Unsupervised Domain Adaptation for EM Image Denoising with Invertible Networks.

Electron microscopy (EM) image denoising is critical for visualization and subsequent analysis. Despite the remarkable achievements of deep learning-based non-blind denoising methods, their performance drops significantly when domain shifts exist between the training and testing data. To address this issue, unpaired blind denoising methods have been proposed. However, these methods heavily rely on image-to-image translation and neglect the inherent characteristics of EM images, limiting their overall denoising performance. In this paper, we propose the first unsupervised domain adaptive EM image denoising method, which is grounded in the observation that EM images from similar samples share common content characteristics. Specifically, we first disentangle the content representations and the noise components from noisy images and establish a shared domain-agnostic content space via domain alignment to bridge the synthetic images (source domain) and the real images (target domain). To ensure precise domain alignment, we further incorporate domain regularization by enforcing that: the pseudo-noisy images, reconstructed using both content representations and noise components, accurately capture the characteristics of the noisy images from which the noise components originate, all while maintaining semantic consistency with the noisy images from which the content representations originate. To guarantee lossless representation decomposition and image reconstruction, we introduce disentanglement-reconstruction invertible networks. Finally, the reconstructed pseudo-noisy images, paired with their corresponding clean counterparts, serve as valuable training data for the denoising network. Extensive experiments on synthetic and real EM datasets demonstrate the superiority of our method in terms of image restoration quality and downstream neuron segmentation accuracy. Our code is publicly available at https://github.com/sydeng99/DADn.

Two-stage error detection to improve electron microscopy image mosaicking.

Large-scale electron microscopy (EM) has enabled the reconstruction of brain connectomes at the synaptic level by serially scanning over massive areas of sample sections. The acquired big EM data sets raise the great challenge of image mosaicking at high accuracy. Currently, it simply follows the conventional algorithms designed for natural images, which are usually composed of only a few tiles, using a single type of keypoint feature that would sacrifice speed for stronger performance. Even so, in the process of stitching hundreds of thousands of tiles for large EM data, errors are still inevitable and diverse. Moreover, there has not yet been an appropriate metric to quantitatively evaluate the stitching of biomedical EM images. Here we propose a two-stage error detection method to improve the EM image mosaicking. It firstly uses point-based error detection in combination with a hybrid feature framework to expedite the stitching computation while maintaining high accuracy. Following is the second detection of unresolved errors with a newly designed metric of EM stitched image quality assessment (EMSIQA). The novel detection-based mosaicking pipeline is tested on large EM data sets and proven to be more effective and as accurate when compared with existing methods.

Perivascular neurons instruct 3D vascular lattice formation via neurovascular contact.

The vasculature of the central nervous system is a 3D lattice composed of laminar vascular beds interconnected by penetrating vessels. The mechanisms controlling 3D lattice network formation remain largely unknown. Combining viral labeling, genetic marking, and single-cell profiling in the mouse retina, we discovered a perivascular neuronal subset, annotated as Fam19a4/Nts-positive retinal ganglion cells (Fam19a4/Nts-RGCs), directly contacting the vasculature with perisomatic endfeet. Developmental ablation of Fam19a4/Nts-RGCs led to disoriented growth of penetrating vessels near the ganglion cell layer (GCL), leading to a disorganized 3D vascular lattice. We identified enriched PIEZO2 expression in Fam19a4/Nts-RGCs. Piezo2 loss from all retinal neurons or Fam19a4/Nts-RGCs abolished the direct neurovascular contacts and phenocopied the Fam19a4/Nts-RGC ablation deficits. The defective vascular structure led to reduced capillary perfusion and sensitized the retina to ischemic insults. Furthermore, we uncovered a Piezo2-dependent perivascular granule cell subset for cerebellar vascular patterning, indicating neuronal Piezo2-dependent 3D vascular patterning in the brain.

The Efficacy of Recombinant Human Type III Collagen in the Treatment of Atrophic Scars and Its Impact on p38 MAPK Signaling Pathway Proteins.

