Ecotoxicology, 9, 383᎐397, 2000 ᮊ 2000 Kluwer Academic Publishers. Manufactured in The Netherlands.

Effects of Endocrine Disruptors on Prosobranch Snails (): in the Laboratory. Part I: Bisphenol A and Octylphenol as Xeno-Estrogens

JORG¨ OEHLMANN,* ULRIKE SCHULTE-OEHLMANN, MICHAELA TILLMANN AND BERND MARKERT Internationales Hochschulinstitut Zittau, Lehrstuhl Umwelt¨erfahrenstechnik, Fachgruppe Human- und Okotoxikologie,¨ Markt 23, D-02763 Zittau, Germany

Abstract. The effects of suspected endocrine disrupting chemicals on freshwater and marine proso- branch species were analysed in laboratory experiments. In this first publication, the responses of the and of the marine prosobranch Nucella lapillus to the xeno-estro- genic model compounds bisphenol AŽ. BPA and octylphenol Ž. OP are presented at nominal concentra- tion ranges between 1 and 100 ␮grL. Marisa was exposed during 5 months using adult specimens and in a complete life-cycle test for 12 months. In both experiments, the xeno-estrogens induced a complex syndrome of alterations in female Marisa referred to as ‘‘superfemales’’ at the lowest concentrations. Affected specimens were characterised by the formation of additional female organs, an enlargement of the accessory pallial sex glands, gross malformations of the pallial oviduct section resulting in an increased female mortality, and a massive stimulation of oocyte and spawning mass production. The effects of BPA and OP were comparable at the same nominal concentrations. An exposure to OP resulted in inverted U-type concentration response relationships for egg and spawning mass production. Adult Nucella from the field were tested for three months in the laboratory. As in Marisa, superfemales with enlarged accessory pallial sex glands and an enhancement of oocyte production were observed. No oviduct malformations were found probably due to species differences in the gross anatomical structure of the pallial oviduct. A lower percentage of exposed specimens had ripe sperm stored in their vesicula seminalis and additionally male Nucella exhibited a reduced length of penis and prostate gland when compared to the control. Because statistically significant effects were observed at the lowest nominal test concentrationsŽ. 1 ␮g BPA or OPrL , it can be assumed that even lower concentrations may have a negative impact on the snails. The results show that prosobranchs are sensitive to endocrine disruption at environmentally relevant concentrations and that especially M. cornuarietis is a promising candidate for a future organismic invertebrate model to identify endocrine-mimetic test compounds.

Keywords: endocrine disruptors; xeno-estrogens; bisphenol A; octylphenol; snails

*To whom correspondence should be addressed: IHI Zittau, Markt 23, D-02763 Zittau, Germany. E-mail: oehlmann@ihi- zittau.de. 384 J. Oehlmann et al.

