The Fall in Leptin Concentration Is a Major Determinant of the Metabolic Adaptation Induced by Caloric Restriction Independently of the Changes in Leptin Circadian Rhythms

Abstract

Context:

Leptin is involved in the hormonal regulation of the reproductive, somatotropic, thyroid, and autonomic axes and ultimately in the regulation of energy balance. In parallel to the metabolic adaptation observed in response to caloric restriction (CR), plasma leptin concentrations are substantially decreased, suggesting a role for this hormone in the drop in energy expenditure beyond that predicted by the changes in body composition (metabolic adaptation).

Aim:

The aim of the study was to explore the changes in 24-h leptin circadian rhythm in response to CR and to investigate the relationship between these changes and metabolic adaptation.

Design:

In a randomized, controlled trial (Comprehensive Assessment of Long-Term Effects of Reducing Intake of Energy), 48 subjects were assigned to a control group or one of three CR groups for 6 months. Leptin concentration was assessed every 30 min for 24 h, and leptin circadian variations were fitted by Cosinor analysis. Sedentary energy expenditure and urinary catecholamine excretion were measured for 24 h in a metabolic chamber.

Results:

Six months of CR decreased body weight by −11.4 ± 0.6% (mean ± sem; P < 0.001). Mean 24-h circulating leptin concentration decreased by −44 ± 3% (P < 0.001), whereas leptin diurnal amplitude slightly increased over the 6 months of CR. CR caused a metabolic adaptation of −126 ± 25 kcal/d (P <0.001) and a significant decrease in urinary norepinephrine (−13 ± 3%) and T₃ concentrations (10 ± 2%). The metabolic adaptation was significantly and independently related to the changes in 24-h leptin (r² = 0 .22, P < 0.01) but not to the changes in leptin amplitude.

Conclusion:

Our results confirm an important role for leptin as an independent determinant of the metabolic adaptation in response to CR.

Significant weight loss induced by caloric restriction (CR) is characterized by reduced thyroid hormones, catecholamine, and leptin concentrations associated with a hypometabolic state, i.e. a drop in energy expenditure (EE) beyond that expected on the basis of changes in fat-free mass (FFM) and fat mass (FM) (1–4). This hypometabolic state referred to as metabolic adaptation persists with reduced body weight over the long term (4), opposes further decreases in body weight (2), and predisposes some individuals to weight regain (2).

Leptin, an adipokine secreted in proportion to body fat stores (5), is involved in the control of energy balance and body composition (6) by regulating both energy intake and expenditure (7). Leptin concentrations have been shown to be predictive of body weight changes (8, 9), and changes in leptin have been shown to be related to the decrease in EE associated with weight loss (10–12). In calorie-restricted animals and humans, exogenous leptin administration has been shown to reverse the metabolic adaptation induced by CR, restoring not only EE (3, 13, 14) but also catecholamine and thyroid hormone concentrations and skeletal muscle efficiency to baseline values (14, 15). This suggests a role for leptin in the regulation of EE during energy deprivation, possibly to protect against excessive loss of body fat stores (6).

The aim of this study was to characterize for the first time the changes in leptin circadian variations in response to 6 months of CR and to explore its relationships with CR-induced metabolic adaptation (16–18). We therefore hypothesized that the metabolic adaptation occurring with weight loss induced by CR was associated with the decrease in plasma leptin. Plasma leptin concentrations are known to follow a diurnal excursion (19, 20). The amplitude of the circadian variation potentially influences the amount of leptin tissues are exposed to and consequently the biological effect of the hormone (21). Because acute changes in the leptin variation amplitude have been reported to be associated with changes in energy intake (21–23) and with changes in body weight gain (23, 24), we further hypothesized that the metabolic adaptation would be associated with a decrease in leptin diurnal amplitude.

Subjects and Methods

Subjects and study design

The Comprehensive Assessment of Long-Term Effects of Reducing Intake of Energy (CALERIE) study was approved by the Pennington Biomedical Research Center Institutional Review Board. As previously reported (16, 18), from a total of 599 screened potential overweight (25 kg/m² ≤ body mass index < 30 kg/m²) participants, 48 subjects provided informed consent to participate in this study and 46 (26 females and 20 males) completed the study.

