2010 Progress Report: Transport/Fate/Ecological Effects of Steroids from Poultry Litter & Evaluations of Existing/Novel Management Strategies
EPA Grant Number:
Transport/Fate/Ecological Effects of Steroids from Poultry Litter & Evaluations of Existing/Novel Management Strategies
Fisher, Daniel J.
, Kane, Andrew S.
, Klauda, Ronald J.
, Staver, Kenneth
, VanVeld, Peter
, Yonkos, Lance T.
Wye Research and Education Center
Maryland Department of Natural Resources
School of Medicine at the University of Maryland
Virginia Institute of Marine Science
EPA Project Officer:
August 1, 2007 through
June 30, 2010
(Extended to July 31, 2012)
Project Period Covered by this Report:
December 1, 2008 through December 1,2010
Fate and Effects of Hormones in Waste from Concentrated Animal Feeding Operations (CAFOS) (2006)
The award was converted to a cooperative agreement to allow participation of US EPA scientists who are currently working on endocrine disruption issues. Dr. Steve Hutchins and his group at the USEPA National Risk Management Research Laboratory in Ada, OK analyzed aqueous samples for steroids using his GC-MS method. Drs. Jim Lazorchak and Tirumuru Reddy at the USEPA Molecular Indicator Research Branch in Cincinnati, OH analyzed fish livers for vitellogenin (Vtg) using a new Vtg mRNA gene expression assay. Dr. Vickie Wilson of the USEPA, NHEERL, Reproductive Toxicology Division in Research Triangle Park, NC analyzed aqueous samples for estrogenicity (EEQ) using a novel in-vitro estrogen-inducible reporter-gene assay.
The work is proceeding as outlined in the original award.
Poultry litter was applied to two 33 acre research watersheds at WREC in 2008, 2009, and 2010. One field was managed using a soil conservation tillage technique called No-till (NT) whereby the field was not tilled following litter application. This is used to eliminate soil loss from a field. The adjacent filed was managed using a Turbo-till technique (TT) whereby the litter was lightly tilled and mixed into the top few inches of soil. This technique is designed to reduce soil loss compared to conventional tillage but also to reduce overland transport of water soluble and particle-bound contaminants following precipitation events. Nitrogen (N), phosphorous (P), estrogen (E2 = 17β-estradiol, E1 = Estrone) and estrogenicity (in vitro assay employing an estrogen sensitive human mammary cell reporter gene complex with results reported as EEQ or 17β-estradiol equivalents) concentrations were measured in run-off from these watersheds following rain events to determine the best tillage technique to reduce overland transport of contaminants from poultry litter amended fields. Total nutrient loads to receiving water bodies were calculated using runoff concentrations associated with discrete hydrograph data (normalized to kg/ha). Peak estrogen concentrations were measured to investigate whether levels exceeded effects thresholds of biota in receiving bodies.
Results and Discussion: During 2008 and 2009 the field managed under TT practices generated substantially less N and P in runoff than the field managed using NT (Figure 1). Initial post- application rains in 2008 caused runoff from the NT field but not the TT field due to the ability of the TT field to absorb more water. Subsequent rains produced similar volumes of runoff from both fields. In general, N tends to be more soluble while P tends to sorb more strongly to sediments. Therefore, vulnerable N leaches more rapidly from litter and surface sediments during initial rains than does P. This explains N loads being much greater in early 2008 runoff from the NT field compared to the TT field but being similar from both fields during later runoff events. Because P is slower to leach from sediments, loads in runoff from the NT field remained above those from the TT field for the entire 2008 season. During 2009 litter application did not occur until the end of May so concentrations of N and P in May runoff reflect residuals from previous field activities. Post application rain events (June) produced higher N and P loads from NT runoff than from TT runoff. Overall N and P loads were lower than the previous year reflecting the impact of multiple small rain events that did not produce runoff. This result was even more dramatic in 2010 where no post-application runoff occurred until July. By this point most N and P had leached below the sediment surface which had smoothed and compacted minimizing nutrient loads in runoff.
