Node-degree aware edge sampling mitigates inflated classification performance in biomedical random walk-based graph representation learning.

Motivation: Graph representation learning is a family of related approaches that learn low-dimensional vector representations of nodes and other graph elements called embeddings. Embeddings approximate characteristics of the graph and can be used for a variety of machine-learning tasks such as novel edge prediction. For many biomedical applications, partial knowledge exists about positive edges that represent relationships between pairs of entities, but little to no knowledge is available about negative edges that represent the explicit lack of a relationship between two nodes. For this reason, classification procedures are forced to assume that the vast majority of unlabeled edges are negative. Existing approaches to sampling negative edges for training and evaluating classifiers do so by uniformly sampling pairs of nodes. Results: We show here that this sampling strategy typically leads to sets of positive and negative examples with imbalanced node degree distributions. Using representative heterogeneous biomedical knowledge graph and random walk-based graph machine learning, we show that this strategy substantially impacts classification performance. If users of graph machine-learning models apply the models to prioritize examples that are drawn from approximately the same distribution as the positive examples are, then performance of models as estimated in the validation phase may be artificially inflated. We present a degree-aware node sampling approach that mitigates this effect and is simple to implement. Availability and implementation: Our code and data are publicly available at https://github.com/monarch-initiative/negativeExampleSelection.

Adipose retinol saturase is regulated by ß-adrenergic signaling and its deletion impairs lipolysis in adipocytes and acute cold tolerance in mice.

OBJECTIVE: Retinol saturase (RetSat) is an endoplasmic reticulum-localized oxidoreductase highly expressed in organs involved in lipid metabolism such as white (WAT) and brown adipose tissue (BAT). Cold exposure was shown to increase RETSAT protein in BAT but its relevance for non-shivering thermogenesis, a process with beneficial effects on metabolic health, is unknown. METHODS: We analyzed the regulation of RetSat expression in white and brown adipocytes and different murine adipose tissue depots upon ß-adrenergic stimulation and cold exposure. RetSat function during the differentiation and ß-adrenergic stimulation of brown adipocytes was dissected by loss-of-function experiments. Mice with BAT-specific deletion of RetSat were generated and exposed to cold. Gene expression in human WAT was analyzed and the effect of RetSat depletion on adipocyte lipolysis investigated. RESULTS: We show that cold exposure induces RetSat expression in both WAT and BAT of mice via ß-adrenergic signaling. In brown adipocytes, RetSat has minor effects on differentiation but is required for maximal thermogenic gene and protein expression upon ß-adrenergic stimulation and mitochondrial respiration. In mice, BAT-specific deletion of RetSat impaired acute but not long-term adaptation to cold exposure. RetSat expression in subcutaneous WAT of humans correlates with the expression of genes related to mitochondrial function. Mechanistically, we found that RetSat depletion impaired ß-agonist-induced lipolysis, a major regulator of thermogenic gene expression in adipocytes. CONCLUSIONS: Thus, RetSat expression is under ß-adrenergic control and determines thermogenic capacity of brown adipocytes and acute cold tolerance in mice. Modulating RetSat activity may allow for therapeutic interventions towards pathologies with inadequate metabolic activity.

Thrifty energy phenotype predicts weight regain in postmenopausal women with overweight or obesity and is related to FGFR1 signaling.

BACKGROUND&AIMS: Long term improvement of body weight and metabolism is highly requested in obesity. The specific impact of weight loss associated temporary negative energy balance or modified body composition on metabolism and weight regain is unclear. METHODS: We randomly assigned 80 post-menopausal women (BMI 33.9 (32.2-36.8)kg/m2) to an intervention (IG) or control group (CG). IG underwent a dietary three month-weight loss intervention followed by a four week-weight maintenance period without negative energy balance. The CG was instructed to keep their weight stable. Phenotyping was performed at baseline (M0), after weight loss (M3), the maintenance period (M4) and 24-month follow-up (M24). Co-primary outcomes were changes of insulin sensitivity (ISIClamp) and lean body mass (LBM). Energy metabolism and adipose gene expression were secondary endpoints. RESULTS: Between March 2012 and July 2015, 479 subjects were screened for eligibility. 80 subjects were randomly assigned to IG (n = 40) or CG (n = 40). The total number of dropouts was 18 (IG: n = 13, CG: n = 5). LBM and ISIClamp were stable in the CG between M0 and M3, but were changed in the IG at M3 (LBM: -1.4 (95%CI -2.2-(-0.6)) kg and ISIClamp: +0.020 (95%CI 0.012-0.028) mg·kg-1·min-1/(mU·l-1)) (p < 0.01 and p < 0.05 for IG vs. CG, respectively). Effects on LBM, ISIClamp, FM and BMI were preserved until M4. Lower resting energy expenditure per LBM (REELBM) at M3 and stronger difference of REELBM between M3 and M4 (ΔREELBM-M3M4), which indicates a thrifty phenotype, were positively associated with FM regain at M24 (p = 0.022 and p = 0.044, respectively). Gene set enrichment analysis revealed a relationship of this phenotype to weight loss-induced adaption of adipose FGFR1 signaling. CONCLUSION: Negative energy balance had no additional effect on insulin sensitivity. FGFR1 signaling might be involved in the adaption of energy expenditure to temporary negative energy balance, which indicates a thrifty phenotype susceptible to weight regain. TRIAL REGISTRATION: ClinicalTrials.gov number: NCT01105143, https://clinicaltrials.gov/ct2/show/NCT01105143, date of registration: April 16th, 2010.

