| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Departments of Endocrinology and Process Sciences, Genentech, Inc., South San Francisco, California 94080
Address all correspondence and requests for reprints to: Dr. R. G. Clark, Mail Stop 37, Department of Endocrinology, Genentech, Inc., 390 Point San Bruno Boulevard, South San Francisco, California 94080. E-mail: rossc{at}gene.com
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In a series of studies, normal and streptozotocin (STZ) diabetic male Wistar rats (180–210 g), fasted for 6 or 16 h, were injected with rhIGFBP-1 (iv, 80–500 µg/rat). rhIGFBP-1 did not affect blood glucose acutely but did stimulate insulin release in normal rats (5 min post injection; PBS, 103.5 ± 8.5; rhIGFBP-1 (500 µg), 166.8 ± 15.7; rhIGFBP-1 (100 µg); 151.4 ± 14.1% initial). rhIGFBP-1 pretreatment, in normal and diabetic rats, reduced the hypoglycemic response to rhIGF-I (diabetic rats after 20 min: PBS, 103.4 ± 11.4; BP-1 (500 µg) + rhIGF-I (50 µg), 97.6 ± 3.6; rhIGF-I, 48.2 ± 4.3% initial) but did not affect the hypoglycemic response to des(1–3)IGF-I or insulin (0.5 U/kg).
These studies show that rhIGFBP-1 causes insulin release, has a minimal effect on blood glucose, and inhibits the hypoglycemic effect of rhIGF-I. These data suggest that endogenous IGF-I tonically suppresses insulin secretion and imply that aberrant IGFBP levels or reduced IGF-I bioactivity may lead to chronic hyperinsulinemia.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The function of IGFBP-1 in vivo is poorly understood. The administration of purified human IGFBP-1 to rats has been shown to cause an acute, but small, increase in blood glucose (8). Such effects of IGFBP-1 are presumed to be due to a modulation of endogenous IGF activity. The regulation of IGFBP-1 is somewhat better understood. It has been proposed (9) that, when blood glucose rises and insulin is secreted, IGFBP-1 is suppressed, allowing a slow increase in free IGF-I levels that might assist insulin action on glucose transport. Such a scenario places the function of IGFBP-1 as a direct regulator of blood glucose (8, 9). Such data and hypotheses support the idea that IGF-I and IGFBP-1 are directly involved in the acute regulation of glucose metabolism.
Earlier studies (8) with purified natural IGFBP-1 were limited by the supply of protein. The availability of recombinant human IGFBP-1 (rhIGFBP-1) has allowed us to perform a series of studies investigating its effects in vivo in the rat. These new studies support a role for IGFBP-1 in glucose regulation and suggest that a primary role of IGFBP-1, and of endogenous IGF-I, is as a regulator of insulin secretion.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Animals
Male Wistar rats 6–8 weeks of age, 180–210 g, or female dw/dw
rats (10) of 100–140 g (Charles River Laboratories, Hollister, CA and
Wilmington, MA, respectively) were group housed under conditions of
controlled light and temperature and given water and a standard rodent
chow diet (Purina rodent chow no. 5001) ad libitum. Diabetes
was induced in some of these animals by streptozotocin (65 mg/kg, ip,
Sigma Chemical Co., St. Louis, MO). After 4–5 days, one drop of blood
was taken from a tail vein and blood glucose measured by a Lifescan One
Touch II blood glucose monitor (Lifescan, Inc., Milpitas, CA). Only
rats with blood glucose levels greater than 200 mg/dl were considered
to be diabetic and thus chosen for further study.
Surgery
In a series of experiments, normal or diabetic rats were
anesthetized by a single ip injection of ketamine (62.5 mg/kg, Ketaset,
Fort Dodge Laboratories, Inc., Fort Dodge, IA) and xylazine (12.5
mg/kg, Rompun, Miles Inc., Shawnee Mission, KS); then, indwelling
catheters were placed in the right jugular vein and exteriorized
dorsally to the shoulder. After surgery, animals were housed
individually and the jugular catheters were tested for patency and
flushed daily with sterile heparinized saline (100 U/ml).
