Unit of Functional and Adaptive Biology

SET-domain containing protein 2 (SETD2) is the primary methyltransferase responsible for generating H3K36me3, an epigenetic mark that is essential for transcriptional regulation and chromatin integrity. SETD2 mutations are frequently observed in various cancers and tend to cluster within its catalytic SET domain. Despite the clinical relevance of SETD2 missense mutations in cancer, their biochemical and structural consequences remain insufficiently characterized. Here, we present the enzymatic and structural characterization of the SETD2 L1609P mutant enzyme identified in leukemia. The L1609 residue is located in the SET domain within a conserved hydrophobic pocket that is involved in substrate H3K36 recognition. Interestingly, site-directed mutagenesis of residues within this hydrophobic pocket leads to SETD2 enzyme variants with either decreased or increased H3K36me3 methyltransferase activity, suggesting that cancer mutations affecting the L1609 residue could result in a loss- or gain-of-function enzyme variant. Using molecular and cellular approaches, we show that the SETD2 L1609P mutant exhibits reduced H3K36 methyltransferase activity, decreased protein stability, and poor cellular expression. Consistently, the crystal structure of the SETD2 L1609P in complex with a H3K36M peptide shows remodeling of the active site. These findings support the pivotal role of SETD2 inactivation and subsequent disruption of H3K36me3 deposition in oncogenesis, particularly in hematologic malignancies. Our study provides the first mechanistic and three-dimensional protein structure information on how SETD2-associated cancer mutations can lead to altered H3K36 methyltransferase activity.

Ageing is traditionally conceived as a continuous process of progressive physiological decline. However, recent evidence across species suggests that ageing may instead proceed through distinct phases. Using state-of-the-art statistical methods, we develop a rigorous analysis of longitudinal survival data from 1,159 individually tracked female Drosophila melanogaster. This data-driven analysis leads us to introduce a new parametric model of transition rates within the two-phase ageing framework. Flies were monitored using the Smurf assay, which detects increased intestinal permeability through leakage of an ingested blue dye, and is a strong biological marker of ageing. The Smurf phenotype identifies a sharp transition from a non-Smurf state to a Smurf state that precedes death. Our results yield three key findings. First, the Smurf transition rate follows a Gompertz-Makeham law, increasing exponentially with age. Second, contrary to previous constant-rate assumptions, newly transitioned Smurf flies exhibit remarkably high mortality - approximately 40% die within 24 hours - followed by an exponential decline in death rate that stabilises to a lower constant baseline. Third, we identified a mild but statistically significant negative dependence between time spent non-Smurf and subsequent Smurf lifespan. Our best-fit model captures a potential bimodal nature of mortality curves using simple, biologically interpretable functions. Validation using data from two mouse strains confirms the broader applicability of this framework. These results establish a quantitative foundation for the two-phase ageing paradigm and highlight a critical period of vulnerability immediately following the physiological transition to frailty.

Le polypeptide insulinotrope glucose-dépendant (GIP) a été la première incrétine identifiée et joue un rôle essentiel dans le maintien de la tolérance au glucose chez l’homme en bonne santé. Jusqu’à récemment, le GIP a été éclipsé par l’autre incrétine, le peptide-glucagon-like-1 (GLP-1) à l’origine d’une famille de nouveaux médicaments très efficaces contre le diabète et l’obésité. Dans cette revue, nous présentons un aperçu actualisé des multiples facettes de la biologie du GIP. Ses principales actions comprennent l’augmentation de la sécrétion d’insuline stimulée par le glucose, la modulation du stockage lipidique et de la lipolyse dans le tissu adipeux, la réduction du poids corporel par suppression de l’appétit à médiation centrale, le soulagement des nausées d’origine médicamenteuse, la diminution de l’inflammation périphérique et centrale, le renforcement de la formation osseuse et l’amélioration de la cognition. Ce peptide pléiotrope est maintenant un candidat prometteur pour le développement de nouvelles pharmacothérapies adaptées non seulement aux troubles du métabolisme énergétique, glucidique et lipidique, mais aussi aux troubles osseux et neurodégénératifs. The glucose-dependent insulinotropic polypeptide (GIP) was the first incretin identified and plays a critical role in maintaining glucose tolerance in healthy humans. Until recently, GIP has been overshadowed by the other incretin, glucagon-like peptide-1 (GLP-1), which has spawned a family of highly effective new drugs for diabetes and obesity. In this review, we provide an updated overview of the multifaceted biology of GIP. Its major actions include increasing glucose-stimulated insulin secretion, modulating lipid storage and lipolysis in adipose tissue, reducing body weight through centrally mediated appetite suppression, alleviating drug-induced nausea, decreasing peripheral and central inflammation, and promoting cognition and bone formation. This pleiotropic peptide is now a promising candidate for the development of new pharmacotherapies adapted not only to disorders of energy, carbohydrate and lipid metabolism, but also to neurodegenerative and bone disorders.

