Prostate cancer(PCa)is the second most common malignancy among men globally.The Fu-Zheng-Yi-Liu(FZYL)Formula has been widely utilized in the treatment of PCa.This study investigates whether the FZYL Formula can inhibi...Prostate cancer(PCa)is the second most common malignancy among men globally.The Fu-Zheng-Yi-Liu(FZYL)Formula has been widely utilized in the treatment of PCa.This study investigates whether the FZYL Formula can inhibit PCa by tar-geting the TAMs/CCL5 pathway.We conducted in vitro co-cultures and in vivo co-injections of PCa cells and TAMs to mimic their in-teraction.Results showed that the FZYL Formula significantly reduced the proliferation,colony formation,subpopulations of PCSCs,and sphere-formation efficacy of PCa cells,even in the presence of TAM co-culture.Additionally,the Formula markedly decreased the migration,invasion,and epithelial-mesenchymal transition(EMT)of PCa cells induced by TAMs.The FZYL Formula also reversed M2 phenotype polarization in TAMs and dose-dependently reduced their CCL5 expression and secretion,with minimal cytotoxicity observed.Mechanistic studies confirmed that the TAMs/CCL5 axis is a critical target of the FZYL Formula,as the addition of exogen-ous CCL5 partially reversed the formula’s inhibitory effects on PCSCs self-renewal in the co-culture system.Importantly,the Formula also significantly inhibited the growth of PCa xenografts,bone metastasis,and PCSCs activity in vivo by targeting the TAMs/CCL5 pathway.Overall,this study not only elucidates the immunomodulatory mechanism of the FZYL Formula in PCa therapy but also highlights the TAMs/CCL5 axis as a promising therapeutic target.展开更多
首先回顾了计算视觉发展的历史,介绍了神经元、多层感知机和反向传播等人工神经网络的基本知识以及卷积神经网络的发展史及其卷积、池化等基本操作;讨论了Alex Net、VGGNet、Goog Le Net和Res Net等经典卷积神经网络结构,并重点介绍了Ca...首先回顾了计算视觉发展的历史,介绍了神经元、多层感知机和反向传播等人工神经网络的基本知识以及卷积神经网络的发展史及其卷积、池化等基本操作;讨论了Alex Net、VGGNet、Goog Le Net和Res Net等经典卷积神经网络结构,并重点介绍了CapsNet;总结了卷积神经网络在图像分类、语义分割、目标检测以及图像生成等领域的研究进展;最后提出了卷积神经网络研究所面临的挑战以及对CapsNet未来研究的展望。展开更多
基金supported by the National Natural Science Foundation of China(No.82274512)Guangzhou Science and Technology Project(No.202201020327)+1 种基金Collaborative basic and clinical Innovation project between Guangdong Hospital of Chinese Medicine and the School of Biomedical Sciences of the Chinese University of Hong Kong(No.YN2018HK02)Guangdong basic and Applied basic Research Fund(No.2023A1515110708).
文摘Prostate cancer(PCa)is the second most common malignancy among men globally.The Fu-Zheng-Yi-Liu(FZYL)Formula has been widely utilized in the treatment of PCa.This study investigates whether the FZYL Formula can inhibit PCa by tar-geting the TAMs/CCL5 pathway.We conducted in vitro co-cultures and in vivo co-injections of PCa cells and TAMs to mimic their in-teraction.Results showed that the FZYL Formula significantly reduced the proliferation,colony formation,subpopulations of PCSCs,and sphere-formation efficacy of PCa cells,even in the presence of TAM co-culture.Additionally,the Formula markedly decreased the migration,invasion,and epithelial-mesenchymal transition(EMT)of PCa cells induced by TAMs.The FZYL Formula also reversed M2 phenotype polarization in TAMs and dose-dependently reduced their CCL5 expression and secretion,with minimal cytotoxicity observed.Mechanistic studies confirmed that the TAMs/CCL5 axis is a critical target of the FZYL Formula,as the addition of exogen-ous CCL5 partially reversed the formula’s inhibitory effects on PCSCs self-renewal in the co-culture system.Importantly,the Formula also significantly inhibited the growth of PCa xenografts,bone metastasis,and PCSCs activity in vivo by targeting the TAMs/CCL5 pathway.Overall,this study not only elucidates the immunomodulatory mechanism of the FZYL Formula in PCa therapy but also highlights the TAMs/CCL5 axis as a promising therapeutic target.
文摘随着移动边缘计算(mobile edge computing,MEC)技术的不断演进发展,大量的用户设备分散在边缘服务器密集部署的各个区域内。然而,在任务时延与资源受限的前提下,如何选择合适的服务器进行任务卸载,仍然是一个具有挑战性的难题。研究用户-服务器关联、卸载比例以及资源分配的联合优化问题,在考虑需求和服务异构性下最小化系统能耗。该问题被建模为混合整数非线性规划问题,并分解为用户-服务器关联子问题、卸载率和资源分配子问题进行求解。对于第一个子问题,在同时考虑通信质量与服务类型条件下,利用改良的带权匈牙利算法(Kuhn-Munkres matching algorithm,K-M)实现用户-服务器关联。为了处理第二个高度非凸问题,提出一种有效的双层算法,内层采用拉格朗日对偶法得到计算与通信资源分配;外层采用一维搜索方法得到卸载比例。最后,利用块坐标下降技术交替求解两个子问题,直到收敛。仿真结果表明,与随机算法、贪婪算法和带权匈牙利匹配-本地计算(Kuhn-Munkres matching and local computing,KM-LC)算法相比,文中所提出的算法能有效降低系统能耗。
文摘二维码在产品防伪、追溯等领域中广泛应用,数字水印不仅能有效保护其所属权,而且不影响其正常使用。水印在二维码中具有不可见性、鲁棒性。首先对二维码图像进行分块,并对每个分块进行离散余弦变换(Discrete Cosine Transform,DCT);其次对每一分块再取中心块进行奇异值分解(Singular Value Decomposition,SVD);最后将经Logistic和Arnold变换加密后的水印图像置入到奇异值分解后的特征值中,形成置入水印的二维码。二维码置入水印的逆过程实现了水印的提取。实验证实了二维码中水印能有效抵抗图像多种攻击,具有良好的鲁棒性、安全性和不可见性。
文摘首先回顾了计算视觉发展的历史,介绍了神经元、多层感知机和反向传播等人工神经网络的基本知识以及卷积神经网络的发展史及其卷积、池化等基本操作;讨论了Alex Net、VGGNet、Goog Le Net和Res Net等经典卷积神经网络结构,并重点介绍了CapsNet;总结了卷积神经网络在图像分类、语义分割、目标检测以及图像生成等领域的研究进展;最后提出了卷积神经网络研究所面临的挑战以及对CapsNet未来研究的展望。