Experiments
Rats were allowed to recover from the surgery for from 2–5 days
before being used in the following experiments. On the day of
experimentation, after a 6- or 16-h fast, a 30-cm polyethylene
extension tube was connected to the exteriorized catheter to allow the
remote withdrawal of blood or sample injection and so that experiments
could be performed with minimal stress to the animals. In all
experiments, at least 2 basal blood samples (150 µl-200 µl) were
taken from the rats before test substances or saline were administered.
Further blood samples were then taken at frequent intervals and plasma
separated and frozen for subsequent measurements of glucose or insulin.
Fluid volume was replaced by adding 150 µl-200 µl of 5U/ml
heparinized saline after each blood sample.
After the basal blood samples, recombinant human IGFBP-1 (80–500 µg) or its excipient (PBS) were given iv via the jugular catheter. In some experiments, rhIGF-I (50 µg iv), des(1, 2, 3)IGF-I (50 µg iv) or insulin (0.5 U/kg iv) were then given 10–40 min later. In some studies, glucose tolerance tests were also performed by injecting dextrose (2.0 g/kg, IP, Abbot Labs, North Chicago, IL).
Materials
Recombinant human IGF-I was clinical grade vialed material (10
mg/ml in citrate buffer, pH 5.4), human des(1, 2, 3)IGF-I was produced at
Genentech (4.8 mg/ml in 100 mM acetate buffer, pH 4.5),
whereas human insulin was purchased from Eli Lilly & Co. (Indianapolis,
IN, regular Iletin, 100 U/ml).
rhIGFBP-1.
Human IGFBP-1 cDNA was cloned by Xiaojian Huang
(Genentech) using standard techniques and expressed in mammalian CHO
cells. The CHO cell line DP12 (CHO K1 dhfr-, EP 307, 247) was
transfected with the plasmids pSVI6B7-hBP1 (coding for IGFBP-1) and
pSVI7D-DHFR (coding for the selectable marker) using the calcium
phosphate coprecipitation technique. Stable clones were screened for
expression of IGFBP-1 using rabbit anti-BP1 polyclonal serum. A high
producing clone was suspension adapted and scaled up to a 100-liter
production fermenter. The production media were DMEM/Ham’s F-12-based
formulation supplemented with glucose, amino acids, glycine,
hypoxanthine, thymidine, recombinant human insulin, hydrolyzed peptone,
pluronic F68, gentamycin, lipids, and trace elements. Cells were
separated from the supernatant using tangential flow filtration
generating the harvested cell culture fluid.
The initial purification step was the capture of the IGFBP-1 using IGF-I bound to an affinity resin (Affi-Prep 10 Support by Bio-Rad, Richmond, CA). IGF-Affigel bound approximately 1 mg IGF-I per ml of resin. This gave a capacity of only 1 mg IGFBP-1/ml resin at 4 CV/h. However, using Sterogene’s Actigel Superflow (Sterigene, Arcadia, CA), with IGF-I as ligand, an IGF-Sterogene resin (3.2 x 13 cm, 104 ml column) could be made that bound 7 mg IGF-I per ml resin. A flowrate of 10 CV/h was used with 4 mg IGFBP-1 per ml of resin. Cycles were performed on this column at 2–8 C. Fractions were collected after eluting with 0.5 M HAc, 15% isopropanol, pH 3.0 and analyzed by Polymer Labs reverse phase assay (PLRP) for BP-1. To remove the isopropanol, S-Sepharose was used as a buffer exchange medium. Hydrophobic interaction chromatography was performed using a Phenyl Toyopearl resin (TosoHaas, Montgomeryville, PA) to remove putative BP-1 fragments. The final steps were to concentrate and then to dialyze the Phenyl pool, which contained (NH4)2SO4, into PBS. To concentrate, a 5-kDa regenerated cellulose membrane was used on a Genenstak. A total of 0.3 square feet of membrane was flushed with 500 ml 20 mM phosphate, pH 7, at a feed rate of 500 ml/min. The phenyl pool was concentrated 10 times. Because the volume of retentate was small, this was dialyzed vs. PBS using a 6–8 kDa Spectrapor dialysis bag 100x for 4 h thrice to remove the salt. Lastly, the solution was sterile filtered using a 0.22 um Sterivex GV and aliquoted into vials for storage at -70 C. At each step, the purification was monitored using reverse phase HPLC (Polymer Labs) with a gradient of 20–50% acetonitrile over 9 min. In addition, a nonreduced SDS-PAGE gel was run using 1 µg/lane and then stained with a modified Daiichi silver stain.
