| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Developmental Endocrinology Branch, NICHD, NIH, Bethesda, Maryland 20892
Address all correspondence and requests for reprints to: Jeffrey Baron, M.D., National Institutes of Health, Building 10, Room 10N262, 10 Center Drive MSC 1862, Bethesda, Maryland 20892-1862. E-mail: Baronj{at}cc1.nichd.nih.gov
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We conclude that, in the rabbit, fasting induces a rapid depletion of growth plate chondrocytes and inhibition of longitudinal bone growth. These effects appear to be mediated by systemic endocrine mechanisms; circulating IGF-1 levels are diminished because of hepatic resistance to GH. In contrast, the local, paracrine IGF-1 system in growth plate does not appear to contribute to the growth inhibition but instead appears to be up-regulated by fasting.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Inadequate caloric intake inhibits longitudinal bone growth (5, 6, 7). This suppression of growth plate function could arise from down-regulation of either systemic IGF-1 acting in an endocrine fashion or of local IGF-1 acting in a paracrine fashion. There is evidence that undernutrition does affect the systemic GH-IGF-1 system. In rats, fasting diminishes both GH production (8) and hepatic GH sensitivity (9, 10). The diminished GH sensitivity is characterized by decreases in hepatic GH receptor messenger RNA (mRNA) levels, GH binding, IGF-1 mRNA levels, and circulating IGF-1 levels (9, 10, 11, 12, 13). The nutritional regulation of rat hepatic IGF-1 production may be mediated by insulin (14) and by bioavaibility of amino acids (15). In humans, undernutrition may also induce a state of hepatic GH insensitivity. Circulating IGF-1 levels decrease as do levels of GH binding protein that represents a circulating fragment of the GH receptor (16). Unlike the rat, fasting humans have increased circulating GH levels (17, 18).
In contrast, there is little information about the effects of fasting on the local, growth plate IGF-1 system. Thus it is not known whether this system contributes to the inhibition of longitudinal growth during undernutrition. This study was therefore designed to investigate the mechanisms by which undernutrition suppresses longitudinal bone growth, focusing on the roles of systemic and local IGF-1. We hypothesized that growth plate, like liver, is rendered GH insensitive by fasting and that this decreased GH sensitivity contributes to the decreased linear growth associated with caloric deprivation. To test this hypothesis, we subjected growing rabbits to 48 h of fasting and measured the resulting GH receptor mRNA and IGF-1 mRNA levels in growth plate and, for comparison, in liver, kidney, and muscle. Serum GH, GH binding protein (GHBP), and IGF-1 levels were also measured. To confirm that the fast inhibited longitudinal bone growth, we measured tibial growth velocity and quantitated the histological changes at the growth plate.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Growth rate and histological effects were examined at 24 h and 48 h of fasting (n = 8 per group). The growth rate determined at 24 h does not represent an instantaneous growth rate but rather an average growth rate over the first day. To investigate the mechanisms affecting this growth rate, we therefore picked a time point (15 h) within this first 24-h time period. Similarly, to investigate mechanisms affecting the average growth during the second 24 h, we picked a time point (30 h) within that time period. Thus, an additional group of animals was studied at each of three time points: 0, 15, and 30 h of fasting (n = 6 per group). In these animals, serum GH, IGF-1, and GHBP were measured. Liver, kidney, and muscle samples were frozen in liquid nitrogen and stored at -80 C until RNA preparation. The distal femurs and proximal tibiae were isolated, cut longitudinally in half, fractured through the proximal tibial and distal femoral growth plates, and submerged for 4 min in a solution containing 4 M guanidine thiocyanate, 25 mM sodium citrate (pH 7.0), and 0.1 M 2-mercaptoethanol. The growth plate cartilage was then scraped from the underlying bone, frozen in liquid nitrogen, and stored at -80 C (19).
