Effects of Liqi Tongbian decoction on gut microbiota, SCFAs production, and 5-HT pathway in STC rats with Qi Stagnation Pattern.

Slow transit constipation (STC) is a common and debilitating condition characterized by delayed colonic transit and difficulty in fecal expulsion, significantly impacting patients' physical and mental wellbeing as well as their overall quality of life. This study investigates the therapeutic potential of Liqi Tongbian Decoction (LTD) in the treatment of STC, especially in cases involving the context of Qi stagnation, through a multifaceted approach involving the modulation of intestinal flora and short-chain fatty acids (SCFAs). We employed a rat model of STC with Qi Stagnation Pattern, established using the "loperamide + tail-clamping provocation method," to explore the effects of LTD on fecal characteristics, intestinal motility, and colonic pathology. Importantly, LTD exhibited the ability to increase the richness, diversity, and homogeneity of intestinal flora while also modulating the composition of microorganisms. It significantly increased the production of SCFAs, especially butyric acid. Moreover, LTD exerted a substantial influence on the synthesis of serotonin (5-HT) by modulating the expression of tryptophan hydroxylase (TPH) and interacting with the 5-HT4 receptor (5-HT4R), resulting in enhanced colonic motility. Correlation analyses revealed a positive correlation between certain bacterial genera, such as Lachnospiraceae_NK4A136 spp. and Clostridiales spp. and the concentrations of butyric acid and 5-HT. These results suggest a mechanistic link between microbiome composition, SCFAs production, and 5-HT synthesis. These findings highlight the potential of LTD to alleviate STC by facilitating a beneficial interplay among intestinal flora, SCFAs production, and 5-HT-mediated colonic motility, providing novel insights into the management of STC with Qi Stagnation Pattern.

Equivariant Line Graph Neural Network for Protein-Ligand Binding Affinity Prediction.

Binding affinity prediction of three-dimensional (3D) protein-ligand complexes is critical for drug repositioning and virtual drug screening. Existing approaches usually transform a 3D protein-ligand complex to a two-dimensional (2D) graph, and then use graph neural networks (GNNs) to predict its binding affinity. However, the node and edge features of the 2D graph are extracted based on invariant local coordinate systems of the 3D complex. As a result, these approaches can not fully learn the global information of the complex, such as the physical symmetry and the topological information of bonds. To address these issues, we propose a novel Equivariant Line Graph Network (ELGN) for binding affinity prediction of 3D protein-ligand complexes. The proposed ELGN firstly adds a super node to the 3D complex, and then builds a line graph based on the 3D complex. After that, ELGN uses a new E(3)-equivariant network layer to pass the messages between nodes and edges based on the global coordinate system of the 3D complex. Experimental results on two real datasets demonstrate the effectiveness of ELGN over several state-of-the-art baselines.

A dual diffusion model enables 3D molecule generation and lead optimization based on target pockets.

Structure-based generative chemistry is essential in computer-aided drug discovery by exploring a vast chemical space to design ligands with high binding affinity for targets. However, traditional in silico methods are limited by computational inefficiency, while machine learning approaches face bottlenecks due to auto-regressive sampling. To address these concerns, we have developed a conditional deep generative model, PMDM, for 3D molecule generation fitting specified targets. PMDM consists of a conditional equivariant diffusion model with both local and global molecular dynamics, enabling PMDM to consider the conditioned protein information to generate molecules efficiently. The comprehensive experiments indicate that PMDM outperforms baseline models across multiple evaluation metrics. To evaluate the applications of PMDM under real drug design scenarios, we conduct lead compound optimization for SARS-CoV-2 main protease (Mpro) and Cyclin-dependent Kinase 2 (CDK2), respectively. The selected lead optimization molecules are synthesized and evaluated for their in-vitro activities against CDK2, displaying improved CDK2 activity.

MARS: a motif-based autoregressive model for retrosynthesis prediction.