Introduction as endocrine modulators on freshwater and ma- rine prosobranch species in the laboratory. Most Recent reports have shown that a number of of the results were obtained during a research xenobiotics in the environment are capable of project for the German Federal Environmental interfering with the normal endocrine function in AgencyŽ. project code 216 02 001r04 between a variety of and also in humans. Some of September 1997 and May 2000. The objective was the reported effects of suspected endocrine dis- to develop an organismic invertebrate test system ruptors in humans include decreased sperm for the simultaneous identification of either an- counts, increased cases of breast, testicular and drogen- or estrogen-mimicking chemicals. Re- other forms of reproductive cancers, genital ab- cently, gonochoristic prosobranchs were rated as normalitiesŽ. e.g. hypospadia, cryptorchidism , pre- the most promising candidates for this purpose mature puberty in females, and increased cases of next to insects and crustaceansŽ deFur et al., endometriosisŽ. Gist, 1998 . The overwhelming 1999. . This first publication is focussed on the majority of the studies on the effects of hormone- effects of two suspected xeno-estrogens, bisphe- mimetic industrial chemicals were focussed on nol AŽ. BPA and octylphenol Ž. OP . The next findings in vertebrates. More detailed informa- papers will be dedicated to the xeno-androgen tion about the effects on and mechanisms of triphenyltin and the xeno-antiandrogen vinclo- action in invertebrates has only been obtained zolin. from a few cases although invertebrates represent The first evidence that bisphenol A and more than 95% of the known species in the alkylphenols could be estrogenic was published in kingdomŽ. deFur et al., 1999 . The limited the 1930s on the basis of feeding experiments number of examples for endocrine disruption in with BPA and 4-propylphenol to ovariectomised invertebrates is partially due to the fact that their ratsŽ. Dodds and Lawson, 1936, 1938 . More re- hormonal systems are rather poorly understood in cent research has highlighted the implications of comparison with vertebrates. Deleterious en- these effects. The growth of human breast cancer docrine changes following an exposure to certain cellŽ. MCF-7 cultures is affected by octylphenol compounds may therefore easily be missed or at concentrations as low as 0.1 ␮M20Ž.␮grL simply be unmeasurable at present, even though a Ž. number of field investigations and laboratory e.g. Soto et al., 1991 . Estrogenic effects have studies show that endocrine disruption has proba- also been shown in tissue and cell culture experi- bly occurredŽ. for review: deFur et al., 1999 . The ments with rainbow trout hepatocytes, chicken example of tributyltinŽ. TBT compounds and their embryo fibroblasts and in mammalian estrogen masculinising effects in about 150 species of receptor assaysŽ Jobling and Sumpter, 1993; White prosobranch molluscs shows that apparently triv- et al., 1994; Yamakoshi et al., 2000. . ial biochemical changesŽ inhibition of aromatase Bisphenol A is manufactured for the plastics activity according to Bettin et al.Ž. 1996 , can have industry as an intermediate in the production of drastic effects up to the population and commu- polycarbonate and epoxy resins. A smaller amount nity levels by a final sterilisation of affected fe- is used in the manufacture of thermopaper, tyres, males. According to Matthiessen and GibbsŽ. 1998 dental composites and sealantsŽ. BUA, 1997 . The there is no reason to suppose that such far-re- total BPA consumption in Germany was approxi- aching changes are in any sense unique. The main mately 163,000 t in 1994. In 1993, the correspond- endocrine effects of TBT in these molluscs are ing amounts for Western Europe were 347,000 t, the induction of imposex, an additional formation for the USA 552,000 t and Japan 214,000 tŽ BUA, of male sex characters like penis andror vas 1997. . deferens on femalesŽ Gibbs et al., 1987; BPA is not readily biodegradableŽ less than Oehlmann, 1994. and of intersex which is charac- 1% transformation in 28 days in sewage treat- terised by a modification or supplanting of female ment plants according to Howard, 1989. although by male sex organsŽ. Bauer et al., 1995, 1997 . with sufficiently adapted microorganisms, it can This publication is the first in a series which be eliminated to more than 90% in the laboratory investigates effects of compounds suspected to act and in industrial sewage treatment plants. Under Endocrine Effects on Prosobranchs. I: Xeno-Estrogens 385 environmental conditions neither hydrolysis nor of tens of ␮grL of a wide range of alkylphenolic photolysis is likelyŽ. BUA, 1997 . compoundsŽ. Ahel et al., 1994b . Low levels were There is evidence from estrogen receptor reported for drinking water in the USA, with a Ž.Greim, 1998 and MCF-7 assays Ž Krishnan et al., total concentration of alkylphenols of almost 1993. that BPA exhibits an estrogen-mimetic ac- 1 ␮grL and a concentration of OP2EO of 32 tion at concentrations as low as 2᎐5 ␮grL. The ngrLŽ. Clark et al., 1992 . results for in vivo studies with mammals are con- flictingŽ. see Discussion . Materials and methods The occurrence of BPA in aquatic ecosystems is summarised by RippenŽ. 1999 . In effluents from The experiments were performed with two dif- sewage treatment plants in BerlinŽ. Germany ferent gonochoristic prosobranch species, the concentrations of up to 160 ngrL were detected freshwater ramshorn snail, Marisa cornuarietis in 1997Ž mean: 80 ngrL; median: 60 ngrL; ns Ž.Mesogastropoda: , and the ma- 12. . The highest concentrations in rivers were rine dogwhelk Nucella lapillus ŽNeogastropoda: reported from JapanŽ. Tokyo region with values Muricidae. . Marisa specimens came from our own between 10 and 1,900 ngrL. In smaller rivers laboratory breeding stock which was built up with near Berlin concentrations of up to 410 ngrL specimens obtained from the breeding stock of were reported for 1997Ž mean: 23 ngrL; median: Aquazoo DusseldorfŽ. Germany in 1991. Dog- 6ngrL; ns41. . In the same river system sedi- ¨ whelks came directly from the field and were ments were contaminated with - 5᎐150 ␮g collected at Mean Melen, Brittany, in March 1999. BPArkgŽ.Ž dry wt. mean and median: 42 ␮grkg; ´´ For all laboratory experiments, a 24 h ns19.Ž. . In the Rhine estuary The Netherlands Ž.weekends: 48 h semi-static renewal system in 60 concentrations were between -10 and 119 ngrL litre glass aquaria filled with tap waterŽ. for Marisa in 1989. The measured bioconcentration factors or artificial seawaterŽ. for Nucella; salinity 35‰ for BPA in carp Ž.-100 , calculated values Ž BCF and provided with an Eheim power filter was up to 366. and the log P value of 3.32᎐3.4 ow used. The tests were performed under constant suggest a low bioaccumulation potential in aquatic conditions with a temperature of 22"1ЊC for organismsŽ. BUA, 1997 . freshwater and 14"1ЊC for marine snails; the NonylphenolŽ. NP and octylphenol are the most light dark cycle was adjusted to 12:12 h. important high production volume alkylphenols Three different series of exposure experiments with nonylphenol ethoxylatesŽ. NPnEO taking ap- were conducted with the test compounds bisphe- proximately 80% of the world market, and nol AŽ. BPA, Merck Schuchardt, Germany and octylphenol ethoxylatesŽ. OPnEO representing octylphenolŽ. OP, Merck Schuchardt, Germany : the remaining 20%Ž. White et al., 1994 . The alkylphenol ethoxylates are used in a variety of 1. Marisa P() parental generation -test: Adult industrial processes, including wool washing, but Marisa cornuarietis of comparable age were no longer in domestic detergents in the EU. These exposed to nominal concentrations of 1, 5, 25, compounds are biodegraded by removal of ethoxy and 100 ␮g BPA or OPrL for 5 months, groups, producing less biodegradable products, including a solvent controlŽ. ethanol . Thirty including alkylphenols like nonyl- and octylphenol specimens from each group were collected for which frequently persist through sewage treat- analysis at the beginning of the experiment ment and in riversŽ. Ahel et al., 1994a,b . The and at monthly intervals. estimated annual consumption of alkylphenol 2. Marisa LC() life-cycle -test: The spawning ethoxylates was 13,500 t in Germany in 1986 masses with eggs produced by the adult Ž.BUA, 1991 and 18,500 t in the UK in 1992 ramshorn snails in the solvent control and in Ž.CES, 1993 . the 1 and 100 ␮g BPA or OPrL groups during The total emission of octylphenol into U.K. the Marisa P-test were further exposed to these waters was estimated to be about 300 kg in 1998 nominal concentrations over a period of 12

Ž.Environment Agency, 1998 . Rivers in Switzer- months until the hatched F1 specimens were land have been found to contain concentrations one year old. They reached sexual maturity in 386 J. Oehlmann et al.

their 8th month. Thirty specimens from each Standard statistical analyses of the resultsŽ e.g. group were collected for analysis at an age of analyses of covarianceŽ. ANCOVA with multiple 6, 8, and 12 months. Additionally, the hatching comparison of samples according to TukeyŽ low

