Liquid Chromatography–Tandem Mass Spectrometry Analysis of Human Adrenal Vein 19-Carbon Steroids Before and After ACTH Stimulation

Abstract

Context:

A broad analysis of adrenal gland-derived 19-carbon (C₁₉) steroids has not been reported. This is the first study that uses liquid chromatography–tandem mass spectrometry to quantify 9 C₁₉ steroids (androgens and their precursors), estrone, and estradiol in the adrenal vein (AV) of women, before and after ACTH stimulation.

Objective:

The objective of this study was to define the adrenal androgen metabolome in women before and after ACTH infusion.

Design:

This was a retrospective study.

Patients:

Seven women, aged 50.4 ± 5.4 years, with suspected diagnosis of an adrenal aldosterone-producing adenoma were included in the study.

Methods:

AV and iliac serum samples were collected before and after administration of ACTH (15 minutes). AV samples were analyzed using for concentrations of 9 unconjugated C₁₉ steroids, estrone, and estradiol. Dehydroepiandrosterone sulfate (DHEA-S) was quantified by radioimmunoassay.

Results:

AV levels of DHEA-S were the highest among the steroids measured. The most abundant unconjugated C₁₉ steroids in AV were 11β-hydroxyandrostenedione (11OHA), dehydroepiandrosterone (DHEA), and androstenedione (A4). ACTH significantly increased the adrenal output of 9 of the 12 steroids that were measured. ACTH increased the mean AV concentration of DHEA-S by 5-fold, DHEA by 21-fold, A4 by 7-fold, and 11OHA by 5-fold. 11β-Hydroxytestosterone and testosterone were found to be potent androgen receptor agonists when tested with an androgen-responsive cell reporter model.

Conclusion:

The current study indicates that the adrenal gland secretes primarily 3 weak androgens, namely DHEA, 11OHA, and A4. Active androgens, including testosterone and 11β-hydroxytestosterone, are also produced but to a lesser degree.

The adrenal glands are the primary source of androgens and androgen precursors in reproductive-aged and postmenopausal women. The human adrenal gland has been shown to produce a variety of C₁₉ steroids that include dehydroepiandrosterone (DHEA), dehydroepiandrosterone sulfate (DHEA-S), androstenedione (A4), androstenediol, and 11β-hydroxyandrostenedione (11OHA) (1–3). Although these C₁₉ steroids have little androgenic activity, several studies have shown that they provide a pool of circulating precursors for peripheral conversion to more active androgens (1). Most research related to adrenal production of C₁₉ steroids has focused on DHEA, DHEA-S, and 11OHA, and these steroids are sometimes referred to as adrenal androgens (4–6). Because they are mainly derived from the adrenal, these steroids are often used as biomarkers of hyperactivity of the adrenal in women with androgen excess (7–10), but they are not bioactive androgens. Although there is overall agreement that the adrenal gland is a key provider of androgens in women, the overall adrenal androgen steroid metabolome and its responsiveness to corticotropin has not been well studied.

In the present study, we quantified 10 C₁₉ steroids in adrenal vein (AV) samples of women before and after ACTH stimulation. Our results demonstrate that 11OHA, DHEA, and A4 are the most abundant unconjugated C₁₉ steroids in the AV both before and after ACTH infusion. Interestingly, ACTH significantly increased the adrenal output of 9 of the 12 steroids measured, including the active androgens, testosterone and 11β-hydroxytestosterone (11OHT). These findings support an important role for the adrenal gland in the production of C₁₉ steroids (both precursors and active androgens) in normal women.

Subjects and Methods

Additional details on the methods used and statistical analyses are found in Supplemental Materials and Methods published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org.

Patient selection

This study was performed in women only because the physiological contribution of the ovary to the androgen pool is negligible compared with that of the testes. This fact enabled us to better define the specific contribution of the adrenal gland to the circulating androgen pool. Blood samples were collected from 7 female patients (Supplemental Table 1) with primary aldosteronism at Tohoku University Hospital from 2007 to 2008 via adrenal vein sampling (AVS). The protocol for establishing the diagnosis for primary aldosteronism has been reported in detail previously (11). Informed consent was obtained from all the study patients undergoing AVS. The study was approved by the Institutional Review Boards of Tohoku University School of Medicine and Georgia Health Sciences University.

