DEAD-box ATPase 10 (DDX10), a prominent RNA-binding protein in the DDX family, has a critical function in cancer progression. Nevertheless, its well-defined mechanisms in oral squamous cell carcinoma (OSCC) are still not well understood. Here, we identify that DDX10 is substantially increased in OSCC, which is positively correlated with poor prognosis and malignant behavior. Mechanistically, we found that DDX10 had physical interaction with Rab27b by undergoing phase separation. Knockdown of DDX10 inhibited Rab27b-mediated exosome secretion and the expression of programmed cell death-ligand 1 (PD-L1) within its contents. Furthermore, knocking down DDX10 could restore the function and infiltration of T cells, hence inhibiting the progression of OSCC. These findings highlight that the oncogenic role of DDX10 in promoting exosomal PD-L1 secretion via phase separation with Rab27b has been preliminarily validated in T cell exhaustion in OSCC. A potential strategy for improving OSCC immunotherapy may involve the inhibition of DDX10.

Increasing evidence indicates that oligodendrocyte (OL) numbers and myelin as a dynamic cellular compartment perform a key role in the maintenance of neuronal function. Inhibiting white matter (WM) demyelination or promoting remyelination has garnered interest for its potential therapeutic strategy against ischemic stroke. Our previous work has shown that low-intensity pulsed ultrasound (LIPUS) could improve stroke recovery. However, it is unclear whether LIPUS can maintain WM integrity early after stroke or promote late WM repair. This study evaluated the efficacy of LIPUS on WM repair and long-term neurologic recovery after stroke. Male adult C57BL/6 mice underwent a focal cerebral ischemia model and were randomized to receive ultrasound stimulation (30 min once daily for 14 days). The effect of LIPUS on sensorimotor function was assessed by modified neurological severity score, rotarod test, grip strength test, and gait analysis up to 28 days after stroke. We found that ischemic stroke-induced WM damage was severe on day 7 and partially recovered on day 28. LIPUS prevented neuronal and oligodendrocyte progenitor cell (OPC) death during the acute phase of stroke (d7), protected WM integrity, and reduced brain atrophy and tissue damage during the recovery phase (d28). To further confirm the effect of LIPUS on remyelination, we assessed the proliferation and differentiation of OPCs. We found that LIPUS did not increase the number of OPCs (PDGFRα⁺ or NG2⁺), but markedly increased the number of newly produced mature OLs (APC⁺) and myelin protein levels. Mechanistically, LIPUS may promote OL maturation and remyelination by down-regulating the interleukin-17A/Notch1 signaling pathway. In summary, LIPUS can protect OLs and neurons early after stroke and promote long-term WM repair and functional recovery. LIPUS will be a viable strategy for the treatment of ischemic stroke in the future.

Docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) have been proposed to benefit cardiometabolic health. However, the relationship between the intake of DHA and EPA and type 2 diabetes (T2D) risk remains equivocal, and the effects of DHA and EPA on skeletal muscle, the primary organ for glucose metabolism, merit further investigation. Here, we show that habitual fish oil supplementation was associated with a 9% lower T2D risk and significantly interacted with variants at GLUT4 in a prospective cohort of 48,358 people with prediabetes. Muscular metabolome analysis in the animal study revealed that DHA and EPA altered branched-chain amino acids, creatine, and glucose oxidation-related metabolites, concurrently with elevated muscular glycogen synthase and pyruvate dehydrogenase contents that promoted glucose disposal. Further myotube investigation revealed that DHA and EPA promoted muscular GLUT4 translocation by elevating Rab GTPases and target-SNARE expression. Together, DHA and EPA supplementation provides a promising approach for T2D prevention through targeting muscular glucose homeostasis, including enhancing GLUT4 translocation, glycogen synthesis, and aerobic glycolysis.

