Variable Expression of the V1 Vasopressin Receptor Modulates the Phenotypic Response of Steroid-Secreting Adrenocortical Tumors*

GIORGIO ARNALDI+, JEAN-MARIE GASC, YVES DE KEYZER, MARIE-LAURE RAFFIN-SANSON, VÉRONIQUE PERRAUDIN, JEAN-MARC KUHN, MARIE-CHARLES RAUX-DEMAY, JEAN-PIERRE LUTON, ERIC CLAUSER, AND XAVIER BERTAGNA

Groupe d’Etudes en Physiopathologie Endocrinienne, INSERM CJF 9208, Institut Cochin de Génétique Moléculaire; Laboratoire d’Explorations Fonctionnelles Endocriniennes, Hôpital Trousseau; and INSERM U-36, Collège de France; Paris; and Endocrinologie et Maladies Métaboliques, CHU de Rouen, Rouen, France

ABSTRACT

We studied the putative role of the vasopressin receptors in the phenotypic response of steroid-secreting adrenocortical tumors. A retrospective analysis of a series of 26 adrenocortical tumors respon- sible for Cushing’s syndrome (19 adenomas and 7 carcinomas) showed that vasopressin (10 IU, im, lysine vasopressin) induced an ACTH- independent cortisol response (arbitrarily defined as a cortisol rise above baseline of 30 ng/ml or more) in 7 cases (27%). In comparison, 68 of 90 patients with Cushing’s disease (76%) had a positive cortisol response. We then prospectively examined the expression of vaso- pressin receptor genes in adrenocortical tumors of recently operated patients (20 adenomas and 19 adrenocortical carcinomas). We used highly sensitive and specific quantitative RT-PCR techniques for each of the newly characterized human vasopressin receptors: V1, V2, and V3. The V1 messenger ribonucleic acid (mRNA) was detected in nor- mal adrenal cortex and in all tumors. Its level varied widely between 2.0 × 102 and 4.4 × 105 copies/0.1 µg total RNA, and adenomas had significantly higher levels than carcinomas, although there was a large overlap. Among the 6 recently operated patients who had been

subjected to the vasopressin test in vivo, the tumor V1 mRNA levels were higher in the 4 responders (9.5 × 103 to 5.0 × 104) than in the 2 nonresponders (2.0 × 102 and 1.8 × 103). One adenoma that had a brisk cortisol response in vivo, also had in vitro cortisol responses that were inhibited by a specific V1 antagonist. In situ hybridization showed the presence of V1 mRNA in the normal human adrenal cortex where the signal predominated in the compact cells of the zona re- ticularis. A positive signal was also present in the tumors with high RT-PCR V1 mRNA levels; its distribution pattern was heterogeneous and showed preferential association with compact cells. RT-PCR stud- ies for the other vasopressin receptors showed a much lower signal for V2 and no evidence for V3 mRNA. We could not establish whether the V2 mRNA signal observed in normal and tumoral specimens was present within adrenocortical cells or merely within tissue vessels.

We conclude that the vasopressin V1 receptor gene is expressed in normal and tumoral adrenocortical cells. High, and not ectopic, ex- pression occurs in a minority of tumors that become directly respon- sive to vasopressin stimulation tests. (J Clin Endocrinol Metab 83: 2029-2035, 1998)

A LTERED membrane transduction plays a pathogenetic role in endocrine tumors, as shown in recent examples where mutated Gs proteins or G protein-coupled transmem- brane receptors are constitutively activated (1-3).

Adrenocortical tumors have provided yet another patho- genetic concept: that of illegitimate receptor expression with the demonstration that gastric inhibitory polypeptide (GIP) receptor expression was responsible for unusual and rare cases of steroid-secreting adrenocortical tumors, the activity of which was triggered by food intake (4, 5). Other cases were recently reported where cortisol-secreting tumors responded

Received August 4, 1997. Revision received February 18, 1998. Ac- cepted February 26, 1998.

Address requests for reprints to: Dr. Xavier Bertagna, Clinique des Maladies Endocriniennes et Métaboliques, Hopital Cochin, 24 rue du Fg St. Jacques, 75014 Paris, France. E-mail: xavier.bertagna@cch. ap-hop-paris.fr.

* This work was supported in part by INSERM CJF 9208, the INSERM Réseau de Recherche Clinique Comète, the Fondation de France, and the Plan Hospitalier de Recherche Clinique.

+ Recipient of a fellowship from the Assciazone Italiana Ricerca sul Cancro and a Poste Vert from INSERM.

in an unanticipated manner to such agents as ß2-adrenergic agonists or vasopressin (6-8).

