Idiopathic hirsutism: local and peripheral expression of (CYP19A)and5a-reductase (SRD5A1 and SRD5A2)

A. Okay Caglayan, M.D.,a Munis Dundar, M.D.,a Fatih Tanriverdi, M.D.,b Nuran A. Baysal, M.D.,b Kursad Unluhizarci, M.D.,b Yusuf Ozkul, Ph.D.,a Murat Borlu, M.D.,c Cem Batukan, M.D.,d and Fahrettin Kelestimur, M.D.b a Department of Medical Genetics, b Department of Endocrinology, c Department of Dermatology, and d Department of Obstetrics and Gynecology, School of Medicine, Erciyes University, Kayseri, Turkey

Objective: To evaluate idiopathic hirsutism etiology via molecular studies testing peripheral and local aromatase and 5a-reductase expression. Design: Assessment of the expression of messenger RNA (mRNA) for type 1 and 2,5a-reductase isoenzyme (SDR5A1, SDR5A2) and aromatase (CYP19A) in dermal papillae cells and peripheral blood mononuclear cells. Setting: University hospital. Patient(s): 28 untreated idiopathic hirsute patients and 20 healthy women (controls). Intervention(s): Human skin biopsies and peripheral venous blood. Main Outcome Measure(s): SDR5A1, SDR5A2, CYP19A in skin biopsies and peripheral blood. Result(s): A statistically significant reduction of SRD5A1, SRD5A2, and CYP19A gene expression was found in the dermal papillae cells and peripheral blood mononuclear cell between the study and control group. Conclusion(s): Further study, including protein expression and activity assays, are warranted to character- ize the paradoxically low gene expression levels of local 5a-reductase and aromatase in women with idiopathic hirsutism. (Fertil Steril 2011;96:479–82. 2011 by American Society for Reproductive Medicine.) Key Words: 5a-reductase, aromatase, expression, idiopathic hirsutism

Idiopathic hirsutism (IH) is defined as the presence of hirsutism in region of the skin and in the peripheral blood of women with conjunction with regular ovulation and normal levels, idiopathic hirsutism and healthy controls. Our results provide a condition found in only 5% to 15% of consecutive hirsute women insight into the regulation of androgen action in human hair (1). Little information is available regarding the pathogenesis of IH. follicles by local androgen modification at the target cell level. Increased peripheral 5a-reductase activity and androgen receptor gene polymorphisms have been postulated to explain the pathogene- MATERIALS AND METHODS sis of IH (2–4). Recent studies have shown that IH may be associated Patients with some degree of insulin resistance and an increased tendency for Forty-five women with idiopathic hirsutism and 24 control women matched glucose intolerance notably in obese women with IH (5). for race and ethnicity, ranging in age from 16 to 35 years, were selected for The dermal papilla plays a key role in the control of hair growth, the study. The criterion for case selection was having IH plus giving informed and dermal papilla cells (DPCs) seem to be the primary target of an- consent for skin biopsies. None of the control women had ovarian dysfunc- drogen regulation of hair growth (6, 7). Therefore, an alteration in tion, hirsutism, hyperandrogenemia, or had been taking any drugs known to the local regulation of 5a-reductase activity could cause of interfere with androgen, , or gonadotropin serum levels for at least 6 idiopathic hirsutism (1). We additionally investigated peripheral months before the study. The purpose of the protocol was explained to all and local CYP19A (aromatase cytochrome P450) gene expression participants, and informed consent was obtained before the study initiation. This study was approved by the ethics committee and the institutional review to detect potential local imbalances between the androgen and board of Erciyes University Medical School. estrogen levels in women with normal serum androgen levels The diagnosis of IH was based on the presence of hirsutism (modified a (8, 9). We profiled the expression of 5 -reductase and aromatase Ferriman-Gallwey score >8), regular ovulatory menstrual cycles, normal messenger RNA (mRNA) in DPCs from the lower abdominal ovarian morphology, and a normal serum androgen profile, including total and free , , and dehydroepiandrosterone sulfate (DHEAS). We also measured serum levels of follicle-stimulating hormone Received December 24, 2010; revised May 6, 2011; accepted May 13, (FSH), luteinizing hormone (LH), 17a-hydroxyprogesterone (17-OHP), 2011; published online June 15, 2011. and thyroid hormone in all patients. Normal androgen levels were defined A.O.C. has nothing to disclose. M.D. has nothing to disclose. F.T. has according to the reference values of the commercial kits: serum free testoster- nothing to disclose. N.A.B. has nothing to disclose. K.U. has nothing < < < to disclose. Y.O. has nothing to disclose. M.B. has nothing to disclose. one 3.18 pg/mL, androstenedione 3 ng/mL, and DHEAS 3,330 ng/mL. C.B. has nothing to disclose. F.K. has nothing to disclose. Thyroid dysfunction, hyperprolactinemia, nonclassic adrenal hyperplasia, Funded by Erciyes University Research Funding and the Scientific and polycystic ovary syndrome (PCOS), and adrenal/ovarian tumors were Technological Research Council of Turkey (TUB€ _ITAK, Project No: excluded by appropriate blood tests and ultrasonography of the ovaries and SBAG-106S170). adrenal glands. None of the women had a history of taking medication known Reprint requests: Munis Dundar, M.D., Head, Department of Medical to interfere with carbohydrate or to cause hirsutism. Ovulation Genetics, Erciyes University Medical Faculty, 38039, Kayseri, Turkey was confirmed by the day-24 serum progesterone levels (>5 nmol/L) in (E-mail: [email protected]). the patients and healthy controls. The patients and the healthy women

