- Email: [email protected]
Alterations in androgen conjugate levels in women and men with alopecia
FERTILITY AND STERILITY
Vol. 62, No.4, October 1994
Copyright e 1994 The American Fertility Society
Printed on acid-free paper in U. S. A.
Alterations in androgen conjugate levels in women and men with alopecia
RichardS. Legro, M.D.*t Enrico Carmina, M.D.:j: Frank Z. Stanczyk, Ph.D.*
Elisabet Gentzschein, B.S.* Rogerio A. Lobo, M.D.§
University of Southern California School of Medicine, Los Angeles, California, and Cattedra di Endocrinologia, Universita di Palermo, Palermo, Italy
Objective: To assess levels of androgen metabolites thought to reflect, at least in part, peripheral androgen activity in women with androgenic alopecia and men with premature balding in an effort to determine if a common abnormality exists. Design: Prospective study in various groups of women and men. Setting: Reproductive Endocrine Clinic at our university medical center. Patients: Ten normal ovulatory female controls and 50 hyperandrogenic women divided on the basis of hirsutism and alopecia as follows: [1] 8 hirsute women with androgenic alopecia; [2] 12 nonhirsute women with androgenic alopecia; [3] 18 hirsute women without androgenic alopecia; and [4]12 nonhirsute women without androgenic alopecia. Ten normal men and 10 young premature balding men matched for age and weight also were compared. Intervention: Blood was obtained from all subjects. Main Outcome Measure: Comparison of blood hormone levels in the various groups. Results: Serum T, androstenedione, and DHEAS were similarly elevated in hyperandrogenic women with and without alopecia, compared with controls. The female groups were then divided on the basis of hirsutism. Hirsute groups with and without alopecia had similarly elevated levels of unconjugated 3a-androstanediol, 3a-androstanediol glucuronide, 3a-androstanediol sulfate, androsterone glucuronide, and androsterone sulfate compared with controls. In the nonhirsute groups, androgenic alopecia patients were compared with hyperandrogenic females and cycling controls. The androgenic alopecia patients had elevated levels of 3a-androstanediol (0.75 ± 0.12 versus 0.46 ± 0.1 and 0.41 ± 0.1 nmol/L), 3a-androstanediol sulfate (200 ± 31 versus 79.6 ± 6 and 67.0 ± 4.0 nmol/L), elevated ratios of 3a-androstanediol sulfate:3a-androstanediol (267 ± 49 versus 170 ± 20 and 164 ± 49 nmol/L), elevated ratios of 3a-androstanediol sulfate:3a-androstanediol glucuronide (32.2 ± 6 versus 10.8 ± 1 and 10.0 ± 1) and lower ratios of 3a-androstanediol glucuronide:3a-androstanediol glucuronide (8.3 ± 1.8 versus 17 ± 1.7 and 15.2 ± 1.6 nmol/L). In men the premature balding group had lower levels of 3a-androstanediol glucuronide compared with the male controls (29.8 ± 4.4 versus 15.2 ± 1.6 nmol/L). Also, the ratio of 3a-androstanediol glucuronide:3a-androstanediol was significantly decreased, whereas the ratio of 3a-androstanediol sulfate:3a-androstanediol glucuronide was elevated. Conclusions: These data provided evidence confirming that enhanced Sa-reductase activity occurs in androgenic alopecia but also suggests that a disorder of androgen conjugation, favoring sulfurylation over glucuronidation, may be a characteristic feature of scalp hair loss. Fertil Steril 1994;62:744-50 Key Words: 5a-Reductase, alopecia, hyperandrogenism, steroid conjugates Received September 10, 1993; revised and accepted May 19, 1994. *Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, University of Southern California School of Medicine. t Present address: Division of Reproductive Endocrinology, Pennsylvania State University Hospital, Hershey, Pennsylvania.
744
Legro et al.
Androgen conjugates in alopecia
:j: Cattedra di Endocrinologia, Universita di Palermo. § Reprint requests: Rogerio A. Lobo, M.D., Department
of Obstetrics and Gynecology, Los Angeles County and University of Southern California Medical Center, Women's and Children's Hospital, Room 1M2, 1240 North Mission Road, Los Angeles, California 90033 (FAX: 213-2263026).
Fertility and Sterility
-l Although alopecia can be due to different causes ranging from systemic diseases to scalp autoimmunity, the most common form is male pattern baldness or alopecia androgenetica or androgenic alopecia (1). This process, which may be extremely gradual, is a transition from terminal to vellus hair (2) and is generally associated with normal androgen levels, although in some studies patients have had hyperandrogenemia. Up to 38% of female patients with alopecia have been found to have elevated androgen levels (3); alopecia, however, is not as common among hyperandrogenic women. In male patients, there have been reports of increased T levels in saliva (4), increased serum androstenedione (A) (5), and in a group of young affected males, elevated levels of DHEAS (6). However, discounting a rapid and progressive virilization because of an androgen-secreting tumor or exogenous androgens, elevated androgen levels are uncommonly linked with androgenic alopecia, which most commonly follows a hereditary pattern. Because of this, it has been hypothesized that androgen metabolism is modified in androgenic alopecia. In men with baldness, scalp hair 5a-reductase activity has been found to be increased (7). Accordingly, it has been proposed that increased scalp 5a-reductase activity may lead to higher intrafollicular dihydrotestosteorne levels that, in turn, inhibit hair follicle adenylate cyclase activity (1, 7). One way to gain further insight into the pathogenesis of androgenic alopecia is to evaluate circulating levels of androgen metabolites. Recently, we have shown that different manifestations of hyperandrogenism, specifically hirsutism and acne, exhibit different profiles of circulating androgen metabolites that reflect, at least in part, peripheral androgen action (8-10). Accordingly, in this study we have measured several serum C 19 5a- reduced metabolites in men with premature balding and in female patients with androgenic alopecia with and without other manifestations of androgen excess. Levels of the androgen metabolites were compared in various control groups. Because generalized increases in 5a-reductase activity may also lead to hirsutism in women, we have included hirsute and nonhirsute women, with and without alopecia, as additional controls. These patients all have hyperandrogenemia of a similar degree. We hypothesize that alopecia may be characterized by a different pattern of elevated c19 5a metabolites compared with hirsutism and that a similar metabolic fate may also be present in men with premature balding. Vol. 62, No.4, October 1994
MATERIALS AND METHODS Subjects
Twenty premenopausal women (aged 23.2 ± 1 years [mean ± SE], range 18 to 30 years) affected by androgenic alopecia were studied. This diagnosis was made by a dermatologist after a careful examination and the elimination of other dermatologic conditions. Eight patients also had hirsutism in conjunction with alopecia. All had increased levels of at least one circulating androgen (total T, unbound T, A, or DHEAS). None of the women previously had received hormonal treatment for their alopecia or hyperandrogenism. The women with androgenic alopecia had diffuse central alopecia with retention of the frontotemporal hairline and the severity of alopecia, as assessed by the Hamilton method, was 4 or greater (on a scale of 0 to 8) (11). No patient had evidence of masculinization or clitorimegaly. None of the patients had evidence of thyroid or adrenal dysfunction. Specifically, 21-hydroxylase deficiency was excluded based on normal early morning serum 17-hydroxyprogesterone levels. Cushing's disease also was excluded on the basis of a normal 24-hour urine for free cortisol. None of the patients were taking steroids or anticonvulsants. Ten normal women (aged 21.1 ± 1.3 years) and 30 hyperandrogenic women (aged 21.5 ± 0.6 years) without alopecia and of similar weight as those with androgenic alopecia were chosen as comparison groups or controls. The normal women were normoandrogenic women with regular menstrual cycles. They were without stigmata of hyperandrogenism. Hyperandrogenic controls presented with menstrual irregularity. The etiology of the hyperandrogenism is compatible with a diagnosis of polycystic ovarian syndrome that we refer to as hyperandrogenic chronic anovulation. Among the 30 hyperandrogenic women without alopecia, 18 were hirsute and 12 were not. Hirsutism was assessed by a modification of the Ferriman-Gallwey method (12), and only patients with a score of 8 or more were considered hirsute. Thus, despite hyperandrogenemia in all 30 women, the presence of hirsutism in only 18 may be explained by the existence of increased skin 5a-reductase activity in these women (1, 8). All patients and controls were studied during the midfollicular phase of a spontaneous cycle (days 5 to 8) or after a P-induced menses. Male patients with premature baldness (aged 27.7 ± 0.5, range 26 to 30, 100% ideal body weight Legro et al. Androgen conjugates in alopecia
745
· Table 1 Clinical Data and Androgen Levels in 10 Female Controls, 30 Female Hyperandrogenic Patients Without Alopecia and 20 Females Androgenic Alopecia Patients*
Female controls Hyperandrogenic no alopecia Androgenic alopecia
Age
IBW
y
%
21.1 ± 1.3 21.5 ± 0.6 23.2 ± 1.1
107 ± 3 114 ± 3.1 111 ± 3.9
* Values are means ± SE. t P < 0.01 versus controls.
