| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Services d’Endocrinologie (O.C., J.V., M.M., I.B.), Chirurgie Générale et Thoracique (P.C.), Pathologie Cellulaire (F.L.-M.), Biochimie A (S.P.B., E.M.C.), Centre Hospitalier Universitaire, F-38043 Grenoble; and INSERM U-244, Département de Biologie Moléculaire et Structurale, CEA/G (P.L., E.M.C., G.D., J.-J.F.), F-38054 Grenoble, France
Address all correspondence and requests for reprints to: Dr. Olivier Chabre, Service d’Endocrinologie, Centre Hospitalier Universitaire, 38043 Grenoble, France. E-mail: olivier chabre{at}ujf-grenoble.fr
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
However, the roles of these ectopic receptors in adrenocortical tumorigenesis remains unclear. In the normal adrenal cortex, ACTH regulates both cortisol secretion and trophicity. It acts through activation of a G protein-coupled receptor that stimulates adenylate cyclase (10). It has thus been assumed that activation of ectopic receptors could have similar effects on the adrenocortical cells and could be responsible for both hypercortisolism and tumorigenesis. To date, however, no data have been provided on the potential effect of these receptors on adrenocortical cell proliferation. Temporary suppression of the activation of hormone receptors ectopically expressed in adrenals has been attempted by different means: inhibition of GIP secretion by somatostatin (2, 3, 5) or ß2-adrenergic receptor blockade by propanolol (1). These treatments resulted in the inhibition of cortisol secretion, but did not induce any measurable regression of adrenal hyperplasia or tumors. Thus, the relationship between ectopic expression of hormone receptors and the development of adrenal hyperplasia or tumors remains to be addressed.
We report here the study of a patient suffering from a food-dependent, ACTH-independent, Cushing’s syndrome related to a single adrenocortical adenoma. In vivo cortisol secretion was stimulated by any type of food intake, but not by insulin, iv glucose, or orthostatism. In vitro cortisol secretion by the dispersed tumor cells was stimulated by GIP, but not by glucagon-like peptide 1 (GLP-1). The tumor cells also responded to ACTH both in vivo and in vitro, suggesting that they had retained functional ACTH receptors despite suppression of ACTH secretion by hypercortisolism. Thus, the tumor cells provided a unique model to compare the mechanism of action of an unconventional stimulator, GIP, to that of the physiological regulator of adrenal cortex secretion and proliferation, ACTH.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Case report
A 32-yr-old woman was admitted for suspicion of hypercortisolism
because of a 10-kg weight gain in the upper part of the body, rounding
of the face, hirsutism, acnea, secondary amenorrhea, and high blood
pressure (160/90 mm Hg). All signs had developed during the previous
year. A standard 1-mg dexamethasone overnight suppression test had
showed a plasma cortisol value of 121 nmol/L at 08 h. However,
because of the severe clinical features, further exploration was
decided. Free urinary cortisol excretion was 1700 nmol/24 h (normal,
120–250), and circadian variations in serum cortisol were remarkable
because of low values in the morning and peaks always after meals (Fig. 1A
). Plasma ACTH was below or very close
to the detection limit (0.5 pmol/L) at all times, and the standard 2-mg
dexamethasone suppression test did not prevent the peaks of plasma
cortisol or diminish free urinary cortisol excretion. The
ACTH-independent stimulation of cortisol secretion followed any kind of
meal: mixed, oral glucose (100 g), fat-based (490 Cal; 82% fat, 16%
carbohydrate, and 2% protein), and protein-based (490 Cal; 87%
protein, 8% carbohydrate, and 5% fat) meals. When an overnight fast
was performed for 19 h, the morning and afternoon peaks of
cortisol secretion were suppressed (Fig. 1B). Intravenous injection of
glucose (100 g/3 h) had no effect on cortisol secretion. Injection of
10 IU insulin induced hypoglycemia, but no stimulation of cortisol or
ACTH secretion. Measurements of plasma GIP concentrations showed that
preoperative values of plasma cortisol and GIP were correlated (r
= 0.9; Fig. 1, A and B). Subcutaneous injection of 500 µg of the
somatostatin analog octreotide 45 min before a meal blunted the
postprandial stimulation of GIP (160% vs. 400% without
octreotide) and cortisol (134% vs. 280%). An iv GIP
stimulation test was proposed to the patient, but she refused
consent.
|
). Four months after
surgery, ACTH secretion was restored, and cortisol levels were
correlated to ACTH (r = 0.93), but not to GIP (r = -0.38)
levels (Fig. 1D).