MOTIVATION: Retrosynthesis is a critical task in drug discovery, aimed at finding a viable pathway for synthesizing a given target molecule. Many existing approaches frame this task as a graph-generating problem. Specifically, these methods first identify the reaction center, and break a targeted molecule accordingly to generate the synthons. Reactants are generated by either adding atoms sequentially to synthon graphs or by directly adding appropriate leaving groups. However, both of these strategies have limitations. Adding atoms results in a long prediction sequence that increases the complexity of generation, while adding leaving groups only considers those in the training set, which leads to poor generalization. RESULTS: In this paper, we propose a novel end-to-end graph generation model for retrosynthesis prediction, which sequentially identifies the reaction center, generates the synthons, and adds motifs to the synthons to generate reactants. Given that chemically meaningful motifs fall between the size of atoms and leaving groups, our model achieves lower prediction complexity than adding atoms and demonstrates superior performance than adding leaving groups. We evaluate our proposed model on a benchmark dataset and show that it significantly outperforms previous state-of-the-art models. Furthermore, we conduct ablation studies to investigate the contribution of each component of our proposed model to the overall performance on benchmark datasets. Experiment results demonstrate the effectiveness of our model in predicting retrosynthesis pathways and suggest its potential as a valuable tool in drug discovery. AVAILABILITY AND IMPLEMENTATION: All code and data are available at https://github.com/szu-ljh2020/MARS.

Functional outcome following spontaneous intracerebral hemorrhage between ultrasound guidance endoscopic surgery and conventional craniotomy: A retrospective population-based study.

OBJECTIVES: Ultrasound guidance endoscopic surgery (ES) has been widely used in the treatment of cerebral hemorrhage in recent years, but relevant research articles are still scarce. Our study aims to investigate the effect of ES compared with conventional craniotomy (CC) on the postoperative complications, and prognosis of patients with intracerebral hemorrhage. MATERIALS AND METHODS: The clinical data of 1201 patients with ICH treated in our hospital from January 2017 to January 2020 were collected. The t-test, Chi-squared test and Fisher's exact test were used to analyze the clinical baseline data. Among 1021 spontaneous ICH patients, 193 patients who underwent hematoma evacuation were included in the present analysis. RESULTS: The Glasgow Outcome Scale (GOS) score at 6 months had a favorable prognosis in ES group (p = 0.003). ES group had fewer postoperative complications compared with CC group. Operating time and intraoperative blood loss were significantly lower in ES group than CC group (p = 0.001 and p = 0.002). CONCLUSIONS: Our study revealed that receiving ES improved the prognosis of ICH patients. Additionally, endoscopic surgery diminishes operative time, and intraoperative blood loss and reduces the incidence of postoperative complications.

DHX38 enhances proliferation, metastasis, and EMT progression in NSCLC through the G3BP1-mediated MAPK pathway.

