| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
,
Milan S. Petakov,
Sanja Rai
evi,
Dragan Mici,
Jelena Marinkovi,
Carlos Dieguez,
Felipe F. Casanueva and
Vera Popovi
Institute of Endocrinology, Diabetes, and Diseases of Metabolism (S.S.D., M.S.P., S.R., D.M., V.P.) and Institute of Social Medicine and Statistics (J.M.), School of Medicine, University Clinical Center, Beograd, SR Yugoslavia; and Departments of Physiology (C.D.) and Medicine (F.F.C.), School of Medicine and Complejo Hospitalario Universitario de Santiago de Compostela, Santiago de Compostela University, Santiago de Compostela, Spain
Address correspondence and requests for reprints to: F. F. Casanueva, M.D., Ph.D., Department of Medicine, Endocrine Unit, San Francisco Street, P.O. Box 563, E-15780, Santiago de Compostela, Spain. E-mail: melage{at}uscmail.usc.es
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Control subjects and acromegalics had similar BMI, but patients with active acromegaly had significantly lower mean leptin level (mean ± SEM; in men, 2.6 ± 0.4 vs. 7.1 ± 1.1 µg/L, P = 0.003; in women, 16.0 ± 3.4 vs. 23.5 ± 3.1 µg/L; P = 0.036). Mean 8-h leptin correlated with BMI (r = 0.57, P = 0.007, in controls; r = 0.70, P = 0.001, in patients). In stepwise regression analysis with mean 8-h leptin as a dependent variable, BMI (P < 0.001) and gender (P = 0.01) in acromegalics entered the equation, whereas in control subjects gender, free fatty acids, insulin, and age accounted for 99.3% in leptin variability. After surgery, BMI did not change significantly; and glucose (P = 0.014), GH (P < 0.001), and IGF-I (P < 0.001) levels together with AUCins (P = 0.002) decreased, whereas mean leptin concentration rose significantly and attained normal levels (4.1 ± 0.8 µg/L, P = 0.028) in acromegalic men and (23.6 ± 4.7 µg/L, P = 0.003) in acromegalic women. Correlation between leptin level and BMI was preserved after surgery (r = 0.62, P = 0.005). In stepwise regression analysis, free fatty acids (P = 0.04) contributed to 26.8% of the variance in corrected-leptin (for BMI and gender). Leptin concentration peak height and interpeak nadir level rose significantly (P = 0.033 and P = 0.037) after surgery by Cluster analysis, without significant changes in leptin pulse frequency and incremental peak amplitude. Nocturnal rise of leptin (mathematically described by a cubic curve) was characterized by an acrophase just after midnight, before and after surgery. The amplitude and the average leptin concentration of the cubic fit increased significantly after surgery (P = 0.028 and P < 0.001).
In conclusion in acromegalic patients: 1) leptin secretion maintains the pulsatility and nocturnal rise; 2) the gender-based leptin differences are preserved; 3) GH-IGF-I normalization leads to a rise in leptin that is not related to changes in BMI; and 4) the possible role of rise in leptin levels when assessing clinical and metabolic outcome of therapy in acromegalic patients deserves additional studies.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Twenty patients with active acromegaly due to pituitary tumor (6 males and 14 females, aged 25–65 yr) and 20 age-, gender-, and BMI-matched control subjects (8 males and 12 females, aged 27–65 yr) participated in this study, after providing written informed consent approved by the Ethical Committee. Patients were admitted at the Institute of Endocrinology, Diabetes, and Metabolic Diseases (University Clinical Center, Belgrade, Yugoslavia). Acromegaly was diagnosed by classical clinical features, elevated basal levels of GH (mean of four-points day curve), IGF-I, and nonsuppressibility of GH during an oral glucose tolerance test (OGTT). Computerized tomography and/or magnetic resonance imaging of the sellar region confirmed the presence of pituitary macroadenoma in all patients. Anterior pituitary function was normal and was assessed before and 60 days after trans-sphenoidal surgery. Normal levels of basal cortisol and normal response of the hypothalamo-pituitary-adrenal axis during an insulin tolerance test excluded ACTH deficiency. All patients had normal concentrations of T4, T3, and TSH, as well as gonadotropin and estradiol/testosterone levels. Pre- and postoperative PRL levels were in normal range in all patients. None of the patients was hypopituitary after surgery.
