Illicit Upregulation of Serotonin Signaling Pathway in Adrenals of Patients With High Plasma or Intra-Adrenal ACTH Levels

Abstract

Context

In the human adrenal, serotonin (5-HT), released by mast cells stimulates corticosteroid secretion through activation of type 4 serotonin receptors (5-HT₄R). In primary pigmented nodular adrenocortical disease cells, activation of the cAMP/protein kinase A (PKA) pathway by PRKAR1A mutations triggers upregulation of the 5-HT synthesizing enzyme tryptophan hydroxylase (TPH) and the 5-HT₄, 5-HT₆, and 5-HT₇ receptors. Because ACTH stimulates cortisol secretion through activation of PKA, adrenocortical tissues exposed to sustained stimulation by ACTH may harbor increased expression of TPH and 5-HT_4/6/7 receptors.

Objective

To investigate the effects of long-term ACTH stimulation on the serotonergic pathway in adrenals of patients with high plasma or intra-adrenal ACTH levels.

Methods

Adrenal tissues were obtained from patients with Cushing disease, ectopic secretion of ACTH [paraneoplastic Cushing syndrome; (paraCS)], 21-hydroxylase deficiency (21-OHD), primary bilateral macronodular adrenal hyperplasia with intra-adrenal ACTH presence, or cortisol-producing adenomas. TPH and 5-HT_4/6/7 receptor expression was investigated using RT-PCR and immunochemistry in comparison with normal adrenals. Primary cultured adrenocortical cells originating from a patient with paraCS were incubated with 5-HT and 5-HTR agonists/antagonists.

Results

TPH and/or 5-HT_4/6/7 receptors were overexpressed in the different types of tissues. In paraCS cultured cells, the cortisol response to 5-HT was exaggerated compared with normal adrenal cells and the stimulatory action of 5-HT was reduced by 5-HT₄R antagonist.

Conclusion

Our results indicate that prolonged activation of the cAMP/PKA pathway by ACTH induces an aberrant serotonergic stimulatory loop in the adrenal cortex that likely participates in the pathogenesis of corticosteroid hypersecretion.

The serotonin signaling pathway is upregulated in adrenal tissues exposed to high levels of ACTH or harboring constitutive activation of the cAMP protein kinase A pathway.

ACTH released by pituitary corticotrophs, stimulates cortisol production by the adrenal cortex through binding to the type 2 melanocortin receptor (MC2R) and activation of the cAMP/protein kinase A (PKA) signaling pathway (1). Primary adrenal Cushing syndrome is caused by adrenocortical adenomas and hyperplasias that overproduce cortisol independently of pituitary ACTH, which is usually suppressed by cortisol excess. Interestingly, recent studies have shown that cortisol-producing neoplasias frequently display somatic or germline mutations that affect proteins of the cAMP/PKA pathway leading to constitutive activation of PKA. These mutations, which include gain-of-function mutations of the MC2R, GNAS, PRKACA, and PRKACB genes and inactivating mutations of the PRKAR1A, PDE11A and PDE8B genes (2–6), mimic the action of ACTH to stimulate glucocorticoid production, providing a molecular basis for the pathogenesis of primary adrenal Cushing syndrome. As a matter of interest, it has also been shown that cortisol secretion by adrenocortical adenomas and hyperplasias could be stimulated by both locally produced ACTH (7) and aberrantly expressed membrane receptors such as the serotonin (5-hydroxytryptamine; 5-HT) receptors types 4 (5-HT₄), 6 (5-HT₆), and 7 (5-HT₇) (8–11). 5-HT itself appears to be abnormally synthesized by a subpopulation of adrenocortical cells leading to formation of an illicit autocrine/paracrine regulatory loop that likely participates in the pathogenesis of hypercortisolism (10, 12). In particular, the occurrence of an illicit intra-adrenal serotonergic stimulatory mechanism has been described in primary pigmented nodular adrenocortical disease (PPNAD) caused by PRKAR1A mutation (10), suggesting that activation of PKA by the causative genetic defect may trigger upregulation of the 5-HT signaling pathway in neoplastic adrenocortical tissues. Consistent with this assumption, it was noted that inactivation of PRKAR1A expression in the human adrenocortical carcinoma cell line H295R results in overexpression of tryptophan hydroxylase (TPH), the key enzyme for 5-HT production, and the 5-HT₄, 5-HT₆, and 5-HT₇ receptors (10). In addition, TPH inhibitors were found to reduce in vitro cortisol production by PPNAD tissue explants (10), indicating that these compounds may represent a new therapeutic approach in the clinical management of PPNAD-associated hypercortisolism.