AVS

After baseline samples were simultaneously obtained from both adrenal veins, a second set of blood samples were collected from the same sites 15 minutes after the iv administration of 0.25 mg (10 IU) of ACTH (11). Successful adrenal vein cannulation was confirmed from cortisol levels in the adrenal venous samples after ACTH stimulation, which raised cortisol to levels greater than 5 times those found in the vena cava (data not shown) (11). Unilateral hypersecretion of aldosterone was diagnosed based on the aldosterone to cortisol ratio after ACTH loading, with a cutoff value of 2.6 (11). For this study, steroid levels were quantified in the iliac vein and AV plasma from the nontumorous contralateral adrenal gland.

Measurements by liquid chromatography–tandem mass spectrometry (LC-MS/MS) and radioimmunoassay (RIA)

Methods for quantification of C₁₉ and C₁₈ steroids by LC-MS/MS are based on previously published methods as described by Nakamura et al (12). An expanded section on measurements by LC-MS/MS and RIA with statistical analyses is provided in the Supplemental Materials and Methods. Representative LC-MS/MS chromatograms for DHEA, androstenedione, 11OHA, testosterone, 11OHT, and 11-ketotestosterone (11KT) are provided as Supplemental Figure 1.

Cell culture and transactivation assays

MDA-kb2 cells were plated in duplicate at 150 000 cells per well in 24-well dishes. After 24 hours, the cells were treated with increasing concentrations of selected C₁₉ steroids including testosterone, 11OHT, 11KT, A4, 11OHA, and 11-ketoandrostenedione (11KA) with ethanol as the carrier solvent. As suggested by Blake et al (13) after treatment for 16 hours, the cell lysates were used for transactivation assays as described previously (14).

Microarray analysis to study the steroidogenic enzymes in the human adrenal gland

RNA from 5 normal adult adrenal glands was hybridized to an Affymetrix human HG-U133+2 oligonucleotide microarray set, and microarray analysis was performed as described previously (15).

Results

AV C₁₉ steroid and estrogen levels

Before ACTH stimulation, the most abundant AV steroid was DHEA-S (3827 ± 1317 nmol/L) (Table 1). Among the unconjugated C₁₉ steroids, the adrenal glands secreted 11OHA (157 ± 96.2 nmol/L), DHEA (125 ± 56.9 nmol/L), and A4 (79.0 ± 46.9 nmol/L) in substantial amounts (Table 1). The relative production of these steroids, as a percentage of the total mass of the measured unconjugated steroids, was 42% for DHEA, 37% for 11OHA, and 19% for A4. After 15 minutes of ACTH administration, DHEA was the major unconjugated adrenal C₁₉ product (63%) (Supplemental Table 2) followed by 11OHA (21%) and A4 (14%) (Supplemental Table 2). The most abundantly produced AV C₁₉ steroids after ACTH infusion were DHEA-S (18 266 ± 3842 nmol/L), DHEA (2659 ± 666 nmol/L), 11OHA (811 ± 260 nmol/L), and A4 (585 ± 199 nmol/L) (Table 1). The C₁₉ steroids that were increased in AV after ACTH stimulation were DHEA (21-fold), A4 (7-fold), androstenediol (18-fold), and 11OHA (5-fold) (Table 1). After ACTH treatment, the more potent androgens (eg, testosterone and 11OHT) were found to be 7- and 5.5-fold increasingly abundant in the AV compared with basal conditions, whereas estrone was increased by 3.5-fold (Table 1). 11KT and dihydrotestosterone (DHT) did not increase significantly in the AV after ACTH infusion.

Table 1.