A
Total T
DHEAS
nmol!L
5.3 ± 0.7 14.7 ± 0.7t 13.9 ± 0.7t
Jimol/L
1.03 ± 0.1 2.74 ± 0.2t 2.65 ± 0.1t
4.5 ± 0.5 6.9 ± 0.5:j: 7.8 ± 0.8:j:
:j: P < 0.05 versus controls.
[IBW]) were age and weight matched to control nonbalding males (age 29.1 ± 0.9, range 25 to 33, 99% IBW). None of the balding men were receiving any therapy for alopecia. None of the subjects were affected by adrenal or thyroid disease. Balding males were judged according to the Hamilton method (11), and scores ranged from 4 to 8. Study Design
Our study design for the female groups was based on clearly delineating diverse causes of a common phenotype. Although we hypothesized increased 5a-reductase activity as the cause of androgenic alopecia, androgenic alopecia can also result from an elevated production of androgen precursors (primarily ovarian based on our screening). Therefore, we sought hyperandrogenic women without alopecia as a comparison group in addition to normoandrogenic cycling women (Table 1). Additionally, our main outcome variables, serum androgen metabolites of 5a-reductase activity, can be elevated not only by increased scalp activity but also by a generalized increased in skin activity as is the case in hirsutism (Table 2) (8). Therefore, when examining these metabolites, we sought to control for this
confounding variable by subdividing the probands and hyperandrogenic patient groups on the basis of hirsutism (Fig. 1 for the hirsute groups and Figure 2 for the nonhirsute groups). The same female normal controls with normal female hair distribution were used in both latter comparisons. Thus, Figures 1 and 2 represent a subdivision of Table 2 on the basis of hirsutism. Finally, we explored ratios of dihydrotestosterone metabolites and conjugates in the groups in which we hypothesized their production was primarily scalp dependent: the nonhirsute groups (Fig. 3). It was not necessary to further subdivide our male groups because both control males and those with premature balding had similar levels of circulating androgen precursors and similar peripheral hair distributions. Protocol
All patients had fasting blood obtained between 8:00 and 10:00 A.M. after at least 0.5 hour of recumbent rest. Sera from all patients were separated and stored at -20°C until assayed for total T, A, and DHEAS by established RIA techniques (13-15). Unbound Twas measured only in men by ammonium sulfate precipitation of sex hormone-binding
Table 2 Serum Androgen Metabolites in 10 Female Controls, 30 Hyperandrogenic Patients Without Alopecia, and 20 Female Androgenic Alopecia Patients
Females controls Hyperandrogenic no alopeciat Androgenic alopeciat
3a*
3a-G*
3a-S*
AoG*
Ao S*
0.41 ± 0.03 0.67 ± 0.07:j: 0.69 ± 0.06:j:
6.2 ± 0.47 13.2 ± 1.3:j: 11.1 ± 2.3§
67.2 ± 4.2 125.5 ± 15.9:j: 197.6 ± 27.3§11
73 ± 4.6 151.1 ± 15.4:j: 158.7 ± 27.8:j:
2,038 ± 200 2,724 ± 297§ 3,618 ± 686:j:
* 3a, 3a-androstanediol; 3a-G, 3a-androstanediol glucuronide; 3a-S, 3a-androstanediol sulfate; Ao G, androsterone glucuronide; Ao S, androsterone sulfate. Values are nmol/L means ± SE. t The hyperandrogenic and alopecia groups have been further subdivided on the basis of hirsutism in Figure 1 (hirsute) and Figure 2 (nonhirsute). 746
Legro et al.
Androgen conjugates in alopecia
:j: P < 0.01 versus controls. § P < 0.05 versus controls. II P < 0.05 versus hyperandrogenic.
Fertility and Sterility
globulin as previously described (15). In addition, the following 5a- reduced androgens were measured: [1] unconjugated 3a-androstanediol; [2] 3aandrostanediol glucuronide; [3] 3a-androstanediol sulfate; [4] androsterone glucuronide; and [5] androsterone sulfate. These assays employed specific hydrolysis of the conjugates, followed by extraction, celite column chromatography, and RIA as previously described (8, 16, 17). In all cases, intraassay and interassay coefficients of variations did not exceed 10% and 12%, respectively.
3a --llndrostanediol (3al
C
NH-H
Androsterone Glucuronide IRoGI
NH-RR
"pc .01vs. controls "p c.OSV$. nonhii'SUtehyperanctogenic
NH-H
NH-RR
NH-H
NH-RR
3a-Rndrostanediol Glucuronide (3a-GI
Statistical Analyses
Comparisons between female groups were carried out using the analysis of variance and the KruskalW allis test. Differences between the male groups were compared using an unpaired t- test.
