In Vitro Cellular & Developmental Biology - Animal

, Volume 48, Issue 4, pp 216–228

E-Screen evaluation of sugar beet feedstuffs in a case of reduced embryo transfer efficiencies in cattle: the role of phytoestrogens and zearalenone

Authors

    • USDA, Agricultural Research ServiceRed River Valley Agricultural Research Center, Biosciences Research Laboratory
  • M. S. Mostrom
    • Veterinary Diagnostic Services DepartmentNorth Dakota State University
  • E. M. Lenneman
    • USDA, Agricultural Research ServiceRed River Valley Agricultural Research Center, Biosciences Research Laboratory
    • Formerly of North Dakota State University
Article

DOI: 10.1007/s11626-012-9489-9

Cite this article as:
Shappell, N.W., Mostrom, M.S. & Lenneman, E.M. In Vitro Cell.Dev.Biol.-Animal (2012) 48: 216. doi:10.1007/s11626-012-9489-9

Abstract

The E-Screen assay was used to evaluate the estrogenicity of sugar beet by-products obtained from a dairy farm experiencing low success rates of embryo transfer. The beet tailings had ~3-fold the estradiol equivalents of the pelleted beet pulp (3.9 and 1.2 μg estradiol equivalents or E2Eq/kg dry matter, respectively). Whole sugar beets, sugar beet pellets, and shreds from several Midwest US locations were also evaluated by E-Screen. All pellets examined were found to have some estrogenic activity (range ~0.1–2.0 μg E2Eq/kg DM) with a mean of 0.46 μg/kg dry matter and median of 0.28 μg/kg dry matter. Relative E2Eq ranked as follows: pellets > shreds > most unprocessed roots. Using recommended feeding levels and conservative absorption estimates (10%), the estrogenic activity in the original samples could result in blood estradiol equivalents ≥ those found at estrus (10 pg/mL, cows). Chemical analyses revealed no known phytoestrogens, but the estrogenic mycotoxin, zearalenone, was found in 15 of 21 samples. Of significance to those using the E-Screen are our findings that contradict previous reports: ß-sitosterol has no proliferative effect and genistein’s glucuronidated form—genistin—is equal to genistein in proliferative effect. The latter is the result of deconjugation of genistin to genistein in the presence of fetal bovine serum (determined by LC MSMS). These data show the usefulness and caveats of the E-Screen in evaluation of feedstuffs, and indicate a potential for sugar beet by-products to contain zearalenone at concentrations that may impact reproduction.

Keywords

MCF-7EstrogenicitySubfertilityZearalenoneSugar beets

Introduction

Our laboratory used the E-Screen, an in vitro assay originally developed to evaluate the estrogenicity of various pure chemicals (Soto et al., 1995), on samples that had been submitted to the North Dakota State Veterinary Diagnostic Services Laboratory from a Minnesota dairy farm experiencing low embryo transfer success rates. Since the assay’s development, the matrices evaluated for estrogenicity have expanded to include environmental water samples, animal wastes, and estrogenic vegetables such as soybeans (Shappell, 2006; Shappell et al., 2010, Zhang et al., 2007). The submitted feed samples were sugar beet tailings (unprocessed beet parts and associated dirt, discarded by the processing plants, but available for retrieval by farms for animal feed) and pelleted beet shreds (the dried beet pulp material post-sugar extraction).

Review of the literature for reproductive effects associated with feeding of sugar beets led to a publication that reported estrogenic compounds in sugar beet leaves ascribed to ß-sitosterol by thin layer chromatography, infrared spectroscopy, and melting point analysis (Elghamry et al. 1971). Estrogenic activity was assessed by measurement of uterine weights in immature female mice 24 h after subcutaneous injections of leaf extracts. Elghamry et al. (1971) cites four studies dated 1958 to 1969 (three in non-English journals) reporting estrogenic activity in sugar beet plants or silage, and subfertility and genital tract abnormalities in cattle fed these materials. One study (Rigelnik, 1968) reported that feeding cows greater than 35 kg of sugar beet silage per day caused a 30% increase in sterility. With this background information, samples from the case of subfertility submitted to the Veterinary Diagnostic Laboratory were also analyzed for ß-sitosterol, selected common phytoestrogens, and mycotoxins. As submitted samples were found to be estrogenic by E-Screen, sugar beet by-products were collected from several Midwestern regions to evaluate if estrogenicity was unique to the original submitted samples or a more widespread occurrence.

Materials and Methods

Samples.

Sugar beet tailings and pellets collected from the farm with embryo transfer failures were from a 2009 harvest that had been stored in unsheltered, ground bunkers and submitted to the Veterinary Diagnostic Services Department at North Dakota State University in March of 2010. Additional samples consisting of “shreds” and “pellets” (dried sugar beet pulp post-sugar extraction, pelletized in the later case) were collected early in the 2010 sugar beet harvest (September 29, 2010 to October 7, 2010) from regional sugar beet processing companies (Red River Valley of ND and MN, 6 plants designated “1” to “6”; and Saginaw Bay, MI area, 4 plants designated “9” to “12”). Sample sizes are reported in the Appendix Table 1. Mean pellet sample size was 1.4 ± 0.69 kg S.D. (from nine locations), with the two smallest being 0.53 and 0.60 kg. Shred samples were 4.0 ± 0.43 kg S.D. (from three locations). Unprocessed beets were obtained from fields in the Red River Valley (September 25, 2010 in ND, 4 samples, near plant #2; and Novermber 1, 2010 in MN, 3 samples, between plants #3 and #6).

Biochemical analyses.

