Evaluation of the Male Pubertal Onset Assay to Detect Testosterone and Steroid Biosynthesis Inhibitors in CD Rats

M. S. Marty,1 J. W. Crissman,2 and E. W. Carney Toxicology and Environmental Research and Consulting, The Dow Chemical Company, 1803 Building, Midland, Michigan 48674

Received August 21, 2000; accepted January 2, 2001 (EDSTAC) was established and proposed a battery of tests The male pubertal onset assay has been recommended by the designed to detect compounds altering estrogen, androgen, or Endocrine Disrupter Screening and Testing Advisory Committee thyroid function (EDSTAC, 1998). The mammalian in vivo (EDSTAC) as an alternate Tier I screening assay to detect poten- components of the recommended Tier I in vivo battery consist tial endocrine-active chemicals (EACs). Recently, this assay was evaluated by several laboratories using a variety of dosing of the uterotrophic assay, the female pubertal onset assay, and schemes. This study used a 30-day dosing period to confirm and the Hershberger assay, but alternative assays have been pro- extend previous work on the assay’s ability to detect steroid bio- posed. One alternative battery replaces the female pubertal synthesis inhibitors. Weanling male rats were dosed by gavage assay and the Hershberger assay with the 14-day intact male from 21 to 50 days of age with vehicle (0.5% methocel) or chem- assay. icals from the following EAC classes: an androgen (testosterone The other alternate assay under consideration is the male propionate [TP], 0.1 or 0.4 mg/kg/day), a broad-spectrum steroid pubertal onset assay. According to the EDSTAC report, 33- biosynthesis inhibitor (ketoconazole [KETO], 24 mg/kg/day), a day-old male rats are dosed with test material daily by gavage 5 -reductase inhibitor (finasteride [FIN], 20 or 80 mg/kg/day), a ␣ for 20 days, and the age and weight at which these rats attain moderately specific aromatase inhibitor (testolactone [TL], 220 mg/kg/day), or a highly specific aromatase inhibitor (fadrozole puberty is measured. Balano-preputial separation (PPS) serves [FAD], 0.6 or 6.0 mg/kg/day). None of these treatments altered as a biomarker of puberty onset. Body weights are monitored relative thyroid weights. However, TL, KETO, and FIN were throughout the dosing period, and once dosing is complete, rats positive for endocrine activity based on decreases in one or more are euthanized and the weights of the testes, epididymides, reproductive or accessory sex gland organ weights. Of these three levator ani-bulbocavernosus muscles (LABC), ventral prostate inhibitors, only TL significantly increased the age at PPS, indicat- and seminal vesicles with coagulating glands are recorded. At ing that PPS was less sensitive for detecting these EACs. Based on necropsy, a serum sample is collected for the analysis of serum its profile of effects, TL may have been detected as an antiandro- hormone levels, including thyroid stimulating hormone (TSH) gen. TP and FAD were negative in this assay, even at doses that and thyroxine (T4). Additional hormone analyses (i.e., testos- caused effects in other studies. With TP, oral administration terone, estradiol, leutinizing hormone [LH], prolactin [PRL], limited assay sensitivity such that higher TP doses would be needed for detection. FAD decreased body weight gains, but did and triiodothyronine [T3]) are optional. Testes, epididymides, not significantly alter any other assay end points; thus, the capac- and thyroid glands are examined histopathologically. Weights ity of this assay to detect aromatase inhibitors remains in question. and histology for the liver, kidneys, adrenals, and pituitary are Key Words: endocrine disruption; endocrine modulation; optional. EDSTAC; puberty; pubertal onset; preputial separation; testoster- PPS, an androgen-dependent process, serves as an easily one; steroid inhibitors; thyroid. measured, external indicator of pubertal onset in male rats. PPS normally occurs at approximately 43.6 Ϯ 1.0 days of age in Sprague-Dawley rats (Clark, 1999). Aside from androgens, Government mandates passed in 1996 (Food Quality Pro- agents operating through a variety of mechanisms may alter tection Act and the Safe Drinking Water Act) require the U.S. pubertal assay end points (reviewed in Stoker et al., 2000). Environmental Protection Agency to institute a screening pro- PPS has been used as a biomarker of endocrine function for gram to evaluate whether xenobiotics can modulate endoge- many years (Korenbrot et al., 1977), yet a standard protocol for nous endocrine function. Toward this goal, the Endo- the male pubertal assay has not been firmly established or crine Disrupter Screening and Testing Advisory Committee validated. This study was designed as an initial evaluation of this assay’s ability to detect known endocrine-active com-

1 To whom correspondence should be addressed. Fax: (517) 638-9863. pounds (EACs). The EACs used in this study encompassed E-mail: [email protected]. different modes of action and included an androgen (testoster- 2 Current address: Dow Corning Corporation, Midland, MI 48686. one propionate [TP]), a broad-spectrum steroid biosynthesis

