DAN-PD: Domain adaptive network with parallel decoder for polyp segmentation.

Endoscopy is essential for polyp diagnosis and prevention of colorectal cancer. Many deep learning methods have been proposed to perform automatic semantic segmentation of polyps in endoscopic images. However, labeled training images are always scarce, and the styles of endoscopic images from different medical centers vary greatly. The annotation of medical images requires much effort, and how to make more efficient utilization of the existing labeled data is becoming an increasingly critical issue. Considering the characteristics of polyp segmentation tasks and the need for generalization, we proposed a novel method named DAN-PD based on the Vision Transformer. Moreover, we devised the Teacher Parallel Encoder (TPE) and the Domain-Aware Parallel Decoder (DAPD) for the model. Our design innovatively introduces Unsupervised Domain Adaptation (UDA) methods and adversarial learning strategies to the polyp segmentation task. We conducted four transfer learning experiments with three public polyp image datasets to examine the model's performance. The results shows that our proposed method is ahead of other methods in all experiments and reaches the state-of-the-art level.

Resistance to trophic neurite outgrowth of sensory neurons exposed to insulin.

Insulin offers trophic support through receptors expressed widely on peripheral neurons. In this work, we studied whether peripheral sensory neurons demonstrate resistance to its trophic properties, a property relevant during type 2 diabetes mellitus or following supraphysiological therapy. Insulin receptors were not only localized to neuronal membranes and cytoplasm but also had a unique, previously unrecognized localization to neuronal nuclei. We confirmed that nanomolar doses increased neurite outgrowth of adult sensory neurons, but in response to micromolar doses of insulin, even following a brief 2-h exposure, survival and outgrowth of neurites were blunted. Neurons exposed to picomolar insulin concentrations in their media for 5 days had resistance to the impact of later nanomolar doses of insulin. Using a stripe assay seeded with insulin, neurites were more likely to reject higher doses of insulin. Insulin down-regulated mRNAs of the insulin receptor ß subunit and up-regulated levels of GSK-3ß, both potential mechanisms of insulin resistance, while down-regulating the protein expression of pAkt and pGSK-3ß. Overall, these studies identify neuronal nuclear targeting of insulin and evidence for insulin-induced resistance to its trophic properties. The findings have implications for the understanding of the actions of insulin in the treatment of diabetes and neurological disorders.

Guiding adult Mammalian sensory axons during regeneration.

Misdirection of axons after nerve injury impairs successful regeneration of adult neurons. Investigations of axon guidance in development have provided an understanding of pathfinding, but their relevance to regenerating adult axons is unclear. We investigated adult mammalian axon guidance during regeneration after peripheral nerve injury and focused on the effects of the prototypic guidance molecule nerve growth factor (NGF). Adult rat sensory neurons from dorsal root ganglia that expressed the NGF receptor tropomyosin-related kinase A (trkA) were presented with a point source of NGF in vitro. Naive trkA neurons had no net turning response to NGF, but if they had been preconditioned by a peripheral nerve transection in vivo before culturing, their growth cones were attracted toward the NGF gradient. A laminin substrate was required for this behavior and an anti-trkA antibody interrupted turning. These data demonstrate that injured adult mammalian axons can be guided as they regenerate. Moreover, despite the downregulation of trkA mRNA and protein levels within the dorsal root ganglion after injury, sensory neurons retain and increase trkA protein at the injury site where the regenerating axons are found. This may enhance the axonal response to NGF and allow guidance along an NGF gradient created in vivo in the distal nerve stump.

Late onset Tay-Sachs disease in mice with targeted disruption of the Hexa gene: behavioral changes and pathology of the central nervous system.

Tay-Sachs disease is an autosomal recessive neurodegenerative disease resulting from a block in the hydrolysis of GM2 ganglioside, an intermediate in ganglioside catabolism. The mouse model of Tay-Sachs disease (Hexa -/-) has been described as behaviorally indistinguishable from wild type until at least 1 year of age due to a sialidase-mediated bypass of the metabolic defect that reduces the rate of GM2 ganglioside accumulation. In this study, we have followed our mouse model to over 2 years of age and have documented a significant disease phenotype that is reminiscent of the late onset, chronic form of human Tay-Sachs disease. Onset occurs at 11-12 months of age and progresses slowly, in parallel with increasing storage of GM2 ganglioside. The disease is characterized by hind limb spasticity, weight loss, tremors, abnormal posture with lordosis, possible visual impairment, and, late in the disease, muscle weakness, clasping of the limbs, and myoclonic twitches of the head. Immunodetection of GM2 ganglioside showed that storage varies widely in different regions, but is most intense in pyriform cortex, hippocampus (CA3 field, subiculum), amygdala, hypothalamus (paraventricular supraoptic, ventromedial and arcuate nuclei, and mammilary body), and the somatosensory cortex (layer V) in 1- to 2-year-old mutant mice. We suggest that the Tay-Sachs mouse model is a phenotypically valid model of disease and may provide for a reliable indicator of the impact of therapeutic strategies, in particular geared to the late onset, chronic form of human Tay-Sachs disease.