Interactive attack-defense for generalized person re-identification.

Generalized Person Re-Identification (GReID) aims to develop a model capable of robust generalization across unseen target domains, even with training on a limited set of observed domains. Recently, methods based on the Attack-Defense mechanism are emerging as a prevailing technology to this issue, which treats domain transformation as a type of attack and enhances the model's generalization performance on the target domain by equipping it with a defense module. However, a significant limitation of most existing approaches is their inability to effectively model complex domain transformations, largely due to the separation of attack and defense components. To overcome this limitation, we introduce an innovative Interactive Attack-Defense (IAD) mechanism for GReID. The core of IAD is the interactive learning of two models: an attack model and a defense model. The attack model dynamically generates directional attack information responsive to the current state of the defense model, while the defense model is designed to derive generalizable representations by utilizing a variety of attack samples. The training approach involves a dual process: for the attack model, the aim is to increase the challenge for the defense model in countering the attack; conversely, for the defense model, the focus is on minimizing the effects instigated by the attack model. This interactive framework allows for mutual learning between attack and defense, creating a synergistic learning environment. Our diverse experiments across datasets confirm IAD's effectiveness, consistently surpassing current state-of-the-art methods, and using MSMT17 as the target domain in different protocols resulted in a notable 13.4% improvement in GReID task average Rank-1 accuracy. Code is available at: https://github.com/lhf12278/IAD.

Reinforced GNNs for Multiple Instance Learning.

Multiple instance learning (MIL) trains models from bags of instances, where each bag contains multiple instances, and only bag-level labels are available for supervision. The application of graph neural networks (GNNs) in capturing intrabag topology effectively improves MIL. Existing GNNs usually require filtering low-confidence edges among instances and adapting graph neural architectures to new bag structures. However, such asynchronous adjustments to structure and architecture are tedious and ignore their correlations. To tackle these issues, we propose a reinforced GNN framework for MIL (RGMIL), pioneering the exploitation of multiagent deep reinforcement learning (MADRL) in MIL tasks. MADRL enables the flexible definition or extension of factors that influence bag graphs or GNNs and provides synchronous control over them. Moreover, MADRL explores structure-to-architecture correlations while automating adjustments. Experimental results on multiple MIL datasets demonstrate that RGMIL achieves the best performance with excellent explainability. The code and data are available at https://github.com/RingBDStack/RGMIL.

Focus Affinity Perception and Super-Resolution Embedding for Multifocus Image Fusion.

Despite the fact that there is a remarkable achievement on multifocus image fusion, most of the existing methods only generate a low-resolution image if the given source images suffer from low resolution. Obviously, a naive strategy is to independently conduct image fusion and image super-resolution. However, this two-step approach would inevitably introduce and enlarge artifacts in the final result if the result from the first step meets artifacts. To address this problem, in this article, we propose a novel method to simultaneously achieve image fusion and super-resolution in one framework, avoiding step-by-step processing of fusion and super-resolution. Since a small receptive field can discriminate the focusing characteristics of pixels in detailed regions, while a large receptive field is more robust to pixels in smooth regions, a subnetwork is first proposed to compute the affinity of features under different types of receptive fields, efficiently increasing the discriminability of focused pixels. Simultaneously, in order to prevent from distortion, a gradient embedding-based super-resolution subnetwork is also proposed, in which the features from the shallow layer, the deep layer, and the gradient map are jointly taken into account, allowing us to get an upsampled image with high resolution. Compared with the existing methods, which implemented fusion and super-resolution independently, our proposed method directly achieves these two tasks in a parallel way, avoiding artifacts caused by the inferior output of image fusion or super-resolution. Experiments conducted on the real-world dataset substantiate the superiority of our proposed method compared with state of the arts.

A two-stage fine-tuning method for low-resource cross-lingual summarization.

