Liver is a complex organ, capable of incredible regenerative capacity in the mammalian body. In addition to well-known hepatocytes, liver is composed of a complex cell microenvironment, with each cell type contributing to homeostasis, regeneration and disease. Modelling of the liver in vitro has been hampered by 2D cell lines being too simple and whole animal models too complex to underpin the mechanisms of liver biology. Liver organoids are emerging as a great tool to study liver (patho)physiology. Increasing in complexity, organoids have successfully been used to understand the cellular interplay in genetic and metabolic liver diseases, as well as liver cancer. This review will discuss the advancement in modelling cell complexity, maturation and liver architecture in a Petri dish. It will also provide a state-of-the-art perspective of what is still lacking and how the models can be further improved. Finally, we will conclude with our personal vision of where the field is going.

Heme-containing enzymes represent one of nature's most versatile catalytic platforms, capable of mediating a broad spectrum of redox and group-transfer reactions. This review provides a comprehensive overview of the structural and mechanistic principles that underpin the bioengineering of heme proteins for both natural and abiological transformations. Here, we discuss that the core of reactivity of such metalloenzymes is the iron-porphyrin cofactor, whose tunable oxidation and spin states enable the formation of diverse high-energy intermediates, including iron-oxo, carbene, and nitrene species. Using recent bioengineering applications, we also show that the catalytic behavior of these intermediates is not solely determined by intrinsic cofactor chemistry but is profoundly influenced by the surrounding protein scaffold. Key factors such as axial ligation, second coordination sphere interactions, hydrogen-bonding networks, and local electric fields collectively govern substrate binding, intermediate stabilization, and reaction selectivity. The review further discusses a few applications of de novo protein design and artificial metalloenzymes that provide unprecedented control over active-site architecture, allowing the creation of highly robust and tunable catalysts.

The p21-activated kinases (PAKs) are conserved serine/threonine kinases that couple cytoskeletal dynamics to diverse cellular processes. Despite emerging roles for PAK2 in development and disease, its regulated phosphorylation programs remain poorly defined. Here, we performed quantitative phosphoproteomic profiling of WT and Pak2[-/-] mouse embryos at embryonic day 9.5 and identified 151 differential phosphopeptide features, corresponding to 119 phosphosites and 116 PAK2-associated phosphoproteins. Differential events were predominantly downregulated and enriched for serine-proline motifs. Functional analyses highlighted modules involved in nuclear RNA processing, chromatin/DNA regulation, translation/RNA-binding, cytoskeletal signaling, and neurodevelopment. Finally, bulk RNA-seq and single-cell transcriptomic analyses integrated the genes encoding these phosphoproteins into cellular and developmental programs, providing transcriptomic context for the PAK2-associated post-translational events in early embryogenesis.

How do eukaryotic genomes acquire new functional parts, such as regulatory elements and long non-coding RNAs? Here, we propose that these parts arise from non-functional precursors primarily through non-adaptive processes that proliferate when selection is weak. Borrowing concepts from Markov chain theory, we represent each stage along the junk-to-function continuum as a discrete state with defined transition probabilities. This formalism makes clear why selection cannot act on future function, and why a part's current biochemical activity may be unrelated to its evolutionary past. We show how, under certain conditions, new intermediate states appear that increase the forward flux from junk DNA to new functional parts. These conditions-weak selection, abundant epistasis and quality control processes, and the accumulation of messiness-are characteristic of many eukaryotic genomes, allowing them to become more complex as new functional parts emerge.

Although chronic psychosocial stress is a well-established independent risk factor for cardiovascular disease (CVD), the underlying mechanisms linking stress to CVD remain incompletely understood. Here, we review clinical and preclinical research documenting how stress can activate innate immune cells. Based on these findings, we propose that trained innate immunity is one possible mechanism responsible for the lasting effects of stress on the innate immune system, and hence contributes to the association between stress and CVD. Trained immunity is defined as a persistent hyperinflammatory phenotype in innate immune cells, mediated by metabolic and epigenetic reprogramming of the bone marrow myeloid progenitor cells. As we describe herein, accumulating experimental evidence shows that chronic psychosocial stress can induce a maladaptive trained immunity program that accelerates atherosclerosis. By advancing our understanding of stress-induced trained immunity, we will move one step closer to designing new therapeutic, behavioral, and lifestyle strategies to counter maladaptive immune activation and reduce risk for CVD, one of the greatest causes of morbidity and mortality worldwide.