0015-0282/$36.00 Fertility and Sterility Vol. 96, No. 2, August 2011 479 doi:10.1016/j.fertnstert.2011.05.040 Copyright ª2011 American Society for Reproductive Medicine, Published by Elsevier Inc. TABLE 1 Sequence of the primers and probes and Tm values related to CYP19A, SRD5A1, and SRD5A2 genes and expected sizes of amplified fragments for expression analysis.

Gene Sequence (50L30) Tm Size of fragment

CYP19A Sense GAGGGACAGGAAGAGGAGAAG 71.9 108 bp Antisense CTATACAGGCTAAGCAGAGTGAAA Probe TGAACGATCCAAGCAAACTCTCCCAGAATCgttca SRD5A1 Sense GTGTATGCTGATGACTGGGTAA 71.4 120 bp Antisense TCCTGGTTTTCTGAGATTCCTTAG Probe CCACAAGCCAAAACCTATTAGAAAACGGGGacttgtgg SRD5A2 Sense GGCAATATCCAAATAATGAGTAGTGT 70.3 109 bp Antisense GGGCTTCTGCTGTACTTCATAT Probe TGCTTGCCTCCACCAGATGTCCACAagca

Caglayan. Aromatase 5a-reductase hirsutism. Fertil Steril 2011.