Only the original beet tailing sample associated with reproductive failures was analyzed for ß-sitosterol by Covance Laboratories, Madison, WI (AOAC Method, 994.10, 2000). Isoflavone analysis of all samples was performed by Iowa State University—Food Science & Human Nutrition Laboratory, Ames, IA, and included: daidzein, genistein, glycitein, as well as their malonyl and acetyl derivatives; biochanin A, formononetin, and coumestrol. Samples were solvent extracted in duplicate and analyzed by LC HPLC (Murphy et al., 1999). Detection limits ranged from 2 (genistein) to 10 (glycitein) nmoles/g dry matter. In addition to isoflavones, samples were analyzed for mycotoxins (vomitoxin, zearalenone, aflatoxin B1, fumonisins, and 15 additional trichothecenes) using the GC/MS method described by Groves et al. 1999. Practical limits of quantitation were 0.5 mg/kg on dry material for all mycotoxins except aflatoxin B1 (0.02 mg/kg) and fumonisin (2.0 mg/kg).

Beet and beet by-product extraction for E-Screen.

Dry matter determination was performed on pellets, shreds, and tailings and whole beet samples at 60°C. Dried samples were ground using a Stein Mill (Seedburo Equipment Company, Chicago, IL), sieved with a US #12 sieve, 1.7 mm, and extracted essentially as described by Bajer et al. (2007) for isoflavinoid analysis. Briefly, 1 g of dried sample was sonicated (47kHZ, Bransonic Ultrasonic Sonicator, Model 3210, Danbury, CT) in 25 mL of 60% acetonitrile, at 60°C for 1 h, with shaking every 10 min. Particulate material was removed by centrifugation (4°C for 15 min at 2,100×g) followed by passage of the supernatant through solvent-rinsed glass wool. After evaporation of the acetonitrile under a stream of N2 at 20°C, the aqueous portion was diluted to 50 mL with nanopure water. Aqueous extracts were concentrated on solid phase extraction columns (SPE, OASIS HLB, 200 mg, 5 cc, Waters, Milford, MA). Columns were sequentially eluted using 3 mL each of acetone, methanol, acetonitrile, ethyl acetate, methylene chloride, and tert-butyl-methyl ether. The combined solvent fractions were evaporated under nitrogen (37°C) and dried residues were stored at −20°C until resuspension for E-Screen analysis. Extraction efficiencies were evaluated initially by fortifying pellets and tailings with ß-sitosterol (357 μg/g dry matter), and later with either estradiol (20 ng/g dry matter) or zearalenone (1.91 μg/g dry matter) as ß-sitosterol displayed no estrogenic activity in the MCF-7 cell proliferation assay.

E-Screen.

The MCF-7 BOS, estrogen-dependent cell line (derived from a human mammary epithelial carcinoma, graciously provided by Drs. Ana Soto and Carlos Sonnenschein, Tufts University, Boston, MA) was used to evaluate beet sample extracts for estrogenicity relative to 17 ß-estradiol (17 ß-E2, Sigma Chemicals, St. Louis, MO) as previously described (Shappell 2006). Briefly, extracts from beet and beet by-product extracts were resuspended in water and diluted in cell culture media. Twenty-four hours post-plating, steroid-containing medium (fetal bovine serum, FBS) was removed from MCF-7 BOS cells and replaced with steroid-free medium (no phenol red and charcoal dextran-stripped, CD-FBS) containing diluted extracts. Cells were incubated for 5 d, fixed with trichloroacetic acid, stained for protein with sulforhodamine B, solubilized in buffer, and absorbance read at 490 nm. Estradiol equivalents were determined based on a regression analysis of the 17 ß-estradiol curve from the same experiment. Whole beet, pellet, and tailing extracts were tested over a wide range of dilutions, and linear ranges were found at 1:1,000 to1:5,000 of 0.5 g DM extracted into a final volume of 200 uL. Standard curves of pure chemicals were prepared from stocks dissolved in ethanol followed by dilution in cell culture media. Each chemical was tested in a minimum of three separate experiments. Limits of quantitation for estrogenicity were ~14 pg E2Eq/g DM. Specificity of estrogenicity in positive samples was confirmed by E2-receptor antagonist ICI 182,780 (Tocris, Ellisville, MO) as described by Rassmussen and Nielsen 2002.

Examination of genistin purity and stability.

In our laboratory, the estrogenicity of genistein (Stein) and genistin (Tin, the 7-D glucoside of genistein) were very similar, which is in contrast to data on estrogenicity of conjugated estrogens published by Gadd et al. (2010). Therefore Stein (synthetic, 99.1% purity by HPLC, Sigma Chemicals) and Tin (soy extract, >99.5% purity by HPLC, LC Laboratories, Woburn, MA) were analyzed for impurities or breakdown products by UHPLC TQD MS2 (Appendix). Because the presence of impurities did not account for the similar estrogenicity of Stein and Tin, experiments were performed to determine if Tin was hydrolyzed to its aglycone, Stein, under various conditions. Sterile incubations included: genistin (3 × 10−8 M) in buffer (conditions for enzymatic deconjugation—0.6 M sodium acetate, pH 4.5, room temperature, T0 versus T24h); incubated in nanopure water, E-screen control media ± steroid-stripped fetal bovine serum and with and without cells for 0 and 5 d at 37°C (Table 2). Incubation conditions were scaled up from 96 well plates to T25 flasks, using 2.34 × 105 cells and 5 mL of media. At the end of 5 d, media were collected and cells were rinsed with PBS (pH 7.4, 37°C) followed by lysis in ice-cold nanopure water. Cells were removed from media samples by centrifugation (500×g) and added to cell lysates, which were maintained at −80°C until analysis. Post-incubation media (with and without cells, but without genistin) were spiked with genistin at harvest for assessment of potential post-harvest degradation and processing/extraction losses. Samples were processed by SPE as described above, followed by UPLC TQD MS2.

