The prostate cancer drug enzalutamide shortens anogenital distance in male rat offspring by blocking the androgen receptor

Background: Enzalutamide is a non-steroidal anti-androgen drug used to treat prostate cancer. It is a potent androgen receptor (AR) antagonist, with an in vitro Lowest Observed Effect Concentration (LOEC) of 0.05 μM. In this study, we wanted to assess its utility as a model compound for future mechanistic studies aimed at delineating mechanism-of-action of anti-androgenic effects in the developing fetus. Methods: Enzalutamide in vitro activity was tested using an Androgen receptor reporter assay (AR-EcoScreenTM) and a steroidogenesis assay (H295R assay). For in vivo characterization, pregnant Sprague-Dawley rats were exposed to 10 mg/kg bw/day enzalutamide from gestational day 7-21. At gestational day 21, enzalutamide exposure concentrations were measured both in amniotic fluids and fetal plasma, alongside Anogenital distance (AGD). Fetal testes were collected and for testosterone measurements and gene expression profiling. Results: Enzalutamide was a strong AR antagonist in vitro and we also observed disrupted androgen synthesis in the H295R steroidogenic assay with a LOEC of 3.1 μM. In utero exposure resulted in about 20% shorter anogenital distance (AGD) in male fetuses., as well as signs of dysregulated expression of the steroidogenic genes Star, Cyp11a1 and Cyp17a1 in the fetal testes at gestational day 21. Intra-testicular testosterone levels were unaffected. Conclusions: Based on these observations, together with in vitro LOECs and the fetal plasma levels of enzalutamide, we propose that the effect on male AGD was caused by AR antagonism rather than suppressed androgen synthesis. Due to the characteristic mechanism of action of enzalutamide, we suggest to use it as a new model compound in research on anti-androgenic environmental chemicals.

5 bicalutamide target AR transcriptional activity by interfering with recruitment of coactivators to the transcriptional complex (16,17), enzalutamide targets three key stages of AR signaling: blocking androgen binding, inhibiting translocation of activated AR and inhibiting binding of activated AR to the DNA (14). Thus, enzalutamide is a potent, specific inhibitor of androgen signaling.
We recently completed a study using the 5α-reductase inhibitor finasteride as a model compound in an effort to tease out some of the underlying molecular mechanisms driving effects on male AGD (18). Based on in vitro data we found that enzalutamide apart from blocking the AR also inhibited androgen synthesis in vitro.
To follow up on this, we tested enzalutamide in an in utero exposure study based on its known antiandrogenic effect in order to investigate, if the compound affected AGD and, if so, which mechanism was underlying the effect. We found that the AGD of the late gestation male fetuses were significantly shorter and that gene expression levels of steroidogenic enzymes in the fetal testis at GD21 were affected, albeit without significantly affecting the intra-testicular testosterone levels. Thus, we conclude that enzalutamide causes it AGD effect by antagonizing the AR.
AR-EcoScreen™ assay 6 The antagonistic effects of enzalutamide on AR were investigated using the   was used with an injection volume of 100 µl, measuring in ESI-mode using methanol, and 1 mM ammonia in water as the mobile phases (gradient flow rate was 0.4 ml/min). For the other hormones an Ascentis Express C 8 column (2.1 × 100 mm, 2.7 µm) was used with an injection volume of 100 µl, measuring in ESI-/ESI + mode 9 with acetonitrile and 0.1% formic acid in water as the mobile phases (gradient flow rate was 0.25 ml/min). Ten hormones: testosterone, androstenedione, dehydroepiandrosterone (DHEA), corticosterone, cortisol, pregnenolone, progesterone, 17α-OH-progesterone, estradiol and estrone were quantified. The limit of quantification (LOQ) was 0.1 ng/ml for corticosterone, 1.0 ng/ml for DHEA and pregnenolone, 0.02 ng/ml for testosterone and androstenedione, 0.05 ng/ml for 17α-OH-progesterone, and 0.01 ng/ml for all the other hormones. For quantification, external calibration standards were run before and after the samples at levels of 0.1, 0.2, 0.5, 1.0, 2.0, 5.0 and 20 ng/ml, with 5.0 ng/ml internal standards: (testosterone-d2, methyltestosterone-d3, progesterone-c2, and estradiol-d3). The mass spectrometer was an EVOQ Elite Triple Quadropole Instrument from Bruker (Bremen, Germany) and the UPLC system was an Ultimate 3000 system with a DGP- Denmark). They were placed in an animal room with controlled environmental conditions: 12 hr light-dark cycles with light starting at 9 pm, temperature 22 ± 1 °C, humidity 55 ± 5%, 10 air changes per hr.
All animals were fed a standard diet with Altromin 1314 (soy-and alfalfa-free, Altromin GmbH, Lage, Germany). Acidified tap water (to prevent microbial growth) in PSU bottles (84-ACBTO702SU Tecniplast) were provided ad libitum. The PSU bottles and cages as well as the aspenwood shelters (instead of plastic) were used to eliminate any risk of migration of bisphenol A that could potentially confound the study results. From GD7-21, dams were weighed daily and dosed by oral gavage by qualified animal technicians with a stainless steel probe 1.2 × 80 mm (Scanbur, Karlslunde, Denmark) with either vehicle control (corn oil) or enzalutamide (10 mg/kg bw/day) at a constant volume of 2 ml/kg bw per day. All animals were decapitated (guillotined) under CO2/O2-anesthesia at GD21.