BACKGROUND: Non-small cell lung cancer (NSCLC) is a prevalent and aggressive malignancy with limited therapeutic options. Despite advances in treatment, NSCLC remains a major cause of cancer-related death worldwide. Tumor heterogeneity and therapy resistance present challenges in achieving remission. Research is needed to provide molecular insights, identify new targets, and develop personalized therapies to improve outcomes. METHODS: The protein expression level and prognostic value of DHX38 in NSCLC were explored in public databases and NSCLC tissue microarrays. DHX38 knockdown and overexpression cell lines were established to evaluate the role of DHX38 in NSCLC. In vitro and in vivo functional experiments were conducted to assess proliferation and metastasis. To determine the underlying molecular mechanism of DHX38 in human NSCLC, proteins that interact with DHX38 were isolated by IP and identified by LC-MS. KEGG analysis of DHX38-interacting proteins revealed the molecular pathway of DHX38 in human NSCLC. Abnormal pathway activation was verified by Western blot analysis and immunohistochemical (IHC) staining. A molecule-specific inhibitor was further used to explore potential therapeutic targets for NSCLC. The pathway-related target that interacted with DHX38 was verified by co-immunoprecipitation(co-IP) experiments. In cell lines with stable DHX38 overexpression, the target protein was knocked down to explore its complementary effect on DHX38 overexpression-induced tumor promotion. RESULTS: The protein expression of DHX38 was increased in NSCLC, and patients with high DHX38 expression levels had a poor prognosis. In vitro and in vivo experiments showed that DHX38 promoted the proliferation, migration and invasion of human NSCLC cells. DHX38 overexpression caused abnormal activation of the MAPK pathway and promoted epithelial-mesenchymal transition (EMT) in tumours. SCH772984, a novel specific ERK1/2 inhibitor, significantly reduced the increases in cell proliferation, migration and invasion caused by DHX38 overexpression. The co-IP experiments confirmed that DHX38 interacted with the Ras GTPase-activating protein-binding protein G3BP1. DHX38 regulated the expression of G3BP1. Knocking down G3BP1 in cells with stable DHX38 overexpression prevented DHX38-induced tumor cell proliferation, migration and invasion. Silencing G3BP1 reversed the MAPK pathway activation and EMT induced by DHX38 overexpression. CONCLUSION: In NSCLC, DHX38 functions as a tumor promoter. DHX38 modulates G3BP1 expression, leading to the activation of the MAPK signaling pathway, thus promoting tumor cell proliferation, metastasis, and the progression of epithelial-mesenchymal transition (EMT) in non-small cell lung cancer.

A novel CACNA1S gene variant in a child with hypokalemic periodic paralysis: a case report and literature review.

BACKGROUND: The CACNA1S gene encodes the alpha 1 S-subunit of the voltage-gated calcium channel, which is primarily expressed in the skeletal muscle cells. Pathogenic variants of CACNA1S can cause hypokalemic periodic paralysis (HypoPP), malignant hyperthermia susceptibility, and congenital myopathy. We aimed to study the clinical and molecular features of a male child with a CACNA1S variant and depict the molecular sub-regional characteristics of different phenotypes associated with CACNA1S variants. CASE PRESENTATION: We presented a case of HypoPP with recurrent muscle weakness and hypokalemia. Genetic analyses of the family members revealed that the proband had a novel c.497 C > A (p.Ala166Asp) variant of CACNA1S, which was inherited from his father. The diagnosis of HypoPP was established in the proband as he met the consensus diagnostic criteria. The patient and his parents were informed to avoid the classical triggers of HypoPP. The attacks of the patient are prevented by lifestyle changes and nutritional counseling. We also showed the molecular sub-regional location of the variants of CACNA1S which was associated with different phenotypes. CONCLUSIONS: Our results identified a new variant of CACNA1S and expanded the spectrum of variants associated with HypoPP. Early genetic diagnosis can help avoid diagnostic delays, perform genetic counseling, provide proper treatment, and reduce morbidity and mortality.

Association between Sleep and Alzheimer's Disease: A Bibliometric Analysis from 2003 to 2022.

INTRODUCTION: Alzheimer's disease (AD) often presents with sleep disorders, which are also an important risk factor for AD, affecting cognitive function to a certain extent. This study aimed to reveal the current global status, present hotspots, and discuss emerging trends of sleep and AD using a bibliometric approach. METHODS: Research and review articles related to sleep and AD from 2003 to 2022 were extracted from the Web of Science Core Collection. VOSviewer 1.6.18.0, Scimago Graphica, and CiteSpace 6.2.R2 were used to map the productive and highly cited countries, institutions, journals, authors, references, and keywords in the field. RESULTS: Overall, 4,008 publications were included in this bibliometric analysis. The number of publications and citations showed an increasing trend over the past two decades. The USA and China had the largest and second largest, respectively, number of publications and citations and cooperated with other countries more closely. Ancoli-Israel Sonia published the most papers, and Holtzman David M was co-cited most frequently. The most productive journal was Journal of Alzheimer's Disease, and Neurology was the most frequently cited journal. The risk factors, ß-amyloid (Aß), tau, neuroinflammation, astrocytes, glymphatic system, orexin, functional connectivity, and management have been the main research directions of researchers over the past few years and may be the future trend of valuable research. CONCLUSION: We identified hotspots and emerging trends including risk factors, Aß, tau, neuroinflammation, the glymphatic system, orexin, and management, which may help identify new therapeutic targets and improve clinical efficacy of sleep and AD.

