| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles: Hormones and Reproductive Health |
Departments of Obstetrics and Gynecology (R.A., R.P., L.R.B.), Medicine (R.A.), and Biostatistics (L.M.F.), University of Alabama, Birmingham, Alabama 35233; and Department of Gynecology and Obstetrics, The Johns Hopkins University College of Medicine (H.A.Z.), Baltimore, Maryland 21287
Address all correspondence and requests for reprints to: Ricardo Azziz, M.D., Department of Obstetrics and Gynecology, University of Alabama, 618 South 20th Street, OHB 549, Birmingham, Alabama 35233-7333. E-mail: razziz{at}uabmc.edu
Abstract
Excess adrenal androgen (AA) levels are observed in 25–50% of women
with the polycystic ovary syndrome (PCOS), and AA excess in PCOS may
represent selection bias. Thus, it is possible that AA secretion among
the general population is highly variable, and that those women who are
predisposed to secreting greater amounts of AA have a greater
probability of having PCOS. We now hypothesize that the levels of AAs
are highly variable among normal nonhyperandrogenic women, and that
this heterogeneity is the result of a variable response of AAs to ACTH
stimulation. To test this hypothesis we prospectively studied the
response of dehydroepiandrosterone (DHA) and cortisol (F) to a 60-min
acute stimulation with ACTH-(1–24) in 56 healthy eumenorrheic
nonhirsute healthy women with a mean age of 28.9 yr (range, 20–37 yr.)
and a mean body mass index (BMI) of 29.2 kg/m2 (18.2–46.2
kg/m2). Baseline samples and poststimulation samples were
assayed for DHA and F. The basal and ACTH-stimulated levels of DHA, but
not those of F, were negatively correlated with age, although neither
the basal nor ACTH-stimulated responses of DHA and F varied with BMI.
After controlling for age, the basal F level was negatively correlated
to its net increment (i.e. 
F; r = -0.54;
P < 0.001), whereas there was no significant
relationship between basal DHA and DHA. We also compared the
intersubject variability (coefficient of variation) for basal and
stimulated levels of DHA and F. For basal (DHA0), 60 min
(DHA60), and net increment in (DHA) DHA levels, the
coefficients of variation were 67.9%, 61.4%, and 76.0%,
respectively; for F0, F60, and F, they were
40.4%, 16.9%, and 31.3%, respectively. The variance in DHA was
significantly higher, and the variance in F60 was
significantly lower than that in all other variables; DHA0,
DHA60, F0, and F had similar variances.
In conclusion, in our population of healthy reproductive-aged women we
observed that both basal and ACTH-stimulated levels of DHA after
ACTH-(1–24) stimulation had significantly greater intersubject
variance (
60–70%) compared with the basal and poststimulation
levels of F (15–40%). These data support the hypothesis that among
normal women, AA (i.e. DHA) levels are highly variable
compared to those of F. In addition, the intersubject variability in
DHA levels is at least in part due to a variable response of AAs to
ACTH stimulation. Whether the AA excess frequently observed in PCOS is
due to the greater risk of those women with higher AA levels, basally
and after ACTH stimulation, remains to be confirmed.
HYPERANDROGENISM, generally in the form of the polycystic ovary syndrome (PCOS), affects approximately 4% of unselected reproductive-aged women (1). Although most of these patients suffer from excess ovarian androgen secretion, excess adrenal androgen (AA) levels are also observed in 25–50% of women with PCOS (2, 3, 4). The underlying etiology for the AA excess frequently observed in PCOS remains unclear. However, AA excess in PCOS may represent selection bias. Thus, it is possible that AA secretion among the general population is highly variable, and that the women who are genetically predisposed to secrete greater amounts of AA will have a greater probability of having PCOS. High AA secretion would then represent another of the many factors potentially predisposing to or increasing the risk for the development of PCOS.
In support of this hypothesis, various investigators have reported that circulating levels of the adrenal androgens dehydroepiandrosterone (DHA) and DHA sulfate (DHS) appear to be under significant genetic control (5, 6). In fact, some investigators have noted that the genetic influence on AA levels may be greater in women than in men (7). Other data also support the suggestion that the circulating AA levels in women are predetermined to a significant degree. For example, the relative circulating level of DHS in postmenopausal women appears to vary little over time. Thus, individuals who had higher levels of DHS early in menopause tended to always have higher levels relative to other menopausal women and vice versa independent of the age-related decline in AA levels normally observed (8). Overall, it appears that AA levels are highly individualized, and that this variability may be under significant genetic influence.
