IGF2BP2 promotes head and neck squamous carcinoma cell proliferation and growth via the miR-98-5p/PI3K/Akt signaling pathway.

Introduction: As a N6-methyladenosine reader protein, Insulin-like growth factor 2 mRNA-binding protein 2 (IGF2BP2) is a critical player in tumor progression and metastasis. However, its specific function in head and neck squamous carcinoma (HNSCC) has yet to be determined. The present study aimed to determine the role of IGF2BP2 in HNSCC. Methods: The expression of IGF2BP2 in HNSCC was analyzed using The Cancer Genome Atlas (TCGA) dataset and detected in HNSCC tissues and cells, respectively. Gain- and loss- of function methods were employed to study the effects of IGF2BP2 on HNSCC cell proliferation and tumorigenesis in vitro and in vivo. MicroRNAs (miRNAs) regulating IGF2BP2 were predicted using online tools and confirmed experimentally. Results: We showed augmented IGF2BP2 expression in HNSCC, which correlated with poor clinical outcomes. Functional studies showed that IGF2BP2 promoted HNSCC cell proliferation by facilitating cell cycle progression while inhibiting apoptosis. We further demonstrated that IGF2BP2 could enhance HNSCC cell tumorigenesis in vivo. Mechanistically, our data revealed that miR-98-5p could directly target IGF2BP2. The interplay between IGF2BP2 and miR-98-5p is essential to drive the progression of HNSCC via the phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K)-protein kinase B (Akt) pathway signaling pathway. Discussion: The current study revealed the oncogenic role of IGF2BP2 and provided insights into its potential mechanism in HNSCC tumorigenesis. Additionally, IGF2BP2 might represent a promising therapeutic target and serve as prognostic biomarker in patients with HNSCC.

Exploring personalization via federated representation Learning on non-IID data.

Federated Learning (FL) can learn a global model across decentralized data over different clients. However, it is susceptible to statistical heterogeneity of client-specific data. Clients focus on optimizing for their individual target distributions, which would yield divergence of the global model due to inconsistent data distributions. Moreover, federated learning approaches adhere to the scheme of collaboratively learning representations and classifiers, further exacerbating such inconsistency and resulting in imbalanced features and biased classifiers. Hence, in this paper, we propose an independent two-stage personalized FL framework, i.e., Fed-RepPer, to separate representation learning from classification in federated learning. First, the client-side feature representation models are learned using supervised contrastive loss, which enables local objectives consistently, i.e., learning robust representations on distinct data distributions. Local representation models are aggregated into the common global representation model. Then, in the second stage, personalization is studied by learning different classifiers for each client based on the global representation model. The proposed two-stage learning scheme is examined in lightweight edge computing that involves devices with constrained computation resources. Experiments on various datasets (CIFAR-10/100, CINIC-10) and heterogeneous data setups show that Fed-RepPer outperforms alternatives by utilizing flexibility and personalization on non-IID data.

Cellular reaction gene regulation network for swarm robots with pattern formation maneuvering control.

Self-organized pattern formation enables swarm robots to interact with local environments to self-organize into intricate structures generated by gene regulatory network (GRN) control methods without global knowledge. Previous studies have reported that it is challenging to maintain pattern formation stability during maneuvering in the environment due to local morphogenetic reaction rules. Motivated by the mechanism of the GRN in multi-cellular organisms, we propose a novel cellular reaction gene regulatory network (CR-GRN) for pattern formation maneuvering control. In CR-GRN, a cellular reaction network is creatively proposed to depict the robots, environment, virtual target pattern, and their interaction to generate emergent swarm behavior in multi-robot systems. A novel diffusion equation is proposed to simulate the process of morphogen diffusion among cells to ensure stable adaptive pattern generation. In addition, genes, proteins, and morphogens are used to define the internal and external states of cells and form a feedback regulation network. Simulation experiments are conducted to validate the proposed method. The results show that the CR-GRN can satisfy the requirements of turning curvature and maintain the robot's uniformity based on the proposed algorithm. This proves that robots using the CR-GRN can cooperate more effectively to cope in a complicated environment, and maintain a stable formation during maneuvering.