Estrogen concentrations in runoff followed the same trend as nutrients (Table 1). In 2008 the first rain event after application produced copious runoff from the NT field but none from the TT field. Measured E2 and E1 concentrations were 9.2 ng/L and 32 ng/L, respectively. Estrogenicity, which should closely approximate E2, was only 2.6 ng/L. As this is a live cell assay, it is vulnerable to poor performance and under-estimation in instances where tested materials are acutely toxic. Ammonia levels of 25 mg/L in the NT runoff likely produced cytotoxicity explaining the low EEQ results compared to measured E2. A second 2008 rain event produced runoff from both fields with estrogen concentrations being 2x to3x higher in NT runoff than TT runoff. Estrogen concentrations were generally lower in 2009 than 2008. Over multiple 2009 rain events estrogen measures (where detectable) were consistently higher in NT runoff compared to TT runoff. Preliminary data from 2010 suggest slightly higher estrogen levels in NT runoff than TT runoff but analyses are, as yet, incomplete.
b) Runoff was collected from the NT watershed during the first rain event in 2008 and 2009 discussed in a) above. In addition, water from the retention pond receiving runoff from the NT watershed was also collected in 2008. Large batches (~600 L) were collected from the NT discharge flume and directly from the retention pond. This water was placed in 800 L circular tanks (mesocosms), and aerated. In 2008, the flume collected water was diluted 5 times due to high levels of ammonia caused by the applied litter. Ammonia was sufficiently low in 2009 flume collected water that dilution was not necessary. Adult male fathead minnows (Pimephales promelas) were loaded into these mesocosms and exposed for various time periods to examine effects to aquatic biota. The fish were screened for both vitellogenin (Vtg) mRNA gene expression and blood plasma Vtg. Water from the mesocosms was sampled periodically to investigate temporal changes in estrogens and estrogenicity.
Results and Discussion: Estrogen concentrations and fathead minnow Vtg analyses for 2008 and 2009 mesocosm studies are given in Figures 2 4. Initial E2 concentration was low in the 2008 NT flume runoff mesocosm and dropped below detection within 4 days (Figure 2). The E1 concentration, however, started at ~ 6 ng/L, climbed to > 30 ng/L, and declined slowly over the next week remaining over 5 ng/L at 16 d. Estrogenicity mirrored this trend, climbing for several days before declining slowly over the entire sample interval. As E1 is a less potent estrogen than E2, it is not surprising that EEQ values were lower than measured E1 concentrations. Significant increases in Vtg gene expression were found in fish exposed from Day 0 to Day 4. Gene expression was elevated in some individual fish from other exposure intervals (4 8 D, 0 9 D) but highly variable within treatments and not found to differ from controls. Likewise, several individual fish had elevated plasma Vtg levels but responses within treatments were variable and did not differ from controls.
In the 2008 retention pond mesocosm initial E2 of 1.8 ng/L climbed slightly on day 2 before falling below detection on day 4 (Figure 3). Estrone remained fairly constant through day 8. Estrogenicity followed the trend seen in the flume mesocosm, increasing over several days then slowly decreasing over the remaining interval. Unlike the flume mesocosm, peak EEQ (Day 4) actually exceeded E1 and E2. Measured estrogens were insufficient to explain estrogenicity. A dose response in Vtg gene expression occurred with responses from fish exposed 4 8 d and 0 9 d differing significantly from controls. Plasma Vtg was slower to respond with elevated levels only encountered in the 0 9 d exposed fish.
Results from the 2009 flume mesocosms indicated values similar but somewhat lower than those from 2008 (Figure 4). 17β-estradiol was low but remained detectable for the entire 40 d study interval. The E1 concentration started at 4.2 ng/L, climbed to 9.8 ng/L and declined slowly over the next several weeks. Estrogenicity mirrored this trend, starting at 2.2 ng/L, climbing to 6.3 ng/L at day 4 and decreasing slowly but remaining detectable throughout the 40 d study interval. No significant increases in Vtg gene expression were found in fish exposed during 0 7 d or 7 14 d intervals. Significantly elevated plasma Vtg was detected in fish exposed 0 7 d (samples from 7 14 d were not analyzed).
