocytes, are commonly observed. These cells are highly stressed and dysfunctional, and their presence correlates with more advanced stages of MASLD. In parallel, studies have shown that cellular senescence also plays a role in the progression of MASLD. Numerous similarities between these two cellular states have been identified, leading us to hypothesize that ballooned hepatocytes may, in fact, represent senescent hepatocytes. To test this hypothesis, in vitro cellular models reproducing metabolic stress conditions (hepatocyte ballooning and senescence) will be developed, enabling a detailed comparison through multi-omics analyses. In addition, we aim to investigate the impact of these cells on fibrosis and inflammation.

Finally, these findings will be validated using an in vivo model of MASH. A senescence model based on a cell cycle inhibitor has already been established and exhibits all characteristic markers of senescence. Two additional in vitro models currently under development: one combining lipid exposure with senescence induction, and another designed to induce hepatocyte ballooning.

2 – Exploring the involvement of macrophage phosphatidylethanolamine in obesity-associated metaflammation

Dominique GUERRINI

Team: SLIM, UMR 1166

Aims: Adipose tissue macrophage (ATM)-mediated inflammation is a major component in the development
of metabolic disorders associated with obesity and diabetes. Importantly, lipid membrane remodeling
in ATMs plays a key role in their functions and inflammatory phenotype. Numerous studies reported
alterations of phosphatidylethanolamine (PE) metabolism in metabolic diseases; however,
pathophysiological consequences of this observation are unknown. The aim of our project isto explore
the impact of PE metabolism, by modulating PCYT2 (the key enzyme in the PE biosynthesis pathway),
on lipid homeostasis and inflammatory response in macrophagesin the context of metabolic disorders
associated with obesity.

Methods: Consequences of alterations of PE metabolism were evaluated using three different approaches:
(i) RNAseq of ATM from obese patients after bariatric surgery, (ii) transcriptomic and lipidomic
analyses of ATM from mice fed hyperlipidemic diets, and (iii) culture of primary human monocytederived macrophages(HMDM) or THP-1 cells metabolically activated by palmitate, insulin and glucose
where PCYT2 expression was inhibited by siRNA.

Results: (i) RNAseq analysis of ATM from obese patients revealed that PCYT2 expression was correlated with
metabolic and inflammatory pathways. (ii) Furthermore, in hyperlipidemic mice, we found elevated
concentrations of PE in ATM which were strongly associated with insulin-resistance. (iii) Upon PCYT2
inhibition in macrophages, PE abundance decreased, leading to profound membrane lipid remodeling
and reduction of lipid storage as well as an alteration of the anti-inflammatory omega-3/omega-6
fatty-acids ratio. Under these conditions, PCYT2 inhibition was negatively correlated with oxidative
phosphorylation and inflammatory pathways. These modifications were accompanied by a decrease
of ATP production rate and reactive oxygen species production.

Conclusions: PE may play a central role in regulating both inflammatory and metabolic responsesin ATM. Our future
work aims to decipher mechanisms, governed by PCYT2, linking meta-inflammation in ATM to
metabolic disorders associated with obesity and insulin response

3 – Microbiota-Derived 12-Ketolithocholic Acid Drives Atherogenesis through Induction of Pro-inflammatory T Helper 1 Cells

Kaidi ZHANG

Team: SLIM, UMR 1166

Authors: Kaidi Zhang1, Hua Wang1, Carolina Neves1, Marie Lhomme2, Ashna Fathima3, Tong Ye1, Trinath Jamma3, Resist-PP Consortium#, Maryse Guerin1, Wilfried Le Goff1, Joe-Elie Salem4,5, Antonio Gallo1,5, Pukar Kc1, Philippe Lesnik1

1 Sorbonne Université, Inserm, UMR_S1166, Research Institute of Cardiovascular Disease, Metabolism and Nutrition, Faculté de Santé – Hôpital Pitié-Salpêtrière, Paris, France.

2 Foundation for Innovation in Cardiometabolism and Nutrition (ICAN), ICAN OMICS, Paris, France.

3 Birla Institute of Technology and Science, Pilani, Hyderabad Campus, Jawahar Nagar, Medchal, Hyderabad, Telangana, India.

4 Centre d’Investigation Clinique Paris-Est CIC-1901, Hôpital de la Pitié-Salpêtrière, AP-HP, Paris, France.

5 Hôpital Pitié-Salpêtrière, Assistance Publique Hôpitaux de Paris (AP-HP), Paris, France.

#RESIST-PP Consortium : Philippe Lesnik, Thomas Kufer, Stephan C. Bischoff, Mathias Chamaillard, Pablo Pelegrin, Maryse Guerin, Martin Larsen, Maharajah Ponnaiah, Joe-Elie Salem.

Background and Aims: As metabolic derivatives of cholesterol and the microbiota, secondary bile acids have been shown to modulate immune processes by influencing T-helper (Th) 17 and T-regulatory (Treg) cells. Notably, 12-keto lithocholic acid (12-ketoLCA), a structurally distinct secondary bile acid derived from deoxycholic acid (DCA), is present in the plasma and gut at nanomolar and micromolar concentrations, respectively. It has also been associated with various pathological states, including cardiovascular disorders. Hence, we sought to investigate whether 12-ketoLCA directly contributes to atherosclerotic lesion progression and to elucidate the potential underlying mechanisms.

Methods: Ldlr-/- mice fed a 1% cholesterol diet were treated daily with 12-ketoLCA, DCA (negative control) or vehicle via oral gavage for 10 days (short-term/immune phenotyping) and 8 weeks (long-term/atherosclerosis development). Cholesterol and triglyceride levels were monitored throughout the entire treatment period. Aortic root lesions were assessed using Oil-Red O staining, while immune phenotypes in various tissues were evaluated by flow cytometry. Bile acid levels were determined using LC-MS based approaches, and microbiota composition was measured by 16S sequencing.

Results: Following 8 weeks of bile acids administration in mice, 12-ketoLCA treatment exhibited a 76% increase in aortic lesion area compared to DCA (p=0.0179) and a 61% increase relative to vehicle controls (p=0.0014), and a slight increase in total cholesterol was only observed within long-term bile acids treatment group. Notably, 12-ketoLCA treated mice exhibited elevated CD4+IFNγ+ Th1 cell populations in mesenteric lymph nodes (p=0.0356), along with monocyte and neutrophil were unaffected among 8 weeks treatment. In the short-term experiment, 12-ketoLCA initiates Th1 cell populations in peripheral blood. In vitro, 12‑ketoLCA treated bone marrow derived dendritic cells reprogrammed naïve CD4+ T cells to a Th1 phenotype dependent on the vitamin D receptor, while direct treatment of T cells alone had no effects. Clinically, in familial hypercholesterolemia cohort, 12-ketoLCA levels were significantly higher in patients with carotid plaques than in those without. Meanhwhile, the 12-ketoLCA level correlated with an enhanced Th1 signature in dyslipidemic cohort, with its levels being modulated by statin therapy.