Participants were randomized to one of four experimental groups for 6 months: 1) control (weight maintenance diet based on an American Heart Association Step 1 diet), 2) CR (25% calorie restriction from baseline EE), 3) CR+EX = 12.5% CR with exercise (12.5% increase in EE by structured exercise), and 4) LCD (low calorie diet, i.e. 890 kcal/d until a 15% reduction in body weight followed by a weight maintenance diet). The group assignment was stratified to ensure even distributions of sex and body mass index in the four groups. Physiological testing was performed over a 5-d inpatient stay in the institutional clinic at baseline as well as during week 24 (month 6) of the study. Participants were provided with all food during baseline and while undergoing physiological testing based on individual energy requirements and treatment group assignment (16, 17). Body weight was measured in the morning in an overnight fasting state, after voiding. Whole-body fat content was measured using dual-energy x-ray absorptiometry (Hologics QDR 4500 A; Bedford, MA). FM and FFM were calculated from percent body fat assessed by dual-energy x-ray absorptiometry and the measured body weight.

Energy expenditure

Sedentary EE was assessed over 24 h in a metabolic chamber (16, 17). Sleeping EE was calculated between 0200 and 0500 h when no spontaneous physical activity was detected by microwave motion sensors. All urine was collected and acidified (10 ml HCl) for subsequent determination of urinary excretion of nitrogen (ANTEK, Houston, TX) and epinephrine and norepinephrine (Bio-Rad, Philadelphia, PA). Metabolic adaptation was defined as measured minus predicted 24-hr EE (MetAd_{24 h}) and sleeping EE (MetAd_sleep) at month 6. Predicted values of 24-h and sleeping EE were calculated on the basis of baseline values by a stepwise multivariate regression with FM, FFM, age, and sex as independent variables as previously reported (16).

Twenty-four-hour leptin profiles

After exiting the metabolic chamber, a 24-h blood sampling protocol was performed. Beginning at 0800 h, blood samples were taken every 30 min to assess leptin circadian rhythms. Similarly to the chamber stay, meals were consumed at 0900, 1130, and 1700 h. Lights were turned off from 2030 to 0700 h, and physical activity was limited. Plasma leptin concentration was measured by RIA (Linco, St. Charles, MO). Circadian variations of plasma leptin concentration were determined by cosinor analysis, and the following descriptors were derived algebraically: the mesor (mean 24 h leptin concentration), the amplitude (mean deviation between the peak and nadir and the mesor), and the acrophase (time to peak). Amplitude was then expressed as relative to the leptin mesor.

Statistical analysis

All data are expressed as mean ± sem. Statistical analysis were performed using the R software (25). Differences between sexes were assessed by unpaired Student t tests. To account for the large interindividual variability in plasma leptin, an analysis of covariance with repeated-measures over time, and baseline values as covariates was used to test for the effects of treatment (CR groups vs. control), time (months 0 and 6), and their interactions. The effect of sex was assumed to be accounted for by including baseline values as covariates. Post hoc multiple comparisons were performed using the Bonferonni adjustment. Measures of association between single variables were obtained by Pearson's product-moment correlations. Significance was considered at a P < 0.05.

Results

At baseline, in controls and CR groups (CR, CR+EX, and LCD), body weight was 81.5 ± 1.8 vs. 82.3 ± 2.9 kg and percent body fat 31.3 ± 1.8 and 32.2 ± 1.3%. Details about the changes in body composition have already been published (16). Briefly, body weight (−11.4 ± 0.6%), FM (−26.8 ± 1.7%), and FFM (−4.5 ± 0.3%) were significantly (P < 0.001) decreased in all three intervention groups at month 6, whereas they remained stable in the control group.

Twenty-four-hour leptin profiles

The overall effect of the different treatments on leptin circadian variations is displayed in Table 1. Leptin mesor and fasting leptin were significantly decreased in response to CR (P < 0.001) and over time (P < 0.001) but independently of the type of CR and not differently between females and males (Table 1). The percent change in leptin mesor was strongly and negatively associated to the baseline leptin mesor at month 6 (r² = 0.61, P < 0.001, respectively). Controlling for baseline leptin mesor, the changes in leptin mesor, were positively and significantly related to the changes in fat mass (P < 0.04). Leptin amplitude was slightly but significantly (P < 0.01) increased by CR (Table 1) but not differently between males and females. The changes in leptin amplitude were inversely related to the changes in leptin mesor (r² = 0.10, P = 0.07) and to the absolute change in fat mass (r² = 0.11, P = 0.05). The leptin acrophase remained unchanged (Table 1). In controls, leptin circadian variations remained unchanged.