success of the F1 generation was recorded. n.Ž. or Student-Newman-Keuls high n , H test 3. Nucella test: Adult Nucella lapillus were ex- ŽKruskal-Wallis test with multiple comparison of posed to nominal concentrations of 1, 25, and samples according to Nemenyi. , ␹ 2 test, and Weir 100 ␮g BPA or OPrL for 3 months, including test for classified values. were performed accord- a solvent controlŽ. glacial acetic acid . Thirty ing to WeberŽ. 1972 and Lozan´ Ž. 1992 using the specimens from each group were collected for computer programme StatEasy for Windows NT. analysis at the beginning of the experiment and at monthly intervals. Results and discussion During the experiments mortality and produc- tion of spawning masses with the number of eggs Marisa P() parental generation -test in each of the aquaria were recorded at daily intervals. As the fecundity parameters could not During the first series of laboratory experiments be assessed for individual females but only for the with adult Marisa cornuarietis a complex syn- single experimental groups in a tank with a known drome of morphological and physiological alter- number of females following the analyses of spec- ations occurred, resulting in an enhancement of imens, no measures of variabilityŽ e.g. standard spawning mass and egg production, malforma- deviation or standard error. can be calculated for tions of the female genital system, and probably these endpoints. also in a higher female mortality. We refer to this All specimens were narcotised prior to analysis syndrome as the ‘‘induction of superfemales.’’ Su- Ž2.5% MgCl2 in distilled water for Marisa,7% perfemales are female specimens with additional MgCl2 for Nucella.. The individual shell and female sex organs, e.g. a second vagina with vagi- aperture height were measured to the nearest nal opening to the mantle cavityŽ. Fig. 1 andror 0.1 mm before the shell was cracked and the snail an enlargement of the pallial accessory sex glands was removed. The presence, normal appearance, Ž.albumen and capsule gland, Figs. 2a,b . and extensionŽ. to the nearest 0.1 mm of all sex The normal morphological and histological organs was checked as well as the occurrence of structure of the male and female genital system oocytes and sperm in the genital system and of of the ramshorn snail and of pathomorphological visible excrescences on genital and other organs changes during imposex development following a with a dissection microscope. Additionally, the TBT exposure is documented by Schulte-Oehl- VDSIŽ vas deferens sequence indexsmean value mann et al.Ž. 1994, 1995 . Superfemales were not of imposex stages in a sample with values of 0 to observed before in our laboratory although more 3in Marisa cornuarietis and 0 to 6 in Nucella than 8,000 specimens have been analysed since lapillus. as a measure of the imposex intensity in 1992. Additional female sex organs, like in the a sample was calculatedŽ for details see Oehlmann ramshorn snail specimen documented in Fig. 1, et al., 1991; Schulte-Oehlmann et al., 1995. . occurred only exceptionally as individual cases For histopathological analyses of the gonads during the experiments, but the albumen and during the Marisa P-test, 6 male and 6 female capsule glands were enlarged in all females in the specimens from each sample were fixed in BPA and OP experimental groups irrespective of Carnoy’s and Bouin’s fluid, respectively, and then the concentration of the xeno-estrogens. The di- preserved in ethanol. After embedding in para- rect comparison of the pallial gland complex of a plast, serial sectionsŽ. 5᎐7 ␮m were made and typical female from the control groupŽ. Fig. 2a stained with haemalun-chromotrope. The sections with a specimen in a xeno-estrogen treated group were analysed using an image analysis system Ž.Fig. 2b shows that primarily the volume of the Ž.Optimas 5.2, Optimas Cooperation coupled with capsule and albumen gland increased but the an Olympus microscopeŽ. BX 50 . maximum extension or length of these organs, Endocrine Effects on Prosobranchs. I: Xeno-Estrogens 387

Figure 1. Marisa cornuarietis. Photograph of a superfemale with an additional vagina including female opening. Abbreviations: an, anus; c, capsule gland; g, gill; v12 , original vagina in normal position; v , additional vagina on other side of the rectum. which was measured during the experiments, was pared to the control during the experiments, more or less unaffectedŽ no statistically signifi- counted by successful egg laying. This is not only cant differences in extension; H test, p)0.05. . true if the cumulative numbers were recorded This hypertrophy of the two female glands was during the experimentsŽ. Fig. 4 but also when the not the ultimate effect of the superfemale syn- corresponding values per female were analysed drome. In a number of the BPA- and OP-treated Ž.Fig. 5 . An ANCOVA analysis with multiple females a rupture of the pallial oviduct occurred comparison of samples according to Student- as shown in Fig. 2c, irrespective of the applied Newman-Keuls reveals that the spawning mass concentration. The incidence of this phenomenon and egg production in all BPA- and OP-treated was 3.73% three months after start of the experi- groups was significantly higher than in the control ment. The opening was typically found at the groupŽ. p-0.01 . For the OP experiment, an in- transition zone between the albumen and capsule verted U-type concentration response relation- gland and in some of these superfemales a pro- ship exists with the highest concentrationŽ 100 ␮g trusion of spawning massŽ egg capsules filled with OPrL. evoking comparable effects like the sec- eggs and intracapsular fluid. was found at the ond lowestŽ.Ž 5 ␮gOPrL no significant differ- ruptureŽ. Fig. 2d . ences in Student-Newman-Keuls test. . The lower Additionally, in all experimental groups ex- effectiveness of the highest concentration of 100 posed to BPA and OP some females with abortive ␮grL compared to the second highestŽ. 25 ␮grL capsule masses in the lumen of the albumen was much less pronounced in the BPA test series gland were observedŽ. Fig. 3 . These capsules ac- giving only little evidence for a comparable in- cumulated in the tubular part of the gland which verted U-type concentration response relation- continues in the capsule gland. Therefore, the ship as for OP. abortive mass was found in the same region where The detailed investigations of TBT effects the rupture of the oviduct occurs. demonstrated in a number of muricid gastropod The protrusion of spawning masses and the speciesŽ. e.g. Nucella lapillus, Ocinebrina aciculata accumulation of abortive egg capsules in the a comparable rupture of the pallial oviduct as oviduct gave an indication for the underlying observed in BPA and OP exposed ramshorn snails causes for the enlargement of the female pallial due to a blockade of the vaginal opening by glands and the rupture of the oviducts. It was the proliferating vas deferens tissue and a resulting enhancement of the spawning mass and egg pro- accumulation of abortive capsule masses. In such duction in all BPA- and OP-treated groups com- affected populations an increased female mortal- 388 J. Oehlmann et al.