Standard 1-h ACTH stimulation tests with iv injection of 250 µg
ACTH-(1–24) (Cosyntropin) were also performed 8 h preoperatively
and 5 days or 4 months postoperatively. They showed that 79% of the
cortisol response to ACTH was lost after removal of the tumor. Four
months after surgery, basal and ACTH-stimulated cortisol secretion of
the controlateral gland had increased (Table 1). Eight months after surgery another
test was performed at 11 h (2 h after a normal breakfast), and
cortisol values were 174 and 334 nmol/L before and after ACTH
injection. Maximal stimulation of cortisol remains lower than normal
(>580 nmol/L), indicating persistent hypotrophy of the remaining
adrenal gland.
|
ACTH-(1–24) (Cosyntropin, Synacthene) was obtained from Ciba (Basel, Switzerland). Human GIP and human GLP-1-(7–36) amide (GLP-1) were purchased from Bachem Biochimie (France); carbamylcholine chloride (carbachol) and forskolin were obtained from Sigma (St. Louis, MO). Plasma ACTH-(1–39) was measured by immunoradiometric assay, and GIP by RIA with commercial kits [Nichols Institute (San Juan Capistrano, CA) and Peninsula Laboratories (Belmont, CA), respectively]; cortisol was determined by RIA using antiserum from Endocrine Sciences (Calabasa, CA); cAMP was measured by RIA with a cAMP antiserum given to us by Dr. José Saez (Lyon, France); inositol 1,4,5-trisphosphate was determined by RIA using [3H]D-myo-inositol 1,4,5,-trisphosphate (IP3; Amersham, Les Ullis, France).
Cell isolation and culture
Six grams of fresh human tumor were freed of fat and sliced (0.5-mm section) with a Steady-Riggs microtome. The slices were washed three times in medium A [Ham’s F-12-DMEM (1:1) containing 10 mmol/L HEPES, 14 mmol/L NaHCO3, and antibiotics (20 U/mL penicillin, 50 µg/mL streptomycin, and 20 U/mL nystatin)]. The tissue was supplemented with 50 mL medium A containing 3 mg/mL collagenase A (Boehringer Mannheim, Indianapolis, IN) and 0.1 mg/mL deoxyribonuclease (Sigma) and incubated for 30 min at 37 C with stirring. The suspension was then filtered on sterile gauze, and the volume was completed to 50 mL with medium A containing 10% horse serum and 2.5% FCS before centrifugation for 10 min at 400 x g. The cell pellets were washed with the same medium and further purified on a Percoll (Pharmacia, Uppsala, Sweden) gradient (50% Percoll in Ham’s F-12 medium) preformed by centrifugation during 30 min at 62,000 x g. The cells were layered on this gradient, then centrifuged for 15 min at 2,000 x g. Three zones appeared on the gradient: the top one contained cellular fragments, the bottom one contained erythrocytes, and the middle one contained adrenal cells. This latter fraction was aspirated and washed twice with medium A containing serum. It contained 40 x 106 cells; 10 x 106 cells were used in suspension, and 30 x 106 cells were seeded in multiwell dishes and cultured for 24 h in Ham’s F-12-DMEM (1:1) containing 10% FCS with antibiotics, 10 µg/mL transferrin, 10 µg/mL insulin, and 10-4 mol/L vitamin C. After 24 h of culture, the medium was changed to serum-free Ham’s F-12-DMEM. Normal human adrenocortical cells were prepared with the same procedure using a fragment of adrenal removed for pheochromocytoma.
Cortisol production
Freshly dispersed human cells (105 cells in 1 mL) were incubated for 90 min in Ham’s F-12 with various hormones. Then, cortisol secreted into the medium was measured by RIA. Cultured cells were seeded in 16-mm wells at a density of 80 x 103 cells/well. Twenty-four hours later, the medium was removed and replaced by Ham’s F-12 medium during 2 h. Cortisol production was measured by RIA.
cAMP production
Freshly dispersed cells (105 cells in 1 mL) were incubated in KRGH medium (120 mmol/L NaCl, 4.8 mmol/L KCl, 1.2 mmol/L KH2PO4, 2.5 mmol/L CaCl2, 20 mmol/L HEPES, and 2 g/L glucose, pH 7.4) containing 1 mmol/L 3-isobutyl-1-methylxanthine (IBMX). Cells were incubated for 20 min with various effectors, and the reaction was stopped by the addition of 2 mL ethanol. After centrifugation, the supernatant was evaporated, sodium acetate buffer (50 mmol/L; pH 5.6) was added, and cAMP was measured by RIA. Cultured cells were seeded in 16-mm wells at a density of 80 x 103 cells/well. On day 1, medium was removed and replaced by KRGH containing 1 mmol/L IBMX, and a 20-min incubation with the various effectors was performed. The reaction was stopped by the addition of 2 mL ethanol. cAMP was measured by RIA.