Taking into account all these observations, we have therefore hypothesized that in patients with adrenal disorders associated with high plasma ACTH concentration, overstimulation of the cAMP/PKA pathway caused by hyperactivation of the MC2-R could also lead to the appearance of an intra-adrenal 5-HT regulatory loop that may contribute to the pathogenesis of corticosteroid excess. We have thus investigated abnormal expression of the 5-HT–synthesizing enzyme TPH and 5-HT_4/6/7 receptors in adrenal tissues removed from patients with congenital adrenal hyperplasia linked to 21-hydroxylase enzyme deficiency (21-OHD) in whom plasma ACTH levels are elevated in response to cortisol deficiency (13) and patients with ACTH-dependent Cushing syndrome caused by pituitary corticotroph adenomas (Cushing disease; CD) or ectopic ACTH-secreting tumors (paraneoplastic Cushing syndrome; paraCS). Activation of the cAMP/PKA pathway also occurs in cortisol-secreting primary adrenal disorders that, similar to PPNAD, are associated with suppressed plasma ACTH levels. In fact, primary bilateral macronodular adrenal hyperplasias (PBMAH) that result from ARMC5 mutations in a substantial fraction of patients (14), have been shown to contain clusters of ACTH-producing cells that stimulate cortisol secretion in an autocrine/paracrine fashion through the MC2R (7). In addition, cortisol-producing adenomas (CPAs) globally harbor molecular features of PKA activation that result from somatic mutations positively affecting the cAMP/PKA pathway (2, 3, 15). We have thus examined expression of the 5-HT signaling pathway in PBMAH samples and CPAs. Last, we have studied in vitro the effect of 5-HT on cortisol production by adrenocortical cells from one patient with ACTH-dependent Cushing syndrome.

Patients and Methods

Tissue collection

Patients were recruited from seven French centers (Rouen, Caen, Cochin, Kremlin-Bicêtre, La Pitié-Salpêtrière, Grenoble, Lyon) and from the National Institutes of Health (Bethesda, MD). Adrenal tissues were obtained from 5 patients with congenital adrenal hyperplasia caused by genetically proven 21-OHD, 13 patients with CD caused by ACTH-secreting pituitary adenoma, 3 patients with paraCS, 5 patients with PBMAH, and 10 patients with CPA. Clinical and biological data from each patient are summarized in Tables 1 to 3. Adrenal glands were collected after surgery and immediately dissected by the pathologists. For 2 patients with 21-OHD, 6 patients with CD, and 10 patients with adrenocortical adenoma, adrenal explants were frozen and kept at −80°C before RNA extraction for real-time RT-PCR analysis. For 1 of the patients with paraCS and 2 patients with CPA, adrenal fragments were immersed in culture medium for functional studies. For all patients, tissues were fixed in formalin for histological study. Nine normal adrenals (control tissues) were obtained from patients undergoing expanded nephrectomy for kidney cancer. The protocol of collection of the tissues and the experimental procedures were approved by the ethic committees, and informed consent was obtained from all subjects. Human adrenal polyA⁺ RNA were also used as control tissue (Takara, Saint-Germain en Laye, France).

Table 1.

Clinical and Biological Data of Patients With 21-OHD Prior to Adrenalectomy

		21-OHD
Patient		1	2	3	4	5
Sex		F	M	F	M	M
Age, y		7	48	34	43	39
Clinical presentation		SV	SV	SW	SW	SW
Plasma ACTH 0800h, pg/mL, n = (10–48)		13	29	17	914	666
Plasma 17-OHP, ng/mL		128 (N < 1)	72 (N < 2)	118 (N < 5.0)	167 (N < 2.8)	207 (N < 2.3)
Plasma Δ4-androstenedione, ng/mL		3.4 (N < 0.45)	nd	9.7 (N < 3.1)	nd	57 (N < 1.7)
Plasma testosterone, ng/mL		2.41 (N < 0.3)	2.18 (N = 3–9)	2.1 (N < 0.48)	5.2 (N = 2–10)	12.9 (N = 3–9)
Plasma renin, ng/L		12 (N:7–40)	101 (N = 7–40)	35 (N = 5–28)	93.2 (N = 2–59)	54.5 (N = 3–33)
Treatment	Hydrocortisone (mg/d)	20		30	25	15
	Prednisone (mg/d)					5
	Fludrocortisone (µg/d)	50		50	100
Indication of adrenalectomy		Inability to control hyperandrogenism	Pseudotumoral enlargement of adrenal glands	Inability to control hyperandrogenism	Pseudotumoral enlargement of adrenal glands	Inability to control hyperandrogenism

nd, not done; SV, simple virilization; SW, salt wasting.

Table 2.

Clinical, Biological, and Radiological Data of Patients With CD and paraCS

	Cushing Disease													paraCS
Patient	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21
Sex	M	F	F	M	F	F	F	F	H	F	F	F	F	M	M	M
Age, y	79	54	41	64	50	67	66	55	59	41	37	16	47	25	18	66
Plasma ACTH 0800h, pg/mL, (N = 7-63)	100	25	91	192	141	13.6	20.4	59	24	132	100	40	92	310	73	549
Plasma cortisol 0800h, μg/L, (N = 90-308)	270	278	282	769	224	225	87.7	233	188	121	280	220	180	460	844	735
Plasma cortisol 2000h or late night, μg/L	272	263	147	789	199	187	132	125	184	181	120	180	160	250	757	656
Urinary cortisol, μg/d, (N < 80)	519	1246	154	15,406	149	115	134	116	134	153	300	336	880	1390	18567	6144
Maximal size of adrenal glands at CT scan, mm
Left	29	43	24	36	25	51	37	22	34	60	15	20	norm	norm		62
Right	30	38	35	33	30	34	36	30	20	51	15	18	norm	norm		55
Plasma ACTH response to CRH testing, % from basal level	79	55	nd	nd	+200	+150	+180	nd	+300	+30	120	nd	nd	0	103	0
Plasma cortisol response to CRH testing, % from basal level	29	55	nd	nd	+59	+52	+95	nd	+60	+12	43	nd	nd	0	102	0
Plasma cortisol level after low-dose dexamethasone suppression test, μg/L	229	253	62	nd	29	43	84	63	134	nd	150	180	210	nd	nd	679
Cortisol response to high dose dexamethasone suppression test, % from basal level	nd	−42	nd	nd	−10	nd	−31	−15	−17	−6	−28	−9	−50	−24	nd	−67
Preoperative treatment of hypercortisolism
Mifepristone	X
Ketoconazole				X										X	X
Metyrapone		X													X
Pasireotide			X
Cabergoline													X
Mitotane											X

Abbreviations: nd, not done; norm, normal size.

