Although it is well known that estrogenic steroidal hormones are able to affect the sexual development and reproduction of fish at low concentrations, no data on environmental effects of the class of progestogenic hormones are available yet. Synthetic gestagens (progestins) are a component in oral contraceptives. Upon their use, a fraction of the progestins will be excreted via urine into the aquatic environment. On the basis of their pharmacological action in mammals, it is supposed that fish reproduction is the most sensitive endpoint for the progestin treatment. In order to test this assumption, the effects of two progestins currently marketed in contraceptive formulations, levonorgestrel (LNG) and drospirenone (DRSP), were investigated in adult fathead minnows (Pimephales promelas) following an Organization for Economic Cooperation and Development 21-d fish reproduction screening assay draft protocol with additional end points. Levonorgestrel was tested at measured concentrations of 0.8, 3.3, and 29.6 ng/L, and DRSP at concentrations of 0.66, 6.5, and 70 μL. Both tested progestins caused an inhibition of reproduction. For LNG, this occurred at concentrations of 2:0.8 ng/L, no no-observed-effect concentration (NOEC) could be defined. Higher concentrations resulted in masculinization of females with de novo synthesis of nuptial tubercles. Drospirenone treatment, however, affected the reproductive success of fathead minnow at concentrations of 6.5 μL and higher with a clear dose-response relationship and a NOEC of 0.66 μL, which is above environmentally relevant concentrations.
Steroidal sex hormones such as estrogens and gestagens are used widely in oral contraceptives and hormone replacement therapy and may enter the aquatic environment via wastewater effluents. It was shown that estrogenic hormones can affect fish at low concentrations in sexual development and reproduction after long-term exposure [1,2]. Knowledge of the environmental effects of progestogenic hormones is scarce, although they represent an important class of the pharmaceutical active ingredients in hormonal medicines.
In humans, natural progesterone is an important regulator during oocyte maturation and pregnancy via progesterone receptor binding. Synthetic gestagens are able to inhibit ovulation and proliferation of the endometrium and thus are applied for contraception. Furthermore, progestins control estrogen-induced endometrial hyperplasia, which is an important pharmacological activity of progestins provided in hormone replacement therapy . Marketed progestins may vary with regard to interaction with other steroid hormone receptors depending on the molecular core structure from which they were derived (progesterone, testosterone, or spironolactone) [4,5]. Derivatives of 19-nortestosterone (go-nane [levonorgestrel] and the estrane group [norethisterone]) show androgenic activity, whereas derivatives of 19-norprogesterone (pregnanes and norpregnanes) are said to interact little with other steroid receptors than the progesterone receptor (PR). However, the 17-hydroxyprogesterone derivatives (cyproterone acetate and medroxyprogesterone acetate) reportedly have potent antiandrogenic or slight androgenic action . Drospirenone as a derivate of spironolactone rather exhibits an antimineralcorticoid and slight antiandrogenic profile .
The role of endogenous progesterone in environmental organisms, and particularly in fish, is the subject of a number of investigations. Progesterone is an important steroidogenic mediator of oocyte growth and maturation as well as spermatogenesis and sperm maturation in teleosts [6,7]. The most important progesterone in fish is 17α,20β-dihydroxy-4-pregnen-3-one (17α,20β-DP), which is obtained by conversion from 17α-hydroxyprogesterone under influence of gonadotropins from the pituitary. Knowledge of cross-reactivity of progestins in fish is scarce, however.
Hormonal medicines have a content of progestins that is higher than that of estrogens by a factor of 3 to 100, depending on the formulation and therapeutic application. Therefore, it can be assumed that environmental concentrations of progestins caused by human excreta are in the same range or higher than those of estrogenic hormones, which can be detected in the low nanogram per liter range in effluents and consequently occur generally in sub-nanogram per liter concentrations in surface waters. Although data on environmental concentrations of progestins are rare, it is expected that the environmentally relevant concentrations are in the range of a few nanograms per liter or below. Medroxyprogesterone was measured and detected in municipal wastewater effluent samples up to 15 ng/L and surface waters up to 1 ng/L [8,9].