Vasopressin has long been used as an investigative tool for Cushing’s syndrome with the concept that only patients with the Cushing’s disease would respond to the test with a cor- tisol rise secondarily induced by the direct action of vaso- pressin at the pituitary level (9-13). Hence, the apparently illegitimate response of cortisol in patients with a primary, ACTH-independent, adrenocortical tumor.

Vasopressin is classically associated with vasoconstriction, antidiuretic, and ACTH-releasing activities. Vasopressin is also involved in metabolic actions and in cell growth and differentiation. Three types of vasopressin receptors have been characterized and recently cloned [V1 (or V1a), V2, and V3 (or V1b)] (14-17). The presence of V1 receptors in adrenal gland (18) has raised the possibility that vasopressin can directly stimulate adrenal steroidogenesis in both physio- logical and pathological conditions (19, 20). Recent studies have shown that vasopressin stimulated cortisol and aldo- sterone secretion in vitro from normal human adrenal frag- ments and cultured cells via activation of V1 receptors (21, 22). These data suggested that vasopressin receptor expres-

sion (at least of the V1 type) in the adrenal cortex was not an illegitimate phenomenon.

We hypothesized that an apparently aberrant vasopressin responsiveness of steroid-secreting adrenocortical tumors could result either from overexpression of the V1 receptor or from the illegitimate expression of other vasopressin recep- tor subtypes. Analysis of the various vasopressin receptors subtypes was performed by RT-PCR competition-based quantitation and by in situ hybridization for the V1 type. We show that a minority of vasopressin-responsive adrenocor- tical tumors express large amounts of the V1 vasopressin receptor.

Subjects and Methods

Patients

Retrospective study of vasopressin responsiveness in a series of patients with Cushing’s syndrome. We retrospectively examined the plasma cortisol and ACTH responses of 116 patients submitted to a vasopressin stim- ulation test (see below) in the last 24 yr at our institution. The test was performed at the initial work-up of the patients, who were all hyper- cortisolic (increased urinary cortisol excretion and/or lack of suppres- sion under the low dose dexamethasone suppression test). All patients with adrenocortical tumors (n = 26) had undetectable and/or unre- sponsive ACTH plasma levels (<5 pg/mL). All patients with Cushing’s disease (n = 90) had typical laboratory findings and/or proven corti- cotroph adenomas after pituitary surgery.

The vasopressin stimulation test was performed as previously described (10); 10 IU vasopressin (lysine vasopressin, Ciba-Geigy, Lausanne, Switzerland) were administered im at 0800 h, and blood was drawn at 0, 15, 30, 45, and 60 min. The response to vasopressin was measured by substracting the cortisol value at 0 min from the highest value between 15-60 min. A positive response was arbitrarily defined as equal to or higher than 30 ng/mL.

Tissue studies in recently operated patients with adrenocortical tumors. Tissue specimens were obtained from 39 patients with adrenocortical tumors (20 adenomas and 19 carcinomas). The status of each tumor was further assessed by measuring the IGF-II messenger ribonucleic acid (mRNA) content (23). As expected, all adenomas (except 1) had normal contents, whereas most carcinomas (15 of 19) had elevated values (data not shown).

Six patients had undergone a vasopressin stimulation test; there were four responders (three adenomas and and one carcinoma) and two nonresponders (one adenoma and one carcinoma).

Specimens of normal adrenocortical tissue were obtained by careful dissection of glands in three patients who had undergone unilateral adrenalectomy for nonsecreting tumors.

The tissues at the time of surgery were immediately frozen in liquid nitrogen and stored at -80 C until nucleic acid extraction. The pathol- ogist dissected the tumor and provided a sample of fresh, homogeneous, nonnecrotic tumor tissue. Large blood vessels and fibrotic areas were excluded during the preparation of each specimen.

RNA extraction and RT-PCR analysis

Total RNA was isolated by ultracentrifugation in guanidinium iso- thiocyanate-CsCl followed by treatment with ribonuclease-free deoxyri- bonuclease I. RT-PCR was performed according to standard protocols. Briefly, random primed complementary DNA (cDNA) were synthesized from 1 µg total RNA with 200 U Moloney murine leukemia virus reverse transcriptase (BRL, Gaithersburg, MD), and 10% of the cDNA reaction was directly used for each PCR. The same cDNA was used for PCR amplification of glyceraldehyde-3-phosphate dehydrogenase (G3PDH).