Statistical analysis.

The sample size of sugar beets and their by-products in this study was small and results of statistical analysis should be regarded with caution. Using an ANOVA, differences in mean estrogenicity (E2Eq) across sample type (whole beets, shreds, and pellets) were determined. Because the variances were unequal across sample types, Proc Mixed was used and separate variance parameters were estimated for each sample type.

Regional differences in E2Eq or zearalenone of pellets were evaluated by t test with unequal variances. A regression analysis was performed across all sample types to test for a relationship between zearalenone and E2Eq using REG, Model 1.

One half of effective concentration (EC50) values of compounds was compared using a mixed model analysis of variance with experiment as a random effect in the model. Chemicals were tested in a minimum of four separate experiments. Data were analyzed using Proc Mixed, allowing unequal variances between compounds. Tukey’s contrasts were used for post-hoc comparisons of means. All statistical analyses were done using SAS V9.2 (SAS Institute, Inc., Cary, NC).

Results and Discussion

E-Screen analysis of beets and by-products.

The beet tailing and pulp pellet samples submitted to the NDSU Diagnostic Laboratory were estrogenic in the E-Screen assay. In triplicate extractions, tailings contained ~3 times the estrogenicity of beet pulp pellets (3.9 ± 1.27 versus 1.2 ± 0.16 ng E2Eq/g DM, respectively). The greater variability of the tailing samples is to be expected as it is a heterogeneous sample in comparison to the pellets. E-Screen data from the follow-up samples collected in the 2010 fall harvest are presented in Fig. 1, grouped by geographical region of collection. Estrogenicity of beet pulp pellets ranged from 0.09 to 1.98 ng E2Eq/g DM. It was not possible to obtain both pellet and shred samples from most processing facilities, as most facilities are dedicated to the production of pellets or shreds, but not both. Of the shred samples obtained from 3 different facilities, two had almost identical estrogenic activity (0.113 ± 0.0215 and 0.114 ± 0.0137 ng E2Eq/g DM) even though one came from the Red River Valley, ND and one from the Saginaw Bay MI region. The third sample contained roughly double the others (0.217 ± 0.0303 ng E2Eq/g DM). While one beet sample had estrogenic activity similar to the shreds sample (0.233 ng E2Eq/g DM), the other beet samples contained levels lower than the limits of quantitation (values < 0.014 ng E2Eq/g DM were recorded as 0). For statistical analysis of the very limited data set, one pellet sample was an outlier pellet (1.98 ng E2Eq/g DM) and therefore statistical inferences discussed immediately below were made without the outlier. Pellets had higher estrogenic activity than shreds (p < 0.05) and whole beets (p < 0.001); while shreds tended to have more estrogenicity than whole beets (p = 0.09). A trend (p = 0.10) for regional differences in estrogenicity of products was found by analysis of the t test with Satterthwaite’s unequal variance. Mean extraction efficiency for estradiol was 123% based on expected E2Eq of fortified pellets (samples obtained from processing plants and selected due to lack of estrogenic activity). E-Screen was inadequate for assessment of ß-sitosterol extraction efficiency, as the compound was not estrogenic in the assay. Because later chemical analysis indicated the estrogenic mycotoxin zearalenone was present in the original samples submitted by the veterinarian to the Diagnostic Laboratory, zearalenone’s extraction efficiency in pellets was tested, and found to be 96%.
https://static-content.springer.com/image/art%3A10.1007%2Fs11626-012-9489-9/MediaObjects/11626_2012_9489_Fig1_HTML.gif
Figure 1.

Estrogenicity of sugar beets and their by-products by region. E-Screen, an in vitro assay that assesses estrogen-dependent proliferation of human mammary carcinoma cells (MCF-7 BOS), was performed on organic extracts of samples. Cells were exposed for 5 d, fixed, protein stained with sulforhodamine B (SRB), solubilized, and absorbance read at 490 nm. Estradiol equivalents were calculated relative to the cellular proliferative response obtained using a 17 ß-estradiol standard curve (1 × 10−13 M to 1 × 10−9 M). The X-axis represents individual processing plants, with the exception of 7, where whole sugar beets were taken from a field located ~15 miles of plant 2, and 8, a field located between plants 3 and 6 in MN. Samples with no estrogenicity were plotted as negative values. Pellets versus shreds versus unprocessed roots; letters above bars indicate difference at p < 0.05, b’ difference, p = 0.09.

Phytoestrogen evaluation by E-Screen

β-sitosterol. Literature dating back to the 1960s and 1970s report impaired reproductive performance associated with the use of sugar beets as a feed source and ß-sitosterol was proposed as at least one of the chemicals responsible (Elghamry et al. 1971). As cited by Elghamry et al. (1971), German scientists documented abnormalities in the genital tracts of cows fed large quantities of sugar beet leaves and cystic ovaries when fed silage. Therefore we tested ß-sitosterol at concentrations of 1.5 × 10−5 to 1 × 10−8 M, and we found no proliferative activity. Upon repeating the experiments with β-sitosterol purchased from another supplier (both synthetic and extracted from soybeans), we measured the same results—no activity or proliferative response in MCF-7 cells (Fig. 2 and Table 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs11626-012-9489-9/MediaObjects/11626_2012_9489_Fig2_HTML.gif
Figure 2.