Table 1.

Changes in leptin circadian rhythm variations, urinary catecholamine excretion, and thyroid hormones after 6 months of caloric restriction

	Baseline	Month 6
Control
Mesor (ng/ml)	23.6 ± 4.8	21.2 ± 4.3
Amplitude (% mesor)	17.4 ± 2.0	15.8 ± 1.7
Acrophase (h)	14.8 ± 0.3	14.6 ± 0.5
r2 (fitting of data)	0.66 ± 0.07	0.68 ± 0.06
Fasting leptin (ng/ml)	20.8 ± 4.1	17.3 ± 3.3
Epinephrine (nmol/d)	79 ± 19	57 ± 9
Norepinephrine (nmol/d)	260 ± 35	286 ± 47
T3 (ng/dl)	144 ± 8	149 ± 7
T4 (μg/dl)	7.2 ± 0.4	7.3 ± 0.4
Calorie restricted
Mesor (ng/ml)	17.8 ± 3.8	11.3 ± 2.4a
Amplitude (% mesor)	20.2 ± 1.9	22.8 ± 1.7
Acrophase (h)	14.8 ± 0.5	14.4 ± 0.6
r2 (fitting of data)	0.73 ± 0.05	0.74 ± 0.03
Fasting leptin (ng/ml)	15.6 ± 3.1	8.2 ± 1.8a
Epinephrine (nmol/d)	66 ± 14	73 ± 13
Norepinephrine (nmol/d)	264 ± 25	228 ± 20
T3 (ng/dl)	139 ± 7	130 ± 7
T4 (μg/dl)	7.4 ± 0.4	7.5 ± 0.5
Calorie restricted + exercise
Mesor (ng/ml)	21.3 ± 3.9	14.1 ± 3.6a
Amplitude (% mesor)	17.4 ± 1.6	17.9 ± 1.7
Acrophase (h)	14.1 ± 0.4	13.6 ± 0.5
r2 (fitting of data)	0.62 ± 0.07	0.58 ± 0.05
Fasting leptin (ng/ml)	18.8 ± 3.4	8.7 ± 2.1b
Epinephrine (nmol/d)	77 ± 17	67 ± 5
Norepinephrine (nmol/d)	332 ± 29	275 ± 22
T3 (ng/dl)	135 ± 6	120 ± 6
T4 (μg/dl)	7.4 ± 0.4	7.0 ± 0.4
Low-calorie diet
Mesor (ng/ml)	19.1 ± 3.0	10.1 ± 1.9a
Amplitude (% mesor)	18.6 ± 1.8	22.3 ± 2.5
Acrophase (h)	14.4 ± 0.5	13.8 ± 0.5
r2 (fitting of data)	0.70 ± 0.03	0.68 ± 0.05
Fasting leptin (ng/ml)	14.3 ± 2.1	6.6 ± 1.0b
Epinephrine (nmol/d)	58 ± 5	60 ± 9
Norepinephrine (nmol/d)	274 ± 32	216 ± 25
T3 (ng/dl)	156 ± 9	133 ± 6
T4 (μg/dl)	8.1 ± 0.5	6.9 ± 0.1

Leptin, T₃, and T₄ data can be converted to SI units by multiplying by 1, 0.0154, and 12.87, respectively. Values are shown as means ± sem.

^aP < 0.001 vs. baseline.

^bP < 0.01 vs. baseline.

Urinary catecholamines and thyroid hormones

Twenty-four-hour urinary epinephrine excretion was unchanged by CR (P = NS, Table 1). Urinary norepinephrine excretion was significantly decreased by CR (P < 0.01) over time (P < 0.05), with no difference between females and males (Table 1). No difference was found between calorie-restricted groups with regard to norepinephrine excretion. Changes in urinary catecholamine excretion were not correlated to the changes in leptin variations. As reported previously, there was a significant effect of CR on plasma T₃ and T₄ (Table 1) (16). The change in plasma T₃ was significantly related to the change in leptin (P = 0.01), controlling for baseline leptin.