IP3 determination
Freshly dispersed cells were incubated in KRGH at 37 C for 10 or 30 s with different effectors. The reaction was stopped by the addition of 0.2 mL perchloric acid. After neutralization with KOH, IP3 was measured in the supernatant by RIA.
DNA synthesis
DNA synthesis in human adrenocortical tumor cells or in bovine adrenocortical fasciculata-reticularis cells (BAC cells) was assessed in triplicate wells by measurement of [3H]thymidine incorporation. Cells were seeded in 16-mm wells at a density of 80 x 103 cells/well. After 24 h of culture in serum-supplemented medium, the cells were serum starved for 3 days in Ham’s F-12 containing 0.1% albumin and antibiotics. Quiescent cells were then incubated in fresh Ham’s F-12 containing 0.1% albumin and various hormones for the indicated periods of time. Then, 0.25 µCi [3H]thymidine (SA, 87 Ci/mmol) was added to each well during the last 3 h of incubation. Radioactivity incorporated into trichloroacetic acid- insoluble material was measured by scintillation counting.
Mitogen-activated protein (MAP) kinase assay
This assay was performed as previously described (11). Briefly, 24 h after seeding, the cells were serum starved for 48 h in F-12 medium containing 0.1% BSA before stimulation with various hormones. After stimulation, the medium was removed, and the cells were scrapped off and homogenized in a lysis buffer. The cell extract was centrifuged, and the supernatant was analyzed on a MonoQ Sepharose (Pharmacia, Piscataway, NJ) microcolumn with stepwise elution by increasing salt concentration. Western blotting analysis was performed on the elution fractions using a rabbit anti-p42mapk-p44mapk antiserum directed against a synthetic peptide from the C-terminus of rat p44mapk (gift from Dr. Jacques Pouyssegur, Nice, France), and phosphorylation of myelin basic protein was measured in the fractions containing p42mapk-p44mapk kinase immunoreactivity.
RT-PCR
Ribonucleic acid (RNA) from adrenal tumors or from other tissues was purified by a modification of the method of Chomczynski (12) using the total RNA isolation system from Promega (Charbonnieres, France). RNA (5 µg) was treated for 30 min at 37 C and for 5 min at 90 C with RQI deoxyribonuclease (Promega), and first strand complementary DNAs (cDNAs) were generated using 200 U reverse transcriptase (Superscript II, Life Technologies, Grand Island, NY) and 0.2 µg random hexamer DNA primers for 50 min at 37 C and for 15 min at 75 C. Control reactions without reverse transcriptase were performed for each RNA sample. cDNA (0.5 µg) was then PCR amplified in a final volume of 25 µL containing 2.5 U Taq DNA polymerase (Promega), 0.2% dimethylsulfoxide, and 14 pmol of each oligonucleotide primer. The amplification parameters were 94 C (2 min), then 35 or 42 cycles at 94 C (1 min), 55 C (1 min), and 72 C (2 min). For the human GIP receptor, two pairs of primers were designed, based on published sequences (13): sense 1 (nucleotides 99–123), TCACGATGACTACCTCTCCGATCC; antisense 1 (nucleotides 571–594), CGCCTGAACAAA-CTCAAGATGAGC; sense 2 (nucleotides 546–564), TCTCTCGCCACACTGCTGC; and antisense 2 (nucleotides 1008–1027), CAAGATGGTCATGAGGATGG.
For the human ACTH receptor, one pair of primers was designed, based on published sequence EMBL X65633: sense (nucleotides 753–774), GACTGTCCTCGTGTGGTTTTGC, and antisense (nucleotides 1012–990), ATGATGTCATCGGCTGTGGTTTC. The amplification parameters were 94 C (2 min), then 25 or 30 cycles at 94 C (1 min), 55 C (1 min), and 72 C (2 min).
To ensure semiquantitative results, the number of PCR cycles for each set of primers and probes was determined to be in the linear range of amplification. In addition, all cDNA samples were adjusted to yield equal amplification of a fragment of ribosomal protein L27 cDNA (14) as internal standard. Amplified products were separated by agarose gel electrophoresis (2%).