Upper brainstem cholinergic neurons project to ascending and descending circuits.

BACKGROUND: Based on their anatomical location, rostral projections of nuclei are classified as ascending circuits, while caudal projections are classified as descending circuits. Upper brainstem neurons participate in complex information processing and specific sub-populations preferentially project to participating ascending or descending circuits. Cholinergic neurons in the upper brainstem have extensive collateralizations in both ascending and descending circuits; however, their single-cell projection patterns remain unclear because of the lack of comprehensive characterization of individual neurons. RESULTS: By combining fluorescent micro-optical sectional tomography with sparse labeling, we acquired a high-resolution whole-brain dataset of pontine-tegmental cholinergic neurons (PTCNs) and reconstructed their detailed morphology using semi-automatic reconstruction methods. As the main source of acetylcholine in some subcortical areas, individual PTCNs had abundant axons with lengths up to 60 cm and 5000 terminals and innervated multiple brain regions from the spinal cord to the cortex in both hemispheres. Based on various collaterals in the ascending and descending circuits, individual PTCNs were grouped into four subtypes. The morphology of cholinergic neurons in the pedunculopontine nucleus was more divergent, whereas the laterodorsal tegmental nucleus neurons contained richer axonal branches and dendrites. In the ascending circuits, individual PTCNs innervated the thalamus in three different patterns and projected to the cortex via two separate pathways. Moreover, PTCNs targeting the ventral tegmental area and substantia nigra had abundant collaterals in the pontine reticular nuclei, and these two circuits contributed oppositely to locomotion. CONCLUSIONS: Our results suggest that individual PTCNs have abundant axons, and most project to various collaterals in the ascending and descending circuits simultaneously. They target regions with multiple patterns, such as the thalamus and cortex. These results provide a detailed organizational characterization of cholinergic neurons to understand the connexional logic of the upper brainstem.

Hierarchical graph learning for protein-protein interaction.

Protein-Protein Interactions (PPIs) are fundamental means of functions and signalings in biological systems. The massive growth in demand and cost associated with experimental PPI studies calls for computational tools for automated prediction and understanding of PPIs. Despite recent progress, in silico methods remain inadequate in modeling the natural PPI hierarchy. Here we present a double-viewed hierarchical graph learning model, HIGH-PPI, to predict PPIs and extrapolate the molecular details involved. In this model, we create a hierarchical graph, in which a node in the PPI network (top outside-of-protein view) is a protein graph (bottom inside-of-protein view). In the bottom view, a group of chemically relevant descriptors, instead of the protein sequences, are used to better capture the structure-function relationship of the protein. HIGH-PPI examines both outside-of-protein and inside-of-protein of the human interactome to establish a robust machine understanding of PPIs. This model demonstrates high accuracy and robustness in predicting PPIs. Moreover, HIGH-PPI can interpret the modes of action of PPIs by identifying important binding and catalytic sites precisely. Overall, "HIGH-PPI [ https://github.com/zqgao22/HIGH-PPI ]" is a domain-knowledge-driven and interpretable framework for PPI prediction studies.

DeepLogic: Joint Learning of Neural Perception and Logical Reasoning.

