Letrozole increases ovarian growth and Cyp17a1 gene expression in the rat ovary.

After 3 weeks of treatment, letrozole-treated rats had gained more weight than the control rats ( P <.001), and this difference progressively increased until the end of the study ( Fig. 1 ). Most of control rats had a normal 4-day estrous cycle. However, all letrozole-treated rats were entirely acyclic, as reflected by the presence of leukocytes in the vaginal smears, which is the predominant cell type of the diestrous phase, consistent with “pseudo-diestrous.” Ovarian weight was greater (179.1 ± 9.463 mg vs. 115 ± 5.19 mg; P <.001) in the letrozole group than in controls. Similarly, letrozole induced a significant increase in other organ weights, such as heart, liver, spleen, and kidney. Interestingly, adrenal gland and uterus were the only organs that had significantly lower weight in animals receiving letrozole treatment ( Table 1 ). To determine the effects of letrozole on ovarian morphology, the ovaries were sectioned and stained with hematoxylin and eosin and finally evaluated under a conventional microscope. As demonstrated in Figure 2 follicles of rats treated with letrozole had markedly greater thickness of theca-interstitial cells than in the control group. To evaluate the effects of letrozole on individual hormones, the levels of A and T were evaluated in plasma samples using liquid chromatography–mass spectrometry, and LH concentration was measured in serum by a sensitive two-site sandwich immunoassay. As shown in Figure 3A , letrozole dramatically increased androgen levels, inducing a 521- and 244-fold increase (both at P <.001) in both A and T levels, respectively. Similarly, letrozole induced a 2.8-fold increase ( P <.001) in LH concentrations compared with the control group. To investigate whether letrozole affects the expression of the key genes involved in the regulation of steroidogenesis, quantitative real-time PCR was performed in ovarian tissue specimens. As presented in Figure 3B , letrozole induced a significant decrease in Star, Cyp11a1 , and Hsd3b1 mRNA expression, respectively, by 35% ( P <.05), 23% ( P <.05), and 45% ( P <.001) compared with controls. Surprisingly, letrozole exerted a 57-fold increase ( P <.001) in Cyp17a1 transcripts over the control group. To determine whether letrozole affects DNA synthesis, rat theca cells were cultured for 48 hours in the absence or presence of increasing doses of letrozole (0.1–1 µM). The exposure of theca cells to letrozole did not alter cell proliferation at any of the tested concentrations (data not shown). To evaluate the effect of letrozole on Cyp17a1 mRNA expression, quantitative RT-PCR reactions were performed. Letrozole concentrations ranging from 0.1 to 1 µM did not affect Cyp17a1 mRNA levels (data not shown). To evaluate whether the effect of letrozole in the rat ovary is mediated by reduction of aromatization and/or loss of estrogen-induced intraovarian negative feedback loop on theca cell steroidogenesis via modulation of Cyp17a1 expression, mRNA levels of this gene and concentrations of P, A, androsterone, and E 2 were evaluated in cocultures of theca-interstitial and granulosa cells. Letrozole had no effect on either Cyp17a1 mRNA expression or steroid production in theca-granulosa cocultures (data not shown).

The present study has demonstrated that administration of letrozole in intact female rats [1] increases ovarian size and thickness of the theca-interstitial cell layer; [2] reduces Star, Cyp11a1 , and Hsd3b1 mRNA levels; [3] dramatically upregulates androgen production and Cyp17a1 mRNA expression; and [4] increases LH levels; but [5] letrozole in cultures of rat theca-interstitial cells has no direct effect on cell proliferation or Cyp17a1 mRNA expression; and [6] letrozole does not exert any significant effect on Cyp17a1 mRNA expression or steroid production in granulosa-theca cocultures. Letrozole-treated rats showed a significant increase in organ weights, such as heart, kidney, liver, ovary, and spleen compared with the control group, whereas a lower weight was observed for adrenals and uterus. It is likely that letrozole may alter organ weight by inhibiting the aromatization of androgens and hence increasing the level of androgens and decreasing the level of estrogens. Consequently, it leads to anabolic/proliferative effects on androgen-sensitive tissues and the opposite effect on estrogen-dependent organs. Indeed, androgenic stimulation has been previously shown to account for cardiomyocyte and hepatocyte proliferation ( 22 , 23 ), marked renal hypertrophy ( 23 ), and theca-interstitial cell hyperplasia ( 11 ). Conversely, the inhibitory effect of letrozole on weight of the uterus is in agreement with previous studies, whereby letrozole treatment decreased the fibroid and polyp size, owing to letrozole-induced decrease in estrogen production ( 24 ). However, the underlying mechanism by which letrozole decreases adrenal size is