Sepsis remains a leading global cause of death, with immune heterogeneity’s molecular mechanisms poorly understood. This study analyzed 1,862 human peripheral blood samples, constructing a molecular interaction disturbance network that first revealed the network biology underlying sepsis immune heterogeneity. We identified 3 sepsis subtypes with marked different prognostic characteristics, with the C1 subtype showing the worst prognosis characterized by severe CD4⁺ T cell exhaustion—validated across 10 independent cohorts. Integrating single-cell transcriptomics from over 450,000 cells, proteomics, and functional validation, we identified monocyte-derived matrix metalloproteinase-9 (MMP9) as a key regulator driving CD4⁺ T cell dysfunction. Mechanistically, MMP9 modulates T cell exhaustion through dual mechanisms: promoting leukocyte-associated immunoglobulin-like receptor-1 (Lair-1) aggregation on T cell membranes, directly inhibiting zeta chain of T cell receptor associated protein kinase 70 (ZAP70) phosphorylation in T cell receptor signaling, while impairing Ca²⁺-release-activated Ca²⁺ channel function and intracellular calcium clearance, causing calcium dysregulation that blocks nuclear factor of activated T cells (NFAT) activation and nuclear translocation. The selective MMP9 inhibitor MMP9-in-1 effectively reversed T cell dysfunction, restored calcium homeostasis and NFAT nuclear translocation, markedly enhanced CD4⁺ T cell interleukin-2 and interferon-γ production, and reduced programmed cell death protein 1 expression. This work establishes a comprehensive translational framework from molecular network disturbances to clinical phenotypes, advancing sepsis immunopathophysiology understanding and providing effective targets for the treatment of sepsis patients.

Sepsis is a life-threatening disease characterized by systemic inflammation and endothelial barrier dysfunction, leading to multiorgan failure and high mortality. This review provides a comprehensive overview of the pathological mechanisms underlying sepsis-induced endothelial dysfunction, focusing on glycocalyx degradation, endothelial cell death, increased vascular permeability, and coagulopathy. During sepsis, the endothelial glycocalyx (EG) is disrupted, leading to increased vascular permeability and impaired microcirculation. Endothelial cells undergo various forms of cell death, including apoptosis, pyroptosis, ferroptosis, and autophagy, which are driven by inflammation and oxidative stress. These processes are further complicated by the activation of coagulation pathways and the formation of intravascular thrombi. The interaction between endothelial cells and immune cells amplifies the inflammatory response, contributing to the persistence of systemic inflammation. We also discuss emerging therapeutic strategies aimed at protecting endothelial cells, modulating inflammation, and improving coagulation function, including glycocalyx protectants, anti-inflammatory agents, anticoagulants, and endothelial repair mechanisms. Future research should focus on translating these therapeutic approaches into clinical practice to improve outcomes in patients with sepsis.

Vasculogenic mimicry (VM) is a microcirculation pattern that has a crucial effect on hepatocellular carcinoma (HCC) metastasis. In this study, leveraging the GeneCard and The Cancer Genome Atlas databases, we identified ATPase inhibitory factor 1 (IF1) as a potential regulator of VM formation. Our research findings indicate that IF1 can promote HCC cell tube formation in vitro and enhance HCC VM and lung metastasis in vivo. Transcriptome sequencing combined with in vivo experiments revealed that IF1 knockdown elevates miR-20a-3p expression. Lentivirus-mediated miR-20a-3p overexpression reversed IF1-induced VM. Dual-luciferase reporter gene assays showed that estrogen receptor 1 (ESR1) acts as a transcription factor of the miR-20a-3p precursor. Further mechanistic studies revealed that excessive reactive oxygen species accumulation caused by IF1-induced mitochondrial metabolic reprogramming can inhibit ESR1 expression by promoting DNA methylation of its promoter. G protein subunit alpha Z (GNAZ), a miR-20a-3p target protein, can promote VM by phosphorylating components of the ERK pathway. Collectively, these results delineate a novel IF1/ESR1/miR-20a-3p/GNAZ axis in HCC VM and metastasis, providing potential therapeutic targets.