EE and metabolic adaptation

Twenty-four-hour and sleeping EE were significantly decreased in all three caloric-restricted groups (not shown, P < 0.001), whereas it remained unchanged in controls. At month 6, a metabolic adaptation was observed in calorie-restricted groups only (16) when measured over 24 h and during sleep (MetAd_{24 h} −126 ± 25 kcal/d, MetAd_sleep −83 ± 17, kcal/d, P < 0.001) .

At month 6, MetAd_{24 h} was positively and significantly related to the percent change in leptin mesor (r² = 0.22, P < 0.01) (Fig. 1A) and to the absolute change in mesor controlling for baseline leptin concentration (r² = 0.20, P = 0.03). Similarly, MetAd_sleep was related to the percent decrease in leptin mesor (r² = 0.20, P < 0.01) (Fig. 1B) and to the absolute change in mesor controlling for baseline leptin concentration (r² = 0.31, P < 0.01). At month 6, the metabolic adaptation was also related to the relative decrease in norepinephrine excretion (MetAd_{24 h}: r² = 0.1, P = 0.06 and MetAd_sleep: r² = 0.20, P < 0.01). In contrast, the metabolic adaptation was unrelated to the changes in leptin amplitude.

Fig. 1.

Relationship between the changes in leptin mesor and the metabolic adaptation as calculated from 24-h EE (MetAd[_{24 h}], A) and sleep EE (MetAd[_sleep], B) measures. Gray squares, Calorie-restriction group; black squares, calorie restriction + exercise; gray triangle, low-calorie diet.

To further delineate the determinants of the metabolic adaptation in response to CR, we performed a multivariate regression analysis between the metabolic adaptation at month 6 and changes in leptin mesor, urinary norepinephrine, and thyroid hormone concentrations. The change in leptin mesor was the only significant and independent determinant of MetAd_{24 h} (P = 0.03). Changes in leptin mesor (P = 0.04) as well as change in norepinephrine excretion (P = 0.03) were additional significant predictors of the MetAd_sleep.

Discussion

The main finding of this study is that the decline in 24-h leptin in response to CR is a significant and independent determinant of metabolic adaptation induced by CR. These results indicate that the drop in plasma leptin concentration is probably underlying the additional reduction in energy expenditure observed in some individuals following diet-induced weight loss.

A role for blunted leptin levels, along with organochlorine compounds, in the metabolic adaptation has been previously demonstrated in obese individuals undergoing weight loss (26). Indeed, the reversal of the hypometabolic phenotype induced by calorie restriction with recombinant leptin treatment in subjects undergoing weight loss (3, 14) and those with leptin deficiency (13) supports a role for leptin in the mechanisms related to the metabolic adaptation. Moreover, changes in leptin have also been associated with changes in EE in caloric restricted subjects (10, 11). The fact that leptin administration restores thyroid hormone and catecholamine levels suggests a central effect of leptin (3). Low circulating leptin blunts the leptin signaling cascade at the level of the hypothalamus and triggers an exaggerated drop in energy expenditure beyond that expected for the changes in body composition (6). Yet the exact mechanism by which a sustained decrease in leptin induces the metabolic adaptation is unknown. Because the changes in leptin and T₃ are related and the changes in leptin are independently predictive of the metabolic adaptation, we speculate leptin may also exert peripheral effects in skeletal muscle. In support of this hypothesis, a decrease in the energy cost of physical activity beyond that predicted by the changes in body composition was associated with lower plasma leptin concentrations after weight loss (1), and this improved energy efficiency was reversed by exogenous leptin administration (14). Taken together, these observations indicate that a potential mechanism by which CR induces a metabolic adaptation is through leptin-mediated changes in skeletal muscle efficiency, known to be improved during CR (27).

In conclusion, this study reports that the metabolic adaptation occurring during CR is related to the fall in mean 24-h leptin concentration, independently of changes in leptin diurnal excursion amplitude. The hypothalamus and the skeletal muscle seemingly respond to the fall in leptin by decreasing the thyroid and sympathetic axes and by increasing the mitochondrial efficiency, respectively.