Use of canalicular membrane vesicles (CMVs) from rats, dogs, monkeys and humans to assess drug transport across the canalicular membrane.

INTRODUCTION: A novel application of a Ultrafree filter cartridge/centrifugation method was evaluated to determine uptake in canalicular membrane vesicles (CMVs) from SD rats, beagle dogs, cynomolgus monkeys (common safety species in the pharmaceutical industry) and humans to assess biliary transport. METHODS: CMVs prepared from fresh livers of rats, dogs, monkeys and humans (four donors) were characterized for enrichment, basolateral and Golgi contamination and orientation. The presence of MRP2 and p-glycoprotein (P-gp) were confirmed by Western blots. Uptake of [3H]-leukotriene C4 (LTC4) and [3H]-estradiol-17beta-d-glucuronide (E2-Gluc) was determined at a low substrate concentration and/or by kinetic measurements (K(m) and V(max)). Correlation of in vitro data with in vivo findings was achieved by determining the biliary clearance of E2-Gluc in rats after a single i.v. dose and with literature in vivo data for LTC4. RESULTS: CMVs were highly enriched and minimally contaminated based on marker enzyme activities. Uptake clearance among different species varied by approximately ten-fold (rat > dog = human > monkey) for LTC4 and less than two-fold for E2-Gluc. The lower uptake of LTC4 by human than rat CMVs may be attributed to a higher Km value for human than rat CMVs. Uptake of LTC4 or E2-Gluc by human CMVs showed little inter-subject variability (2-5-fold). Differences in in vitro uptake clearance (10-fold) between LTC4 and E2-Gluc in rat CMVs seemed to correlate with differences in their biliary clearance (4-fold) in rats, consistent with LTC4 and E2-Gluc being a high and a low clearance substrate, respectively. DISCUSSION: A novel application of a Ultrafree filter cartridge/centrifugation method was developed to determine uptake in CMVs from different preclinical animal safety species and humans, and may represent a useful approach to study the mechanism of biliary excretion during drug discovery and development.

3,3'-diindolylmethane, a major condensation product of indole-3-carbinol, is a potent estrogen in the rainbow trout.

Indole-3-carbinol (I3C), a compound found in Brassica vegetables has been widely studied for its chemopreventive properties. I3C has been shown to block tumor initiation and promotion; however, it also acts as a tumor promoter. I3C and some of its acid condensation products, particularly 3,3'-diindolylmethane (I33'), have exhibited antiestrogenic properties. We report that I33' acts as an estrogen in the rainbow trout liver in vitro and in vivo by inducing vitellogenin (Vg), a well-characterized biomarker for estrogens. Precision-cut liver slices from male rainbow trout, Oncorhynchus mykiss, were incubated at 14 degrees C for 96 h in media containing I3C, I33', or a mixture of I3C acid condensation products (RXN) (0-250 microM). I33' and RXN increased Vg levels in rainbow trout liver slices by over 300- and 20-fold, respectively, vs vehicle. The efficacy of I33' induction of Vg was comparable to 17 beta-estradiol (E(2)) with 2500-fold less potency. I33' and E(2) cotreatment resulted in additive Vg induction. Tamoxifen completely inhibited I33'-induced Vg induction, suggesting that Vg induction by I33' is entirely through the estrogen receptor. In vivo, juvenile male rainbow trout were fed I3C, RXN (0-2000 mg/kg), or I33' (0-250 mg/kg) for 2 weeks. At 2000 mg/kg, I3C induced Vg by over 100,000-fold compared to controls, which was comparable to 5 mg/kg 17 beta-estradiol (the dose resulting in maximum induction). I33' was five times as potent as I3C with equal efficacy. The potency of RXN was only 5% of I3C. Again, I33' and E(2) cotreatment resulted in additive Vg induction. I33' may have accounted for Vg increases observed in trout fed I3C as it is present in liver after oral dosing at concentrations (70 microM) expected to maximally induce Vg. In trout, results in vitro and in vivo document that I33' is estrogenic, consistent with our hypothesis that I3C promotes liver cancer in trout by estrogenic pathways.

Determining relative estrogenicity by quantifying vitellogenin induction in rainbow trout liver slices.

Precision cut tissue slices, by modeling the entire organ, are a valuable tool for studying protein induction or inhibition by test chemicals. This manuscript describes parameters to quantify relative estrogenicity of chemicals in rainbow trout liver slices by measuring vitellogenin (Vg) induction, a well-characterized biomarker of estrogen receptor signal transduction. Hank's medium (phenol-red free) supplemented with Hepes, sodium bicarbonate, and 1% bovine serum albumin was utilized. The experimental parameters were optimized using 1000 nM 17beta-estradiol, a potent estrogen in rainbow trout that induces Vg production in vivo. The addition of trout serum and retention of the media was essential, probably to allow for the accumulation of Vg in the slices and media. Histological examination and ATP analyses indicated no toxicity in control or 17beta-estradiol-treated liver slices after 120 h. Induction was 4-fold greater with 25% serum containing media compared to media with 10% serum. We observed Vg induction as great as 500-fold over controls at 96 h in liver slices and media containing 25% serum and 1000 nM 17beta-estradiol. Controls without 17beta-estradiol, incubated in media with 10 or 25% serum, exhibited no detectable Vg production, indicating that the induction seen was not from the media or serum. We observed that 48 h was required for significant Vg induction in the media and liver slices. Maximum induction in slices occurred at 96 h, whereas media Vg levels continued to increase to 120 h, suggesting a time delay between Vg production and excretion by the liver. The feasibility of this model to detect weak environmental estrogens was determined with 0-250 microM o,p'DDE and bisphenol A. Both compounds induced Vg in this model with EC50 values of 10(4) and 2x10(5) higher than E(2), respectively. Our results indicate the importance of media, serum, and time selection for optimal Vg induction. This model allows for the determination of relative estrogenicity of chemicals in a controlled in vitro system while utilizing the advantages of precision cut slice technology.