Neural-symbolic learning, aiming to combine the perceiving power of neural perception and the reasoning power of symbolic logic together, has drawn increasing research attention. However, existing works simply cascade the two components together and optimize them isolatedly, failing to utilize the mutual enhancing information between them. To address this problem, we propose DeepLogic, a framework with joint learning of neural perception and logical reasoning, such that these two components are jointly optimized through mutual supervision signals. In particular, the proposed DeepLogic framework contains a deep-logic module that is capable of representing complex first-order-logic formulas in a tree structure with basic logic operators. We then theoretically quantify the mutual supervision signals and propose the deep&logic optimization algorithm for joint optimization. We further prove the convergence of DeepLogic and conduct extensive experiments on model performance, convergence, and generalization, as well as its extension to the continuous domain. The experimental results show that through jointly learning both perceptual ability and logic formulas in a weakly supervised manner, our proposed DeepLogic framework can significantly outperform DNN-based baselines by a great margin and beat other strong baselines without out-of-box tools.

Three-dimensional high-resolution anorectal manometry: A comparative pilot study with X-ray defaecography.

Three-dimensional high-resolution anorectal manometry (3DHRAM) is a new technique that can explore anorectal disorders and provide interesting topographic data for the diagnosis of pelvic floor disorders such as paradoxical puborectalis syndrome (PPS). Our object was to evaluate whether 3DHRAM can reliably diagnose PPS already diagnosed with X-ray defaecography, which is considered to be the gold standard. All patients being tested in our department for dyschezia by 3D-HRAM and X-ray defecography were eligible for the study. The 3DHRAM results were compared with X-ray defecography. The sensitivity, specificity, and positive and negative predictive values were calculated for various 3DHRAM criteria to propose a diagnostic strategy for PPS. Twenty-three patients presented with PPS on X-ray defaecography. On 3DHRAM, according to our diagnostic strategy, the kappa value was 0.706, with a positive predictive value of 71.88% [95% CI, 53.02-85.60], a specificity of 80.43% [95% CI, 65.62-90.13], a sensibility of 95.83% [95% CI, 76.98-99.78], and area under curve value was 0.922. In this study, 3DHRAM was used to diagnose PPS with the same degree of reliability as X-ray defaecography, and we confirmed its use in the diagnosis of pelvic floor disorders. Further studies will be necessary to define classifications for these new anatomic data from 3DHRAM.

RetroComposer: Composing Templates for Template-Based Retrosynthesis Prediction.

The main target of retrosynthesis is to recursively decompose desired molecules into available building blocks. Existing template-based retrosynthesis methods follow a template selection stereotype and suffer from limited training templates, which prevents them from discovering novel reactions. To overcome this limitation, we propose an innovative retrosynthesis prediction framework that can compose novel templates beyond training templates. As far as we know, this is the first method that uses machine learning to compose reaction templates for retrosynthesis prediction. Besides, we propose an effective reactant candidate scoring model that can capture atom-level transformations, which helps our method outperform previous methods on the USPTO-50K dataset. Experimental results show that our method can produce novel templates for 15 USPTO-50K test reactions that are not covered by training templates. We have released our source implementation.

Dissection of the long-range projections of specific neurons at the synaptic level in the whole mouse brain.

Through synaptic connections, long-range circuits transmit information among neurons and connect different brain regions to form functional motifs and execute specific functions. Tracing the synaptic distribution of specific neurons requires submicron-level resolution information. However, it is a great challenge to map the synaptic terminals completely because these fine structures span multiple regions, even in the whole brain. Here, we develop a pipeline including viral tracing, sample embedding, fluorescent micro-optical sectional tomography, and big data processing. We mapped the whole-brain distribution and architecture of long projections of the parvalbumin neurons in the basal forebrain at the synaptic level. These neurons send massive projections to multiple downstream regions with subregional preference. With three-dimensional reconstruction in the targeted areas, we found that synaptic degeneration was inconsistent with the accumulation of amyloid-ß plaques but was preferred in memory-related circuits, such as hippocampal formation and thalamus, but not in most hypothalamic nuclei in 8-month-old mice with five familial Alzheimer's disease mutations. Our pipeline provides a platform for generating a whole-brain atlas of cell-type-specific synaptic terminals in the physiological and pathological brain, which can provide an important resource for the study of the organizational logic of specific neural circuits and the circuitry changes in pathological conditions.

