Factors which predict variance in weight change

National Institute Of Diabetes And Digestive And Kidney Diseases

Investigators

Linked publications, trials & patents

Abstract

To investigate how energy expenditure and substrate oxidation rate changes with over and underfeeding the following studies have been conducted or are underway. In one study, after careful calibration of weight maintenance EE, individuals underwent a series of measurements of 24-hour EE in a respiratory chamber in which they are fasting or overfed (by 200% of weight maintenance needs) with diets that vary in macronutrient content. The dietary macronutrient content varied by fat, protein and carbohydrate content. In addition, behavioral, metabolic (e.g. core body temperature) and hormonal tests were performed to investigate the mechanism of the changes in EE and substrate oxidation rates. These individuals were followed up long term (up to 7 years) to investigate whether these energy expenditure phenotypes predicted longer term weight change. As part of these studies, we have demonstrated the reproducibility of the changes in energy expenditure with fasting and with overfeeding. We enrolled 96 participants in the core portion of this study. Consumption of high carbohydrate and high protein overfeeding diets increased EE the most, while low protein overfeeding increased EE the least. We confirmed the previous observation that individuals with a greater increase in EE with overfeeding have less decline in EE with fasting. This was confirmed in multiple analyses adjusting for confounders and using principal components analyses. These results indicate a clear EE (metabolic) phenotype in response to diet that can be identified as potential thrifty versus spendthrift phenotypes. After 6 months, we found that several EE phenotypes predicted weight gain: 1. Individuals with greater decrease in EE with fasting 2. Individuals with less increase in EE with low protein overfeeding 3. Individuals with greater increase in EE with high carbohydrate overfeeding. In longer term follow-up (up to 2 years) we have been able to demonstrate that increased 24hEE during low protein (<3% protein) overfeeding did predict less weight gain at 1 year. None of the other diets predicted weight change. These data indicate that low protein overfeeding is a valuable tool for uncovering metabolic phenotypes with clinical implications and that in free living conditions EE changes as measured in our study are likely overpowered by food availability and activity differences. Our data also demonstrated that macronutrient composition largely determines fuel preference (carbohydrate versus lipid oxidation rates) and that extrinsic dietary factors account for approximately 20% of the variance in these measurements. However, there is a strong intra-individual component to fuel preference with carbohydrate oxidation rates. In addition, lesser metabolic flexibility (those who decreased their 24hRQ and lipid oxidation rates less) determined during 24 hours of high fat overfeeding was associated with greater weight gain at one year follow-up. These results were confirmed whether the change in RQ or lipid oxidation rates were used or when the results were adjusted for the energy balance RQ or lipid oxidation rate. In further examining the thrifty versus spendthrift phenotypes, we have found that the greater drop in 24hEE with fasting, is due to a higher energy balance 24hEE rather than a lower fasting 24hEE. Thus, thrifty subjects have a higher energy balance 24hEE compared to spendthrift. Spendthrift subjects increase their EE during sleep. Furthermore, during protein imbalance overfeeding, thrifty individuals have blunted increase in 24hEE. These analyses fundamentally recharacterizes the prevailing model of the thrifty vs. spendthrift phenotype as we have defined it. We have also extended the thrifty versus spendthrift phenotype to EE changes with cold exposure. Individuals who have greater increase in 24hEE during mild cold exposure (~19°C) have less decrease in 24hEE with fasting. We investigated the mechanisms underlying these energy expenditure changes to fasting and overfeeding. One measure of the ability to dissipate heat is changes in core body temperature. Core body temperature correlated with changes in EE in response to fasting, such that individuals with lower core body temperature have a greater decrease in EE with fasting. Core body temperature increases with overfeeding, and in the setting of a high fat diet, this increase is associated with an increase in EE. We have also examined the role of the sympathetic nervous system. Urinary epinephrine excretion increases with fasting. We have found that there is an association between the increase in urinary epinephrine excretion and change in EE during fasting such that individuals who decrease fasting less have greater increase in urinary epinephrine excretion. Urinary dopamine excretion also increased during fasting and low protein overfeeding and was associated with changes in pancreatic polypeptide indicating a shift in sympathetic/parasympathetic tone with protein deprivation. Fibroblast growth factor 21 (FGF-21) is secreted by the liver and increases EE in rodent models. In humans FGF-21 increases with low protein diets. FGF-21 increased substantially (by ~300%) following our two different low protein overfeeding diets and that increased change in FGF 21 concentrations correlated with greater %EE increased with low protein overfeeding. Moreover, the greater increase in FGF-21 concentrations were associated with less weight gain at 6 months. Most significantly, this increase in FGF-21 mediated the association between greater increase in %EE with low protein overfeeding and weight change at 6 months indicating that FGF-21 is a good target for weight loss treatment. FGF-21 also decreases with 24h of fasting and cold exposure. We have consistently found that less decrease in FGF-21 is associated with less decline in EE during sleep (a surrogate of resting EE). This was also true during cold exposure, less decline in FGF-21 was associated with relatively greater increase in Sleeping EE during mild cold exposure. We have examined additional hormonal targets demonstrating an inverse association between increase in ghrelin concentrations and 24hEE during fasting. Thyroid hormones did change with fasting and with low protein and high protein overfeeding diets and were not related to changes in energy expenditure. We have recently completed large scale metabolomics measurements from plasma collected prior to and following each 24hEE session. These measurements demonstrated a coordinated response of predominantly fatty acid metabolites associated with alterations in substrate oxidation rates. This work will lay the groundwork for future precision nutrition studies. As increased adiposity may insulate against trans-abdominal heat loss which may increase TEF, we investigated the effect of central insulation on the EE and TEF changes