Loss of Trim28 in muscle alters mitochondrial signalling but not systemic metabolism.

Type 2 diabetes mellitus (T2DM), a condition characterised by insulin resistance (IR) and skeletal muscle mitochondrial abnormalities, is a leading cause of death in developed societies. Much work has postulated that improving pathways linked to mitochondrial health, including autophagy, may be a potential avenue to prevent or treat T2DM. Given the recent data indicating a role for tripartite motif-containing 28 (TRIM28) in autophagy and mitochondrial pathways, we investigated whether muscle-specific deletion of TRIM28 might impact on obesity, glucose tolerance, and IR in mice. We studied two different muscle-specific (MCK-cre and ACTA1-cre-ERT2) TRIM28 knockout models, which were phenotyped during and after being fed a chow or high-fat diet (HFD). Whilst muscle-specific deletion of TRIM28 in both models demonstrated alterations in markers of mitochondrial activity and autophagy in skeletal muscle, we did not observe major impacts on the majority of metabolic measures in these mice. Specifically, we demonstrate that deletion of TRIM28 in skeletal muscle of mice during (MCK-cre) or post-development (ACTA1-cre-ERT2) does not prevent HFD-induced obesity or glucose intolerance. These findings are consistent with those reported previously in relation to autophagy and mitochondria in other cell types, and thus warrant further study into the biological role TRIM28 has in relation to mitochondrial function.

Development of Standard X-Ray Beams for Calibration of Radiobiology Cabinet and Conformal Irradiators.

This work seeks to develop standard X-ray beams that are matched to radiobiology X-ray irradiators. The calibration of detectors used for dose determination of these irradiators is performed with a set of standard X rays that are more heavily filtered and/or lower energy, which leads to a higher uncertainty in the dose measurement. Models of the XRad320, SARRP, and the X-ray tube at the University of Wisconsin Medical Radiation Research Center (UWMRRC) were created using the BEAMnrc user code of the EGSnrc Monte Carlo code system. These models were validated against measurements, and the resultant modeled spectra were used to determine the amount of added filtration needed to match the X-ray beams at the UWMRRC to those of the XRad320 and SARRP. The depth profiles and half-value layer (HVL) simulations performed using BEAMnrc agreed to measurements within 3% and 3.6%, respectively. A primary measurement device, a free-air chamber, was developed to measure air kerma in the medium energy range of X rays. The resultant spectra of the matched beams had HVL's that matched the HVL's of the radiobiology irradiators well within the 3% criteria recommended by the International Atomic Energy Agency (IAEA) and the average energies agreed within 2.4%. In conclusion, three standard X-ray beams were developed at the UWMRRC with spectra that more closely match the spectra of the XRad320 and SARRP radiobiology irradiators, which will aid in a more accurate dose determination during calibration of these irradiators.

Nonclosedness of sets of neural networks in Sobolev spaces.

We examine the closedness of sets of realized neural networks of a fixed architecture in Sobolev spaces. For an exactly m-times differentiable activation function ρ, we construct a sequence of neural networks [Formula: see text] whose realizations converge in order-(m-1) Sobolev norm to a function that cannot be realized exactly by a neural network. Thus, sets of realized neural networks are not closed in order-(m-1) Sobolev spaces Wm-1,p for p∈[1,∞). We further show that these sets are not closed in Wm,p under slightly stronger conditions on the mth derivative of ρ. For a real analytic activation function, we show that sets of realized neural networks are not closed in Wk,p for anyk∈N. The nonclosedness allows for approximation of non-network target functions with unbounded parameter growth. We partially characterize the rate of parameter growth for most activation functions by showing that a specific sequence of realized neural networks can approximate the activation function's derivative with weights increasing inversely proportional to the Lp approximation error. Finally, we present experimental results showing that networks are capable of closely approximating non-network target functions with increasing parameters via training.

Deletion of Trim28 in committed adipocytes promotes obesity but preserves glucose tolerance.

The effective storage of lipids in white adipose tissue (WAT) critically impacts whole body energy homeostasis. Many genes have been implicated in WAT lipid metabolism, including tripartite motif containing 28 (Trim28), a gene proposed to primarily influence adiposity via epigenetic mechanisms in embryonic development. However, in the current study we demonstrate that mice with deletion of Trim28 specifically in committed adipocytes, also develop obesity similar to global Trim28 deletion models, highlighting a post-developmental role for Trim28. These effects were exacerbated in female mice, contributing to the growing notion that Trim28 is a sex-specific regulator of obesity. Mechanistically, this phenotype involves alterations in lipolysis and triglyceride metabolism, explained in part by loss of Klf14 expression, a gene previously demonstrated to modulate adipocyte size and body composition in a sex-specific manner. Thus, these findings provide evidence that Trim28 is a bona fide, sex specific regulator of post-developmental adiposity and WAT function.