ACTH-producing neoplasia: P6-P18, pituitary; P19, bronchial carcinoid; P20, pancreatic carcinoma; P21, small cell lung cancer.

Table 3.

Clinical, Biological, Genetic, and Radiological Data of Patients With PBMAH and CPA

	PBMAH					CPA
Patient	22	23	24	25	26	27	28	29	30	31	32	33	34	35	36
Sex	F	M	M	F	F	F	F	F	F	F	F	F	F	F	F
Age, y	66	46	73	50	44	66	33	52	62	29	41	42	55	42	32
Plasma ACTH 0800h, pg/ml, N = 7–63	<5	<5	<5	9	<3	<5	<5	<5	<5	<5	6.2	<5	<5	<5	2
Plasma cortisol 0800h, μg/L, N = 90–308	250	292	224	248	40	134	180	119	166	188	65	68	163	147	129
Plasma cortisol 2000h or late night, μg/L	189	275	188	84	100	147	196	134	123	213	87	69	151		146
Urinary cortisol, μg/d, N < 80	178	300	159	39	283	150	283	31	99	371	44	9.7	136	225	170
Maximal size of adrenal glands at CT scan, mm
Left	37	35	80	80	65		22	21	34	30		25	35		30
Right	32	25	90	60	32	20								22
Plasma ACTH response to CRH testing, % from basal level	na	na	na	nd	na	na	0	0	nd	23	112	372	0	na	nd
Plasma cortisol response to CRH testing, % from basal level	na	na	na	nd	na	na	0	0	nd	−3	+38	43	−7	na	nd
Plasma cortisol level after low-dose dexamethasone suppression test, μg/L	na	na	na	nd	42	130	na	na	na	208	85	82	172	176	135
Cortisol response to high-dose dexamethasone suppression test, % from basal level	+8	0	0	+70	na	na	+31	+34	+4	+2	+6.1	+9.1	+16	na	nd
Germline ARMC5 mutation	no	nd	nd	yes	no	nd
Somatic PRKACA mutation	nd					no	no	no	no	yes	yes	yes	yes	nd	nd
Preoperative treatment of hypercortisolism
Ketoconazole	X	X	X
Metopirone
Mitotane		X

Abbreviations: nd, not done; na, not applicable.

RT-PCR

Total RNAs were extracted from adrenal samples derived from patients 1, 2, 10 to 15, and 28 to 33, by using Tri-Reagent (Sigma-Aldrich, Saint-Quentin Fallavier, France) and purified with the Nucleospin RNA II isolation kit (Macherey-Nagel, Hoerdt, France). The concentration of total RNAs was measured on Nanodrop 2000c (ThermoScientific, Villebon-Sur-Yvette, France). Purified RNAs and polyA⁺ RNAs were converted into cDNA by using ImProm-II RT kit (Promega, Charbonnières-les-Bains, France). Real-time PCR amplifications were performed using SYBR Green PCR Master Mix (Applied Biosystems, Villebon-Sur-Yvette, France) in a QuantStudio 3 Real-Time PCR System (Applied Biosystems) with specific primers (10). Samples were analyzed in duplicate in two different experiments. Quantification of cDNAs was normalized to cyclophilin (PPIA) gene expression. Results were analyzed using 2^-ΔΔCt method.

Immunohistochemistry

Sections from formalin-fixed paraffin-embedded adrenals were first deparaffinized and rehydrated. Sections underwent antigen retrieval treatment by heating at 95°C in 10 mM citrate buffer (pH = 6; 20 minutes) or Tris-EDTA SDS 0.05% (pH = 9; 20 minutes). They were treated with a peroxidase blocking reagent (Dako Corporation, Les Ulis, France). The tissue sections were incubated for 60 minutes with primary antibodies against 5-HT (AHP522, Serotec, 1/40), 5-HT₄ (LS-A2685, LS-BIO, 25 µg/mL; bs-12054R, Interchim, 1/250), 5-HT₆ (ab101911, Abcam, 1/1200), and 5-HT₇ (S0320, Sigma-Aldrich, 1/100) receptors, TPH1 (HPA022483, Sigma-Aldrich, 1/50), DAB2 (sc-13982, Santa Cruz, 1/200), ACTH (AP15336PU, Acris, 1/100), CYP11B1 (1/200), and CYP11B2 (1/500) provided by Dr. Celso Gomez-Sanchez (16). They were then incubated with antirabbit immunoglobulins streptavidin-biotin complex coupled to peroxidase (Dako Corporation). Immunoreactivities were revealed with diaminobenzidine (Dako Corporation). Sections were finally counterstained with hematoxylin, mounted in Eukitt (Kindler GmbH and Co., Freibourg, Germany), cover-slipped, and examined on an Eclipse E600D microscope (Nikon, Les Ulis, France) on PRIMACEN, the Cell Imaging Platform of Normandy (Rouen University). A histological score was used to evaluate the staining intensity (0: no staining; 1: weak; 2: moderate; 3: strong staining), with moderate and strong staining being considered as positive.