E-Screen comparison of various chemicals. In order to validate estrogenic responsiveness of the MCF-7 cells in our laboratory, known phytoestrogens and the mycotoxin zearalenone, standard curves of pure chemicals were tested using the E-Screen described in Fig. 1. Proliferation/estrogenicity is proportional to the increased protein reflected by increased SRB absorbance (A 490 nm). Each compound was tested in a minimum of three separate experiments. This graph represents typical experiments, with error bars representing standard deviations (n = 5 wells).

Table 1.

Structure and estradiol equivalents

https://static-content.springer.com/image/art%3A10.1007%2Fs11626-012-9489-9/MediaObjects/11626_2012_9489_Tab1_HTML.gif

aThese compounds were not found in original submitted samples, with the exception of β-sitosterol, present at 357 μg/g dry matter, and zearalenone 0.24 and 0.6 ug/g dry matter

bRelative estrogenic activity on a molar basis, with estradiol equivalents (E2Eq) for estradiol = 1. In our laboratory, all compounds except genistein and genistin had different relative E2Eqs, <0.05, by Tukey contrasts

cSoto et al. 1995 by coulter counter

dWelshons et al. 1990 by DNA

eGutendorf and Westendorf 2001 by MTT (typographical error in table, compound cited as genistin, but only genistein was tested)

fMorito et al., 2001 by SRB

gMatsumura et al. 2008 by coulter counter

hSitosterol literature: Gutendorf and Westendorf (2001) by MTT, Ju et al. (2004) by MTT, and Mellanen et al. (1996) by MTT

Literature reports on the estrogenicity of ß-sitosterol have been inconsistent. Using soybean-derived ß-sitosterol Ju et al. (2004) reported that concentrations of 1 to 150 μM of ß-sitosterol stimulated proliferation of MCF-7 cells using the MTT assay (99% purity by HPLC, source identified as African potato extract by Anthony Almada, ImagiNutrition, Laguna Niguel, CA, personal communication). Gutendorf and Westendorf (2001) also reported proliferation of MCF-7 cells with ß-sitosterol treatment using the MTT assay. In contrast, Mellanen et al. (1996) reported ß-sitosterol to have no estrogenic activity with MCF-7 cells; and yet T-47-D cells (an estrogen-independent breast cancer cell line) were responsive at 1 μM, and male juvenile rainbow trout produced vitellogenin when fed ß-sitosterol. Estrogenic impurities of β-sitosterol used by Mellanen et al (1996) cannot be discounted though because purity was cited as only 91%. Another group using MCF-7 cells found ß-sitosterol inhibited cell proliferation at 10 nM and 1 μM (16% inhibition and 38% inhibition, respectively; extracts from a Malaysian herbaceous plant, purity not reported, Chai et al. 2008). Awad et al. (2007) also reported inhibitory response in MCF-7 cells with 8 or 16 μM synthetic ß-sitosterol (95.7% purity by HPLC, Sigma Chemicals, Peter Bradford, SUNY, Buffalo, NY, personal communication), as well as an increase in apoptotic enzyme activity (caspase-8) and cell surface death receptor (Fas). It is unclear why ß-sitosterol has been reported as having both inhibitory and proliferative effects on MCF-7 cells at similar concentrations. Two possibilities are (1) impurities of test compounds or (2) problems with the MTT assay, a spectrophotometric assay that depends on reduction of a colorimetric substrate, MTT, by succinate dehydrogenase. All reports of proliferative, or estrogenic, responses to β-sitosterol in MCF-7 cells used the MTT assay. For the MTT assay to be valid, a linear relationship between enzymatic activity and cell number must exist. This relationship can be altered if test compounds result in metabolic shifts that decrease the available subtrate (succinate). This situation has been implicated when DNA and cell number fail to correlate with MTT values (caveats of the MTT assay were discussed previously by Shappell, 2003).

Genistin vs genistein, purity, and genistin stability testing. In our laboratory, standard curves of genistein and genistin repeatedly yielded similar (p > 0.05) E2Eq responses in the E-Screen assay (Fig. 2, Tin 4 experiments, Stein 6 experiments). The purity of the genistein and genistin standard was examined as even a 1% impurity of Stein in the Tin standards, could account for the elevated estrogenicity of Tin. Stocks of synthetic Stein and Tin derived from soy product were analyzed by mass spectrometry (UPLC TQD MS2) in order to assess if impurities in each preparation were present. Analysis of a 1,000 pg/μL Tin standard indicated that it was essentially pure, with a molar ratio of ~11,600 moles Tin to 1 mole Stein (Fig. 3a). In contrast, the Stein standard contained approximately 6% Tin on a mole basis (Fig. 3b).
https://static-content.springer.com/image/art%3A10.1007%2Fs11626-012-9489-9/MediaObjects/11626_2012_9489_Fig3_HTML.gif
Figure 3.

Purity evaluation of genistin and genistein. In contrast to literature reports and expectations for conjugated versus unconjugated estrogenic compounds, our laboratory found genistin and genistein to have equivalent estrogenicity by E-Screen. Therefore the compounds were tested for purity by UPLC TQ MS2. At 100 pg/μL injection volumes, genistin (Tin) was found to be present in genistein (Stein); yet even at 10-fold higher injection concentrations (1,000 pg/μL), there were only traces of Stein (a). The retention time of Stein is 0.93 min, m/z = 269.1, and total ion count abundance was 30 (left top panel, based on two confirmatory ion fragments) while Tin from the same injection with a retention time of 0.5 min, m/z of 431.0, and had a total ion count abundance of 109,344. In contrast, the 1,000 pg/μL injection of Stein found Tin to be present at approximately 10% of the total ion count abundance of Stein (10,823 versus 126,140). Therefore the E-Screen activity seen when cells were exposed to the conjugated form of Stein (Tin) could not be explained by the presence of a large Stein impurity.

