Abstract

Background

Fat mobilization in adipose tissue (AT) has a specific timing. However, circadian rhythms in the activity of the major enzyme responsible for fat mobilization, hormone-sensitive lipase (HSL), have not been demonstrated in humans.

Objective

To analyze in a cross-sectional study whether there is an endogenous circadian rhythm in HSL activity in human AT ex vivo and whether rhythm characteristics are related to food timing or fasting duration.

Methods

Abdominal AT biopsies were obtained from 18 severely obese participants (age: 46 ± 11 years; body mass index 42 ± 6 kg/m2) who underwent laparoscopic gastric bypass. Twenty-four-hour rhythms of HSL activity and LIPE (HSL transcript in humans) expression in subcutaneous AT were analyzed together with habitual food timing and night fasting duration.

Results

HSL activity exhibited a circadian rhythm (P = .023) and reached the maximum value at circadian time 16 (CT) that corresponded to around midnight (relative local clock time. Similarly, LIPE displayed a circadian rhythm with acrophase also at night (P = .0002). Participants with longer night fasting duration >11.20 hours displayed almost double the amplitude (1.91 times) in HSL activity rhythm than those with short duration (P = .013); while habitual early diners (before 21:52 hours) had 1.60 times higher amplitude than late diners (P = .035).

Conclusions

Our results demonstrate circadian rhythms in HSL activity and may lead to a better understanding of the intricate relationships between food timing, fasting duration and body fat regulation.

Hormone-sensitive lipase (HSL) is the major enzyme responsible for the release of free fatty acids from adipose tissue (AT) (1). Its main function is to mobilize stored fat that is the primary source of energy for most tissues. HSL activity is acutely increased by catecholamines and inhibited by insulin via phosphorylation–dephosphorylation reactions at specific amino acid residues (2-4). Obesity is associated with decreased expression and function of adipocyte HSL (4) and consequently with a reduction of fat mobilization, which contributes to the development and maintenance of the expanded AT mass (5-7).

There are specific times of day and night that are characteristic for either mobilizing or accumulating fat in adipose tissue stores (8). A temporal order in daily patterns of gene expression, enzyme activity, or hormone secretion may be responsible for these processes and crucial for AT to either accumulate or mobilize fat at the proper time. This phenomenon is known as temporal compartmentalization (9). Accordingly, previous studies by our group and others have shown that the expressions of multiple metabolic genes in AT (approximately 2% of the adipose transcriptome) display circadian rhythms (10-12).

Insulin sensitivity, which is related to body fat accumulation, has an endogenous circadian rhythm ex vivo, with maximum insulin sensitivity at noon and minimum at midnight (13), while a rhythmic expression of murine HSL transcript (hsl) together with rhythmic lipolytic activity has also been demonstrated in cultured AT explants (14). Furthermore, a higher amplitude (ie, the difference in the level between peak and trough values) of HSL activity rhythm is associated with a higher mobilization of lipids from AT in mice (14). Nevertheless, the direct evidence for the existence of circadian rhythms in HSL activity and in LIPE (encoding HSL in humans) in AT is still lacking in humans.

Ingested food provides the main source for fat uptake in AT. The timing of eating, particularly of high-energy content meals, may be decisive in fat storage or mobilization, and changes in this timing can have metabolic consequences for the development of obesity (15). In addition, unusual eating times can induce a disruption of the circadian system, leading to metabolic alterations and health problems (16). Late eating has been related to increased obesity or difficulties in weight loss (17, 18). By contrast, recent studies show that time-restricted eating (TRE), which allows one to consume ad libitum energy intake within a defined window of time, with fasting durations of at least 8 hours, is related to decreased obesity and to a better weight loss prognosis (19, 20).

We hypothesize that in human AT there is a circadian rhythm in HSL activity and that the amplitude of this rhythm decreases with late eating and increases with longer fasting duration at night. Evidence in support of these hypotheses would provide potential novel molecular mechanisms underlying the established connections between food timing, TRE, and body weight regulation.

We hereby present results on 24-hour patterns of HSL activity (pHSL/HSL) in ex vivo AT in 18 participants who underwent gastric bypass surgery for obesity. We have studied the potential association between this 24-hour pattern and habitual meal timing and fasting duration at night. Since HSL protein expression is at least in part determined by pretranslational factors (21), we also analyzed 24-hour gene expression of LIPE.

Materials and Methods

Subjects characteristics

In a cross-sectional study, abdominal AT biopsies were obtained from severely obese women (n = 9) and men (n = 9), age 46 ± 11, and body mass index (BMI) 42 ± 6 kg/m2 who underwent laparoscopic gastric bypass surgery at the General Surgery Service of Virgen de la Arrixaca University Hospital (Murcia, Spain). Characteristics of the population are represented in Table 1. None of the subjects included where under insulin therapy.

Table 1.

General characteristics of the participants, lifestyle habits, and adipose tissue characteristics

