Phyllostachys nigra (Lodd. ex Lindl.) derived polysaccharide with enhanced glycolipid metabolism regulation and mice gut microbiome.

This study focuses on the characterization and regulation of glycolipid metabolism of polysaccharides derived from biomass of Phyllostachys nigra (Lodd. ex Lindl.) root (PNr). The extracts from dilute hydrochloric acid, hot water, and 2 % sodium hydroxide solution were characterized through molecular weight, gel permeation chromatography, monosaccharides, Fourier transform infrared, and nuclear magnetic resonance spectroscopy analyses. Polysaccharide from alkali extraction and molecular sieve purification (named as: PNS2A) exhibited optimal inhibitory of 3T3-L1 cellular differentiation and lowered insulin resistance. The PNS2A is made of a hemicellulose-like main chain of →4)-ß-D-Xylp-(1→ that was connected by branches of 4-O-Me-α-GlcAp-(1→, T-α-D-Galp-(1→, T-α-L-Araf-(1→, →2)-α-L-Araf-(1→, as well as ß-D-Glcp-(1→4-ß-D-Glcp-(1→ fragments. Oral delivery of PNS2A in diabetes mice brought down blood glucose and cholesterol levels and regulated glucose and lipid metabolism. PNS2A alleviated diabetes symptoms and body weight and protected liver and kidney function in model animals by altering the gut microbiome. Polysaccharides can be a new approach to develop bamboo resources.

Information theory-guided heuristic progressive multi-view coding.

Multi-view representation learning aims to capture comprehensive information from multiple views of a shared context. Recent works intuitively apply contrastive learning to different views in a pairwise manner, which is still scalable: view-specific noise is not filtered in learning view-shared representations; the fake negative pairs, where the negative terms are actually within the same class as the positive, and the real negative pairs are coequally treated; evenly measuring the similarities between terms might interfere with optimization. Importantly, few works study the theoretical framework of generalized self-supervised multi-view learning, especially for more than two views. To this end, we rethink the existing multi-view learning paradigm from the perspective of information theory and then propose a novel information theoretical framework for generalized multi-view learning. Guided by it, we build a multi-view coding method with a three-tier progressive architecture, namely Information theory-guided heuristic Progressive Multi-view Coding (IPMC). In the distribution-tier, IPMC aligns the distribution between views to reduce view-specific noise. In the set-tier, IPMC constructs self-adjusted contrasting pools, which are adaptively modified by a view filter. Lastly, in the instance-tier, we adopt a designed unified loss to learn representations and reduce the gradient interference. Theoretically and empirically, we demonstrate the superiority of IPMC over state-of-the-art methods.

A Unified Model Using Distantly Supervised Data and Cross-Domain Data in NER.

Named entity recognition (NER) systems are often realized by supervised methods that require large hand-annotated data. When the hand-annotated data is limited, distantly supervised (DS) data and cross-domain (CD) data are usually used separately to improve the performance. The distantly supervised data can provide in-domain dictionary information, and the hand-annotated cross-domain information can be provided by cross-domain data. These two types of information are complemental. However, there are two problems required to be solved before using directly. First, the distantly supervised data may contain a lot of noise. Second, directly using cross-domain data may degrade performance due to the distribution mismatching problem. In this paper, we propose a unified model named PARE (PArtial learning and REinforcement learning). The PARE model can simultaneously use distantly supervised data and cross-domain data as external data. The model uses the partial learning method with a new label strategy to better handle the noise in distantly supervised data. The reinforcement learning method is used to alleviate the distribution mismatching problem in cross-domain data. Experiments in three datasets show that our model outperforms other baseline models. Besides, our model can be used in the situation where no hand-annotated in-domain data is provided.

[Antitumor effect of gene therapy with recombinant adenovirus-mediated human interleukin-2 gene transfer and its mechanism].

OBJECTIVE: To explore the transfection, expression, antitumor effect and mechanism of a replication-deficient adenovirus vector expressing human IL-2 gene (advhIL-2) in a murine H22 hepatocellular carcinoma model. METHODS: KM mice bearing tumor 100-200 mm3 were injected intratumorally with advhIL-2, adv-LacZ or with PBS alone. The tumor evolution was recorded every 3 days. The transfection and expression of the recombinant adenovirus were examined with X-gal staining and RT-PCR respectively. The splenic LAK and CTL activities were assayed with 51Cr 4 hours release. The infiltration of CD4+ and CD8+ T cells in the tumors were analyzed with immunofluorescence. RESULTS: It was found that the recombinant adenovirus can bring about in vivo effective transfection and expression. The continuosly expressive time of advhIL-2 is longer than 12 days. AdvhIL-2 has dose-dependent anti-tumor effect. The mice received a dose of 2 x 10(9) pfu advhIL-2 divided into halves for two injections developed tumor more slowly and survived much longer than the mice treated with PBS (P < 0.01). Moreover, advhIL-2 increased the splenic LAK and CTL activities, CD4+ and CD8+ T cell infiltration in the H22 tumors significantly. CONCLUSION: Adenovirus-mediated hIL-2 gene treatment has significant antitumor activity in pre-established hepatocellular carcinoma in mice. Antitumor immunity might be responsible for the therapeutic effect.