Conclusion: In summary, the data demonstrate that 12‑ketoLCA accelerates atherosclerotic plaque growth by reprogramming dendritic cells to polarize the immune system toward a Th1dominant, pro-inflammatory state.

4 – hIAPP aggregation and toxicity are modulated in vitro by other amyloid proteins

Guillaume FOURNIER

Team: Lipodystrophies, Metabolic and Hormonal Adaptations, and Aging UMRS_938, Saint-Antoine Research Center (CRSA)

Guillaume Fournier1,2, Cécile Feuillie1, Lucie Khemtemourian1, Ghislaine Guillemain2
1 Univ. Bordeaux, CNRS, Bordeaux INP, CBMN, UMR 5248, F-33600 Pessac, France. 2Sorbonne Université, Institut Hospitalo-Universitaire, INSERM UMR_S938, Institute of Cardiometabolism and Nutrition (ICAN), Centre de Recherche de St-Antoine (CRSA), 27 Rue de Chaligny, 75012, Paris, France.

Summary: Type 2 diabetes is characterized by the progressive loss of pancreatic β-cell function, associated with amyloid deposits formation in the islets of Langerhans. These toxic aggregates contain human Islet Amyloid Polypeptide (hIAPP), a peptide co-secreted with insulin. During amyloid formation, soluble monomers progressively assemble into fibrils through distinct nucleation and growth steps. Primary nucleation corresponds to the initial oligomerization of monomers into small aggregates, which grow after by elongation (addition of monomers to the fibril ends). Existing fibrils can also accelerate the aggregation process through secondary nucleation (side chain elongation) at their surface or by fragmentation, a step which will generate new small aggregates, able to grow.

In vivo, hIAPP mechanisms are poorly understood, but in vitro experiments have shown that aggregation can be modulated by different factors. Interestingly, analyses of pancreatic aggregates from diabetic patients revealed that within the pancreatic islet, hIAPP colocalize with other amyloid proteins (amyloid-β, which is only produced in the brain, and α-synuclein, respectively involved in Alzheimer’s and Parkinson’s diseases). These observations suggest that different amyloid proteins can interact through cross-seeding events. The questions we then asked is whether these cross-interactions could influence and even accelerate hIAPP aggregation and toxicity?

We first investigated the in vitro cross-fibrillation of hIAPP in the presence of seeds (little fibrils) from other amyloid-forming proteins: Aβ, α-synuclein, and H18R-hIAPP (synthetic mutant peptide less prone to aggregate). If the morphology of the fibrils is identical in each condition, hIAPP fibril formation is strongly accelerated by α-synuclein and hIAPP, mildly by Aβ and not by H18R-hIAPP. Finally, kinetic analysis was used to elucidate the aggregation mechanism and suggests different aggregation mechanisms, with secondary nucleation or fragmentation dominating depending on the type of seeds used.

To conclude, amyloid aggregates accelerate hIAPP fibril formation without changing fibril morphology, but kinetic analysis indicates that different amyloid seeds appear to drive hIAPP aggregation via distinct pathways. As perspectives, we will investigate the role of cross-fibrils on cellular toxicity and determine their atomic structure by cryo-electron microscopy to assess cross-seeding.

5 – Prevention of the harmful effects of HIV integrase inhibitors on adipose tissue by incretin receptor agonists : study of the metabolic and immune components  

Wided METHENNI

Team: Lipodystrophies, Metabolic and Hormonal Adaptations, and Aging UMRS_938, Saint-Antoine Research Center (CRSA)

6 – Impact of an alternating high fat diet on atrial remodeling

Raoul MANUEL

Team: Cellular and molecular plasticity in cardiovascular diseases UMRS_1166-ICAN, Cardiovascular and Metabolic Diseases

7 – Role of the PDGFRalpha pathway in obesity-induced hypertension

Erika LAURO

Team: Cellular and molecular plasticity in cardiovascular diseases UMRS_1166-ICAN, Cardiovascular and Metabolic Diseases

Introduction: Metabolic diseases are associated with small artery remodeling and increased vascular resistance, contributing to arterial hypertension. This remodeling involves structural reorganization of the vascular wall, including smooth muscle cell (SMC) expansion. Platelet-derived growth factor (PDGF) signaling is a key regulator of vascular remodeling, and elevated circulating PDGF-AA levels have been reported in obesity and type 2 diabetes in both patients and experimental models. This suggests that activation of its receptor, PDGFRα, may contribute to vascular alterations, potentially through the recruitment and proliferation of PDGFRα-expressing perivascular progenitor cells.

Objectives: We aimed (1) to determine the role of PDGFRα signaling and PDGFRα⁺ progenitor cells in small artery remodeling, and (2) to assess their contribution to obesity-associated hypertension in mice.

Methods: We used a tamoxifen-inducible mouse model of constitutive PDGFRα activation (RosaCRE-ERT2xPDGFRα+/(S)K), as well as transgenic mouse models with genetic deletion of the pathway. Both male and female mice were exposed to either normal diet (ND) or high-fat diet (HFD, 60% fat).

Blood pressure was measured by the tail-cuff method and by carotid catheter. Small mesenteric arteries reactivity was studied using ex vivo arteriography. Vascular remodeling was measured using immunofluorescence on isolated mesenteric arteries. Expression of genes and proteins of interest were measured using RT-PCR and Western-blotting.

Results: In male mice, constitutive PDGFRα activation for 6 weeks led to a 25% increase in blood pressure (87±24 mmHg vs 65±9 mmHg in controls, p<0.05) and a 25% decrease in cardiac output in male mice (p<0.05), without affecting other cardiac parameters. This was associated with hypertrophic remodeling of small mesenteric arteries, characterized by increased media thickness and cross-sectional area, without changes in vascular reactivity. Larger vessels were unaffected. Echo-doppler analyses showed an increased resistivity index  of the renal artery (+13% p<0.05) but not of the aorta, suggesting an increased resistance of renal small arteries.