285 286 MARTY, CRISSMAN, AND CARNEY inhibitor (ketoconazole [KETO]), a 5␣-reductase inhibitor (fin- litter to ensure that littermates were not assigned to the same treatment group. asteride [FIN]) that blocks the conversion of testosterone to the Ten males were assigned to each treatment group and animals were dosed daily more potent androgen dihydrotestosterone (DHT), and two by gavage as previously described (Marty et al., in press). Because control males from the first series of experiments were similar to control males from aromatase inhibitors that block the synthesis of C18 estrogens the second series of experiments (mean weanling weight: 51.7 Ϯ 2.7 vs. from C19 androgens (testolactone [TL] and fadrozole [FAD]). 52.7 Ϯ 4.6 g; body weight at PPS: 237.1 Ϯ 18.6 vs. 237.3 Ϯ 21.1 g; age at When possible, these compounds were administered at dose PPS: 43.8 Ϯ 1.5 vs. 45.0 Ϯ 2.3 days of age; body weight at study termination: levels reported to cause endocrine-mediated effects by the oral 259.6 Ϯ 20.2 vs. 251.1 Ϯ 23.1 g), results from the controls were combined route in previous studies. With TP, no previous studies were prior to analysis. The present study differed from the protocol proposed by EDSTAC in found in which oral administration to intact juvenile male rats several ways. First, there is uncertainty as to whether a 20-day dosing period was performed; therefore, doses were selected by increasing is sufficient (Ashby and Lefevre, 2000); therefore, animals were dosed for 30 the concentration of TP that stimulated endocrine changes days beginning at 21 days of age. A 30-day dosing period had been recom- when using an alternate dosing route. Specifically, Freitag and mended previously (Kelce and Wilson, 1997). Additional changes included a) Do¨cke (1987) reported decreased testicular weights when ad- the weight of the LABC was not recorded due to the redundancy of this end point, because the weights of several other tissues (e.g., epididymides, prostate, ministering 50 ␮g TP/kg/day subcutaneously during the pre- seminal vesicles) are androgen dependent. Furthermore, the LABC is some- and peripubertal periods. With KETO, oral administration at 24 what laborious to dissect during a large-scale screening assay; a) whole- mg/kg/day to male rats for 30 days decreased sperm number prostate weights were measured rather than ventral prostate weights, to reduce and percent motile sperm and increased percent abnormal variability introduced during the dissection process; c) after fixation, thyroid sperm (Vawda and Davies, 1986). The dose of TL (220 mg/ weights were recorded in order to monitor thyroid size; and d) histological evaluations were done on the thyroid, but not the testes and epididymides. The kg/day) was greater than the dose required to inhibit aromatase optional measurements of serum testosterone, DHT, and liver weights were in humans but was not anticipated to produce antiandrogenic added to this study. In our initial work with this assay, we focused on puberty effects (Vigersky et al., 1982). FIN (10 mg/kg/day) given by onset, the reproductive organs, and the thyroid; hence, accessory sex organ and gavage for 14 or 20 days decreased seminal vesicle, levator ani, liver weights were not collected for compounds run in the first experiment, and ventral prostate weights in rats (Di Salle et al., 1993; which included low-dose TP, KETO, low-dose FIN, and TL. Ultimately, this omission did not alter the outcome of the study, because accessory sex gland Haüsler et al., 1996) and significantly decreased fertility at 80 weights were collected at higher dose levels of TP and FIN, and effects of mg/kg/day (Wise et al., 1991). The dose of the aromatase KETO and TL were detected without inclusion of these end points. inhibitor FAD was based on previous reports, wherein FAD at Observations. All animals were examined at least once per day throughout 6.0 mg/kg/day severely decreased the number of estrous cycles the study for clinical signs of toxicity. Body weights were recorded daily and (Nunez et al. 1996), delayed puberty onset, and decreased used to adjust dose volume. Body weights were compared statistically at 21, uterine weight in juvenile female rats (Marty et al., 1999). 27, 34, 41, and 50 days of age, and body weight gains were analyzed during the intervals of 21–27, 21–34, 21–41, and 21–50 days of age. Males were examined daily for PPS beginning at 35 days of age. On the day MATERIALS AND METHODS that PPS was achieved, age and body weights of the affected animals were recorded. If an animal had not achieved PPS by 50 days of age, that animal was Animals. Litters of CD௡ (Sprague-Dawley–derived) rats were received arbitrarily assigned a value of 51 days of age. This artificial value was applied from Charles River Laboratories (CRL; Portage, MI). At CRL, pups from to one animal in each of the KETO and low-dose TP groups and to two animals contemporary litters were cross-fostered into study litters to reach the required in the high-dose FIN group. number of male pups per litter. Cross-fostered pups from the same litter were Pathology. Approximately 24 h after the final dose, animals were not placed in multiple study litters, in order to control for litter effects at the weighed, anesthetized with methoxyflurane, and given a limited gross nec- time of randomization into treatment groups. Once received at 14 days of age, ropsy. The testes, epididymides, seminal vesicles, prostate, and liver were pups were housed and maintained under laboratory conditions as previously removed and weighed. Seminal vesicle weights included seminal fluid. Thy- described (Marty et al., in press). roid glands were removed, fixed in neutral, phosphate-buffered 10% formalin, Chemicals. TP and FIN were purchased from Sigma (St. Louis, MO). weighed, and examined histologically. Terminal serum samples were collected KETO was from ICN Biomedicals (Costa Mesa, CA), and TL tablets were for analysis of testosterone and DHT. Hormone assays were conducted by purchased from Bristol-Myers-Squibb (Princeton, NJ). Novartis Pharmaceuti- Anilytics, Inc. (Gaithersburg, MD). cals Corporation (Summit, NJ) generously donated FAD for these experiments. Statistical analysis. Age at PPS, body weight at PPS, body weights, serum Compound purity was estimated to be Ն 98% by its respective chemical hormone data, and absolute and relative organ weights were evaluated by supplier. Bartlett’s test for equality of variances (Winer, 1971). Due to heterogeneity of Experimental design. In the first series of experiments, 10 litters, each variances, some data transformations were required prior to analyses. Data that containing 12 pups and 1 lactating dam, were used. Each litter comprised six were log transformed included thyroid weights for TP-, KETO- and TL-treated male and six female pups, so that both the male and female pubertal onset animals, relative thyroid weights for KETO- and TL-treated groups, relative assays could be conducted simultaneously (female assay results in Marty et al., seminal vesicle weights for TP- and FIN-treated animals, and serum testos- 1999). In the first experiment, males were treated with vehicle, low-dose TP, terone and/or DHT levels in the TP, KETO, and FAD treatment groups. low-dose FIN, KETO, and TL. In the second experiment, 11 litters containing Heterogeneity of variance required that inverse weights for relative testicular, 10 male pups per litter were used. In this experiment, rats were treated with epididymal, and thyroid weights be used for TP data. After Bartlett’s test, age high-dose TP, high-dose FIN, and FAD. A second manuscript with male at PPS, body weight at PPS, serum hormone data, body weights, and absolute pubertal onset assay results using different chemicals is reported separately organ weights were evaluated by an analysis of covariance (ANCOVA) with (Marty et al., in press). Pups were weighed at weaning (21 days of age) and body weight at weaning as the covariate, as was recommended in the EDSTAC randomized into treatment groups such that each group had approximately report (EDSTAC, 1998). Relative organ weights were evaluated by analysis of equal mean body weights and variances. Furthermore, pups were blocked by variance (ANOVA). In accordance with the recommendations in the EDSTAC MALE PUBERTAL ASSAY AND STEROID INHIBITORS 287