Total population (n = 18)
CharacteristicsMeanSD
Age (years)4611
BMI (kg/m2)426
Total body fat (impedance; %)41.068.87
Metabolic syndrome traits
 Waist circumference (cm)126.5715.32
 Triglycerides (mmol/L)1.510.82
 Glucose (mmol/L)6.431.87
 HDL-c (mmol/L)1.260.46
 Diastolic blood pressure (mmHg)82.786.18
 Systolic blood pressure (mmHg)138.2215.16
 MetS Score3.181.38
Lifestyle habits
Meal timing
 Breakfast onset (hh:mm)8:511:03
 Lunch onset (hh:mm)14:400:36
 Dinner onset (hh:mm)21:430:49
Sleep characteristics
 Sleep onset (hh:mm)0:291:05
 Sleep duration (hours)7.151.44
Fasting duration
 Fasting duration (hours)10.781.44
Adipose tissue
Total fat cell size (μm)97.628.72
Rhythm characteristics
HSL activity
 Average (AU)1.190.58
 Amplitude (fold change)0.160.10
 Acrophase (hh:mm)23:325:49
 Percent of rhythmicity (%)25.8816.97
LIPE expression
 Average (AU)1.040.13
 Amplitude (fold change)0.430.32
 Acrophase (hh:mm)22:123:52
 Percent of rhythmicity (%)44.5428.66
Total population (n = 18)
CharacteristicsMeanSD
Age (years)4611
BMI (kg/m2)426
Total body fat (impedance; %)41.068.87
Metabolic syndrome traits
 Waist circumference (cm)126.5715.32
 Triglycerides (mmol/L)1.510.82
 Glucose (mmol/L)6.431.87
 HDL-c (mmol/L)1.260.46
 Diastolic blood pressure (mmHg)82.786.18
 Systolic blood pressure (mmHg)138.2215.16
 MetS Score3.181.38
Lifestyle habits
Meal timing
 Breakfast onset (hh:mm)8:511:03
 Lunch onset (hh:mm)14:400:36
 Dinner onset (hh:mm)21:430:49
Sleep characteristics
 Sleep onset (hh:mm)0:291:05
 Sleep duration (hours)7.151.44
Fasting duration
 Fasting duration (hours)10.781.44
Adipose tissue
Total fat cell size (μm)97.628.72
Rhythm characteristics
HSL activity
 Average (AU)1.190.58
 Amplitude (fold change)0.160.10
 Acrophase (hh:mm)23:325:49
 Percent of rhythmicity (%)25.8816.97
LIPE expression
 Average (AU)1.040.13
 Amplitude (fold change)0.430.32
 Acrophase (hh:mm)22:123:52
 Percent of rhythmicity (%)44.5428.66
Table 1.

General characteristics of the participants, lifestyle habits, and adipose tissue characteristics

Total population (n = 18)
CharacteristicsMeanSD
Age (years)4611
BMI (kg/m2)426
Total body fat (impedance; %)41.068.87
Metabolic syndrome traits
 Waist circumference (cm)126.5715.32
 Triglycerides (mmol/L)1.510.82
 Glucose (mmol/L)6.431.87
 HDL-c (mmol/L)1.260.46
 Diastolic blood pressure (mmHg)82.786.18
 Systolic blood pressure (mmHg)138.2215.16
 MetS Score3.181.38
Lifestyle habits
Meal timing
 Breakfast onset (hh:mm)8:511:03
 Lunch onset (hh:mm)14:400:36
 Dinner onset (hh:mm)21:430:49
Sleep characteristics
 Sleep onset (hh:mm)0:291:05
 Sleep duration (hours)7.151.44
Fasting duration
 Fasting duration (hours)10.781.44
Adipose tissue
Total fat cell size (μm)97.628.72
Rhythm characteristics
HSL activity
 Average (AU)1.190.58
 Amplitude (fold change)0.160.10
 Acrophase (hh:mm)23:325:49
 Percent of rhythmicity (%)25.8816.97
LIPE expression
 Average (AU)1.040.13
 Amplitude (fold change)0.430.32
 Acrophase (hh:mm)22:123:52
 Percent of rhythmicity (%)44.5428.66
Total population (n = 18)
CharacteristicsMeanSD
Age (years)4611
BMI (kg/m2)426
Total body fat (impedance; %)41.068.87
Metabolic syndrome traits
 Waist circumference (cm)126.5715.32
 Triglycerides (mmol/L)1.510.82
 Glucose (mmol/L)6.431.87
 HDL-c (mmol/L)1.260.46
 Diastolic blood pressure (mmHg)82.786.18
 Systolic blood pressure (mmHg)138.2215.16
 MetS Score3.181.38
Lifestyle habits
Meal timing
 Breakfast onset (hh:mm)8:511:03
 Lunch onset (hh:mm)14:400:36
 Dinner onset (hh:mm)21:430:49
Sleep characteristics
 Sleep onset (hh:mm)0:291:05
 Sleep duration (hours)7.151.44
Fasting duration
 Fasting duration (hours)10.781.44
Adipose tissue
Total fat cell size (μm)97.628.72
Rhythm characteristics
HSL activity
 Average (AU)1.190.58
 Amplitude (fold change)0.160.10
 Acrophase (hh:mm)23:325:49
 Percent of rhythmicity (%)25.8816.97
LIPE expression
 Average (AU)1.040.13
 Amplitude (fold change)0.430.32
 Acrophase (hh:mm)22:123:52
 Percent of rhythmicity (%)44.5428.66

AT biopsies were obtained from subcutaneous AT at the end of the surgical procedure which finished between approximately 12:00 and 13:00 hours for all participants. Samples were collected from the upper left side of the abdomen. Anthropometry, meal timing, metabolic syndrome (MetS) traits, and sleep characteristics were assessed.

Protocols were approved by the Institution Review Board of Virgen de la Arrixaca University Hospital, and participants signed a written informed consent before biopsies were obtained.

Anthropometry and metabolic syndrome characteristics.

Weight was determined in participants wearing light clothes and bare-footed using a digital electronic weighing scale. Height was determined using a Harpender digital stadiometer (range 0.70-2.05 m) with the subject standing and the head in the Frankfurt plane. From these data, the BMI was calculated according to the formula: weight/height2 (kg/m2). Total body fat (%) was measured by bioimpedance with a TANITA Model TBF-300 (TANITA Corporation of America, Arlington Heights, IL, USA). Waist circumference was measured at the umbilicus level (22).

Fasting plasma concentrations of triacylglycerides, glucose, and high-density lipoprotein cholesterol (HDL-c) were determined with commercial kits (Roche Diagnostics, Mannheim, Germany). Arterial pressure was measured with a mercury sphygmomanometer while seated for at least 10 minutes. For each subject, a MetS score was calculated as the number of components of MetS, based on thresholds for waist circumference, fasting glucose, triacylglycerides, HDL-c, and systolic or diastolic blood pressure with a maximum value of 5 points (23).

Lifestyle habits

Meal timing, sleep characteristics, and fasting duration.