The effects of exogenously administered progestins in fish have not been described yet. As a hypothesis for this research, it was assumed, based on the hormonal action in humans, that fish exposed to progestins should primarily show effects on gamete maturation (i.e., that fertility and fecundity [reproductive end points] would be disturbed).
For the assessment of reproductive effects of progestins, two representatives were selected according to their interactions with different steroidal receptors. Levonorgestrel (LNG; 17β-hydroxy-18-methyl-19-nor-17α-pregn-4-en-20-yn-3-on) is derived from 19-nortestosterone and has a low but substantial affinity to the androgen receptor and therefore an androgenic partial effect. In contrast, the more recently introduced progestin drospirenone (DRSP; 6β,7β;15β,16β-dimethylen-3-oxo-17α-pregn-4-en-21,17-carbolacton) as a derivative of 17β-spironolactone has antimineralocorticoid and slight antiandrogenic activity  and a pharmacodynamic profile similar to that of endogenous progesterone [11,12].
Fathead minnow (Pimephales promelas) represents a species widely used in standard testing protocols and provides a sensitive model for determining the environmental effects on reproductive function in fish.
A set of studies with LNG and DRSP was conducted to assess the described pharmacological properties for an aquatic vertebrate species.
MATERIALS AND METHODS
Experimental setup of the 21-d fish reproduction assay
To investigate the reproductive effects, three 21-d fish reproduction assays with LNG and one with DRSP were conducted following a draft guideline of the Organization for Economic Cooperation and Development  but focussing on additional reproductive in contrast to biomarker end points and a published pair-breeding test approach for endocrine-active compounds .
Fathead minnows were bred in the Laboratory of Ecotoxicology, Nonclinical Drug Safety, Bayer Schering Pharma AG, Berlin, Germany. Originally the breeding stock was obtained from Osage Catfisheries.
Four replicate pairs of adult fathead minnows were used for each test concentration of the two investigated progestins and for the control group, each pair held separately in a tank. The pairs were chosen by random allocation. Only adult fish in good health and free from any apparent malformation were used. Labeled 20-L glass aquaria were used as test vessels. Each test vessel contained one spawning tile, which could easily be examined and exchanged when eggs were present.
The fish were exposed to three test concentrations or to the tap water control for a period of 21 d after a pre-exposure phase of at least 21 d under flow-through conditions. The flow rate of 3.5 L/h and replicate were regulated by peristaltic pumps with four channels (Heidolph). The light/dark rhythm was adjusted to 16:8 h.
The hydrographical parameters were recorded weekly. The oxygen concentration was >60% of air saturation value, the pH value was in the range of 6.5 to 8.5, and the temperature was maintained at 25 ± 1°C during the test.
Mortalities, behavior, and visible morphologic abnormalities were recorded daily during the pre-exposure and exposure phase. Egg numbers, clutches per pair, and sizes of egg clutches were recorded daily, as well as morphology including changes in secondary sexual characteristics and behavior during the exposure phase.
The test compounds levonorgestrel, (abbreviated as LNG, Chemical Abstracts Service [CAS] 797–63–7, International Union of Pure and Applied Chemistry [IUPAC] name (-)-13-ethyl-17-hydroxy-18,19-dinor-17α-pregn-4-en-20-yn-3-one, purity 100.2%, solubility in water 1.33 mg/L at pH 6.8) and drospirenone (abbreviated as DRSP, CAS 67392–87–4, IUPAC name 6β,7β; 15β, 16β-dimethylene-3-oxo-17α-pregn-4-ene-21,17-carbolactone, purity 99.6%, solubility in water 9.41 mg/L) were obtained from Bayer Schering Pharma AG.
A range finding test (RFT) was performed at concentrations from 1 to 100 ng/L. Even at the lowest concentration, a strong inhibitory effect on fish reproduction was observable. Therefore, the RFT was followed by a study with lower concentrations of 0.2, 2.0, and 20.0 ng/L. Like the RFT, all concentrations caused a decrease in reproductive success. Because of limitations in the chemical analysis of sub-ng/L levels of LNG and the lack of a clear dose-response relationship, the results are not presented here.