Origin of human tissues
Human tissues used as controls for GIP receptor or ACTH receptor RNA expression have the following origin. Cerebral tissue, collected from epileptic surgery, was a piece of temporal cortex outside epileptic foci. Neuropathological examination clearly indicated the absence of any tumoral process. Spleen tissue, collected from splenectomy for lymphoma, was a sample exempt of tumoral process. Adrenal tissues considered normal were adrenal cortex adjacent to a pheochromocytoma (one sporadic and one related to a germinal 634 mutation of the ret protooncogene); adrenocortical tumors were collected from unilateral adrenalectomy for Cushing’s syndrome (food independent), and adrenocortical hyperplasia tissues were collected from bilateral adrenalectomy for paraneoplastic ACTH-dependent Cushing’s syndrome. All tissues originate from patients who underwent surgery in the University Hospital of Grenoble, France.
Hybridization with labeled internal oligonucleotidic probe
Hybridization was performed with a 20-mer oligonucleotide probe
located in exon 9 of GIP receptor messenger RNA (mRNA; position
910–929 bp) specifically labeled with [
-32P]ATP by T4
polynucleotide kinase (Life Technologies). Agarose gel was transferred
under vacuum onto a Hybond N membrane (Amersham) in 4 N
NaOH for 1 h, and the membrane was washed twice with 1 x SSC
(0.15 mol/L NaCl and 15 mmol/L sodium citrate). Hybridization with the
labeled probe was performed overnight at 42 C in 5 x SSC, 0.1%
SDS, and 5 x Denhardt’s solution [prepared as described
previously (15)]. The blots were washed twice at room temperature in
2 x SSC, followed by two washes at 42 C with 2 x SSC and
0.1% SDS. Radiolabeled bands were visualized on a ß-imager
(PhosphorImager, Molecular Dynamics, Sunnyvale, CA).
In situ hybridization
In situ hybridization was performed using a 48-mer
oligonucleotide probe located in exon 4 of the GIP receptor (position
276–323) and with a 43-mer oligonucleotide probe located in exon 5
(position 387–430). The oligonucleotides were labeled using terminal
deoxynucleotide transferase (Boehringer Mannheim) and
[
-35S]deoxy-ATP (New England Nuclear-DuPont, Boston,
MA; 1250 Ci/mmol). The tissue sections were incubated with labeled
probe in the presence or absence of a 100-fold excess of unlabeled
probe as previously described (16). Slides were exposed for 28 days at
4 C before revelation and counterstaining with toluidine blue.
Statistics
Data are reported as the mean of triplicate determinations ± SD. All experiments in BAC cells were performed at least three times in an independent fashion. Statistical analysis of the raw data was performed by ANOVA followed by appropriate post-hoc tests (Student’s t test and Scheffe’s F test). Unless otherwise indicated, values are taken as significant for P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Tumor cells, normal human adrenocortical cells, and bovine adrenocortical cells were obtained by enzymatic dispersion and put in suspension or maintained in primary culture for 72 h. Measurements of cortisol and second messenger synthesis were performed on both cell suspension and primary culture, whereas measurement of [3H]thymidine incorporation and MAP kinase activity were performed only on primary cultures.
Cortisol secretion and second messenger synthesis
In tumor cells, cortisol secretion was stimulated by ACTH,
forskolin, and GIP, whereas GLP-1 had no effect. Stimulation by GIP was
maximal at the lowest concentration (0.1 nmol/L) tested in cell
suspension (Fig. 2A) and showed a
dose-response curve with a maximal response at 1 nmol/L in primary
culture (Table 2). In normal cells either
in suspension or in culture, cortisol production was stimulated by 1
nmol/L ACTH (6- or 3-fold, respectively), whereas no stimulation was
elicited by 1 or 10 nmol/L GIP (data not shown). Stimulation of
cortisol secretion by GIP was correlated with a stimulation of cAMP
production similar to that obtained by ACTH or forskolin, whereas GLP-1
had no effect (Fig. 2B and Table 2). IP3 production was stimulated by
carbamylcholine (carbachol), an agonist for the acetylcholine
m1-muscarinic receptor, but by neither GIP nor ACTH (Table 3).
|
|
|
Quiescent tumor cells were treated with serum, ACTH, GIP, or
GLP-1, and incorporation of [3H]thymidine was assessed
24, 30, and 36 h later. We found that serum induced an 8- to
9-fold stimulation of [3H]thymidine incorporation. Both
ACTH and GIP induced a 3-fold stimulation at 36 h, whereas GLP-1
had no significant effect (Fig. 3). When
BAC cells and normal human adrenocortical cells were tested under the
same experimental conditions as the human tumor cells, GIP had no
effect, whereas ACTH stimulated [3H]thymidine 2- and
3.3-fold, respectively (data not shown).
|
MAP kinases are serine/threonine protein kinases whose activity is
essential to the control of cell proliferation (17). They can be
activated by signals originating from either tyrosine kinase or G
protein-coupled receptors (18). This prompted us to investigate the MAP
kinase responses to ACTH and GIP in the adrenal tumor cells. Quiescent
tumor cells were stimulated by serum, ACTH, or GIP. Activation of
p42-p44 MAP kinases was measured 8 min (acute response) and 120 min
(sustained response) after addition of the stimulus. Serum appeared to
stimulate MAP kinase activity 7-fold after 8 min and 4.5-fold after 120
min, whereas both ACTH and GIP induced 5- and 4-fold stimulations after
8 min and 120 min, respectively (Fig. 4).