Mechanistically, early proliferation of perivascular cells was observed after 4 days, followed by increased SMC content after 2 weeks, suggesting recruitment and differentiation of progenitor cells.

In contrast, female mice did not show increased blood pressure upon PDGFRα activation. However, under HFD, 50% of female mice developed hypertension (mean arterial pressure of 104±37 mmHg vs 83±21 mmHg for ND), which was completely prevented by genetic deletion of the pathway, without affecting metabolic parameters. Analyses of vascular remodeling in this model are ongoing.

Conclusion: These findings identify PDGFRα signaling as a key regulator of small artery remodeling and blood pressure in metabolic disease. Our results suggest that activation of this pathway promotes the recruitment of perivascular progenitor cells, contributing to hypertrophic remodeling and increased vascular resistance. Targeting PDGFRα signaling may represent a novel therapeutic strategy to prevent hypertension associated with obesity.

8 – Machine learning approach to identify clusters of patients with distinct cardiac chronotropic response during incremental exercise

Ali RAMMAL

Team: Platform I/O

Ali Rammal, Rose Tchala Sare, Christel Kerdiles, Richard Isnard, Michel Zeitouni, Maharajah Ponnaiah, Stéphane N. Hatem

Background. The chronotropic cardiac response is the primary mechanism by which cardiac output increases during exercise; its alteration is central to exercise intolerance in cardiopathies. Heart rate (HR) trajectory during incremental exercise is complex and continuously regulated by the autonomic nervous system. Despite its clinical relevance, no standardized computational framework exists to stratify patients by chronotropic phenotype at scale from CPET data.

Purpose. To develop a machine-learning pipeline that extracts algorithmically derived HR trajectory parameters and identifies clinically meaningful chronotropic phenotypes from ramp CPET data.

Methods. 2,589 ramp CPETs performed at Pitié-Salpêtrière Hospital (2017–2025) were analysed. Three parameters were algorithmically derived from each HR signal: (i) tHR: time of occurrence of a slope change in the HR trajectory, detected via segmented regression with Davies test validation and anchored to the first ventilatory threshold (VT₁); (ii) αHR: angle of slope change at tHR, quantifying the direction and magnitude of HR inflection, with sinusoidal transformation applied to preserve directionality and limit leverage of extreme angles; and (iii) εHR: a dimensionless elasticity-based metric of chronotropic responsiveness. Two models were evaluated: a full model incorporating αHR, εHR, tHR, and VO₂max, and a reduced model using αHR, εHR, and tHR exclusively, requiring no gas analysis. Optimal cluster number was determined by silhouette score and dendrogram height gap. Agglomerative hierarchical clustering (Euclidean distance, complete linkage) was applied to identify exercise phenotypes. A Random Forest classifier (500 trees, stratified 70/15/15 split) was trained to enable stable phenotype assignment across datasets.

Results. Two stable clustering solutions were identified: k=3 and k=4. Across both solutions, a consistent severity gradient emerged. A small but clinically distinct group (3–4%) was characterized by strongly negative αHR, severely attenuated HR slope after VT₁, negative εHR, and the lowest VO₂max, representing the most pathological chronotropic phenotype. An intermediate group exhibited earlier breakpoints, moderate HR deflection, and mixed αHR directionality. The largest group (43–44%) presented a near-linear HR trajectory with the latest tHR, highest εHR, and highest VO₂max, reflecting preserved chronotropic reserve. The Random Forest achieved 94.3% accuracy with the full model and 92.0% with the reduced model. External validation in an independent cohort of 98 patients with cardiac dysfunction (atrial fibrillation, pacemaker, heart transplant) confirmed phenotype reproducibility and clinical relevance.

Conclusion. A machine-learning pipeline applied to three algorithmically derived HR trajectory parameters identifies three to four clinically meaningful chronotropic phenotypes, stratifying patients from severe chronotropic incompetence to preserved exercise response. The direction and timing of the HR slope change at VT₁, captured by αHR and tHR, emerge as key discriminating axes. Reliable phenotype assignment is achievable without gas analysis, enabling automated chronotropic profiling in any standard exercise laboratory.

9 – An Innovative AI Method for Three-Dimensional Analysis of Aortic Wall Deformation: A Multicenter, Multi-Vendor Evaluation

Tom DA SILVA FARIA 

Team: LIB

Tom Da Silva-Faria1,$, Jonas Leite1, Alban Redheuil1,2, Lan-Anh Nguyen3, Khaoula Bouazizi-Verdier1,2, Jia Guo1, Thomas Dietenbeck1, Kevin Bouaou1,2, Sophia Houriez–Gombaud-Saintonge1,2, Umit Gencer3, Elie Mousseaux3, Gilles Soulat3, Emilie Bollache1,2, Nadjia Kachenoura1,2
1Sorbonne Université, INSERM, CNRS, Laboratoire d’Imagerie Biomédicale (LIB), F-75006 Paris, France
2Institute of Cardiometabolism and Nutrition (ICAN), F-75013 Paris, France
3APHP, Hôpital Européen Georges Pompidou, Université de Paris Cité, PARCC, INSERM, F-75015 Paris, France
$tom.da_silva-faria@sorbonne-universite.fr

Objectives: Accurately tracking tissue displacement over time in cardiovascular imaging is essential for advanced applications such as motion-compensated reconstruction [1] and functional assessment [2-4]. While 4D flow MRI offers unprecedented concomitant volumetric and temporal coverage, its usefulness for straightforward wall tracking is challenged by relatively low spatio-temporal resolution along with complex, non-rigid motion of cardiac and aortic structures. This highlights the need for advanced computational methods to reliably estimate displacement fields. In the context of a broader project on fully automated time-resolved aortic segmentation from 4D flow MRI, we compared an unsupervised deep-learning based registration method (VoxelMorph [5]) with a conventional diffeomorphic demons algorithm [6] to estimate aortic wall motion.

