ToxSci Advance Access originally published online on April 2, 2008
Toxicological Sciences 2008 104(1):228-230; doi:10.1093/toxsci/kfn067
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Comments on "Evaluation of Estrogenic Activities of Aquatic Herbicides and Surfactants Using a Rainbow Trout Vitellogenin Assay"
Dow AgroSciences LLC, Indianapolis, IN 46268, T: 317-337-3137, E-mail: vjkramer{at}dow.com
Received February 6, 2008; accepted March 2, 2008
In reading the paper, "Evaluation of estrogenic activities of aquatic herbicides and surfactants using a rainbow trout vitellogenin assay" (Xie et al., 2005
, Toxicol. Sci. 87, 391–398), we have identified a number of issues that require additional detail or possible correction of the information and conclusions as presented.
A major technical issue concerns the use of "juvenile" trout combined with very limited replication, which calls into question the validity of the conclusions derived from the experimental results. The authors apparently did not account for an underlying and uncontrollable source of variability that confounds interpretation of these results. The sex of each individual trout used in the studies was not identified. Instead, the trout are described only as: "Juvenile rainbow trout (standard length: 11.5 ± 2.2 cm)." The Toxicological Sciences article also references their work on 17β-estradiol in Xie et al. (2004), which describes the trout used as "Mixed-sex juvenile rainbow trout...." The use of juvenile trout in the Toxicological Sciences article creates the likelihood that the presence of juvenile females could greatly skew the baseline mean response in any subset of fish selected for a treatment. The authors used very small sample sizes, N = 6 (three replicates of two fish each). With such a small sample size it is highly likely that some treatments could receive, five males and one female, or five females and one male, or four females and two males or even all males or all females. An uneven distribution of male and female juvenile trout would greatly bias any results. The potential presence of juvenile female rainbow trout is critical to study interpretation because females exhibit blood vitellogenin levels up to 1000 times that of male rainbow trout (males = 0.38 µg/ml; juvenile females = 423 µg/ml; Bon et al., 1997). Because the sex of the juvenile fish was apparently not known or controlled as part of the experimental design, it cannot be excluded that the results could be due solely to the presence of more females in any particular treatment relative to controls. Furthermore, the vitellogenin results as presented in the Toxicological Sciences article are clouded by the attempted normalization of the vitellogenin levels in the blood to the total protein in the blood, that is, ng Vtg/mg protein. This normalization was unnecessary, departs from common practice (Bon et al., 1997; Thorpe et al., 2000, 2001), obscures their data from comparison with other published values, and does not correct for the inherent variability caused by the use of mixed sex fish.
A second major technical issue is that the Toxicological Sciences article appears to suggest that the estradiol exposures for the calculation of the EEQs (estradiol equivalent concentrations) occurred contemporaneously with the pesticide exposures reported in this study. However, it appears they occurred previous to this study and were not conducted in a similar manner that would allow them to be used for EEQ calibration of the pesticide exposures. Xie et al. (2005) report that estradiol EEQ calculations were based on exposures consisting of a daily static renewal exposure for 7 days under a 16:8-h light:dark cycle and five unspecified estradiol concentrations. However, the authors reference Xie et al. (2004) as the method by which EEQ values were derived, which describes a daily static renewal exposure for 14 days under a 14:10-h light:dark cycle using four estradiol concentrations ranging from 0.5 to 100 ng/l. The most critical difference between these methods is the exposure duration of 14 days in Xie et al. (2004) instead of 7 days in the Toxicol. Sci. work. Vitellogenin responses may differ greatly over time (Thorpe et al., 2000) and thus an accurate calibration of the vitellogenin response cannot be obtained by comparing responses taken at different times. Furthermore, the vitellogenin responses to the five concentrations of 17-β estradiol were not presented in the Toxicol. Sci. paper. Two additional issues potentially invalidating the use of the EEQ calculations are (1) the calibration curve presented in Xie et al. (2004) was drawn through a single point, and (2) the response levels reported in Toxicological Sciences (up to 80 ng vitellogenin/mg plasma protein) were not sufficiently great to rise above the baseline of the data presented in Xie et al. (2004). Therefore, this calibration curve cannot be used to predict an equivalent estradiol concentration in the Toxicological Sciences paper.