NH-H
NH-RR
3a-Rndrostanediol Sulfate (3a-SI
RESULTS
Table 1 depicts the clinical characteristics of female controls, hyperandrogenic and androgenic alopecia patients, together with their levels of total T, unbound T, A, and DHEAS. There were no difAndrosterone Glucuronide IRoGI
1 C
H-H
H-AA
"pc0.01vs. hirsu•hyperandrogenieand hirsute alopecia
C
H-H
H-AA
•pc0 .01vs. t-~rsu•hyper..-.dfogeniclfld
-olopocia
3a-Rndrostanediol Glucuronide (3a-GI
C
H-H
H-AA
H-H
H-AA
"pc0.01vs.hirlutlhyperandrogenic:IW'Id hirSIAialopecia
~a-Rndrostanedioi-Sulfate (~a-S) 225~-------::,----,
s:
200 175 150 125
~1~~
50 25 0 C
H-H
H-AA
"p<0.01vs. hirsutthyperandrogenk:.-.d hii'SlMabpeeta
Figure 1 Serum androgen metabolites (mean ± SE) in 10 female controls (C), 18 hirsute hyperandrogenic females (H-H), and 8 hirsute androgenic alopecia females (H-AA). Vol. 62, No.4, October 1994
C
NH-H
NH-RR
"pc .Ol vscontr~s
"p c.05vs. not-hirsutehyperandrogenic
Figure 2 Serum androgen metabolites (mean ± SE) in 10 female controls (C), 12 nonhirsute hyperandrogenic females (NH-H), and 12 nonhirsute alopecia androgenetica females (NH-AA).
ferences in age or in the percentage of IBW between the groups. Both the androgenic alopecia and hyperandrogenic groups had elevated androgen levels compared with controls. Ther_e were no differences in these androgens between the androgenic alopecia and hyperandrogenic groups. Table 2 depicts the androgen conjugate levels in the three groups. Both hyperandrogenic and androgenic alopecia patients had higher levels of serum 3a-androstanediol, 3a-androstanediol sulfate, 3a-androstanediol glucuronide, androsterone glucuronide, and androsterone sulfate compared with normal controls. Serum 3a-androstanediol sulfate was higher in androgenic alopecia compared with hyperandrogenic patients. Depicted in Figure 1 are the values of the androgen metabolites for all the hirsute women and controis. The hirsute women without androgenic alopecia and hyperandrogenic androgenic alopecia patients had significantly (P < 0.01) higher serum values of 3a-androstanediol, 3a-androstanediol glucuronide, 3a-androstanediol sulfate, and androsterone glucuronide compared with the controls. Legro et al.
Androgen conjugates in alopecia
747
[
NH-H
NH-HH
• p < .05 vs. controls and nonhirsule hyperandrogenic
3a-S/~
Ratios were calculated between 3a-androstanediol and the 3a-androstanediol metabolites in the nonhirsute groups as depicted in Figure 3. Theratio of 3a-androstanediol sulfate:3a-androstanediol was higher (P < 0.05) in nonhirsute androgenic alopecia than that of both normal controls and nonhirsute hyperandrogenic patients. The most marked difference among the ratios was the ratio of 3a-androstanediol sulfate:3a-androstanediol glucuronide, which was significantly higher (P < 0.01) in nonhirsute androgenic alopecia than in the other two groups (Fig. 3). There were no differences in androgen levels between normal and premature balding males (Table 3). However, there was a significant decrease (P < 0.01) in levels of serum 3a-androstanediol glucuronide in the premature balding group. The ratio of 3a-androstanediol glucuronide:3a-androstanediol androstanediol was significantly lower in the balding group (P < 0.01) (Fig. 4). The ratio of 3aandrostanediol sulfate:3a-androstanediol glucuronide, on the other hand, was significantly higher in the premature balding group (P < 0.03) (Fig. 4).
DISCUSSION
[
NH-H
NH-HH
• p < .01 vs. controls and nonhirsute hyperandrogenic
Figure 3 Ratios (mean ± SE) between circulating 3a-androstanediol and the 3a-androstanediol conjugates in 10 female controls (C), 12 nonhirsute hyperandrogenic females (NH-H), and 12 nonhirsute alopecia androgenetic females (NH-AA).
No differences were apparent between the hirsute women without androgenic alopecia and hyperandrogenic androgenic alopecia groups. Figure 2 compares the nonhirsute patients with hyperandrogenic and androgenic alopecia patients with controls. The nonhirsute androgenic alopecia group had higher serum levels of 3a-androstanediol (P < 0.01) and 3a-androstanediol sulfate (P < 0.01) compared with controls and the hyperandrogenic group (P < 0.05 and P < 0.01, respectively). Serum androsterone sulfate, although not statistically different, was higher in nonhirsute androgenic alopecia compared with the nonhirsute hyperandrogenic and control groups. Compared with these elevations in the sulfate metabolites, there were no differences in any of the glucuronide metabolites. 748
Legro et al.
Androgen conjugates in alopecia
The pathogenesis of androgenic alopecia is not well understood, although it is generally accepted that it may be determined by an increase of scalp 5a-reductase activity (1, 6, 8). Recently, Montalto et al. (18) reported that in a group of female androgenic alopecia patients who had increased levels of unbound T, C 19 sulfate conjugates (5a-androstane3/3,17/3-diol sulfate and 3a-androstanediol sulfate) were increased in comparison with normals. They suggested that these serum androgen metabolites may be excellent markers of female androgenic alopecia and that alternative metabolic pathways of androgens, such as the metabolic pathways of DHEA, may be important in the pathogenesis of female androgenic alopecia. However, circulating levels of androgen metabolites are determined not only by peripheral enzymatic activity but also by availability of androgen precursors (8, 16, 17). Moreover, increased C 19 sulfate conjugates are not exclusive of female androgenic alopecia patients because we have observed increased serum levels of 3a-androstanediol sulfate in hyperandrogenic female patients who have hirsutism but not alopecia (8). Because of this, we speculated that a better insight into scalp androgen metabolism could be gained by comparing the values of circulating androgen metabolites in female androFertility and Sterility
Table 3
Androgen Levels in 10 Control Males and 10 Males With Premature Balding Unbound T
T
AoS*
3a*
3a-S*
3a-G*
3.7 ± 0.5 5.5 ± 0.4
16.9 ± 1.1 17.6 ± 2.2
2,230 ± 180 2,140 ± 350
0.80 ± 0.9 0.76 ± 0.7
129 ± 8.8 138 ± 20.0
29.8 ± 4.4 15.2 ± 1.6t
Male controls Premature balding
* Ao S, androsterone sulfate; 3a, 3a-androstanediol; 3a-S, 3aandrostanediol sulfate; 3a-G, 3a-androstanediol glucuronide. Values are nmol/L means ± SE.
genic alopecia patients not only with normoandrogenic controls but also with hyperandrogenic controls who have similar levels of androgenic precursors. Moreover, because the presence of hirsutism may influence the circulating values of metabolites (7), we included two groups of hyperandrogenic controls, hirsute and nonhirsute. This
3a-6/3a
C
PB
• p < .a 1 ••· .:ontrala
3a-S/3a
C
PB
3a-S/3a-6
• p c .05 ••· contrals
Figure 4 Ratios (mean± SE) between circulating 3a-androstanediol and their possible precursors in 10 male controls (C) and 10 premature balding controls (PB). Vol. 62, No.4, October 1994
t P < 0.01 versus controls.
