| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Departments of Endocrinology (M.A., R.K.S., C.H.), Immunology (C.H.J., B.T.), and Pathology (J.B.L., H.D.S.), Odense University Hospital, DK-5000 Odense C, Denmark
Address all correspondence and requests for reprints to: Marianne Andersen, M.D., Ph.D., Department of Endocrinology, Odense University Hospital, 5000 Odense C, Denmark. E-mail: m.andersen{at}dadlnet.dk
Abstract
Immunohistochemical analysis of the distribution of human fetal antigen 1 (FA1) in adult human tissues has demonstrated a strong association between FA1 and (neuro)endocrine structures. In the anterior pituitary gland FA1 was colocalized with GH, and the present study was performed to evaluate a possible relationship between GH and FA1. FA1 and GH levels were measured during a 24-h period at 20-min intervals. In contrast to the known GH peaks during 24-h sampling, there was no detectable FA1 peak. The FA1 responses to placebo were not significantly different from the responses to the combination of pyridostigmine and GHRH. No significant difference was found between basal FA1 (nanograms per ml) levels [median (minimum–maximum)] in healthy adults [n = 40; 28.6 ng/ml (12.5–72.0)], acromegalic patients [n = 11; 31.0 ng/ml (21.6–56.3)], and patients with GH deficiency [n = 22; 32.1 ng/ml (13.4–108.7)].
FA1 levels were significantly reduced, in the six of seven acromegalic GH responders to octreotide, from [median (minimum–maximum)] 30.6 ng/ml (20.0–43.1) to 20.3 (13.9–30.2; P < 0.02). There was no significant change during placebo. FA1 levels were significantly increased compared with placebo values during 3 months of GH therapy. The increase in FA1 levels was significantly higher than the change during placebo (P < 0.003).
In conclusion, a common secretory and stimulatory pathway for FA1 and GH in healthy adults has been ruled out. However, we found that pharmacologically induced changes in GH levels during weeks to months had a corresponding direct or indirect effect on FA1 levels in patients with GH deficiency or acromegaly. However, a direct effect of octreotide on FA1 levels, independent of GH levels, has not been ruled out.
HUMAN FETAL ANTIGEN 1 (FA1) was originally
identified in and purified from human second trimester amniotic fluid
(1, 2, 3, 4). Analysis of the protein structure revealed that
FA1 was a heterogeneous glycoprotein with a molecular mass of
26.2–31.8 kDa and that the polypeptide backbone of 225–262 amino acid
residues included 6 epidermal growth factor (EGF)-like motifs (3, 4). FA1 is synthesized as a larger membrane-associated precursor
defined by a cDNA that has been referred to as adrenal-specific cDNA
(human pG2) (5), 
-like (mouse and human dlk)
(6) preadipocyte factor-1 [mouse (7), and
rat pref-1 (8)], and zona glomerulosa-specific factor
(rat) (9). These cDNAs are essentially identical and
define a protein with an extracellular domain (i.e. FA1),
followed by short juxtamembrane, transmembrane, and intracellular
regions.
The EGF repeats of FA1 are closely related to those within the
/Notch/Serrate family of transmembrane proteins originally described
in Drosophila melanogaster. In Drosophila,
and Serrate are ligands for Notch, and the cell fate of
Notch-expressing pluripotent cells depends in part on which ligand was
bound (10). Interestingly, although and Serrate
previously were thought to act as transmembrane proteins only, an
in vivo-generated soluble form of
(DlEC) acting as a Notch agonist has recently
been reported (11). Although the FA1 precursor does not
represent a vertebrate homolog for either of these invertebrate
proteins, FA1 might act in a manner similar to the /Notch/Serrate
system, as both membrane-associated and soluble FA1 have been shown to
be involved in the differentiation/proliferation of various cell types
(7, 12, 13, 14). These effects are in line with some of the
effects of GH. The dual effector theory of GH action states that GH has
both a direct differentiating effect and an indirect proliferating
effect through IGF-I (14A ).