ACTH stimulates the release of both AAs and glucocorticoids in vivo (9) and in vitro (10). Hence, it is possible that the elevated AA levels found in PCOS patients with AA excess reflect an exaggerated response of these steroids to ACTH, and that this adrenal response to ACTH is under genetic control. In fact, we previously reported that AA excess in PCOS patients is related to an exaggerated secretory response of the adrenal cortex to ACTH for both DHA and androstenedione, but not to altered pituitary responsivity to CRH or to increased sensitivity of the adrenal cortex to ACTH (11).
Overall, it is then possible that AA excess among PCOS patients represents selection bias, such that women who are genetically predisposed to secrete greater amounts of AA will be at greater risk of developing PCOS. This suggestion would require that AA secretion among the general population be highly variable. We now hypothesize that compared with cortisol (F), the levels of AAs are highly variable among normal nonhyperandrogenic women. We further hypothesize that this heterogeneity in AA secretion is the result of a variable response of AAs to ACTH stimulation. To test these hypotheses we prospectively studied the responses of DHA and F to acute stimulation with ACTH-(1–24) in healthy eumenorrheic nonhirsute women.
Subjects and Methods
Subjects
We recruited 56 eumenorrheic women with regular menstrual cycles every 26–32 days, without evidence of hirsutism, with a negative family history for endocrine disorders, and taking no medications, including oral contraceptives. All underwent acute adrenal stimulation testing, as outlined below, after appropriate written and informed consent was obtained according to the guidelines of the joint committee on clinical investigation of The Johns Hopkins Hospital and the institutional review board of University of Alabama (Birmingham, AL).
Study protocol
Acute adrenal stimulation was performed as previously described (9). In brief, all studies were performed between 0800–1030 h in the fasting state during the follicular phase (days 3–8) of the menstrual cycle. Dexamethasone was not administered before the study so that resting basal steroid levels could be assessed. Three baseline samples were obtained 15 min apart and mixed to form the 0 min (basal) sample. Immediately thereafter 1 mg ACTH-(1–24) (Cortrosyn, Organon, West Orange, NJ) was administered iv over 60 s, and blood was sampled 60 min later. Serum was separated and stored at -20 C until assayed.
We previously reported that 1 mg ACTH-(1–24), iv, elicits a maximum response of adrenocortical steroids regardless of body weight (9). Furthermore, we noted that the steroid response to this dose of ACTH-(1–24) is highly reproducible over time (9).
Hormonal measures
Baseline samples were assayed for DHS, DHA, and F, and DHA and F were also measured in the 60 min samples. Serum samples from all patients were batched for analysis, and hormonal assays were performed at one time.
DHS was measured by direct RIA (Diagnostic Products, Los Angeles, CA), and DHA was measured by RIA using a single antibody, separating free from bound steroid with dextran-coated charcoal (Radioassay Systems Laboratories, Inc., Carson, CA). The intraassay coefficients of variation (CVs) were 4.1%, 6%, and 7% for DHS, and 9.8%, 4.7%, and 4.3% for DHA for low, medium, and high values, respectively. F levels were determined by RIA as previously described (12), except that tritiated, rather than iodinated, F was used, and dextran-coated charcoal was used to separate antibody-bound and free steroid. The intraassay CVs were 4.3% and 6.7% for high and low values, respectively.
Statistical analysis
In addition to determining the steroid levels at 0 min (basal
level or steroid0) and 60 min (maximal response
or steroid60) after ACTH-(1–24) administration,
the net change (net increment) in hormone levels was calculated
(steroid). Correlations were established using the Pearson
correlation coefficient analysis. The CVs were compared by the
Duncan’s multiple range test (
= 0.05) and Fischer’s least
significant difference test.
Results
The mean age of our main study group consisting of 56 study
subjects was 28.9 yr (range, 20–37 yr) with a mean body mass index
(BMI) of 29.2 kg/m2 (range, 18.2–46.2
kg/m2). Their mean basal and stimulated values
(and ranges) before and after ACTH-(1–24) stimulation are depicted in
Table 1
and Fig. 1. The ACTH-stimulated F levels in our
subjects fell within the range previously reported for normal
individuals (13, 14).
|
|
The basal levels of DHA (i.e.
DHA0) and its response to ACTH (i.e.
DHA60 and DHA) were negatively correlated
(decreased) with age (r = -0.46, -0.58, and -0.52;
P < 0.0001–0.0009, respectively). Alternatively,
neither F0, F60, nor F
varied with subject age. Furthermore, neither basal nor ACTH-stimulated
responses of DHA and F varied with BMI. The circulating level of DHS
was positively correlated with the DHA0 and
DHA60 values (r = 0.47 and 0.35;
P < 0.008 and 0.02, respectively), but not with
DHA. Alternatively, DHS levels were not correlated to either the
basal level of F or its response to ACTH.
