Biomaterial Mechanical Microenvironment and Tissue Functional Regeneration： From Static Regulation to Dynamic Remodeling

Organoid technology has become a highly promising in vitro model in biomedical research, but its traditional culture relies on animal-derived basement membrane matrices, which suffer from batch-to-batch variability, undefined biochemical composition, and poorly controllable mechanical properties. Recent studies have confirmed that matrix stiffness, viscoelasticity, topography, and temporal mechanical dynamics play regulatory roles equally as critical as biochemical signals in stem cell self-organization, lineage specification, and functional maturation of organoids. This review systematically summarizes the progress in understanding how extracellular matrix (ECM) mechanical properties regulate organoid formation. Focusing on the YAP/TAZ factors and the integrin receptors, which engage in crosstalk with biochemical signals to synergistically determine cell fate. We then elaborate how specific mechanical parameters, such as stiffness, viscoelasticity, and topography, influence each distinct stage of organoid development, from initial aggregation and compaction, through proliferative expansion, to lineage-committed differentiation and terminal functional maturation. Appropriate stiffness facilitates efficient aggregation and symmetry breaking, while excessive rigidity impairs proliferation and accelerates cellular senescence. In hepatic organoid expansion where only stiffness matching liver tissue (~6.2 kPa) supported robust growth. Viscoelastic stress relaxation promotes integrin clustering, enhances focal adhesion maturation, and stimulates endogenous ECM deposition, thereby supporting self-organized crypt-villus architecture and basement membrane assembly. Notably, dynamic softening of hydrogels enables optimal crypt budding, whereas rapid stress relaxation permits nephron segment coiling and drives vascular organoid maturation toward arteriole differentiation, highlighting stress relaxation as a key temporal parameter. Topographical cues provide contact guidance that directs polarized outgrowth, lumen formation, and neuronal orientation, all of which are essential for achieving organotypic architecture. The embedded 3D printing of complex neural networks enables precise control over branching geometries, facilitating studies of morphogenesis under spatial constraints. The engineered mechanical microenvironments have been successfully applied to recapitulate disease phenotypes: stiff matrices induce fibrosis and epithelial-mesenchymal transition in liver organoids, promote chemoresistance in cholangiocarcinoma. Furthermore, we summarize engineering strategies for both natural and synthetic hydrogels aimed at creating mechanically defined culture platforms. Natural biomaterial modifications achieve spatial separation of mechanical constraints from biochemical support, enabling self-limited compressive stimulation within a defined time window. Synthetic polymers allow independent tuning of stiffness, ligand density, and viscoelasticity through adjustable crosslinking and dynamic non-covalent interactions, thus providing orthogonal control over multiple mechanical parameters. Dynamic microenvironments that mimic progressive stiffening as observed in development, tumor evolution, or fibrosis can be constructed using composites that undergo spontaneous structural transitions. In the future, critical challenges include deciphering the synergistic rules governing mechanical-biochemical interplay, developing spatiotemporally programmable hydrogels that enable on-demand modulation of local mechanics, establishing comprehensive databases of tissue-specific mechanical parameter windows, and advancing standardized high-throughput screening platforms to accelerate preclinical validation and clinical translation. Moreover, integrating real-time mechanical sensors and non-invasive imaging techniques would allow dynamic monitoring of intracellular forces and tissue-level stress during culture. Ultimately, this review aims to provide a theoretical foundation for deepening our understanding of mechano-biological coupling mechanisms and for rationally designing high-fidelity, physiologically relevant organoid culture platforms that fully harness the regulatory power of mechanical cues.

Tissue repair requires the reconstruction of functional vascular networks to restore oxygen and nutrient delivery, eliminate metabolic waste, and maintain tissue homeostasis. Although biochemical factors such as growth factors and cytokines have been extensively studied in vascular regeneration, increasing evidence indicates that mechanical signals from the extracellular matrix (ECM) are equally important regulators of angiogenesis. Matrix stiffness, as a major biomechanical property of the tissue microenvironment, provides critical information that influences endothelial cell (EC) behavior and vascular remodeling through mechanotransduction. However, the biological effects of matrix stiffness are not universally defined as either pro- or anti-angiogenic, but are highly dependent on tissue context, cellular composition, disease progression, and the dynamic changes occurring during tissue repair. Therefore, understanding how matrix stiffness regulates vascular regeneration is essential for developing more precise strategies for tissue engineering and regenerative medicine. In this review, we summarize recent advances in the regulation of angiogenesis by matrix stiffness during tissue repair from multiple perspectives. First, we discuss the direct effects of matrix stiffness on EC behaviors, including differentiation, morphological remodeling, migration, proliferation, barrier maintenance, and vascular network formation. ECs sense mechanical alterations in the surrounding matrix through various mechanosensitive structures, including integrins, focal adhesion complexes, and mechanosensitive ion channels such as Piezo1 and TRPV proteins. These mechanical signals are subsequently transmitted through intracellular pathways involving FAK, RhoA/ROCK, MAPK, and Hippo-YAP/TAZ signaling, leading to cytoskeletal reorganization and transcriptional regulation. Through these mechanisms, matrix stiffness regulates endothelial functional states and determines the balance between regenerative vascular formation and pathological vascular remodeling. Beyond the direct regulation of ECs, we further highlight the importance of stiffness-mediated intercellular communication within the regenerative microenvironment. Matrix stiffness can regulate the behavior and secretory profiles of vascular-associated cells, including mesenchymal stem cells and macrophages, thereby indirectly affecting endothelial function through paracrine mechanisms. Changes in matrix mechanics influence the secretion of angiogenic factors, inflammatory cytokines, and extracellular vesicles from these cells, creating a mechanical regulation network that coordinates vascular regeneration. This perspective extends the traditional concept of matrix stiffness from a physical support structure to an active regulatory signal that integrates multiple cellular responses during tissue repair. Furthermore, we discuss the tissue-specific effects of matrix stiffness in vascular regeneration across different organs, including the aorta, lung, liver, and heart. Although excessive matrix stiffening is frequently associated with fibrosis and vascular dysfunction, the consequences of mechanical changes vary considerably among tissues. Moderate stiffness alterations may support endothelial activation and vascular stabilization in certain regenerative contexts, whereas persistent pathological stiffening can promote inflammation, endothelial dysfunction, and impaired vascular repair. These findings emphasize that the biological meaning of “soft” and “stiff” microenvironments should be interpreted according to specific tissue and experimental conditions rather than as universal mechanical classifications. Finally, we summarize current challenges and future perspectives in this field. A major limitation is the lack of standardized mechanical characterization among different studies, as stiffness measurements are influenced by material properties, testing methods, and experimental conditions, limiting direct comparison across research systems. Moreover, most existing models fail to fully reproduce the dynamic and viscoelastic properties of native ECM during tissue regeneration. Future studies should combine advanced biomaterials with standardized mechanical analysis, organoid models, and multi-omics approaches to establish more accurate mechanical regulatory maps. Overall, this review proposes that matrix stiffness is not merely a structural feature of tissues, but an active biological signal that regulates vascular regeneration through coordinated mechanotransduction and multicellular interactions. A deeper understanding of stiffness-mediated vascular regulation will provide new theoretical insights and therapeutic opportunities for improving tissue repair outcomes.

Objective This study aims to investigate the regulatory effects of microcapsules with different matrix stiffness on the survival, proliferation, adhesion, and osteogenic differentiation of mesenchymal stem cells (MSCs), to further explore the relationship between matrix stiffness and MSCs osteogenic behavior, and to evaluate the in vivo histological responses and osteogenic-related protein expression of the microcapsules using an animal subcutaneous implantation model.Methods In this study, three-dimensional carbon nanofiber (CNF)/sodium alginate (SA) microcapsules with different matrix stiffness were constructed. By adjusting the CNF content, three groups of microcapsules with stiffness values of 27.05, 35.90, and 52.40 kPa were prepared, and their physicochemical properties, including morphology, mechanical properties, particle size, chemical structure, and degradation properties, were characterized. The effects of these microcapsules with different stiffness on the survival, adhesion, and osteogenic differentiation of MSCs were evaluated. Live/dead staining was used to observe cell survival and distribution in the microcapsules, while alkaline phosphatase (ALP) activity and the expression of osteopontin (OPN) and osteocalcin (OCN) were examined to assess osteogenic differentiation. Immunofluorescence staining was further performed to observe the expression of Yes-associated protein (YAP) and integrin β1 under different stiffness conditions. In addition, a rat subcutaneous implantation model was used to evaluate the in vivo tissue compatibility of the microcapsules. Histological staining and immunohistochemical analysis were carried out to examine local tissue responses and the expression of osteogenic-related proteins.Results The results showed that all three groups of microcapsules exhibited regular and uniform morphology and similar particle size, and all groups showed good degradability. Except for the differences in mechanical properties, no obvious differences were found in the other physicochemical properties among the three groups. Live/dead staining indicated that all groups of microcapsules had good cytocompatibility. MSCs in the low-stiffness group were mostly round, whereas cells in the medium- and high-stiffness groups showed more obvious spreading. Higher matrix stiffness was more favorable for MSC adhesion, growth, and survival. ALP staining and quantitative analysis showed that ALP activity increased significantly with increasing matrix stiffness. The expression of the osteogenic-related proteins OPN and OCN also increased with increasing stiffness, and the high-stiffness group showed the strongest positive signals. Immunofluorescence results further indicated that YAP and integrin β1 fluorescence signals tended to increase in MSCs with increasing matrix stiffness, suggesting that matrix stiffness may influence MSC osteogenic differentiation through stiffness-related cell adhesion and mechanotransduction. In vivo experiments showed that all three groups of microcapsules exhibited good histocompatibility after subcutaneous implantation, although different degrees of degradation and local tissue response were observed. Immunohistochemical analysis showed that the expression levels of OPN and OCN increased with increasing matrix stiffness, and the high-stiffness group exhibited higher expression of these osteogenic-related proteins.Conclusion CNF endows SA microcapsules with tunable matrix stiffness. While maintaining good biocompatibility, the high-stiffness microenvironment can effectively promote the adhesion and osteogenic differentiation of MSCs, accompanied by increased ALP activity and enhanced OPN and OCN expression. The increased YAP and integrin β1 fluorescence signals in the high-stiffness group further indicate that matrix stiffness is closely related to cellular mechanical responses. These findings provide a three-dimensional cell carrier with potential application value for MSC delivery and bone tissue regeneration based on mechanical microenvironment regulation.

Objective Peripheral nerve injury is a common clinical condition that can impair the regulatory function of target organs, leading to functional disability; severe injuries may even result in lifelong disability. Therefore, how to effectively promote the morphological and functional reconstruction of injured peripheral nerves has become a critical scientific issue that urgently needs to be addressed. This study biomimicked the highly oriented structure and electrophysiological properties of peripheral nerves and investigated the feasibility and effectiveness of combining electrical stimulation (ES) with oriented conductive substrates for the repair of injured peripheral nerves at the cellular level.Methods Polydimethylsiloxane (PDMS) substrates with oriented microgroove structures of varying widths (0, 5, 10, and 20 μm) were fabricated using silicon wafers with micropatterned grooves as templates. Graphene-based conductive patterned substrates (GCPSs) with oriented microgrooves were then constructed via soft lithography and designated as GCPS 0, GCPS 5, GCPS 10, and GCPS 20 according to the groove width. The GCPSs were prepared from collagen/polycaprolactone (PCL) matrices containing 1.0 wt% graphene, and their morphology, wettability, and conductivity were characterized. Mesenchymal stem cells (MSCs) viability, proliferation, adhesion, and orientation on the substrates were evaluated before ES treatment. Then, ES at a frequency of 2 Hz and field strengths of 10, 20, or 50 mV/cm were applied to the cells for 10 min daily, over 3 or 7 consecutive days. Neural differentiation was assessed by morphological observation, immunofluorescence staining, and analysis of neural lineage-related markers.Results The oriented microgroove PDMS substrates supported long-term MSCs growth, guided cells to align along the groove direction, and promoted a morphological transition toward neural-like cells, exhibiting the capacity to induce MSCs differentiation toward the neural lineage. Microgroove structures effectively guided the alignment of MSCs and promoted the development of a more elongated, neural-like morphology, with the most pronounced morphological changes observed on grooves of 10 and 20 μm in width. Meanwhile, the incorporation of graphene endowed the GCPSs with favorable electrical conductivity, though the groove width did not significantly affect the conductive performance. The GCPSs showed good cytocompatibility with MSCs, and cell proliferation on these substrates exhibited a time-dependent pattern, with no significant effect observed from groove width. The inductive effect of ES combined with GCPSs on MSCs differentiation toward specific neural lineages was co-regulated by ES intensity (0, 10, 20, and 50 mV/cm) and microgroove width. ES at 10 mV/cm and 20 mV/cm promotes MSCs differentiation toward neurons on microgrooved substrates of various widths (5, 10, and 20 μm); under the same ES intensities, microgrooves with widths of 5 μm and 10 μm are more favorable for MSCs differentiation toward glial cells/SCs. Notably, low-intensity ES (10 or 20 mV/cm) combined with GCPS 10 significantly enhanced the neural differentiation of MSCs. Neurotrophic factor secretion levels by MSCs cultured on various GCPS substrates after 3 d of ES application was further detected. There were no significant differences in nerve growth factor-β (NGF-β) and brain-derived neurotrophic factor (BDNF) concentrations among the GCPS 5, GCPS 10, and GCPS 20 substrates, and these levels were not affected by variations in ES intensity. Under ES at 10 mV/cm, the glial cell line-derived neurotrophic factor (GDNF) concentration in each GCPS substrate was higher than that in the corresponding non-ES group on the same substrate. However, as ES intensity increased, the GDNF concentration decreased to varying degrees.Conclusion This study demonstrates that the combined strategy of “oriented topographical structure+ES” can effectively promote the neural differentiation of MSCs and holds promise for accelerating the repair process of peripheral nerve injuries.

Parkinson’s disease (PD) is a progressive neurodegenerative disorder characterized by the loss of dopaminergic neurons in the substantia nigra pars compacta and abnormal accumulation of α-synuclein. Its pathogenesis involves multiple interconnected processes, including oxidative stress, mitochondrial dysfunction, endoplasmic reticulum stress, neuroinflammation, and impaired protein homeostasis. Extracellular signal-regulated kinase 1/2 (ERK1/2), a major component of the mitogen-activated protein kinase signaling network, plays an important role in neuronal survival, synaptic plasticity, stress responses, and cell death. However, accumulating evidence indicates that ERK signaling in PD is not unidirectional, as both insufficient activity and sustained overactivation have been associated with neuronal dysfunction and disease progression. In this review, we systematically summarize the evidence for the pathological and protective roles of ERK signaling in PD. Insufficient ERK activity may impair neurotrophic support, synaptic plasticity, mitochondrial homeostasis, and neuronal adaptation to stress, thereby increasing the vulnerability of dopaminergic neurons. In contrast, persistent and excessive ERK activation may promote neuroinflammation, oxidative stress, mitochondrial dysfunction, apoptosis, autophagy-lysosomal impairment, and abnormal α-synuclein accumulation. Conversely, moderate and transient ERK activation induced by neurotrophic factors, bioactive compounds, or exercise may enhance neuronal survival, antioxidant defense, synaptic plasticity, and functional recovery. These apparently contradictory findings suggest that ERK activation itself cannot be simply classified as either neuroprotective or neurotoxic. Based on the available evidence, we propose the concept of an “ERK activity window” to explain the context-dependent effects of ERK signaling in PD. This hypothesis suggests that ERK activity needs to remain within an appropriate dynamic range to maintain neuronal homeostasis and adaptive responses. ERK activity below this range may weaken neurotrophic support and stress resistance, whereas prolonged and high-intensity activation may exceed the adaptive capacity of neurons and glial cells, thereby promoting inflammatory and degenerative processes. Importantly, this functional window should not be regarded as a fixed threshold. Its biological effects are likely determined by the intensity and duration of activation, cell type, subcellular localization, upstream stimuli, interaction with other signaling pathways, and disease stage. This framework may therefore help reconcile the apparently inconsistent findings among different PD models and experimental conditions. From a therapeutic perspective, these findings indicate that ERK should not necessarily be considered a target for simple activation or inhibition. Instead, restoring ERK signaling toward an appropriate functional state may represent a more rational strategy. Exercise is of particular interest because it can simultaneously influence neurotrophic signaling, oxidative stress, inflammation, synaptic plasticity, and ERK activity, potentially promoting adaptive remodeling of ERK signaling according to pathological conditions. Nevertheless, current evidence is predominantly derived from cellular and animal studies, and the dynamic, cell-specific, and spatial characteristics of ERK signaling remain insufficiently defined. Blood-brain barrier penetration, long-term safety, preservation of physiological ERK functions, and the lack of reliable biomarkers for monitoring brain ERK activity also limit clinical translation. Importantly, the “ERK activity window” remains a theoretical framework, and direct clinical evidence that ERK modulation can alter the progression of human PD is currently lacking. Future studies should therefore integrate longitudinal analysis, cell-specific and spatial approaches, and clinical biomarker validation to determine whether precise regulation of ERK signaling can provide a viable strategy for disease-modifying intervention in PD.

The abnormal accumulation, structural change, and spread of tau protein in the brain are now considered important factors in the progression of Alzheimer’s disease (AD) and many other tau-related neurodegenerative diseases. Clinical and pathological studies have shown that the amount and distribution of abnormal tau protein are closely related to synaptic damage, neuronal loss, and cognitive decline. Therefore, tau protein has become an important target for disease-modifying therapy. However, tau is difficult to target because it has no fixed structure, lacks a clear active site, undergoes many post-translational modifications, and forms different abnormal structures in different diseases and disease stages. Therefore, the effective and selective removal of harmful tau species is both a major opportunity and a major challenge in drug discovery for neurodegenerative diseases. Proteolysis-targeting chimeras (PROTACs) offer a new strategy to deal with this problem. Unlike traditional inhibitors, PROTACs do not simply block protein function. Instead, they bring the target protein close to an E3 ubiquitin ligase, which leads to ubiquitination and degradation of the target protein by the proteasome. This event-driven mechanism may be especially useful for disease-related proteins such as tau protein, which are hard to inhibit with traditional small molecules. In recent years, tau protein-targeting PROTACs and related degradation strategies have shown promising effects in cell and animal models. These effects include reducing abnormal tau protein levels, lowering synaptic toxicity, regulating disease-related signaling pathways, and improving behavioral or cognitive outcomes. These studies support targeted tau protein degradation as a promising therapeutic approach, although many problems still need to be solved before clinical use. In this review, we summarize recent progress in the development of tau protein-targeting PROTACs and discuss the main factors that affect their design and activity. We focus on the choice of tau protein-binding ligands, their binding sites, and their ability to recognize abnormal tau protein rather than normal tau protein. We also discuss the selection of E3 ligase recruiters, linker length and structure, ternary complex formation, and structure-activity relationships that influence degradation potency and selectivity. In addition, we introduce new strategies beyond traditional proteasomal degradation, including the regulation of tau protein phosphorylation and autophagy-based clearance. These approaches may be more suitable for different tau species or different stages of disease. A major challenge for tau protein-targeting PROTACs is delivery to the central nervous system. Many PROTACs have high molecular mass, large polar surface area, flexible structures, and may be removed by efflux transporters. As a result, they often have poor ability to cross the blood-brain barrier and may not reach enough free drug levels in the brain. Therefore, the therapeutic potential of tau protein-targeting PROTACs depends not only on their ability to degrade tau protein in vitro, but also on whether they can achieve long-lasting, selective, and safe target engagement in important brain regions after practical administration. Future studies should focus on improving selectivity for abnormal tau protein, understanding E3 ligase function in neural cells, designing brain-penetrant molecules, developing better delivery methods, and testing these drugs in models that better reflect human disease. Overall, tau protein-targeting PROTACs are a promising and fast-developing strategy for treating tau-related diseases. However, their successful clinical translation will require progress in chemical biology, medicinal chemistry, neuroscience, drug delivery, and translational medicine.

Glycosylation constitutes a critical, highly intricate, and diverse co- and post-translational modification characterized by the covalent attachment of sugar moieties to proteins, lipids, or small RNAs. Complex glycans are ubiquitously present across nearly all living organisms. Despite their prevalence, our comprehension of glycan diversity remains limited, likely due to the inherent challenges associated with elucidating their structural complexity. Among post-translational modifications, protein glycosylation is notably prevalent, involving the enzymatic transfer of oligosaccharides to specific amino acid residues by glycosyltransferases. This modification plays a pivotal role in modulating protein function, including participation in various biological and biochemical recognition processes. G protein-coupled receptors (GPCRs), encoded by approximately one thousand genes, share a conserved architecture comprising seven transmembrane helices interconnected by three intracellular and three extracellular loops. Representing the largest family of human membrane proteins, GPCRs regulate a vast array of physiological and pathological processes, thereby constituting the most extensive class of therapeutic targets. In mammalian systems, the majority of GPCRs undergo glycosylation predominantly at their extracellular N-terminus or extracellular loops, primarily through N-linked and O-linked glycosylation. These covalent carbohydrate modifications exert profound effects on multiple facets of GPCR biology, including endoplasmic reticulum folding and quality control, membrane trafficking and surface expression, receptor internalization and recycling or degradation, ligand binding and signal transduction, biased signaling, and receptor dimerization. Notably, glycosylation can exert both positive and negative regulatory influences on these processes. Dysregulation of GPCR glycosylation has been implicated in a range of pathological conditions, such as cancer, diabetes, and neurological disorders, underscoring its physiological and pathological significance. Aberrant glycosylation patterns may lead to dysfunctional receptor signaling, thereby contributing to disease progression. Consequently, elucidating the precise roles of glycosylation in GPCR function not only enhances fundamental understanding of receptor biology but also facilitates the development of novel therapeutic strategies targeting glycosylation pathways or specific GPCR glycoforms. Although the functional consequences of glycosylation have been investigated for several decades, delineating their exact structural underpinnings has remained challenging due to the intrinsic flexibility and heterogeneity of carbohydrate structures. Recent advances in structural biology, particularly single-particle cryo-electron microscopy (cryo-EM) and X-ray crystallography, have catalyzed a paradigm shift in our understanding of GPCR glycosylation. The integration of these sophisticated structural techniques with biochemical approaches has redefined glycosylation as a complex structural element integral to receptor function. These developments provide a molecular framework for comprehending how glycans influence receptor pharmacology and open new avenues for the rational design of glyco-engineered biologics and allosteric modulators targeting specific extracellular motifs. This mini-review concentrates on glycosylation within the GPCR superfamily, summarizing the impact of glycosylation modifications on receptor function and regulatory mechanisms, recent progress in the structural biology of GPCR glycosylation, and prospective directions for future research in this field.

Objective Alcohol use disorder (AUD) is a chronic, relapsing brain disorder characterized by compulsive alcohol drinking and impaired control over alcohol use. Its pathogenesis involves functional abnormalities in neural circuits underlying reward, cognitive control, and emotion regulation. Current clinical assessments primarily rely on psychometric scales, which rely on subjective experience, are susceptible to recall bias and concealment, and provide limited insight into the aberrant neural responses underlying AUD. Addiction involves state-dependent, non-stationary neural dynamics, necessitating analytical approaches capable of capturing the time-varying features of brain states. Microstate analysis characterizes the spatiotemporal dynamics of brain activity by identifying brief, recurrent, quasi-stable spatial patterns in electroencephalography (EEG), and thus provides a useful tool for investigating neural mechanisms in resting-state and task states. Previous studies have shown that microstate parameters can reflect addiction-related pathological craving and functional impairment. However, the alcohol cue-induced temporal reorganization of microstates in patients with AUD remains to be investigated. In this study, we applied microstate analysis to investigate aberrant microstate patterns in patients with AUD in response to alcohol cue exposure.Methods Twenty-one patients with AUD admitted to Shandong Daizhuang Hospital from October 2023 to January 2024 were enrolled. All participants met the ICD-11 diagnostic criteria. A visual cue exposure paradigm containing alcohol cues was used. EEG data were recorded sequentially during eyes-open resting state, neutral cue exposure, and alcohol cue exposure, with eyes-closed resting-state data collected after each condition for microstate analysis. Subjective craving intensity was assessed using a visual analogue scale (VAS) before and after cue exposure.Results The microstate topographies across conditions were largely consistent with the four topographies of microstate classes A–D. Compared with the resting state, after exposure to alcohol cues, AUD patients showed a significant increase in the mean duration (t=4.341, P<0.001), coverage (t=12.645, P<0.001), and occurrence (t=2.740, P=0.030) of microstate B, while the mean duration (t=–5.196, P<0.001) and coverage (t=–7.623, P<0.001) of microstate A significantly decreased. Under neutral cues, the occurrence (t=5.806, P<0.001), mean duration (t=6.541, P<0.001), and coverage (t=9.496, P<0.001) of microstate C significantly increased. In addition, under alcohol cues, the transition probabilities between microstates A and B and between microstates C and D were significantly increased (A→B: t=12.401, P<0.001; B→A: t=2.637, P=0.024; C→D: t=3.390, P=0.007; D→C: t=3.690, P=0.004), whereas the bidirectional transition probabilities between microstates A and C were significantly decreased (A→C: t=–7.861, P<0.001; C→A: t=–3.253, P=0.008).Conclusion Alcohol cues can induce the retrieval of past drinking experiences and pathological craving in patients with AUD. These cue-elicited responses are reflected in EEG microstate parameters, which could serve as potential electrophysiological markers of cue reactivity in AUD and provide a basis for understanding the neural mechanisms underlying alcohol addiction.

无

Bacterial biofilms are major causes of chronic infections and medical-device-associated infections. Their persistence and resistance to antibiotics and host immunity impose substantial clinical and economic burdens. Non-invasive, dynamic, and quantitative characterization of biofilms is therefore essential for anti-infective research, drug development, and clinical risk assessment. Electrochemical impedance spectroscopy (EIS) has been widely used for biofilm monitoring because it is real-time, in situ, and non-destructive. However, EIS interpretation remains dominated by equivalent circuit models, which simplify biofilms into ideal electrical elements. Although equivalent circuit models are intuitive and convenient, they have fundamental limitations: fitted parameters often lack clear physical meaning; different circuit topologies can fit the same spectrum; model selection is subjective; and the models have limited predictive power and cannot directly retrieve intrinsic biophysical properties. These limitations restrict EIS from advancing from status monitoring to mechanistic interpretation. Transmission-line models, dielectric-spectrum theory, and machine-learning approaches have been attempted, but a unified theoretical framework that quantitatively links the multi-dimensional physical properties of biofilms to broadband impedance responses is still lacking. This review systematically examines the progress and bottlenecks of biofilm impedance analysis and proposes a physics-driven paradigm. The central hypothesis is that, within an appropriate frequency window, an interfacial biofilm can be approximated as a quasi-homogeneous dielectric membrane. This approximation does not deny microscopic heterogeneity; rather, it assumes that when the characteristic spatial scale of the applied field is much larger than that of local cellular aggregates, pores, and extracellular polymeric substances, the biofilm behaves as a continuous dielectric medium governed mainly by dielectric polarization and ionic migration rather than by faradaic electrochemical reactions. On this basis, a minimal operable physical model is constructed using Maxwell""s equations, dielectric relaxation theory, and thermodynamic constraints. The model adopts an electrode-biofilm-electrolyte geometry with appropriate boundary conditions and a frequency-dependent complex permittivity. Combined with physical priors and Bayesian inversion, it aims to retrieve a minimal set of intrinsic quantities from EIS data: biofilm thickness, biomass, conductivity, dielectric constant, and interfacial adhesion-related equivalent electromagnetic characteristics. These parameters can be independently cross-validated by confocal laser scanning microscopy, quartz crystal microbalance, conductivity measurements, and microfluidic mechanical assays, and can be correlated with biological features such as metabolic activity, three-dimensional structure, mechanical properties, interfacial stability, and stress resistance. Compared with conventional equivalent circuit models, the proposed framework offers clearer physical definitions, improved inversion uniqueness through physical constraints such as Kramers-Kronig consistency, predictive capability through the evolution of physical parameters, and compatibility with independent non-electrochemical validation. Because multi-parameter inversion from limited-band EIS data is inherently ill-posed, the framework requires physical priors, multi-frequency measurements, orthogonal experimental calibration, and multimodal constraints to reduce parameter coupling and non-uniqueness. The validity of the quasi-homogeneous approximation must be further defined for highly heterogeneous, disrupted, or non-continuous biofilms, and the retrieved parameters must be cross-checked across strains and culture conditions. Nevertheless, this physics-based paradigm may enable EIS to provide quantitative physical portraits of biofilms rather than merely indicating whether biofilms are present or increasing. It may also be extended to cell monolayers, stem-cell layers, organoid sheets, and other two-dimensional interfacial living systems, offering a new theoretical route for spectroscopic characterization of biological interfaces.

Parvoviruses are small, non-enveloped, single-stranded DNA viruses. Unlike most oncolytic platforms, several members inherently prefer rapidly proliferating tumor cells, not through genetic engineering but because their life cycle depends on mitotic host cells. Their non-structural protein 1 (NS1) is the principal effector of this natural oncolytic activity. Integrating current knowledge of its structure, tumor-killing mechanism and translational development, we argue that NS1 constitutes a distinct class of oncolytic agents: a structurally definable and engineerable cytotoxic effector whose antitumor value does not depend on viral replication. Structurally, NS1 comprises three modules acting in concert: an N-terminal HUH endonuclease domain, a central helicase/ATPase domain, and C-terminal transactivation domain. These endow NS1 with site-specific DNA endonuclease and helicase activities as well as pro-apoptotic activity, forming the molecular basis of the intrinsic oncolytic effect of the genus parvovirus. Mechanistically, the N-terminal endonuclease introduces single-strand breaks into host genomic DNA; separately, the C-terminal transactivation domain (TAD2) activates ataxia-telangiectasia mutated and Rad3-related kinase (ATR), which phosphorylates CDC25C at Ser216 and thereby inactivates the cyclin B/CDK1 complex, imposing a p53-independent G2 arrest. Because p53 is the most frequently mutated gene in the majority of human malignancies, this p53 independence is central to the broad applicability of NS1, and p53-deficient tumor cells are not less sensitive to NS1 than their wild-type counterparts. In parallel, mitochondrial damage induced by NS1 promotes sustained opening of the mitochondrial permeability transition pore (mPTP), collapse of the mitochondrial membrane potential, release of cytochrome c, an increased Bax/Bcl-2 ratio and ultimately caspase-9/3/7-dependent apoptosis. When apoptotic cells are not promptly cleared, secondary necrosis releases damage-associated molecular patterns (DAMPs) including HMGB1, extracellular ATP, HSP70 and tumor-associated antigens; notably, H-1PV infection elevates extracellular ATP release in a dose-dependent manner. These DAMPs, together with cGAS-STING activation triggered by mitochondrial DNA leakage, convert local killing into systemic immune surveillance: macrophages polarize from M2 towards an inflammatory M1 phenotype through TLR4 and P2X7 signaling and secrete TNF-α, IL-12 and IL-1β together with the chemokines CXCL9/CXCL10 that recruit natural killer (NK) cells and CD8⁺ T cells, while modulation of natural cytotoxicity receptor (NCR) ligands sensitizes NS1-expressing tumor cells to NK-mediated killing. The translational evidence forms a four-tier chain: broad cytotoxicity across human tumor cell lines; in vivo efficacy in rodent tumor models; clinical activity and safety in canine spontaneous tumors, which owing to their genetic heterogeneity and intact immune system predict human outcomes better than rodent models; and, for H-1PV, two ParvOryx trials (a phase I/IIa study in glioblastoma and a phase II study in unresectable metastatic pancreatic ductal adenocarcinoma) demonstrating a favorable safety profile with evidence of immune activation. Four interlocking bottlenecks still separate NS1 from routine clinical use: host immune clearance of delivery vectors, insufficient tumor-targeted delivery, the manufacturing paradox arising from NS1 cytotoxicity to producer cells, and tumor resistance and immune evasion. We therefore propose 3 viewpoints. First, the field is moving from oncolytic whole viruses towards a defined NS1 effector drug: non-viral delivery of NS1-encoding plasmids by lipoplexes or engineered exosomes would decouple NS1 cytotoxicity from viral replication and alleviate the constraints of immune clearance, biosafety and manufacturing. Second, companion-dog spontaneous tumor trials constitute the most leveraged translational bridge and should serve as the mandatory pre-registration platform for NS1-based agents. Third, the intrinsic immunogenicity of NS1-induced cell death makes NS1 a natural partner for immune checkpoint inhibitors, and cold-to-hot conversion of the tumor microenvironment should guide rational combination design. This integrated picture should inform the development of NS1-based oncolytic drugs over the next five to ten years.