Materials and methods: We studied 142 individuals (55 women, age 57 ± 16 years; healthy volunteers n =55, patients with type 2 diabetes =30, myocardial infarction =32, ascending aorta aneurysm =25) from two centers using different MRI scanners and field strengths (1.5T and 3T). All underwent thoracic aortic 4D flow MRI (18-50 cardiac time phases). A VoxelMorph-based model (Figure 1) was trained to register each 𝐼𝑡 cardiac phase to the high-contrast peak systolic 𝐼𝑠 magnitude phase (Figure 2 A.), using a composite loss combining bidirectional imaging volume similarity and smoothness constraints on the predicted 3D displacement fields 𝜙⃗ 𝑠→𝑡 and 𝜙⃗ 𝑡→𝑠. The model was trained on 500 (𝐼𝑡, 𝐼𝑠) volumes pairs derived from 100 exams, and evaluated on all concerned cardiac phases of the remaining 42 independent exams. Structural similarity index (SSIM [7]), local normalized cross-correlation (LNCC [5]), and mean squared error (MSE) were computed between each 𝐼𝑡 phase, and both the 𝐼𝑠°𝜙⃗ 𝑠→𝑡 registered and 𝐼𝑠 non-registered phases, the latter representing baseline distance. Metrics were averaged across the cardiac cycle and compared with those obtained using diffeomorphic demons. An example of aortic displacement field is illustrated in Figure 2 B.

Results: Mean SSIM of 0.75 ± 0.09, LNCC of 0.75 ± 0.08, and MSE of 13.7 ± 6.3 × 103, obtained on baseline non-registered data were significantly improved when applying VoxelMorph: SSIM increased to 0.85 ± 0.06, LNCC increased to 0.86 ± 0.06 and MSE decreased to 6.2 ± 3.3 × 103. Diffeomorphic demons achieved smaller improvements (SSIM =0.79 ± 0.09, LNCC =0.80 ± 0.09, MSE =9.0 ± 5.1 × 103) (Figure 2 C.).

Conclusion: VoxelMorph-based unsupervised registration captured aortic wall motion in 4D flow MRI more accurately than diffeomorphic demons. It provided physiologically plausible displacement fields for all cardiac phases relative to peak systole, enabling assessment of 3D aortic deformation—a relevant and novel index of arterial stiffness. Moreover, this unsupervised training strategy, which does not rely on anatomy-specific annotations, could be applied to any other dynamically moving structure.

10 –

Sonia SLABLAB

Team: Metabolic disorders, diabetes, and comorbidities, UMR 1166

1 – Untreated secondary hypertension is associated with distinct retinal microvascular remodelinbg assessed by adaptive optics imaging analysis

Federico BERNARDO ROSSI

Department : Nutrition

Federico Bernardo Rossi¹, Michel Paques², Maharajah Ponnaiah³, Antonio Gallo⁴

¹PhD Program in Arterial Hypertension and ARHYVAB, University of Padua, Padua, Italy;

²Centre Hospitalier National d’Ophtalmologie des Quinze-Vingts, Sorbonne University, Paris, France.

³Institute of Cardiometabolism and Nutrition ICAN, Paris, France.

⁴University Hospitals Pité Salpêtrière, Sorbonne University, Paris, France.

Objective: Secondary hypertension is characterized by chronic exposure to hormonal excess that may induce microvascular injury beyond blood pressure elevation alone. Adaptive optics retinal imaging enables direct in vivo quantification of retinal arteriolar structure with near-histological resolution. We hypothesized that untreated secondary hypertension is associated with a distinct microvascular remodeling phenotype compared with essential hypertension.

Design and Method: We prospectively enrolled patients with secondary hypertension due to primary aldosteronism, Cushing syndrome, or pheochromocytoma/paraganglioma, and patients with essential hypertension undergoing adaptive optics retinal imaging. Primary outcomes were retinal wall-to-lumen ratio and wall cross-sectional area. Propensity scores were derived using multivariable logistic regression including demographic and clinical covariates. Patients with secondary hypertension were matched 1:3 to essential hypertension controls without replacement using a caliper of 0.2 standard deviations of the logit of the propensity score, with exact matching on sex. Outcomes were analyzed using generalized estimating equations with robust standard errors to account for clustering within matched sets. A prespecified subgroup analysis evaluated patients examined prior to surgical treatment.

Results: The unmatched cohort included 60 patients with secondary hypertension and 957 with essential hypertension. Fifty-seven patients with secondary hypertension were matched to 162 controls. In the matched cohort (n = 219), mean age was 55.8 ± 12.4 years and mean systolic blood pressure was 133 ± 17 mmHg. In the cohort of patients who received surgical treatment, secondary hypertension was independently associated with baseline lower wall-to-lumen ratio (β = −0.049, 95% confidence interval −0.071 to −0.026; p < 0.001) and higher wall cross-sectional area (β = +432.5, 95% confidence interval +382.1 to +482.9; p < 0.001). However, no association secondary hypertension with wall-to-lumen ratio or wall cross-sectional area was seen in the overall matched cohort.

Conclusions: These findings suggest a specific hypertrophic remodeling pattern related to hormonal excess that may partially regress after causal treatment in the subgroup of patients with secondary hypertension assigned to surgery.

2 – Unsupervised Phenotypic Clustering of Cardiovascular Risk Profiles in Systemic Lupus Erythematosus

Lévy AZOULAY 

Team : LIB

L.-D. AZOULAY¹,², N. KACHENOURA², S. BOUSSOUAR³, E. CHARPENTIER³, N. PASI³, L.-A. NGUYEN³, T. BROUSSAUD³, A. GIRON², J. LEITE², N. AIT-ABDALLAH¹, M. PHA¹, M. HIɹ, A. MATHIAN¹, M. PINETON DE CHAMBRUN¹, M. PAPO¹, F. COHEN¹, J. HAROCHE¹, A. REDHEUIL³, Z. AMOURA¹

¹ Hôpital Pitié-Salpêtrière, Médecine Interne 2, Paris, France ² INSERM, Laboratoire d’Imagerie Biomédicale, Paris, France ³ Hôpital Pitié-Salpêtrière, Imagerie Cardio-Thoracique, IHU ICAN, Paris, France

Introduction: Systemic lupus erythematosus (SLE) is a systemic autoimmune disease primarily affecting young women[1]. An extensive and robust body of data has demonstrated the existence of an accelerated atherosclerosis phenomenon in SLE patients[2]. Despite this increased risk profile, optimal cardiovascular (CV) risk prediction and prevention strategies remain yet to be defined[3].

This study aimed to identify distinct cardiovascular risk phenotypes in SLE using an unsupervised cluster analysis, and to compare the risk of incident cardiovascular events and subclinical atherosclerosis progression in the identified subgroups.

Methods: Consecutive SLE patients who underwent a comprehensive cardiovascular risk assessment at the French National Referral Center for SLE between 2014 and 2024 were retrospectively included.