FIG. 1. Mean body weight gains (X Ϯ SD) for CD rats treated for 30 days with various endocrine-active compounds. Body weights were monitored daily from days 21 to 50 and were analyzed statistically for the intervals (A) 21–27 and 21–34 days of age; and (B) 21–41 and 21–50 days of age. Body weight gains for animals treated with KETO, TL, and FAD differed from controls at one or more of the time intervals examined. Asterisks (*) mark values signiﬁcantly different from controls at p Ͻ 0.05; n Ն 9 animals per treatment group.

final report (1998), significant differences (p Ͻ 0.05) were examined using 8.8%, respectively (Fig. 2; control terminal body weight was least square means (LSM) to compare the vehicle to treatment groups during 255.4 Ϯ 21.6 g). Neither body weight gains nor terminal body post hoc comparisons. In most cases, analyses were grouped based on test weights differed significantly with TP or FIN treatments. compound such that each analysis contained results from control animals compared with results from animals treated with one test compound. The level of statistical significance for all analyses was set a priori at ␣ ϭ 0.05. Preputial Separation The mean ages and body weights for PPS in the various RESULTS treatment groups are illustrated in Figures 3A and 3B. Control animals achieved PPS at 44.4 Ϯ 2.0 days of age and at a mean Observations, Body Weights, and Body Weight Gains body weight of 237.2 Ϯ 19.4 g. In an evaluation of assay sensitivity to androgenic com- There were no significant treatment-related clinical observations in any of the treatment groups. One animal from the low-dose FAD group was removed from study due to ocular enlargement unrelated to treatment. Mean body weight gain data are shown in Figures 1A and 1B. Males treated with TL and both doses of FAD exhibited reduced body weight gains at one or more of the time intervals evaluated. Animals treated with TL had body weight gains that were 6, 8, and 12% less than control animals during the intervals from 21–34, 21–41, and 21–50 days of age, respectively. Low-dose FAD–treated animals had significantly decreased body weight gains over 21–41 days of age (8% lower than controls), and animals treated with both doses of FAD had 14% decreases in body weight gains over the entire dosing period. KETO significantly decreased body weight gains during the 21–41 day interval (7% less than controls), but body weight gains for these animals were not significantly different FIG. 2. Final mean body weights of 51-day-old male CD rats treated with various endocrine-active compounds for 30 days. TL- and FAD-treated ani- over the entire dosing period (21–50 days of age). Terminal mals had significantly decreased terminal body weights when compared with body weights were decreased significantly in TL-treated and vehicle-treated control animals. Asterisks (*) denote values significantly dif- low- and high-dose FAD–treated animals by 7.3, 10.7, and ferent from controls at p Ͻ 0.05; n Ն 9 animals per treatment group. 288 MARTY, CRISSMAN, AND CARNEY

FIG. 3. Mean age (A) and body weight (B) at which PPS was first detected in male CD rats treated with various endocrine-active substances (X Ϯ SD). Animals were dosed by oral gavage beginning at 21 days of age and examined daily for PPS beginning at 35 days of age. Low-dose FIN significantly accelerated PPS, whereas TL significantly delayed PPS. Animals given low-dose FAD achieved PPS at a significantly lower body weight than control animals. Asterisks (*) denote values statistically different from the control group at p Ͻ 0.05; n Ն 9 animals per treatment group. pounds, neither dose of TP significantly altered the age or body 0.001), whereas FAD did not alter the age at puberty onset, but weight at PPS. However, the broad-spectrum steroid biosyn- animals achieved PPS at a lower body weight (215.6 Ϯ 33.5 at thesis inhibitor KETO (24 mg/kg/day) delayed puberty onset to 0.6 mg/kg/day, p ϭ 0.023; 220.1 Ϯ 16.2 at 6.0 mg/kg/day, p ϭ 46.1 Ϯ 2.5 days of age, although this delay was not statistically 0.070). The mean body weights at puberty onset did not differ identified (p ϭ 0.051). FIN, a 5␣-reductase inhibitor, signifi- in KETO-, FIN-, or TL-treated animals. cantly accelerated PPS at 20 mg/kg/day (42.9 Ϯ 1.4, p ϭ 0.041), but delayed this end point at 80 mg/kg/day (45.6 Ϯ 3.0; Liver Weights p ϭ 0.074). Mixed results were seen with the aromatase inhibitors, TL (220 mg/kg/day) and FAD (0.6 or 6.0 mg/kg/ Absolute and relative liver weights, optional end points in day). TL significantly delayed the age at PPS (47.0 Ϯ 1.6; p ϭ the male pubertal assay, are illustrated in Figure 4. Low-dose