For the definitive study, the lowest test concentration was selected on the basis of the predicted environmental concentration (not higher than 2 ng/L; calculated according to the European Medicines Agency ; http://www.emea.europa.eu). The highest concentration was selected following the RFT.
For the preparation of the test solutions with nominal concentrations of 2.0, 20.0, and 100.0 ng/L tap water, a stock solution was made. The stock solution was prepared by dissolving the compound in dechlorinated tap water, ultra-sonicating for 30 min, and mixing for 24 h. For the different test solutions, the stock solution was diluted with tap water or used undiluted for the highest concentration. The test solutions were filled into syringes (50 ml; BD Plastipak). With the exposure phase first, the test solutions were added to the mixing vessels via the syringe pumps (Infors HT). The syringes were exchanged every 48 h. The correctness of the pumping speed of 1 ml/h was checked at every exchange of the syringes and adjusted if necessary.
Nominal concentrations of 2.0, 20.0, and 200.0 μL DRSP were prepared. Due to the low aqueous solubility of the compound, triethylene glycol (Merck) was used as the solvent. For the preparation of the test solutions, a stock solution was prepared by dissolving the test compound in triethylene glycol, ultrasonicating for 30 min, and mixing for 24 h. For the different treatments, the stock solution was further diluted with triethylene glycol or used undiluted for the highest concentration. These diluted stock solutions were filled into syringes and dosed as described above. For animal welfare reasons, only a vehicle control (triethylene glycol, 0.8 mg/L) without a dilution water control was used because no effect was expected from triethylene glycol.
The gonads of two male and female fish per group were examined by histopathology. After anesthesis with ethyl-3-amonibenzoate methansulfonate (MP Biomedicals), whole animals were fixed in neutral buffered formalin (10%; Sigma Aldrich), dehydrated in graded ethanol and xylene, paraffinized in a tissue processor (Tissue-Tek VIP, Miles Scientific), and embedded in paraffin wax. Cross sections of 3 to 5 μm in thickness were stained with hematoxylin and eosin. These sections were examined microscopically to determine the reproductive condition of the fish (staging according to Leino ).
Additionally, ovaries and testes were analyzed histomorphometrically (Software IQ Easy Measure 1.4.1, INTEQ) similar to the procedure described by Van der Ven et al. . All oocytes present on two sections per female were counted and categorized into the stages oogonia, previtellogenic oocytes, and vitellogenic oocytes. For males, testes were examined from two slides by counting and categorizing the spermatogenic cysts (spermatogonia, spermatocytes, and mature spermatozoa) at three fields of view at 40× magnification.
To determine the content and stability of the stock solutions, samples for analysis were taken directly after preparation and at the end of the exposure period. Additionally, samples from each test solution and from controls were analyzed taken at weeks 1, 2, and 3 of exposure. All solutions were stored at −18°C until analysis.
Analytical determination was carried out by Dr. U. Noack Laboratories.
For LNG, analytical evaluation was carried out via high-performance liquid chromatography (HPLC)–tandem mass spectrometry (2695 Alliance separation module, Waters with a mass-selective detector, Micromass Quattro Premier™ API, Waters) using external standards. Analysis was performed in gradient mode on a C-18 column (X-Terra RP C18, 3.5 μm, 100 × 4.6 mm, SN:01673809213666, Waters). Levonorgestrel was detected in positive mode after electrospray ionization via the mass transitions m/z 313.42 > 109.30 (quantifier) and 313.42 > 91.32 (qualifier). A stock solution of LNG (100 mg/L) in methanol was diluted with mobile phase (50% methanol/50% water containing 0.1% formic acid) to obtain a linear calibration function in this concentration range. The coefficient of correlation (r2) was at least >0.998.
The limit of quantification (LOQ) for LNG was set to 1.5 ng/L due to study requirements. For method validation, replicates of dechlorinated tap water spiked with 1.5 and 15 ng/L (10·LOQ) LNG were analyzed. Mean recoveries of 98 ± 7% for LOQ level and 91 ± 4% for 10·LOQ level were obtained. Two blank replicates of dechlorinated tap water were analyzed analogously to demonstrate that LNG did not occur in the used tap water. The mean recoveries of the fortification samples were between 70 and 110%. Filtrated test samples from the aquaria were treated as described for the fortification samples prior to analysis.