Ferroptosis is a form of regulated cell death driven by iron-dependent lipid peroxidation, characterized by morphological and biochemical features and is mechanistically distinct from apoptosis, necroptosis, and classical necrosis. It is widely involved in tumor initiation and progression, metabolic reprogramming, therapeutic responses, and both primary and acquired drug resistance. The tumor microenvironment frequently imposes multiple stresses, including oxidative stress, nutrient deprivation, hypoxia, and chemotherapeutic insults. In response, tumor cells employ post-translational modifications to rapidly alter the abundance, catalytic activity, and subcellular localization of target proteins, thereby acquiring adaptive survival advantages. The ubiquitin-proteasome system (UPS) is a central system for maintaining protein homeostasis in eukaryotic cells. Through the coordinated action of the ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzyme (E2), and ubiquitin-protein ligase (E3) cascade and deubiquitinase (DUB), it enables substrate-specific and reversible regulation of protein function. In the tumor microenvironment, E3 ubiquitin ligases and DUBs jointly regulate the protein stability and functions of core ferroptosis effectors, including solute carrier family 7 member 11 (SLC7A11), glutathione peroxidase 4 (GPX4), ferroptosis suppressor protein 1 (FSP1), and acyl-CoA synthetase long-chain family member 4 (ACSL4). These regulators influence three major pathways: iron homeostasis, lipid peroxidation, and cellular antioxidant defense. This regulation reshapes the capacity of tumor cells to withstand oxidative damage and pharmacological stress. Consequently, the UPS has emerged as a critical post-translational regulatory node that determines ferroptosis sensitivity in tumor cells. Although previous reviews have discussed ferroptosis in cancer and ubiquitination in tumor biology, an integrated understanding of how the dynamic balance between E3 ligases and DUBs orchestrates tumor ferroptosis, and how these mechanisms may be therapeutically exploited, remains needed. This review focuses on the UPS-mediated regulatory network of tumor ferroptosis. We summarize the molecular mechanisms by which E3 ligases and DUBs regulate iron homeostasis, lipid metabolic reprogramming, and antioxidant defense, and emphasize that this regulation is markedly context-dependent and bidirectional. We also systematically evaluate the translational potential of various proteostasis-targeting tools to modulate ferroptosis. In addition to proteolysis-targeting chimeras (PROTAC) and deubiquitinase-targeting chimeras (DUBTAC), we discuss emerging technologies such as molecular glue, autophagy-targeting chimeras (AUTAC), and lysosome-targeting chimeras (LYTAC) in the context of tumor ferroptosis. These modalities offer distinct strategies for either degrading or stabilizing ferroptosis-related proteins. Furthermore, we explore the sensitizing value of targeting the ubiquitination-ferroptosis axis in combination with chemotherapy, radiotherapy, and immunotherapy, and analyze the theoretical rationale for and translational barriers to such combination regimens. Currently, clinical translation in this field remains constrained by multiple bottlenecks: the ubiquitination sites and chain types of ferroptosis-related substrates have not been fully characterized; the impact of tumor heterogeneity on UPS-ferroptosis regulatory circuits lacks systematic investigation; clinically validated predictive biomarkers for identifying patients likely to benefit are still scarce; and the in vivo safety evaluation framework for UPS-targeting small-molecule agents remains to be established. To address these limitations, we propose integrating ubiquitinomics, lipidomics, and single-cell multi-omics to construct panoramic regulatory maps, developing tumor microenvironment-responsive delivery systems to optimize the therapeutic window, and outlining directions for translating basic mechanistic findings into precision cancer therapy. Addressing these challenges will require collaborative efforts across chemistry, biology, and clinical oncology. In summary, this review aims to deepen the understanding of proteostasis-mediated ferroptosis regulation in tumors and to provide theoretical references for the development of novel anticancer interventions.

Social interaction is a continuously unfolding process in which partners adjust their coordination in response to changing goals, task phases, roles, strategies, and ongoing feedback. Such coordination is also shaped by prior interaction experience. Inter-brain synchrony (IBS), defined as the temporal coordination of neural activity across individuals during social interaction, provides an important window into the neural mechanisms underlying interpersonal coordination. Existing research shows that IBS varies across interaction contexts and changes over the course of an interaction. Prior interactions, repeated training, and external interventions may also influence subsequent patterns of IBS. This review synthesizes evidence for these changes and their relationships with behavioral coordination and interaction outcomes. It conceptualizes IBS flexibility along two interrelated dimensions —adaptability and experientiality—distinguished primarily by the factors on which IBS changes depend. Adaptability refers to dynamic adjustments in IBS in response to current interaction goals, task phases, role allocation, strategies, and feedback. These adjustments may manifest as increases or decreases in synchrony, transitions between distinct IBS states, or reconfiguration of inter-brain networks. Experientiality refers to the influence of prior interaction experience or external interventions on subsequent IBS. Such effects may manifest as post-task residual synchrony, changes across repeated interactions or training sessions, and alterations induced by noninvasive brain stimulation or multi-brain neurofeedback. Changes within an ongoing interaction may reflect both current interaction demands and accumulated experience. Identifying experience-dependent effects therefore requires distinguishing the phase in which experience is acquired from a subsequent assessment phase, while accounting for current interaction demands. Much of the available evidence comes from cross-sectional comparisons or assessments at only two time points; direct evidence of persistence, recurrence in later interactions, or generalization across partners and contexts remains limited. At the mechanistic level, several processes may contribute to these two dimensions of flexibility. Metastability may serve as an overarching dynamical principle, allowing interactive systems to shift between integration and segregation rather than remain in a fixed state of high synchrony. Within this framework, adaptability may be driven by mutual prediction and prediction-error updating, which support real-time changes in coupling strength and inter-brain state transitions. Experientiality may reflect history-dependent reconfiguration of this metastable system through accumulated co-occurrence, relational learning, reactivation, offline consolidation, and selective retention of functionally relevant coupling patterns, thereby reshaping subsequent coupling tendencies, state-transition patterns, and recurrent coordination configurations. It should be noted that these proposed mechanisms draw primarily on theoretical frameworks derived from research on learning and plasticity within individual brains and require validation in future hyperscanning studies. This review also summarizes quantitative measures relevant to the two dimensions of IBS flexibility. Adaptability can be captured by temporal-dynamic indices, such as sliding-window trajectories, variability, state transitions, directionality, time-lagged coupling, and dynamic network organization. Experientiality can be quantified using measures of post-task residual coupling, pre-post differences in IBS, delayed retention or reinstatement, and longitudinal trajectories of growth or maintenance. These measures provide statistical descriptions of IBS patterns but are insufficient, on their own, to support inferences about adaptability or experientiality. Such inferences require linking observed patterns of IBS change to specific interaction events and behavioral changes. Guided by this framework, future research should combine dynamic analyses, tracking at multiple time points, and experimental manipulations to examine how IBS flexibility relates to interaction demands, behavioral adjustments, and interaction outcomes, as well as whether experience-related effects are retained or transferred across different interaction contexts. Such work may deepen our understanding of the neural mechanisms underlying dynamic interpersonal coordination and provide a theoretical basis for research and applications in educational and clinical settings.

Objective Cysteine S-carboxyethylation is a recently identified post-translational modification with potential biological relevance, but its experimental identification is costly and time-consuming. Existing computational approaches mainly rely on sequence information and may not fully exploit positional differences around the central cysteine or the complementary representation capabilities of heterogeneous neural networks. This study aimed to develop an effective sequence-based model for prioritizing potential cysteine S-carboxyethylation sites.Methods We developed Cys-DHE, a dual-branch heterogeneous deep ensemble model for cysteine S-carboxyethylation site prediction. The benchmark dataset contained 1 204 training samples, including 602 positive and 602 negative samples, and an independent test set containing 913 samples, including 83 positive and 830 negative samples. Each sample was represented by a 41-aa sequence window centered on the target Cys. Binary-weight encoding (BWE) was used to combine amino-acid identity with a distance-dependent weighting mechanism, thereby emphasizing residues closer to the central site. The resulting 41×20 feature matrix was simultaneously input into two complementary branches: CNN-BiLSTM-Attention for extracting local patterns and bidirectional contextual dependencies, and CNN-Transformer for modeling cross-position interactions and long-range dependencies. The prediction probabilities of the two branches were integrated using fixed weights of 0.75 and 0.25, respectively, determined according to five-fold cross-validation results on the training set. Model performance was evaluated using sensitivity (Sn), specificity (Sp), accuracy (Acc), Matthews correlation coefficient (MCC), AUROC, and AUPR, and was further compared with individual branches and conventional machine-learning methods.Results In five-fold cross-validation, Cys-DHE achieved Sn, Sp, Acc, MCC, AUROC, and AUPR values of 0.680 3, 0.842 7, 0.751 9, 0.526 2, 0.825 7, and 0.856 0, respectively. Compared with the individual branches, the fixed-weight ensemble improved Sn, Acc, AUROC, and AUPR, indicating complementary information between the two network architectures. Cys-DHE also outperformed SVM, RF, XGBoost, and LightGBM across the evaluated metrics under the same cross-validation setting. On the independent test set, Cys-DHE achieved an Acc of 89.60%, Sn of 69.88%, Sp of 91.57%, MCC of 0.508 8, AUROC of 0.865 4, and AUPR of 0.485 2. Compared with DLBWE-Cys, Cys-DHE showed higher Sp, Acc, and MCC, although Sn, AUROC, and AUPR were slightly lower. Additional local Cys analysis showed that positive sites were more frequently accompanied by additional Cys residues in their local sequence neighborhoods. Among samples containing an additional Cys within the 41-aa window, the median distance from the central Cys to the nearest additional Cys was approximately 5 residues in positive samples and 10 residues in negative samples, with a significant difference (P=1.35×10??).Conclusion Cys-DHE effectively integrates position-aware sequence encoding with complementary heterogeneous deep-learning architectures and provides balanced performance for cysteine S-carboxyethylation site prediction. The model is intended as a sequence-based candidate prioritization tool for narrowing potential modification sites before subsequent structural analysis and experimental validation. Sequence proximity should not be interpreted as direct evidence of three-dimensional spatial proximity or disulfide-bond formation. Future work will incorporate structural descriptors, solvent accessibility, spatial neighborhood information, and reliable disulfide-bond annotations, while expanding experimentally validated datasets to further improve model robustness, interpretability, and generalization across proteins. An online prediction server is available athttp://www.lzzzlab.top/Cys-DHE/.

Systemic chronic inflammation associated with advancing age is closely linked to the onset and progression of multiple age-related diseases. This persistent inflammatory state, commonly termed inflammaging, reflects complex interactions between accumulated cellular damage and dysregulated immune responses. Inflammasomes are multiprotein complexes of the innate immune system that respond to cellular stress and damage and amplify inflammatory signaling. Their sustained or recurrent activation may contribute to the establishment and maintenance of inflammaging. This review focuses on the NLRP3 inflammasome and summarizes its activation mechanisms, regulatory pathways, and pathological significance within the aging microenvironment. Particular attention is given to neurodegenerative diseases, osteoarthritis, atherosclerosis, and type 2 diabetes, emphasizing shared activation mechanisms and tissue-specific effects. Although initiating stimuli differ among these conditions, several molecular events can contribute to NLRP3 inflammasome activation. Priming mediated by nuclear factor-κB (NF-κB) promotes the expression of NLRP3 and pro-interleukin-1β, whereas lysosomal damage, mitochondrial dysfunction, reactive oxygen species accumulation, mitochondrial DNA release, and potassium efflux can participate in activation pathways. These events are interconnected rather than obligatory sequential steps in every setting. Inflammasome assembly promotes caspase-1 activation, which facilitates interleukin-1β and interleukin-18 maturation and gasdermin D cleavage, potentially leading to pyroptosis. The relative contributions of these processes depend on cell type, tissue context, and disease stage. Within aging tissues, persistent damage signals and impaired cellular homeostasis may favor repeated inflammasome activation, establishing a cycle of inflammatory amplification and further tissue injury. However, the pathological consequences vary across organs. In neurodegenerative diseases, inflammasome signaling is associated with neuroinflammation and neuronal vulnerability; in osteoarthritis, it may contribute to joint inflammation and cartilage degeneration. In atherosclerosis, it is implicated in vascular inflammation and plaque progression, whereas in type 2 diabetes, it is linked to metabolic inflammation and impaired glucose homeostasis. These distinctions underscore the importance of interpreting shared molecular mechanisms within specific tissue environments. The review further discusses the therapeutic potential of targeting inflammasomes to mitigate inflammaging and associated disease processes. Strategies include inhibiting NLRP3 activation, modulating upstream stress and priming pathways, and limiting downstream inflammatory signaling or pyroptotic cell death. By interrupting shared mechanisms of inflammatory amplification, these approaches may offer opportunities for coordinated intervention across multiple age-related disorders. Nevertheless, mechanistic overlap does not imply uniform therapeutic responses, and benefits observed in experimental models should not be equated with established clinical efficacy or evidence of delayed human aging.Successful clinical translation will require careful consideration of disease stage, tissue specificity, treatment timing, and the consequences of prolonged interference with innate immune defenses. Long-term safety is particularly important for older populations and preventive applications. The limited availability of validated biomarkers of inflammasome activity also complicates patient stratification, assessment of target engagement, and monitoring of therapeutic responses. Future research should clarify causal relationships, identify contexts in which inflammasome inhibition is most beneficial, and establish reliable translational endpoints. Overall, inflammasomes represent an important link between aging-associated damage and chronic inflammatory amplification, supporting their investigation as shared targets for early intervention in multiple age-related diseases, while highlighting the need for disease-specific evaluation and rigorous assessment of benefits and risks.

Bloodstream infection (BSI) is a severe systemic infectious disease with rapid disease progression and high in-hospital mortality. The widespread and irrational use of antibiotics in recent decades has continuously driven the emergence of drug-resistant strains, resulting in a year-by-year increase in the detection rate of multidrug-resistant pathogens in BSI patients. In this context, early and accurate diagnosis is particularly critical for guiding rational antimicrobial therapy, preventing disease deterioration and optimizing patient long-term prognosis. As the recognized gold standard for clinical BSI diagnosis, conventional blood culture remains the most reliable testing method, yet it has prominent inherent technical limitations, including long incubation cycles and slow turnaround times. Generally, the positive detection time of routine blood culture exceeds 12 h after venous blood collection. Each hour of delayed antimicrobial intervention significantly increases the risk of adverse clinical outcomes and patient mortality. Moreover, the lag of laboratory test results easily leads to blind empirical antibiotic abuse, which further promotes the spread of drug-resistant bacteria and substantially increases the difficulty of nosocomial infection prevention and control in clinical institutions. The primary report of blood culture, covering Gram staining results and positive alert time, can effectively distinguish Gram-positive cocci, Gram-negative bacilli and yeast-like fungi, enabling timely and targeted early empirical anti-infection treatment before the acquisition of definitive species identification and antimicrobial susceptibility data. This review systematically summarizes effective strategies to accelerate blood culture detection efficiency from two core dimensions: clinical process optimization and blood culture bottle product improvement. At the clinical process level, standardized blood collection procedures, strict specimen delivery within 2 h, satellite blood culture laboratory mode and hierarchical reporting system can significantly shorten specimen transportation and pre-processing time, thereby effectively advancing the release of primary detection reports. These optimized clinical strategies require almost no additional equipment investment and are highly feasible and promotable for grassroots and community hospitals with limited laboratory resources. In terms of product modification, the optimization of medium formula, the addition of high-efficiency antibiotic adsorption resins, and the improvement of CO₂ sensing membranes and gas-isolated bottle structures can effectively promote the resuscitation and proliferation of dormant or inhibited pathogenic microorganisms. Such consumable improvements can shorten the turnaround time of primary reports by 1–4 h and are fully compatible with mainstream automatic blood culture detection platforms widely used in current clinical practice. This review also comprehensively sorts out a variety of emerging BSI detection technologies, including FISH (fluorescence in situ hybridization), PCR (polymerase chain reaction), MALDI-TOF-MS(matrix-assisted laser desorption/ionization time-of-flight mass spectrometry), direct whole-blood pathogen detection, next-generation sequencing, T2MR(T2 magnetic resonance), biomarker detection and machine learning-assisted identification. The study deeply analyzes the unique technical advantages and practical clinical limitations of each detection technology, and clarifies that the large-scale clinical promotion and popularization of these novel detection methods are restricted by high operating costs, incomplete pathogen coverage, nucleic acid contamination risk, blood matrix interference and the lack of stable routine phenotypic drug susceptibility testing capacity. Although emerging detection techniques possess prominent rapid-detection advantages, traditional blood culture is still clinically irreplaceable because it isolates viable pathogens, which can then be subjected to Gram staining and phenotypic antimicrobial susceptibility testing to provide accurate results for precise antibiotic selection. Overall, blood cultures remain irreplaceable. So speeding up blood cultures is still an important way to address the pain points in BSI detection.

Skin tissue regeneration requires coordinated responses from immune cells, keratinocytes, fibroblasts, vascular endothelial cells, and other cell populations. Wound closure restores tissue continuity, but it does not necessarily indicate recovery of the epidermal barrier, dermal architecture, skin appendages, or tissue function. Chronic wounds and pathological scars are distinct manifestations of dysregulated repair: the former are commonly associated with persistent inflammation and impaired tissue formation, whereas the latter involve excessive extracellular matrix deposition and fibrosis. Their distinct pathological settings need to be considered when interpreting mechanisms and evaluating interventions. Network pharmacology integrates information on drug constituents, candidate targets, signaling pathways, and disease phenotypes to generate testable hypotheses about multicomponent and multitarget drug actions. It can narrow the range of candidates for subsequent studies, but associations derived from databases, network topology, functional enrichment, or molecular docking do not establish direct target engagement or a causal contribution to tissue regeneration. This review summarizes the biological basis and representative regulatory networks of skin wound repair and tissue regeneration, outlines the main steps and limitations of network pharmacology, and examines how it can be integrated with other techniques. Transcriptomics, proteomics, phosphoproteomics, and metabolomics can compare predicted networks with molecular responses measured in treated samples. Single-cell sequencing and spatial omics can further locate candidate targets in specific cell populations and wound regions, while sampling across repair stages helps assess their temporal context. Machine-learning methods can rank candidate targets for experimental follow-up, provided that model performance is evaluated using appropriate validation procedures. Biochemical and biophysical approaches, including direct binding assays and measurements of protein activity, can test whether a predicted drug-target relationship is supported at the molecular level. Target inhibition or genetic perturbation, together with rescue experiments where feasible, can then assess whether the target contributes to an observed drug response. Cell-based assays and animal studies can relate these molecular findings to processes such as re-epithelialization, angiogenesis, extracellular matrix remodeling, and restoration of skin structure. Together, these approaches can prioritize candidate targets and relate predicted drug actions to molecular, cellular, and tissue-level observations. Agreement between network predictions and experimental observations can support a proposed mechanism, but it does not, by itself, establish a causal role for the target. Database coverage and target-prediction algorithms can bias candidate selection; commonly used oral bioavailability and drug-likeness thresholds may be poorly suited to topical treatment; and the composition and tissue exposure of multicomponent preparations are not always known. In addition, target effects may vary among wound types, cell populations, anatomical regions, and repair stages. Two-dimensional cell cultures and animal models cannot fully reproduce human chronic wounds, and promising experimental results do not by themselves demonstrate clinical benefit. Together, these limitations may affect the reliability and applicability of network predictions. Future work should standardize data sources and reporting, account for pathological and spatiotemporal context, and use direct target-engagement and causal-intervention experiments to test key hypotheses. Drug dose, route of administration, local exposure, formulation stability, and safety should be evaluated alongside healing outcomes. Beyond wound area and inflammatory changes, endpoints should include barrier recovery, tissue architecture, relevant skin appendages, and functional properties when the injury model permits. A connected evidence pathway from network prediction through molecular and biological validation to clinically relevant evaluation may provide a more reliable basis for studying drugs intended to improve skin wound repair and tissue regeneration.

The deconvolution of contributors from mixed biological samples represents a long-standing core technical challenge in the field of forensic genetics, as the accuracy of this process directly impacts case investigation efficiency and judicial fairness. Based on whether the deconvolution relies on software as a boundary, this paper systematically reviews the two existing types of strategies for deconvoluting mixed biological samples. (1) Pre-processing strategies based on biological experiments, which achieve physical separation and independent detection of cells from different sources through microscope-guided single-cell isolation, or accomplish differentiation prior to detection via specific physicochemical treatments. Among these, physical separation based on differences in cell morphology mainly employs cell capture techniques characterized by low-to-medium throughput but high precision, using visualization tools such as microscopes for single-cell isolation, followed by specialized detection of the isolated cells. Chemical separation based on differences in biochemical properties mainly targets different body fluids and can directly link a body fluid to its contributing source, including differential lysis, as well as coding-region single nucleotide polymorphisms (cSNPs) and some compound markers. (2) post-processing strategies based on computational biology, which rely on bioinformatics algorithms to analyze mixed DNA testing data, enabling precise discrimination of contributor components. Among these, deconvolution based on probabilistic models mainly depends on probabilistic genotyping software (PGS), currently including software that applies two core algorithms: Markov chain Monte Carlo methods (MCMC) and maximum likelihood estimation (MLE). While clustering-based deconvolution relies on single-cell RNA sequencing (scRNA-seq) and single-cell ATAC sequencing (scATAC-seq), in which each cell is labeled with a distinct barcode, thereby allowing deconvolution of the resulting data and classification of each cell into its cell population of origin. On this basis, this paper summarizes the academic debates, advantages, and disadvantages of the two deconvolution strategies, evaluates their applicable scenarios and limitations in forensic practice, and proposes a graded application strategy based on sample quality assessment, as well as a pathway for the integrated use of the different deconvolution strategies, taking into account the diverse conditions of forensic exhibits and limited resources in routine forensic casework: for sperm-vaginal fluid mixed stains, differential lysis is preferentially adopted to obtain single-source male and female profiles. If same-sex contributors are involved, PGS is further used for deconvolution. For mixed samples that are conventional in amount and well preserved, PGS can be used for post-processing deconvolution after routine STR profiling. For mixed samples that are trace, aged, and poorly preserved, it is necessary to introduce linear amplification via transposon insertion (LIANTI) or primary template-directed amplification (PTA) for whole genome amplification (WGA) on the basis of isolated cells, thereby creating conditions for obtaining STR profiles. Finally, this paper looks forward to the future development of rapid testing instruments for mixed forensic samples and the promising applications of artificial intelligence in mixture deconvolution: the former includes the integration of microfluidic chips, the field deployment of portable nanopore sequencing and its combined use with PGS, while the latter includes the application of artificial intelligence technologies such as deep learning models and multimodal large language models, aiming to provide a methodological reference for the accurate deconvolution of mixed biological samples.

Objective Colorectal cancer liver metastasis (CRLM) is a leading cause of mortality in patients with colorectal cancer (CRC). Although local immunity centered on CD8⁺ T cells plays a critical role in suppressing tumor progression, the unique immunotolerant microenvironment of liver metastases often induces heterogeneous functional alterations in these cells. However, the lineage trajectory and molecular mechanisms underlying phenotypic remodeling of specific CD8⁺ tissue-resident memory T cells during CRLM progression remain incompletely understood. Therefore, this study aimed to investigate the lineage evolution and phenotypic remodeling characteristics of specific CD8⁺ tissue-resident memory T cells in the CRLM microenvironment, explore their potential communication with microenvironmental cells, and establish a clinical prediction and molecular subtyping strategy based on the perturbation profile of this cell subset.Methods This study integrated three single-cell RNA sequencing (scRNA-seq) datasets of CRC, spanning normal tissue, polyps, primary tumors, and liver metastases. Tissue preference analysis, T-cell state scoring, trajectory inference, and transcriptional regulatory analysis were performed to identify and characterize key CRLM-associated cell subsets. CellChat and MultiNicheNet were further employed to investigate intercellular communication and potential ligand-receptor interactions. Subsequently, differentially expressed genes of the identified subset were used to construct a gene interaction perturbation network in the cancer genome atlas (TCGA) cohort, followed by consensus clustering for molecular subtyping and independent validation in a gene expression omnibus (GEO) cohort. Finally, immunotherapy response and drug sensitivity analyses were performed to evaluate differences in immune microenvironmental characteristics and potential therapeutic responses among molecular subtypes.Results Single-cell analysis successfully identified a key CRLM-associated KLRB1⁺ CD8⁺ tissue-resident memory T-cell subset (CD8-Trm-KLRB1), which exhibited prominent tissue-resident and immunoregulatory features, together with relatively low cytotoxic activity. Pseudotime analysis suggested that CD8⁺ T cells may undergo a continuous state transition from GZMB⁺ CD8⁺ effector T cells (CD8-Teff-GZMB), through GZMK⁺ CD8⁺ effector memory T cells (CD8-Tem-GZMK), to CD8-Trm-KLRB1 cells, accompanied by stress-related functional remodeling. Antigen presentation and T-cell receptor activity analyses suggested that this state transition was not simply attributable to insufficient antigen presentation or complete loss of T-cell receptor signaling. Cell-cell communication analysis indicated enhanced potential immunoregulatory communication between neutrophils and CD8-Trm-KLRB1 cells in the CRLM microenvironment, with the resistin (RETN)-adenylyl cyclase-associated protein 1 (CAP1)-mediated RETN-CAP1 ligand-receptor interaction being particularly prominent. Meanwhile, the CD48 molecule (CD48)-programmed cell death 1 (PDCD1)-mediated CD48-PDCD1 ligand-receptor interaction associated with CD8-Trm-KLRB1 cells also showed an increasing trend. The clinical molecular subtypes (subtype 1 and subtype 2) established based on the characteristics of this cell subset showed high consistency across TCGA and GEO cohorts. Subtype 2, which was associated with poorer prognosis, exhibited higher tumor immune dysfunction and exclusion (TIDE) scores and stronger immune exclusion features, while also showing a lower half-maximal inhibitory concentration (IC50) for axitinib, suggesting greater potential sensitivity to anti-angiogenic agents and a possible greater clinical benefit from such treatment.Conclusion This study reveals dynamic functional remodeling of CD8-Trm-KLRB1 cells during CRLM progression and suggests that their state may be regulated by upstream signals from the metastatic microenvironment, including neutrophil-associated RETN-CAP1 signaling, while their interactions with downstream immune cells through immunoregulatory communication such as CD48-PDCD1 may contribute to local immune microenvironmental remodeling. The molecular subtyping framework based on the perturbation characteristics of CD8-Trm-KLRB1-associated genes stratifies patients according to distinct prognostic and immune microenvironmental features and suggests potential differences in therapeutic response, providing a new perspective for understanding the CRLM microenvironment and developing individualized treatment strategies.

Protein disulfide isomerase A1 (PDIA1; hereafter PDI) is best known as an endoplasmic reticulum oxidoreductase, but vascular injury and cellular activation generate extracellular and cell-surface PDI pools that contribute to thrombus formation. Extracellular PDI supports platelet recruitment and activation, thrombin generation, and fibrin formation, and genetic or pharmacological attenuation of PDI suppresses thrombosis in several experimental models. Extracellular PDI is therefore being explored as an antithrombotic target, supported by experimental studies in which PDI-directed interventions suppress thrombosis with limited changes in measured hemostatic endpoints. This review examines antithrombotic PDI ligands through two recurrent druggable regions: the CGHC-containing catalytic regions of the a and a′ domains and the non-catalytic b′/x region involved in substrate recognition and conformational regulation. Catalytic-region ligands act directly on thiol-disulfide exchange and include both covalent and reversible inhibitors. Their pharmacology depends not only on electrophilic reactivity, but also on local recognition around the active site, site preference, and reversibility. Because the catalytic motifs are relatively conserved among vascular thiol isomerases, controlling the thiol isomerase family activity profile is an important design consideration for this class. The b′/x region provides a different route to PDI modulation. The b′ domain contains the major hydrophobic substrate-binding pocket, whereas the adjacent x-linker participates in interdomain conformational coupling. Studies of rutin, Bepristat, and other mapped ligands identify this region as a recurrent recognition site, with H256 repeatedly emerging as a determinant of ligand sensitivity. Binding at the b′/x region can alter access of protein substrates, reposition the x-linker, and redistribute PDI conformational states. These effects help explain why some ligands inhibit reactions involving macromolecular substrates while preserving or even enhancing activity toward small-molecule disulfides. Current structure-activity studies also indicate that substituents directed toward the pocket exterior may provide opportunities to adjust polarity, membrane permeability, and exposure while retaining the interactions required for PDI modulation. Across ligand classes, antithrombotic activity is shaped by substrate type, catalytic reaction, thiol isomerase family profile, and cellular compartment and is therefore not fully captured by biochemical potency in a single reductase assay. For extracellular PDI drug development, membrane permeability and distribution are especially relevant because the intended pharmacological pool is extracellular or cell-surface PDI. The identity of the active chemical species after dosing also requires attention, particularly for compounds that undergo metabolism or interconversion in vivo. Interface mapping should therefore be considered together with substrate-dependent functional assays, family activity profiling, compartmental distribution, and pharmacokinetic exposure rather than with enzyme IC?? values alone. Human studies with isoquercetin provide an initial translational link between PDI-directed pharmacology and thrombosis-related pathways. Oral treatment has reduced platelet-dependent thrombin generation, increased plasma PDI-inhibitory activity, and altered related biomarkers in selected patient populations. These studies establish measurable pharmacodynamic effects in humans and provide a basis for outcome-focused evaluation of clinical antithrombotic efficacy and bleeding risk. Current evidence supports extracellular PDI as a mechanistically distinct antithrombotic target and points to two complementary medicinal-chemistry strategies. Catalytic-region ligands require coordinated control of reactivity, site preference, and thiol isomerase family activity profile, whereas b′/x ligands offer opportunities to tune substrate recognition, conformational regulation, and compartmental exposure.