FIG. 4. Mean liver weights (X Ϯ SD) were collected for each treatment group to examine the potential for increased liver metabolism and enhanced hormone clearance. Low-dose FAD significantly decreased absolute liver weight, but relative liver weight was not changed with this treatment. FIN (80) significantly increased relative liver weight, suggesting that enzyme induction has occurred with this treatment. Liver weights were not collected for animals dosed with low-dose TP, low-dose FIN, KETO, or TL. Asterisks (*) mark significantly different liver weights p Ͻ 0.05; n Ն 9 animals per treatment group. MALE PUBERTAL ASSAY AND STEROID INHIBITORS 289

FIG. 5. Mean reproductive organ weights (X Ϯ SD) in CD rats following a 30-day treatment with endocrine-active compounds. TL significantly decreased absolute testes weight in comparison with control animals (A), whereas relative testes weights were not altered among any of the treatment groups (B). Absolute epididymal weights were decreased significantly with KETO, TL, and FIN (C). These same treatments altered relative epididymal weights (D). Asterisks (*) denote values significantly different from controls at p Ͻ 0.05; n Ն 9 animals per treatment group.

FAD significantly decreased absolute liver weights compared 0.396 Ϯ 0.035 g and 0.156 Ϯ 0.020 g/100 g body weight (Figs. with liver weights in control animals (Fig. 4A; control liver 5C and 5D). KETO and FIN (both doses), compounds which weight was 8.842 Ϯ 1.003 g); however, this decrease appeared inhibit androgen synthesis, and the aromatase inhibitor TL to be secondary to decreased body weights (Fig. 4B). Relative significantly decreased absolute epididymal weights by 17.4, liver weights were significantly increased by 7.6% in animals 13.4, 14.9, and 21.2%. These compounds also reduced relative treated with high-dose FIN. Note that liver weights were not epididymal weights. The more specific aromatase inhibitor recorded in our initial experiments with the male pubertal onset FAD, and the androgen TP, were without effect on this param- assay; therefore, there are no values for liver weights in ani- eter. mals treated with low-dose TP, low-dose FIN, KETO, or TL. Accessory Sex Gland Weights Reproductive Organ Weights Monosson et al. (1999) and Asbhy and Lefevre (2000) Control animals in this study had mean absolute and relative recently examined the relationship between accessory sex testicular weights of 2.603 Ϯ 0.2312 g and 1.024 Ϯ 0.108 gland weight, body weight, and age in peripubertal male rats. g/100 g body weight, respectively (Fig. 5). Only the aromatase These investigators consider absolute accessory sex gland inhibitor TL caused a significant (6.6%) decrease in absolute weights in juvenile animals to be most relevant. In this study, testicular weight (p ϭ 0.046). both absolute and relative accessory sex gland weights are Mean paired epididymal weights in control animals were presented for completeness (Figs. 6A–6D). 290 MARTY, CRISSMAN, AND CARNEY

FIG. 6. Absolute (A) and relative (B) prostate weights (X Ϯ SD) were decreased significantly in male CD rats treated for 30 days with high-dose FIN. This same treatment altered absolute (C) and relative (D) seminal vesicle weights. Note that seminal vesicle weights include seminal fluid. Accessory sex organ weights were not collected for low-dose TP, KETO, low-dose FIN, or TL groups. Asterisks (*) mark values significantly different from controls at p Ͻ 0.05; n Ն 9 animals per treatment group.

Of the steroid biosynthesis inhibitors tested, only FIN (80 statistical analyses, the interanimal variability inherent in these mg/kg/day) significantly reduced prostate weight by 30.4% and hormonal end points limited their utility. For example, KETO seminal vesicle weight by 64.0% compared with control values reduced mean serum testosterone by 43% and FIN reduced (controls: 0.598 Ϯ 0.071 g for prostate and 0.472 Ϯ 0.099 g for mean serum DHT by 45% compared with their respective seminal vesicles; p ϭ 0.0001 in both cases). None of the other controls (control testosterone ϭ 2.12 Ϯ 1.54 ng/ml; control treatments significantly altered accessory sex gland weights. DHT ϭ 0.62 Ϯ 0.47 ng/ml); however, these changes were not Note that in our initial experiment with the male pubertal onset statistically identified. Furthermore, FIN increased mean serum assay, accessory sex gland weights were not recorded for the testosterone by 30%, although this change was not significant low-dose TP and FIN groups, and the KETO and TL treatment (p ϭ 0.208). groups. Despite this variability, serum hormone data revealed two significant differences. First, there was a significant reduction Serum Testosterone and DHT Levels (64%) in mean serum testosterone with 0.1 mg TP/kg/day, but To examine the effectiveness of the steroid biosynthesis high-dose TP did not affect serum testosterone levels. Second, inhibitors, levels of serum testosterone and DHT were ana- the aromatase inhibitor TL significantly reduced serum testos- lyzed in samples collected 24 h after administration of the final terone by 50% (p ϭ 0.032), whereas the more specific aro- dose of test material (Fig. 7). Despite transforming data for matase inhibitor FAD was without effect (p ϭ 0.757). MALE PUBERTAL ASSAY AND STEROID INHIBITORS 291