Clinical protocol
The evaluation of patients included pre- and postoperative determinations of basal GH (mean value of the four-points day curve), IGF-I, and lipid levels. Insulin, glucose, and GH concentrations were determined during standard 75-g 2-h OGTT. The mean 8-h leptin concentration for each subject was measured from a sample formed by pooling of equal aliquots from that subject’s study samples collected hourly beginning at 2200 h until 0600 h the next morning. Additionally, in 10 acromegalics (3 males and 7 females) who fulfilled criteria for postoperative biochemical remission of disease (normal GH and IGF-I levels and GH < 1µg/L after glucose loading), we assessed leptin pulsatility during the night for 8 h encompassing 2200 h through 0600 h on the next day with 10-min sampling intervals. Control subjects underwent the same protocol, except that OGTT and assessment of leptin pulsatility were not performed. Sera for GH, IGF-I, insulin, and leptin were stored at -20C until analysis.
Assays
Plasma glucose was determined by the glucose oxidase method (Glucose Autoanalyzer; Beckman Coulter, Inc., Fullerton, CA). Plasma high-density lipoprotein (HDL) concentration was obtained after precipitation of low-density lipoprotein (LDL) and very LDL with dextrane sulfate and magnesium chloride. Commercial enzymatic methods were used in determination of serum cholesterol (Monotest; Roche Molecular Biochemicals, Penzberc, Germany) and triglycerides (Peridecrome; Roche Molecular Biochemicals), and the serum LDL cholesterol level was calculated by the Friedwald formula (26, 27, 28, 29). The interassay coefficient of variation (CV) was less than 3.8%, 5.0%, and 2.5% for total cholesterol, HDL, and triglycerides, respectively. The colorimetric method was used to determine plasma concentration of free fatty acids (FFA), as described before (30).
GH was determined by solid phase two-site fluorometric assay based on direct sandwich technique with two monoclonal antibodies directed against two different epitopes of human GH molecule (Delfia; Wallac, Inc. Oy, Turku, Finland). The minimal detection limit was 0.011 µg/L, and intra- and interassay CV were less than 5.0% and 6.3%.
Total IGF-I was performed using polyclonal RIA (INEP, Zemun, SR Yugoslavia) after acid ethanol extraction. The sensitivity of assay was 0.2 µg/L; the within in and between CV were 3.2% and 9.1%, respectively. Insulin levels were determined by RIA (INEP), with a lower limit of sensitivity of 3.0 mIU/L, and average intra- and interassay CV were 4.7% and 7.2%, respectively.
Leptin levels were measured by RIA (Human Leptin RIA Kit; Linco Research, Inc., St. Louis, MO). The lowest level of leptin that could be detected by this assay was 0.5 µg/L. The intra- and interassay CV were 8.3% and 6.2%, respectively, at a concentration of 4.9 µg/L. All serum samples from an individual were run in the same assay and in duplicates.
Pulse analysis
We used Cluster, a computerized pulse analysis algorithm to identify statistically significant leptin pulses using a quadratic function variance model. This is model-free discrete peak detection method that does not require assumptions for half-life of a hormone, basal secretion, and secretory burst waveform (31). Cluster parameters were two points for test nadir and two points for test peaks. Pooled t statistics for significant upstrokes and downstrokes was 2.0 in each case. The following pulse attributes were determined: frequency (number of peaks per 8 h), interpeak interval, maximal peak height (highest absolute serum leptin concentration attained within the peak), incremental peak amplitude (algebraic difference between maximal peak height and prepeak nadir), area under the peak (AUC), and interpulse nadir concentration.
Calculations and statistical analyses
AUC for insulin (AUCins) during OGTT was calculated using the trapezoid formula.
To determine the nocturnal rise of leptin in 10 acromegalics, we used nonlinear regression to estimate parameters of a best-fit pattern function to data, performing the nonlinear regression procedure in SPSS, Inc. for Widows, version 7.5. A cubic curve was fitted through all data to obtain individual parameter estimates: acrophase (the time between reference time and time of peak value), mesor (the average value of a cubic curve fitted to the data; the mesor and mean 8-h leptin concentration during the night are equivalent), and the amplitude of best-fit pattern are defined as 50% of the difference between its global maximum and its global minimum and is expressed for each profile (it is given in concentration unites as well as in percentage of mesor).