Objective To address the lack of transposon classification methods specifically designed for plant genomic data and the need to further improve the overall classification performance of existing methods, we developed a high-performance plant transposon classification model.Methods Plant transposon sequences were collected and curated from the PlantRep database. After data cleaning, class merging, and removal of duplicate sequences, a classification dataset containing 268 525 sequences and 13 superfamily classes was constructed and divided into training, validation, and test sets at a ratio of 8:1:1. We designed DeepPlantTE by integrating multi-scale convolution, the large kernel attention mechanism, bidirectional long short-term memory (BiLSTM), and gated attention pooling for plant transposon classification at the superfamily level. The input sequence length was set to 3 107 bp, which was determined before training according to the 90th percentile of sequence lengths in the training set. Model performance was evaluated using accuracy, precision, recall, and F1-score, with precision, recall, and F1-score calculated using macro averaging. DeepPlantTE was evaluated on a self-constructed test set, a common-category subset, InpactorDB, and MnTEdb, and was compared with a convolutional neural network (CNN), a long short-term memory network (LSTM), BiLSTM, a gated recurrent unit (GRU), a bidirectional gated recurrent unit (BiGRU), DeepTE, TERL, TEclass2, and Terrier. The common-category subset was used for comparison with existing specialized tools, whereas InpactorDB and MnTEdb were used to assess the generalization ability of the model on independent data. In addition, loss function comparisons, stratified five-fold cross-validation, sequence length analysis, and ablation experiments were conducted to comprehensively evaluate model performance across different data sources, training settings, input sequence lengths, and model architectures.Results DeepPlantTE achieved a superfamily classification accuracy of 89.65% on the self-constructed test set, with macro-averaged precision, recall, and F1-score values of 86.92%, 85.00%, and 85.88%, respectively. On the common-category subset, the model achieved accuracies of 92.15% at the superfamily level and 95.43% at the order level. On InpactorDB, the accuracies at the order and superfamily levels were 99.27% and 97.07%, respectively, whereas the corresponding accuracies on MnTEdb were 96.54% and 92.94%. The loss function comparison showed that standard cross-entropy achieved the highest accuracy, precision, and F1-score, with values of 89.65%, 86.92%, and 85.88%, respectively. However, its recall was 85.00%, which was lower than the recall values obtained with weighted cross-entropy and weighted Focal Loss, which were 85.81% and 85.41%, respectively. In five-fold cross-validation, the mean accuracy, precision, recall, and F1-score were 89.29%±0.15%, 86.41%±0.38%, 84.60%±0.54%, and 85.42%±0.38%, respectively. Sequence length analysis showed that an input length of 3 107 bp produced the highest accuracy, precision, and F1-score, whereas an input length of 4 000 bp produced the highest recall. In the stepwise ablation experiment, the model accuracy decreased from 89.65% to 80.13%. The independent ablation experiments showed that replacing or removing BiLSTM, the large kernel attention mechanism, or the embedding layer resulted in varying degrees of performance degradation. Among these components, BiLSTM and the large kernel attention mechanism had relatively more pronounced effects on model performance.Conclusion DeepPlantTE can effectively improve plant transposon classification performance and demonstrates good accuracy, stability, and generalization ability. It can provide a more reliable tool for the automated analysis of repetitive sequences in plant genomes.

Tumor angiogenesis is a core pathological driver of malignant growth, invasion, and metastasis, and represents a major obstacle to effective anticancer therapy. Conventional antiangiogenic agents, including neutralizing monoclonal antibodies and small-molecule tyrosine kinase inhibitors, are constrained by drug resistance, on-target toxicity, and the compensatory activation of alternative pro-angiogenic pathways. Targeted protein degradation (TPD) has emerged as a promising strategy that may overcome these limitations. Unlike traditional inhibitors that operate through an occupancy-driven mechanism, degraders function through an event-driven catalytic mechanism in which a single degrader molecule can catalytically promote the ubiquitination and subsequent proteasomal or lysosomal degradation of multiple molecules of the same target protein, thereby eliminating the target protein together with all of its functions rather than merely blocking a single active site. This review systematically delineates the pathophysiological basis of tumor angiogenesis, encompassing the angiogenic switch driven by hypoxia-inducible factor-1α (HIF-1α), the aberrant structural and functional features of the tumor vasculature, and the self-reinforcing vicious cycles that sustain malignant progression. We then categorize the principal TPD modalities, with particular emphasis on proteolysis-targeting chimeras (PROTACs) and lysosome-targeting chimeras (LYTACs), and describe their design principles and mechanisms of action. PROTACs recruit E3 ubiquitin ligases to direct intracellular targets to the ubiquitin-proteasome system for degradation, whereas LYTACs engage lysosome-targeting receptors to route extracellular and membrane proteins to the lysosome, thereby expanding the degradable target space beyond the intracellular proteome. We further summarize recent advances in the application of TPD to antiangiogenic intervention at four distinct levels. First, at the signaling initiation level, degradation of vascular endothelial growth factor (VEGF) and its receptor vascular endothelial growth factor receptor 2 (VEGFR2) by aptamer-engineered PROTACs, antibody-based LYTACs, and small-molecule VEGFR2 degraders blocks the upstream trigger of the angiogenic cascade. Second, at the signal transduction level, degradation of focal adhesion kinase (FAK), B-Raf proto-oncogene serine/threonine kinase (BRAF), phosphodiesterase 4 (PDE4), and bromodomain-containing protein 4 (BRD4) interrupts intracellular signaling cascades and transcriptional responses. Third, at the microenvironment level, degradation of proviral integration site for Moloney murine leukemia virus (PIM) kinases curtails the pro-angiogenic signals secreted by tumor-associated macrophages. Fourth, multi-target synergistic strategies, exemplified by bispecific degraders that simultaneously target human epidermal growth factor receptor 2 (HER2) and VEGF-A, represent a preliminary exploration of cooperative intervention. Representative degraders reviewed herein elicit diverse antiangiogenic phenotypes, ranging from the direct inhibition of endothelial cell proliferation, migration, and tube formation to reductions in microvessel density and the suppression of vasculogenic mimicry. Finally, we discuss the major challenges that remain, including the precise matching of degradation kinetics with the vascular normalization window, target selectivity and tissue specificity, and the activation of compensatory networks and the emergence of resistance. Since the application of TPD in this setting remains at an early stage, the clinical translation of these approaches will require careful optimization of degradation efficiency, spatiotemporal control, and delivery strategies. We also envision future directions toward immunotherapy-based combination regimens and precision-degradation approaches guided by single-cell and spatial transcriptomics. In our view, TPD should not be regarded merely as a substitute for conventional antiangiogenic drugs, but rather as a molecular tool capable of systematically remodeling the tumor vascular microenvironment. When precise control of degradation efficiency and spatiotemporal regulation are integrated with a thorough understanding of the heterogeneity and compensatory capacity of tumor vessels, the full clinical potential of this strategy can be realized.

Renal fibrosis is the common pathological endpoint of many forms of chronic kidney disease (CKD), yet the mechanisms determining effective repair versus persistent fibrotic remodeling remain incompletely understood. Peritubular capillaries (PTCs) are not merely passive targets of fibrotic injury, but form a dynamic vascular microenvironment linking regional perfusion, oxygen delivery, tubular metabolism, endothelial barrier integrity, pericyte support, immune-cell trafficking, and extracellular matrix (ECM) remodeling. We propose that PTC abnormalities should not be interpreted simply as reduced capillary density, but as a multidimensional disturbance of microvascular structure and function. Structural vascular preservation does not necessarily indicate effective microcirculation: anatomically detectable capillaries may still exhibit impaired perfusion, barrier dysfunction, abnormal endothelial-pericyte interactions, disturbed oxygen metabolism, or reduced network connectivity. Conversely, increased angiogenesis does not necessarily signify successful vascular repair if vessels remain immature, poorly perfused, or inadequately integrated into the existing network. To organize these abnormalities, we conceptualize PTC remodeling into four predominant but nonexclusive states: endothelial barrier disruption with inflammatory adhesion; endothelial-pericyte uncoupling with no-reflow; dysregulated angiogenesis with network simplification; and capillary rarefaction with fibrotic matrix locking. These states are not sequential stages of CKD. They may coexist within the same kidney, overlap within the same anatomical region, and vary with disease etiology, injury phase, vascular compartment, and treatment exposure. We further propose six dimensions for evaluating PTC status—structural presence, functional perfusion, barrier integrity, pericyte coupling, oxygen metabolism, and network connectivity—so that vascular quantity, function, and reparative capacity are assessed separately rather than inferred from a single endothelial marker or capillary-density measurement. Mechanistically, these states converge on a bidirectional microvascular-fibrotic feedback loop. Barrier disruption facilitates leukocyte recruitment and alters vascular paracrine signaling. Endothelial-pericyte uncoupling compromises capillary stability, whereas abnormal pericyte contraction may increase local vascular resistance and contribute to no-reflow before overt capillary loss. Dysregulated angiogenesis may generate vessels that fail to mature or connect to an effective perfusion network. Progressive capillary rarefaction compromises oxygen delivery, while sustained tubular transport demand and mitochondrial dysfunction may aggravate regional oxygen supply-demand mismatch. Injured tubules, immune cells, endothelial cells, pericytes, and fibroblasts communicate through soluble mediators, extracellular vesicles, direct cell contact, and cell-matrix signaling. Importantly, fibrosis is not merely a downstream consequence of microvascular injury. Altered ECM may increase diffusion distance, modify mechanotransduction, impair endothelial survival and migration, and restrict capillary reopening and regeneration. We use the term matrix locking to describe this sustained constraint on vascular repair, while emphasizing that current evidence does not establish irreversible disease in humans. This framework also has therapeutic implications. PTC-directed treatment should not be reduced to increasing vessel number or uniformly promoting angiogenesis. Instead, intervention should be matched to the dominant and potentially reversible abnormality: preserving endothelial barrier function; restoring endothelial-pericyte stability and reducing pathological microvascular resistance; normalizing angiogenesis so that vascular growth is accompanied by maturation, barrier competence, network integration, and effective perfusion; and, in advanced rarefaction, relieving matrix-related constraints while supporting vascular regeneration. The context- and time-dependent effects of signaling through VEGF, Tie2, PDGF-B/PDGFRβ, eNOS/NO, and related pathways indicate that the same molecular signal may exert different effects depending on disease context and timing. Therapeutic efficacy should therefore be evaluated by restoration of effective microcirculation and tissue function rather than by isolated molecular changes or capillary counts alone. Despite increasing mechanistic evidence, the translational basis for this framework remains limited. Most causal and interventional data derive from animal or cell models, whereas human evidence is largely limited to pathological observations, patient-sample associations, and prognostic studies. Currently available circulating and urinary candidate biomarkers generally lack sufficient specificity for PTC injury or state assignment, and no standardized quantitative thresholds have been externally validated. The four-state model should therefore be regarded as a hypothesis-generating analytical framework rather than a clinicopathological classification system. Future studies should integrate spatially resolved tissue phenotyping, blood and urine biomarkers, and dynamic microcirculatory assessment in longitudinal cohorts. They should test reproducibility, assess prognostic value beyond conventional pathology, estimated glomerular filtration rate, and proteinuria, and determine whether changes in PTC status precede or mediate therapeutic benefit. Linking reproducible microvascular phenotypes to defined mechanisms, treatment response, and long-term renal outcomes will be essential to determine whether state-directed assessment and restoration of the PTC microenvironment can ultimately inform patient stratification and therapeutic decision-making.

Objective Cryo-focused ion beam (cryo-FIB) milling is an essential technique for the in situ structural analysis of frozen-hydrated specimens. It allows researchers to prepare site-specific lamellae within a cryo dual-beam scanning electron microscope (SEM), and the prepared lamellae are subsequently transferred to a transmission electron microscope (TEM) for tomographic data collection. However, in practical operation, both adjusting the coincidence point of the electron beam (EB) and ion beam (IB) fields of view and relocating lamellae after sample transfer are time-consuming processes. Current automated packages cannot provide a rapid solution for accurate coincidence point alignment. For this reason, we developed SA-FIB (semi-automatic FIB thinning tool), a fast positioning program dedicated to cryo-FIB milling.Methods SA-FIB was developed on the Thermo Fisher Scientific Aquilos 2 cryo dual-beam SEM platform using the Autoscript scripting package, and consists of three functional modules: coincidence point adjustment with milling-site recording, a milling assistant that supports both automated and manual operation, and a map module for on-site annotation. Coincidence point adjustment is completed semi-automatically, with no requirement for stage tilting. When the stage tilt and rotation angles remain unchanged, the working distances of the EB and IB, as well as the stage tilt angle, can be saved as a reusable reference. Next, the target is centered and focused in the IB field of view for coarse correction via the “add point” function, then moved to the center of the EB field of view for fine correction through the “E correct” function, which requires at most two rounds of correction. We used Saccharomyces cerevisiae and mouse sperm as test specimens. After sputter coating and gas-injection coating, lamellae were milled via a gradient beam-current process (30 kV; 0.5 nA to 50 pA), thinning from an initial size of approximately 13 μm×2 μm to a final thickness of less than 200 nm. We also measured the positioning errors and grid-loading errors of the two microscopes.Results Adjusting and recording a single coincidence site takes only 15 to 20 s, and all sites are automatically labeled on the low-magnification overview map of the grid. In semi-automatic mode, one site is automatically thinned to approximately 500 nm within 10 to 15 min, and then manually thinned to below 200 nm; after the lamella is transferred to the TEM, the map allows the target lamella to be located within a few minutes. In fully automatic mode, since the coincidence point has already been adjusted, the three preparation steps of AutoTEM Cryo (Eucentric tilt, Milling angle and Image acquisition) can be skipped, saving approximately three minutes per site. The positioning errors of Aquilos 2 and Titan Krios G3i are about 1 mm and 100 nm respectively, while the repeated grid-loading errors are 12 to 26 mm for the SEM and 60 to 141 mm for the TEM. Since the commonly used 200-mesh grids have a mesh size of approximately 100 mm, the largest TEM loading error exceeds this 100-mm criterion and makes coordinate-based targeting unfeasible; accordingly, SA-FIB only provides a relative-position map.Conclusion SA-FIB enables rapid coincidence point adjustment and milling-site annotation without stage tilting, simplifies the cryo-FIB milling workflow, facilitates lamella localization after sample transfer, and can be used either as a stand-alone aid for manual milling or in combination with fully automatic packages such as AutoTEM Cryo. It provides a convenient solution for cryo in situ structural analysis.

Ischemic stroke accounts for approximately 80% of all stroke cases worldwide and remains a leading cause of long-term neurological disability and mortality. Intravenous thrombolysis and endovascular thrombectomy constitute the mainstay of acute reperfusion therapies, yet their clinical utility is severely constrained by narrow therapeutic time windows. Even when large-vessel recanalization is successfully achieved, a considerable proportion of patients still suffer poor functional outcomes due to futile reperfusion and ischemia-reperfusion injury. Historically, preclinical studies have largely focused on direct neuronal protection against calcium overload, oxidative stress and ferroptosis. Unfortunately, most promising neuroprotective agents that exhibited robust benefits in rodent models have failed to demonstrate satisfactory efficacy in phase-Ⅱ and phase-Ⅲ clinical trials. One critical contributor to such translational failure is the disregard for the pathological disturbance within the cerebral vascular microenvironment. Instead of merely regarding neuronal death as an autonomous cellular event, accumulating evidence indicates that sequential biochemical disruption of perivascular microenvironment acts as an initiating amplifier that propagates damage toward brain parenchyma after ischemic insult. The cerebral vascular microenvironment centers on brain microvessels and comprises endothelial cells, endothelial glycocalyx, basement membrane, pericytes, astrocytic end-feet, perivascular microglia, as well as interstitial ions, inflammatory mediators, reactive oxygen species and extracellular matrix components. Distinct from the concept of neurovascular unit which emphasizes functional coupling between neurons, glia and vessels, the vascular microenvironment highlights dynamic biochemical homeostasis in perivascular interstitial space. Following cerebral ischemia, injured brain tissue can be spatially divided into the irreversibly damaged infarct core, metabolically-stressed ischemic penumbra and mildly-affected peripheral regions. These three zones lack rigid anatomical boundaries and evolve dynamically depending on reperfusion timing, presenting divergent features in blood perfusion, blood-brain barrier integrity, cellular viability and immune-inflammatory profiles. Four interrelated pathological cascades jointly disrupt microenvironmental homeostasis: calcium-overload-triggered cytotoxic cascades; blood-brain barrier breakdown resulting from glycocalyx shedding, tight-junction dissociation and matrix metalloproteinase-driven basement-membrane degradation; inflammatory-immune dysregulation dominated by microglial phenotypic reprogramming; and mutual crosstalk between coagulation disturbance and neuroinflammation that exacerbates secondary brain injury. Multiple molecular layers orchestrate these pathological alterations. Excessive activation of HIF-1α, NF-κB and MAPK/ERK signaling pathways promotes barrier leakage, pro-inflammatory factor secretion and extracellular-matrix remodeling. Epigenetic modulation mediated by non-coding RNAs including miRNAs and lncRNAs reshapes vascular microenvironment via three major patterns: modulating endothelial-pericyte barrier integrity, switching microglia-astrocyte inflammatory phenotypes, and driving metabolic reprogramming among multiple cell populations within microdomains. In addition, metabolic shift toward glycolysis among endothelial cells, microglia, astrocytes and compromised neurons establishes self-amplifying vicious cycles of energy depletion and oxidative stress in perivascular niches. Targeted theranostic approaches oriented toward vascular microenvironment bring new opportunities for precision medicine of ischemic stroke. Magnetic nanoprobes and near-infrared fluorescent probes respond to elevated reactive oxygen species, pH shift and inflammatory signals, enabling in situ visualization of pathological signatures of microenvironment and risk evaluation of hemorrhagic transformation. Function-tailored nanodelivery systems exert multi-modal protective effects: reactive oxygen species-responsive polymeric micelles deliver calcium chelators to block calcium-overload-initiated injury cascades; nanozymes scavenge reactive oxygen/nitrogen species and revert microglial pro-inflammatory polarization; biomimetic membrane-coated nanoparticles accumulate at micro-thrombosis sites to break thrombo-inflammatory loops and mitigate no-reflow phenomenon; ferroptosis-inhibitor-loaded nanocarriers eliminate excessive labile iron within perivascular toxic microdomains. During subacute and recovery phases, spatiotemporally controlled release of angiogenic factors by nanocarriers facilitates maturation of newly-formed microvessels rather than blind vascular sprouting, which helps reconstruct intact blood-brain barrier and restore microenvironmental homeostasis. Nevertheless, substantial translational bottlenecks still impede clinical transformation of these nanomedicines, including potential biosafety risks caused by in vivo material accumulation, protein-corona-induced impairment of targeting capacity, intrinsic blood brain barrier transport obstacles, lack of unified quality-control criteria for large-scale manufacture, reliance on high-end clinical imaging instruments, and prominent interspecies discrepancies between experimental animals and human patients. On the basis of summarizing spatiotemporal pathological characteristics, multi-level regulatory mechanisms and cutting-edge microenvironment-oriented diagnostic and interventional strategies, the present review advocates a conceptual shift from isolated neuronal protection to comprehensive reconstruction of perivascular vascular microenvironment homeostasis. Such perspective provides theoretical foundations for expanding therapeutic time windows and inspires future development of novel therapeutic modalities for ischemic stroke.

Objective Abnormal intestinal angiogenesis and dysregulation of the mucosal immune microenvironment are important mechanisms underlying the occurrence and progression of ulcerative colitis (UC). This study aimed to identify biomarkers associated with angiogenesis in UC through integrated bioinformatics analysis and to validate their expression patterns and effects on vascular endothelial cell function through in vivo and in vitro experiments.Methods Six UC-related datasets were retrieved from the GEO database. GSE48959, GSE59071, and GSE73661 were batch-corrected using the ComBat algorithm and used as the discovery cohort, while GSE47908, GSE92415, and GSE206285 were used as independent external validation cohorts. Differentially expressed genes were identified using the limma package, and a weighted gene co-expression network was constructed using WGCNA. The DEGs, WGCNA core module genes, and angiogenesis-related genes were intersected to obtain candidate genes. Multiple machine learning methods were further integrated to screen candidate features, followed by DNN and SHAP-based analysis of candidate gene feature contributions. PPI network topology analysis was further performed to identify hub genes, and the CIBERSORT algorithm was used to evaluate candidate genes and immune infiltration features. In addition, a DSS-induced mouse model of colitis and a TNF-α-stimulated HUVECs inflammatory model were established for experimental validation, combined with hematoxylin and eosin staining, immunofluorescence staining, Western blotting, wound healing assays, Transwell invasion assays, and Matrigel tube formation assays.Results A total of 481 upregulated and 285 downregulated genes were identified in the discovery cohort, and these genes were mainly enriched in pathways related to leukocyte migration and the complement and coagulation cascades. WGCNA identified the yellow module as the module most strongly associated with UC. The intersection of DEGs, WGCNA core module genes, and angiogenesis-related genes yielded 26 candidate genes. Four machine learning algorithms further identified C2, TIMP1, and TGM2 as common candidate features, while DNN-SHAP analysis and PPI network topology analysis further highlighted TIMP1 as a key hub gene. After multi-step screening, TIMP1 was ultimately identified as a key candidate gene. Validation across multiple datasets showed that TIMP1 was significantly upregulated in UC tissues and exhibited favorable discriminatory performance (all AUCs>0.85). In the DNN model, TIMP1 showed a relatively high feature contribution, supporting its potential relevance to UC classification. CIBERSORT analysis indicated that TIMP1 expression was positively correlated with neutrophils and resting NK cells, but negatively correlated with regulatory T cells and CD8? T cells. In vivo and in vitro experiments further confirmed that TIMP1 expression was markedly increased under inflammatory conditions. In DSS-induced colitis mice, increased CD31-positive vascular signals were accompanied by elevated TIMP1 expression and reduced VE-cadherin expression. Moreover, TIMP1 knockdown significantly inhibited endothelial cell migration, invasion, and tube formation, accompanied by decreased expression of VEGFA, VEGFR-2, and VEGFR-3.Conclusion TIMP1 may participate in UC-associated abnormal angiogenesis and changes in vascular endothelial cell function under inflammatory conditions, highlighting its potential as a biomarker and a candidate for further therapeutic investigation.

Cryo-electron tomography (cryo-ET) is a cornerstone technique in structural biology, enabling the three-dimensional visualization of macromolecular complexes within their native, near-physiological state. As structural biology transitions from isolated cells toward intact multicellular organisms, complex tissues, and clinical specimens, specimen thickness represents the primary physical barrier. Biological specimens exceeding a few hundred nanometers induce severe electron scattering, drastically degrading image contrast and high-resolution interpretability during transmission electron microscopy. Cryogenic focused ion beam (cryo-FIB) micromachining has established itself as the indispensable approach to overcome this constraint. By selectively ablating vitrified material above and below the target plane, cryo-FIB produces uniform, electron-transparent lamellae without ice crystallization, chemical dehydration, or mechanical sectioning artifacts. This review systematically synthesizes biological cryo-FIB development across three core dimensions: ion beam sources, correlative targeted localization, and automated processing workflows. In dual-beam systems, ion sources dictate the balance between bulk excavation throughput and delicate surface finishing. Biological cryo-FIB primarily employs two major classes of ion sources: liquid-metal ion sources (LMIS) and inductively coupled plasma sources. Traditional gallium-based LMIS operates via field evaporation, delivering high source brightness, small virtual source sizes, and ultra-fine probe diameters under low currents. These properties establish Ga? LMIS as the empirical benchmark for precision boundary sculpting, fine milling, and final low-damage polishing. Conversely, plasma FIB (pFIB) ionizes working gases—such as xenon or argon—yielding beam currents orders of magnitude higher than LMIS. This enables exceptionally rapid volumetric sputtering, making pFIB ideally suited for large-volume trenching, high-pressure frozen blocks, and serial volume imaging. Combined with preparation strategies like the Waffle method and cryogenic lift-out, these complementary sources extend biological cryo-FIB from monolayer cells to thick tissues and multicellular specimens. To achieve reliable milling of rare, transient, or deeply embedded targets, cryogenic correlative light and electron microscopy (cryo-CLEM) has developed rapidly. The methodology has evolved from standalone platforms—relying on external optical imaging, fiducial markers, and cross-instrument coordinate transformations—toward in-chamber integrated systems and coincident multi-beam configurations. These integrated platforms eliminate transfer-induced contamination risks and mechanical drift, achieve sub-micrometer registration accuracy, and permit real-time fluorescence monitoring to verify target retention during milling. Furthermore, multimodal optical interrogation—incorporating reflected light, thin-film interference, and quantitative fluorescence—facilitates non-destructive, pre-TEM assessments of lamella thickness and morphology. In parallel, software automation has progressed from simple parameter execution to high-throughput, multi-site unattended milling pipelines. Traditional manual fabrication is labor-intensive and constrained by operator subjectivity. To meet expanding data demands in in situ structural biology, modern automated workflows coordinate site revisiting, drift tracking, stress-relief trenching, and progressive multi-stage milling. By standardizing operational protocols and integrating automated image-registration feedback, these systems substantially reduce operator dependence while enhancing fabrication reproducibility and yield. Moreover, automation establishes a streamlined interface bridging upstream targeted coordinate transfer directly with downstream automated transmission electron microscopy data acquisition. Looking forward, biological cryo-FIB is advancing toward intelligent, closed-loop milling frameworks driven by machine-learning feedback and real-time image analysis, enabling autonomous adjustment of beam parameters and cutting geometries. Concurrently, exploring hybrid multi-species ion regimens will combine the rapid bulk ablation of plasma sources with the gentle polishing capabilities of fine ion probes, while standardized, automated lift-out workflows will unlock routine preparation of bulk biological tissues. By uniting hardware innovations, correlative targeting, and intelligent automation, biological cryo-FIB will further accelerate the transition of in situ structural biology from localized cellular snapshots toward the routine, systematic exploration of intact physiological tissue microenvironments.

Inflammatory bowel disease (IBD) is a chronic disorder characterized by recurrent mucosal inflammation. Although biologics and small-molecule therapies can effectively suppress inflammation, only 13%-34% of patients achieve deep histological remission, and most experience relapse within 2-3 years. Existing biomarkers, including C-reactive protein (CRP), fecal calprotectin, and endoscopic measures, have limited ability to detect subclinical molecular abnormalities, suggesting that disease-driving factors may persist even during clinical remission. Recently, Pang et al. reported a study in Science that investigated the role of epithelial cell-death signaling using a prospective adult IBD cohort comprising 80 participants and more than 900 biopsy specimens. By integrating bulk and spatial transcriptomics, patient-derived organoids, and 24-36 months of longitudinal follow-up, the authors characterized epithelial cell-death programs associated with IBD persistence. They found that the necroptosis-associated marker pRIPK3 could be detected in the absence of overt histological inflammation and persisted during advanced therapy. IFN-γ and TNF reshaped the transcriptional landscape of specific colonic epithelial cells, inducing an inflammatory macrophage-like program termed cluster B and triggering RIPK1-independent RIPK3 signaling. This was followed by NOS2/iNOS-mediated mitochondrial apoptosis driving colonocyte death, whereas PUMA-dependent apoptosis contributed to intestinal stem-cell loss. Higher baseline pRIPK3 levels were associated with subsequent disease activity, although their independent predictive value remains to be established. These findings suggest that persistent IBD may reflect not only residual immune inflammation but also inflammation-induced pathological epithelial states. Future therapeutic strategies may therefore need to extend beyond inflammation control toward “molecular repair”, including targeting noncanonical RIPK3 signaling, modulating iNOS/PUMA-dependent mitochondrial apoptosis, and developing relapse-risk stratification strategies based on tissue biomarkers such as pRIPK3. Such approaches may help shift the therapeutic goal from “clinical remission” toward “molecular healing”.

Objective Methylthioadenosine phosphorylase (MTAP) is found to be frequently deleted in malignant cancers, but the role of MTAP underlying metastasis and stemness of liver cancer remains far from clear. 5-fluorouracil (5-Fu) is clinically used in treating advanced liver cancer, however, chemoresistance is one of the major barriers impeding its efficacy. Liver cancer stem cells are considered to be the reservoir for tumor progression and chemoresistance. This study aims to explore the role of MTAP in regulating stemness and chemoresistance of 5-Fu in liver cancer. It is designed to elucidate the molecular mechanisms underlying MTAP-manipulated stemness. How MTAP influences metastasis and chemoresistance against 5-Fu in liver cancer is also elaborated.Methods The correlation of MTAP expression with stemness marker CD133 was evaluated by flow cytometry. Colony formation assays were conducted to analyze the tumorigenic ability. Western blot assays were applied to delineate the molecular mechanisms underlying MTAP-regulated stemness. Trans-well assays were conducted to analyze the metastatic potential of liver cancer cells. In vivo experiments were performed using subcutaneous xenograft and lung metastasis model on nude mice. Tumor size and weight were analyzed upon 5-Fu treatment. Metastatic nodules on lungs were evaluated. Immunohistochemical and TUNNEL assays were performed on slides obtained from tumor tissues.Results MTAP overexpression inhibited expression of CD133 and colony formation, which is further ablated by MTAP inhibitor. MTAP deprivation augmented stemness and colony formation, which is reversed by protein arginine methyltransferase 5 (PRMT5) overexpression but not its catalytically inactive counterparts. Critically, SRY-box transcription factor 2 (SOX2) was found to be a downstream regulator of MTAP/PRMT5, leading to stemness of liver cancer cells. MTAP overexpression suppressed SOX2 expression, while MTAP inhibition augmented expression of SOX2. Critically, MTAP overexpression inhibited migration and invasion of liver cancer cells. In vivo experiments in animal model found that MTAP deprivation promotes tumor growth and metastatic nodules of liver cancer. Additionally, MTAP overexpression inhibits tumor growth. Moreover, MTAP deprivation contributes to chemoresistance shown by increased tumor size and weight treated by 5-Fu. In contrast, MTAP overexpression led to further decline in tumor size and weight elicited by 5-Fu treatment. Accordingly, SOX2 was inhibited in tumor tissues after MTAP overexpression. Marker of proliferation Ki67 (Ki67) expression was dampened while apoptosis was found in MTAP-overexpressed tumor tissues.Conclusion In this study, plethora of experimental studies collectively demonstrated that MTAP deprivation enhances stemness of liver cancer by regulating SOX2, exerting anti-cancer and chemoresistant role against 5-Fu in liver cancer. It is also worth noting that MTAP inhibition promoted metastasis of liver cancer. MTAP overexpression augmented tumor shrinking elicited by 5-Fu and SOX2 was involved in this process. These results concluded that MTAP/PRMT5/SOX2 signaling axis was involved in regulating stemness and chemoresistance of liver cancer. These findings help to exploit novel therapies by targeting stemness in MTAP-depleted liver cancer. It shed novel light on unveiling MTAP-regulated stemness in antagonizing 5-Fu treatment in liver cancer. It also adds mounting evidence to the theoretical basis for accomplishing personalized chemotherapy of liver cancer patients by targeting MTAP or its downstream PRMT5/SOX2.