Dendritic complexity change in the triple transgenic mouse model of Alzheimer's disease.

Alzheimer's disease (AD) is an irreversible, neurodegenerative disease that is characterized by memory impairment and executive dysfunction. However, the change of fine structure of neuronal morphology remains unclear in the AD model mouse. In this study, high-resolution mouse brain sectional images were scanned by Micro-Optical Sectioning Tomography (MOST) technology and reconstructed three-dimensionally to obtain the pyramidal neurons. The method of Sholl analysis was performed to analyze the neurons in the brains of 6- and 12-month-old AD mice. The results showed that dendritic complexity was not affected in the entorhinal cortex between 6-month-old mice and 12-month-old mice. The dendritic complexity had increased in the primary motor cortex and CA1 region of hippocampus of 12- month-old mice compared with 6-month-old mice. On the contrary, dendritic complexity in the prefrontal cortex was decreased significantly between 6-month-old and 12-month-old mice. To our knowledge, this is the first study to provide high-resolution brain images of triple transgenic AD mice for statistically analyzing neuronal dendrite complexity by MOST technology to reveal the morphological changes of neurons during AD progression.

Past climate change and recent anthropogenic activities affect genetic structure and population demography of the greater long-tailed hamster in northern China.

The genetic diversity and the spatial structure of a species are likely consequences of both past and recent evolutionary processes, but relevant studies are still rare in East Asia where the Pleistocene climate has unique influences. In this study, we examined the impact of past climate change and recent anthropogenic activities on the genetic structure and population size of the greater long-tailed hamster (Tscherskia triton), an agricultural rodent pest species in northern China. DNA sequence data of 2 mitochondrial genes and genotypic data of 11 microsatellite DNA loci from 41 populations (545 individuals) were gathered. Phylogenetic and population genetic analyses, as well as species distribution modeling and coalescent simulations, were conducted to infer its historical and demographic patterns and processes. Two deeply diverged mitochondrial clades were recovered. A small one was restricted to the Shandong Peninsula while the main clade was further divided into 3 geographic clusters by their microsatellite DNA genotypes: Northwest, North-center and Northeast. Divergence dating indicated a Middle-to-Late Pleistocene divergence between the 2 clades. Demographic analysis indicated that all 3 and pooled populations showed consistent long-period expansions during last glacial period; but not during the Holocene, probably due to the impact of climate warming and human disturbances. Conflicting patterns between mtDNA and microsatellite markers imply an anthropogenic impact on North-center populations due to intensified agricultural cultivation in this region. Our study demonstrated that the impact of past glaciation on organisms in East Asia significantly differs from that of Europe and North America, and human activity is an important factor in determining the genetic diversity of a species, as well as its spatial structure.

Novel AroA with high tolerance to glyphosate, encoded by a gene of Pseudomonas putida 4G-1 isolated from an extremely polluted environment in China.

Glyphosate has been used globally as a safe herbicide for weed control. It inhibits 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase (AroA), which is a key enzyme in the aromatic amino acid biosynthetic pathway in microorganisms and plants. A Pseudomonas putida strain, 4G-1, was isolated from a soil heavily contaminated by glyphosate in China. Its AroA-encoding gene (aroA) has been cloned, sequenced, and expressed in Escherichia coli. Phylogenetic analysis revealed that this AroA belongs neither to class I nor to class II AroA enzymes. When compared with E. coli AroA, 4G-1 AroA shows similar values for K(m)[PEP], K(m)[S3P], and specific enzyme activity. Moreover, 4G-1 AroA exhibits high tolerance to glyphosate, which indicates a protein with a high potential for structural and functional studies of AroA in general and its potential usage for the generation of transgenic crops resistant to the herbicide.