Aberrant proximal tubule DNA methylation underlies phenotypic changes related to kidney dysfunction in patients with diabetes.

Epigenetic mechanisms are considered to contribute to diabetic nephropathy by maintaining memory of poor glycemic control during the early stages of diabetes. However, DNA methylation changes in the human kidney are poorly characterized, because of the lack of cell type-specific analysis. We examined DNA methylation in proximal tubules (PTs) purified from patients with diabetic nephropathy and identified differentially methylated CpG sites, given the critical role of proximal tubules in the kidney injury. Hypermethylation was observed at CpG sites annotated to genes responsible for proximal tubule functions, including gluconeogenesis, nicotinamide adenine dinucleotide synthesis, transporters of glucose, water, phosphate, and drugs, in diabetic kidneys, whereas genes involved in oxidative stress and the cytoskeleton exhibited demethylation. Methylation levels of CpG sites annotated to ACTN1, BCAR1, MYH9, UBE4B, AFMID, TRAF2, TXNIP, FOXO3, and HNF4A were correlated with the estimated glomerular filtration rate, whereas methylation of the CpG site in RUNX1 was associated with interstitial fibrosis and tubular atrophy. Hypermethylation of G6PC and HNF4A was accompanied by decreased expression in diabetic kidneys. Proximal tubule-specific hypomethylation of metabolic genes related to HNF4A observed in control kidneys was compromised in diabetic kidneys, suggesting a role for aberrant DNA methylation in the dedifferentiation process. Multiple genes with aberrant DNA methylation in diabetes overlapped genes with altered expressions in maladaptive proximal tubule cells, including transcription factors PPARA and RREB1. In conclusion, DNA methylation derangement in the proximal tubules of patients with diabetes may drive phenotypic changes, characterized by inflammatory and fibrotic features, along with impaired function in metabolism and transport.NEW & NOTEWORTHY Cell type-specific DNA methylation patterns in the human kidney are not known. We examined DNA methylation in proximal tubules of patients with diabetic nephropathy and revealed that oxidative stress, cytoskeleton, and metabolism genes were aberrantly methylated. The results indicate that aberrant DNA methylation in proximal tubules underlies kidney dysfunction in diabetic nephropathy. Aberrant methylation could be a target for reversing memory of poor glycemic control.

Highly Efficient Blue Fluorescent Organic Light-Emitting Devices Based on λ5-Phosphinine Derivatives.

Although λ5-phosphinine derivatives are known as a promising class of blue fluorescent emitters, those photoluminescent quantum yield (PLQY) values have been reached up to 92 %, however, only a few examples have been explored as an emitter for blue organic light-emitting device (OLED), and the external quantum efficiency (EQE) has been below 2.4 % so far. In this study, we newly developed two types of blue λ5-phosphinine derivatives namely CN-COCF3 and CO2Me-CHO, and investigated the photophysical properties in the solid states. The photophysical analyses in solid state films suggested that the strong electron-accepting nature of these λ5-phosphinine derivatives caused the inferior PLQY values, and the exciplex formation with the host and neighboring materials should be avoided to improve the device efficiency. By choosing suitable host and neighboring materials with deep ionization potentials, we successfully realized efficient blue fluorescent OLEDs with EQE of over 4 % and CIE (0.14, 0.18). This is among the best in λ5-phosphinine-based blue OLEDs so far.

Emergence of integrated behaviors through direct optimization for homeostasis.

Homeostasis is a self-regulatory process, wherein an organism maintains a specific internal physiological state. Homeostatic reinforcement learning (RL) is a framework recently proposed in computational neuroscience to explain animal behavior. Homeostatic RL organizes the behaviors of autonomous embodied agents according to the demands of the internal dynamics of their bodies, coupled with the external environment. Thus, it provides a basis for real-world autonomous agents, such as robots, to continually acquire and learn integrated behaviors for survival. However, prior studies have generally explored problems pertaining to limited size, as the agent must handle observations of such coupled dynamics. To overcome this restriction, we developed an advanced method to realize scaled-up homeostatic RL using deep RL. Furthermore, several rewards for homeostasis have been proposed in the literature. We identified that the reward definition that uses the difference in drive function yields the best results. We created two benchmark environments for homeostasis and performed a behavioral analysis. The analysis showed that the trained agents in each environment changed their behavior based on their internal physiological states. Finally, we extended our method to address vision using deep convolutional neural networks. The analysis of a trained agent revealed that it has visual saliency rooted in the survival environment and internal representations resulting from multimodal input.