Because the Stein impurities in Tin could not account for the E-Screen results, further experiments were conducted to determine its stability under cell culture conditions, the results of which are presented in Table 2. Initially, Tin was incubated in the absence or presence of cells in complete media (including 10% charcoal-dextran stripped fetal bovine serum, CD-FBS) at 37°C for the 5 d period required for E-Screen testing; in buffer (24 h room temperature); or added to conditioned media (5 d exposure to cells) just prior to extraction. Tin remained unchanged in incubations of buffer (24 h), but when Tin was fortified in conditioned complete culture media just prior to extraction, trace amounts of Stein were found. When incubations occurred in the presence of cells, almost a complete loss of Tin (~85% loss) was found, with an approximate 75% conversion to Stein on a mole basis. No deuterated internal standard was used, therefore quantitations are approximate. In a subsequent experiment, the effect of CD-FBS on Tin was tested to establish if the serum in the media was responsible for the Tin loss. In these experiments, Tin recovered after 5 d in nanopure water or media at 37°C was ~96% to 127% of that of 0 d and no Stein was detected. In contrast, after 5 d in the presence of CD-FBS (10% serum in water) the molar ratio of Tin/Stein was 0:134 in one incubation and 8:129 in another second incubation (Fig. 4) These results indicate that lability of Tin, in the presence of cell culture media, is likely the result of deconjugation by enzymes present in the FBS. Conversion of Tin to Stein by FBS and/or cells would explain the similarity in E2Eq obtained for the two compounds. These results are less surprising when reviewing literature that reported genistin conversion to its aglycone form in incubations with human saliva or intestinal cell-free extracts from mice (Allred et al., 2001).
Table 2.

Genistin stabilitya in buffer or cell culture media ± MCF-7 cells

Treatments

Tin

Stein

Buffer + Tin, day 0 or day 1

++++

Nanopure water + Tin, day 0 or day 5b

++++

Media without FBS + Tin, day 0 or day 5

++++

Cells + Media for 5 d → + Tin at harvest

+

FBS + Tin, day 0b

++++

FBS + Tin, day 5b

+

+++

Cells + complete media + Tin, 5 db

+

+

Tin genistin, Stein genistein

aSamples extracted by SPE and analyzed by UPLC TQD MS2, triplicate injections, quantitation based on external standards. Because no internal standards were used, proportionality of compounds is represented by the relative number of “+”s. Negative control sample blanks for all combinations were run (no genistin in incubations) to ensure ion fragments were due to genistin and not other compounds. Neither genistin nor genistein was present in negative control samples

bSamples in duplicate

https://static-content.springer.com/image/art%3A10.1007%2Fs11626-012-9489-9/MediaObjects/11626_2012_9489_Fig4_HTML.gif
Figure 4.

Genistin stability. To investigate the possibility of conversion of genistin to genistein under culture conditions, stability experiments were performed. Genistin mixed with cell culture media (including 10% charcoal dextran-stripped fetal bovine serum) and incubated at 37°C for either 0 or 5 d prior to extraction by solid phase extraction was analyzed by UHPLC TQD MS2. See Fig. 3 legend for retention times and m/z for compounds. In one injection of day 0 Tin sample, total ion count abundance of Tin was 2,862, while Stein was 105 (a). After 5 d (b), Tin total ion count abundance was below limits of quantitation of the lowest standard (0.5 pg/μL on column), while the total ion count abundance of Stein was 10,644.

As with ß-sitosterol, reports of relative estrogenic activity of genistein and its glycoside, genistin, are contradictory. While Morito et al. (2001) found Tin had a greater proliferative effect than Stein on MCF-7 cells, the reverse was true for the human estrogen receptor ß (ER-ß), that is Tin had lower affinity than Stein for the receptor. Compounds tested in that study were isolated from soy isoflavones with or without enteric bacteria digestion, and chemical purities were not indicated. Early reports (Peterson and Barnes, 1991) found inhibition of proliferation in MCF-7 cells treated with 18.5 μM Stein and stimulation of proliferation with 9.25 μM Tin. In contrast, Cherdshewasart and Sriwatcharakul (2008) reported that Tin and Stein had similar proliferative effects on MCF-7 cells after a 3-d incubation; again the source and purity of test compounds were not reported. Addition of a S-9 fraction to the phytoestrogens increased the proliferative response of the cells (and to 17ß-E2), but the relative proliferation of Tin and Stein was not different. Similarly, Mayr et al. (1992), reported that MCF-7 cells treated with Tin produced one half the 52 KD 35S-Met labeled protein in SDS-PAGE of Stein as assessed by autoradiography. Typically glycoside conjugates are assumed to have less biological activity than their corresponding aglycones. Gadd et al. (2010) published the relative E-Screen estrogenicity of several glucuronidated estrogens and their unconjugated parent estrogens; the potency of estriol-3-glucruonide was 1% of estriol, and the potency of 17-β-estradiol-3 glucuronide was 0.04% of 17-β-estradiol. It is unclear why genistin is more susceptible to deconjugation than estrogen conjugates in an E-Screen assay. Not only do these findings raise a cautionary flag to users of E-Screen to consider such possibilities, but these findings may have in vivo implications. Genistin absorbed from the gastrointestinal tract into the blood stream could possibly be deconjugated in the blood to Stein, causing a concomitant increase in estrogenic activity.

Daidzein and formononetin.