Drospirenone was analyzed via liquid chromatographytandem mass spectrometry (LC-MS/MS, Acquity™ Ultra Performance LC, with Acquity™ Ultra Performance LC triple-quadrupole MS/MS detector, Waters) in gradient mode on a C-18 column via electrospray ionization in positive mode with the mass transitions m/z 366.97 > 97.00 (quantifier) and m/z 366.97 > 91.00 (qualifier). A stock solution of the standard in methanol (100 mg/L) was prepared. Samples for calibration in the range from 0.5 to 100 μL were obtained by dilution with the mobile phase (50% 0.1% formic acid in HPLC-grade water/50% 0.1% formic acid in methanol). The coefficient of correlation was at least >0.9998.
The LOQ for DRSP was set to 2 μL due to study requirements. For method validation, replicates of dechlorinated tap water spiked with 2 and 20 μL (10·LOQ) DRSP were analyzed. Mean recoveries of 94 ± 9% for LOQ level and 108 ± 3% for 10·LOQ level were obtained. Two blank replicates of dechlorinated tap water were analyzed analogously. The mean recoveries of the fortification samples were between 70 and 110%.
The primary variables were the total count of eggs over 3 weeks and the number of eggs per clutch. To allow the comparison of the pretreatment and treatment period, which had different lengths, the pretreatment values of the numbers of eggs and the numbers of clutches were normalized to the length of the treatment period for the descriptive statistics.
For the analyses of the total count of eggs, generalized estimating equations with a negative binomial link were used. An appropriate offset was used to allow for difference in the length of the pretreatment and the treatment period.
Estimates for differences of the number of eggs per clutch were obtained through a general mixed model using separate variance components for each treatment group in order to deal with the heteroscedasticity. Consequently, Satterthwaite degrees of freedom were used in the calculation of the tests and confidence intervals. The respective pretreatment value was included as a covariate in the model. No adjustments for multiplicity were performed.
Analytical results for levonorgestrel
Measured concentrations differed from the nominal. At the lowest concentration, 40% of the nominal concentration was found (0.8 ng/L); the mid and high concentration solutions showed average recovery rates of 17% (3.3 ng/L) and 30% (29.6 ng/L). Therefore, the measured concentrations were used for interpretation of the study results. In all exposure groups, the recovery rate increased from the first sampling until the end of the study. All stock solutions showed an adequate recovery rate and were stable during exposure time.
Analytical results for drospirenone
Measured concentrations differed from the nominal. The mid and high concentration solutions showed average recovery rates of 33% (6.5 μL) and 35% (70 μL); the low concentration (2 μL) was below the limit of quantification. Therefore, the low concentration was extrapolated at 0.66 μL according to a recovery rate of 33%. Accordingly, the measured and extrapolated concentrations were used for interpretation of the study results. All stock solutions showed an adequate recovery rate and were stable during the exposure time.
Reproductive success for levonorgestrel
The presented results are supplemented by the earlier RFT. In this RFT with 1, 10, and 100 ng/L (nominal), even the lowest concentration caused an inhibition of egg deposition toward the end of the exposure phase. The definitive study confirmed the results of the RFT at similar concentrations (0.8, 3.3, and 29.6 ng/L). The eggs per pair and days under LNG treatment are given in Figure 1A. Figure 1B shows that at the concentrations of 0.8 and 3.33 ng/L a reduced number of eggs were deposited in the first week and ceased almost completely from after week 2. A significant decrease in total egg deposition compared with that of the control group was observed for all treatment groups (p ≤ 0.0001; Fig. 1A). The treatment resulted in behavioral changes in male fishes, such as lack of interest for their spawning tiles as well as aggressive behavior. Furthermore, all treated females showed dose-dependent morphologic changes, such as an increase of abdominal girth above 0.8 ng/L and stronger pigmentation of fins and skin at 29.6 ng/L. In the highest RFT concentration of 100 ng/L, a strong masculinization of the secondary sex characteristics of each of the four treated female fish was observed (Fig. 2), demonstrated by darker pigmentation and de novo synthesis of nuptial tubercles, a typical male secondary sex characteristic.