An unsupervised analysis was performed using a hierarchical clustering algorithm on the following clinical, biological and imaging variables as input:

• Age, sex, SLE duration

• Antiphospholipid syndrome (APS)

• Hypertension, diabetes, body-mass index (BMI), smoking

• Estimated glomerular filtration rate (eGFR)

• Fasting lipid profile, including Total cholesterol (TC), High-density lipoprotein (HDL), Low-density lipoprotein (LDL) and Triglycerides (TG)

• Coronary artery calcium (CAC) score

• Epicardial adipose tissue (EAT) volume

The primary outcome was incident CV event (including acute coronary syndrome, stroke and symptomatic peripheral artery disease[PAD]). The secondary outcome was follow-up CAC scores.

Results: A total of 226 patients were included (91% females, 45±13 years). SLE duration was 13 ± 9 years. Prior organ involvement included articular manifestations in 202 patients (89%), skin manifestations in 151 patients (67%), lupus nephritis in 76 patients (34%) and pleural or pericardial effusions in 50 patients (22%). APS was present in 23 patients (10%). A total of 213 patients (94%) were on hydroxychloroquine, 151 (67%) on corticosteroids, and 115 (51%) were receiving or had received conventional synthetic disease-modifying antirheumatic drugs.

Three clusters were identified and compared (Figure 1A, Table 1).

• Cluster 1 (N=123) included young female patients with very few CV risk factors.
• Cluster 2 (N=78) included older patients with a longstanding SLE course and high LDL.
• Cluster 3 (N=25) included middle-aged patients with cardiometabolic disorders.

Primary outcome

After a median follow-up of 8 [4–10] years, 13 events occured (8 acute coronary syndromes, 4 strokes, 1 PAD). No events were observed in cluster 1 (0%), compared with 6 (8%) in cluster 2 and 7 (28%) in cluster 3. On survival analysis, the risk of incident CV events significantly differed between cluster (Log-rank P<0.001, Figure 1B).

Secondary outcome

A subset of 95 patients had a follow-up cardiac CT after a mean follow-up duration of 4±4 years. Follow- up CAC scores significantly differed across the three clusters (cluster 1: 14±52, cluster 2: 133±303, cluster 3: 293±571, P=0.004).

Identification of cluster drivers

A post-clustering decision tree identified (i) age ≥ 48 years-old, (ii) diabetes mellitus, and (iii) epicardial adipose tissue volume ≥ 71 ml/m2 as key determinants of cluster membership with a sensitivity of 73% and a specificity of 81%.

Conclusions: Cardiovascular risk in SLE comprises three subgroups with distinct phenotypes, risks of incident events and subclinical atherosclerosis progression.

These results suggest that a multimodal assessment focusing on cardiometabolic risk factors may help addressing the unmet need of cardiovascular risk stratification in SLE patients.

Beyond stratification, the weight of cardiometabolic risk factors in this study supports the evaluation of GLP1-R agonists in this population.

References:

[1] Hoi A, Igel T, Mok CC,et al. Systemic lupus erythematosus. The Lancet. 2024;403:2326–38. doi: 10.1016/S0140-6736(24)00398-2

[2] Conrad N, Verbeke G, Molenberghs G, et al. Autoimmune diseases and cardiovascular risk: a population-based study on 19 autoimmune diseases and 12 cardiovascular diseases in 22 million individuals in the UK. The Lancet. 2022;400:733–43. doi: 10.1016/S0140-6736(22)01349-6

[3] Drosos GC, Vedder D, Houben E, et al. EULAR recommendations for cardiovascular risk management in rheumatic and musculoskeletal diseases, including systemic lupus erythematosus and antiphospholipid syndrome. Ann Rheum Dis. 2022;81:768–79. doi: 10.1136/annrheumdis-2021-221733

3 – EAT/ coronaires

Paul RITVO

Team: LIB

4 – Strain myocardite

Paul GARIGOUX

Team: LIB

5 – Deep Learning Analysis to Evaluate Right Atrial Volume in Atrial Fibrillation

Lucile DUCOUSSO-BALOUZET

Department: Cardiology

Right Atrial (RA) cavities and remodeling in atrial
fibrillation (AF) remain insufficiently explored using
transthoracic echocardiography (TTE), while current
manual measurements are time-consuming.
Automatic segmentation of the RA with AI, through
Deep Learning (DL) approaches, could significantly
improve its analysis. The aim of this study was to
evaluate the feasibility and accuracy of AI-based
automatic RA segmentation using biplane TTE to
provide reliable RA volumes (RAV) and enhance
the assessment of RA remodeling in patients with
AF.

6 – Beyond Ejection Fraction: Unsupervised Clustering-Based Phenotyping of Heart Failure in the EDUSTRA Cohort

Lucile DUCOUSSO-BALOUZET

Department: Cardiology

Heartfailureisclassicallystratifiedbyejectionfraction(HFpEF,HFmrEF,HFrEF),aclassificationbasedonasingleparameterthatdoesnotreflectthefullcomplexityofglobalmyocardialfunction.Strainparametersprovideamoredetailedassessmentofmyocardialdeformationandmayrepresentmorediscriminatingphenotypicmarkers.TheaimofthisstudywastoapplyanunsupervisedclusteringapproachtopatientsincludedintheEDUSTRAcohorttoexplorethenaturalphenotypicstructureofthispopulation.

7 – The AIM-SO score identifies a hidden metabolic and functional severity gradient in Sarcopenic Obesity: A UK Biobank study

Vittoria ZAMBON

Department: nutrition

Authors: Vittoria Zambon Azevedo1,2 Jean-Michel Oppert3,4 Vlad Ratziu1,2,4

Affiliations:

(1) Foundation for Innovation in Cardiometabolism and Nutrition, IHU ICAN, Paris, France; (2) Assistance Publique Hôpitaux de Paris (AP-HP), Service d’Hépatologie et Gastro-entérologie, Hôpital Pitié-Salpêtrière, Paris, France; (3) Assistance Publique Hôpitaux de Paris (AP-HP), Service de Nutrition, Hôpital Pitié-Salpêtrière, Paris, France; (4) Sorbonne Université, Paris, France.

Keywords: Sarcopenic obesity; Diagnosis assessment; Population-based study; Myosteatosis; Liver-muscle axis; Cardiometabolic outcomes.

Background and Aims: Sarcopenic obesity (SO) is defined by the coexistence of high adiposity and reduced muscle mass and function, increasing the risk of metabolic impairment. However, SO remains clinically underdiagnosed due to a lack of consensus and the masking effect of a seemingly normal BMI. Having previously validated the AI-assisted AIM-SO score in a bariatric cohort, we aimed to apply it to a large population-based cohort to 1) validate this tool against muscle function decline, 2) explore liver-muscle crosstalk, and 3) evaluate cardiometabolic risk across a severity gradient of SO.