Objective: To investigate the effects of recombinant human type III collagen on atrophic scars and its impact on the p38 mitogen-activated protein kinase (p38MAPK) signaling pathway. Methods: A total of 94 patients with atrophic scars admitted to our hospital from March 2020 to October 2022 were selected as subjects and evenly divided into a control group and an observation group. The control group (n = 47) received carbon dioxide fractional laser treatment, while the observation group (n = 47) was treated with recombinant human type III collagen dressings in addition to the laser treatment. Clinical efficacy, scar conditions, skin physiological parameters, serum levels of p38MAPK pathway-related proteins, and inflammatory markers were compared between the two groups. Results: The overall effective rate in the observation group was 95.74%, significantly higher than 74.47% in the control group (P < .05). Before treatment, there was no significant difference in Vancouver Scar Scale (VSS) scores, stratum corneum hydration, and transepidermal water loss between the two groups (P > .05). After treatment, the VSS score in the observation group was significantly lower than in the control group (P < .05). Similarly, prior to treatment, there were no significant differences in serum levels of mitogen-activated protein kinase 1 (MEK1), mitogen-activated protein kinase 2 (MEK2), extracellular signal-regulated kinase 1 (ERK1), and extracellular signal-regulated kinase 2 (ERK2), interleukin-10 (IL-10) and tumor necrosis factor-alpha (TNF-α) between the two groups (P > .05). After treatment, levels of MEK1, MEK2, ERK1, ERK2, IL-10, and TNF-α in the observation group were significantly lower than those in the control group (P < .05). Conclusion: Recombinant human type III collagen significantly improves the treatment of atrophic scars, effectively ameliorating scar conditions and skin physiology. It also regulates the p38MAPK signaling pathway and reduces inflammation.

A review of the pharmacological action and mechanism of natural plant polysaccharides in depression.

Depression is a prevalent mental disorder. However, clinical treatment options primarily based on chemical drugs have demonstrated varying degrees of adverse reactions and drug resistance, including somnolence, nausea, and cognitive impairment. Therefore, the development of novel antidepressant medications that effectively reduce suffering and side effects has become a prominent area of research. Polysaccharides are bioactive compounds extracted from natural plants that possess diverse pharmacological activities and medicinal values. It has been discovered that polysaccharides can effectively mitigate depression symptoms. This paper provides an overview of the pharmacological action and mechanisms, intervention approaches, and experimental models regarding the antidepressant effects of polysaccharides derived from various natural sources. Additionally, we summarize the roles and potential mechanisms through which these polysaccharides prevent depression by regulating neurotransmitters, HPA axis, neurotrophic factors, neuroinflammation, oxidative stress, tryptophan metabolism, and gut microbiota. Natural plant polysaccharides hold promise as adjunctive antidepressants for prevention, reduction, and treatment of depression by exerting their therapeutic effects through multiple pathways and targets. Therefore, this review aims to provide scientific evidence for developing polysaccharide resources as effective antidepressant drugs.

Corrigendum: Learning the heterogeneous representation of brain's structure from serial SEM images using a masked autoencoder.

[This corrects the article DOI: 10.3389/fninf.2023.1118419.].

Novel receptor, mutation, vaccine, and establishment of coping mode for SARS-CoV-2: current status and future.

Since the outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and its resultant pneumonia in December 2019, the cumulative number of infected people worldwide has exceeded 670 million, with over 6.8 million deaths. Despite the marketing of multiple series of vaccines and the implementation of strict prevention and control measures in many countries, the spread and prevalence of SARS-CoV-2 have not been completely and effectively controlled. The latest research shows that in addition to angiotensin converting enzyme II (ACE2), dozens of protein molecules, including AXL, can act as host receptors for SARS-CoV-2 infecting human cells, and virus mutation and immune evasion never seem to stop. To sum up, this review summarizes and organizes the latest relevant literature, comprehensively reviews the genome characteristics of SARS-CoV-2 as well as receptor-based pathogenesis (including ACE2 and other new receptors), mutation and immune evasion, vaccine development and other aspects, and proposes a series of prevention and treatment opinions. It is expected to provide a theoretical basis for an in-depth understanding of the pathogenic mechanism of SARS-CoV-2 along with a research basis and new ideas for the diagnosis and classification, of COVID-19-related disease and for drug and vaccine research and development.

The Effect of Bergenin on Isonicotinic Acid Hydrazide and Rifampicin-Induced Liver Injury Revealed by RNA Sequencing.

