| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles |
Charles A. Dana Research Institute and the Harvard-Thorndike Laboratory of the Beth Israel Deaconess Medical Center, Division of Endocrinology and Metabolism, Beth Israel Deaconess Medical Center, Harvard Medical School (C.S.M., M.Z., R.J.C., J.S.F.), Boston, Massachusetts 02215; Department of Endocrinology and Metabolism, Gulhane School of Medicine (M.O., S.C., I.C.O.), Etlik-Ankara 06018, Turkey; Clinical Neuroendocrinology Branch, National Institutes of Health (A.B.N., P.N., V.L., M.-L.W., P.W.G., J.L.), Bethesda, Maryland 20892; Department of Biomathematics (M.A.S., R.M.E.) and Clinical Pharmacology Program and Department of Psychiatry and Biobehavioral Sciences (J.V.), University of California–Los Angeles, School of Medicine, Los Angeles, California 90095; and Department of Medicine and General Clinical Research Center, University of Virginia Health Sciences Center (J.L.), Charlottesville 22908-0202
Address correspondence and requests for reprints to: Julio Licinio, M.D., Department of Psychiatry, UCLA Gonda Center 3357A, 695 Charles Young Drive South, Los Angeles, California 90095-1761. E-mail licinio{at}ucla.edu
Abstract
Leptin signals the status of energy reserves to the brain. Leptin stimulates biosynthesis of TRH in vitro and influences the activity of the hypothalamic-pituitary-thyroid axis in vivo in rodents. Because blood levels of both leptin and TSH display diurnal variation with a distinct nocturnal rise, we sought to determine whether a relationship exists between fluctuations in circulating leptin and TSH.
We measured serum leptin and TSH levels every 7 min for 24 h in five healthy men and found that both leptin and TSH levels are highly organized and pulsatile. A similar pattern of leptin and TSH rhythms was observed, with TSH and leptin levels reaching a nadir in late morning and a peak in the early morning hours. Importantly, cosinor analysis on the absolute leptin and TSH levels revealed a statistically significant fit for a 24-h period and the two hormones showed similar probabilities of rhythm and superimposable peak values. Furthermore, this study shows a strong positive Pearson correlation between the 24-h patterns of variability of leptin and TSH in healthy subjects. Finally, the ultradian fluctuations in leptin levels showed pattern synchrony with those of TSH as determined by cross-correlation analysis, by cross-approximate enthropy and Bayessian analysis applied independently. To further explore whether these associations could reflect an underlying regulation of TSH secretion by leptin, we also studied frequently sampled leptin and TSH levels in four brothers, members of a family with leptin deficiency (one normal homozygote, two heterozygotes, and one leptin-deficient homozygote). Leptin levels of the homozygous leptin-deficient subject are detectable but bioinactive, and the rhythm of his TSH is disorganized. 24-h pattern of leptin and TSH variability in the heterozygous subjects, although significantly correlated, showed a weaker correlation compared with the strong correlation in the normal subjects.
These data are consistent with the possibility that leptin may regulate TSH pulsatility and circadian rhythmicity, but interventional studies are needed to definitively prove whether leptin regulates the minute-to-minute oscillations and ultradian rhythm of TSH levels.
LEPTIN, AN ADIPOCYTE-DERIVED hormone, has been implicated in the regulation of food intake and metabolism (1), and regulates neuroendocrine function in rodents (2). More specifically, leptin administration can blunt the food deprivation induced increase of cortisol and decrease in thyroid and gonadal hormone levels in mice (2). Interestingly, leptin’s secretion is pulsatile with a nocturnal rise (3, 4, 5, 6, 7), and the circadian/ultradian variations in plasma leptin levels are inversely related to those of ACTH and cortisol (5) in young men and display pattern synchrony with those of both LH and estradiol in young women (4). Thus, it seems that leptin oscillations are synchronous to those of the hypothalamic-pituitary-adrenal and gonadal axes.
Circulating concentrations of another pituitary hormone, TSH, also show significant circadian and ultradian variation (8, 9, 10). More specifically, TSH is secreted by the thyrotroph cells of the pituitary in a pulsatile manner with increases in pulse amplitude and frequency at night (11, 12, 13, 14, 15). Although in vitro and in vivo data demonstrate that leptin regulates TRH messenger RNA (mRNA) in the paraventricular nucleus as well as TSH secretion in response to starvation in rodents (2, 16, 17, 18, 19, 20, 21), it remains unknown whether leptin has a similar action in humans and whether the ultradian and the minute to minute dynamics of leptin levels are closely associated with those of TSH.
