Metabolic response to fasting predicts weight gain during low-protein overfeeding in lean men: further evidence for spendthrift and thrifty metabolic phenotypes

Abstract

Background: Greater increase in 24-h energy expenditure (24EE) during overfeeding and smaller decrease in 24EE during fasting ("spendthrift" metabolic phenotype) are associated with more weight loss during sustained caloric restriction in overweight subjects.

Objectives: The aim of this study was to investigate whether these acute metabolic responses can also predict weight gain during sustained overfeeding in lean individuals.

Methods: Seven lean men participated in this study. Prior to overfeeding, 24EE responses to fasting and 200% normal-protein overfeeding were measured in a whole-room indirect calorimeter. Volunteers underwent 6 wk of 150% low-protein (2%) overfeeding followed by another wk of weight-maintaining diet, during which 24EE was revaluated. Body composition, 24EE, and various hormone concentrations, including fibroblast growth factor 21 (FGF21), were assessed at baseline, at wk 1, 3, and 6 of the overfeeding period, and 1 wk following overfeeding through the use of dual-energy X-ray absorptiometry, indirect calorimetry, and ELISA. Cumulative energy surplus was calculated from 24EE, daily physical activity, and direct measurements of calories of nutrient intake, feces, and urine by bomb calorimetry.

Results: The average weight gain during 6 wk of low-protein overfeeding was 3.8 kg (6.1%, min: +2.5%, max: +8.0%). During 24-h fasting at baseline, 24EE decreased on average (mean ± SD) by 158 ± 81 kcal/d (P = 0.007). Subjects with less 24EE decrease during fasting (more metabolically spendthrift individuals) gained less weight (r = -0.84, P = 0.03), less fat mass (r = -0.81, P = 0.049), and stored less calories (r = -0.91, P = 0.03) during overfeeding. Following overfeeding, increased 24EE above requirements for achieved body size was associated with less weight and fat mass gain (r = -0.78, P = 0.04) and with the increase in 24EE during 200% normal-protein overfeeding measured at baseline (r = 0.91, P = 0.005). Serum FGF21 concentrations increased up to 44-fold during overfeeding (P < 0.0001).

Conclusions: Low-protein overfeeding may be an important tool to identify metabolic phenotypes (spendthrift compared with thrifty) that characterize susceptibility to weight gain. This trial was registered at clinicaltrials.gov as NCT00687115.

Keywords: FGF21; adaptive thermogenesis; bomb calorimetry; leptin; low-protein overfeeding; metabolic adaptation; metabolic phenotype; obesity; spendthrift; thrifty.

Published by Oxford University Press on behalf of the American Society for Nutrition 2019.

Figures

FIGURE 1

Study design. 24EE, 24-h energy expenditure; CATE, 24-h urinary catecholamine measurements; DXA, total body dual-energy X-ray absorptiometry; EB1, 24EE assessment to determine weight-maintaining energy needs; EB2, 24EE assessment in energy balance; EB3, 24EE assessment in energy balance at new weight; FST, 24EE assessment during fasting; HOR, hormonal measurements; LP-OF, 24EE assessment during low-protein overfeeding; OF, 24EE assessment during 200% normal-protein overfeeding; WMD, weight-maintaining diet; * = in random order.

FIGURE 2

Changes in absolute body weight (A), percentage body weight (B), FM (C), and FFM (D) during 6 wk of daily 150% low-protein overfeeding. Data presented as weekly change (every lefthand figure) with gray lines indicating individual participants and error bars indicating SDs; and change from baseline to post-overfeeding (every righthand figure) with connected dots indicating individual participants. The dashed vertical line of each lefthand figure represents the end of the overfeeding period. P values were calculated with a paired t test. FM, fat mass; FFM, fat-free mass; Post-OF, post-overfeeding period.

FIGURE 3

Inverse relations between the decrease in 24EE during fasting at baseline and body weight gain (A) and FM gain (B) following the 6-wk overfeeding period. Scatter plots displaying results determined with the use of Pearson correlation. Data presented for 6 subjects due to 1 invalid assessment of 24EE during fasting at baseline. 24EE, 24-h energy expenditure; FM, fat mass.

FIGURE 4

Relations between metabolic adaptation in the post-overfeeding period and body weight and FM gain after 6 wk of low-protein overfeeding. (A) During low-protein overfeeding, 24EE increased by 356 kcal/d at wk 1 (95% CI: 257, 456 kcal/d; P  = 0.0006), 302 kcal/d at wk 3 (95% CI: 86, 535 kcal/d; P  = 0.005), and 390 kcal/d at wk 6 (95% CI: 301, 478 kcal/d; P = 0.0003) above requirements for achieved body size. 24EE did not significantly differ from baseline in the post-overfeeding period but showed a high interindividual variability (+166 kcal/d; 95% CI: −32, 362 kcal/d; P  = 0.09). Measured minus predicted 24EE at each time point was compared with baseline by 1-sample Student's t test. Data presented for n  = 5 subjects at wk 1, 3, and 6 due to limited calorimeter availability and logistic issues experienced during the overfeeding period. Inverse associations between metabolic adaptation and (B) body weight gain and (C) increase in FM. Scatter plots displaying results determined with the use of Pearson correlation. 24EE, 24-h energy expenditure; FM, fat mass; Post-OF, post-overfeeding period.