allowed us to compare not only the larger groups but also subgroups with one another. In evaluating the profile of serum androgen metabolites, we observed that both the androgenic alopecia and hyperandrogenic group as a whole had increased levels of all androgen metabolites (3a-androstanediol, 3a-androstanediol glucuronide, 3a-androstanediol sulfate, androsterone glucuronide, and androsterone sulfate) in comparison with normal subjects. The inactivity of the androgen metabolites to distinguish between the androgenic alopecia and hyperandrogenic group is probably because of the contribution of androgen metabolites from nonscalp hirsute skin among the latter group. These data partially confirm the previous observations of Montalto et al. (18). When nonhirsute androgenic alopecia patients were compared with the nonhirsute hyperandrogenic group, they had increased levels not only of 3a-androstanediol sulfate but also of 3a-androstanediol, suggesting a general increase of 3a-androstanediol production in the scalp. Particularly interesting was the observation that, in spite of this presumed increased activity, 3a-androstanediol glucuronide was normal. This suggests that in androgenic alopecia patients, scalp 5a-reductase activity may be increased but that 3a-androstanediol is preferentially sulfurylated, perhaps because of reduced glucuronyltransferase activity. Support for this hypothesis are the data finding that in nonhirsute androgenic alopecia patients the ratio of 3aandrostanediol glucuronide:3a-androstanediol was decreased, yet that of 3a-androstanediol sulfate:3aandrostanediol was increased, and the ratio of 3aandrostanediol sulfate:3a-androstanediol glucuronide was markedly increased. The data in men are different but may be related. Although precursor androgen levels did not differ in the two groups of matched men, serum 3a-androstanediol glucuronide levels were lower in premature balding men. Similarly, the ratio of 3a-androstanediol glucuronide:unbound T and 3a-androstanediol glucuronide:3a-androstanediol ratios were lower in premature balding males. The ratio between unLegro et al. Androgen conjugates in alopecia
749
bound T and various C 19 conjugates is used in men to reflect Sa-reductase activity. In this study, our finding of a lower 3a-androstanediol glucuronide:unbound T in premature balding men should not be interpreted as this group having reduced Sareductase activity, but that 3a-androstanediol glucuronide formation may be different in premature balding. Although the 3a-androstanediol sulfate:3aandrostanediol ratio was not significantly higher in the balding group, the 3a-androstanediol sulfate:3aandrostanediol glucuronide ratio was significantly higher. Taken together, these data might suggest that in premature balding there is a reduction in the formation of 3a-androstanediol glucuronide and, therefore, a shift in the metabolic pathway toward 3a-androstanediol sulfate. The data pointing to this conjecture are clearer in the nonhirsute groups and less obvious in the men and hirsute groups, perhaps because of a more generalized increase in Sa-reduced metabolites in men and hirsute women. Compared with nonhirsute women, skin Sa-reductase activity is clearly much more elevated in men and hirsute women. In terms of the pathophysiology of scalp hair loss, it is not possible to determine if a shift away from glucuronidation and toward sulfurylation explains scalp hair loss or is the result of it. Because C 19 glucuronides are cleared from the circulation faster than the sulfates, our observed changes may merely reflect alterations in metabolism in this state of enhanced Sa-reductase activity. However, we may hypothesize here that one determinant of scalp loss in men and women may be a shift in conjugation of Sa-reduced androgens. In conclusion, our data confirm that androgenic alopecia is associated with increased scalp Sa-reductase activity among hyperandrogenic women. Women with increased androgen levels, however, constitute only a third of patients with alopecia (3). Therefore, more studies involving normoandrogenic patients will be necessary before the pathogenic cause of female androgenic alopecia can be attributed to increased Sa-reductase activity. Further studies of premature balding in the male will also serve to highlight the relative contribution of alterations in Sa-reductase activity and in the metabolism of C19 conjugates. Nevertheless, our data would suggest that a shift in C 19 metabolism away from glucuronation and toward sulfurylation may be characteristic of hair loss among hyperandrogenic individuals.
750
Legro et al.
Androgen conjugates in alopecia
REFERENCES 1. Lobo RA. Hirsutism, alopecia, and acne. In: Becker KL, Rebar RW, editors. Principles and practice of endocrinology and metabolism. Philadelphia: J.B. Lippincott Company, 1990:834-48. 2. Ludwig E. Classification of the types of androgenetic alope· cia (common baldness) occurring in the female sex. Br J Med 1977;97:247-54. 3. Futterweit W, Teh HC, Kingsley P. The prevalence of hy· perandrogenism in 109 consecutive women presenting with diffuse alopecia. JAm Acad Dermatol1988;19:831-6. 4. Sawaya ME, Honig LS, Hsia SL. Increased androgen binding capacity in sebaceous glands in scalp of male pattern baldness. J Invest Dermatol1989;92:91-5. 5. Schmidt JB, Lindmaier A, Spona J. Hormonal parameters in androgenetic hair loss in the male. Dermatologica 1991;182:214-7. 6. Pitts RL. Serum elevation of dehydroepiandrosterone sulfate associated with male pattern baldness in young men. J Am Acad Dermatol1987;16:571-3. 7. Schweikert HU, Wilson JD. Regulation of human hair growth by steroid hormones. I. Testosterone metabolism in isolated hairs. J Clin Endocrinol Metab 1974;38:811-9. 8. Matteri RK, Stanczyk FZ, Gentzschein EE, Delgado C, Lobo RA. Androgen sulfate and glucuronide conjugates in nonhirsute and hirsute women with polycystic ovarian syndrome. Am J Obstet Gynecol 1989;161:1704-9. 9. CarminaE, Stanczyk FZ, Matteri RK, Lobo RA. Serum androsterone conjugates differentiate between acne and hirsutism in hyperandrogenic women. Fertil Steril1991;55:872-6. 10. CarminaE, Lobo RA. Evidence for increased androsterone metabolism in some normoandrogenic women with acne. J Clin Endocrinol Metab 1993;76:1111-4. 11. Hamilton JB. Patterned long hair in man: types and incidence. Ann NY Acad Sci 1951;53:708-28. 12. Hatch R, Rosenfield RL, Kim MH, Tredway D. Hirsutism: implication, etiology, and management. Am J Obstet Gynecol 1981;140:815-30. 13. Lobo RA, Granger L, Goebelsmann U, Mishell DR Jr. Elevations in unbound serum estradiol as a possible mechanism for inappropriate gonadotropin secretion in women with PCO. J Clin Endocrinol Metab 1981;52:156-8. 14. Lobo RA, Kletzky OA, Kaptein EM, Goebelsmann U. Prolactin modulation of dehydroepiandrosterone sulfate secretion. Am J Obstet Gynecol 1980;138:632-6. 15. Stumpf PG, Nakamura RM, Mishell DR Jr. Changes in physiologically free circulating estradiol and testosterone during exposure to levonorgestrel. J Clin Endocrinol Metab 1981;52:138-43. 16. Matteri RK, Stanczyk FZ, Cassidenti DL, Paulson RJ, Lobo RA. The ovarian contribution to peripherally derived serum C 19 conjugates. J Clin Endocrinol Metab 1992; 75:768-72. 17. Thompson DL, Horton N, Rittmaster RS. Androsterone glucuronide is a marker of adrenal hyperandrogenism in hirsute women. Clin Endocrinol (Oxf) 1990;32:283-92. 18. Montalto J, Whorwood CB, Funder JW, Yong ABW, Callan A, Davies HE, et a!. Plasma C19 steroid sulphate levels and indices of androgen bioavailability in female pattern androgenic alopecia. Clin Endocrinol (Oxf) 1990;32:1-12.