We studied five normal men to examine the hypothesis that pulsatile release of leptin and TSH exhibit pattern coupling, or synchronicity, by measuring plasma concentrations of these hormones every 7 min over a 24-h period. First, we used model-free Cluster analysis to quantify the pulsatile mode of leptin and TSH release. In complementary assessments of the nonpulsatile features of leptin and TSH secretion we applied Bayessian and cosinor analysis to quantitate the two hormones’ nyctohemeral rhythmicity and used the approximate enthropy (ApEn) statistic to estimate the pattern orderliness of their release process. To evaluate whether the ultradian fluctuations in leptin levels show pattern synchrony with those of TSH we used cross-correlation analysis and cross-ApEn independently. To further explore whether leptin’s ultradian and circadian patterns are causally related to those of TSH, we studied the pattern of leptin and TSH secretion in four brothers, one of them being a leptin-deficient homozygote, two heterozygotes, and one normal homozygote.
Materials and Methods
Clinical research protocol
Five young, healthy, Caucasian men were studied after giving informed consent for the clinical research protocol. Subjects had an age range of 22–26 yr, an average body mass index (BMI; weight in kilograms divided by height in meters squared) of 22.7 ± 0.14 kg/m2, and an average percentage of fat of 17.6 ± 4.19. In addition, we studied four brothers, members of a family with several leptin-deficient subjects. Volunteers with a history of medical illness, obesity, smoking, or substance abuse were excluded. One of the brothers (BMI, 54.33 and 42.8% of fat) was a leptin-deficient homozygote (-/-), two (BMI, 24.91 and 24.49, respectively, with 14.8% and 17.7% of fat) were heterozygotes (±), and one was a normal (+/+) subject (BMI, 19.53). The study subjects did not take any prescribed or over-the-counter medications, dietary supplements, or hormones and had not undertaken transmeridian travel of more than three time zones for at least 30 days before the study.
The clinical research protocol and the diet that subjects were on during their hospitalization have been described in detail previously (3). Briefly, the study subjects were admitted to the Clinical Research Center 48 h before initiation of the study to be acclimated to the Clinical Research Center environment. The study subjects were exposed to light from 0700–2300 h and darkness from 2300–0700 h, during which time they slept. Sleep was monitored by the NightCap apparatus (Healthdyne Technologies, Marietta, GA) (4, 5). On the day of blood collection, samples were withdrawn every 7 min for 1442 min beginning at 0800 h via an indwelling catheter that had been inserted the evening before the frequent blood sampling day. Thus, a total of 207 blood samples were obtained from each subject, and sera were frozen in -70 C for later assays. During the frequent blood sampling day subjects were allowed to walk from their beds to the bathroom and to an adjacent hospital dayroom to maintain a level of physical activity comparable to baseline.
Hormonal assays
Total human leptin was measured by RIA, as described previously (3, 5). The sensitivity of the assay was 0.2 ng/mL. TSH was measured by a commercially available RIA (Diagnostics Systems Laboratories, Inc., Webster, TX). The sensitivity of the assay was 0.03 µIU/mL; the intra-assay coefficient of variation was less than 4%, and interassay coefficient of variation was less than 9.2%.
Pulse analysis: CLUSTER
To assess possible changes in plasma leptin and TSH concentration pulse parameters during daytime (0800–1700 h, which is characterized by lower leptin levels) and nighttime (2300–0800 h, which is characterized by higher leptin levels), we used CLUSTER (3, 22), a largely model-free computerized pulse analysis algorithm to identify statistically significant pulses in relation to dose-dependent measurement error in each hormone time series. The details of the protocol for this analysis have been previously described (3, 5).
Nyctohemeral (24-h) rhythmicity: cosinor analysis
The presence of diurnal rhythms was tested with single and population cosinor as well as multiple components analysis (23) using ChronoLab 3.0.3 (24, 25). This procedure fits the data to a cosine function of a fixed anticipated period and calculates an estimate of the MESOR (Midline Estimating Statistic of Rhythm), which represents the mean value of the rhythmic function; amplitude, which is the difference between the peak and the MESOR of the fitted curve; acrophase, the time lag expressed in radians from a reference time point (in our case from midnight) of the top point in the fitted curve; and telophase, the time lag from the reference point of the lowest point in a fitted curve, which occurs exactly 12-h after the acrophase. The probability of rhythm was tested by an F-test, according to the null hypothesis of zero amplitude. Tests for normality and independence of residuals and homogeneity of their variance were also used (23, 24, 25).