This material was also tested for its ability to bind rhIGF-I in a competitive binding assay. This assay involved coating the rhIGFBP-1 (160 ng/ml) onto 96-well plates overnight, blocking the plates for 1 h with 0.5% BSA, adding rhIGF-I (250–0.08 ng/ml) for 1 h, then adding 20K cpm of 125I-IGF-I and incubating for 2 h before washing and counting.
Measurements
Plasma glucose concentrations were measured using a Chem 1A
serum chemistry analyzer (Technicon Instruments Corp., Tarrytown, NY).
Insulin was measured by rat specific RIA (Linco Research Inc., St.
Charles, MO).
Statistical comparisons were by one-way ANOVA followed by Duncan’s multiple range test. Means and SE are shown.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
with chromatograms at these steps
shown in Fig. 1 and gels of the material shown in Fig. 2. The advantage of affinity purification is shown in
Table 1 by: 1) the volume being reduced from 87 liters to 1 liter; 2)
the greater than 50% recovery of rhIGFBP-1; and 3) the chromatograms
shown in Fig. 1. The phenyl toyopearl step appeared to successfully
remove IGFBP-1 fragments (Fig. 1) but was associated with a loss of
50% of the material (Table 1). Overall, 142 mg of rhIGFBP-1 was
recovered from the 548 mg of rhIGFBP-1 in the culture fluid, an overall
recovery of about 25%. Figure 2 shows an SDS-PAGE gel of the material
after the phenyl toyopearl and diafilter steps. After diafiltering the
rhIGFBP-1 is present as essentially one peak by reverse phase HPLC
(Fig. 1) and as one band by SDS page (Fig. 2). Amino-acid analysis
showed a close correlation between theoretical and actual amino acid
composition and gave a protein concentration of 1.753 mg/ml compared
with 2.0 mg/ml by reverse phase assay. Finally, amino acid sequencing
indicated that the final bulk was 98% IGFBP-1. This purifed rhIGFBP-1
was the material used in all the experiments described below with the
amount of protein administered to the rats determined by the protein
concentration.
|
|
|
. This plot shows that radiolabeled IGF-I bound to the
rhIGFBP-1 and could be displaced by unlabeled rhIGF-I.
|
rhIGFBP-1: effect on blood glucose
We next chose to repeat the studies reported by Lewitt
et al. (8), who used normal male rats and found a transient
hyperglycemia after the administration of rhIGFBP-1. Figure 4, A and B, shows, in each panel, the combined data from
two separate experiments (n = 13–19/group) where animals were
fasted for either 16 h (Fig. 4A) or 6 h (Fig. 4B) and given
an injection of excipient or rhIGFBP-1 at a dose of 500 µg (Fig. 4B)
or 250 µg (Fig. 4A) at time zero. In both groups of rats, there was a
small progressive decline in blood glucose of 5–10% that was most
likely due to a dilution effect of the saline (200 µl per sample), we
administered after each blood sample having a cumulative effect to
dilute the blood. Another contributing factor could be that there was
an initial transitory stress response due to the handling or blood
sampling from the rats. Figure 4A is a direct repeat of the data
reported by Lewitt et al. (8), who fasted rats for 16 h
before administering rhIGFBP-1. In our studies, there was no sign of
the hyperglycemia in rhIGFBP-1-treated rats, so we ran the studies
shown in Fig. 4B, to discover if, with a shorter fast, and therefore a
higher basal insulin level, rhIGFBP-1 would have an effect on blood
glucose. As shown in Fig. 4B, the glucose values in rats treated with
excipient or rhIGFBP-1 were almost identical.
|
rhIGFBP-1: effect on the hypoglycemia induced by rhIGF-I
Figure 5 shows three experiments studying the
effect of pretreatment with rhIGFBP-1 on the hypoglycemia induced by
rhIGF-I in normal male rats (Fig. 5, A and B) and in diabetic rats
(Fig. 5C). In Fig. 5A, normal male rats (n = 3–4 rats/group) were
fasted for 6 h, then given rhIGFBP-1 (500 µg) or excipient at
time zero. After 10 min, all animals received 50 µg of rhIGF-1, which
induced a greater than 50% fall in blood glucose (Fig. 5, A and B).