Figure 2. Marisa cornuarietis. Photographs of a control femaleŽ. a and of BPA or OP treated superfemales Ž b᎐c . with opened mantle cavities. InŽ. b the enlargement of the capsule and albumen gland is indicated by the broken line in comparison to the control female inŽ. a . In Ž. c a rupture in the wall of the pallial oviduct occurs Ž arrow . and in Ž. d additionally a protrusion of the spawning mass is visible at the ruptureŽ. arrow . Abbreviations: a, albumen gland; an, anus; c, capsule gland; g, gill; v, vagina. Endocrine Effects on Prosobranchs. I: Xeno-Estrogens 389

Marisa LC() life-cycle -test

The hatching success of the F1 generation during the Marisa LC-test was not affected by the two test compounds. In general, BPA and OP pro-

duced comparable effects in the F1 generation as already demonstrated for the P-test above. Only slight differences regarding a higher incidence of sterilised superfemales due to a rupture of the oviduct and a less striking enhancement of spawn- ing mass and egg production at the highest con- centration were observedŽ. Fig. 7 . The first specimens with oviduct ruptures were already found when the first females reached sexual maturity at an age of 6 months in the two Figure 3. Marisa cornuarietis. Photographs of an albumen ␮ r gland from a control animalŽ. above and a BPA treated tested OP concentrationsŽ. 1 and 100 g L . The femaleŽ. below with an abortive capsule mass opened by a incidence was 5.0% for the lower and 9.1% for longitudinal section. Abbreviations: ac, abortive capsule mass; the higher concentrationŽ mean for both OP el, extension of albumen gland lumen; gp, glandular pouches treatments: 6.5%; mean for BPA: 0%. . At an age of the albumen gland; t, tubular part of the albumen gland. of eight months the percentage of females with these oviduct malformations increased to 15.4% in the 100 ␮gOPrL groupŽ mean: 7.7% for OP and 0% for BPA. . At the end of the experiment, ity and a consequent shift of the sex ratio in when all females were sexually mature, superfe- favour of males was observedŽ Gibbs et al., 1987; males with oviduct ruptures were also found for Oehlmann et al., 1996. . Therefore, it can be as- the first time in the two BPA exposure groups sumed that not all Marisa superfemales with an during the Marisa LC-test. The incidences are oviduct rupture were found at the monthly analy- 5.2% for the lower and 11.8% for the higher BPA ses of the samples with thirty specimens per group Ž.mean: 8.3% , 7.4% for the lower and 10.0% for as it is likely that these females will not survive the higher OP concentrationŽ. mean: 8.1% . Al- for long after the rupture occurred. The mortality though more sterilised superfemales than in the data presented for the Marisa P-test in Fig. 6 P-test were found during the Marisa LC-test and show clearly a statistically significant increase in these specimens did only occur in the xeno- all BPA and OP exposed groups of snails com- estrogen treated groups in both experiment se- pared to the control. Because the two test com- ries, the differences were not statistically signifi- pounds exhibit in general a very low acute toxicity cant Ž ␹ 2 test; p)0.05. . As during the P-test, the Ž.BUA, 1991, 1997 , the elevated mortality seems mortality was also statistically significantly higher not to be a direct effect of BPA and OP, but a in all xeno-estrogen exposed cohorts of the LC- result of the induction of superfemales with a testŽ. range: 14.8᎐21.5% when compared to the rupture in the oviduct. It was not possible to sex controlŽ. 8.6% between month 6 and the end of the dead specimens due to their rapid decay but the experiment Ž ␹ 2 test; p-0.05. . in all experimental groups, a shift of the sex ratio As already stated, the enhancement of the in favour of males was detectable when compared spawning mass and egg production was less to the control although these differences were not marked for the F1 generation in the highest con- statistically significant Ž ␹ 2 test, p)0.05. . centration of BPA and OP although statistically The histopathological analyses of the gonads significant when compared to the control of both sexes during the Marisa P-test gave no ŽANCOVA analysis with multiple comparison of indication that spermiogenesis or oogenesis were samples according to Tukey; p-0.05. . The differ- affected by either BPA or OP in the applied ences between the two xeno-estrogen concentra- concentration range. tions were not significant. The reduced effect of 390 J. Oehlmann et al.

Figure 4. Marisa cornuarietis. Cumulative numbers of spawning massesŽ. a, c and eggs Ž. b, d produced by all females of the single experimental groups exposed to BPAŽ. a, b and OP Ž. c, d during the Marisa P-test. Exposure groups: Ž.ؒ solvent control, Ž.ⅷ 1 ␮grL, Ž.B 5 ␮grL, Ž.' 25 ␮grL, Ž.q 100 ␮grL.

␮ r 100 g BPA or OP L in the F1 can be inter- mentioned up-regulation of androgen receptors preted as a loss of sensitivity which might be due as a result of a long term exposure to high xeno- to a down-regulation of the estrogen receptors. It estrogen concentrations offers a possible explana- has been described for a number of vertebrate tion for the enhanced VDSI values. The un- species that a long term application of estrogen changed endogenous testosterone titres are agonists resulted in a down-regulation of estrogen supposed to have this effect as the androgen can receptorsŽ. reduced numbers andror sensitivity bind to a higher number or more sensitive andro- and simultaneously in an up-regulation of andro- gen receptors in the tissue. gen receptorsŽ. higher numbers andror sensitivity Ž.Marquardt and Schafer,¨ 1994 . Nucella test At the end of the Marisa LC-test, when all females were sexually mature, the imposex inten- Also during the three months experiment series sities increased in the 100 ␮grL exposure groups with the dogwhelk Nucella lapillus an induction of of BPA and OPŽ. Fig. 8 . This virilisation was only superfemales was observed with an increase of significant for BPAŽ Weir test for classified val- oocyte productionŽ. Fig. 9 and a hypertrophy ues, p-0.001. and was not observed during the of the pallial female sex glandsŽ. Fig. 10 . In con- P-test with adult ramshorn snails. The above trast to Marisa, no gross malformations of the Endocrine Effects on Prosobranchs. I: Xeno-Estrogens 391

Figure 6. Marisa cornuarietis. MortalityŽ in relative and abso- lute values. in the experimental groups during the Marisa P-test. Asterisks indicate statistical significant differences to control Ž ␹ 2 test. : ଙ,p-0.05; ଙଙ,p-0.01; ଙଙଙ,p-0.005.