Growth rate determination
Before fasting, two metal pins (30 gauge; 5 mm in length) were
placed in each proximal tibial metaphysis. For pin placement, the
animals were sedated with im glycopyrolate (0.02 mg), ketamine
hydrochloride (33 mg/kg), and xylazine (6.6 mg/kg) and given a single
prophylactic im dose of penicillin-G (44,000 IU/kg). Using sterile
technique, the pins were inserted percutaneously into the bony
metaphysis. The pins were directed posteriorly, parallel to and several
mm distal to the proximal tibial growth plate. Serial radiographs of
the tibiae were obtained in duplicate after 0, 24, and 48 h of
full feeding or fasting, as previously described (20). For each
radiograph, animals were sedated with an intramuscular injection of
ketamine hydrochloride (33 mg/kg) plus xylazine (6.6 mg/kg). The
radiographs were examined under a dissecting microscope. Using a
micrometer, we measured the distance between the more proximal pin and
the proximal margin of the proximal tibial epiphysis (20). The observer
was blinded to the dietary regimen.
Histological examination
After 48 h of full feeding or fasting, the animals were
killed, and the proximal tibiae were prepared for histological
examination. Serial 6-µm sections were cut parallel to the tibia and
stained with Masson trichrome stain, as previously described (21). The
stained tissue sections were examined by a single observer blinded to
the dietary regimen.
GH, GHBP, and IGF-1 assays
Serum was obtained just before sacrifice for measurement of GH
and GHBP. Rabbit serum GH hormone was measured using an RIA (Dr.
A. F. Parlow, Pituitary Hormone and Antisera Center, Torrance,
CA).
Total GHBP was measured in triplicate by an assay based on radioimmunoprecipitation of the GH/GHBP complex with an anti-GHBP antibody (Endocrine Sciences, Calabasas Hills, CA) as previously described (19, 22). Briefly, 25 µl of serum were incubated overnight at 4 C with excess [125I]hGH, and a monoclonal antibody (mAb) directed against the extracellular domain of the GH receptor (mAb 263) (23). The trimolecular complex (mAb 263-GHBP-[125I]hGH) was separated from the free [125I]hGH by precipitation with goat-antimouse immunoglobulin and polyethylene glycol. Normal rabbit serum has a GHBP concentration of 1340 pmol/liter with an association constant of 1.2 nM-1 (19). Ten nanograms per ml rabbit GH lowered the apparent GHBP concentration by 13%; 20 ng/ml rabbit GH lowered the apparent GHBP concentration by 26%. In each sample, GHBP measurements were corrected for endogenous rabbit GH.
Rabbit serum IGF-1 was measured after acid-ethanol extraction by RIA using a polyclonal antibody and IGF-II to block any residual IGFBPs (24).
RNA preparation
For liver, kidney and muscle samples, total RNA was prepared by
homogenization in a guanidinium isothiocyanate solution (RNAzol,
Cinna/Biotecx, Friendswood, TX), followed by phenol/chloroform
extraction and 2-propanol precipitation. The growth plate tissue
(50–100 mg) was homogenized in 0.5 ml of a solution containing 4
M guanidine thiocyanate, 25 mM sodium citrate
(pH 7.0), and 0.1 M 2-mercaptoethanol. Samples were diluted
with 1.5 ml of a solution containing 19 mM Tris-HCl (pH
7.0) and 0.72 g/liter proteinase K (ICN Biomedicals Inc., Aurora, OH).
After a 20-min digestion at room temperature, the samples were
extracted with 1 ml phenol and 1 ml chloroform. After collection of the
aqueous phase, 0.2 ml 3 M sodium acetate (pH 5.2) and 2 ml
2-propanol were added. RNA was precipitated at -20 C for 3 h and
collected by centrifugation (9, 500 x g for 30 min).