variety of known endocrine-active agents. Male rats were treated with test compound from 21 through 50 days of age, examined for pubertal onset, and organ weights (testes, epididymides, prostate, seminal vesicles, thyroid, and liver) were measured at 51 days of age. Serum testosterone and DHT were measured. Results from this study were compared with results obtained with other proposed Tier I assays (Tables 1 and 2). One factor of critical importance in the pubertal male screening assay is the effect of body weight on assay end points. In this study, TL significantly delayed age at PPS, yet animals had approximately equal body weights to control animals at the time PPS was achieved. This scenario poses the question as to whether the delay in pubertal onset was mediated by changes in rate of growth. It seems unlikely that the effect of TL on pubertal onset is solely mediated by changes in rate of growth, FIG. 7. Mean serum levels of testosterone and DHT (X Ϯ SD) showed a large degree of interanimal variability; however, serum testosterone concen- because TL also affected reproductive and accessory sex gland trations were decreased statistically with low-dose TP and TL. Serum DHT organ weights and serum testosterone levels. However, in was not significantly affected with any of the treatments. Asterisks (*) indicate instances where body weight gains prior to puberty are altered, values significantly different from controls at p Ͻ 0.05; n Ն 9 animals per the effect of body weight should be considered in view of other treatment group. assay end points. To test the ability of the male pubertal assay to detect Thyroid Weights and Histology androgens, males were dosed with 0.1 or 0.4 mg TP/kg/day. TP Both FAD doses affected absolute thyroid weight by Ն failed to significantly alter body weight gains, the age or body 12.2% (Fig. 8A; control thyroid weight was 18.8 Ϯ 2.9 mg). weight at PPS, or androgen-dependent organ weights. The lack However, this difference was lost when relative thyroid of effect on puberty onset agrees with work by Freitag and weights were examined (Fig. 8B). Thyroid glands from study Do¨cke (1987), who reported that puberty was not altered in rats animals also were evaluated histologically, but no treatment- treated from 30 days of age through PPS with subcutaneous related effects were noted. injections of 0.1 mg TP/kg/day. Serum testosterone was decreased significantly at low-dose TP. This effect on serum DISCUSSION testosterone appears to be either transient or artifactual due to the lack of a dose response and the presence of a slight, These experiments were undertaken to evaluate the sensi- nonsignificant increase in testicular weight, which is inconsis- tivity and specificity of the male pubertal assay to detect a tent with reduced testosterone production. Furthermore, de-

FIG. 8. Mean absolute thyroid weights (X Ϯ SD) were decreased signiﬁcantly in male CD rats treated with FAD (A); however, when terminal body weight was considered, there was no effect on relative thyroid weights by FAD or any other treatment (B). Asterisks (*) denote values signiﬁcantly different from controls at p Ͻ 0.05; n Ն 9 animals per treatment group. 292 MARTY, CRISSMAN, AND CARNEY

TABLE 1 Results from the 14-day Intact Male Assay Using Similar EACs

Compound Testes Epididymides Accessory sex gland Prostate Seminal vesicles Testosterone DHT Estradiol

Testosteronea —— 111111 Ketoconazoleb,c — 222— 22— Finasteridec —— 222 — 2 — Anastrozolec —— 2 — 2 ——2

Note. Alterations in other hormones (prolactin, gonadotropins, thyroid stimulating hormone, and thyroid hormones) not included in this table. See original references. aData from O’Connor et al., 2000. bData from Cook et al., 1997. cData from O’Connor et al., 1998. creases in serum testosterone and DHT were not accompanied (Table 2). Interestingly, these investigators did not find signif- by reduced epididymal or accessory sex gland weights, al- icant changes in epididymal weights with 25 mg KETO/kg/day though these organs are androgen dependent (Cook et al., at 35–36 or 54–55 days of age when it was administered for 14 1993). Taken together, these data suggest that the male puber- or 20 days, respectively. This may suggest that 49–51 days of tal onset assay was unable to detect TP at doses up to 0.4 age is a highly sensitive time point in which to measure mg/kg/day. Because androgens have been linked with alter- changes in epididymal weight. ations in several end points measured in the male pubertal KETO caused a 43% reduction in serum testosterone and a assay, it seems likely that higher doses of TP would have been 21% decrease in serum DHT. These androgens are known to detected. contribute to the maintenance of epididymal weight (Awoniyi Results for testosterone using an alternate Tier I assay, the et al., 1993; George et al., 1989). It seems likely that a greater 14-day intact male assay, are presented in Table 1. In this change in serum hormone concentrations would have been assay, intraperitoneal injection of testosterone at dose levels detected if the animals had been necropsied at a time point of Ն 0.5 mg/kg/day produced significant organ weight and closer to the administration of the last dose of KETO instead of hormonal effects. These dose levels are higher than those 24 h later (Wang et al., 1992). Because of the rapid recovery employed in the present study and utilize a route of exposure from KETO treatment, many studies administer multiple doses likely to result in greater activity. of KETO per day to maintain suppression of testosterone levels Using the broad-based steroid biosynthesis inhibitor KETO (English et al., 1986; Irsy and Koranyi, 1990; Trachtenberg, at 24 mg/kg/day, the age at PPS was increased slightly (not 1984). In the final version of the EDSTAC male pubertal assay statistically identified) by 1.7 days. KETO had no effect on (EDSTAC, 1998), animals are sacrificed on the same day that testicular weight, but significantly reduced both absolute and the final dose of test material is administered, which presum- relative epididymal weights. This result confirms a previous ably would result in more uniform steroid hormone suppres- finding, where epididymal weight was decreased at 49–50 days sion. The weight of evidence for KETO endocrine activity of age following a 14-day treatment with 15 mg KETO/kg/day would have been improved with accessory sex gland weight

TABLE 2 Summary of Pubertal Male Assays Conducted in Another Laboratory

Compound Age at PPS Testes weight Epididymal weight Seminal vesicles Prostate

Ketoconazolea —b —b 2b 2c 2c Finasteridea NE 22b 2b 2b CPAa,d NE 2b 2b 2c 2c Anastrozolea,d NE —b —b 11