Regularity of the two hormone series: ApEn
ApEn is a large-scale, translation-invariant, and model-independent regularity statistic developed to quantify the orderliness of sequential measures, such as hormonal time series over a 24-h period (5). Larger ApEn values correspond to greater randomness (irregularity). The basic derivation and calculation of ApEn have been presented previously (5). In addition, to evaluate relative regularity of TSH and leptin concentrations, we used the cross-ApEn. This measure quantifies the conditional regularity or synchrony of point-by-point variations across two time series, as described previously (5).
Cross-correlation analysis
Cross-correlation was calculated after lagging the concentration time series of one hormone relative to the concentration time series of the other hormone. Cross-correlation was carried out at variable lags that are the times in minutes separating the consecutive samples in the paired hormone series of interest, as described previously (5). Significant cross-correlation values for the study group at any particular lag were tested against the null hypothesis of purely random associations as previously described (5).
Bayesian mixed effect model
To examine differences in the baseline levels and diurnal
pulsatility of serum leptin and TSH between normal homozygotes,
heterozygotes and leptin-deficient homozygotes, we constructed a
Bayesian nonlinear, hierarchical model (26). As with our
earlier cosinor analysis, this model assumes that each observed leptin
or TSH measurement yit for individual
i = 1, 2, ... N at time t follows basic
sinusoidal behavior, such that
 | (1) |
i is the phase shift (in radians
from 8AM) and 
N(0,
2).
We further assume that the individual effects
ai, bi and
i are, a priori,
independent, and normally distributed about their unknown grand
population means A, B and 
with unknown variances
2A, 2B,
2C. We take vague, yet completely proper,
priors on these unknowns and fix the support of
i to lie within (0,
) to avoid
confounding it with the sign of ai. This model
assumes a fixed period of 24 h. As ai approaches 0 for an individual, the periodic term in this model vanishes and we are left with random variation about the baseline. Examining whether the mean value of ai for a subset of individuals equals 0 tests whether or not this group exhibits periodic oscillation. Likewise, examining whether the mean values of bi for different subsets of individuals equals 0 tests whether baseline values differ between groups.
Due to our small sample size, a Bayesian approach overcomes the shortfall of a traditional nonlinear, mixed effects model that relies on asymptotics to make inference.
We fit our model using the standard Metropolis-Hastings-within-Gibbs
algorithm (27), breaking up our parameter space in blocks
(A, ai, 2A,), (B,
bi, 2B), (C,
ci, 2C),
(2). We use random walk drivers with their
variances adjusted to give appropriate acceptance rates to improve
mixing, and check convergence using multiple runs with starting values
drawn directly from our priors.
Following the Bayesian paradigm, we identify a parameter or function of parameters as significantly different from 0 if the 95% Bayesian credible interval for that parameter or function does not lie over 0. We determine the 95% Bayesian credible interval from the posterior of the parameter or function given the data. This is the Bayesian equivalent to setting a P value of 0.05 to designate significance.
Statistical analysis of mean hormone levels. Comparisons were made by using Student’s t test or ANOVA. When multiple comparisons were made a protected value of P < 0.01 was used to determine statistical significance, as described previously (4). Results are presented as the mean ± SE. Statistical significance was accepted for P less than 0.05.
Results
Mean leptin concentrations
The mean 24-h leptin concentration was 5.26 ± 0.091 ng/mL in the group of normal volunteers. The mean 24-h immunoreactive but bioinactive leptin concentration was 0.41 ± 0.002 ng/mL in the leptin-deficient homozygote (-/-; percentage of fat mass, 42.8%). The mean 24-h leptin concentrations were 1.07 ± 0.03 ng/mL and 0.78 ± 0.02 ng/mL in the two heterozygotes (+/-; percentage of fat masses, 17.7% and 14.8%, respectively), and 2.84 ± 0.09 ng/mL in the normal homozygote (+/+; normal range for percentage of fat mass, 12–18%).
