Cold seep ecosystems serve as critical hubs in marine carbon cycling through methane emissions and organic matter processing. While terrestrial lignin constitutes a major fraction of persistent organic carbon in cold seep sediments, its microbial transformation pathways in deep-sea cold seep environments remain unexplored. Here, we present the first comprehensive analysis of lignin distribution across sediment horizons at the Haima cold seep, coupled with a multi-omics investigation of microbial lignin metabolism. Laboratory enrichment of sediment communities employing lignin as the exclusive carbon substrate revealed substantial microbial community restructuring dominated by Burkholderiales, Pseudomonadales, and Rhizobiales lineages. Integrated omics resolved 2-tiered metabolic cascades: (a) enzymatic depolymerization via dyP-type peroxidases and LigEFG-mediated β-aryl ether cleavage, targeting syringyl and diarylpropane subunits; (b) funneling of aromatic intermediates through 4,5-/3,4-PDOG (protocatechuate dioxygenase) pathways into central carbon metabolism. Although direct methanogenesis was undetected, methylotrophic potential was evidenced through methane cycle gene expression patterns by lignin decomposers. Phylogenetic surveys further demonstrated the global prevalence of lignin decomposers across 12 major cold seep systems. These decomposers showed marked divergence in enzymatic repertoires compared to degraders from other ecosystems. Our findings establish 3 paradigm shifts: (a) The turnover rates of terrestrial organic carbon are likely underestimated in deep-sea ecosystems; (b) microbial consortia employ combinatorial enzymatic strategies distinct from terrestrial decomposition regimes; (c) methyl shunting from lignin breakdown primes methanogenic precursors, revealing cryptic linkages between refractory carbon cycling and greenhouse gas reservoirs.

Radiotherapy (RT) is a cornerstone treatment for non-small cell lung cancer (NSCLC), but its efficacy is often limited by immune suppression in the tumor microenvironment. In this study, we identified SPP1 as a key factor up-regulated after RT, mainly expressed by immunosuppressive macrophages. Single-cell RNA sequencing and in vivo models showed that SPP1⁺ macrophages inhibit CD8⁺ T cell infiltration and correlate with poor prognosis. Targeting SPP1 in macrophages enhanced RT efficacy, reduced tumor burden, and restored antitumor immunity. In summary, combining RT with SPP1⁺ macrophage-targeted intervention may serve as a promising strategy to overcome immune-mediated radioresistance and enhance therapeutic efficacy in NSCLC.

Abrupt shifts, referred to as critical transitions, are frequently observed in complex biological systems, characterized by marked qualitative changes occurring from one stable state to another through a pre-transitional/critical state. Pinpointing such critical states, along with the signaling molecules, can provide valuable insights into the fundamental mechanisms of intricate biological processes. However, the identification and early warning of the critical state remains a challenge, particularly in model-free cases with high-dimensional single-cell data, where traditional statistical methods often prove inadequate due to the inherent sparsity, noise, and heterogeneity of the data. In this study, we propose a novel quantitative method, cell-specific causal network entropy (CCNE), to infer the specific causal network for each cell and quantify dynamic causal changes, thereby enabling the identification of critical states in complex biological processes at the single-cell level. We validated the accuracy and effectiveness of the proposed approach through numerical simulations and 5 distinct real-world single-cell datasets. Compared to existing methods for detecting critical states, the proposed CCNE exhibits enhanced effectiveness in identifying critical transition signals. Moreover, CCNE score is a computational tool for distinguishing temporal changes in cellular heterogeneity and demonstrates satisfactory performance in clustering cells over time. In addition, the reliability of CCNE is further emphasized through the functional enrichment and pathway analysis of signaling molecules.

