Browse Articles

Oligoclonal expansion of atypical Vδ2− γδ T cells in Good’s Syndrome

Abstract Good’s syndrome is a rare adult-onset immunodeficiency characterized by thymoma, hypogammaglobulinemia, B-cell lymphopenia, and T-cell dysfunction. Despite well-characterized defects in conventional immune subsets, the impact of this disorder on unconventional T cells, including γδ T cells, remains largely unexplored. In this study, we analyse γδ T cells in 10 patients with Good’s syndrome using immunophenotyping, functional assays, and T-cell receptor (TCR)δ repertoire profiling of peripheral blood and thymoma tissue. Our analyses reveal a pronounced expansion of the Vδ2⁻ γδ T-cell compartment, composed primarily of Vδ1⁺, Vδ3⁺ and the exceptionally rare Vδ8⁺ subsets. The Vδ2⁻ cells are characterized by an activated and effector phenotype and a private and oligoclonal TCRδ repertoire. The thymoma tissue contains distinct clonotypes compared to circulation, suggesting clonal focusing in response to the tumor. Together, our findings show that γδ T-cell perturbations are integral characteristics of Good’s syndrome and broaden our understanding of immune dysregulation in this acquired immunodeficiency.

The inflammatory “death-to-life” loop in VEXAS syndrome

Benchmarking large language models for cell-free RNA diagnostic biomarker discovery

Abstract Large language models can synthesize biomedical knowledge, parse vast amounts of data, and generate code, positioning them as promising tools for biomarker discovery from high-throughput omics data. Here, we benchmark six models from OpenAI, Anthropic, and Google on plasma cell-free RNA datasets spanning three clinical cohorts: Kawasaki disease versus multisystem inflammatory syndrome in children, active tuberculosis versus symptomatic respiratory controls, and myalgic encephalomyelitis/chronic fatigue syndrome versus sedentary controls. We evaluate literature-guided nomination of diagnostic gene panels for downstream machine learning and autonomous construction of end-to-end classifiers from raw count matrices to held-out test predictions. Despite prompt adherence issues, model-nominated panels recapitulate canonical immune pathways and outperform random panels across cohorts, even matching differential gene expression baselines in the tuberculosis cohort. End-to-end automation proves feasible but is model- and task-dependent. One model approaches conventional performance for Kawasaki disease versus multisystem inflammatory syndrome in children, whereas performance decreases for tuberculosis and myalgic encephalomyelitis/chronic fatigue syndrome cohorts. These findings delineate current capabilities and limitations of large language models in diagnostics and open a path for their future use in biomarker discovery.

A step towards guided complement-targeted therapy in ITP?

Integrating microbial cell factory with new-to-nature photobiocatalysis for de novo biosynthesis of d-homotryptophan

Transfusing HPA-mismatched platelets to mimic fetomaternal hemorrhage elicits fetal/neonatal alloimmune thrombocytopenia

Abstract Fetal/neonatal alloimmune thrombocytopenia (FNAIT) is a bleeding disorder in which maternal antibodies target fetal and neonatal platelet alloantigens, most commonly human platelet alloantigen-1a (HPA-1a), resulting in fetal and neonatal thrombocytopenia severe enough to cause life-threatening organ bleeds, such as intracranial hemorrhage. Hemolytic disease of the fetus and newborn (HDFN) is an analogous disease caused by maternal exposure to fetal red blood cell (RBC) alloantigens, most commonly because of postpartum fetomaternal hemorrhage (FMH), which can be prevented by prophylactic administration of fetal RBC-specific antibodies. Unlike HDFN, the events that trigger FNAIT are unknown and can occur during first pregnancies, making FNAIT difficult to predict and prevent. Herein, we investigated the ability of in utero FMH to induce maternal alloimmunization to HPA-1a and cause FNAIT in a preclinical model. Transfusion of HPA-1a+ platelets into wild-type (WT) mice in numbers representing moderate and severe FMH in humans induced production of equivalent levels of HPA-1a–specific antibodies in nonpregnant mice and mice pregnant with WT or HPA-1a+ fetuses, causing FNAIT in the latter. Administration to pregnant females of the HPA-1a–specific monoclonal antibody (mAb) RLYB212/mAb 26.4 prevented FMH-induced maternal alloimmunization to HPA-1a and FNAIT in genetically susceptible pups. In mice pregnant with HPA-1a+ fetuses but not exposed to FMH, administration of RLYB212/mAb 26.4 did not cause FNAIT. Together, these findings identify in utero FMH as a potential trigger for maternal alloimmunization to fetal HPA-1a, and provide proof of concept that prophylactic administration of HPA-1a–specific antibodies may safely and effectively prevent FMH-induced FNAIT in at-risk pregnancies.