Objective Cryo-focused ion beam (cryo-FIB) milling combined with cryo-electron tomography (cryo-ET) has become the leading approach for in situ structural studies of cells and tissues. A persistent bottleneck, however, hampers its workflow efficiency and throughput: the fluorescence signal that is used to identify the target structure during milling is rarely carried forward into the electron-microscopy (EM) data-collection stage. After a lamella is transferred to the transmission electron microscope (TEM), relocating the exact sub-cellular feature previously highlighted by fluorescence still depends on manual, experience-based procedures—typically importing the fluorescence and low-magnification EM maps into generic image editors, aligning them by eye, and estimating the target position before picking it in the acquisition software. This fragmentation is inefficient and error-prone. Meanwhile, existing correlative light and electron microscopy (CLEM) instruments target fluorescence only during milling, and mainstream TEM acquisition programs (SerialEM, Thermo Fisher Tomography 5, Leginon, JEOL JADAS-Tomo) provide no interface to import or overlay a fluorescence layer. The objective of this work was therefore to develop a lightweight, user-friendly tool that bridges fluorescence targeting and TEM tilt-series acquisition, so that fluorescently labelled structures can be located and collected conveniently.Methods We developed LamFluoView, a GUI-based CLEM site-selection program designed specifically for the TEM data-collection environment. LamFluoView reads the SerialEM low-magnification montage directly (without exporting it), loads the corresponding cryo-fluorescence image, and registers the two modalities by feature matching of the lamella contour, supplemented by interactive rotation, scaling, and transparency adjustment. Once aligned and locked, the user marks regions of interest (ROIs) directly on the high-contrast fluorescence image; the program converts each ROI in real time into the corresponding specimen-stage coordinates of the electron microscope and pushes them to the SerialEM Navigator list, enabling a seamless transition from targeting to automated, high-resolution tilt-series acquisition. Registration parameters are retained across a batch, so only translation is needed to register sibling lamellae. Registration accuracy was quantified using 100-nm fluorescent beads (n=115): red-channel spots were localized by Gaussian surface fitting and compared with their EM positions. Biological validation was performed on MitoTracker-labelled mitochondria and Pericentrin-mCherry-labelled centrosomes.Results Gaussian fitting of the fluorescent beads gave a mean localization deviation of ±56.19 nm in X and ±58.93 nm in Y (n=115), well within the requirement for targeting micrometre- to hundred-nanometre-scale sub-cellular structures on milled lamellae. In the biological tests, LamFluoView correctly co-localized MitoTracker mitochondria and Pericentrin-mCherry centrosomes with their corresponding EM morphology. A centrosome selected via the fluorescence channel produced a tomographic reconstruction with clearly resolved centriolar microtubule architecture, confirming that the tool facilitates the identification of low-abundance subcellular structures within the lamellae.Conclusion LamFluoView extends CLEM from the milling stage into the data-collection stage and resolves a long-standing cryo-ET difficulty—that fluorescence information is not readily usable for selecting data acquisition sites—by turning fluorescence-guided selection into a one-click, coordinate-accurate operation. Coupled with high-throughput cryo-PFIB milling, it is particularly suited to rapidly screening and structurally resolving rare, low-abundance structures such as centrioles, cell-junction complexes, and viral entry sites. LamFluoView is open-source and currently interfaces with SerialEM; a floating, transparency-toggle mode also permits partial fluorescence navigation over other acquisition software. Together, these advances help drive in situ cryo-EM toward precise, high-throughput data acquisition.

Objective Cryo-electron tomography (cryo-ET) enables three-dimensional visualization of macromolecular complexes, organelles, and cellular structures in near-native states and has therefore become an important approach for in situ structural biology. However, extending cryo-ET from isolated cells and relatively small specimens to large biological tissues remains technically challenging. The major obstacles include insufficient vitrification of water within thick specimens, the difficulty of identifying and extracting specific regions of interest from large frozen samples, the low efficiency and poor controllability of cryo-lamellae preparation, and the mechanical instability of lamellae during transfer and cryo-transmission electron microscopy (cryo-TEM) imaging. These limitations substantially restrict the application of cryo-ET to large tissue samples. In this study, we developed a high-efficiency workflow, termed HULF (high-pressure freezing, ultramicrotome cryo-trimming, cryo-Lift-Out, and cryo-plasma focused ion beam milling), for preparing cryo-lamellae from large biological tissues and facilitating subsequent in situ cryo-ET analysis.Methods Spider and silkworm silk glands were selected as representative large biological tissue models. Two high-pressure freezing strategies were initially evaluated. In the first strategy, tissue samples were loaded into customerized high-pressure freezing carriers and vitrified using 20% dextran solution as the filling and cryoprotective medium with a Wohlwend HPF Compact 03 high-pressure freezer. The vitrified samples were subsequently subjected to ultramicrotome cryo-trimming using a Leica FC7 cryo-ultramicrotome to remove excess tissue and carrier material and to generate sample strips with a thickness of approximately 30 μm and a depth of approximately 50 μm. In the second strategy, a grid-based high-pressure freezing system using 2-dimethylpentane as the filling medium was evaluated in combination with cryo-Lift-Out, which facilitated direct visualization and localization of the region of interest. For both strategies, cryo-plasma focused ion beam (cryo-pFIB) milling was subsequently used for tissue segmentation and progressive thinning. Based on the advantages and limitations identified from these two strategies, we established the HULF workflow. In this workflow, large tissue specimens were first vitrified using the dextran-based carrier high-pressure freezing strategy, followed by ultramicrotome cryo-trimming to reduce the specimen volume and expose a well-defined tissue region. Cryo-Lift-Out was then employed to selectively extract the region of interest and transfer the block onto a receiver grid, thereby avoiding repeated manipulation of the original carrier. The extracted tissue block was segmented and progressively thinned by cryo-pFIB milling using a sequential rough-milling, fine-milling, and polishing strategy, ultimately producing cryo-lamellae with a target thickness of approximately 150 nm. In addition, we optimized the receiver-grid configuration by replacing conventional bare copper grids with copper grids coated with a gold support film to improve mechanical support and electrical conductivity and to reduce cryo-lamellae damage during sample transfer and cryo-TEM examination. The quality and structural integrity of the resulting lamellae were assessed by cryo-TEM, and their suitability for structural analysis was further validated by cryo-ET tilt-series acquisition and three-dimensional tomographic reconstruction.Results The comparison of the two preliminary strategies revealed distinct advantages and limitations. The carrier-based high-pressure freezing strategy using 20% dextran solution provided stable vitrification of the silk gland tissue and was therefore suitable for preserving the ultrastructural integrity of relatively large tissue specimens. However, direct cryo-pFIB milling of the trimmed sample strips presented difficulties in specimen transfer and manipulation, which limited the efficiency and reproducibility of cryo-lamellae preparation. In contrast, the grid-based high-pressure freezing strategy using 2-dimethylpentane facilitated exposure and identification of the tissue surface and therefore simplified localization of the target region. Nevertheless, its vitrification performance was less stable in large biological specimens, and crystalline ice formation was observed in some preparations, compromising subsequent cryo-ET imaging.These observations led to the central design principle of the HULF workflow: rather than relying on a single preparation strategy to simultaneously achieve optimal vitrification, target localization, specimen extraction, and cryo-lamellae fabrication, each technical step should be assigned to the approach that provides the greatest advantage for that specific stage of the workflow. Accordingly, HULF combines the stable vitrification provided by dextran-based high-pressure freezing with the efficient target extraction enabled by cryo-Lift-Out and the precise thinning capability of cryo-pFIB milling. Ultramicrotome cryo-trimming further reduces the volume of the large frozen specimen and generates a geometrically defined sample strip, which substantially facilitates subsequent cryo-Lift-Out. The extracted tissue blocks could then be transferred to receiver grids and progressively milled to approximately 150 nm, producing lamellae suitable for cryo-ET data acquisition. A further limitation identified during optimization was the mechanical fragility of lamellae supported by bare copper grids. Large cracks and cryo-lamellae fractures could occur during grid manipulation and transfer to the cryo-TEM, potentially resulting from mechanical forces and vibration applied to the unsupported grid structure. Replacing bare copper grids with copper grids coated with a gold support film provided additional mechanical support for the lamellae and improved the stability of the sample during transfer and cryo-TEM examination. Lamellae prepared using the optimized configuration exhibited good structural integrity and relatively low ice contamination. Importantly, lamellae remained intact after prolonged storage in liquid nitrogen and subsequent reloading for cryo-TEM examination, indicating improved stability during storage and transfer. Finally, HULF-prepared lamellae successfully supported cryo-ET tilt-series acquisition, and three-dimensional tomographic reconstruction clearly revealed ordered fibrous structures within spider silk gland tissue, demonstrating the applicability of the workflow to in situ structural investigation of large biological tissues.Conclusion We established HULF, a cryo-lamella preparation workflow for large biological tissue samples. By integrating high-pressure freezing, ultramicrotome cryo-trimming, cryo-Lift-Out, and cryo-pFIB milling, HULF modularly combines the key steps of sample vitrification, target-region extraction, and efficient thinning, thereby improving the efficiency, stability, and reproducibility of cryo-lamellae preparation from large tissue samples. This workflow provides a practical technical strategy and methodological framework for in situ structural studies of large biological tissues using cryo-ET.

Chemical cross-linking coupled with mass spectrometry (CXMS/XL-MS) has emerged as a core technology for three-dimensional protein structure elucidation and protein-protein interaction characterization. The molecular design of chemical cross-linkers is a key determinant of the detection capability and application scope of this technology. In this review, we systematically summarize recent advances in chemical cross-linker development from three interconnected perspectives: expansion of amino acid residue coverage, adaptation to complex biological systems, and advancement toward in vivo cross-linking in living cells. In terms of residue targeting, the repertoire has expanded from traditional lysine-reactive chemistries (e.g., NHS esters, imidoesters, and ortho-phthalaldehyde) to carboxyl, guanidino, sulfhydryl, phenolic hydroxyl, and imidazole groups, thereby enabling targeting of Asp/Glu, Arg, Cys, Tyr, and His. Representative chemistries include carbodiimides, sulfonyl ynamides, hydrazides, diazo groups, glyoxals, maleimides, haloacetamides, vinyl sulfones, disulfides, urazoles, methylated urazoles and sulfonyl fluorides. Combined with semi-specific photo-cross-linkers that rely on reactive groups such as aryl azides, diazirines, and benzophenones, as well as multi-cross-linker combination strategies, these advances have significantly improved cross-linking coverage depth and spatial restraint density. To adapt to complex biological samples, the introduction of enrichable, cleavable, and multifunctional integrated cross-linkers has effectively alleviated the problems of low abundance and difficult identification of cross-linked peptides in complex mixtures. Enrichment strategies include biotin-avidin, phosphonate-IMAC, and dimethylpiperidine-antibody systems, as exemplified by reagents such as Leiker, PhoX, DPST, and various click-chemistry-based probes. Cleavable modules include chemical (disulfide, azo) and MS-cleavable (sulfoxide, Asp-Pro peptide bond, urea-linked butyrate moiety, glycosidic bond) motifs, while integrated designs combine enrichment, cleavage, and membrane permeability. We also summarize representative proteome-level CXMS datasets reported over the past three years, highlighting how cross-linker chemistry, sample preparation, mass spectrometry platforms, and search engines jointly determine identification depth. In terms of application scenarios, three major routes—membrane-permeable cross-linkers, nanoparticle-mediated targeted delivery, and site-specific in vivo cross-linking via unnatural amino acids (UAAs)—have collectively enabled the transition from in vitro cross-linking to in situ analysis in living cells and subcellular organelles. Site-specific incorporation of UAAs includes photo-cross-linking systems (e.g., SiTomics, Photo-Ile, Photo-Leu, Photo-Met) and proximity-enabled chemical cross-linking strategies (e.g., SuFEx, PFEx). These approaches capture native conformations, transient and weak interactions, and compartment-specific interactions that are difficult to observe in vitro, thereby revealing the unique biological value of in vivo CXMS. We also discuss challenges in delivery, reaction, and identification, as well as the complementarity of CXMS with proximity labeling, cryo-EM, and AlphaFold-based structural modeling. Future advances in in vivo cross-linking will require progress along three dimensions: capturing dynamic interactions with temporal resolution, extending to tissues and living animals, and targeting membraneless organelles. Integrating these data with single-cell spatial omics, live-cell imaging, and artificial intelligence-based prediction will advance protein-protein interaction research toward dynamic, in situ structural systems biology. Finally, we discuss future directions and highlight that, driven by emerging demands for probing hetero-biomolecular interactions (e.g., protein-nucleic acid, protein-glycan, and protein-lipid interactions), higher-order topological structures of multiprotein assemblies, and artificial intelligence-assisted structural modeling, cross-linker design is shifting from a concept-driven to a data- and application-driven paradigm. This transition is expected to further propel CXMS toward higher precision and broader applicability in the analysis of increasingly complex biological systems.

Bacterial traceability is of great significance for investigating public safety incidents and curbing the spread of diseases. It can provide important clues for identifying contamination sources, clarifying transmission routes, and supporting biosafety and public health responses. Traditional bacterial traceability techniques, such as molecular biology and spectroscopy methods, are able to achieve bacterial species identification and strain typing. However, they still face obvious limitations when required to accurately determine the geographical origin of bacteria. In many cases, they cannot effectively answer key questions such as where a target bacterium comes from and under what conditions it was cultivated. Stable isotope analysis technology, as an important tool in modern science, has unique advantages including high stability, good repeatability, and strong geographical tracing capability. These advantages enable it to effectively overcome the limitations of traditional methods and demonstrate distinctive application value in the field of bacterial geographical traceability. This paper systematically reviews the basic theory of stable isotope analysis technology and its two core tracing pathways, namely tracing based on natural isotope fingerprints and tracing based on stable isotope probing. Focusing on three core application dimensions, namely revealing bacterial metabolic pathways, constructing stable isotope fingerprint maps, and inferring growth environments, it summarizes the latest research progress of this technology in bacterial traceability. It further explains its application value in inferring culture water sources and geographical origins, identifying culture medium types, analyzing extreme environments, and identifying the sources of functional microbial communities. Through these applications, stable isotope analysis can link bacterial metabolic characteristics, cultivation conditions, and environmental backgrounds with isotopic information, thereby providing new technical support for bacterial source tracing and related biosafety work. At the same time, this paper summarizes the key bottlenecks currently faced by the application of this technology. First, stable isotope signals are susceptible to interference from environmental factors and experimental procedures, which may affect the accuracy and reliability of traceability results. Second, the identification efficiency for specific microbial groups remains low, especially in complex or extreme environments. Third, a standardized stable isotope fingerprint database system is still incomplete, which limits data comparison, sharing, and practical application across laboratories and regions. These problems restrict the broader promotion and precise application of stable isotope analysis in bacterial traceability and indicate that further methodological and systematic efforts are still needed. Finally, the paper discusses future optimization directions from the perspectives of combined detection of multiple elements, integration with high-throughput sequencing, automated and precise detection, and multidisciplinary innovation. It is expected that these developments will improve the analytical efficiency, resolution, and standardization of stable isotope analysis technology, expand its application boundaries in bacterial traceability, and provide stronger support for research and practice in related fields. Such progress will help stable isotope analysis play a greater role in public safety, biosafety, and infectious disease prevention and control.

Diabetic foot ulcers (DFUs) are a major complication of diabetes, associated with persistent infection, frequent recurrence, amputation, and increased mortality. Despite standard care, including debridement, infection control, offloading, and restoration of tissue perfusion, many wounds fail to progress beyond the inflammatory phase. This persistent impairment suggests that local pathology alone may not fully explain delayed healing. Gut microbiota dysbiosis has emerged as a potential systemic contributor through interactions among intestinal barrier integrity, microbial metabolites, immune responses, vascular function, and the wound microenvironment.This review integrates clinical observations and experimental findings to examine how dysbiosis may contribute to impaired DFU healing through the gut–skin (wound) axis. It organizes the available evidence into an interconnected framework linking barrier disruption and metabolic disturbances to persistent inflammation,immune imbalance, endothelial dysfunction, tissue hypoperfusion, and wound biofilm formation, while distinguishing direct DFU evidence from findings extrapolated from diabetes and other disease models. Disruption of the intestinal barrier may facilitate the translocation of lipopolysaccharide and other microbial products into the circulation, sustaining low-grade inflammation through Toll-like receptor 4/nuclear factor-κB signaling and NLRP3 inflammasome activation. Concurrently, reduced short-chain fatty acid availability and altered bile acid and tryptophan-derived metabolite profiles may weaken epithelial protection and disturb immune homeostasis. These alterations involve receptor-mediated signaling and epigenetic regulation, including free fatty acid receptors, histone deacetylases, bile acid receptors, and the aryl hydrocarbon receptor. Their combined effects may impair regulatory T-cell responses, delay macrophage transition toward reparative phenotypes, and disrupt neutrophil function, thereby preventing timely inflammatory resolution. Systemic inflammation and metabolic dysfunction may further reinforce endothelial injury, compromise microcirculation, and inhibit effective angiogenesis. The resulting hypoxic, inflammatory wound environment favors microbial persistence and may support biofilm development. Within polymicrobial biofilms, quorum sensing and extracellular polymeric matrices enhance antimicrobial tolerance and immune evasion, sustaining a feedback loop of persistent colonization, inflammation, and defective repair. However, direct transfer of specific intestinal microorganisms or virulence factors to DFU wounds through the circulation has not been established; the proposed gut–biofilm connection currently rests primarily on indirect inflammatory, immunological, and metabolic mechanisms. Potential adjunctive strategies include dietary modification, probiotics, prebiotics, synbiotics, postbiotics, and targeted antibiofilm approaches. These interventions should be evaluated alongside, rather than as replacements for, established DFU care. Their clinical translation remains constrained by heterogeneous formulations, limited DFU-specific trials, uncertain safety profiles in vulnerable patients, and insufficient evidence for meaningful healing outcomes. Candidate biomarkers reflecting barrier dysfunction, microbial metabolism, and systemic inflammation may support future stratification, but validated thresholds and associations with clinical outcomes remain lacking. Overall, the gut microbiota provides a plausible mechanistic perspective on DFU chronicity, but current evidence does not establish the direction or causality of the entire proposed pathway. Future studies should combine longitudinal microbiome and metabolomic profiling with standardized wound assessment, rigorous control of antibiotic exposure and other confounders, and mechanistic validation. Randomized trials embedded in standard treatment pathways are needed to determine whether microbiota-targeted adjuncts improve healing, reduce recurrence and amputation, and offer acceptable safety and cost-effectiveness. This framework supports testable hypotheses and stratified intervention design without overstating current clinical readiness.

Achieving a systematic understanding of embryonic development is a central goal of developmental biology. Digital embryo approaches provide an emerging framework for integrating multimodal developmental data toward digital reconstruction and predictive modeling of embryonic development. In this review, a bibliometric analysis was conducted to characterize research trends and emerging topics in digital embryo research and related fields from 2010 to March 2026. A total of 9 770 articles were retrieved from the Web of Science, of which 8 622 articles were retained after topic-based filtering. Keyword analysis identified scRNA-seq as the predominant technology, with spatial transcriptomics and multi-omics technologies emerging as major research themes. Early embryonic development was a major biological focus, while blastoids, gastruloids, and retinal organoids were prominent in vitro models. Based on bibliometric results, we synthesize recent advances across five essential pillars: single-cell multi-omics atlases, spatial omics landscapes, in vitro developmental models, bioinformatics analysis, and artificial intelligence virtual cells (AIVCs), highlighting their distinct contributions to digital embryo research. Single-cell omics technologies characterize embryonic states at molecular resolution and reveal both coordinated and decoupled dynamics across omics layers. Spatial omics preserve cellular location and tissue context, extending developmental atlases from molecular-state annotation toward spatiotemporal reconstruction. In vitro models provide experimentally tractable platforms for direct observation, controlled perturbation, and mechanistic validation, while supplying complementary data for digital embryo modeling. Multi-omics integration, spatial mapping and alignment, and three-dimensional reconstruction constitute key computational tasks in digital embryo construction. AIVCs learn representations of cell states and developmental dynamics, thereby supporting the prediction of perturbation responses and cell fates. We group current methods for in silico perturbation prediction into three recurring strategies: latent-state representation learning, biological network-informed modeling, and distributional transport, involving frameworks such as (variational) autoencoders, graph neural networks, optimal transport, and flow matching. Recent AIVC approaches are increasingly being applied to developmental biology, demonstrating the potential to model developmental trajectories, reconstruct spatiotemporal dynamics, and predict perturbation-induced changes. Overall, digital embryo research is transitioning from spatiotemporal multi-omics atlas construction toward predictive virtual-cell modeling. At the data level, major challenges include incomplete spatiotemporal coverage of developmental atlases, limited multimodal datasets, and a scarcity of perturbation data. AIVCs remain constrained by limited interpretability and out-of-distribution generalization, as well as a lack of standardized benchmarking and systematic experimental validation. We propose AIVC modeling of preimplantation embryos as a plausible near-term objective, with the potential to advance developmental biology and reproductive medicine.

Phased array technology operates through the high-precision electronic regulation of excitation phase and amplitude across multi-channel array elements to achieve inertia-free beam steering, spatial focusing, and dynamic scanning. Originally developed in radar and wireless telecommunications, the technology has progressively expanded into medical diagnostics and therapeutics, emerging as a foundational pillar of precision medicine. To prevent conceptual generalization and establish rigorous physical boundaries, this review explicitly delineates three technical tiers: strictly defined phased arrays governed by coherent wave interference; multi-coil electromagnetic systems guided by spatial field modulation concepts; and generalized array-based architectures employing multi-channel spatial encoding or high-speed temporal scanning. In diagnostic imaging, strictly defined acoustic and microwave phased arrays have enhanced spatial and temporal resolution, accelerating the advancement of flexible wearable ultrasound patches, microelectromechanical piezoelectric arrays, and non-contact vital sign radar sensors. Concurrently, generalized array frameworks have provided critical architectural support for mainstream clinical modalities: multi-channel magnetic resonance imaging (MRI) receive coil arrays exploit localized spatial sensitivity profiles for parallel acquisition acceleration and signal-to-noise ratio enhancement, rather than active beam synthesis; meanwhile, distributed cold-cathode X-ray source arrays in static computed tomography (CT) replace mechanical gantry rotation with microsecond-to-millisecond temporal electronic switching, substantially mitigating motion artifacts while fundamentally differing from wave-interference phased arrays. In precision therapeutics, phased array systems leverage precise energy deposition capabilities to achieve non-invasive or minimally invasive intervention in deep-seated lesions. Focused ultrasound (FUS) and microwave hyperthermia systems combine physical matching media, medical-image-based tissue modeling, and multi-channel adaptive phase compensation to overcome severe beam refraction and wavefront aberrations across heterogeneous tissues such as skull and breast. Furthermore, multi-coil electromagnetic systems apply phased-array spatial field shaping principles to quasi-static or low-to-medium frequency fields; by optimizing the pulse repetition frequency of rapid transient magnetic fields in transcranial magnetic stimulation (TMS) and managing high-frequency currents in magnetic hyperthermia (MHT), these arrays achieve localized field enhancement and preserve scalp safety via spatial vector superposition. Translating these advanced array systems from laboratory validation into clinical practice encounters significant physical and engineering hurdles. At the physical and computational level, high-density array integration intensifies inter-element mutual coupling, while phase aberration corrections across complex tissues substantially increase the computational burden, data throughput, and optimization dimensionality. At the engineering and regulatory level, massive high-power transmit modules introduce critical thermal management bottlenecks, while industrial manufacturing repeatability across thousands of channels, long-term biocompatibility of flexible materials, and the interpretability of artificial intelligence algorithms remain essential hurdles for medical device approval. Looking forward, the biomedical phased array domain is evolving toward portable wearable integration, intelligent adaptive beamforming, and multimodal closed-loop theranostics. Deep learning algorithms demonstrate strong potential to mitigate optimization overhead and execute real-time aberration corrections. Concurrently, edge computing helps reconcile channel density with thermal dissipation constraints, and the synergistic integration of real-time imaging feedback with adaptive energy delivery will redefine personalized non-invasive precision intervention.

Metabolism and immunity are deeply interconnected, with increasing evidence demonstrating that cellular metabolic states shape immune cell functions and inflammatory responses. Beyond their classical roles as sources of energy and biosynthetic substrates, several metabolic intermediates have emerged as functional regulators of immune signaling pathways. However, whether metabolic intermediates can directly modify immune signaling proteins and regulate immune responses through such mechanisms remains poorly understood. A recent study by Zuo et al. revealed that pyruvate, a terminal product of glycolysis, possesses an unrecognized protein-modifying function that regulates type I interferon (IFN-I) signaling. The study demonstrated that glycolysis-derived pyruvate induces STAT1 pyruvylation at lysine 201 (Lys201), disrupting STAT1-STAT2 complex formation and attenuating IFN-I-mediated antiviral responses. Furthermore, genetic blockade of STAT1 pyruvylation through site-directed mutation enhanced antiviral immunity in vivo, establishing a direct mechanistic link between metabolic pathways and innate immune signaling regulation. This discovery expands the current understanding of immunometabolism by demonstrating that metabolic intermediates can function as chemical messengers to directly reshape immune signaling networks. Subsequent studies have further demonstrated that pyruvylation is a widespread protein modification dynamically regulated by the glycolytic state, with its writing and erasing mechanisms as well as its role in transcriptional regulation beginning to be elucidated. Further investigation into the selective regulatory mechanisms, substrate landscape, and pathophysiological functions of pyruvylation will help clarify how metabolic signals are encoded and decoded, and deepen our understanding of how metabolic states influence immune homeostasis and disease progression.

Objective Treadmill exercise intervention was evaluated for its effects on addictive behavior, learning and memory, and hippocampal neuroinflammation in methamphetamine (MA)-addicted mice, and transcriptomic screening was performed to identify candidate molecules involved in exercise-mediated attenuation of MA-related neuronal injury.Methods Two-month-old male C57BL/6 mice were randomly assigned to four groups: control (C), exercise control (CE), methamphetamine (M), and methamphetamine plus exercise (ME). Mice in the M and ME groups received intraperitoneal injections of MA (1 mg/kg) once daily for 7 consecutive days to establish the addiction model, whereas mice in the C and CE groups received equal volumes of saline. Subsequently, mice in the CE and ME groups underwent treadmill exercise at 12 m/min for 60 min/d for 2 weeks. Conditioned place preference (CPP), the Y-maze, and the Morris water maze (MWM) were used to assess addictive behavior, spatial recognition memory, and spatial learning and memory, respectively. Hippocampal neuronal ultrastructure was examined by transmission electron microscopy (TEM), and microglia-related responses were evaluated by immunofluorescence staining for ionized calcium-binding adapter molecule 1 (Iba-1) in the hippocampal CA1 region. Previously generated and publicly available RNA-seq data from brain tissue of the C, M, and ME groups were reanalyzed, with three biological samples per group. Differentially expressed genes (DEGs) were re-screened using the criteria of P<0.05 and fold change>1.5 or <0.67, followed by gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. NLR family CARD domain containing 5 (Nlrc5) was further validated by reverse transcription-quantitative polymerase chain reaction (RT-qPCR), and Western blot was used to determine the protein expression of NLRC5, nuclear factor kappa B (NF-κB), interleukin-1β (IL-1β), postsynaptic density-95 (PSD-95), and synaptophysin.Results Treadmill exercise significantly attenuated MA-induced conditioned place preference and improved learning and memory performance in both the Y-maze and MWM tests. Exercise increased exploration time and distance in the novel arm of the Y-maze, shortened escape latency, and increased the number of platform crossings in the MWM. Average swimming speed during the probe trial did not differ significantly among groups, suggesting that the MWM differences were not attributable to obvious differences in swimming performance. TEM revealed that MA exposure induced mitochondrial swelling, disrupted cristae, reduced organelle abundance, and nuclear envelope abnormalities in hippocampal neurons, whereas these ultrastructural changes were alleviated after exercise. Immunofluorescence staining showed that MA increased Iba-1 fluorescence intensity in the hippocampal CA1 region, while treadmill exercise significantly reduced this change. Transcriptomic analysis identified 160 DEGs associated with MA exposure and 58 DEGs associated with exercise intervention, with 9 overlapping DEGs. KEGG enrichment analysis showed that MA-related DEGs were enriched in the NOD-like receptor signaling pathway. Among the overlapping DEGs, Nlrc5, which was downregulated after MA exposure and restored following exercise intervention, was selected as a candidate gene for further validation. RT-qPCR confirmed that treadmill exercise reversed the MA-induced decrease in Nlrc5 mRNA expression. Western blot analysis further showed that treadmill exercise increased NLRC5, PSD-95, and synaptophysin protein expression while reducing NF-κB and IL-1β expression.Conclusion Two weeks of treadmill exercise attenuated MA-induced conditioned place preference and ameliorated learning and memory impairment, alleviated hippocampal neuroinflammation and neuronal ultrastructural damage, and restored the expression of synapse-related proteins. Transcriptomic screening combined with molecular validation demonstrated that exercise intervention was accompanied by restoration of NLRC5 expression, reduced NF-κB and IL-1β expression, and increased PSD-95 and synaptophysin expression. These findings suggest that NLRC5 may represent a candidate molecule involved in exercise-mediated amelioration of MA-related neuronal injury. However, the present results mainly demonstrate coordinated changes in NLRC5 expression, inflammatory indicators, and cognitive outcomes rather than a direct causal relationship, and the specific role of NLRC5 requires further functional validation.