Cross-lingual summarization (CLS) is the task of condensing lengthy source language text into a concise summary in a target language. This presents a dual challenge, demanding both cross-language semantic understanding (i.e., semantic alignment) and effective information compression capabilities. Traditionally, researchers have tackled these challenges using two types of methods: pipeline methods (e.g., translate-then-summarize) and end-to-end methods. The former is intuitive but prone to error propagation, particularly for low-resource languages. The later has shown an impressive performance, due to multilingual pre-trained models (mPTMs). However, mPTMs (e.g., mBART) are primarily trained on resource-rich languages, thereby limiting their semantic alignment capabilities for low-resource languages. To address these issues, this paper integrates the intuitiveness of pipeline methods and the effectiveness of mPTMs, and then proposes a two-stage fine-tuning method for low-resource cross-lingual summarization (TFLCLS). In the first stage, by recognizing the deficiency in the semantic alignment for low-resource languages in mPTMs, a semantic alignment fine-tuning method is employed to enhance the mPTMs' understanding of such languages. In the second stage, while considering that mPTMs are not originally tailored for information compression and CLS demands the model to simultaneously align and compress, an adaptive joint fine-tuning method is introduced. This method further enhances the semantic alignment and information compression abilities of mPTMs that were trained in the first stage. To evaluate the performance of TFLCLS, a low-resource CLS dataset, named Vi2ZhLow, is constructed from scratch; moreover, two additional low-resource CLS datasets, En2ZhLow and Zh2EnLow, are synthesized from widely used large-scale CLS datasets. Experimental results show that TFCLS outperforms state-of-the-art methods by 18.88%, 12.71% and 16.91% in ROUGE-2 on the three datasets, respectively, even when limited with only 5,000 training samples.

[miRNA-128-3p inhibits malignant behavior of glioma cells by downregulating KLHDC8A expression].

OBJECTIVE: To determine whether miRNA-128-3p regulates malignant biological behavior of glioma cells by targeting KLHDC8A. METHODS: Dual-luciferase reporter assays, qRT-PCR and Western blotting were used to verify the targeting of miRNA-128-3p to KLHDC8A. Edu assay, flow cytometry, Transwell assay and would healing assay were used to determine the effects of changes in miRNA-128-3p and KLHDC8A expression levels on malignant behavior of glioma cells. Rescue experiment was carried out to verify that miRNA-128-3p regulated glioma cell proliferation, apoptosis, invasion and migration by targeting KLHDC8A. RESULTS: The expression level of KLHDC8A was significantly increased in high-grade glioma tissue and was closely related to a poor survival outcome of the patients. Overexpression of KLHDC8A promoted glioma cell proliferation, migration and invasion, and miRNA-128-3p overexpression inhibited proliferative and metastatic capacities of glioma cells. Mechanistically, KLHDC8A expression was directly modulated by miRNA-128-3p, which, by targeting KLHDC8A, inhibited malignant behavior of glioma cells. CONCLUSION: Upregulation of miRNA-128-3p inhibits uncontrolled growth of glioma cells by negatively regulating KLHDC8A expression and its downstream effectors, suggesting that the miRNA-128-3p-KLHDC8A axis may serve as a potential prognostic indicator and a therapeutic target for developing new strategies for glioma treatment.

GPR30 agonist G1 combined with hypothermia alleviates cognitive impairment and anxiety-like behavior after subarachnoid hemorrhage in rats.