were studied in the follicular phase (days 2 to 9) of their cycles. Serum sam- examined by melt curve analysis and agarose gel electrophoresis after each ples for assessing hormone levels were drawn after an overnight fast in the real-time PCR run. For analysis of quantitative results, the relative quantifi- follicular phase and were stored at 20C until assayed. cation method was performed by software provided by Corbett. Serum FSH (ACS 180; Bayer), LH (ACS 180; Bayer), and estradiol (ACS 180; Bayer) levels were determined by an automated chemiluminescence system; free and total testosterone (DSL-4900; Diagnostic Systems Labora- Statistical Analysis tories, Inc.), DHEAS (Immunotech), and androstenedione (Immunotech) All analyses were performed using the Statistical Package for the Social Sci- levels were measured by radioimmunoassay (RIA); the – ences (SPSS, Inc.). Continuous variables with normal distribution (age, body binding globulin (SHBG) level was measured by immunoradiometric assay mass index) were analyzed as mean values standard deviation and ranges; (Orion Diagnostica, Finland), using commercial kits. continuous variables without normal distribution were analyzed as median values and interquartile range. As continuous variables were without normal distribution, we analyzed differences by Mann-Whitney U test. Correlations Preparation of Hair Follicle Subunits among Ferriman-Gallwey score in the subgroup of hirsute women were Punch biopsies were taken under local anesthesia from the midline subumbil- determined by using Spearman’s test. P<.05 was considered statistically sig- ical area of healthy volunteers and hirsute patients. Intact anagen hair follicles nificant. For categorical variables, differences were analyzed by means of the were isolated from the subcutaneous fat according to standard procedures chi-square test and Fisher’s exact test when appropriate. (10) and were placed in Ham’s F-10 medium supplemented with 2 mM L-glutamine, penicillin (100 IU/mL), and streptomycin (100 mg/mL) and containing diethylpyrocarbonate and were transported to the laboratory for RESULTS assays. Total RNA was immediately extracted from the isolated papilla. Fifteen patients and four controls did not meet our selection criteria and were disqualified from the study. The isolation of dermal papil- Quantitative Real-Time RT-PCR Protocol lae and/or RNA extraction was unsuccessful for technical reasons in two cases, so the final study included of 28 patients and 20 controls. Total RNA was extracted from patient and control dermal papillae with Mo- The patients and controls did not differ in mean age (21.7 5.94 and bio (BiOstic Blood Total RNA Isolation Kit) and was extracted from periph- eral blood mononuclear cell (PMBCs) with Qiagene (RNeasy midi kit, Qiagen), according to the manufacturers’ instructions. The RNA was dissolved in diethylpyrocarbonate-treated, double-distilled TABLE 2 water and was stored at 80C. We used the First Strand cDNA Synthesis Epidemiologic characteristics of the patients with Kit (Fermentas) to reverse transcribe 2 mg of total RNA, according to the manufacturer’s instructions. The RNA integrity was verified by examining idiopathic hirsutism and controls. the real-time reverse-transcriptase polymerase chain reaction (RT-PCR) product of b-actin mRNA. The resultant complementary DNA (cDNA) Patients with Controls [ [ from each woman was submitted to PCR by use of primer pairs for Characteristic IH (± SD) (n 28) (± SD) (n 20) b SRD5A1, SDR5A2, and CYP19A (target genes) and -actin (reference Gender Female Female gene). The primer sequences are listed in Table 1. Race Caucasian Caucasian m Quantitative PCR reactions were performed in duplicate with 5 L of di- Ethnicity Turkish Turkish m luted cDNA template in a 20- L reaction containing Precision Mastermix Age (y) 21.7 5.94 22.6 0.4 (PrimerDesign), 300 nM primers, and 200 nM probe using the Rotor Gene BMI (kg/m2) 26.6 5.75 24.5 0.2 6000 Real-Time PCR Machine (Corbett). We used the TaqMan system and mFG score 17.1 4.36 1.9 0.6 internal positive and negative controls for each gene and designed primers and probes with Primer Express software (PrimerDesign Ltd) (Table 2). Note: BMI ¼ body mass index; IH ¼ idiopathic hirsutism; mFG ¼ mod- The PCR conditions were initial denaturation for 10 minutes at 95C, 45 am- ified Ferriman-Gallwey score; SD ¼ standard deviation. plification cycles (15 seconds at 95 C, 30 seconds at 50 C, and 15 seconds at Caglayan. Aromatase 5a-reductase hirsutism. Fertil Steril 2011. 72C), and final denaturation for 10 minutes at 72C. Product specificity was