normally not produce egg capsules when the spec- imens are transferred from the field into the laboratory. Under these constant conditions, the produced oocytes are digested in the ingestion gland. In the field, the raise of oocyte production demonstrated in Fig. 9 will result in a consequent stimulation of egg capsule formation. If capsule production in dogwhelks is comparably excessive increased as in Marisa by xeno-estrogens, the Figure 5. Marisa cornuarietis. Comparison of spawning mass sexual products might congest in the distal part of Ž.a and egg numbers per female Ž. b relative to the control group produced during the Marisa P-test. As both parameters the capsule gland leading to comparable individ- could not be assessed for individual females but only for the ual and population effects as in the final stages of single experimental groupsŽ. with a known number of females imposex development when females are sterilised the bars are calculated on the basis of mean values. Conse- due to a blockade of the oviduct by proliferating quently, it is not possible to present a measure of variability vas deferens tissue: accumulation of abortive cap- within a group. sule masses, distension and rupture of the pallial oviduct, and consequently an increased female oviduct occurred like the slit-like ruptures at the mortality. anterior region of the albumen gland as described The accessory female pallial sex glands in Nu- above and consequently also no female sterilisa- cella lapillus Žalbumen, ingestion and capsule tion nor a higher overall or female mortality. The gland. were enlarged in all BPA and OP groups main reason for this is the morphological differ- compared to the control already after the first ence in the structure of the pallial oviduct section and second month of exposure although these in the ramshorn snail and in dogwhelks which do differences were not statistically significant. As not exhibit a comparable bottleneck passage for already stated for Marisa, primarily the volume of the sexual products in the proximal section of the the capsule and albumen gland appeared to be oviductŽ Oehlmann et al., 1988; Schulte-Oehl- increased but the maximum extension of these mann et al., 1994. . But beyond this it has also to organs, which was exclusively measured at these be taken into account that Nucella lapillus does time points, was more or less unaffected. There- 392 J. Oehlmann et al.

Figure 7. Marisa cornuarietis. Cumulative numbers of spawning massesŽ. a, c and eggs Ž b, d . produced by the experimental groups ⅷ ␮ r ؒ exposed to BPAŽ. a, b and OP Ž. c, d during the Marisa LC-test Ž F1 generation . . Exposure groups: Ž.solvent control, Ž.1 g L, Ž.q 100 ␮grL. fore, at the end of the 3 months experiment not time of year when the female gland complex is only the extension of the three pallial glands was smaller. measuredŽ Fig. 10a with results for the capsule The increase in weight of the female pallial gland. but also the weight of the entire complex gland complex under the influence of xeno- Ž.Fig. 10b with significant differences between all estrogens is a very easily measurable endpoint groups which received the two xeno-estrogens which opens the possibility to use prosobranch and the controlŽ Kruskal-Wallis test with multiple snails in a comparable manner as ovariectomised comparison of samples according to Nemenyi; rodents for a kind of invertebrate uterotrophic p-0.001. . It has to be considered that the test assay. animals were sampled during the breeding season Contrary to the findings for Marisa cornuari- in March and the test was started in April when etis, the males were affected by the two test the female glands still attain almost their maxi- compounds in Nucella lapillus. Already after the mum natural sizeŽ. Oehlmann, 1994 . The effects first month of exposure, the percentage of males of the two test compounds on the pallial oviduct with sperm stored in the vesicula seminalis mass would be probably more obvious at another dropped to 64.7᎐75.0% in the BPA and 63.6᎐ Endocrine Effects on Prosobranchs. I: Xeno-Estrogens 393

Figure 8. Marisa cornuarietis. Imposex intensities, measured as the vas deferens sequence indexŽ. VDSI , during the Marisa

LC-testŽ. F1 generation at an age of 6 Ž.Ž white bars , 8 hatched bars.Ž. and 12 months black bars . The asterisk indicates a statistical significant difference to controlŽ Weir test for classi- fied values, p-0.001. .

88.9% in the OP-treated groups while the corre- sponding value was 100% for the control males. These differences were statistically significant Ž ␹ 2 test; p-0.05. with the exception of the 1 ␮g OPrL groupŽ. 88.9% . At the end of the second and third month, none of the xeno-estrogen ex- posed males had sperm in this storage organ but

Figure 10. Nucella lapillus. Comparison of the length of the capsule glandŽ. a and the weight of the female pallial glands Ž.b at the end of the Nucella test Ž 3 months of exposure . . Mean values and standard deviation are shownŽ sample sizes: control: 21; 1 ␮g BPArL: 15; 25 ␮g BPArL: 19; 100 ␮g BPArL: 20; 1 ␮gOPrL: 16; 25 ␮gOPrL: 21; 100 ␮gOPrL: 22 females. . Asterisks indicate statistical significant differ- ences to controlŽŽ H test Kruskal-Wallis test .. : ଙଙଙ,p- 0.001.

still 18.8%Ž. month 2 or 12.5% Ž. month 3 in the control. This finding is an indication that BPA and OP might advance sexual repose at the end of the breeding seasons. Figure 9. Nucella lapillus. Relative numbers of females with Additionally, the length of the male penis and oocytes in the oviduct during the Nucella test with BPA and of the prostate gland were significantly reduced OP after 2Ž. white bars and 3 months Ž hatched bars . of when compared to the controlŽ ANOVA with exposure. Asterisks indicate statistical significant differences Ž ␹ 2 . ଙଙ - ଙଙଙ - multiple comparison of samples according to Stu- to control test : ,p 0.01, ,p 0.001. The - differences for the first month sample were not statistically dent-Newman-Keuls, p 0.05. . In Figure 11 the different Ž ␹ 2 test, p)0.05. . results after three months are presented but this 394 J. Oehlmann et al.