When tested under the same conditions, BAC cells also showed a
significant activation of MAP kinase by ACTH (2-fold; data not
shown).
|
GIP and ACTH receptor mRNA expressions were assessed by
semiquantitative RT-PCR of tumor RNA. For GIP receptor, two pairs of
primers were used. The first pair (S1-AS1) allows amplification of a
fragment that includes part of exon 2, exons 3–5, and a part of exon
6. This fragment encodes for the predicted extracellular domain of the
receptor, transmembrane domain I, and part of intracellular loop I.
Amplification of tumoral cDNA with S1-AS1 showed three bands (Fig. 5A), one of the expected size (495 bp)
and two smaller bands of approximately 390 and 290 bp. Sequencing of
the amplification products revealed that all three bands contained GIP
receptor cDNA sequences; the larger one was the expected entire
fragment, whereas the smaller bands were lacking exon 4 or both exons 4
and 3. The second pair (S2-AS2) allowed amplification of a fragment
that includes part of exon 6 and exons 7–10. It encodes the region of
the receptor comprised between transmembrane segments I and V.
Amplification of tumoral cDNA with S2-AS2 showed a weak band of the
expected size (481 bp) and a smaller, but intense, band around 420 bp
(Fig. 5B). Sequencing of the amplification products showed that the
smaller band was constituted of exons 6, 7, 8, and 10, but lacked exon
9. The presence of exon 9 in the larger band was confirmed by specific
hybridization (Fig. 5C).
|
, A and B).
No bands were detected from normal adrenal cortex, even after 42 cycles
of amplification (data not shown).
For the ACTH receptor, a fragment of 259 bp, which encodes the
transmembrane domains I and II and the first intracellular loop, was
amplified by RT-PCR. After 25 cycles of amplification, ACTH receptor
cDNA was detected in normal adrenal, but not in the tumor or in the
adjacent hypotrophic adrenal tissue (data not shown). After 30 cycles
of amplification, ACTH receptor RNA was detected in the tumor but at a
lower level than in a normal adrenal or pathological adrenals
responsible for ACTH-dependent Cushing’s syndrome or ACTH-independent
and food-independent Cushing’s syndrome (Fig. 6).
|
In situ hybridization was performed on tumor slices
using a radiolabeled GIP receptor cDNA probe. Hybridization signal was
clearly detected on adrenocortical tumor cells, but was not observed in
lymphocytic aggregates or in lipomatous metaplasia (Fig. 7).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(21). Both
types of mutations mimic a constant stimulation of the thyroid cells by
TSH and thus provide a good molecular explanation for the TSH
independence of thyroid toxic adenoma cells. Similarly, as ACTH
receptor also activates adenylate cyclase (10) through coupling to
Gs, one could expect to observe activating mutations of the
ACTH receptor or Gs in adrenal Cushing’s adenoma.
However, despite extensive search, no activating mutation of the ACTH
receptor has been reported in adrenal tumors (22, 23).
Gs mutations are also generally not found in human
adrenocortical neoplasms (24) and were reported in only two cases of
Cushing’s syndrome: an adrenal adenoma in a patient carrying the rare
McCune-Albright syndrome (25) and adrenal hyperplasia in an infant
(26). Mutations of another G protein -subunit, Gi2,
which, on the contrary, drives inhibition of adenylate cyclase, had
originally been reported in 3 of 8 adrenocortical tumors (21). However,
this finding was not confirmed by all subsequently published work on
the topic, with no mutations of Gi2 found in a total of
85 adrenal tumors (24, 27, 28). Therefore, the molecular mechanisms
leading to adrenal Cushing’s adenoma remain largely unknown, even if
some cases (25, 26) are consistent with the hypothesis that activation
of the adenylate cyclase pathway may be involved. In adrenal Cushing’s syndrome, ACTH-independent cortisol secretion is usually very constant over time, consistent with the hypothesis that the adrenal tumor or hyperplasia responsible for hypercortisolism could be related to a molecular event mimicking a constant ACTH stimulation. However, the observation, in two patients, that cortisol secretion varied throughout the day as a function of food intake led to the recognition that in some cases, cortisol secretion was not truly autonomous but was controlled by a hormone other than ACTH, such as the gut hormone GIP (2, 3). This concept could later be extended to other hormones, such as vasopressin (4, 7, 8), and catecholamines (1). In some of these cases, it could be demonstrated that the mechanism underlying the abnormal hormonal control of cortisol secretion was ectopic expression (or overexpression) of the corresponding receptor in the pathological adrenal (1, 5, 9). Except for three cases (5, 6, 7), all patients reported with adrenal Cushing’s syndrome linked to ectopic receptor expression had bilateral adrenal hyperplasia, which suggests that abnormal expression of the receptor was present in all adrenal cells.