Sample barcoding-based multiplex single-cell and single-nucleus sequencing (sc/sn-seq) offers substantial advantages by reducing costs, minimizing batch effects, and identifying artifacts, thereby advancing biological and biomedical research. Despite these benefits, universal sample barcoding has been hindered by challenges such as inhomogeneous expression of tagged biomolecules, limited tagging affinity, and insufficient genetic insertion. To overcome these limitations, we developed Toti-N-Seq, a universal sample multiplex method, by tagging Toti-N-glycan on cell surfaces or nuclear membranes via our engineered streptavidin–Fbs1 GYR variant fusion protein, which could be used not only for sc-seq but also for sn-seq. Instead of targeting lipids or proteins, we focused on targeting the ubiquitous N-glycans found on any species with accessible membranes, which minimizes the exchange between barcoded samples and avoids biased barcoding. Our technology can be broadly applied to multiple species and nearly all eukaryotic cell types, with an overall classification accuracy of 0.969 for sc-seq and of 0.987 for sn-seq. As a demonstration with clinical human peripheral blood mononuclear cells, our Toti-N-Seq achieved rapid one-step sample preparation (<3 min) for easily scaling up while keeping high fidelity of sample ratios, removing artifacts, and detecting rare cell populations (~0.5%). Consequently, we offer a versatile platform suitable for various cell types and applications.

Stimuli-responsive materials have shown promising applications in the areas of sensing, bioimaging, information encryption, and bioinspired camouflage. In particular, multi-stimuli-responsive materials represent a hot topic due to their modulated properties under multiple stimuli. Herein, we successfully developed multi-stimuli-responsive inks and a series of complex multi-stimuli-responsive 3-dimensional (3D) structures were fabricated via digital light processing 3D-printing technology. Notably, these complex 3D structures show shape memory, fast-response photochromic and thermochromic behavior, and excellent repeatability due to the combination of photochromic molecules (4-(2,2-bis(4-fluorophenyl)vinyl) benzyl methacrylate) and thermochromic pigments. Furthermore, a programmable encrypted box that changes colors and morphology by controlling temperature and ultraviolet irradiation was designed and printed, and this encrypted box exhibits strong security using OpenCV-based image recognition technology. This strategy provides a promising approach for the design of multi-stimuli-responsive materials and complex encryption systems in the future.

Background: A century ago, a mystery between a virus and Parkinson’s disease (PD) was described. Owing to the limitation of human brain biopsy and the challenge of electron microscopy in observing virions in human brain tissue, it has been difficult to study the viral etiology of PD. Recent discovery of virobiota reveals that viruses coexist with humans as symbionts. Newly developed transcriptomic sequencing and novel bioinformatic approaches for mining the encrypted virome in human transcriptome make it possible to study the relationship between symbiotic viruses and PD. Nevertheless, whether viruses exist in the human substantia nigra (SN) and whether symbiotic viruses underlie PD pathogenesis remain unknown. Methods: We collected current worldwide human SN transcriptomic datasets from the United States, the United Kingdom, the Netherlands, and Switzerland. We used bioinformatic approaches including viruSITE and the Viral-Track to identify the existence of viruses in the SN of patients. The comprehensive RNA sequencing-based virome analysis pipeline was used to characterize the virobiota in the SN. The Pearson’s correlation analysis was used to examine the association between the viral RNA fragment counts (VRFCs) and PD-related human gene sequencing reads in the SN. The differentially expressed genes (DEGs) in the SN between PD patients and non-PD individuals were used to examine the molecular signatures of PD and also evaluate the impact of symbiotic viruses on the SN. Findings: We observed the existence of viruses in the human SN. A dysbiosis of virobiota was found in the SN of PD patients. A marked correlation between VRFC and PD-related human gene expression was detected in the SN of PD patients. These PD-related human genes correlated to VRFC were named as the virus-correlated PD-related genes (VPGs). We identified 3 bacteriophages (phages), including the Proteus phage VB_PmiS-Isfahan, the Escherichia phage phiX174, and the Lactobacillus phage Sha1, that might impair the gene expression of neural cells in the SN of PD patients. The Proteus phage VB_PmiS-Isfahan was a common virus in the SN of patients from the United Kingdom, the Netherlands, and Switzerland. VPGs and DEGs together highlighted that the phages might dampen dopamine biosynthesis and weaken the cGAS-STING function. Interpretation: This is the first study to discover the involvement of phages in PD pathogenesis. A lifelong low symbiotic viral load in the SN may be a contributor to PD pathogenesis. Our findings unlocked the black box between brain virobiota and PD, providing a novel insight into PD etiology from the perspective of phage–human symbiosis.