The results indicate that the compound affects reproductive functions, such as inhibition of egg laying after 1 week of exposure. Moreover, masculinization of female secondary sex characteristics was observed from 29.6 ng/L onwards.
Gonad histology for levonorgestrel
Figure 3 shows the gonads from the definitive study. The evaluation indicates an androgenic effect on male and female reproductive tissue. In comparison with the control, at the high concentration group an increase of the frequency of atretic follicles from 0.7 to 6.1% and of late vitellogenic oocytes from 58.0 to 29.8 was observed. The percentage of oogonia decreased by half (Table 1). The ovary is overfilled with mature oocytes with an increased span of the abdomen (Fig. 3B).
Males showed an increase of mature spermatids and testes size (Fig. 3D). At the highest concentration, tubuli appeared enlarged, and the lumen was filled with mature spermatids. The germinal epithelium was thickened, and therefore the interstitial tissue appeared compressed. Moreover, the pressure seemed to cause atrophy of interstitial Leydig cells and Sertoli cells. The average amount of spermatocysts per field of view decreased from 98 in the control to 58.5 at the highest concentration. There was no obvious shift in the relative percentage of the different stages from the control to the highest concentration. However, the number of examined slides per group was too low for statistical analysis.
Reproductive success for drospirenone
In contrast to LNG, a decrease in reproductive success under DRSP treatment was observable at significantly higher concentrations (Fig. 4A). The lowest tested concentration (0.66 μL) showed no inhibitory effect on fish reproduction. The applied concentrations of 6.5 and 70 μL (mid and high concentrations) resulted in a complete inhibition of egg deposition from week 2 of the exposure phase onward (Fig. 4B). Additionally, there was a lack of breeding behavior of male fish, and aggressiveness was observed. Female fishes seemed to occasionally develop a darker pigmentation of skin and fins and an increased size of waist. The solubilizing agent triethylene glycol did not have any visible effect on fitness or reproduction of the control animals because egg numbers were in the range typical for this fish species .
For the egg deposition at the mid and high concentrations, there was a significant proportional change (p < 0.0001) (Fig. 4A). The comparisons to vehicle were also statistically significant for the mid (p = 0.0005) and high concentrations (p = 0.0001).
Gonad histology for drospirenone
The females from the mid and high concentrations showed enlarged ovaries with a high percentage of degenerating vitellogenic oocytes. In addition, a high resorption activity (macrophages) and several necrotic and apoptotic processes such as the degeneration of the follicular wall were observed. In comparison with the vehicle control, the histomorphometric evaluation revealed an increase of the frequency of degenerating atretic follicles (preovulatory atretic follicles) from 3.0 to 52.0% and a decreasing amount of primary growth oocytes from 34.2 to 17.8% and of mature vitellogenic stage oocytes from 26.0 to 2.2% (Table 2).
Table Table 1.. Ovaries of control and high treatment groups; percentages of different maturational stages (%). The low and mid concentrations of levonorgestrel treatment were not examined
The testes histology at the mid and high concentrations revealed an increase in the number of mature spermatids and testes size (Fig. 5). At the highest concentration, tubuli appeared enlarged, and the lumen was filled with mature spermatids. The germinal epithelium was thickened, and therefore the interstitial tissue appeared compressed. Moreover, the pressure seemed to cause atrophy of interstitial Leydig cells and Sertoli cells. The average amount of counted spermatocysts per field of view decreased from 95.8 in the control to 51 cysts in the high concentration group. The relative percentage of the different stages shifted from control to the high concentration group. The number of spermatogonia cysts counted decreased from the control to the highest concentration from 20.8 to 16%, while the number of mature spermatid cysts was increasing from 22.5 to 39.1%.
The variations between nominal and measured concentrations (average recovery rates for both steroids of approximately 30%) of the test solutions are most probably caused by adsorption of the steroidal compounds to the surfaces of the flow-through test system. By the nature of the test design, a prolonged saturation phase for the test system is not possible, because the fish are required to remain undisturbed in their aquaria after the preexposure phase. Due to the fact that all stock solutions showed a satisfactory recovery rate and the pumps were regularly checked, a dosage problem can be excluded.