FIGURE 5

Relations between metabolic adaptation during the post-overfeeding period and changes in baseline 24EE during 200% normal-protein overfeeding (A) and fasting (B). The graphs are scatter plots displaying results determined from the Pearson correlation. In panel B, data are presented for 6 subjects because of 1 invalid assessment of 24EE during fasting at baseline. 24EE, 24-h energy expenditure.

FIGURE 6

Changes in urinary catecholamines during and following 6 wk of daily 150% low-protein overfeeding. (A) On average, epinephrine concentration did not change during and following low-protein overfeeding as compared with baseline values (all P  > 0.3). (B) Norepinephrine concentration increased on average by 2.1-fold at wk 1 (95% CI: 0.97-fold, 3.2-fold; P  = 0.06), by 2.6-fold at wk 3 (95% CI: 1.4-fold, 3.9-fold; P  = 0.004), by 2.4-fold at wk 6 (95% CI: 1.3-fold, 3.5-fold; P = 0.01), and was comparable with baseline values in the post-overfeeding period (mean: 1.7-fold; 95% CI: 0.6-fold, 2.8-fold; P  = 0.13). Fold change values were calculated for each subject as the ratio between values at each time point divided by the baseline value. At each time point, black circles show individual fold change values and bars show average fold change. Significance was determined with Student's t test. The average absolute change is reported at each time point along with P values for fold-change. Data presented for: n  = 6 subjects (baseline, wk 6, and post-overfeeding), n  = 5 subjects (wk 1), and n  = 4 subjects (wk 3). Post-OF, post-overfeeding period.

FIGURE 7

Changes in serum fT3 (A), serum fT4 (B), leptin (C), and leptin adjusted for FM (D) during and following 6 wk of daily 150% low-protein overfeeding. (A) Serum fT3 remained unchanged during and following low-protein overfeeding. (B) Serum fT4 decreased on average by 20.6% after 6 wk of overfeeding (95% CI: −32.4%, −8.7%; P = 0.005) and remained 23.0% lower than baseline in the post-overfeeding period (95% CI: −33.9%, −12.2%; P  = 0.002). Percentage change values of fT3 and fT4 were calculated for each subject as the difference between the absolute values at each time point minus the baseline absolute value, and then multiplied by 100. At each time point, black circles show individual percentage change values and bars show average percentage change. Significance was determined with Student's t test. (C) Serum leptin concentration increased on average by 1.8-fold at wk 3 (95% CI: 1.4-fold, 2.2-fold; P  = 0.002), 3.0-fold at wk 6 (95% CI: 2.3-fold, 3.8-fold; P = 0.0009), and remained 2.8-fold higher than baseline in the post-overfeeding period (95% CI: 1.7-fold, 3.8-fold; P = 0.006). Fold change values were calculated for each subject as the ratio between values at each time point divided by the baseline value. At each time point, black circles show individual fold change values and bars show average fold change. Significance was determined with Student's t test. (D) Serum leptin adjusted for FM remained unchanged during and following overfeeding. Measured minus predicted serum leptin at each time point was compared with baseline by 1-sample Student's t test. The average absolute increase is reported at each time point along with P values for percentage change or fold-change. Data are presented for n  = 5 subjects at wk 1, 3, and 6 during the overfeeding period. In panels A, B, and C, the average absolute increase is reported at each time point along with P values for percentage change or fold-change. FM, fat mass; fT3, free triiodothyronine 3; fT4, free triiodothyronine 4; Post-OF, post-overfeeding period.

FIGURE 8

Changes in serum FGF21 (A) during low-protein overfeeding, and its relation to weight gain after 6 wk of low-protein overfeeding (B) and metabolic adaptation in the post-overfeeding period (C). (A) Serum FGF21 concentration increased on average by 25-fold at wk 3 (95% CI: 11-fold, 39-fold; P  = 0.007) and 44-fold at wk 6 (95% CI: 29-fold, 58-fold; P = 0.0006) during the low-protein overfeeding period and remained 3.5-fold higher during the post-overfeeding period (95% CI: 1.3-fold, 5.7-fold; P = 0.03). Fold change values were calculated for each subject as the ratio between values at each time point divided by the baseline value. At each time point, black circles show individual fold change values and bars show average fold change. Significance was determined with Student's t test. (B) Association between change in serum FGF21 from baseline to overfeeding wk 6 and body weight gain. (C) Association between change in serum FGF21 from baseline to the post-overfeeding period and metabolic adaptation. Data are presented for n  = 6 subjects at wk 3 and wk 6 during the overfeeding period and for n = 7 subjects in the post-overfeeding period. Scatter plots displaying results determined with the Pearson correlation. FGF21, fibroblast growth factor 21; Post-OF, post-overfeeding period.

FIGURE 9

Exemplary body weight time courses of 2 individuals with different metabolic phenotype during the 6 wk of daily 150% low-protein overfeeding. Data are presented as daily weight change as a percentage of the body weight on the first day of the overfeeding period. The individual with the thriftiest metabolic phenotype (i.e., the subject with the greatest decrease in baseline 24EE during fasting equal to −253 kcal/d) gained twice as much weight as the individual with the most spendthrift metabolic phenotype, who had a decrease in baseline 24EE during fasting equal to −4 kcal/d. 24EE, 24-h energy expenditure.