A comparison of our own results with other reports in the literature is difficult because only a few studies have investigated the hormone- mimetic effects of BPA, OP or related xeno- estrogens in invertebrates so far. The results for the standard invertebrate test organism in aquatic ecotoxicology, the water flea Daphnia magna are conflicting. Zou and FingermanŽ. 1997 found a reduced moulting frequency in their experiments with xeno-estrogens but these findings could not be confirmed by CaspersŽ. 1998 for BPA. He criticises the suitability of this endpoint for the assessment of endocrine-mimetic properties of test compounds as it is at least doubtful whether or not steroids have a functional role in crus- taceansŽ. deFur et al., 1999 . Our findings are supported by the investiga- tions of Andersen et al.Ž. 1999b who revealed an increase of egg production if the copepod Acartia tonsa was exposed to either 23 ␮g 17ß-estradiolrL or 20 ␮g BPArL. In fathead minnows Ž Pimephales promelas. low aqueous concentrations of 0.05 ␮g 4-nonylphenolrL also resulted in an enhance- ment of egg productionŽ. Giesy et al., 2000 . These authors observed a comparable inverted U-type concentration response relationship as it was found for the endpoints spawning mass and egg production in our OP experiments with Marisa. Zou and FingermanŽ. 1999 found a reduced chitobiase activity in the fiddler crab, Uca pugila- Figure 11. Nucella lapillus. Comparison of the penisŽ. a and tor, when exposed to 10 mg OPrL for 7 days. prostate gland lengthŽ. b at the end of the first exposure 3 Although the reduced chitobiase activity might month. Mean values and standard deviation are shownŽ sam- ple sizes: control: 16; 1 ␮g BPArL: 12; 25 ␮g BPArL: 17; 100 result in a slowing of moulting it seems at least ␮g BPArL: 12; 1 ␮gOPrL: 18; 25 ␮gOPrL: 15; 100 ␮g questionable whether this effect is endocrine- OPrL: 11 males. . Asterisks indicate statistical significant dif- mediated. ferences to controlŽ ANOVA with multiple comparison of In contrast to the small number of studies on . ଙ - samples according to Student-Newman-Keuls : ,p 0.05. xeno-estrogen effects on invertebrates, there are numerous publications investigating the re- effect was observable already four and eight weeks sponses of BPA and OP in in-vitro tests or tests after the start of the experiment. During copula- with vertebrates, but partially with conflicting re- tion, the male penis is inserted into the bursa sults. Andersen et al.Ž. 1999a analysed the estro- copulatrix of the female which is positioned in the genicity of 20 chemicals in 8 different short-term distal section of the pallial oviduct. Therefore, a assays. While BPA induced an estrogenic re- minimum extension of the penis is necessary for a sponse in all assays, the results for OP varied successful transfer of sperm. The reduced penis among assays in the single laboratories indicating and prostate gland length as well as possible that additional standardisation of the test proto- advancement of sexual repose might have nega- cols is required. Lutz and KloasŽ. 1999 report a tive implications for the reproductive success of comparable binding affinity of both compounds to males also in the field. the estrogen receptor of the amphibian Xenopus Endocrine Effects on Prosobranchs. I: Xeno-Estrogens 395 lae¨is.If X. lae¨is was exposed to low concentra- effect concentration. values might be still lower. tions of either BPA or OP during larval develop- Under certain circumstances, like in ramshorn ment, a significantly higher number of female snails due to their specific morphological struc- phenotypes occurred in the adults compared to ture of the pallial oviduct, these alterations can controlsŽ. Kloas et al., 1999 . be expected to exhibit serious consequences at vom Saal et al.Ž. 1998 stated that OP and BPA the population level. The occurrence of superfe- reduce the daily sperm production in male off- males with a rupture in their oviduct is a very spring of mice fed these chemicals from gestation striking parallel case to the deleterious effects of day 11 to 17 at doses as low as 2 ␮grkg body tributyltin compounds on muricid gastropods in weight. These results could not be confirmed by many coastal regions of the world. The mechanis- Ashby et al.Ž. 1999 . Additionally, Takao et al. tical studies of imposex and intersex induction in Ž.1999 found that a prepubertal and pubertal oral prosobranchs showed clearly that steroids play an administration of BPA resulted in a decrease of important role in this group of invertebrates and plasma free testosterone levels and alterations in that even some of the basic regulatory mecha- the differentiation of the male reproductive tract. nisms known for vertebrates are presentŽ Bettin Howdeshell et al.Ž. 1999 report an increased post- et al., 1996; Matthiessen and Gibbs, 1998. . We natal body weight gain and an advancement of hope that the results communicated in this paper puberty in female CF-1 mice whose mothers were and in the next publications dedicated to xeno- treated with 2.4 ␮g BPArkg on days 11 to 17 of androgens and antiandrogens will help to demon- gestation. strate that invertebrates, like vertebrates, are sen- Laws et al.Ž. 2000 tested the estrogenic activity sitive to endocrine disruption at environmentally of BPA and OP in comparison to other xeno- relevant concentrations and additionally that they estrogens, ethinylestradiol, and 17ß-estradiol in are provided with complex feed back mechanisms feeding experiments with prepubertal and in their steroid metabolism. ovariectomised adult ratsŽ 3-day uterotrophic as- say. . They showed that both compounds were estrogenic at daily oral doses of 50᎐200 mgrkg Acknowledgements Ž.OP and 100᎐200 mgrkg Ž BPA . . Sharpe et al. Ž.1995 report a reduced testicular size and sperm The experiments with Marisa cornuarietis were production in male offspring of Wistar rats which financially supported by the Federal Environmen- were exposed to low doses of BPA during gesta- tal Agency, BerlinŽ. R&D project 216 02 001r04 . tion and lactation via the drinking water but these We would also like to thank Constanze Hannich, results could not be confirmed by Cagen et al. Ulrike Schneider and Christina Schmidt for their Ž.1999 . excellent technical assistance and two anonymous Although the relevance of all reported effects reviewers for their helpful comments. of BPA and OP in vertebrates was not checked for Marisa cornuarietis and Nucella lapillus, either because the experimental design was unsuited for References this purposeŽ e.g. advancement of puberty; modu- Ahel, M., Giger, W. and Koch, M.Ž. 1994a . Behaviour of lation of mating behaviour. or the samples are alkylphenol polyethoxylate surfactants in the aquatic envi- not yet analysedŽ e.g. spermiogenesis and oogene- ronmentᎏI. Occurrence and transformations in sewage ᎐ sis in F specimens. , it is obvious that a number treatment. Water Res. 28, 1131 42. 1 Ahel, M., Giger, W. and Schaffner, C.Ž. 1994b . Behaviour of of responses are comparable. This is true espe- alkylphenol polyethoxylate surfactants in the aquatic envi- cially for the stimulation of egg production and ronmentᎏII. Occurrence and transformation in rivers. the positive uterotrophic effect in both species as Water Res. 28, 1143᎐52. well as for the reduction of penis and prostate Andersen, H.R., Andersson, A.M., Arnold, S.F., Autrup, H., gland length in male Nucella. Another important Barfoed, M., Beresford, N.A., Bjerregaard, P., Chris- tiansen, L.B., Gissel, B., Hummel, R., Jorgensen, E.B., point is that both snail species exhibit already Korsgaard, B., Le Guevel, R., Leffers, H., McLachlan, J., marked effects at the lowest nominal test concen- Moller, A., Nielsen, J.B., , N., Oles-Karasko, A., trations indicating that LOECŽ lowest observed Pakdel, F., Pedersen, K.L., Perez, P., Skakkebaek, N.E., 396 J. Oehlmann et al.