The pellet was washed with 75% ethanol, dried, and resuspended in 200
µl water. An equal volume of 8 M LiCl was added. The RNA
was reprecipitated at -20 C overnight and collected by centrifugation
(10,700 x g for 30 min). After washing with 75%
ethanol, the pellet was suspended in 30 µl 10 mM Tris (pH
7.6) and 1 mM EDTA. RNA concentration was approximated
initially by measuring the absorbance at 260 nm. The integrity of RNA
preparations was confirmed by visual inspection of ethidium
bromide-stained 28S and 18S ribosomal RNA bands after electrophoresis
of 15-µg total RNA through 1.25% agarose–2.2 M
formaldehyde gels. Ribosomal RNA content was quantitated from a
photograph of the ethidium bromide-stained gel by computer-assisted
video densitometry using the program Image 1.49 (Dr. W. Rasband, NIH,
Bethesda, MD; Ref.25). The intensity of the ribosomal band was
proportional to the RNA content throughout the range used in these
studies (data not shown).
Antisense RNA probes
The GH receptor RNA probe was transcribed from a 370-bp
EcoRI/AccI fragment of the rabbit GH receptor
complementary DNA (4; kindly provided by Dr. W. Wood, Genentech, South
San Francisco, CA) subcloned into EcoRI- and
AccI-digested pGEM-3Z (Promega, Madison, WI). The sequence
encodes part of the extracellular domain, the transmembrane domain, and
part of the intracellular domain of the receptor. This plasmid was
linearized with HindIII, and, using T7 RNA polymerase, a
32P-labeled 404-base antisense riboprobe was transcribed,
containing 370 bases complementary to the rabbit GH receptor mRNA and
34 bases corresponding to the vector.
To prepare a rabbit IGF-1 probe, exon 3 of the rabbit IGF-1 gene was amplified by PCR using as template rabbit genomic DNA, prepared from the liver of a New Zealand White rabbit. The oligonucleotide primers represented the first 21 bases and the last 21 bases of exon 3 of the human IGF-1 gene (5' primer ACAAGCCCACAGGGTATGGCT and 3' primer GACATGCCCAAGACCCAGAAG) (26). A single major product was purified, inserted into a plasmid pCRII (Invitrogen Corp, San Diego), and sequenced. The predicted amino acid sequence showed 96% identity with the human and rat IGF-1 exon 3 (26, 27). This fragment was subcloned in pGEM-4Z (Promega, Madison, WI). The plasmid was linearized with XbaI, and, using T7 RNA polymerase, a 32P-labeled approximately 320-bases antisense riboprobe was transcribed, containing approximately 175 bases complementary to the rabbit IGF-1 mRNA and approximately 145 bases corresponding to the vector.
RNase protection assay
Ten- to fifteen-microgram aliquots of total RNA were combined
with 320,000 cpm of GH receptor and IGF-1 probes and hybridized at 45 C
for 16 h as previously described (19). As an internal control for
potential losses during processing, an equal amount (
3000 cpm) of
32P end-labeled, AviII-digested pGEM-3Z DNA
(Promega, Madison, WI) (1.7 and 1.0 kb in size) was added to each RNA
sample at the beginning of the assay. After RNase digestion, the
protected probe was denatured and electrophoresed through 8%
polyacrylamide-8 M urea gels. The bands corresponding to
protected probes and to the pGEM-3Z fragments were quantitated by
storage phosphor autoradiography using a PhosphorImager apparatus
and ImageQuant software (Molecular Dynamics, Sunnyvale, CA). RNA losses
were not significantly different in the fasted vs. control
groups (data not shown). Each assay was performed at least twice. The
mean of the replicates was used for data analysis.