Four sets of specific oligodeoxynucleotides were designed: V1, 5’-CGGCTTCATCTGCTACAACATC-3 and 5’-CGAGTCCTTCCA- CATACCCGT-3’; V2, 5’-GCTTGGGCCTTCTCGCTCCTT-3 and 5’- GCACAAAGGGCGCCCCTTCCA-3’; V3, 5’-ACCCCCACAGCAGG- CAAGG-3 and 5’-TCTCGGGTCAGCAGCATCAA-3’; and G3PDH,

5’-ATCCCATCACCATCTTCCAG and AGGGATGATGTTCTGGAGA- GC-3’.

To ensure complementary DNA-specific amplification, primers were chosen either flanking an intron [V1 and V3 (personal data)] or over- lapping the junction of two exons (for V2; Fig. 1a).

After 36 cycles of amplification (30 for V1 and 22 for G3PDH) of 45 s at 94 ℃, 45 s at 57-62 C, and 1 min at 72 C, 20% of the PCR products were analyzed on a 2% agarose gel, blotted onto a nylon membrane (Amer- sham, les Ulis, France), and probed with internal end-labeled oligonu-

FIG. 1. RT-PCR for vasopressin receptors. a, Schematic representation of the V1, V2, and V3 cDNAs (V, intron; , coding region; , trans- membrane domain) and the respective primers (=), oligonuleotides for hybridization (4), and expected fragments of RT-PCR amplification b, Map and scheme for the construction of the competitive V1 mRNA. Mutated V1 cDNA was generated by insertion of 80 bp within a region of V1 gene amplified. Corresponding RNA was then synthesized. c, Ex- ample of quantitative RT-PCR of V1 mRNA. The upper panel shows the gel analysis of RT-PCR reactions with a narrow range of mutant RNA amounts. The lower band in the agarose gel represents amplified DNA fragment derived from tissue sample V1 mRNA, and the upper band represents the superimposition of two bands from the competitor mu- tated V1 mRNA. At the equivalent point of signal intensities between the two bands, the amounts of target V1 mRNA are the same as the copy number of competitor. This point is calculated as follows: a, plotting the logarithm of the ratio of mutant over patient DNA, expressed in densi- tometric units, as a function of the logarithm of the copies of mutant V1 mRNA present in the same reaction; and b, extrapolating the value for x = 0 (3.87 in the present case).

a) V1

505 bp

2

3

4

5

2

4.

V2

439 bp

2

3

4

5

7

0

7%

V3

285 bp

2

3

4

6

3

un

b) V1

505 bp

2.

3

4

5

8

7

-

1

1

80 bp

Tth111

V1

505 bp

1

9

4

3

=

1

V1 mutant

585 bp

2

3

4

5

80 bp

8

I

+

c)

Dilutions of V1 mutant ( RNA copy number) RNA patient

9,7x102 3,9x103 1,5x1046,2x104 2,5x105

0, 1 µg 0, 1 µg 0, 1 µg 0, 1 µg 0, 1 µg

V1 mutant

585 bp 505 bp

V1

6

V1 RNAm (log copy number / 0,1 µg RNA Tot

y = 3,8754 + 1,2681x

R^2 = 0,923

-

5

.

.

4

.

3

.

2

-1

M

Log (V1 mut / V1 patient)

0

1

2

cleotides (Fig. 1a). The nature of the fragments was also confirmed by enzyme restriction analysis.

Quantitative / competitive RT-PCR for the V1 mRNA (Fig. 1, b and c)

Briefly, mutated V1 cDNA was generated by insertion of 80 bp (de- rived from pUC18) within a region of the V1 gene comprised between the specific PCR primers used. Corresponding RNA was synthesized using a RNA transcription kit. The concentration of mutant transcript was determined by measuring adsorbance at 260 nm, and quality was checked by agarose gel electrophoresis. In a first series of reactions, fixed amounts of tumoral RNA were mixed with increasing amounts of syn- thetic RNA ranging from 103-107 copies. After 27 cycles of amplification (45 s at 94 ℃, 45 s at 60 C, and 1 min at 72 C) and a final step of 10 min at 72 C, 40% of the PCR products were separated on ethidium bromide- stained 3% agarose gel. A second series of reactions was then performed with a narrowed range of mutant RNA dilutions, spaced by 4-fold only, to allow for precise RNA copy number determination for each individual tumor. After agarose gel separation and ethidium bromide staining, the products were quantified under UV light using NIH Image software. The logarithm of the ratio of band intensities within each lane was plotted against the logarithm of the copy number of synthetic RNA added per reaction. This function was linear. The quantity of target message was determined where the ratio of synthetic/ authentic V1 band intensities was equal to 1 (Fig. 1c). Results are presented in arbitrary units (1U corresponding to 1 copy V1 mRNA per 0.1 µg total RNA).

