Constrained combinatorial optimization (CCO) problems are prevalent across various fields and represent key challenges in computational science and engineering. Although numerous classical and quantum algorithms have been proposed to tackle these problems, substantial limitations still persist. Classical algorithms exhibit exponential computational complexity growth with scale and persistent vulnerability to local minima traps. Quantum computing, while offering theoretical advantages through global superposition, faces practical barriers such as short decoherence times and current hardware limitations. To address these challenges, we propose a quantum-inspired fast algorithm for solving CCO problems. Our approach enhances global search capability with superposition encoding and avoids constraint-induced local minima​​ via a project–feedback strategy. Particularly, our method aligns with mature electronics manufacturing and demonstrates a proof-of-concept implementation in classical systems, indicating high-efficiency potential for solving constrained optimization problems.

Peripheral artery disease (PAD) confers elevated risk for major adverse cardiovascular events (MACE), yet accurate risk stratification remains a challenge, particularly among patients with advanced disease necessitating endovascular revascularization. This study aimed to improve the prediction of MACE in a clearly defined high-risk PAD population (hospitalized patients undergoing endovascular intervention) by identifying novel protein biomarkers and developing a robust risk model. We prospectively analyzed blood samples from 164 hospitalized PAD patients scheduled for endovascular revascularization, employing untargeted plasma proteomics and metabolomics. Differential protein and metabolite profiles were compared between patients with and without subsequent MACE. Several proteins, including MMP3, MMP19, and PRB2, were markedly elevated in patients who developed MACE. A proteomics-based risk model incorporating these biomarkers achieved high discriminative accuracy (area under the curve > 0.80) for identifying individuals at increased risk. Metabolomic profiling revealed additional pathway alterations, notably involving tryptophan and glycogen metabolism, which provided mechanistic insights into cardiovascular complications but were not directly incorporated into the prediction model. This study demonstrates that integrating protein biomarkers markedly improves risk stratification in advanced PAD patients undergoing surgical intervention. The findings offer promising tools for early detection and enable more personalized management for this high-risk subgroup, while also deepening understanding of disease pathophysiology. However, further validation in larger and more diverse prospective cohorts is warranted before these findings can be broadly applied in clinical practice.

Chimeric antigen receptor T (CAR-T) cell therapy, a groundbreaking technology in tumor immunotherapy, has demonstrated unprecedented potential in the field of autoimmune diseases in recent years. This article provides a systematic review of the developmental trajectory and core concepts of CAR-T-cell therapy in autoimmune diseases, emphasizing its conceptual evolution from traditional “killing” strategies to “precision immune remodeling”. Leveraging multitarget approaches (e.g., CD19 and B-cell maturation antigen) and chimeric autoantigen receptor technology, it achieves efficient elimination of pathogenic B cells, plasma cells, and autoreactive T cells, along with profound remodeling of the immune microenvironment, thereby inducing long-term disease remission and restoring immune tolerance. Nevertheless, unresolved challenges still exist in monotherapy strategies, such as antigen escape, nontumor toxicity of emerging targets, limited in vivo persistence, high production costs, and immune reconstitution imbalance. Future research ought to concentrate on the development of multitarget/logic-gated chimeric antigen receptor constructs, the optimization of chimeric antigen receptor architecture and nonviral delivery systems, the validation of the long-term safety of universal CAR-T cells, the customization of personalized treatment regimens, and the exploration of mechanisms for modulating the immune microenvironment. This review emphasizes that CAR-T-cell therapy shows potential for initiating a new era of personalized, mechanism-driven treatment for autoimmune diseases, offering crucial insights for clinical translation.

Knee osteoarthritis (KOA) is clinically characterized by recurrent and progressively worsening episodes; however, its underlying pathological mechanisms remain incompletely understood. Phenotypic changes and the specific migration of circulating monocytes may be key factors contributing to the recurrence and progressive exacerbation of synovial inflammation in KOA. In this study, we report a specific subtype of circulating monocytes in KOA that highly express interleukin 1A (IL1A), IL1R1, tumor necrosis factor, CXCL12, CXCL3, CXCL2, CCL20, and G0S2. These monocytes exhibit a trained immune phenotype, and their CXCR4-dependent migration to the synovium exacerbates synovial inflammation. HMGB1 (high mobility group protein B1) primes the trained immunity of this subtype, resulting in increased chromatin accessibility, and the transcriptional storage of multiple proinflammatory factors enables them to exhibit a more positive inflammatory response when restimulated by HMGB1. In addition, we find that MyD88, downstream of HMGB1, recruits cytoplasmic interferon regulatory factor 1 to the nucleus and then stabilizes interferon regulatory factor 1 in chromatin through SUMOylation of tripartite motif containing 28 to exert transcriptional activity and epigenetic enhancement of proinflammatory factor expression. Finally, we found that KOA inflammation was effectively attenuated by blocking HMGB1. Taken together, this study reveals the mechanism by which trained immunity of circulating monocytes promotes the progression of synovial inflammation, supporting a future therapeutic approach in the management of KOA.