To assess the effects of the two marketed progestins, LNG and DRSP, we performed 21-d reproductive tests in fish. For LNG, the reproductive success was significantly different from the tap water control group at all concentrations. Due to the high variability in reproductive output and the limitations in chemical analysis, it was not possible to determine a no-observed-effect concentration (NOEC).
Table Table 2.. Ovaries of control and drospirenone-treated females; percentages of different maturational stages (%)
For DRSP at the microgram per liter range, there was also a significant decrease in the reproductive success at the mid and high concentrations, whereas the lowest concentration had no inhibitory effect. Therefore, the NOEC was determined with 2 μL (nominal concentration). Because the low concentration could not be quantified, an interpolation can be based on the recovery rates of the mid and high concentrations that were approximately 33%. Assuming a similar recovery rate of the low concentration, the actual concentration of the NOEC would be approximately 0.66 μL. The effects on fertility in the mid and the high concentration groups were also reflected at the tissue level by histologic changes in the ovary and testes.
In addition to the observed effects on reproductive success, LNG-treated female fish showed masculinization, expressed in de novo development of the nuptial tubercles at the nominal test concentration of 100 ng/L in the RFT. In the definitive test, females were also masculinized at the highest concentration (29.6 ng/L), expressed in dark coloration of skin and fins, but not as strong as in the RFT. Nuptial tubercles are a morphological characteristic for reproductively active male fathead minnows. The exposure of these fish to androgens can result in de novo synthesis of nuptial tubercles in females [18,19]. Therefore, the occurrences of the tubercles in females treated with LNG is suggestive of an androgenic partial effect of this hormone. Levonorgestrel as a nortestosterone-derived progestin is reported to bind the androgen receptor with 58% of the affinity of the natural ligand . Drospirenone exhibits an antiandrogenic activity (one-third of that of cyproterone acetate, the most potent antiandrogenic progestin) by preventing the androgen receptor from androgen binding [4,10,20]. Therefore, as expected, no masculinization of female fish occurred even under the highest DRSP treatment.
The pre-exposure fecundity in both studies (14.7/20.7 eggs/reproductive day) seems to be at the lower end of normal reproductive activity of 50 to 200 eggs every 3 to 4 d in fathead minnows (14.6–58.3 eggs/reproductive day [18,19]). Due to the high biological variations in reproduction of permanent spawners, an experimental setup with a high number of replicates may derive statistically more conclusive results.
It has been shown that in carp, as in most of the teleost fish, 17β-estradiol is the mediator of oocyte growth, whereas the progestin 17α,20β-DP acts as a maturation-inducing hormone via stimulating the germinal vesicle breakdown that is part of the preovulatory maturation process . Similar to the endocrine regulation in mammals, the release of the stimulating gonadotropins from the pituitary involves a complex interplay of internal and external stimuli that can be disturbed by chemical contamination of the aquatic environment . These described physiological processes can also result in morphological changes of gonadal tissue in fish.
In the present studies, follicle atresia was observed by destabilization of the vitelline envelope and degeneration of the nucleus and the yolk material in LNG- and DRSP-exposed female fish. Such preovulatory atretic follicles are usually resorbed by macrophages, which phagocytose the developed oocytes . In control females, a complete range of oocytes including developing and mature stages (oogonia, previtellogenic oocytes, and vitellogenic oocytes) was present. The percentage of atretic follicles in the control females of both studies was low (0–3.5%). Depending on the phase of the spawning cycle, the relative presence of various stages can be variable . In fractional spawners, atretic follicles of a percentage up to 12% may be recognized as a normal process during long spawning periods .