INTRODUCTION: This study aimed to investigate the treatment effect of G protein-coupled receptor 30 (GPR30) agonist G1 combined with hypothermia (HT) on cognitive impairment and anxiety-like behavior after subarachnoid hemorrhage (SAH) in rats. METHODS: Fifty male rats were randomly assigned to one of five groups: Sham group, SAH group, SAH + G1 group, SAH + HT group, and SAH + G1 + HT group. The SAH rat model was established by modified endovascular puncture in all groups except the Sham group. Neurological function after the operation was assessed by Garcia scoring. The degree of rat cerebral edema was determined using dry-wet weighing method on the 28th day after operation. Moreover, the behavioral test was performed on rats on the 4th and 28th days after operation. RESULTS: Compared with Sham group, the Garcia score of each SAH rat model group decreased significantly on the first day and thereafter increased gradually. However, the recovery rate of each treatment group was higher than the SAH group (no treatment), and the Garcia score of SAH + G1 + HT group was much higher than the SAH group on the seventh day after operation. In addition, each treatment group could obviously reduce the cerebral edema degree of SAH rats, among which rats in SAH + G1 + HT group had lower cerebral edema degree than SAH + G1 group and SAH + HT group. Behavioral test results showed that the combination of GPR30 agonist G1 and HT markedly improved the learning and memory ability of SAH rats, alleviated their anxiety- and emotion-related behavior, and enhanced their social interaction. CONCLUSION: GPR30 agonist G1 combined with HT reduces cognitive impairment and anxiety-like behavior in rats with SAH.

LncRNA XIST Exacerbates Oxygen-Glucose Deprivation/Reoxygenation-Induced Cerebral Injury Through the miR-25-3p/TRAF3 Axis.

Ischemic stroke causes lethal damage to the brain. Identifying key regulators of OGD/R-induced cerebral injury is important for developing novel therapies for ischemic stroke. HMC3 and SH-SY5Y cells were treated with OGD/R as an in vitro ischemic stroke model. Cell viability and apoptosis were determined via CCK-8 assay and flow cytometry. Inflammatory cytokines were examined by ELISA. Luciferase activity was measured for evaluating the interaction of XIST, miR-25-3p, and TRAF3. Bcl-2, Bax, Bad, cleaved-caspase 3, total caspase 3, and TRAF3 were detected via western blotting. HMC3 and SH-SY5Y cells showed increased XIST expression and decreased miR-25-3p expression following OGD/R. Importantly, silencing of XIST and overexpression of miR-25-3p reduced apoptosis and inflammatory response following OGD/R. Furthermore, XIST worked as a miR-25-3p sponge, and miR-25-3p targeted TRAF3 to suppress its expression. Moreover, the knockdown of TRAF3 ameliorated OGD/R-induced injury. Loss of XIST-mediated protective effects was reversed by overexpression of TRAF3. LncRNA XIST exacerbates OGD/R-induced cerebral damage via sponging miR-25-3p and enhancing TRAF3 expression.

Novel methylation-related long non-coding RNA clinical outcome prediction method: the clinical phenotype and immune infiltration research in low-grade gliomas.

Background: Recent studies have suggested that long non-coding RNAs (lncRNAs) may play crucial role in low-grade glioma; however, the underlying mechanisms linking them to epigenetic methylation remain unclear. Methods: We downloaded expression level data for regulators associated with N1 methyladenosine (m1A), 5-methyladenine (m5C), and N6 methyladenosine (m6A) (M1A/M5C/M6A) methylation from the Cancer Genome Atlas-low-grade glioma (TCGA-LGG) database. We identified the expression patterns of lncRNAs, and selected methylation-related lncRNAs using Pearson correlation coefficient>0.4. Non-negative matrix dimensionality reduction was then used to determine the expression patterns of the methylation-associated lncRNAs. We constructed a weighted gene co-expression network analysis (WGCNA) network to explore the co-expression networks between the two expression patterns. Functional enrichment of the co-expression network was performed to identify biological differences between the expression patterns of different lncRNAs. We also constructed prognostic networks based on the methylation presence in lncRNAs in low-grade gliomas. Results: We identified 44 regulators by literature review. Using a correlation coefficient greater than 0.4, we identified 2330 lncRNAs, among which 108 lncRNAs with independent prognostic values were further screened using univariate Cox regression at P< 0.05. Functional enrichment of the co-expression networks revealed that regulation of trans-synaptic signaling, modulation of chemical synaptic transmission, calmodulin binding, and SNARE binding were mostly enriched in the blue module. The calcium and CA2 signaling pathways were associated with different methylation-related long non-coding chains. Using the Least Absolute Shrinkage Selector Operator (LASSO) regression analysis, we analyzed a prognostic model containing four lncRNAs. The model's risk score was 1.12 *AC012063 + 0.74 * AC022382 + 0.32 * AL049712 + 0.16 * GSEC. Gene set variation analysis (GSVA) revealed significant differences in mismatch repair, cell cycle, WNT signaling pathway, NOTCH signaling pathway, Complement and Cascades, and cancer pathways at different GSEC expression levels. Thus, these results suggest that GSEC may be involved in the proliferation and invasion of low-grade glioma, making it a prognostic risk factor for low-grade glioma. Conclusion: Our analysis identified methylation-related lncRNAs in low-grade gliomas, providing a foundation for further research on lncRNA methylation. We found that GSEC could serve as a candidate methylation marker and a prognostic risk factor for overall survival in low-grade glioma patients. These findings shed light on the underlying mechanisms of low-grade glioma development and may facilitate the development of new treatment strategies.