Cell culture

Adrenals were sliced and stirred for 45 minutes at 37°C in cell culture medium containing collagenase type 1A (60 mg/mL; Sigma-Aldrich) and desoxyribonuclease 1 type 4 (4 mg/mL; Sigma-Aldrich). Dispersed cells were suspended in 50% DMEM and 50% Ham-F12 (Life Technologies, Marly-le-Roi, France) supplemented with 5% fetal bovine serum (Sigma-Aldrich), 1% antibiotic-antimycotic solution (Life Technologies), and 1% insulin-transferrin-selenium solution (BD Biosciences, San Jose, CA). Cells were plated on 24-well culture dishes. Adrenocortical cells were cultured at 37°C in a 5% carbon dioxide-95% air atmosphere with 100% relative humidity. The cell culture medium was changed 24 hours after plating. Cultured cells obtained from patient 20 with paraCS were treated either with ACTH 1-24 (Synacthen, Novartis Pharma, France), 5-HT (Sigma-Aldrich), BIMU-8 (Sigma-Aldrich), AS-19 (Lille, France) alone or in combination with GR113808, SB258585 or SB269970 (Sigma-Aldrich).

Adrenal explants from two patients with CPA (patients 35 and 36) were cut into fragments of equal size and weight (∼100 mg). The tissue fragments were successively incubated with DMEM for three hours, and then with DMEM alone or DMEM added with 4-chloro-dl-phenylalanine (10⁻⁵ M) for three hours. All culture media were collected for cortisol level measurement.

Radioimmunoassays

Cortisol and dehydroepiandrosterone (DHEA) concentrations in cell culture medium were measured by radioimmunoassay procedures using specific antibodies [(17) and Sigma-Aldrich, respectively]. As previously described (17), the sensitivity of the cortisol assay was 15 pg. Cross-reactivity of antibodies with test compounds was <0.01%.

Statistical analysis

Steroid secretion and mRNA levels are expressed as mean ± SEM. Statistical analysis were performed by using a Mann-Whitney test to compare the mRNA levels measured in control and pathological tissues, respectively (Prism 6, GraphPad Software). No statistical analysis of the data obtained in 21-OHD tissues was possible because of the low number of samples included in this group.

Results

The 5-HT signaling pathway in adrenal tissues exposed to high circulating levels of ACTH

5-HT synthesis in adrenal tissues

TPH1 mRNA is detected in the normal human adrenal at variable levels and appears overexpressed in adrenal samples removed from patients with 21-OHD (4.3 ± 0.9) or CD (3.8 ± 0.4 vs 1.2 ± 0.3 in normal adrenal; P = 0.0008), whereas TPH2 mRNA was not detectable (data not shown). As previously reported, TPH1 is exclusively detected in mast cells localized in zona glomerulosa (ZG) from normal adrenals (NA; Fig. 1) (10). In adrenals of three of the five patients with 21-OHD, a diffuse and homogeneous TPH1 immunoreactivity was observed in the subcapsular region of the cortex (Fig. 1). In addition, the cortex displayed histological disorganization. In particular, small TPH1-positive compact cells were densely packed and organized as hairpin-shaped cords in the outer layer, whereas no clusters of cells arranged as glomeruli could be seen under the capsule. TPH1 immunostaining was also visualized in spongiocytic adrenocortical cells structured in cords in the inner cortex, corresponding to zona fasciculata (ZF), and/or located in hyperplastic nodules in adrenals of the two remaining patients with 21-OHD (data not shown).

Figure 1.

Localization of TPH1 and 5-HT in the adrenal tissues in respect to cortex zonation. Representative microphotographs showing the distributions of TPH1, 5-HT, DAB2, aldosterone synthase (CYP11B2), and 11β-hydroxylase (CYP11B1) immunoreactivities in NA, 21-OHD (patient 1), and CD (patient 6) samples. Arrows indicate either TPH1 or 5-HT immunoreactive cells in NA. Ca, capsule; ZF, zona fasciculata.

In adrenals from patients with CD, TPH1 immunoreactivity was detected in adrenocortical cells in the outer and inner regions of the cortex for 11 of 13 patients (Fig. 1). Groups of TPH1-positive cells organized as rosette structures were visualized under the capsule. Only 1 patient with CD exhibited no adrenal TPH1 staining. TPH1 immunoreactivity was also visualized in the medulla for 9 patients (data not shown).

As expected, TPH1-positive cells were also labeled by 5-HT antibodies in NA, 21-OHD, and CD adrenal tissues. Conversely, no TPH1 and 5-HT labeling could be detected in adrenals from patients with paraCS (data not shown).

