Circadian Variation of Serum Leptin and Adipose Tissue Changes in Children

Summary

Higher morning serum leptin values are associated with larger adipose tissue gains in children; however, it is unclear if leptin circadian variation is itself associated with adipose tissue changes during growth. We studied the association of circadian variation in leptin with change in total body fat mass (TBFM), total body percentage fat (%FM), and trunk fat mass (TrFM). Baseline serum samples for leptin were obtained every 3h for 24h from 130 children (baseline age 9.6±2.5y; 51.1% male; BMI-Z 1.59) with mean follow-up of 11.1±4.0y and underwent dual-energy x-ray absorptiometry. ANCOVA models examined change in TBFM, %FM, or TrFM as dependent variables and number of years of follow-up, sex, race, baseline age, pubertal status, initial visit body composition, and initial visit serum leptin circadian variables (maximal diurnal leptin [acrophase], diurnal amplitude, and percentage change of amplitude) as independent factors. Although initial visit mesor (24h average) leptin was positively associated with initial visit TBFM (r²=0.78, p<0.001), %FM (r²=076, p<0.001), and TrFM (r²=0.71, p<0.001), none of the circadian leptin variables studied was significantly associated with change in TBFM, %FM, or TrFM. We found no evidence that circadian variation in serum leptin concentrations during childhood is associated with long-term changes in children’s adiposity.

1. INTRODUCTION

Leptin is an adipocyte-derived hormone that is important for energy balance, hypothalamic neuroendocrine regulation, and reproductive function and is positively associated with body fat mass.¹ Leptin concentrations differ by sex, with females having higher leptin concentrations than males independent of fat mass² even in early childhood.³ Although there is a clear positive association between serum leptin and fat mass assessed cross-sectionally, the relationship between serum leptin and change in body weight over time is less well-established.

High baseline serum leptin concentrations, independent of baseline fat mass, have been reported in some studies to predict greater BMI and fat mass over time in children at high risk for adult obesity^{4, 5} and in unselected children and adolescents.^6–10 However some studies of children have suggested that low baseline leptin may also be associated with weight gain over time.^{6, 11, 12} Data have been similarly mixed in adults with some studies showing an association between baseline leptin concentration and development of obesity while others showing no significant association.^13–18

In addition to variation in serum leptin concentrations among individuals, there is variation in leptin within an individual throughout the day. Circulating leptin concentrations have a circadian rhythm that is synchronized by the mammalian biological clock in the hypothalamic suprachiasmatic nucleus, with minimum values around noon to mid-afternoon and maximum values after midnight in humans.¹⁹ Disruption of this normal circadian variation in serum leptin has been hypothesized to be associated with human weight gain.²⁰ More generally, circadian disruption is linked to greater adiposity.²¹ For example, mice with a homozygous mutation in the Clock gene, which expresses a key transcription factor for the molecular circadian clock within pacemaker neurons of the suprachiasmatic nucleus, have attenuated diurnal feeding rhythms, increased adiposity, and features of metabolic syndrome such as hyperlipidemia, hyperglycemia, and hepatic steatosis.²² In addition, Matkovic et al²⁰ found the height of the plasma leptin nocturnal rise in a cohort of 25 adolescent females was inversely proportional to percentage body fat acquisition over a 6-month interval, suggesting that the height of the nocturnal rise of plasma leptin might help protect or predispose individuals to future short-term weight gain. In the present study, we investigated the association of circadian variables in serum leptin such as amplitude (difference between single highest and single lowest leptin value) with change in total body fat mass, percent fat mass, and trunk fat mass in a cohort of children taking part in a prospective, longitudinal study. We also measured the association of other variables commonly reported in circadian biology such as acrophase (single highest concentration), and percentage change in amplitude (difference between maximal and minimum leptin concentration divided by minimum leptin) with change in total body fat mass, percent fat mass, and trunk fat mass.

2. METHODS

2.1. Study Participants

A convenience sample of 141 healthy children (73 males, 68 females) aged 6 to 12 years were recruited for an inpatient visit between June 1996 and July 2011 by mailings sent to Montgomery County, MD, Fairfax County, VA, and Washington D.C., advertisements in local newspapers, and from referral by local physicians for participation in a prospective longitudinal observational study to characterize risk factors for obesity. The study ceased data collection in April 2018. Volunteers qualified for inclusion if they had high risk for adult obesity because they had obesity at presentation (BMI for age and sex ≥95^th percentile per NHANES I age-, sex-, and race- specific data²³) or were of normal weight (BMI between the 5^th and 85^th percentile per NHANES I age-, sex-, and race- specific data²³) with both parents having overweight (BMI ≥25 kg/m²), reported all 4 grandparents to be Non-Hispanic Black or Non-Hispanic White, and were in good general health. Standardized BMI (BMIz) was determined according to current CDC growth charts.²⁴ Exclusion criteria included significant physical or mental health issues as determined by history and physical examination, use of medications impacting weight, and pregnancy. The study was approved by the National Institute of Child Health and Human Development (NICHD) Institutional Review Board and registered at https://www.clinicaltrials.gov as NCT00001522. Each participant gave written assent, and each parent gave written consent for participation.