Bone Marrow Mesenchymal Stem Cell-Derived Exosomal KLF4 Alleviated Ischemic Stroke Through Inhibiting N6-Methyladenosine Modification Level of Drp1 by Targeting lncRNA-ZFAS1.

Ischemic stroke has become a serious public health problem that causes high rates of death and disability. Bone marrow mesenchymal stem cell (BMSC)-derived exosomes have shown promising therapeutic results in IS, while the underlying mechanisms need further investigation. Cell and mice models were established through oxygen-glucose deprivation/reoxygenation (OGD/R) treatment and middle cerebral artery occlusion (MCAO)/reperfusion. Exosomes were isolated from BMSCs. Related gene and protein expression was measured by qRT-PCR, Western blotting, and immunofluorescence analysis. The biological functions of treated cells and tissues were analyzed by MTT, ELISA, JC-1, flow cytometry, TTC staining, or TUNEL staining. The interaction of KLF4/lncRNA-ZFAS1 promoter and lncRNA-ZFAS1/FTO was measured by ChIP, dual-luciferase reporter, or RIP assays. The m6A levels of Drp1 were measured by MeRIP-PCR. Mitochondrial staining and transmission electron microscopy (TEM) were used to evaluate the mitochondrial morphology in N2a cells and brain tissues. BMSC-derived exosomes increased the viability of neuronal cells treated with OGD/R while decreasing LDH release, oxidative stress, mitochondrial injury, and apoptosis. Furthermore, these effects were abolished by knockdown of exosomal KLF4. KLF4 increased lncRNA-ZFAS1 through binding to its promoter. LncRNA-ZFAS1 overexpression suppressed the m6A levels of Drp1 and reversed the promoting effect of exosomal KLF4 silencing on mitochondrial injury and the imbalance of mitochondrial dynamics by targeting FTO. Exosomal KLF4 alleviated the infarct area, neuronal injury, and apoptosis in MCAO mice through lncRNA-ZFAS1/FTO/Drp1 axis. BMSC-derived exosomal KLF4 promoted lncRNA-ZFAS1 expression to repress Drp1 m6A modification by targeting FTO, thus reducing mitochondrial dysfunction and alleviating neuronal injury in ischemic stroke.

Exosomal ZEB1 Derived from Neural Stem Cells Reduces Inflammation Injury in OGD/R-Treated Microglia via the GPR30-TLR4-NF-κB Axis.