Lipotoxicity is a common pathological process underlying obesity, insulin resistance, type 2 diabetes, metabolic dysfunction-associated steatotic liver disease (MASLD), and cardiometabolic abnormalities. Rather than simply reflecting an increase in total lipid content, lipotoxicity arises from a mismatch among lipid uptake, storage, mobilization, oxidation, and organelle quality control, which shifts fatty acids away from safe storage and efficient utilization toward the sustained accumulation of toxic lipid intermediates. Current evidence therefore suggests that the protective effects of exercise against lipotoxicity should be understood as a systematic remodeling of lipid fate rather than merely as a reduction in fat deposition. Exercise can promote the channeling of fatty acids into relatively safe triacylglycerol stores and dynamic lipid droplet pools, while improving lipid droplet size, subcellular localization, perilipin (PLIN) protein decoration, and lipid droplet turnover. At the same time, exercise enhances lipid droplet–mitochondria functional coupling, mitochondrial fatty acid oxidation, respiratory capacity, and metabolic flexibility, thereby increasing the probability that mobilized fatty acids are efficiently oxidized instead of being diverted toward diacylglycerols, ceramides, acylcarnitines, and oxidized lipids. Exercise also modulates endoplasmic reticulum stress, oxidative stress, inflammatory signaling, autophagy/lipophagy, mitophagy, and inter-organ communication, linking lipid handling to broader organelle homeostasis and tissue protection. These adaptations are tissue specific. In skeletal muscle, exercise-induced changes in intramyocellular lipid content should be interpreted together with lipid droplet quality, subcellular distribution, PLIN-associated remodeling, and mitochondrial oxidative capacity, which helps explain the athlete""s paradox. In the liver, exercise affects lipid input, lipid droplet clearance, lipophagy, mitochondrial oxidation, and stress-response pathways, providing a mechanistic basis for its beneficial effects in MASLD and related metabolic disorders. In the heart, the protective effects of exercise depend more strongly on the balance between fatty acid uptake, lipid mobilization, and mitochondrial oxidative capacity, and excessive lipid redistribution under certain metabolic conditions may increase cardiac lipotoxic stress. Adipose tissue further contributes by regulating free fatty acid overflow and systemic lipid supply, while muscle-, liver-, adipose-, vascular-, and gut-derived signals participate in cross-organ metabolic coordination. Importantly, the evidence base is heterogeneous: human studies mainly support changes in lipid droplet phenotype, insulin sensitivity, metabolic flexibility, and tissue function, whereas animal and cellular studies provide stronger mechanistic evidence for specific molecular pathways and organelle interactions. Therefore, mechanistic findings from experimental models should not be directly equated with human exercise responses. This review integrates current evidence from skeletal muscle, liver, heart, and adipose tissue and proposes "lipid fate remodeling" as a unifying framework for exercise-mediated protection against lipotoxicity. Future studies should move beyond static measurements of total lipid content and incorporate lipid molecular species, subcellular localization, lipid flux, organelle interactions, mitochondrial function, and individualized exercise prescription. Such an approach may improve mechanistic interpretation and help identify when exercise promotes adaptive lipid handling and when exercise-induced lipid mobilization may exceed tissue oxidative capacity. Accordingly, the translational goal of exercise prescription should be to restore the dynamic matching among lipid input, lipid droplet buffering, controlled lipolysis, mitochondrial oxidation, and organelle quality control, rather than to maximize lipid mobilization or fat oxidation in isolation.

Mitochondria, as the central organelle responsible for cellular energy supply and metabolic regulation, play a crucial role in maintaining cellular homeostasis through their quality control mechanisms. Mitochondrial quality control abnormalities can lead to cellular damage and dysfunction, directly contributing to the development of aging-related diseases such as cancer and cardiovascular diseases. A kinase anchoring protein 1 (AKAP1), as an anchoring protein localized to the outer mitochondrial membrane, also participates in regulating important biological processes such as mitochondrial structure and function, mitochondrial dynamics, and mitophagy. Recent studies show that AKAP1 interacts with ubiquinone oxidoreductase core subunit S1 (NDUFS1) and manganese superoxide dismutase (MnSOD) to maintain mitochondrial cristae structure, membrane potential, and respiratory function, and exerts antioxidant stress resistance. Furthermore, AKAP1 not only promotes mitochondrial fusion, but also interacts with estrogen receptor β (ERβ), protein kinase A (PKA), calcineurin (CaN), tribbles pseudo kinase 3 (TRIB3), and dynamin-related protein 1 (Drp1) to promote or inhibit mitochondrial fission. In addition, AKAP1 binds to PKA and activates the protein kinase B (AKT)/ endothelial nitric oxide synthase (eNOS) signaling pathway to inhibit excessive mitophagy. This article reviews the role and mechanism of AKAP1 in regulating mitochondrial quality control, aiming to provide valuable insights for the development and application of AKAP1-targeted drugs and exercise interventions. Firstly, as an outer mitochondrial membrane protein, AKAP1 is involved in maintaining the stability of mitochondrial crista structure, and its gene deletion will cause the loss and breakage of crista. Secondly, AKAP1 is also involved in maintaining mitochondrial membrane potential and oxidative phosphorylation function. mRNAs of short proteins, especially eukaryotic-origin electron transport chain (ETC) components, are specifically recruited by AKAP1 in a translation-independent manner that depends on mRNA splicing. AKAP1 loss can lower ETC levels, but its gene overexpression can reduce the production of reactive oxygen species (ROS) and enhance the mitochondrial membrane potential, the activity of respiratory chain complex I, and respiratory function. Thirdly, AKAP1 participates in the positive regulation of mitochondrial fusion, but its role in mitochondrial fission may be dependent on specific cell types, as it can promote or inhibit the mitochondrial fission by increasing phosphorylation of Drp1^Ser637. The effect and mechanism of AKAP1 on regulating other phosphorylation sites of Drp1 (e.g. Drp1^Ser616) and mitochondrial fusion proteins (e.g. mitofusion1/2 (Mfn1/2), OPA1) are unclear. Further studies on mitochondrial fusion/fission imbalance-related diseases (such as cardiovascular diseases and breast cancer) can be carried out to clarify the mechanism of AKAP1 in the pathological process and the relationship between AKAP1 and mitochondrial dynamic proteins in different tissues. Finally, AKAP1 can activate the AKT/eNOS signaling pathway and interact with PKA under the conditions of PTEN-induced putative kinase protein 1 (PINK1) gene deletion, thereby inhibiting excessive mitophagy. Future research can explore whether there are more mitophagy-related proteins (such as NIX, BNIP3, FUNDC1, etc.) involved and their association with AKAP1 by constructing tissue-specific PINK1 single knockout and PINK1/AKAP1 double knockout models. In addition, the molecular mechanism by which AKAP1 mediates the translocation of NDUFS1, PKA, and Drp1 from cytoplasm to mitochondria is unclear. In future research, it is necessary to elucidate how AKAP1 specifically recognizes its interacting proteins based on the cellular environment.

Directed cell migration is a fundamental biological process underlying embryonic development, tissue regeneration, wound healing, immune surveillance, and tumor metastasis. Among the various guidance cues that orient cell movement, chemotaxis (directed migration along chemical gradients) and electrotaxis (directed migration in response to electric fields) represent two physiologically important modes of directional migration that often coexist and interact in vivo. Phosphatidylinositol 3-kinase (PI3K), a highly conserved family of lipid kinases, plays a central role in both processes through the PI3K-PTEN-PIP₃ signaling axis. Class I PI3Ks catalyze the phosphorylation of phosphatidylinositol-4,5-bisphosphate (PIP₂) to generate phosphatidylinositol-3,4,5-trisphosphate (PIP₃) at the plasma membrane, which serves as a docking platform for pleckstrin homology (PH) domain-containing proteins such as Akt/PKB, PDK1, and Rac guanine nucleotide exchange factors, thereby coordinating actin cytoskeleton remodeling and leading-edge protrusion. The lipid phosphatase PTEN antagonizes PI3K activity by dephosphorylating PIP₃, and its asymmetric localization to the rear of migrating cells reinforces the front-rear PIP₃ gradient. In chemotaxis, PI3K is typically activated downstream of G protein-coupled receptors (GPCRs) or receptor tyrosine kinases (RTKs), with Ras GTPases mediating the recruitment and activation of PI3K at the leading edge. Although the PI3K-PIP₃ pathway is not absolutely essential for gradient sensing, it significantly amplifies directional signals, promotes persistent migration, and cooperates with parallel pathways such as TORC2-PKB and PLA₂ to ensure robust chemotactic responses. Distinct PI3K isoforms also exhibit non-redundant functions in chemotaxis: PI3Kγ and PI3Kδ are critical for immune cell migration, whereas PI3Kα and PI3Kβ predominantly mediate growth factor-stimulated motility. In electrotaxis, the upstream activation of PI3K is less well understood and may involve membrane potential changes, ion channel activity, and electrophoresis-driven redistribution of membrane proteins, rather than direct ligand-receptor binding. Importantly, the relative contribution of PI3K to electrotaxis varies considerably across cell types, and pharmacological or genetic inhibition of PI3K can either suppress or, in some contexts, enhance directional migration, reflecting the complex and context-dependent roles of distinct PI3K isoforms. Recent studies in Dictyostelium discoideum have revealed that individual PI3K isoforms play non-redundant and even opposing roles: PI3K1 and PI3K2 exert a suppressive effect on electrotaxis through Akt and ERK, whereas other isoforms may be required for maintaining directional responses. This review systematically compares the upstream activation mechanisms, polarity establishment, isoform specificity, and downstream effector networks of PI3K in chemotaxis versus electrotaxis. We highlight that despite divergent upstream sensing mechanisms, both migration modes converge on the PI3K-PTEN-PIP₃ module as a shared polarity-regulating platform. We further discuss the potential integration of chemical and electrical cues at the level of PI3K signaling and propose that the PI3K-PTEN-PIP₃ axis functions as a candidate integration node for multiple directional inputs. Finally, we outline key unresolved questions, including the molecular basis of electric field sensing, the precise functions of individual PI3K isoforms in electrotaxis, and how cells prioritize and integrate simultaneous directional cues, which may inform therapeutic strategies for wound healing, regenerative medicine, and cancer metastasis.

Objective Accurate inference of distant kinship relationships from genomic data is a fundamental challenge in population genetics, forensic identification, and pedigree reconstruction. Traditional identity-by-descent (IBD)-based methods, while widely adopted, suffer from limited predictive power at distant degrees due to the exponential decay of shared genomic segments. To address these bottlenecks, this study proposes and systematically evaluates a hybrid analytical framework that integrates IBD segment estimation with machine learning (ML) classification, aiming to substantially improve both the accuracy and computational efficiency of multi-class kinship degree prediction across a wide range of relatedness (first to ninth degree).Methods To achieve these objectives, the study was designed around a comprehensive comparative evaluation involving four distinct analytical methods applied to both simulated and real-world genotype data.The simulation family was generated by the ped-sim software for three independent families. The real family consists of five independent Chinese families.The first method comprised the conventional Phased IBD method, implemented through the standard Eagle+GERMLINE+ERSA workflow, in which Eagle performed haplotype phasing, GERMLINE detected IBD segments, and ERSA estimated kinship degree based on total shared segment length. The second method, designated as Phased IBD-ML, utilized the same phasing and IBD detection procedures but replaced conventional rule-based estimation with an XGBoost classifier trained on IBD-derived features. The third method employed the conventional Phase-free IBD method using identity-by-state (IBIS) software, which infers IBD segments directly from unphased genotype data without requiring prior phasing. The fourth method, Phase-free IBD-ML, applied the same XGBoost classifier to IBIS-extracted features.Results In simulated families, the IBD-XGBoost models outperformed conventional methods in predicting relationships from first to ninth degree, improving accuracy by an average of 12.82% (Phased) and 6.49% (Phase-free) and reducing false negatives (FN) by 8.45% and 4.63%, respectively. After statistical analysis, it was found that all the performance improvements reached a significant difference level (P<0.05). On real-world family data, the Phased IBD-XGBoost model achieved an average accuracy of 70.04%, a confidence interval accuracy (CIA: predicting kinship degrees to within ±1) of 94.8%, and the FN of 0.81% in identifying first to ninth degrees. Compared to the conventional Phased methods, these metrics represent improvements of 12.7%, 12.19%, and 13.06%, respectively, also achieving a significant improvement level (P<0.05).The Phase-free IBD-XGBoost model, while delivering slightly lower overall recognition performance (accuracy: 67.47%, CIA: 93.33%, FN: 2.18%), completed analysis in just 15 min, a 95% reduction in time compared to the 6 h required by the Phased model (P<0.05).Conclusion In conclusion, this study provides strong evidence demonstrating that combining IBD analysis with the XGBoost machine learning model offers a powerful and flexible approach for predicting distant kinship. This method significantly outperforms traditional IBD-based approaches in terms of accuracy and computational efficiency. The proposed hybrid analysis method leverages the data's adaptive learning capabilities, overcoming the inherent limitations of traditional threshold rules, and provides flexible options for phased or non-phased implementation to meet diverse practical needs. These findings clearly carry substantial practical implications for forensic identification, complex pedigree reconstruction, and large-scale genomic screening, thereby offering a robust, efficient, and versatile analytical tool for both routine casework scenarios and population-level investigations.

Objective Limited sample sizes in natural anti-aging compound datasets, class imbalance, and redundancy among high-dimensional molecular descriptors can restrict model performance in virtual screening. These problems make it difficult to identify informative molecular features and reliably recognize minority-class active compounds. This combination of data scarcity and imbalance increases the difficulty of establishing robust decision boundaries for compound screening. This study aimed to develop a tabular learning method tailored to small-sample and class-imbalanced settings for virtual screening of natural anti-aging compounds, with emphasis on improving active-compound recognition and candidate ranking.Methods Sq-TabPFN was developed by introducing sequential attention into the Tabular Prior-data Fitted Network (TabPFN). Across successive decision steps, the attention module generates instance-specific sparse masks that emphasize informative molecular descriptors while progressively suppressing redundant features. Thus, feature selection is integrated into model learning rather than performed as a fixed preprocessing step. To further address class imbalance, a strategy termed Wasserstein-Based Dual Secure Sampling (WBDSS) was proposed. WBDSS first removes potentially noisy majority-class samples near the decision boundary and identifies reliable minority-class samples according to model confidence. Representative root samples are then selected from the safe minority subset, and their neighboring samples are weighted according to Wasserstein distance to guide minority-class interpolation. The resulting rebalanced training set is used to fine-tune Sq-TabPFN, allowing adaptive feature selection and distribution-aware resampling to jointly strengthen minority-class learning while reducing the influence of noisy or unreliable samples.Results WBDSS-optimized Sq-TabPFN achieved the best overall screening performance among the compared models. Across 10 runs with different random seeds, its mean F1 score and PR_AUC were 0.557 and 0.792, respectively. Compared with TabPFN, F1 and PR_AUC increased by 19.5% and 21.1%, respectively, indicating improved recognition and ranking of active compounds under class imbalance. Statistical testing indicated that the improvement in PR_AUC was unlikely to result from random variation. Under scaffold-based splitting, F1 and PR_AUC were 0.54 and 0.79, respectively, showing that the model retained favorable predictive performance when structural overlap between the training and test sets was restricted. Experiments on two benchmark molecular classification datasets further supported the cross-task applicability of Sq-TabPFN. The optimized model was subsequently applied to the CMAUP natural product database for virtual screening and identified 10 candidate compounds. Five candidates were supported by published evidence or plausible aging-related mechanisms. Molecular docking analysis suggested potential binding between the candidate compounds and targets associated with anti-aging effects. These results support complementary roles for sequential attention in feature utilization and WBDSS in minority-class representation.Conclusion By integrating sequential-attention-based feature selection with WBDSS-based data rebalancing, the proposed method simultaneously addresses descriptor redundancy and class imbalance in small-sample molecular screening. WBDSS-optimized Sq-TabPFN improves the recognition and ranking of potentially active compounds and provides an effective computational approach for virtual screening of natural anti-aging compounds. The method can help narrow the candidate space and support the prioritization of compounds for subsequent experimental investigation. The screening and docking results are computational predictions rather than direct evidence of anti-aging activity and therefore require further biological validation.

Identifying drugs that reverse disease-associated transcriptomic features has been a well-established strategy in drug repurposing, yet its potential for de novo drug discovery remains largely untapped. Xing et al. recently presented the GPS (Gene Expression Profile Predictor on Chemical Structures) platform in Cell, which employs deep learning to directly predict compound-induced transcriptomic changes from chemical structures and enables virtual profiling of nearly seven million compounds for reversal score-driven screening. In hepatocellular carcinoma and idiopathic pulmonary fibrosis, GPS successfully identified novel compound series with favorable cellular selectivity and in vivo efficacy, demonstrating the feasibility of a closed-loop pipeline from virtual screening to lead optimization. This commentary systematically dissects GPS’s core innovations—the RCL(Robust Collaborative Learning) framework and Z-RGES (Z-score normalised Reverse Gene Expression Score), and benchmarks them against contemporaneous models including TranSiGen and MiTCP, delineating their complementary strengths and shared bottlenecks. Furthermore, we discuss three key frontiers—cross-context generalization assessment, multimodal validation integrating structural biology, and ADMET-incorporated multi-objective optimization—that shall guide the paradigm shift from repurposing known drugs to generative de novo design in transcriptomics-driven drug discovery.

Aromatic amino acids, including tryptophan, phenylalanine, and tyrosine, are not only fundamental protein building blocks but also metabolic precursors of neuroactive molecules such as serotonin, dopamine, and norepinephrine, thereby regulating gut-brain communication, immune homeostasis, and metabolic health. Their therapeutic potential in neuropsychiatric, neurodegenerative, and metabolic diseases is substantial, but conventional supply routes are constrained by the low efficiency and poor optical purity of chemical synthesis and plant extraction, while oral administration faces hepatic first-pass metabolism, poor blood-brain barrier penetration, and off-target toxicity. Microbial cell factories offer a green route to convert renewable carbon sources such as glucose and glycerol into high-value aromatic derivatives. This review first outlines microbial biosynthetic routes. From tryptophan, one route proceeds via tryptophan hydroxylase to 5-hydroxytryptophan, then to serotonin, and further to melatonin; another proceeds through the kynurenine pathway, branching to neuroprotective kynurenic acid and neurotoxic quinolinic acid. Phenylalanine can be converted to phenethylamine or, via phenylpyruvate and phenylacetaldehyde, to 2-phenylethanol. Tyrosine is hydroxylated to L-3,4-dihydroxyphenylalanine (L-DOPA), which is decarboxylated to dopamine and further hydroxylated to norepinephrine. These routes establish the material basis for in situ synthesis by engineered probiotics. The review then analyzes key bottlenecks and engineering breakthroughs. Rate-limiting enzymes such as tryptophan hydroxylase, tyrosine hydroxylase, dopamine β-hydroxylase, and thyroid peroxidase suffer from poor prokaryotic expression, inclusion body formation and low catalytic efficiency. Cytochrome P450 systems are limited by inefficient electron transfering from P450 reductase. To address these problems, rational design guided by molecular dynamics, directed evolution coupled with biosensors, substrate-channel engineering, and reconstruction of electron transport chains have been applied. Cofactor supply is another central constraint. BH4 is absent in common chassis and is oxidized during hydroxylation; therefore, de novo BH4 synthesis and regeneration, together with reduced nicotinamide adenine dinucleotide phosphate (NADPH), reduced flavin adenine dinucleotide (FADH2), and S-adenosylmethionine (SAM) supply and recycling, must be systematically engineered. Metabolic network reprogramming, including relief of feedback inhibition, removal of competing branches, and transporter-mediated efflux, further improves flux. Through these coordinated strategies, compounds such as 5-hydroxytryptophane (5-HTP), L-DOPA, dopamine, and phenethylamine have reached gram- to tens-of-grams-per-liter titers. However, the review emphasizes that in vitro fermentation cannot be directly extrapolated to in vivo therapy. A considerable capability gap separates high-yield fermentation from live biotherapeutic products. After oral administration, engineered probiotics confront the intestinal ecosystem, including oxygen gradients, pH fluctuations, gastric acid, bile salts, digestive enzymes, nutrient scarcity, and the absence of selective antibiotics. The review dissects four interwoven challenges: genetic stability; metabolic adaptability and environmental stress; secretion and absorption efficiency; and biosafety and clinical translation. Corresponding strategies include chromosomal integration, multi-copy integration, active stabilization circuits, encapsulation and adhesion engineering, oxygen-insensitive pathway design, efflux system mining, and biosafety locks. Importantly, these four dimensions are not isolated but mutually constrained, so live biotherapeutic product design must balance them systematically rather than optimizing a single property. Looking forward, the review proposes that artificial intelligence and machine learning will accelerate custom enzyme design; synthetic microbial consortia will modularly partition pathways to reduce metabolic burden and resolve cofactor conflicts; physical signals such as ultrasound, visible or near-infrared light, and alternating magnetic fields will enable spatiotemporally precise release; and targeted delivery to extraintestinal tissues will expand applications. Collectively, this review provides a theoretical framework and technical roadmap for moving aromatic amino acid derivatives from green biomanufacturing toward next-generation living therapeutics.

Traumatic brain injury (TBI) initiates secondary injury processes that include neuroinflammation, oxidative stress, blood-brain barrier disruption, cerebral edema, synaptic dysfunction, and cell death. These interacting processes contribute to persistent motor, cognitive, and neurobehavioral impairments. This review aims to summarize the major techniques, functional effects, and mechanistic networks of non-invasive neuromodulation in animal models of TBI. It also evaluates current limitations and future directions from the perspective of causal evidence. Published animal studies were qualitatively synthesized by stimulation modality, injury model, functional outcome, and mechanistic target, with particular attention to whether pathway involvement was supported by causal interventions or only by correlated changes. In recent years, transcranial magnetic stimulation (TMS), transcranial electrical stimulation (tES), low-intensity transcranial ultrasound stimulation (LITUS), and non-invasive vagus nerve stimulation have been investigated in animal models of TBI. TMS primarily regulates cortical excitability and activity-dependent network reorganization through magnetically induced electric fields. Different forms of tES modulate neuronal membrane potentials, corticospinal excitability, and inflammatory responses through weak electrical currents. LITUS provides relatively high spatial precision and penetration depth. It can affect both cortical and deep brain structures. Non-invasive vagus nerve stimulation acts indirectly through vagal afferent pathways and associated central and peripheral regulatory networks. Although these techniques have different physical inputs and stimulation targets, their effects may converge on several interconnected mechanistic hubs. These hubs include neurotrophic signaling and synaptic remodeling, neuroimmune regulation, endogenous antioxidant defense, and neurovascular unit protection. Brain-derived neurotrophic factor (BDNF)-tropomyosin receptor kinase B (TrkB)-cyclic adenosine monophosphate response element binding protein (CREB) signaling is associated with neuronal survival, synaptic protein expression, cortical excitability, and neural circuit reorganization. LITUS studies have provided interventional support for this pathway through BDNF neutralization. A LITUS study has provided interventional support for BDNF/TrkB signaling through BDNF neutralization. Evidence from repetitive TMS and transcranial direct current stimulation is mainly based on parallel molecular, electrophysiological, structural, and behavioral changes. Neuroimmune regulation commonly involves reduced microglial activation and shifts toward repair-associated microglial phenotypes. It also involves decreased production of pro-inflammatory mediators. Transcranial pulsed current stimulation and low-intensity focused ultrasound stimulation have been linked to orexin-A (OX-A)/orexin receptor type 1 (OX1R) signaling. This signaling may inhibit nuclear factor-κB (NF-κB) and the NOD-like receptor family pyrin domain-containing 3 (NLRP3) inflammasome. Receptor antagonism and inflammasome reactivation have strengthened the causal evidence for this pathway. However, OX-A/OX1R-related evidence currently covers only specific electrical and ultrasound paradigms. It should not be regarded as a universal mechanism shared by all four neuromodulation techniques. Antioxidant effects have been associated with nuclear factor erythroid 2-related factor 2 (Nrf2)/heme oxygenase-1 (HO-1) signaling. They are also associated with reduced reactive oxygen species production and inhibition of neuronal death. This evidence has mainly been established in studies of transcutaneous auricular vagus nerve stimulation (taVNS). Neurovascular effects may include preservation of tight-junction and endothelial proteins, reduction of cerebral edema, improvement of cerebral perfusion, and promotion of angiogenesis. However, the evidence supporting these mechanisms is unevenly distributed among stimulation techniques. Some studies have provided causal support through receptor antagonism, pathway activation or inhibition, neurotrophic factor neutralization, and chemogenetic manipulation. Most available evidence still relies on simultaneous changes in molecular expression, histopathology, and functional outcomes. Direct comparisons among stimulation paradigms are also scarce. Differences in animal species, injury models, skull conditions, stimulation doses, intervention windows, and outcome measures further limit cross-study comparisons. Overall, non-invasive neuromodulation shows the potential to promote functional recovery after TBI by modulating interconnected repair mechanisms. However, current evidence therefore cannot establish the overall superiority of a single technique. It also cannot identify a mechanism that is indispensable across all stimulation paradigms. Future studies should standardize injury and stimulation parameters. Behavioral, electrophysiological, imaging, histological, and molecular outcomes should be assessed together. Direct comparative studies, cell-type-specific manipulation, rescue experiments, long-term safety assessments, and adaptive closed-loop stimulation are also needed. These approaches may strengthen causal inference and improve clinical translation. This review summarizes the major techniques, functional effects, and mechanistic networks of non-invasive neuromodulation in animal models of TBI. It also evaluates current limitations and future directions from the perspective of evidence hierarchy.

Cartilage organoids (COs), novel three-dimensional in vitro tissue models derived from stem-cell-driven self-assembly, are capable of recapitulating partial tissue architecture, extracellular matrix composition and key cellular biological behaviors of native articular cartilage. They offer valuable research alternatives to conventional two-dimensional cell culture and animal experiments for deciphering pathological mechanisms of osteoarthritis (OA), performing high-throughput drug screening, and implementing regenerative repair of cartilage defects. Conventional two-dimensional monolayer cell culture cannot reconstruct the complex in vivo cartilage microenvironment characterized by hypoxia, dynamic mechanical loading, and intricate crosstalk between cells and extracellular matrix. Under such culture conditions, chondrocytes tend to undergo dedifferentiation and lose phenotypic characteristics specific to hyaline cartilage. In addition, animal models are constrained by inter-species discrepancies, obvious individual heterogeneity and relatively high experimental costs, which often limit the reliable translation of pre-clinical findings to human clinical scenarios. As promising alternative research tools, COs can maintain stable chondrocyte phenotypes. By applying pro-inflammatory stimuli or excessive mechanical stress, researchers can reconstruct OA-mimicking pathological microenvironments in vitro, which facilitates the exploration of critical molecular events such as chondrocyte hypertrophy, cellular senescence and extracellular matrix degradation. Meanwhile, COs possess translational prospects for regenerative therapy through direct transplantation into articular cartilage defect sites, providing a potential cell-based strategy for patients suffering from irreversible cartilage damage. Patient-derived COs further enable personalized pathological simulation, making it feasible to evaluate individual-specific drug responses ex vivo and avoid the confounding interference from inter-individual biological variation. Such personalized models also hold great promise for uncovering patient-specific disease signatures that cannot be readily observed in universal experimental systems. At present, in vitro fabrication of COs mainly contains two technical routes: scaffold-free self-assembly and scaffold-assisted culture. Available seed cell sources include induced pluripotent stem cells, multiple subtypes of mesenchymal stem cells as well as primary chondrocytes. Each cell source exhibits respective advantages and limitations in terms of chondrogenic differentiation efficiency, risk of pathological hypertrophic transformation and immunogenic property. Nevertheless, several prominent bottlenecks still hinder further progress of COs research. It remains technically difficult to faithfully reproduce temporospatial biochemical concentration gradients and physiological dynamic mechanical microenvironment of native cartilage. When COs grow above a certain millimeter scale, insufficient diffusion of nutrients and oxygen frequently gives rise to necrosis of cells located in the central core region. Long-term in vitro culture also increases risks of chondrocyte dedifferentiation and aberrant hypertrophy. Furthermore, the absence of an integrated quality-control system combining histological, biochemical and biomechanical evaluation indicators severely impedes experimental reproducibility and subsequent clinical translation across different laboratories. Many existing protocols remain confined to laboratory-scale operation and are hard to scale up for future clinical demands. This review summarizes developmental patterns and physiological microenvironment features of articular cartilage, compares merits and drawbacks of different seed cell types and construction protocols, and dissects four core restrictive challenges. Multiple optimization approaches are elaborated, including biomaterial modification, bioreactor cultivation, microfluidic organ-on-a-chip systems and cadherin-functionalized microspheres. Future research directions represented by hybrid organoid-chip composite platforms are also discussed. This work intends to advance standardized fabrication of COs and promote their broad applications in basic mechanistic research, high-throughput drug evaluation and regenerative medicine.

<b>Objective</b> Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by persistent deficits in social communication and restricted, repetitive behaviors. Increasing evidence suggests that these clinical manifestations are associated with atypical structural maturation of the brain’s triple-network system, including the default mode network (DMN), central executive network (CEN), and salience network (SN). However, how these networks exhibit distinct developmental trajectories remains insufficiently understood. In this study, we aimed to investigate network-specific structural maturation patterns in ASD using a BrainAGE framework based on voxel-wise gray matter density.<b>Methods</b> Structural magnetic resonance imaging (sMRI) data from 135 typically developing controls (TC) and 154 individuals with ASD were obtained from the Autism Brain Imaging Data Exchange (ABIDE) database. Voxel-based morphometry (VBM) was used to derive whole-brain gray matter density maps. Network-specific features were extracted from DMN, CEN, and SN templates. Feature selection was performed exclusively in the TC group using Pearson correlation analysis between voxel-wise gray matter density and chronological age, and voxels with <i>P</i><0.001 were retained as age-related features. The ASD group used the corresponding voxel locations selected from the TC group to ensure identical feature space. Brain age prediction was performed using an Individual-Level Weighted Multi-Layer Perceptron (ILWMLP) BrainAGE model. The model was trained and internally validated in the TC group using leave-one-out cross-validation (LOOCV), and subsequently applied to the ASD group to estimate network-specific brain age. BrainAGE (delta age) was calculated as the difference between predicted brain age and chronological age. Group differences in developmental trajectories were assessed using Fisher <i>Z</i> test. A two-way mixed analysis of variance (ANOVA) was further conducted to compare atypical developmental patterns across networks. In addition, general linear model (GLM) analysis with age as a continuous variable was performed to validate the robustness of the findings.<b>Results</b> The BrainAGE model showed significant predictive performance in all three networks (DMN: <i>r</i>=0.600; CEN: <i>r</i>=0.778; SN: <i>r</i>=0.742). Compared with TC participants, individuals with ASD exhibited significantly altered age-dependent brain maturation patterns across all networks. Fisher <i>Z</i> test revealed significant group differences in developmental trajectories within DMN, CEN, and SN (all <i>P</i><0.05). Importantly, the three networks demonstrated distinct developmental profiles. The DMN showed persistent delayed maturation across development, with the greatest deviation during adolescence. The CEN exhibited a dynamic pattern, shifting from delayed maturation in early development to relatively accelerated maturation in later stages. The SN showed progressively increasing delayed maturation with age. Two-way mixed ANOVA further confirmed significant differences in atypical developmental patterns among networks. GLM analysis yielded consistent results, supporting the robustness of the observed effects.<b>Conclusion</b> This study demonstrates that structural maturation abnormalities in ASD are not globally uniform but exhibit strong network specificity within the triple-network system. Although the DMN, CEN, and SN are functionally integrated, each network follows a distinct developmental trajectory during brain maturation. These findings extend the BrainAGE framework from global brain aging to network-level structural development and provide novel evidence for heterogeneous neurodevelopmental alterations in ASD. This network-specific perspective may contribute to the identification of imaging biomarkers and improve understanding of individualized brain maturation patterns in neurodevelopmental disorders.

