Regulators of Food Intake

National Institute Of Diabetes And Digestive And Kidney Diseases

Investigators

Linked publications, trials & patents

Abstract

Energy intake and dietary macronutrient composition is the most difficult part of the energy balance equation to measure. In a natural history study of factors which predict food intake and using an inpatient computerized vending machine system, food intake has been measured over 340 individuals. This study has yielded crucial information about ad libitum food intake and generated important new hypotheses to be tested in this field. We have developed two new outpatient studies to understand variations in ad libitum intake. One plans to recruit 800 participants and will involve collection of multiple “inputs” including physiologic, behavioral, environmental (including gut microbiome measurements) and genetic with ad libitum intake as measured by a buffet meal as the outcome. This will be a data driven study using machine learning to identify novel pathways and determinants of ad libitum intake. The second will investigate whether an unpredictability schema as assessed by current and childhood experiences is associated with overeating using an outpatient ad libitum buffet meal. This study will also investigate whether unpredictability is related to resting metabolic rate and measures of glucose regulation. Understanding the potential confounders of energy intake is vital to this field. We have found that both fat free mass (adjusted for height) and 24-hour energy expenditure are associated with ad libitum food intake. In mediation analysis, energy expenditure accounts for over 80% of the explained variance in energy intake indicating that energy expenditure rather than fat free mass drives energy intake. Thus, increases in metabolic rate (as a weight loss intervention) may have the paradoxical effect of stimulating excess caloric intake. In fact, in further analyses, we have found that energy intake adjusted for energy expenditure predicts weight gain. We were able to quantify how much intake affects the protective effect of EE on weight gain calculating that for every 0.46kg/100 kcal higher EE there is a counterbalance of -23 kg/100 kcal from drive to increased intake. We have completed a study investigating this “energy sensing” link. Participants had measures of energy intake measured after 24 hours of fasting and energy balance during warm (24°C) and cold (19°C); participants also ate ad libitum in our metabolic chamber under the same temperature conditions. We found that following 24 hours of fasting in our respiratory chamber, individuals whose respiratory quotient and carbohydrate oxidation rates dropped less ate more following fasting. This implies that individuals with less metabolic flexibility to fasting have greater drive to energy intake. Energy intake did not increase following cold conditions compared to warm conditions but did increase during cold temperature versus warm temperature conditions. Greater increase in respiratory quotient (RQ) and carbohydrate oxidation rates during cold exposure were associated with increased energy intake following cold exposure but not with changes in energy expenditure. This indicates an effect of cold exposure on concurrent ad libitum intake, but a residual influence of cold exposure on energy intake via an alteration of fuel preference. Previous studies (including our own) have incompletely assessed the variation in energy deficit and surplus to over and underfeeding. A complete assessment requires simultaneous assessment of intake (as exact calories), energy expenditure and calorie loss in urine and stool all of which contribute to the variation in energy deficit or surplus during calorie restriction or overfeeding respectively. We have developed a study that will precisely calculate inter-individual energy deficit or surplus and whether this will predict ad libitum intake. As currently planned, participants will be fed an energy balance, 50% calorie restriction and 200% overfeeding diet in random order while residing for 3 days in our metabolic chamber. Urine and stool calorie loss will be collected as well. Each diet period will be followed by 3 day ad libitum period also conducted in metabolic chamber. The hypotheses are that 1) inter-individual variability in energy deficit will be lower during calorie restriction vs. overfeeding consistent with the Dual Intervention Point model and 2) inter-individual energy deficit or surplus will be associated with subsequent ad libitum intake. The ad libitum vending machine model has demonstrated additional important determinants of ad libitum intake and weight change. We have found that calories from soda intake predicted future weight gain indicating a role for sweetened beverages as contributor to the current obesity epidemic. In work elucidating hormonal associations with ad libitum intake, we found that participants with higher fasting fibroblast growth factor 21 concentrations consume less soda. Higher fasting GLP-1 concentrations were associated with lower carbohydrate and lower intake of high simple sugar/high fat foods and. higher urinary dopamine a potential marker of reward-based food selection was associated with greater drive to eat and increased energy intake. We have also demonstrated that dietary protein has a negative regulatory effect on short term (meal to meal) protein and total calorie intake, but not on daily intake. Consumption of high protein and high fat diet can lead to increased acid accumulation. This can be assessed by measuring the anion gap. We found that anion gap was associated with greater adiposity, higher energy intake and energy expenditure. This indicates that higher dietary acid loads have downstream metabolic effects. We examined secular changes in energy intake on our vending machine paradigm over the past 20 years finding that there has been a gradual decrease in energy intake over this period. We also investigated how urinary fructose and sucrose correlated with sugar intake to further validate these as objective biomarkers of intake. To understand behavioral and psychosocial factors, we have also investigated