Ischemic stroke (IS) is the most common type of stroke and the second leading cause of death overall. Neural stem cells play protective roles in IS, but the underlying mechanism remains to be determined. Neural stem cells (NSC) were obtained from the fetal brain tissue of C57BL/6J mice. NSC-derived exosomes (NSC-Exos) were identified in the conditioned medium. Internalization of NSC-Exos was analyzed by fluorescence microscopy. In vitro microglia ischemic stroke injury model was induced using oxygen glucose deprivation/re-oxygenation (OGD/R) method. Cell viability and inflammation were analyzed by MTT, qPCR, ELISA and Western blotting assay. Interaction between ZEB1 and the promoter of GPR30 was verified by luciferase assay and chromatin immunoprecipitation. NSC-Exos prevented OGD/R-mediated inhibition of cell survival and the production of inflammatory cytokines in microglia cells. NSC-Exos increased ZEB1 expression in OGD/R-treated microglia. Down-regulation of ZEB1 expression in NSC-Exos abolished NSC-Exos' protective effects on OGD/R-treated microglia. ZEB1 bound to the promoter region of GPR30 and promoted its expression. Inhibiting GPR30 reversed NSC-Exos effects on cell viability and inflammation injury in OGD/R-treated microglia. Our study demonstrated that NSC exerted cytoprotective roles through release of exosomal ZEB1,which transcriptionally upregulated GPR30 expression, resulting in a reduction in TLR4/NF-κB pathway-induced inflammation. These findings shed light on NSC-Exos' cytoprotective mechanism and highlighted its potential application in the treatment of IS.

CircUCK2 regulates HECTD1-mediated endothelial-mesenchymal transition inhibition by interacting with FUS and protects the blood-brain barrier in ischemic stroke.

Ischemic stroke (IS) has become a cerebrovascular disease of widespread concern. Overexpression of circUCK2 alleviates neuronal damage in IS. However, the specific regulatory mechanisms of circUCK2 are not fully understood. In this study, we found that circUCK2 and HECT domain E3 ubiquitin ligase 1 (HECTD1) were downregulated in IS models in vitro and in vivo. Overexpression of circUCK2 or HECTD1 inhibited endothelial-mesenchymal transition (EndoMT) and protected the blood-brain barrier (BBB) in transient middle cerebral artery occlusion mice from damage. It was further discovered that circUCK2 regulated HECTD1 expressions by interacting with fused in sarcoma (FUS). Moreover, FUS overexpression partially restored the effect of circUCK2 on EndoMT, and overexpression of HECTD1 weakened the effect of FUS on EndoMT. Collectively, circUCK2 upregulates the expression of HECTD1 by combining with FUS and inhibits EndoMT to alleviate BBB damage in IS both in vivo and in vitro.

OIP5-AS1 Inhibits Oxidative Stress and Inflammation in Ischemic Stroke Through miR-155-5p/IRF2BP2 Axis.

BACKGROUND: Ischemic stroke is a very dangerous disease with high incidence, fatality and disability rate in human beings. Massive evidence has indicated that oxidative stress and inflammation are intimately correlated with progression of ischemic stroke. Additionally, LncRNAs were reported to be involved in ischemic stroke. Here, we aim to explore the effects and molecular mechanism of lncRNA OIP5-AS1 on oxidative stress and inflammation in ischemic stroke. METHODS: HMC3 and SH-SY5Y cells were under the condition of oxygen-glucose deprivation/reoxygenation (OGD/R) treatment to establish cell models of ischemic stroke. Commercial kits were employed to detect the indicators of oxidative stress including ROS, MDA and SOD. The expression of OIP5-AS1, miR-155-5p and IRF2BP2 mRNA was determined using RT-qPCR. The protein levels of inflammatory factors including TNF-α, IL-1ß and IL-6 and IRF2BP2 were assessed by western blot and/or ELISA. Luciferase activity assay was employed to validate their correlations among OIP5-AS1, miR-155-5p and IRF2BP2. RESULTS: In OGD/R-induced HMC3 and SH-SY5Y cells, the expression of OIP5-AS1 and IRF2BP2 was reduced while miR-155-5p was elevated. OGD/R induction promoted oxidative stress and inflammatory response in HMC3 and SH-SY5Y cells, while OIP5-AS1 or IRF2BP2 sufficiency as well as miR-155-5p inhibitor attenuated OGD/R-induced these influences. In addition, IRF2BP2 knockdown abolished the suppressive impacts of OIP5-AS1 overexpression on oxidative stress and inflammatory response in OGD/R-induced HMC3 and SH-SY5Y cells. Mechanistically, OIP5-AS1 enhanced IRF2BP2 expression via sponging miR-155-5p. CONCLUSION: OIP5-AS1 suppressed oxidative stress and inflammatory response to alleviate cell injury caused by OGD/R induction in HMC3 and SH-SY5Y cells through regulating miR-155-5p/IRF2BP2 axis, which might offer novel targeted molecules for ischemic stroke therapy.