All participants were daytime workers so the biological night coincided with the circadian night. They were asked for their habitual sleep habits and food schedule by the same interviewer. The questions included habitual time to bed, number of awakenings during the night, time of sleep offset, sleep duration, and time of starting each of the 3 main meals of the day (breakfast, lunch, and dinner). Participants were classified as early (n = 9) and late dinners (n = 9) using the median (21:52 hours) and as early (n = 9) and late lunch eaters (n = 9) (median 14:40 hours ) as previously reported (24, 25).

Fasting duration (at night) was calculated by the following formulae: timing of the last meal of the previous night (dinner offset) minus timing of the first meal of the next day (breakfast offset). Participants were classified as long (n = 9) or short fasting duration (n = 9) using the median (11.20 hours).

Adipose tissue

Fat cell size.

Adipocyte diameter was determined according to Sjöström et al. (26). In order to measure fat cell size with the Sjöström technique, it is necessary to have a large amount of tissue and that it is well preserved and has not stretched or deformed during the surgery process. Due to the large amount of fat required to perform the explants we only had enough tissue for cell size determinations in 9 subjects. To determine technical and operator reliability of these assessments, variability was examined from duplicate measures in several participants (n = 7) in the same slice, in 2 different slices from the same AT sample, and in slices observed by different operators. The correlation factors were 0.99, 0.99, and 0.94 respectively.

HSL signaling assay

Culture conditions.

In general, key experimental procedures that may have reset the peripheral clock were designed to coincide with the morning times, including surgery and the time of the medium change. Therefore, the timings of an individual’s sleep/wake schedule and adipose tissue culture coincided.

Indeed, immediately after the surgery (always in the morning), biopsies were placed in medium and within half an hour were set at 37°C for approximately 24 hours in a humidified atmosphere containing 7% CO2 in culture dishes. Samples of approximately 1500 mg were cut into pieces of 1 to 2 mm3 in order to enhance the contact of AT with the medium and were placed in 2.5 mL of Dulbecco’s modified Eagles Medium (DMEM) supplemented with 10% fetal bovine serum (GIBCO), glucose (4.5 g/L), and a mixture of penicillin–streptomycin–glutamine (PSG from GIBCO #10378-016).

On the next day at 08:00 hours, to coincide with the approximate habitual wake time, the medium was changed to Dulbecco’s modified Eagles Medium supplemented with 1 g/L glucose and a mixture of PSG (without fetal bovine serum). One large adipose sample from the subcutaneous depot was equally divided into 6 parts, 1 part for each of the 6 time points at the following circadian times (CT0 wake time/medium change [08:00 hours], CT4 [12:00 hours], CT8 [16:00 hours], CT12 [20:00 hours], CT16 [00:00 hours], and CT20 [04:00 hours]).

Every 4 hours, AT explants were collected in cryotubes and frozen at –80°C for later analysis of the phosphorylated HSL/total HSL (pHSL/tHSL; tHSL as loading control) ratio by Western blot and expression of LIPE. We used the phosphorylated HSL/total HSL as a measure of HSL activity because it has been shown that ex vivo Protein kinase A phosphorylation of recombinant HSL can increase the activity of the enzyme by 100% (27)

Western blot.

Explants were homogenized in RIPA buffer (0.1% sodium dodecyl sulfate, 0.1% sodium deoxycholate, 1% Triton X-100 in phosphate-buffered saline) with protease and phosphatase inhibitor cocktails (Sigma-Aldrich, St. Louis, MO, and Santa Cruz Biotechnology, Inc., respectively). The homogenates were sonicated and soluble proteins were retrieved by centrifugation for 20 minutes, 4°C, 7000g.

Protein concentration was measured in the supernatants using the bicinchoninic acid assay. Laemmli sample buffer (Bio-Rad Laboratories, Inc., USA) was added and samples were heated at 95°C for 5 minutes. Samples were then separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes, which were then blocked in TBS-T (tris-buffered saline, containing 0.1% Tween, and 5% nonfat dried milk) and immunoblotted using a 1:1000 dilution of Ser 660 rabbit monoclonal anti-pHSL antibody (Phospho-HSL (S660) #4126S, Cell Signaling Technology, Danvers, MA). This antibody measures Ser650 in humans, 1 of the 2 major determinants of protein kinase A–mediated activation of HSL against lipid substrates (28).

After washing the membranes, they were incubated with horseradish peroxide–conjugated goat antirabbit antibody (Sigma-Aldrich) in TBS-T. Antibody binding was visualized using the Pierce chemiluminescent Western blotting detection system (PerkinElmer, Madrid, Spain). Membranes were then blocked again and reprobed with 1:1000 dilution of rabbit polyclonal anti-HSL antibody (#4107, CellSignaling Technology).

Densitometry.

To quantify the densities of the pHSL and tHSL, band densitometry was performed using Image J, version 1.44 (National Institutes of Health, Bethesda, MD). The phosphorylated form of HSL was normalized to the respective levels of total HSL. In order to normalize the values of samples run in different blots, the same control sample was run in all the gels (Fig. 1).

HSL signaling in subcutaneous adipose tissue. Phosphorylated HSL (pHSL) and total HSL (tHSL) were measured by Western blot in cultured tissue samples obtained every 4 hours for 24 hours at the following times (CT0 wake time/medium change [08:00 hours], CT4 [12:00 hours], ZT8 [16:00 hours], CT12 [20:00 hours], CT16 [00:00 hours], and CT20 [04:00 hours]). (A) pHSL/tHSL across the day and night, after densitometric quantification (mean ± standard error of the mean). (B) An example of the Western blot performed for pHSL and tHSL determinations in adipose tissue.
Figure 1.

HSL signaling in subcutaneous adipose tissue. Phosphorylated HSL (pHSL) and total HSL (tHSL) were measured by Western blot in cultured tissue samples obtained every 4 hours for 24 hours at the following times (CT0 wake time/medium change [08:00 hours], CT4 [12:00 hours], ZT8 [16:00 hours], CT12 [20:00 hours], CT16 [00:00 hours], and CT20 [04:00 hours]). (A) pHSL/tHSL across the day and night, after densitometric quantification (mean ± standard error of the mean). (B) An example of the Western blot performed for pHSL and tHSL determinations in adipose tissue.