Methods: A cross-sectional analysis included 20,634 UK Biobank participants with available dual-energy X-ray absorptiometry (DXA), abdominal magnetic resonance imaging (MRI), and clinical assessments. Participants were stratified by AIM-SO score into three stages: stage 0 (without SO, n=14,553), stage 1 (morphological SO, n=5,297), and stage 2 (SO with reduced handgrip strength, HGS, n=784). Multivariate regressions and ANCOVA models were adjusted for sex, age, and lifestyle factors.

Results: AIM-SO identified 6,081 (29.5%) participants with SO (44.1% men, 11.3% women), who exhibited high visceral fat, myosteatosis, reduced muscle mass and strength (all p<0.001). While SO prevalence dropped to 3.8% (n=784) when associated with reduced HGS (stage 2), the current ESPEN-EASO diagnosis criteria detected less than 2.0%. By requiring loss of function prior to body composition assessment, current criteria miss individuals with adverse morphological and metabolic phenotypes (identified by stage 1) who have not yet reached severe functional decline. Exploring the liver-muscle axis, myosteatosis increased the odds of muscle weakness independently of BMI (men: OR=1.15, and women: OR=1.08, p<0.001). Both hepatic steatosis and myosteatosis were independent predictors of muscle weakness, overriding the protective effect of lean mass (p≤0.001). A clear severity gradient emerged across stages 0, 1 and 2. Despite an unchanged mean BMI (~28.7 kg/m², p=0.895), we observed a stepwise decline in muscle mass (24.1 → 20.1 → 19.4 kg) and strength (34.6 → 30.5 → 16.5 kg), alongside rising body fat (35.5 → 40.1 → 40.4 %) and myosteatosis (7.5 → 8.6 → 9.0 %).

This progression was accompanied by a linear rise in diabetes (8.0 → 10.0 → 16.0%) and cardiovascular events (9.0 → 13.0 → 19.0 %), all p<0.001. SO was a strong independent risk factor for liver (OR=3.36, 95% CI: 1.18–9.52) and kidney (OR=1.48, 95% CI: 1.11–1.98) events in men, and cardiovascular events (OR=1.41, 95% CI: 1.12–1.76) in women, all p≤0.02.

Conclusion: The AIM-SO score identifies high-risk metabolic phenotypes overlooked by current diagnostic criteria that rely on functional loss over body composition. Stage 1 (morphological SO) represents a critical window for intervention where metabolic damage and myosteatosis are already established, yet muscle strength remains partially preserved. AIM-SO enables early risk stratification, providing the opportunity to intervene before metabolic damage progresses to severe functional loss.

• CARDI-HACK

Tania MEDALI

Innovation & Research Valorization Division – IHU ICAN

Hypertrophic cardiomyopathy (HCM) is a common inherited myocardial disease and a leading cause of sudden cardiac death in young individuals. However, risk stratification remains challenging due to significant clinical variability and the complexity of genetic factors. The CardI-HACK project aims to address this limitation by developing an artificial intelligence (AI)-based model integrating both clinical and genomic data. To achieve this, a multicenter cohort of approximately 500 patients was assembled, including comprehensive clinical, imaging, and genetic datasets, all anonymized in compliance with GDPR regulations. A data challenge framework was implemented, enabling multidisciplinary teams to develop and evaluate predictive models using standardized training and test datasets. This approach facilitated the development of several AI-based models and improved the identification of high-risk patient profiles, while promoting innovation and robust model comparison. Overall, integrating clinical and genetic data through AI represents a promising strategy to enhance risk prediction and support more personalized management in patients with HCM.

• ICONIC

Louise MEYFROIT

Innovation & Research Valorization Division – IHU ICAN

The ICONIC project, born from the collaboration between two national infrastructures—IHU ICAN and CONSTANCES—aims to create a large, certified non-invasive imaging biobank linked to the CONSTANCES cohort. This collection of MRI and ultrasound images, complemented by a biobank, will be analyzed to characterize heart, vessel, and liver function in the French population across all ages and cardiometabolic risk profiles. These innovative data, currently non-existent in France, will enable the creation of a reference imaging cohort for both academic and industrial research on common and rare diseases. The project will thus contribute to the development of a cardiovascular imaging sector by providing resources to French academic and industrial stakeholders (pharmaceuticals, medical devices, software, data, imaging), opening numerous opportunities for national and international public-private partnerships.

• MEDITWIN

Arturo HERNANDEZ-CERVANTES

Innovation & Research Valorization Division – IHU ICAN

Antonio Gallo, Arturo Hernandez-Cervantes, Stéphane Commans, Stéphane Hatem, Philippe Giral, Stéphane Barritault, Maud Decraene, Jeanne Haidar, Bertrand Cariou, Alban Redheuil, Nadjia Kachenoura, Khaoula Bouazizi, Marie l’Homme, Farid Ichou, Maharajah Ponnaiah, Pauline Paul Stéphen Raj, Philippe Lesnik, Maryse Guerin and Wilfried Le Goff

The Meditwin project aims to create personalized virtual twins of organs, metabolism, and tumors to enhance diagnosis and treatment. These digital models will allow doctors to simulate various disease progression scenarios, helping to optimize patient care in neurology, cardiology, and oncology.

Funded by iDEMO (Bpifrance, 2024), the project is led by a consortium that includes seven University Hospital Institutes (IHUs), the Nantes University Hospital (CHU de Nantes), INRIA, and several partner startups. It is coordinated by Dassault Systèmes and organized around eight use cases, one of which focuses on familial hypercholesterolemia (WP1.UC3).

This particular use case (WP1.UC3) is being developed in collaboration with CHU Nantes, ICAN, the Vision Institute (IHU Foresight), Dassault Systèmes, and INRIA. It aims to improve the monitoring of patients at high cardiovascular risk by integrating clinical, biological, and imaging data.

Initially, the models will be trained using retrospective data from the FH-CALC and SAFIR patient cohorts. To refine and validate the predictive methods developed, additional heterogeneous data will be needed—specifically from patients with familial hypercholesterolemia (FH) (ATHERO-FH and DESTINY-FH) and high-risk individuals (ERICAH and Predirisk). Over the past year, significant progress has been made in the development of this use case.