Further immunohistochemical experiments were performed with ZG and ZF specific markers to analyze the distribution of TPH1/5-HT labelings according to the histological and functional zonation of the adrenal cortex in patients with 21-OHD and CD. The large subcapsular region intensely positive for TPH1 or 5-HT in cortex of patients with 21-OHD was also immunoreactive to disabled homolog 2 (DAB2), a ZG marker, and aldosterone synthase (encoded by CYP11B2), a crucial enzyme for aldosterone synthesis (Fig. 1). As expected, TPH1/5-HT immunoreactive cell cords in ZF were positive for 11β-hydroxylase (encoded by CYP11B1), an enzyme involved in cortisol synthesis. In adrenals of patients with CD, DAB2 labeling appeared discontinuous and only rare groups of CYP11B2-positive cells, that resemble the aldosterone-producing cell clusters described in the normal adrenal, were detected in the subcapsular region of the cortex. There was no clear overlap between TPH1/5-HT and DAB2 or CYP11B2 stainings. In addition, CYP11B1-immunoreactive cells were largely distributed throughout the outer cortex, except in APCCs and in ZF. Globally, these data indicate that, depending on the types of tissues, 5-HT may be released by aldosterone- and/or cortisol-producing cells.

Expression of the 5-HT_4/6/7 receptors in adrenal tissues

As previously shown, NA tissues were found to express HTR4 mRNA and 5-HT₄R immunoreactivity was intensely detected in the ZG and only weakly observed in the ZF. In addition, the 5-HT₄R was visualized in the zona reticularis (ZR) and 5-HT (10⁻⁹ to 10⁻⁶ M) stimulated DHEA production by cultured normal adrenocortical cells in a dose-dependent manner (Fig. 2A–2D). However, the efficacy of 5-HT to activate DHEA (E_max = 133%) production was lower than those formerly observed for aldosterone and cortisol (18, 19). Consistent with the previous detection of HTR4 mRNA in normal adrenal chromaffin cells (20), intense 5-HT₄R immunoreactivity was also observed in the adrenal medulla (Fig. 2D).

Figure 2.

Expression of serotonin receptor type 4 in the adrenal tissues. (A) Expression levels of HTR4 mRNA in NA, 21-OHD, and CD tissues. (B) Effect of increasing concentrations of 5-HT on aldosterone (■), cortisol (●), and DHEA (○) secretion by cultured normal adrenocortical cells. (C–I) Distribution of 5-HT₄ immunoreactivities in NA (C, D), 21-OHD (E, F; patient 2), CD (G, H; patient 7), and paraCS (I; patient 20) samples. Immunostaining is principally visualized in spongiocytic cells both at the cytoplasmic and membrane levels. Infiltration of the paraCS tissue by metastatic cells (meta; I) is seen on the right side of the dotted line. (J) ACTH immunoreactivity in metastatic cells invading the adrenocortical tissue from patient 20. Ca, capsule.

Alternatively, HTR4 mRNAs were detected at high levels in the 21-OHD tissues from patients 1 and 2. Strong 5-HT₄R immunostaining could be visualized in the inner cortex from 4 of 5 patients with 21-OHD, whereas ZG labeling was weak and limited to some adrenocortical cells distributed throughout the subcapsular region (Fig. 2E and 2F). HTR4 mRNA levels were similar in CD and NA samples (Fig. 2A). However, at variance with NA, CD adrenal tissues exhibited 5-HT₄R immunostaining in limited groups of cells discontinuously distributed under the capsule, contrasting with intense immunoreactivity observed in some cell cords in ZF (Fig. 2G and 2H). In paraCS tissues, 5-HT₄R immunopositive adrenocortical cells were principally observed in the vicinity of clusters of metastatic cells that were shown to produce ACTH (Fig. 2I and 2J).

As previously published (10), HTR6 mRNA was not detectable and HTR7 mRNA was weakly expressed in NA tissues (Fig. 3A and Fig. 4A). Moreover, neither HT₆R nor 5-HT₇R immunoreactivity could be observed in steroidogenic cells of NA tissues (Fig. 3B and Fig. 4B). On the contrary, HTR6 and HTR7 mRNAs were found in 21-OHD adrenocortical tissues. In addition, 5-HT₆R (3/5 patients) and 5-HT₇R (4/5 patients) immunostainings were visualized in clusters of spongiocytic cells in the subcapsular region and inner cortex (Fig. 3C–3D and Fig. 4C–4D). Similarly, HTR6 and HTR7 mRNA were overexpressed in adrenals from patients with CD. Immunopositive cells for 5-HT₆R (8/13 patients) and-5-HT₇R (10/13 patients) were organized in clusters or cords in the ZF (Fig. 3E–3F and Fig. 4E). In the remaining three patients with CD, 5-HT₇R immunolabeling was exclusively detected in the outer cortex (Fig. 4F). In the adrenal tissue from patients with paraCS, both 5-HT₆R and 5-HT₇R were present in patients 19 and 20 (Fig. 3G–3H and Fig. 4G–4H).

Figure 3.

Expression of serotonergic receptor type 6 in the adrenal tissues. (A) Expression levels of HTR6 mRNA in 21-OHD and CD tissues. (B) Absence of 5-HT₆R immunoreactivity in NA. (C–H) Distribution of 5-HT₆R immunoreactivities in 21-OHD (C, D; patient 2), CD (E, F; patient 15), and paraCS (G, patient 19; H, patient 20) samples. Ca, capsule.