Atresia can also be caused by environmental stress  or by lacking a spawning opportunity. Before starting the egg deposition, females need to be stimulated by males occupying the spawning tile. On that account, male aggressiveness and missing interest in the spawning tile could additionally be responsible for the absent breeding behavior. Massive follicle atresia is also reported by Van der Ven et al.  to be related to high concentrations of 17β-estradiol by inhibition of vitellogenesis through inhibition of gonadotropin production (negative feedback). This negative feedback at the pituitary may result in inhibition of gonadotropin II release, which is the mediator of ovulation (similar to mammalian luteinizing hormone) . In this regard, the measurement of plasma vitellogenin concentrations might be useful.
In testes, the observed changes indicate a proliferating effect with a retention of spermiation for both progestins, however, at a different concentration range.
As described by Nagahama and Yaron [6,25], the steroidogenic pathway in male teleosts shifts at the beginning of spermiation from androgen to 17α,20β-DP, which is necessary for the final maturation and motility of sperm. The production of 17α,20β-DP in sperm needs to be stimulated by a precursor steroid (17α-hydroxyprogesterone) via gonadotropin from the pituitary-gonadal axis.
The 17α,20β-DP is also needed for the acquisition of sperm motility by increasing the pH of the sperm duct. The exogenously administered progestin obviously caused the above described hormonal shift in an even exaggerated way, thus increasing the number of mature spermatids and the abundance of germinal tissue. Even though sperm motility was not an end point in the present study, the histological findings in the progestin-exposed male reproductive tissue indicate an impairment of motility because the mature spermatids were not released and there was no spawning behavior. This may be due to a negative feedback with lacking gonadotropin II release signal, similar to the discussed effects in females. Under the influence of an excess of progesterone, the maturation of spermatocysts would proceed, but the sperm cells would remain immotile, and the onset of spermiation would be inhibited due to the lacking gonadotropin II signal.
In comparison, DRSP seems to be a thousandfold less potent in affecting fish reproduction than LNG. There are no pronounced differences in the lipophilic properties (log POWn-octanol/water partition coefficient; LNG, 3.55; DRSP, 3.08) of both progestins. As described by Elger et al. , the relative binding affinity of LNG to the human uterine progesterone receptor is higher than that of DRSP (compared with promegestone; progesterone, 30%; DRSP, 20%; LNG, 250%). Therefore, different efficacies in inhibition of ovulation in the rat were found (50% effective dose; LNG, 0.01–0.03 mg; DRSP, 0.3–1.0 mg/animal, administered subcutaneously) .
The uptake of steroid compounds in fish gills is mediated by a plasma sex steroid binding globulin (SHBG) that is produced in the liver. In zebrafish, SHGB is shown to have a high affinity to several natural and synthetic steroids with a preference for testosterone and estradiol. In cases of 17α-ethinylestradiol and LNG, the binding affinities of the synthetic steroids exceed those of the natural ligands. The relative binding affinity of zebrafish SHBG to LNG is reported to be 52% of that of 5α-dihydrotestosterone, whereas the affinity to 11-ketotestosterone, the active androgen in zebrafish, is lower (28%) . Thus, the transport of LNG is comparable or even better than that of other male steroids, which may explain the high sensitivity of fathead minnow with regard to the androgenic partial effect of LNG.
In conclusion, the results of the presented studies with both progestins, LNG and DRSP, indicate a similar effect on the inhibition of the reproductive function of fathead minnow, though at markedly different concentration ranges. For LNG, no NOEC could be determined. Additionally, from a concentration of 3.3 ng/L LNG, clear masculinization was appearing most probably due to the androgenic partial effect of this hormone. On the basis of a theoretical estimation (according to the European Medicines Agency ), the environmental concentrations of LNG can reach a level of 2 ng/L. Consequently, the results showed that effects on fish caused by LNG may be of environmental relevance.
For DRSP, a NOEC was determined at a nominal concentration of 2 μL (0.66 μL). This value is above the predicted environmental concentrations, which are expected to be in the low nanogram per liter range.
Currently, a full life cycle test with LNG and further research on the 21-d reproduction assay is ongoing to further assess the effects of LNG also in comparison with other progestins.
The authors thank their colleagues of the ecotoxicology and pathology laboratories of Bayer Schering Pharma AG for technical support and Anna-Lena Frisk and Jakob Walter for supporting the histopathologic examination. This research is part of a PhD thesis funded by Bayer Schering Pharma AG.