Since the skin is the largest organ of the human body, skin injuries may result in substantial social and economic burdens. Skin wound management remains a substantial challenge due to the complex nature of the healing process involving inflammation, tissue proliferation, and remodeling. Coordinated interactions between cells, extracellular matrices, and cytokines drive regeneration, while cell metabolism—an essential pillar of physiological activity—regulates key biological processes like proliferation, angiogenesis, immune responses, and tissue regeneration. Modulating metabolic pathways provides a promising alternative to conventional therapies. However, understanding of the role of cell metabolism in wound healing remains to be fully elucidated. Further exploration of their cross talk will not only establish a robust theoretical foundation for innovative treatments but also open avenues for promoting scar-minimizing regenerative healing and bringing new prospects for clinical practice.

Hepatic lipid dysregulation drives metabolic dysfunction-associated steatotic liver disease (MASLD); nonetheless, the precise regulatory mechanisms remain incompletely elucidated. In this study, we examine the function of EVA1A, a known hepatocellular carcinoma tumor suppressor, in hepatic lipid metabolism. Hepatic EVA1A was markedly down-regulated in individuals diagnosed with MASLD, as well as in mice subjected to a high-fat diet. Hepatocyte-specific knockout of Eva1a in mice resulted in significant hepatic steatosis, accompanied by disrupted fatty acid metabolism, marked by increased fatty acid uptake and compromised β-oxidation, while hepatic Eva1a overexpression reversed these metabolic changes and largely alleviated fatty liver in ob/ob mice. Mechanistically, EVA1A deficiency activates mTORC1 (mechanistic target of rapamycin complex 1)–PPARγ2 (peroxisome proliferator-activated receptor γ2) signaling to up-regulate CD36 transcription. Concurrently, it transcriptionally represses the S-depalmitoylase APT1 while enhancing palmitoyl acyltransferases ZDHHC4/5, boosting CD36 palmitoylation. This dual action promotes CD36 plasma membrane localization for fatty acid uptake, reducing its mitochondrial distribution and impairing β-oxidation. Collectively, these results establish EVA1A as an essential regulator of hepatic lipid homeostasis, coordinating fatty acid uptake and β-oxidation by modulating CD36 expression and palmitoylation. Therefore, targeting the EVA1A–CD36 axis represents a promising therapeutic strategy for MASLD.

The development of an immunosuppressive microenvironment is a critical factor in stomach carcinogenesis. Polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) serve a pivotal function in mediating immune suppression. However, the precise mechanisms underlying PMN-MDSCs infiltration into the tumor immune microenvironment (TIME) and their immunosuppressive functions remain poorly understood. In this investigation, we observed that PMN-MDSCs were up-regulated during stomach carcinogenesis, with gastric cancer (GC) cells secreting CCL26 to promote the infiltration of PMN-MDSCs into the TIME via the CX3CR1 receptor. The infiltrating CX3CR1⁺ PMN-MDSCs secreted transforming growth factor-β1 (TGF-β1), which, through the SMAD2/3/SNAI1 signaling pathway, further sustained the secretion of CCL26 by GC cells, establishing a positive feedback loop. Additionally, CX3CR1⁺ PMN-MDSCs suppressed mTOR signaling via TGF-β1 and competed with CD8⁺ T cells for glucose, disrupting glycolysis and leading to T cell exhaustion. Notably, inhibition of CX3CR1 reduced the infiltration of PMN-MDSCs, improved anti-PD-1 therapeutic efficacy, and suppressed tumor growth. In conclusions, this study illustrated that the CCL26–CX3CR1 axis mediates a positive feedback loop between GC cells and PMN-MDSCs, promoting CD8⁺ T cell exhaustion and tumorigenesis. Targeting CX3CR1 on PMN-MDSCs could serve as a potential therapeutic strategy to prevent stomach carcinogenesis.

Organic piezoelectric materials have garnered increasing interest for biomechanical applications; however, their practical appeal has been limited by low piezoelectricity and low mechanical compliance. Recently, liquid–liquid interface polar engineering has emerged as a promising strategy to overcome these limitations. By utilizing polarity asymmetry at the liquid–liquid interface, this strategy enables precise control over molecular assembly and phase separation, thereby inducing high piezoelectric polarization of organic materials while maintaining their intrinsic mechanical compliance. This breakthrough paves the way for developing organic piezoelectric materials that exhibit both high piezoelectricity and high mechanical compliance, advancing their potential in biomechanical sensing, actuating, and energy harvesting.