Fertility and Sterility
Vol. 62, No.4, October 1994
Copyright e 1994 The American Fertility Society
Printed on acid-free paper in U. S. A.
Alterations in androgen conjugate levels in women and men with alopecia
RichardS. Legro, M.D.*t Enrico Carmina, M.D.:j: Frank Z. Stanczyk, Ph.D.*
Elisabet Gentzschein, B.S.* Rogerio A. Lobo, M.D.§
University of Southern California School of Medicine, Los Angeles, California, and Cattedra di Endocrinologia, Universita di Palermo, Palermo, Italy
Objective: To assess levels of androgen metabolites thought to reflect, at least in part, peripheral androgen activity in women with androgenic alopecia and men with premature balding in an effort to determine if a common abnormality exists. Design: Prospective study in various groups of women and men. Setting: Reproductive Endocrine Clinic at our university medical center. Patients: Ten normal ovulatory female controls and 50 hyperandrogenic women divided on the basis of hirsutism and alopecia as follows: [1] 8 hirsute women with androgenic alopecia; [2] 12 nonhirsute women with androgenic alopecia; [3] 18 hirsute women without androgenic alopecia; and [4]12 nonhirsute women without androgenic alopecia. Ten normal men and 10 young premature balding men matched for age and weight also were compared. Intervention: Blood was obtained from all subjects. Main Outcome Measure: Comparison of blood hormone levels in the various groups. Results: Serum T, androstenedione, and DHEAS were similarly elevated in hyperandrogenic women with and without alopecia, compared with controls. The female groups were then divided on the basis of hirsutism. Hirsute groups with and without alopecia had similarly elevated levels of unconjugated 3a-androstanediol, 3a-androstanediol glucuronide, 3a-androstanediol sulfate, androsterone glucuronide, and androsterone sulfate compared with controls. In the nonhirsute groups, androgenic alopecia patients were compared with hyperandrogenic females and cycling controls. The androgenic alopecia patients had elevated levels of 3a-androstanediol (0.75 ± 0.12 versus 0.46 ± 0.1 and 0.41 ± 0.1 nmol/L), 3a-androstanediol sulfate (200 ± 31 versus 79.6 ± 6 and 67.0 ± 4.0 nmol/L), elevated ratios of 3a-androstanediol sulfate:3a-androstanediol (267 ± 49 versus 170 ± 20 and 164 ± 49 nmol/L), elevated ratios of 3a-androstanediol sulfate:3a-androstanediol glucuronide (32.2 ± 6 versus 10.8 ± 1 and 10.0 ± 1) and lower ratios of 3a-androstanediol glucuronide:3a-androstanediol glucuronide (8.3 ± 1.8 versus 17 ± 1.7 and 15.2 ± 1.6 nmol/L). In men the premature balding group had lower levels of 3a-androstanediol glucuronide compared with the male controls (29.8 ± 4.4 versus 15.2 ± 1.6 nmol/L). Also, the ratio of 3a-androstanediol glucuronide:3a-androstanediol was significantly decreased, whereas the ratio of 3a-androstanediol sulfate:3a-androstanediol glucuronide was elevated. Conclusions: These data provided evidence confirming that enhanced Sa-reductase activity occurs in androgenic alopecia but also suggests that a disorder of androgen conjugation, favoring sulfurylation over glucuronidation, may be a characteristic feature of scalp hair loss. Fertil Steril 1994;62:744-50 Key Words: 5a-Reductase, alopecia, hyperandrogenism, steroid conjugates Received September 10, 1993; revised and accepted May 19, 1994. *Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, University of Southern California School of Medicine. t Present address: Division of Reproductive Endocrinology, Pennsylvania State University Hospital, Hershey, Pennsylvania.
744
Legro et al.
Androgen conjugates in alopecia
:j: Cattedra di Endocrinologia, Universita di Palermo. § Reprint requests: Rogerio A. Lobo, M.D., Department
of Obstetrics and Gynecology, Los Angeles County and University of Southern California Medical Center, Women's and Children's Hospital, Room 1M2, 1240 North Mission Road, Los Angeles, California 90033 (FAX: 213-2263026).
Fertility and Sterility
-l Although alopecia can be due to different causes ranging from systemic diseases to scalp autoimmunity, the most common form is male pattern baldness or alopecia androgenetica or androgenic alopecia (1). This process, which may be extremely gradual, is a transition from terminal to vellus hair (2) and is generally associated with normal androgen levels, although in some studies patients have had hyperandrogenemia. Up to 38% of female patients with alopecia have been found to have elevated androgen levels (3); alopecia, however, is not as common among hyperandrogenic women. In male patients, there have been reports of increased T levels in saliva (4), increased serum androstenedione (A) (5), and in a group of young affected males, elevated levels of DHEAS (6). However, discounting a rapid and progressive virilization because of an androgen-secreting tumor or exogenous androgens, elevated androgen levels are uncommonly linked with androgenic alopecia, which most commonly follows a hereditary pattern. Because of this, it has been hypothesized that androgen metabolism is modified in androgenic alopecia. In men with baldness, scalp hair 5a-reductase activity has been found to be increased (7). Accordingly, it has been proposed that increased scalp 5a-reductase activity may lead to higher intrafollicular dihydrotestosteorne levels that, in turn, inhibit hair follicle adenylate cyclase activity (1, 7). One way to gain further insight into the pathogenesis of androgenic alopecia is to evaluate circulating levels of androgen metabolites. Recently, we have shown that different manifestations of hyperandrogenism, specifically hirsutism and acne, exhibit different profiles of circulating androgen metabolites that reflect, at least in part, peripheral androgen action (8-10). Accordingly, in this study we have measured several serum C 19 5a- reduced metabolites in men with premature balding and in female patients with androgenic alopecia with and without other manifestations of androgen excess. Levels of the androgen metabolites were compared in various control groups. Because generalized increases in 5a-reductase activity may also lead to hirsutism in women, we have included hirsute and nonhirsute women, with and without alopecia, as additional controls. These patients all have hyperandrogenemia of a similar degree. We hypothesize that alopecia may be characterized by a different pattern of elevated c19 5a metabolites compared with hirsutism and that a similar metabolic fate may also be present in men with premature balding. Vol. 62, No.4, October 1994
MATERIALS AND METHODS Subjects
Twenty premenopausal women (aged 23.2 ± 1 years [mean ± SE], range 18 to 30 years) affected by androgenic alopecia were studied. This diagnosis was made by a dermatologist after a careful examination and the elimination of other dermatologic conditions. Eight patients also had hirsutism in conjunction with alopecia. All had increased levels of at least one circulating androgen (total T, unbound T, A, or DHEAS). None of the women previously had received hormonal treatment for their alopecia or hyperandrogenism. The women with androgenic alopecia had diffuse central alopecia with retention of the frontotemporal hairline and the severity of alopecia, as assessed by the Hamilton method, was 4 or greater (on a scale of 0 to 8) (11). No patient had evidence of masculinization or clitorimegaly. None of the patients had evidence of thyroid or adrenal dysfunction. Specifically, 21-hydroxylase deficiency was excluded based on normal early morning serum 17-hydroxyprogesterone levels. Cushing's disease also was excluded on the basis of a normal 24-hour urine for free cortisol. None of the patients were taking steroids or anticonvulsants. Ten normal women (aged 21.1 ± 1.3 years) and 30 hyperandrogenic women (aged 21.5 ± 0.6 years) without alopecia and of similar weight as those with androgenic alopecia were chosen as comparison groups or controls. The normal women were normoandrogenic women with regular menstrual cycles. They were without stigmata of hyperandrogenism. Hyperandrogenic controls presented with menstrual irregularity. The etiology of the hyperandrogenism is compatible with a diagnosis of polycystic ovarian syndrome that we refer to as hyperandrogenic chronic anovulation. Among the 30 hyperandrogenic women without alopecia, 18 were hirsute and 12 were not. Hirsutism was assessed by a modification of the Ferriman-Gallwey method (12), and only patients with a score of 8 or more were considered hirsute. Thus, despite hyperandrogenemia in all 30 women, the presence of hirsutism in only 18 may be explained by the existence of increased skin 5a-reductase activity in these women (1, 8). All patients and controls were studied during the midfollicular phase of a spontaneous cycle (days 5 to 8) or after a P-induced menses. Male patients with premature baldness (aged 27.7 ± 0.5, range 26 to 30, 100% ideal body weight Legro et al. Androgen conjugates in alopecia
745
· Table 1 Clinical Data and Androgen Levels in 10 Female Controls, 30 Female Hyperandrogenic Patients Without Alopecia and 20 Females Androgenic Alopecia Patients*
Female controls Hyperandrogenic no alopecia Androgenic alopecia
Age
IBW
y
%
21.1 ± 1.3 21.5 ± 0.6 23.2 ± 1.1
107 ± 3 114 ± 3.1 111 ± 3.9
* Values are means ± SE. t P < 0.01 versus controls.