whether socioeconomic status and food insecurity are associated with energy intake. We found that those with higher food insecurity scores, eat more during the ad libitum period. Moreover, individuals with food insecurity have metabolic risk factors for increased energy intake and weight gain namely higher 24h RQ, higher carbohydrate oxidation, lower lipid oxidation rates and lower fasting concentrations of the appetite suppressing hormone, GLP-1. We have continued investigating novel dietary biomarkers in our study investigating stable isotopes ratios. The study was a multi-factorial design in which dietary pattern varied by meat, soda, and fish content. Plasma (collected every 2 weeks) and hair were analyzed for changes in the stable isotopes C13 and N15. In both plasma and hair, carbon isotope enrichment was associated with meat and soda intake, but more strongly with meat intake, while nitrogen isotope enrichment was associated with fish intake. Stable isotope ratios in specific amino acids and fatty acids may provide even more specific food signatures. We found that carbon isotope ratio (CIR) in plasma alanine had a high sensitivity and specificity for soda intake, while CIR for leucine demonstrated the same for meat intake. We also found that amino acid nitrogen isotope ratios (NIR) can denote diet patterns. NIR in amino acid leucine denoted fish intake, while NIR in amino acid threonine was lower in meat and fish intake. We have also found specific long chain fatty acid signatures that denote meat intake. The CIR of plasma saturated and mono-unsaturated fatty acids are sensitive and specific measures of dietary meat intake, but not SSB or fish intake. We measured the total fatty acid concentration in each participant as well. Increased relative abundances of plasma odd-chain fatty acids i-15:0, 17:0, and 17:1 was associated with meat intake, and higher relative abundances of plasma EPA and DHA (20:5n-3 and 22:6n-3) were associated with fish. In this study, we also found important changes in metabolism notably that those who consumed fish for 12 weeks increased their 24h energy expenditure while those who consumed soda decreased their 24h energy expenditure indicating a role of diet pattern (or quality) on EE independent of macronutrient composition. We are conducting a follow-up study investigating how dose of soda intake (none vs. medium vs. high dose) during an eucaloric diet over a 12-week period affects stable isotope ratios, additional dietary biomarkers (long chain fatty acid profiles, metabolomics), gut microbiome and energy expenditure. This study is currently enrolling participants. An understudied component of food intake is regulation of thirst and thirst perception. Our clinical trial comparing water intake and thirst perception in lean versus individuals with obesity following a 24 hour fast without fluid intake and a 3% saline challenge found that adjusted for body size, lean and obese individuals drank the same amount of water following these challenges, but thirst perception in obese was lower as was their copeptin response to both challenges. Leptin may mediate this lower co-peptin response, but we have not found additional evidence that other hormones purportedly involved in water balance (aldosterone, atrial and brain natriuretic peptide or apelin) differ between lean and obese individuals during these tests. We also demonstrated that disinhibition, a measure of overeating in response to various stimuli, is associated with greater perceive thirst in lean individuals but less those in individuals with obesity. Adherence is also a crucial component of success for diet induced weight loss. We completed a study investigating whether dietary adherence differed in lean versus individuals with obesity on weight maintaining versus calorie reduction diet. We found that there was no difference in adherence of hunger measures between the groups. These results indicate that difficulty with dietary adherence is not characteristic of individuals with obesity nor driven by hunger associated with dieting. However, we did find that overall levels of perceived stress, anhedonia and food insecurity measured prior to the diet period predicted lower adherence independent of percent body fat. In this study, we used ecological momentary assessment (EMA) as a daily adherence measure. These measures were sent randomly to participants twice daily. We found that affect affected adherence as measured by EMA, such that those with overall negative affect were less adherent. This raises the question of what measures might improve adherence. One way to decrease perceived stress and improve affect is to allow increased flexibility around the diet. Thus, we have designed a follow-up study to investigate the effect of flexible vs. rigid dietary instructions on adherence scores. We have also completed an on-line study investigating how the COVID-19 pandemic has affected eating behaviors and weight gain. We recruited participants across the world who answered questions about COVID, eating behavior and weight history. Participants are asked to complete questionnaires monthly for 12 months. A subset also participated in an ecological momentary assessment substudy that assessed more immediate daily impacts of COVID pandemic and eating behavior. We also have an in person substudy measuring weight and height in individuals at baseline, 6 and 12 months that can be correlated with the online assessments. In terms of the main study aims, we found that fear of covid was associated with maladaptive eating behaviors and weight even in the latter stages of the pandemic. The association with maladaptive eating behaviors but not weight was mediated by measures of psychological stress. In secondary analyses, we have also found that childhood and adult measures of unpredictability were associated with uncontrolled eating, self-reported weight and weight change, but this was mediated by intuitive eating (the self-reported ability to “listen to” natural hunger and fullness cues). We have also found that poor sleep quality is associated with higher scores on measures of eating disorders (binge eating and night eating).