The first datasets from the FH-CALC cohort have been successfully obtained and curated, including metagenomic, lipidomic, and metabolomic data, as well as imaging-derived metrics such as coronary artery calcium (CAC) scores and coronary CT angiography measurements, alongside structured clinical datasets. In parallel, regulatory approvals have been secured for the DESTINY-FH cohort, with the inclusion of the first patients scheduled for June of this year. The protocol for the ERICAH cohort has been completed, and the first patients have already been enrolled in the ATHERO-FH cohort.

• EpOS-LT as a central pilar for translational research

Stéphane COMMANS
Innovation & Research Valorization Division – IHU ICAN

EpOS-LT is a key clinical cohort for translational research in Liver Diseases and provides a unique translational framework bridging clinical care, biological research, and AI-based innovation.

It accelerates precision medicine for metabolic liver diseases, enabling the development of predictive models, non-invasive diagnostics, and personalized patient management.

Metabolic-Associated Fatty Liver Disease (MAFLD) is considered the liver manifestation of metabolic syndrome, and is now one of the most common causes of chronic liver disease. Steatosis on ultrasound is found in about 30% of individuals in the general population and transaminases are elevated in about 8%. The histological spectrum of MAFLD covers (a) Metabolic-Associated fatty liver (MAFL), (b) steatohepatitis (MASH), which combines lesions of lobular inflammation and hepatocyte ballooning, (c) fibrosis, and (d) cirrhosis. Liver-related mortality is determined primarily by the occurrence of cirrhosis, which is responsible for episodes of decompensation, sepsis, and hepatocellular carcinoma. Hepatocellular carcinoma (HCC) is the sixth most common cancer in the world by incidence and the fourth most common by mortality. European data collected in the North of England during the decade 2000-2010 showed that metabolic risk factors are found in 35% of patients newly diagnosed with HCC, but there are no effective screening policies and is becoming a major medical issue. The obesity epidemic is closely associated with the increasing prevalence and severity of MAFLD.

However, there are uncertainties regarding the natural history and prognosis of MAFLD. In particular, the stratification of patients in whom there is significant inter-individual variability (in particular the severity and speed of disease progression, cases of regression), the prognosis of progression, the progression to cancer, the role of comorbidities or complications, the impact of a therapeutic intervention (drug or surgery) on the course of the disease.

The main objective of the EpOS-LT cohort is to identify predictive factors (histological, genetic, clinical, biological and imaging) of the progression of MAFLD to the most advanced stage (compensated or decompensated cirrhosis, whether or not liver transplantation is required, liver cancer, death), within a retrospective and prospective cohort of well-characterized patients covering the entire spectrum of MAFLD severity patients, followed in the gastroenterology department of the Pitié-Salpêtrière hospital group. The endpoint is the correlation between the identified predictors (individually or combined) and the evolution/severity of liver damage, the occurrence of liver events (progression to cirrhosis, decompensation of cirrhosis, hepatocellular carcinoma, liver transplantation, liver-related death) or non-liver-related events (cardiovascular and metabolic diseases). The analyses will include artificial intelligence tools best suited to integrate heterogeneous complex data (including imaging and OMICS), as well as to improve the reading of biopsies currently dependent not only on sampling and inter-observer variability but also on a purely qualitative or semi-quantitative evaluation of the lesions of interest (inflammation, fibrosis).

• STEER Study: Lower Risk of Cardiovascular Events with Semaglutide vs. Tirzepatide in Real-World Settings

NOVO NORDISK

1 – Smart Gold Nano-Probes for Non-Invasive Diagnosis of Liver Fibrosis

Valeria PISTORIO

Team: Biliary and Fatty Liver Diseases – UMR 938

Valeria Pistorio1, Julien Hédou1, Killian Lucas3, Melissa De Souza3, Carolina Neves2, Virginie Steunou1, Edmond Gravel3, Laurent Devel4, Thierry Poynard1, Philippe Lesnik2 and Jérémie Gautheron1.

1 Sorbonne Université, Inserm UMRS_938, Centre de Recherche Saint-Antoine (CRSA) Paris 75012 France.
2 Sorbonne Université, INSERM UMRS-1166, Institute of Cardiometabolism and Nutrition (ICAN), 75013 Paris, France. 3 Université Paris-Saclay, CEA, INRAE, Département Médicaments et Technologies pour la Santé (DMTS), SCBM, 91191 Gif-sur-Yvette, France. 4 CEA, INRAE, Médicaments et Technologies pour la Santé (MTS), SIMoS, Université Paris-Saclay, 91191, Gif-sur-Yvette, France.

Summary: Chronic liver diseases represent a major global health burden, with liver fibrosis being a key determinant of clinical outcome1. It is now well established that fibrosis regression is an active, tightly regulated biological process involving specialized cellular programs and proteolytic systems. In the context of extracellular matrix remodeling and resolution of fibrosis, MMP-9 (gelatinase), MMP-12 (macrophage elastase) and MMP-13 (collagenase) play an important role in the resolution of hepatic fibrosis2,3. However, the molecular mechanisms underlying fibrosis regression remain insufficiently characterized4.

Current diagnostic approaches mainly rely on invasive procedures or static measurements of fibrosis burden, and fail to capture the dynamic processes underlying extracellular matrix (ECM) remodeling and fibrosis regression5. There is therefore a critical need for non-invasive tools capable of monitoring fibrosis evolution in real time. The FibrosOr project aims to develop innovative diagnostic tools based on smart, clearable hybrid organic–gold nanostructures (hAuNS) designed to detect matrix metalloproteinase (MMP) activity, notably of MMP-9, MMP-12, and MMP-13, in vivo. These nano-probes act as activity-based sensors: upon proteolytic cleavage by target MMPs, they release renal-clearable nanoparticles that can be detected in urine through a simple colorimetric readout, providing a non-invasive and functional measure of intrahepatic ECM remodeling.

To support the development and validation of these probes, we use two complementary murine models of liver fibrosis and regression: carbon tetrachloride (CCl₄)-induced hepatocellular injury and DDC diet-induced cholestatic fibrosis. In both models, fibrosis was induced and allowed to regress upon injury withdrawal, and characterized using histological and molecular approaches. These models exhibit distinct ECM remodeling dynamics and differential regulation of MMP-9, MMP-12, and MMP-13, highlighting the importance of injury-specific contexts in fibrosis resolution. In parallel, data from transcriptomics, proteomics, lipidomics have been generated, and further Joint omics integration analysis (MOFA+) will be performed in order to identify both model specific and regression specific mechanisms.