Our patient presented a severe ACTH-independent hypercortisolism and a right unilateral adrenal mass. Surprisingly, cortisol levels were low in the morning and showed a diurnal variation coincident with food intake. A causal relationship between food intake and cortisol secretion was suggested by the observation that fasting prevented cortisol secretion. All subsequent in vivo explorations (including the lack of effect of iv glucose, insulin, or orthostatism; the positive effect of glucidic, lipidic, proteic, or mixed meals; and the prevention of food-induced cortisol secretion by somatostatin pretreatment) pointed to an abnormal control of cortisol secretion by a stimulus linked to food intake, whatever its type. There are only two known gut hormones whose secretion is stimulated by food intake: GIP and GLP-1. Both hormones are incretins, as their main function is to stimulate insulin secretion by the pancreatic islet ß-cells. In our patient, serum GIP levels were highly correlated with serum cortisol, which made this hormone a likely candidate for controlling adrenal tumor cortisol secretion. However, for ethical reasons (refusal of the patient), it was not possible to test the in vivo effect of GIP or GLP-1. Besides food, cortisol secretion could also be stimulated by synthetic ACTH-(1–24) (Cosyntropin). It should be stressed, however, that this sensitivity to ACTH could not be responsible for hypercortisolism, as endogenous secretion of ACTH was totally suppressed.
The relationship between the adrenal mass and the food-induced hypercortisolism was demonstrated by surgery; right adrenalectomy was followed by cure of hypercortisolism, which was replaced by profound hypocortisolism, with loss of the responses to both food and ACTH stimulation. Pathological examination showed a single adrenal adenoma, with hypotrophic adrenal cortex tissue adjacent to it. This suggested that the adenoma was solely responsible for both food-induced hypercortisolism and the cortisol response to ACTH-(1–24). Four months after surgery, ACTH secretion had recovered, and cortisol secretion, still insufficient, was strictly correlated to ACTH levels, but no longer to GIP levels. This confirmed that the controlateral adrenal gland was not sensible to food intake, and that it had reacted normally first to suppression, and then to restoration of ACTH secretion. Thus, the conclusion of the in vivo observations is that the adrenal adenoma cells were sensitive both to a factor linked to food intake, presumably GIP or GLP-1, and to ACTH.
The in vitro experiments were designed to study the potential mechanisms of action of GIP and GLP-1 on the adenoma cells, with ACTH as a positive control. Cortisol secretion by cells in suspension proved to be sensitive to stimulation by GIP and ACTH, but not by GLP-1, an observation that was reproduced in plated cells. The response to GIP seemed more sensitive in suspended cells, but was observed in both conditions at 0.1 nmol/L, which is in the range of GIP serum concentrations in the patient and normal subjects (3). This result allowed to compare the mechanisms of action of GIP and ACTH in the tumor cells, with GLP-1 as a negative control. We show that GIP, like ACTH, stimulates the production of cAMP. This is consistent with the observation that GIP stimulates cAMP production in GIP-sensitive tissues (16) and in cells transfected with the GIP receptor (13, 16, 29). The rat GIP receptor was also shown to stimulate intracellular Ca2+ when transfected in HEK-293 cells (16) or COS-7 cells (29). In the latter report it was shown that Ca2+ was released from intracellular stores, presumably after liberation of IP3. This suggests that the GIP receptor can stimulate both adenylate cyclase and phospholipase C, like other G protein-coupled receptors from the same subfamily (30, 31). However, this is not the case in human adrenal adenoma cells, as production of IP3 could be detected after stimulation by carbachol, but not by GIP. Such differences in coupling between transfected cells and tissues has been observed with other G protein-coupled receptors and could result from differences in the levels of expression of G proteins (32) or other components of the signal transduction pathway.