The hypoglycemic effect of rhIGF-I was completely inhibited by
pretreatment 10 min early with 500 µg of rhIGFBP-1 (Fig. 5A). In a
second study (n = 5–6 rats/group), both the dose of rhIGFBP-1 and
the time before the injection of rhIGF-I were changed. The lower dose
of rhIGFBP-1 (250 µg) only partially inhibited the hypoglycemic
effect of 50 µg of rhIGF-I (Fig. 5B). This partial effect of the
lower dose of rhIGFBP-1 could be due to a longer delay (20
vs. 10 min in Fig. 5A) before the injection of rhIGF-I.
Therefore, rhIGFBP-1 showed activity in vivo as it could
either completely inhibit (Fig. 5A) or attenuate (Fig. 5B) the
hypoglycemic effect of an injection of rhIGF-I. In the study shown in
Fig. 5C, diabetic rats (n = 3–5 rats/group) were fasted for
6 h and given 500 µg of rhIGFBP-1, then an injection of PBS or
rhIGF-I 10 min later. By itself, 500 µg of rhIGFBP-1 had no immediate
effect on blood glucose but completely inhibited the hypoglycemic
effect of 50 µg of rhIGF-I.
|
shows an experiment
where 500 µg rhIGFBP-1 was given at time zero followed 10 min later
by an injection of 50 µg of des(1, 2, 3)IGF-I. By itself, rhIGFBP-1 had
no effect on blood glucose and did not inhibit the hypoglycemic effect
of des(1, 2, 3)IGF-I. Similar data were obtained in normal rats (data not
shown). Figure 6B shows an experiment where diabetic rats (3/group)
were given an IV injection of PBS or rhIGFBP-1 (500 µg), then 20 min
later given an IV injection of insulin (0.5 U/kg). The hypoglycemic
effect of insulin was not affected by pretreatment with rhIGFBP-1.
Similar data were obtained in normal rats (data not shown). These
experiments support the concept that the glycemic effects of rhIGFBP-1
are mediated by it binding IGF-I.
|
, A and
B, shows the effects of rhIGFBP-1 on plasma glucose and insulin levels
in two independent experiments. The first study used two doses of
rhIGFBP-1 (100 and 500 µg); the second study used only the high dose
(500 µg). In both studies, a group of rats treated with PBS served as
the control. The blood glucose data confirms the data shown in Fig. 4
;
there were no changes in blood glucose after dosing with rhIGFBP-1
alone. However, an acute, prolonged, and dose-related rise in blood
insulin levels followed the injection of rhIGFBP-1. The rise in insulin
in the first study (Fig. 7A) was statistically significant after 5 and
10 min for high dose rhIGFBP-1 and for low dose rhIGFBP-1 after 5 min,
whereas in the second study the increase was statistically
significant after 5 min.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In addition to administering rhIGFBP-1 to normal rats, we also used diabetic rats. In diabetic rats or humans (12), insulin levels are low compared with IGF-I levels that remain comparatively normal. If endogenous IGF-I did acutely help regulate blood glucose, then an injection of rhIGFBP-1 might be expected to have a greater effect in a diabetic rat where the contribution of IGF-I to glucose regulation might be higher than in a normal rat. However, in diabetic rats we could find no acute effect of rhIGFBP-1 on blood glucose. We administered rhIGFBP-1 and then at varying times later administered other agents. It is clear that up to 20 min later rhIGFBP-1 inhibited the hypoglycemic effect of rhIGF-I.