With the exception of Tin, the relative potency of the other phytoestrogens tested in MCF-7 cells was similar to reports in the literature (Table 1). The order of potency was genistein ~ genistin > daidzein > formononetin. It should always be remembered that relative estrogenicity of compounds is assay-dependent, such that in vivo responses may not reflect in vitro rankings, nor are in vitro rankings from one assay type interchangeable with others.

Analysis of samples for isoflavones and mycotoxins.

No formononentin, known soy or alfalfa/clover isoflavones, or coumestrols were detected in any samples, with the exception of trace quantities of analytes presumed to be daidzin and acetyldaidzin. Concentrations of mycotoxins in beet by-product samples are reported in Table 3. With the exception of the original tailing sample, which contained 0.4 ug g−1 vomitoxin, zearalenone was the only mycotoxin found in the beet by-products. A regression analysis was performed across all sample types from 2010 between zearalenone and E2Eqs. The pellet outlier omitted was omitted as the studentized residual was 12.05 (anything greater than 2 is considered an outlier in a regression). Under these conditions, a significant linear relationship was found with an R2 = 0.53 (p = 0.0006). If the original tailing and pellet samples and the outlier pellet sample are included, the R2 = 0.76 (p < 0.0001). Because values are so low the correlation results should be interpreted with caution. Future studies should evaluate sugar beet by-products with higher levels of zearalenone/mycotoxin contamination to determine if the relationship to estrogenicity by E-Screen holds. Zearalenone E2Eqs in Table 3 are not exact matches for E-Screen E2Eqs. Factors that will result in the disparity include: extremely low values in either analysis inherently contain less precision and extracts from sugar beet by-products will contain a mixture of chemicals potentially exerting both positive and negative effects on the cells in the E-Screen.
Table 3.

Zearalenone, relative E2Eqs, and E-Screen E2Eqs per gram of DM sample

Pellet sample

ZEAa (μg g−1)

ZEA E2Eqb (ng g−1)

E-Screen E2Eq (ng g−1)

Shred sample

ZEA (μg g−1)

ZEA E2Eq (ng g−1)

E-Screen E2Eq (ng g−1)

Whole beet sample

ZEA (μg g−1)

ZEA E2Eq (ng g−1)

E-Screen E2Eq (ng g−1)

Original Pellet

0.24

2.4

1.24

    

Original Tailings

0.6

6.0

3.86

#1 RRV

0.05

0.5

0.27

        

#2 RRV

0.25

2.5

0.40

    

#2a RRV (7)

0.05

0.5

0.23

#3 RRV

0.05

0.5

0.29

#3 RRV

0.05

0.5

0.11

#2b RRV (7)

0

0

0.02

#4 RRV

0.24

2.4

0.42

#4 RRV

0.20

2.0

0.22

#2c RRV (7)

0

0

0.06

#5 RRV

0.40

4.0

1.98

    

#2d RRV (7)

0

0

0.02

#6 RRV

0.05

0.5

0.27

    

#6a RRV (8)

0

0

0

#9 SB

0.20

2.0

0.16

#12 SB

0.05

0.5

0.11

#6b RRV (8)

0

0

0

#10 SB

0.27

2.7

0.28

    

#6c RRV (8)

0

0

0.02

#11 SB

0.05

0.5

0.09

        

aValues from 0.05 to 0.09 μg g−1 zearalenone (ZEA) were below limits of quantitation, but not detection, and therefore they were assigned a value of 0.05

bTo obtain estradiol equivalents from zearalenone, the zearalenone concentrations are multiplied by 0.01, as zearalenone is 1/100th as estrogenic as 17 ß-E2 in the E-Screen

cNumber in parentheses correlates to Region number on X-axis of Fig. 3

Historical feeding of sugar beet and their by-products and adverse reproductive effects.

In the Journal of Animal Science, reports of feeding sugar beet by-products date back to 1915 (Morton 1915), and within the past decade, research has included a feeding study of by-products from Roundup Ready sugar beets (Hartnell et al. 2005). Van der Peet-Schwering et al (2004) found that feeding sugar beet pulp on an ad libitum basis at up to 45% of diet (as fed, not dry matter −DM) to pigs was without detrimental effect on interval of weaning to estrus, return to estrus after 1st insemination, farrowing rate, litter size, and piglet weights over three parities. Growing beef steers fed pressed beet pulp at 20% or 40% of the ration (DM basis) for 21 d had body weights (BWs) that were 95% and 92%, respectively, of BWs of steers fed no beet pulp (Bauer et al. 2007). Pelleted beet pulp replacement of corn in the rations of lactating dairy cows resulted in a reduction in DM intake when fed at 24% of DM, but neither milk production nor composition was negatively impacted (Voelker and Allen 2003). From review of extension publications, feeding of sugar beets or their by-products to horses, beef and dairy cattle, and pigs appears to be an accepted practice (Hagstrom 2008; Extension.org 2008; Buffum and Griffith 1902; Lardy 2006). Cautions are typically mentioned concerning the risk of too high a percentage of dietary intake of beet by-products due to choking hazards of whole or partial beets, requirement for rehydration of shreds for horses, and storage of wet pulp/silage to avoid spoilage.