Bergenin (BER), a natural component of polyphenols, has a variety of pharmacological activities, especially in improving drug metabolism, reducing cholestasis, anti-oxidative stress and inhibiting inflammatory responses. The aim of this study was to investigate the effects of BER on liver injury induced by isonicotinic acid hydrazide (INH) and rifampicin (RIF) in mice. The mice model of liver injury was established with INH (100 mg/kg)+RIF (100 mg/kg), and then different doses of BER were used to intervene. The pathological morphology and biochemical indicators of mice were detected. Meanwhile, RNA sequencing was performed to screen the differentially expressed genes and signaling pathways. Finally, critical differentially expressed genes were verified by qRT-PCR and Western blot. RNA sequencing results showed that 707 genes were significantly changed in the INH+RIF group compared with the Control group, and 496 genes were significantly changed after the BER intervention. These differentially expressed genes were mainly enriched in the drug metabolism, bile acid metabolism, Nrf2 pathway and TLR4 pathway. The validation results of qRT-PCR and Western blot were consistent with the RNA sequencing. Therefore, BER alleviated INH+RIF-induced liver injury in mice. The mechanism of BER improving INH+RIF-induced liver injury was related to regulating drug metabolism enzymes, bile acid metabolism, Nrf2 pathway and TLR4 pathway.

Whole-transcriptome sequencing with ceRNA regulation network construction and verification in glioblastoma.

OBJECTIVES: To explore the key genes involved in the occurrence and development of glioblastoma (GBM) by analyzing whole-transcriptome sequencing and biologic data from GBM and normal cerebral cortex tissues and to search for important noncoding RNA (ncRNA) molecular markers based on the competitive endogenous RNA (ceRNA) network. METHODS: Ten GBM and normal cerebral cortex tissues were collected for full transcriptome sequencing, screened for differentially expressed (DE) mRNAs, miRNAs, lncRNAs, and circRNAs, and subjected to bioinformatic analysis. We constructed a Protein-Protein Interaction (PPI) network and a circRNA/lncRNA-miRNA-mRNA regulatory network and identified them using reverse transcription-quantitative polymerase chain reaction (RT-qPCR). Finally, The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA) databases were used to validate and conduct a survival analysis of the target genes. RESULTS: A total of 5341 DEmRNAs, 259 DEmiRNAs, 3122 DElncRNAs, and 2135 DEcircRNAs were identified. Enrichment analysis showed that target genes regulated by DEmiRNA, DElncRNA, and DEcircRNA were closely related to chemical synaptic transmission and ion transmembrane transport. A PPI network analysis screened 10 hub genes that directly participate in tumor cell mitosis regulation. In addition, the ceRNA composite network showed that hsa-miR-296-5p and hsa-miR-874-5p were the central nodes of the network, and the reliability of relevant key molecules was successfully verified through RT-qPCR identification and the TCGA database. The CGGA database survival analysis produced 8 DEmRNAs closely related to GBM patient survival prognosis. CONCLUSIONS: This study revealed the important regulatory functions and molecular mechanisms of ncRNA molecules and identified hsa-miR-296-5p and hsa-miR-874-5p as key molecules in the ceRNA network. They may play an important role in GBM pathogenesis, treatment, and prognosis.

Learning the heterogeneous representation of brain's structure from serial SEM images using a masked autoencoder.

Introduction: The exorbitant cost of accurately annotating the large-scale serial scanning electron microscope (SEM) images as the ground truth for training has always been a great challenge for brain map reconstruction by deep learning methods in neural connectome studies. The representation ability of the model is strongly correlated with the number of such high-quality labels. Recently, the masked autoencoder (MAE) has been shown to effectively pre-train Vision Transformers (ViT) to improve their representational capabilities. Methods: In this paper, we investigated a self-pre-training paradigm for serial SEM images with MAE to implement downstream segmentation tasks. We randomly masked voxels in three-dimensional brain image patches and trained an autoencoder to reconstruct the neuronal structures. Results and discussion: We tested different pre-training and fine-tuning configurations on three different serial SEM datasets of mouse brains, including two public ones, SNEMI3D and MitoEM-R, and one acquired in our lab. A series of masking ratios were examined and the optimal ratio for pre-training efficiency was spotted for 3D segmentation. The MAE pre-training strategy significantly outperformed the supervised learning from scratch. Our work shows that the general framework of can be a unified approach for effective learning of the representation of heterogeneous neural structural features in serial SEM images to greatly facilitate brain connectome reconstruction.

A Convenient All-Cell Optical Imaging Method Compatible with Serial SEM for Brain Mapping.