IGF-I can bind to the insulin receptor, but it has been proposed that most of the metabolic effects of IGF-I are mediated by type 1 IGF receptors (13). Therefore, administered rhIGFBP-1 most likely affects insulin release by binding to and inhibiting the ability of IGF-I to bind and act at the IGF-I receptor rather than the insulin receptor (14). However, for other IGFBPs, especially IGFBP-3, there are reports of binding protein receptors and therefore IGFBP actions that are independent of IGF-I or its receptors (15). To address the mode of action of rhIGFBP-1, we studied its effect on the hypoglycemic response to rhIGF-I, des (1, 2, 3)IGF-I, or insulin. We used des(1, 2, 3)IGF-I as it binds poorly to hIGFBP-1 (16), and insulin as it is does not bind to the IGFBPs. Pretreatment with rhIGFBP-1 inhibited the hypoglycemic effect of rhIGF-I, but not that of des(1, 2, 3)IGF-I or insulin. This confirms that the in vivo effects of rhIGFBP-1 are probably due to it binding to IGF-I and inhibiting access to IGF receptors.
Animal studies have shown that rhIGFBP-1 inhibits the hypoglycemic effect of rhIGF-I (8), body growth (17), plasma insulin, and the responses to exogenous rhIGF-I (this paper). In two papers, an enhancing effect of IGFBP-1 on IGF-I action was found after the local application to wounds in rats (18) and in mice and rabbits (19). In these wound-healing models, IGFBP-1 by itself had either no activity (18) (16) or a stimulatory effect (19). In a recent review, in vitro data were described (4) as showing IGFBP-1 to have both inhibitory and stimulatory effects on IGF action. In summary, in vivo the effect of systemic IGFBP-1 administration has uniformly been found to inhibit IGF action. A release of insulin caused by an injection of rhIGFBP-1 inhibiting endogenous IGF activity is in line with this consensus.
There is evidence that endogenous IGF-I assists in the regulation of intermediary metabolism with actions on both insulin secretion and glucose transport. But, in the rat (20), infusions of low-dose rhIGF-I (0.39 nmol/h) do not augment glucose uptake in peripheral tissues. However, such low doses of rhIGF-I (0.39 nmol/h), although not sufficient to significantly alter peripheral glucose transport, directly and dramatically reduced insulin secretion. Much higher doses (1.96 nmol/h) were needed before rhIGF-I affected peripheral glucose uptake (20). In vitro data reinforces this idea as 40-fold lower concentrations of rhIGF-I (26 pmol/liter) are needed to affect insulin secretion (21) compared with the concentrations (1 nmol/liter) needed to affect glucose transport (22, 23). In some in vitro studies (21), insulin secretion is affected by very low concentrations of rhIGF-I (2 ng/ml) that are well in the range believed to be free in blood under physiological conditions (12, 24, 25). Therefore, a rise in insulin is a logical outcome after rhIGFBP-1 administration, assuming that rhIGFBP-1 reduces the bioavailability of rhIGF-I and that endogenous free IGF-I levels are low (12). It has recently been proposed (12) that there are three fractions of IGF in blood, the total amount, an easily dissociable free fraction (2% of the total) and a free fraction that can be ultrafiltered (1% of the total). Which of these pools of IGF, particularly the proposed two free pools of IGF, are able to bind to IGF receptors, or are accessible to IGFBP-1, is as yet unclear. It is not clear how the IGFs affect insulin secretion. The in vivo administration of rhIGF-I clearly inhibits insulin secretion, but in vitro the effects of rhIGF-I on insulin production have been less clear-cut(14). In humans (1), as in the rat, low doses of rhIGF-I can reduce insulin secretion without changing blood glucose. The preservation of normal glucose levels in the face of a reduction of insulin could be due to rhIGF-I, simultaneous with its effect on the pancreas, causing an increase in insulin sensitivity. Another explanation could be that the inhibitory effect of rhIGF-I on the pancreas is balanced by a direct stimulatory effect on peripheral glucose uptake into tissues without affecting insulin sensitivity (26). Similar explanations can be used to explain the rise in insulin secretion caused by the administration of IGFBP-1.