The basal levels of both DHA and F (i.e.
DHA0 and F0, respectively)
were positively correlated to their maximal poststimulation values
(i.e. DHA60 and
F60; r = 0.71; P < 0.0001
and r = 0.46; P < 0.0004, respectively). In
contrast, the basal F level was negatively correlated to its net
increment or change (i.e. F; r = -0.53;
P < 0.001), although DHA0 was
positively correlated to the DHA value (r = 0.39;
P < 0.006).
The relationship between the basal levels of DHA and F and their
respective responses to ACTH-(1–24) were then reanalyzed, controlling
for subject age. The basal levels of F and F remained negatively
correlated (r = -0.54; P < 0.0001), whereas the
positive relationship between DHA0 and DHA was
no longer significant (r = 0.20; P = 0.17). Figure 2 depicts the relationship between the
basal levels of DHA and F and their respective increments, controlling
for subject age.
|
In the main study group of 56 subjects we first calculated the
intersubject CVs for each of the six variables under study. For
DHA0, DHA60, and DHA,
the CVs were 67.9%, 61.4%, and 76.0%, respectively; for
F0, F60, and F, they
were 40.4%, 16.9%, and 31.3%, respectively. We then compared the CVs
of these variables by Duncan’s multiple range test (with an =
0.05). The CV of DHA was significantly higher than those of all
other variables, the CV of F60 was significantly
lower than those of all other variables, and
DHA0, DHA60,
F0, and F had similar CVs. Similar findings
were obtained when the CVs of DHA0,
DHA60, DHA, F0,
F60, and F were compared by Fischer’s least
significant difference test (i.e. least squares means
method).
The fractional difference from the mean of the group for each subject
was also calculated (i.e. the observed subject value minus
the group mean, with the difference divided by the group mean). As
before, comparison of these differences by fischer’s least significant
difference test indicated that the maximal F response (i.e.
F60) had significantly less intersubject
variability than all other variables (P <
0.04–0.0001), except F. In contrast, the CV of DHA was
significantly greater than that of all other variables
(P < 0.02–0.0001), except
DHA0.
Discussion
Overall, in our population of healthy reproductive-aged women, both basal and maximal levels of DHA after ACTH-(1–24) stimulation had significantly greater intersubject variance (i.e. 60–70%) compared to the basal level and the response of total F (i.e. 15–40%). These data support the hypothesis that among normal women, AA (i.e. DHA) levels are highly variable, at least compared to those of F. Hence, it is possible that the presence of AA excess in 25–30% of PCOS patients simply reflects selection bias, such that women who are genetically predisposed to secrete greater amounts of AAs will have a greater probability of having the hyperandrogenic disorder PCOS.
It is unlikely that the difference in intersubject variability between DHA and F is due to a greater intrinsic variability of the response to ACTH within subjects over time. In fact, we previously reported that the intersubject variability of the response to ACTH-(1–24) stimulation was equally stable for DHA and F (i.e. area under the response curve studied for up to 6 months), with CVs of 16% and 7%, respectively (9). It is also possible that the lesser variability of F after ACTH stimulation, compared with that of DHA, is due to the greater degree of binding of F by corticosteroid-binding globulin compared with the lesser binding of DHA by sex hormone-binding globulin (15). However, studying a separate cohort of 18 healthy women, we observed that the intersubject variability of unbound F after ACTH-(1–24) stimulation remained low at 26% (Azziz, R., unpublished data).
The intersubject uniformity in the maximal response of total F to stimulation was reflected by the fact that the higher the basal level of F, the lesser its net increment (i.e. change) in response to stimulation. Alternatively, after controlling for the effect of age, there was no relationship between the basal DHA level and its net increment after ACTH-(1–24) stimulation. Hence, the intersubject variability in DHA levels is at least in part due to a variable response of AAs to ACTH stimulation.