Caesarean sections GD 21
Dams were decapitated (guillotined) under CO 2 /O 2 -anesthesia at GD21 and fetuses were collected by caesarean section. The dams were exposed 1h ± 15 min before decapitation in the same order as Caesarean sections were performed to adjust for the chemical analysis of maternal blood, fetal blood and amniotic fluid. Uteri were taken out and weighed, and the number of live fetuses, resorptions, and implantations were registered. Body weights of the fetuses were recorded prior to decapitation (by a scissor). Maternal trunk blood was collected and transferred to heparin-coated vials. Trunk blood from all fetuses was collected and transferred to heparin-coated vials and pooled for each gender within each litter. Blood samples were kept on ice and centrifuged at 4000 rpm, 4 ºC for 10 min. Plasma was transferred to new tubes and stored at -80 ºC. Amniotic fluid was collected from all 11 fetuses, pooled within each litter, snap frozen in liquid nitrogen and subsequently stored at -80 ºC. AGD was measured as the distance between the genital papilla and the anus by the same, blinded technician using a stereomicroscope with a micrometer eyepiece. The AGD index (AGDi) was calculated by dividing AGD by the cube root of the body weight. Fetal testes were isolated by dissection under a stereomicroscope and LC was performed on a Dionex Ultimate 3000 RS (Thermo Scientific, CA) with a Poroshell SB C-18 (100 × 2.1 mm, 2.7 µm particle size) column held at 30 °C (Agilent technologies, Walbron, Germany). The solvent system consisted of A: 2.5 mM ammonium hydroxide + 0.1% formic acid in water and B: acetonitrile. Solvent programming were: 2% B from 0 to 1 min followed by a linear gradient to 95% B to 14 min, isocratic 95% B from 14 to 16 min followed by reversal to initial conditions to 16.1 min and re-equilibration of the column to 20 min. The flow rate was 0.3 ml/min from 0 to 1 min followed by a linear gradient to 0.4 ml/min to 14 min, which was held to 16 min followed by reversal to initial conditions.

Synthesis of cDNA and RT-qPCR analysis
Protocols were essentially as previously described (23). Briefly, total RNA was extracted from GD21 testis (n = 12/group) using RNeasy mini kit and on-column Rn01751069_mH, Rps18 Rn01428913_gH, and Sdha Rn00590475_m1. In addition, primers and probes Cyp17a1, Cyp11a1 and Star were designed in our lab (24). The following cycling conditions were used: an initial step of 95 °C for 20 sec followed by 45 two-step thermal cycles of 95 °C for 1 sec and 60 °C for 20 sec. The relative transcript abundance was calculated using the 2 − ΔCT method using Rps18 and Sdha as normalizing genes.

Statistics
Data from the AR-Eco Screen and H295R assay were analyzed by one-way ANOVA followed by Dunnett's post hoc test in GraphPad Prism 5 (GraphPad Software, San Diego California, USA). Results are presented as mean ± SEM for the three independent experiments. One measurement of DHEA in the H295R assay was lost so the DHEA data presented are from two independent experiments.
In vivo data on maternal parameters, fetal body weight, AGD and AGDi, were analyzed by one-way ANOVA followed by Dunnett's post hoc test, using SAS® (SAS Enterprise Guide 6.1, SAS Institute, Inc., Cary, NC, USA). AGD was analyzed using fetal weight as a covariate and fetal body weights were analyzed using the number 15 of offspring per litter as covariate. For all analyses, the litter was the statistical unit. Statistical analyses were adjusted using litter as an independent, random and nested factor. For data presentation, group mean ± SEM was calculated from 6 litters/group based on litter means.
Analysis of enzalutamide concentrations in plasma and amniotic fluid as well as RT-qPCR data and intra-testicular testosterone levels was performed with student's ttest in GraphPad Prism 8 (GraphPad Software, San Diego California, USA). In cases of non-normal distribution or non-equal variance between groups, data was logtransformed prior to analysis, while the graphs still represent the untransformed data. For data presentation, mean ± SEM was calculated from 6 litters/group (maternal plasma, amniotic fluid and fetal plasma), 1 testicle from 6 fetuses /group (fetal intra-testicular testosterone) and 1 testicle from 12 fetuses/group (RT-qPCR).