Effects of ACTH on AV Concentration of C₁₉ Steroids and Estrogens in Women

Steroid	Pre-ACTH, nmol/L	Post-ACTH, nmol/L	P Value	Fold Change
DHEA	125 ± 56.9	2659 ± 666	<.05	21.2
DHEA-S	3827 ± 1317	18 266 ± 3842	<.05	4.8
Androstenediol	1.15 ± 0.24	20.7 ± 6.46	<.05	18.0
Androstenedione	79.0 ± 46.9	585 ± 199	<.05	7.4
11OHA	157 ± 96.2	811 ± 260	<.05	5.2
11KA	0.99 ± 0.33	3.18 ± 0.63	<.05	3.2
Testosterone	0.78 ± 0.26	5.71 ± 1.42	<.05	7.3
11OHT	0.48 ± 0.17	2.62 ± 0.74	<.05	5.5
11KT	0.39 ± 0.09	0.49 ± 0.11	.35	1.3
DHT	0.10 ± 0.03	0.10 ± 0.03	.40	1.0
Estrone	0.08 ± 0.03	0.29 ± 0.06	<.05	3.5
Estradiol	0.05 ± 0.03	0.05 ± 0.03	.61	1.0

Values are expressed as mean ± SEM (n = 7 women). The fold change increase after ACTH infusion is calculated by dividing the mean post-ACTH concentrations by the mean pre-ACTH concentrations.

^aLC-MS/MS was used to quantify the above C₁₉ steroids (except DHEA-S which was analyzed by RIA) and estrogens in the adrenal vein before andafter 15 min of iv ACTH administration.

To visualize individual variation in adrenal steroidogenesis and physiology, the 4 predominant AV C₁₉ steroids, namely DHEA, DHEA-S, 11OHA, and A4 as well as the more active androgens, testosterone and 11OHT, were plotted for each subject (Supplemental Figure 2). The line graphs denote the steroid concentrations (nanomoles per liter) in the AV serum of 7 individuals subjected to pre- and post-ACTH stimulation (15 minutes), with each line denoting 1 individual.

C₁₉ steroid concentrations in AV vs iliac vein

Nine C₁₉ steroid hormones were measured from serum collected in the iliac and AV before and after 15 minutes of ACTH stimulation. Before ACTH stimulation, the unconjugated C₁₉ steroids, DHEA, A4, 11OHA, and 11OHT as well as the conjugated steroid DHEA-S were present at significantly higher (21-, 42-, 83-, 2-, and 1.7-fold, respectively) levels in the AV compared with the iliac vein (Supplemental Table 3). However, after ACTH stimulation, 8 of the 9 measured steroids, including androstenediol and 11KA and the more potent androgens, namely testosterone and 11OHT, were found to be significantly higher in the AV. The steroids with the highest differential between the AV and the periphery after ACTH administration were DHEA (253-fold), 11OHA (250-fold), and A4 (216-fold) (Supplemental Table 3). Testosterone and its 11β-hydroxylated derivative 11OHT were found to be 13- and 9.5-fold more abundant in the post-ACTH AV compared with the iliac vein.

Effect of adrenal-derived C₁₉ steroids on androgen receptor (AR) activation

To better understand the physiological function of the adrenal-derived C₁₉ steroids, 6 of the steroid products were tested for their ability to activate the AR in MDA-kb2 cells. The MDA-kb2 cell model has transgenic expression of the human AR and an androgen-responsive luciferase reporter. The steroids indicated were assayed for androgen activity over a wide range of concentrations (Figure 1). Luciferase activity was significantly increased at 0.1 nmol/L testosterone (2.0 ± 0.1-fold, P < .05), maximum induction was observed at 30 nmol/L (16.4 ± 1.3-fold), and induction remained level at 100 nmol/L (Figure 1A). The EC₅₀ for testosterone was 0.52 nmol/L (Figure 1C). Luciferase activity for 11OHT and 11KT increased significantly, beginning at a concentration of 1 nmol/L for each steroid (P < .05). This activity continued to increase to 30 nmol/L for 11OHT, after which there was a slight decrease in the assayed activity (Figure 1A). For 11KT, the activity continued to increase and reached a plateau at 30 nmol/L (Figure 1A). For A4, the luciferase activity increased significantly, beginning at 30 nmol/L (2.8 ± 0.7-fold, P < .05), and it continued to increase in a concentration-dependent manner until it reached 1000 nmol/L (Figure 1B). 11KA also exhibited a dose-dependent increase in responsiveness to 1000 nmol/L (Figure 1B). 11OHA did not show any effect on luciferase activity even at 1000 nmol/L (Figure 1B). Similarly, DHEA did not increase the AR reporter activity (data not shown).

Figure 1.