The high degree of variation in the AA response to ACTH-(1–24) compared to that of F indicates that although ACTH is an important determining factor of F secretion, this hormone appears to be only partially responsible for AA secretion. In fact, other as yet unknown factors seem to play an important, and possibly determinant, role in the regulation of AA biosynthesis. Various clinical examples illustrate the dichotomous secretion of AAs and glucocorticoids. For example, in patients with myotonic dystrophy the circulating level of AAs and their response to ACTH are markedly reduced, although the level and response of F are either normal or markedly elevated (16). In addition, with little change in glucocorticoid secretion, the levels of AAs increase during adrenarche and progressively decline with age (17). Finally, in our present study a greater individual DHA response (i.e. net increment) to ACTH stimulation did not directly predict higher basal levels of either DHA or DHS. Hence, it appears that control of F and AA secretion frequently diverges, and that factors other than ACTH are important determinants of AA biosynthesis.
The mechanism(s) underlying the variability of the DHA response to ACTH
remains unknown. Variations in extraadrenal factors, such as degree of
insulinemia or amount of the putative cortical androgen-secreting
hormone (18), may play a role in determining the
variability in AA secretion. However, it is more likely that the
intersubject variability in AA secretion reflects an intraadrenal
process. For example, differences in the responsivity of AAs and F to
acute ACTH stimulation may reflect differences in the relative cell
mass of the zonae fasciculata and reticularis of the adrenal cortex.
Another promising mechanism accounting for the intersubject variability
in AA secretion is variation in the function of P450c17. This enzyme
possesses both 17-hydroxylase and 17,20-lyase activities; the latter
activity is critical for androgen biosynthesis. Importantly, the
17-hydroxylase and 17,20-lyase activities of P450c17 are
differentially regulated. For example, 17,20-lyase activity can be
modified by variations in the amounts of the cofactors
P450-oxidoreductase (19, 20) and cytochrome
b5 (21, 22, 23), by changes in the
structural elements that allow P450c17 to interact with
P450-oxidoreductase (24), and by the degree of P450c17
phosphorylation (25). Hence, genetic or environmental
factors that modulate 17,20-lyase activity are likely to play a key
role in controlling AA secretion.
Our results suggest that there is significant populational heterogeneity in the adrenal secretion of DHA in response to ACTH, whereas there is little intersubject variability for maximum F secretion in response to ACTH. Teleologically this steroidogenic arrangement makes sense, because an insufficient or excessive F response to ACTH can be life-threatening, whereas variations in the secretion of AAs are not. Thus, evolutionary pressures may favor tight control of F production, but much more lax production of AAs. In conclusion, our study in a population of healthy nonhyperandrogenic women indicates that the levels of AAs are highly variable between subjects and that this heterogeneity may be the result in part of a variable response of AAs to ACTH stimulation. Hence, it is possible that the presence of AA excess in 25–30% of PCOS patients simply reflects selection bias, such that women who are genetically predisposed to secrete greater amounts of AAs will have a greater probability of having this hyperandrogenic disorder PCOS. Nonetheless, this hypothesis remains to be confirmed.
Footnotes
1 This work was supported in part by NIH Grant RO1-HD-29364 (to
R.A.), GCRC Grant M01-RR00032, and Grant N0014-96-I-0255 from the
Office of Naval Research (to C.R.P.). 
Received August 23, 2000.
Revised November 15, 2000.
Accepted November 22, 2000.
References
-hydroxylase and 17,20 lyase activities of P450c17:
Contributions of serine106 and P450 reductase. Endocrinology. 132:2498–2506.[Abstract]
-hydroxylase-17,20-lyase
(CYP17) by cytochrome b5: endocrinological and
mechanistic implications. Biochem J. 308:901–908.
This article has been cited by other articles:
 |
|
 |
|
  L. Marinelli, E. Trevisi, L. Da Dalt, M. Merlo, G. Bertoni, and G. Gabai Dehydroepiandrosterone secretion in dairy cattle is episodic and unaffected by ACTH stimulation J. Endocrinol., September 1, 2007; 194(3): 627 - 635. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Zhou, I. M. Bird, D. A. Dumesic, and D. H. Abbott Adrenal Hyperandrogenism Is Induced by Fetal Androgen Excess in a Rhesus Monkey Model of Polycystic Ovary Syndrome J. Clin. Endocrinol. Metab., December 1, 2005; 90(12): 6630 - 6637. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Reisch, M. Slawik, O. Zwermann, F. Beuschlein, and M. Reincke Genetic influence of an ACTH receptor promoter polymorphism on adrenal androgen secretion Eur. J. Endocrinol., November 1, 2005; 153(5): 711 - 715. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. O. Yildiz, K. S. Woods, F. Stanczyk, A. Bartolucci, and R. Azziz Stability of Adrenocortical Steroidogenesis over Time in Healthy Women and Women with Polycystic Ovary Syndrome J. Clin. Endocrinol. Metab., November 1, 2004; 89(11): 5558 - 5562. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