Sonnenschein, C., Soto, A.M., Sumpter, J.P., Thorpe, S.M. Environment Agency.Ž. 1998 . Endocrine-disrupting substances and Grandjean, P.Ž. 1999a . Comparison of short-term es- in wildlife: A review of the scientific evidence and strategic trogenicity tests for identification of hormone-disrupting response? Environment Agency, Bristol. Publication No. chemicals. En¨iron. Health Perspect. 107,Ž. Suppl. 1 , 89᎐108. HO-11r97-100-B-BANP. Andersen, H.R., Halling-Sorensen, B. and Kusk, K.O.Ž. 1999b : Gibbs, P.E., Bryan, G.W., Pascoe, P.L. and Burt, G.R.Ž. 1987 . A parameter for detecting estrogenic exposure in the cope- The use of the dog-whelk, Nucella lapillus, as an indicator pod Acartia tonsa. Ecotoxicol. En¨iron. Saf. 44,56᎐61. of tributyltinŽ. TBT contamination. J. Mar. Biol. Ass. U.K. Ashby, J., Tinwell, H. and Haseman, J.Ž. 1999 . Lack of effects 67, 507᎐23. for low dose levels of bisphenol A and diethylstilbestrol on Giesy, J.P., Pierens, S.L., Snyder, E.M., Miles-Richardson, S., the prostate gland of CF1 mice exposed in utero. Regul. Kramer, V.J., Snyder, S.A., Nichols, K.M. and Villeneuve, Toxicol. Pharmacol. 30, 156᎐66. D.A.Ž. 2000 . Effects of 4-nonylphenol on fecundity and Bauer, B., Fioroni, P., Ide, I., Liebe, S., Oehlmann, J., Stroben, biomarkers of estrogenicity in fathead minnows Ž Pime- E. and Watermann, B.Ž. 1995 . TBT effects on the female phales promelas.. En¨iron. Toxicol. Chem. 19, 1368᎐77. genital system of Littorina littorea: A possible indicator of Gist, G.L.Ž. 1998 . National Environmental Health Association tributyltin pollution. Hydrobiologia 309,15᎐27. position on endocrine disrupters᎐adopted July 2, 1997. J. Bauer, B., Fioroni, P., Schulte-Oehlmann, U., Oehlmann, J. En¨iron. Health 60,21᎐3. and Kalbfus, W.Ž. 1997 . The use of Littorina littorea for Greim, H.Ž. 1998 . Hormonahnlich¨ wirkende Stoffe in der tributyltinŽ. TBT effect monitoringᎏResults from the Umwelt. Nachr. Chem. Tech. Lab. 46,63᎐6. German TBT survey 1994r1995 and laboratory experi- Howard, P.H.Ž. 1989 . Handbook of En¨ironmental Fate and ments. En¨iron. Pollut. 96, 299᎐309. Exposure Data for Organic Chemicals. Vol. I: Large Produc- Bettin, C., Oehlmann, J. and Stroben, E.Ž. 1996 . TBT-induced tion and Priority Pollutants. Chelsea: Lewis Publ. imposex in marine neogastropods is mediated by an in- Howdeshell, K.L., Hotchkiss, A.K., Thayer, K.A., Vanden- creasing androgen level. Helgolander¨ Meeresunters. 50, bergh, J.H. and vom Saal, F.S.Ž. 1999 . Exposure to bisphe- 299᎐317. nol A advances puberty. Nature 401, 763᎐4. BUA.Ž. 1991 . Nonylphenol. Stoffbericht des Beratergremiums fur¨ Jobling, S. and Sumpter, J.P.Ž. 1993 . Detergent components in umweltrele¨ante AltstoffeŽ. BUA . Weinheim: VCH. sewage effluent are weakly oestrogenic to fish: An in vitro BUA.Ž. 1997 . Bisphenol A Ž2,2-Bis-4- Žhydroxyphenyl . propan .. study using rainbow trout Ž.Oncorhynchus mykiss hepato- Stoffbericht 203 des Beratergremiums fur¨ umweltrele¨ante cytes. Aquat. Toxicol. 27, 361᎐72. AltstoffeŽ. BUA . Stuttgart: S. Hirzel. Kloas, W., Lutz, I. and Einspanier, R.Ž. 1999 . Amphibians as a Cagen, S.Z., Waechter, J.M., Dimond, S.S., Breslin, W.J., model to study endocrine disruptors: II. Estrogenic activity Butala, J.H., Jekat, F.W., Joiner, R.L., Shiotsuka, R.N., of environmental chemicals in vitro and in vivo. Sci. Total Veenstra, G.E. and Harris, L.R.Ž. 1999 . Normal reproduc- En¨iron. 225,59᎐68. tive organ development in Wistar rats exposed to bisphe- Krishnan, A.V., Starhis, P., Permuth, S.F., Tokes, L. and nol A in the drinking water. Regul. Toxicol. Pharmacol. 30, Feldman, D.Ž. 1993 . Bisphenol A: An estrogenic substance 130᎐9. is released from polycarbonate flasks during autoclaving. Caspers, N.Ž. 1998 . No estrogenic effects of bisphenol A in Endocrinology 132, 2279᎐86. Daphnia magna STRAUS. Bull. En¨iron. Contam. Toxicol. Laws, S.C., Carey, S.A., Ferrell, J.M., Bodman, G.J. and 61, 143᎐8. Cooper, R.L.Ž. 2000 . Estrogenic activity of octylphenol, CES.Ž. 1993 . Uses, fate and entry to the environment of nonylphenol, bisphenol A and methoxychlor in rats. Toxi- nonylphenol ethoxylates. Report for the Department of col. Sci. 54, 154᎐67. the Environment UK by Consultants in Environmental Lozan,´¨ J.L.Ž. 1992 . Angewandte Statistik fur Naturwis- Sciences Ltd, Beckenham. senschaftler. Berlin, Hamburg: Parey. Clark, L.B., Rosen, R.T., Hartman, T.G., Louis, J.B., Suffet, Lutz, I. and Kloas, W.Ž. 1999 . Amphibians as a model to study I.H., Lippincott, R.L. and Rosen, J.D.Ž. 1992 . Determina- endocrine disruptors: I. Environmental pollution and es- tion of alkylphenol ethoxylates and their acetic acid trogen receptor binding. Sci. Total En¨iron. 225,49᎐57. derivates in drinking water by particle beam liquid chro- Marquardt, H. and Schafer,¨ S.G.Ž.Ž. eds . 1994 . Lehrbuch der matographyrmass spectrometry. Int. J. En¨iron. Anal. Toxikologie. Mannheim, Leipzig, Wien, Zurich:¨ Bibli- Chem. 147, 167᎐80. ographisches Institut Wissenschaftsverlag. deFur, P.L., Crane, M., Ingersoll, C. and Tattersfield, L.Ž. eds . Matthiessen, P. and Gibbs, P.E.Ž. 1998 . Critical appraisal of Ž.1999 . Endocrine disruption in invertebrates: Endocrinol- the evidence for tributyltin-mediated endocrine disruption ogy, testing, and assessment. In Proceedings of the Work- in mollusks. En¨iron. Toxicol. Chem. 17,37᎐43. shops on Endocrine Disruption in In¨ertebrates,12᎐15 Dec. Oehlmann, J.Ž. 1994 . Imposex bei Muriciden Ž Gastropoda, 1998, Noordwijkerhout, The Netherlands. Pensacola, FL: Prosobranchia., eine okotoxikologische¨ Untersuchung zu SETAC Press. TBT-Effekten. Gottingen,¨ Cuvillier. Dodds, E.C. and Lawson, W.Ž. 1936 . Synthetic estrogenic Oehlmann, J., Stroben, E. and Fioroni, P.Ž. 1988 . Zur agents without the phenanthrene nucleus. Nature 137, 996. Anatomie und Histologie des Fortpflanzungssystems von Dodds, E.C. and Lawson, W.Ž. 1938 . Molecular structure in Nucella lapillus Ž.L., 1758 . Zool. Anz. 221, 101᎐16. relation to oestrogenic activity. Compounds without a Oehlmann, J., Stroben, E. and Fioroni, P.Ž. 1991 . The morpho- phenanthrene nucleus. Proc. Royal Soc. London B. 125, logical expression of imposex in Nucella lapillus ŽLIN- 222᎐32. NAEUS.ŽGastropoda: Muricidae . . J. Moll. Stud. 57, 375᎐90. Endocrine Effects on Prosobranchs. I: Xeno-Estrogens 397