The mammalian brain, with its complexity and intricacy, poses significant challenges for researchers aiming to understand its inner workings. Optical multilayer interference tomography (OMLIT) is a novel, promising imaging technique that enables the mapping and reconstruction of mesoscale all-cell brain atlases and is seamlessly compatible with tape-based serial scanning electron microscopy (SEM) for microscale mapping in the same tissue. However, currently, OMLIT suffers from imperfect coatings, leading to background noise and image contamination. In this study, we introduced a new imaging configuration using carbon spraying to eliminate the tape-coating step, resulting in reduced noise and enhanced imaging quality. We demonstrated the improved imaging quality and validated its applicability through a correlative light-electron imaging workflow. Our method successfully reconstructed all cells and vasculature within a large OMLIT dataset, enabling basic morphological classification and analysis. We also show that this approach can perform effectively on thicker sections, extending its applicability to sub-micron scale slices, saving sample preparation and imaging time, and increasing imaging throughput. Consequently, this method emerges as a promising candidate for high-speed, high-throughput brain tissue reconstruction and analysis. Our findings open new avenues for exploring the structure and function of the brain using OMLIT images.

Cudraflavone B induces human glioblastoma cells apoptosis via ER stress-induced autophagy.

BACKGROUND: Glioblastoma (GBM) is the most common malignant intracranial tumor with a low survival rate. However, only few drugs responsible for GBM therpies, hence new drug development for it is highly required. The natural product Cudraflavone B (CUB) has been reported to potentially kill a variety of tumor cells. Currently, its anit-cancer effect on GBM still remains unknown. Herein, we investigated whether CUB could affect the proliferation and apoptosis of GBM cells to show anti-GBM potential. RESULTS: CUB selectively inhibited cell viability and induced cell apoptosis by activating the endoplasmic reticulum stress (ER stress) related pathway, as well as harnessing the autophagy-related PI3K/mTOR/LC3B signaling pathway. Typical morphological changes of autophagy were also observed in CUB treated cells by microscope and scanning electron microscope (SEM) examination. 4-Phenylbutyric acid (4-PBA), an ER stress inhibitor, restored the CUB-caused alteration in signaling pathway and morphological change. CONCLUSIONS: Our finding suggests that CUB impaired cell growth and induced cell apoptosis of glioblastoma through ER stress and autophagy-related signaling pathways, and it might be an attractive drug for treatment of GBM.

Improving the Neural Segmentation of Blurry Serial SEM Images by Blind Deblurring.

Serial scanning electron microscopy (sSEM) has recently been developed to reconstruct complex largescale neural connectomes, through learning-based instance segmentation. However, blurry images are inevitable amid prolonged automated data acquisition due to imprecision in autofocusing and autostigmation, which impose a great challenge to accurate segmentation of the massive sSEM image data. Recently, learning-based methods, such as adversarial learning and supervised learning, have been proven to be effective for blind EM image deblurring. However, in practice, these methods suffer from the limited training dataset and the underrepresentation of high-resolution decoded features. Here, we propose a semisupervised learning guided progressive decoding network (SGPN) to exploit unlabeled blurry images for training and progressively enrich high-resolution feature representation. The proposed method outperforms the latest deblurring models on real SEM images with much less ground truth input. The improvement of the PSNR and SSIM is 1.04 dB and 0.086, respectively. We then trained segmentation models with deblurred datasets and demonstrated significant improvement in segmentation accuracy. The A-rand (Bogovic et al. 2013) decreased by 0.119 and 0.026, respectively, for 2D and 3D segmentation.

J-domain protein chaperone circuits in proteostasis and disease.

The J-domain proteins (JDP) form the largest protein family among cellular chaperones. In cooperation with the Hsp70 chaperone system, these co-chaperones orchestrate a plethora of distinct functions, including those that help maintain cellular proteostasis and development. JDPs evolved largely through the fusion of a J-domain with other protein subdomains. The highly conserved J-domain facilitates the binding and activation of Hsp70s. How JDPs (re)wire Hsp70 chaperone circuits and promote functional diversity remains insufficiently explained. Here, we discuss recent advances in our understanding of the JDP family with a focus on the regulation built around J-domains to ensure correct pairing and assembly of JDP-Hsp70 machineries that operate on different clientele under various cellular growth conditions.

Combined effects of vitamin C and cold atmospheric plasma-conditioned media against glioblastoma via hydrogen peroxide.

