Accuracy of methods of sex steroid determination
Ellis, T.; Scott, A.P.; Waring, C.; Margiotta-Casaluci, L.; Pottinger, T.G.; Morris, S.; Sumpter, J.P.. 2011 Accuracy of methods of sex steroid determination. Defra, 28pp. (CEH Project Number: C03801)Before downloading, please read NORA policies.
N015247CR.pdf - Published Version
Reproductive hormones (estrogenic and androgenic steroids) enter natural water bodies from various sources, e.g. human sewage, farmed livestock and vertebrate wildlife. A previous Defra research study (SF0241 Impacts of intensive in-river aquaculture on wild salmonids) reported very high concentrations of such steroids in two UK rivers (the River Test and R. Avon in Hampshire/Wiltshire) in the vicinity of trout farms. The mean reported concentrations of the natural steroids 11-ketotestosterone (11-KT), testosterone (T) and oestradiol (E2) ranged from 4 to 79 ng/L. The maximum level observed in an individual sample was 145 ng/L for 11-KT. Furthermore, the subject brown trout and rainbow trout farms were implicated as a source of the steroids, as mean levels reported “downstream” of the trout farms were 1.3 to 6 fold higher than those “upstream”. Nevertheless, the steroid concentrations reported both upstream and downstream of the trout farms would be of great concern for wild fish in these rivers due to potential for endocrine disruption, i.e. when exogenous substances negatively impact the endogenous hormone systems and reproductive function of the organism. The synthetic sex steroid ethinyl-estradiol (EE2) was also measured in the SF0241 study. EE2 originates from human prescription medicine (the contraceptive pill) via sewage treatment works, and has been studied in rivers due to its endocrine disruptive activity. Levels of EE2 reported in SF0241 were one to two orders of magnitude lower than the natural steroids. Mean values were in the range of 0.13 – 0.30 ng/L, and were not found to be elevated downstream of the trout farms. The very high levels of the natural sex steroids that were reported were not only of concern because of potential adverse effects on resident fish, but were also surprisingly high. In the SF0241 study, all steroid levels in acquired water samples had been measured using commercial Enzyme-ImmunoAssay (EIA) kits (also referred to as Enzyme-Linked Immuno-Sorbent Assay, ELISA kits). However, the kits used to measure the natural steroids used a different enzyme component in the assay to the EE2 kits. Environmental water samples are typically concentrated before assay, and it was speculated that other compounds in the river water may have interfered with the enzyme stage in the kits for the natural steroids. It was therefore hypothesised that the very high concentrations of the natural steroids reported in SF0241 were false positives due to interference in the EIAs. Due to concern over the possible inaccuracy of the EIA kits (and trout farms as a source of endocrine disrupting compounds), Defra’s Chemicals and Nanotechnology Division funded this project (CB0427) to examine the “Accuracy of methods of sex steroid determination”. The aim of the project was to test the above hypothesis by repeating the sampling and sample processing, and then assaying for the steroids using the EIA kits and an additional method – radioimmunoassay (RIA) - expected to be less susceptible to interference. Water samples were collected from the two trout farm sites, one on the R. Test and one on the R. Avon, between January and June 2010 and extracted (C18 solid phase extraction (SPE) followed by extract clean-up with aminopropyl SPE). Additional “spiked” and “blank” water samples were also prepared and processed. Four independent laboratories conducted EIAs and RIAs (3 laboratories per assay technique) for the four steroids (11-KT, T, E2, EE2) using replicate aliquots of the same 44 samples, each aliquot representing ca 1 L of river water. Participating laboratories (other than the lead laboratory) conducted the assays blind (i.e. were unaware of the sample details) and returned the calculated steroid concentrations to the project leaders for collation. A few anomalous results were questioned and, upon investigation, were found to be due to human errors and were subsequently corrected. There was broad agreement between the EIA and RIA measurements for all four steroids showing that the EIA kits did not generate erroneously high values. The base hypothesis for the project, i.e. that some EIA kits generate erroneously high values, was therefore rejected. The two assay methods were comparable for accuracy and precision. Recovery of steroids using the SPE methodology was examined using water samples spiked with known amounts of steroids. It was found that the recovery efficiency varied between samples and steroids. Higher recoveries were evident for the oestrogens (mean recovery 69% and 67% for E2 and EE2 respectively) than for the androgenic compounds (mean recovery 29% and 46% for 11KT and T respectively). Measured steroid concentrations in river water samples were all <0.6 ng/L. The measured concentrations cannot be considered definitive, as they are uncorrected for recovery efficiency. Nevertheless, they are considered low, unlikely to be of concern for endocrine disruption, and demonstrate that the water steroid levels reported in SF0241 were not typical of river steroid concentrations in 2010. The levels of 11KT, T, E2 and EE2 in the rivers were respectively 1300, 70, 90 and 3 times lower than reported in SF0241. In addition to the major difference in measured levels between the two studies: • the relative concentrations also differed, with 11KT being the lowest rather than highest as reported in SF0241. • river steroid levels were higher in Spring rather than the Winter period, being opposite to the seasonal effect reported in SF0241 • river EE2 levels in 2010 were too low to be detectable by EIA (equivalent to <0.05 ng/L). • the clear elevation in river water concentrations of the natural steroids associated with the fish farms reported in SF0241 was not evident in the present study. It was not possible to provide a definitive explanation for the differences found between the present results and those reported within SF0241. Eleven alternate hypotheses are discussed that could have contributed to the divergent results. The available evidence points towards miscalculation and assaying errors within SF0241 as the probable cause of differences. However, this hypothesis could not be definitively accepted because the raw assay data and the calculations from which the SF0241 results derived had not been retained. Published information on steroid output from fish farms is presented. It is suggested that there is no urgent requirement to further examine the steroid output from UK fish farms. Although the current research did indicate a possible 0.13 ng/L increase in testosterone in the immediate outflow of one farm, this was questionable and, if real, is below concentrations that cause endocrine disruption and would be further diluted in the receiving channel and then main river. During this project, a number of potential sources of error associated with the measurement of steroids from samples were identified and are discussed. It is suggested that guidance for quality control could be developed. Future research into river water steroids could also include: • comparisons of the recovery efficiencies of different steroids from water samples, and a full optimisation and validation of the most appropriate extraction methodology for the different steroids • a comparison of ‘modelled’ versus ‘measured’ river water steroid levels to examine whether the exclusion of other sources of steroids, e.g. agricultural livestock, is an important omission.
|Item Type:||Report (UNSPECIFIED)|
|Programmes:||CEH Topics & Objectives 2009 onwards > Water|
|Funders/Sponsors:||Department of Environment, Food and Rural Affairs (Defra)|
|Additional Keywords:||steroid hormones, endocrine disruption, androgen, estrogen, fish farm|
|NORA Subject Terms:||Biology and Microbiology
|Date made live:||19 Oct 2011 09:11|
Actions (login required)