A
Total T
DHEAS
nmol!L
5.3 ± 0.7 14.7 ± 0.7t 13.9 ± 0.7t
Jimol/L
1.03 ± 0.1 2.74 ± 0.2t 2.65 ± 0.1t
4.5 ± 0.5 6.9 ± 0.5:j: 7.8 ± 0.8:j:
:j: P < 0.05 versus controls.
[IBW]) were age and weight matched to control nonbalding males (age 29.1 ± 0.9, range 25 to 33, 99% IBW). None of the balding men were receiving any therapy for alopecia. None of the subjects were affected by adrenal or thyroid disease. Balding males were judged according to the Hamilton method (11), and scores ranged from 4 to 8. Study Design
Our study design for the female groups was based on clearly delineating diverse causes of a common phenotype. Although we hypothesized increased 5a-reductase activity as the cause of androgenic alopecia, androgenic alopecia can also result from an elevated production of androgen precursors (primarily ovarian based on our screening). Therefore, we sought hyperandrogenic women without alopecia as a comparison group in addition to normoandrogenic cycling women (Table 1). Additionally, our main outcome variables, serum androgen metabolites of 5a-reductase activity, can be elevated not only by increased scalp activity but also by a generalized increased in skin activity as is the case in hirsutism (Table 2) (8). Therefore, when examining these metabolites, we sought to control for this
confounding variable by subdividing the probands and hyperandrogenic patient groups on the basis of hirsutism (Fig. 1 for the hirsute groups and Figure 2 for the nonhirsute groups). The same female normal controls with normal female hair distribution were used in both latter comparisons. Thus, Figures 1 and 2 represent a subdivision of Table 2 on the basis of hirsutism. Finally, we explored ratios of dihydrotestosterone metabolites and conjugates in the groups in which we hypothesized their production was primarily scalp dependent: the nonhirsute groups (Fig. 3). It was not necessary to further subdivide our male groups because both control males and those with premature balding had similar levels of circulating androgen precursors and similar peripheral hair distributions. Protocol
All patients had fasting blood obtained between 8:00 and 10:00 A.M. after at least 0.5 hour of recumbent rest. Sera from all patients were separated and stored at -20°C until assayed for total T, A, and DHEAS by established RIA techniques (13-15). Unbound Twas measured only in men by ammonium sulfate precipitation of sex hormone-binding
Table 2 Serum Androgen Metabolites in 10 Female Controls, 30 Hyperandrogenic Patients Without Alopecia, and 20 Female Androgenic Alopecia Patients
Females controls Hyperandrogenic no alopeciat Androgenic alopeciat
3a*
3a-G*
3a-S*
AoG*
Ao S*
0.41 ± 0.03 0.67 ± 0.07:j: 0.69 ± 0.06:j:
6.2 ± 0.47 13.2 ± 1.3:j: 11.1 ± 2.3§
67.2 ± 4.2 125.5 ± 15.9:j: 197.6 ± 27.3§11
73 ± 4.6 151.1 ± 15.4:j: 158.7 ± 27.8:j:
2,038 ± 200 2,724 ± 297§ 3,618 ± 686:j:
* 3a, 3a-androstanediol; 3a-G, 3a-androstanediol glucuronide; 3a-S, 3a-androstanediol sulfate; Ao G, androsterone glucuronide; Ao S, androsterone sulfate. Values are nmol/L means ± SE. t The hyperandrogenic and alopecia groups have been further subdivided on the basis of hirsutism in Figure 1 (hirsute) and Figure 2 (nonhirsute). 746
Legro et al.
Androgen conjugates in alopecia
:j: P < 0.01 versus controls. § P < 0.05 versus controls. II P < 0.05 versus hyperandrogenic.
Fertility and Sterility
globulin as previously described (15). In addition, the following 5a- reduced androgens were measured: [1] unconjugated 3a-androstanediol; [2] 3aandrostanediol glucuronide; [3] 3a-androstanediol sulfate; [4] androsterone glucuronide; and [5] androsterone sulfate. These assays employed specific hydrolysis of the conjugates, followed by extraction, celite column chromatography, and RIA as previously described (8, 16, 17). In all cases, intraassay and interassay coefficients of variations did not exceed 10% and 12%, respectively.
3a --llndrostanediol (3al
C
NH-H
Androsterone Glucuronide IRoGI
NH-RR
"pc .01vs. controls "p c.OSV$. nonhii'SUtehyperanctogenic
NH-H
NH-RR
NH-H
NH-RR
3a-Rndrostanediol Glucuronide (3a-GI
Statistical Analyses
Comparisons between female groups were carried out using the analysis of variance and the KruskalW allis test. Differences between the male groups were compared using an unpaired t- test.