Note. Note that exposure times in these studies varied from the present study, in which males were dosed from 21 to 50 days of age. NE, not evaluated in the studies by Ashby and Lefevre. aData from Ashby and Lefevre, 2000. bDenotes similar results with the present study. TL results are compared with the antiandrogen CPA (cyproterone acetate) and FAD results are compared with the aromatase inhibitor anastrozole. cThese parameters were not measured in the present study for KETO or TL. dData from Ashby and Lefevre, 1997. MALE PUBERTAL ASSAY AND STEROID INHIBITORS 293 data (Table 2), which were not collected for this compound in onstrated that a high dose of TL (75 mg/day) competed with the present study. In the 14-day intact male assay, KETO was DHT for androgen-receptor binding and produced antiandro- detected due to hormonal changes at 25 mg/kg/day, and at 125 genic effects in male rats. Although the dose level of TL used mg/kg/day decreases in androgen-sensitive organs weights in the present study was less than that used by Vigersky and were seen (Table 1). colleagues, the results of this study on puberty onset and The endocrine activity of FIN, a 5␣-reductase inhibitor that reproductive organ weights are consistent with the effects blocks the conversion of testosterone to DHT, was readily reported for other antiandrogens (Ashby and Lefevre, 1997, detected with the male pubertal assay. Animals given high- 2000; Dhar et al., 1983; Dhar and Setty, 1990a,b; Gray et al., dose FIN (80 mg/kg/day) had a slight increase in age and body 1999; Monosson et al., 1999). Although the decreased repro- weight at puberty, a response profile that may be indicative of ductive organ weights by TL concur with the antiandrogen a true endocrine-active compound. Unlike other male pubertal hypothesis (see results for CPA in Table 2), serum testosterone studies (Table 2), FIN had no effect on testicular weights; levels were decreased in TL-treated animals, a finding which is however, epididymal, prostate, and seminal vesicle weights contrary to other studies with antiandrogens in which eleva- were significantly decreased. These results were predicted, tions in serum testosterone were observed (Dhar and Setty, because prostate and seminal vesicle growth are under DHT 1990b; O’Connor et al., 1998). This difference may be related control (Wise et al., 1991). Furthermore, these findings are to the time interval (24 h) between antiandrogen dosing and consistent with previous reports (Tables 1 and 2). Interestingly, serum sample collection for testosterone analysis. mean age at PPS was increased in high-dose FIN-treated ani- Our inability to detect the specific aromatase inhibitor fadro- mals (not significant), yet males achieved PPS at a significantly zole with the male pubertal assay differs from previous work younger age than control animals with low-dose FIN treatment. with anastrozole (Table 2). Similar to FAD, anastrozole is The reason for this response is not known. The rationale for considered to be highly specific for aromatase inhibition accelerated PPS is further complicated by the serum hormone (Dukes et al., 1996). Although the reason for this discrepancy data, whereby high-dose FIN increased serum testosterone and is not readily apparent, the study by Ashby and Lefevre used a low-dose FIN decreased serum testosterone relative to control different strain of rats and an alternate dosing duration (14 animals. One hypothesis is that the relative proportion of days), with animals sacrificed 24 h after the final dose at 36–37 testosterone to DHT influenced the achievement of PPS, a or 49–50 days of age. As shown in Table 1, aromatase inhib- possibility given that the male reproductive system is shifting itors have different effects in adult rats, decreasing relative from DHT to testosterone as the dominant androgen during this seminal vesicle and accessory sex gland weights. period (Ojeda and Urbanski, 1994). Alternatively, the body To examine the requirement to control for weaning weight weights of low-dose FIN-treated animals were the same as the during statistical analyses, data from these experiments also control animals at the time they reached puberty, suggesting were analyzed by analysis of variance (ANOVA) with Dun- that these FIN-treated animals had slightly faster growth rates. nett’s test and compared with the results using ANCOVA and Indeed, although not statistically identified, the body weight LSM analyses. Although most analyses yielded similar results gains and final body weights of these animals were slightly whether ANOVA or ANCOVA was used, four differences higher than weights in control animals. Lastly, the lack of a were detected with the ANCOVA analysis that were not de- readily apparent explanation for the acceleration of PPS by tected using ANOVA and Dunnett’s test. Specifically, low-dose FIN allows the possibility that this result is artifac- ANOVA failed to identify a) significant decreases in terminal tual. Despite this unusual finding, the variety of effects with body weight (p ϭ 0.12) in TL-treated animals and b) decreased FIN treatment readily suggests endocrine activity (Table 2). absolute testes weight (p ϭ 0.11) with TL treatment. Further- The capacity of the male pubertal assay to detect aromatase more, Dunnett’s test failed to detect c) statistically significant inhibitors is somewhat uncertain, particularly in light of the differences in age at PPS for low-dose FIN–treated animals and limited role of estrogens in male pubertal onset. In this study, d) decreased absolute thyroid weights for both doses of FAD, the two aromatase inhibitors (TL and FAD) produced different although the overall ANOVA analyses were significant in each results. Although both compounds significantly decreased case (p ϭ 0.03). A TL-induced decrease in testicular weight body weight gains, a finding reported previously for aromatase was consistent with other effects produced by TL treatment, inhibitors (Nunez et al., 1996; Vanderschueren et al., 1997), which were detected by ANOVA analyses. Although an accel- TL also delayed PPS, decreased absolute testicular and epidid- eration in PPS with low-dose FIN was not identified with ymal weights, and reduced serum testosterone levels. FAD did Dunnett’s test, this analysis identified significantly decreased not significantly alter reproductive or accessory sex gland epididymal weight with low-dose FIN. The effect of FAD on weights or serum hormone concentrations. The overall re- absolute thyroid weight was not detected by Dunnett’s test, but sponse differences between TL, a less-specific aromatase in- this decrease was secondary to decreased body weights. Nei- hibitor, and FAD, a highly specific aromatase inhibitor, suggest ther ANOVA nor ANCOVA detected differences in relative that TL is not operating solely, if at all, through aromatase thyroid weight with FAD treatment. These statistical discrep- inhibition. A previous report by Vigersky et al. (1982) dem- ancies related to the use of Dunnett’s test, which controls type 294 MARTY, CRISSMAN, AND CARNEY