The two groups (adenomas and carcinomas) were compared by one- way ANOVA and unpaired Student’s test. Statistics were performed after logarithmic transformation of the data.

In situ hybridization

Normal and pathological tissues embedded in paraffin were used for in situ hybridization. Tissues were fixed in 4% buffered paraformalde- hyde immediately after removal from the patient. After 24 h in the fixative solution, tissues were washed in 70% ethanol, dehydrated, and embedded in paraffin. Sections (7 um) were cut and mounted on si- lanated histological slides.

The full-length cDNA (2.2 kilobases) of the V1 receptor was subcloned in pBluescript KS+ at the EcoRI site. After linearization by Xbal and Sall, the V1 receptor cDNA was transcribed by T3 or T7 RNA polymerases (Boehringer Mannheim, Mannheim, Germany) in the presence of [35S]UTP (Amersham) to generate, respectively, the antisense and sense probes. The detailed protocol for in situ hybridization has been recently published (24).

Perifusion experiments

The perifusion system technique was previously described (25). Briefly, a fragment of the tumor obtained at surgery was immersed in 200 mL DMEM and rapidly transported to the laboratory. The tumor tissue was diced into small pieces (1-2 mm3), rinsed three times with fresh medium (DMEM), mixed with Bio-Gel P2 (Bio-Rad, Richmond, CA) and transferred into polystyrene columns delimited by two Teflon pestles. The perifusion chambers were supplied with DMEM at constant flow rate (260 µL/min) at 37 C and pH 7.4. The perifusion medium was continuously gassed with a 95% O2-5% CO2 mixture. The tumor slices were allowed to stabilize for 2 h before any substance was administered. Test substances were dissolved in gassed DMEM and infused into columns at the same flow rate as DMEM alone. Effluent fractions were collected at 5-min intervals and kept at -20 C until assay. Vaso- pressin and [d(CH2)5,Tyr(OMe)2]arginine vasopressin ([d(CH2)5,Tyr (OMe)2]AVP) were purchased from Sigma Chemical Co. (St. Louis, MO). DMEM was supplied by Life Technologies (Grand Island, NY).

Cortisol concentrations were determined in all fractions collected by RIA using antisera developed in our laboratory (25). The antibodies showed significant cross-reaction with 11-deoxycortisol (5%), but the cross-reactivities were much lower with corticosterone and 11-deoxy- corticosterone (0.28% and 0.10%, respectively). The antibodies exhibited very low cross-reaction (<0.01%) with all other steroids tested. The assay was sensitive enough to detect 1 pg cortisol. The intra- and interassay coefficients of variation were 2% and 8%, respectively.

Plasma assays

ACTH was measured by a highly specific immunoradiometric assay with a detection limit of 5 pg/mL (ELSA ACTH, CIS-Bio International, Gif-sur-Yvette, France), and plasma cortisol levels were measured using a competitive binding assay as previously described (26).

Results

Retrospective study of vasopressin responsiveness in a series of patients with Cushing’s syndrome (Fig. 2)

Sixty-seven of the 90 patients with Cushing’s disease (76%) had a positive (≥30 ng/ml) cortisol response to vasopressin stimulation. Seven of the 26 patients with Cushing’s syn- drome due to a primary cortisol-secreting adrenocortical tu- mor (27%) had a positive cortisol response to vasopressin stimulation. This response was ACTH independent; in fact, plasma ACTH levels were undetectable and remained un- responsive (<5 pg/mL). Among the 7 responsive adreno-

FIG. 2. Cortisol response to vasopressin in patients with Cushing's syndrome. Individual plasma cortisol increase after vasopressin ad- ministration (10 IU lysine vasopressin, im) in 26 patients with ad- renocortical tumors (left) and 90 patients with Cushing's disease (right). Responders (plasma cortisol increase ≥30 ng/ml) are shown with bold symbols.

500

Plasma cortisol increase (ng/ml)

400

300

200

100

0

30

-100

Adrenal Cushing’s tumors disease

FIG. 3. V1 receptor mRNA in normal and tumoral adrenals. Tissue contents are shown as copy number per 0.1 µg total RNA on a log scale in 20 adenomas, 19 carcinomas, and 3 normal glands. The tumors of the 6 patients who had been subjected to in vivo vasopressin stimu- lation are indicated, with a solid circle for the four responders and a dotted circle for the two nonresponders.

P< 0.01

106

105

V1 mRNA

(copy number / 0.1 µg total RNA )

%

104.

103

102.