Helicobacter pylori stands as a preeminent human-specific pathogen and the primary causative agent of chronic gastritis, peptic ulcer disease, and gastric adenocarcinoma. Central to its pathogenicity is the capacity to establish persistent colonization within the harsh acidic milieu of the stomach, a process in which bacterial adhesion to the gastric epithelium represents the indispensable initiating step. The first line of defense against microbial intrusion is mounted by gastric mucosal mucins and their elaborate O-glycosylation modifications, which together constitute both a physicochemical barrier and an immunomodulatory interface. The gel-forming mucins MUC5AC and MUC6, secreted by surface foveolar and glandular epithelial cells respectively, together with the membrane-tethered MUC1, are heavily decorated with structurally diverse O-glycans. The proper presentation of terminal glycan determinants—including Lewis b (Le^b), sialyl-Lewis x/a (sLe^x/sLe^a), and ABO histo-blood group antigens—is fundamental to maintaining mucosal integrity and resisting pathogen invasion. However, Helicobacter pylori infection profoundly disrupts this equilibrium by altering the expression levels and glycosylation patterns of host gastric epithelial mucins, thereby reshaping the very landscape it seeks to exploit. This review synthesizes current understanding of how Helicobacter pylori orchestrates mucosal colonization through a multivalent glycan-adhesin recognition system. The bacterium deploys an array of outer-membrane adhesins—most notably BabA and SabA—to specifically engage host gastric mucosal receptors, including Le^b, sialylated glycans, and related structures. By binding these carbohydrate ligands, Helicobacter pylori anchors itself firmly within the mucus layer and along the epithelial surface, effectively converting host defensive glycans into molecular footholds that facilitate initial attachment and long-term persistence. We posit that this glycan-adhesin crosstalk constitutes not only the gateway to initial colonization but a sustained pathogenic axis that actively perpetuates chronic inflammation and co-opts host glycosylation pathways throughout the infection course. These glycan-adhesin interactions not only determine bacterial tropism and colonization efficiency but also enable Helicobacter pylori to adapt to dynamic changes within the gastric microenvironment. Beyond this adhesive strategy, Helicobacter pylori infection provokes extensive mucosal and bacterial responses on two parallel fronts. On the host side, Helicobacter pylori-derived products and epithelia-derived chemokines drive robust neutrophil recruitment, which amplifies the local inflammatory response. On the bacterial side, under diverse environmental stresses encountered during chronic colonization—including, among others, nutrient deprivation, aerobic exposure, sublethal antibiotic pressure, proton-pump inhibitor intervention, and altered intragastric pH—the spiral bacilli undergo profound intracellular remodeling and progressively transition into dense coccoid forms, enabling the pathogen to transit into viable-but-non-culturable (VBNC) coccoid forms that exhibit enhanced resilience and tolerance to environmental stressors. Furthermore, lipopolysaccharide (LPS), an essential component of the Helicobacter pylori outer membrane, undergoes extensive glycosylation modifications that allow the bacterium to mimic host Lewis blood group antigens. This glycan camouflage serves a dual pathological role. On one front, it enables the bacterium to evade innate immune detection, thereby sustaining chronic persistent infection. On another, the presentation of host-like antigenic determinants provokes the production of cross-reactive autoantibodies, which mistakenly target autologous gastric epithelial components and exacerbate mucosal inflammatory injury. These interconnected mechanisms—adhesin-mediated colonization, inflammation-driven bacterial adaptation, and LPS-mediated immune subversion—act in concert to drive the progressive amplification of gastric mucosal inflammation and the stepwise evolution of precancerous lesions along the Correa cascade. Despite advances in understanding Helicobacter pylori pathogenesis, current clinical management remains encumbered by significant limitations. Conventional antibiotic-based eradication regimens are increasingly compromised by rising global rates of bacterial resistance to key drugs such as clarithromycin and metronidazole, the intrinsic difficulty of eradicating VBNC spherical forms, and persistently high post-treatment recurrence rates in endemic settings. In this context, anti-adhesion therapy targeting glycan-adhesin interactions has emerged as a promising alternative strategy. By competitively blocking bacterial attachment rather than directly killing bacteria, glycan-based inhibitors may reduce selective pressure for resistance development and provide a novel approach for controlling persistent infections. We therefore propose that anti-adhesion therapy should be reconceptualized as a longitudinal intervention targeting the entire colonization-persistence-progression continuum, rather than merely an adjunct to conventional eradication. However, several challenges remain before glycan-targeted therapies can be translated into clinical practice. The structural complexity and heterogeneity of host glycans, individual variations in gastric glycosylation patterns, and the difficulty in achieving highly specific targeting represent major obstacles. Future research integrating glycomics, microbiome analysis, structural biology, artificial intelligence-based prediction models, and precision medicine approaches may facilitate the identification of optimized glycan mimetics and personalized anti-adhesion therapies. Such advances will promote the transition of glycan-based intervention strategies from mechanistic exploration toward practical clinical applications, offering new opportunities for the prevention and treatment of Helicobacter pylori-associated gastric diseases.

Lactylation is an emerging post-translational modification (PTM) that connects lactate metabolism with epigenetic regulation and protein-functional remodeling. Since histone lactylation was identified in 2019, accumulating evidence has shown that lactylation is not restricted to chromatin regulation, but also modifies non-histone proteins involved in protein stability, DNA repair, immune escape, metabolic adaptation, fibrosis, and tissue remodeling. However, current studies remain highly heterogeneous, and a central unresolved question is how the same lactate-derived modification generates distinct biological outcomes across cell types and disease contexts. In this review, we summarize the molecular machinery of lactylation, including writers, erasers, candidate readers, donor-generating routes, and histone and non-histone substrates, and propose a four-layer lactylation circuit as an integrative analytical framework. The framework comprises four regulatory layers: Layer I, metabolic input, which determines lactate production, transport, and compartmentalization; Layer II, donor substrate supply, which includes a putative lactyl-coenzyme A (Lac-CoA)-dependent route and an alanyl-tRNA synthetase 1 (AARS1)-dependent direct lactate-sensing route; Layer III, enzymatic and spatial control, which governs substrate selectivity through writers, erasers, and local molecular organization; and Layer IV, decoding and PTM crosstalk, which involves candidate reader proteins and competition or cooperation among lactylation, acetylation, ubiquitination, and other lysine modifications. We further discuss emerging mechanistic issues, including local Lac-CoA generation, Lac-CoA versus acetyl-coenzyme A competition, AARS1-mediated lactyl-adenylate formation, structural recognition of lactyl-lysine by BRG1, DPF2, TRIM33, and BRD9, and site-specific non-histone lactylation of proteins such as p53 and MRE11. These mechanisms suggest that lactylation output is determined not simply by bulk lactate abundance, but by the coordinated effects of donor availability, enzyme selectivity, subcellular localization, substrate identity, and downstream decoding. We then apply this framework to three major disease domains. In neurodegenerative diseases, lactylation links microglial glycolytic reprogramming with inflammatory and neurodegenerative outputs. In cancer, distinct lactylation circuits regulate proliferation, immune evasion, DNA repair, and metabolic reprogramming. In metabolic and cardiovascular diseases, chronic metabolic stress is translated into fibrosis, steatosis, endothelial-to-mesenchymal transition, vascular remodeling, calcification, or tissue repair. To provide complementary systems-level support, we analyzed publicly available transcriptomic datasets from gastric cancer, Alzheimer""s disease, and metabolic dysfunction-associated steatotic liver disease (MASLD). Differential-expression and pathway-enrichment analyses revealed partially overlapping metabolic, inflammatory, and tissue-remodeling signatures across these disease contexts. Nevertheless, we emphasize that bulk mRNA-level changes do not directly indicate protein lactylation abundance, site-specific modification stoichiometry, or causal lactylation mechanisms. Finally, we summarize therapeutic strategies targeting different circuit layers, including lactate dehydrogenase and monocarboxylate transporters at the metabolic-input level, ACSS2-dependent donor supply, p300/CBP-associated enzymatic control, and downstream PD-1/PD-L1 or G6PD/pentose phosphate pathway programs. These strategies should currently be regarded as lactylation-related rather than lactylation-specific interventions. The proposed four-layer circuit therefore serves as a hypothesis-generating framework rather than a fully validated universal mechanism. Future studies integrating isotope tracing, subcellular acyl-CoA metabolomics, site-resolved lactylomics, structural biology, reader-domain perturbation, and single-cell or spatial multi-omics in disease-relevant models will be essential to define causal circuit topology, identify context-specific therapeutic vulnerabilities, and determine whether lactylation can ultimately be translated into a clinically actionable regulatory axis.

<b>Objective</b> To address the challenge of accurately predicting the activity of long-sequence sgRNAs in prokaryotic organisms, where complex secondary structure folding and substantial variability in target cleavage efficiency hinder precise prediction, this study developed a dual-channel binary classification deep neural network model.<b>Methods</b> A total of more than 130 000 sgRNA sequences were integrated from three prokaryotic datasets, namely Cas9, eSpCas9, and knoRecA_Cas9. The editing efficiency median of each dataset was used as the threshold to convert continuous activity values into balanced binary labels (high-efficiency and low-efficiency sgRNAs). The core architecture of the proposed model consists of two parallel feature extraction channels. The local feature channel employs continuous embedding and multi-scale convolutional layers (with kernel sizes of 3, 5, and 7) to capture sequence-specific local motifs. The global feature channel incorporates a pretrained RNA foundation model (RNA-FM) to extract high-dimensional semantic representations with 640 dimensions, followed by a nonlinear compression layer to reduce the feature dimension to 64, thereby capturing global evolutionary information and RNA spatial folding tendencies. The extracted features from the two channels are concatenated and fed into a fully connected classification head. During training and prediction, L2 regularization, dropout, and a five-model ensemble soft-voting strategy were applied to improve model robustness and generalization performance.<b>Results</b> Evaluation on the 3 independent test datasets demonstrated that the proposed model achieved <i>AUC </i>values of 0.795 1, 0.858 2, and 0.668 2 on the Cas9, eSpCas9, and knoRecA_Cas9 datasets, respectively, with an average <i>AUC</i> of 0.773 8. The proposed model consistently outperformed conventional machine learning classifiers, including GBDT, LightGBM, XGBoost, and one-hot encoded CNN models. Furthermore, compared with representative RNA sequence-based models, including RNASwan-seq and RNABERT, our model achieved higher <i>AUC </i>values across all three independent test datasets, with average <i>AUC</i> improvements of approximately 2.91% and 1.20%, respectively. These results indicate that the proposed framework can more effectively integrate local sequence patterns with global RNA semantic information, thereby improving the robustness of sgRNA activity prediction in complex prokaryotic systems. Ablation experiments and t-SNE visualization further confirmed that the dual-channel feature fusion strategy enhanced the model’s representational capacity. Deep interpretability analyses, including DeepSHAP, nucleotide substitution heatmaps, and paired-site masking analysis, revealed that the model successfully captured the importance patterns of the PAM-proximal seed region, nucleotide preference at key positions, and potential nonlinear interactions within the core targeting region.<b>Conclusion</b> The proposed dual-channel fusion model exhibited superior predictive performance compared with existing state-of-the-art approaches across three independent test datasets. By effectively integrating local sequence features with global RNA semantic representations, this framework provides a robust and generalizable computational approach for predicting sgRNA activity in long-sequence prokaryotic CRISPR systems. The three benchmark datasets used in this study, together with the source code for model implementation and evaluation, are publicly available at <ext-link ext-link-type="uri" xlink:href="https://github.com/yaoyueyao4-cmyk/sgRNA-project">https://github.com/yaoyueyao4-cmyk/sgRNA-project</ext-link>.

Acute ischemic stroke is a major cerebrovascular disease with high disability and mortality rates. Although clinical intravenous thrombolysis and mechanical thrombectomy can restore blood perfusion in ischemic brain tissue, they are highly likely to induce secondary cerebral ischemia-reperfusion injury (CI/RI) after blood reperfusion. The pathological process of CI/RI is extremely complex, presenting as a cascading damage network where oxidative stress outbreaks, mitochondrial dysfunction, excessive neuroinflammation, disruption of the blood-brain barrier structure, imbalance of the neurovascular unit homeostasis, and multiple programmed cell death pathways are intertwined. Various pathological links positively feedback on each other, continuously expanding the scope of cerebral infarction, causing irreversible apoptosis of neurons, and ultimately leading to long-term neurological deficits. At present, commonly used neuroprotective small-molecule drugs in clinical practice, such as edaravone and butylphthalide, have several drawbacks. These include poor blood-brain barrier penetration and low drug enrichment concentration at the lesion site; insufficient lesion targeting and obvious off-target toxic side effects of systemic administration; a narrow treatment time window and the inability of single-target drugs to block multi-pathway cross-linked damage. These limitations have made it difficult to translate a large number of basic research findings into clinical applications. The nanodelivery system, with its four core advantages of tunable physicochemical properties, flexible surface functionalization, integration of multifunctional modules, and response to the pathological microenvironment, can efficiently cross the damaged blood-brain barrier through endocytosis, accumulate directionally in the ischemic penumbra, and release drugs in a targeted manner in response to characteristic signals such as low pH, high reactive oxygen species, and high matrix metalloproteinases at the lesion site. This provides a new technical approach for achieving both precise treatment and early molecular diagnosis of CI/RI. This paper systematically reviews the core pathological mechanisms of CI/RI and outlines six major mainstream nanodrug delivery systems, namely polymer-based (synthetic polymers, natural biopolymers), lipid-based (traditional lipid carriers, intelligent lipid nanoparticles), inorganic nanozyme materials, cell membrane biomimetic nanocarriers, natural extracellular vesicles, and carrier-free self-assembled nanoformulations. It compares the material construction characteristics, brain-targeting mechanisms, neuroprotective advantages, and inherent limitations of each type of carrier. It also elaborates on the recent cutting-edge research progress of nanodrug delivery systems in three directions: targeted single-pathway intervention, multi-pathological pathway collaborative blockade, and multimodal visual imaging diagnosis of CI/RI. Moreover, it analyzes the current core practical bottlenecks restricting the translation of nanodrugs from the laboratory to clinical applications, including long-term in vivo biosafety risks, insufficient lesion targeting accuracy, great difficulty in large-scale and standardized preparation, and the lack of a unified evaluation system for drug efficacy and safety. By integrating frontier research cases of co-delivery of multimodal imaging probes and neuroprotective drugs, this paper distinguishes the functional differences between single-therapeutic nanocarriers and integrated diagnosis and treatment platforms, and clarifies how imaging modules achieve precise lesion localization and dynamic disease course monitoring, providing a clear development context for the research on visual and precise intervention of stroke. From the cross-disciplinary perspectives of biochemistry, materials science, and pharmacy, this paper prospects the development ideas of the next-generation intelligent, highly biocompatible, and integrated diagnosis and treatment nanoplatforms, as well as the promotion of the combined application of nanodelivery technology with clinical thrombolysis and thrombectomy protocols. It is expected to provide systematic and comprehensive theoretical support for the structural optimization, basic research and development, and clinical translation of novel nanodrugs for cerebral ischemia-reperfusion injury.

Thyroid hormones (THs) play a crucial role in the regulation of energy metabolism, participating in various physiological processes such as cellular metabolism, nutrient transport, and energy homeostasis. In recent years, multiple clinical studies have demonstrated that disturbed THs homeostasis is closely associated with the onset and progression of various liver diseases. Nevertheless, the clinical observations cannot be fully explained merely by the canonical metabolic-regulatory mechanisms of THs. Therefore, the non-metabolic biological effects of THs have driven the attentions of researchers, and THs-mediated immune responses are considered to contribute to the hepatic immunometabolic homeostasis and progression of liver diseases. Liver is a vital organ integrating metabolism and local immune defense. Under physiological conditions, hepatic metabolic signals and immune responses collaborate to sustain intrahepatic microenvironmental homeostasis. However, under pathological circumstances such as lipid overload, inflammatory stimulation, and toxin-induced injury, the thyroid hormone/thyroid hormone receptor (TH/THR) signaling axis is prone to functional dysregulation. The dysregulation of TH/THR contributes to liver injury through two interconnected mechanisms. On the one hand, it directly disturbs hepatic metabolic homeostasis and impairs metabolic function of hepatocytes, thereby promoting hepatic lipid accumulation and lipotoxicity. On the other hand, it remodels the phenotype and function of certain intrahepatic immune cells. For instance, impaired thyroid hormone receptor β (THRβ) signaling in hepatocytes exacerbates lipid deposition and lipotoxic injury; meanwhile, pathologically elevated intrahepatic T3 levels may alter Kupffer cell polarization, promoting a pro-inflammatory phenotype and increasing the production of pro-inflammatory mediators, thereby amplifying hepatic inflammation. Clinical evidences further indicate that thyroid dysfunction, dysregulated intrahepatic deiodinase expression, and low-T3 syndrome are associated with disease severity and poor prognosis in patients with metabolic dysfunction-associated steatohepatitis (MASH), liver fibrosis, and hepatocellular carcinoma (HCC). Accordingly, elucidating the metabolic and immunomodulatory mechanisms governed by TH/THR signaling in the liver will provide a more comprehensive understanding of the pathogenesis of liver diseases and inform the development of immunometabolism-targeted therapies. In addition to the classical metabolic effects and signaling pathways of THs, this review systematically summarizes the roles and molecular mechanisms of the TH/THR signaling in immune regulation, including innate and adaptive immune cells. First, THs regulate the activation, polarization, and functional reprogramming of innate and adaptive immune cells through both genomic and nongenomic mechanisms, thereby contributing to immune homeostasis within tissues. On this basis, this review further discusses the core molecular mechanisms by which perturbed TH/THR signaling participates in the progression of major chronic liver diseases, including MASH, liver fibrosis, and HCC, with special emphasis on the regulatory effects of THs on Kupffer cells, the key hepatic innate immune cells. Currently, targeting liver-enriched THRβ has emerged as a promising therapeutic strategy for liver diseases, and multiple highly-selective THRβ agonists have progressed into different phases of clinical development. In this review, the latest research progress and clinical value of THRβ-targeted agents are systematically outlined. Among them, Resmetirom has demonstrated clinically meaningful improvements in hepatic steatosis and fibrosis and was approved by the US Food and Drug Administration (FDA) for adults with noncirrhotic MASH and moderate-to-advanced fibrosis. In summary, this review elaborates the dual regulatory functions of the TH/THR signaling in metabolic and immune homeostasis, as well as its critical biological significance in the progression of chronic liver diseases. It also discusses the clinical translational potential of THRβ-targeted drugs such as natural products and functional components in traditional Chinese medicine. This review will provide scientific references for in-depth mechanistic research on liver diseases and the development and optimization of novel immunometabolism-targeted intervention regimens.

The sensitive scalp is a complex syndrome characterised by abnormal nerve sensations, such as stinging, itching, burning, and tightness, and diminished tolerance to physical or chemical stimuli that would not typically cause such unpleasant responses, with its pathology involving physical barrier destruction at the stratum corneum level, neural hyperreactivity and sensory nerve sensitization, and local immunological dysregulation and chronic low-grade inflammation. The homeostasis of the scalp microenvironment plays a crucial regulatory role in maintaining the hair follicle growth cycle, yet this “soil” factor has long been underappreciated in hair loss research, which has traditionally centered on intrinsic follicular defects, such as genetic predisposition or androgen-driven mechanisms. In this review, we summarized the potential mechanisms underlying the imbalance in the “barrier-neural-immune” network in sensitive scalps. Sensitive scalp is further manifested as impaired skin barrier, including increased transepidermal water loss and abnormal lipid secretion. These alterations are accompanied by increased surface pH, excessive secretion of sebum rich in free fatty acids, reduced microbial diversity, and dysbiosis characterized by a reduction in Cutibacterium and an increase in Staphylococcus colonization. Mechanistically, we also analyzed how neurogenic inflammation, mediated by substances such as substance P (SP) and calcitonin gene-related peptide (CGRP), along with persistent low-grade inflammation, disrupts the microenvironment of the hair follicle stem cell niche through mechanisms involving mechanosensitive ion channels (e.g., Piezo1), oxidative stress, and cytokine cascades. When the barrier is compromised, the weakened stratum corneum may facilitate the penetration of external stimuli, sensitizing cutaneous nerve endings. Those hyperresponsive nerves then fire off neuropeptides like substance P and CGRP. These peptides activate mast cells and other immune cells via the MrgprB2 receptor, prompting degranulation and a wave of pro-inflammatory mediators that kick off neurogenic inflammation. At the same time, oxidative stress and inflammatory cytokines drive a persistent low-grade inflammation that directly damages the hair follicle stem cell niche. Physical compression from shrinking hair shafts activates Piezo1 channels, setting off calcium influx and caspase-3-dependent apoptotic signaling. We primarily considered the interactions between inflammatory signaling and the sensitization of the androgen receptor (AR) pathway, as well as the inhibition of the Wnt/ β-catenin pathway, which may subsequently contribute to hair follicle miniaturization and disruption of the growth cycle. The inflammatory microenvironment appears to boost AR expression in dermal papilla cells, making them more sensitive to dihydrotestosterone (DHT). This heightened sensitivity may promote the expression of transforming growth factor-β1 (TGF-β1), TGF-β2, and Dickkopf-1 (DKK1), which suppress the Wnt/β-catenin pathway—essential for follicle growth and regeneration—potentially accelerating follicular miniaturization and stalling the growth cycle. Coupled with direct inflammatory apoptosis, oxidative stress, and circadian disruption, these routes may collectively speed up hair follicle stem cell depletion and diffuse hair thinning. These findings point to the need for more mechanistic and translational research to test the proposed pathways and explore therapies that target the microenvironment in sensitive scalp-related hair loss. Ultimately, we hope to provide theoretical support for the intervention of hair loss by restoring the steady state of the scalp microenvironment and combining with new precision delivery technologies such as microneedles.

Bone remodeling is essential for maintaining skeletal integrity and mineral homeostasis and depends on the coordinated activities of osteoclasts, osteoblasts, and osteocytes. Osteoclast-mediated bone resorption must remain closely coupled with osteoblast-mediated bone formation, while osteocytes continuously sense mechanical loading and changes in the local and systemic environment. Reactive oxygen species are involved in these physiological processes, particularly in the regulation of osteoclast differentiation and osteoblast function. Once oxidative stress exceeds the capacity of cellular antioxidant systems, however, this balance may be disturbed, leading to enhanced bone resorption, impaired bone formation, and deterioration of bone quality. In recent years, ferroptosis has attracted increasing attention in studies of bone metabolism. Ferroptosis is characterized by iron-dependent accumulation of lipid peroxides, but ferroptosis-related alterations in bone tissue may occur before obvious cell death becomes detectable. Changes in the labile iron pool, membrane lipid composition, antioxidant capacity, and mitochondrial metabolism can already influence the biological behavior of bone cells at an earlier stage. On this basis, ferroptosis susceptibility may provide a useful way to describe a reversible metabolic state in which cells become increasingly vulnerable to lipid peroxidation while widespread membrane damage and terminal cell death have not yet occurred. Such a state should not be identified by a single molecular marker. A more appropriate assessment would combine changes in intracellular iron handling, lipid peroxidation, antioxidant defense, cell viability, bone-cell function, and the response to ferroptosis-related interventions. Measurements of Fe²⁺ or the labile iron pool, C11-BODIPY oxidation, 4-HNE and MDA, GPX4 activity or expression, GSH/GSSG balance, LDH release, and PI staining may therefore provide complementary information when interpreted together with functional changes and rescue experiments. Several metabolic processes may contribute to the development of ferroptosis susceptibility in the bone. Disturbances in iron uptake, ferritin storage, ferritinophagy, and ferroportin-mediated iron export can increase the availability of redox-active iron. At the same time, ACSL4- and LPCAT3-dependent incorporation of polyunsaturated fatty acids into membrane phospholipids increases the pool of oxidizable lipid substrates. Reduced activity of the system Xc^--GSH-GPX4 pathway further weakens the ability of cells to remove lipid hydroperoxides, while FSP1-CoQ and mitochondrial DHODH-CoQ systems provide additional antioxidant protection. Mitochondrial dysfunction, inflammation, and abnormalities in calcium-phosphate and endocrine regulation may further increase this metabolic pressure. Importantly, the consequences are unlikely to be identical among bone-cell populations. In osteoclasts, increased iron availability and oxidative stress may strengthen RANKL-NFATc1-associated differentiation and resorptive activity. In osteoblasts, persistent lipid peroxidation may interfere with RUNX2-related differentiation, collagen synthesis, matrix production, and mineralization before extensive cell loss occurs. Osteocytes, because of their long lifespan and specialized location within the mineralized matrix, may be particularly sensitive to chronic metabolic stress, resulting in impaired mechanosensing and disturbed RANKL- or FGF23-related feedback. Evidence linking these changes with osteoporosis, diabetes-related bone disorders, chronic kidney disease-mineral and bone disorder, inflammatory bone loss, and particle-induced osteolysis is increasing, although the strength of evidence differs among disease models. Changes in ROS, MDA, GPX4, or other individual indicators alone are insufficient to demonstrate ferroptosis-related susceptibility, and stronger conclusions require evidence of iron dependence, lipid peroxidation, temporal sequence, and reversibility. Emerging single-cell, spatial, proteomic, lipidomic, and metabolomic approaches may help identify susceptible cell populations and bone-remodeling regions more precisely, although the mineralized matrix and the effects of sample processing remain important technical limitations. Overall, ferroptosis susceptibility may represent an intermediate metabolic state linking disturbances in iron and redox homeostasis with progressive uncoupling of bone resorption and formation. Clarifying its cell-specific features, temporal development, and reversible window may improve our understanding of how chronic metabolic stress contributes to bone remodeling disorders and may also provide a basis for earlier and more targeted intervention.

Objective Stroke remains the leading cause of acquired disability in adults worldwide, imposing a substantial socioeconomic burden on patients, families, and healthcare systems. Among the various neurological deficits caused by stroke, motor dysfunction——particularly contralateral hemiplegia——is the most common and debilitating consequence, and it predominantly results from damage to the cortical motor conduction pathways. When motor cortex neurons or their projecting fibers are damaged, the distal axons and their associated myelin sheaths undergo a series of progressive, programmed degeneration, fragmentation, and eventual phagocytic clearance. This pathological process is termed Wallerian degeneration (WD). However, the temporal progression of WD following ischemic stroke and the associated alterations in the local glial microenvironment remain incompletely understood. The present study was designed to systematically elucidate the time course of corticospinal tract (CST) axonal WD after ischemic stroke in mice, and to simultaneously characterize the dynamic pathological changes of astrocytes, microglia, oligodendrocyte lineage cells, and myelinating oligodendrocytes during the WD process.Methods Eight-week-old male C57BL/6J mice were randomly assigned to control and ischemic stroke groups. A focal cortical ischemic stroke model was established in the right sensorimotor cortex using the photothrombotic method. This method induces localized intravascular thrombosis in the target cortical region through precisely controlled interaction between the photosensitizer and laser light of specific wavelength. At day 1 post-surgery, 2,3,5-triphenyltetrazolium chloride (TTC) staining was performed to evaluate infarct volume, hematoxylin and eosin (HE) staining for histopathological examination, and Nissl staining for assessment of neuronal injury. For axonal degeneration time-course analysis, rAAV-CAG-mCherry anterograde viral tracing was employed to specifically label CST neurons. Cervical spinal cord sections at the CST region were collected at day 3, 5, 7, 14, and 60 post-injury. Morphological changes of mCherry-positive CST fibers were observed under fluorescence microscopy, and the fluorescence signal area fraction within the defined region was quantitatively analyzed to evaluate the dynamic process of axonal degeneration. For glial response assessment, immunofluorescence staining against Iba1 (microglia), GFAP (astrocytes), Olig2 (oligodendrocyte lineage cells), and APC (myelinating oligodendrocytes) was performed at day 7, 14, and 60 post-surgery. The numbers of positive cells were counted to quantitatively analyze the dynamic changes in each glial population at different time points after ischemia, thereby revealing the evolution of the glial microenvironment during WD.Results The photothrombotic method stably and reliably established an ischemic brain injury model confined to the sensorimotor cortex. Viral tracing revealed that the proportion of mCherry-positive CST fibers was significantly reduced at day 7 after ischemic injury to the neuronal soma. Immunofluorescence staining demonstrated that during WD, the numbers of microglia and astrocytes in the CST region were significantly increased, whereas the numbers of oligodendrocyte lineage cells and myelinating oligodendrocytes showed no significant changes.Conclusion WD of CST axons is observed at day 7 post-ischemic stroke in a delayed manner. This degenerative process selectively activates microglia and astrocytes, while leaving the number of oligodendrocyte lineage cells largely unaffected.

In human assisted reproductive technology (ART), embryos experience an extremely high rate of developmental arrest (over 50%) during the 5-day developmental window prior to blastocyst formation, which constitutes a major bottleneck limiting clinical pregnancy rates. A recent paper published in Cell titled “Two distinct causes contribute to the low efficiency of human pre-implantation development” reveals the two independent causes underlying this phenomenon. Utilizing a custom-developed high-throughput dual-view live-cell light-sheet microscope, the research team achieved the first continuous live-cell imaging of human and non-human primate embryos with high spatiotemporal resolution for up to 120 h. The study discovered that embryos are highly prone to errors during the second mitosis. Multipolar spindle assembly, triggered by centriole overduplication, leads to chromosome segregation errors and micronuclei formation, which subsequently induces early developmental arrest (prior to the 8-cell stage). Conversely, late developmental arrest (on day 4) is primarily driven by the endoplasmic reticulum (ER) stress response. Furthermore, the study demonstrated that transient treatment with a PLK4 inhibitor (centrinone) can effectively prevent spindle defects at the 2-cell stage. This research not only reshapes our understanding of the developmental dynamics of human pre-implantation embryos but also provides new insights into the mechanisms underlying human pre-implantation developmental arrest and offers valuable references for future studies on assisted reproduction and potential intervention strategies.