associated with overfeeding. However, despite earlier evidence of an important role of abdominal heat dissipation in determining DIT and TEF, we did not find any changes in EE during overfeeding with additional abdominal insulation. Using 18 fluorodeoxyglucose positron emission computed tomographic (PET-CT) scans, we investigated whether those with visualized brown fat after cold exposure individuals had evidence of brown fat stimulation after being overfed by 200% of their energy needs using a high fat normal protein diet while in our metabolic chamber. We found no evidence of activation of brown fat with overfeeding following a high fat overfeeding, indicating that brown fat does not mediate the increased energy expenditure associated with overfeeding. We have found evidence of brown activation with high carbohydrate overfeeding likely due to high carbohydrate diets induction of the sympathetic nervous system (SNS) activation. In support of this, we found that those with both lower urinary epinephrine excretion and free T4 concentrations during thermoneutrality had increased brown fat activation indicating lower SNS and thyroid tone define individuals with a greater capacity to activate brown fat. We also found an association between greater peak cold induced BAT and less decrease in EE with fasting (spendthrift phenotype) indicating a physiologic role for BAT in these phenotypes. Metabolic responses to diet interventions may also be associated with behavioral or cognitive measures. We found that individuals with greater physical anhedonia scores had less decrease in 24hEE with fasting (were more spendthrift) but that this effect was mediated by depressive symptom scores. We also found that fasting FGF21 concentrations higher scores on a test the Iowa Gambling task (a test of decision making). We also investigated the presence of thrifty and spendthrift EE phenotypes in controlled inpatient intervention studies involving caloric restriction (in obese participants) and overfeeding (in lean participants), During these admissions, EE was measured during over and underfeeding series of metabolic and behavioral testing (including biopsies of muscle and fat) was conducted prior to any intervention. Participants then underwent 6 weeks of an inpatient dietary protocol involving underfeeding (for overweight and obese individuals) or low protein overfeeding (for lean, “obesity” resistant individuals). During the inpatient study, all aspects of food intake, energy expenditure, and energy loss are carefully measured to determine if differences in weight gain or loss can be attributed to recruitment of adaptive thermogenesis or other factors. Consistent with the presence of a thrifty versus spendthrift EE phenotype even in obese individuals, we found that less decrease in 24hEE with fasting was associated with greater weight loss in this controlled inpatient study. We also found that less decrease in 24hE with fasting, predicted less weight gain in those overfed (150% of weight maintaining energy needs) a low protein diet for 6 weeks. FGF21 increased substantially during low protein overfeeding and remained elevated (~3 times the baseline concentration) one week following the overfeeding period. We collected muscle and adipose tissue biopsies on these participants prior to and following weight loss, and performed RNA sequencing. In muscle we found that expression of uncoupling proteins 2 and 3 showed the greatest change with weight loss. These proteins dissipate proton gradients across mitochondrial membranes and are plausible candidates for regulation of energy expenditure. Individuals with greater decrease in UCP2 loss less weight and had lower 24hEE at the end of the 6 weeks of caloric restriction. This may also be mediated by changes in FGF-21. In adipose tissue, fat cell size fall into a small and large size distribution. We found decrease in size of large cells and a change in the nadir between the distribution points indicate that triglyceride loss during weight loss occurs exclusively in large cells. We also performed measurement of 24hEE during these interventions. This allowed us to look at 24hEE changes during the diet period. We found that less decline in EE after one week of calorie restriction was associated with greater energy deficit and greater weight loss. The degree of decline in 24hEE was consistent in individuals across the diet period and was present one week following the diet after participants were on a weight maintaining diet. This indicates that metabolic adaptation to caloric restriction determines variability of weight loss and is an intrinsic individual trait. In addition, it appears that lower 24hEE measured during energy balance prior to caloric restriction predicts less weight regain following the weight loss intervention. If 24hEE is a driver for appetite (as we have demonstrated in other studies), this may provide a physiologic explanation for difficulty with maintaining weight loss. The accumulated evidence from these studies continues to confirm that these EE and substrate oxidation rate phenotypes do play a role in predicting weight change certainly during controlled inpatient studies and during at least short term (6-12 months) follow-up. We have learned a great deal of information about the mechanism of these energy expenditure changes finding important roles for the sympathetic nervous system (particularly epinephrine) and circulating hormones FGF-21 and ghrelin. Furthermore, we have fundamentally redefined the concept of these phenotypes demonstrating that they hinge on differences in 24hEE during energy balance. Given the importance of these EE phenotypes and in particular the phenotype related to changes with low protein overfeeding and metabolic flexibility, we have developed two new protocols involving prolonged stays on our domiciled unit to further investigate these findings. One will further explore the question of our analysis regarding the recharacterization of our phenotypes; namely that the higher 24hEE during energy balance in thriftier participants. We have developed a dietary dose response study with different levels of under and overfeeding using standard and low protein diets to understand when thrifty versus spendthrift individuals diverge in their responses. This study began recruitment in May 2023 and is ongoing. The other study will further explore the role of fuel preference (e.g. the ability to increase fat oxidation in response to a high fat diet) on weight, fat mass, organ specific lipid deposition (liver, pancreas and muscle), insulin resistance (whole body, hepatic and adipose tissue) and insulin secretion. Participants will be fed a eucaloric high fat diet while residing in our metabolic chamber to precisely measure cumulative lipid balance. Then they will be fed a high fat over feeding diet for 4 weeks. We will investigate if the cumulative lipid balance predicts changes in weight, fat mass, organ specific fat deposition, insulin resistance and secretion.