Chiral superfluorescence from perovskite superlattices at room temperature

Molecular regulation and physiological role of GOLPH3-mediated Golgi retention

Abstract The Golgi complex serves as the central hub of the biosynthetic pathway, where anterograde and retrograde trafficking converge. How cargo and Golgi-resident proteins traverse this organelle has long been debated. Recent studies have identified a molecular machinery that sorts resident proteins into retrograde-directed COPI vesicles during cisternal maturation. Golgi phosphoprotein 3 (GOLPH3) is a key component of this system; however, its physiological relevance and regulatory mechanisms remain poorly defined. Here, we show that GOLPH3 depletion in mice alters both protein and lipid glycosylation, causes partially penetrant embryonic lethality, and severely impairs growth and bone mineralization. At the molecular level, we find that GOLPH3 is regulated by functionally antagonistic S-acylation events that control the topology of its membrane association. To mediate retrograde trafficking of Golgi-resident glycosyltransferases, GOLPH3 must bind their cytosolic tails. This occurs via a negatively charged surface region, which is correctly oriented only in one of the S-acylated GOLPH3 conformations. Together, these findings reveal a lipid-mediated regulatory mechanism for intra-Golgi trafficking and establish the critical role of GOLPH3 in vertebrate development.

Minimal residual disease in CLL: when does it really matter?

Influenza coinfection inhibits control of mycobacterial infection in a human challenge model

Abstract Mycobacterium tuberculosis infection is a dynamic continuum. Clinical outcomes reflect complex host-pathogen interactions. Epidemiological and animal studies have suggested influenza coinfection as a risk factor for progression from contained infection to active disease, but human studies have been lacking. Using a whole blood luminescent mycobacterial growth inhibition assay within a human influenza challenge study, we show that influenza infection reduces immunological control of mycobacterial growth. Transcriptome-wide RNA sequencing, cytokine and cellular analyses of subjects’ blood before and after influenza infection reveal that innate immune pathways, including type 1 interferon signalling, are activated by influenza but their subsequent responsiveness to mycobacteria is reduced, with multiple genes’ responses to BCG lux infection repressed by influenza coinfection. Our data suggest that influenza infection impairs immune mechanisms that contain mycobacterial growth and may be a risk factor for tuberculosis (TB) disease. Influenza vaccination might offer high risk, high prevalence populations protection against TB disease.

Anticoagulation with mechanistically distinct FXI/FXIa antibodies amrecibart (REGN9933A2) and cenvacibart (REGN7508Cat)

Thrombosis is a major contributor to global morbidity and mortality. Current standards of care target the extrinsic and/or common pathways of coagulation, effectively inhibiting thrombosis but also increasing bleeding risk, highlighting the unmet need for additional treatment options. Genetic deficiency in factor XI (FXI), a component of the intrinsic pathway, reduces thrombosis risk without spontaneous bleeding. We generated 2 FXI monoclonal antibodies (mAbs) with distinct profiles to provide new approaches to anticoagulation. Cenvacibart (REGN7508Cat) targets the catalytic domain to completely block FXI activity (induced by FXIIa or FXIa in the intrinsic pathway or thrombin in an intrinsic/common pathway amplification loop), thereby maximizing anticoagulation; amrecibart (REGN9933A2) targets the apple 2 domain of FXI/FXIa to specifically prevent FXI activity induced by FXIIa-delivering perhaps less anticoagulation but with potentially lower bleeding risk. We evaluated the anticoagulant effects of both mAbs in vitro in human/non-human primate plasma, in vivo in non-human primates, and healthy volunteers. Both mAbs inhibited intrinsic pathway-triggered coagulation, assessed by activated partial thromboplastin time (aPTT); cenvacibart exhibited a greater increase in aPTT versus amrecibart or other FXI-targeted inhibitors. Neither amrecibart nor cenvacibart affected the extrinsic pathway, assessed by prothrombin time (PT). In non-human primates, both mAbs prevented thrombosis without increasing bleeding. In first-in-human studies, both mAbs were generally well tolerated and dose-dependently inhibited intrinsic pathway-triggered coagulation, with durable aPTT prolongation without affecting PT. Amrecibart and cenvacibart may offer tailored therapies for patients with different bleeding risk profiles. The trials are registered at www.clinicaltrials.gov as #NCT05102136 and #NCT05603195.

An electron-density point-cloud framework for robust protein-ligand interaction prediction

Abstract Accurate protein-ligand affinity prediction typically depends on precise 3D coordinates, limiting robustness when structures are low-resolution or predicted. We introduce E-CloudBind, a framework that fuses electron-density point clouds with intrinsic molecular graphs to model non-covalent and covalent interactions without relying on sub-ångström accuracy. Ligand electron densities are obtained by semi-empirical quantum calculations, whereas protein pockets are represented by van der Waals-guided Gaussian point clouds, a physically motivated proxy that preserves interaction geometry while tolerating coordinate noise. Point-cloud encoders capture local non-covalent patterns and a heterogeneous graph neural network integrates them with covalent features for affinity regression. Across PDBbind splits and out-of-distribution scenarios, E-CloudBind matches or exceeds leading sequence-, graph- and structure-based baselines, with markedly reduced sensitivity to resolution and to experimental-versus-predicted proteins. Case studies further illustrate atom-level interpretability and large-scale virtual screening. By decoupling interaction learning from exact coordinates, E-CloudBind enables robust structure-based modeling on heterogeneous conditions.