Immunohistochemical analysis of the distribution of FA1 in adult human
tissues has demonstrated a strong association between FA1 and
(neuro)endocrine structures, i.e. the anterior pituitary
gland (15) and the adrenal gland, where a profound
staining of glomerulosa cells was observed (3), and the
insulin-producing ß-cells of the islets of Langerhans
(16). Recently, FA1 has been located within the sex
hormone-producing Leydig cells of the testis and the theca interna and
Hilus cells of the ovary (17). In the anterior pituitary
gland FA1 was colocalized with GH (15), and the present
study was performed to evaluate a possible relationship between GH and
FA1. Colocalization per se may be a reason for studying
cosecretion. In acromegaly the colocalization of 
-subunit and GH has
been studied. GHRH stimulated -subunit release both in
vitro and in vivo in patients with elevated -
subunit levels, indicating corelease of GH and -subunit
(18). A similar parallelism of the responses of GH and
-subunit to TRH and bromocriptine has been shown by
Ishibashi et al. (19). Hofland et
al. (19A ) studied the secretion of GH, PRL, and
-subunit in vitro. In human fetal (20) and
adult pituitaries (21) it has been reported that GH and
PRL can be stored and secreted by the same normal pituitary cells. We
found that patients with acromegaly and hyperprolactinemia will
normalize PRL levels during octreotide treatment (22). We
studied the physiological secretion of FA1 and GH in females only, as
basal FA1 levels did not differ significantly between the sexes, and
there was no correlation to age beyond adolescence (23).
We used a known potent GH stimulus, the combination of pyridostigmine
(PD) and GHRH, to examine a possible costimulation of FA1. The
influence of pathophysiologically high GH levels on FA1 was studied in
acromegalic patients. In patients with GH deficiency (GHD) the lack of
a possible source of FA1 was evaluated. The effect of GH therapy on FA1
levels was studied in patients with GHD. In patients with acromegaly, a
possible direct or indirect effect of octreotide was studied.
Materials and Methods
24-h FA1 and GH profiles
We studied 10 lean healthy women [mean age (minimum–maximum), 25.9 yr (23–33)]; the mean ± SEM body mass index (BMI) was 20.4 ± 1.7 kg/m2. None of the participants had any medical problems or was taking any drugs. They all had regular menstrual cycles, and blood sampling took place in the follicular phase. A heparinized iv cannula (1.1 x 25 mm; Introducer, Carmeda, Täby, Sweden) was placed in a forearm vein. Blood withdrawal began at least 30 min after venipuncture through a nonthrombogenic catheter (ConFlo system, Carmeda) inserted through the cannula and connected to a peristaltic pump (Swemed Lab Pump, Carmeda). The blood samples were drawn continuously. The flow rate was 3 ml/h, with a shift to a new vial after a 20-min interval. The subjects were hospitalized and offered a standardized menu at 0900, 1200, and 1730 and a snack between meals at 1500 and 2000 h. The menu contained 7000–8000 KJ/d comprised of a maximum 30% lipids, 50–60% complex carbohydrates, and 15–18% proteins. The automatic blood-sampling system allowed the participants to sleep during the night while blood was sampled. The subjects were not allowed to sleep during the day, but they were allowed to sit or walk around in a quiet regimen.
Basal and stimulated GH and FA1
In 40 normal adults [38.3 yr (22–58)], the mean BMI was
22.4 ± 0.3 kg/m2. All IGF-I values were
normal (24), and the mean IGF-I level was 224 ± 11
µg/liter. According to age, sex, and use/nonuse of oral
contraceptives, 5 groups of 8 individuals were selected, as shown in
Table 1
, making a total of 40 subjects.
In 7 of these subjects a placebo test was performed. None of the
participants had any medical problems or were taking drugs other than
oral contraceptives; the premenopausal women using oral contraceptives
had been using them for at least 1 yr. They were all within 10% of
ideal body weight. The participants attended the clinic after fasting
overnight, and during the experiment they were not permitted to smoke,
eat, sleep, or drink anything but tap water. A cannula was inserted
into an antecubital vein for blood sampling and administration of GHRH
(Groliberin, Pharmacia Biotech, Uppsala, Sweden)/placebo.
At 0900 h, 120 mg PD (Mestinon, Hoffmann-La Roche Inc., Basel, Switzerland) or placebo was administered
orally, and 60 min later, at 1000 h (time zero), 1 µg/kg BW GHRH
or placebo was administered iv as a bolus (25). Serum
samples for the measurement of GH and FA1 were taken at -60, 0, 20,
30, 60, and 90 min.
|
Unstimulated FA1 was determined in 11 acromegalic patients, 5 females, and 6 males [50.4 yr (26–71)]; the BMI was 26.7 kg/m2 ± 1.1. All patients had active acromegaly in accordance with the consensus report (26), and all had a pituitary adenoma verified by computed tomography. The effect of octreotide (Sandostatin, Novartis, Basel, Switzerland) on FA1 was studied in 7 acromegalic patients: 4 females and 3 males [51.2 yr (26–71)]. Octreotide was compared with placebo in a randomized, double blind, cross-over trial. Treatment consisted of octreotide or saline given as placebo administered sc into the thigh three times daily (0630, 1330, and 2130 h). Each treatment trial consisted of 4 wk, with a washout period of 12 wk. A diurnal GH profile was determined before and at the end of each 4-wk period; in each patient serum GH was measured at 0800, 1000, 1200, 1400, 1600, and 2000 h (27). Serum samples for FA1 were determined at wk 0 and 4 during both octreotide and placebo treatments.