During habitation on low-Earth-orbit (LEO) space stations, astronauts are exposed to a combination of orbital microgravity, space radiation, an enclosed and artificially controlled cabin environment, altered sleep-wake rhythms, and sustained operational and psychosocial demands. These factors can modify sensorimotor inputs, body-fluid distribution, arousal state, and task-related neural processing, thereby affecting brain function and behavior. Because evidence from real spaceflight remains limited by small samples, heterogeneous mission durations, and different measurement time points, this review integrates findings from real human spaceflight with relevant ground-based analog studies while treating real orbital studies as the primary source of evidence. We focus on electroencephalography (EEG) and event-related potentials (ERPs), structural and functional neuroimaging, and sensorimotor and cognitive outcomes, with particular attention to the direction, anatomical distribution, temporal course, and possible physiological significance of the reported changes. EEG studies indicate that long-duration spaceflight is associated with increased resting-state beta activity, particularly in sensorimotor and frontotemporal regions, together with state-dependent alterations in beta-band functional connectivity. In contrast, alpha source power and functional connectivity associated with the default mode network are reduced during long-duration missions, and some of these effects may persist into the early postflight period. Task-related electrophysiological findings further suggest that sustained mental workload can weaken frontocentral theta-temporoparietal gamma phase-amplitude coupling, whereas in-flight decision-making is accompanied by reduced P300 amplitude and reorganization of feedback-related potentials. Short-duration missions, however, have shown relative preservation of posterior dominant rhythms and some ERP measures, indicating that mission duration, recording state, and task demands substantially influence the observed electrophysiological phenotype. Neuroimaging provides complementary evidence at structural and network levels. Long-duration spaceflight has been associated with upward brain displacement, ventricular and cerebrospinal fluid expansion, regional gray- and white-matter remodeling, and alterations in white-matter structural connectivity. Functional magnetic resonance imaging (fMRI) studies further indicate reorganization of vestibular-sensorimotor, visual, and default mode networks. Some network changes recover after return to Earth, whereas others may remain detectable for months. Importantly, most MRI measurements are obtained after landing rather than in orbit; therefore, postflight findings may reflect a mixture of in-flight adaptation, the rapid transition back to 1 g, and subsequent readaptation rather than a single spaceflight effect. Behavioral evidence is most consistent for sensorimotor function. Early after return, astronauts frequently show impaired postural control, gait, spatial orientation, and coordination, consistent with reweighting among vestibular, visual, and proprioceptive information. By comparison, simple cognition and some critical operational tasks are often preserved. More subtle or high-load processes-including sustained attention, feedback evaluation, temporal processing, error monitoring, and metacognitive calibration may be more sensitive than global cognitive performance. This dissociation between neural change and preserved behavior suggests that altered brain activity should not automatically be interpreted as functional deterioration. Integrating the available modalities, we propose that LEO space-station habitation is better characterized by system-specific and stage-dependent neural adaptation than by a generalized increase or decline in brain function. The most coherent cross-modal pattern involves sensorimotor reorganization: reduced gravitational reference and weight-bearing feedback under microgravity may drive reweighting of vestibular, visual, and proprioceptive inputs, accompanied by changes in beta activity and vestibular-sensorimotor networks and followed by transient postflight sensorimotor deficits. A second major process involves fluid-related structural remodeling, in which cephalad fluid shifts and altered hydrostatic gradients may contribute to changes in cerebrospinal fluid spaces, ventricular volume, and brain tissue configuration. Resting-state and cognitive electrophysiological changes likely represent additional adjustments in network regulation and resource allocation, but their functional interpretation remains more dependent on sleep, workload, recording conditions, and individual differences. Future studies should move beyond isolated pre-post comparisons and group averages toward standardized longitudinal designs spanning preflight, multiple in-flight time points, and postflight recovery. Synchronized assessment of brain function, behavior, sleep, and key environmental parameters should be used to characterize individual adaptation trajectories and determine which neural changes reflect successful compensation, which predict delayed recovery, and which have genuine value for brain-health risk prediction and individualized countermeasures.

Conjugation chemistry is a major determinant of the structural homogeneity, stability, pharmacokinetics, manufacturability, and therapeutic index of antibody-drug conjugates (ADCs). Conventional conjugation through solvent-accessible lysine residues or cysteines generated by reduction of interchain disulfide bonds has enabled the development of many clinically successful ADCs. However, these approaches can produce mixtures that differ in drug-to-antibody ratio (DAR) and attachment site, complicating product characterization and, in some settings, contributing to variable stability, clearance, efficacy, and off-target toxicity. This review examines how advances in conjugation technology address these limitations and how the choice of conjugation strategy influences chemistry, manufacturing, and controls (CMC), preclinical performance, and clinical translation. For clarity, the principal conjugation approaches are organized into five broadly defined and partially overlapping categories: nonspecific conjugation; enzyme-mediated site-specific conjugation; click chemistry and bioorthogonal conjugation, including chemoenzymatic glycan remodeling; photocatalytic site-selective conjugation; and conjugation based on genetically encoded noncanonical amino acids. This framework is intended to facilitate comparison rather than to represent mutually exclusive technological generations. For each category, we compare the origin and accessibility of the conjugation site, reaction mechanism, operating conditions, achievable drug-to-antibody ratio (DAR) distribution, requirement for antibody engineering, product homogeneity, linker stability, scalability, and stage of development. Enzyme-mediated approaches involving sortase A, microbial transglutaminase, formylglycine-generating enzyme, or ligase-dependent conjugation enable controlled modification at defined residues or engineered recognition motifs. Click and bioorthogonal reactions, particularly strain-promoted azide-alkyne cycloaddition, permit rapid and selective coupling under mild conditions and can be combined with glycan remodeling or orthogonal reactive handles to generate homogeneous, dual-payload, or multi-payload ADCs. Photocatalytic site-selective approaches provide temporal control over bioconjugation reactions but remain predominantly at the methodological or preclinical stage. Genetic code expansion enables the precise incorporation of reactive noncanonical amino acids, although limited expression efficiency, manufacturing complexity, and production cost remain important barriers to broader application. Comparative analysis indicates that no single conjugation platform is optimal for every ADC. Greater site control and a narrower DAR distribution can improve product consistency and may reduce the rapid clearance or aggregation associated with some highly loaded species. These advantages, however, must be balanced against antibody engineering requirements, enzyme or reagent supply, reaction yield, purification burden, analytical complexity, scale-up feasibility, and regulatory expectations. Moreover, clinical performance cannot be attributed to conjugation chemistry alone. Target biology, antigen density and heterogeneity, antibody properties, linker cleavage, payload potency and permeability, DAR, dosing regimen, and patient selection jointly determine efficacy and safety. Advantages observed in platform studies should therefore not be assumed to confer clinical benefit without appropriate comparative evidence. Representative approved and clinical-stage ADCs are considered to illustrate both the contribution and the limits of conjugation strategy in clinical development. Persistent translational challenges include premature payload release, off-target uptake, dose-limiting hematologic or organ toxicities, tumor heterogeneity, impaired internalization and lysosomal processing, drug-efflux mechanisms, and resistance to payload-induced cell death. These issues emphasize the need to integrate conjugation-site selection with linker and payload design rather than optimize each component in isolation. Future development is expected to combine site-controlled conjugation with hydrophilic linkers, dual-payload or bispecific architectures, conditionally activated release mechanisms, and biomarker-guided patient stratification. Expansion into autoimmune and infectious diseases will place additional demands on selectivity, tolerability, and manufacturing consistency. Overall, rational selection of conjugation chemistry, supported by rigorous analytical characterization and disease-specific pharmacology, should help translate structurally defined ADCs into products with more predictable quality and wider therapeutic windows.

Consciousness, which underpins human perception, cognition, and behavioural responses, has long been a major focus of neuroscience and medicine. Alterations in consciousness occur not only during reversible processes such as sleep and anaesthesia, but also in pathological conditions, including disorders of consciousness (DOC) and brain injury. With recent advances in the synchronized acquisition and analysis of multimodal physiological signals, such as electroencephalography (EEG) and electrocardiography (ECG), the dynamic interactions between the heart and brain have become increasingly apparent. A growing body of evidence indicates that heart-brain interactions are closely associated with the maintenance and transition of consciousness states, offering a novel perspective for elucidating their underlying mechanisms and supporting relevant clinical assessments. Current approaches to heart-brain interaction analysis for consciousness-state assessment can be broadly grouped into three categories: ECG-based methods, EEG-based methods, and heart-brain coupling analysis. These approaches characterize heart-brain interactions during changes in consciousness from the complementary perspectives of autonomic regulation, cortical interoceptive processing, and cross-system information transfer, respectively. Nevertheless, this field remains at an early stage. Most studies have focused on a single experimental context or a fixed analytical method, with insufficient systematic consideration of the applicability and limitations of different approaches. Moreover, the mechanisms underlying the emergence and alteration of consciousness remain incompletely understood. Accordingly, this review examines heart-brain interaction methods employed in recent consciousness-state assessment studies in accordance with progressively deeper analytical perspectives, and systematically evaluates the advantages and limitations of these methods across distinct consciousness-related conditions. Heart rate variability (HRV) is one of the classical ECG-based indices used to investigate heart-brain interactions. It refers to fluctuations in consecutive interbeat intervals over a specified time period and is regarded as a physiological measure of autonomic nervous system regulation that may indirectly reflect the strength of heart-brain interactions. Depending on the analytical dimension, HRV analysis is commonly divided into time-domain, frequency-domain, and complexity-index analyses. These approaches characterize autonomic activity in terms of temporal fluctuations, the distribution of frequency components, and nonlinear dynamical properties, respectively. Although extensive research supports their physiological interpretation, they cannot characterize directional interactions from the heart to the brain. Among EEG-based methods for heart-brain interaction analysis, EEG activity directly related to cardiac fluctuations has received particular attention. One of the most widely used measures is the heartbeat-evoked potential (HEP). HEP-based approaches are considered to directly reflect the capacity of the cerebral cortex to process cardiac afferent signals and may be especially advantageous for characterizing rapidly changing consciousness levels, such as those observed during sleep. However, they do not comprehensively capture dynamic bidirectional information exchange between the brain and heart and are susceptible to cardiac artefacts, respiration, and variation in preprocessing procedures. Heart-brain coupling analysis quantifies functional associations between ECG and EEG through cross-system interaction measures, thereby assessing dynamic interactions between the heart and brain. It emphasizes information transfer, feedback regulation, and cross-timescale coordination between the two systems, offering a more comprehensive perspective for investigating changes in consciousness during sleep, anaesthesia, and DOC. However, the neurophysiological mechanisms and physiological interpretation of individual computational indices have not yet been fully established, and the robustness of their results and their potential for clinical generalization require further validation. Finally, this review summarizes the commonalities and differences among the three classes of methods across research domains involving different consciousness states. It further argues that heart-brain interaction research requires deeper investigation at both the mechanistic level of consciousness and the level of clinical application. Current challenges and key issues are identified, and future directions and feasible pathways for heart-brain interaction research are discussed, with the aim of providing new perspectives for researchers in related fields.

Objective To address the diagnostic lag, low consistency, and subjective biases associated with traditional clinical scales (such as HAMD and DSM-5) and psychiatric interviews, this study aimed to establish objective, non-invasive electrophysiological biomarkers for major depressive disorder (MDD). By leveraging eyes-closed resting-state electroencephalography (rsEEG), which offers superior patient compliance and eliminates task-related cognitive confounding compared to task-state paradigms, we systematically investigated the spatial distribution patterns, disease-course-related non-linear evolution, and temporal robustness of the θ/β ratio (TBR) and θ/α ratio (TAR). Specifically, we aimed to map how these frequency-band ratios reflect neuropathological deficits in top-down cognitive control, cortical arousal, and emotional regulation across different severity tiers of depression, validating rsEEG as a reliable tool for clinical grading and screening.Methods A total of 154 right-handed participants (57 males, aged (41.02±16.93) years) were enrolled, following the exclusion of 12 subjects from an initial cohort of 166 due to environmental noise or poor electrode contact. Participants were stratified into healthy controls (HC, n=51), mild (n=40), moderate (n=21), and severe depression (n=42) groups based on HAMD, MMSE, and MoCA scales. After resting quietly for 30 min, eyes-closed rsEEG was recorded for 10 min using 19 electrodes positioned according to the international standard 10-20 system (sampling rate: 200 Hz; contact impedance<10 kΩ; bilateral mastoids A1/A2 reference) in a quiet, dimly lit room. Following offline filtering and manual artifact rejection of head movements, EMG, and ECG noises, a continuous, stable 3-min window was extracted for comparative analysis. Relative power spectral densities for the theta (4–7 Hz), alpha (8–12 Hz), and beta (13–30 Hz) bands were calculated to derive TBR and TAR across nine cortical regions (left/right anterior, middle, posterior regions, and midline Fz, Cz, Pz). Data were evaluated using 4 (groups)×9 (brain regions) two-way ANOVA with Bonferroni-adjusted post-hoc comparisons.Results Two-way ANOVA for TBR revealed highly significant main effects of group, brain region, and their interaction (all P<0.001). Abnormal TBR values were primarily observed in the frontal and temporal regions and exhibited significant disease-course-related changes. Specifically, TBR values were significantly higher in the mild and severe depression groups than in the healthy control and moderate depression groups (P<0.01), suggesting that depression-related EEG characteristics may undergo complex and dynamic changes as the disorder progresses, which may be modulated by clinical heterogeneity (such as the presence of comorbid anxiety and recurrent depression in the moderate group). TAR effectively differentiated depressive states and showed significant hemispheric asymmetry specifically prominent in the central (Cz, P<0.001) and posterior (Pz, P=0.019) regions, indicating a central-posterior regional bias in depression-related brain dysfunction. Stability analysis showed that TAR maintained stable between-group characteristics with only 3 min of EEG recording. In contrast, the healthy control""s TBR falsely exhibited a midline-dominant distribution in the 3-min segment, and its characteristic off-axis spatial asymmetry became stable only with a 10-min recording. These findings indicate that TAR has greater sensitivity detection and stronger temporal robustness than TBR.Conclusion TBR and TAR are sensitive indicators for assessing depression severity and associated neurophysiological alterations. Elevated TBR may indicate impaired cognitive control and reduced cortical arousal, whereas the spatial lateralization and nonlinear changes in TAR may reflect specific features of disease progression, potentially corresponding to interhemispheric information processing imbalance. Given its stability in recordings as short as 3 min, TAR may have considerable potential for efficient clinical assessment.

The drug resistance in the treatment of solid tumors is closely related to abnormalities in the tumor vascular microenvironment (TVM). The TVM is dynamically composed of multiple stromal, vascular, and immune cell populations together with extracellular matrix. Its structural disorder and functional imbalance can lead to abnormal angiogenesis, perfusion disorders, limited drug delivery, and insufficient immune cell infiltration, thereby creating a microenvironment conducive to tumor progression and treatment resistance. Therefore, the coordinated regulation of TVM through vascular normalization and immune remodeling has become an important research direction for improving the treatment effect of solid tumors. Tumor necrosis factor superfamily member 14 (TNFSF14), also known as LIGHT (homologous to lymphotoxin, inducible, competes with herpes simplex virus (HSV) glycoprotein D for herpesvirus entry mediator (HVEM), expressed by T lymphocyte), is an important molecule connecting tumor vascular remodeling and immune regulation. LIGHT can bind to the lymphotoxin β receptor (LTβR) to activate the NF-κB (nuclear factor kappa B)-related signaling pathway, regulate the functions of vascular endothelial cells and stromal cells, and induce the expression of adhesion molecules and chemokines such as VCAM-1(vascular cell adhesion molecule 1), ICAM-1(intercellular adhesion molecule 1), CCL19 (C-C motif chemokine ligand 19), CCL21 (C-C motif chemokine ligand 21), and CXCL13 (C-X-C motif chemokine ligand 13), promoting the formation of high endothelial venules (HEVs). HEVs can improve the vascular access of immune cells into tumor tissues and further promote the formation of tertiary lymphoid structures (TLS), providing a structural basis for the recruitment of effector T cells and sustained anti-tumor immune responses. At the same time, LIGHT interacts with HVEM to provide costimulatory signals to T cells, enhancing their activation, proliferation, and cytotoxic function. Thus, LIGHT can link tumor vascular normalization, immune cell recruitment, and T cell activation through the dual actions mediated by LTβR and HVEM, forming a coordinated regulation model of “vascular remodeling-immune infiltration-immune activation”, providing new molecular targets for the treatment of solid tumors. Despite LIGHT’s good anti-tumor potential, its therapeutic application is still limited by factors such as insufficient stability in vivo, short half-life, and potential off-target effects due to systemic exposure. In recent years, researchers have developed various strategies for the precise delivery of LIGHT, including fusion proteins, bacterial vectors, viral vectors, nanoparticles, chimeric antigen receptor T cells (CAR-T), and mesenchymal stem cells (MSCs), to enhance the enrichment, expression, or sustained release of LIGHT in tumor sites through different approaches, thereby improving its therapeutic efficacy and safety. This article systematically reviews the structural characteristics, biological functions, and molecular mechanisms mediating TVM remodeling by LIGHT, with a focus on the signal transduction mediated by LTβR and HVEM and their roles in tumor immune regulation. Subsequently, it systematically outlines the molecular mechanisms by which LIGHT regulates the tumor vascular microenvironment and its intrinsic connection with immune cell recruitment and activation, from the perspectives of TVM abnormalities, vascular normalization, HEV formation, and TLS induction. Furthermore, it categorizes and summarizes precise delivery strategies such as LIGHT fusion proteins, bacterial vectors, viral vectors, nanoparticles, CAR-T cells, and MSCs according to delivery vectors and modes of action, and compares their functional characteristics, antitumor effects, and potential advantages and limitations. On this basis, the article further discusses the key issues that LIGHT therapeutic strategies still face in terms of targeting, controllability, in vivo stability, safety, and clinical translation, and anticipates the potential value of its combination with immune checkpoint inhibitors, anti-angiogenic therapy, and other tumor treatment methods. This article aims to deepen the understanding of the antitumor mechanism of LIGHT from the perspective of synergistic regulation of tumor vascular microenvironment remodeling and immune activation, providing theoretical basis and new research ideas for the design of solid tumor treatment strategies based on precise regulation of TVM and the clinical translation of LIGHT-related treatment systems.

Osteoarthritis (OA) is a common chronic degenerative joint disease and a major cause of pain and disability, particularly in middle-aged and older adults. OA is now recognized as a whole-joint disorder involving articular cartilage, synovium, subchondral bone, the meniscus, the infrapatellar fat pad, and the local inflammatory microenvironment. In addition to mechanical factors, metabolic abnormalities such as obesity, dyslipidemia, insulin resistance, and hyperglycemia are increasingly recognized as contributors to OA onset and progression. Obesity can increase mechanical joint loading, whereas metabolic abnormalities alter cellular metabolism and inflammatory responses within joint tissues. However, the molecular mechanisms by which metabolic stress leads to persistent cellular dysfunction and tissue damage remain incompletely understood. Protein acylation links cellular metabolism to chromatin regulation and protein function. Its dynamic balance is influenced by acyl-donor availability, cellular redox status, mitochondrial activity, glycolytic flux, and the activities of acyltransferases and deacylases. In OA, mitochondrial dysfunction, enhanced glycolysis, lactate accumulation, altered acyl-CoA metabolism, and dysregulated NAD⁺-dependent deacylation may reshape histone and non-histone protein acylation. These changes have been implicated in inflammatory gene expression, extracellular matrix metabolism, mitochondrial quality control, chondrocyte senescence, ferroptosis susceptibility, and synovial immune dysfunction. Recent studies increasingly focus on defined acylation sites, target proteins, regulatory enzymes, and cell-type-specific effects. Acetylation and deacetylation are involved in OA-related processes. sirtuin 1 (SIRT1)-mediated regulation of nuclear factor kappa-B (NF-κB) p65 acetylation is associated with inflammatory and catabolic responses, whereas histone deacetylase 3 (HDAC3)-dependent control of Parkin acetylation links matrix stiffening to excessive mitophagy and chondrocyte senescence. SIRT3-related pathways connect nicotinamide adenine dinucleotide (NAD⁺)-dependent deacetylation with mitochondrial DNA integrity, mitochondrial renewal, and metabolic remodeling, while SIRT6 has been implicated in inflammation and cellular senescence. Lactylation has emerged as another important direction in OA research. histone H3 lysine 18 (H3K18) lactylation is associated with glycolysis-related transcription through the lactate dehydrogenase A (LDHA)-H3K18la-triosephosphate isomerase 1 (TPI1) axis. UGDH K6 lactylation impairs glycosaminoglycan synthesis and promotes matrix damage, whereas H4K12 lactylation has been linked to chondrocyte senescence and autophagy inhibition. In the synovial microenvironment, cluster of differentiation 11b (CD11b) lactylation in macrophages impairs efferocytosis and contributes to sustained inflammation under hyperglycemic conditions. In contrast, H3K56 lactylation has been associated with increased type II collagen synthesis in a post-traumatic OA model, indicating that lactylation effects depend on the modified site, target, cell state, and disease context. SIRT5-related malonylation and succinylation provide additional links between acylation and metabolic dysfunction in OA. Increased protein malonylation and reduced SIRT5 activity are associated with disturbed glycolysis, mitochondrial respiration, and carbon metabolism in chondrocytes. SIRT5-mediated acyl-CoA synthetase long-chain family member 4 (ACSL4) desuccinylation has been linked to reduced lipid peroxidation and ferroptosis susceptibility. By contrast, OA-specific evidence for crotonylation remains limited and centers on crotonylation-related metabolic genes such as glutaryl-CoA dehydrogenase (GCDH) rather than defined functional acylation sites. This review summarizes current evidence linking metabolic dysregulation to acylation regulation in OA. Collectively, available studies indicate that shared metabolic disturbances reshape distinct acylation programs across joint cell types, producing modification- and site-specific effects on cellular function and tissue homeostasis. However, evidence strength varies among acylation types, particularly regarding defined functional sites and in vivo causality. Further site-specific and in vivo studies are needed to clarify which acylation changes actively contribute to OA progression and to evaluate their mechanistic and therapeutic relevance.

Objective DNA methylation and histone modifications jointly regulate gene expression. In this study, co-regulation refers to their combined contribution to gene expression changes. We analyzed multi-omics data from HepG2 cells and in vitro differentiated hepatocytes. This study aimed to identify key DNA methylation sites and histone modifications. A further aim was to construct a co-regulation index to quantify their combined contribution on gene expression.Methods We collected ChIP-seq data for 11 histone modifications, whole-genome bisulfite sequencing data, and poly(A)+ RNA-seq data from the ENCODE database. Differentially expressed genes and differentially methylated CpG sites were identified by using DESeq2 and methylKit. The promoter regions were defined as 2 000 bp upstream and 2 000 bp downstream of the transcription start site. It was divided into 80 bins of 50 bp. Bins 3 to 12 flanking transcription start site were selected as candidate bins. The methylation differences between HepG2 cells and in vitro differentiated hepatocytes in these bins were evaluated with a Wilcoxon rank-sum test. Hypomethylated sites in up-regulated genes and hypermethylated sites in down-regulated genes were retained as key sites. A 1 000 bp region centered on each key site was divided into five 200 bp segments. Histone modification signals were calculated for each segment. DNA methylation differences and histone modification signal differences were used as features in the XGBoost classifier. We evaluated all 2 047 combinations of the 11 histone modifications and DNA methylation by five-fold cross-validation. We also constructed a co-regulation index. This index combined the change in each epigenetic feature with the direction and strength of its correlation with gene expression. The index values were converted to Z-scores. Regions with Z-scores above 0 were selected for further analysis. The stability of this classification was assessed by 1 000 bootstrap iterations. We then used RNA-seq and clinical data from patients with hepatocellular carcinoma in TCGA to analyze genes whose promoters contained these regions. Univariate Cox regression, LASSO regression, and multivariable Cox regression were used to identify survival-associated candidate genes.Results We identified 4 676 differentially expressed genes. Among them, 1 537 genes were up-regulated and 3 139 genes were down-regulated in HepG2 cells relative to in vitro differentiated hepatocytes. The selected promoter bins had higher mean absolute methylation differences than the remaining bins in the up-regulated gene group (p=1.49×10^-4). We identified 1 182 promoter hypomethylation sites in 495 up-regulated genes and 1 075 promoter hypermethylation sites in 1 083 down-regulated genes. Among the 11 histone modifications, the models combining DNA methylation with H3K27ac, H3K9ac, or H3K4me3 showed the highest predictive performance. A model containing these three histone modifications and DNA methylation achieved the AUC of 0.900 9. After the co-regulation index was added to the prediction model, the AUC increased from 0.900 9 to 0.928 7. A total of 882 regions with Z-scores above 0 were identified in the promoters of 322 genes. Bootstrap analysis showed a mean classification consistency of 99.38%. These genes were enriched in signal regulation, cell adhesion, tissue development, the PI3K-Akt signaling pathway, and ECM-receptor interaction. Among the 322 genes, 43 genes were differentially expressed in TCGA hepatocellular carcinoma samples. Seven genes were associated with survival in univariate Cox analysis. FBXW10 and B3GALT2 were retained after LASSO and multivariable Cox analyses. The two-gene risk model achieved AUC values of 0.759, 0.834, and 0.914 for 1-, 3-, and 5-year survival, respectively.Conclusion H3K4me3, H3K9ac, and H3K27ac showed the strongest co-regulation with DNA methylation in the hepatic cell models. The co-regulation index quantified the combined contribution of DNA methylation and histone modifications to gene expression changes. The incorporation of the co-regulation index increased the model AUC from 0.900 9 to 0.928 7. The index was used to identify key co-regulated regions. Further analysis with TCGA data identified FBXW10 and B3GALT2 as survival-associated candidate genes. These findings will be helpful for further analyzing and understanding the co-regulation of DNA methylation and histone modifications in cancer research.

Objective To design a wearable multi-frequency electrical impedance tomography (mfEIT) system for regional lung ventilation analysis and to preliminarily evaluate its feasibility for acquiring regional relative ventilation information in a lung ventilation model and under wearable conditions.Methods A 16-electrode wearable mfEIT system was developed based on a Cortex-M7 microcontroller. Multi-frequency sinusoidal excitation signals were generated using a wideband mirror constant current source, and boundary voltages were acquired through a high-speed differential acquisition circuit. The voltage signals were preprocessed using target-bin discrete Fourier transform (DFT) demodulation and fourth-order infinite impulse response (IIR) low-pass filtering and were then transmitted to a personal computer (PC) via Bluetooth. EIT image reconstruction was performed using EIDORS. Relative ventilation signals were further extracted from the region of interest (ROI), and relative flow-volume loops were constructed to describe dynamic changes in regional lung ventilation. A lung ventilation model experiment was conducted to assess the signal acquisition performance, normalized lung volume parameter extraction, and the identification of uneven ventilation between the left and right lungs. Both matched ventilation and left–right uneven ventilation conditions were established to examine whether the reconstructed images and regional signals could reflect controlled differences in ventilation distribution. A continuous acquisition demonstration under wearable conditions was further performed by a single healthy participant to observe the ability of the system to continuously record relative ventilation changes in a practical wearable setting.Results The system achieved multi-frequency excitation over a range of 1 kHz to 500 kHz. The maximum aggregate acquisition rate of complete boundary-voltage frames was 40 fps, with each complete frame comprising 208 valid voltage measurements acquired at one excitation frequency. Under four-frequency frame-by-frame scanning, the effective acquisition rate was approximately 10 fps for each excitation frequency. Under a 1 mA excitation current, the average signal-to-noise ratio (SNR) remained above 42 dB within the 500 kHz excitation bandwidth, and reached 61.4 dB at 180 kHz. The total voltage fluctuation during continuous acquisition was less than 1%. These results indicated that the system maintained stable boundary-voltage acquisition performance over a wide excitation-frequency range and during continuous operation. In the lung ventilation model experiment, the errors of normalized tidal volume, normalized inspiratory reserve volume, and normalized expiratory reserve volume extracted by mfEIT were 5.24%, 10.31%, and 9.29%, respectively, compared with the bidirectional flowmeter measurements. In the uneven ventilation between the left and right lungs, the relative vital capacity of the right lung measured by EIT was 54.76% of that of the left lung, with an error of 3.53% compared with the flowmeter result. The single-participant wearable demonstration showed that ventilation-related signals could be continuously acquired at the different excitation frequencies and that changes in relative tidal volume, respiratory cycle, and relative flow-volume loop morphology could be tracked during sitting, light exercise, and standing. These observations were descriptive and did not establish diagnostic accuracy or absolute pulmonary-function quantification.Conclusion The developed wearable mfEIT system integrates wideband excitation, rapid boundary-voltage acquisition, embedded signal processing, wireless transmission, EIT reconstruction, and ROI-based functional analysis within a compact platform. Its main contribution is the establishment of an end-to-end pathway that converts multi-frequency electrical measurements into interpretable regional relative ventilation information, enabling ventilation distribution and temporal dynamics to be considered within a unified framework. The model results support the identification of regional ventilation differences, while the single-participant experiment demonstrates the feasibility of continuous wearable acquisition. The system therefore provides a technical basis for continuous regional relative ventilation monitoring. Further comparisons with standard pulmonary-function measurements and reference EIT systems in larger populations are required before clinical assessment or absolute regional pulmonary-function quantification can be considered.

β-arrestin serves as a core functional effector governing signal transduction of G protein-coupled receptors (GPCRs), exerting dual roles as a signal terminator and molecular scaffold. However, the molecular mechanism underlying how β-arrestin responds to diverse GPCRs and interacts with hundreds of effector proteins to mediate a broad spectrum of physiological functions remains poorly understood. A landmark original study recently published in Nature overturned the conventional view of β-arrestin as a static scaffold protein. Endogenously expressed β-arrestin undergoes liquid-liquid phase separation (LLPS) to form highly dynamic biomolecular condensates, which orchestrate compartmentalized assembly and fine-tuning of GPCR signaling complexes. This work establishes a novel paradigm centered on β-arrestin condensates for GPCR signal regulation and reshapes the molecular framework of hierarchically organized intracellular signaling networks.