2.2. Outcomes

During an overnight admission at the NIH Clinical Center, baseline serum samples were obtained after an overnight fast using intravenous catheters every 3 hours for 24 hours for measurement of leptin with samples obtained at 0900, 1200, 1500, 1800, 2100, 2400, 0300, and 0600. Participants were prescribed a weight-maintenance diet and meals were delivered according to the hospital standard schedule, but consumption was not tightly controlled. Participants were instructed to sleep at their usual bedtimes at the NIH Clinical Center. Serum leptin was measured by commercially available assays (Linco Diagnostics, St. Charles, MO, or Mayo Medical Laboratories New England, Wilmington, MA). The functional sensitivity for the leptin assays was 0.4–0.5 ng/ml; intraassay and interassay variabilities were less than 8% and less than 18%, respectively.

The 24-hour leptin profiles were analyzed for acrophase leptin (single highest concentration), amplitude (difference between single highest and single lowest leptin value), and percentage change (difference between maximal and minimum leptin concentration divided by minimum leptin). Body composition was measured by dual-energy X-ray absorptiometry (DXA), by QDR 2000 pencil beam or 4500A/Discovery fan beam densitometers (Hologic Inc., Marlborough, MA) at baseline and annually or semiannually thereafter, with adjustments for known differences in these machines as previously described.⁴ Data from the Hologic QDR 2000 instrument in the pencil beam mode were analyzed using software version 5.64. Data from the Hologic QDR4500/Discovery instruments were analyzed with pediatric software versions 11.2, 12.1, 12.4.2, 12.7.3, 13.2, 13.4, and 13.4.2. No simultaneously-obtained data from children were available from the different DXA machines, so their data could not be actively compared. DXA scans were scheduled annually for duration of participation in the study. For this analysis, only baseline and final DXA scans were analyzed for each participant. Change at final follow-up in total body fat mass, percent fat mass, and trunk fat mass from initial visit scan by DXA were examined over a follow-up interval intended to last up to 15 years, through young adulthood (maximum age 27y). The change was calculated by performing direct subtractions of initial visit from follow-up visit results.

2.3. Statistical analysis

Data for leptin and fat mass were log-transformed as necessary for heteroskedasticity. Unadjusted linear regressions were used to assess correlations between initial visit diurnal leptin variables (mesor [24h mean] leptin, diurnal leptin amplitude, and diurnal leptin percentage change) and body composition.

One of the listed secondary hypotheses of the longitudinal study protocol proposed that the size of the nocturnal rise of leptin would be correlated with future weight gain. ANCOVA models were run using SPSS 27 with change in total body fat mass, percent fat mass, or total trunk fat mass as the dependent variables and number of years of follow-up, sex, race, baseline age, pubertal stage (categorized as prepuberty, early- to mid- puberty or late-puberty²⁵), initial body composition (baseline total body fat mass, percent fat mass, or total trunk fat mass), and baseline serum leptin circadian variables (acrophase leptin, leptin amplitude, and percentage change in leptin) as the independent factors. Analyses were also run for a variable similar to that used by Matkovic et al,²⁰ comparing leptin percentage rise at night (defined as average of 9PM, midnight, 3 AM and 6 AM leptin) over daytime leptin (defined as average of 9 AM, noon, 3 PM, and 6 PM leptin), for the leptin increase between 6 PM and 12 midnight, and for minimal daily leptin concentration. No interaction effects were modeled. A nominal p-value of <0.05 was considered significant.

3. RESULTS

130 children (baseline age 9.6±2.5 years; 51.1% male; 64% non-Hispanic White; 36% non-Hispanic Black; mean BMI 23.8 kg/m²; mean BMI-Z 1.59; 54% with obesity, 16% with overweight) completed diurnal leptin sampling and were followed for a mean of 11.1±4.0 years (range, 3.2–18.9 years).

The 24-hour profiles of leptin concentrations of this study’s participants were consistent with previous studies^{19, 20} revealing a nocturnal increase in circulating leptin concentrations and a nadir in the morning/early afternoon (Figure 1A). Baseline mesor leptin was positively associated with baseline percent fat mass (Figure 1B, r²=0.76, p<0.001), trunk fat mass (Figure 1C, r²=0.71, p<0.001), and total body fat mass (Figure 1D, r²=0.78, p<0.001). There were similar findings for diurnal leptin amplitude (Figures 1E, 1F, and 1G, ps<0.001) and acrophase leptin (Figures 1K, 1L, and 1M, ps<0.001). but percentage change in diurnal leptin amplitude was not significantly associated with these variables (Figures 1H, 1I,and 1J, ps>0.05).