We sought, then, to evaluate the potential effects of GIP on tumor cell proliferation. We found that in human adenoma cells, ACTH also had a significant stimulatory effect on DNA synthesis, although weaker than the effect of serum, and that this effect was reproduced by GIP. In BAC cells and normal human adrenocortical cells, GIP had no effect, whereas the stimulatory effect of ACTH was observed. This correlates well with the trophic effect of ACTH on normal adrenal cortex in vivo (33, 34) and suggests that GIP may also have a trophic effect on the tumor cells in vivo. It must be stressed that under different experimental conditions, ACTH was first reported to have a paradoxical inhibitory effect in vitro on BAC cell DNA synthesis and proliferation (35, 36, 37). However, a stimulatory effect has been observed under specific conditions (38), indicating that the in vitro effect of ACTH on adrenocortical cell DNA synthesis is dependent on the experimental protocol. Due to the limited amount of tumor material we could not test whether stimulation of DNA synthesis by GIP was correlated with an increase in tumor cell number. However, this was the case for BAC and normal human adrenocortical cells treated with ACTH under the same protocol; we measured increases in the cell number of, respectively, 40% and 30% after 96 h, compared to a 100% increase with serum (data not shown). Such an increase in cell number should represent the sum of the stimulation of cell proliferation and the inhibition of apoptosis (37).
cAMP has tissue-specific effects on growth, differentiation, and gene expression. In most cell types, cAMP inhibits cell proliferation and MAP kinases. By contrast, in some cells, it is stimulatory (39, 40, 41). Calleja et al. have shown that the effect of cAMP on MAP kinase stimulation depends on the cell type (42). In BAC cells, the effect of ACTH or cAMP on MAP kinase depends on the culture conditions: inhibitory in the conditions we had previously described (11) and stimulatory in the conditions used here. When we measured MAP kinase activity in the human adenoma cells under the same conditions, a weak stimulation by ACTH and GIP was observed. The correlation between ACTH or GIP effects on DNA synthesis and MAP kinase activity does not demonstrate that activation of MAP kinase is a necessary step in the stimulation of DNA synthesis in the human adenoma cells. However, it does provide other evidence that GIP has an effect similar to ACTH on these cells.
These data pointed to the presence of functional ACTH and GIP receptors in the tumor cells. The genes of rat (16) and human (13) GIP receptor have been cloned. In the rat, GIP receptor RNA expression is not limited to ß-cells of pancreatic islets; it is highly expressed in the brain and more weakly in different peripheral organs, including adrenal cortex, but it is not expressed in the spleen (16). RT-PCR amplification of GIP receptor mRNA demonstrated the presence of transcripts of this gene in the tumor, whereas the adjacent hypotrophic adrenal tissue and the pathological adrenals, either responsible for autonomous cortisol hypersecretion or linked to ACTH-dependent cortisol hypersecretion, did not appear to express this gene. GIP receptor RNA expression was also detected in normal human brain, but not in the spleen. Remarkably, different transcripts for the GIP receptor were detected in the tumor: the full-length transcript and smaller transcripts lacking exon 3, 4, or 9. This finding, which is similar to that made by Lacroix and collaborators (9), shows that in the tumor, the abnormal expression of the GIP receptor gene is associated with abnormal splicing of its mRNA. However the biological significance of the short transcripts is probably limited, at least concerning the transcript lacking exon 9, as this exon codes for a significant part of the putative transmembrane domain IV; the disruption of this domain is expected to profoundly alter the receptor function. In situ hybridization experiments confirmed that, as expected, the expression of the GIP receptor was restricted to the tumor adenoma cells. ACTH receptor gene was also expressed in the tumor cells, as shown by RT-PCR of ACTH receptor mRNA, although to a lesser degree than in the adjacent normal tissue. Such a low level of expression should, however, be sufficient to explain the response to pharmacological doses of 1–24 ACTH.
In conclusion, we have demonstrated that an adrenal adenoma responsible for food-dependent Cushing’s syndrome is characterized by a high level of expression of functional GIP receptors, and we provide several results suggesting that expression of this receptor may play a role in the development of the tumor. First, the expression of the GIP receptor gene was restricted to the adenoma cells; it was not detected in the adjacent adrenal tissue, and it should also be absent in the remaining controlateral gland, which is not sensitive to GIP. Therefore, unlike in food-dependent Cushing’s syndrome associated with bilateral hyperplasia (9), the genetic abnormality is not present in all of the adrenal cells of the patient and must result from a postzygotic event. The nature of this event remains to be defined, but it could result from a somatic mutation in a segment of DNA controlling GIP receptor gene expression. Secondly, GIP receptor expression can fully account for the peculiar food-dependent cortisol secretion of the tumor. Thirdly, stimulation of tumor cell DNA synthesis by GIP suggests that GIP receptor may also play a role in tumor cell development. Finally, all measured effects of GIP on the tumor cells were similar to the effects of ACTH. In particular, GIP induced an increase in cAMP production. This is in favor of an important role for activation of the ACTH signal transduction pathway in adrenal Cushing’s adenoma development.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received April 6, 1998.