The normal development of spermatogonial stem cells (SSCs) is essential for maintaining male fertility. Nevertheless, signaling molecules and pathways regulating the fate decisions of human SSCs remain elusive. We have reported that Opa interacting protein 5 (OIP5) is involved in controlling human SSC self-renewal and apoptosis. Notably, we found that insulin-like growth factor 2 mRNA-binding protein 1 (IGF2BP1) interacted with OIP5 in human SSCs. Our RNA sequencing (RNA-seq) and immunofluorescence–fluorescence in situ hybridization (IF-FISH) identified heme oxygenase 1 (HMOX1) as a downstream target of IGF2BP1 in human SSCs. Methylated RNA immunoprecipitation-binding qPCR (MeRIP-qPCR) and RNA stability assays revealed that IGF2BP1 could bind to HMOX1 mRNA and enhance its stability and HMOX1 expression level. Functional assays demonstrated that IGF2BP1 or HMOX1 silencing resulted in the decreased expression levels of ferroptosis-associated genes (e.g., SLC7A11 and SLC3A2), the increased intracellular reactive oxygen species (ROS) level and ferrous ion (Fe²⁺) content, and the impaired stemness maintenance of human SSCs. Additionally, we observed that IGF2BP1-mediated HMOX1 stabilization activated system Xc⁻, thereby inhibiting ferroptosis of human SSCs. Furthermore, IGF2BP1 gene variants were positively correlated with the occurrence of non-obstructive azoospermia (NOA). Collectively, these results imply a critical role of IGF2BP1/HMOX1 mRNA/system Xc⁻ axis in regulating human SSC ferroptosis, autophagy, and stemness maintenance. This study thus provides novel insights into the regulatory mechanisms of human SSC fate determinations and offers new targets for treating male infertility.

Quantum machine learning stands poised as a forefront application for near-term quantum devices, addressing scalability challenges posed by classical computers in handling large datasets. Gaussian boson sampling (GBS), an intricate quantum algorithm deemed computationally infeasible for classical counterparts, represents a substantial leap forward in computational tasks. However, to date, the benefits of GBS-assisted quantum unsupervised machine learning are not experimentally demonstrated. Here, we present the first experimental implementation of quantum unsupervised machine learning using the GBS protocol with a universal programmable integrated photonic chip. The experimental system contains 16 squeezing sources, a universal programmable unitary matrix network of 16 modes, and a multi-channel single-photon detector, producing substantial output data crucial for 2 typical types of unsupervised tasks: feature extraction and generative network. Compared to classical approaches, the study demonstrates quantum-enhanced capability in extracting complex features from high-dimensional spaces and improved performance in generating arbitrary curve points and reconstructing handwritten digit images. This work not only underscores the potential of GBS in expressing high-dimensional features but also charts a path toward practical implementations within scalable, dimension-enhanced quantum unsupervised machine learning frameworks. The quantum unsupervised machine learning paradigm, offering theoretical acceleration and reduced training parameters for high-dimensional datasets, shows significant promise for advancing real-world applications of quantum technologies.

A microwave-coupled biphasic solvent system was developed for steerable liquefaction of pentoses and hierarchic separation of furfural by leveraging the dielectric properties of γ-valerolactone (GVL)/NaCl aqueous solution under microwave irradiation, during which more than 85.38 mol% of furfural yield was obtained from xylan at 140 °C for 20 min, compared to that of 78.1 mol% at 150 °C for 120 min under conventional heating conditions. Strikingly, the separation efficiency was improved under microwave conditions with the partition coefficients (R) of furfural ranging from 31.1 to 35.68, compared with 22.71 to 29.33 under conventional heating conditions. Moreover, the coupled effect on the substrate from the microwave was also explored. Notably, high heating power during the initial hydrolysis phase enhances the depolymerization of xylan glycosidic bonds, resulting in the rapid release of 87.89 mol% xylose from xylan. Meanwhile, the application of low heating power during the dehydration process effectively facilitates the production of furfural, which is promptly extracted into the organic phase, thereby minimizing product loss. Besides, the feasibility of the process was assessed using real biomass as the substrate, demonstrating low energy consumption attributed to the efficient microwave absorption by GVL and the NaCl aqueous solution. This approach offers a viable strategy for the efficient and scalable production of furfural.