Hitting AML where it breathes: the peroxisome fat trap

Switchable adhesion of phase-transition eutectogels with integrated machine learning-enhanced intelligent adhesion sensing

Abstract Switchable adhesion underpins emerging technologies in robotics, microelectronics, and biomedical engineering. However, achieving switchable surface adhesion that can adapt to substrates with varying material compositions and surface roughness, while simultaneously enabling real-time and wireless monitoring of adhesion strength, poses a substantial challenge. Here, we present a eutectogel-based system that integrates electrothermally switchable adhesion with wireless sensing capability for in situ monitoring of adhesion forces. The switching mechanism is systematically elucidated through a combination of mechanical analysis and molecular-level characterization. The integration of machine-learning assisted adhesion sensing with dynamic gripping and locomotion enables safer and smarter robotic operation in adhesion joints, smart grippers and climbing robots. Demonstrations in adhesion-aware sensing, robotic grasping, and wall climbing validate the system’s practical utility, establishing a pathway toward next-generation intelligent adhesive interfaces that are both adaptive and self-perceptive.

Inflammatory cell death and monocyte dysfunction in VEXAS syndrome

Abstract VEXAS (vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic) syndrome is a severe adult-onset autoinflammatory disease caused by somatic mutations in the UBA1 gene, disrupting cytoplasmic ubiquitin-activating enzyme E1 function in hematopoietic progenitors. Its pathogenesis remains poorly understood, particularly the mechanisms by which UBA1 mutations disrupt myeloid cell function in the context of inflammatory stimuli. Here, we combine a genetically engineered THP-1 monocytic model with ex vivo analyses of blood and tissue samples from patients with VEXAS syndrome to investigate the consequences of the canonical UBA1M41V mutation. We show that UBA1-mutated monocytes exhibit tumor necrosis factor α (TNF-α)–induced cell death, characterized by receptor-interacting serine/threonine-protein kinase 1 (RIPK1) phosphorylation, and mixed lineage kinase domain-like– and caspase-8–mediated cell death. Importantly, we extend these findings to patient-derived CD14+ sorted cells, confirming that these cells undergo aberrant apoptotic and necroptotic cell death. Mechanistically, activation of these cell death pathways appears to be promoted by defective NF-κB–dependent transcriptional responses and reduced cFLIP(L) expression following TNF-α stimulation. UBA1-mutated monocytes also display blunted cytokine responses to Toll-like receptor (TLR) agonists despite preserved TLR expression, linked to an impaired NF-κB response. UBA1M41V-derived macrophages exhibit a proinflammatory transcriptional profile with increased chemokine secretion that promotes monocyte recruitment. In addition, these UBA1-mutated macrophages display impaired efferocytosis due to lysosomal dysfunction. Together, these findings reveal a pathogenic axis in VEXAS syndrome linking UBA1 loss of function and defective ubiquitination to RIPK1-mediated inflammatory cell death, impaired antimicrobial signaling, and defective resolution mechanisms. Our study provides novel mechanistic insights into the myeloid dysfunction underlying inflammation and cytopenia in VEXAS syndrome and supports the therapeutic targeting of inflammatory cell death pathways.

Lethal permeabilization of host bacteria to small-molecule compounds during phage penetration

Bridging the gap to pregnancy for modeling FNAIT prophylaxis

Misplaced-dipole engineered repairable fluoropolymer elastomer for flexible perovskite solar cell with excellent thermal-mechanical cycling resistance

Bone marrow endothelial injury by venetoclax and azacitidine

Controls on natural hydrogen generation during serpentinization of mantle rocks

Abstract Mantle rocks undergoing serpentinization can generate significant amounts of natural hydrogen, yet the rates and controlling processes remain poorly understood. Here, we constrain the possible hydrogen generation rates in two distinct mantle rock types, the fertile lherzolites of the Western Pyrenees and the depleted harzburgites of Northern California, to relatively low rates of ~0.1 to ~0.5 tonnes H₂ yr⁻¹ km⁻³ of reactive rock. When integrated over the full reactive volumes, this corresponds to total production rates of ~300 to ~600 tonnes H₂ yr⁻¹. By combining three-dimensional geophysical inversion with numerical modelling of fluid-rock processes, we show that hydrogen generation rates are mainly limited by H₂ saturation in the fluid and reaction kinetics. Under these constraints, hydrogen generation in mantle-derived serpentinization systems proceeds slowly, making rapid large-scale replenishment unlikely and suggesting that large, economically relevant accumulations, would require timescales of thousands to tens of thousands of years to develop.