These results will frameworks to characterize fibrosis regression dynamics and to identify relevant proteolytic signatures across injury contexts. Altogether, these findings will support the development and optimization of MMP-sensitive nano-probes, and pave the way for their application as non-invasive tools to monitor liver fibrosis evolution.

References:
1 Wynn, Thomas A. « Fibrotic disease and the TH1/TH2 paradigm. » Nature Reviews Immunology 4.8 (2004): 583-594.
2 Duarte, Sergio, et al. « Matrix metalloproteinases in liver injury, repair and fibrosis. » Matrix Biology 44 (2015): 147-156.
3 Geervliet, Eline, and Ruchi Bansal. « Matrix metalloproteinases as potential biomarkers and therapeutic targets in liver diseases. » Cells 9.5 (2020): 1212.

2 – Development of a centralized echocardiographic analysis: from protocols to the development of artificial intelligence tools.

Unit: Cardiology

Laurie SOULAT-DUFOUR

3 – Lean individuals with NAFLD have more severe liver disease and poorer clinical outcomes (NASH-CO Study)

Oumarou NABI

Team: Biliary and Fatty Liver Diseases – UMR 938

Oumarou Nabi1,2,3,4, Nathanaël Lapidus1,2, Karine Lacombe1,2, Jerome Boursier5,6, Victor de Lédinghen7, Marie Zins8,9, Marcel Goldberg9,10, Lawrence Serfaty1,4,11

1-Sorbonne Université, Paris, France,

2-Institute Pierre Louis Epidemiology and Public Health, Epidemiology, Paris, France

3-Washington University in St. Louis, Saint Louis, USA

4-Inserm UMR S938/ Centre de recherche Saint-Aintoine, Paris, France

5-centre hospitalier universitaire d’Angers, Angers, France

6-University of Angers, Angers, France

7-Hospital Center University De Bordeaux, Bordeaux, France

8-Paris City University, Paris, France

9-Inserm UMS 11, Paris, France,

10-Université Versailles Saint-Quentinen-Yvelines, Guyancourt, France,

11-Arlin Alsace Hospitals Academics De Strasbourg, Strasbourg, France

Background: While obesity is a major risk factor, Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD) has also been reported in lean subjects. In this clinical setting, severity of liver injury and clinical outcomes are debated and few studies have been conducted at a general population level. This study aimed to assess prevalence, characteristics and mortality outcomes of lean MASLD in the French adult population.

Methods: The study population consisted of 127,291 participants from the nationwide CONSTANCES cohort. After exclusion of subjects with excessive alcohol consumption, viral hepatitis or other liver diseases, 110,120 were analyzed. Noninvasive diagnosis of MASLD and advanced fibrosis was performed using the combination of Fatty Liver Index and Forns Index. Outcomes analyzed were liver-related events, hepatocellular carcinoma (HCC), cardiovascular disease (CVD), extrahepatic cancer (EHC), chronic kidney disease (CKD), liver transplantation and overall mortality. The median follow-up was 30 months.

Results: The median follow-up was 3.58 years. The prevalence of MASLD was 18.3% in overall population and 5.3% (95%CI:5.2–5.4) in lean subjects, while 16.3% (95%CI:15.7-16.8) of MASLD subjects were lean. Despite their better metabolic profile, the prevalence of advanced fibrosis was significantly higher in lean than in non-lean MASLD (3.7%vs.1.7%, respectively, p<0.01). Among MASLD subjects and after adjustment for demographics, metabolic risk factors and lifestyle, lean status was associated with advanced fibrosis (OR=1.26, 95%CI:1.20-1.65, p=0.005), an increased risk of liver-related events (adjusted HR=5.84,95%CI:4.03-8.46), chronic kidney disease (adjusted HR=2.49,95% CI: 1.49-4.16), and overall mortality (adjusted HR=3.01, 95%CI: 2.21-4.11). Liver related events and overall mortality were related to the severity of fibrosis, both in lean and non-lean MASLD subjects, whatever the usual risk factors.

Conclusion: This large, prospective, community-based cohort provides real-world evidence that lean MASLD subjects have more progressive liver disease and poorer clinical outcomes, independent from the usual risk factors and lifestyle. Because of the fairly low prevalence of MASLD in lean subjects, MASLD screening should target those patients with metabolic abnormalities and/or unexplained cytolysis. The poorer prognosis of MASLD in lean individuals may encourage the inclusion of these patients in clinical trials.

4 – Development and validation of a cardiac MRI protocol for assessing myocardial viability in explanted pig hearts (initial results from MANDRAGORE)

Elodie BERG

Unit: Cardiology

Objective: Heart transplantation remains the standard of care for end-stage heart failure, but early graft dysfunction compromises outcomes. Myocardial viability at the time of implantation remains inadequately assessed. Cardiac magnetic resonance imaging (CMRI) techniques may offer a noninvasive approach to characterizing myocardial integrity. The primary objective was to validate a cardiac MRI protocol for assessing myocardial viability prior to transplantation.

Materials and Methods: Thirty porcine hearts were divided into three groups: brain-dead donors (BDD), donors after circulatory death (DCD), and an ischemic group serving as a positive control for myocardial injury. All hearts were preserved under static hypothermia. Acquisitions were repeated every hour for four hours. Quantitative T1 mapping and proton magnetic resonance spectroscopy (¹H-MRS) were used to analyze structural and metabolic changes over time.

Results: Quantification of intramyocardial metabolites using ¹H-MRS was successfully performed in 28 of 30 hearts. T1 mapping revealed homogeneous myocardial protection. In the DBD group, T1 increased linearly over time (Y = 28.05X + 1472). A similar but slower increase was observed in the DCD group (Y = 17.53X + 1524). The ischemic group exhibited the steepest slope, reflecting rapid tissue damage (Y = 38.46X + 1388). Lactate accumulation differed among the groups: it was fastest in the DCD group (slope 0.01082), followed by the DBD group (0.01035). The ischemic group had higher initial concentrations but slower kinetics (0.009235).

Conclusions: This study demonstrates the feasibility of a rapid multiparametric cardiac MRI protocol capable of characterizing myocardial alterations ex vivo, paving the way for an objective assessment of graft viability prior to transplantation.

5 – News and innovations in the field of cardio-renal-hepato-metabolic (CRHM) disorders

Charlotte Crépin, Medical Lead at Boehringer Ingelheim and medical expert

BOEHRINGER INGELHEIM