BMSC-Derived Exosomal Egr2 Ameliorates Ischemic Stroke by Directly Upregulating SIRT6 to Suppress Notch Signaling.

Exosomes generated by BMSCs contribute to functional recovery in ischemic stroke. However, the regulatory mechanism is largely unknown. Exosomes were isolated from BMSCs. Tube formation, MTT, TUNEL, and flow cytometry assays were applied to examine cell angiogenesis, viability, and apoptosis. Protein and DNA interaction was evaluated by ChIP and luciferase assays. LDH release into the culture medium was examined. Infarction area was evaluated by TTC staining. Immunofluorescence staining was applied to examine CD31 expression. A mouse model of MCAO/R was established. BMSC-derived exosomes attenuated neuronal cell damage and facilitated angiogenesis of brain endothelial cells in response to OGD/R, but these effects were abolished by the knockdown of Egr2. Egr2 directly bound to the promoter of SIRT6 to promote its expression. The incompetency of Egr2-silencing exosomes was reversed by overexpression of SIRT6. Furthermore, SIRT6 inhibited Notch signaling via suppressing Notch1. Overexpression of SIRT6 and inhibition of Notch signaling improved cell injury and angiogenesis in OGD/R-treated cells. BMSC-derived exosomal Egr2 ameliorated MCAO/R-induced brain damage via upregulating SIRT6 to suppress Notch signaling in mice. BMSC-derived exosomes ameliorate OGD/R-induced injury and MCAO/R-caused cerebral damage in mice by delivering Egr2 to promote SIRT6 expression and subsequently suppress Notch signaling. Our study provides a potential exosome-based therapy for ischemic stroke.

Robust Graph Regularized NMF with Dissimilarity and Similarity Constraints for ScRNA-seq Data Clustering.

The notable progress in single-cell RNA sequencing (ScRNA-seq) technology is beneficial to accurately discover the heterogeneity and diversity of cells. Clustering is an extremely important step during the ScRNA-seq data analysis. However, it cannot achieve satisfactory performances by directly clustering ScRNA-seq data due to its high dimensionality and noise. To address these issues, we propose a novel ScRNA-seq data representation model, termed Robust Graph regularized Non-Negative Matrix Factorization with Dissimilarity and Similarity constraints (RGNMF-DS), for ScRNA-seq data clustering. To accurately characterize the structure information of the labeled samples and the unlabeled samples, respectively, the proposed RGNMF-DS model adopts a couple of complementary regularizers (i.e., similarity and dissimilar regularizers) to guide matrix decomposition. In addition, we construct a graph regularizer to discover the local geometric structure hidden in ScRNA-seq data. Moreover, we adopt the l2,1-norm to measure the reconstruction error and thereby effectively improve the robustness of the proposed RGNMF-DS model to the noises. Experimental results on several ScRNA-seq datasets have demonstrated that our proposed RGNMF-DS model outperforms other state-of-the-art competitors in clustering.

WalkGAN: Network Representation Learning With Sequence-Based Generative Adversarial Networks.