Glioblastoma is the most lethal intracranial malignant tumor, for which the five-year overall survival rate is approximately 5%. Here we explored the therapeutic combination of vitamin C and plasma-conditioned medium on glioblastoma cells in culture and as subcutaneous or intracranial xenografts in mice. The combination treatment reduced cell viability and proliferation while promoting apoptosis, and the effects were significantly stronger than with either treatment on its own. Similar results were obtained in the two xenograft models. Vitamin C appeared to upregulate aquaporin-3 and enhance the uptake of extracellular H2O2, while the combination treatment increased intracellular levels of reactive oxygen species including H2O2 and activated the JNK signaling pathway. The cytotoxic effects of the combination treatment were partially reversed by the specific JNK signaling inhibitor SP600125. Our results suggest that the combination of vitamin C and plasma-conditioned medium has therapeutic potential against glioblastoma, and they provide mechanistic insights that may help investigate this and other potential therapies in greater depth.

The effects of taraxasterol on liver fibrosis revealed by RNA sequencing.

Effective treatment of liver fibrosis remains a challenging medical problem. Taraxasterol (TAR) has anti-inflammatory, anti-tumor and hepatoprotective effects. Studies have shown that TAR has good biological activity against liver injury induced by various factors. However, the anti-fibrotic effect of TAR and its mechanism are never clarified. The purpose of this study was to investigate the effects of TAR in liver fibrosis and to reveal its possible mechanism by RNA sequencing. Our results suggested that TAR attenuated CCl4-induced hepatocyte necrosis, inflammatory infiltration and ECM deposition. TAR inhibited the levels of ALT, AST, ALP, γ-GT, LN, HA, PC III and IV-C in serum and TNF-α, IL-6, IL-1ß and MDA in liver. In addition, TAR increased the activities of SOD and GSH-Px in liver. RNA sequencing analysis of liver tissues revealed that CCl4 and TAR significantly altered 4,155 genes and 2,675 genes, respectively. TAR reversed changes in ECM-related genes. More specifically, TAR mediated the expression of genes related to the activation of the Hippo pathway, while inhibiting the expression of genes related to the activation of HIF-1α, TGF-ß/Smad, and Wnt pathways. In the validation experiments, the qRT-PCR results showed that the expression levels of Yap1, Tead3, Hif1α, Vegfa, Tgfß1, Want3a, and Ctnnb1 mRNA were consistent with the RNA sequencing results. The Western blot results showed that TAR inhibited the levels of TGF-ß1 and p-Smad2. In addition, the results in vitro were consistent with those in vivo. Therefore, we concluded that TAR improved CCl4-induced liver fibrosis by regulating Hippo, HIF-1α, TGF-ß/Smad and Wnt pathways.

Synthesis and Characterization of Fly Ash-Based Geopolymers Activated with Spent Caustic.

The spent caustic with strong alkali first replaced the alkali activator to prepare the geopolymer. The influence of spent caustic to the geopolymer was characterized through compressive strength measurement, XRD, MIP analysis and NMR, and the immobilization efficiency of organic in geopolymer was evaluated through the measurement of total organic carbon (TOC). The results show that the spent caustic can partially replace the alkali activator to prepare the geopolymer, and it shows a better performance than that which was activated with pure NaOH solution when the alkalinity is between 4 mol and 14 mol. The organic matter in the spent alkali can be effectively fixed in the geopolymer, which will hinder the geopolymerization in the initial stage of the polymerization reaction but has little effect on the chemical structure and mechanical properties of the final product. With the degree of alkalinity increasing, the immobilization efficiency is improved, and the maximum can reach 84.5%. The organics in the spent caustic will hinder geopolymerization at the initial stage but has little effect on the chemical structure and mechanical property of the final product. This study proposes a new method for the recycling of spent caustic, which also reduces the preparation cost of geopolymers.

Icariin exhibits protective effects on cisplatin-induced cardiotoxicity via ROS-mediated oxidative stress injury in vivo and in vitro.