still unknown, although a potential androgen receptor–dependent inhibitory effect of T might be involved, as described by others ( 25 ). Despite the growing evidence that letrozole plays a major role in the treatment of several estrogen-dependent gynecologic disorders, little is known about its effects on ovarian physiology. Letrozole-induced increases in rat ovarian size as well as theca-interstitial cell hyperplasia have been demonstrated during in vivo studies evaluating the role of letrozole in inducing both ovarian and metabolic features of polycystic ovary syndrome ( 11 , 26 ). However, other authors did not observe any changes in ovarian weight when administering letrozole to rats, which was likely due to a short exposure time to lower letrozole doses together with oral administration ( 12 ). In the present study, ovarian weight and thickness of the theca-interstitial cell layer were significantly higher in letrozole-exposed rats than in controls. We hypothesize that this effect on ovarian growth may be mediated by either direct action on the ovary or via action at the level of the hypothalamo–pituitary axis. A direct letrozole-induced proliferative effect has been reported in cultures of human endometrium ( 27 ). However, the potential mechanisms by which letrozole could exert its anabolic effect on theca-interstitial cells, and thus ovarian size, are not known. Bajetta and colleagues have shown that daily administration of letrozole at doses of 0.5 or 2.5 mg increases insulin-like growth factor 1 (IGF-1) levels in postmenopausal women with advanced breast cancer, and these findings have been confirmed by recent studies using letrozole and another aromatase inhibitor, anastrozole ( 28 , 29 ). In light of these findings, one may speculate that letrozole-induced proliferative effect might be mediated through up-regulation of growth factors belonging to the IGF family. Indeed, mRNA encoding both IGF-1 and IGF-1 receptor has been detected in human ( 30 , 31 ), bovine ( 32 ), and rat theca cells ( 33 , 34 ), suggesting that letrozole might stimulate expression and synthesis of IGF-1 and thus may be inducing growth of theca-interstitial cells. Although the mediation of letrozole effects by IGF-1 cannot be excluded, the lack of a letrozole-induced stimulatory effect on theca-interstitial cell proliferation in vitro indicates that a direct effect of letrozole on growth of the theca-interstitial cell compartment is unlikely. However, one cannot exclude the possibility that such effects may occur upon longer exposure to letrozole. Indeed, Manneras et al. ( 11 ) demonstrated the presence of theca hyperplasia in the ovaries of letrozole-treated rats after 11–13 weeks of drug administration. In our in vivo experiment, a letrozole-induced increase in body weight was observed only after more than 40 days of letrozole treatment; it is therefore possible that letrozole may also induce delayed but direct effects on growth of tissues, including ovaries. Alternatively, the letrozole-induced growth-inducing effect on the theca-interstitial cell compartment and ovarian morphology may be due to indirect mechanisms. First, a letrozole-induced increase of intraovarian androgens may affect follicle development and ovarian growth. Several studies have shown the presence of androgen receptor in theca cells of human ( 35 ), primates ( 36 ), and rats ( 37 ). Thus, it is conceivable that letrozole-induced overproduction of androgens might induce androgen receptor–mediated trophic effects on theca cell proliferation. In a model closer to humans, intact monkeys treated with androgens had increased numbers of preantral and small antral follicles as androgens stimulated theca and granulosa cell proliferation and inhibited apoptosis ( 36 ). Second, reduction of estrogen production due to letrozole-induced aromatase inhibition could enhance LH secretion by releasing a negative feedback of estrogens on the hypothalamus and pituitary ( 12 ), serving as a trophic stimulus to theca-interstitial cell growth that leads to overproduction of androgens. Indeed, letrozole treatment induced an increase in LH levels in our in vivo experiment, suggesting that a letrozole-induced increase in ovarian growth may be due to stimulation of the hypothalamic–pituitary axis. Thus, letrozole-induced androgen excess via inhibition of aromatase together with an increase in LH secretion due to letrozole-induced release of the negative feedback on the hypothalamic–pituitary axis might induce a trophic action on theca cell proliferation. The present in vivo study showed that letrozole increased androgen production and reduced the mRNA expression of several genes involved in steroidogenesis except for Cyp17a1 , which was greatly enhanced in letrozole-exposed rats compared with controls. We speculate that an androgen-dependent feedback inhibition on steroidogenic enzyme expression within the ovary may account for the above-mentioned decrease in mRNA expression. Indeed, an