The Mesoscopic Connectome of the Cholinergic Pontomesencephalic Tegmentum.

The pontomesencephalic tegmentum, comprising the pedunculopontine nucleus and laterodorsal tegmental nucleus, is involved in various functions via complex connections; however, the organizational structure of these circuits in the whole brain is not entirely clear. Here, combining viral tracing with fluorescent micro-optical sectional tomography, we comprehensively investigated the input and output circuits of two cholinergic subregions in a continuous whole-brain dataset. We found that these nuclei receive abundant input with similar spatial distributions but with different quantitative measures and acquire similar neuromodulatory afferents from the ascending reticular activation system. Meanwhile, these cholinergic nuclei project to similar targeting areas throughout multiple brain regions and have different spatial preferences in 3D. Moreover, some cholinergic connections are unidirectional, including projections from the pedunculopontine nucleus and laterodorsal tegmental nucleus to the ventral posterior complex of the thalamus, and have different impacts on locomotion and anxiety. These results reveal the integrated cholinergic connectome of the midbrain, thus improving the present understanding of the organizational structure of the pontine-tegmental cholinergic system from its anatomical structure to its functional modulation.

Towards Accurate and Compact Architectures via Neural Architecture Transformer.

Designing effective architectures is one of the key factors behind the success of deep neural networks. Existing deep architectures are either manually designed or automatically searched by some Neural Architecture Search (NAS) methods. However, even a well-designed/searched architecture may still contain many nonsignificant or redundant modules/operations (e.g., some intermediate convolution or pooling layers). Such redundancy may not only incur substantial memory consumption and computational cost but also deteriorate the performance. Thus, it is necessary to optimize the operations inside an architecture to improve the performance without introducing extra computational cost. To this end, we have proposed a Neural Architecture Transformer (NAT) method which casts the optimization problem into a Markov Decision Process (MDP) and seeks to replace the redundant operations with more efficient operations, such as skip or null connection. Note that NAT only considers a small number of possible replacements/transitions and thus comes with a limited search space. As a result, such a small search space may hamper the performance of architecture optimization. To address this issue, we propose a Neural Architecture Transformer++ (NAT++) method which further enlarges the set of candidate transitions to improve the performance of architecture optimization. Specifically, we present a two-level transition rule to obtain valid transitions, i.e., allowing operations to have more efficient types (e.g., convolution → separable convolution) or smaller kernel sizes (e.g., 5×5 → 3×3). Note that different operations may have different valid transitions. We further propose a Binary-Masked Softmax (BMSoftmax) layer to omit the possible invalid transitions. Last, based on the MDP formulation, we apply policy gradient to learn an optimal policy, which will be used to infer the optimized architectures. Extensive experiments show that the transformed architectures significantly outperform both their original counterparts and the architectures optimized by existing methods.

Supervised Learning Achieves Human-Level Performance in MOBA Games: A Case Study of Honor of Kings.

We present JueWu-SL, the first supervised-learning-based artificial intelligence (AI) program that achieves human-level performance in playing multiplayer online battle arena (MOBA) games. Unlike prior attempts, we integrate the macro-strategy and the micromanagement of MOBA-game-playing into neural networks in a supervised and end-to-end manner. Tested on Honor of Kings, the most popular MOBA at present, our AI performs competitively at the level of High King players in standard 5v5 games.

Graph Convolutional Module for Temporal Action Localization in Videos.