I error with multiple comparisons (␣ ϭ 0.05), may reflect the Cook, J. C., Mullin, L. S., Frame, S. R., and Biegel, L. B. (1993). Investigation more conservative nature of this type of analysis or the mar- of a mechanism for Leydig cell tumorigenesis by linuron in rats. Toxicol. ginal nature of these findings. Overall, the use of either Appl. Pharmacol. 119, 195–204. ANOVA or ANCOVA yielded similar results with regard to Cook, J. C., Kaplan, A. M., Davis, L. G., and O’Connor, J. C. (1997). Development of a tier I screening battery for detecting endocrine-active detecting EACs. compounds (EACs). Regul. Toxicol. Pharmacol. 26, 60–68. Overall, the male pubertal assay is capable of detecting Dhar, J. D., Srivastava, S. R., and Setty, B. S. (1983). Effect of flutamide, a chemicals that operate through a variety of mechanisms. nonsteroidal antiandrogen on functional maturation of the epididymis of rat. KETO, TL, and FIN were deemed positive for endocrine Exp. Clin. Endocrinol. 82, 140–144. activity in the current male pubertal assay study; however, TP Dhar, J. D., and Setty, B. S. (1990a). Changes in testis, epididymis and other and FAD were considered negative. The negative designation accessory organs of male rats treated with anandron during sexual matura- for TP assumes that altered serum testosterone alone was tion. Endocr. Res. 16, 231–239. insufficient to trigger a Tier II test in the absence of other Dhar, J. D., and Setty, B. S. (1990b). Effect of a nonsteroidal antiandrogen, effects. This interpretation seems reasonable, given the vari- anandron, on the reproductive system and fertility in male rats. Contracep- tion 42, 121–138. ability in serum hormone data and the idea that hormone data Di Salle, E., Giudici, D., Briatico, G., Ornati, G., and Panzeri, A. (1993). represent a single time point, whereas organ weights represent Hormonal effects of turosteride, a 5␣-reductase inhibitor, in the rat. J. a culmination of events over time. Hypothetically, higher doses Steroid Biochem. Mol. Biol. 46, 549–555. of TP would be detectable with this assay, but the ability of this Dukes, M., Edwards, P. N., Large, M., Smith, I. K., and Boyle, T. (1996). The assay to identify aromatase inhibitors remains questionable. preclinical pharmacology of “Arimidex” (Anastrozole; ZD1033) – A potent, With regard to assay specifics, the present study suggests selective aromatase inhibitor. J. Steroid Biochem. Mol. Biol. 58, 439–445. that PPS was not the most sensitive end point for detection of EDSTAC Final Report. (1998). Endocrine Disrupter Screening and Testing these endocrine-active compounds. Changes in the age at PPS Advisory Committee Final Report. U.S. Environmental Protection Agency. were observed only at concentrations that altered reproductive/ Internet access at URL: http://www.epa.gov/oscpmont/oscpendo/history/ finalrpt.htm. accessory sex gland organ weights. In most cases, organ weight English, H. F., Santner, S. J., Levine, H. B., and Santen, R. J. (1986). Inhibition changes observed in this study deviated from control values of testosterone production with ketoconazole alone and in combination with by Ͻ 22%, with the exception of marked effects on accessory a gonadotropin releasing hormone analogue in the rat. Cancer Res. 46, sex gland weights by FIN. This may reflect the adaptability of 38–42. the male endocrine system to maintain homeostasis. Measure- Freitag, J., and Do¨cke, F. (1987). Differential effects of chronic testosterone ment of steroid hormones had limited value in this study due to treatment on the onset of puberty in male rats. Exp. Clin. Endocrinol. 90, interanimal variability. A greater number of animals per treat- 361–364. ment group would be needed in future studies to improve the George, F. W., Johnson, L., and Wilson, J. D. (1989). The effect of a 5 utility of these data. Additional work is needed to establish alpha-reductase inhibitor on androgen physiology in the immature rat. Endocrinology 125, 2434–2438. criteria for a positive response in the male pubertal onset assay, Gray, L. E., Jr., Ostby, J., Monosson, E., and Kelce, W. R. (1999). Environ- especially in light of the apical nature of its end points. Finally, mental antiandrogens: Low doses of the fungicide vinclozolin alter sexual it is worth noting that the compounds used in the present study differentiation of the male rat. Toxicol. Ind. Health 15, 48–64. were strong endocrine-active agents. The capacity of the male Haüsler, A., Allegrini, P. R., Biollaz, M., Batzl, Ch., Scheidegger, E., and pubertal onset assay to detect weakly active environmental Bhatnagar, A. S. (1996). CGP 53153: A new potent inhibitor of 5␣- agents, including aromatase inhibitors, requires further inves- reductase. J. Steroid Biochem. Mol. Biol. 57, 187–195. tigation. Irsy, G., and Koranyi, L. (1990). Neuroendocrinological effects of ketoconazole in rats. Acta Endocrinol. (Copenh). 122, 409–413. REFERENCES Kelce, W. R., and Wilson, E. M. (1997). Environmental antiandrogens: De- velopmental effects, molecular mechanisms, and clinical implications. J. Ashby, J., and Lefevre, P. A. (1997). The weanling male rat as an assay for Mol. Med. 75, 198–207. endocrine disruption: Preliminary observations. Reg. Toxicol. Pharmacol. Korenbrot, C. C., Huhtaniemi, I. T., and Weiner, R. I. (1977). Preputial 26, 330–337. separation as an external sign of pubertal development in the male rat. Biol. Ashby, J., and Lefevre, P. A. (2000). The peripubertal male rat assay as an Reprod. 17, 298–303. alternative to the Hershberger castrated male rat assay for the detection of Marty, M. S., Crissman, J. W., and Carney, E. W. (1999). Evaluation of the anti-androgens, oestrogens and metabolic modulators. J. Appl. Toxicol. 20, EDSTAC female pubertal assay in CD rats using 17␤-estradiol, steroid 35–47. biosynthesis inhibitors, and a thyroid inhibitor. Toxicol. Sci. 52, 269–277. Awoniyi, C. A., Reece, M. S., Hurst, B. S., Faber, K. A., Chandrashekar, V., Marty, M. S., Crissman, J. W., and Carney, E. W. (2001). Evaluation of the and Schlaff, W. D. (1993). Maintenance of sexual function with testosterone male pubertal assay’s ability to detect thyroid inhibitors and dopaminergic in the gonadotropin-releasing hormone-immunized hypogonadotropic infer- agents. Toxicol. Sci. 60, 63–76. tile male rat. Biol. Reprod. 49, 1170–1176. Monosson, E., Kelce, W. R., Lambright, C., Ostby, J., and Gray, L. E., Jr. Clark, R. L. (1999). Endpoints of reproductive system development. In An (1999). Peripubertal exposure to the antiandrogen fungicide, vinclozolin, Evaluation and Interpretation of Reproductive Endpoints for Human Risk delays puberty, inhibits the development of androgen-dependent tissues, and Assessment, pp. 27–62. International Life Sciences Institute, Health and alters androgen receptor function in the male rat. Toxicol. Ind. Health 15, Environmental Science Institute, Washington, D.C. 65–79. MALE PUBERTAL ASSAY AND STEROID INHIBITORS 295

