Dietary supplementation with fermentable fiber suppresses adiposity and the associated parameters of metabolic syndrome. Microbiota-generated fiber-derived short-chain fatty acids (SCFAs) and free fatty acid receptors including GPR43 are thought to mediate these effects. We find that while fermentable (inulin), but not insoluble (cellulose), fiber markedly protected mice against high-fat diet (HFD)-induced metabolic syndrome, the effect was not significantly impaired by either inhibiting SCFA production or genetic ablation of GPR43. Rather, HFD decimates gut microbiota, resulting in loss of enterocyte proliferation, leading to microbiota encroachment, low-grade inflammation (LGI), and metabolic syndrome. Enriching HFD with inulin restored microbiota loads, interleukin-22 (IL-22) production, enterocyte proliferation, and antimicrobial gene expression in a microbiota-dependent manner, as assessed by antibiotic and germ-free approaches. Inulin-induced IL-22 expression, which required innate lymphoid cells, prevented microbiota encroachment and protected against LGI and metabolic syndrome. Thus, fermentable fiber protects against metabolic syndrome by nourishing microbiota to restore IL-22-mediated enterocyte function.

Diet-induced obesity is often modeled by comparing mice fed high-fat diet (HFD), which is made from purified ingredients, vs. normal chow diet (NCD), which is a low-fat assemblage of relatively unrefined plant and animal products. The mechanism by which HFD promotes adiposity is complex but thought to involve low-grade inflammation and altered gut microbiota. The goal of this study was to investigate the extent to which HFD-induced adiposity is driven by fat content vs. other factors that differentiate HFD vs. NCD. Mice were fed NCD, HFD, or other compositionally defined diets (CDD), designed to mimic NCD and/or explore the role of HFD components. A range of metabolic parameters reflecting low-grade inflammation and adiposity were assayed. Relative to NCD, HFD, and to a lesser, but, nonetheless, significant extent, CDD induced increased adiposity, indicating both lipid content and other aspects of HFD are obesogenic. Moreover, HFD and CDD induced a rapid and marked loss of cecal and colonic mass. Such CDD-induced effects were not affected by adjusting dietary protein levels/types but could be largely eliminated by exchanging insoluble fiber (cellulose) for soluble fiber (inulin). Replacing cellulose with inulin in HFD also protected mice against decreased intestinal mass, hyperphagia, and increased adiposity. Such beneficial effects of inulin were microbiota dependent, correlated with elevated fecal short-chain fatty acid levels analyzed via 1H-NMR-based metabolomics and were partially recapitulated by administration of short-chain fatty acid. HFD-induced obesity is strongly promoted by its lack of soluble fiber, which supports microbiota-mediated intestinal tissue homeostasis that prevents inflammation driving obesity and metabolic syndrome.

Serum levels of the adipocyte hormone leptin are increased in proportion to body fat stores as a result of increased production in enlarged fat cells from obese subjects. In vitro studies indicate that insulin and glucocorticoids work directly on adipose tissue to upregulate in a synergistic manner leptin mRNA levels and rates of leptin secretion in human adipose tissue over the long term. Thus, the increased leptin expression observed in obesity could result from the chronic hyperinsulinemia and increased cortisol turnover. Superimposed upon the long-term regulation, nutritional status can influence serum leptin over the short term, independent of adiposity. Fasting leads to a gradual decline in serum leptin that is probably attributable to the decline in insulin and the ability of catecholamines to decrease leptin expression, as observed in both in vivo and in vitro studies. In addition, increases in serum leptin occur approximately 4-7 h after meals. Increasing evidence indicates that insulin, in concert with permissive effects of cortisol, can increase serum leptin over this time frame and likely contributes to meal-induced increases in serum leptin. Further research is required to elucidate the cellular and molecular mechanisms underlying short- and long-term nutritional and hormonal regulation of leptin production and secretion.

We investigated the in vitro regulation of leptin expression in adipose tissue from severely obese women and men before and after culture with insulin (7 nM) and/or dexamethasone (25 nM). Leptin mRNA and leptin secretion were two- to threefold higher in subcutaneous vs. omental adipose tissue before culture. Dexamethasone transiently increased leptin mRNA approximately twofold in both depots after 1 day of culture [ P < 0.01 vs. basal (no hormone control)], but leptin secretion was only increased in omental adipose tissue ( P < 0.005 vs. basal). Insulin did not increase leptin mRNA in either depot but increased leptin secretion ∼1.5- to 3-fold in subcutaneous tissue throughout 7 days of culture ( P < 0.05 vs. basal). The combination of insulin and dexamethasone increased leptin mRNA and leptin secretion approximately two- to threefold in both depots at day 1( P < 0.005 vs. basal or insulin) and maintained leptin expression throughout 7 days of culture. We conclude that insulin and glucocorticoid have depot-specific effects and function synergistically as long-term regulators of leptin expression in omental and subcutaneous adipose tissue from obese subjects.

The advent of deep learning has recently led to great successes in various engineering applications. As a prime example, convolutional neural networks, a type of feedforward neural network, now approach human accuracy on visual recognition tasks like image classification and face recognition. However, here we will show that feedforward neural networks struggle to learn abstract visual relations that are effortlessly recognized by non-human primates, birds, rodents and even insects. We systematically study the ability of feedforward neural networks to learn to recognize a variety of visual relations and demonstrate that same-different visual relations pose a particular strain on these networks. Networks fail to learn same-different visual relations when stimulus variability makes rote memorization difficult. Further, we show that learning same-different problems becomes trivial for a feedforward network that is fed with perceptually grouped stimuli. This demonstration and the comparative success of biological vision in learning visual relations suggests that feedback mechanisms such as attention, working memory and perceptual grouping may be the key components underlying human-level abstract visual reasoning.