C₁₉ steroid effects on AR activity. MDA-kb2 cells, containing human androgen receptor and an androgen receptor-driven luciferase reporter, were treated with increasing concentrations of testosterone, 11OHT, and 11KT (A) and A4, 11OHA, and 11KA (B). After 16 hours, luciferase activity was measured in cell lysates. Data are presented as mean fold induction compared with basal controls ± SEM (n > 4 independent experiments). C, EC₅₀ for the tested C₁₉ steroids.

Steroidogenic enzymes involved in androgen and estrogen synthesis in the human adrenal gland

Based on the above results, we propose a novel pathway in the human adrenal gland leading to the production of the unique adrenal androgens (Figure 2A). Microarray data for the genes encoding the various steroidogenic genes suggest that the human adrenal expresses high levels of the needed steroidogenic enzymes, namely 17α-hydroxylase/17,20-lyase, steroid sulfotransferase 2A1, type 2 3β-hydroxysteroid dehydrogenase, and 11β-hydroxylase (CYP11B1), and the cofactor cytochrome b₅. The enzyme, type 5 17β-hydroxysteroid dehydrogenase, which is able to produce androstenediol and testosterone from DHEA and A4, respectively (16), is present in substantial quantities in the adrenal gland. Other steroidogenic enzymes such as type 1 (HSD11B1) and type 2 11β-hydroxysteroid dehydrogenase (HSD11B2) are weakly expressed in the adrenal. Transcripts for the enzymes, type 2 17β-hydroxysteroid dehydrogenase, type 3 17β-hydroxysteroid dehydrogenase, and type 2 5α-reductase are hardly expressed in the adrenal. Microarray data also suggested that the major enzyme involved in estrogen synthesis, namely aromatase, is expressed negligibly in the adrenal (Figure 2B).

Figure 2.

Microarray analysis of steroidogenic enzymes involved in androgen (A) and estrogen (B) synthesis in the human adrenal glands. Dotted lines indicate that the reaction appears to be negligible in the adrenal. The enzyme transcripts present at the highest levels in the adrenal are marked in red and those with negligible expression are marked in green. The steroids that are not produced by the adrenal are marked in blue. AKR1C3, type 5 17β-hydroxysteroid dehydrogenase; CYB5, cytochrome b₅; CYP17, 17α-hydroxylase/17,20-lyase; CYP19, aromatase; HSD3B2, type 2 3β-hydroxysteroid dehydrogenase; HSD17B1, type 2 17β-hydroxysteroid dehydrogenase; HSD17B2, type 2 17β-hydroxysteroid dehydrogenase; HSD17B3, type 3 17β-hydroxysteroid dehydrogenase; SRD5A2, type 2 5α-reductase; SULT2A1, steroid sulfotransferase 2A1.

Discussion

The adrenal cortex is an important source of androgens and androgen precursors in both premenopausal and postmenopausal women. Although adrenal C₁₉ steroid biosynthesis has been studied over the past 5 decades, the spectrum of the steroids examined in these investigations has been limited owing to the available methodologies. Our study focused on better defining the adrenal androgen metabolome and its response to ACTH with the use of LC-MS/MS. LC-MS/MS has emerged as a technique suited to identify panels of steroids in complex samples such as blood, urine, saliva, and tissues such as breast and placenta (17, 18). Herein, we used LC-MS/MS to quantify nine unconjugated C₁₉ steroids and 2 estrogens in human AV samples before and after ACTH stimulation. Our findings suggest that the adrenal releases a panel of ACTH-regulated C₁₉ steroids, some of which are potent activators of the AR.