Maintenance of protein homeostasis requires coordination between protein synthesis and degradation, yet whether autophagy directly controls translational machinery remains unclear. Here we explored whether autophagy maintains translational fidelity. Proteomic analysis of ubiquitin-enriched fractions and p62-associated proteins in autophagy-deficient tissue confirmed previous studies identifying ribosomal proteins and RNA-binding factors among candidate autophagy cargo. Pharmacological or genetic impairment of autophagy increased translational errors, particularly during oxidative stress. Components of the translational machinery localised to LC3-positive autophagic structures and accumulated in human brain tissue affected by neurodegeneration. These findings support a hypothesis in which autophagy preserves protein synthesis quality by removing damaged translational machinery. We term this proposed mechanism translophagy, potentially linking autophagy dysfunction, oxidative stress and the aberrant proteins in neurodegenerative disease.

Peptides and antibodies occupy complementary therapeutic niches. Peptides offer compact size and high-affinity recognition of difficult targets, whereas antibodies provide exquisite specificity, long serum half-life, and effector functions. Here we review strategies that merge these modalities by engineering peptides into immunoglobulins and antibody fragments. We discuss peptide grafting into hypervariable and framework loops, terminal fusion to antibody chains, Fc-peptide fusion proteins, and bioconjugation approaches. Across these formats, we examine examples that illustrate the importance of insertion site, peptide topology, and linker design for folding, activity, and developability. We highlight how antibody fusion and conjugation with natural and engineered peptides can confer novel target binding, conditional activation, altered biodistribution, improved pharmacokinetics, or intracellular delivery, and we discuss how disulfide-rich and macrocyclic peptides have enriched the design space. Finally, we also point out key translational challenges, including manufacturability, stability, heterogeneity, and immunogenicity, and propose that advances in protein design and chemical bioconjugation will enable next-generation therapeutics at the interface of antibodies and peptides.

Thiol-based redox switches utilize the unique nucleophilicity of cysteine and selenocysteine to dynamically link real-time cellular redox fluctuations to metabolic regulation and signaling pathways. Capturing the precise atomic-level thermodynamic and kinetic mechanisms driving these oxidative modifications has long been limited by the chemical instability of transient intermediates and the immense computational costs of classical molecular dynamics simulations. However, recent advancements in chemoselective small-molecule probes now allow for the high-purity trapping and enrichment of specific sulfenic and sulfinic acid states. In parallel, a paradigm shift toward machine learning, graph neural networks, and protein language models has bypassed traditional computational bottlenecks, enabling high-throughput, proteome-wide predictions of redox switches in seconds. Furthermore, emerging data reveal that these redox modifications do not merely alter well-structured proteins, but actively dictate conditional folding transitions and structural transformations within intrinsically disordered proteins and biomolecular condensates. Here, we provide a comparative overview of these dual experimental and computational advancements and highlight how the integration of generative diffusion models could facilitate the real-time simulation of conditional, multi-state structural ensembles across the redox proteome.

Scientific publishing is undergoing a systemic breakdown. Thousands of journals now collect article processing charges (APCs)-increasingly paid from public research funds-while providing minimal editorial oversight. This erodes standards, distorts incentives and threatens public trust in science. APC payments to six major publishers exceeded 8 billion dollars between 2019 and 2023, with almost no transparency over how these fees are set or spent. This cannot continue. We propose that public and institutional funds support journals-whether through APCs, subscriptions or bundled agreements-only where they are accredited against a transparent, enforceable quality standard. Implemented in stages and led by funding bodies and learned societies, such accreditation would narrow the space in which predatory journals operate without resorting to blacklists.

Encapsulins are self-assembling protein nanocompartments that selectively encapsulate cargo proteins to perform specialized functions. The Myxococcus xanthus encapsulin system incorporates three cargo proteins, of which EncB and EncC are classified as ferritin-like proteins with ferroxidase activity; however, their functional distinction remains unclear. Here, we investigate the individual contributions of EncB and EncC to iron mineralization within the encapsulin shell. We show that EncB and EncC produce iron-containing minerals with differing phosphorus content and that reducing the number of cargo proteins increases ferrous iron influx into the encapsulated mineral core. These findings provide the first functional comparison of two ferritin-like co-encapsulated cargo proteins and suggest that cargo composition may allow fine-tuning of mineral properties in response to cellular requirements.