Results from 2008 and 2009 suggest that there are a number of dynamic processes occurring. At a minimum, conjugated E1 is becoming deconjugated and free E1 and E2 are degrading. This explains the rapid increase in E1 during the first few days followed by the longer, slower decrease. During the first few days there is an abundance of conjugated E1 relative to free E1 so deconjugation exceeds degradation leading to an increase in E1. Later, as conjugated E1 is depleted, degradation exceeds deconjugation so we see a decline in E1. The same processes may be happening with E2 but low initial concentrations near detection limits make trends difficult to discern. These processes are both the result of microbial action. Therefore, differences in initial microbial communities within runoff (NT flume - soil dominated) and receiving bodies (NT retention pond - soil and aquatic constituents) might explain differences in measured estrogens and estrogenic activity.
c) A total of 92 1st-order stream sites were selected for sampling for estradiol from five watersheds on the Delmarva Peninsula in 2004 for a project funded by the Harry R. Hughes Center for Agro-Ecology, Inc. These watersheds included the Pocomoke River, Choptank River, Nanticoke River, Marshyhope Creek and Wicomoco River. All are in areas of high poultry litter usage. Approximately 60% of the sites sampled had detectable estradiol concentrations using an RIA analytical method (≥ 18 ng/L). Resident adult fish and frogs were examined from 15 Delmarva sites for evidence of endocrine disruption in the late summer/fall of 2005 and the spring/summer of 2006. Sites with the highest estradiol concentrations and or multiple detections of estradiol from the 2004 survey were chosen as sample sites. A total of 12 fish and 3 frog species were captured by electrofishing. In 2005 and 2006, 397 and 82 individual organisms were collected, respectively, including 12 male largemouth bass (Micropterus salmoides). Endpoints for endocrine disruption were Vtg in male fish blood plasma or intersex in male fish (testicular oocytes). No Vtg was detected in any species sampled. Only one species, the largemouth bass, showed any evidence of testicular oocytes. Of the 12 male bass sampled, only 2 had testicular oocytes (17%) and the severity of intersex was quite low, especially compared to earlier intersex reported by Dr. Blazer (USGS) in the Potomac and Shenandoah Rivers. Due to the relative scarcity of largemouth bass in small order streams on the Delmarva Peninsula, it was decided to sample lakes and larger rivers in 2008 and 2009. These lakes act as drainage areas/sinks for water from large agricultural areas that receive poultry litter as fertilizer. Sampling for bass was conducted using an electrofishing boat. The watersheds of these lakes and rivers are currently being analyzed to establish land-use patterns (i.e., agriculture vs. urbanized).
Results and Discussion: During 2008, approximately ten male fish were collected from each of six Delmarva lakes (Table 2). Collection activity occurred during late spring (MD sites) and early summer (DE sites). Occurrence of testicular oocytes ranged from 33% to 100% with a mean across all sites of 63% (37/59). Severity, estimated by number of oocytes/histological section, was generally very low. In most fish, only one or several immature oocytes were found within approximately 12 examined testes sections. The most severe case was in a fish from Tuckahoe Lake with 34 oocytes encountered. Efforts are underway to reconcile oocytes enumeration methods to generate TO severity indices according to the methods of Blazer et al. (2007).
Table 3 reports results of 2009 largemouth bass collections from Tuckahoe Lake and from the Pocomoke River in May (pre-spawn) and again in August (post-spawn). Occurrence of intersex varied in Tuckahoe Lake fish from 0% (May) to 33% (August) but was constant at 20% in the Pocomoke River fish collected on both sample occasions. Overall occurrence was 18% (6/34), less than 1/3 the frequency encountered in 2008. Severity was also low with no fish having more than 4 oocytes detected.
d) In 2009 and 2010, we cooperated with Dr. Joseph Love of the Maryland Department of Natural Resources to sample gonads from male largemouth bass tournament mortalities for testicular oocytes. This was in conjunction with the Departments ongoing Largemouth/ Smallmouth Bass Tournament Survey designed to help characterize the bass populations in the State. Sample containers filled with formalin were provided to DNR personnel. Various tissues of recently dead fish were sampled. In conjunction with their normal tissue sampling, DNR personnel removed the gonads from male bass, put them in the sample container for preservation and returned them to Dr. Yonkos for histological examination for testicular oocytes.
Results and Discussion: In 2009, 20 fish were examined and no testicular oocytes were found. These fish were caught in the Potomac River and in Mattawoman Creek, a tributary of the Potomac, during the summer. In 2010, 24 tournament mortality fish caught in the Potomac River during spring tournaments were examined. Testicular oocytes were only found in one fish (4%).