Figure 4.

Expression of serotonergic receptor type 7 in the adrenal tissues. (A) Expression levels of HTR7 mRNA in NA, 21-OHD, and CD tissues. (B–H) Distribution of 5-HT₇R immunoreactivities in NA (B), 21-OHD (C, D; patient 2), CD (E, patient 16; F, patient 11), and paraCS (G, patient 19; H, patient 20) samples. Infiltration of the paraCS tissue by metastatic cells (H) is seen on the right side of the dotted line. **P < 0.01. Ca, capsule; meta, metastatis.

Effect of 5-HT on cortisol production from a paraCS tissue

To evaluate whether activation of 5-HT receptors could regulate steroidogenesis, we have examined the effects of 5-HT and diverse synthetic ligands of the 5-HT_4/6/7 receptors on cortisol secretion by adrenocortical cells removed from patient 20 with paraCS. ACTH (10⁻¹⁰ M), which was used as a positive control, induced a +644% increase in cortisol production. 5-HT (10⁻⁹ to 10⁻⁵ M) dose-dependently stimulated cortisol secretion by cultured adrenocortical cells (Fig. 5) with higher efficacy (E_max = 264%) and potency (EC₅₀ = 1.4 nM) than in normal adrenocortical cells (10, 18). The 5-HT4R agonist BIMU-8 and the 5-HT7R agonist AS-19 induced modest increases in cortisol release, reaching 145% and 127%, respectively, with potencies of 4.7 nM and 11.7 nM (Fig. 5A). The potency of 5-HT was not influenced by the 5-HT₇R antagonist SB-269970 (10⁻⁷) but was weakly reduced by the 5-HT₆R antagonist SB258585 (10⁻⁷) and highly decreased by the 5-HT₄R antagonist GR-113808 (10⁻⁷) (Fig. 5B).

Figure 5.

Effect of 5-HT and 5-HT receptor ligands on cortisol production from a paraCS adrenal. (A) Effects of increasing concentrations of 5-HT (●), the 5-HT₄R agonist BIMU8 (○) and the 5-HT₇R agonist AS-19 (□) on cortisol secretion by cultured adrenocortical cells derived from patient 20 (left). Right, cortisol response to ACTH (10⁻¹⁰ M, control). (B) Potencies (EC₅₀) of 5-HT on cortisol secretion by cultured adrenocortical cells derived from normal and paraCS adrenals. 5-HT was administrated alone or in combination with the 5-HT₄R antagonists GR113808 (10⁻⁷ M), the 5-HT₆R antagonist SB258585 (10⁻⁷ M), or the 5-HT₇R antagonist SB269970 (10⁻⁷ M). B, basal level.

The 5-HT signaling pathway in PBMAH and its relation to ACTH-producing cells

To examine the impact of intra-adrenal ACTH on the occurrence of 5-HT signaling pathway in PBMAH samples, consecutive sections of adrenal samples from five patients with PBMAH were labeled with ACTH antibodies and antibodies to TPH1 and the 5-HT_4/6/7 receptors. As previously reported (7), clusters of ACTH-producing cells were observed in all samples (Fig. 6A and 6B). TPH1 and 5-HT₄R immunoreactivities were essentially detected in regions enriched in ACTH-positive cells (Fig. 6C–6F). The distribution of the 5-HT₆ and 5-HT₇ receptors was more diffuse and apparently not correlated with the presence of ACTH-positive cells (Fig. 6G and 6H).

Figure 6.

Immunohistochemical detection of ACTH and the 5-HT signaling pathway in PBMAH tissues. Representative microphotographs showing the distribution of ACTH (A, B; patient 22), TPH1 (C, D; patient 22), 5-HT₄R (E, F; patient 22), 5-HT₆R (G; patient 22) and 5-HT₇R (H, patient 24) stainings in PBMAH tissues.

The 5-HT signaling pathway in cortisol-producing adrenocortical adenomas

To investigate further the influence of activation of the cAMP/PKA pathway on adrenal 5-HT signaling, we have studied expression of TPH1 and the 5-HT_4/6/7 receptors in a series of 10 CPAs, a type of adrenocortical tumors known to display diverse genetic alterations leading to constitutive activation of the PKA pathway (3, 6, 15). The levels of TPH1 and HTR4 mRNAs were similar in NA and CPA samples (Fig. 7). Conversely, expression of the HTR6 and HTR7 mRNAs was increased in CPAs in comparison with NA. In addition, CPAs were found to intensely express TPH1 and the three types of receptors with variable distributions throughout the tissues.

Figure 7.

Expression of tryptophan hydroxylase type 1 and 5-HT₄, 5-HT₆, and 5-HT₇ receptors in CPA tissues. Expression levels of (A) TPH1,(B) HTR4, (C) HTR6, and (D) HTR7 mRNA in NA and CPA tissues. (E–L) Representative microphotographs showing the distributions of TPH1 (E, patient 28; F, patient 33) and 5-HT₄ (G, patient 27; H, patient 31), 5-HT₆ (I, patient 28; J, patient 31) and 5-HT₇ (K, patient 27; L, patient 31) receptors in adrenocortisol-producing adenomas. Ca: capsule. M, effect of the TPH inhibitor PCPA (10⁻⁵ M) on cortisol secretion by CPA explants. Mean (±SEM) cortisol production by four distinct explants from patients 35 and 36 incubated without (DMEM) or with PCPA. *P < 0.05; ***P < 0.001.