Statistical analysis
For each sample, the activity of the RNase-protected probe was
normalized to the ribosomal RNA content and to the intensity of the
corresponding pGEM-3Z bands. The mean GH receptor and IGF-1 mRNA levels
of the control group in each experiment were set equal to 100%. Data
are presented as mean ± SEM. Comparisons between fed
and fasted rabbits were made, where appropriate, by ANOVA followed by
posthoc Fisher protected least significant difference tests
or by unpaired two-tailed Student’s t test with the
Bonferroni correction. Data were log transformed before analysis, where
appropriate.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
).
|
). The
number of proliferative chondrocytes per column was significantly
decreased by fasting (38 ± 2 vs. 45 ± 2;
P = 0.05) as was the number of hypertrophic
chondrocytes per column (18 ± 1 vs. 23 ± 1;
P < 0.001). Fasting also caused a decrease in the
height of the terminal hypertrophic chondrocyte, i.e. the
chondrocyte adjacent to the metaphyseal border (23 ± 1
vs. 30 ± 1 µm; P < 0.001).
|
). The serum GHBP showed no significant
differences among the three groups (1980 ± 250 vs.
2300 ± 270 vs. 2590 ± 400 pmol/liter in the fed,
15-h fast, 30-h fast groups, respectively; Fig. 3B). Serum IGF-1
decreased significantly after 15 and 30-h fasting compared with
controls (353 ± 24 vs. 139 ± 30 vs.
105 ± 25 ng/ml in the fed, 15-h fast, 30-h fast groups,
respectively; P < 0.001, Fig. 3C).
|
and 5).
|
|
).
In kidney, fasting did not significantly affect GH receptor mRNA levels
(100 ± 4% vs. 97 ± 4 vs. 95 ±
6 in the fed, 15-h fast, 30-h fast groups, respectively; Fig. 5).
In muscle tissue, fasting significantly increased muscle GH receptor
mRNA levels compared with controls (100 ± 15% vs.
221 ± 51 vs. 295 ± 41 in the fed, 15-h fast,
30-h fast groups, respectively; P < 0.05; Fig. 5).
IGF-1 mRNA
In growth plate, IGF-1 mRNA levels were increased significantly in
the group fasted for 15 h compared with controls
(P < 0.02). The increase at 30 h did not reach
statistical significance (100 ± 14% vs. 165 ±
15 vs. 128 ± 21 in the fed, 15-h fast, 30-h fast
groups, respectively; Figs. 4 and 6).
|
).
In the kidney, IGF-1 mRNA levels were very low and did not change after
15 h but decreased after 30 h fasting compared with the
controls (100 ± 5% vs. 98 ± 7 vs.
76 ± 4 in the fed, 15-h fast, 30-h fast groups, respectively;
P < 0.02; Fig. 6).
In the muscle, IGF-1 mRNA levels were very low and did not change after
fasting compared with the controls (100 ± 11% vs.
130 ± 9 vs. 96 ± 9 in the fed, 15-h fast, 30-h
fast groups, respectively; Fig. 6).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Chronic undernutrition has previously been shown to inhibit longitudinal bone growth (5, 6, 7). Here we show that inhibition occurs after only 48 h of fasting. Growth rate is proportional to the number of proliferative chondrocyte divisions per unit time and to the height of the terminal hypertrophic chondrocyte (28). Thus, the observed decreases in the number of proliferative chondrocytes and in the height of the terminal hypertrophic chondrocytes may both have contributed to the decreased growth rate.
To explore the underlying mechanism, we measured serum GH, GHBP, and IGF-1. During the fast, serum GH levels increased and thus could not explain the growth inhibition. Serum GH levels also increase with undernutrition in humans, sheep, cows, and pigs (17, 29, 30, 31). Thus, in this respect, the rabbit provides a better model for human GH physiology than does the rat, which shows decreased serum GH levels with fasting (8). In this study, a single GH measurement was made in each animal. Because GH secretion is pulsatile in the rabbit (32), this value probably reflects a mean serum concentration but does not distinguish between changes in pulse amplitude or frequency. Although frequent GH sampling has not been reported for the rabbit during fasting, humans show increases in both pulse amplitude and frequency. The observed rabbit GH levels may also have been increased by the stress of blood drawing in both the fed and fasted groups.