Singular Values for ReLU Layers.

Despite their prevalence in neural networks, we still lack a thorough theoretical characterization of rectified linear unit (ReLU) layers. This article aims to further our understanding of ReLU layers by studying how the activation function ReLU interacts with the linear component of the layer and what role this interaction plays in the success of the neural network in achieving its intended task. To this end, we introduce two new tools: ReLU singular values of operators and the Gaussian mean width of operators. By presenting, on the one hand, theoretical justifications, results, and interpretations of these two concepts and, on the other hand, numerical experiments and results of the ReLU singular values and the Gaussian mean width being applied to trained neural networks, we hope to give a comprehensive, singular-value-centric view of ReLU layers. We find that ReLU singular values and the Gaussian mean width do not only enable theoretical insights but also provide one with metrics that seem promising for practical applications. In particular, these measures can be used to distinguish correctly and incorrectly classified data as it traverses the network. We conclude by introducing two tools based on our findings: double layers and harmonic pruning.

Fenofibrate Rescues Diabetes-Related Impairment of Ischemia-Mediated Angiogenesis by PPARα-Independent Modulation of Thioredoxin-Interacting Protein.

Fenofibrate, a peroxisome proliferator-activated receptor α (PPARα) agonist, reduces lower limb amputations in patients with type 2 diabetes. The mechanism is, however, unknown. In this study, we demonstrate that fenofibrate markedly attenuates diabetes-related impairment of ischemia-mediated angiogenesis. In a murine model of hindlimb ischemia, daily oral fenofibrate treatment restored diabetes-impaired blood flow recovery, foot movement, hindlimb capillary density, vessel diameter, and vascular endothelial growth factor signaling to nondiabetic levels in both wild-type and PPARα-knockout mice, indicating that these fenofibrate effects are largely PPARα independent. In vitro, fenofibric acid (FFA) rescued high glucose-induced (25 mmol/L) impairment of endothelial cell migration, tubulogenesis, and survival in a PPARα-independent manner. Interestingly, fenofibrate in vivo and FFA in vitro reversed high glucose-induced expression of thioredoxin-interacting protein (TXNIP), an exquisitely glucose-inducible gene previously identified as a critical mediator of diabetes-related impairment in neovascularization. Conversely, adenoviral overexpression of TXNIP abrogated the restorative effects of FFA on high glucose-impaired endothelial cell function in vitro, indicating that the effects of FFA are mediated by TXNIP. We conclude that fenofibrate rescues diabetic impairment in ischemia-mediated angiogenesis, in large part, by PPARα-independent regulation of TXNIP. These findings may therefore explain the reduction in amputations seen in patients with diabetes treated with fenofibrate.

Skeletal muscle-specific overexpression of heat shock protein 72 improves skeletal muscle insulin-stimulated glucose uptake but does not alter whole body metabolism.

AIMS: The induction of heat shock protein 72 (Hsp72) via heating, genetic manipulation or pharmacological activation is metabolically protective in the setting of obesity-induced insulin resistance across mammalian species. In this study, we set out to determine whether the overexpression of Hsp72, specifically in skeletal muscle, can protect against high-fat diet (HFD)-induced obesity and insulin resistance. MATERIALS AND METHODS: An Adeno-Associated Viral vector (AAV), designed to overexpress Hsp72 in skeletal muscle only, was used to study the effects of increasing Hsp72 levels on various metabolic parameters. Two studies were conducted, the first with direct intramuscular (IM) injection of the AAV:Hsp72 into the tibialis anterior hind-limb muscle and the second with a systemic injection to enable body-wide skeletal muscle transduction. RESULTS: IM injection of the AAV:Hsp72 significantly improved skeletal muscle insulin-stimulated glucose clearance in treated hind-limb muscles, as compared with untreated muscles of the contralateral leg when mice were fed an HFD. Despite this finding, systemic administration of AAV:Hsp72 did not improve body composition parameters such as body weight, fat mass or percentage body fat, nor did it lead to an improvement in fasting glucose levels or glucose tolerance. Furthermore, no differences were observed for other metabolic parameters such as whole-body oxygen consumption, energy expenditure or physical activity levels. CONCLUSIONS: At the levels of Hsp72 over-expression reported herein, skeletal muscle-specific Hsp72 overexpression via IM injection has the capacity to increase insulin-stimulated glucose clearance in this muscle. However, upon systemic injection, which results in lower muscle Hsp72 overexpression, no beneficial effects on whole-body metabolism are observed.