Reward processes have traditionally been ascribed to dopamine (DA)-associated circuits. While external stimuli, such as food and drugs of abuse, are activators of DA-neuron activity, growing evidence indicates that interoceptive signals also play a critical role. Among these, the gut-brain vagal axis has emerged as a key regulator, although its precise contribution to mesolimbic DA signaling and behavior remains unclear. Here, we combine complementary ex vivo and in vivo approaches across multiple scales to show that gut-brain vagal tone is essential for gating mesolimbic DA system activity and functions, modulating DA-dependent molecular and cellular processes, and scaling both food- and drug-induced reinforcement. These findings challenge the traditional brain-centric view of reward processing, supporting a more integrated model in which vagus-mediated interoceptive signals intrinsically shape motivation and reinforcement. By uncovering the influence of gut-brain vagal communication on DA functions, this work provides insights into the neurobiology of adaptive and maladaptive reward, with broad relevance for eating disorders and addiction.

AimsDeveloping and evaluating target-specific machine learning (ML) scoring functions (SFs) to enhance structure-based virtual screening for Poly (ADP-ribose) polymerase 1 (PARP1), a key DNA repair enzyme and validated target in BRCA-mutated cancers. Existing PARP1 inhibitors face limitations in selectivity and toxicity, motivating ML-driven strategies for identifying improved chemotypes [1].MethodsBioactivity data from ChEMBL were combined with property-matched decoys generated using DeepCoy. Ligands were docked into the PARP1 crystal structure with Smina, and training and test sets were prepared. Five ML algorithms, random forest, extreme gradient boosting, support vector machines (SVM), and neural network models, were trained in regression and classification modes using multiple featurisation techniques, including protein-ligand extended connectivity (PLEC) fingerprints, GRID, and Morgan fingerprints, as described in our recent publication [2]. Models were benchmarked against generic SFs using enrichment factor (EF1%), normalised EF1% (NEF1%) and precision-recall curves.Results and conclusionsTarget-specific ML SFs consistently surpassed generic approaches, with a SVM regressor trained on PLEC fingerprints achieving the best performance (EF1% = 38.8, NEF1% = 0.764). Models retained strong predictive ability on a structurally dissimilar test set, confirming generalisation and recovery of diverse inhibitors. These results demonstrate that ML-tailored SFs substantially improve virtual screening for PARP1 and provide an effective route toward novel, selective anticancer compounds.

The rapid expansion of available three-dimensional protein structures-derived from both experimental techniques and bioinformatic predictions-offers unprecedented opportunities for drug discovery, particularly for targets that have historically been difficult to characterize. However, the effective analysis of these increasingly complex and voluminous structural datasets remains a major challenge. Efficient representations of protein conformations are essential to facilitate large-scale comparison, structural classification, and functional interpretation in therapeutic contexts. The concept of structural alphabets, introduced by Pr S. Hazout in 1999, provides a robust and scalable framework to represent local protein backbone conformations using a limited set of recurring structural motifs. This representation enables a one-dimensional encoding of three-dimensional protein structures that retains essential geometric features beyond secondary structure, while allowing systematic and interpretable analyses. In this review, we focus on HMM-SA, a structural alphabet constructed using a hidden Markov model. HMM-SA defines 27 structural motifs, including 18 regions specifically dedicated to loops, and captures the statistical dependencies between them. We present a detailed overview of the HMM-SA framework, and of the computational tools derived from this structural alphabet, developed to explore protein function, conformational variability, and the structural determinants of molecular recognition. The utility of HMM-SA is illustrated through a case study on HIV-2 protease (PR2), a critical enzyme in antiretroviral drug development. By analyzing PR2 structural asymmetry, ligand-induced conformational changes, and mutation-driven alterations, we highlight the ability of HMM-SA-based methods to identify key structural features involved in ligand specificity and resistance mechanisms, thereby advancing therapeutic target analysis.

Insulin deficiency caused by the loss of β cells and/or impaired insulin secretion is a key factor in the pathogenesis of type 2 diabetes (T2D). The restoration of β cell number and function is thus a promising strategy to combat diabetes. Dual-specificity tyrosine-regulated kinase 1A (DYRK1A) has been shown to regulate human β cell proliferation. DYRK1A inhibitors are potential therapeutic tools, due to their ability to induce β cell proliferation. However, their anti-diabetic effects in the complex setting of type 2 diabetes remains unexplored. The aim of this study was to determine the impact of chronic DYRK1A inhibition on the remission of diabetes in pre-diabetic and overtly diabetic Goto-Kakizaki (GK) rats. We assessed the impact of in vivo treatment with a DYRK1A inhibitor, Leucettinib-92, on β cell proliferation and insulin secretion in GK rats. Further, we evaluated the effects of long-term Leucettinib-92 treatment on the whole-body glucose metabolism in overtly diabetic GK rats through the assessment of fasting and post-absorptive glycemia, glucose tolerance and insulin sensitivity. Short-term in vivo treatment of prediabetic GK rats with Leucettinb-92 stimulated β cell proliferation in vivo, and sustainably prevented the development of overt hyperglycemia. Long-term treatment of adult GK rats with established diabetes increased the β cell mass and reduced basal hyperglycemia. Leucettinib-92 treatment also improved glucose tolerance, and glucose-induced insulin secretion in vivo. We show that DYRK1A inhibition restores the β cell mass and function in a preclinical model of T2D, leading to the improvement of body's global glucose homeostasis.