480 Caglayan et al. Aromatase 5a-reductase hirsutism Vol. 96, No. 2, August 2011 TABLE 3 different androgen sensitivities of hair follicles at different body sites. The dermal papilla contains androgen receptors and the Hormone characteristics of the patient group. 17b-hydroxysteroid dehydrogenase (17b-HSD) and 5a- reductase isotypes type 1 and 2. Both of these isotypes convert Patients with Control group testosterone to the more active androgen 5a-, Hormone IH (± SD) (n [ 28) (± SD) (n [ 20) which can in turn be metabolized to 5a- by the FSH (mIU/mL) 5.4 2.22 4.9 1.5 action of 17b-HSD and further converted to 5a-androsterone by LH (mIU/mL) 3.1 1.52 3.88 2.4 the enzyme 3b-hydroxysteroid dehydrogenase (3b-HSD). Addition- E2 (pg/mL) 48.8 24.68 46.2 19.7 ally, testosterone can be metabolized to estradiol via a sequence of Total T (ng/dL) 29.2 15.51 32.4 10.1 reactions involving the enzyme aromatase. FT (pg/mL) 1.5 0.52 1.24 0.82 A well-known hypothesis suggests that the activity of 5a-reduc- DHEAS (ng/mL) 2,106.2 1,103.4 2,600.9 500.7 tase enzyme is increased in skin with hirsutism. Previously, Serafini 17-OHP (ng/mg) 0.5 0.7 0.74 0.3 and Lobo (11) suggested that 5a-reductase skin activity is increased SHBG (nmol/L) 13.8 8.7 22.1 14.5 by an androgen-independent mechanism that may be genetically A (ng/mL) 1.5 0.7 1.9 0.2 determined. Our study was designed to assess the expression of Note: 17-OHP ¼ 17a-hydroxyprogesterone; A ¼ androstenedione; 5a-reductase type 1 and type 2 mRNA in blood and dermal papilla ¼ ¼ ¼ DHEAS dehydroepiandrosterone sulfate; E2 estradiol; FSH cells from the lower midline abdominal skin in women with hirsut- follicle-stimulating hormone; fT ¼ free testosterone; IH ¼ idiopathic hirsutism; LH ¼ luteinizing hormone; SD ¼ standard deviation; ism. Our results showed decreased expression of SRD5A1 and SHBG ¼ sex hormone–binding globulin; T ¼ testosterone. SRD5A2 in the peripheral blood and dermal papillae cells of patients versus controls. It has to be emphasized that gene expression and Caglayan. Aromatase 5a-reductase hirsutism. Fertil Steril 2011. enzyme activity are not equivalent. Previously, Sultan et al. (12) and Oliveira et al. (13) used RT-PCR 22.6 0.4 years, respectively) or body mass index (26.6 5.75 and to investigate SRD5A1 and SRD5A2 expression in plucked scalp 24.5 0.2 kg/m2, respectively) values, whereas the modified hairs from hirsute patients (13 idiopathic and 20 with PCOS) and Ferriman-Gallwey score was statistically significantly (P<.001) controls (10 men and 15 nonhirsute women). Their study did not de- higher in the patients with IH (17.1 4.3) than in the healthy women tect SRD5A2 expression in hair samples and found no differences in (1.9 0.6) (Table 2). The hormonal profile of the patient group is SRD5A1 expression between any of the groups. With these results, shown in Table 3. they argued that SRD5A1 gene expression is not responsible for The mRNA expression levels of SRD5A1, SRD5A2, and CYP19A differences in hair growth in normal men, normal women, or hirsute were statistically significantly lower in tissue and PMBCs from pa- patients, with the caveat that more detailed studies including other tients when compared with controls (Table 4). Within the patient follicular compartments would be necessary (13). group, the mRNA expression levels of SRD5A1, SRD5A2, and There are several differences in study design that can explain the CYP19A genes were statistically significantly higher in tissue than contrast between our results and those of Oliveira et al. (13). They in PMBCs (P<.05) (Table 4). There was no statistically significant analyzed SRD5A expression in plucked anagen hairs but not in the correlation between serum hormone values and SRD5A1, SRD5A2, connective tissue sheath, the lower bulb, the dermal papilla cells, or CYPA1 gene expression in patient tissues or PBMCs. and the sebaceous gland; we isolated the dermal papillae. The scalp hair follicle cells that they studied are generally not considered to be a main target for hirsutism, which is why we focused on subumbil- DISCUSSION ical hair. In addition, we used a more sensitive molecular method to are important regulators of human hair growth, but the assess gene expression, real-time PCR rather than reverse- mechanism of androgen action is not completely understood. transcriptase PCR. Our study also included a larger number of IH Assuming that the specific response of hair follicles to androgens patients (28 vs. 13). Alternatively, androgens may have different depends on androgen metabolism in the target cells, differences in effects on gene transcription, translation, or posttranslational intracellular androgen metabolism may be responsible for the processes in various body sites or even within same follicle.