androgen-induced negative feedback on Star mRNA expression has been previously shown in rat Leydig cells ( 38 ), indicating that androgens may regulate steroidogenesis at the crucial, rate-limiting step of cholesterol transfer to the inner mitochondrial membrane. In addition, Star expression has been shown to be upregulated by estrogen ( 39 ), suggesting that the decreased estrogen output in letrozole-exposed rats may account for letrozole-induced decrease in Star mRNA expression. Thus, these findings suggest that steroid hormone-induced feedback on steroidogenic enzymes may lead to auto/paracrine mechanisms through which steroid hormones can regulate their own production in ovarian tissue. In the ovary, androgens are synthesized in LH-stimulated theca cells and play an important role in preovulatory follicular development and maintenance of oocyte viability ( 40 ). The Cyp17a1 gene is the rate-limiting step in androgen biosynthesis, encoding the microsomal cytochrome P450c17α, which is a single enzyme with dual activity: 17α-hydroxylase and 17,20-lyase. In the present study we demonstrated a profound up-regulation of both androgen and LH levels together with an increase in Cyp17a1 mRNA expression in letrozole-exposed rats. Previously it has been shown that granulosa cell-derived estrogens modulate theca cell steroidogenesis via a short negative-feedback loop within the rodent follicle ( 41 ). Thus, we speculate that the letrozole-induced increase in androgen levels may be due not only to the direct effect of letrozole on the blockade of aromatization but also to a release of the negative-feedback loop of estrogens on theca cell steroidogenesis. Indeed, in vitro rodent follicle cultures have shown that estrogen receptor-α mediates an intraovarian negative feedback loop on theca cell steroidogenesis via modulation of Cyp17a1 mRNA expression ( 42 ). However, the lack of effect of letrozole on Cyp17a1 mRNA expression and on steroid production in granulosa-theca cocultures does not support the idea of the estrogen-induced intraovarian negative feedback loop on theca cell steroidogenesis. Therefore, the present findings suggest that it is unlikely that a letrozole-induced decrease of estrogen levels might stimulate theca cell androgen production by abrogating the negative feedback of estrogen on Cyp17a1 mRNA expression in a paracrine manner. Although the ovary represents the major source of androgens, other endocrine tissues are equipped with steroidogenic enzymes to produce androgens and thus might contribute to the letrozole-induced increase in androgen production observed in our in vivo study. For example, Manneras et al. ( 11 ) demonstrated that letrozole-treated rats showed increased weights of the main adipose tissue depots. Adipose tissue has been shown to possess the enzymatic capability to produce androgens, including 17α-hydroxylase activity ( 43 ). Therefore, a contribution of adipose tissue in letrozole-induced increase in androgen production cannot be ruled out. In conclusion, the present study provides new insights into the effects of letrozole on ovarian growth and steroidogenesis using in vivo and in vitro rat models. Letrozole stimulates ovarian growth and greatly up-regulates androgen production and Cyp17a1 mRNA expression, the key gene regulating the androgen biosynthesis pathway, and the effect is likely related to increased LH level. It is important to note that these effects were observed after long-term exposure to letrozole and that the above conclusions should not be extrapolated to short-term and intermittent exposure, such as that occurring in women undergoing ovulation induction.

Three Wistar dams, each with 10 female pups, were obtained from Charles River Laboratories. Pups were raised with the lactating dam (not the biological mother of all the pups) until 21 days of age and then housed two per cage under controlled conditions (21–24°C, 55%–65% humidity, 12-hour light/12-hour dark cycle). Rats were fed standard commercial food and tap water ad libitum. At 21 days of age, rats were randomly divided into two experimental groups (control [n = 14] and letrozole [n = 14]) and implanted SC with 70-day continuous-release pellets (Innovative Research of America) containing 28 mg of letrozole (daily dose, 400 µg) (Novartis Pharma) or placebo. The dose of letrozole was chosen according to a previous study ( 12 ). The control group received identical pellets lacking the bioactive molecule. The animals were anesthetized using isoflurane before the SC insertion of the pellet. Rats were weighed every week from 21 days of age. Estrous cycle stage was determined microscopically with Giemsa staining of the predominant cell type in vaginal smears obtained daily from the ninth week of age to the end of the experiment, considering cycles with duration of 4 to 5 days to be regular ( 14 , 15 ). The study was concluded after 10 weeks of exposure to letrozole, when the rats were 13 weeks of