RNA extraction and LIPE expression.

Total RNA was extracted from adipocyte samples (100 mg) using RNeasy Lipid tissue kit (Qiagen, Hilden, Germany) in accordance with the manufacturer’s recommendations. RNA concentration and quality were measured by NanoDrop 2000 spectrophotometer (ThermoFisher Scientific). A total of 50 ng of RNA was reverse transcribed into cDNA using iScript cDNA synthesis kit (Qiagen) and random hexamer primers (Invitrogen, Carlsbad, CA). PCR conditions and primers for detection of LIPE mRNA and 18S rRNA transcripts have been described previously (29).

Data were obtained as Ct values according to the manufacturer’s guidelines and used to determine ΔCt values (ΔCt = Ct of the target gene – Ct of the housekeeping gene (18S) of each sample. Fold changes of gene expression were calculated by the 2−ΔΔCt method. Afterwards, in each participant we expressed each of the 6 CTs (CT0 wake time/medium change) along the day relative to the median, and we finally normalized the expression in each subject with the average of the 6 CT values to get the normalized data.

Statistical analyses.

To assess circadian rhythmicity in a variable for a subject, “cosinor analysis” was performed to fit the data of each subject with 24-hour waveforms (sinPhase and cosPhase in the JMP); then the timing of the peak (acrophase), the amplitude of the 24-hour rhythm (ie, the difference in the level between peak and trough values), and the percentage of rhythmicity (represents the percentage of the variance of the data that may be explained by the model) were obtained. To obtain the group average of circadian rhythm for a variable, data of all subjects were pooled and analyzed using a mixed effects model in which the cosinor analysis was used and subject was included as a random intercept. We did not find any interaction between diabetic and nondiabetic subjects for the main outcomes, and therefore we combined them together.

Additional Pearson’s correlation analyses were performed to assess the potential associations between rhythm characteristics and fasting duration at night and timing of food intake.

Analysis of variance was used to determine significant differences in HSL activity rhythm characteristics, such as amplitude, acrophase, and percentage of rhythmicity, between long and short fasting duration, early or late dinner, and early or late lunch, split by the median. Further adjustments were performed for total body fat mass, fat cell diameter, sleep duration, breakfast and dinner timing, when necessary.

CirWave (version 14.0, developed by R.A. Hut; available from https://www.euclock.org) was used for the cosinor analysis. The other statistical analyses were carried out by using IBM SPSS Statistics for Windows (version 20.0; Armonk, NY, USA). The level of significance for all statistical tests and hypotheses was set at P < .050.

Results

Characteristics of the participants are presented in Table 1. Timing of breakfast, lunch, and dinner (mean [range]) were 8:51 hours [07:02-11:00]; 14:40 hours [13:30-16:00], and 21:43 hours [20:00-23:00] respectively. Fasting duration was 13.21 hours [7.75-13.17]. Fat cell size was of 97.62 μm (85.32-115.39 μm).

Adipose tissue explants exhibited circadian rhythms in HSL activity (P = .023) (Fig. 2A); interestingly, HSL activity in AT reached its maximum value (acrophase) at CT around midnight (Table 1).

HSL activity and LIPE expression in subcutaneous AT, circadian rhythms, and phase map. Timing is represented both in circadian time (CT, bottom x-axis) and relative local clock time (top x-axis). (A) presence of significant circadian rhythms in HSL activity (pHSL/tHSL) in the total population (n = 18). HSL activity is represented in double plotted graphs of a 24-hour sinusoidal curve. Data (% normalized) are reported as means ± standard error of the mean (SEM) and represented by black dots. The solid line represents the 24-hour sinusoidal curve of the total population; (B) Rhythmic expression of LIPE in adipose tissue explants. Data (% normalized) are reported as means ± SEM and represented by black dots. The solid line represents the 24-hour sinusoidal curve of the total population; (C) phase map of acrophases imputed from cosinor analyses of HSL activity and LIPE expression for total population. Values are shown as mean values in dots, and upper and lower limits (95% of confidence) in solid lines.
Figure 2.

HSL activity and LIPE expression in subcutaneous AT, circadian rhythms, and phase map. Timing is represented both in circadian time (CT, bottom x-axis) and relative local clock time (top x-axis). (A) presence of significant circadian rhythms in HSL activity (pHSL/tHSL) in the total population (n = 18). HSL activity is represented in double plotted graphs of a 24-hour sinusoidal curve. Data (% normalized) are reported as means ± standard error of the mean (SEM) and represented by black dots. The solid line represents the 24-hour sinusoidal curve of the total population; (B) Rhythmic expression of LIPE in adipose tissue explants. Data (% normalized) are reported as means ± SEM and represented by black dots. The solid line represents the 24-hour sinusoidal curve of the total population; (C) phase map of acrophases imputed from cosinor analyses of HSL activity and LIPE expression for total population. Values are shown as mean values in dots, and upper and lower limits (95% of confidence) in solid lines.

A similar pattern was found in LIPE expression (Fig. 2B), which also showed robust circadian rhythms (P = .0002) and reached LIPE acrophase also at night, but slightly earlier than HSL activity acrophase (around 22:00 hours for LIPE vs 23:30 hours for HSL activity) (Table 1).

When we studied adipocyte size, we found that it was positively associated with the total levels of HSL activity (24-hour average) (r = 0.75; P = .034) and with the amplitude of HSL activity rhythm (r = 0.76; P = .030) (n = 9) (30).

Furthermore, the amplitude of HSL activity rhythm was positively correlated with the reported fasting duration (r = 0.57; P = .013) (Fig. 3A). Indeed, in those participants who had a long fasting duration, classified as those with fasting duration ≥11.20 hours (median), with an average of 11.95 ± 0.73, the amplitude of the rhythm was almost double (91% higher) the amplitude in those who had the short fasting duration (averaging 9.6 ± 0.91 hours) with amplitudes of 0.23 ± 0.03 versus 0.09 ± 0.03, respectively P = .027 (Fig. 3B).