NH-H
NH-RR
3a-Rndrostanediol Sulfate (3a-SI
RESULTS
Table 1 depicts the clinical characteristics of female controls, hyperandrogenic and androgenic alopecia patients, together with their levels of total T, unbound T, A, and DHEAS. There were no difAndrosterone Glucuronide IRoGI
1 C
H-H
H-AA
"pc0.01vs. hirsu•hyperandrogenieand hirsute alopecia
C
H-H
H-AA
•pc0 .01vs. t-~rsu•hyper..-.dfogeniclfld
-olopocia
3a-Rndrostanediol Glucuronide (3a-GI
C
H-H
H-AA
H-H
H-AA
"pc0.01vs.hirlutlhyperandrogenic:IW'Id hirSIAialopecia
~a-Rndrostanedioi-Sulfate (~a-S) 225~-------::,----,
s:
200 175 150 125
~1~~
50 25 0 C
H-H
H-AA
"p<0.01vs. hirsutthyperandrogenk:.-.d hii'SlMabpeeta
Figure 1 Serum androgen metabolites (mean ± SE) in 10 female controls (C), 18 hirsute hyperandrogenic females (H-H), and 8 hirsute androgenic alopecia females (H-AA). Vol. 62, No.4, October 1994
C
NH-H
NH-RR
"pc .Ol vscontr~s
"p c.05vs. not-hirsutehyperandrogenic
Figure 2 Serum androgen metabolites (mean ± SE) in 10 female controls (C), 12 nonhirsute hyperandrogenic females (NH-H), and 12 nonhirsute alopecia androgenetica females (NH-AA).
ferences in age or in the percentage of IBW between the groups. Both the androgenic alopecia and hyperandrogenic groups had elevated androgen levels compared with controls. Ther_e were no differences in these androgens between the androgenic alopecia and hyperandrogenic groups. Table 2 depicts the androgen conjugate levels in the three groups. Both hyperandrogenic and androgenic alopecia patients had higher levels of serum 3a-androstanediol, 3a-androstanediol sulfate, 3a-androstanediol glucuronide, androsterone glucuronide, and androsterone sulfate compared with normal controls. Serum 3a-androstanediol sulfate was higher in androgenic alopecia compared with hyperandrogenic patients. Depicted in Figure 1 are the values of the androgen metabolites for all the hirsute women and controis. The hirsute women without androgenic alopecia and hyperandrogenic androgenic alopecia patients had significantly (P < 0.01) higher serum values of 3a-androstanediol, 3a-androstanediol glucuronide, 3a-androstanediol sulfate, and androsterone glucuronide compared with the controls. Legro et al.
Androgen conjugates in alopecia
747
[
NH-H
NH-HH
• p < .05 vs. controls and nonhirsule hyperandrogenic
3a-S/~
Ratios were calculated between 3a-androstanediol and the 3a-androstanediol metabolites in the nonhirsute groups as depicted in Figure 3. Theratio of 3a-androstanediol sulfate:3a-androstanediol was higher (P < 0.05) in nonhirsute androgenic alopecia than that of both normal controls and nonhirsute hyperandrogenic patients. The most marked difference among the ratios was the ratio of 3a-androstanediol sulfate:3a-androstanediol glucuronide, which was significantly higher (P < 0.01) in nonhirsute androgenic alopecia than in the other two groups (Fig. 3). There were no differences in androgen levels between normal and premature balding males (Table 3). However, there was a significant decrease (P < 0.01) in levels of serum 3a-androstanediol glucuronide in the premature balding group. The ratio of 3a-androstanediol glucuronide:3a-androstanediol androstanediol was significantly lower in the balding group (P < 0.01) (Fig. 4). The ratio of 3aandrostanediol sulfate:3a-androstanediol glucuronide, on the other hand, was significantly higher in the premature balding group (P < 0.03) (Fig. 4).
DISCUSSION
[
NH-H
NH-HH
• p < .01 vs. controls and nonhirsute hyperandrogenic
Figure 3 Ratios (mean ± SE) between circulating 3a-androstanediol and the 3a-androstanediol conjugates in 10 female controls (C), 12 nonhirsute hyperandrogenic females (NH-H), and 12 nonhirsute alopecia androgenetic females (NH-AA).
No differences were apparent between the hirsute women without androgenic alopecia and hyperandrogenic androgenic alopecia groups. Figure 2 compares the nonhirsute patients with hyperandrogenic and androgenic alopecia patients with controls. The nonhirsute androgenic alopecia group had higher serum levels of 3a-androstanediol (P < 0.01) and 3a-androstanediol sulfate (P < 0.01) compared with controls and the hyperandrogenic group (P < 0.05 and P < 0.01, respectively). Serum androsterone sulfate, although not statistically different, was higher in nonhirsute androgenic alopecia compared with the nonhirsute hyperandrogenic and control groups. Compared with these elevations in the sulfate metabolites, there were no differences in any of the glucuronide metabolites. 748
Legro et al.
Androgen conjugates in alopecia
The pathogenesis of androgenic alopecia is not well understood, although it is generally accepted that it may be determined by an increase of scalp 5a-reductase activity (1, 6, 8). Recently, Montalto et al. (18) reported that in a group of female androgenic alopecia patients who had increased levels of unbound T, C 19 sulfate conjugates (5a-androstane3/3,17/3-diol sulfate and 3a-androstanediol sulfate) were increased in comparison with normals. They suggested that these serum androgen metabolites may be excellent markers of female androgenic alopecia and that alternative metabolic pathways of androgens, such as the metabolic pathways of DHEA, may be important in the pathogenesis of female androgenic alopecia. However, circulating levels of androgen metabolites are determined not only by peripheral enzymatic activity but also by availability of androgen precursors (8, 16, 17). Moreover, increased C 19 sulfate conjugates are not exclusive of female androgenic alopecia patients because we have observed increased serum levels of 3a-androstanediol sulfate in hyperandrogenic female patients who have hirsutism but not alopecia (8). Because of this, we speculated that a better insight into scalp androgen metabolism could be gained by comparing the values of circulating androgen metabolites in female androFertility and Sterility
Table 3
Androgen Levels in 10 Control Males and 10 Males With Premature Balding Unbound T
T
AoS*
3a*
3a-S*
3a-G*
3.7 ± 0.5 5.5 ± 0.4
16.9 ± 1.1 17.6 ± 2.2
2,230 ± 180 2,140 ± 350
0.80 ± 0.9 0.76 ± 0.7
129 ± 8.8 138 ± 20.0
29.8 ± 4.4 15.2 ± 1.6t
Male controls Premature balding
* Ao S, androsterone sulfate; 3a, 3a-androstanediol; 3a-S, 3aandrostanediol sulfate; 3a-G, 3a-androstanediol glucuronide. Values are nmol/L means ± SE.
genic alopecia patients not only with normoandrogenic controls but also with hyperandrogenic controls who have similar levels of androgenic precursors. Moreover, because the presence of hirsutism may influence the circulating values of metabolites (7), we included two groups of hyperandrogenic controls, hirsute and nonhirsute. This
3a-6/3a
C
PB
• p < .a 1 ••· .:ontrala
3a-S/3a
C
PB
3a-S/3a-6
• p c .05 ••· contrals
Figure 4 Ratios (mean± SE) between circulating 3a-androstanediol and their possible precursors in 10 male controls (C) and 10 premature balding controls (PB). Vol. 62, No.4, October 1994
t P < 0.01 versus controls.