Effects of Weight Loss on Adipose and Muscular Neuropilin 1 mRNA Expression in Obesity: Potential Implication in SARS-CoV-2 Infections?

INTRODUCTION: Neuropilin 1 (NRP-1) is a novel co-receptor promoting SARS-CoV-2 infectivity. Animal data indicate a role in trans-endothelial lipid transport and storage. As human data are sparse, we aimed to assess the role of NRP-1 in 2 metabolic active tissues in human obesity and in the context of weight loss-induced short- and long-term metabolic changes. METHODS: After a standardized 12-week weight reduction program, 143 subjects (age >18; body mass index ≥27 kg/m2, 78% female) were randomized to a 12-month lifestyle intervention or a control group using a stratified randomization scheme. This was followed by 6-month follow-up without any intervention. Phenotyping was performed before and after weight loss, after 12-month intervention and after subsequent 6 months of follow-up. Tissue-specific insulin sensitivity was estimated by HOMA-IR (whole body and mostly driven by liver), insulin sensitivity index (ISI)Clamp (predominantly skeletal muscle), and free fatty acid (FFA) suppression during hyperinsulinemic-euglycemic clamp (FFASupp) (predominantly adipose tissue). NRP-1 mRNA expression was measured in subcutaneous adipose tissue (NRP-1AT) and skeletal muscle (NRP-1SM) before and after weight loss. RESULTS: NRP-1 was highly expressed in adipose tissue (7,893 [7,303-8,536] counts), but neither NRP-1AT nor NRP-1SM were related to estimates of obesity. Higher NRP-1AT was associated with stronger FFASupp (r = -0.343, p = 0.003) and a tendency to higher ISIClamp (r = 0.202, p = 0.085). Weight loss induced a decline of NRP-1AT but not NRP-1SM. This was more pronounced in subjects with stronger reduction of adipose ACE-2 mRNA expression (r = 0.250; p = 0.032) but was not associated with short- and long-term improvement of FFASupp and ISIClamp. CONCLUSION: NRP-1AT is related to adipose insulin sensitivity in obesity. Weight loss-induced decline of NRP-1AT seems not to be involved in metabolic short- and long-term improvements after weight loss. However, weight loss-induced reduction of both NRP-1AT and ACE-2AT indicates a lower susceptibility of adipose tissue for SARS-CoV-2 after body weight reduction.

Fetuin-B, a potential link of liver-adipose tissue cross talk during diet-induced weight loss-weight maintenance.

BACKGROUND/OBJECTIVES: Numerous hepatokines are involved in inter-organ cross talk regulating tissue-specific insulin sensitivity. Adipose tissue lipolysis represents a crucial element of adipose insulin sensitivity and is substantially involved in long-term body weight regulation after dietary weight loss. Thus, we aimed to analyze the impact of the hepatokine Fetuin-B in the context of weight loss induced short- and long-term modulation of adipose insulin sensitivity. SUBJECTS/METHODS: 143 subjects (age > 18; BMI ≥ 27 kg/m2) were analyzed before (T-3) and after (T0) a standardized 12-week dietary weight reduction program. Afterward, subjects were randomized to a 12-month lifestyle intervention or a control group. After 12 months (T12) no further intervention was performed until 6 months later (T18) (Maintain-Adults trial). Tissue-specific insulin sensitivity was estimated by HOMA-IR (predominantly liver), ISIClamp (predominantly skeletal muscle), and free fatty acid suppression during hyperinsulinemic-euglycemic clamp (FFASupp) (predominantly adipose tissue). Fetuin-B was measured at all concomitant time points. RESULTS: Circulating Fetuin-B levels correlated significantly with estimates of obesity, hepatic steatosis as well as HOMA-IR, ISIClamp, FFASupp at baseline. Fetuin-B decreased during dietary weight loss (4.2 (3.5-4.9) vs. 3.8 (3.2-4.6) µg/ml; p = 2.1 × 10-5). This change was associated with concomitant improvement of HOMA-IR (r = 0.222; p = 0.008) and FFASupp (r = -0.210; p = 0.013), suggesting a particular relationship to hepatic and adipose tissue insulin sensitivity. Weight loss induced improvements of insulin resistance were almost completely preserved until months 12 and 18 and most interestingly, the short and long-term improvement of FFASupp was partially predicted by baseline level of Fetuin-B. CONCLUSIONS: Our data suggest that Fetuin-B might be a potential mediator of liver-adipose cross talk involved in short- and long-term regulation of adipose insulin sensitivity, especially in the context of diet-induced weight changes. TRIAL REGISTRATION: ClinicalTrials.gov number: NCT00850629, https://clinicaltrials.gov/ct2/show/NCT00850629 , date of registration: February 25, 2009.