Role of 5-HT on cortisol production in cortisol-producing adenomas

To explore whether intra-adrenal 5-HT could control steroidogenesis, we have measured cortisol levels in the incubation medium of CPA explants exposed to the TPH inhibitor p-chlorophenylalanine (PCPA). Incubation of CPA fragments with PCPA (10⁻⁵ M) decreased cortisol production by 44% (Fig. 7M).

Discussion

In the normal human adrenal gland, 5-HT is exclusively synthesized and produced by subcapsular mast cells and, after its release, mainly stimulates aldosterone secretion through a paracrine mechanism involving the 5-HT₄ receptor (21, 22). Conversely, in PPNAD related to PRKAR1A mutations, activation of the cAMP/PKA pathway triggers overexpression of TPH2 and the 5-HT_4/6/7 receptors in adrenocortical cells, leading to formation of an intra-adrenal serotonergic stimulatory loop that is involved in cortisol hypersecretion (10). We have thus examined whether this paracrine process could be observed in adrenal tissues exposed to persistently high plasma ACTH levels.

In 21-OHD and CD adrenals, TPH1 was abnormally expressed in steroidogenic cells indicating that persistently high plasma ACTH levels upregulate expression of the enzyme. In this respect, it is interesting to note that activation of the hypothalamo-pituitary corticotropic axis induced by chronic stress provokes expression of THP1 in the rat adrenal cortex (23). At variance with PPNAD samples, TPH2 mRNA was undetectable in 21-OHD and CD adrenals. It is possible that preferential expression of TPH2 rather than TPH1 in PPNAD may be because of neuroendocrine differentiation of PPNAD cells, as published elsewhere (24). The distribution patterns of TPH1 within the cortex were different between 21-OHD and CD adrenals. In fact, immunoreactivity for TPH1 was intense and extensively spread throughout the subcapsular region in 21-OHD adrenals but appeared more diffusely disseminated in CD adrenal cortices. These observations suggest that, in addition to ACTH, angiotensin II, the plasma levels of which are elevated in response to mineralocorticoid deficiency in 21-OHD patients, may also favor ectopic production of TPH1 in ZG cells, as shown by the overlap between TPH1 and DAB2/CYP11B2 labelings.

The observation that TPH1 is absent in the adrenals removed from patients with paraCS who are exposed to high concentrations of ACTH (resulting both from markedly elevated plasma levels and intra-adrenal production of the hormone in two cases), could suggest that the stimulatory action of ACTH on TPH1 expression is attenuated at greater concentration values. However, this possibility needs to be confirmed in a larger series of tissues.

Illicit expression of TPH1 in the adrenal cortex indicates that adrenocortical cells may be able to synthesize 5-HT because l-amino acid decarboxylase, which allows conversion of 5-hydroxytryptophan to 5-HT, is known to be expressed in the adrenal gland as in many organs (25). As expected, we could observe the presence of detectable amounts of 5-HT in TPH1-positive cells. It cannot be excluded that 5-HT contained in adrenocortical cells may also partly originate from the plasma through an uptake process as the 5-HT transporter SERT has been shown to be expressed in the adrenal cortex (10).

Ectopic synthesis of 5-HT in 21-OHD and CD samples suggests that 5-HT may stimulate corticosteroid secretion by a paracrine mode of action. We have thus investigated expression of 5-HT receptors in the tissues by focusing on the types 4, 6, and 7 that had previously been shown to be upregulated in PPNAD. Globally, expression of the 5-HT₄R was increased in 21-OHD and CD tissues, as well as paraCS samples. The level of 5-HT₄R mRNA in CD was similar to that of the NA but immunohistochemistry revealed that, at variance with control tissues, 5-HT₄R immunoreactivity is markedly decreased in the ZG but abnormally detected in the ZF in CD adrenals. The fact that 5-HT₄R labeling was particularly intense in adrenocortical cells surrounding metastatic ACTH-producing cells in paraCS tissues strongly supports a causative role of corticotropin in 5-HT₄R upregulation. The 5-HT₆ and 5-HT₇ receptors, which are not physiologically present in the human cortex, were also found to be expressed at variable levels in the tissues with the exception of the 5-HT₆R in the one paraCS samples. Activation of PKA by ACTH in adrenocortical cells may thus also trigger 5-HT₆ and 5-HT₇ receptor expression, as shown in PPNAD. In addition, we have examined the 5-HT signaling pathway in PBMAH tissues which are known to contain islets of ACTH-producing cells (7) and CPAs that globally harbor a molecular signature of PKA activation (3, 6, 15). Our data indicate that in PBMAH, high expression of TPH1 and the 5-HT₄R is associated with the presence of ACTH-positive cells indicating that ACTH may upregulate these two proteins through an autocrine/paracrine process. This hypothesis is supported by previous findings showing that in PBMAH, ACTH-positive cells and the surrounding adrenocortical cells express the MC2 receptor (7). In addition, ACTH has been shown to stimulate its own synthesis in these cells (26). Conversely, expression of the 5-HT₆ and 5-HT₇ receptors in the PBMAH samples appears less dependent on local synthesis of ACTH. It is possible that upregulation of HTR6 may be favored by hypersecretion of cortisol, as suggested by the identification of glucocorticoid responsive elements in the promoter of the gene (27). On the other hand, overexpression of 5-HT₇R may be rather regarded as an additional trait of pseudogonadal differentiation of the adrenal cortex observed in PBMAH tissues (7). Indeed, at variance with the adrenal gland, gonads have been shown to physiologically express the 5-HT₇R (28, 29). We observed that the 5-HT signaling pathway is upregulated in CPAs in both PRKACA-mutated and PRKACA nonmutated tumors, a finding that could suggest that in some tumors, mechanisms other than activation of PKA could be responsible for overexpression of TPH1 and 5-HT_4/6/7 receptors. However, it should also be noticed that genomic studies have shown that activation of the cAMP/PKA pathway is a hallmark of CPAs and can result from various somatic mutations affecting not only PRKACA but also several genes encoding diverse proteins of the pathway such as GNAS and CREB1 (4, 15).

