Search

Abstract The invasion-metastasis cascade in head and neck squamous cell carcinoma (HNSCC) is predominantly caused by the interaction between tumor cells and tumor microenvironment, including hypoxia as well as stromal cells. However, the mechanism of hypoxia-activated tumor-stroma crosstalk in HNSCC metastasis remains to be deciphered. Here, we demonstrated that HIF1α was upregulated in HNSCC specimens compared with adjacent normal tissues, whose overexpression was associated with lymph node metastasis and predicted unfavorable prognosis. HIF1α expression correlated positively with the levels of miR-5100 as well as α-SMA, the marker of CAFs. Hypoxia/HIF1α regulated transcriptionally miR-5100 to promote the degradation of its target gene QKI, which acts as a tumor suppressor in HNSCC. Hypoxic HNSCC-derived exosomal miR-5100 promoted the activation of CAFs by orchestrating QKI/AKT/STAT3 axis, which further facilitated HNSCC metastasis. Additionally, miR-5100 derived from plasma exosomes indicated HNSCC malignant progression. In conclusion, our findings illuminate a novel HIF1α/miR-5100/QKI pathway in HNSCC metastasis, and suggest that miR-5100 might be a potential biomarker and therapeutic target for HNSCC.

Abstract Ubiquitination is a reversible process that not only controls protein synthesis and degradation, but also is essential for protein transport, localization and biological activity. Deubiquitinating enzyme (DUB) dysfunction leads to various diseases, including cancer. In this study, we aimed to explore the functions and mechanisms of crucial DUBs in head and neck squamous cell carcinoma (HNSCC). Based on bioinformatic analysis and immunohistochemistry detection, YOD1 was identified to be significantly downregulated in HNSCC specimens compared with adjacent normal tissues. Further analysis revealed that reduced YOD1 expression was associated with the malignant progression of HNSCC and indicated poor prognosis. The results of the in vitro and in vivo experiments verified that YOD1 depletion significantly promoted growth, invasion, and epithelial-mesenchymal transition in HNSCC. Mechanistically, YOD1 inhibited the activation of the ERK/β-catenin pathway by suppressing the ubiquitination and degradation of TRIM33, leading to the constriction of HNSCC progression. Overall, our findings reveal the molecular mechanism underlying the role of YOD1 in tumor progression and provide a novel potential therapeutic target for HNSCC treatment.

Introduction Very preterm (VPT) infants may experience varying degrees of neurodevelopmental challenges. Lack of early markers for neurodevelopmental disorders may delay referral to early interventions. The detailed General Movements Assessment (GMA) could help us to identify early markers for VPT infants at risk of atypical neurodevelopmental clinical phenotype in the very early stage of life as soon as possible. Preterm infants with high risk of atypical neurodevelopmental outcomes will have the best possible start to life if early precise intervention in critical developmental windows is allowed.Methods and analysis This is a nationwide, multicentric prospective cohort study that will recruit 577 infants born

The main task of cross-modal analysis is to learn discriminative representation shared across different modalities. In order to pursue aligned representation, conventional approaches tend to construct and optimize a linear projection or train a complex architecture of deep layers, yet it is difficult to compromise between accuracy and efficiency on modeling multimodal data. This paper proposes a novel graph-embedding learning framework implemented by neural networks. The learned embedding directly approximates the cross-modal aligned representation to perform cross-modal retrieval and image classification combining text information. Proposed framework extracts learned representation from a graph model and, simultaneously, trains a classifier under semi-supervised settings. For optimization, unlike previous methods based on the graph Laplacian regularization, a sampling strategy is adopted to generate training pairs to fully explore the inter-modal and intra-modal similarity relationship. Experimental results on various datasets show that the proposed framework outperforms other state-of-the-art methods on crossmodal retrieval. The framework also demonstrates convincing improvements on the new issue of image classification combining text information on Wiki dataset.