Temporal action localization, which requires a machine to recognize the location as well as the category of action instances in videos, has long been researched in computer vision. The main challenge of temporal action localization lies in that videos are usually long and untrimmed with diverse action contents involved. Existing state-of-the-art action localization methods divide each video into multiple action units (i.e., proposals in two-stage methods and segments in one-stage methods) and then perform action recognition/regression on each of them individually, without explicitly exploiting their relations during learning. In this paper, we claim that the relations between action units play an important role in action localization, and a more powerful action detector should not only capture the local content of each action unit but also allow a wider field of view on the context related to it. To this end, we propose a general graph convolutional module (GCM) that can be easily plugged into existing action localization methods, including two-stage and one-stage paradigms. To be specific, we first construct a graph, where each action unit is represented as a node and their relations between two action units as an edge. Here, we use two types of relations, one for capturing the temporal connections between different action units, and the other one for characterizing their semantic relationship. Particularly for the temporal connections in two-stage methods, we further explore two different kinds of edges, one connecting the overlapping action units and the other one connecting surrounding but disjointed units. Upon the graph we built, we then apply graph convolutional networks (GCNs) to model the relations among different action units, which is able to learn more informative representations to enhance action localization. Experimental results show that our GCM consistently improves the performance of existing action localization methods, including two-stage methods (e.g., CBR [15] and R-C3D [47]) and one-stage methods (e.g., D-SSAD [22]), verifying the generality and effectiveness of our GCM. Moreover, with the aid of GCM, our approach significantly outperforms the state-of-the-art on THUMOS14 (50.9 percent versus 42.8 percent). Augmentation experiments on ActivityNet also verify the efficacy of modeling the relationships between action units. The source code and the pre-trained models are available at https://github.com/Alvin-Zeng/GCM.

Disturbance-immune weight sharing for neural architecture search.

Neural architecture search (NAS) has gained increasing attention in the community of architecture design. One of the key factors behind the success lies in the training efficiency brought by the weight sharing (WS) technique. However, WS-based NAS methods often suffer from a performance disturbance (PD) issue. That is, the training of subsequent architectures inevitably disturbs the performance of previously trained architectures due to the partially shared weights. This leads to inaccurate performance estimation for the previous architectures, which makes it hard to learn a good search strategy. To alleviate the performance disturbance issue, we propose a new disturbance-immune update strategy for model updating. Specifically, to preserve the knowledge learned by previous architectures, we constrain the training of subsequent architectures in an orthogonal space via orthogonal gradient descent. Equipped with this strategy, we propose a novel disturbance-immune training scheme for NAS. We theoretically analyze the effectiveness of our strategy in alleviating the PD risk. Extensive experiments on CIFAR-10 and ImageNet verify the superiority of our method.

Effects of Qing Hua Chang Yin on lipopolysaccharide­induced intestinal epithelial tight junction injury in Caco­2 cells.

Disruption of the intestinal mucosal barrier integrity is a pathogenic process in inflammatory bowel disease (IBD) development, and is therefore considered a drug discovery target for IBD. The well­known traditional Chinese formulation Qing Hua Chang Yin (QHCY) has been suggested as a potential therapeutic agent for the treatment of ulcerative colitis. However, the possible underlying molecular mechanisms regarding its therapeutic effect remain unclear. Consequently, the present study investigated the effects of QHCY on lipopolysaccharide (LPS)­induced loss of intestinal epithelial barrier integrity in vitro using the Caco­2 cell model of intestinal epithelium. QHCY reversed the LPS­induced decrease in transepithelial electrical resistance and significantly alleviated the increased fluorescently­labeled dextran 4 flux caused by LPS. Moreover, QHCY upregulated the mRNA and protein expression levels of occludin, zona occludens­1 and claudin­1 in LPS­exposed Caco­2 cells. In conclusion, QHCY was able to protect intestinal epithelial barrier integrity following an inflammatory insult; the protective effects of QHCY may be mediated by modulation of the expression of tight junction proteins.