Thyroid function
Early morning free T3 was 4.1 pg/mL in the leptin-deficient homozygote, 3.3 pg/mL in the heterozygote, and 2.15 pg/mL in the normal homozygote, with the normal range varying from 2.3–4.2 pg/mL. Early morning free T4 was 1.3 ng/dL, 1.2ng/dL, and 1.2 ng/dL in the -/-, -/+, and +/+ subjects, respectively. Early morning TSH levels were 2.6 IU/L in the leptin-deficient homozygote, with a normal range of 0.35–5.5IU/L. TSH levels were 1.7 and 1.9 IU/L for the -/+ and +/+, subjects respectively. TRH test showed that basal TSH levels were 3.2 IU/L in the -/- subject, with peak levels of 17.0 IU/L and 7.7 IU/L after 90 min. The thyroid gland was normal, and antithyroid antibodies were negative in homozygous, heterozygous, and normal subjects.
Circadian rhythms in normal subjects and in a family with leptin deficiency
The circadian pattern of leptin release was similar in all five
normal subjects. Leptin levels display a nadir after the onset of
sampling in the morning and increase thereafter to a peak that is
reached in the early morning hours, as previously described and as
illustrated in Fig. 1
. A strikingly
similar pattern of TSH secretion was observed in this study with TSH
levels reaching a nadir in the morning and a peak in the early morning
hours (Fig. 1). Importantly, cosinor analysis on the absolute leptin
and TSH levels revealed a significant fit for a 24-h period in all
normal subjects. Both hormones showed similar probabilities of rhythm
and superimposable peak values. Also acrophase times were placed at
virtually the same time point in the early morning (leptin, 0157
h; TSH, 0218 h). The data obtained were well represented by the
cosine curve, and maximum values did not occur at appreciably different
time lags from the acrophase and its confidence limits, as tested by
multiple components analysis. Based on these findings, we can state
that leptin and TSH show virtually identical circadian rhythms and
possible ultradian variations in TSH rhythm seem not to obscure the
24-h trends (Fig. 1).
|
Ultradian rhythms in normal subjects
Both leptin and TSH levels are highly organized and pulsatile in
healthy Caucasians. Furthermore, this study shows a strong positive
Pearson correlation between the 24-h patterns of variability in leptin
and TSH (Fig. 2). Synchronicity between
leptin and TSH time series was assessed independently by
cross-correlation analysis and by cross-ApEn. We found statistically
significant cross-correlation values between leptin and TSH during the
time windows of -119 to +140 min (P < 0.01), and
these results were confirmed by ApEn analysis. To independently
determine pattern synchronicity between leptin and TSH, we also
determined the cross ApEn between these two hormones as described
previously (5). The mean random cross-ApEn, which would
reflect unrelatedness between these two hormones by pattern consistency
or synchrony was 2.268 ± 0.030. The significantly reduced
(nonrandom) cross-ApEn values between leptin and TSH (average of
1.816 ± 0.245, P < 0.02) demonstrate that there
is synchronicity or pattern coupling between these two hormones in
healthy Caucasians. Thus, using two independent methods, we show
significant temporal association between the 24-h rhythms of leptin and
those of TSH.
|
In this family, it is clear that in its homozygous form, leptin gene mutation results in detectable but bioinactive leptin levels. In addition, although all four brothers are within the ApEn range obtained by studying the five normal unrelated volunteers, it is obvious that the leptin rhythmicity in the leptin-deficient homozygote (-/-) is clearly disorganized. In addition, leptin rhythmicity in the normal brother is more organized than that in the two heterozygotes brothers, which in turn is relatively more organized than the secretion in the leptin deficient homozygote (-/-). Similarly, organization of their TSH rhythms, as reflected by the ApEn statistic, parallels the organization of leptin in the four brothers. Finally, it seems that leptin gene status influences cross-ApEn in a similar way. Thus, leptin and TSH levels are highly correlated in the five unrelated normal men and the normal (+/+) brother, but the respective correlations, although significant, are weaker in the heterozygous (+/-) brothers and least structured in the leptin-deficient subject (-/-).
Furthermore, the 24-h patterns of leptin and TSH variability are
strongly correlated in healthy subjects (P < 0.01,
Fig. 2
) and in the normal (+/+) brother (r = 0.83,
P < 0.001). Interestingly, although the 24-h patterns
of leptin and TSH variability are significantly correlated in the
heterozygote (+/-) brothers, the corresponding Pearson correlation
coefficients are smaller (r = 0.37 and .28, P <
0.01). In contrast, the corresponding Pearson correlation coefficients
are nonsignificant in the homozygote (-/-) brother (r =
0.11, P = NS).