In the current study, serum GHBP was not changed significantly by fasting. In humans, serum GHBP concentrations decrease with chronic dietary restriction (16), but not with acute fasting (33). A similar result has been observed in pigs, whereas rat GHBP falls with fasting (11, 34). However, rat GHBP derives from alternative splicing of the GH receptor gene, whereas in rabbits and humans, the GHBP is derived from the GH receptor by proteolytic cleavage (35). Because GHBP is the principal binding protein for GH, our data suggest that unbound GH is also increased during fasting in the rabbit.
Serum IGF-1 levels were significantly decreased during fasting. This decrease provides one mechanistic explanation for the inhibition in longitudinal bone growth and the histologic changes in the growth plate. The circulating IGF-1 levels were significantly decreased by 15 h of fasting, and yet the average growth velocity during the first day of fasting was only affected mildly, suggesting a lag time between the endocrine change and the cellular response. Since the liver is believed to be the principal source of circulating IGF-1 (36), the observed decrease in hepatic IGF-1 mRNA levels may be responsible for the decrease in circulating IGF-1 levels. The decreases in liver IGF-1 mRNA levels and in circulating IGF-1 levels occurred despite increased GH levels, suggesting that the liver is rendered GH resistant. Similar hepatic GH resistance has been observed in rats (7, 12, 13) and humans during fasting (17, 18).
In rats, this fasting-induced GH resistance is associated with decreased hepatic GH receptor mRNA expression (13). In contrast, we did not observe a significant change in hepatic GH receptor mRNA expression in the fasted rabbit. Similarly, GH receptor mRNA levels in bovine liver do not change with chronic caloric restriction (37). Thus, the mechanism of fasting-induced hepatic GH resistance may differ among mammalian species. While we did not measure GH receptor protein levels directly, the absence of change in the mRNA level and in the GHBP level (a proteolytic product of the GH receptor) indirectly suggest that the receptor density is not regulated by fasting of this duration in the rabbit.
In contrast to the systemic IGF-1 system, the local, growth plate IGF-1 system did not appear to be down-regulated. Neither GH receptor mRNA nor IGF-1 mRNA levels decreased with fasting in the growth plate, suggesting that the changes in the local growth plate GH receptor-IGF-1 system do not contribute to the inhibition in bone growth. To the contrary, both GH receptor mRNA and IGF-1 mRNA increased with fasting. We did not measure GH receptor expression at the protein level. However, GH receptor mRNA and protein levels correlate in other systems (38, 39). Thus, GH receptor protein expression may also be increased by fasting. A similar increase in GH receptor mRNA occurs with glucocorticoid excess, suggesting that these changes may represent a compensatory mechanism occurring in response to growth inhibition (19). Alternatively, fasting may have increased circulating glucocorticoid concentrations, thus inducing increased GHR and IGF-1 mRNA levels. Our data do not exclude the possibility that translation of IGF-1 transcript could be decreased or that local IGF-1 protein degradation could be increased during fasting. Alternatively, resistance to IGF-1 could develop through altered expression of IGF receptor or IGF binding proteins, thus contributing to the decreased growth rate.
The fasting-induced changes in GH receptor mRNA and IGF-1 mRNA levels in the kidney paralleled those observed in liver; GH receptor was unchanged, whereas IGF-1 mRNA decreased with the more prolonged fast, a result similar to that observed in the rat (13).
GH receptor mRNA levels in muscle increased with fasting with a magnitude and time course similar to that observed in growth plate. However, unlike growth plate, muscle IGF-1 mRNA did not significantly change during fasting. This pattern contrasts with that observed in the rat, in which both muscle GH receptor and IGF-1 mRNA decreased with fasting (12, 13). These data again emphasize that the nutritional regulation of this system is highly species specific.
We conclude that, in the rabbit, fasting induces a rapid depletion of growth plate chondrocytes and inhibition of longitudinal bone growth. Our data suggest that these effects are mediated by systemic endocrine mechanisms; circulating IGF-1 levels are diminished because of hepatic resistance to GH. In contrast, the local, paracrine IGF-1 system in growth plate does not appear to contribute to the growth inhibition but instead appears to be up-regulated by fasting.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Commissioned Officers in the United States Public Health
Service.