Our third concern regards the conclusions of "greater than additive" and "less than additive" response. These conclusions were apparently not supported by tests of statistical significance comparing responses of mixture exposures with responses predicted by mixture toxicity theory (Brian et al., 2005; Olmstead and LeBlanc, 2005) and also with response levels of the individual materials tested alone. Failure to demonstrate a statistically significant difference provides no basis to reject a null hypothesis of a simple additive response. The authors did not present any of these calculations and therefore there is no basis to independently conclude that the responses were "greater than additive" or "less than additive." For example, the vitellogenin data for triclopyr tested alone is not shown. This baseline data is necessary to determine whether the observed responses in the binary exposures are significantly different from the observed responses of the materials tested alone. It is also noted that the error bars for the triclopyr mixture treatments (Fig. 5) appear to overlap both with the predicted response and with zero. More puzzling is their conclusion that the data from the "field setting" support the claim that triclopyr causes "greater than additive" effects in combination with an alkylphenol-containing surfactant. Given that triclopyr was not detected (< 2 ng/l) in the pond water samples that were tested, this claim is not supportable. It is not enough to suggest that because the pond was previously treated with a triclopyr-containing herbicide, it necessarily follows that triclopyr, either alone or in combination with other materials, elicited the effects in their laboratory test system.
Overall, the methodological issues described above suggest that the results and conclusions in the Toxicological Sciences article are unsupportable. In particular, we believe that the authors' assertion that 2,4-dichlorophenoxyacetic acid (2,4-D) is estrogenic in rainbow trout is unsupported by the data presented and is not consistent with published information on 2,4-D. Unfortunately, Xie et al. also defend their conclusions by incorrectly claiming literature support for 2,4-D estrogenicity while at the same time failing to cite literature inconsistent with their conclusions. Xie et al. erroneously cite only one paper purporting to support their conclusion (Nishihara et al., 2000). This study tested over 500 chemicals in a yeast two-hybrid assay and specifically reported that 2,4-D was not estrogenic. In addition, the authors failed to describe the findings of at least 10 other publications, all of which reported negative results for estrogenicity (Blair et al., 2000; Hurst and Sheahan, 2003; Hwang, 2002; Jungbauer and Beck, 2002; Jung et al., 2004; Kojima et al., 2004; Lin and Garry, 2000; Petit et al., 1997; Soto et al., 1995; Vonier et al., 1996).
Our final concern is with what appears to be discrepancies between the Toxicological Sciences article and previously released versions of the data and conclusions made public by the San Francisco Estuary Institute (SFEI) (http://www.sfei.org/sfeireports.htm) in reports released as part of the Aquatic Pesticides Monitoring Program (APMP) operated by SFEI from 2002 to 2004. Two reports were published: an interim report of Phase 2 of the project "APMP2" (Siemering, 2004a) and a final report "APMP3" (Siemering, 2004b). The dates of finalization of the APMP reports (April 2004 and January 2005) pre-date that of the Toxicol. Sci. manuscript submission. Comparison of these reports with that published in Toxicological Sciences revealed a number of differences in the depictions of the methods, data and conclusions. For example, the Toxicological Sciences article apparently presents recalculation of the data and/or statistics, resulting in significant changes in the relative size of the error bars depicted in Figure 1 versus the same data as presented in APMP2 Figures 4 and 5 (pp. 72–73). Asterisks indicating statistically significant estrogenic responses appear in Figure 1 that are not present in Figures 4 and 5 of APMP2. This apparent re-analysis led to strikingly different conclusions regarding estrogenicity of 2,4-D. In the APMP2 report, it is stated (p. 72) regarding vitellogenin induction, "...2,4-D caused a fourfold increase in induction over animals exposed to untreated water." In Xie et al. with reference to the same data the authors state, "...2,4-D had a 93-fold increase in plasma vitellogenin compared to the control fish." No explanation is apparent for the dramatically different descriptions of the same data. Another difference is found under "Range finding experimental exposure" in the Toxicological Sciences article (p. 392), in which it is stated, "The measured concentrations for pesticides glyphosate, 2,4-D, diquat dibromide, and triclopyr, were 0.11, 1.64, 2.07, and 1.25 mg/l, respectively." (Emphasis added.) The SFEI publications report instead that the concentrations were measured in the stock solutions (APMP3, p. 5-3, Table 1). There is no measured data provided in the Toxicological Sciences article (nor in the SFEI reports) to demonstrate that the exposure tank water samples were analyzed for concentrations of the test chemicals.
REFERENCES
Blair RM, Fang H, Branham WS, Hass BS, Dial SL, Moland CL, Tong W, Shi L, Perkins R, Sheehan DM. The estrogen receptor relative binding affinities of 188 natural and xenochemicals: Structural diversity of ligands. Toxicol. Sci. (2000) 54:138–153.