The persistent hyperglycemic microenvironment in diabetic wounds causes dysfunction of repair cells, resulting in impaired angiogenesis and disorganized extracellular matrix (ECM) deposition. Small extracellular vesicles (sEVs) derived from hypoxia-challenged chorionic plate mesenchymal stem cells (CP-MSC-sEVs^Hypo) have been explored as novel acellular therapeutics. However, the effects of oxygen tension during culturing parent cells on the pro-regenerative efficacy and mechanisms of CP-MSC-sEVs have not been systematically investigated. In addition, their rapid clearance and limited interaction with target cells at wound sites constrained therapeutic efficacy. In this study, oxygen tension was first optimized systematically during the precondition of CP-MSCs to improve the pro-regenerative properties of the derived sEVs^Hypo. Then, sEVs^Hypo-engineered nanofiber scaffolds were fabricated to achieve the sustained release of sEVs^Hypo at wound sites by polydopamine (PDA)-mediated interfacial adhesion. In vitro experiments revealed that sEVs obtained under 5% oxygen tension (sEVs^Hypo-5%) exhibited the most pronounced angiogenic and collagen-regenerative performance. Furthermore, small RNA sequencing and bioinformatic analyses revealed that miR-21-5p was the most abundant miRNA in sEVs^Hypo-5%, and functional rescue assays validated that miR-21-5p was the key mediator of endothelial activation and ECM remodeling. To improve the bioavailability of sEVs^Hypo-5%, PDA was leveraged to immobilize sEVs^Hypo-5% onto a biocompatible, ECM-mimicking poly(ε-caprolactone) nanofiber scaffold to achieve high loading efficacy and sustained release. In a full-thickness diabetic wound model, the sEVs^Hypo-5%-loaded nanoscaffold accelerated wound closure and achieved superior pro-healing effects on D14 via enhancing neovascularization and ECM deposition, as confirmed by histological and immunofluorescence analyses. Collectively, this sEVs^Hypo-5%-loaded nanoscaffold achieved efficient immobilization and sustained delivery of sEVs^Hypo-5% to promote angiogenesis and matrix repair by endogenously delivering miR-21-5p at diabetic wound sites, with strong potential for clinical translation.

4-Chromanone is a biologically important heterocyclic scaffold found in natural products and pharmaceuticals. Traditional syntheses of 2-substituted 4-chromanones often rely on high-temperature peroxide initiation, transition-metal catalysis, or complex photocatalytic systems, limiting functional-group tolerance and sustainability. Herein, we report a Giese-type strategy for direct C–H functionalization of chromone derivatives, where chromones act as both substrates and hydrogen atom transfer (HAT) reagents to active inert C(sp³)–H bonds and P–H bonds under mild, metal-free conditions. Upon light irradiation, chromones form triplet states that mediate HAT for diverse substrates, including ethers, esters, alcohols, alkanes, sulfides, and arylphosphines. The resulting radicals are captured by ground-state chromones, efficiently affording 2-substituted 4-chromanones. Mechanistic studies and density functional theory calculations support the radical-mediated pathway, and the method avoids the use of metal catalysts, external redox mediators, or photocatalysts, highlighting its scalability and sustainable potential.

The utility of silk protein spans from East Asian traditions to promising applications in modern biomaterials science. While its bioactive effect is widely accepted, its molecular mechanisms and the supporting data have remained elusive. This study reveals the anti-photoaging mechanism of silk protein enzymatic hydrolysate (SEH), identifying key peptides SO1 from silk seroin and SC1 from silk sericin. Mechanistically, SO1 targets PDGFRβ to inhibit the NF-κB/ERK/SASP cellular senescence axis, while SC1 targets TRPV4 to suppress MITF-driven melanin synthesis. Additionally, the SO1 and SC1 combination suppresses key hallmarks of photoaging. It reduced senescence, confirmed by diminished β-galactosidase activity and SASP, while concurrently mitigating oxidative stress and hyperpigmentation, and promoting cell migration. Further network pharmacology analysis reveals a multi-component interaction network connecting bioactive silk peptides, target receptors, and downstream signaling pathways involved in mitigating photoaging. This integrative mechanism was defined as the “Silk Peptides Mesh”, with SO1 and SC1 identified as representative core peptides within this system. Finally, a 28-d skin clinical trial involving 30 participants confirms these findings, showing that a cream containing SO1 and SC1 improves photoaging-related skin indicators. Importantly, this work establishes a paradigm shift from conventional single-molecule approaches to a multi-component network understanding of bioactive peptide functionality.