Oehlmann, J., Fioroni, P., Stroben, E. and Markert, B.Ž. 1996 . and Hashimoto, K.Ž. 1999 . Exposure with the environmen- TributyltinŽ. TBT effects on Ocinebrina aciculata ŽGastro- tal estrogen bisphenol A disrupts the male reproductive poda: Muricidae. : Imposex development, sterilization, sex tract in young mice. Life Sci. 65, 2351᎐7. change and population decline. Sci. Total En¨iron. 188, vom Saal, F.S., Cooke, P.S., Buchanan, D.L., Palanza, P., 205᎐23. Thayer, K.A., Nagel, S.C., Parmigiani, S. and Welshons, Rippen, G.Ž. 1999 . Handbuch Umweltchemikalien. Stoffdaten, W.V.Ž. 1998 . A physiologically based approach to the study Pruf¨ ¨erfahren, Vorschriften, 3rd ed., 49th supplement issue. of bisphenol A and other estrogenic chemicals on the size Landsberg, Ecomed. of reproductive organs, daily sperm production, and behav- Schulte-Oehlmann, U., Fioroni, P., Oehlmann, J. and Stroben, ior. Toxicol. Ind. Health 14, 239᎐60. E.Ž. 1994 . The genital system of Marisa cornuarietis ŽGas- Weber, E.Ž. 1972 . Grundriß der biologischen Statistik, 7th ed. tropoda, Ampullariidae.ᎏA morphological and histologi- Stuttgart: Fischer. cal analysis. Zool. Beitr. N.F. 36,59᎐81. White, R., Jobling, S., Hoare, S.A., Sumpter, J.P. and Parker, Schulte-Oehlmann, U., Bettin, C., Fioroni, P., Oehlmann, J. M.G.Ž. 1994 . Environmentally persistant alkylphenolic and Stroben, E.Ž. 1995 . Marisa cornuarietis ŽGastropoda, compounds are estrogenic. Endocrinology 135, 175᎐82. Prosobranchia. : a potential TBT bioindicator for freshwa- Yamakoshi, Y., Otani, Y., Fujii, S. and Endo, Y.Ž. 2000 . ter environments. Ecotoxicology 4, 372᎐84. Dependence of estrogenic activity on the shape of the Sharpe, R.M., Fisher, J.S., Millar, M.M., Jobling, S. and 4-alkyl substituent in simple phenols. Biol. Pharm. Bull. Sumpter, J.P.Ž. 1995 . Gestational and lactational exposure 23, 259᎐61. of rats to xenoestrogens results in reduced testicular size Zou, E. and Fingeman, M.Ž. 1997 . Synthetic estrogenic agents and sperm production. En¨iron. Health Perspect. 103, do not interfere with sex differentiation but do inhibit 1136᎐43. molting of the Cladoceran Daphnia magna. Bull. En¨iron. Soto, A.M., Justicia, H., Wray, J.W. and Sonnenschein, C. Contam. Toxicol. 58, 596᎐602. Ž.1991 . p-Nonylphenol, an estrogenic xenobiotic released Zou, E. and Fingeman, M.Ž. 1999 . Effects of estrogenic agents from ‘modified’ polystyrene. En¨iron. Health Perspect. 92, on chitobiase activity in the epidermis and hepatopancreas 167᎐73. of the fiddler crab, Uca pugilator. Biol. Ecotoxicol. En¨iron. Takao, T., Nanamiya, W., Nagano, I., Asaba, K., Kawabata, K. Saf. 42, 185᎐90.