Higher levels of plasminogen activator inhibitor-1 (PAI-1) circulate in the blood of people with obesity and predict hyperlipidemia and metabolic dysfunction-associated steatotic liver disease in studies predominantly conducted in women. To understand the roles of PAI-1 in women with obesity, we generated inducible whole-body PAI-1-depleted female mice (Pai-1[i-KD]). After high-fat diet feeding, Pai-1[i-KD] and littermate controls did not differ in body weight or energy balance. However, Pai-1[i-KD] females accumulated smaller inguinal WAT adipocytes. Spectral flow cytometry revealed reduced adipocyte progenitors and macrophages in inguinal WAT. Pai-1[i-KD] mice also exhibited decreased liver cholesterol. These preclinical findings argue that elevated PAI-1 alters subcutaneous adipose tissue biology and contributes to pathogenic phenotypes of obesity.

ErbB ligands activate distinct receptor dimers that regulate cell proliferation, but how they influence cell-cycle commitment remains unclear. Here, we show that EGF and HRG induce different modes of G1/S progression in ErbB2-amplified BT474 breast cancer cells. Unexpectedly, HRG, despite activating the more potent ErbB2-ErbB3 heterodimer, failed to accelerate G1/S progression; instead, EGF promoted earlier passage through the restriction point and faster S-phase entry. Mechanistically, EGF drove rapid and coordinated cell-cycle progression through ERK-FOS signaling, whereas HRG induced sustained MYC activity, elevated transcriptional heterogeneity, and caused extensive rewiring of cell-cycle regulatory networks. Together, these findings reveal that ligand-specific ErbB signaling controls not only the timing of cell-cycle entry but also the transcriptional architecture that governs proliferative fate decisions.

Aging is a complex biological process characterized by gradual impairment of organ function and increased susceptibility to disease. Among the systems most affected, the gut and liver are particularly important because of their role in metabolism, detoxification, and immune regulation. These two organs are closely connected through the portal vein and bile ducts, enabling continuous, bidirectional communication that coordinates homeostasis. Understanding how gut-liver communication changes with age is therefore critical, as even subtle alterations over time can have widespread effects on overall physiology and the progression of age-related disorders. In this review, we examine current knowledge on how aging shapes gut and liver physiology, highlighting changes in microbial composition, barrier function, and hepatic metabolism. We then discuss how sex influences the gut-liver axis during aging. Finally, we explore how disruptions in the gut-liver axis contribute to the development of age-related conditions.

Autophagy is an evolutionarily conserved cellular quality control pathway that responds to the metabolic state of the cell, and its dysregulation has been broadly associated with metabolic disorders like diabetes mellitus. Among different types of autophagy, mitophagy or the selective autophagic clearance of dysfunctional mitochondria has emerged as particularly relevant in pancreatic β-cell biology and its pathophysiology. Recent advances in functional genomics and animal studies implicate the autophagy/mitophagy pathway components as effector transcripts at diabetes risk loci, providing a new rationale for investigation of genetic determinants of autophagy/mitophagy in β cells. In this review, we take a β-cell centric perspective to examine the evidence for autophagy/mitophagy across four clinically relevant diabetes categories (type 1, type 2, gestational, and monogenic diabetes) and discuss the functional significance and complexity of these pathways in β-cell failure.

Photosystem II (PS II) is a water-plastoquinone oxidoreductase. This photosystem has a short half-life (< 1 h) which is driven by repair of oxidative damage to the D1 protein. Reactive oxygen species (ROS) produced by the photosystem under stress conditions oxidatively modify key residues, which trigger the sequential disassembly of the photosystem. We hypothesize that the principal target for ROS damage of D1 is [332]His. This residue is a ligand to Mn1 of the Mn4CaO5 cluster and is the first residue exhibiting oxidative modification during photoinactivation. Recently, cryo-EM studies indicate that PS II monomers containing oxidatively modified [332]His lose their manganese clusters and the extrinsic proteins associated with the photosystem, both hallmarks of the earliest stages of PS II turnover/repair.