The advantage of examining tournament mortality fish is that information can be gathered from an otherwise wasted resource. Research effort and cost are transferred to tournament participants reducing cost and man-power requirements to the project. The disadvantage is that efforts lack experimental control. Specific areas are not targeted and animals succumbing to fishing stress (i.e., tournament mortalities) may not representative of overall populations. In addition, fish collected are generally larger and therefore older, which may also not reflect overall population trends. Examination of tournament mortality fish for TO may be a useful preliminary tool for selecting river systems to target for further investigation, but is probably not adequate in-and-of- itself to assess the health of fish populations within those systems.
e) In the summer of 2010, a laboratory experiment was conducted to examine the estrogenic effects of three different poultry litters on fathead minnows. The litters were from Perdue, Allens and Tysons, the major poultry integrators on the Delmarva Peninsula. One of the litters tested was the Perdue litter that was applied to the WREC watersheds to conduct the 2010 tillage management experiments discussed in a) above. Aqueous preparations of each litter were prepared and added to large aquaria. Initial 20 L solutions of each litter were mixed in aged aerated well water at 2.5 g/L. These solutions were then diluted 5 fold to achieve 100 L final aqueous exposure treatments containing soluble/suspendable contaminants at 0.5 g litter/L. Resulting litter exposure concentrations were similar to those from previous laboratory and mesocosm experiments and , based on measurement of ammonia, compared well with concentrations seen in runoff studies discussed in a) above (less than 2008 runoff, greater than 2009 runoff). As with the 2008 NT runoff mesocosm study, dilution was necessary to ensure that ammonia concentrations were below lethal levels. Litter solutions were allowed to equilibrate overnight (16 h) before adult male fathead minnows were added. The experiment was run static without treatment renewal to better mimic conditions in a natural system. Water samples were taken periodically over 28 days to track changes in estrogen concentrations over time (E1, E2, and EEQ). Samples were filtered and distributed to various collaborating laboratories for estrogen and estrogenicity measurement. Fish were sacrificed at Day 9 for examination of plasma Vtg and Vtg mRNA. Initial estrogen analyses (GC/MS/MS) have been completed as have measurement of blood plasma Vtg and VtgG mRNA. Estrogen samples are currently being analyzed by LC/MS/MS and RIA. Estrogenicity (EEQ) samples are currently being analyzed using an in-vitro estrogen-inducible reporter-gene assay.
Results and Discussion: Based on GC/MS/MS estrogen analyses, all three litter samples produced the same response as that seen in runoff mesocosm studies from 2008 and 2009. Initial E1 levels were low but increased substantially before decreasing over the remaining study interval (Figure 5). Concentrations of E2 followed the same pattern but to a lesser degree reflecting lower initial concentrations. Based on the Perdue litter treatment (which was subsampled at frequent intervals) the time to peak E1 was Day 6, somewhat later than in mesocosm studies. This may reflect differences in microbial communities between field runoff and laboratory generated litter treatments. Additional chemical analyses and in vitro estrogenicity measurements are not yet available.
All three litters induced Vtg gene expression and protein production in male fathead minnows. Tysons, with the highest measured estrogens of the three litters, induced the greatest vitellogenic response. Perdue and Allens had similar responses despite apparently higher estrogen levels in the Allens treatment. Additional chemical analysis and results from estrogenicity assays may help explain this disparity.
2) There has been no change of key personnel involved in the project. The only personnel changes have been addition of the EPA collaborators described above.
3) In the summer of 2010, a No Cost Amendment was issued for this Cooperative Agreement. The no cost amendment allowed for collaboration with a recently awarded Harry R. Hughes Center for Agro-Ecology, Inc (USDA) contract to study the fate and effects of antibiotics in poultry litter. This new contract payed for the application of litter to the WREC experimental watersheds to study the effects of tillage practices on antibiotic runoff. Because of this extra litter application, scientists at WREC and our EPA collaborators on the CAFO cooperative agreement were able to study an additional season of steroid runoff from these fields. Results from these extra studies are covered in the 2010 portions of 1(a) and 1(b) above. In addition, this new Agro-Ecology funding allowed further laboratory experiments to study changes in antibiotic concentrations and antibiotic resistance over time. Our CAFO group was able to use this same experiment to further study degradation of steroids and changes in estrogenicity over time, and differences in Vtg responses in fish from different poultry litters. Preliminary results from these laboratory studies are covered in 1(e) above.