Revised June 10, 1998.
Accepted June 16, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-mediated MAP kinase activation by
a common signalling pathway. Nature. 376:781–784.[CrossRef][Medline]
genes as a cause of toxic thyroid
adenomas. J Clin Endocrinol Metab. 82:2695–2701.2A-adrenergic receptor to multiple
G-Proteins: a simple approach for estimating receptor-G-protein
coupling efficiency in a transient expression system. J Biol Chem. 269:5730–5734. subunits of heterotrimeric G proteins stimulate
the mitogen-activated protein kinase pathway in COS-7 cells. J
Biol Chem. 269:7851–7854.This article has been cited by other articles:
 |
|
 |
|
  N M Albiger, G Occhi, B Mariniello, M Iacobone, G Favia, A Fassina, D Faggian, F Mantero, and C Scaroni Food-dependent Cushing's syndrome: from molecular characterization to therapeutical results Eur. J. Endocrinol., December 1, 2007; 157(6): 771 - 778. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Louiset, V. Contesse, L. Groussin, D. Cartier, C. Duparc, G. Barrande, J. Bertherat, H. Vaudry, and H. Lefebvre Expression of Serotonin7 Receptor and Coupling of Ectopic Receptors to Protein Kinase A and Ionic Currents in Adrenocorticotropin-Independent Macronodular Adrenal Hyperplasia Causing Cushing's Syndrome J. Clin. Endocrinol. Metab., November 1, 2006; 91(11): 4578 - 4586. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Lampron, I. Bourdeau, P. Hamet, J. Tremblay, and A. Lacroix Whole Genome Expression Profiling of Glucose-Dependent Insulinotropic Peptide (GIP)- and Adrenocorticotropin-Dependent Adrenal Hyperplasias Reveals Novel Targets for the Study of GIP-Dependent Cushing's Syndrome J. Clin. Endocrinol. Metab., September 1, 2006; 91(9): 3611 - 3618. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. L. Mazzuco, O. Chabre, N. Sturm, J.-J. Feige, and M. Thomas Ectopic Expression of the Gastric Inhibitory Polypeptide Receptor Gene Is a Sufficient Genetic Event to Induce Benign Adrenocortical Tumor in a Xenotransplantation Model Endocrinology, February 1, 2006; 147(2): 782 - 790. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Baldacchino, S. Oble, P.-O. Decarie, I. Bourdeau, P. Hamet, J. Tremblay, and A. Lacroix The Sp transcription factors are involved in the cellular expression of the human glucose-dependent insulinotropic polypeptide receptor gene and overexpressed in adrenals of patients with Cushing's syndrome J. Mol. Endocrinol., August 1, 2005; 35(1): 61 - 71. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. M. Swords, S. Aylwin, L. Perry, J. Arola, A. B. Grossman, J. P. Monson, and A. J. L. Clark The Aberrant Expression of the Gastric Inhibitory Polypeptide (GIP) Receptor in Adrenal Hyperplasia: Does Chronic Adrenocorticotropin Exposure Stimulate Up-Regulation of GIP Receptors in Cushing's Disease? J. Clin. Endocrinol. Metab., May 1, 2005; 90(5): 3009 - 3016. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Bertherat, V. Contesse, E. Louiset, G. Barrande, C. Duparc, L. Groussin, P. Emy, X. Bertagna, J.-M. Kuhn, H. Vaudry, et al. In Vivo and in Vitro Screening for Illegitimate Receptors in Adrenocorticotropin-Independent Macronodular Adrenal Hyperplasia Causing Cushing's Syndrome: Identification of Two Cases of Gonadotropin/Gastric Inhibitory Polypeptide-Dependent Hypercortisolism J. Clin. Endocrinol. Metab., March 1, 2005; 90(3): 1302 - 1310. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Groussin, K. Perlemoine, V. Contesse, H. Lefebvre, A. Tabarin, P. Thieblot, J. L. Schlienger, J. P. Luton, X. Bertagna, and J. Bertherat The Ectopic Expression of the Gastric Inhibitory Polypeptide Receptor Is Frequent in Adrenocorticotropin-Independent Bilateral Macronodular Adrenal Hyperplasia, but Rare in Unilateral Tumors J. Clin. Endocrinol. Metab., May 1, 2002; 87(5): 1980 - 1985. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Lacroix, N. N'Diaye, J. Tremblay, and P. Hamet Ectopic and Abnormal Hormone Receptors in Ad |