Amplitude of HSL activity and association with fasting duration and meal timing. Amplitude is presented as % normalized, and fasting duration, dinner timing, and lunch timing in hours. (A) Correlation between fasting duration at night of the participants and amplitude of HSL activity rhythms in the adipose tissue explants (n = 18). (B) in participants who had a long fasting duration ≥11.20 hours (median), the amplitude was double compared with in those that had short fasting duration *P < .05 (n = 18). Data are reported as means ± standard error of the mean (SEM). (C) Dinner timing correlated negatively to the amplitude of HSL activity (n = 9) after adjustment by fat cell size and total body fat. (D) Habitual early dinner eaters (dinner timing <21:52 hours had higher amplitude than late dinner eaters *P < .05 (n = 9). Data are reported as means ± SEM. (E) Inverse correlation between amplitude of LIPE with lunch timing (n = 9) after adjustment by fat cell size and by total body fat. (F) Habitual early lunch eaters (lunch timing <14:40 hours had higher amplitude than late lunch eaters. *P < .05 (n = 9). Data are reported as means ± SEM.
Figure 3.

Amplitude of HSL activity and association with fasting duration and meal timing. Amplitude is presented as % normalized, and fasting duration, dinner timing, and lunch timing in hours. (A) Correlation between fasting duration at night of the participants and amplitude of HSL activity rhythms in the adipose tissue explants (n = 18). (B) in participants who had a long fasting duration ≥11.20 hours (median), the amplitude was double compared with in those that had short fasting duration *P < .05 (n = 18). Data are reported as means ± standard error of the mean (SEM). (C) Dinner timing correlated negatively to the amplitude of HSL activity (n = 9) after adjustment by fat cell size and total body fat. (D) Habitual early dinner eaters (dinner timing <21:52 hours had higher amplitude than late dinner eaters *P < .05 (n = 9). Data are reported as means ± SEM. (E) Inverse correlation between amplitude of LIPE with lunch timing (n = 9) after adjustment by fat cell size and by total body fat. (F) Habitual early lunch eaters (lunch timing <14:40 hours had higher amplitude than late lunch eaters. *P < .05 (n = 9). Data are reported as means ± SEM.

Dinner timing correlated negatively with the amplitude of HSL activity rhythm after adjusting for fat cell size and total body fat mass (r = –0.760; P = .047) (Fig. 3C). Similarly, the habitual early dinner eaters, classified as those having dinner before 21:52 hours (median), had significantly higher amplitude (60% higher) than late dinner eaters (0.24 ± 0.02 vs 0.15 ± 0.02, P = .035)) (Fig. 3D).

The amplitude of the circadian rhythms in LIPE expression correlated inversely with the timing of lunch after adjusting for fat cell size and total body fat mass (r = –0. 83; P = .020) (Fig. 3E) and was therefore lower in late lunch eaters, classified as those having lunch from 14:40 hours (median), than in early lunch eaters (a trend; P = .072) (Fig. 3F).

The percentage of rhythmicity in HSL activity, which refers to the percentage of variability explained by the model (cosine curve), was negatively associated with dinner timing (r = –0.72; P = .028) (later dinner lower rhythmicity) and positively with fasting duration (a trend, r = 0.66; P = .051) (longer fasting duration higher rhythmicity) (30).

No significant associations were found between the acrophase of the rhythm (ie, the timing of the peak) and the timing of food intake or fasting duration.

Discussion

This is the first study to demonstrate the presence in human subcutaneous AT of a circadian rhythm in HSL activity that reaches its maximum (acrophase) around midnight. This endogenous (circadian) rhythm persisted ex vivo even approximately 2 days after surgery and thus was intrinsically generated by the local clocks, independent of the influences of the central clock.

In the current work we determined HSL activity using the pHSL/tHSL ratio because it has been shown that ex vivo protein kinase A phosphorylation of recombinant HSL can increase the activity of the enzyme by 100% (27).

It is of interest that the amplitude of the rhythm was positively correlated with fasting duration at night and negatively with late dinner eating. Expression of LIPE was also negatively correlated with the timing of lunch intake. As a result, those participants who had a longer fasting displayed approximately double the amplitude (1.91 times) in HSL activity than those with short fasting duration, while those having an early dinner had 1.6-fold higher amplitude than late dinner eaters.

The characteristics of the HSL activity rhythms were similar to those of LIPE expression. However, the maximum expression of the gene was slightly earlier, 1 hour, and 20 minutes before HSL activity, the time used to translate and activate the protein by phosphorylation (31). HSL protein is primarily regulated by post-translational mechanisms; nevertheless, pretranslational mechanisms (including transcription) are important under certain physiological settings (21).

The maximum HSL activity was at CT16, which corresponds to midnight (relative local clock time), which suggests that the circadian clock is able to predict the lack of energy in the following hours, when the individual is usually sleeping and not eating, and increases HSL activity to provide enough energy for the night. Similarly, microdialysis studies in subcutaneous AT of participants under ambulatory conditions following standardized meals show that AT lipolytic activity increases at night (8). Adipose tissue is the major delivery source of nonesterified fatty acids, so while several factors control the circulating nonesterified fatty acid level in the day (specially meal timing) AT lipolysis is the major determinant at night (8).

The current data confirm previous results in animal models suggesting that the circadian clock within the adipocyte may play a significant role in body fat mobilization and accumulation (14). Since the current study has been performed ex vivo, the daily rhythm in HSL activity seems to be intrinsic to AT (circadian) and is not dependent on the action of the main clock located in the suprachiasmatic nucleus.

Cortisol rhythm is known to regulate HSL activity in adipose tissue (32). Furthermore, the highest levels of HSL activity concurred with the lowest levels of insulin sensitivity as shown in a previous experiment (13) and as expected due to their opposite metabolic function: HSL’s function is to mobilize fat while insulin’s function is to accumulate.