Metabolic impact of weight loss induced reduction of adipose ACE-2 - Potential implication in COVID-19 infections?

BACKGROUND & AIMS: Angiotensin converting enzyme (ACE)-2 is a modulator of adipose tissue metabolism. However, human data of adipose ACE-2 is rarely available. Considering that, ACE-2 is believed to be the receptor responsible for cell entry of SARS-CoV-2, a better understanding of its regulation is desirable. We therefore characterized the modulation of subcutaneous adipose ACE-2 mRNA expression during weight loss and the impact of ACE-2 expression on weight loss induced short- and long-term improvements of glucose metabolism. METHODS: 143 subjects (age > 18; BMI ≥ 27 kg/m2) were analyzed before and after a standardized 12-week dietary weight reduction program. Afterwards subjects were randomized to a 12-month lifestyle intervention or a control group (Maintain-Adults trial). Insulin sensitivity (IS) was estimated by HOMA-IR (as an estimate of liver IS) and ISIClamp (as an estimate of skeletal muscle IS). ACE-2 mRNA expression (ACE-2AT) was measured in subcutaneous adipose tissue before and after weight loss. RESULTS: ACE-2AT was not affected by obesity, but was reduced in insulin resistant subjects. Weight loss resulted in a decline of ACE-2AT (29.0 (20.0-47.9) vs. 21.0 (13.0-31.0); p = 1.6 ∗ 10-7). A smaller reduction of ACE-2 AT (ΔACE-2AT) was associated with a larger improvement of ISIClamp (p = 0.013) during weight reduction over 3 months, but not with the extend of weight loss. The degree of changes in insulin resistance were preserved until month 12 and was also predicted by the weight loss induced degree of ΔACE-2AT (p = 0.011). CONCLUSIONS: Our data indicate that subcutaneous adipose ACE-2 expression correlates with insulin sensitivity. Weight loss induced decline of subcutaneous adipose ACE-2 expression might affect short- and long-term improvement of myocellular insulin sensitivity, which might be also relevant in the context of ACE-2 downregulation by SARS-CoV-2. TRIAL REGISTRATION: ClinicalTrials.gov number: NCT00850629, https://clinicaltrials.gov/ct2/show/NCT00850629, date of registration: February 25, 2009.

Association between Subcutaneous Adipose Tissue Inflammation, Insulin Resistance, and Calorie Restriction in Obese Females.

The worldwide epidemic of overweight and obesity has led to an increase in associated metabolic comorbidities. Obesity induces chronic low-grade inflammation in white adipose tissue (WAT). However, the function and regulation of both innate and adaptive immune cells in human WAT under conditions of obesity and calorie restriction (CR) is not fully understood yet. Using a randomized interventional design, we investigated postmenopausal overweight or obese female subjects who either underwent CR for 3 mo followed by a 4-wk phase of weight maintenance or had to maintain a stable weight over the whole study period. A comprehensive immune phenotyping protocol was conducted using validated multiparameter flow cytometry analysis in blood and s.c. WAT (SAT). The TCR repertoire was analyzed by next-generation sequencing and cytokine levels were determined in SAT. Metabolic parameters were determined by hyperinsulinemic-euglycemic clamp. We found that insulin resistance correlates significantly with a shift toward the memory T cell compartment in SAT. TCR analysis revealed a diverse repertoire in SAT of overweight or obese individuals. Additionally, whereas weight loss improved systemic insulin sensitivity in the intervention group, SAT displayed no significant improvement of inflammatory parameters (cytokine levels and leukocyte subpopulations) compared with the control group. Our data demonstrate the accumulation of effector memory T cells in obese SAT and an association between systemic glucose homeostasis and inflammatory parameters in obese females. The long-standing effect of obesity-induced changes in SAT was demonstrated by preserved immune cell composition after short-term CR-induced weight loss.