Enzalutamide acts as an AR antagonist in vitro
The AR agonistic and antagonistic potential of enzalutamide were investigated using the AR-EcoScreen™ assay (19). Enzalutamide did not show any agonistic activity ( Fig. 1A) but showed AR antagonistic activity at all concentrations (p < 0.001) between 0.05-12.5 µM (Fig. 1B). The lowest observed effect concentration (LOEC) was 0.05 µM while the IC 50 value was 0.1 µM. We confirmed that the reduction in luciferase activity was not due to cytotoxicity following enzalutamide exposure (Fig. 1C).

Enzalutamide affects steroidogenesis in vitro
The H295R Steroidogenesis Assay (20) was used to test whether enzalutamide affects the synthesis of ten sex steroid hormones (Fig. 2). All steroid hormones, except for the two estrogens estrone and estradiol, were affected by enzalutamide.
Pregnenolone levels were slightly increased with a LOEC of 3.1 µM. However, no increase was seen at concentrations of 25 µM and above. The progestagens, progesterone and 17α-OH-progesterone were decreased with LOECs of 1.8 and 6.3 µM, respectively. The levels of the two androgens, androstenedione and testosterone were decreased, both with LOECs of 3.1 µM. By contrast, the adrenal androgen and precursor of the other androgens, dehydroepiandrosterone (DHEA), was increased with a LOEC of 0.8 µM. The corticosteroids, corticosterone and cortisol were generally decreased with LOECs of 6.3 and 25 µM, respectively, although cortisol was increased at 0.8 µM.
Enzalutamide was present in different biological compartments and male AGD was shorter in exposed animals at GD21 Pregnant Sprague Dawley rats were exposed to 10 mg/kg bw/day enzalutamide from GD7 to GD21 and maternal as well as fetal parameters were investigated at GD21.
The distribution of enzalutamide in the different biological compartments (i.e. maternal plasma, amniotic fluid and fetal plasma) was determined at GD21 by HPLC-MS/MS. In maternal plasma the concentration was 1002 ± 145 nM, while it was 285 ± 46 nM in amniotic fluid, and 282 ± 53 nM and 122 ± 21 nM in the plasma of the female and male fetuses, respectively (Fig. 3). We measured maternal body weight, weight gain (GD 7-21), and uterus weight to determine if enzalutamide exposure resulted in maternal toxicity, but observed no treatment-related significant differences between the groups or any signs of maternal toxicity (Table 1). In addition, the number of fetuses as well as fetal weights of both males and females were similar between the two groups (Table 1). Attesting to the anti-androgenic potential of enzalutamide, we found that both AGD and AGDi was 19% shorter (p < 0.001) in exposed males compared to control males (Fig. 4, Table 1). There were no significant differences in either AGD or AGDi between exposed females and control females (Table 1).
Enzalutamide affects gene expression of steroidogenic enzymes in male fetal testis, without affecting intra-testicular testosterone levels Based on our observations that enzalutamide affects steroidogenesis in vitro, we assessed expression levels of key genes encoding steroidogenic enzymes in the male fetal testis at GD21. There was an upregulation of Star (p < 0.001), Cyp11a1 (p < 0.05), Cyp17a1 (p < 0.05) and a downregulation of Nr5a1 (p < 0.05) following in utero exposure to enzalutamide (Fig. 5). The expression levels of the two hydroxysteroid dehydrogenase genes, Hsd3b1 and Hsd17b1, were unchanged, as were the germ cell marker gene Ddx4 and Sertoli cell marker Sox9 (Fig. 5). Next, we measured the intra-testicular testosterone levels, but found no significant difference between control males and exposed males (Fig. 6). Finally, histopathological assessment of H&E stained fetal testis sections showed no adverse alterations to testis histology (Fig. 7).