The 24-hour average levels of HSL activity were positively correlated with the fat cell size, results which are in agreement with previous data in humans that show that large adipocytes have increased lipolytic capacity, probably due to the enrichment of HSL, and other lipolytic enzymes (33). Interestingly, we also found a novel association between the amplitude of HSL activity and fat cell size towards a higher amplitude with higher adipocyte size suggesting also a higher mobilization of lipids, consistent with studies in mice (14, 34).

Fasting refers to any period of abstention from food, it is usually applied to an interval of time longer than the normal 8 hours of sleep (34). In the studied participants, night fasting varied between ~8 and ~13 hours and differences in this duration were related to HSL activity in AT. Longer fasting at night associated with increases in the amplitude of HSL activity, which suggests a higher mobilization of lipids from AT (14). Previous studies performed in animal models have shown that HSL activity is increased in response to fasting through the activity of several fasting lipolytic hormones such as catecholamines, glucagon, and adrenocorticotropin (2, 3).

In this line, Wilkinson et al. (20) observed significant reductions in body weight from baseline when volunteers had a night fasting duration of 14 hours. Our current results show that those participants who had a fasting duration of ~12 hours on average (long fasting duration) had double the amplitude of HSL activity in AT compared with those having a short fasting duration of ~9 hours, which may give a mechanistic explanation to the body fat mobilization when following a TRE diet (20). Nevertheless, as tissue was incubated ex vivo, further interventional studies should be performed in vivo to assess the real impact of fasting duration on AT HSL activity rhythm.

Results suggest that those participants who have a similar body fat content and fat cell size, when eating late at night (later than 21:52 hours) may have more difficulties to mobilize fat. Having a late meal may prevent the natural increase in HSL activity during those night hours, due to the acute entrance of energy to the body with food intake. Previously, having a late dinner (or eating late at night) has been associated with increased risks of obesity (17), dyslipidemia (35, 36), hyperglycemia (24, 37), and metabolic syndrome (38, 39).

Similarly, those who had a later lunch showed a reduced amplitude of LIPE expression, which could affect HSL production and activity (14). Changes in LIPE amplitude may be involved in previous data showing that those participants who have a late lunch (after 15:00 hours) have more difficulties than early eaters in losing weight with a dietary treatment (25) or bariatric surgery for obesity (40).

One limitation of the current work is the relatively low number of participants (n = 18) and that the study is performed in severe obese individuals and perhaps not directly translatable to the general population. However, these studies that include ~150 explants of fresh AT for culture are difficult to perform and limited to obese individuals, due to the AT quantity needed. Other limitations are that the habitual timing of intake was assessed by questionnaires and that the detected HSL rhythm may be influenced by dietary habits before surgery, culture conditions, and the surgical process.

In the current study, key experimental procedures that may have reset the peripheral clock were designed to coincide with the morning times, including the surgery, and the time of the medium change; therefore, the timings of an individual’s sleep/wake schedule and adipose tissue culture coincided. The significant association between the amplitude of HSL rhythms and the timing of eating and fasting duration suggests that the rhythms in HSL (in culture) reflect the habits of the individual (in vivo). Nevertheless, there is no direct evidence that adipose tissue explants cultured ex vivo are the same as in vivo adipose tissue and the interpretation of matching ex vivo experiments with the in vivo situations, such as hour of the day, should be done with caution.

In summary, we have demonstrated that HSL activity in human AT shows an endogenous circadian rhythm, with maximum activity at midnight; and that the amplitude of this rhythm increases with long fasting duration and decreases with late eating. These results may lead to a better understanding of the intricate relationships between food timing, fasting duration, and body fat and body weight regulation.

Abbreviations

    Abbreviations
     
  • AT

    adipose tissue

  •  
  • BMI

    body mass index

  •  
  • CT

    circadian time

  •  
  • Ct

    cycle threshold

  •  
  • HDL-c

    high-density lipoprotein cholesterol

  •  
  • HSL

    hormone-sensitive lipase

  •  
  • LIPE

    HSL transcript in humans

  •  
  • MetS

    metabolic syndrome

  •  
  • pHSL

    phosphorylated HSL

  •  
  • tHSL

    total HSL

  •  
  • TRE

    time-restricted eating

Acknowledgments

Financial Support: This work has been supported in part by the Spanish Government of Investigation, Development and Innovation (SAF2017-84135-R to M.G., SAF2017-88457-R to O.M., AGL2017-85270-R OM) including FEDER cofunding; the Autonomous Community of the Region of Murcia through the Seneca Foundation (20795/PI/18) to M.G., Junta de Andalucía (CTS235, CTS164) to O.M. and F.S.M., and NIDDK R01DK105072 to M.G.. F.A.J.L.S. was supported in part by NIDDK R01DK105072 and NHLBI R01HL140574. M.A.A. was supported by the Ministry of Education, Culture and Sport (Spain). CIBERehd is funded by Instituto de Salud Carlos III. K.H. was supported in part by RF1AG059867 and National Institute of Aging (NIA) RF1AG064312.

Author Contributions: M.G. conceptualized acquired funding, designed research, administrated the project, wrote the paper. O.M. and F.A.J.L.S. designed research and reviewed the paper, and performed protein analyses. K.H., and C.Z. analyzed data. M.A.A., C.Z., A.K., and J.L., performed research. F.S.M., P.A., and M.R. contributed analytic tools, intellectual advice, and reviewed the paper.

Additional Information

Disclosure Summary: There are no financial conflicts of interest to disclose except for F.A.J.L. Scheer who has received lecture fees from Bayer HealthCare (2016), Sentara HealthCare (2017), Philips (2017), Vanda Pharmaceuticals (2017), and Pfizer Pharmaceuticals (2018).

Data Availability

All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.

References

1.

Rydén
 
M
,
Jocken
 
J
,
van Harmelen
 
V
, et al.  
Comparative studies of the role of hormone-sensitive lipase and adipose triglyceride lipase in human fat cell lipolysis
.
Am J Physiol Endocrinol Metab.
 