TABLE 4 SRD5A1. SRD5A2, and CYP19A gene expression values in tissue and peripheral blood mononuclear cells (PBMC) in patient and control groups.

SRD5A1/b-actin Median SRD5A2/b-actin Median CYP19A/b-actin Median Gene (25p–75p) 3 10L3 (25p–75p) 3 10L3 (25p–75p) 3 10L3

Tissue expression Control 246.84 3,358.27 1,953.12 Patient 7.523 4.06 6.16 P value <.001 <.001 <.001 PBMC expression Control 84.29 956.06 452.7 Patient 1.51 1.78 2.41 P value <.001 <.001 <.001

Caglayan. Aromatase 5a-reductase hirsutism. Fertil Steril 2011.

Fertility and Sterility 481 Ska1ba et al. (14) investigated mRNA expression of SRD5A1 and some patients with IH present with a very mild form of insulin SRD5A2 in dermal papillae from the lower abdominal skin in 42 resistance. patients with PCOS and IH. They found higher levels of SRD5A1 Another aim of our study was to assess the expression of aromatase mRNA than SRD5A2 mRNA in both groups, and also they found mRNA in peripheral blood and dermal papilla cells from the lower a positive correlation between both SRD5A gene isotypes and midline abdominal skin in women with hirsutism. We found statisti- concentrations of free serum testosterone. Based on these results, cally significant differences in aromatase gene expression in the tissue they argued that testosterone increases the mRNA level of both and PMBCs between the patient and the control groups. Unluhizarci SRD5A gene isotypes mRNA level in dermal papillae from the lower et al. (5) previously found that the estradiol/free testosterone ratio, abdominal regions in patients with hirsutism. Our results did not rep- which is a marker of aromatase activity, is lower in patients with IH licate the positive correlation between the free serum testosterone when compared with healthy women. One aim of our study was to and the SRD5A expression and decreased SRD5A gene. Both our evaluate aromatase activity at the tissue level, and we demonstrated study and Skalba et al. (14) study need further validation of decreased aromatase gene expression activity in the skin and blood mRNA expression results. of women with IH. It is well known that the pilosebaceous unit may It is believed that the presence and activity of androgen receptors, synthesize androgens de novo from or by locally growth hormone, insulin-like growth factor, insulin, thyroid hor- converting circulating weak androgen to more potent one (5, 20). mones, and play an important role in the development More studies with a larger number of patients should be performed of hirsutism, especially in its idiopathic form (15–17). An altered to investigate the role of decreased 5a-reductase and aromatase gene regulation of hair growth—independent of androgens but linked to expression and 5a-reductase enzyme activity in women with IH. By an increased production of local hormones or mediators—has establishing the pathogenetic mechanisms underlying IH, new been hypothesized. Recently, two studies have shown that insulin therapeutic strategies may play an important role in offering more resistance, a well-known hallmark of PCOS, may also be present effective therapies. Improved treatment of hirsutism awaits advance- in women with idiopathic hirsutism (5, 18). However, at least in ments in the understanding of the interaction between androgens most patients, this association seems to be linked to the presence and other determinants of hair-follicle development. of polycystic ovaries (increased ovarian size) (19); therefore, these reports may just confirm that a mild insulin resistance is present in Acknowledgments: The authors thank Victoria Clark for her contribution in women with ovulatory PCOS (19, 20). It cannot be excluded that editing the manuscript.