To detect differences between control subjects and patients with acromegaly, we used the Mann-Whitney test. Wilcoxon Signed Rank test was used to compare parameters before and after surgery in patients with acromegaly. Relationships between variables were sought by Pearson’s correlation coefficient. Stepwise multivariate linear regression analysis was used to identify the independent effects of variables associated with variation in mean 8-h leptin concentrations. The regression coefficients generated by this analysis indicates the slope of the association between the dependent variable and specified independent variable obtained with or without prior adjustment for other independent variables in the model. The SE represents the variability in this association, and the significance is reflected by P value. The model R2 indicates the percentage of variance in the dependent variable that is accounted for by independent variables included in the model. Data are given as mean ± SEM. P < 0.05 was considered statistically significant. Calculations were performed using SPSS, Inc. for Windows, version 7.5.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All patients completed the protocol. The duration of the disease
was estimated to be 7.8 ± 0.8 yr. BMI and basal levels of GH,
IGF-I, and insulin in control subjects and patients with acromegaly are
shown in Table 1
. There were no
significant differences in BMI and basal insulin concentrations among
two groups in either men or women. As expected, GH and IGF-I levels in
both male and female acromegalic patients were significantly higher
than those in normal subjects (P < 0.001 for GH and
IGF-I in both men and women).
|
. In
acromegalic men and women glucose concentration was significantly
higher than in control subjects (6.5 ± 0.7 vs.
4.3 ± 0.2 mmol/L; P = 0.02).
|
Changes in GH, IGF-I, insulin, and mean 8-h leptin concentrations,
as well as in BMI values, after surgery are summarized in Table 1
. Mean
serum GH concentrations in acromegalic men and women significantly
decreased (P = 0.028 and P = 0.001,
respectively). After glucose loading, all patients except, three men,
did not suppress GH levels to less than 1 µg/L. Serum IGF-I levels
decreased significantly after surgery in both men and women
(P = 0.04 and P = 0.002, respectively)
and were in the age- and BMI-matched normal range (P =
0.3 for men and P = 0.7 for women). Altogether GH
(P < 0.001) and IGF-I (P < 0.001)
decreased significantly and only three men did not achieve criteria for
short-term cure of acromegaly. Insulin concentrations decreased after
surgery. In all patients with acromegaly, AUCins
during OGTT was significantly reduced after surgery (13291.5 ±
1801.4 vs. 4466.9 ± 507.8
mUL-1·min; P = 0.002). BMI did
not change after surgery in both male and female patients
(P = 0.3 and P = 0.9), thus
postoperative values were within range of control men and women
(P = 0.8 and P = 0.4,
respectively).
Metabolic characteristics of patients with acromegaly after
trans-sphenoidal surgery are reported in Table 2. Altogether, in
patients with acromegaly, after surgery fasting glucose levels
decreased significantly from 6.5 ± 0.7 to 4.8 ± 0.3 mmol/L
(P = 0.014). Serum total cholesterol and LDL fraction
did not change after surgery. In contrast to LDL, cholesterol HDL
fraction rose significantly in men (P = 0.008), as well
as in women (P = 0.023). Triglyceride concentrations
significantly decreased after surgery in both male and female patients
(P = 0.028 and P = 0.019,
respectively). Finally, FFA levels did not change after surgery in both
men and women and remained in the range of control subjects.
Serum leptin
Mean 8-h leptin concentrations in control subjects and
acromegalics before surgery are presented in Table 1. Acromegalic
patients showed significantly lower mean leptin values in comparison
with age- and BMI-matched control subjects for men and women
(P = 0.001 and P = 0.036,
respectively). There was no significant difference in leptin levels
between pre- and postmenopausal women in either the control or
acromegalic group. Together, in men and women, mean concentrations of
leptin positively correlated with BMI (r = 0.576,
P = 0.007, in control subjects; r = 0.704,
P = 0.001, in acromegalics). Stepwise multivariate
linear regression analysis was used to assess the determinants of
preoperative mean leptin concentrations using age, BMI, and basal GH,
basal insulin, IGF-I, serum lipids, and FFA levels as explanatory
variables (Table 3). Gender, FFA
concentration, insulin level, and age accounted for 99.3% of the
variance in mean leptin concentration in control subjects. In patients
with acromegaly, BMI and gender (Table 3) explained a large proportion
(67.0%) of mean leptin variation.
|
and Fig. 1, mean leptin levels significantly
rose after surgery in both men and women (P = 0.028 and
P = 0.003, respectively). In comparison with control
subjects postoperative leptin values achieved those of age- and
BMI-matched control subjects (P = 0.06 and
P = 0.8 for men and women, respectively). In an overall
model assessing variables that contribute independently to the
variation in postoperative mean leptin concentrations in patients with
acromegaly, a large proportion (65.1%) of the variation was explained
by BMI and gender (Table 3). After adjusting postoperative mean leptin
concentrations for influences of BMI and gender (known as being
confounding factors for leptin levels), corrected values of mean leptin
concentrations were obtained (Ladj). Thus,
another stepwise regression analysis model was tested, with
Ladj as a dependent variable (Table 3). In
this model, FFA contributed to 26.8% of the variance in corrected mean
leptin levels independently of other variables.