Pulmonary arterial hypertension (PAH) is a devastating disease characterized by perivascular inflammation, immune dysregulation, and vascular remodeling. Recent studies have unveiled a potential link between the gut microbiome and PAH pathogenesis, suggesting that microbial dysbiosis and increased intestinal permeability may contribute to the inflammatory pathology in PAH and ultimately disease progression. This perspective highlights the emerging evidence of the role of leaky gut in PAH, the interplay between microbiota-induced immune responses, and the activation of endogenous retroviruses like human endogenous retrovirus K. Understanding these complex interactions opens new interdisciplinary avenues for research and therapeutic interventions, potentially transforming PAH management through microbiome-targeted strategies.

Gecko-inspired dry adhesives have shown great potential in the field of robotics. However, there is still a large gap between current artificial adhesive-based grippers and natural geckos, especially in terms of precise and fast control of adhesion, which is an important capability for robotic gripper systems, since the targets to be gripped may vary in size and weight (including thin, fragile, soft, and deformable), and manipulation must be fast to meet high productivity requirements. Here, we propose a robotic gripper that is able to switch adhesion rapidly (in less than 0.5 s) to grasp and release objects of various sizes and weights (such as glass substrates, fragile silicon wafers, and deformable polyethylene terephthalate films) by mimicking the self-peeling behavior of gecko toe pads. The gripper retains the fast and stable manipulation of the conventional mechanical gripper, which is more reliable and has a higher load capacity than stimulus-responsive switchable adhesives. Systematic experimental and theoretical studies provide insights into the construction and analysis of the self-peeling model and mechanism to identify certain crucial parameters affecting the self-peeling behavior. Furthermore, a strategy for active adhesion control (i.e., precise adhesion modulation) is integrated by introducing a preset peeling angle θ_B, providing the gripper with a quantitative criterion for adjusting the adhesion strength (0 to 82.77 kPa) according to the requirements of practical applications. The gripper has great potential to be an alternative end-operating gripper for robotic systems, opening an avenue for the development of robotic manipulation.

Proteins play a critical role in biology and biopharma due to their specificity and minimal side effects. Predicting the effects of mutations on protein stability is vital but experimentally challenging. Deep learning offers an efficient solution to this problem. In the present work, we introduced ProstaNet, a deep learning framework that predicts stability changes resulting from single- and multiple-point mutations using geometric vector perceptrons–graph neural network for 3-dimensional feature processing. For training ProstaNet, we meticulously crafted ProstaDB, a comprehensive and pristine thermodynamics repository, including 3,784 single-point mutations and 1,642 multiple-point mutations. We also created thermodynamic looping for enlarging the limited data size of multiple-point mutation and applied an innovative clustering method to generate a standard testing set of multiple-point mutation. Besides, we identified residue scoring as the most important encoding method in protein properties prediction. With these innovations, ProstaNet accurately predicts thermostability changes for both single-point and multiple-point mutations without showing any bias. ProstaNet achieves an accuracy of 0.75, outperforming existing methods for single-point mutation prediction, including ThermoMPNN (0.63), PoPMuSiC^sym (0.66), MUPRO (0.52), and FoldX (0.71). ProstaNet also achieves a 1.3-fold increase in accuracy compared to FoldX for multiple-point mutation predictions. Validated by experiment, 4 out of 5 single-point mutation predictions (80%) and all multiple-point mutation predictions (100%) for HuJ3 mutants were accurate, demonstrating the potential benefits of ProstaNet for protein engineering and drug development.