In summary, study of this unique family demonstrates that the levels of TSH ApEn reflect those of leptin ApEn, which in turn is closely associated with leptin gene status. These data, albeit limited, point to a temporal association between the organization of the 24-h TSH and leptin levels that, when considered in the context of previous in vitro and in vivo data, may indicate that TSH secretion could be influenced by leptin’s circulating levels.
Bayesian model. Figure 3
presents the observed leptin and TSH levels and the Bayesian model fits
for the leptin-deficient subjects (one homozygous, +/+; two
heterozygous, -/+; one homozygous leptin deficient, -/-). In each
time series, the solid line represents the mean model
response, whereas the dashed lines trace out the 95%
Bayesian prediction interval for the model. Large diurnal leptin
oscillations occur in the +/+ group and smaller oscillations are still
visible in the -/+ subjects. The diurnal oscillations of the normal
subjects studied are similar to those of the +/+ subject (data not
shown). However, little, if any, periodic behavior is noticeable in the
homozygous leptin-deficient subject. Except for this subject, maximal
leptin and TSH levels occur at night. The former subject’s TSH levels
achieve a maximum in the early morning. These observations are
consistent with the estimated model parameters we report in Table 1.
|
|
reports the mean and 95% credible intervals for the amplitude,
baseline and phase parameters for both leptin and TSH for all three
groups of members of the leptin-deficient family. The table also
reports estimates of the relative amplitude. Relative amplitude is a
measure of oscillation that adjusts for varying baselines. Examining
leptin results first, the estimates of amplitude and relative amplitude
for the +/+ and -/+ groups are significantly nonzero, signifying
diurnal oscillation. However, the 95% credible interval of the
amplitude and relative amplitude for the -/- group falls squarely
over zero, implying no significant diurnal behavior. Baseline estimates
for all groups are significantly different from each other with the +/+
group expressing the highest levels of leptin and the homozygous
leptin-deficient subject expectedly the lowest. As the leptin levels in
the homozygous leptin-deficient subject do not significantly oscillate,
it is impossible to estimate this phase. Referring to the TSH model
fits in Table 1, all three groups possess diurnal variation as
expressed through significantly nonzero estimates of amplitude and
relative amplitude. The phase of oscillation is also significantly
different in the -/- group from the other two groups. The homozygous
leptin-deficient subject’s TSH levels reach a maximum at approximately
0900 h (0830h to 0930 h, 95% credible interval). Finally,
returning to Fig. 1, the leptin and TSH time series also exhibit short
time-scale pulsatile behavior, in addition to the diurnal variation.
This effect is especially pronounced in the TSH data from the first USA
+/+ subject, where we observe several short pulses consistent with the
estimated diurnal response. We plan to develop further models to
capture both behaviors. Discussion
Circulating levels of leptin, an adipocyte-derived hormone that reflects the amount of energy stored in adipose tissue, are highly organized and have a distinct diurnal and circadian rhythm (3, 5, 6, 7). Previously, we examined the associations between the pulsatile secretion of leptin and that of gonadotrophins and ACTH (3, 5). As leptin levels increase at night, leptin pulsatile secretion becomes more organized, and the fluctuations of plasma leptin concentrations are synchronous with those of LH and estradiol (5). Moreover, the patterns of synchrony of leptin/LH and leptin/estradiol are more orderly at night (3, 5). We, therefore, suggested that the nocturnal rise in leptin concentrations may contribute to the nighttime changes in LH pulsatility patterns. The ultradian fluctuations of leptin levels also show a pattern synchrony with those of ACTH as well as LH and estradiol as determined by cross-ApEn (3, 5).
A circadian variation has also been shown for TSH (8, 11, 12, 13, 14, 15, 29). TSH levels rise late in the evening, peak between 2300 and 0400 h and decline thereafter to a nadir which occurs before noon (11, 12, 13, 14). In addition, circulating TSH levels display an ultradian rhythm with an increase of TSH pulse amplitude and frequency at night (8, 11, 12, 13, 14). In this study, we assessed whether circulating levels of TSH show a pattern synchrony with circulating leptin levels. We confirm, in a group of young healthy Caucasians, that frequently sampled serum leptin and TSH concentrations have circadian and ultradian release patterns with quantifiable levels of orderliness that are significantly different from random (3). Furthermore, we demonstrate that leptin and TSH have almost identical circadian rhythms. Finally, by using two independent methods to assess temporal linkage, cross- correlation analysis and cross-ApEn, we demonstrate that the ultradian fluctuations in leptin levels show pattern synchrony with those of TSH in normal subjects. In addition to normal subjects, we have studied members of a family with genetic leptin deficiency. Data from members of this family indicate that leptin deficiency is closely associated with dysregulated patterns of TSH pulsatile and circadian rhythm. These data suggest that leptin may have a role in regulating TSH secretion in humans.