GH substitution of patients with GHD
GHD was diagnosed using a PD-GHRH test (see above) as well as an insulin tolerance test (ITT). The ITT was performed after an overnight fast; actrapid (0.15 U/kg BW) was administered iv. If the patient had an established insufficiency of the pituitary-adrenal axis, 0.10 U/kg BW actrapid was used instead. Serum samples for GH were determined at 0, 30, 60, 90, and 120 min. During the ITT all patients had blood glucose levels below 2.2 mmol/liter and clinical symptoms of hypoglycemia.
Unstimulated FA1 levels were determined in 22 patients [8 females and 14 males; 48.3 yr (20–61)], with a mean BMI of 26.6 ± 0.8 kg/m2. The 22 patients included in a randomized clinical trial of GH treatment vs. placebo for 12 months [10 patients: 4 females and 6 males; 51.5 yr (38–61)] were randomized for GH treatment; one of the patients left the study after 3 months. Twelve patients [4 females and 8 males (51 yr (20–54)] were randomized for placebo (28). FA1 levels were determined at 0, 3, 6, 9, and 12 months.
The samples used in this study were obtained in connection with other investigations. The Declaration of Helsinki II was observed, and the local ethical committee has approved the study. All subjects were volunteers, and they signed an informed consent document before taking part in the study.
Biochemical analysis
Serum FA1 was quantified using a sandwich ELISA technique based on polyclonal anti-FA1 antibodies purified by immunospecific affinity chromatography. The technique and assay parameters have previously been described in detail (23). The assay is free of interference with rheumatoid factors and does not cross-react with human EGF (23) or GH, PRL, hCG, LH, FSH, or TSH (29). During the study, interassay coefficients of variation based on three quality controls were below 5%. Serum GH was determined by RIA (Pharmacia Biotech). We evaluated the comparability between two GH assays: a polyclonal assay (Pharmacia Biotech) and a monoclonal assay (Delfia). We found that it was possible to use a conversion factor to alternate between the two methods without losing diagnostic power for the GH measurements: RIA = Delfia x 1.59 (30). The level of detection for the assay was 0.4 mU/liter, the intra- and interassay coefficients of variation at 0.5 mU/liter were 14% and 10%, respectively, and at 22 mU/liter they were 4% and 5%, respectively. Serum IGF-I was measured after extraction with HCl/ethanol (30 µl serum in 750 µl). After centrifugation the supernatant was further diluted 1:40 in assay buffer. It was then determined using an immunofluorometric sandwich assay with two monoclonal antibodies following the Delfia principle and using an AutoDelfia reader (Wallac, Inc., Turku, Finland). The sensitivity limit was 2.5 ng/liter. Intra- and interassay coefficients of variation were lower than 1.9% and 8.6%, respectively. The determination was made in a single assay (24).
Statistics
Data are presented for each individual and as the median (minimum–maximum) or the mean ± SEM; P < 5% was considered statistically significant. The Friedman two-way ANOVA was carried out to compare more than two paired groups; if an overall significant P value was found, relevant pairs of data were compared (31). The Wilcoxon signed rank test was used for testing paired differences between data from placebo and active treatment, and the Mann-Whitney U test was used for comparing differences between two groups. Spearman’s rank correlation was used to test any association between numerical values.
Results
The diurnal rhythm
FA1 and GH levels were measured during a 24-h period at 20-min
intervals. The albumin and FA1 concentrations revealed similar
fluctuations; therefore, the diurnal FA1 and GH variations shown in
Fig. 1 have been corrected for apparent
fluctuations in albumin concentrations. In contrast to the known GH
peaks during 24-h sampling, there was no detectable FA1 peak.
|
There was no significant difference between the unstimulated FA1
levels in the five groups of normal adults (n = 40; Table 1
). No
significant difference was found between basal FA1 (nanograms per ml)
levels [median (minimum–maximum)] in healthy adults [n = 40;
28.6 ng/ml (12.5–72.0)], acromegalic patients [n = 11; 31.0
ng/ml (21.6–56.3)], and patients with GH deficiency (GHD; n =
22; 32.1 ng/ml (13.4–108.7); Fig. 2].