2007
;
292
(
6
):
E1847
-
E1855
.

2.

Steinberg
 
D
,
Mayer
 
SE
,
Khoo
 
JC
, et al.  
Hormonal regulation of lipase, phosphorylase, and glycogen synthase in adipose tissue
.
Adv Cyclic Nucleotide Res.
 
1975
;
5
:
549
-
568
.

3.

Khoo
 
JC
,
Steinberg
 
D
,
Thompson
 
B
,
Mayer
 
SE
.
Hormonal regulation of adipocyte enzymes. The effects of epinephrine and insulin on the control of lipase, phosphorylase kinase, phosphorylase, and glycogen synthase
.
J Biol Chem.
 
1973
;
248
(
11
):
3823
-
3830
.

4.

Large
 
V
,
Reynisdottir
 
S
,
Langin
 
D
, et al.  
Decreased expression and function of adipocyte hormone-sensitive lipase in subcutaneous fat cells of obese subjects
.
J Lipid Res.
 
1999
;
40
(
11
):
2059
-
2066
.

5.

Bougnères
 
P
,
Stunff
 
CL
,
Pecqueur
 
C
,
Pinglier
 
E
,
Adnot
 
P
,
Ricquier
 
D
.
In vivo resistance of lipolysis to epinephrine. A new feature of childhood onset obesity
.
J Clin Invest.
 
1997
;
99
(
11
):
2568
-
2573
.

6.

Jensen
 
MD
,
Haymond
 
MW
,
Rizza
 
RA
,
Cryer
 
PE
,
Miles
 
JM
.
Influence of body fat distribution on free fatty acid metabolism in obesity
.
J Clin Invest.
 
1989
;
83
(
4
):
1168
-
1173
.

7.

Jocken
 
JW
,
Goossens
 
GH
,
van Hees
 
AM
, et al.  
Effect of beta-adrenergic stimulation on whole-body and abdominal subcutaneous adipose tissue lipolysis in lean and obese men
.
Diabetologia.
 
2008
;
51
(
2
):
320
-
327
.

8.

Hagström-Toft
 
E
,
Bolinder
 
J
,
Ungerstedt
 
U
,
Arner
 
P
.
A circadian rhythm in lipid mobilization which is altered in IDDM
.
Diabetologia.
 
1997
;
40
(
9
):
1070
-
1078
.

9.

Tahara
 
Y
,
Shibata
 
S
.
Chrono-biology, chrono-pharmacology, and chrono-nutrition
.
J Pharmacol Sci.
 
2014
;
124
(
3
):
320
-
335
.

10.

Gómez-Abellán
 
P
,
Gómez-Santos
 
C
,
Madrid
 
JA
, et al.  
Circadian expression of adiponectin and its receptors in human adipose tissue
.
Endocrinology.
 
2010
;
151
(
1
):
115
-
122
.

11.

Garaulet
 
M
,
Ordovás
 
JM
,
Gómez-Abellán
 
P
,
Martínez
 
JA
,
Madrid
 
JA
.
An approximation to the temporal order in endogenous circadian rhythms of genes implicated in human adipose tissue metabolism
.
J Cell Physiol.
 
2011
;
226
(
8
):
2075
-
2080
.

12.

Christou
 
S
,
Wehrens
 
SMT
,
Isherwood
 
C
, et al.  
Circadian regulation in human white adipose tissue revealed by transcriptome and metabolic network analysis
.
Sci Rep.
 
2019
;
9
(
1
):
2641
.

13.

Carrasco-Benso
 
MP
,
Rivero-Gutierrez
 
B
,
Lopez-Minguez
 
J
, et al.  
Human adipose tissue expresses intrinsic circadian rhythm in insulin sensitivity
.
Faseb J.
 
2016
;
30
(
9
):
3117
-
3123
.

14.

Shostak
 
A
,
Meyer-Kovac
 
J
,
Oster
 
H
.
Circadian regulation of lipid mobilization in white adipose tissues
.
Diabetes.
 
2013
;
62
(
7
):
2195
-
2203
.

15.

Garaulet
 
M
,
Gómez-Abellán
 
P
.
Timing of food intake and obesity: a novel association
.
Physiol Behav.
 
2014
;
134
:
44
-
50
.

16.

Bandín
 
C
,
Scheer
 
FA
,
Luque
 
AJ
, et al.  
Meal timing affects glucose tolerance, substrate oxidation and circadian-related variables: a randomized, crossover trial
.
Int J Obes (Lond).
 
2015
;
39
(
5
):
828
-
833
.

17.

Xiao
 
Q
,
Garaulet
 
M
,
Scheer
 
FAJL
.
Meal timing and obesity: interactions with macronutrient intake and chronotype
.
Int J Obes (Lond).
 
2019
;
43
(
9
):
1701
-
1711
.

18.

Bandín
 
C
,
Martinez-Nicolas
 
A
,
Ordovás
 
JM
, et al.  
Differences in circadian rhythmicity in CLOCK 3111T/C genetic variants in moderate obese women as assessed by thermometry, actimetry and body position
.
Int J Obes (Lond).
 
2013
;
37
(
8
):
1044
-
1050
.

19.

Dashti
 
HS
,
Scheer
 
FAJL
,
Saxena
 
R
,
Garaulet
 
M
.
Timing of food intake: identifying contributing factors to design effective interventions
.
Adv Nutr.
 
2019
;
10
(
4
):
606
-
620
.

20.

Wilkinson
 
MJ
,
Manoogian
 
ENC
,
Zadourian
 
A
, et al.  
Ten-hour time-restricted eating reduces weight, blood pressure, and atherogenic lipids in patients with metabolic syndrome
.
Cell Metab.
 
2020
;
31
(
1
):
92
-
104.e5
.

21.

Large
 
V
,
Arner
 
P
,
Reynisdottir
 
S
, et al.  
Hormone-sensitive lipase expression and activity in relation to lipolysis in human fat cells
.
J Lipid Res.
 
1998
;
39
(
8
):
1688
-
1695
.

22.