Adenomas

Carcinomas

Normal adrenals

cortical tumors, there were 5 adenomas and 2 adrenocortical carcinomas.

Tissue studies in recently operated patients with adrenocortical tumors

RT-PCR detection of vasopressin receptors in tissue RNA. Ex- pression of the V1 receptor gene was easily detected (20 cycles PCR), giving the expected signal of 505 bp in all normal adrenal glands and all tumoral tissues. The expected 439-bp fragment was amplified with the V2 primers in most adrenal samples and in 1 human kidney that was used as a positive control (data not shown). However, the amount of V2 RT- PCR product appeared considerably lower than that of V1 RT-PCR product, as 36 cycles of PCR were needed. The signals of V1 and V2 were not detected without the RT reaction. Thus, normal adrenal gland and most adrenal tu- mors express both V1 and V2 receptors, at least at the mRNA level. In contrast, RT-PCR V3 signal was absent in all tissues, whether normal or tumoral, even after 38 cycles.

Quantitative/competitive RT-PCR of V1 mRNA (Fig. 3). RT-PCR competition-based quantitation was performed for V1 mRNA to further validate the results of comparative RT- PCR. Mean V1 mRNA was significantly decreased in carci- nomas compared to adenomas (2.2 × 103 vs. 1.0 × 104; P < 0.01). However, the levels of V1 gene expression showed a wide range of variation in both tumor groups.

In vivo vasopressin tests had been performed in six patients of this prospective series; they showed that tumor levels of

V1 mRNA correlated with clinical findings; V1 mRNA levels were low in the two nonresponders (1.8 × 103 in one ade- noma; 2.0 × 102 in one carcinoma) and high in the four responders (9.5 × 103, 3.4 × 104, and 5.0 × 104 in three adenomas; 4.6 × 104 in one carcinoma).

Effect of vasopressin on cortisol secretion by tumor tissue in vitro (Fig. 4). The effect of vasopressin on cortisol production was directly studied in vitro on perifused adrenocortical tumor fragments of a patient whose adenoma had high levels of V1 mRNA (5.0 × 104) and who had a brisk cortisol response to vasopressin in vivo (Fig. 4). Exposure of tumor tissue to vasopressin (10-7 mol/L) induced cortisol secretion that reached a maximum relative increase over baseline (202 ± 27%) within 30 min. To investigate the type of receptor in- volved in the stimulatory action of vasopressin, tumor frag- ments were exposed to the nonapeptide in the presence of the [d(CH2)5,Tyr(Ome)2]AVP, a selective V1 vasopressin recep- tor antagonist. As shown in Fig. 4 vasopressin-evoked cortisol secretion was totally abolished by [d(CH2)5,Tyr (Ome)2]AVP.

In situ hybridization (Fig. 5). In situ hybridization was per- formed to further characterize the cell type(s) in adrenal gland that expressed the V1 receptor. The results showed that V1 mRNA was present in the normal adrenal cortex, mainly in the compact cells of the zona reticularis and less in the zona glomerulosa and fasciculata. In situ hybridization demon- strated a very high intensity signal in tumoral samples of two patients who had high tissue contents in V1 mRNA by RT-

In vivo response

In vitro response

FIG. 4. In vivo and in vitro vasopressin responsiveness in the same patient. In vivo response: in vivo vasopressin test showing the plasma cortisol response and undetectable ACTH (O; < 5 pg/mL). In vitro response: in vitro perifusion of the tumor fragments showing the brisk response to vasopressin (10-7; left) and its blockade by the specific V1 antagonist [d(CH2)5, Tyr(OMe)2]AVP (10-6; right). The patterns represent the mean (±SEM) of three independent perifusions. Each point is the mean of two consecutive 5-min fractions. The basal cortisol release (100%) was calculated as the mean of the fractions just preceding the administration of the secretagogue.

V1 antagonist

VP

VP

VP

250

Plasma cortisol (ng/ml)

300

200

Plasma ACTH (pg/ml)

Released cortisol

(% of baseline)

200

·

150

100

100

50

50

0

9

P

Q

9

o

0

0

0

15

30

45

60

0

60

120

180

240

0

60

120

180

240

300

TIME (min)

TIME (min)