Received April 21, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. C. Lasseter, L. Shaughnessy, D. Cummings, J. C. Pezzullo, W. Wargin, R. Gagnon, J. Oliva, and G. Kosutic Ghrelin Agonist (TZP-101): Safety, Pharmacokinetics and Pharmacodynamic Evaluation in Healthy Volunteers: A Phase I, First-in-Human Study J. Clin. Pharmacol., February 1, 2008; 48(2): 193 - 202. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Baron Growth Hormone Therapy in Childhood: Titration Versus Weight-Based Dosing? J. Clin. Endocrinol. Metab., July 1, 2007; 92(7): 2436 - 2438. [Full Text] [PDF]
|
|
|
|
 |
|
 T. C. R. Prickett, G. K. Barrell, M. Wellby, T. G. Yandle, A. M. Richards, and E. A. Espiner Response of plasma CNP forms to acute anabolic and catabolic interventions in growing lambs Am J Physiol Endocrinol Metab, May 1, 2007; 292(5): E1395 - E1400. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. M. Luque, Z. H. Huang, B. Shah, T. Mazzone, and R. D. Kineman Effects of leptin replacement on hypothalamic-pituitary growth hormone axis function and circulating ghrelin levels in ob/ob mice Am J Physiol Endocrinol Metab, March 1, 2007; 292(3): E891 - E899. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. Luque, S. Park, and R. D. Kineman Severity of the Catabolic Condition Differentially Modulates Hypothalamic Expression of Growth Hormone-Releasing Hormone in the Fasted Mouse: Potential Role of Neuropeptide Y and Corticotropin-Releasing Hormone Endocrinology, January 1, 2007; 148(1): 300 - 309. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Gat-Yablonski, T. Ben-Ari, B. Shtaif, O. Potievsky, O. Moran, R. Eshet, G. Maor, Y. Segev, and M. Phillip Leptin Reverses the Inhibitory Effect of Caloric Restriction on Longitudinal Growth Endocrinology, January 1, 2004; 145(1): 343 - 350. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Nilsson, J. Falk, E. M. Ritzen, J. Baron, and L. Savendahl Raloxifene Acts as an Estrogen Agonist on the Rabbit Growth Plate Endocrinology, April 1, 2003; 144(4): 1481 - 1485. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. M. Naranjo, S. Yakar, M. Sanchez-Gomez, A. U. Perez, J. Setser, and D. LERoith Protein Calorie Restriction Affects Nonhepatic IGF-I Production and the Lymphoid System: Studies Using the Liver-Specific IGF-I Gene-Deleted Mouse Model Endocrinology, June 1, 2002; 143(6): 2233 - 2241. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Nagaya, M. Kojima, M. Uematsu, M. Yamagishi, H. Hosoda, H. Oya, Y. Hayashi, and K. Kangawa Hemodynamic and hormonal effects of human ghrelin in healthy volunteers Am J Physiol Regulatory Integrative Comp Physiol, May 1, 2001; 280(5): R1483 - R1487. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.B. Ballinger, C. Camacho-Hubner, and N.M. Croft Growth failure and intestinal inflammation QJM, March 1, 2001; 94(3): 121 - 125. [Full Text] [PDF]
|
|
|
|
 |
|
 A B Ballinger, O Azooz, T El-Haj, S Poole, and M J G Farthing Growth failure occurs through a decrease in insulin-like growth factor 1 which is independent of undernutrition in a rat model of colitis Gut, May 1, 2000; 46(5): 695 - 700. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. Bollag, Q. Zhong, P. Phillips, L. Min, L. Zhong, R. Cameron, A. L. Mulloy, H. Rasmussen, F. Qin, K. H. Ding, et al. Osteoblast-Derived Cells Express Functional Glucose-Dependent Insulinotropic Peptide Receptors Endocrinology, March 1, 2000; 141(3): 1228 - 1235. [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 |