Bacterial infections pose a major challenge today due to the rise of drug resistance in pathogenic bacteria, creating an urgent need for the development of nonantibiotic therapies. Metal nanomaterials are increasingly recognized as promising antibacterial agents due to their unique physical and chemical properties, which offer strong antibacterial capabilities and broad-spectrum activity against drug-resistant bacteria. Unlike traditional antibacterial therapies, metal nanomaterials rarely induce drug resistance because of their diverse antibacterial mechanisms, which destroy bacteria by directly damaging bacterial cells or generating oxidative stress. Metal nanomaterials have proven effective in treating various drug-resistant bacterial infections. This perspective highlights recent advances in metal nanomaterials for antimicrobial applications. Firstly, bacterial infections and the current dilemma are introduced. Next, recent progress in the antibacterial activity of metal nanomaterials against drug-resistant bacteria is summarized, along with the challenges in their antimicrobial applications. Finally, future prospects and remaining challenges associated with the use of metal nanomaterials to treat drug-resistant bacteria are discussed.

Droplet microfluidics is a rapidly evolving technology enabling precise control and manipulation of small-volume droplets, typically ranging from picoliters to nanoliters, offering important potential for biomedical applications. By generating highly uniform droplets with size variation below 5% and at high frequencies exceeding 10,000 droplets per second using techniques such as flow focusing, this approach facilitates high-throughput experimentation with minimal reagent consumption. These features make droplet microfluidics invaluable for single-cell analysis, drug screening, and disease diagnostics. Recent advancements in integrating droplet microfluidics with biological and clinical workflows have expanded possibilities for personalized medicine, early disease detection, and high-resolution cellular assays. This review provides an overview of recent progress in droplet microfluidics, focusing on key techniques for droplet generation, manipulation, and detection. It explores their applications in cutting-edge biomedical research, including single-cell analysis, 3-dimensional cell culture, drug development, and cancer research. Additionally, we discuss current challenges, such as improving reproducibility, scalability, and system integration, and outline promising future directions to fully realize the potential of droplet microfluidics in biomedicine.

The health problem of infertility has garnered increasing attention, prompting a deeper understanding of its causes. The broad-spectrum and nonselective herbicide glufosinate ammonium (GLA) is widely used in many countries. Previous studies have demonstrated the reproductive toxicity of GLA, but its potential toxic mechanisms remain unclear. Here, mice, Sertoli cells, and Leydig cells were used to create GLA preconditioning models. Results showed that GLA exposure caused morphological and functional damage of sperm. Concurrently, our study revealed that GLA, similar to Erastin, could induce ferroptosis in Sertoli and Leydig cells, as indicated by the dose-dependent increases of intracellular iron levels, lipid peroxidation, and cell death. Additionally, both the lipid ROS scavenger Fer and the iron chelator deferiprone were found to mitigate GLA-induced cell death. Intriguingly, our findings suggested that GLA-induced ferroptosis was dependent on autophagy, as the use of pharmacological inhibitors (3-methyladenine, chloroquine, and bafilomycin A1) or autophagy-related gene 5 gene knockout markedly reduced ferroptosis induced by GLA. We also demonstrated that nuclear receptor coactivator 4 (NCOA4)-mediated ferritinophagy, which involves the autophagic degradation of the primary intracellular iron storage protein ferritin, is essential for GLA-induced ferroptosis by showing that NCOA4 knockdown decreased intracellular iron levels and attenuated lipid peroxidation, eventually alleviating GLA-induced cell death. Moreover, we observed that inhibition of the AMP-activated protein kinase–Unc-51-like kinase 1 (AMPK-ULK1) pathway activity by knockdown of AMPK expression markedly reduced the mitochondrial reactive oxygen species (mtROS) level and alleviated GLA-induced ferroptosis. Collectively, GLA induced excessive mtROS production through activation of the AMPK-ULK1 pathway, triggering excessive autophagy that ultimately led to ferroptosis via NCOA4-mediated ferritinophagy.