4) The major objectives of the original project have been completed. As outlined in 3) above, additional CAFO steroid work has been completed under a No Cost Amendment. This work was conducted in conjunction with a new Harry R. Hughes Center for Agro-Ecology, Inc (USDA) project funded in 2010 to study the environmental transport and fate of antibiotics from poultry litter used as fertilizer and the effects of different tillage practices on antibiotic runoff.
Results for Years 2 and 3 are presented in Section 1: a) through e) above. In summary, results indicate: (a) Tillage practices can have a dramatic effect on runoff of water-soluble contaminants from poultry litter amended watersheds. The Turbo-till method, which buries some of the litter after application and roughens the soil surface, reduced steroid concentrations (Estrone - E1 and 17-β estradiol E2) and estrogenicity (EEQ) in runoff compared to runoff from a No-till field. This Turbo-till practice also reduced the runoff of nitrogen and phosphorous compared to No-till application of litter; (b) Results from laboratory microcosms containing runoff from poultry litter amended fields suggest that there are a number of dynamic processes occurring that effect E1, E2, and Vtg in fathead minnows. At a minimum, conjugated E1 is becomes deconjugated and free E1 and E2 degrades. These processes are both the result of microbial action. Changes in the relative concentrations of E1 and E2 over time due to this microbial activity might explain differences in measured estrogens and, therefore, estrogenic activity since E2 is much more bioactive and causes greater induction of Vtg in fish than E1; (c) In 2008, occurrence of testicular oocytes in male largemouth bass from six lakes and ponds on the Delmarva Peninsula ranged from 33% to 100%, with a mean across all sites of 63%. Severity, estimated by number of oocytes/histological section, was generally very low. This is the first finding of intersex on the peninsula. Resampling of bass in 2009 from Tuckahoe Lake showed the occurrence of intersex varied from 0% (May) to 33% (August). Sampling from a large river on the peninsula (Pocomoke River) showed a constant occurrence of intersex (20%) in bass collected on both sample occasions. Severity was also low as in 2008, with no male fish having more than 4 oocytes detected; (d) In 2009, 20 male largemouth bass from fishing tournaments were examined and no testicular oocytes were found. These fish were caught in the Potomac River and in Mattawoman Creek, a tributary of the Potomac. In 2010, 24 tournament mortality fish caught in the Potomac River during spring tournaments were examined. Testicular oocytes were found in 4%; (e) In 2010, laboratory experiments using different poultry litters and fathead minnows produced similar responses to those seen in runoff mesocosm studies from 2008 and 2009. Initial E1 levels were low but increased substantially before decreasing over the remaining study interval. Concentrations of E2 followed the same pattern but to a lesser degree reflecting lower initial concentrations. Based on the Perdue litter treatment, which was subsampled at frequent intervals, the time to peak E1 was Day 6, somewhat later than in mesocosm studies. This may reflect differences in microbial communities between field runoff and laboratory generated litter treatments. All three litters examined in 2010 induced Vtg gene expression and blood plasma protein production in male fathead minnows. Tysons, with the highest measured estrogens of the three litters, induced the greatest vitellogenic response. Perdue and Allens had similar responses despite apparently higher estrogen levels in the Allens treatment.
All activities proposed in the original cooperative agreement have been completed. In addition, additional CAFO steroid work proposed in conjunction with the new Harry R. Hughes Center for Agro-Ecology, Inc. (USDA) antibiotics project funded in 2010 has been mostly completed. Estrogen samples from these 2010 experiments are currently being analyzed by LC/MS/MS and RIA for comparison with the GC/MS analyses already completed. Estrogenicity (EEQ) samples are currently being analyzed using an in-vitro estrogen-inducible reporter-gene assay. These EEQ samples will be compared to the estrogen analyses to determine relationships between the change in estrone and 17-β estradiol concentrations over time and estrogenicity. Estrogenicity in fish as measured by Vtg induction will be compared to the EEQ results to determine the relationship between an aquatic endpoint and a human cell line endpoint.
Blazer VS, Iwanowicz LR, Iwanowicz DD, Smith DR, Young JA, Hedrick JD, Foster SW, Reeser SJ. Intersex (testicular oocytes) in smallmouth bass Micropterus dolomieu from the Potomac River and selected nearby drainages. Journal of Aquatic Animal Health 2007;19:242-253.
No journal articles submitted with this report: View all 9 publications for this project
Progress and Final Reports:
2008 Progress Report
2009 Progress Report
2011 Progress Report