FIG. 5. In situ hybridization of V1 mRNA. Distribution of V1 receptor mRNA by in situ hybridization in nor- mal adrenal gland (A-C) and tumors (D-G). In the normal adrenal gland (A), V1 is detected mainly in the zona re- ticularis (ZR) and also in the zona fas- ciculata (ZF), but at a much lower level of intensity. The zona glomerulosa (ZG) shows only a background level of label- ing, comparable to that obtained with the sense probe (B). Histological stain- ing (hematoxylin-eosin) of a neighbor- ing section (see inset in A) confirms the identification of ZR and ZF (C). In an adenoma with a moderately elevated level of V1 receptor (9.5 × 103), a spotty pattern of hybridization shows that only single cells or small clusters ex- press the V1 receptor (D). Histological staining of this tumor (E) shows that positive cells look like ZR cells and not like spongiocytes, i.e. ZF cells. In a car- cinoma with higher levels of V1 receptor mRNA (3.3 × 104), the V1 is highly ex- pressed in large, histologically homoge- neous nodules of undifferentiated cells (F and G) that are different from either ZR or ZF cells. A, B, D, and F, Darkfield illumination; bar = 100 um. C, E, and G, Brightfield illumination; bar = 20 um.

A

B

C

D

E

F

e

PCR [9.5 × 103 in one adenoma (Fig. 5D); 3.3 × 104 in one carcinoma (Fig. 5F)]. The specific V1 expression was localized in adrenocortical tumoral cells, particularly in compact cells with a nonhomogeneous pattern in the same tumor (Fig. 5, D and F).

Discussion

Several studies in animals and man have demonstrated that adenocortical tumors can express ectopic receptors and, therefore, exhibit abnormal or unanticipated responses to stimulation by various hormones such as catecholamines, LH/hCG, TSH, PRL, glucagon, GH (27-30), and, quite re- cently, GIP (4, 5). The concept has emerged that steroid- secreting adrenocortical tumors may not be really autono-

mous but, rather, are under the influence of abnormal, ectopic or overexpressed, membrane receptors.

In contradiction with a long accepted dogma, we and others have described cases of vasopressin-responsive ad- renocortical tumors in patients with ACTH-independent hy- percortisolism (6-8). We show here that it occurs in a mi- nority of such tumors (27%), which can be either benign or malignant. This figure is less than that reported by another group (7), who found 100% responders under a similar stim- ulatory test. In our retrospective study we were careful to check that all of our patients were hypercortisolic and had no ACTH response. We arbitrarily chose 30 ng/ml as a signif- icant increase; a recent study by Dickstein et al. (13) showed that a positive cortisol response to vasopressin (arbitrarily

defined as a 20% relative increase) was present in 74% of patients with Cushing’s disease. Using their criteria, our figures would not change (26% for adrenocortical tumors and 76% for Cushing’s diseases), giving an almost identical rate of responders for the group of patients with Cushing’s disease as in Dickstein’s study (13).

It is now well established that vasopressin can directly stimulate steroidogenesis in normal animal and human ad- renocortical cells via the activation of V1 receptors (19-22, 31). We hypothesized that the likelihood of a tumor being responsive to vasopressin would be directly correlated with the level of expression of a eutopic V1 receptor in the tumor. We indeed showed that only those patients with high V1 mRNA levels were responsive to vasopressin in vivo. In gen- eral, higher values were found in benign tumors, but an occasional carcinoma had both extremely high tumor V1 mRNA content and a brisk cortisol response to vasopressin in vivo.

Because V1 receptor gene expression occurs in many dif- ferent cell types, including in vessels, it was necessary to demonstrate its presence in adrenocortical cells. Two studies of human adrenocortical tumors have shown a direct action of vasopressin in vitro on cortisol secretion and on intracy- tosolic calcium movements (8, 22). We show here the direct action of vasopressin on perifused fragments from an ade- noma responsible for Cushing’s syndrome and show its blockade by a specific V1 antagonist.

Using in situ hybridization, we show the presence of V1 mRNA directly in human adrenocortical cells and reveal its nonhomogeneous pattern of distribution. In the normal ad- renal cortex, a V1 mRNA signal is present in all three zonas, yet it shows a preferential association with the compact cells of the zona reticularis. This cell type is supposedly respon- sible for preferential androgens production and may partic- ipate in adrenocortical cell regeneration. These features might have some connection with the proposed growth ac- tion of vasopressin on adrenal cortex (32). In situ V1 mRNA signals were easily detected in some tumors selected for their high contents, as determined by quantitative RT-PCR. In the two studied tumors, the V1 mRNA signal showed a highly heterogeneous pattern, but in both cases, it was absent in clear cells and appeared specifically associated with compact cells: in the adenoma with a spotty pattern (Fig. 5D), and in the carcinoma in large homogeneous nodules made of un- differentiated cells (Fig. 5F). These in situ data also confirmed that V1 mRNA was present in tumor cells and not in vessels.