No correlation was found between basal FA1 levels and age, BMI, or
IGF-I levels in either healthy adults or patients with acromegaly or
GHD. The basal FA1 levels in patients with GHD were not correlated to
peak GH responses to either the PD-GHRH test or the ITT. Three patients
had unmeasurable GH levels in response to the ITT; these patients had
high basal FA1 levels. The three patients had normal basal IGF-I levels
(data not shown) (24, 25).
|
The combination of PD and GHRH significantly stimulated GH levels
in 40 healthy adults. The FA1 responses to placebo were not
significantly different from the responses to the combination of PD and
GHRH (Table 2).
|
GH levels were significantly reduced by octreotide in six of seven
patients by 62% (48–87), only patient 2 did not significantly respond
to octreotide treatment. In the six responders the FA1 levels were
significantly reduced during octreotide treatment from [median
(minimum–maximum)] 30.6 (20.0–43.1) to 20.3 (13.9–30.2;
P < 0.02). The percent reduction was 32% (16–48).
There was no significant change during placebo (Fig. 3). IGF-I levels were significantly
reduced during octreotide therapy from 392 µg/liter (213–693) to
153.5 µg/liter (86–492; P < 0.02). There was no
significant change during placebo. There was no significant correlation
between the changes induced in GH/IGF-I levels and FA1 levels in
acromegalic patients.
|
FA1 levels were significantly increased compared with those in the
placebo group during 3 months of GH therapy (Fig. 4). The increase in FA1 levels was
significantly higher than the change during placebo (P
< 0.003, 
change in FA1 levels during GH vs. placebo).
The basal values were not significantly different in the two groups
(P > 0.05). During 6, 9, and 12 months of GH therapy,
the effect on FA1 levels was reduced (data not shown). There was no
significant correlation between IGF-I changes and FA1 changes during 6
and 12 months of therapy in patients with GHD. GH was not measured
during therapy.
|
We have previously reported the localization of FA1 within the adult anterior pituitary gland. Immunohistochemical staining revealed a colocalization of FA1 and GH in the somatotroph cells. However, the two proteins were present in different intracellular compartments, as judged by their different staining pattern (15), and in contrast to GH, FA1 staining was most intense in the Golgi zone (15). The Golgi-associated localization of FA1 within the somatotroph cells is in agreement with the posttranslational modifications, i.e. glycosylations, of this protein (4) and strongly suggests that our immunohistochemical findings reflect FA1 expression within this type of cell. A recent study has reported the presence of FA1 in the developing human adenohypophysis gland as well (32). FA1 is synthesized as a transmembrane protein, which is subsequently cleaved at the part nearest the membrane (4, 12), thus releasing soluble FA1 to the extracellular compartment. This proteolytic cleavage probably also takes place at the somatotroph cell membrane. FA1 has been identified in culture supernatants from organotypic slice cultures of GH-secreting pituitary adenomas (Larsen, J. B., personal communication). Although FA1 and GH are present in distinct compartments within the somatotroph cell, we considered the possibility of cosecretion of these two proteins. However, in our in vivo data the physiological secretion of FA1 and GH did not correspond to either stimulated or unstimulated conditions. In healthy adults GH is released in pulses from the somatotroph cells in the pituitary gland and is regulated primarily by GHRH, a 44-amino acid peptide that is essential to the initiation of the GH pulse, and by somatostatin, an inhibitory peptide that modulates the amplitude of the GH pulse (33). PD, a cholinesterase inhibitor, potentiates GHRH-induced GH release (25). The known cholinergic stimulation of GH release may act through an inhibition of somatostatin (34). The combination of PD and GHRH only stimulated GH, whereas FA1 levels were unchanged compared with placebo. During the 24-h sampling, FA1 levels were slightly higher in the early morning hours, possibly reflecting the lower glomerular filtration rate during this time of the day (23). The 24-h secretion of GH and FA1 did not correspond in healthy females. These findings excluded a common secretory and stimulatory pathway for GH and FA1 in healthy adults. We have ruled out that basal FA1 levels were of somatotroph cell origin only, as 3 GH-deficient patients with unmeasurable GH response to the insulin tolerance test and panhypopituitarism had high FA1 levels. The high GH levels in patients with acromegaly were not reflected in high FA1 levels. As their basal FA1 levels corresponded to the levels in healthy adults, a possible explanation for the normal basal values may be that FA1 is cleared from blood by the kidneys (23), and both GH and IGF-I increase plasma flow and the glomerular filtration rate (35, 36). However, changes in GH levels over weeks to months in patients with GHD or acromegaly significantly influenced FA1 levels. In 6 of 7 acromegalic patients with a significant reduction in GH levels during 4 wk of octreotide therapy, FA1 levels were reduced as well. The reduction of FA1 levels may be due to a reduction of GH levels or IGF-I levels per se; alternatively, it may be caused by a direct effect of octreotide on FA1 secretion from the known FA1 sources. In patients with GHD we found that FA1 levels significantly increased during 3 months of GH therapy. The changes in GH levels may directly or indirectly influence FA1 levels, indirectly through parallel changes in IGF-I levels. Patients with GHD have no or very few GH-secreting cells, excluding a direct effect of GH therapy on the GH-secreting cells in these patients. In isolated rat pancreatic islets GH and PRL have been shown to increase pref 1 mRNA levels (8).