Objective Conventional pulmonary function testing remains the primary clinical approach for evaluating respiratory dysfunction, but its reliability is strongly dependent on subject cooperation and test execution. In addition, global spirometric indices provide limited information on regional ventilation heterogeneity, which is important for understanding the spatial characteristics of pulmonary functional abnormalities. To address these limitations, this study proposed a dual-branch fusion method based on electrical impedance tomography (EIT) for the identification of pulmonary function abnormalities and evaluated its feasibility in screening and auxiliary classification tasks.Methods A total of 507 valid EIT-pulmonary function examination samples were included, comprising 244 samples with normal pulmonary function and 263 samples with abnormal pulmonary function. During forced expiration, thoracic EIT boundary voltage signals were collected, and corresponding conductivity image sequences were reconstructed to represent both respiratory dynamic changes and intrapulmonary ventilation distribution. For each valid examination sample, 64 consecutive frames from the expiratory phase were selected as the model input. A dual-branch architecture was then constructed, in which one branch used boundary voltage signals to characterize temporal respiratory dynamics, while the other branch used reconstructed conductivity images to describe spatial ventilation patterns. Residual network (ResNet) models with different depths, including ResNet18, ResNet50, and ResNet101, were employed for feature extraction. The extracted voltage and image features were concatenated to generate a unified spatiotemporal representation. Several machine learning classifiers, including support vector machine, k-nearest neighbor, decision tree, linear discriminant analysis, and Stacking ensemble learning, were further applied for classification. Five-fold cross-validation was used to evaluate model performance in binary classification between normal and abnormal pulmonary function, three-class classification among normal, obstructive, and non-obstructive pulmonary function patterns, and four-class classification involving multiple ventilatory dysfunction types. Accuracy and macro-averaged F1-score were used as the main evaluation metrics. Feature extraction, feature standardization, classifier training, and model evaluation were independently completed within each fold to reduce the risk of information leakage.Results On average, the fused voltage-conductivity channel outperformed the single voltage and single conductivity-image channels. The average accuracy and F1-score of the fusion channel reached 0.835 ± 0.053 and 0.765 ± 0.073, respectively, representing improvements of 6.5% and 9.7% compared with the voltage-only channel. Among the different network depths, ResNet18 achieved the best overall performance, suggesting that a relatively shallow residual network may be more suitable for learning EIT-based spatiotemporal features, probably because the key pathological information in EIT is mainly reflected by low-frequency temporal changes and large-scale regional ventilation differences rather than fine image textures. Among the classifiers, the Stacking strategy achieved the highest comprehensive performance, with an accuracy of 0.830 8 ± 0.059 9 and an F1-score of 0.821 9 ± 0.060 7, indicating that ensemble learning can further improve the robustness of multi-class recognition. The overall advantage of the fusion channel also supports the complementary value of temporal boundary-voltage dynamics and spatial conductivity-distribution information in characterizing pulmonary function abnormalities.Conclusion These findings demonstrate that dual-branch fusion of EIT voltage signals and conductivity images can effectively characterize spatiotemporal features associated with abnormal pulmonary ventilation. By integrating temporal respiratory dynamics with regional ventilation distribution, the proposed framework reduces dependence on a single predefined EIT index and provides a more comprehensive representation of pulmonary function abnormalities. The proposed method offers a non-invasive and radiation-free approach for pulmonary function abnormality screening, auxiliary classification, and bedside dynamic assessment, and may support the future development of wearable or portable EIT-based respiratory monitoring systems.

Objective Community-acquired pneumonia (CAP) is a common infectious disease encountered in emergency departments, and timely severity stratification is essential for early risk identification and individualized treatment. Chest CT can accurately evaluate the extent of pulmonary involvement; however, its use for repeated bedside assessment is limited by radiation exposure, cost, limited portability, and the need for patient transport. In contrast, EIT is a radiation-free, noninvasive, and bedside functional imaging modality that can continuously monitor regional ventilation. Nevertheless, its relatively low spatial resolution limits its ability to directly characterize the anatomical extent of pulmonary lesions. Therefore, this study constructed a CT shape-prior-guided 3D CT-EIT composite image and explored its value for CAP severity assessment.Methods A total of 60 adult patients with CAP who were admitted to the Emergency Department of the Third Hospital of Hebei Medical University between October 2023 and October 2024 were retrospectively included. According to a five-lobe CT score ranging from 0 to 25, the patients were divided into a light pneumonia (LP) group and a moderate-to-severe pneumonia (MSP) group. Patients with CT scores of 0-8 were classified into the LP group (n=29), whereas those with CT scores of 9-25 were classified into the MSP group (n=31). Clinical indices, including oxygenation index (PaO₂/FiO₂), fibrinogen, and pneumonia severity index (PSI), were collected. The 3D EIT composite image based on the individual CT images of each patient was constructed, and the percentage of ventilation blocked regions (PVBR), ventilation distribution center error (VDCE), and global ventilation delay index (GVDI) were established. Receiver operating characteristic (ROC) curves were plotted to evaluate the discriminative performance, and Pearson correlation coefficient (r) was used to assess the correlation between 3D EIT ventilation indices and CT score.Results There were no statistically significant differences in baseline characteristics between the two groups (P>0.05). Compared with the LP group, the MSP group showed lower PaO₂/FiO₂ (166 mmHg vs. 298 mmHg, P<0.01), higher fibrinogen (4.59 g/L vs. 3.90 g/L, P<0.01), higher PSI (114.23 vs. 94.52, P<0.01), higher PVBR (39.33 vs. 26.33, P<0.001), higher VDCE (42.21 vs. 35.33, P<0.05), and higher GVDI (0.48 vs. 0.38, P<0.001). Among the conventional clinical indices, PaO₂/FiO₂ showed the highest area under the ROC curve (AUC=0.723). Among the 3D EIT ventilation indices, PVBR showed the best performance, with an AUC of 0.775 and a specificity of 96.6%. CT score was negatively correlated with PaO₂/FiO₂, but positively correlated with fibrinogen, PSI, and 3D EIT ventilation indices. Among these, CT score showed the strongest correlation with PVBR (r=0.696).Conclusion 3D CT-EIT composite imaging is feasible for rapid bedside functional assessment of CAP severity. PVBR, VDCE, and GVDI provide complementary information on the extent of ventilation blockage, the spatial redistribution of ventilation, and the temporal delay in ventilation, respectively. Among these indices, PVBR showed the best discriminatory performance and the strongest correlation with CT-defined lesion severity. Although 3D EIT cannot replace CT for structural diagnosis, it may serve as a practical radiation-free functional complement for bedside severity stratification, serial ventilation monitoring, and early clinical decision-making in patients with CAP.

Neurodegenerative diseases (NDs), particularly Alzheimer’s disease (AD) and Parkinson’s disease (PD), and cardiovascular diseases (CVDs), including heart failure (HF) and coronary artery disease (CAD), are major chronic disorders associated with population aging. Their frequent coexistence in clinical practice suggests that they may not progress independently, but may instead be linked through bidirectional pathological communication within the brain-heart axis. Epidemiological evidence underscores this clinical burden: systematic reviews and meta-analyses have reported pooled relative risks for dementia of approximately 1.27 among individuals with a history of CAD and 1.60 among those with HF, while the pooled prevalence of cognitive impairment and dementia in patients with HF has been estimated at approximately 41.42% and 19.79%, respectively. This review synthesizes current evidence on the mechanistic links, disease-state manifestations, biomarkers, and potential therapeutic strategies involved in the comorbidity between NDs and CVDs from the perspective of the brain-heart axis. The brain and heart communicate through autonomic, neuroendocrine, neuroimmune, vascular, and humoral pathways. Sympathetic activation, parasympathetic withdrawal, vagal dysfunction, cerebral hypoperfusion, endothelial injury, blood-brain barrier (BBB) alterations, peripheral immune activation, and circulating inflammatory mediators may jointly contribute to cross-organ pathological signaling. Oxidative stress and inflammatory response are not specific to the brain-heart axis; rather, they may act as candidate amplifying processes within this network. Mitochondrial dysfunction and NADPH oxidase activation can increase the generation of reactive oxygen species and reactive nitrogen species, which may activate nuclear factor-κB, the NLRP3 inflammasome, and mitogen-activated protein kinase/c-Jun N-terminal kinase signaling. These pathways can promote the release of interleukin-1β, interleukin-6, tumor necrosis factor-α, and other inflammatory mediators, which may in turn aggravate mitochondrial injury, endothelial dysfunction, immune cell activation, and further oxidant generation. Cytokines, chemokines, oxidized lipids, mitochondrial DNA, damage-associated molecular patterns, and extracellular vesicles may transmit these signals between the cardiovascular and central nervous systems, whereas autonomic reflexes provide an additional route for bidirectional amplification. Distinct disease states illustrate different manifestations of this network. The association between AD and CVDs may involve vascular risk factors, impaired amyloid-β clearance, abnormal cholesterol homeostasis, cerebral microvascular injury, neuroinflammation, and endothelial dysfunction. HF-related cognitive impairment may be linked to reduced cardiac output, recurrent cerebral hypoperfusion, hypoxia-related signaling, BBB dysfunction, circulating inflammation, and impaired autonomic feedback. In PD, cardiovascular autonomic dysfunction and abnormalities in heart rate and blood pressure regulation may interact with vascular injury, reduced cerebral perfusion, inflammation, and oxidative stress. However, the relationship between PD and CAD remains heterogeneous and may be modified by age, ethnicity, medication use, metabolic factors, and coexisting diseases. Biomarker evaluation should therefore move beyond isolated indicators. Traditional markers, including malondialdehyde, superoxide dismutase, glutathione, C-reactive protein, and interleukin-6, reflect systemic oxidative or inflammatory burden but lack sufficient specificity to distinguish primary neurodegeneration, cardiovascular injury, or a comorbid state involving both. Greater clinical value may be obtained from cross-system panels that combine oxidative stress and inflammatory markers with cardiac stress or injury biomarkers, such as N-terminal pro-B-type natriuretic peptide; glial or neuronal biomarkers, such as glial fibrillary acidic protein and neurofilament light chain; BBB- or endothelial injury-related biomarkers; and omics-derived candidates, including microRNAs, long non-coding RNAs, DNA methylation signatures, metabolites, and proteins. Multi-omics integration and artificial intelligence-assisted analysis may support risk stratification, disease monitoring, and prediction of treatment response. However, these approaches require standardized analytical platforms, validation in independent cohorts, and careful distinction between associative findings and causal mechanisms. Potential interventions include antioxidant, anti-inflammatory, mitochondria-targeted, and multi-target combination strategies, together with emerging approaches such as nanodelivery, CRISPR/Cas9- or CRISPR/dCas9-based regulation, mesenchymal stem cell-mediated mitochondrial transfer, and ergothioneine supplementation. Nevertheless, most available evidence is indirect, preclinical, or derived from other disease indications, and no single drug or technology can currently be regarded as specific to brain-heart axis comorbidity. The clinical relevance of these strategies will depend on whether they can be matched to dominant pathological phenotypes and improve both neurological and cardiovascular outcomes. Overall, comorbidity between NDs and CVDs is more appropriately understood as a heterogeneous cross-organ regulatory network than as a disorder driven by a single pathway. Future research should prioritize multicenter longitudinal cohorts, standardized clinical phenotyping, validation of cross-system biomarker panels, and mechanism-guided interventions in clearly defined patient subgroups, with integrated assessment of cognitive function, cardiovascular events, quality of life, long-term safety, and treatment response.

Meiotic recombination is a core molecular event that generates genetic diversity in sexually reproducing organisms, and is of great significance for both genome evolution and proper chromosome segregation. Meiotic recombination occurs in two forms: crossover (CO) and non-crossover (NCO). Although NCO events are far more abundant than CO events, they have been difficult to detect. Consequently, previous research has largely focused on CO, leaving the characteristics and functional roles of NCO poorly understood. A recent study published in Nature used large-scale family-based whole-genome sequencing data to construct a high-resolution recombination map. This map revealed the mutational features induced by NCO—the dominant form of recombination—along with their sex-specific differences, and quantified, for the first time, the contribution of NCOs to de novo mutations and age-related effects, representing an important milestone in the field of meiotic recombination. Specifically, compared with paternal NCOs, maternal NCOs are fewer in number but span longer regions; they accumulate continuously with increasing maternal age and occur largely outside programmed recombination hotspots. NCOs also represent a major source of de novo mutations, with approximately 11% of maternal de novo mutations being attributable to NCOs. This study not only provides a high-resolution map of human meiotic recombination as a resource for future research, but also offers, for the first time, a genome-wide perspective on the contribution of meiotic NCO recombination to genetic diversity.

<b>Objective</b> Accurate prediction of anticancer drug response plays an important role in precision medicine and drug discovery. However, existing computational approaches often suffer from insufficient integration of heterogeneous biological data and limited generalization to unseen drugs. To address these challenges, this study proposes DeepFusion-CDR, a multimodal deep learning framework that integrates drug-target interaction information with multi-omics characteristics of cancer cell lines to improve the accuracy and generalization of anticancer drug response prediction.<b>Methods</b> A drug-target interaction (DTA) prediction model based on the Kolmogorov-Arnold Network (KAN) was first developed to capture complex nonlinear relationships between drugs and target proteins. The pretrained model was then used to generate drug target activity profiles, which provide biologically meaningful representations of drug mechanisms of action. On the drug side, molecular fingerprints, molecular graph structures, and drug target activity profiles were integrated as complementary multi-view features to comprehensively characterize the chemical and biological properties of drugs. On the cell-line side, multi-omics data, including genomic mutations, gene expression profiles, and DNA methylation information, were incorporated to describe the molecular heterogeneity of cancer cells. Subsequently, a multimodal deep fusion architecture was designed to jointly model the interactions between drug features and cell-line features, enabling accurate prediction of drug responses. The proposed framework was evaluated on the Genomics of Drug Sensitivity in Cancer (GDSC) dataset under different experimental settings, and its robustness and generalization capability were further assessed using the independent Cancer Therapeutics Response Portal version 2 (CTRPv2) dataset.<b>Results</b> Experimental results on the GDSC dataset demonstrated that DeepFusion-CDR consistently outperformed several state-of-the-art baseline methods and achieved superior predictive performance across multiple evaluation scenarios. The model demonstrated strong generalization capability, particularly in challenging settings involving unseen drugs. Validation on the independent CTRPv2 dataset further confirmed the model’s ability to generalize across different datasets. In addition, ablation studies revealed that both drug target activity profiles and multi-omics features made significant contributions to prediction performance. Removing any major feature modality led to a measurable decrease in model accuracy, indicating that different data sources provide complementary information for drug response prediction. These findings highlight the importance of multimodal collaborative learning and demonstrate the effectiveness of incorporating drug-target interaction knowledge into predictive modeling.<b>Conclusion</b> DeepFusion-CDR achieves accurate prediction of anticancer drug responses by integrating drug mechanism-related information with multi-omics characteristics of cancer cell lines. The incorporation of drug target activity profiles provides complementary biological information beyond traditional chemical structure descriptors and enhances the predictive capability of the model for unseen drugs. By combining drug-target interaction knowledge with heterogeneous biological data within a unified deep learning framework, DeepFusion-CDR offers a promising computational approach for precision medicine, drug repurposing, and anticancer drug discovery.

Objective Multi-omics integration and analysis remain a major challenge in biomedical research. These tasks often require extensive coding skills and specialised bioinformatics expertise, which many biological and medical researchers do not have. Deep learning has emerged as a powerful approach for predictive modelling and data-driven discovery. However, most existing platforms focus only on traditional statistical analysis or basic machine learning. They do not combine deep learning support with standard statistical workflows. Moreover, none of them provide intelligent assistance to help with parameter tuning, error diagnosis, or result interpretation. To address this gap, we introduce AutoMATA, a fully code-free platform that streamlines multi-omics analysis from expression data to predictive modelling. AutoMATA is designed to lower the technical barrier for experimental biologists and clinical researchers who want to use advanced deep learning methods but lack programming experience.Methods AutoMATA's architecture is built around three core modules. The data processing and normalisation module automates essential steps such as gene and protein ID conversion and data normalisation. This module also allows users to integrate multi-omics data. The statistical analysis and visualisation module offers important and commonly used functions including differential expression analysis, principal component analysis, correlation analysis, and pathway enrichment for both GO and KEGG. Users can adjust thresholds and choose output formats for publication. The deep learning module provides twelve neural network architectures covering supervised, unsupervised, and semi-supervised learning. These include Multilayer Perception (MLP), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Transformer, Autoencoder, Variational Autoencoder (VAE), Radial Basis Function Neural Network (RBFNN), Self-Organising Map (SOM), DeepCluster, Pseudo-Labeling, and Ladder Network. Users can customise key settings including epoch, regularisation method, regularisation weight, dropout rate, and feature selection method. Two training strategies (train-validation-test split and stratified K-fold) are available. In addition, two Artificial Intelligence (AI) agents, DeepSeek and Qwen, are integrated into the platform. These agents answer user questions about parameter suggestion, task failure diagnosis, and result interpretations based on the actual task context and platform logic.Results We demonstrate AutoMATA in three case studies using public datasets. First, for colon adenocarcinoma data, AutoMATA reproduces gold-standard differential expression and clustering results. It correctly identifies known upregulated genes such as CXCL3 and CXCL8, and visualises enriched pathways. This confirms that AutoMATA's statistical module works as reliably as existing tools. Second, using multi-omics data for bladder, pancreatic, and stomach cancers, AutoMATA's deep learning module predicts cancer recurrence with good performance. Recurrent neural networks achieve the best results, with accuracy above 79% for stomach cancer. The platform also reproduces known biomarkers such as CDK6 in bladder cancer. Third, for breast cancer subtype classification, AutoMATA achieves 88.2% accuracy using a tuned RNN model. This performance is better than traditional machine learning methods like logistic regression and random forest. The AI agents provide on-demand assistance, helping users understand why certain models perform better, suggest parameter adjustments, and explain the biological meaning of the output. This makes the platform especially useful for non-experts.Conclusion By offering advanced multi-omics integration, comprehensive statistical analysis, flexible deep learning modelling options, and built-in AI agents, AutoMATA empowers researchers to extract meaningful biological insights and build robust predictive models without writing any code. The platform and source code are freely accessible athttps://github.com/ABILiLab/AutoMATA.

Objective In order to address the challenge of rapid diagnosis in pulmonary diseases, this paper proposes a cross-modal fusion method based on the Cross-modal Transformer (CMT) that integrates respiratory sounds (RS) and electrical impedance tomography (EIT), with the aim of improving the accuracy and robustness of multi-classification tasks. RS encodes the acoustic characteristics of the airways via the Meyrieh spectrogram, whilst EIT depicts regional lung ventilation distribution through spatio-temporal image sequences. The two modalities are naturally complementary in terms of functional and spatial information, providing a physiological basis for multimodal fusion diagnosis.Methods A dual-branch feature extraction framework was constructed, employing Convolutional Neural Networks (CNNs) to extract local features from RS time-frequency spectra and EIT ventilation images, whilst utilising Bidirectional Long Short-Term Memory Networks (BiLSTMs) to model the temporal dependencies across modalities. The development of a transformer-based cross-modal attention fusion module represents a significant advancement in the field. The module utilises a multi-head self-attention mechanism and gated convolutional units to achieve deep semantic alignment and complementary information fusion between RS acoustic features and EIT spatial ventilation features. The model employs an end-to-end joint optimisation strategy, with a cross-entropy loss function supervising the overall training process, and utilises a time-synchronisation mechanism to ensure strict alignment of the dual-modal inputs within the respiratory cycle. The proposed CMT method was subjected to systematic experimental validation on two datasets: The BRACETS dataset (three-class classification, 795 samples) and the CleftPalate dataset (two-class classification, 549 samples) are the focus of this study. Quantitative evaluation was performed using accuracy, balanced accuracy (BAcc) and macroF1 score.Results On the BRACETS dataset, the CMT method achieved an accuracy of 87.21%, a balanced accuracy (BAcc) of 88.24%, and a macroF1 score of 87.40%. In comparison to the optimal baseline method, DCNN, the macroF1 score exhibited an enhancement of 8.73 percentage points, thereby substantiating a substantial performance superiority. The findings of the ablation experiments suggest that the three core modules – CNN, BiLSTM and Transformer – all contribute to performance enhancements. Upon the removal of these modules, the MacroF1 score decreased by 1.84%, 1.13% and 2.20%, respectively. Among these, the Transformer cross-modal fusion module had the most significant impact, validating its crucial role in the interaction of heterogeneous modal information. Hyperparameter sensitivity analysis indicates that the optimal parameter configuration is a sequence length of T=128 and a feature dimension of d=128, achieving a good balance between classification performance and computational efficiency. Sequences that are insufficiently extensive or dimensions that are unduly limited impede the capacity to adequately represent temporal dynamics, whilst excessively protracted sequences may engender superfluous information. On the CleftPalate dataset, the CMT method achieved an accuracy of 96.58%, a BAcc of 96.60%, and a MacroF1 of 96.56%, representing a further improvement of 2.51 percentage points compared to the DCNN. This result validates the model's generalisation capability across different data distributions and task scenarios. The fusion representation learned by CMT has been shown to exhibit tighter intra-class cohesion and clearer inter-class separation boundaries, as evidenced by feature visualisation and case studies (Smith et al., 2022). The system has the capacity to adaptively aggregate effective features when the reliability of multimodal information is uneven, and can maintain correct classification even when ambiguity exists in a single modality.Conclusion The proposed method effectively achieves deep fusion of heterogeneous modal information from RS and EIT, fully exploiting the physiological complementary relationship between respiratory airflow acoustic features and regional lung ventilation distribution. The model displays excellent classification performance and stability across various data distributions and task scenarios, thus providing a novel technical pathway for the non-invasive intelligent diagnosis of pulmonary diseases.

Objective Frozen shoulder (FS) is a debilitating musculoskeletal disorder characterized by persistent inflammation and progressive fibrosis of the glenohumeral joint capsule, leading to pain and severely restricted range of motion. Given the limited efficacy of current pharmacotherapies, there is an urgent need for novel anti-fibrotic agents. Guyanxiao Tincture (GYX), a traditional Chinese herbal formula, has shown clinical benefits in alleviating FS symptoms. However, its bioactive constituents and the molecular mechanisms underlying its therapeutic effects remain poorly defined. This study aimed to systematically evaluate the therapeutic efficacy of GYX against FS and to elucidate the mechanisms by which GYX attenuates capsular fibrosis.Methods A Sprague-Dawley (SD) rat model of FS was established by immobilizing the unilateral shoulder with a plaster cast for 21 d. Following model induction, animals were administered with GYX or its serum-identified bioactive component, icariside F2. The absorbed prototype compounds of GYX in FS rat serum were profiled using ultraperformance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-Q-TOF-MS). Therapeutic effects were assessed by measuring shoulder range of motion (ROM), histopathological evaluation of the capsule via hematoxylin-eosin staining, and quantification of transforming growth factor-β1 (TGF-β1) content by ELISA. To explore the underlying mechanisms, transcriptomic profiling of the shoulder capsule was performed using RNA sequencing, and differentially expressed genes were validated by real-time quantitative PCR. Protein expression and localization were examined through immunohistochemistry, immunofluorescence, and Western blot. Protein-protein interaction networks were constructed to identify key regulatory hubs. In vitro, primary shoulder capsule fibroblasts were stimulated with TGF-β1 to induce a fibrotic phenotype and then treated with icariside F2. The expression of alpha cardiac muscle 1 (ACTC1), α-smooth muscle actin, and collagen I was subsequently measured.Results Serum pharmacochemistry analysis identified icariside F2 as the predominant circulating bioactive component of GYX in FS rats. Both GYX and icariside F2 treatment significantly improved the ROM and ameliorated histopathological lesions, including the attenuation of synovial hyperplasia, inflammatory infiltration, and excessive collagen deposition in the shoulder capsule. Consistently, elevated TGF-β1 levels in the model group were markedly reduced after the interventions. RNA sequencing revealed that actin ACTC1 was dramatically upregulated in the fibrotic capsule, accompanied by the enrichment of fibrosis-associated pathways such as TGF-β signaling, focal adhesion, and ECM-receptor interaction. Protein-protein interaction network analysis demonstrated that ACTC1 directly interacts with core fibrotic genes, including Col1a1, Col1a2, Thbs2, and Fn1. Critically, icariside F2 administration significantly reversed the aberrant overexpression of ACTC1 and suppressed the activation of the aforementioned fibrosis-related pathways in vivo. In TGF-β1-stimulated fibroblasts, icariside F2 dose-dependently downregulated ACTC1 at both the mRNA and protein levels and concurrently decreased the expression of the fibrotic markers α-smooth muscle actin and collagen I. Immunofluorescence and Western blot further confirmed that icariside F2 attenuated stress fiber formation and collagen production in vitro.Conclusion GYX and its bioactive component icariside F2 effectively alleviate the progression of frozen shoulder by inhibiting capsular fibrosis. Mechanistically, the therapeutic action is functionally linked to the suppression of ACTC1 and its downstream profibrotic signaling cascade, highlighting ACTC1 as a promising therapeutic target for the management of frozen shoulder.

Small-molecule therapeutics and chemical probes remain indispensable in modern biomedical research and drug discovery. However, with the rapid expansion of chemical space and the increasing diversification of biological target classes, high-throughput and material-efficient screening technologies are facing growing demands. Small-molecule microarrays (SMMs) provide a miniaturized and spatially addressable platform in which thousands to tens of thousands of compounds are immobilized on a solid surface and screened in parallel against proteins, cell lysates, or nucleic acid structures. Since the last comprehensive review of this field in 2014, SMM technology has undergone substantial methodological and application-oriented development; however, these advances have often been reported in a fragmented manner and still require systematic integration. This review therefore clarifies the terminology and scope of SMM and systematically summarizes its recent advances, with particular attention paid to four interconnected dimensions: surface chemistry, interface microenvironment modulation, signal detection, and applications. Regarding surface chemistry, immobilization methods are organized into tag-based and broad-spectrum strategies. Tag-based strategies use reactive handles or affinity tags, such as covalent tags, biotin, or fluorous tags, to achieve defined attachment and controllable molecular display, but require prior modification and may mask key pharmacophores. Broad-spectrum strategies exploit intrinsic functional groups or physicochemical properties, including isocyanate coupling, ultraviolet-activated photo-capture, and polymer-based immobilization. They show broader compatibility with diverse libraries, natural products, and approved drugs, but may increase ligand heterogeneity and nonspecific background. This classification provides references for selecting suitable surface construction strategies in SMM studies. Interface microenvironment modulation is another factor affecting the SMM performance. Because immobilized small molecules are displayed on solid substrates, steric hindrance and restricted conformational freedom may reduce target accessibility. Flexible linkers, polyethylene glycol spacers, oligonucleotide tethers, and quasi-three-dimensional polymer layers have therefore been introduced to increase ligand-substrate distance, alleviate steric constraints, and preserve solution-like binding behavior. This review also summarizes signal detection strategies, including fluorescence-based labeling, HaloTag-assisted readouts, and label-free technologies such as oblique-incidence reflectivity difference and surface plasmon resonance imaging. In applications, SMM has expanded from purified protein screening to more diverse biological contexts. Cell lysate-based screening enables interrogation of unstable, difficult-to-purify, or context-dependent targets, whereas structure-oriented screening against RNA/DNA motifs has extended SMM into nucleic acid-targeted discovery. These advances allow SMM to address challenging target classes, including transcription factors, intrinsically disordered proteins, membrane-associated proteins, and protein-protein interactions, targeted protein degradation systems, and higher-order nucleic acid structures. Compared with activity-based high-throughput screening or DNA-encoded library selection, SMM provides a direct, amplification-free binding readout and can capture weak interactions, although rigorous validation remains essential. Finally, this review discusses SMM in fragment-based drug discovery (FBDD) and artificial intelligence-assisted drug design (AIDD). In FBDD, SMM offers a parallel, low-consumption format for detecting weak fragments–target interactions and identifying fragment hits for validation and optimization. In AIDD, SMM can generate binding fingerprints, including fluorescence intensities, signal-to-noise ratios, Z-scores, and apparent affinity parameters in concentration-gradient designs. These datasets may support virtual screening, hit prioritization, binding landscape construction, functional group clustering, and structure–activity relationship inference. Overall, SMM has evolved into a versatile screening and data-generation platform, providing a methodological engine for expanding the druggable target space and accelerating early-stage discovery of chemical probes and lead compounds.

Adipose tissue macrophages (ATMs) are crucial immunomodulatory factors in the adipose tissue (AT) microenvironment, playing an irreplaceable role in maintaining the balance of the local immune system and regulating metabolic homeostasis. Under obese conditions, the excessive accumulation of lipids leads to abnormal expansion of adipose tissue, which further disrupts the homeostasis of the local microenvironment, including the imbalance of inflammatory factors, the occurrence of oxidative stress, and the damage of microcirculation. As an essential immune cell population in adipose tissue, ATMs are deeply involved in the occurrence and progression of metabolic disorders and multiple obesity-related diseases, and their functional abnormalities are closely related to the initiation and development of adipose tissue inflammation and systemic metabolic disorders. The interaction between ATMs and adipokines, including adiponectin, leptin, resistin and retinol-binding protein 4 (RBP4), acts as the core immunological and molecular mechanism linking adipose tissue inflammation to continuous disease deterioration, and this interaction is also a key research focus in the field of obesity-related diseases in recent years. This review discusses in detail the facilitating effect of the AT inflammatory microenvironment on the differentiation of peripheral monocytes into ATMs, elaborating on the specific molecular mechanisms by which various inflammatory mediators and abnormal metabolic products in the inflammatory microenvironment induce the differentiation of peripheral monocytes into functional ATMs. It also focuses on the regulatory mechanisms of chemokine-mediated ATM polarization and recruitment, including the specific roles of key chemokines like monocyte chemoattractant protein-1 (MCP-1) and C-X-C motif chemokine ligand 10 (CXCL10) in mediating the recruitment of ATMs to adipose tissue, as well as the molecular pathways that regulate the switch between M1 pro-inflammatory phenotype and M2 anti-inflammatory phenotype of ATMs. In addition, this review explores the specific process by which ATMs trigger AT inflammation by secreting various pro-inflammatory factors such as tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and IL-1β, which further amplify the local inflammatory response and disrupt the metabolic homeostasis of adipose tissue. Furthermore, it analyzes the specific binding patterns and molecular characteristics of adiponectin, leptin, resistin and RBP4 with ATM surface receptors, and clarifies how the activation of downstream immune signaling pathways, such as Janus kinase-signal transducer and activator of transcription (JAK-STAT), nuclear factor-κB (NF-κB), and phosphatidylinositol 3-kinase-protein kinase B (PI3K-AKT), triggered by these binding processes, induces the occurrence and development of various obesity-related diseases, including insulin resistance, cancer, metabolic-associated steatohepatitis (MASH) and cardiovascular diseases. This paper highlights the unique and crucial role of the interaction between ATMs and adipokines in the occurrence and progression of obesity-related diseases, emphasizing that this interaction is a key link connecting local adipose tissue inflammation to systemic metabolic disorders. It further summarizes the potential therapeutic strategies targeting the ATM-adipokine axis, and comprehensively illustrates three major clinical application barriers, namely biological complexity, technical bottlenecks, and clinical translation obstacles, in the process of applying these therapeutic strategies. For each barrier, this review puts forward corresponding feasible solutions and clear research directions, which not only enrich the theoretical system of the ATM-adipokine axis in the field of obesity research, but also provide a solid theoretical basis and practical ideas for the clinical intervention, prevention and treatment of obesity and its related diseases.