In contrast to the swine study (Van der Peet-Schwering et al. 2004) which appears to be the only recent study investigating the effect of beet feeding on reproduction, the literature of the 1950–1970s contains several reports documenting negative outcomes on reproduction associated with sugar beet feeding. Elghamry et al (1971) cited Rigenik (1968, thesis unavailable) with respect to the fact that daily feeding of >35 kg of sugar beet silage to cows increased sterility 30%, with increased rates of cystic ovaries and delayed ovulation. Elghamry et al. (1971) found that fractionated sugar beet leaf extracts, when injected subcutaneously into immature female mice, for 3 d resulted in an increase in uterine weights 24 h post-treatment. While estrogenic activity was isolated from two fractions, estrogenicity in one fraction was identified as ß-sistosterol using mass spectroscopic and ultraviolet spectral methods. Injections of 5 μg of crystallized “ß-sitosterol” into a ~10 g mouse resulted in a 55% increase in uterine weight over control animals. A 100 μg dose increased uterine weights by 144% and estradiol equivalents were calculated at 1:8,400 based in injections with estradiol. In a second paper by this group (Grunert et al. 1969), mice were dosed with methanol-extracted sugar beet plant parts, and fresh and dry beet silage, either subcutaneously or orally. Beet roots did not have estrogenicity, but in increasing order, uterine weights increased in mice dosed with extracts of dry silage, petioles, fresh silage, and leaf blades. They also reported a lack of estrogenic effect from beet roots, findings substantiated by our data.

Zearalenone in sugar beets.

The presence of zearalenone on sugar beets has been documented (Bugbee 1982; Bosch and Mirocha 1992; Burlakoti et al. 2008), but is not often considered a problem. This most likely reflects the relatively limited worldwide production of sugar beets compared to other agricultural products on which this mycotoxin is found (corn, soybean, wheat, rice, distiller grains). Fusarium spp., known to produce zearalenone, were isolated from 31% of the rotten sugar beet tissues sampled from a Minnesota harvest (Bugbee, 1982), while the first report of zearalenone associated with sugar beets and beet fibers from Fusarium-infected samples was collected in northwestern Minnesota from 1987 to 1989 (Bosch and Mirocha, 1992). In this report, 31 of 75 fiber samples were positive for zearalenone content, ranging from 13 to 4,650 μg/kg dry weight. Our findings of zearalenone in beet pulp pellets and shreds support these findings. A 2010 survey of worldwide mycotoxin contamination of animal commodities (Rodrigues and Naehrer 2011) reported global incidence of positive samples as 42% for zearalenone (3,349 total samples, 11,195 analyses). In North America the incidence was 52%, second only to South America with an incidence of 57%. Ranking of incidence by commodity types was as follows: corn gluten meal (95%), dried distillers grains with solubles (87%), rice/bran (55%), and corn (46%). Sugar beet by-products were not included in the report, but were included with a study of the mycotoxin profiles of Fusarium spp. (Burlakoti et al., 2008). A known pathogen—Gibberella zeae—the anamorph of Fusarium was isolated from barley, wheat, potatoes, and sugar beets with concomitant zearalenone production. Additional data obtained from the author (T.B. Adhikari, ND State University, Fargo, ND, personal communication) documented all 6 isolates from sugar beets produced zearalenone in rice cultures (1 to 16 ppm).

Reproductive effects of zearalenone.

While reports of negative reproductive effects of feeding sugar beet by-products are scarce, zearalenone’s negative impact on reproduction are better documented (review, Diekman and Green, 1992), though often at concentrations of zearalenone 5- to 10-fold higher than detected in the beet samples submitted by the veterinarian. The review cites gilts and sows fed 5 mg/kg zearalenone had extended interestrous cycles and 9 mg/kg resulted in decreased sperm motility in boars. Most recently Jiang et al. (2012) found serum concentrations of testosterone, estradiol, LH, and FSH decreased and prolactin increased in gilts receiving 1 to 3 mg/kg dietary zearalenone for 18 d. In sheep, 3 mg zearalenone/d per ewe or ~2 mg/kg dietary zearalenone (DM basis), depressed ovulatory rates, and lambing percentages (Smith et al. 1990). Dietary exposure of ~8 to 16 mg/kg zearalenone increased the percentage of ewes that were anovulatory. There appears to be a window of sensitivity to the effects of zearalenone, as these concentrations were without effect on pregnancy number and lambing rate when fed from 5 to 10 d postmating. This window of zearalenone sensitivity may be critical when performing embryo transfer. The sensitivity of dairy cattle to zearalenone is indicated by Mirocha et al. (1968), who reported infertility associated with a fungal estrogen F-2 (later to be identified as zearalenone) in hay.

Zearalenone degradation, absorption, metabolism, and estrogenic activity.

Although the measured concentrations of zearalenone were what would generally be considered relatively low and possibly, clinically “insignificant,” zearalenone is the only known xenoestrogen detected in the sugar beets and sugar beet by-products analyzed in this study. There are currently no United States tolerance limits for zearalenone in feed stuffs, though the Food and Agriculture Organization of the United Nations reported that 16 countries have set limits ranging from 50 to 1,000 μg/kg wet weight of maize and other cereals (Worldwide Regulations for Mycotoxins, 2004. The EU has published a “guidance value of 0.5 mg zearalenone/kg wet weight of complimentary and complete feeding stuffs (12% moisture content) for dairy cattle” (KYPRIANOU, Markos 2006). The low concentrations of zearalenone cannot be ruled out as potential contributor to the measured estrogenic activity, and it is therefore relevant to briefly review what is known about zearlaneone degradation in the rumen, absorption, and hepatic metabolism, as well as the estrogenicity of zearalenone and its metabolites, especially in cattle. In monogastrics, the oral bioavailability of zearalenone was 28% in rats (Mallis et al. 2003) and >45% in pigs (Biehl et al. 1993). In vitro degradation studies in rumen fluid found only 49 to 63% of zearalenone had been degraded after a 24-h incubation (Macri et al. 2005). In only 5 h, the metabolites α and ß-zearalenol were present. There have been at least two reports of zearalenone and the zearalenol isomers in the bile or livers of cattle fed the naturally occurring mycotoxin (Kennedy et al. 1998; Kleinova et al. 2002). In cannulated cows zearalenone ingestion resulted in formation of both α and ß zearalenol, with all in approximately equal proportions: α-zearalenol 30%/ß-zearalenol 40%/zearalenone 30% in the duodenal contents (Danicke et al. 2005). Miles et al. (1996) reported that sheep metabolize zearalenone to α- and ß-zearalenol and zearalanol, by identification of urinary metabolites by GC-MS analysis.