The mechanism underlying TSH pulsatility and circadian rhythmicity, and
how leptin may influence its rhythmicity remains largely unknown,
however. Although dopamine may affect TSH pulse amplitude, this
neurotransmitter has not been shown to affect TSH pulse frequency
(14). Neither serotonin (30), melatonin
(31), nor 
-adrenergic pathways play significant roles
in determining TSH circadian rhythmicity (32). The fact
that caloric restriction results in a decrease in basal or
TRH-stimulated TSH levels, TSH pulse amplitude and nocturnal increase
of TSH (33, 34, 35) suggest that a factor that reflects
nutritional status and energy stores could be responsible for these
effects of caloric restriction. Because deficient nocturnal surge of
TSH has also been observed in central hypothyropidism
(29), it is reasonable to speculate that leptin, an
adipocyte-derived hormone that communicates information on the amount
of energy stored in adipose tissue to the hypothalamus, could also
regulate TSH pulsatile and circadian rhythm. More importantly, it has
been previously shown that leptin’s pulsatility is synchronous to that
of gonadotropins. Since the pulsatilities of TSH and gonadotropins are
concordant, a common hypothalamic pulse generator may regulate
circulating concentrations of both hormones (8, 36). This
hypothesis is further supported by recent experimental evidence.
ObRb mRNA is expressed in the paraventricular nucleus (PVN)
(37), and leptin can increase the expression of the proTRH
gene in the hypothalamic PVN (17), thus regulating the
hypothalamic-pituitary-thyroid axis (17). Leptin increases
ProTRH biosynthesis and TRH release by direct actions on TRH neurons or
indirectly by altering neuropeptides that themselves regulate the TRH
neurons (18, 19). Both NPY neurons (38), and
-MSH and AgRP expressing neurons have nerve fibers that project from
the arcuate nucleus to the PVN and are in close proximity to the
Pro-TRH-expressing neurons (39, 40). These and other
neuropeptides could play a role in modulating the role of TRH neurons
by mediating the effects of leptin and/or independently of leptin.
Hypophysiotropic TRH neurons in the PVN are also subjected to negative
regulation by serum thyroid hormone levels (41)
(42). When serum thyroid hormone levels drop, TRH
secretion is increased and results in increased TSH secretion.
Conversely, increase of serum thyroid levels suppresses secretion of
PVN TRH and results in reduced TSH secretion (42). This
regulatory system is altered during fasting, however, because decreased
serum thyroid levels in response to fasting are associated with a
reduction in both TRH and TSH secretion (43, 44). It has
been previously shown that leptin administration can prevent the fall
in thyroid hormone in response to starvation in rodents
(2) (17, 37) and that this is associated with
maintenance of TRH mRNA expression in the PVN (17). These
observations suggest that leptin plays an important role in regulating
TRH biosynthesis in the PVN and thus TSH and possibly,
T3 and T4 levels.
It is also possible that leptin, in addition to its action at the hypothalamic level, could act directly on the thyrotrope to influence TSH secretion. Recently both leptin and leptin-receptor expression has been demonstrated in TSH expressing cells of rat and mouse anterior pituitaries (20, 46, 47) and can potentially influence its secretion.
Thus, several mechanisms exist by which leptin could regulate TSH pulsatile and circadian secretion. However, it is also possible that both leptin and TSH rhythms are regulated by a common hypothalamic pulse generator or that leptin influences TSH rhythmicity indirectly (i.e. by regulating the rhythmicity of other hormone levels such as LH or cortisol) (3, 5). It has been previously proposed that serum estrogen and/or androgen levels may affect serum TSH response to TRH (48, 49) and may influence TSH secretion (50). Because circadian and pulsatile changes in TSH secretion are not influenced by cortisol secretion (51), it seems unlikely that the underlying association between leptin and cortisol is responsible for the association between leptin and TSH rhythms observed in this study. The full spectrum of factors influencing TSH rhythmicity remains to be fully elucidated.