We found no evidence for V3 gene expression in normal or tumoral adrenal cortex. In contrast, we found a definite, although very low, RT-PCR signal for the V2 receptor in the normal gland and in most tumors. A previous study showing the lack of effect of the V2-specific agonist 1 deamino-8-D- arginine vasopressin (desmopressin) on a vasopressin- responsive adrenocortical tumor suggests that this receptor type is not involved with steroid secretion (6). It is more likely that this slight signal reflects the presence of V2 receptors in tumor-associated vessels.

A spectrum of more or less spectacular endocrine abnor- malities can be observed when structurally normal mem- brane receptors are overexpressed. Extreme overexpression of ß2-adrenergic receptors in the heart of transgenic mice

leads to abnormal function even in the absence of ligand (33, 34); apparently “illegitimate” expression of GIP or B2-adre- nergic receptors in adrenocortical tumors leads to cortisol oversecretion when the “illegitimate” receptor is triggered by physiological (“legitimate”) conditions such as food in- take, postural changes, or even stress. At the end of this spectrum, we show here that eutopic (legitimate?) V1 gene expression can modulate the phenotype of steroid-secreting adrenocortical tumors, at least in response to pharmacolog- ical manipulations.

Acknowledgments

We are indebted to Prof. Y. Chapuis, who operated on the patients, and to Dr. A. Louvel for the pathological evaluation. We thank Mrs. M. Le Scouarnec for her expert secretarial assistance.

References

1. Spiegel AM. 1996 Mutations in G proteins and G protein-coupled receptors in endocrine disease. J Clin Endocrinol Metab. 81:2434-2442.

2. Dhanasekaran N, Heasley LE, Johnson GL. 1995 G protein-coupled re- ceptor systems involved in cell growth and oncogenesis. Endocr Rev. 16:259-270.

3. Mulligan LM, Kwok JBJ, Healey CS, et al. 1993 Germ-line mutations of the RET proto-oncogene in multiple endocrine neoplasia type 2A. Nature. 363:458-460.

4. Lacroix A, Bolte E, Tremblay J, et al. 1992 Gastric inhibitory polypeptide- dependent cortisol hypersecretion-a new cause of Cushing’s syndrome. N Engl J Med. 327:969-973.

5. Reznik Y, Allali-Zerah V, Chayvialle J, et al. 1992 Food dependent Cushing’s syndrome mediated by aberrant adrenal sensitivity to gastric inhibitory polypeptide. N Engl J Med. 327:981-986.

6. Lacroix A, Bolte E, Tremblay J, Rousseau G, Bouvier M, Hamet P. 1997 Propranolol therapy for ectopic B-adrenergic receptors in adrenal Cushing’s syndrome. N Engl J Med. 337:1429-1434.

7. Horiba N, Suda T, Aiba M, et al. 1995 Lysine-vasopressin stimulation of cortisol secretion in patients with adrenocorticotropin-independent ma- cronodular adrenal hyperplasia. J Clin Endocrinol Metab. 80:2336-2341.

8. Perraudin V, Delarue C, de Keyzer Y, et al. 1995 Vasopressin responsive adrenocortical tumor in a mild Cushing’s syndrome: in vivo and in vitro studies. J Clin Endocrinol Metab. 80:2661-2667.

9. Contreras P, Araya V. 1996 Overnight dexamethasone pre-treatment improves the performance of the lysine-vasopressin test in the diagnosis of Cushing’s syndrome. Clin Endocrinol (Oxf). 44:703-710.

10. Raux MC, Binoux M, Luton J-P, Gourmelen M, Girard F. 1975 Studies of ACTH secretion control in 116 cases of Cushing’s syndrome. J Clin Endocrinol Metab. 40:186-197.

11. Bertagna X, Raux-Demay MC, Guilhaume B, Girard F, Luton J-P. 1995 Cush- ing’s disease. In: Melmed S, ed. The pituitary. Oxford: Blackwell; 478-545.

12. Favrod-Coune C, Raux-Demay M, Proeschel M, Bertagna X, Girard F, Luton J-P. 1993 Potentiation of the classic ovine corticotrophin-releasing hormone stimulation test by the combined administration of small doses of lysine vasopressin. Clin Endocrinol (Oxf). 38:405-410.

13. Dickstein G, De Bold CR, Gaitan D, et al. 1996 Plasma corticotropin and cortisol responses to ovine corticotropin-releasing hormone (CRH), arginine- vasopressin (AVP), CRH plus AVP, and CRH plus metyrapone in patients with Cushing’s disease. J Clin Endocrinol Metab. 81:2934-2941.