DISCUSSION
Enzalutamide is a second generation prostate cancer drug used to treat men with mCRPC (13). In this study, we explored the use of enzalutamide as a potential model compound or for investigations of anti-androgenic actions of environmental chemicals. Our particular interest in this regard is the use of target-specific compounds to delineate the molecular mechanisms driving the development of anogenital tissues. The purpose is to obtain better knowledge on the utility of AGD as a general biomarker for fetal anti-androgenic effects.
Enzalutamide is designed to specifically target the AR (14). In our AR reporter gene assay the IC 50 for AR antagonism was calculated to be 0.05 µM, which is close to the IC 50 of 0.03-0.05 µM previously reported (14,25). Our results thereby confirmed that enzalutamide acts as an AR antagonist, without having any AR agonistic activity. Because of these specific antagonistic properties, we performed an in utero exposure study in rats, with focus on effects on the male AGD.
We measured the concentration of enzalutamide in maternal plasma, amniotic fluid and fetal plasma at GD 21 in Sprague Dawley rats and found it present in all three compartments. These data clearly show that enzalutamide can transfer across the placenta, albeit the maternal concentration was about four times higher than in the fetal compartments. Notably, we also measured almost twice the concentration of enzalutamide in male plasma compared to female. It remains unclear why this difference was observed, but sex differences in pharmacokinetics is a welldocumented phenomenon, both in humans and animals (26). Furthermore, the fetal plasma concentration of 0.1-0.3 µM is twice or more than the LOEC of 0.05 µM for AR antagonism observed in vitro. Taken together, these data show that an exposure level of 10 mg/kg bw/day during gestation is appropriate to reach biologically active enzalutamide levels in the fetus and reduce fetal androgen signaling. As a reduced AGD caused by fetal androgen insufficiency is a well-known marker in both humans and rodents, we believe that the observed effects are translatable to humans.
Fetal androgen insufficiency, caused by low androgen levels or blockage of AR signaling, can prevent the male perineum to develop properly and result in a short AGD (8). One previous report shows that enzalutamide induces shortening of male AGD in mice, although they do not state how much shorter AGD is compared to controls (27). We found the average AGD to be 19% shorter in enzalutamide exposed males than in control males. This effect is less pronounced than previously observed with the prostate cancer drug and AR antagonist flutamide, which shortened male AGD by 19% at a dose of only 2 mg/kg (28), and between 35-43% in the 6-8 mg/kg dose range (28,29). Conversely, enzalutamide has a greater effect on male AGD than some of the pesticides known to interfere with AR, such as vinclozolin and procymidone, where doses of 50-100 mg/kg are required to induce a similar shortening of AGD (28,(30)(31)(32). Because of the greater potency and the known mode of action of selective pharmaceuticals like flutamide, finasteride or enzalutamide, they are often better suited as model compounds than less specific environmental chemicals when performing studies aimed at characterizing mechanisms of effects.
Apart from blocking AR action, enzalutamide was also found to disrupt hormone synthesis in the H295R in vitro steroidogenesis assay. Progestagen, androgen and corticosteroid synthesis was downregulated, whereas estrogen synthesis was unaffected. Dehydroepiandrosterone (DHEA), on the other hand, was markedly elevated. It remains unclear exactly how enzalutamide cause these effects in the steroidogenic pathway, but judging from the hormone profile, we speculate that 3βhydroxysteroid dehydrogenase is inhibited. With the decrease in androstenedione and testosterone, a decrease in estrogens could have been expected. It is, however, possible that sulfotransferase activity, which is responsible for metabolism of estrogens (33), was reduced. Reduced sulfotransferase activity could equalize the estrogen levels and result in unchanged levels. Another possible explanation could be induction of CYP19 as enzalutamide has previously been shown to induce other CYPs, specifically CYP2C9, CYP2C19 and CYP3A4 (34). These in vitro assays together with physiologically-based kinetic modeling have the potential to be able to predict 20 AGD effects in the future, thereby minimizing animal testing.
Since steroidogenesis was affected in vitro, we also investigated the fetal testis for signs of steroidogenic disruption. We saw increased expression of the key steroidogenic genes Star, Cyp11a1 and Cyp17a1. We initially speculated that this could be a compensatory response to reduced testosterone levels. However, we did not detect any significant reduction in intra-testicular testosterone levels. Thus, at the functional level the effects of enzalutamide on the fetal testis may be diminishable compared to its effect on AR activation in peripheral tissues. In support of this, the in vitro LOEC on testosterone synthesis was 3.1 µM, a concentration that is much higher than the actual measured fetal levels of 0.1 µM.
Nevertheless, our results point to some degree of disturbance to fetal testis function, but more studies are needed to elucidate the exact mechanism of action of enzalutamide on the steroidogenesis pathway.

CONCLUSION
In summary, enzalutamide has strong AR antagonistic effects both in vitro and in vivo. While we observed weak inhibition of steroid synthesis in the rat, this is likely of minor importance in the mode of action, as intra-testicular testosterone levels were not affected. Since enzalutamide is used to treat prostate cancer patients, fetuses will most likely not be exposed to the compound under normal circumstances; albeit, environmental exposure cannot be ruled out. The effects of enzalutamide on the developing male fetus are therefore not of immediate concern, but confirms that enzalutamide is a valuable model compound for future studies on effects of AR antagonists.  Enzalutamide does not affect intra-testicular testosterone at GD 21: Intra-testicular testoster Enzalutamide does not affect fetal testis histology: Cross-sections of formalin-fixed GD21 tes

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.