Upregulation of the 5-HT signaling pathway may have important potential implications in the pathogenesis of steroid hypersecretion. In fact, the 5-HT_4/6/7 receptors are positively coupled to the cAMP/PKA pathway, which is a major activator of corticosteroid synthesis in the ZF and ZR (30). As a matter of fact, we observed that cortisol production from a paraCS adrenal tissue was strongly stimulated by 5-HT in vitro principally through activation of 5-HT₄R. In addition, the 5-HT₆R may participate in adrenocortical tumor expansion through its positive coupling to the mTOR pathway (31) the activation of which is known to decrease apoptosis in adrenocortical cells (32). Most importantly, activation of the 5-HT signaling pathway may contribute to the pathogenesis of adrenal androgen hypersecretion in 21-OHD. In fact, we show that 5-HT is able to stimulate DHEA from cultured normal adrenocortical cells presumably through the eutopic 5-HT₄R that is detected in the ZR through immunohistochemistry in agreement with previous studies using binding of synthetic radioligands on adrenal slices (33). In addition, androgen levels are difficult to decrease in some patients with 21-OHD even by administrating high doses of glucocorticoid. In these cases, the resistance of androgen secretion to ACTH suppression is likely indicative of the involvement of an intra-adrenal stimulatory process similar to the serotonergic pathway. Because the pathological adrenal tissues abnormally express different combinations of 5-HT receptors, antagonizing one single type of receptors may probably only induce partial suppression of steroid excess. It seems thus more promising, in a therapeutic perspective, to target adrenal 5-HT synthesis via administration of peripheral TPH inhibitors, which are now used in the treatment of carcinoid syndrome (34, 35). In support of this strategy, our data show that PCPA decreases cortisol production by CPA explants in vitro.

Collectively, our data obtained from adrenal tissues exposed to high levels of ACTH suggest that corticotropin is able to induce the 5-HT signaling pathway in the adrenal cortex. We cannot rule out the hypothesis that the medications received by the patients prior to adrenal surgery may have influenced upregulation of the 5-HT signaling pathway. However, occurrence of an illicit serotonergic regulatory loop was observed in both treated and nontreated patients. In addition, our findings are consistent with the data obtained in rats, in which stress not only induces activation of the hypothalamo-pituitary corticotropic axis but also triggers abnormal synthesis of 5-HT and overexpression of 5-HT receptors in the adrenal cortex (23). This mechanism is probably aimed at potentiating the glucocorticoid response to stress through recruitment of intra-adrenal stimulatory signals whose action adds to that of circulating ACTH. The serotonergic pathway has also been shown to represent an inducible paracrine system in other organs such as pancreatic islets. In mice, the increase in β-pancreatic cell mass observed during pregnancy involves upregulation of TPH and 5-HT receptors in response to prolactin and placental lactogen stimulation to enhance insulin secretion and maintain glucose homeostasis (36). The adrenal gland may thus constitute another example of the importance of 5-HT as a paracrine factor in peripheral organs that also include bone, placenta, the mammary gland, and the adipose tissue (37–40). Finally, our results suggest that inhibition of adrenal 5-HT synthesis via administration of peripheral TPH inhibitors may represent an original treatment of corticosteroid hypersecretion as an alternative to adrenal surgery, especially in female patients with 21-OHD and persistently high androgen levels despite glucocorticoid suppressive treatment or patients with CPA. Overexpression of 5-HT receptors in CD adrenals also indicates that 5-HT reuptake inhibitors, which may be used to treat depressive syndrome associated with hypercortisolism, should be avoided in this condition.

Glossary

Abbreviations:

5-HT

serotonin

5-HT₄R

type 4 serotonin receptors

21-OHD

21-hydroxylase deficiency

CD

Cushing disease

CPA

cortisol-producing adenoma

DHEA

dehydroepiandrosterone

MC2R

type 2 melanocortin receptor

NA

normal adrenal

paraCS

paraneoplastic Cushing syndrome

PBMAH

primary bilateral macronodular adrenal hyperplasias

PCPA

p-chlorophenylalanine

PKA

protein kinase A

PPNAD

primary pigmented nodular adrenocortical disease

TPH

tryptophan hydroxylase

ZG

zona glomerulosa

ZF

zona fasciculata

ZR

zona reticularis