Network representation learning, also known as network embedding, aims to learn the low-dimensional representations of vertices while capturing and preserving the network structure. For real-world networks, the edges that represent some important relationships between the vertices of a network may be missed and may result in degenerated performance. The existing methods usually treat missing edges as negative samples, thereby ignoring the true connections between two vertices in a network. To capture the true network structure effectively, we propose a novel network representation learning method called WalkGAN, where random walk scheme and generative adversarial networks (GAN) are incorporated into a network embedding framework. Specifically, WalkGAN leverages GAN to generate the synthetic sequences of the vertices that sufficiently simulate random walk on a network and further learn vertex representations from these vertex sequences. Thus, the unobserved links between the vertices are inferred with high probability instead of treating them as nonexistence. Experimental results on the benchmark network datasets demonstrate that WalkGAN achieves significant performance improvements for vertex classification, link prediction, and visualization tasks.

I2Transformer: Intra- and Inter-Relation Embedding Transformer for TV Show Captioning.

TV show captioning aims to generate a linguistic sentence based on the video and its associated subtitle. Compared to purely video-based captioning, the subtitle can provide the captioning model with useful semantic clues such as actors' sentiments and intentions. However, the effective use of subtitle is also very challenging, because it is the pieces of scrappy information and has semantic gap with visual modality. To organize the scrappy information together and yield a powerful omni-representation for all the modalities, an efficient captioning model requires understanding video contents, subtitle semantics, and the relations in between. In this paper, we propose an Intra- and Inter-relation Embedding Transformer (I2Transformer), consisting of an Intra-relation Embedding Block (IAE) and an Inter-relation Embedding Block (IEE) under the framework of a Transformer. First, the IAE captures the intra-relation in each modality via constructing the learnable graphs. Then, IEE learns the cross attention gates, and selects useful information from each modality based on their inter-relations, so as to derive the omni-representation as the input to the Transformer. Experimental results on the public dataset show that the I2Transformer achieves the state-of-the-art performance. We also evaluate the effectiveness of the IAE and IEE on two other relevant tasks of video with text inputs, i.e., TV show retrieval and video-guided machine translation. The encouraging performance further validates that the IAE and IEE blocks have a good generalization ability. The code is available at https://github.com/tuyunbin/I2Transformer.

miR-138-5p Inhibits the Growth and Invasion of Glioma Cells by Regulating WEE1.

BACKGROUND: Accumulating evidence has demonstrated the role of differentially expressed miRNAs in glioma progression. Our previous bioinformatics analyses revealed a role of miR-138-5p in glioma. miR-138-5p was decreased in various tumors, and He et al. found that miR-138-5p had an inhibitory effect on glioma cells in 2021. However, the role of miR-138-5p in the development of glioma and the underlying mechanism is unknown. In this study, we explored whether miR-138-5p affects the biology of glioma by regulating WEE1 expression. METHODS: miR-138-5p and WEE1 G2 checkpoint kinase (WEE1) RNA and protein expression levels in glioma tissues were detected with qRT-PCR and western blotting, respectively. The effects of miR-138-5p and WEE1 on glioma cell migration and invasion were investigated using Transwell assays. CCK-8 assay was used to measure the effects of miR-138-5p and WEE1 on glioma cell proliferation. The mortality of glioma cells transfected with miR-138-5p and WEE1 was measured with flow cytometry. The relationship between miR-138-5p and WEE1 was explored using a luciferase reporter analysis. RESULTS: Functional studies indicated that overexpression of miR-138-5p suppressed cell proliferation, migration, and invasion and promoted death in glioma cell lines. WEE1 was identified as a target of miR-138-5p, and overexpression of miR-138-5p significantly suppressed the levels of WEE1. Moreover, reintroduction of WEE1 partially abrogated miR-138-5p-induced suppression of motility and invasion in glioma cells. CONCLUSION: The low expression of miR-138-5p in glioma suggests a tumor suppressor role for this miRNA. miR-138-5p suppresses glioma progression by regulating WEE1. These data provide new insights into the molecular mechanism of glioma.

LncRNA CEBPA-AS1 knockdown prevents neuronal apoptosis against oxygen glucose deprivation/reoxygenation by regulating the miR-455/GPER1 axis.