REFERENCES

1. Azziz R, Carmina E, Sawaya ME. Idiopathic hirsut- 8. Carmina E, Lobo RA. Polycystic ovaries in women 15. Akiyama M, Smith LT, Holbrook KA. Growth factor ism. Endocr Rev 2000;21:347–62. with normal menses. Am J Med 2001;111:602–6. and growth factor receptor localization in the hair 2. Sawaya MA, Salita AR. Androgen receptor polymor- 9. Bulun SE, Sebastian S, Takayama K, Suzuki T, follicle bulge and associated tissue in human fetus. J phisms (CAG repeat lengths) in androgenetic alopecia, Sasano H, Shozu M. The human CYP19 (aromatase Invest Dermatol 1996;106:391–6. hirsutism and acne. J Cutan Med Surg 1998;3:9–15. P450) gene: update on physiologic roles and genomic 16. Danilenko DM, Ring BD, Pierce GF. Growth factors 3. Legro RS, Shahbahrami B, Lobo RA, Kovacs BW. organization of promoters. J Biochem Mol and cytokines in hair follicle development and Size polymorphisms of the androgen receptor among Biol 2003;86:219–24. cycling: recent insights from animal models and the female Hispanics and correlation with androgen char- 10. Philpott MP, Green MR, Kealey T. Human hair potentials for clinical therapy. Mol Med Today acteristics. Obstet Gynecol 1994;83:701–6. growth in vitro. J Cell Sci 1990;97:463–71. 1996;2:460–7. 4. Faredin I, Toth I. Metabolism and concentration of 11. Serafini P, Lobo RA. Increased 5-a reductase in idio- 17. Philpott MP, Sanders DA, Kealey T. Effects of insulin androgenic in the abdominal skin of women pathic hirsutism. Fertil Steril 1985;43:74–8. and insulin-like growth factors on cultured human with idiopathic hirsutism. Acta Med Hung 1984;41: 12. Sultan C, Lumbroso S, Paris F, Jeandel C, hair follicle: IGF-I at physiologic concentrations is 19–34. Terouanne B, Belon C, Audran F, et al. Disorders of an important regulator of hair follicle growth 5. Unluhizarci K, Karababa Y, Bayram F, Kelestimur F. androgen action. Semin Reprod Med 2002;20: in vitro. J Invest Dermatol 1994;102:857–61. The investigation of insulin resistance in patients with 217–28. 18. Frajndlich R, Spritzer PM. Association between ovar- idiopathic hirsutism. J Clin Endocrinol Metab 13. Oliveira IO, Lhullier C, Brum IS, Spritzer PM. The ian volume and serum insulin levels in ovulatory pa- 2004;89:2741–4. 5a-reductase type 1, but not type 2, gene is expressed tients with idiopathic hirsutism. Fertil Steril 6. Randall VA, Thornton MJ, Hamada K, in anagen hairs plucked from the vertex area of the 2005;83:1561–4. Messenger AG. Androgen action in cultured dermal scalp of hirsute women and normal individuals. 19. Carmina E, Longo RA, Rini GB, Lobo RA. Pheno- papilla cells from human hair follicles. Skin Pharma- Braz J Med Biol Res 2003;36:1447–54. typic variation in hyperandrogenic women influences col 1994;7:20–6. 14. Skalba P, Dabkowska-Huc A, Kazimierczak W, the finding of abnormal metabolic and cardiovascular 7. Hamada K, Thornton MJ, Laing I, Messenger AG, Samojedny A, Samojedny MP, Che1micki Z. Content risk parameters. J Clin Endocrinol Metab 2005;90: Randall VA.The metabolism of testosterone by dermal of 5-alpha-reductase (type 1 and type 2) mRNA in 2545–9. papilla cells cultured from human pubic and axillary dermal papillae from the lower abdominal region in 20. Chen W, Thiboutot D, Zouboulis CC. Cutaneous hair follicles concurs with hair growth in 5a-reductase women with hirsutism. Clin Exp Dermatol androgen metabolism: basic research and clinical per- deficiency. J Invest Dermatol 1996;106:1017–22. 2006;31:564–70. spectives. J Invest Dermatol 2002;119:992–1007.

482 Caglayan et al. Aromatase 5a-reductase hirsutism Vol. 96, No. 2, August 2011