Figure 1. Baseline Serum Leptin Diurnal Variation and Associations between Baseline Leptin Diurnal Variables and Baseline Body Composition.

A: Mean variation of serum leptin in study participants across 24 hours demonstrates expected increases in leptin at night. B-D: Associations between baseline mesor (mean 24h) leptin and baseline percent total body fat mass (B), trunk fat mass (C), and total body fat mass (D). E-G: Associations between baseline diurnal leptin amplitude (difference between single highest and single lowest leptin value) and baseline percent total body fat mass (E), trunk fat mass (F), and total body fat mass (G). H-J: Associations between baseline diurnal leptin percent change (difference between maximal and minimum leptin concentration divided by minimum leptin) and baseline percent total body fat mass (H), trunk fat mass (I), and total body fat mass (J). K-M: Associations between baseline diurnal acrophase leptin (single highest concentration) and baseline percent total body fat mass (K), trunk fat mass (L), and total body fat mass (M).

In the longitudinal analyses of change in adiposity for both log fat mass and percent fat mass, as expected, sex was associated with change in adiposity, with females gaining more adipose tissue (Supplemental Figure; p’s<0.001). Age at baseline was also associated with change in total fat mass, with younger children gaining more fat mass (p’s<0.005). Baseline fat mass percentage was associated with change in fat mass percentage. However, none of the main circadian leptin variables studied (acrophase leptin, percentage change, and amplitude) were significantly associated with change in total body fat mass, trunk fat mass, or percent fat mass (all p’s>0.085; Figure 2). Similarly, none of the analyses for leptin percentage rise at night vs. day (ps > .489) or for leptin increase between 6 PM and 12 midnight (ps > .353) were statistically significant. For minimal leptin (log transformed), there was a marginal positive association with change in log trunk fat mass (β=0.18, p=.034), but no significant associations with change in log total fat mass or percentage fat mass (ps>0.32).

Figure 2: Prospective Associations of Baseline Diurnal Leptin Variables with Longitudinally-Assessed Changes in Body Composition.

There were no significant associations between baseline diurnal leptin amplitude, diurnal leptin percentage change in amplitude, and acrophase leptin concentration for change in total fat mass (A, B, and C), change in trunk fat mass (D, E, and F) or change in percent fat mass (G, H, and I). Data have been adjusted for covariates (number of years of follow-up, sex, race, baseline age, pubertal stage, and initial body composition).

4. DISCUSSION

In this study, baseline mesor leptin was positively associated with baseline percent fat mass, total body fat mass, and trunk fat mass. In addition, diurnal leptin amplitude and acrophase leptin were positively associated with baseline adiposity measures. However, none of the main circadian serum leptin variables studied (acrophase leptin, amplitude, percentage amplitude change, 6pm-12MN leptin change, or mean day-night differences, were associated with change in total body fat mass, trunk fat mass, or percent fat mass over time in children at high risk for obesity.

Prior reports in children have examined if baseline morning serum leptin is associated with BMI or adiposity change.^4–12 Our finding of a positive association between minimal leptin (which occurs during the day) and gain in trunk fat mass is consistent with prior findings. However, few studies have investigated the effect of leptin’s circadian rhythmicity on weight change. In a cohort of adolescent females,²⁰ the height of the plasma leptin nocturnal rise was reported to be inversely proportional to percentage body fat acquisition over a 6 month interval, suggesting that the height of the nocturnal rise of plasma leptin might help protect or predispose individuals to future weight gain. As changes in percentage fat may be affected by differences in lean tissues like muscle and bone, as well as by changes in body water, the results from changes in percent fat measurements may not necessarily be conclusive. Furthermore, short-term associations may not be confirmed in longer-term studies. We did not find a significant link between baseline acrophase leptin or amplitude and fat mass gain in our study participants over a mean of 11.1±4.0 years of follow-up.

Strengths of our study include the relatively long duration of participant follow-up and use of direct measures of adipose mass. Limitations of the study include the relatively small sample size that precluded separate analyses in children with healthy weight and overweight/obesity, the use of multiple DXA scanners in the study as older machines became unavailable, and the lack of data on exact sleep and wake times of participants, which limits our ability to confirm all had a normal night’s sleep on the day we measured nocturnal rise in serum leptin (with blood draws at midnight, 3 AM, and 6 AM).

The present data provide no evidence for a role of diurnal leptin variation in children’s gains in fat mass. Further research is needed on what roles circadian variation of leptin serves for regulation of children’s body weight.