Based on the current study, we provide evidence of a novel pathway in the human adrenal, leading to the production of several adrenal androgens and precursor steroids (Figure 2A). Microarray data suggest that the human adrenal expresses high levels of the steroidogenic enzymes that are obligatory for DHEA and DHEA-S synthesis, namely 17α-hydroxylase/17,20-lyase and sulfotransferase 2A1, and the cofactor cytochrome b₅. The high expression levels agree well with the dominant production of DHEA-S seen in our steroid analysis. Our array data indicate that the adrenal does not express the testicular enzyme responsible for A4 to testosterone conversion, namely type 3 17β-hydroxysteroid dehydrogenase, but does express high levels of 17β-hydroxysteroid dehydrogenase, which can produce testosterone from A4. In addition, low expression of type 2 17β-hydroxysteroid dehydrogenase indicates that the adrenal milieu favors the unidirectional conversion of DHEA and A4 into their respective products, A4 and testosterone. The high expression of CYP11B1 allows the adrenal to produce the 11β-hydroxylated products of A4 and testosterone. This phenomenon is also indicated by the significant increase in 11OHA and 11OHT concentrations after ACTH stimulation. The enzymes HSD11B1 and HSD11B2 are moderately expressed in the adrenal, thereby suggesting the possibility of the bidirectional conversion between 11OHA and 11KA and confirmed by increased 11KA concentration after ACTH stimulation. The low expression of type 2 5α-reductase also supports the inability of the adrenal to produce DHT, which is confirmed by the low AV concentrations of DHT after ACTH stimulation.

Perhaps the most interesting finding of the current study relates to the 11β-hydroxylated C₁₉ steroids, 11OHA and 11OHT. 11OHA was previously suggested to arise from the adrenal gland as a result of the 11β-hydroxylation of A4 by the adrenal-specific enzyme CYP11B1 (8, 19). Our findings confirm that 11OHA is abundantly produced by the human adrenal and is under the regulation of ACTH. There is little known regarding the source or physiological role of 11OHT in humans although its circulating levels have been measured using RIA (20). 11OHT and 11KT do have a physiological role in a variety of teleost where they represent active androgens (21, 22). In fish, it is probably HSD11B2 that converts 11OHT to 11KT, but in our studies the human adrenal did not appear to produce 11KT (22–24). These testosterone derivatives appear to be synthesized in small amounts by mouse testis (22). Our findings establish that 11OHA and 11OHT are human adrenal products, of which 11OHT should be considered a direct adrenal androgen product, whereas 11OHA is abundantly produced and can act as a precursor for peripheral and target tissue production of more active hormones.

To evaluate the ability of the various C₁₉ steroids found in the AV to modulate the activity of the AR, MDA-kb2, a cell line with endogenous AR that stably expresses an androgen-responsive mouse mammary tumor virus-luciferase reporter gene, was used (14). The 11β-hydroxy derivatives of A4 and testosterone were selected for testing along with their precursor steroids. Although 11OHA is one of the most abundant steroids in the AV, it had almost no androgenic activity. On the other hand, although 11OHT was secreted in smaller quantities, it was a much more potent androgen. Based on our findings, testosterone activates the AR at concentration that is approximately 300-fold less than A4, whereas A4 requires 30-fold higher levels than 11OHT and 11KT to provoke an AR response. Our study suggests that the basal AV levels of A4 could stimulate the AR moderately. Previous data using the MDA-kb2 cell line suggested that EC₅₀ for testosterone was 0.51 nmol/L (13), which was consistent with our data. Thus, it is interesting to note that the basal AV levels of testosterone as determined by our study are sufficient to elicit an AR response. In addition, the peripheral conversion of DHEA and A4 to testosterone and DHT may also strengthen their androgenic activity. Yazawa et al (22) transfected the CV-1 cells transiently with an AR expression vector to compare the AR activation abilities of testosterone, 11KT, 11OHA, and A4. Their findings suggested that 11KT was effectively as strong as testosterone in eliciting a response. However, our data suggest that even though 11KT is a potent androgen, it does not activate the AR as strongly as testosterone. Nevertheless, both studies suggest that A4, 11KA, and 11OHA are poor activators of the AR.

In summary, LC-MS/MS profiling of AV samples from women demonstrated that the adrenal gland synthesizes a panel of C₁₉ steroids, including the potent androgens testosterone and 11OHT. In addition we showed that AV levels of DHEA-S, before and after ACTH infusion, are the highest among those of the steroids measured. As for the unconjugated steroids, 11OHA, DHEA, and A4 are the most abundant in the AV samples before and after ACTH stimulation. The study represents the first use of LC-MS/MS for broad-based analysis of this number of C₁₉ steroids in AV samples before and after ACTH administration and also suggests that this method will provide excellent methodology to further define the androgen metabolome of the human adrenal. Further investigations will be needed to clarify the role of the potent C₁₉ steroids as adrenal biomarkers in conditions of androgen excess.