age. All treatments and procedures were carried out in accord with accepted standards of human animal care as outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and a protocol approved by the Institutional Animal Care and Use Committee at the University of California, Davis. After 10 weeks the rats were killed in diestrous phase by intra-cardiac perfusion of 0.9% saline under anesthesia using ketamine and xylazine (75/5 mg/kg, IP); trunk blood was collected, and plasma was stored at −20°C until assayed. The heart, liver, pancreas, adrenals, spleen, kidneys, uterus, and ovaries were excised. One half ovary from each animal was frozen and stored at −80°C for subsequent use. One ovary from each animal was fixed in 10% formalin, embedded in paraffin wax, and then sectioned serially at 5-µm thickness. Sequential sections were mounted and stained by the hematoxylin and eosin procedure. The specimens were evaluated under ×400 magnification. Total RNA was isolated from the ovarian specimen and from theca-interstitial cells using the MagMAX-96 Total RNA Isolation Kit (Applied Biosystems) and the King Fisher robot (Thermo Scientific). Reverse transcription of total RNA to complementary DNA (cDNA) was performed using the High Capacity cDNA Reverse Transcription Kit for reverse transcription–polymerase chain reaction (RT-PCR) (Applied Biosystems). The PCR assays were set up in 28-µL volumes, consisting of 5 µL cDNA, 4.5 µL forward and 4.5 µL reverse 900n Mprimers, and 14 µL of 2× SYBR Green PCR Master Mix (Applied Biosystems). Quantitative real-time PCR reactions were performed in triplicate using the ABI 7300 Real-Time PCR System (Applied Biosystems). Separate cDNA dilutions were included in each PCR run to generate standard curves. Data were analyzed using SDS 1.4 software (Applied Biosystems). The relative amount of target messenger RNA (mRNA) was expressed as a ratio normalized to hypoxanthine phosphoribosyltransferase ( Hprt ). The primer sequences were as follows: rat Star forward (5′-GCC TGA GCA AAG CGG TGT C-3′) and reverse (5′-CTG GCG AAC TCT ATC TGG GTC TGT-3′); rat Cyp11a1 forward (5′-GCT GGA AGG TGT AGC TCA GG-3′) and reverse (5′-CAC TGG TGT GGA ACA TCT GG-3′); rat Hsd3b1 forward (5′-CCA GAA ACC AAG GAG GAA T-3′) and reverse (5′-CCA GAA ACC AAG GAG GAA T-3′); rat Cyp17a1 forward (5′-ACT GAG GGT ATC GTG GAT GC-3′) and reverse (5′-CCG TCA GGC TGG AGA TAG AC-3′); and rat Hprt forward (5′-TTG TTG GAT ATG CCC TTG ACT-3′) and reverse (5′-CCG CTG TCT TTT AGG CTT TG-3′). Androstenedione and Twere obtained from Steraloids, whereas testosterone-d3 was obtained from Cerillient. Acetonitrile and methanol were high-performance liquid chromatography (HPLC) grade and obtained from Burdick and Jackson. Acetone, isopropanol, and ammoniumhydroxide were Optima grade and obtained from Fisher. Formic acid was American Chemical Society (ACS) grade and obtained from EMD. Each sample was directly assayed; the following extraction procedure was applied to each specimen. Each sample aliquot (300 µL) was placed in a 2.0-mL autosampler vial and spiked with 150 µL of internal standard solution (i.e., androsteneione-d7 and testosterone-d3). Detection and quantitation of all analytes was accomplished using selective reaction monitoring. We developed a novel turbulent flow chromatography (TFC) HPLC–tandem mass spectrometry (MS/MS) method that allowed the simultaneous detection of A and T. It consists of a high-pressure liquid chromatography instrument configuration multiplexing Thermo Aria TLX-2 TFC (two loading pumps and two eluting pumps; Shimadzu LC-10AD) system and an autosampler outfitted with a 300-position Peltier tray, coupled to a Thermo Scientific TSQ Vantage triple quadrupole mass spectrometer, equipped with a heated electro-spray ionization source. The instrument was controlled using Aria software (version 1.6.1). A Thermo Cyclone P extraction column (0.5 × 50 mm, 60-µm particle size) was used for online sample extraction of diluted serum, and HPLC separation was carried out by a 2.1 × 100-mm, 3-µm particle size ACE C18 column protected by a reverse-phase guard cartridge (Mac-Mod) contained within a Hot Pocket column heater (Thermo Scientific). Precursor and product ions for each target analyte were chosen for selective reaction monitoring transitions, and the related parameters for the different analytes were isolated by HPLC separation according to the following mobile-phase gradient: solvent A, water containing 0.1% formic acid; B, methanol; C, acetonitrile/isopropyl alcohol/acetone 60/30/10; D, water/acetonitrile (98/2 vol/vol) with 0.1% ammonium hydroxide. Detection and quantification used select reaction monitoring LC-MS/MS transitions of initial precursor ions for A and T mass to charge ratio (m/z) 287.2, 291.4, and 315.2, respectively. The responses for the major product ions for each of the analytes were plotted, and peaks at the proper retention time were integrated using LCQuan (Thermo Scientific). This