It has been previously demonstrated that most children with either leptin (52, 53) or leptin receptor mutations (54) have hypothalamic hypothyroidism (52, 55). In contrast to these observations, all adult leptin-deficient homozygous and heterozygous adult patients in this study have normal thyroid hormone levels but abnormal pattern of TSH secretion. It is possible that as leptin-deficient subjects grow other factors that regulate TSH secretion, the nature of which remains to be elucidated, assume a more important role in TSH regulation and maintenance of normal circulating thyroid hormone levels.
Based on these data, we propose that leptin regulates pulse patterns and absolute levels of leptin may contribute to the regulation of fluctuations in TSH levels in a fashion similar to the possible role of leptin in regulating fluctuations in the levels of ACTH or LH (3). Based on these data, it appears that impaired circadian rhythm of TSH is not absolutely necessary for maintenance of normal thyroid hormone levels in adults, since the homozygote -/- subject had normal thyroid hormone levels, despite disorganized TSH circadian rhythm. A limitation of the current study is the lack of data from morbidly obese subjects that could have been used as controls for the leptin deficient homozygote subject. Additional studies including morbidly obese subjects as controls, as well as patients with anorexia nervosa (56, 57) may be warranted.
This role of leptin provides a link between nutritional status and the episodic activity of the hypothalamic-pituitary-peripheral axes and may have important implications for neuroendocrine function in humans. Despite the highly statistically significant associations shown in this study and the data from leptin-deficient subjects strongly suggesting that leptin may play a direct role in regulating TSH rhythmicity, cross-sectional and observational data cannot prove causality. Alternatively, these data also support a permissive role for leptin in the regulation of hypothalamic-pituitary-thyroid axis.
Thus, interventional studies to fully prove causality and basic research studies to elucidate the underlying mechanism are clearly needed. Such experimental studies administering either leptin itself or leptin agonists and antagonists could definitively prove the causal direction of the associations observed in this study and could contribute to the elucidation of the mechanisms underlying the associations between the dynamics of leptin secretion and neuroendocrine function in humans.
Acknowledgments
We thank Dr. S. Refetoff for critically reviewing the manuscript.
Footnotes
1 Supported by the Boston Obesity and Nutrition Research Award
(P30-DK-46200) and by Beth Israel Hospital General Clinical Research
Center Grant M01-RR01032. 
2 Supported by a Howard Hughes Medical Institute Predoctoral
Fellowship and NIH Training Grant GM08042.
Received November 14, 2000.
Revised March 19, 2001.
Accepted March 27, 2001.
References
-Melanocyte-stimulating hormone is contained in nerve terminals
innervating thyrotropin-releasing hormone-synthesizing neurons in the
hypothalamic paraventricular nucleus and prevents fasting-induced
suppression of prothyrotropin-releasing hormone gene expression. J
Neurosci. 20:1550–1558.This article has been cited by other articles:
 |
|
 |
|
  W. Russell, R. F. Harrison, N. Smith, K. Darzy, S. Shalet, A. P. Weetman, and R. J. Ross Free Triiodothyronine Has a Distinct Circadian Rhythm That Is Delayed but Parallels Thyrotropin Levels J. Clin. Endocrinol. Metab., June 1, 2008; 93(6): 2300 - 2306. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Landa, S. I. Garcia, M. L. Schuman, A. Burgueno, A. L. Alvarez, F. E. Saravia, C. Gemma, and C. J. Pirola Knocking down the diencephalic thyrotropin-releasing hormone precursor gene normalizes obesity-induced hypertension in the rat Am J Physiol Endocrinol Metab, May 1, 2007; 292(5): E1388 - E1394. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Ortega, N. Pannacciulli, C. Bogardus, and J. Krakoff Plasma concentrations of free triiodothyronine predict weight change in euthyroid persons Am. J. Clinical Nutrition, February 1, 2007; 85(2): 440 - 445. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. S. Farooqi and S. O'Rahilly Genetics of Obesity in Humans Endocr. Rev., December 1, 2006; 27(7): 710 - 718. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Reinehr, G. de Sousa, and W. Andler Hyperthyrotropinemia in Obese Children Is Reversible after Weight Loss and Is Not Related to Lipids J. Clin. Endocrinol. Metab., August 1, 2006; 91(8): 3088 - 3091. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Chan, G. Matarese, G. K. Shetty, P. Raciti, I. Kelesidis, D. Aufiero, V. De Rosa, F. Perna, S. Fontana, and C. S. Mantzoros Differential regulation of metabolic, neuroendocrine, and immune function by leptin in humans PNAS, May 30, 2006; 103(22): 8481 - 8486. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Kok, F. Roelfsema, M. Frolich, A. E. Meinders, and H. Pijl Spontaneous Diurnal Thyrotropin Secretion Is Enhanced in Proportion to Circulating Leptin in Obese Premenopausal Women J. Clin. Endocrinol. Metab., November 1, 2005; 90(11): 6185 - 6191. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Misra, K. K. Miller, K. Kuo, K. Griffin, V. Stewart, E. Hunter, D. B. Herzog, and A. Klibanski Secretory dynamics of leptin in adolescent girls with anorexia nervosa and healthy adolescents Am J Physiol Endocrinol Metab, September 1, 2005; 289(3): E373 - E381. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. G. Smith Development of Growth Hormone Secretagogues Endocr. Rev., May 1, 2005; 26(3): 346 - 360. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. W. Kok, F. Roelfsema, S. Overeem, G. J. Lammers, M. Frolich, A. E. Meinders, and H. Pijl Altered setting of the pituitary-thyroid ensemble in hypocretin-deficient narcoleptic men Am J Physiol Endocrinol Metab, May 1, 2005; 288(5): E892 - E899. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. G. Smith, L. Betancourt, and Y. Sun Molecular Endocrinology and Physiology of the Aging Central Nervous System Endocr. Rev., April 1, 2005; 26(2): 203 - 250. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Spiegel, R. Leproult, M. L'Hermite-Baleriaux, G. Copinschi, P. D. Penev, and E. Van Cauter Leptin Levels Are Dependent on Sleep Duration: Relationships with Sympathovagal Balance, Carbohydrate Regulation, Cortisol, and Thyrotropin J. Clin. Endocrinol. Metab., November 1, 2004; 89(11): 5762 - 5771. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. T. Gibson, I. S. Farooqi, M. Moreau, A. M. DePaoli, E. Lawrence, S. O'Rahilly, and R. A. Trussell Congenital Leptin Deficiency Due to Homozygosity for the {Delta}133G Mutation: Report of Another Case and Evaluation of Response to Four Years of Leptin Therapy J. Clin. Endocrinol. Metab., October 1, 2004; 89(10): 4821 - 4826. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Bluher and C. S. Mantzoros The Role of Leptin in Regulating Neuroendocrine Function in Humans J. Nutr., September 1, 2004; 134(9): 2469S - 2474S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I.S. Farooqi and S. O'Rahilly Monogenic Human Obesity Syndromes Recent Prog. Horm. Res., January 1, 2004; 59(1): 409 - 424. [Abstract] [Full Text]
|
|
|
|
 |
|
 D. M. Keenan, F. Roelfsema, N. Biermasz, and J. D. Veldhuis Physiological control of pituitary hormone secretory-burst mass, frequency, and waveform: a statistical formulation and analysis Am J Physiol Regulatory Integrative Comp Physiol, September 1, 2003; 285(3): R664 - R673. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I.A. Harsch, P.C. Konturek, C. Koebnick, P.P. Kuehnlein, F.S. Fuchs, S. Pour Schahin, G.H. Wiest, E.G. Hahn, T. Lohmann, and J.H. Ficker Leptin and ghrelin levels in patients with obstructive sleep apnoea: effect of CPAP treatment Eur. Respir. J., August 1, 2003; 22(2): 251 - 257. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. A. Oral, E. Ruiz, A. Andewelt, N. Sebring, A. J. Wagner, A. M. Depaoli, and P. Gorden Effect of Leptin Replacement on Pituitary Hormone Regulation in Patients with Severe Lipodystrophy J. Clin. Endocrinol. Metab., July 1, 2002; 87(7): 3110 - 3117. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Chan, S. Bluher, N. Yiannakouris, M. A. Suchard, J. Kratzsch, and C. S. Mantzoros Regulation of Circulating Soluble Leptin Receptor Levels By Gender, Adiposity, Sex Steroids, and Leptin : Observational and Interventional Studies in Humans Diabetes, July&nbs |