The non-aromatizable androgen, dihydrotestosterone, induces antiestrogenic responses in the rainbow trout.

In order to satisfy government mandates, numerous studies have been performed categorizing potential endocrine disrupting chemicals as (anti)estrogens or (anti)androgens. We report here that dihydrotestosterone (DHT), a potent, non-aromatizable androgen receptor agonist, induces antiestrogenic responses through direct and/or indirect modulation of vitellogenin (Vg), steroid hormone and total cytochrome P450 levels. DHT and two weak, aromatizable androgens, DHEA and androstenedione (0.05-50 mg/kg per day), were fed to juvenile trout for 2 weeks. DHEA and androstenedione significantly increased blood plasma Vg by up to 30- and 45-fold, respectively (P<0.05, t-test). 17beta-Estradiol (E2) increases were also observed with both androgens, albeit with lower sensitivity. DHT markedly decreased Vg and E2 levels, suggesting that DHEA and androstenedione increased Vg and E2 via conversion to E2 and not by estrogen receptor agonism. DHEA and androstenedione had no effect on total cytochrome P450 content, while DHT significantly decreased P450 content in a dose dependent fashion. These results indicate that alterations in metabolism mediated by androgen receptor binding may be responsible for the Vg and E2 decreases by DHT. In an attempt to decipher between receptor and non-receptor androgenic mechanisms of the observed DHT effects, DHT (0, 50 or 100 mg/kg per day) and flutamide (0-1250 mg/kg per day), an androgen receptor antagonist, were fed to juvenile rainbow trout for 2 weeks. Flutamide alone was as effective as DHT in decreasing E2 and Vg levels in males but did not significantly reverse DHT induced Vg decreases in either sex (P>0.05, F-test). DHT decreases in total P450 content were partially attenuated in males by flutamide co-treatment, but not females, suggesting a partial androgenic mechanism to the P450 decreases as well as a fundamental sex difference responding to androgen receptor binding. Moreover, flutamide alone decreased P450 content by up to 30% in males and 40% in females. These effects may be mediated through direct androgen receptor binding irrespective of whether the binding is agonistic or antagonistic. This study indicates that androgen receptor agonists/antagonists can elicit significant antiestrogenic effects that may not necessarily be mediated through classic receptor binding mechanisms and signal transduction pathways.

Rainbow trout, Oncorhynchus mykiss, as a model for aromatase inhibition.

The feasibility of utilizing rainbow trout, Oncorhynchus mykiss, as an alternative model for studying the inhibition of aromatase (CYP 19) was investigated. The suppression of estrogen-dependent tumors by aromatase inhibitors has been important in the treatment of breast cancer. Estrogens, estrogen precursors and xenoestrogens have been found to promote liver cancer in the trout model. A steroid, 4-hydroxy-4-androstene-3,17-dione (4-OHA), and non-steroids, aminoglutethimide (AG) and Letrozole (CGS 20267), all of which are known aromatase inhibitors in rats and humans, were examined in vitro for activity in trout ovarian microsomes. Aromatase activity was quantified as the release of 3H2O from the conversion of [3H]-4-androstene-3,17-dione to 17beta-estradiol and estrone. Trout ovarian microsomes exhibited activity between 39-60 fmol mg(-1) min(-1) with a calculated Vmax of 71.1 fmol mg(-1) min(-1) when incubated at 25 degrees C with 200 nM 4-androstene-3,17-dione (K(M) = 435 nM). Significant inhibition by 4-OHA up to 80% was seen at 1.5 microM. At 2000 microM, AG decreased aromatase activity by up to 82%. Letrozole reduced aromatase activity a maximum of 90% in a dose-dependent manner, but the Ki (2.3 microM) was 1000-fold higher than reported in human trials. Indole-3-carbinol and some of its derivatives, two DDE isomers and four flavones (except alpha-naphthoflavone) at 1000 microM did not significantly inhibit aromatase in vitro. Letrozole and clotrimazole, fed to juvenile rainbow trout at doses up to 1000 ppm for 2 weeks, were not effective in suppressing dehydroepiandrosterone (DHEA) induced increases in vitellogenin and 17beta-estradiol levels. These results document that trout aromatase is sensitive to inhibition in vitro by known inhibitors of the mammalian enzyme. The mechanism(s) for lack of inhibition in vivo is currently unknown and must be further investigated in order to develop a trout model for studying the role of aromatase in carcinogenesis.