|
illustrates the best
fit-patterns of 8-h serum leptin concentration profiles obtained from
10 patients with acromegaly before and after surgery. Average leptin
concentrations during the night (value equivalent to mesor)
significantly rose after treatment in both male (from 2.7 ± 0.03
to 4.5 ± 0.06 µg/L; P < 0.001) and female
patients (15.8 ± 0.1 to 19.0 ± 0.2 µg/L;
P < 0.001). Thus, mesor values were significantly
higher after trans-sphenoidal surgery in all patients (11.9 ± 3.4
vs. 14.6 ± 2.3 µg/L; P = 0.037),
(Fig. 2ab). The absolute amplitude of the cubic fit rose significantly
after surgery, from 1.7 ± 0.4 to 2.6 ± 0.5 µg/L
(P = 0.028), whereas the relative amplitude (expressed
as a percentage of mesor) did not significantly change (15.1 ±
2.1 vs. 19.9 ± 2.0%; P = 0.1). The
time pattern of the 8-h profile was not affected by surgery, as
evidenced by a similar acrophase before and after treatment (43.0
± 30.6 vs. 55.0 ± 45.3 min. after midnight;
P = 0.9) (Fig. 2ab).
|
illustrates serum leptin
concentration profiles obtained at 10-min intervals in acromegalic men
and women during the night before and 2 months after trans-sphenoidal
surgery. By Cluster analysis, as shown in Table 4, leptin pulse frequency and interpeak
interval were similar at baseline and after surgery in all patients.
The maximal serum leptin concentration peak height increased
significantly (P = 0.033), but the incremental peak
amplitude remained at baseline level (Table 4). The serum leptin
concentration peak areas rose after surgery, but this was not
significant (P = 0.06). The rise of average leptin
concentration after surgery, therefore, can be explained by the
significant elevation of interpeak nadir leptin concentration,
P = 0.037 (Table 4).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
GH has long been known to be associated with changes in body composition. GH-deficient children and adults have abnormal body composition, with increased fat mass and decreased fat-free mass together with higher leptin levels, in comparison with BMI-, age-, and gender-matched healthy subjects, whereas acromegalics have increased lean body mass and decreased percentage of body fat, suggesting that leptin is a marker of body composition rather than fat mass alone. Treatment with recombinant human GH of adults with adult-onset of GH deficiency, as well as children with GH deficiency, resulted in marked decrease in fat mass and in leptin levels (13, 14, 15, 16, 32, 35, 36). Recently, it has been shown that ob mRNA is markedly suppressed in hypophysectomized rats and that this effect could not be reversed by GH infusion. When these rats were treated with IGF-I a further suppression of ob mRNA was found (37). One study in humans failed to demonstrate any direct effect of GH on circulating leptin concentration (38). Lower levels of serum leptin in active acromegaly, therefore can be explained, at least in part, with increased lean body mass and decreased percentage of body fat (39).
Elevated fasting glucose levels with increased insulin secretion in
patients with active acromegaly were also demonstrated in this study.
The effects of insulin on leptin secretion are controversial. Although
acute short-term insulin infusions (
3 h) have no stimulatory effect
(40, 41, 42), longer hyperinsulinemic clamps resulted in increased leptin
concentrations in some (18, 43), but not all, studies (17, 44). We have
shown that after correction of hyperinsulinemia in patients with
insulinoma leptin levels decrease (45). There are several papers
showing that insulin action may be direct, and it is possible to see it
in vitro. The controversy arises because there is a
stimulation of leptin secretion but not ob mRNA synthesis
(17, 19). Chronic hyperactivity of the GH-IGF-I axis in patients with
acromegaly induces a state of insulin resistance, both in the liver and
in the periphery. These patients display hyperinsulinemia and increased
glucose turnover in basal and postabsorptive states (46, 47). GH excess
in normal and insulin-deficient men increases FFA and ketone body
production via stimulation of lipolysis (48), and these effects could
have been due to impairment of insulin sensitivity (49). Thus, there is
a possibility that the rate of glucose uptake, which is the rate limit
for triglyceryde storage in adipose cells, rather than insulin and
glucose per se, may represent the signal for the regulation
of leptin gene expression (50). The state of impaired insulin action
and/or secretion in acromegaly, therefore, may contribute in further
lowering of leptin levels. Another interpretation of these data is that
leptin travels with changes in body composition. Because chronic
hyperactivity of GH and IGF-I axis affects both the liver and
peripheral tissues it may be important to measure leptin besides the
liver-derived IGF-I (used as a parameter of cure). This is supported by
a recent study of Sjogren et al. (51), who have shown that
liver-derived IGF-I is not required for postnatal body growth in mice
with the complete inactivation of the IGF-I gene in the
hepatocytes.