14. Thibonnier M, Auzan C, Madhun Z, Wilkins P, Mattera-Berti L, Clauser E. 1994 Molecular cloning, sequencing, and functionnal expression of a cDNA encoding the human Vla vasopressin receptor. J Biol Chem. 269:3304-3310.

15. Birnbaumer M, Seibold A, Gilbert S, et al. 1992 Molecular cloning of the receptor of human antidiuretic hormone. Nature. 357:333-335.

16. de Keyzer Y, Auzan C, Lenne F, et al. 1994 Cloning and characterization of the human V3 pituitary vasopressin receptor. FEBS Lett. 356:215-220.

17. Sugimoto T, Saito M, Mochizuki S, Watanabe Y, Hashimoto S, Kawashima H. 1994 Molecular cloning and functional expression of a cDNA encoding the human V1b vasopressin receptor. J Biol Chem. 269:27088-27092.

18. Jard S. 1985 Vasopressin receptors. Front Horm Res. 13:89-104.

19. Balla T, Enyedi P, Spat A, Antoni F. 1985 Pressor-type vasopressin receptors in the adrenal cortex: properties of binding and effects on phosphoinositide metabolism and aldosterone metabolism. Endocrinology. 117:421-423.

20. Bird IM, Nicol M, Williams BC, Walker SW. 1990 Vasopressin stimulates cortisol secretion and phosphoinositide catabolismin cultured bovine adrenal fasciculata/reticularis cells. J Mol Endocrinol:109-116.

21. Guillon G, Trueba M, Joubert D, et al. 1995 Vasopressin stimulates steroid

secretion in human adrenal glands: comparison with angiotensin-II effect. Endocrinology. 136:1285-1295.

22. Perraudin V, Delarue C, Lefebvre H, Contesse V, Kuhn JM, Vaudry H. 1993 Vasopressin stimulates cortisol secretion from human adrenocortical tissue through activation of V1 receptors. J Clin Endocrinol Metab. 76:1522-1528.

23. Gicquel C, Bertagna X, Schneid H, et al. 1994 Rearrangements at 11p15 locus and overexpression of IGF2 in human adult adrenocortical tumors. J Clin Endocrinol Metab. 78:1444-1453.

24. Sibony M, Commo F, Callard P, Gasc JM. 1995 Enhancement of mRNA in situ hybridization signal by microwave signal. Lab Invest. 73:586-591.

25. Lefebvre H, Contesse V, Delarue C, et al. 1992 Serotonin-induced stimulation of cortisol secretion from human adrenocortical tissue is mediated through activation of a 5-HT4 receptor subtype. Neuroscience. 47:999-1007.

26. Laudat MH, Cerdas S, Fournier C, Guiban D, Guilhaume B, Luton J-P. 1988 Salivary cortisol measurement: a practical approach to assess pituitary-adrenal function. J Clin Endocrinol Metab. 66:343-348.

27. Katz MS, Kelly TM, Dax EM, Pineyro MA, Partilla JS, Gregerman RI. 1985 Ectopic ß-adrenergic receptors coupled to adenylate cyclase in human adre- nocortical carcinomas. J Clin Endocrinol Metab. 60:900-908.

28. Schorr I, Rathnam P, Saxena BB, Ney RL. 1971 Multiple specific hormone receptors in the adenylate cyclase of an adrenocortical carcinoma. J Biol Chem. 246:5806-5811.

29. Williams LT, Gore TB, Lefkowitz RJ. 1977 Ectopic ß-adrenergic receptor binding sites. J Clin Invest. 59:319-324.

30. Matsukura S, Kakita T, Sueoka S, et al. 1980 Multiple hormone receptors in the adenylate cyclase of human adrenocortical tumors. Cancer Res. 40:3768-3771.

31. Van Wijk PA, Rijnberk A, Croughs RJM, Wolfswinkel J, Selman PJ, Mol JA. 1994 Responsiveness to corticotropin-releasing hormone and vasopressin in canine Cushing’s syndrome. Eur J Endocrinol. 130:410-416.

32. Payet N, Deziel Y, Lehoux JG. 1984 Vasopressin: a potent growth factor in adrenal glomerulosa cells in culture. J Steroid Biochem. 20:449-454.

33. Milano CA, Allen LF, Rockman HA, et al. 1994 Enhanced myocardial function in transgenic mice overexpressing the ß2-adrenergic receptor. Science. 264:582-586.

34. Bond RA, Leff P, Johnson TD, et al. 1995 Physiological effects of inverse agonists in transgenic mice with myocardial overexpression of the ß2-adre- noreceptor. Nature. 374:272-276.