software was used to generate calibration curves and quantitate the analytes in all samples. The concentrations of A and T in each sample (e.g., calibrators, quality control, and unknowns) were determined by an internal standard method using the peak area ratio and linear regression analysis. The responses for A and T were linear and gave correlation coefficients ( R 2 ) of 0.99 or better. Determination of serum level of rat LH was performed by the Center for Research in Reproduction, School of Medicine, University of Virginia using LH sandwich assay using monoclonal antibodies against bovine LH (no. 581B7) and against the human LH-β subunit (no. 5303; Medix Biochemica) ( 16 , 17 ). Reference LH (rat) was provided by Dr. A. F. Parlow (National Hormone and Peptide Program). In experiments evaluating possible direct actions of letrozole on ovarian theca-interstitial cells, ovaries were obtained from intact young rats not treated with letrozole, and theca-interstitial cells were purified as described previously ( 18 , 19 ). Theca-interstitial cells were incubated in 24-well fibronectin-coated plates at a density of 400,000 cells per well for steroido-genesis and in 96-well fibronectin-coated plates at a density of 35,000 cells per well for proliferation assay. The cultures were carried out for 48 hours at 37°C in an atmosphere of 5% CO 2 humidified air in serum-free McCoy's 5A culture medium supplemented with 1% antibiotic/antimycotic mix, 0.1% bovine serum albumin, and 2 mM l -glutamine. The cells were incubated in the absence (control) or in the presence of letrozole (0.1–1 µM). The concentrations of this compound were selected on the basis of previous studies ( 20 , 21 ). All cultures were carried out in the presence of ovine LH (5 ng/mL). All the above chemicals were purchased from Sigma Chemical except for letrozole, which was purchased from Novartis Pharma, and LH, which was obtained from the National Hormone and Pituitary Program at the Harbor-UCLA Medical Center. To evaluate possible effects of letrozole on proliferation of cells, the purified theca-interstitial cells were incubated for 48 hours in 96-well fibronectin-coated plates at a density of 35,000 cells/well in the absence (control) or in the presence of letrozole (0.1–1 µM) and supplemented with LH (5 ng/ mL). The extent of DNA synthesis was determined through a thymidine incorporation assay. Radiolabeled [ 3 H] thymidine (1 µCi per well) was added to the cells 24 hours before the culture was stopped. Subsequently, the cells were harvested with a multiwell cell harvester (PHD Harvester, Model 290; Cambridge Technology), and radioactivity was measured in a liquid scintillation counter (Wallac 1409; PerkinElmer). The total number of viable cells was estimated with the use of a CellTiter-Blue Cell Viability Assay (Promega). This assay involves the conversion of resazurin to resorufin by metabolically active cells, resulting in the generation of a fluorescent product at the excitation wavelength 544 nm and the emission wavelength 590 nm that is proportional to the number of viable cells. Fluorescence was determined with the use of a microplate reader (Fluostar Omega; BMG). To validate the assay, a standard curve with a known number of cells was generated, and a linear correlation was verified ( r 2 = 0.99, P <.001). Each experiment was repeated three times (eight replicates per experiment). Twenty-four-well fibronectin-coated plates with transwell inserts (6.5-mm diameter) with 8.0-µm pore size polycarbonate membrane (Transwell Permeable Supports; Corning Life Sciences) were used for granulosa-theca cell coculture experiments. The membranes of precooled inserts were coated using 40 µL of Matrigel (ECM gel, growth factor reduced, without phenol red, from Engelberth-Holm-Swarm mouse sarcoma; Sigma-Aldrich) diluted to a final protein concentration of 0.3 mg/mL with cold McCoy's media. Matrigel was used to mimic basement membrane layer and to facilitate communication of theca and granulosa compartments. The granulosa cells were transferred to the transwell inserts (200,000 cells per transwell insert), and the theca cells (200,000 cells per well) were placed in the lower chambers of the wells. Subsequently, the cells were cultured for 48 hours without (control) or with letrozole (0.1–1 µM). Luteinizing hormone (5 ng/mL) and FSH (30 ng/ml) were added to all cell cultures. After 48 hours media were collected for steroids evaluation, and RNA isolation and cDNA synthesis were conducted separately for granulosa and theca cells. Determination of expression of Cyp17a1 was performed as described above by quantitative real-time PCR using Hprt as a reference gene. Each experiment was repeated three times. Statistical analysis was performed using JMP 9.0 software (SAS). Data are presented as the mean ± SEM. Means were compared by Student's t test. When appropriate, data were logarithmically transformed. A value of P <.05 was considered statistically significant.