Serum leptin concentrations can be influenced by gonadal function, as well. Estrogens have an up-regulatory role in leptin secretion in women as recently reported that they stimulate leptin secretion from human omental adipose tissue (20). However, we could not find a significant difference in leptin levels between pre- and postmenopausal women in either patients with acromegaly or healthy subjects. Sexual dimorphism in leptin secretion, which is seen in normal subjects (20, 52, 53), is preserved in patients with acromegaly.
We have also demonstrated that leptin concentrations in plasma over 8 h during the night in patients with active acromegaly are characterized by a nocturnal rise, with peak levels occurring shortly after midnight. This pattern of leptin secretion was unaffected by trans-sphenoidal surgery. Pooled leptin levels, as well as average leptin concentration during the night (mesor), significantly rose after surgery. The rise of leptin concentration can be explained by increased interpeak nadir leptin concentrations. The circadian rhythm of leptin is preserved in GH-deficient hypopituitary adults and in patients with perinatal stalk-transection syndrome (11, 12). These studies, together with ours, suggest that nyctohemeral fluctuation in plasma leptin is not influenced by GH-IGF-I axis activity. This is consistent with the following studies: Simon et al. (54), who showed that both slight circadian component and sleep modulate plasma leptin levels that interact in normal condition, and Schoeller et al. (55), who demonstrated that the nocturnal rise in serum leptin is directly linked to meal timing and may represent a delayed "postprandial" rise in leptin.
Taken together, although leptin concentrations increase after surgery, GH, IGF-I, insulin and fasting glucose levels decrease. Thus, both the decrease in IGF-I (due to normalization of GH secretion) and improved insulin sensitivity can participate in postoperative elevation in leptin concentrations (17, 37). When interpreting these results it should be mentioned that BMI did not change after surgery in acromegalics and that in both patients with acromegaly after surgery and in normal subjects multiple regression analysis showed that serum leptin levels were positively associated with FFA. There is a possibility that plasma FFA in the setting of normal GH and insulin secretion may have a role in regulation of leptin secretion. This is consistent with a recent report that in rat skeletal muscle and adipose tissue synthesis of ob mRNA can be induced by metabolic substrates (including FFA) for hexosamine biosynthetic pathway (56).
A collateral finding of this study was significant change in lipid profiles. Cholesterol HDL fraction rose while triglyceride levels decreased after GH normalization. This may have beneficial effect on increased cardiovascular risk in acromegalics (57).
In conclusion, we have found the preservation of gender differences in leptin levels in acromegaly. The hyperactive GH-IGF-I axis is associated with low serum leptin concentrations, but the nocturnal rise and pulsatility are preserved. These findings indicate that GH and IGF-I could influence leptin levels but are not involved in the genesis of leptin pulsatility or the circadian rhythm. The rise in serum leptin levels during short-term remission of acromegaly is not due to changes in BMI. Because BMI might not be a surrogate for body composition, additional studies are necessary to define the clinical significance of postoperative leptin rise in patients with acromegaly.
Received July 22, 1999.
Revised September 27, 1999.
Accepted October 1, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S Pekic, M Doknic, D Miljic, M Joksimovic, J Glodic, M Djurovic, C Dieguez, F Casanueva, and V Popovic Ghrelin test for the assessment of GH status in successfully treated patients with acromegaly. Eur. J. Endocrinol., May 1, 2006; 154(5): 659 - 666. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Blain, A. Vuillemin, F. Guillemin, R. Durant, B. Hanesse, N. de Talance, B. Doucet, and C. Jeandel Serum Leptin Level Is a Predictor of Bone Mineral Density in Postmenopausal Women J. Clin. Endocrinol. Metab., March 1, 2002; 87(3): 1030 - 1035. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Marzullo, C. Buckway, K. L. Pratt, A. Colao, J. Guevara-Aguirre, and R. G. Rosenfeld Leptin Concentrations in GH Deficiency: The Effect of GH Insensitivity J. Clin. Endocrinol. Metab., February 1, 2002; 87(2): 540 - 545. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. S. Randeva, R. D. Murray, K. C. Lewandowski, C. J. O'Callaghan, R. Horn, P. O'Hare, G. Brabant, E. W. Hillhouse, and S. M. Shalet Differential Effects of GH Replacement on the Components of the Leptin System in GH-Deficient Individuals J. Clin. Endocrinol. Metab., February 1, 2002; 87(2): 798 - 804. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