Nunez, S. B., Blye, R. P., Thomas, P. M., Reel, J. R., Barnes, K. M., Malley, Vanderschueren, D., van Herck, E., Nijs, J., Ederveen, A. G., De Coster, R., J. D., and Cutler, G. B., Jr. (1996). Recovery of reproductive function in rats and Bouillon, R. (1997). Aromatase inhibition impairs skeletal modeling and treated with the aromatase inhibitor fadrozole. Reprod. Toxicol. 10, 373– decreases bone mineral density in growing male rats. Endocrinology 138, 377. 2301–2307. O’Connor, J. C., Cook, J. C., Slone, T. W., Makovec, G. T., Frame, S. R., and Vawda, A. I., and Davies, A. G. (1986). An investigation into the effects of Davis, L. G. (1998). An ongoing validation of a tier I screening battery for ketoconazole on testicular function in Wistar rats. Acta Endocrinol. detecting endocrine-active compounds (EACs). Toxicol. Sci. 46, 45–60. (Copenh) 111, 246–251. O’Connor, J. C., Davis, L. G., Frame, S. R., and Cook, J. C. (2000). Evaluation Vigersky, R. A., Mozingo, D., Eil, C., Purohit, V., and Bruton, J. (1982). The of a tier I screening battery for detecting endocrine-active compounds 1 antiandrogenic effects of ⌬ -testolactone (Teslac) in vivo in rats and in vitro (EACs) using the positive controls testosterone, coumestrol, progesterone, in human cultured ﬁbroblasts, rat mammary carcinoma cells, and rat prostate and RU486. Toxicol. Sci. 54, 338–354. cytosol. Endocrinology 110, 214–219. Ojeda, S. R., and Urbanski, H. F. (1994). Puberty in the rat. In The Physiology Wang, J. M., Wu, X. L., You, W., Ling, L. X., Wu, J., and Zhang, G. Y. (1992). of Reproduction. (E. Knobil, J. D. Neill, G. S. Greenwald, C. L. Markert, Pharmacokinetic and pharmacodynamic studies of the effect of ketocon- and D. W. Pfaff, Eds.), Vol. 2, pp. 363–409. Raven Press, New York. azole on reproductive function in male rats. Int. J. Androl. 15, 376–384. Stoker, T. E., Parks, L. G., Gray, L. E., and Cooper, R. L. (2000). Endocrine- nd disrupting chemicals: Prepubertal exposures and effects on sexual matura- Winer, B. J. (1971). Statistical Principles in Experimental Design, 2 ed. tion and thyroid function in the male rat. A focus on the EDSTAC recom- McGraw-Hill Book Company, Inc., New York. mendations. Crit. Rev. Toxicol. 30, 197–252. Wise, L. D., Minsker, D. H., Cukierski, M. A., Clark, R. L., Prahalada, S., Trachtenberg, J. (1984). The effects of ketoconazole on testosterone produc- Antonello, J. M., MacDonald, J. S., and Robertson, R. T. (1991). Reversible tion and normal and malignant androgen dependent tissues of the adult rat. decreases of fertility in male Sprague-Dawley rats treated orally with ﬁn- J. Urol. 132, 599–601. asteride, a 5␣-reductase inhibitor. Reprod. Toxicol. 5, 337–346.