Metabolic enzyme dysregulation promotes hepatocellular carcinoma (HCC) progression through metabolic reprogramming and lysine acetylation. Glutaryl-CoA dehydrogenase (GCDH), a key enzyme in lysine metabolism, has been demonstrated to play an essential role in modulating lysine crotonylation, which impacts the progression of HCC. However, the specific mechanisms by which GCDH influences lysine acetylation in HCC have not been completely clarified. In this study, GCDH was found to be acetylated at lysine 438 by acetyltransferase P300 and deacetylated by HDAC1. GCDH K438 acetylation was critical for its tumor-suppressive function in HCC cells. Overexpression of GCDH led to elevated levels of reactive oxygen species (ROS) and reduced oxidative phosphorylation (OXPHOS), thereby triggering ATR/Chk1-mediated DNA damage repair dysfunction and promoting autophagy in HCC cells. Furthermore, our investigation demonstrated that decreased GCDH expression was markedly associated with shorter overall survival in HCC patients and served as an independent prognostic indicator. Collectively, our findings demonstrate that the acetylation of GCDH at lysine 438 (K438), mediated by P300 and HDAC1, plays a vital role in the tumor-suppressive activities of HCC cells. GCDH inhibits HCC progression through ROS-mediated DNA repair dysfunction and autophagy.

The terahertz (THz) frequency range, situated between microwave and infrared radiation, has emerged as a pivotal domain with broad applications in high-speed communication, imaging, sensing, and biosensing. The development of topological THz metadevices represents a notable advancement for photonic technologies, leveraging the distinctive electronic properties and quantum-inspired phenomena inherent to topological materials. These devices enable robust waveguiding capabilities, positioning them as critical components for on-chip data transfer and photonic integrated circuits, particularly within emerging 6G communication frameworks. A principal advantage resides in the capacity to maintain low-loss wave propagation while effectively suppressing backscattering phenomena, a critical requirement for functional components operating at higher frequencies. In parallel, by leveraging advanced materials such as liquid crystals, plasma, and phase-change materials, these devices facilitate real-time control over essential wave parameters, including amplitude, frequency, and phase, which augments the functionality of both communication and sensing systems, opening new avenues for THz-based technologies. This review outlines fundamental principles of topological components and reconfigurable metadevices operating at THz frequencies. We further explore emerging strategies that integrate topological properties and reconfigurability, with a specific focus on their implementation in chip-scale photonic circuits and free-space wavefront control.

Language model (LM), a foundational algorithm in the development of capable artificial intelligence, has been widely explored, achieving remarkable attainment. As research advances, large language models (LLMs) have emerged by pretraining transformer-based models on large-scale corpora. These models showed great zero-shot and few-shot learning capabilities across a variety of tasks, attracting widespread attention from both academia and industry. Despite impressive performance, LLMs still tackle challenges in addressing complex real-world scenarios. Recently, the advent of DeepSeek has reignited intense interest among researchers. In this paper, we provide a concise development history of LLM and discuss current challenges and future perspective. In practice, we focus on 4 crucial aspects of LLMs, including emergent abilities, human alignment, retrieval augmented generation, and applications in specific domains.

Vascular disease is the leading cause of death worldwide. Predicting the burden of vascular disease and identifying modifiable key risk factors are critical for developing effective prevention strategies. This study aimed to project the global and regional burden of peripheral artery disease (PAD) from 2021 to 2050, with a specific focus on the impact of modifiable key risk factors and the potential benefits of their management. Compared to the 2021 Global Burden of Disease Study (GBD 2021), the number of PAD cases worldwide is projected to increase by 220% by 2050, reaching a staggering 360 million (95% uncertainty interval, 270 to 450). Age-standardized mortality is expected to double, while disability-adjusted life years (DALYs) are forecasted to rise from 19.7 to 33.1 per 100,000. Among individuals aged ≥65 years, PAD prevalence is projected to surge to 21.7% in women and 14.8% in men. Moreover, over 50% of PAD cases are expected to occur in low- and middle-income countries (LMICs). Metabolic diseases are anticipated to be the primary drivers of the rising PAD burden, with diabetes playing a key role in increasing PAD prevalence and severity. By effectively managing metabolic risk factors, age-standardized prevalence could be reduced by 36%, mortality by 17%, and DALYs by 10%. As metabolic risks, particularly diabetes, continue to rise alongside population aging, the global PAD burden is expected to increase substantially, especially in LMICs. Importantly, proactive metabolic risk management strategies have the potential to markedly alleviate the burden of vascular disease and reduce the growing geographic health disparities.