Renal metabolic behavior constitutes a critical evaluation parameter during drug development, toxicity assessment, and biodistribution. However, the inherent biological architecture of existing animal and cell-based models fails to recapitulate the authentic metabolic processes and spatial distribution patterns of drugs in the human kidney at a macroscopic level. Therefore, developing a standardized human-derived kidney model was necessary. In this work, the ex vivo tumor-bearing kidney (ETK) model was established using kidneys from 50 clinically diagnosed patients after the radical surgery. The standardized ETK model included the following: (a) selected according to the proportion of preserved normal tissue, vascular status suitable for ETK perfusion, and overall structural integrity of the kidneys; (b) surgical preparation exposed the renal pelvis and enabled catheterization of the renal artery and ureter; (c) perfused with hypothermic (4 to 10 °C) saline at a flow rate of 70 ml/min within perfusion system; and (d) ETK detection efficiency and drug biodistribution were assessed using biological sample indicators. Results showed that ETK maintained intact glomerular and tubular microstructures throughout perfusion and produced stable urine output with preserved filtration capacity. When treated with different kinds of nephrotoxic agents (cisplatin, gentamicin, and cephalosporins), the ETK model reproduced drug-specific injury patterns and exhibited biochemical changes and histological features that were consistent with real clinical observations. In drug-distribution analysis, indocyanine green (ICG) showed a negative fluorescence tumor contrast and enriched fluorescence in the normal kidney in the ETK model, which were consistent with real clinical observations, whereas mouse orthotopic renal cancer models showed positive ICG tumor accumulation, indicating species-dependent differences in renal drug distribution. Overall, the ETK model provided a structurally preserved and functionally responsive human-derived renal system for investigating renal drug metabolism, nephrotoxicity, tumor imaging, and drug-delivery strategies.

High-glucose microenvironment in diabetic wound exacerbates the risk of bacterial infections and perpetuates chronic inflammation, thereby complicating wound healing. To meet these challenges, we develop antibacterial agents by integrating probiotics with nanozymes targeting high glucose levels and multidrug-resistant bacterial infections in diabetic wound healing. We engineer probiotic Lactobacillus plantarum (LP) with Fe-based nanozymes containing glucose oxidase (GOx) as antibacterial agents (LP@FeG) for diabetic wound therapy. The bioactivity of LP ferments glucose in the wounds, thereby reducing local pH through lactic acid production. Besides, GOx catalyzes the conversion of glucose into gluconic acid and hydrogen peroxide, further lowering local pH. In addition, GOx stimulates the peroxidase-like activity of Fe-based nanozymes, thereby generating reactive oxygen species that attack the cellular membrane, facilitating the entry of Fe²⁺/Fe³⁺ into bacteria and inducing bacterial death through ferroptosis and lipid peroxidation. In vivo studies demonstrate that LP@FeG exhibits prolonged sterilization of methicillin-resistant Staphylococcus aureus with an almost 100% antibacterial rate, marked anti-inflammatory, and proangiogenic effects in diabetic wounds. This strategy resolves both the high glucose level and bacterial infections in wound healing, being promising for the treatment of diabetic wounds.

SUN1/2, core components of the linker of nucleoskeleton and cytoskeleton complex, transmit extracellular mechanical forces to nuclear lamina and chromatin. However, their role in regulating peripheral chromatin in mechanosensing and mechanoadaptation remains unclear. Using CRISPR/Cas9-mediated knockout of Sun1 or Sun2 in myoblasts, we identified a SUN1/2-dependent mechano-feedback loop. SUN1/2 depletion down-regulates genes for cell adhesion (e.g., integrin alpha-4) and for mechanotransduction (e.g., cell division cycle 42 and Ras homolog family member A). The primary mechanism involves redistribution of heterochromatin from nuclear periphery to the nucleoplasm and remodeling of lamina-associated domains (LADs), as an adaptive response to the loss of SUN proteins. Furthermore, lamin A/C acts as a key downstream effector, consistently modulating adhesion-related gene expression through the remodeling of LADs. Functionally, knockout of either Sun1/2 or Lmna aggravates differentiation defects in C2C12 myoblasts and abolishes adaptive responses to mechanical cues. This study provides proof of concept that nuclear mechanotransduction proteins can modulate cellular mechanoadaptation via a mechano-feedback loop, which coordinates LAD reorganization with the expression of upstream mechanotransduction genes.