There are two independent reports of IGFBP-1 being overexpressed in mice. One group (27) has reported no clear effect of overexpressing human-IGFBP-1 on glucose homeostasis in mice. Another group reports that mice overexpressing a rat IGFBP-1 transgene are hyperinsulinemic (28). It is unclear why these two groups obtained different phenotypes in their IGFBP-1 transgenic mice. The phenotype, in terms of glucose metabolism, of mice homozygous for a deletion of IGFBP-1 is awaited with interest.
The phosphorylation state of IGFBP-1 is regulated by hormones (29), and there is some evidence that phosphorylation might affect its activity. For example, when applied to a wound, dephosphorylated rhIGFBP-1 plus rhIGF-I improved wound strength, whereas phosphorylated hIGFBP-1 plus rhIGF-I had no effect (18). Therefore, it is possible that the effects of phosphorylated hIGFBP-1, such as that used by Lewitt et al. (8) may be different than the effects of the nonphosphorylated rhIGFBP-1 used in the current studies.
The present experiments suggest that endogenous IGF-I tonically inhibits insulin secretion and suggest that aberrant IGFBP-1 levels may reduce IGF-I activity and contribute to excess insulin secretion from a nonstimulated pancreas. Humans with diabetes have disregulated IGFBPs that may affect IGF-I bioactivity and lead to increased insulin secretion and perhaps in the long-term to hyperinsulinemia and glucose intolerance. It is relevant that IGF-I increases insulin sensi-tivity (30) and has beneficial glycemic effects in humans with diabetes (31). Our data clearly show that the administration of rhIGFBP-1 affects insulin secretion suggesting that IGF-I may play a significant role in the regulation of glucose metabolism.
|
|
Acknowledgments |
|---|
Received October 9, 1996.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  T. Degenhardt, M. Matilainen, K.-H. Herzig, T. W. Dunlop, and C. Carlberg The Insulin-like Growth Factor-binding Protein 1 Gene Is a Primary Target of Peroxisome Proliferator-activated Receptors J. Biol. Chem., December 22, 2006; 281(51): 39607 - 39619. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Sandhu, J. M. Gibson, A. H. Heald, D. B. Dunger, and N. J. Wareham Association between Insulin-Like Growth Factor-I: Insulin-Like Growth Factor-Binding Protein-1 Ratio and Metabolic and Anthropometric Factors in Men and Women Cancer Epidemiol. Biomarkers Prev., January 1, 2004; 13(1): 166 - 170. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. B. Wheatcroft, M. T. Kearney, A. M. Shah, D. J. Grieve, I. L. Williams, J. P. Miell, and P. A. Crossey Vascular Endothelial Function and Blood Pressure Homeostasis in Mice Overexpressing IGF Binding Protein-1 Diabetes, August 1, 2003; 52(8): 2075 - 2082. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Le Roith, C. Bondy, S. Yakar, J.-L. Liu, and A. Butler The Somatomedin Hypothesis: 2001 Endocr. Rev., February 1, 2001; 22(1): 53 - 74. [Abstract] [Full Text]
|
|
|
|
 |
|
 Y. Dubaquie, D. L. Mortensen, A. Intintoli, D. A. Hogue, G. Nakamura, P. Rancatore, P. Lester, M. D. Sadick, E. Filvaroff, P. J. Fielder, et al. Binding Protein-3-Selective Insulin-Like Growth Factor I Variants: Engineering, Biodistributions, and Clearance Endocrinology, January 1, 2001; 142(1): 165 - 173. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. van Buul-Offers, M. van Kleffens, J. G. Koster, D. J. Lindenbergh-Kortleve, M. G. Gresnigt, S. L. S. Drop, C. M. Hoogerbrugge, R. J. Bloemen, J. A. Koedam, and J. W. van Neck Human Insulin-Like Growth Factor (IGF) Binding Protein-1 Inhibits IGF-I-Stimulated Body Growth but Stimulates Growth of the Kidney in Snell Dwarf Mice Endocrinology, April 1, 2000; 141(4): 1493 - 1499. [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 |
