This study provides new insights and evidences that high level expression of GBP1 is significantly correlated with progression and prognosis in GBMs. Furthermore, transfection of GBP1 revealed its regulation on migration and invasiveness of glioma cells, decreasing sensitivity of chemotherapeutic agent, shortening survival of tumor-bearing animals. These data demonstrate that GBP1 may serve as a novel prognostic biomarker and a potential therapeutic target for gliomas.
Nucleolar and spindle-associated protein (NUSAP1) is a microtubule and chromatin-binding protein that stabilizes microtubules to prevent depolymerization, maintains spindle integrity. NUSAP1 could cross-link spindles into aster-like structures, networks and fibers. It has also been found to play roles in progression of several cancers. However, the potential correlation between NUSAP1 and clinical outcome in patients with glioblastoma multiforme (GBM) remains largely unknown. In the current study, we demonstrated that NUSAP1 was significantly up-regulated in GBM tissues compared with adult non-tumor brain tissues both in a validated cohort and a TCGA cohort. In addition, Kaplan-Meier analysis indicated that patients with high NUSAP1 expression had significantly lower OS (P = 0.0027). Additionally, in the TCGA cohort, NUSAP1 expression was relatively lower in GBM patients within the neural and mesenchymal subtypes compared to other subtypes, and associated with the status of several genetic aberrations such as PTEN deletion and wild type IDH1. The present study provides new insights and evidence that NUSAP1 over-expression was significantly correlated with progression and prognosis of GBM. Furthermore, knockdown of NUSAP1 revealed its regulation on G2/M progression and cell proliferation (both in vitro and in vivo). These data demonstrate that NUSAP1 could serve as a novel prognostic biomarker and a potential therapeutic target for GBM.
Zero-shot cross-domain slot filling alleviates the data dependence in the case of data scarcity in the target domain, which has aroused extensive research. However, as most of the existing methods do not achieve effective knowledge transfer to the target domain, they just fit the distribution of the seen slot and show poor performance on unseen slot in the target domain. To solve this, we propose a novel approach based on prototypical contrastive learning with a dynamic label confusion strategy for zero-shot slot filling. The prototypical contrastive learning aims to reconstruct the semantic constraints of labels, and we introduce the label confusion strategy to establish the label dependence between the source domains and the target domain on-the-fly. Experimental results show that our model achieves significant improvement on the unseen slots, while also set new state-of-the-arts on slot filling task. 1
We investigated whether glioma stem-like cells (GSCs) malignantly transformed bone marrow mesenchymal stem cells (tBMSCs) in the tumor microenvironment. Transplantation of enhanced green fluorescence protein (EGFP)-labeled BMSCs into irradiated athymic nude mice was followed by intracranial injection of red fluorescent protein-expressing glioma stem-like cells (SU3-RFP-GSCs). Singly cloned EGFP-BMSCs, harvested from the intracranial tumors showed TERT overexpression, high proliferation, colony formation and invasiveness in Transwell matrigel assays. Transfection of normal BMSCs with TERT (TERT-BMSCs) enhanced proliferation, colony formation and invasiveness, though these characteristics remained lower than in tBMSCs. The tBMSCs and TERT-BMSCs showed high surface expression of CD44, CD105, CD29 and CD90 and an absence of CD31, CD34, CD45, and CD11b, as in normal BMSCs. Alizarin red S and oil red O staining confirmed tBMSCs and TERT-BMSCs transdifferentiated into osteocytes and adipocytes, respectively. When normal BMSCs were indirectly co-cultured in medium from SU3-RFP-GSCs, they exhibited increased growth and proliferation, suggesting paracrine factors from GSCs induced their malignant transformation. Tumorigenicity assays in athymic nude mice showed that transplanted tBMSCs and TERT-BMSCs generated 100% and 20% subcutaneous tumors, respectively, while normal BMSCs generated no tumors. GSCs thus induce malignant transformation of BMSCs by activating TERT expression in BMSCs.
A novel fluorescence derivatization method combined with HPLC was developed to detect the activity of caspase-3 and -8 in two cell lines (Hela cells and A549 cells) which were activated by low temperature-assisted ultraviolet irradiation (LT-UV), mitomycin C (MMC) and camptothecin during the apoptosis, respectively. Two peptide substrates for either caspase-3 or -8 were designed, of which peptide fragments were obtained by enzymatic modification, followed by fluorescence derivatization. A single fluorescent product was formed when a peptide was heated at 120 °C for 10 min in a neutral aqueous medium (pH 7.0) containing catechol, sodium periodate and sodium borate. Commercial kits for detecting the activity of caspase-3 and -8 were used as a control. The relative activity of the caspases detected by fluorescence derivatization was similar to that obtained by commercial kits, which indicated that the novel method is reliable. The activity assays of recombinant human caspases showed that the novel method provided higher selectivity than that of commercial kits, which proved it to be more accurate for determining the activity of caspases in apoptosis.
Zero-shot cross-domain slot filling alleviates the data dependence in the case of data scarcity in the target domain, which has aroused extensive research. However, as most of the existing methods do not achieve effective knowledge transfer to the target domain, they just fit the distribution of the seen slot and show poor performance on unseen slot in the target domain. To solve this, we propose a novel approach based on prototypical contrastive learning with a dynamic label confusion strategy for zero-shot slot filling. The prototypical contrastive learning aims to reconstruct the semantic constraints of labels, and we introduce the label confusion strategy to establish the label dependence between the source domains and the target domain on-the-fly. Experimental results show that our model achieves significant improvement on the unseen slots, while also set new state-of-the-arts on slot filling task. 1
One less addressed issue of deep reinforcement learning is the lack of generalization capability based on new state and new target, for complex tasks, it is necessary to give the correct strategy and evaluate all possible actions for current state. Fortunately, deep reinforcement learning has enabled enormous progress in both subproblems: giving the correct strategy and evaluating all actions based on the state.In this paper we present an approach called orthogonal policy gradient descent(OPGD) that can make agent learn the policy gradient based on the current state and the actions set, by which the agent can learn a policy network with generalization capability. we evaluate the proposed method on the 3D autonomous driving enviroment TORCS compared with the baseline model, detailed analyses of experimental results and proofs are also given.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.