Bon E, Barbe U, Nunez Rodriguez J, Cuisset B, Pelissero C, Sumpter JP, LeMenn F. Plasma vitellogenin levels during the annual reproductive cycle of the female rainbow trout (Oncorhynchus mykiss): Establishment and validation of an ELISA. Comp. Biochem. Physiol. (1997) 117B:75–84.[CrossRef][Medline]
Brian J, Harris CA, Scholze M, Backhaus T, Booy P, Lamoree M, Pojana G, Jonkers N, Runnalls T, Bonfà A, et al. Accurate prediction of the response of freshwater fish to a mixture of estrogenic chemicals. Environ. Health Perspect. (2005) 113:721–728.[Web of Science][Medline]
Hurst MR, Sheahan DA. The potential for oestrogenic effects of pesticides in headwater streams in the UK. Sci. Total Environ. (2003) 301:87–96.[CrossRef][Medline]
Hwang U-G. Effect of 2,4-dichlorophenoxyacetic acid on vitellogenin synthesis and E2-ER binding affinity or hepatocytes in rainbow trout (Oncorhynchus mykiss). Hanguk Yangsik Hakhoechi (2002) 15:31–37. (English abstract).
Jung J, Ishida K, Nishihara T. Anti-estrogenic activity of fifty chemicals evaluated by in vitro assays. Life Sci. (2004) 74:3065–3074.[CrossRef][Web of Science][Medline]
Jungbauer A, Beck V. Yeast reporter system for rapid determination of estrogenic activity. J. Chromatogr. B (2002) 777:167–178.[CrossRef]
Kojima H, Katsura E, Takeuchi S, Niiyama K, Kobayashi K. Screening for estrogen and androgen receptor activities in 200 pesticides by in vitro reporter gene assays using Chinese hamster ovary cells. Environ. Health Perspect. (2004) 112:524–531.[Web of Science][Medline]
Lin N, Garry VF. In vitro studies of cellular and molecular developmental toxicity of adjuvants, herbicides, and fungicides commonly used in Red River Valley, Minnesota. J. Toxicol. Environ. Health A (2000) 60:423–439.[CrossRef][Medline]
Nishihara T, Nishikawa J, Kanayama T, Dakeyama T, Saito K, Imagawa M, Takatori S, Kitagawa Y, Hori S, Utsumi H. Estrogenic activities of 517 chemicals by yeast two-hybrid assay. J. Health Sci. (2000) 46:282–298.
Olmstead AW, LeBlanc GA. Toxicity assessment of environmentally relevant pollutant mixtures using a heuristic model. Int. Environ. Assess. Manage. (2005) 1:114–122.[CrossRef]
Petit F, Le Goff P, Cravedi J-P, Valotaire Y, Pakdel F. Two complementary bioassays for screening the estrogenic potency of xenobiotics; recombinant yeast for trout estrogen receptor and trout hepatocyte cultures. J. Mol. Endocrinol. (1997) 19:321–335.
Siemering G. Aquatic pesticide monitoring project report phase 2(2003) monitoring project report. In: SFEI contribution 108 (2004a) Oakland, CA: San Francisco Estuary Institute.
Siemering G. Aquatic Pesticide Monitoring Program Phase 3 (2004) Monitoring Project. SFEI Contribution 391—Siemering G, ed. (2004b) Oakland, CA: San Francisco Estuary Institute.
Soto AM, Sonnenschein C, Chung KL, Fernandez MF, Olea N, Serrano FO. The E-SCREEN assay as a tool to identify estrogens: An update on estrogenic environmental pollutants. Environ. Health Perspect. Suppl. (1995) 103(S7):113–122.
Thorpe KL, Hutchinson TH, Hetheridge MJ, Scholze M, Sumpter JP, Tyler CR. Assessing the biological potency of binary mixtures of environmental estrogens using vitellogenin induction in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Sci. Technol. (2001) 35:2476–2481.[Medline]
Thorpe KL, Hutchinson TH, Hetheridge MJ, Sumpter JP, Tyler CR. Development of an in vivo screening assay for estrogenic chemicals using juvenile rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. (2000) 19:2812–2820.[CrossRef][Web of Science]
Vonier PM, Andrew CD, Mclachlan JA, Guillette LJ Jr, Arnold SF. Interaction of environmental chemicals with the estrogen and progesterone receptors from the oviduct of the American alligator. Environ. Health Perspect. (1996) 104:1318–1322.[Web of Science][Medline]
Xie L, Sapozhnikova Y, Bawardi O, Schlenk D. Evaluation of wetland and tertiary wastewater treatments for estrogenicity using in vivo and in vitro assays. Arch. Environ. Contam. Toxicol. (2004) 48:81–86.[CrossRef][Web of Science]
Xie L, Thrippleton K, Irwin MA, Siemering GS, Mekebri A, Crane D, Berry K, Schlenk D. Evaluation of estrogenic activities of aquatic herbicides and surfactants using an rainbow trout vitellogenin assay. Toxicol. Sci. (2005) 87:391–398.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||