Ischemic stroke (IS) is a common nervous system disease, which is a major cause of disability and death in the world. In present study, we demonstrated a regulatory mechanism of CCAAT/enhancer binding protein-alpha antisense 1 (CEBPA-AS1) in oxygen glucose deprivation/reoxygenation (OGD/R)-induced SH-SY5Y cells, with a focus on neuronal apoptosis. CEBPA-AS1, miR-455, and GPER1 expressions were evaluated by using qRT-PCR and Western blotting. The binding relationship among CEBPA-AS1, miR-455, and GPER1 was determined by a dual luciferase reporter assay. Neuronal viability and apoptosis were examined using MTT and flow cytometry assays, followed by determination of apoptosis-related factors (caspase 3, caspase 8, caspase 9, Bax, and Bcl-2). CEBPA-AS1 and GPER1 levels were upregulated, and miR-455 level was downregulated in the cell model of OGD/R induced. CEBPA-AS1 knockdown increased SH-SY5Y viability and reduced OGD/R-induced apoptosis. CEBPA-AS1 could act as a sponge of miR-455, and CEBPA-AS1 knockdown was found to elevate miR-455 expression. miR-455 overexpression also promoted SH-SY5Y cell viability and rescued them from OGD/R-induced apoptosis by binding to GPER1. GPER1 overexpression or miR-455 inhibition reversed the anti-apoptotic effect of CEBPA-AS1 knockdown. These findings suggest a regulatory network of CEBPA-AS1/miR-455/GPER1 that mediates neuronal cell apoptosis in the OGD model, providing a better understanding of pathogenic mechanisms after IS.

miRNA-338-3p inhibits glioma cell proliferation and progression by targeting MYT1L.

Glioma is a common and aggressive primary malignant brain tumor. MicroRNAs (miRNAs) play key roles in the post-transcriptional regulation of gene expression. Currently, miRNAs are considered to be useful biomarkers for the diagnosis and prognosis of glioma. Previously, we screened three differentially expressed miRNAs from Gene Expression Omnibus (GEO) database which included miRNA-338-3p. miRNA-338-3p is involved in tumor development in different cancers. However, in glioma, its function and its underlying mechanism remain unclear. We found that overexpression of miRNA-338-3p suppressed cell proliferation, migration, invasion, and promoted apoptosis of glioma in vitro. Myelin transcription factor 1-like (MYT1L) was found to be a direct target of miRNA-383-3p in glioma cells as the expression of MYT1L was inhibited by overexpressing miRNA-338-3p. Additionally, silencing MYT1L produced similar effects as overexpressing miRNA-338-3p in glioma cells. Overexpression of MYT1L also completely attenuated the inhibitory effect induced by miRNA-338-3p overexpression. These results suggest that the miRNA-338-3p/ MYT1L axis plays a critical role in the progression of glioma. Our study delineates one of the complex molecular mechanisms that drive the growth of glioma and may be useful in finding novel prognostic predictors and treatment targets in glioma. AVAILABILITY OF DATA AND MATERIALS: All data generated or analysed during this study are included in this published article.

A Clustering Method of Case-Involved News by Combining Topic Network and Multi-Head Attention Mechanism.

Finding the news of same case from the large numbers of case-involved news is an important basis for public opinion analysis. Existing text clustering methods usually based on topic models which only use topic and case infomation as the global features of documents, so distinguishing between different cases with similar types remains a challenge. The contents of documents contain rich local features. Taking into account the internal features of news, the information of cases and the contributions provided by different topics, we propose a clustering method of case-involved news, which combines topic network and multi-head attention mechanism. Using case information and topic information to construct a topic network, then extracting the global features by graph convolution network, thus realizing the combination of case information and topic information. At the same time, the local features are extracted by multi-head attention mechanism. Finally, the fusion of global features and local features is realized by variational auto-encoder, and the learned latent representations are used for clustering. The experiments show that the proposed method significantly outperforms the state-of-the-art unsupervised clustering methods.