allowed us to compare not only the larger groups but also subgroups with one another. In evaluating the profile of serum androgen metabolites, we observed that both the androgenic alopecia and hyperandrogenic group as a whole had increased levels of all androgen metabolites (3a-androstanediol, 3a-androstanediol glucuronide, 3a-androstanediol sulfate, androsterone glucuronide, and androsterone sulfate) in comparison with normal subjects. The inactivity of the androgen metabolites to distinguish between the androgenic alopecia and hyperandrogenic group is probably because of the contribution of androgen metabolites from nonscalp hirsute skin among the latter group. These data partially confirm the previous observations of Montalto et al. (18). When nonhirsute androgenic alopecia patients were compared with the nonhirsute hyperandrogenic group, they had increased levels not only of 3a-androstanediol sulfate but also of 3a-androstanediol, suggesting a general increase of 3a-androstanediol production in the scalp. Particularly interesting was the observation that, in spite of this presumed increased activity, 3a-androstanediol glucuronide was normal. This suggests that in androgenic alopecia patients, scalp 5a-reductase activity may be increased but that 3a-androstanediol is preferentially sulfurylated, perhaps because of reduced glucuronyltransferase activity. Support for this hypothesis are the data finding that in nonhirsute androgenic alopecia patients the ratio of 3aandrostanediol glucuronide:3a-androstanediol was decreased, yet that of 3a-androstanediol sulfate:3aandrostanediol was increased, and the ratio of 3aandrostanediol sulfate:3a-androstanediol glucuronide was markedly increased. The data in men are different but may be related. Although precursor androgen levels did not differ in the two groups of matched men, serum 3a-androstanediol glucuronide levels were lower in premature balding men. Similarly, the ratio of 3a-androstanediol glucuronide:unbound T and 3a-androstanediol glucuronide:3a-androstanediol ratios were lower in premature balding males. The ratio between unLegro et al. Androgen conjugates in alopecia
749
bound T and various C 19 conjugates is used in men to reflect Sa-reductase activity. In this study, our finding of a lower 3a-androstanediol glucuronide:unbound T in premature balding men should not be interpreted as this group having reduced Sareductase activity, but that 3a-androstanediol glucuronide formation may be different in premature balding. Although the 3a-androstanediol sulfate:3aandrostanediol ratio was not significantly higher in the balding group, the 3a-androstanediol sulfate:3aandrostanediol glucuronide ratio was significantly higher. Taken together, these data might suggest that in premature balding there is a reduction in the formation of 3a-androstanediol glucuronide and, therefore, a shift in the metabolic pathway toward 3a-androstanediol sulfate. The data pointing to this conjecture are clearer in the nonhirsute groups and less obvious in the men and hirsute groups, perhaps because of a more generalized increase in Sa-reduced metabolites in men and hirsute women. Compared with nonhirsute women, skin Sa-reductase activity is clearly much more elevated in men and hirsute women. In terms of the pathophysiology of scalp hair loss, it is not possible to determine if a shift away from glucuronidation and toward sulfurylation explains scalp hair loss or is the result of it. Because C 19 glucuronides are cleared from the circulation faster than the sulfates, our observed changes may merely reflect alterations in metabolism in this state of enhanced Sa-reductase activity. However, we may hypothesize here that one determinant of scalp loss in men and women may be a shift in conjugation of Sa-reduced androgens. In conclusion, our data confirm that androgenic alopecia is associated with increased scalp Sa-reductase activity among hyperandrogenic women. Women with increased androgen levels, however, constitute only a third of patients with alopecia (3). Therefore, more studies involving normoandrogenic patients will be necessary before the pathogenic cause of female androgenic alopecia can be attributed to increased Sa-reductase activity. Further studies of premature balding in the male will also serve to highlight the relative contribution of alterations in Sa-reductase activity and in the metabolism of C19 conjugates. Nevertheless, our data would suggest that a shift in C 19 metabolism away from glucuronation and toward sulfurylation may be characteristic of hair loss among hyperandrogenic individuals.
750
Legro et al.
Androgen conjugates in alopecia
REFERENCES 1. Lobo RA. Hirsutism, alopecia, and acne. In: Becker KL, Rebar RW, editors. Principles and practice of endocrinology and metabolism. Philadelphia: J.B. Lippincott Company, 1990:834-48. 2. Ludwig E. Classification of the types of androgenetic alope· cia (common baldness) occurring in the female sex. Br J Med 1977;97:247-54. 3. Futterweit W, Teh HC, Kingsley P. The prevalence of hy· perandrogenism in 109 consecutive women presenting with diffuse alopecia. JAm Acad Dermatol1988;19:831-6. 4. Sawaya ME, Honig LS, Hsia SL. Increased androgen binding capacity in sebaceous glands in scalp of male pattern baldness. J Invest Dermatol1989;92:91-5. 5. Schmidt JB, Lindmaier A, Spona J. Hormonal parameters in androgenetic hair loss in the male. Dermatologica 1991;182:214-7. 6. Pitts RL. Serum elevation of dehydroepiandrosterone sulfate associated with male pattern baldness in young men. J Am Acad Dermatol1987;16:571-3. 7. Schweikert HU, Wilson JD. Regulation of human hair growth by steroid hormones. I. Testosterone metabolism in isolated hairs. J Clin Endocrinol Metab 1974;38:811-9. 8. Matteri RK, Stanczyk FZ, Gentzschein EE, Delgado C, Lobo RA. Androgen sulfate and glucuronide conjugates in nonhirsute and hirsute women with polycystic ovarian syndrome. Am J Obstet Gynecol 1989;161:1704-9. 9. CarminaE, Stanczyk FZ, Matteri RK, Lobo RA. Serum androsterone conjugates differentiate between acne and hirsutism in hyperandrogenic women. Fertil Steril1991;55:872-6. 10. CarminaE, Lobo RA. Evidence for increased androsterone metabolism in some normoandrogenic women with acne. J Clin Endocrinol Metab 1993;76:1111-4. 11. Hamilton JB. Patterned long hair in man: types and incidence. Ann NY Acad Sci 1951;53:708-28. 12. Hatch R, Rosenfield RL, Kim MH, Tredway D. Hirsutism: implication, etiology, and management. Am J Obstet Gynecol 1981;140:815-30. 13. Lobo RA, Granger L, Goebelsmann U, Mishell DR Jr. Elevations in unbound serum estradiol as a possible mechanism for inappropriate gonadotropin secretion in women with PCO. J Clin Endocrinol Metab 1981;52:156-8. 14. Lobo RA, Kletzky OA, Kaptein EM, Goebelsmann U. Prolactin modulation of dehydroepiandrosterone sulfate secretion. Am J Obstet Gynecol 1980;138:632-6. 15. Stumpf PG, Nakamura RM, Mishell DR Jr. Changes in physiologically free circulating estradiol and testosterone during exposure to levonorgestrel. J Clin Endocrinol Metab 1981;52:138-43. 16. Matteri RK, Stanczyk FZ, Cassidenti DL, Paulson RJ, Lobo RA. The ovarian contribution to peripherally derived serum C 19 conjugates. J Clin Endocrinol Metab 1992; 75:768-72. 17. Thompson DL, Horton N, Rittmaster RS. Androsterone glucuronide is a marker of adrenal hyperandrogenism in hirsute women. Clin Endocrinol (Oxf) 1990;32:283-92. 18. Montalto J, Whorwood CB, Funder JW, Yong ABW, Callan A, Davies HE, et a!. Plasma C19 steroid sulphate levels and indices of androgen bioavailability in female pattern androgenic alopecia. Clin Endocrinol (Oxf) 1990;32:1-12.
Fertility and Sterility