BACKGROUND: Cisplatin-induced cardiotoxicity severely limits its clinical application as an antitumor drug and increases the risk of cardiovascular disease. Icariin (ICA), the main flavonoid isolated from Epimedii Folium, has been demonstrated to have various beneficial effects on cardiovascular disease. However, the protective effect of ICA against cisplatin-induced cardiotoxicity remains unclear. PURPOSE: In present study, we explored the protective action of ICA against cisplatin-induced cardiotoxicity and its possible molecular mechanisms in vitro and in vivo. METHODS: Mice were intraperitoneally injected with cisplatin 4 mg/kg every other day for 7 times to establish myocardial injury model. ICA (15, 30 mg/kg) was administered to mice by gavage for 21 days. H9c2 cells were treated with ICA (3, 6, 12 µM) in the presence or absence of cisplatin (40 µM), and then cell viability, oxidative stress, apoptosis, and mitochondrial function were evaluated. RESULTS: Biochemical index detection and histopathological staining analysis showed that ICA had a good protective effect on cisplatin-induced cardiotoxicity. Cellular experiments showed that ICA inhibited cisplatin-induced oxidative stress in a dose-dependent manner by regulating the levels of glutathione peroxidase (GSH-Px), catalase (CAT), superoxide dismutase (SOD) and malondialdehyde (MDA). ICA could inhibit the expression of NF-κB and the secretion of inflammatory factors, thereby alleviating the inflammatory injury caused by cisplatin. In addition, ICA could alleviate cisplatin-induced myocardial injury by activating SIRT1 and PI3K/Akt signaling pathways and inhibiting MAPKs signaling pathway. CONCLUSION: These results suggest that ICA could attenuate cisplatin-induced cardiac injury by inhibiting oxidative stress, inflammation and apoptosis, laying a foundation for ICA to reduce chemotherapy-induced cardiotoxicity in clinical practice.

Morusin Enhances Temozolomide Efficiency in GBM by Inducing Cytoplasmic Vacuolization and Endoplasmic Reticulum Stress.

Glioblastoma multiforme (GBM) is an aggressive brain tumor with high risks of recurrence and mortality. Chemoradiotherapy resistance has been considered a major factor contributing to the extremely poor prognosis of GBM patients. Therefore, there is an urgent need to develop highly effective therapeutic agents. Here, we demonstrate the anti-tumor effect of morusin, a typical prenylated flavonoid, in GBM through in vivo and in vitro models. Morusin showed selective cytotoxicity toward GBM cell lines without harming normal human astrocytes when the concentration was less than 20 µM. Morusin treatment significantly induced apoptosis of GBM cells, accompanied by the activation of endoplasmic reticulum (ER) stress, and the appearance of cytoplasmic vacuolation and autophagosomes in cells. Then, we found the ER stress activation and cytotoxicity of morusin were rescued by ER stress inhibitor 4-PBA. Furthermore, morusin arrested cell cycle at the G1 phase and inhibited cell proliferation of GBM cells through the Akt-mTOR-p70S6K pathway. Dysregulation of ERs and cell cycle in morusin exposed GBM cells were confirmed by RNA-seq analysis. Finally, we demonstrated the combination of morusin and TMZ remarkably enhanced ER stress and displayed a synergistic effect in GBM cells, and suppressed tumor progression in an orthotopic xenograft model. In conclusion, these findings reveal the toxicity of morusin to GBM cells and its ability to enhance drug sensitivity to TMZ, suggesting the potential application value of morusin in the development of therapeutic strategies for human GBM.

USF1/CD90 signaling in maintaining glioblastoma stem cells and tumor-associated macrophages adhesion.

BACKGROUND: Glioblastoma stem cells (GSCs) and their interplay with tumor-associated macrophages (TAMs) are responsible for malignant growth and tumor recurrence of glioblastoma multiforme (GBM), but the underlying mechanisms are largely unknown. METHODS: Cell viability, stemness, migration, and invasion were measured in GSCs after the knockdown of upstream stimulating factor 1 (USF1). Luciferase assay and chromatin immunoprecipitation qPCR were performed to determine the regulation of CD90 by USF1. Immunohistochemistry and immunofluorescent staining were used to examine the expression of USF1 and GSC markers, as well as the crosstalk between GSCs and TAMs. In addition, the interaction between GSCs and TAMs was confirmed using in vivo GBM models. RESULTS: We show that USF1 promotes malignant glioblastoma phenotypes and GSCs-TAMs physical interaction by inducing CD90 expression. USF1 predicts a poor prognosis for glioma patients and is upregulated in patient-derived GSCs and glioblastoma cell lines. USF1 overexpression increases the proliferation, invasion, and neurosphere formation of GSCs and glioblastoma cell lines, while USF1 knockdown exerts an opposite effect. Further mechanistic studies reveal that USF1 promotes GSC stemness by directly regulating CD90 expression. Importantly, CD90 of GSCs functions as an anchor for physical interaction with macrophages. Additionally, the USF1/CD90 signaling axis supports the GSCs and TAMs adhesion and immunosuppressive feature of TAMs, which in turn enhance the stemness of GSCs. Moreover, the overexpression of CD90 restores the stemness property in USF1 knockdown GSCs and its immunosuppressive microenvironment. CONCLUSIONS: Our findings indicate that the USF1/CD90 axis might be a potential therapeutic target for the treatment of glioblastoma.