While degradation/metabolism is often considered to reduce or destroy biological activity of natural or manmade toxins, this cannot be assumed for zearalenone. The relative estrogenic potency of zearalenol isomers is species and assay dependent. For example, Hagler et al. (1979) found while zearalenone’s uterotropic effect in rats was ~1/3 that found with α-zearalenol, Fusarium cultured on rice produced 14 times greater quantities of zearalenone than α-zearalenol. ß-zearalenol was ~1/3 as active as α isomer in the uterotropic assay (Hagler et al. 1979). Fitzpatrick et al. (1989) reported in oviduct binding assays (from immature female chickens, rats, and pigs) that zearalenone had 10 times the affinity of the α-zearalenol, and 100 times the affinity of ß-zearalenol, when competing with estradiol under saturating conditions. While relative activity of isomers may be variable, absorption of zearalenone is not and from these data, it is clear that in cows both absorption of zearalenone and its conversion to estrogenic metabolites will occur.

Clinical implications—estimating estrogenic concentration in blood from beet consumption.

To estimate the estrogenic effect of feeding beet by-products to cattle, the feeding guidelines of Lardy and Schafer (2008) were used. Recommended feeding of beet pulp at 20% of dietary dry matter to beef cattle would correspond to the consumption of 13.6 kg of wet pulp or ~3.42 kg of DM (wet pulp containing 25.1% DM) in a 500-kg cow. Figure 5 shows a calculation of estimated circulating E2Eq based on the E2Eq provided from sugar beet pellet sample submitted to the Veterinary diagnostic lab and the normal estradiol profiles of a cycling cow (Duke’s Physiology 2004). Using a conservative assumption of only 10% bioavailability for beet EE2 equivalents, both the beet tailing (33 pg/mL) and pellet samples (10 pg/mL) would have provided sufficient E2Eq to compete with or alter circulating estradiol in cycling cows. The dynamic range of serum estradiol is 1 to 10 pg/mL in the estrous cycle of the cow with 10 pg/mL representing maximal concentrations present during estrus (Duke’s Physiology 2004). The calculation shown in Fig. 5 assumes that beet products fed at levels consistent with the recommendation of Lardy and Shafer (2008), however the actual percent of diet that was being fed to the cows at the MN farm with decreased embryo transfer success rates is unknown. The inclusion of these estimations is an attempt to assess whether these concentrations of estrogenic feedstuffs could result in changes in circulating concentrations of “estrogens,” and subsequently impact success rates of embryo transfer recipients. The authors quickly acknowledge the caveats in these calculations, such as assuming bolus delivery of “dose” versus a more extended release upon absorption. The model provides only a starting point to consider the potential impact. It should also be noted that intensive reproductive management procedures, such as embryo transfer, which generally involve superovulation and synchronization protocols, would be expected to be more susceptible to the adverse effects of estrogenic feedstuffs than routine breeding of cattle by bulls. Human guidelines for the provisional maximum tolerable daily intake for zearalenone are 0.5 μg/kg body weight, based on a No Observable Effect Level (NOEL) of 40 μg/kg body weight per day from a 15-d pig study (Joint Committee FAO/WHO, 2004). While the NOEL of ruminants could be expected to be higher, consumption of the original suspect feed at the rate given in the model would result in 1.6 to 4 μg of zearalenone/kg body weight.
https://static-content.springer.com/image/art%3A10.1007%2Fs11626-012-9489-9/MediaObjects/11626_2012_9489_Fig5_HTML.gif
Figure 5.

Model for calculating circulating E2Eq from sugar beet by-product feeding (original pellet sample from veterinarian) and serum estradiol and progesterone profiles of cycling cow adapted from Duke’s Physiology of Domestic Animals (2004).

Conclusions

The sporadic occurrence of conditions which result in fungal contamination of sugar beets and their by-products, and the narrow window in which xenoestrogens may exert adverse effects on animal reproduction may explain the limited testing of sugar beets for zearalenone, phytoestrogens, and other sources of estrogenic activity. Livestock producers feeding sugar beets and their by-products should be aware of a potential for zearalenone contamination and estrogenic activity, which would result in reduced reproductive efficiencies in sensitive species. While such feeding may typically present no risk to animal health, producers should be mindful of climatic and storage conditions which could increase the presence of these toxins. This warning has particular urgency in years when sugar beet cultivating regions experience abnormally high rainfalls.

Acknowledgments

The authors would like to acknowledge the excellent technical assistance of Lloyd Billey and Jason Holthusen (ARS, Fargo, ND); statistical analysis by LuAnn Johnson, financial support of McNair Scholar; provision of MCF-7 BOS cells by Drs. Ana Soto and Carlos Sonnenschein; and contributions of Delvin Salathe, DVM, who provided the original suspect samples. Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture.

Supplementary material

11626_2012_9489_MOESM1_ESM.doc (32 kb)
ESM 1(DOC 31 kb)

Copyright information

© The Society for In Vitro Biology 2012