We have, as yet, few data on FA1 excretion and/or degradation and the possible regulatory role of FA1 levels. However, an important regulatory role of GH on pharmacokinetics would be less likely. Circulating FA1 may be indirectly or directly regulated by changes in GH levels in a number of ways, including transcription, translation, processing, and secretion. Although we have excluded somatotroph cells as the only source of basal FA1, the anterior pituitary gland may be secretory active in healthy adults or in certain patients. In the future the possible pancreatic FA1 source as well as other putative localizations will be evaluated.
The results of several in vitro studies suggest a regulatory role for FA1 as well as its membrane-associated precursor, dlk/pref-1, during the differentiation of specific cell types (12, 13, 14, 32). There are four major splice variants of the pref-1/dlk gene, two of which lack the sequences that are needed for generating the soluble gene product (e.g. FA1) (12, 37, 38). Thus by alternate splicing, the cell might either produce a pref-1/dlk protein that is membrane associated only or a form from which FA1 can be generated. Interestingly, a recent study suggests that that these protein variants play opposite roles during adipogenesis (39). Another recent study has shown that pref-1/dlk expressed in thymic epithelial cells may be involved in supporting thymocyte cellularity (40). Although these and other reports strongly suggest that FA1 may act as an important local modulator of proliferation/differentiation, we have no knowledge regarding the effect of FA1, if any, on distant cells. It seems most likely that serum FA1 levels reflect the activity in dlk/pref-1-expressing cells; however, we cannot exclude that circulating FA1 may be an active component.
GH and FA1 are related topologically in the pituitary gland; however, GH and FA1 might also be functionally related. In conclusion, a common secretory and stimulatory pathway for FA1 and GH in healthy adults has been ruled out. However, we found that pharmacologically induced changes in GH levels during weeks to months had a corresponding direct or indirect effect on FA1 levels in patients with GHD or acromegaly. However, a direct effect of octreotide on FA1 levels, independent of GH levels, has not been ruled out. Further studies both in vitro and in vivo will be necessary to characterize this new member of the EGF superfamily and its possible relation to GH.
Footnotes
Abbreviations: BMI, Body mass index; EGF, epidermal growth factor; FA-1, human fetal antigen 1; GHD, GH deficiency; ITT, insulin tolerance test; PD, pyridostigmine.
Received February 22, 2001.
Accepted July 18, 2001.
References
-like protein in pancreatic islets: molecular cloning and
expression pattern during development and growth of the endocrine
pancreas. Endocrinology 138:3940–3948-
subunit response to growth hormone (GH)-releasing hormone in
patients with active acromegaly. Evidence for -subunit and GH
coexistence in the same tumoral cell. J Clin Endocrinol Metab 61:541–546-subunit by pituitary
tumors. J Clin Endocrinol Metab 64:1187–1193-subunit prolactin release by
cultured pituitary adenoma cells from acromegalic patients: correlation
with GH release. Clin Endocrinol (Oxf) 30:601–611[Medline]
-like (dlk) variants and requires the expression of
membrane-associated dlk. Differentiation 64:103–114[Medline]
This article has been cited by other articles:
 |
|
 |
|
  B. M. Abdallah, M. Ding, C. H. Jensen, N. Ditzel, A. Flyvbjerg, T. G. Jensen, F. Dagnaes-Hansen, J. A. Gasser, and M. Kassem Dlk1/FA1 Is a Novel Endocrine Regulator of Bone and Fat Mass and Its Serum Level Is Modulated by Growth Hormone Endocrinology, July 1, 2007; 148(7): 3111 - 3121. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. M. Abdallah, P. Boissy, Q. Tan, J. Dahlgaard, G. A. Traustadottir, K. Kupisiewicz, J. Laborda, J.-M. Delaisse, and M. Kassem dlk1/FA1 Regulates the Function of Human Bone Marrow Mesenchymal Stem Cells by Modulating Gene Expression of Pro-inflammatory Cytokines and Immune Response-related Factors J. Biol. Chem., March 9, 2007; 282(10): 7339 - 7351. [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 |