Garaulet
 
M
,
Hernández-Morante
 
JJ
,
Tébar
 
FJ
,
Zamora
 
S
.
Anthropometric indexes for visceral fat estimation in overweight/obese women attending to age and menopausal status
.
J Physiol Biochem.
 
2006
;
62
(
4
):
245
-
252
.

23.

Corbalán-Tutau
 
D
,
Madrid
 
JA
,
Nicolás
 
F
,
Garaulet
 
M
.
Daily profile in two circadian markers “melatonin and cortisol” and associations with metabolic syndrome components
.
Physiol Behav.
 
2014
;
123
:
231
-
235
.

24.

Lopez-Minguez
 
J
,
Saxena
 
R
,
Bandín
 
C
,
Scheer
 
FA
,
Garaulet
 
M
.
Late dinner impairs glucose tolerance in MTNR1B risk allele carriers: a randomized, cross-over study
.
Clin Nutr.
 
2018
;
37
(
4
):
1133
-
1140
.

25.

Garaulet
 
M
,
Gómez-Abellán
 
P
,
Alburquerque-Béjar
 
JJ
,
Lee
 
YC
,
Ordovás
 
JM
,
Scheer
 
FA
.
Timing of food intake predicts weight loss effectiveness
.
Int J Obes (Lond).
 
2013
;
37
(
4
):
604
-
611
.

26.

Sjöström
 
L
,
Björntorp
 
P
,
Vrána
 
J
.
Microscopic fat cell size measurements on frozen-cut adipose tissue in comparison with automatic determinations of osmium-fixed fat cells
.
J Lipid Res.
 
1971
;
12
(
5
):
521
-
530
.

27.

Contreras
 
JA
,
Danielsson
 
B
,
Johansson
 
C
,
Osterlund
 
T
,
Langin
 
D
,
Holm
 
C
.
Human hormone-sensitive lipase: expression and large-scale purification from a baculovirus/insect cell system
.
Protein Expr Purif.
 
1998
;
12
(
1
):
93
-
99
.

28.

Krintel
 
C
,
Osmark
 
P
,
Larsen
 
MR
,
Resjö
 
S
,
Logan
 
DT
,
Holm
 
C
.
Ser649 and Ser650 are the major determinants of protein kinase A-mediated activation of human hormone-sensitive lipase against lipid substrates
.
PLoS One.
 
2008
;
3
(
11
):
e3756
.

29.

Mairal
 
A
,
Langin
 
D
,
Arner
 
P
,
Hoffstedt
 
J
.
Human adipose triglyceride lipase (PNPLA2) is not regulated by obesity and exhibits low in vitro triglyceride hydrolase activity
.
Diabetologia.
 
2006
;
49
(
7
):
1629
-
1636
.

30.

Arredondo-Amadori
 
M
,
Zambrano
 
C
,
Kulyté
 
A
, et al.  
Data from: Circadian rhythms in hormone-sensitive lipase in human adipose tissue: relationship to meal timing and fasting duration
. Figshare repository 2011. Deposited 22 June
2020
. doi: 10.6084/m9.figshare.12535256.v1.

31.

Ben-Ari
 
Y
,
Brody
 
Y
,
Kinor
 
N
, et al.  
The life of an mRNA in space and time
.
J Cell Sci.
 
2010
;
123
(
Pt 10
):
1761
-
1774
.

32.

Samra
 
JS
,
Clark
 
ML
,
Humphreys
 
SM
,
Macdonald
 
IA
,
Matthews
 
DR
,
Frayn
 
KN
.
Effects of morning rise in cortisol concentration on regulation of lipolysis in subcutaneous adipose tissue
.
Am J Physiol.
 
1996
;
271
(
6 Pt 1
):
E996
-
1002
.

33.

Laurencikiene
 
J
,
Skurk
 
T
,
Kulyté
 
A
, et al.  
Regulation of lipolysis in small and large fat cells of the same subject
.
J Clin Endocrinol Metab.
 
2011
;
96
(
12
):
E2045
-
E2049
.

34.

Longo
 
VD
,
Panda
 
S
.
Fasting, circadian rhythms, and time-restricted feeding in healthy lifespan
.
Cell Metab.
 
2016
;
23
(
6
):
1048
-
1059
.

35.

Yoshida
 
J
,
Eguchi
 
E
,
Nagaoka
 
K
,
Ito
 
T
,
Ogino
 
K
.
Association of night eating habits with metabolic syndrome and its components: a longitudinal study
.
BMC Public Health.
 
2018
;
18
(
1
):
1366
.

36.

Chen
 
HJ
,
Chuang
 
SY
,
Chang
 
HY
,
Pan
 
WH
.
Energy intake at different times of the day: Its association with elevated total and LDL cholesterol levels
.
Nutr Metab Cardiovasc Dis.
 
2019
;
29
(
4
):
390
-
397
.

37.

Nakajima
 
K
,
Suwa
 
K
.
Association of hyperglycemia in a general Japanese population with late-night-dinner eating alone, but not breakfast skipping alone
.
J Diabetes Metab Disord.
 
2015
;
14
:
16
.

38.

Kutsuma
 
A
,
Nakajima
 
K
,
Suwa
 
K
.
Potential association between breakfast skipping and concomitant late-night-dinner eating with metabolic syndrome and proteinuria in the Japanese population
.
Scientifica (Cairo).
 
2014
;
2014
:
253581
.

39.

Berg
 
C
,
Lappas
 
G
,
Wolk
 
A
, et al.  
Eating patterns and portion size associated with obesity in a Swedish population
.
Appetite.
 
2009
;
52
(
1
):
21
-
26
.

40.

Ruiz-Lozano
 
T
,
Vidal
 
J
,
de Hollanda
 
A
,
Scheer
 
FAJL
,
Garaulet
 
M
,
Izquierdo-Pulido
 
M
.
Timing of food intake is associated with weight loss evolution in severe obese patients after bariatric surgery
.
Clin Nutr.
 
2016
;
35
(
6
):
1308
-
1314
.

This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)